In vivo imaging and quantitative monitoring of autophagic flux in tobacco BY-2 cells.

Autophagy has been shown to play essential roles in the growth, development and survival of eukaryotic cells. However, simple methods for quantification and visualization of autophagic flux remain to be developed in living plant cells. Here, we analyzed the autophagic flux in transgenic tobacco BY-2 cell lines expressing fluorescence-tagged NtATG8a as a marker for autophagosome formation. Under sucrose-starved conditions, the number of punctate signals of YFP-NtATG8a increased, and the fluorescence intensity of the cytoplasm and nucleoplasm decreased. Conversely, these changes were not observed in BY-2 cells expressing a C-terminal glycine deletion mutant of the NtATG8a protein (NtATG8aΔG). To monitor the autophagic flux more easily, we generated a transgenic BY-2 cell line expressing NtATG8a fused to a pH-sensitive fluorescent tag, a tandem fusion of the acid-insensitive RFP and the acid-sensitive YFP. In sucrose-rich conditions, both fluorescent signals were detected in the cytoplasm and only weakly in the vacuole. In contrast, under sucrose-starved conditions, the fluorescence intensity of the cytoplasm decreased, and the RFP signal clearly increased in the vacuole, corresponding to the fusion of the autophagosome to the vacuole and translocation of ATG8 from the cytoplasm to the vacuole. Moreover, we introduce a novel simple easy way to monitor the autophagic flux non-invasively by only measuring the ratio of fluorescence of RFP and YFP in the cell suspension using a fluorescent image analyzer without microscopy. The present in vivo quantitative monitoring system for the autophagic flux offers a powerful tool for determining the physiological functions and molecular mechanisms of plant autophagy induced by environmental stimuli.

Introduction

Macroautophagy (hereafter referred to as autophagy) is an evolutionarily conserved system for degradation of intracellular components through the vacuole/lysosome, which has been shown to play essential roles in the growth, development and survival of eukaryotic cells. Autophagy is a process that starts with the formation of a cup-shaped phagophore in the cytoplasm that becomes a double-membrane structure, called the autophagosome, and engulfs long-lived proteins and organelles.- The autophagosome subsequently fuses with the vacuole/lysosome, and the inner membrane, the autophagic body, is degraded by hydrolytic enzymes.- These changes are called autophagic flux.

Various methods have been used to analyze autophagic flux in plants. Though transmission electron microscopy is effective to directly observe autophagy-related intracellular structures,, it is not suitable for characterization of their in vivo dynamics. A cysteine protease inhibitor E-64 and a vacuolar H+-ATPase inhibitor concanamycin A (CA) allow light microscopic observation of intracellular accumulation of autolysosomes and autophagic bodies, respectively., However, this technique is not suitable for real-time imaging or quantitative monitoring because accumulation of these visible structures takes a long time. Concanamycin A may also affect the activity of normal autophagic flux. Another method uses monodansylcadaverine (MDC) and Lysotracker as a marker probe for acidic compartments to detect autolysosome-like structures., However, these methods are not specific for autophagy, and therefore need to be combined with other techniques.

Over 30 known autophagy-related genes (ATGs) have been identified in yeast, many of which are conserved in most eukaryotes including mammals and plants., Formation of the autophagosomes requires two ubiquitin conjugation-like reactions with ATG12 and ATG8. The C-terminal glycine-residue of ATG8 has been shown to be essential for the conjugation reaction and autophagosome formation., In animal cells, electrophoretic detection of the lipidation of microtubule-associated protein light chain 3 (LC3), a mammalian homolog of ATG8, has been established as a quantitative marker for autophagy. However, this method has not been established in plant cells.

ATG8 fused with the green fluorescent protein (GFP) have enabled the observation of autophagosomes in living cells.,,, The GFP fluorescence can be observed in punctate and ring-shaped structures, which increase under nutrient starvation or stress condition., These structures are transported from the cytoplasm to the vacuole,, and transported GFP-ATG8 protein is processed in vacuolar lumens. The processing of GFP-ATG8 protein has been detected by western blot analysis,, but this method is not suitable for non-invasive real-time detection. The tandemly fused fluorescent proteins have been used for monitoring the mitophagy and the autophagosome maturation in yeast and animals., However, in plant cells, such approaches have not been applied to monitor autophagy, and simple methods for quantification of autophagic flux remain to be developed.

