A Plant Vacuolar Protease, VPE, Mediates Virus-Induced Hypersensitive Cell Death

See allHide authors and affiliations

Science  06 Aug 2004:
Vol. 305, Issue 5685, pp. 855-858
DOI: 10.1126/science.1099859


Programmed cell death (PCD) in animals depends on caspase protease activity. Plants also exhibit PCD, for example as a response to pathogens, although a plant caspase remains elusive. Here we show that vacuolar processing enzyme (VPE) is a protease essential for a virus-induced hypersensitive response that involves PCD. VPE deficiency prevented virus-induced hypersensitive cell death in tobacco plants. VPE is structurally unrelated to caspases, although VPE has a caspase-1 activity. Thus, plants have evolved a regulated cellular suicide strategy that, unlike PCD of animals, is mediated by VPE and the cellular vacuole.

Some regulatory mechanisms that underlie programmed cell death (PCD) are thought to be conserved in animals and plants, and many studies have provided evidence that PCD in both shares components that include caspase activity (14). However, there is only indirect evidence that caspases are present in plants. The identification of a protease exhibiting caspase activity is essential in elucidating the molecular mechanism that operates PCD in plants. In this study, we have identified VPE (5), a protein that has a proteolytic activity toward a caspase-1 substrate (a caspase-1 activity), as a key component underlying hypersensitive cell death, a typical PCD in plants (6). VPE, like a caspase, is a cysteine protease and was originally found as a processing enzyme responsible for the maturation of seed storage proteins (7, 8). Evidence indicates that VPE is responsible for the maturation or activation not only of various vacuolar proteins in plants (5, 911), but also of lysosomal proteins in the mouse (12). Previously, we reported that VPE was up-regulated during cell death in association with leaf senescence and lateral root formation of Arabidopsis (5, 13). This implied that VPE might be involved in plant cell death.

Hypersensitive response (HR) is caused by interactions between plants and incompatible pathogens, in association with rapid and localized cell death (hypersensitive cell death) at the infected sites of host tissues (14). The HR plays a role in preventing the growth and spread of the pathogens into healthy tissues. For this study, we used Nicotiana plants that carried the N resistance gene to tobacco mosaic virus (TMV) to form synchronous lesions that showed hypersensitive cell death in the TMV-infected leaves after a temperature shift (supporting online text and fig. S1). We isolated four VPE cDNAs [NtVPE-1a (DNA Data Bank of Japan, accession no. AB075947), NtVPE-1b (AB075948), NtVPE-2 (AB075949), and NtVPE-3 (AB075950)] from the TMV-infected leaves. Both mRNA and protein levels of VPE were rapidly and transiently increased at an early stage of HR in the leaves (supporting online text and fig. S1). To clarify the effect of VPE on lesion formation, we infected the tobacco leaves with TMV, and then infiltrated each protease inhibitor into the infected region 1 hour before the temperature shift. The characteristic lesion was observed in the TMV-infected region 24 hours after the temperature shift (Fig. 1A). Lesion formation was strongly inhibited by a VPE inhibitor, Ac-ESEN-CHO, and a caspase-1 inhibitor, Ac-YVAD-CHO (Fig. 1A). Other protease inhibitors used were unable to inhibit lesion formation. These results suggest that VPE activity and caspase-1 activity contribute to TMV-induced cell death.

Fig. 1.

VPE exhibits a caspase-1 activity involved in hypersensitive cell death. (A) TMV-induced hypersensitive cell death in tobacco leaves is inhibited by a caspase-1 inhibitor and a VPE inhibitor. A TMV-infected tobacco leaf was infiltrated with the indicated protease inhibitors 1 hour before the temperature shift. The photograph was taken 24 hours after the temperate shift. (B) Top: A TMV-infected leaf infiltrated with or without biotin-xVAD-fmk. Bottom: A biotinylated-inhibitor blot of the biotin-xVAD-fmk-infiltrated leaves (lane 1) and control leaves (lane 2). An asterisk indicates nonspecific signals. (C) A biotinylated-inhibitor blot of the TMV-infected leaves that were pre-incubated with each protease inhibitor before further incubation with biotin-xVAD-fmk (lanes 1 to 7) and an immunoblot of the leaf extracts with antibodies to VPE (lane 8). (D) A biotinylated-inhibitor blot of the leaf extract after immunodepletion with antibodies to VPE (lane 1) or with preimmune serum (lane 2). An asterisk indicates nonspecific signals. (E) Effects of various inhibitors on VPE activity in TMV-infected leaves 3 hours after the temperature shift. Each experiment was repeated three times, and vertical bars represent stabdard error. DMSO, dimethylsulfoxide; Ac-YVAD-CHO, a caspase-1 inhibitor; Ac-ESEN-CHO, a VPE inhibitor; Ac-DEVD-CHO, a caspase-3 inhibitor; Pepstatin A, an aspartic protease inhibitor; PMSF, a serine protease inhibitor; E-64, a papain-type-protease inhibitor.

