Dissection of autophagy in tobacco BY-2 cells under sucrose starvation conditions using the vacuolar H⁺-ATPase inhibitor concanamycin A and the autophagy-related protein Atg8

Abstract

Tobacco BY-2 cells undergo autophagy in sucrose-free culture medium, which is the process mostly responsible for intracellular protein degradation under these conditions. Autophagy was inhibited by the vacuolar H⁺-ATPase inhibitors concanamycin A and bafilomycin A₁, which caused the accumulation of autophagic bodies in the central vacuoles. Such accumulation did not occur in the presence of the autophagy inhibitor 3-methyladenine, and concanamycin in turn inhibited the accumulation of autolysosomes in the presence of the cysteine protease inhibitor E-64c. Electron microscopy revealed not only that the autophagic bodies were accumulated in the central vacuole, but also that autophagosome-like structures were more frequently observed in the cytoplasm in treatments with concanamycin, suggesting that concanamycin affects the morphology of autophagosomes in addition to raising the pH of the central vacuole. Using BY-2 cells that constitutively express a fusion protein of autophagosome marker protein Atg8 and green fluorescent protein (GFP), we observed the appearance of autophagosomes by fluorescence microscopy, which is a reliable morphological marker of autophagy, and the processing of the fusion protein to GFP, which is a biochemical marker of autophagy. Together, these results suggest the involvement of vacuole type H⁺-ATPase in the maturation step of autophagosomes to autolysosomes in the autophagic process of BY-2 cells. The accumulation of autophagic bodies in the central vacuole by concanamycin is a marker of the occurrence of autophagy; however, it does not necessarily mean that the central vacuole is the site of cytoplasm degradation.

Abbreviations

ATG

autophagy-related

DMSO

dimethyl sulfoxide

GFP

green fluorescent protein

3-MA

3-methyladenine.

Introduction

Cells degrade their components in order to maintain and remake themselves. Autophagy is one of the mechanisms responsible for the degradation of these constituents.

Autophagy is a process in which a cell transports a part of its cytoplasm into lysosomes or vacuoles for degradation. It is classified as macroautophagy or microautophagy according to how the part of the cytoplasm is sequestrated and transported into the lysosomes or vacuoles.^1,2 In macroautophagy, a special membrane sac called an isolation membrane first sequestrates a part of the cytoplasm and makes a structure enclosed by a double membrane, an autophagosome. An autophagosome then fuses with a preexisting endosome and/or lysosome to transform to an autolysosome and degrades the enclosed cytoplasm. In microautophagy, a part of the cytoplasm is taken up into the lysosome by invagination and fission of the lysosomal membrane, and is then degraded.

The autophagy-related (Atg) protein Atg8 is a reliable marker of macroautophagy in yeast cells. By the analysis of mutants defective in macroautophagy, around 15 proteins responsible for the formation of autophagosomes have been found.^3-6 Of these proteins, which are designated as autophagy-related proteins, a ubiquitin-like protein Atg8 localizes to the structure called PAS (preautophagosomal structure or phagophore assemble site) and to autophagosomes.⁷ Atg8 protein is thought to work for the elongation and fusion of the membrane of an autophagosome precursor to complete the autophagosome structure.^8-10 It can therefore be used as a marker for the maturation process from a PAS to an autophagosome in yeast cells. The Atg8 protein is conserved among most eukaryotic cells.^11-15 The Arabidopsis genome codes for 9 homologues of the yeast Atg8 protein, which are designated AtAtg8a-i.^16,17 By expressing a fusion protein of one of these Atg8s or tobacco Atg8 and a fluorescent protein such as green fluorescent protein (GFP) or other fluorescent proteins, the emergence and movement of putative autophagosomes have been observed.^18-23

Concanamycin A is a specific inhibitor of vacuolar H⁺-ATPase, and blocks the acidification of organelles depending on this ATPase.^24-26 Concanamaycin A inhibits vacuolar H₊-ATPase activity and the acidification of organelles more strongly than another inhibitor bafilomycin A₁.²⁷ Since acidic organelles function in a variety of cellular processes, they have various effects on cellular physiology.^28-30 In Chara cells, the pH of the central vacuoles is raised by treatment with bafilomycin and concanamycin.^31,32 In tobacco BY-2 cells, concanamycin inhibits the transport of vacuolar resident proteins to the central vacuole, although it does not seem to affect the pH of central vacuoles at the concentrations used in this study.³³ This suggests that the acidification of organelles other than the central vacuole by the H⁺-ATPase is involved in the transport of vacuolar proteins. In the autophagic process of mammalian cells such as rat hepatocytes, the inhibitors block autophagy at the step of transformation from autophagosomes to autolysosomes.³⁴

In plant cells, the pathway of macroautophagy has not been elucidated clearly.³⁵ When transgenic plants expressing Atg8 labeled with fluorescent protein were treated with concanamycin, the structures derived from autophagosomes, which appear to correspond to autophagic bodies in yeast cells, are seen in the central vacuole of Arabidopsis root and hypocotyl cells.^18,20,36 These results have been interpreted as showing that autophagosomes directly fuse with the central vacuole and release their contents, autophagic bodies, into the lumen of the central vacuoles.³⁷ On the other hand, electron microscopic observations have shown the presence of autophagic vacuoles containing partially degraded cytoplasm in Arabidopsis and tobacco cells cultured in sucrose-free medium, suggesting that degradation of cytoplasm starts before autophagosomes fuse with the vacuoles.^38,39 Furthermore, in tobacco BY-2 cells, 2 fates of fluorescent autophagosomes are observed by fluorescence microscopy: one is direct fusion with the central vacuole; the other is interaction with more small vesicles to possibly become autolysosomes.²¹

Tobacco BY-2 cells cultured in sucrose-free medium perform macroautophagy. The addition of a protease inhibitor such as E-64c or leupeptin into the medium blocks the process and causes the accumulation of numerous autolysosomes containing undegraded cytoplasm in the perinuclear cytoplasm.^40,41 The autolysosomes are acidic inside and contain acid phosphatase and protease.³⁹ It has been thought that cysteine protease inhibitors retard the degradation of the cytoplasm enclosed in the autolysosomes, and as a result, autolysosomes containing undegraded cytoplasm accumulate in the cells, probably because of physical interference by accumulating cytoplasm.