Tobacco BY-2 cells are superior for analyzing the intracellular localization and dynamics of proteins and organelles, as well as cell cycle-related phenomena. Four isoforms of ATG8 genes named NtAtg8a, b, c, d have been found from the EST database (http://mrg.psc.riken.go.jp/strc/index.htm). NtAtg8a mRNA has been suggested to be expressed in lag, log and stationary phase cells. NtAtg8b-d have also been obtained from a cDNA library generated from cells treated with several plant hormones or under sucrose starvation conditions.

We here established a non-invasive monitoring system for autophagic flux in tobacco BY-2 cells expressing NtATG8a fused to a variety of fluorescent tags. Simultaneous in vivo imaging of the autophagosome formation, decrease in cytosolic ATG8 and accumulation of ATG8 in the vacuole in living cells allowed characterization of in vivo dynamics of autophagic flux. Furthermore, we introduce a novel simple method to monitor the autophagic activity in living cells by ratiometric fluorescence measurement. These in vivo quantitative monitoring systems of autophagy should provide a powerful tool for characterizing autophagy in plant cells.

Results and Discussion

In vivo imaging of autophagic flux

To visualize the dynamics of the autophagic flux in tobacco BY-2 cells, we generated a transgenic tobacco BY-2 cell line (BY-YA8) stably expressing a YFP-NtATG8a fusion protein under the control of the cauliflower mosaic virus 35S promoter. Under normal growth conditions, YFP fluorescence was detected in the cytoplasm and nucleoplasm of 3-d-old cultured BY-YA8 cells (Fig. 1A, Control). A few punctate signals of YFP-NtATG8a were observed in the cytoplasm. When the BY-YA8 cells were transferred to sucrose-free medium, an increase of punctate signals (Fig. 1A, Starvation) was observed. It reached a plateau at 2–3 h and did not change until 6 h under sucrose-starved conditions (Fig. 1B).

Figure 1. Visualization of sucrose starvation-induced autophagosome formation in tobacco BY-2 cells. (A) Three-day-old BY-2 cells expressing the YFP-NtATG8a construct were incubated in complete (Control) or sucrose-free medium (Starvation) for 3 h. Confocal fluorescence (a, c, e) and differential interference contrast (DIC) images (b, d) were obtained by CLSM. Arrows indicate punctate signals of YFP-NtATG8a. (A-e) Close up of the fluorescence image of YFP-NtATG8a in (A−c). Scale bars: 50 μm (a–d) and 20 μm (e). The data are representative of three experiments. (B) Time course graph of sucrose starvation-induced autophagosome formation. To quantify the levels of autophagic flux, the numbers of YFP punctate signals per 10 cells were counted at the indicated time points. The open and closed circles indicate sucrose-rich and sucrose-free conditions, respectively. Data are the means ± SE of three independent experiments. (C) Effects of PI3K inhibitors on sucrose starvation-induced autophagosome formation. Three-day-old BY-2 cells were incubated in complete or sucrose-free medium for 3 h with 3-MA (5 mM) or Wortmannin (10 μM). D.W. was used as a control. Scale bars: 50 μm. The data are representative of three experiments. (D) Quantitative levels of autophagosome formation in (C). The numbers of YFP punctate signals per 10 cells were counted. The open and closed bars indicate sucrose-rich and sucrose-free conditions, respectively. Data are the means ± SE of three independent experiments.

The phosphoinositide 3-kinase (PI3K) plays an essential role in the formation of the autophagosome. A PI3K inhibitor, 3-methyladenine (3-MA), has been shown to inhibit autophagy in many eukaryotic cells including the BY-2 cells. To confirm if the punctate signals derived from YFP-NtATG8a corresponds to the autophagosome, we tested the effects of several PI3K inhibitors. As shown in Figure 1C, the presence of 3-MA or wortmannin in culture media for 3 h clearly inhibited the number of punctate signals compared with the control (Fig. 1C and D), suggesting that the punctate signals are the autophagosomes.

The C-terminal glycine residue of ATG8 is essential for the association with the autophagosome in all eukaryotic cells, and deletion or point mutation of this glycine residue in the ATG8 protein has been used as a negative control of autophagic flux. The C-terminal glycine residue is conserved in tobacco ATG8 homologs. Thus we also established a transgenic BY-2 cell line (BY-HGA8ΔG) expressing an HA-tagged GFP fused with a C-terminal glycine deletion mutant of the NtATG8a protein (Fig. 2A, NtATG8aΔG). In BY-HGA8ΔG cells, the GFP signal was detected in the cytoplasm and nucleoplasm similarly to the BY-YA8 cells (Fig. 2B, Control). However, when the cells were transferred to sucrose-free medium for 5 h, in contrast to the BY-YA8 cells, no punctate signals were observed (Fig. 2B, Starvation). This indicates that the C-terminal glycine residue of NtATG8a is also essential for the association with the autophagosome in tobacco, and the punctate structures are indeed autophagosomes.

