Ecdysone-mediated up-regulation of the effector caspase DRICE is required for hormone-dependent apoptosis in Drosophila cells.

The Drosophila steroid hormone ecdysone mediates cell death during metamorphosis by regulating the transcription of a number of cell death genes. The apical caspase DRONC is known to be transcriptionally regulated by ecdysone during development. Here we demonstrate that ecdysone also regulates the transcription of DRICE, a major effector caspase and a downstream target for DRONC in the fly. Using RNA interference in an ecdysone-responsive Drosophila cell line, we show that drice up-regulation is essential for apoptosis induced by ecdysone. We also show that drice expression is specifically controlled by the ecdysone-regulated transcription factor BR-C. Combined with previous observations, our results indicate that transcriptional regulation of the components of the core apoptotic machinery plays a key role in hormone-regulated programmed cell death during Drosophila development.

The Drosophila steroid hormone ecdysone mediates cell death during metamorphosis by regulating the transcription of a number of cell death genes. The apical caspase DRONC is known to be transcriptionally regulated by ecdysone during development. Here we demonstrate that ecdysone also regulates the transcription of DRICE, a major effector caspase and a downstream target for DRONC in the fly. Using RNA interference in an ecdysone-responsive Drosophila cell line, we show that drice up-regulation is essential for apoptosis induced by ecdysone. We also show that drice expression is specifically controlled by the ecdysone-regulated transcription factor BR-C. Combined with previous observations, our results indicate that transcriptional regulation of the components of the core apoptotic machinery plays a key role in hormone-regulated programmed cell death during Drosophila development.
Programmed cell death is necessary to delete superfluous cells in metazoans and to maintain homeostasis (reviewed in Refs. 1 and 2). The core cell death machinery, consisting of the BH3-only proteins, BCL-2 family, caspase adaptors, and caspases, is highly conserved and is present in all metazoan cells (reviewed in Refs. [3][4][5][6]. As most components of the cell death machinery are present constitutively within a cell, it is widely believed that execution of apoptosis is primarily regulated post-transcriptionally, that is apoptotic signals somehow feed into and activate preexisting caspase machinery. However, recent data suggest that many components of the core apoptosis machinery, including some caspases, are transcriptionally regulated during cell death and that the levels of the prosurvival and proapoptotic factors in the cell may be crucial to activate the apoptotic program (reviewed in Ref. 4). Consistent with this, there is evidence that various signals such as cytotoxic insults, hormones, and growth factors regulate the activation of the death program by controlling the balance between prosurvival and proapoptotic proteins of the core cell death machinery (4). To understand cell death regulation, it is thus essential to understand the transcriptional control of apoptosis execution.
Steroid hormones are known to regulate cell survival and cell death in many tissues. Drosophila melanogaster is an ideal model system to study steroid hormone-regulated apoptosis as a single steroid hormone, 20-hydroxyecdysone, regulates cell death during development (reviewed in Refs. 5 and 7-9). Ecdysone binds to its heterodimeric EcR/Usp receptor and transcriptionally regulates a number of primary response genes. Waves of ecdysone, produced at various times during fly development, regulate molting, cell proliferation, differentiation, and death in a highly controlled manner (5,(7)(8)(9). During the transition of larva into pupa, an ecdysone pulse toward the end of the third larval instar stage signals puparium formation, followed by a second pulse ϳ12 h later, which initiates head eversion. During this process obsolete larval tissues, such as salivary glands and midgut, are deleted and replaced by adult tissues (5,(7)(8)(9). Cell death in the larval midgut begins in response to the late larval pulse of ecdysone while the salivary glands undergo removal around 15 h later in response to the second hormone pulse (5,(7)(8)(9).
