Visualization of autophagy in Arabidopsis using the fluorescent dye monodansylcadaverine and a GFP-AtATG8e fusion protein

Authors

  • Anthony L. Contento,

    1. Department of Genetics, Development and Cell Biology,
    2. Plant Sciences Institute, and
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yan Xiong,

    1. Department of Genetics, Development and Cell Biology,
    2. Interdepartmental Plant Physiology Program, 253 Bessey Hall, Iowa State University, Ames, IA 50011, USA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Diane C. Bassham

    Corresponding author
    1. Department of Genetics, Development and Cell Biology,
    2. Plant Sciences Institute, and
    3. Interdepartmental Plant Physiology Program, 253 Bessey Hall, Iowa State University, Ames, IA 50011, USA
      (fax 515 294 1337; e-mail bassham@iastate.edu).
    Search for more papers by this author

(fax 515 294 1337; e-mail bassham@iastate.edu).

Summary

Autophagy is a process that is thought to occur in all eukaryotes in which cells recycle cytoplasmic contents when subjected to environmental stress conditions or during certain stages of development. Upon induction of autophagy, double membrane-bound structures called autophagosomes engulf portions of the cytoplasm and transfer them to the vacuole or lysosome for degradation. In this study, we have characterized two potential markers for autophagy in plants, the fluorescent dye monodansylcadaverine (MDC) and a green fluorescent protein (GFP)-AtATG8e fusion protein, and propose that they both label autophagosomes in Arabidopsis. Both markers label the same small, apparently membrane-bound structures found in cells under conditions that are known to induce autophagy such as starvation and senescence. They are usually seen in the cytoplasm, but occasionally can be observed within the vacuole, consistent with a function in the transfer of cytoplasmic material into the vacuole for degradation. MDC-staining and the GFP-AtATG8e fusion protein can now be used as very effective tools to complement biochemical and genetic approaches to the study of autophagy in plant systems.

Introduction

Autophagy is a process that is thought to occur in all eukaryotes in which cells recycle cytoplasmic contents during periods of stress, providing materials for re-use or energy for survival, or during development to promote cellular remodeling (see Huang and Klionsky, 2002 for a review). During autophagy, double membrane-bound structures, called autophagosomes, engulf portions of the cytoplasm and transfer them to the vacuole or lysosome. The outer membrane of the autophagosome fuses with the vacuolar/lysosomal membrane, releasing the inner membrane and its contents into the vacuole/lysosome. Autophagy is involved in the turnover of long-lived proteins and organelles under normal conditions and during periods of nutrient stress (Aubert et al., 1996; Doelling et al., 2002; Seglen and Bohley, 1992).

Nutrient stress-induced autophagy in plants has been studied in maize root tips (Brouquisse et al., 1991; James et al., 1993) and whole maize plants (Brouquisse et al., 1998), and rice (Chen et al., 1994), sycamore (Aubert et al., 1996), tobacco (Moriyasu and Ohsumi, 1996) and Arabidopsis (Contento et al., 2004) suspension cell cultures during sucrose starvation. In starving cells, an increase in the size of the vacuole is evident, with a decrease in cytoplasmic volume and an increase in the activity of vacuolar enzymes (Moriyasu and Ohsumi, 1996). These changes allow survival of the cell or plant for an extended period of time, until a nutrient source is restored.

Our understanding of the autophagy process has been enhanced by the isolation of yeast mutants that are defective in autophagy, leading to the identification of 27 autophagy-related ATG genes (Harding et al., 1996; Klionsky et al., 2003; Thumm et al., 1994; Tsukada and Ohsumi, 1993). Autophagy in yeast is negatively regulated by a Target-of-Rapamycin (TOR) kinase (Noda and Ohsumi, 1998) and is initiated by the ATG1 protein kinase switch and a phosphatidylinositol-3-kinase (Kihara et al., 2001; Matsuura et al., 1997; Straub et al., 1997). Autophagy utilizes a unique conjugation system, generating an Atg5–Atg12 protein conjugate that is required for autophagosome formation and an Atg8-phosphatidylethanolamine conjugate, which controls autophagosome expansion (Ichimura et al., 2000; Klionsky and Ohsumi, 1999; Mizushima et al., 1998a,b). In Arabidopsis, genes encoding potential homologs for many of these yeast autophagy proteins have been found. Arabidopsis knockout mutants in homologs of the yeast autophagy genes ATG9, ATG7, VTI1, and ATG18 showed increased sensitivity to nitrogen deficiency and an early senescence phenotype (Doelling et al., 2002; Hanaoka et al., 2002; Surpin et al., 2003; Xiong et al., 2005), suggesting that autophagy is important in Arabidopsis under these conditions.

Another key to increasing our understanding of the autophagy process has been the development of markers or assays for autophagy. One such marker is the protein Atg8, which is essential for autophagy in yeast. It is involved in lipid recruitment during autophagosome formation upon amino acid starvation (Kirisako et al., 1999) and is the only identified ATG protein that, in addition to being associated with the preautophagosome, remains associated with mature autophagosomes until they reach the vacuole (Ichimura et al., 2000; Kirisako et al., 2000). Immunofluorescence microscopic studies of Atg8 in yeast revealed that it localizes to punctate structures identified as autophagosomes (Kirisako et al., 1999). Electron microscopic studies in rat have shown that that a mammalian Atg8 homolog, microtubule-associated light chain 3 protein (LC3), is also localized to autophagosomes (Kabeya et al., 2000; Legesse-Miller et al., 2000; Mann and Hammarback, 1994). Later studies used green fluorescent protein (GFP)-LC3 as a marker in transgenic mice to examine the variations in autophagy from organ to organ (Mizushima et al., 2004).