In this study, we examined the pathway of autophagy in tobacco BY-2 cells using the vacuolar H⁺-ATPase inhibitor concanamycin and a fusion protein of GFP and AtAtg8. We found that concanamycin has a different effect than the cysteine protease inhibitor E-64c on cellular morphology. We report that concanamycin distorts the pathway of macroautophagy in tobacco BY-2 cells, although it is still useful for the detection of autophagy.

Results

Effect of vacuolar H⁺-ATPase inhibitor concanamycin A on vacuolar morphology of tobacco BY-2 cells

The morphology of tobacco BY-2 cells substantially changes under sucrose starvation (Fig. 1, see also Moriyasu and Ohsumi⁴⁰). In the logarithmic growth phase, the cells contained many cytoplasmic strands (Fig. 1A), which were gradually lost during starvation. Since vacuolar H⁺-ATPase is supposed involved in various cellular processes including autophagy, we first examined the effects of its inhibitor, concanamycin on the morphological changes of BY-2 cells under sucrose starvation. When 100 nM concanamycin was added to sucrose-free culture medium, many particulate structures accumulated in the central vacuoles (Fig. 1B vs. C). Because the structures were moving in the central vacuole in the Brownian manner, and appeared to be particles based on phase contrast microscopy (Fig. 1D, E), we inferred that these intravacuolar structures were not strands that go through the central vacuole but real particles that are dispersed in the lumen of central vacuoles. Concanamycin did not evoke such morphological change at 1 nM, whereas a small amount of accumulation of vacuolar inclusions was observed at 10 nM. We confirmed that bafilomycin A₁, another vacuolar H⁺-ATPase inhibitor, evoked the same morphological change at 5 µM, but not at 1 µM. This seems reasonable because concanamycin A inhibits vacuolar H⁺-ATPase more strongly than bafilomycin A₁.

Figure 1.

Effects of concanamycin on the morphology of tobacco BY-2 cells. Four-day-old BY-2 cells were transferred to sucrose-free culture medium, and the cellular morphology was observed by light microscopy. (A) Immediately after transfer to the medium containing 0.1 µM concanamycin. (B) Cultured for 1 d in medium containing 0.1 µM concanamycin. (C) Cultured for 1 d in medium containing 1% (v/v) DMSO as a solvent control. (D and E) Cells cultured in the presence of concanamycin for 2 d were observed under a light microscope with Nomarski (D) or phase contrast (E) optics. Bar, 20 µm.

Concanamycin at 100 nM did not cause this type of mor-phological change in medium containing sucrose (Fig. 2B vs. A), showing that the accumulation of intravacuolar particles is involved in cellular responses to sucrose starvation. However, at 1 µM, many vacuolar particles appeared even in nutrient-sufficient medium (Fig. 2C), and occasionally aggregated at corners of the central vacuole (Fig. 2D).

Figure 2.

Effects of concanamycin on tobacco BY-2 cells cultured in the presence or absence of sucrose. Four-day-old BY-2 cells were cultured in medium without sucrose (A) or with sucrose (B, C and D) containing 0.1 µM (A and B) or 1 µM (C and D) concanamycin for 1 d. Bar in A and B, 50 µm; bar in C and D, 20 µm.

Concanamycin A increases vacuolar pH

We examined whether concanamycin affects vacuolar acidification under the conditions used in this study. In order to roughly estimate the effects of concanamycin on vacuolar pH, the cells were prestained with the acidotrophic fluorescent dye quinacrine and then kept in sucrose-free medium in the presence or absence of 100 nM concanamycin (Fig. 3). Fluorescence microscopy showed that less quinacrine fluorescence remained in vacuoles in concanamycin-treated cells compared with untreated cells (Fig. 3, E vs. F), suggesting that alkalization of the central vacuole occurred in the presence of concanamycin. Such vacuolar alkalization occurred in less than 3 h at 1 µM concanamycin, whereas it took more than 12 h at 100 nM (data not shown). Such alkalization in the central vacuole also occurred in cells incubated in sucrose-containing medium in the presence of 100 nM concanamycin (Fig. 3D), although no accumulation of vacuolar inclusions occurred (Fig. 3A). This result confirms that concanamycin increases vacuolar pH both in the presence and absence of sucrose in culture medium, thus showing that the difference in cellular responses in the presence and absence of sucrose does not reflect a difference in pH perturbation by concanamycin.

Figure 3.

Effect of concanamycin on vacuolar pH. (A, B and C) Four-day-old tobacco BY-2 cells were kept in 0.1 M sorbitol, 5 mM HEPES-Na (pH 7.5) and 10 µM quinacrine for 5 min to stain their central vacuoles with quinacrine. The cells were then cultured for 1 d in medium with sucrose (A) or without sucrose (B and C) containing 0.1 µM concanamycin (A and B) or 1% (v/v) DMSO as a solvent control (C). (D, E and F) Fluorescence images corresponding to Nomarski images A, B and C, respectively. Bar, 40 µm.