Figure 2. Subcellular localization of the GFP-NtATG8aΔG mutant protein in tobacco BY-2 cells. (A) Schematic diagrams of the C-termini of NtATG8a and its mutant defective in the Gly116 residue. The Gly116 residue of NtATG8a is shown in underlined. (B) Subcellular localization of the GFP-NtATG8aΔG mutant protein in tobacco BY-2 cells. Transgenic BY-2 cells were incubated in complete (Control) or sucrose-free medium (Starvation) for 5 h. These fluorescence images were obtained by CLSM. Scale bar: 20 μm. The data are representative of three experiments.

A vacuolar H+-ATPase inhibitor, CA, induces alkalinization of the vacuole and accumulation of vesicular structures, which are assumed to be autophagic bodies.,, Under sucrose-starved conditions, we also observed similar vesicular structures in the vacuole after CA treatment in both BY-YA8 (Fig. 3A, DIC) and BY-HGA8ΔG cells (Fig. 3B, DIC), suggesting that these vesicular structures observed in BY-2 cells correspond to autophagic bodies.

Figure 3. The accumulation of YFP-NtATG8a protein in vacuoles after treatment with concanamycin A. Three-day-old transgenic BY-2 cells expressing the YFP-NtATG8a protein (A) and the HA-GFP-NtATG8aΔG protein (B) were incubated in sucrose-free medium for 24 h in the presence of CA (1 μM). Control was at a time 0 points. These fluorescence and DIC images were obtained by CLSM. Scale bar: 20 μm. V, vacuole; N, nucleus. The data are representative of three experiments.

To monitor the autophagosome transport from the cytoplasm to the vacuole, we treated the BY-YA8 cells with CA. An accumulation of YFP fluorescence was detected in the vacuole after 24 h in sucrose-free conditions (Fig. 3A). In contrast, no accumulation was detected in BY-HGA8ΔG cells (Fig. 3B), indicating that the YFP-NtATG8a fusion protein as an autophagic marker was transported selectively from the cytoplasm to the vacuole in BY-2 cells. Taking these results together, the BY-YA8 cell line allows us to monitor the dynamics of autophagic flux in living plant cells.

The fluorescence intensity of the YFP-ATG8 protein as a suitable marker for autophagic flux

Though the in vivo imaging of BY-YA8 cells is useful to analyze the dynamics of autophagy, precise and convenient quantification (or counting) of fluorescent punctate signals in the cytoplasm of intact plant cells using fluorescence microscopy is not easy. To investigate the fluorescence intensity of YFP-NtATG8a as a marker for autophagic flux in plant cells, we analyzed the changes of fluorescence intensity in BY-YA8 cells. In sucrose-rich conditions, the YFP fluorescence showed almost no change (Fig. 4A and C, Control). In contrast, the YFP signal gradually decreased in a time-dependent manner after induction of autophagy by sucrose starvation (Fig. 4A and C, Starvation), and the reduction of YFP fluorescence was significantly inhibited by a PI3K inhibitor, 3-MA (Fig. 4A and C, St + 3-MA). In contrast, in BY-HGA8ΔG cells, the intracellular GFP signal was not affected by sucrose starvation for 24 h (Fig. 4B and D, Starvation). These results are consistent with those with transgenic animal cells expressing GFP-LC3 and suggest that YFP-NtATG8a is selectively degraded by autophagy, which can be quantitatively analyzed as a marker for autophagic flux.

Figure 4. Changes of intracellular fluorescence intensity during sucrose starvation in tobacco BY-YA8 cells. Pseudo-colored images of representative tobacco BY-2 cells expressing the YFP-NtATG8a protein (A) and the HA-GFP-NtATG8aΔG protein (B) at the indicated time points. Three-day-old transgenic BY-2 cells were incubated in complete (Control) or sucrose-free medium (Starvation) for 24 h in the presence or absence of 3-MA (5 mM). These images were captured by CLSM and converted to a rainbow palette in the Zeiss LSM Image Browser software. Scale bar: 50 μm. The data are representative of three experiments. (C, D) Quantification of the fluorescence intensity of the cytoplasm and nucleoplasm in tobacco BY-YA8 (C) and BY-GA8ΔG cells (D). These signals were obtained quantitatively using the Image J software. Data are the means ± SE of three independent experiments.