Monodansylcadaverine (MDC) is an autofluorescent amine that specifically stains autophagosomes in mammals (Biederbick et al., 1995; Munafo and Colombo, 2001). In mammalian cells, the specificity of MDC staining is derived both from ion trapping, as autophagosomes are known to be acidic compartments, and interaction with lipid molecules found in high concentration in autophagosomes (Niemann et al., 2000). Biederbick et al. (1995) stained intact human cells with MDC, followed by sucrose gradient fractionation to isolate the stained component. Electron microscopy demonstrated that the MDC-stained organelles were double-membrane autophagosomes. Small, MDC-stained structures were also visible in CHO cells in which autophagy was induced by amino acid starvation (Munafo and Colombo, 2001). Labeling of these structures was prevented in the presence of autophagy inhibitors, indicating that MDC also labels autophagosomes in this cell type. A GFP-LC3 fusion protein was recruited to the MDC-labeled vesicles during amino acid starvation, which confirmed the autophagic nature of these structures (Munafo and Colombo, 2001).

Autophagy has been studied morphologically in plant cells (Aubert et al., 1996; Moriyasu and Ohsumi, 1996), and Arabidopsis knockout mutants in homologs of yeast autophagy genes have been characterized (Doelling et al., 2002; Hanaoka et al., 2002). However, the analysis of autophagy in plants has been hindered by the lack of a convenient marker or assay for the autophagy process. Two markers for autophagosomes have been described in mammalian cells, the fluorescent dye MDC and the Atg8/LC3 protein. Here, we demonstrate that these markers can be used as an assay for autophagy in Arabidopsis, and analyze the conditions under which autophagosomes form in vivo in plant cells.

Results

MDC staining of Arabidopsis suspension cells

To analyze the potential for using MDC as a marker for autophagy in plants, Arabidopsis suspension cells were sucrose-starved for 0, 24, 48 or 72 h to induce autophagy (Contento et al., 2004). Starved cells, or non-starved controls, were stained by incubation with 50 μm MDC for 10 min at room temperature, followed by washing with phosphate-buffered saline (PBS) to remove excess dye, and observed by fluorescence microscopy. In the control cells (+sucrose) very little staining was observed, with only weak staining of the cell wall and/or plasma membrane at all time points. In contrast, after 24 h of starvation cells displayed strong staining of moving, spherical structures within the cell (Figure 1a). The size of the structures varied, and often large and small stained structures could be seen within the same cell; however, their average size increased by 48 h of starvation. Most of the structures were seen in the cytoplasm, but occasionally they appeared to be present within the vacuole, suggesting that they may eventually be delivered to the vacuole (also see Figure 3f). The number of cells with MDC staining decreased after 72 h, possibly due to an increase in cell death at later starvation time points (Contento et al., 2004), although the cells that did stain frequently contained a large number of stained structures compared with the 24 and 48 h time points (Figure 1c). Small MDC-stained structures were only occasionally seen in the control, non-starved cells, and these structures were less intensely stained, smaller and in much fewer number than in the starved samples. Larger structures were never seen in the control cells. The presence of MDC-stained structures predominantly in starved cells suggests that MDC stains autophagosomes in Arabidopsis, as it does in mammalian cells.

Figure 1.

Monodansylcadaverine (MDC) staining of Arabidopsis suspension cells.
(a) Arabidopsis suspension cells sucrose-starved for 0, 24, 48, and 72 h, and non-starved controls, were stained with MDC and analyzed by fluorescence microscopy. Scale bar = 25 μm.
(b) The percentage of cells with MDC-stained structures was determined for suspension cells that had been sucrose-starved for 0, 24, 48, and 72 h. The experiment was repeated three times, with 100 cells counted for each replicate. Error bars represent standard error.
(c) Suspension cells were sucrose-starved for up to 72 h, incubated with 50 mm ammonium chloride for 20 min, and stained with MDC. The number of MDC-stained structures visible per cell cross section was determined, averaging 30–50 cells per sample. White bars indicate ammonium chloride-treated samples, gray bars indicate control (no ammonium chloride) samples.
(d) Suspension cells that had been sucrose-starved for 0, 24, 48, and 72 h were stained with MDC. Fluorescence of suspension cells was measured using a Hitachi F-2000 fluorescence spectrophotometer, with an excitation wavelength of 335 nm and an emission wavelength of 508 nm. The intensity of MDC staining was expressed in relative fluorescence per gram of fresh weight, compared with the 0-h control. Error bars represent standard error.

Figure 3.