To detect the protein that had caspase-1 activity, we developed a biotinylated-inhibitor blot analysis with an irreversible caspase inhibitor, biotin-xVAD-fmk. In this analysis, an enzyme that conjugates with the inhibitor becomes visible on the blot with streptavidin-conjugated horseradish peroxidase. Like Ac-YVAD-CHO, biotin-xVAD-fmk strongly inhibited lesion formation (Fig. 1B, top). Two bands of 40 and 38 kD were detected specifically on the blot of the extract from the biotin-xVAD-fmk–infiltrated leaves (Fig. 1B, bottom). We examined the competitive effects of various inhibitors on in vitro formation of the enzyme-inhibitor complex with the TMV-infected leaf extracts. When Ac-YVAD-CHO was added as a competitor with biotin-xVAD-fmk, the 40- and 38-kD bands completely disappeared on the blot (Fig. 1C, lane 5). Their signal intensities were reduced with Ac-ESEN-CHO (Fig. 1C, lane 7). On the other hand, the 40- and 38-kD proteins were detected on an immunoblot of the TMV-infected leaves with antibodies to VPE (Fig. 1C, lane 8), which corresponded to the intermediate and mature forms of VPE, respectively (15, 16). The immunodepleted extract from the TMV-infected leaf with antibodies to VPE reduced the signal intensities of both the 40- and 38-kD bands on the in vitro biotinylated-inhibitor blot analysis (Fig. 1D), suggesting that the immunodepletion reduced the level of the active enzyme exhibiting caspase-1 activity. These results imply that the proteases that exhibited the caspase-1 activity in the TMV-infected leaves are both forms of VPE. VPE activity was detected in TMV-infected leaves in vitro (Fig. 1E). Biotin-YVAD-fmk completely inhibited the VPE activity at a concentration of 50 μM and reduced it by 64% at a concentration of 5 μM (Fig. 1E, left). The concentration of the inhibitor was comparable to the concentrations necessary to inhibit the authentic human caspase-1. Other inhibitors had no effect on the VPE activity (Fig. 1E, right). These results imply that the abolishment of lesion formation by the VPE and caspase-1 inhibitors was caused by the inhibition of VPE.

For an in planta analysis, we isolated two VPE homologs of Nicotiana benthamiana [NbVPE-1a (AB181187) and NbVPE-1b (AB181188)], which correspond to the most abundant VPEs in N. tabacum (NtVPE-1a and NtVPE-1b) from the TMV-infected leaves. We used the cDNAs for virus-induced gene silencing (VIGS) with a potato virus X vector (pPVX) to silence VPE genes in N. benthamiana, a species well-suited for VIGS (17). The expression of major VPEs was suppressed in four independent VPE-silenced plants that we generated (supporting online text and fig. S2). We measured the VPE activity on Ac-ESEN-MCA in the VPE-silenced and non-silenced leaves that had been treated with or without salicylic acid (Fig. 2A). VPE activity was reduced in all four independent VPE-silenced plants. These results revealed that the VIGS, which was caused by four independent sequences derived from NbVPEs, abolished both VPE expression and activity, indicating that the specificity to VPE was high in the VIGS. Each circle in Fig. 2A represents the VPE and caspase-1 activities in each of the VPE-silenced and non-silenced plants. The level of VPE activity completely paralleled that of the caspase-1 activity in each plant (Fig. 2A). This result and the immunodepletion analysis of VPE (Fig. 1D) show that VPE is the protease with caspase-1 activity. This was in agreement with our previous result that plant VPEs cleave a peptide bond at the C-terminal sides of not only asparagine residues but also aspartic acid residues (15, 1820).

Fig. 2.

VPE deficiency suppresses TMV-induced hypersensitive cell death. (A) Both VPE and caspase-1 activities were measured in the leaves that had been treated with (closed circles) or without (open circles) salicylic acid (SA) for 24 hours, in the VPE-silenced plants (pPVX: VPE, red) and in non-silenced plants including pPVX-inoculated (pPVX, blue) and non-inoculated (no pPVX, green) plants. (B) Non-silenced (pPVX) and VPE-silenced (pPVX: VPE) N. benthamiana plants were infected with TMV on halves of their leaves (indicated by asterisks). The photographs of the plants (left) and the leaves (right) were taken 24 hours after the temperature shift.