We next quantitatively monitored the accumulation of YFP fluorescence in the vacuole in the presence of CA that inhibits vacuolar degradation or quenching of YFP. As discussed above, the fluorescence of YFP-NtATG8a in the cytoplasm and nucleoplasm decreased under sucrose starvation [Fig. 5A, CA (-)]. At the same time, in the absence of CA, the YFP fluorescence in the vacuole showed almost no change under sucrose-free conditions [Fig. 5A and B, CA (-)]. In contrast, in the presence of CA, the fluorescence intensity of YFP in the vacuole increased rapidly in a time-dependent manner [Fig. 5A and B, CA (+) +D.W.], and this increase was strongly inhibited by treatment with 3-MA [Fig. 5A and B, CA (+) +3-MA]. These results indicate that the decrease in fluorescence intensity of YFP-NtATG8a in the cytoplasm/nucleoplasm and its increase in the vacuole in the presence of CA are suitable markers for autophagic flux in vivo in BY-2 cells.

Figure 5. Changes of fluorescence intensity in the vacuoles of tobacco BY-YA8 cells in the presence of concanamycin A. (A) Pseudo-colored images of representative tobacco BY-2 cells expressing the YFP-NtATG8a protein at the indicated time points. Three-day-old transgenic BY-2 cells were transferred to sucrose-free medium at time 0 in the presence or absence of CA (1 μM), and with or without 3-MA (5 mM). DMSO was used as a control of 3-MA. The images were captured by CLSM and converted to a rainbow palette in the Zeiss LSM Image Browser software. Scale bar: 50 μm. The data are representative of three experiments. (B) Quantification of the fluorescence intensity in the vacuoles of tobacco BY-YA8 cells. These signals were obtained quantitatively using the Image J software. Data are the means ± SE of three independent experiments.

A simple and easy method for imaging and quantification of autophagic flux using tandem fluorescent-tagged ATG8 protein

In animal cells, the site of autophagosome formation, the cytoplasm, occupies most of the cell volume. In contrast, in plants as well as yeasts, the vacuole where the autophagosomes are transported occupies substantial space of the cells. In yeast cells, the ALP (alkaline phosphatase) assay to quantify the movement of autophagic substrate from the cytoplasm to the vacuole has been established for quantitative determination of autophagic activities. However, similar method has never been reported in plants.

We showed that both reduction in the cytoplasmic fluorescence and increment in the vacuolar fluorescence of YFP-NtATG8a correlated well with the level of autophagic flux in BY-2 cells (Figs. Four and 5). Therefore, we postulated that quantification of both the cytoplasmic and vacuolar fluorescence of YFP-NtATG8a could provide useful and convenient information for quantification of autophagic flux in living plant cells. However, simultaneous measurement of the cytoplasmic and the vacuolar fluorescence basically requires confocal fluorescence microscopy, which is laborious and time-consuming.

The pH value of the vacuolar lumen in plant cultured cells is around 5.5. The fluorescence intensity of GFP/YFP is pH dependent and GFP/YFP fluorescence often quenches when translocated into the vacuole. In contrast, the fluorescence intensity of RFP is much less affected by pH and therefore RFP-fusion proteins are a useful tool to monitor the translocation of proteins into the vacuole.

Based on the difference of the fluorescence properties between YFP and RFP, we established a transgenic cell line in which ATG8 protein localized in the cytoplasm and the vacuole can be distinguished. First, we fused NtATG8a to a tandem fusion of fluorescent tags, acid-insensitive mRFP1,, and pH-sensitive YFP, and checked the pH sensitivity of the fluorescence of these proteins by a spectrofluorometer. As expected, YFP fluorescence was drastically quenched at lower pH values, while mRFP fluorescence was not affected by pH values (). Nextly, we generated transgenic BY-2 cell lines (BY-HRYA8 and HRYA8ΔG) expressing HA-mRFP-YFP-NtATG8a and NtATG8aΔG fusion proteins (Fig. 6A). Under normal growth conditions, the YFP signal was detected in the cytoplasm and nucleoplasm of 3-d-old cultured BY-HRYA8 cells (Fig. 6B and C, Control, YFP) similarly to the BY-YA8 cells. In contrast, the RFP signal was detected in the cytoplasm, nucleoplasm and weakly in the vacuole (Fig. 6B and C, Control, RFP).