Colocalization of monodansylcadaverine (MDC) staining and GFP-AtATG8e in Arabidopsis protoplasts.
(a) Suspension cell protoplasts expressing GFP-AtATG8e were grown in sucrose-containing medium for 48 h and visualized by fluorescence microscopy.
(b) Suspension cell protoplasts expressing GFP-AtATG8e were incubated in sucrose starvation medium for 48 h.
(c) Non-transformed suspension cell protoplasts were incubated in sucrose starvation medium for 48 h and stained with MDC.
(d) Suspension cell protoplasts expressing GFP-AtATG8e were incubated in sucrose starvation medium for 48 h followed by staining with MDC. The merged image shows co-localization of the fluorescent signals.
(e) Leaf protoplasts expressing GFP-AtATG8e were incubated in sucrose starvation medium for 48 h and stained with MDC. The merged image shows co-localization of the fluorescent signals.
(f) Suspension cell protoplasts expressing GFP-AtATG8e were starved for 48 h followed by staining with FM4-64 for 6 h to label the tonoplast. The merged image shows the presence of an autophagosome inside the vacuole. Arrows indicate autophagosomes (GFP and merged panels) or tonoplast (FM4-64 panel).
(g) Suspension cell protoplasts expressing GFP-AtATG8e were starved for 48 h followed by incubation with FM4-64 for 6 h on ice. Arrows indicate autophagosomes (GFP panel) or FM4-64-labeled plasma membrane (FM4-64 panel). Scale bars = 25 μm for all panels.

Autophagosomes are thought to be acidic compartments, and MDC staining specificity in mammals is derived partly from this property (Niemann et al., 2000). To further characterize the Arabidopsis MDC-stained structures, sucrose-starved suspension cells were pre-incubated with 50 mm ammonium chloride for 20 min to neutralize the lumenal pH, prior to MDC staining. This treatment significantly inhibited MDC staining (Figure 1c), with the average number of autophagosomes visible per cell cross section reduced approximately 10-fold at 24 and 48 h starvation time points. Ammonium chloride treatment does not prevent autophagosome formation, as a GFP-fusion with the Atg8 homolog AtATG8e still labels autophagosomes after treatment (see below; data not shown). This indicates that an acidic lumen is required for labeling with MDC, providing further evidence that MDC specifically labels autophagosomes in Arabidopsis. Interestingly, at 72-h starvation, ammonium chloride was less effective at inhibiting MDC staining. At this time point, autophagosomes were still labeled with MDC in the presence of ammonium chloride, although staining was much weaker than in its absence. Extensive cell death occurs after 72-h sucrose starvation of suspension cells (Contento et al., 2004), potentially by autophagic or type II programmed cell death. Our results indicate that there may be a change in the structure or lipid composition of autophagosomes at this late stage, concurrent with a switch from cell survival to cell death.

Fluorescence spectroscopic analysis of MDC-stained Arabidopsis suspension cells

Fluorescence of MDC-stained suspension cell samples was measured by UV-fluorescence spectroscopy to determine if the difference in signal between starving and non-starving cell samples could be used as a quantitative measure of autophagy. Cells were stained and washed as described above, and the fluorescence, relative to control cells, was determined per gram of fresh weight (Figure 1d). The fluorescence intensity increased more than twofold after 24 h of starvation. The fluorescence decreased slightly after 48 h of starvation, and decreased again after 72 h, but remained significantly above the 0-h levels. Fluorescence levels in the non-starvation control samples did not differ between the time points. The fluorescence intensity of the starving and control samples correlates with the number of MDC-staining cells and intensity of MDC staining as determined by microscopy (Figure 1a,b). These data suggest that fluorescence spectroscopy measurement of MDC-stained cells can be used as a rapid, semiquantitative assay to determine the extent of autophagy in plant suspension cells.

Movement of MDC-stained structures

The MDC-stained structures in starved suspension cells varied in size and movement. Smaller structures showed a rapid and apparently directional movement around the cell, while larger structures moved much more slowly, typically vibrating in place (see Supplementary Material). The speed of the smaller MDC-stained structures in 48-h sucrose-starved suspension cells was measured by determining distance traveled over a period of time. These small spheres moved across the cell with an average speed of approximately 2 μm sec−1, comparable in speed with the movement of organelles such as peroxisomes or Golgi (Boevink et al., 1998; Mano et al., 2002; Mathur et al., 2002).

Characterization of an Arabidopsis Atg8 ortholog

Previous yeast and mammalian studies have shown that Atg8 and its orthologs specifically localize to autophagosomes (Kabeya et al., 2000; Kirisako et al., 1999), and therefore make excellent markers for autophagosomes. The Arabidopsis genome encodes nine potential Atg8 orthologs, at least seven of which are expressed (Doelling et al., 2002; Hanaoka et al., 2002). One of the Arabidopsis ATG8 genes, AtATG8e (At2g45170; APG8e in Hanaoka et al., 2002), shows increased expression by Affymetrix GeneChip analysis during sucrose starvation (Contento et al., 2004), making it a good candidate for an autophagosomal marker in Arabidopsis. The expression increase was verified in suspension cells after 24 and 48 h of sucrose starvation by RT-PCR (Figure 2a) and by Northern blot hybridization using a gene-specific probe (Figure 2b). AtATG8e showed an increase in transcript level after 24 and 48 h of sucrose starvation, compared with low RNA levels in the 0 and 48 h control samples. AtATG8e transcript levels were also determined for whole Arabidopsis plants. One-week-old seedlings were subjected to either nitrogen or carbon starvation. Seedlings were harvested after 2 or 4 days of starvation and total RNA was extracted for Northern blot analysis (Figure 2c). In control seedlings, AtATG8e transcript levels remained low. Transcript levels increased slightly after 2 or 4 days of nitrogen starvation, whereas a significant increase in expression was seen upon sucrose starvation.

Figure 2.