To determine how VPE deficiency affects hypersensitive cell death, both the VPE-silenced and non-silenced plants were infected with TMV on different halves of their leaves. The non-silenced plants formed typical visible lesions in the TMV-infected regions of the leaves 24 hours after the temperature shift (Fig. 2B, pPVX, indicated by asterisks). However, the VPE-silenced plants formed no visible lesions in the TMV-infected regions (Fig. 2B, pPVX:VPE, indicated by asterisks). The VPE deficiency suppressed the hypersensitive cell death in response to TMV infection.

It has been shown that disintegration of the vacuolar membranes is the crucial event in plant cell death (21). For an ultrastructural analysis of the VPE-silenced and non-silenced N. benthamiana plants, we examined the TMV-infected leaves 0, 9, and 24 hours after the temperature shift (Fig. 3, A to G). This was because N. benthamiana leaves started to form lesions at 12 hours, which was later than for N. tabacum leaves. The vacuolar membranes were partially disintegrated in the leaves at 9 hours (Fig. 3, C to E), indicating that the disintegration of the vacuolar membranes occurred in the leaves before visible lesions were formed. The disintegration of the vacuolar membranes continued, resulting in complete vacuolar collapse in association with plasmolysis, and finally the cytoplasmic aggregations were left within the cells (Fig. 3B). However, in the VPE-silenced plants, the vacuoles and vacuolar membranes remained intact, even 24 hours after the temperature shift, and no morphological differences except for chloroplasts were found in the leaves before and after the temperature shift (Fig. 3, F and G). Chloroplasts accumulated starch granules in both the silenced and non-silenced leaves at 24 hours (Fig. 3, B and G).

Fig. 3.

VPE deficiency suppresses vacuolar collapse leading to TMV-induced hypersensitive cell death. (A to G) Morphological changes in the TMV-infected regions of the non-silenced leaves (pPVX) at (A) 0, [(C) to (E)] 9, and (B) 24 hours after the temperature shift, and of the VPE-silenced leaves (pPVX:VPE) at (F) 0 and (G) 24 hours, under the electron microscope. (D) A higher magnification view of the boxed area in (C). Scale bars, 5 μm for (A) to (C), (F), and (G), and 1 μm for (D) and (E). cw, cell wall; pm, plasma membrane; vm, vacuolar membrane; v, vacuole; ch, chloroplast; sg, starch granule. Red triangles indicate the disintegrated parts of vacuolar membranes. (H) The TMV-infected leaves of the non-silenced (pPVX) and VPE-silenced (pPVX:VPE) plants were infiltrated with the vital dye BCECF-AM. Protoplasts were prepared from the leaves at 0 and 9 hours after the temperature shift and then stained with trypan blue. The BCECF fluorescent images (BCECF) and differential interference contrast images after staining with trypan blue (DIC/TB) of the protoplasts were inspected. v, vacuole. (I) Pulsed-field gel electrophoresis of total DNA from TMV-infected leaves of the non-silenced plants at 0 (lane 1), 9 (lane 2), 12 (lane 3), and 24 (lane 4) hours after the temperature shift and of VPE-silenced plants at 24 hours (lane 5).

To investigate vacuolar collapse during hypersensitive cell death, we infiltrated 2′, 7′-bis-(2-carboxyethyl)-5(6)carboxyfluorescein acetoxymethylester (BCECF-AM) into TMV-infected leaves and then prepared protoplasts from the leaves 0, 9, and 12 hours after the temperature shift. We subjected the protoplasts to a diagnosis of vacuolar membrane disintegration with BCECF fluorescence and to a viability assay with trypan blue. Protoplasts that were prepared at 0 hours accumulated BCECF only in the vacuoles (Fig. 3H, pPVX, 0 hours). Protoplasts (n > 100) that were prepared at 9 hours were separated into three staining types (Fig. 3H, pPVX): (I) BCECF-positive and trypan blue–negative (∼70% of the protoplasts), (II) BCECF-negative and trypan blue–negative (∼20%), and (III) BCECF-negative and trypan blue–positive (∼10%). The staining indicates that type I protoplasts were alive and had intact vacuoles. Type II protoplasts were also alive even though the vacuolar membranes had disintegrated (as shown by distribution of BCECF throughout the cell and in the periphery regions inside the cells). Type III protoplasts were dead. At 12 hours after the temperature shift, the number of dead protoplasts in the population was greater (22). The finding that many of the protoplasts with degraded vacuolar membranes were alive suggests that cell death is preceded by vacuolar collapse. All protoplasts from the VPE-silenced leaves were alive and accumulated BCECF in the vacuoles at 9 hours (Fig. 3H, pPVX:VPE) and even at 12 hours (22), which indicates the integrity of the vacuoles in the silenced plants. VPE deficiency suppressed the disintegration of the vacuolar membranes in the TMV-infected leaves. These results suggest that VPE is involved in vacuolar collapse, which triggers hypersensitive cell death.