Figure 6. The delivery of tandem fluorescent-tagged NtATG8a to the vacuole during autophagic flux in tobacco BY-2 cells. (A) Schematic diagrams of the HA-mRFP-YFP-NtATG8a and -NtATG8aΔG fusion proteins. (B–G) Three-day-old BY-2 cells expressing the HA-mRFP-YFP-NtATG8a (B–D) or HA-mRFP-YFP-NtATG8aΔG protein (E, F) were incubated in sucrose-rich medium containing 100 μM BTH (BTH) or DMSO (Control) for 18 h. Arrows indicate punctate signals of YFP-NtATG8a. (g) A merged image close-up of the cell in (f). Arrowheads indicate autophagosome-like structures. These fluorescence images were obtained by CLSM. Scale bar: 20 μm. V; vacuole, N; nucleus. (B, C, E) The data are representative of three experiments. (D, F) Quantification of fluorescence intensity inside the ROIs. The relative intensity levels inside the nuclei of the control were standardized as 100%. These signals were obtained quantitatively using the Image J software. Data are the means ± SE of three independent experiments. (G) Quantification of the intensity ratio (mRFP/YFP) in whole-cell sections of tobacco HRYA8 and HRYA8ΔG cells by CLSM. Quantification was performed by dividing the average value of the YFP signals with the average value of the RFP signals using the Image J software. Data are the means ± SE of three independent experiments. (H) Quantification of the intensity ratio (mRFP/YFP) in whole-cell sections of tobacco HRYA8 and HRYA8ΔG cells using a fluoro-image analyzer. Three-day-old transgenic BY-2 cells were incubated in sucrose-rich or sucrose-free medium for 18 h and then transferred to a six-well plate. The signals of both RFP and YFP were detected with a fluoro-image analyzer.

A recent study has shown that salicylic acid agonist BTH [benzo-(1,2,3)-thiadiazole-7-carbothioic acid S-methyl ester] induces autophagy in Arabidopsis root cells. When BY-HRYA8 cells were treated with 100 μM BTH, punctate signals were observed in the cytoplasm at 12 h in BY-HRYA8 cell lines (Fig. 6B, BTH), indicating that BTH induces autophagy in BY-2 cells.

When autophagy is induced by BTH treatment, the YFP fluorescence of the nucleoplasm decreased similarly to the BY-YA8 cells (Fig. 6C, D; YFP), In contrast, the cytoplasmic RFP fluorescence decreased but the vacuolar RFP fluorescence significantly increased (Fig. 6B and C, BTH, RFP), which is consistent with the results shown in Figure 5. In BY-HRYA8ΔG cells, all the fluorescence signals of YFP and RFP in the cytoplasm or the vacuole were not affected by BTH (Fig. 6E and F), and no accumulation of RFP in the vacuole was observed. These results strongly demonstrate that the decrease in the nucleoplasmic fluorescence of YFP/RFP and the increase in the vacuolar fluorescence of RFP can be good markers for the autophagic flux in BY-2 cells: autophagosomes fuse with the vacuole and ATG8 protein are transported from the cytoplasm to the vacuole.

These results led us to introduce the ratio of the fluorescence of RFP and YFP of the HRYA8 cells as a marker for autophagic flux. First, we quantitatively measured the fluorescence of each cell by confocal microscopy (Fig. 6G). Upon BTH treatment, the ratio of fluorescence (mRFP/YFP) dramatically increased, corresponding to the induction of autophagic flux. In contrast, (mRFP/YFP) was not affected by BTH in HRYA8ΔG cells, in which ATG8 is not translocated to the vacuole through the autophagic flux.

We further simplified the method to noninvasively monitor the fluorescence of RFP and YFP in the cell suspension using a fluoro-image analyzer (Fig. 6H) or fluorescent microscopy (). In HRYA8 cells, when the autophagic flux is activated by sucrose starvation, RFP fluorescence was merely affected, while YFP fluorescence dramatically decreased in a time-dependent manner (), corresponding to the translocation of HA-mRFP-YFP-NtATG8a from the cytoplasm to the vacuole. Accordingly, the mRFP/YFP increased (Fig. 6H and ). In contrast, in HRYA8ΔG cells, fluorescence of RFP, YFP as well as mRFP/YFP ratio was scarcely affected. These results indicate that the ratio of fluorescence (mRFP/YFP) of the HRYA8 cells can be a reliable marker of autophagic activity by monitoring the translocation of HA-mRFP-YFP-NtATG8a from the cytoplasm to the vacuole through autophagic activity. To our knowledge, this is the first demonstration to monitor the activity of autophagic flux by only measuring the fluorescence of the cell suspension without microscopy in plant cells.