Recent data suggest that EcR/Usp and ecdysone-induced transcription factors including ␤FTZ-F1, BR-C, E74, E75, and E93 play a role in ecdysone-mediated cell death in the larval salivary gland and midgut (5, 8, 10 -13). Studies, mostly with salivary glands, indicate that ecdysone controls the up-regulation of a number of proapoptotic genes such as rpr, hid, dark and dronc, and down-regulates the expression of death inhibitors such as diap1 and diap2 (14 -16). Among the seven Drosophila caspases (17), ecdysone is well known to regulate dronc expression (18,19). DRONC is a CED3/caspase-9-like apical caspase that is essential for several programmed cell death pathways in the fly (17,18,20). Our previous data with DRONC suggest that ecdysone-mediated up-regulation of dronc is an important regulator of hormone-dependent cell death in Drosophila cells (19). One of the downstream targets of DRONC is the effector caspase DRICE (17). As DRICE is the major caspase-3-like effector enzyme in Drosophila, we tested whether it is also regulated by ecdysone. We report here that drice up-regulation by ecdysone plays an important role in hormone-dependent cell death.

MATERIALS AND METHODS
Ecdysone Treatment of Salivary Glands-Animals (W1118) were grown on bromphenol blue-supplemented food and staged by the gut clearance technique as previously described (13,21). Late third instar larvae with empty guts were collected, dissected in Schneiders cell medium (Invitrogen), and incubated with or without 1 mM ecdysone (Sigma) for 1 h at 25°C.
Cell Culture-Drosophila l(2)mbn cells, a kind gift from Dr. A. Dorn (Johannes Gutenburg University, Mainz, Germany) (22) were maintained as described previously (19). Cells, at 1 ϫ 10 6 /well, were seeded in six-well plates in triplicate or quadruplicate and allowed to recover for 3 days. Where necessary, ecdysone (10 M) (Sigma) was added for the desired time. For RNA interference (RNAi) 1 experiments, cells were treated with ecdysone until ϳ50% of control cells were apoptotic as the rate of l(2)mbn apoptosis varies between batches.
Apoptosis Detection-To assess apoptosis, cells were stained with 4Ј-6-diamidino-2-phenylindole (DAPI) fluorescent dye. Briefly, cells were fixed and stained with 1 g/ml DAPI, 20% formaldehyde, and mounted 1:1 with 80% glycerol. Changes in nuclear morphology were observed by fluorescent microscopy and apoptosis (%) was calculated as the proportion of cells with condensed chromatin in a total count of at least 300 cells. Data were derived from three or four experiments.
Caspase Assays-Cell lysates were prepared by freeze thawing and clarified by centrifugation at 13,000 rpm for 5 min at 4°C. Equal amounts of lysate were assayed for caspase activity using 100 M VDVAD-amc and DEVD-amc substrates as described previously (23), and the release of amc was measured using a fluorometric plate reader (PerkinElmer Life Sciences) (excitation 385 nm, emission 460 nm).
RT-PCR Analysis-Total RNA was extracted from l(2)mbn cells using TRIzol reagent (Invitrogen) as per manufacturer's protocol. 1-5 g of total RNA was used as a template for cDNA synthesis, in a 20-l reaction with 500 ng of oligo(dT) 18 , using a Superscript II RNase H Ϫ Reverse Transcriptase kit (Invitrogen), according to manufacturer's protocol. Using 1.5 l of 1:3 diluted cDNA template, PCR amplification was performed using appropriate primers in a 50-l reaction employing 27-33 cycles. Drosophila rp49 was used as a control. Aliquots of PCR products (20 l) were electrophoresed on 1.5-2% agarose gel for analysis.

