Unleash the dogs of death: metacaspase 5, microtubules, and hypersensitive response

Abstract

Vitis rupestris metacaspase 5, tethered to microtubules, drives grapevine. Hypersensitive response via calcium-dependent auto-processing, linking cytoskeletal dynamics to defence activation by elicitors. Metacaspase 5 is a key player for the hypersensitive response of grapevine against biotrophic pathogens and must be activated rapidly as to prevent colonisation. This activation is likely to occur through changes in protein activity. By expressing a GFP fusion of metacaspase 5 from Vitis rupestris in tobacco BY-2 cells, we can show that this protein is bound to microtubules and that the overexpressors are more responsive to the cell-death-inducing elicitors, cis-3-hexenal and harpin. The disruption of microtubules and actin filaments by these elicitors can be blocked by inhibitors of dynamic turnover and stabilisation. Stabilisation of microtubules by taxol can mitigate cis-3-hexenal induced mortality. Mutations of the catalytic or the putative microtubule-binding sites of metacaspase 5 can suppress auto-processing of this enzyme in biochemical assay. Likewise, the response to cis-3-hexenal (cell death, induction of salicylate-related gene expression) is suppressed in cells, whilst the cytoplasmic remodelling is retained. Calcium and the sites for catalysis or microtubule binding are required for both auto-processing and enzyme activity. We arrive at a model, where metacaspase 5 is inactive when tethered to microtubules, but becomes unleashed for auto-processing upon defence-mediated microtubule breakdown.

Keywords: Cis-3-hexenal; Vitis rupestris; Hypersensitive response; Metacaspase; Programmed cell death; Tobacco BY-2.

Fig. 1

Subcellular localisation of Vitis rupestris metacaspase 5 (VrMC5) in tobacco BY-2 cells as hosts during G₂/M transition and metaphase. Metacaspases are visualised by a fused GFP reporter, microtubules by immunofluorescence using a secondary antibody conjugated to tetramethylrhodamine. PPB preprophase band, + indicates microtubules not decorated by VrMC5, * indicates VrMC5 decorating a microtubule like beads on a string

Fig. 2

Subcellular localisation of Vitis rupestris metacaspase 5 (VrMC5) in tobacco BY-2 cells as hosts during telophase and cytokinesis. Metacaspases are visualised by a fused GFP reporter, microtubules by immunofluorescence using a secondary antibody conjugated to tetramethylrhodamine. Cp cell plate, rMT radial microtubules emanating from the nuclear envelope of the daughter nucleus

Fig. 3

Biochemical properties of Vitis rupestris metacaspase 5 (VrMC5). A Domain structure of VrMC5 and alignment with the class-II metacaspases AtMC4 from A. thaliana. Colours highlight the respective domains. Inset shows the putative microtubule-binding domain and its similarity to the microtubule-binding domain of human MAP1B. B Co-elution of VrMC5-GFP with de-tyrosinated α-tubulin during chromatography based on the affinity for ethyl-N-phenylcarbamate (EPC), a tubulin-binding herbicide, using increasing ionic stringency for elution by increasing the concentration of KCl

Fig. 4

Overexpression of VrMC5-GFP enhances the response of tobacco BY-2 cells to inducers of programmed cell death. A Mortality in response to as compared to the solvent control (triacetin, 625 ppm, 15 min), cis-3-hexenal (12.5 μM, 15 min) and harpin (30 μg/mL, 48 h) in mitotic cells (3d after sub-cultivation). Data represent the mean ± standard error of four independent biological replicates, ** indicates significant difference value P < 0.01 (Student’s t test). B Response of subcellular localisation of VrMC5-GFP to the same treatments assessed at the same time points by spinning disc confocal microscopy

Fig. 5

Microtubule breakdown and mortality in response to cis-3-hexenal can be suppressed by taxol. A Representative images of microtubules visualised by GFP-NtTUA3 are shown for the control treatment (triacetin, 625 ppm, 10 min), cis-3-hexenal (12.5 μM, 10 min), the stabilising compound taxol at 1 μM for 2 h, and the same taxol pre-treatment, but followed by treatment with cis-3-hexenal (12.5 μM, 10 min). Images show geometric projections of confocal z-stacks collected in the cortical region of the cell. B mortality induced by cis-3-hexenal without or with pre-treatment with taxol as compared to the solvent (triacetin, 625 ppm). Data represent means and standard errors from three independent experiments scoring 1500 individual cells per data point