Characterization of AtATG8e during starvation.
(a) Total RNA was extracted from suspension cells that had been sucrose-starved for 0, 24, and 48 h and from non-starved cells 48 h after transfer to sucrose-containing medium. RT-PCR was performed using primers specific for AtATG8e. Products were visualized on an agarose gel stained with ethidium bromide.
(b) Northern blot hybridizations were performed using RNA samples as above and an AtATG8e-specific probe.
(c) Arabidopsis seedlings were grown on nutrient-rich control medium for 0 and 4 days, grown on medium lacking nitrogen for 2 and 4 days, or grown on medium lacking sucrose in the dark for 2 and 4 days. Total RNA was extracted and analyzed by Northern blotting using an AtATG8e-specific probe.
(d) Protoplasts from Arabidopsis suspension cells expressing either a GFP-AtATG8e fusion protein or GFP alone were lysed and separated into soluble and insoluble fractions. The distribution of GFP and GFP-AtATG8e between the fractions was determined by immunoblot analysis using anti-GFP antibodies.
(e) Protoplasts from Arabidopsis suspension cells expressing a GFP-AtATG8e were incubated for 2 days in the presence (+suc) or absence (−suc) of sucrose, lysed, loaded onto a 25% sucrose step on top of a 60% sucrose cushion, and centrifuged at 95 000 g for 2 h. Fractions were analyzed by immunoblotting using GFP antibodies. Lane 1, load; lane 2, 25% sucrose step; lane 3, 60% sucrose cushion interface.

A GFP-AtATG8e fusion construct was used to transiently transform protoplasts made from Arabidopsis suspension cells. Protein was extracted from transformed cells after incubation in the presence or absence of sucrose and the distribution of GFP-AtATG8e fusion protein between insoluble and soluble fractions was determined, as yeast Atg8 is associated with the autophagosome membrane. Insoluble protein was separated from soluble protein by centrifugation, and GFP-AtATG8e was detected by immunoblotting using anti-GFP antibodies (Figure 2d). Regardless of whether the protoplasts were starved or not, the majority of the GFP-AtATG8e fusion protein was found in the insoluble fraction, while only a small amount was in the soluble fraction, indicating that GFP-AtATG8e may be membrane-associated. In contrast, GFP alone was found mostly in the soluble fraction under the same conditions.

To further confirm the membrane association of GFP-AtATG8e, protoplasts expressing GFP-AtATG8e were incubated for 2 days in the presence or absence of sucrose. Lysed protoplast extracts were loaded onto a 25% sucrose step on top of a 60% sucrose cushion and membranes separated by centrifugation onto the cushion. Immunoblotting using GFP antibodies (Figure 2e) revealed that in the starved samples, the majority of the GFP-AtATG8e was present at the top of the 60% sucrose cushion, indicating that it is likely to be membrane-associated. In contrast, in the presence of sucrose, some GFP-AtATG8e was seen at the cushion, but the majority of the protein was found above the cushion, in the load and upper 25% sucrose fraction. Little protein was found below the 60% sucrose interface in either case (data not shown). As a control, an integral membrane protein, SYP41, was found almost entirely on top of the 60% sucrose cushion (data not shown; Bassham et al., 2000). These results demonstrate that at least a portion of the GFP-AtATG8e is associated with membranes, and that a larger fraction is membrane-bound in starved cells than in non-starved cells.

Protoplasts expressing GFP-AtATG8e were visualized by fluorescence microscopy after 48 h of sucrose starvation (Figure 3b). GFP fluorescence was associated with moving, spherical structures of varying size and speed, similar to the structures found during MDC-staining of protoplasts and suspension cells. At very high levels of GFP-AtATG8e expression, these structures could not be seen, and instead the GFP fluorescence was found throughout the cytoplasm; the amount of plasmid was therefore adjusted to allow moderate expression levels. In non-starved cells, non-motile aggregates of GFP-AtATG8e fluorescence were dispersed throughout the cytoplasm and distinct structures were not visible (Figure 3a). These results indicate that in Arabidopsis, as in other organisms, the Atg8 homolog AtATG8e can be used as a marker for autophagy.

Colocalization of MDC and GFP-AtATG8e fusion protein in Arabidopsis protoplasts

To confirm that MDC specifically stains autophagosomes in Arabidopsis, as it does in mammalian cells, protoplasts expressing GFP-AtATG8e were sucrose-starved for 48 h, followed by staining with MDC. In suspension cell and leaf protoplasts, GFP-AtATG8e and MDC staining colocalized to small, moving structures that were absent in non-starved controls (Figure 3d,e). Single labeling controls confirmed that there is no bleed-through of signal between filters (Figure 3a–c). The rapid movement of the autophagosomes was problematic for imaging; in many cases the autophagosomes moved slightly while changing filters. However, observation of many different protoplasts with double labeling confirmed that GFP-AtATG8e and MDC do in fact label the same structures. We conclude that the MDC-stained vesicles are likely to be autophagosomes, and that both MDC and GFP-AtATG8e can be used as markers for autophagy in Arabidopsis.

