Drosophila IAP1-mediated ubiquitylation controls activation of the initiator caspase DRONC independent of protein degradation

Abstract

Ubiquitylation targets proteins for proteasome-mediated degradation and plays important roles in many biological processes including apoptosis. However, non-proteolytic functions of ubiquitylation are also known. In Drosophila, the inhibitor of apoptosis protein 1 (DIAP1) is known to ubiquitylate the initiator caspase DRONC in vitro. Because DRONC protein accumulates in diap1 mutant cells that are kept alive by caspase inhibition ("undead" cells), it is thought that DIAP1-mediated ubiquitylation causes proteasomal degradation of DRONC, protecting cells from apoptosis. However, contrary to this model, we show here that DIAP1-mediated ubiquitylation does not trigger proteasomal degradation of full-length DRONC, but serves a non-proteolytic function. Our data suggest that DIAP1-mediated ubiquitylation blocks processing and activation of DRONC. Interestingly, while full-length DRONC is not subject to DIAP1-induced degradation, once it is processed and activated it has reduced protein stability. Finally, we show that DRONC protein accumulates in "undead" cells due to increased transcription of dronc in these cells. These data refine current models of caspase regulation by IAPs.

Figure 1. Apoptosis in Uba1 mutant clones is dependent on DRONC and cannot be inhibited by expression of DIAP1.

Shown are eye-antennal imaginal discs from third instar larvae. Posterior is to the right. In each panel, arrows highlight two representative clones. (A) Uba1 mosaic eye-antennal discs labeled for cleaved CASPASE-3 (α-CAS3*) antibody (red). These discs were incubated at 30°C 12 hours before dissection (see Material and Methods). Presence of GFP marks the location of Uba1 clones (see arrow). (B) TUNEL labeling of Uba1 mosaic eye-antennal imaginal discs expressing an RNAi transgene targeting dronc (UAS-droncIR (inverted repeat)) using the MARCM technique (see Material and Methods). Clones are positively marked by GFP. TUNEL-positive cell death is largely blocked by dronc knockdown (B′ and B″). (C) Strong overexpression of diap1 in Uba1 clones (magenta in C′″) fails to rescue the apoptotic phenotype, as visualized by CAS3* labeling (red in C′). Uba1 clones are marked by GFP due to the MARCM technique. Please note that diap1 is so strongly overexpressed in the clones that we had to adjust the settings in such a way that endogenous DIAP1 in wild-type tissue is below the detection limit (C′″). Genotypes: (A) hs-FLP UAS-GFP; FRT42D Uba1^D6/FRT42D tub-Gal80; tub-GAL4. (B) hs-FLP UAS-GFP; FRT42D Uba1^D6/FRT42D tub-Gal80; tub-GAL4/UAS-droncIR. (C) hs-FLP UAS-GFP/UAS-diap1; FRT42D Uba1^D6/FRT42D tub-Gal80; tub-GAL4.

Figure 2. DIAP1 ubiquitylates DRONC in S2 cells.

DRONC C>A–V5 was coexpressed with His-Ub and the indicated DIAP1 constructs in S2 cells. Ubiquitylated proteins were purified and analyzed by immunoblot for ubiquitylated DRONC with V5 antibodies. Co-expression of DIAP1^wt leads to higher molecular weight modification of DRONC, while the RING-ligase inactive mutants (CΔ6, F437A) cannot ubiquitylate DRONC. * marks non-modified DRONC that is due to unspecific DRONC:matrix association.

Figure 3. Overexpression of diap1 does not trigger degradation of DRONC.

Shown is an eye imaginal disc from a third instar larva. Posterior is to the right. diap1-overexpressing clones are marked by absence of GFP and can be detected using anti-DIAP1 antibodies in magenta (A′″). The boundary between diap1-expressing clones and normal tissue is indicated by a white stippled line in (A′). DRONC levels are unchanged (A′). (A) and (A″) are merged images. Genotype: UAS-diap1/hs-FLP; tub>GFP>GAL4.

Figure 4. Loss of DIAP1 in GMR-rpr eye discs does not alter DRONC protein levels.