Fig. 6

Engineering auto-processing of VrMC5-GFP. A Part of the alignment of VrMC5 with class-II metacaspases of A. thaliana to show the positions of the engineered sites. B Structure of the two processing mutants and predicted fragment sizes. C Auto-processing of VrMC5-GFP and the two mutants upon extraction from tobacco BY-2 cells expressing those constructs under control of a CaMV 35S promoter. Total proteins were extracted at day 3 after sub-cultivation and subjected to immunoblot analysis using an anti-GFP antibody. D Functional consequences of auto-processing mutants on cell mortality of VrMC5-GFP in response to cis-3-hexenal (12.5 μM, 15 min). Data represent the mean ± standard error (SE) of four independent biological replicates,** indicates significant differences at P < 0.01 (Student’s t test). E Modulation of self-processing in VrMC5-GFP variants in response to cis-3-hexenal (12.5 μM, 30 min) as compared to the untreated control probed by an anti-GFP antibody

Fig. 7

Auto-processing mutants of MC5 loose inducibility of SA-related transcripts by cis-3-hexenal although cytoplasmic remodelling is retained. A Steady-state transcript levels relative to the untreated control for PHENYLAMMONIUM LYASE A and B (PALA and PALB), ISOCHORISMATE SYNTHASE 1 (ICS1), and PATHOGENESIS-RELATED PROTEIN 1a (PR1a) 30 min after addition of cis-3-hexenal (12.5 μM) in tobacco BY-2 cells expressing either the wild-type VrMC5 or the two auto-processing mutants C139A and R226G. Transcript levels were estimated by real-time qPCR and normalised to EF-1α as the internal standard. Data represent the mean ± standard error (SE) of three independent biological replicates in technical triplicate, asterisks indicate significant differences with*P < 0.05, **P < 0.01 and ***P < 0.001 (Student’s t test). B The specific cytoplasmic remodelling is retained in the auto-processing mutants. Cells visualised at 60 min after addition of either cis-3-hexenal (12.5 μM) or 90 μg/ml harpin as specificity control

Fig. 8

Pretreatment with jasmonic acid (JA, 100 μM, 30 min) mitigates mortality and inducibility of SA-related transcripts by cis-3-hexenal. A Mortality scored 30 min after addition of cis-3-hexenal (12.5 μM) in tobacco BY-2 cells expressing either the wild-type VrMC5 or the two auto-processing mutants C139A and R226G. Data represent mean and standard errors from three independent experimental series counting 500 individual cells per data point and replication. B Steady-state transcript levels relative to the untreated control for PHENYLAMMONIUM LYASE A and B (PALA and PALB), ISOCHORISMATE SYNTHASE 1 (ICS1), and PATHOGENESIS-RELATED PROTEIN 1a (PR1a) 30 min after addition of cis-3-hexenal (12.5 μM). Transcript levels were estimated by real-time qPCR and normalised to EF-1α as the internal standard. Data represent the mean from three independent biological replicates in technical triplicate, asterisks indicate significant differences between the induction by cis-3-hexenal with and without JA pre-treatment at significance levels P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***) based on a Student’s t test)

Fig. 9

Autolysis and enzymatic activity of VrMC5 in vitro depend on calcium. A Recombinantly expressed VrMC5 and its mutants C139A and R226G separated by SDS-PAGE (12% acryl amide) upon visualisation with Coomassie Blue. 10 μg of protein loaded per lane. B VrMC5 processing in vitro over calcium concentration (upper image), and suppression of processing by adding the calcium chelator EGTA in the presence of 5 mM CaCl₂ (lower image), both assessed after 20 min of incubation. C Dependence of enzymatic activity of recombinant VrMC5 on calcium in vitro measured by cleavage of the artificial substrate Boc-GRR-AMC releasing a fluorescent product upon cleavage of the tripeptide GRR after the Arg (R) residue that can be measured fluorometrically

Fig. 10

Self-processing and enzymatic activity of VrMC5 depend on Cys139 and Arg226. A Autolysis of VrMC5 WT and its mutants in vitro after incubation with 5 mM CaCl2. Mixtures were separated on 12% SDS-PAGE and bands were visualised by Coomassie Blue staining. B Enzymatic activity of recombinant VrMC5 and its mutants in vitro either in the absence or the presence of 10 mM CaCl₂. Measured by cleavage of the artificial substrate Boc-GRR-AMC releasing a fluorescent product upon cleavage of the tripeptide GRR after the Arg (R) residue that can be measured fluorometrically. Purified proteins were added at 30–40 nM to the assay. C139A, purified VrMC5C139A; R226G, purified VrMC5R226G