Effect of concanamycin A on the accumulation of vacuolar particles is reversible

When cells accumulating many particles in the central vacuoles, like the cells shown in Figure 1B, were transferred to fresh sucrose-free culture medium in the absence of concanamycin, the particles disappeared within 2 d probably because they were digested (Fig. 4A and C). When the cells were transferred to MS medium containing sucrose, the vacuolar particles disappeared more rapidly (Fig. 4B and D), and concomitantly cells resumed cell division and reconstructed many transvacuolar strands. These results show that the effect of concanamycin on the accumulation of vacuolar particles is reversible and thus suggest that it does not represent some adverse effect of concanamycin, leading cells to death. Taken together, the results show that the central vacuole of a BY-2 cell is capable of digesting parts of the cytoplasm taken up into the vacuole, as shown in an alga, Chara corallina.⁴²

Figure 4.

Morphological changes of tobacco BY-2 cells after removal of concanamycin. BY-2 cells that had been cultured in sucrose-free culture medium containing 0.1 µM concanamycin for 1 d were cultured in the presence (B and D) or absence (A and C) of sucrose for another 1 d (A and B) or 2 d (C and D). Bar, 20 µm.

Electron microscopy confirms that a vacuolar particle is a membrane-enclosed vesicle containing part of the cytoplasm

Electron microscopy showed that the particles accumulating in the central vacuoles in the presence of concanamycin are parts of the cytoplasm (Fig. 5). Each of these cytoplasmic drops seemed to be surrounded by membrane, and cellular organelles such as mitochondria and ER appeared to be preserved in these drops (Fig. 5B, arrow). Figure 5C shows the presence of a membrane encircling a particle (arrowhead). These structures resembled the autophagic bodies of yeast cells, which are formed by the fusion of autophagosomes with the preexisting vacuole, although the structures in BY-2 cells were larger than those in yeast cells. Similar images have been reported in cells of Arabidopsis roots and hypocotyls treated with concanamycin, and in these reports, cytoplasmic drops were thought to be formed in the same manner as in yeast cells.^18,20 Besides membrane-enclosed cytoplasmic drops, membranous structures that seemed to have lost cytoplasm were also seen in the vacuoles (Fig. 5B, arrowhead). Furthermore, in BY-2 cells treated with concanamycin, autophagosomes (Fig. 5D, arrow) and possibly their precursor structures (Fig. 5D, arrowhead) were seen in the cytoplasm. These were approximately 1 to 2 µm in diameter and seemed to be enclosed by a double membrane. In addition to autophagosomes, structures that resembled autophagosomes, but were different from typical autophagosomes in that they were larger or contained more than 2 cytoplasmic drops inside (Fig. 5C, E; arrow), were also seen in the cytoplasm. Here, these structures are called autophagosome-like structures. Autophagosome-like structures are likely to be formed through the swelling of the outer membrane of a single autophagosome or the fusion of 2 or more autophagosomes. Autophagosomes and autophagosome-like structures were rarely seen in control cells that were treated with only dimethyl sulfoxide (DMSO) as a solvent control. These observations suggest that the number of autophagosomes and autophagosome-like structures increase following treatment with concanamycin, and that cytoplasmic drops in the central vacuole are structures released into the central vacuole from autophagosomes and autophagosome-like structures.

Figure 5.

Electron micrographs of tobacco BY-2 cells treated with concanamycin. Four-day-old BY-2 cells were cultured for 16 h in sucrose-free culture medium containing 0.1 µM concanamycin. v, central vacuole. n, nucleus. m, mitochondrion. p, plastid. (A) Many particles in the central vacuole. Bar, 5 µm. (B) Arrows, vacuolar particles containing parts of the cytoplasm. Arrowheads, membranous structures in the central vacuole. Bar, 1 µm. (C) Arrows, autophagosome-like structures; arrowheads, membrane of a vacuolar particle enclosing a mitochondrion but having lost cytosol. Bar, 1 µm. (D) Arrow, autophagosome. Arrowheads point to a putative precursor of an autophagosome. Bar, 1 µm. (E) Arrows, autophagosome-like structures. Bar, 2 µm.

In contrast to the effect of concanamycin, cysteine protease inhibitors such as E-64c and leupeptin cause the accumulation of autolysosomes, which contain the degradation intermediates of parts of the cytoplasm.^39,40 Parts of the cytoplasm enclosed in autolysosomes are mostly degraded and unidentified, although they are occasionally recognizable as originating from mitochondria.³⁹ Thus, the degradation of cytoplasm enclosed in an autolysosome is at a more advanced stage with respect to cytoplasmic degradation than cytoplasm in cytoplasmic drops accumulating in the central vacuoles or in autophagosomes and autophagosome-like structures. This suggests that concanamycin works upstream from E-64c.

Taken together, it seems that concanamycin somehow blocks the maturation of autophagosomes to autolysosomes in the process of autophagy in tobacco BY-2 cells induced under sucrose starvation conditions. As a result of such inhibition, autophagosomes probably accumulate in the cytoplasm and the accumulation of a large number of autophagosomes may result in their unusual fusion with the central vacuole to make cytoplasmic drops in the central vacuole.