Plant PCDs are reported to be accompanied by DNA fragmentation (23, 24). Similarly, cleavage of nuclear DNA into ∼50-kb fragments was detected in the TMV-infected leaves at 12 hours after the temperature shift, and the fragmentation was completed at 24 hours (Fig. 3I, pPVX). In contrast, such fragmentation was completely suppressed in the VPE-silenced plants (Fig. 3I, pPVX:VPE). Vacuolar nucleases have been shown to be induced during plant PCD (21). VPE-mediated vacuolar collapse might be involved in the nuclear DNA fragmentation that occurs during hypersensitive cell death.

An immunoblot analysis with antibodies to TMV coat protein (Fig. 4A) showed that the virus was produced more abundantly in the VPE-silenced plants, which formed no visible lesions, than in the non-silenced plants. We densitometrically determined the levels of the coat protein and ribulose-1,5-biphosphate carboxylase-oxygenase (RuBisCO) in three independent experiments (Fig. 4A, bottom). The level of RuBisCO as a loading control was fixed with less than 7% of error. The virus level in the VPE-silenced leaves increased for 24 hours after the temperature shift. In contrast, increase in the virus level in non-silenced leaves was suppressed after 8 hours. The virus level in the VPE-silenced leaves at 24 hours was definitely more than that in non-silenced leaves. Therefore, VPE was involved in suppression of virus production during the HR. On the other hand, induction of pathenogenesis-related (PR) proteins, PR-1 and PR-2, continued after the temperature shift in the VPE-silenced leaves as in the non-silenced plants (Fig. 4B). The VPE deficiency did not affect the production of the PR proteins, although it affected TMV-induced cell death. These results suggest that PCD and defense-protein induction are not coupled during the HR and that the HR is composed of two independent processes, PCD and defense-protein induction. VPE regulates PCD but not defense-protein induction. There has been a lot of discussion about whether PCD during the HR is really critical for resistance (25). Our results suggest that PCD contributes to resistance to a virus infection.

Fig. 4.

VPE deficiency affects virus proliferation, although it does not affect the production of the PR proteins. (A) An immunoblot showing change in the protein levels of TMV coat protein (CP) in TMV-infected leaves of the non-silenced (pPVX) and VPE-silenced (pPVX:VPE) plants after the temperature shift (top). RuBisCO is a loading control of the immunoblot. The protein levels were densitometrically determined (bottom). Each experiment with a control was repeated with three independent samples, and vertical bars represent standard error. The relative abundance is given as a percentage of the maximum value. (B) Immunoblots showing changes in the protein levels of NbVPE, PR-1, and PR-2 in TMV-infected leaves of the non-silenced (pPVX) and VPE-silenced (pPVX:VPE) plants after the temperature shift.

Although the Arabidopsis genome does not have a caspase family, it has a metacaspase family, which is distantly related to the caspase family. A metacaspase was reported to be involved in the cell death of yeast (26). The gene expression of metacaspase was found in pathogen-infected tomato leaves (27), and a proteolytic activity toward Ac-VEID-MCA (a metacaspase/caspase-6 substrate) was detected in dying suspensor cells of Norway spruce (28). Metacaspases might function in the cytosol during cell death as animal caspases. This is in contrast to the VPE functions in vacuole-mediated cell death. VPE is structurally unrelated to the caspase family, although it has caspase-1 activity and an ability to bind caspase-1 inhibitor.

In animals, dying cells are engulfed by phagocytes. However, in plants, which do not have phagocytes, cells surrounded by rigid cell walls must degrade their materials by themselves. Vacuolar collapse has been shown to trigger degradation of the cytoplasmic structures and lead to cell death (21), although its molecular mechanism is not known. Our findings suggest that VPE functions as a key player in vacuolar collapse–triggered cell death. VPEs are distributed in mono- and dicotyledonous plants. Arabidopsis VPE genes are up-regulated in dying cells during development and senescence of tissues (5, 13). Thus, VPE might regulate various types of PCD in higher plants. Identification of the VPE-target proteins, which are possibly associated with the vacuolar membranes, would help to unravel the molecular mechanism of VPE-mediated vacuolar collapse underlying plant PCD. Because VPE acts as a processing enzyme to activate various vacuolar proteins, it might also convert the inactive hydrolytic enzymes to the active forms, which are involved in the disintegration of vacuoles, to initiate the proteolytic cascade in plant PCD. Understanding the VPE-regulated mechanism, which operates in the early process of the HR, may also lead to practical applications for strengthening disease resistance in crops.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 and S2

References and Notes

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article