To clarify the location of the autophagosomes within the cell, protoplasts expressing GFP-AtATG8e were starved of sucrose for 48 h, followed by staining with the fluorescent dye FM4-64. FM4-64 binds to the plasma membrane and is taken up into cells by endocytosis, eventually reaching the vacuolar membrane after several hours of incubation (Ueda et al., 2001). It can therefore be used as a fluorescent marker for the tonoplast. Starved protoplasts internalized the FM4-64 more slowly than non-starved protoplasts, presumably because of a decreased rate of endocytosis (data not shown). However, after 6 h of incubation, staining of the tonoplast could be seen by confocal microscopy (Figure 3f; arrows in FM4-64 panel). To confirm that the FM4-64 was taken up by endocytosis, rather than non-specific staining of dying cells, starved cells expressing GFP-AtATG8e were also stained with FM4-64 for 6 h on ice to prevent uptake (Figure 3g). The FM4-64 stained the plasma membrane of these cells (arrows in FM4-64 panel) but no internal staining was observed, indicating that internalization in the starved protoplasts is an active membrane trafficking process and the cells are still viable at this time.

While most of the GFP-AtATG8e-labeled structures were present in the cytoplasm, occasionally they were seen inside the vacuole (arrow in Figure 3f, merged panel); this was confirmed by taking a Z-series through the protoplast (data not shown). This suggests that the autophagosomes present in the cytoplasm are eventually delivered to the vacuole for degradation. Interestingly, the autophagosomes within the vacuole were labeled with FM4-64 in addition to GFP-AtATG8e (Figure 3f), whereas those in the cytoplasm were not. It appears that prior to or upon uptake into the vacuole, membrane derived from endosomes or the tonoplast is incorporated into the autophagosomes. In mammalian cells, autophagosomes can fuse with endosomes prior to their final fusion with lysosomes (Lucocq and Walker, 1997), and it is possible that this also occurs in plant protoplasts.

MDC staining of Arabidopsis roots and senescent leaf protoplasts

Ideally, for an autophagosome marker to be most useful, it should work as a reporter for autophagy in intact plants, as well as suspension cells and protoplasts. To determine whether MDC can be used to stain autophagosomes in whole plants, 7-day-old Arabidopsis seedlings were subjected to nutrient starvation. Seedlings were transferred to medium lacking sucrose and grown in the dark to induce sucrose starvation, followed by staining with MDC (Figure 4). Roots were observed using fluorescence microscopy after 0, 2, or 5 days of starvation. The 0-day roots displayed only weak staining of the cell wall and plasma membrane. After 2 days of sucrose starvation, MDC fluorescence signal increased, and MDC-stained motile structures were visible in the cytoplasm of cells along the length of the root, except for the cells of the root tip. After 5 days of sucrose starvation, MDC staining was more intense. The size and fluorescence intensity of the moving structures increased, and the structures were more abundant throughout the root, including the root tip. Control seedlings grown in the light or on sucrose-containing plates did not contain MDC-stained vesicles. Alternatively, seedlings were transferred to medium lacking nitrogen and stained with MDC after 0, 2, or 5 days. Similar results were seen as for sucrose-starved seedlings (data not shown), indicating that autophagy can be induced in Arabidopsis roots by either sucrose or nitrogen starvation.

Figure 4.

Monodansylcadaverine (MDC) staining in Arabidopsis roots.
Seven-day-old Arabidopsis seedlings were subjected to sucrose starvation in the dark for up to 5 days. Starved seedlings or non-starved controls were stained with MDC and roots were analyzed by fluorescence microscopy. Scale bar = 50 μm.

The autophagy genes AtATG9 and AtATG7 are both required for the proper timing of senescence in Arabidopsis, as knockout mutants show a premature senescence phenotype (Doelling et al., 2002; Hanaoka et al., 2002). However, the cellular basis for this phenotype is unclear, and the extent to which autophagy is involved in the senescence process is not known. To provide insight into the role of autophagy in leaf senescence, detached Arabidopsis leaves were incubated in the dark for 2–3 days to initiate senescence and protoplasts were isolated from this senescing tissue. The induction of senescence was confirmed by the increase in expression of two senescence-associated genes, AtSEN1 (At4g35770) and AtYSL4 (At5g4100), after 2 days of dark incubation (Xiong et al., 2005; data not shown). The protoplasts were stained with MDC and viewed using fluorescence microscopy (Figure 5). A similar pattern of MDC staining was observed in the senescent protoplasts as in protoplasts after starvation, with motile-stained autophagosomes visible that were absent in protoplasts from freshly detached leaves. These data suggest that autophagy is induced in detached, senescing leaves at an early time point, before visible signs of senescence such as loss of chlorophyll are evident.

Figure 5.

Monodansylcadaverine (MDC) staining of senescent Arabidopsis protoplasts.
Detached Arabidopsis leaves were incubated in the dark for 0, 1, and 2 days to initiate senescence and protoplasts were isolated from this senescing tissue. Protoplasts were stained with MDC and analyzed by fluorescence microscopy. Scale bar = 25 μm.

Discussion

Autophagy is a process in eukaryotes by which long-lived proteins and organelles are turned over throughout the life cycle of an organism (Seglen and Bohley, 1992). It may be induced during development, periods of environmental stress, or senescence and cell death (Aubert et al., 1996; Doelling et al., 2002). In this study, we have characterized two potential markers for autophagy in plants, the fluorescent dye MDC and a GFP-AtATG8e fusion protein, and shown that they are both likely to label autophagosomes in Arabidopsis.