Concanamycin A inhibits net protein degradation

Treatment with the papain-type cysteine protease inhibitor E-64c, as well as other inhibitors such as antipain and leupeptin, inhibits intracellular cysteine protease activity, and as a result blocks protein degradation through autophagy in tobacco BY-2 cells.⁴⁰ In contrast, concanamycin treatment did not reduce protease activity measured in vitro (Fig. 6A). However, like E-64c, concanamycin blocked net protein degradation during sucrose starvation, which is one of the signs of autophagy in BY-2 cells (Fig. 6B). As shown in a previous study⁴⁰, protein degradation during sucrose starvation is mostly non-selective. We confirmed this result and further found that almost all cellular polypeptides seem to be rescued from degradation by concanamycin (Fig. 6C), as is the case with E-64c. Thus the results confirm the notion that concanamycin works as a blocker of autophagy in tobacco BY-2 cells. Although we found that some specific polypeptides accumulated in the cells treated with concanamycin, we did not address our attention to these polypeptides in this study.

Figure 6.

Effects of concanamycin on cellular protease activity and protein contents. Four-day-old tobacco BY-2 cells were cultured in sucrose-free culture medium containing no (DMSO and EtOH), 0.1, and 1 µM concanamycin (Con) for 1 d (A) or for 2 d (B and C). DMSO or ethanol (EtOH) was added as a solvent control of concanamycin addition. t=0 means immediately after the start of culture in sucrose-free medium. (A) Cells were homogenized and cellular protease activities were measured at pH 5.5 using FTC-casein as a substrate. (B) Cellular protein contents were measured. (C) Cellular proteins from the same volume of culture were separated by SDS-PAGE and stained with silver. 0.1Con, 0.1 µM concanamycin for 2 d; 1Con, 1 µM concanamycin for 2 d. Molecular masses (kD) of marker proteins are shown on the left.

Concanamycin A suppresses accumulation of autolysosomes by E-64c treatment

When cells were treated with E-64c and concanamycin simultaneously, autolysosomes did not accumulate and the cellular morphology was the same as that of cells treated with only concanamycin (Fig. S1). This result strongly supports the notion that the point of action of concanamycin in the route of autophagy is upstream from that of E-64c.

The autophagy inhibitor 3-methyladenine (3-MA) suppresses the effect of concanamycin A

3-MA blocks autophagy induced by sucrose starvation in tobacco BY-2 cells.⁴³ Since the accumulation of autolysosomes is not seen in cells treated with 3-MA, the point of action of 3-MA is thought to be upstream from the formation of autolysosomes.⁴³ If the cytoplasmic drops that accumulate in the central vacuole following concanamycin treatment are structures derived from autophagosomes, 3-MA should abolish the accumulation of these structures. We used light microscopy to examine whether 3-MA blocks the accumulation of cytoplasmic drops in the central vacuole (Fig. S2). When cells were treated with concanamycin together with 3-MA, almost no accumulation of cytoplasmic drops occurred in the central vacuole (Fig. S2). This result supports the notion that vacuolar inclusions are derived from autophagosomes.

GFP-AtAtg8 protein is uniformly distributed in the cytoplasm and recruited to autophagosomes during autophagy

We transformed tobacco BY-2 cells to constitutively express GFP with its C terminus fused to Atg8 protein from Arabidopsis. Four-day-old transgenic cells expressing the GFP-AtAtg8 fusion protein appeared to have GFP fluorescence in much of the cytosol (Fig. 7), suggesting that the expressed proteins are distributed almost uniformly in the cytosol. However, 6 to 24 h after the cells were starved for sucrose, many punctate structures with distinct GFP fluorescence showed up (Fig. 7B and C, arrows). These particles appeared to exist in the cytoplasm. At a higher resolution, some of these structures seemed to be ring-shaped (Fig. 7D, arrows). By 2 d after sucrose deprivation, these fluorescent particles mostly disappeared, and instead, somewhat bright GFP fluorescence was observed in the lumen of the central vacuole (Fig. 7E). These results suggest that GFP-AtAtg8 in the cytosol was recruited to autophagosomes, which are recognized as punctate or ring-shaped structures under fluorescence microscopy, and subsequently moved into the central vacuole during the process of autophagy.

Figure 7.

Morphological changes of transgenic cells expressing GFP-AtAtg8 fusion protein in response to sucrose starvation. (A, B, C, D and E) Four-day-old BY-2 cells expressing GFP-AtAtg8 (A) were cultured under sucrose starvation conditions for 1 d (B, C and D) or 2 d (E), and observed by a confocal laser microscope. (F, G and H) Four-day-old BY-2 cells expressing GFP (F) were cultured under sucrose starvation for 1 d (G) or 2 d (H). For A, B, E, F, G and H, the Nomarski images are on the left; the fluorescence images on the right. Bars represent 20 µm.

GFP fluorescence also appeared to be distributed in the cytoplasm of cells expressing native GFP protein (Fig. 7F). However, no punctate structures having GFP fluorescence appeared even when these cells were transferred to sucrose-deficient medium (Fig. 7G). By 2 d after sucrose deprivation, significant GFP fluorescence was observed in the central vacuole of some cells (Fig. 7H), suggesting that GFP in the cytosol is transported into the central vacuole as a substrate for autophagy.

Autolysosomes accumulated by cysteine protease inhibitors take over GFP fluorescence from autophagosomes

We examined whether the autolysosomes accumulated after treatment with the cysteine protease inhibitor E-64c acquire GFP fluorescence from autophagosomes labeled with GFP-Atg8. The autolysosomes had distinct GFP fluorescence, confirming that they are a successive organelle derived from the autophagosomes (Fig. S3).