When the autophagy is activated in YFP-NtATG8a-expressing cells, the YFP fluorescence in the cytoplasm gradually decrease for 24 h (Fig. 4), while the number of the punctate structures corresponding to the autophagosomes reached a plateau after 2–3 h and did not change afterwards (Fig. 1B). These results indicate that YFP-NtATG8a is continuously translocated from the cytoplasm to the vacuole at least for 24 h after the induction of autophagy by sucrose starvation. Therefore monitoring the decrease in cellular total fluorescence of YFP-NtATG8a could be a rough marker for autophagic flux. However, YFP-NtATG8a fluorescence can be affected by the cellular content of the protein and does not always corresponds to the activity of autophagic flux. In contrast, the present method using the two different fluorescent proteins with different pH sensitivity tandemly fused to NtATG8a enabled us to monitor the autophagic flux by normalizing the fluorescence intensity with RFP fluorescence as an internal standard.

Concluding Remarks

The system we report here applying the tandem fluorescent-tagged ATG8 does not require inhibitors such as CA (or drugs) for non-invasive monitoring of autophagic flux. This is a novel way to quantitatively monitor the autophagic flow, translocation of ATG8 from the cytoplasm to the vacuole and degradation of ATG8 in the vacuole in living plant cells. Furthermore, we introduced a simple method to quantitatively monitor the autophagic flux only by measuring the fluorescence of cell suspension without microscopy. These are simple methods to monitor all the processes of the dynamics of ATG8. Not only starvation of various nutrients but also various biotic and abiotic stresses may trigger autophagy in plants. However, such physiological information is still limited and little is known on what triggers autophagy how. The present novel method should provide a useful tool to characterize the physiological functions and molecular mechanisms of plant autophagy induced by various environmental stimuli, and screening potential chemicals to affect autophagy.

Materials and Methods

Plant cell materials and chemicals

A tobacco BY-2 (Nicotiana tabacum L. cv Bright Yellow 2) suspension was maintained by weekly dilution (1/100) of cells in modified Linsmaier and Skoog (LS) medium, as previously reported. The cell suspension was agitated on a rotary shaker at 95 rpm and 25°C in the dark. Transgenic BY-2 cell lines stably expressing the YFP-NtATG8a (BY-YA8 cells), HA-GFP-NtATG8aΔG (BY-HGA8ΔG cells), HA-mRFP-YFP-NtATG8a (BY-HRYA8 cells) and HA-mRFP-YFP-NtATG8aΔG (BY-HRYA8ΔG cells) fusion proteins were maintained similarly to the non-transformed BY-2 cell line. 3-MA, wortmannin and CA were obtained from Sigma. BTH was obtained from Wako.