Several lines of evidence indicate that MDC is an autophagosome-specific marker in Arabidopsis, as has previously been shown in mammalian cells. First, it labels small, apparently membrane-bound structures of an appropriate size to be autophagosomes. Secondly, these structures are only found in cells under conditions that are known to induce autophagy (i.e. starvation and senescence). Under nutrient-rich conditions, only a very weak, diffuse cytoplasmic MDC labeling is seen. Thirdly, while most of the MDC-labeled structures are present in the cytosol, occasionally they can be seen inside the vacuole, suggesting that they function in the transfer of materials into the vacuole. Fourthly, the MDC-stained structures have an acidic lumen, characteristic of autophagosomes in other species. Fifthly, MDC staining co-localizes in starved cells with GFP-AtATG8e. Atg8 homologs localize to autophagosomes in all species that have been studied, and this appears to be true also in Arabidopsis. Finally, MDC-labeled structures are absent from transgenic plants with reduced levels of AtATG18a, a homolog of yeast Atg18 required for autophagosome formation, even upon starvation or during senescence (Xiong et al., 2005). Preventing autophagosome formation therefore prevents MDC staining, suggesting that MDC does in fact specifically stain autophagosomes in Arabidopsis plants.

The key advantages of using MDC to detect autophagic vesicles in plants are specificity and simplicity. Past research has determined that MDC specifically labels autophagosomes in mammalian systems (Biederbick et al., 1995; Munafo and Colombo, 2001) and in this work we demonstrate that the same is true in Arabidopsis. Staining with MDC is simple, requiring only a brief exposure to the agent and washing to remove excess stain. Previous assays to measure autophagy in plants have involved electron microscopy to identify autophagosomes (Aubert et al., 1996; Moriyasu and Ohsumi, 1996), which is extremely time-consuming and technically difficult, or staining with a lysosomal dye (Takatsuka et al., 2004), which requires the presence of protease inhibitors, and the specificity of this assay is not clear. Studies of autophagy mutants have relied on indirect measures of autophagy such as sensitivity to starvation (Doelling et al., 2002; Hanaoka et al., 2002; Surpin et al., 2003). Using the assay we have described, MDC can be used to quickly stain Arabidopsis cells and whole plants, giving a direct measure of autophagy without the need for inhibitors or the construction of transgenic plants.

The nature of the MDC-stained structures has revealed some important features of the process of autophagy in plants. The size of autophagosomes increased as the duration of starvation increased (Figure 4). The increasing size of autophagosomes may be regulated by the increase in the expression of AtATG8e protein, as Atg8 is responsible for controlling the size of autophagosomes in yeast (Kirisako et al., 1999). Larger autophagosomes may also be able to engulf organelles such as mitochondria and peroxisomes for delivery to the vacuole for degradation (Klionsky and Ohsumi, 1999). The MDC-stained autophagosomes were motile and moved either very slowly, vibrating in place, or very quickly, moving rapidly across the cell at a constant velocity of 2 μm sec−1. These two classes of autophagosome movement were also observed for GFP-AtATG8e-labeled autophagosomes, and appear to be related to the size of the autophagosome; the significance of this finding is not yet known. In mammalian cells, it is known that MDC stains only mature autophagosomes, after acidification of the lumen, whereas GFP-LC3 labels autophagosomes at all stages of formation (Bampton et al., 2005). While the great majority of GFP-AtATG8e-labeled autophagosomes also stain with MDC, it is possible that these reagents label slightly different sub-populations of autophagosomes, and may be useful to distinguish between mutants with defects at different stages of autophagosome formation.

A key observation is that autophagosome formation is induced very rapidly in a detached leaf senescence assay (Figure 5). The role of autophagy in leaf senescence has been somewhat controversial (Matile, 1997), and autophagy mutants show a premature senescence phenotype (Doelling et al., 2002; Hanaoka et al., 2002; Surpin et al., 2003). We demonstrate that autophagy is induced at early time points after leaf detachment, several days before a loss of chlorophyll is evident. This suggests that nutrient remobilization by autophagy is important in maintaining leaf structure and function for as long as possible, potentially to maximize the use of nutrients in seed filling.

We have demonstrated that GFP-AtATG8e is also a useful marker for autophagy in Arabidopsis, and labels the same structures as MDC (Figure 3). GFP-AtATG8e may be more useful for co-localization studies with other proteins than is MDC, as it is more amenable to confocal microscopy. We have had some difficulty in visualizing MDC using confocal microscopy, perhaps due to its weaker signal. The disadvantage of GFP-AtATG8e as a marker is, of course, that transgenic plants or cells must be generated in order to use the marker. GFP fusions with three other Arabidopsis Atg8 orthologs (AtATG8i/At3g15580, AtATG8b/At4g04620 and AtATG8f/At4g16520) were also tested in transient transformation of protoplasts. GFP fluorescence was localized to similar punctate structures as was observed in the GFP-AtATG8e transformants during sucrose starvation (data not shown), suggesting that all of these orthologs localize to autophagosomes. This is not entirely unexpected, as in mammals, all three of the Atg8 orthologs that have been studied are found on autophagosomes (Hemelaar et al., 2003; Kabeya et al., 2004; Tanida et al., 2002). It is possible that expression of the Arabidopsis Atg8 orthologs will be induced by different stimuli, or in different tissues or developmental stages of the plant. The differential expression of ATG8 orthologs during sucrose starvation (Contento et al., 2004), and the difference in expression pattern of AtATG8e in Arabidopsis seedlings during nitrogen and carbon starvation, may support this idea. Yoshimoto et al. (2004) recently demonstrated that AtATG8 and AtATG4 proteins are required for autophagy, supporting our conclusions that ATG8 homologs can be used as an autophagosome marker in Arabidopsis.