The autophagy inhibitor 3-MA blocks the formation of autophagosomes

When 3-MA was applied to culture medium, autophagosomes did not appear (Fig. S4, A). In contrast, in control cells, where water, the solvent for the 3-MA solution, was added to the medium, fluorescent autophagosomes appeared (Fig. S4, B, arrows). Based on counting of the autophagosomes, 3-MA substantially blocked their formation (Fig. S4, E). The result is consistent with our previous result that 3-MA substantially blocked the formation of autolysosomes.⁴³ After 2 d, autophagosomes were seldom seen, and a significant portion of the GFP fluorescence moved into the central vacuole in control cells (Fig. S4, D), whereas such movement did not occur in 3-MA-treated cells (Fig. S4, C). This result suggests that the final destination of autophagosomes is the central vacuole.

GFP-AtAtg8 is localized on cytoplasmic particles accumulating in the central vacuole

To examine GFP-AtAtg8 localization, we cultured transformed cells expressing GFP-AtAtg8 in sucrose-free medium containing concanamycin. Many particles were observed to be moving in a Brownian manner in the central vacuole (Fig. 8A). The lumen of the central vacuole also had GFP fluorescence, but these particles had stronger GFP fluorescence. Not all particles in the vacuole had GFP fluorescence, and in some cells, only a small proportion of the particles had stronger GFP fluorescence pervading the vacuolar lumen (Fig. 8B). It is likely that GFP associated with the particles at first gradually became distributed in the vacuolar lumen. In contrast, in control cells without concanamycin, autophagosomes were seen in the cytoplasm (Fig. 8C). Autophagosomes were presumed to mingle with cytoplasmic particles in the central vacuole in concanamycin-treated cells (Fig. 8A); however, we could not strictly discriminate autophagosomes from many of the other vacuolar particles by fluorescence microscopy. This observation supports the idea that the particles accumulated by concanamycin treatment are the structures formed by the fusion of autophagosomes with the central vacuole, corresponding to autophagic bodies in yeast cells.

Figure 8.

Effects of concanamycin on changes in the morphology of tobacco BY-2 cells expressing GFP-AtAtg8 fusion protein under sucrose starvation conditions. BY-2 cells expressing GFP-AtAtg8 (A, B and C) or VM23-GFP (D and E) were cultured under sucrose starvation condition for 1 d in the presence of 0.1 µM concanamycin (A, B and D). DMSO was used at 1% (v/v) as a solvent control of concanamycin addition (C and E). The Nomarski images are on the left; the fluorescence images on the right. Bars represent 20 µm.

If a vacuolar particle is formed by the fusion of an autophagosome with the central vacuole, the membrane encircling the particle should become the inner membrane of the autophagosome and be distinct from the vacuolar membrane. To try to prove this, we used BY-2 cells that constitutively express a fusion protein of GFP and the vacuolar membrane protein VM23 from radish.⁴⁴ In the transgenic cells, the vacuolar membrane had GFP fluorescence.⁴⁵ When such cells were transferred to sucrose-free culture medium containing concanamycin, many particles formed in the vacuole (Fig. 8D), but fewer had GFP fluorescence. Furthermore, the vacuolar lumen did fluoresce less than GFP-Atg8-expressing cells (Fig. 8D vs. A and B), showing that the membranes surrounding most vacuolar particles are distinct from vacuolar membranes.

Taken together, the results from transgenic cells expressing GFP-Atg8 and VM23-GFP support the notion that vacuolar particles are derived from autophagosomes. It should, however, be noted that a significant portion of the vacuolar particles formed by concanamycin seemed to have GFP-VM23 fluorescence, suggesting that invagination of portions of the vacuole membrane, such as occurs during microautophagy, occurs under these experimental conditions.

GFP-AtAtg8 fusion protein is degraded into GFP in response to sucrose starvation

To confirm that the fusion protein with the correct size is expressed and to examine whether the expressed protein is altered in response to sucrose starvation, we analyzed cellular proteins of transformed cells that had been starved for sucrose (Fig. 9). The pattern of polypeptides stained with silver did not change drastically during 1 d of sucrose starvation (Fig. 9A). Western blotting with anti-GFP antibody detected the presence of only one band of approximately 40 k_D in 4-d-old cells, which are in the logarithmic growth phase (Fig. 9B, arrowhead). Antibody against an N-terminal peptide of Atg8 protein also recognized this band (Fig. 9C, arrowhead). Thus this band was regarded as the fusion protein of 28 k_D GFP and 12 k_D Atg8. During sucrose starvation, the band corresponding to the fusion protein became faint, and instead, a broad band, likely composed of several bands, appeared around 28 k_D. Antibody against Atg8 peptide did not recognize this band around 28 k_D (Fig. 9C). The broad 28 k_D band was likely composed of GFP alone and GFP variants truncated at the N- and/or C-terminus. After 2 d of starvation, the fusion protein nearly disappeared, whereas the intensity of bands corresponding to GFP variants became stronger. Thus, the linkage between GFP and AtAtg8 is cleaved in response to sucrose starvation. Under sucrose starvation, the GFP moiety appeared to remain undegraded 2 d after sucrose deprivation, although synthesis of the fusion protein seemed to stop.

Figure 9.

Cleavage of GFP-AtAtg8 fusion protein in tobacco BY-2 cells under sucrose starvation conditions. BY-2 cells expressing GFP-AtAtg8 fusion protein or only GFP were transferred to sucrose-free culture medium. Immediately, 1 d, and 2 d after transfer (0, 1, and 2, respectively), proteins were extracted from the cells. (A) Cellular proteins were separated and stained with silver. The same amount of proteins was loaded into each lane. Molecular masses (kD) of marker proteins are shown on the left. (B) Western blotting using anti-GFP antibody. Arrowhead, GFP-AtAtg8 band.

