Caspase-8 cleaves its substrates from the plasma membrane upon CD95-induced apoptosis

Abstract

Apoptosis occurs through a tightly regulated cascade of caspase activation. In the context of extrinsic apoptosis, caspase-8 is activated by dimerization inside a death receptor complex, cleaved by auto-proteolysis and subsequently released into the cytosol. This fully processed form of caspase-8 is thought to cleave its substrates BID and caspase-3. To test if the release is required for substrate cleavage, we developed a novel approach based on localization probes to quantitatively characterize the spatial-temporal activity of caspases in living single cells. Our study reveals that caspase-8 is significantly more active at the plasma membrane than within the cytosol upon CD95 activation. This differential activity is controlled by the cleavage of caspase-8 prodomain. As a consequence, targeting of caspase-8 substrates to the plasma membrane can significantly accelerate cell death. Subcellular compartmentalization of caspase-8 activity may serve to restrict enzymatic activity before mitochondrial pathway activation and offers new possibilities to interfere with apoptotic sensitivity of the cells.

Figure 1

(a) Schematic representation of localization probes. (b) Localization of probes before cleavage: cytosol (1), plasma membrane (2), nucleus (3), cytoskeleton (4), endoplasmic reticulum (5) and mitochondria (6). Arrows indicate possibilities of cleavage depending on the localization of the enzyme (denoted by red star). A full color version of this figure is available at Cell Death and Differentiation journal online.

Figure 2

Comparison of specificity for caspase-8 substrates. Cleavage of NES probes linked to EYFP over time, with the cleavage sequences GIETDS for caspase-3 (a), LLEVDG for cFLIP-L (b), LQTDG for BID (c) and PVETDS for caspase-8 (d). The last frame of each sequence corresponds to apoptotic body formation (e). Plots representing the quantification of cleavage of the four probes, as described in Supplementary Material. One can observe that all probes are cleaved with relatively similar rates, starting at the same time and roughly reaching saturation at the time of apoptotic body formation (f). Three examples of cleavage comparison of NES-LQTDG-EBFP2 (dashed line) and NES-DEVDR-YFP (solid line) on real-time scale. Each color corresponds to a co-expressed pair of probe within one cell. (g). Same as (f), average on 15 cells, and scaled at 1 for apoptotic body formation (h). Comparison of apoptosis sensitivity with (left) and without (right) mCherry-XIAP. At time 70 min, cells expressing mCherry-XIAP are alive despite full cleavage of NES-LQTDG-YFP (green), while control cells expressing mCherry are dead (i). Comparison of NES-DEVDR-YFP with (blue square) and without (green circle) mCherry-XIAP. The averaging of control cells was stopped at 40 min, when half of the measured cells were apoptotic (j). Comparison of NES-ELQTDG-EBFP2 with (blue square) and without (green circle) mCherry-XIAP as in (i). Note the overlap of the curves until the time, when half of the control cells were apoptotic. Scale bar: 10 μm

Figure 3

Calnexin (a), MitoNEET (b) and cytokeratin-8 (c) linked to EYFP and mGFP by the sequence LQTDG were compared with NES probes fused to mCherry by the same linker after CD95 induction (d). Plots of normalized cleavage, as in Figure 2e for probes shown in a–c. The three plots corresponding to cytoplasmic probes with no access to plasma membrane appeared similar and demonstrate that these probes are much less efficiently cleaved than the NES probe. (e) Comparison of cleavage of NES (dashed lines) and calnexin (solid lines) probe using the LQTDG sequence with (blue squares) and without (green circles) XIAP overexpression. Note that calnexin-LQTDG follows the same kinetics as in control cells until apoptosis of half of the cells. (f) H2B probe cleavage (blue diamond) compared with NES probe cleavage (red squares) with LQTDG cleavage sequence. (g) Same as (f) with the plasma membrane Myr–Palm-snap3 probe (blue diamond). Scale bar: 10 μm

Figure 4

TNFα+cycloheximide induces activity inside the cytosol. (a) Cleavage comparison of NES-DEVDR-YFP and NES-LQTDG-mCherry, as in Figure 2 with TNFα+cycloheximide induction in HeLa cells. (b) Same as (a) with Myr–Palm-snap3-LQTDG-YFP and NES-LQTDG-mCherry. (c) Same as (a) with NES-LQTDG-YFP and calnexin-LQTDG-mCherry. Scale bar: 10 μm