We have described two complementary assays for studying autophagy in plants. The first is the use of the autophagosome-localized protein GFP-AtATG8e to visualize autophagosomes in transiently transformed protoplasts, and potentially in stably transformed Arabidopsis plants. The major disadvantage of this assay is that transgenic plants must be generated, and therefore the analysis of autophagy mutants, for example, remains time-consuming by this method. The second assay is the use of MDC, an inexpensive, autofluorescent amine, in a simple, specific staining procedure for studying autophagosomes in living plant cells and tissues. MDC can be used to semiquantitatively determine the extent of autophagy in vivo, as a measure of an increase in fluorescence, or to determine the presence or absence of autophagosomes in cells and tissue under nutrient stress and senescence. While further testing is necessary to determine if MDC can be used in other plant species, this assay can now be used to study the induction and inhibition of autophagy in Arabidopsis using chemical agents, environmental and developmental stimuli, and RNAi and knockout mutants. The simplicity of the assay may allow the development of a genetic screen for new Arabidopsis mutants that are defective in autophagy. It is possible that other autofluorescent lysosomotropic agents such as monodansylpentane (Niemann et al., 2001) specifically label autophagosomes, allowing for more variety in the available compounds that can be used to study autophagy. MDC staining can now be used as a very effective tool to complement biochemical and genetic approaches to the study of autophagy in plant systems.

Experimental procedures

Growth of Arabidopsis suspension cell cultures

An Arabidopsis thaliana Columbia-0 suspension cell culture was obtained from Dr S.B. Gelvin and maintained by subculturing weekly into 50 ml of MS medium [Murashige–Skoog Minimal Organics Medium (Gibco BRL, Gaithersburg, MD, USA), 2% (w/v) sucrose, 1 μg ml−1 naphthalene acetic acid (Sigma, St Louis, MO, USA), 50 ng ml−1 kinetin (Sigma)]. Cultures were grown in Erlenmeyer flasks at room temperature, under ambient light, with constant shaking (115 rpm).

Growth of plant materials

Arabidopsis thaliana Columbia-0 plants were grown on soil or solid medium [Murashige–Skoog Vitamin and Salt Mixture (Gibco BRL), 1% (w/v) sucrose (Sigma), 2.4 mm 2-morphinolino-ethanesulfonic acid (MES) (Sigma) and 0.8% phytagar (Gibco BRL)]. All plants were grown under 16 h light.

Sucrose starvation treatment of suspension cell cultures

All starvation time courses were begun using suspension cells 3 days after subculturing, at an approximate cell density 2 × 105 cells ml−1. Cultures were washed three times with either sucrose-containing medium for control samples or medium lacking sucrose for starvation samples. After the third wash, 50 ml of the appropriate medium was added and the cells were grown for up to 72 h on a rotational shaker using the conditions described above.

Nutrient starvation treatment of whole seedlings

Seven-day post-germination seedlings grown on solid medium containing sucrose were transferred to fresh medium for control samples, or to medium lacking sucrose or nitrogen. Control and nitrogen starvation samples were grown in the light under long day conditions, while sucrose starvation samples were grown in darkness. Seedling samples were harvested at 0, 2, 4, and 5 days after transfer.

MDC staining

Arabidopsis cells were stained with a 0.05 mm final concentration of MDC (Sigma) in PBS for 10 min (Biederbick et al., 1995). When staining protoplasts, PBS was supplemented with 0.4 m mannitol. Cells were washed with PBS two times to remove excess MDC. Fluorescence of suspension cells was measured using a Hitachi (Tokyo, Japan) F-2000 fluorescence spectrophotometer, with an excitation wavelength of 335 nm and an emission wavelength of 508 nm. The intensity of MDC staining was expressed in relative fluorescence per gram of fresh weight, as a percentage of the 0-h control. Seedlings were stained by immersion in 0.05 mm MDC in PBS for 10 min, followed by two PBS washes. All samples were kept on ice in the dark after staining. For ammonium chloride treatment, cells were pre-incubated with 50 mm ammonium chloride for 20 min at room temperature, washed twice with PBS, and stained with MDC as above.

Preparation of protoplasts

Protoplasts were prepared from suspension cells (5 days after subculture) or leaf tissue (4 weeks post-germination) by digestion with cellulase (0.3% for cells, 1% for leaf) and macerozyme (0.15% for cells, 0.2% for leaf; Yakult Pharmaceutical, Osaka, Japan; Sheen, 2002). Protoplasts were strained through 70 μm Nylon Mesh (Carolina Biological Supplies, Burlington, NC, USA). Starvation was induced by incubating protoplasts for 48 h in MS medium lacking sucrose, but supplemented with 0.4 m mannitol to maintain isoosmotic conditions. Control samples were grown in MS medium containing sucrose.

Senescence induction of detached leaves

The first and second true leaves of 14-day post-germination seedlings grown on MS plates were detached. The leaves were placed onto filter paper infused with 3 mm MES (pH 5.7) and incubated in the dark for up to 2 days (Weaver et al., 1998). Protoplasts were then isolated from the senescing leaves and stained with MDC.