Plasmid construction

The YFP-NtATG8a fusion construct (pH35YG2) was kindly provided by Prof. Matsuoka (Kyushu University). The HA-GFP-NtATG8aΔG protein was constructed using the PCR-based cloning and Gateway recombination system. Briefly, a GFP-fused NtATG8a fragment was amplified by PCR from constructs containing GFP or NtATG8a with the following primers: 5′-CCG GAATTC ATG GTG AGC AAG GGC GAG GAG CTG T-3′ and 5′-tct ttg aaa gcc ttc ccc atG CCG CCG CCG CCC TTG TAC AGC TCG TCC ATG CCG AGA-3′ for GFP containing the overlapping region of NtATG8a, 5′-gca tgg acg agc tgt aca agG GCG GCG GCG GCA TGG GGA AGG CTT TCA AAG AAG AAT-3′ and 5′-CCG CTCGAGGAGCTC TCA GCT ATT TGC ACG ACC AAA GGT T-3′ for NtATG8a containing the overlapping region of GFP (EcoRI and XhoI/SacI sites are underlined, and lowercase indicates the overlapping region). The PCR products were linked by PCR using the following primers: 5′-CCG GAATTC ATG GTG AGC AAG GGC GAG GAG CTG T-3′ and 5′-CCG CTCGAGGAGCTC TCA GCT ATT TGC ACG ACC AAA GGT T-3′. The resulting product (GFP-NtATG8a) was cut with EcoRI and XhoI and subcloned into the EcoRI-XhoI site of the pGADT7 vector (Life Technologies). To generate an N-terminal HA conjugated GFP-NtATG8aΔG fragment, the HA-GFP-NtATG8a fragment was amplified by PCR using the following primers: 5′-CACCGG ATC CAT GGA GTA CCC ATA CGA CGT ACC AG-3′ and 5′-TCAAAAGGTTTTCTCACTGCTGTAG-3′ to create a deletion in the C-terminus of NtATG8a (CACC sequences for use with the Gateway system are underlined and the initiation and termination codons are in bold). The PCR product was subcloned into the pENTR/SD/D-TOPO vector (Life Technologies) and then cloned into a pK7WG2 vector using the LR clonase reaction. The plasmids pBI121 (Life Technologies)-HA-mRFP-YFP-NtATG8a and pBI121-HA-mRFP-YFP-NtATG8aΔG were constructed by PCR-based cloning methods. YFP-fused NtATG8a and mRFP1 fragments were amplified by PCR from constructs containing the YFP-NtATG8a fragment in pH35YG2 and the mRFP1-AtATG8a fragment in pUC19, which was kindly gifted by Dr. Yoshimoto, with the following primers: 5′-CCG GAATTC ATG GTG AGC AAG GGC GAG-3′ and 5′-CCG CTCGAG GAG CTC TCA GCT ATT TGC ACG ACC AAA GGT T-3′ for YFP-NtATG8a, 5′-CAC CCA TAT GAT GGC CTC CTC CGA G-3′ and 5′-CATATG GGC GCC GGT GGA GTG GCG G-3′ for mRFP1 without the stop codon (EcoRI, XhoI and NdeI sites are underlined and the termination codon is in bold). The resulting product (YFP-NtATG8a) was cut with EcoRI and XhoI, and subcloned into the EcoRI-XhoI site of the pGADT7 vector (Life Technologies), and the mRFP1 fragment was subcloned into the pENTR/SD/D-TOPO vector, cut with NdeI and subcloned into the NdeI site of pGADT7-YFP-NtATG8a, sequentially. The HA-mRFP-YFP-NtATG8a or HA-mRFP-YFP-NtATG8aΔG fragments were amplified by PCR using the following primers: 5′-CAC CGGATC CGA GTA CCC ATA CGA CGT ACC AG-3′ and 5′-CCG CTC GAG GAGCTC TCA GCT ATT TGC ACG ACC AAA GGT T-3′ for HA-mRFP-YFP-NtATG8a, 5′-CCG CTC GAG GAGCTC TCA AAA GGT TTT CTC ACT GCT GTA G-3′ for HA-mRFP-YFP-NtATG8aΔG (BamHI and SacI sites are underlined and the initiation and termination codons are in bold). Finally, the resulting products (HA-mRFP-YFP-NtATG8a and HA-mRFP-YFP-NtATG8aΔG) were cut with BamHI and SacI and then cloned into the BamHI-SacI site of the pBI121 vector (Life Technologies). The plasmid pET-21a (Merck KGaA)-ompA-His-mRFP-YFP-NtATG8aΔG was constructed by PCR-based cloning methods. To generate the signal peptide of ompA gene with a C-terminal His6 tag, two oligonucleotids: 5′-TAA AAA GAC AGC TAT CGC GAT TGC AGT GGC ACT GGC TGG TTT CGC TAC CGT AGC GCA GGC CGC ACA TCA TCA CCA TCA CCA TG-3′ and 5′-GATCCA TGG TGA TGG TGA TGA TGT GCG GCC TGC GCT ACG GTA GCG AAA CCA GCC AGT GCC ACT GCA ATC GCG ATA GCT GTC TTT TTC A-3′ (NdeI and BamHI cohesive are underlined and the initiation codon are in bold) were annealed in the buffer (10 mM TRIS-HCl, 10 mM MgCl2, 1 mM Dithiothreitol, 50 mM NaCl) for 10 min at 95°C, followed by gradual cooling to room temperature for 30 min. The product was subcloned into the NdeI and BamHI sites of pET-21a. The mRFP-YFP-NtATG8aΔG fragment was amplified by PCR from pBI121-HA-mRFP-YFP-NtATG8aΔG with the following primers: 5′-CGGGATCCATGGCCTCCTCCGAGGA-3′ and 5′-CCGCTCGAGGAGCTCTCAAAAGGTTTTCTCACTGCTGTAG-3′ (BamHI and XhoI sites are underlined). Finally, the resulting product was cut with BamHI and XhoI and then cloned into the BamHI-XhoI site of the pET-21a-ompA-His vector.