Also in cells expressing GFP alone, a broad band of approximately 28 k_D was detected, suggesting that the expressed GFP is partially degraded. This band became faint in response to sucrose deprivation, suggesting that net degradation of GFP occurs. Furthermore, a slight shift of the bands to a lower molecular mass was observed, suggesting that GFP in the cytoplasm is transported to lytic compartments by autophagy and further truncated.

In parallel with the effect of 3-MA on morphological changes, 3-MA inhibited the cleavage of GFP-AtAtg8 fusion protein to GFP and Atg8 (Fig. S5). In the absence of 3-MA, the intensity of GFP bands increased from day 1 to day 2. In addition, 3-MA inhibited this increase. Unlike 3-MA, concanamycin did not have any inhibitory effect on the cleavage of GFP-AtAtg8 fusion protein. Rather, concanamycin promoted the accumulation of bands corresponding to GFP in the cells. This is probably due to concanamycin promoting the fusion of autophagosomes with the central vacuoles. Alternatively, it is likely that concanamycin raises the vacuolar pH and inhibits the degradation of GFP in the vacuoles. E-64c neither inhibited cleavage of GFP-AtAtg8 nor enhanced the accumulation of GFP bands, suggesting that the protease responsible for the splitting of the fusion protein is not a papain-type cysteine protease.

Discussion

Our previous and present results obtained using 3 kinds of inhibitors, the autophagy inhibitor 3-MA, the vacuolar H⁺-ATPase inhibitors concanamycin and bafilomycin and the cysteine protease inhibitor E-64c, clearly show a stepwise process of autophagy in tobacco BY-2 cells. All three inhibitors block the net protein degradation that occurs under sucrose starvation conditions, which is consistent with the idea that all 3 inhibitors block autophagy, which occurs under starvation. 3-MA did not cause the accumulation of any special structure under starvation conditions, whereas the accumulation of autophagosome-related structures and autolysosomes was respectively observed following treatment with concanamycin and E-64c. When the cells are treated with E-64c, autolysosomes, in which portions of the cytoplasm appear to already be partially degraded, accumulate. In contrast, when the cells are treated with concanamycin, portions of the cytoplasm with a boundary membrane, which resemble autophagic bodies in yeast cells, are seen in the central vacuole as well as autophagosomes and autophagosome-like structures in the cytoplasm. Cells treated simultaneously with concanamycin and E-64c exhibit the same morphology as those treated with concanamycin alone, supporting the notion that the point of action of concanamycin is upstream from that of E-64c. That 3-MA suppresses both the effects of concanamycin and E-64c suggests that the point of action of 3-MA is upstream from those of concanamycin and E-64c.

Membrane-enclosed parts of the cytoplasm, which are derived from autophagosomes, accumulated in the central vacuole when tobacco BY-2 cells were starved for sucrose in the presence of the vacuolar H⁺-ATPase inhibitor concanamycin. Since vacuolar pH increased following treatment with concanamycin, vacuolar hydrolases that have their optimal pH in the acidic range should be inhibited. Thus if parts of the cytoplasm were transported into the central vacuole in the presence of concanamycin, they should remain undegraded in the central vacuole. Indeed, our results showed that high vacuolar pH is necessary for the retention of parts of the cytoplasm taken up into the central vacuole, since they disappeared after concanamycin had been washed out (Fig. 4).

However, we reason that the primary effect of concanamycin on autophagy is distortion of the normal macroautophagic pathway in tobacco BY-2 cells. The accumulation of portions of the cytoplasm that occurs in the presence of concanamycin is likely to be caused by a block in transformation of autophagosomes to autolysosomes. The accumulation of numerous autophagosomes in the cytoplasm probably leads to the fusion of the autophagosomes with the central vacuole and the accumulation of parts of the cytoplasm in the central vacuole. Thus, we think that the accumulation of many cytoplasmic drops in the central vacuole does not mean that the central vacuole is the compartment where degradation occurs during autophagy in tobacco BY-2 cells, but rather reflects an artifact.

The following observations support the notion that almost no autophagosomes directly fuse with the central vacuole in the normal autophagic pathway of BY-2 cells.

First, in mammalian and yeast cells, parts of the cytoplasm to be degraded accumulate in the compartment for protein degradation during autophagy under condition inhibiting proteolysis. When mammalian cells are treated with a cysteine protease inhibitor under nutrient-starvation conditions, parts of the cytoplasm remain in the lytic compartments of mammalian cells, the lysosomes, and as a result, autolysosomes accumulate in the cells.⁴⁶ In yeast cells undergoing nutrient starvation that are treated with a serine protease inhibitor, phenylmethylsulfonyl fluoride (PMSF), parts of the cytoplasm accumulate in the vacuole, which is the lytic compartment of yeast cells.⁴⁷ In addition to protease inhibitor analyses, mutants that are defective in vacuolar protease activities have been obtained in yeast cells. In these mutants, parts of the cytoplasm, called autophagic bodies, accumulate in the vacuole in the process of autophagy.³ We have shown that tobacco BY-2 cells starved for sucrose in the presence of a protease inhibitor such as leupeptin or E-64c, cytoplasmic particles, which represent partially-degraded parts of the cytoplasm, accumulate in autolysosomes, instead of in the central vacuole.⁴⁰ Thus it is reasonable to suppose that the lytic compartments for autophagy in tobacco BY-2 cells are autolysosomes.