Figure 5

Caspase-6 overexpression does not affect early caspase-8 activity. (a) Comparison by western blotting of caspase-6 expression (expected size: 35 kD) in HeLa-CD95 overexpressing caspase-6 (three left bands) and control cells (three right bands). The three lanes correspond to the three experiments performed with probes, shown on (f) and Supplementary Figures 4b and d. The timeseries (right) shows caspase-6 cleavage (expected size: 15 kD) over time. The vertical line visualizes the time when around 50% of the cells were apoptotic as seen by transmission microscopy before lysing the cells. The contrast of the ladder was enhanced on half of its width for better visibility. (b) Same as (a) in HeLa cells. The three lanes correspond to the three experiments performed with probes, shown on (e) and Supplementary Figures 4a and c. (c) Comparison of time of apoptotic body formation measured by microscopy in HeLa cells overexpressing caspase-6 (blue diamond) and control HeLa cells (red square). Note the overlap between the curves. (d) Same as (c) with HeLa-CD95. Note the 4-min shift of time of death when caspase-6 is expressed. Measurement was performed in the same chamber, excluding difference in ligand treatment. (e) Cleavage comparison of Myr-snap3-LQTDG (red diamond and green triangle) and NES-LQTDG (blue square and orange cross) with (solid line) and without (dashed line) caspase-6 overexpression in HeLa cells as in Figure 2. (f) Same as (e) in HeLa-CD95. Scale bar: 10 μm

Figure 6

Caspase-8 mGFP forms a plasma membrane and cytosolic dotty pattern after induction. (a) HeLa cells knocked-down for caspase-8 and expressing caspase-8-mGFP, monitored after induction. Note the dotty pattern, first exclusively at the periphery (white arrows at 13 min) then also inside the cell (white arrows at 25 min). (b) NB7 stably expressing caspase-8-mGFP monitored until apoptotic body formation after induction. A dotty pattern can be observed exclusively inside the cell (e.g., white arrow at 123 min). (c) Cleavage comparison of NES-LQTDG (upper curves) and calnexin-LQTDG (lower curves) in HeLa-cas8GFP and HeLa-cas8GFP-D210/216/223A-1 as in Figure 2e. (d) Same with HeLa-cas8GFP and HeLa-cas8GFP-D210/216/223A-2. (e) Cleavage comparison of Myr-snap3-LQTDG-mCherry in HeLa-cas8GFP (blue diamond), HeLa-cas8GFP-D210/216/223A-1 (purple square) and HeLa-cas8GFP-D210/216/223A-2 (orange circle) expressing EBFP2-XIAP. Normalization was made to 1, when all probes were cleaved. (f) Same as (e) with NES-LQTDG-mCherry, with normalization with cytosolic intensity at time 0. (g) Same as (e) with calnexin-LQTDG-mCherry, with a normalization with the total intensity at time 0 measured from 3D stacks. Note the reduced amount of calnexin probe cleavage with caspase-8 mutants that cannot be cleaved between their prodomain and catalytic domain. Scale bar: 10 μm. A full color version of this figure is available at Cell Death and Differentiation journal online

Figure 7

BID and caspase-3 targeted to the plasma membrane, and not to ER, accelerate cell death. (a) Cumulative distribution of time of apoptotic body formation for HeLa-CD95 cells overexpressing BID-GFP targeted to the plasma membrane (Myr–Palm, blue diamond) and ER (calnexin, red square), compared with non-transfected cells as control (green triangle). (b) Same as (a) with caspase-3 as substrate for cleavage. In both (a) and (b), plasma membrane targeting (Myr–Palm, diamond) accelerates cell death. (c–e). Comparison of cleavage of NES-DEVDR-YFP reflecting enzymatic activity of caspase-3 and NES-LQTDG-mCherry reflecting enzymatic activity of caspase-8 coexpressed alone (c) or with NES-caspase-3 (d) or Myr–Palm-caspase-3 (e). Note the stepwise shift of enzymatic activity of caspase-3 (blue diamond) in comparison to enzymatic activity of caspase-8 (red square), when caspase-3 is targeted to the cytosol (d) and the plasma membrane (e) compared with the control (c). Plots of probe cleavage are as described in Figure 2e, however, not applying time normalization. Scale bar: 10 μm