(A) Schematic illustration of the GMR-reaper (GMR-rpr) eye imaginal disc from 3^rd instar larvae. The morphogenetic furrow (MF, arrowhead) initiates at the posterior (P) edge of the disc and moves towards the anterior (A) (arrow). The GMR enhancer is active posterior to the MF (bracket) and thus expresses rpr posterior to the MF (red area). (B-B″) Eye disc showing normal protein distribution of DIAP1 (B′) and DRONC (B″). Both DIAP1 and DRONC levels are increased in the MF (arrowhead). (B) is the merged image of DIAP1 and DRONC labeling. (C–C″) Eye discs expressing two copies of GMR-rpr eye disc labeled for DIAP1 (C′) and DRONC (C″). Arrowheads mark the MF. DIAP1 levels are markedly reduced posterior to the MF (C′, arrow). Surprisingly, DRONC protein levels are also reduced (C″, arrow). The brackets indicate the extent of GMR expression. (D–D″) 2×GMR-rpr eye disc labeled for cleaved CASPASE 3 (CAS3*) (D′) and DRONC (D″). DRONC protein levels are reduced in the CAS3*-positive area (arrow). Arrowheads mark the MF. The brackets indicate the extent of GMR expression.

Figure 5. “Undead” diap1 mutant cells trigger transcription of dronc.

Shown are 3^rd instar larval wing (A,B,F) and eye imaginal discs (C,D,E) labeled for DRONC protein levels (blue) and dronc transcriptional activity (red) using the dronc1.33-lacZ reporter (ß-GAL labeling). (A,A′) Co-labeling for DRONC protein (A) and dronc reporter activity (A′) of a wild-type wing disc expressing the dronc1.33-lacZ transgene. (B-B′″) A diap1^22-8s mosaic wing disc expressing p35 under nub-GAL4 control in a dronc1.33-lacZ background. A mutant clone in the wing pouch is highlighted by an arrow in the GFP-only channel (B). DRONC protein (B′) and ß-GAL immunoreactivity as readout of dronc1.33-lacZ activity (B″) are increased in the same cells and overlap (B′″). Please note that the dronc1.33-lacZ reporter produces nuclear ß-GAL, while DRONC protein appears cytoplasmic. (C) GFP expression in the eye imaginal disc indicates the dorsal expression domain (arrow) of the dorsal eye (DE)-GAL4 driver . (D) Increased dronc reporter activity in the dorsal half of the eye imaginal disc (arrow) in undead cells obtained by co-expression of hid and p35 using DE-GAL4. (E) Expression of p35 alone by DE-GAL4 does not induce dronc reporter activity. (F-F″) A diap1^22-8s mosaic wing disc in a dronc1.33-lacZ background which does not express p35. diap1^22-8s mutant clones are marked by the absence of GFP (F). An arrow points to a representative diap1^22-8s clone in the wing pouch. In the same position, neither DRONC protein (F′) nor dronc reporter activity (F″) are increased. Note, that this clone is present in the wing pouch which has the capacity to upregulate DRONC and dronc transcription in the ‘undead’, p35-expressing condition (see panel B″). Genotypes: (A) dronc1.33-lacZ/+. (B) ubx-FLP; nub-GAL4 UAS-p35/dronc1.33-lacZ; diap1^22-8s FRT80/ubi-GFP FRT80. (C) DE-GAL4 UAS-GFP/+. (D) UAS-p35 UAS-hid/dronc1.33-lacZ; DE-GAL4. (E) UAS-p35/dronc1.33-lacZ; DE-GAL4. (F) ubx-FLP; nub-GAL4/dronc1.33-lacZ; diap1^22-8s FRT80/ubi-GFP FRT80.

Figure 6. Ubiquitylation controls processing of DRONC.

(A) Top: schematic outline of the domain structure of DIAP1⁺ (wild-type) and RING-deleted DIAP1^33-1s. Not drawn to scale. Immunoblots of embryonic extracts of stage 6–9 wild-type (wt) and heterozygous diap1^33-1s mutants were probed with anti-DIAP1 (upper panel) and anti-DRONC antibodies (middle panel). The RING-depleted diap1^33-1s allele produces a stable protein (DIAP1^33-1s) that is detectable by its faster electrophoretic mobility (upper panel). In RING-depleted diap1^33-1s embryos a significant portion of processed DRONC is present (middle panel) which likely accounts for the apoptotic phenotype of diap1^33-1s embryos . These extracts were obtained from a cross of heterozygous males and females. Thus, only one quarter of the embryos is homozygous mutant for diap1^33-1s; yet, processed DRONC is easily detectable. The anti-DRONC antibody is specific for the large subunit of DRONC. Lower panel: loading control. (B) Extracts of imaginal discs from wild-type (wt) and mosaic Uba1 imaginal discs were analyzed by immunoblotting using an antibody raised against the small subunit of DRONC. Clones of the temperature sensitive allele Uba1^D6 were induced at the permissive temperature in first larval instar and then shifted to the non-permissive temperature (30°C) during third larval instar 12 hours before dissection (see Material and Methods). This treatment ensures that approximately 50% of the mosaic disc is mutant for Uba1. Although only 50% of the disc tissue is mutant for Uba1, processed DRONC is easily detectable. Lower panel: loading control.