Secondly, concanamycin evoked not only the accumulation of autophagic bodies in the central vacuole, but also the accumulation of autophagosomes and autophagosome-like structures in the cytoplasm. The transformation of autophagosomes to autolysosomes is blocked by bafilomycin A₁ and concanamycin A in mammalian cells.³⁴ Thus it is reasonable to suppose that concanamycin blocks the maturation of an autophagosome to an autolysosome. Although the target H⁺-ATPase of these inhibitors is unknown, it is likely that they inhibit membrane fusion by alkalizing the lumen of lysosomes and endosomes. Thus, autophagy in BY-2 cells may be perturbed because concanamycin alkalizes some acidic organelles related to autophagy in addition to the central vacuole. Although we did not confirm the alkalization of organelles other than the central vacuole, it is likely that other acidic organelles are also alkalized by treatment with concanamycin. It is known that Golgi cisternae in BY-2 cells swell in the same treatment, probably originating from their alkalization.⁴⁸ In BY-2 cells, concanamycin inhibits the transport of proteins to vacuoles although the vacuolar pH is normal.³³

Taking this evidence together strengthens our notion that the main lytic compartments of autophagy in tobacco BY-2 cells are autolysosomes and not the central vacuole. Some autolysosomes fuse with the central vacuole, and final degradation is in the central vacuole, but almost all autolysosomes complete the degradation of the cytoplasm enclosed in them. However, in some plant species, autophagosomes appear to directly fuse with the central vacuole in a normal autophagic process.⁴⁹ Our observation of barley and Arabidopsis root tips treated with the cysteine protease inhibitor E-64d, an esterified form of E-64c, also demonstrated that parts of the cytoplasm accumulate in the central vacuole as well as in newly formed autolysosomes. Why the sites for degradation of cytoplasmic components in autophagy are different in different cell types remains unknown. How the contribution of these 2 organelles to autophagy is regulated in plant cells is an issue for future research.

How transformation of autophagosomes to autolysosomes occurs in tobacco BY-2 cells is not fully understood. In mammalian cells, it proceeds by the fusion of autophagosomes with preexisting lysosomes and/or endosomes. In the case of yeast, newly formed autophagosomes fuse with the preexisting vacuole to form a vacuole containing autophagic bodies. As noted above, in some plant species, autophagosomes directly fuse with the central vacuole, as in yeast cells. For tobacco autophagosomes to be transformed to autolysosomes, there must be many lines of transport of membrane vesicles to autophagosomes or autolysosomes. We previously found that the styryl dye FM4-64 on the plasma membrane flows into autolysosomes probably through an endocytic pathway.⁴⁵ Thus, it is probable that as in mammalian cells, the fusion of autophagosomes with endosomes contributes to the formation of autolysosomes in tobacco BY-2 cells. Our previous results also showed that FM4-64 residing on the membrane of central vacuoles and a protein existing in the lumen of central vacuoles move to autolysosomes under sucrose starvation conditions, suggesting a flow of proteins and membrane lipids from the central vacuole to autophagosomes/autolysosomes. Thus, it is possible that hydrolytic enzymes supplied from the central vacuole to autophagosomes also contribute to the transformation of autophagosomes to autolysosomes. It is therefore conceivable that concanamycin blocks these processes by increasing the internal pH of endosomes and/or the central vacuole. In the absence of E-64, protein degradation continues unabated. If digestion occurs in autolysosomes before their fusion with the central vacuole, the number of empty lysosomes should increase under sucrose starvation in the absence of E-64c. This notion is supported by the previous observation that the number of empty vesicles, probably lysosomes, increases significantly in tobacco cells starved for sucrose in the absence of E-64c.⁴³

In this study, we succeeded in visualizing autophagosomes using a light microscope. The recruitment of GFP-AtAtg8 protein to autophagosomes is a good marker of autophagy in living plant cells as in yeast and mammalian cells. A large amount of GFP-AtAtg8 is expressed in the cells. Comparing the results from morphological analysis and immunoblotting, GFP-AtAtg8 is thought to be degraded into GFP and AtAtg8 moieties as autophagy proceeds. The processing of GFP-AtAtg8 to GFP and AtAtg8 can act as a biochemical marker of autophagy in tobacco cells, and perhaps as a general biochemical marker of autophagy in plant cells.

Materials and Methods

Plant material

Cultured tobacco (Nicotiana tabacum) BY-2 cells were used. The cells were cultured in Murashige and Skoog culture medium containing 3% (w/v) sucrose and 0.2 mg/L 2,4-D as described previously.⁴⁰ The cells were starved for sucrose as described previously.⁴⁰

Chemicals

3-MA was purchased from Sigma. Concanamycin and bafilomycin were obtained from Wako Pure Chemicals Ind., Ltd (Tokyo, Japan). and dissolved in DMSO or ethanol as 100x concentrated solutions. E-64c was from Peptide Institute Inc. (Osaka, Japan). Fluorescein isothiocyanate-labeled casein was prepared⁴¹ according to Twining.⁵⁰

Light microscopy

Light microscopes (OptiPhoto, Nikon and BX-51, Olympus) equipped with fluorescence and Nomarski differential interference contrast optics were used. Photographs were taken using a photomicrographic camera (Microflex UFX-II, Nikon) and negative monochrome (Presto, ISO 400, Fuji) and color (Fujicolor, ISO400, Fuji) film or a digital camera (DS-Fi1, Nikon). For obtaining confocal images, confocal laser microscopes (LSM510, Zeiss) were used. To stain cells with quinacrine, they were kept in 0.1 M sorbitol, 5 mM HEPES-NaOH (pH 7.5) and 10 µM quinacrine at room temperature for about 5 min.