Establishment of tobacco BY-2 cell lines stably expressing the YFP-NtATG8a, HA-GFP-NtATG8aΔG, HA-mRFP-YFP-NtATG8a and HA-mRFP-YFP-NtATG8aΔG constructs

Transformation of tobacco BY-2 cells was performed following An with minor modifications as follows: 4 mL of 3-d-old exponentially growing culture was transferred to 90-mm Petri dishes and incubated at 28°C with 100 µL of fresh overnight-culture of Agrobacterium tumefaciens pGV2260 containing the binary vectors pK7WG2, pH35YG2 or pBI121 (Life Technologies). After a 48 h co-cultivation, the tobacco cells were washed and plated onto LS agar medium with carbenicillin (250 μg mL–1) containing hygromycin (50 μg mL–1) or kanamycin (50 μg mL–1). Every 3–4 weeks, transformants were selected and transferred onto fresh medium for continued selection. Several cell lines were selected by examination of the YFP/GFP/RFP signals by fluorescence microscopy.

Sucrose starvation treatment of tobacco BY-2 cells

Exponentially growing cells (50 mL, 3-d-old) were collected by centrifugation at 100 × g for 5 min. The cell pellets were washed in an equal volume of sucrose-free LS medium, and after three additional washing steps, the cells were re-suspended in 50 mL of sucrose-free LS medium.

Confocal imaging and quantification of the punctate signals as well as intracellular fluorescence intensity

All images were captured with the 20 × and 40 × objectives on a microscope (AXIOobserver.Z1; Zeiss) equipped with a confocal laser scanning head and control systems (LSM5 Exciter; Zeiss). For all experiments, the laser intensity was adjusted to the lowest level that retained a significant signal-to-noise ratio. For counting the numbers of autophagosomes, transgenic tobacco BY-2 cells were mounted on glass slides and the punctate YFP fluorescence signals were counted, and the average was determined from 200 cells per each treatment.

Intracellular fluorescence intensity was quantified using the confocal images. The fluorescence intensity of YFP, GFP and RFP were calculated by the average pixel value of a ROI for 200 cells per each treatment using the Image J software.

Ratiometric quantification of the fluorescence of cell suspension

Three-day-old BY-HRYA8, HRYA8ΔG and non-transformed BY-2 cells were incubated in sucrose-rich or sucrose-free medium for 20 h. The fluorescence of the cell suspension was quantified using either a fluorescence microscope or a fluoro-image analyzer as follows. At various time points, an aliquot of the cell suspensions was transferred to a 96-well plate. Cells were collected in the bottom of well by centrifugation at 100 × g for 5 min, and then both the RFP and YFP fluorescence at each well was captured with a 10 × objective using an inverted fluorescence microscope equipped with a CCD camera (BZ-9000). Alternatively, an aliquot of the concentrated cell suspension was transferred to a six-well plate, and the fluorescence of each well was detected using a 473 nm excitation laser through a 520 nm LP emission filter for YFP and a 532 nm excitation laser and 580 nm emission filter for RFP with a fluoro-image analyzer FLA3000G (GE Healthcare Bio-sciences). In both cases, the ratiometric quantification of the fluorescence was performed by dividing the average value of the YFP signals with the average value of the RFP signals using the Image J software.

Protein expression in E. coli and pH titration

pH titration assays were performed as described by Gjetting et al. with minor modifications as follows: The pET21a-mRFP-YFP-NtATG8aΔG plasmid was transformed into E. coli BL21-AI (Life Technologies). Fresh cultures of E. coli carrying the foreign gene were cultured at 37°C until the A600 of the culture medium reached 0.5. Protein expression was induced with 1 mM IPTG and 0.2% l-arabinose, and cell growth was continued at 37°C for 16 h. The cells were harvested by centrifugation, resuspended in cold-PBS buffer with 1% tween-20, and lysed by the freeze-thaw method in accordance with the supplier’s instructions. For pH titration of mRFP-YFP-NtATG8aΔG, the cell lysate (concentration ~20 mg mL−1) was diluted to 10 volumes with 50 mM phosphate-citrate adjusted to different pH values between pH 7.0 and pH 5.4. Thereafter, the fluorescence signals derived from the cell lysate were detected with a FP6300 fluorescence spectrophotometer (JASCO). The YFP signal was excited at 514 nm and detected at 530 nm, and the mRFP1 signal was excited at 584 nm and detected at 610 nm.