Total RNA isolation, RT-PCR, and Northern blot analyses

Suspension cell and seedling samples were collected, the medium was removed from the cells, and the samples were stored at −80°C until RNA extractions were performed. Total RNA was isolated using a TRIzol extraction method (http://www.science.siu.edu/plant-biology/PLB420/DNA.Techniques/TRIzol.method.html). Northern blot analyses were performed using a probe consisting of a radiolabeled cDNA fragment corresponding to AtATG8e (At2g45170). Hybridization was performed using the manufacturer's protocol for UltraHyb solution at high stringency (Ambion, Austin, TX, USA). RT-PCR was performed using the following AtATG8e-specific primers (5′-AGATCTATGAATAAAGGAAGCATCTTT- 3′ and 5′-TCTAGATTAGATTGAAGAAGCAACGAA -3′).

Transient transformation with GFP-AtATG8e

GFP-fusion constructs were made using a modified pJ4GFP-XB vector (Igarashi et al., 2001) in which the multiple cloning site of the vector was altered to produce N-terminal GFP fusions, instead of C-terminal fusions. The original GFP was replaced with a modified GFP fragment containing a multiple cloning site immediately prior to the stop codon, obtained by PCR using the primers 5′-GGATCCATGGTGAGCAAGGGCGAGGAGCTGTTCA-3′ and 5′-GAGCTCTAGTCTAGAAGCTTAGATCTCTTGTACAGCTCGTCCATGCCGTG-3′. An AtATG8e (At2g45170) cDNA was synthesized by RT-PCR from total RNA from suspension cells that had been sucrose-starved for 48 h, using gene-specific primers 5′-AGATCTATGAATAAAGGAAGCATCTTT-3′ and 5′-TCTAGATTAGATTGAAGAAGCACCGAA-3′. The cDNA was sequenced for verification and ligated into the modified pJ4GFP-XB vector. Protoplasts were transformed using 20 μg of plasmid DNA, according to Sheen (2002). Cells were transferred to control or starvation medium, and were incubated at room temperature, in darkness for 48 h, with 40 rpm orbital shaking. Cells were then stained with MDC and visualized using UV-fluorescence microscopy.

Detection of GFP by immunoblotting

The pJ4GFP-AtATG8e construct or pJ4GFP construct alone were used to transform Arabidopsis suspension cell protoplasts. Protoplasts were starved in medium lacking sucrose for 48 h, or in sucrose-containing medium as a control, and protein was extracted from the transformed cells using cold 0.1 m Tris-HCl (pH 7.5), 0.3 m sucrose, 1 mm EDTA, followed by filtration through cheesecloth. Extracted protein was centrifuged at 125 000 g to separate soluble and insoluble proteins, and both fractions were separated by SDS-PAGE and transferred to nitrocellulose membrane. Rabbit anti-GFP antibodies (Invitrogen, Carlsbad, CA, USA) were used to detect the GFP-AtATG8e present in each fraction.

Alternatively, protoplasts expressing GFP-AtATG8e were incubated for 48 h in the presence or absence of sucrose and protein was extracted by grinding in cold PBS, 1 mm EDTA, 0.1 mm PMSF. Sucrose was added to each sample to a final concentration of 20% (w/v), and 0.5 ml of each sample was loaded onto a 1-ml layer of 25% sucrose in PBS, 1 mm EDTA, overlaying a 60% sucrose cushion. The samples were centrifuged for 2 h at 95 000 g at 4°C. Fractions (0.6 ml) were taken from the top of the sucrose gradient, corresponding to the load, 25% sucrose layer, and 60% sucrose interface. Protein was precipitated from each fraction using trichloroacetic acid and analyzed by immunoblotting using GFP antibodies as above.

Visualization of MDC-stained and GFP-labeled autophagosomes

Both MDC and GFP were visualized in vivo using a Zeiss Axioplan II compound microscope equipped with AxioCam HRC digital imaging system (Carl Zeiss Inc., Göttingen, Germany), for both light and UV-fluorescence microscopy. MDC-stained suspension cells, protoplasts, and tissues were visualized using a DAPI-specific filter, while GFP-AtATG8e-transformed protoplasts were visualized using an FITC-specific filter. GFP/MDC colocalization was confirmed using a DAPI/TRITC/FITC filter. The two images were allocated false green (FITC) and red (DAPI) colors. Image stacks were processed using Adobe Photoshop 6.0 (Adobe Systems, Mountain View, CA, USA). The rate of autophagosome movement was determined using Zeiss Axiovision version 3.0 software.

FM4-64 staining of protoplasts

Protoplasts prepared from suspension cells were transiently transformed with GFP-AtATG8e and starved for 48 h to induce autophagy. Protoplasts were incubated for 20 min at room temperature with a 40 μg ml−1 solution of FM4-64 in MS starvation medium. They were washed three times with the same medium, followed by incubation at room temperature for 6 h with shaking. Confocal microscopy was performed with a Leica TCS/NT confocal microscope (Leica Microsystems, Exton, PA, USA) fitted with a 100-mW argon ion laser (488 nm excitation), a 25-mW krypton laser (568 nm excitation) and tunable emission wavelength collection. A 100× Leica oil immersion objective was used throughout.

Acknowledgements

We thank Drs Tracey Pepper, Margie Carter, and Bob Doyle for assistance with microscopy, and members of the Hill and Miller labs for helpful discussions. This research was supported by the Plant Responses to the Environment Program of the National Research Initiative Competitive Grants Program, US Department of Agriculture (grant no. 2002-35100-12034 to D.C.B.) and by the Iowa State University Plant Sciences Institute.

Supplementary Material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ2396/TPJ2396sm.htm

Figure S1. Movement of MDC-stained autophagosomes.

Figure S2. Movement of GFP-AtATG8e-labeled autophagosomes.

Ancillary