Protease assays

Cells were collected on a glass filter (GF/A, 47 mm in diameter, Whatman) from 3 mL of culture, transferred to a Teflon-homogenizer and homogenized with 1 mL of 0.1 M acetate-NaOH (pH 5.0) buffer containing 28 mM 2-mercaptoethanol.⁴⁰ The homogenate was centrifuged at 15,000 rpm for 10 min. The resultant supernatant (60 µL) and 40 µL 0.5% (w/v) fluorescein isothiocyanate-labeled casein were mixed and incubated at 37°C for 30 min. The protease reaction was stopped by adding 100 µL 10% (w/v) trichloroacetic acid. After centrifugation, fluorescence at 525 nm was measured with an excitation wavelength at 490 nm.

Protein assay, SDS-PAGE and western blotting

Cells in 1 mL of culture medium were collected on a glass filter and transferred to a Teflon homogenizer. They were homogenized with 2 mL of 100 mM HEPES-NaOH (pH 7.5), 1 mM EDTA, 100 µM leupeptin, 1 mM phenylmethylsulfonyl fluoride or 4-(2-aminoethyl)benzenesulfonyl fluoride, and 28 mL 2-mercaptoethanol. The homogenate was centrifuged at 15,000 rpm for 10 min at 4°C. The protein content in the supernatant was measured according to the method of Lowry et al.⁵¹, modified by Bensadoun and Weinstein.⁵² BSA was used as a standard. Proteins in the supernatant were also separated by electrophoresis on SDS-polyacrylamide gradient gels (Ready Gel 10/20, Bio-Rad; or PAG Mini Daiichi 11/14 or 14/15, Daiichi Pure Chemicals, Co., Ltd.) and visualized by silver staining.

For protein gel blotting, proteins in a polyacrylamide gel were electroblotted onto nitrocellulose membrane in a solution consisting 25 mM Tris, 192 mM glycine and 20% methanol. After transfer, the membrane was immersed in 10 mM Tris-HCl (pH 8.0) containing 5% (w/v) skim milk, 150 mM NaCl and 0.05% (w/v) Tween 20 at 4°C overnight. Anti-GFP antibody (GFP Polychlonal Antibody (IgG Fraction) #8363, Clontech Lab., Inc., CA, USA) was diluted 1:1000. The secondary antibody used was anti-rabbit IgG antibody conjugated with alkaline phosphatase (Anti-Rabbit IgG (Fc) AP Conjugate, Promega). Binding of the primary and secondary antibodies to membranes was done in 10 mM Tris-Cl (pH 8.0) containing 150 mM NaCl and 0.05% Tween 20. Nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate were used as substrates for alkaline phosphatase.

Electron microscopy

Cells were fixed with 2% (w/v) glutaraldehyde and 1% (w/v) formaldehyde in 100 mM sodium cacodylate-HCl (pH 6.9) buffer at room temperature for 1 h and at 4°C overnight. After the samples were treated with 1% (w/v) tannic acid for 1 h at room temperature, they were fixed again with 1% (w/v) osmium tetroxide for 2 h at room temperature. The samples were stained en block with 2% (w/v) uranyl acetate for 2 h at room temperature. Thereafter they were dehydrated by an ethanol series and propylene oxide, and embedded in Spurr's resin. Thin sections were made from these blocks and stained with uranyl acetate and lead citrate. They were observed with an electron microscope (H-7000, Hitachi).

Construction of plasmids

RNA was isolated from roots of Arabidopsis seedlings and single-stranded cDNA synthesized using reverse transcriptase. The AtATG8d open reading frame was amplified by PCR using the DNA as templates and the following 2 primers: the sense primer, 5′- tttgtacatggcgattagctccttcaa-3′, comprises the BsrG I site and 20 bases of oligoDNA from the 5′-end of the coding region on the coding strand for AtATG8d DNA; the antisense primer, 5′- ccgcggccgcttagaagaagatcccgaacg-3′, includes the Not I site and 20 bases of oligoDNA from the 5′-end of the coding region on the anticoding strand for AtATG8d DNA. The amplified DNA was digested with BsrG I and Not I, and then ligated into the BsrGI-Not I site of the GFP expression vector⁵³ to make a plasmid expressing GFP-AtAtg8d fusion protein under control of the cauliflower mosaic virus 35S promoter.

The plasmid was cut with Hind III and EcoR I to obtain the DNA fragment, the 35S promoter-synthetic GFP DNA-AtAtg8d-the poly A signal of the nopaline synthase gene. The DNA region from Hind III to EcoR I site of the binary vector pSMAB701 (see Niwa et al.⁵³) was replaced with the DNA fragment. The resulting plasmid was used for Agrobacterium-mediated transformation of tobacco cells.

Sucrose starvation

Four-day-old culture cells were transferred to a centrifuge tube. Cells were collected by centrifugation at 100 × g for 4 min. After the supernatant was removed, cells were washed with culture medium lacking sucrose and resuspended in the original volume of culture medium lacking sucrose. The cell suspension was transferred to a Petri dish and kept at 26 ±1°C with rotation at 110 rpm. E-64c was added to the culture medium from a 1 mM stock solution to make a final concentration 10 µM. 3-MA was added from a 0.1 M stock solution to a final concentration of 5 mM. Concanamycin was added from a 10 µM stock solution to a final concentration 0.1 µM, unless otherwise noted. When these chemicals were added to culture medium, the same volume of solvents used to make the stock solutions was added in parallel to controls.

Funding

This work was supported by KAKENHI [grant number 23120504, 25120704 to Y.M.].

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.