RESULTS
drice Is Up-regulated in Salivary Glands Treated with Ecdysone-It has previously been shown that ecdysone treatment of late second instar larvae salivary glands results in the upregulation of dronc mRNA (18). Because the effector caspases are necessary for salivary gland death to occur (5, 8) we investigated whether the major downstream effector caspase DRICE was also regulated by ecdysone. To determine whether drice mRNA expression is induced by ecdysone, dissected salivary glands from staged late third instar larvae were treated with ecdysone. After a 1-h ecdysone treatment a dramatic increase in drice mRNA levels was observed, demonstrating that ecdysone can induce drice expression in larval salivary glands (Fig.  1A). Caspase assays performed with lysates from ecdysonetreated salivary glands showed elevated caspase activity (Fig.  1B). In particular we observed a 4.5-fold increase in activity on DEVD-amc, which is a substrate for DRICE. Consistent with this observation, DRICE protein levels were elevated in extracts from ecdysone-treated salivary glands (Fig. 1C). These findings suggest the involvement of DRICE in ecdysone-regulated salivary gland death and more importantly demonstrate that drice mRNA is up-regulated by ecdysone.
drice Is Induced by Ecdysone in l(2)mbn Cells-To further investigate the ecdysone-induction of drice transcription and dissect out its transcriptional regulation by various transcrip-tion factors we used the ecdysone-responsive Drosophila cell line l(2)mbn. We first analyzed ecdysone-mediated apoptosis by DAPI staining. As shown in Fig. 2A, apoptotic cells were identified by nuclear fragmentation. In the representative experiment shown in Fig. 2B, low levels of apoptosis were observed at 12-and 24-h post ecdysone treatment. From 36 h the number of cells undergoing apoptosis increased significantly, and by 48 h 70% of cells display condensed nuclei. Northern blot analysis indicated that l(2)mbn cells have low levels of drice transcript. Up-regulation of the drice transcript was evident 6 h after treatment and continued to increase with treatment time, being greatest at 48 h (Fig. 2C). Caspase activity on DEVD-amc and VDVAD-amc substrates was observed from 24 and 12 h, respectively, and subsequently increased with time (Fig. 2D). In addition, increased DRICE processing is observed around 36 h, and the accumulation of processed DRONC is evident at 12 h (Fig. 2E). The results with l(2)mbn cells confirm that drice is transcriptionally regulated by ecdysone treatment. In addition, our observations suggest that following ecdysone treatment of l(2)mbn cells, the increase in drice transcript levels correlates with increased DRICE and its processed form, DEVD-specific caspase activity, and apoptosis.
DRICE Is Essential for Execution of Apoptosis in l(2)mbn Cells following Ecdysone Treatment-We next wanted to determine the importance of DRICE in ecdysone-induced apoptosis of l(2)mbn cells. To do this we used RNAi to ablate gene function. Because DCP-1 is also an important effector caspase in Drosophila it was of interest to test whether there was compensation between the two key effector caspases, DRICE and DCP-1. We also tested whether knockdown of both drice and dcp-1 could prevent apoptosis of l(2)mbn cells following ecdysone treatment. As a comparison we silenced the expression of the initiator caspase DRONC, shown previously to be important in ecdysone-induced l(2)mbn death (19). The knockdown of drice, dcp-1, both drice and dcp-1, or dronc by RNAi was confirmed by RT-PCR analysis using total RNA prepared from ecdysone-treated cells (Fig. 3A). At a time when ϳ50% of control cells were undergoing apoptosis we were unable to detect significant levels of apoptosis in drice dsRNA-treated cells (Fig.  3B). The rate of ecdysone-induced apoptosis in dcp-1 knockdown cells was reduced to ϳ50% of controls (Fig. 3B) and was reduced to ϳ30% of controls in dronc knockdown cells (Fig. 3B). These results show that DRICE, DCP-1, and DRONC are all required for ecdysone to efficiently induce l(2)mbn apoptosis; however, ablation of drice had the most profound effect on ecdysone-mediated programmed cell death. FIG. 1. drice is up-regulated in salivary glands treated with ecdysone. Salivary glands were dissected from staged wild type late third instar larvae and were incubated for 1 h at 25°C with 1 mM ecdysone or left untreated. A, to assess induction of drice, first strand cDNA was synthesized, and RT-PCR was performed using gene-specific oligos. rp49 levels were used as loading control. B, protein was prepared from homogenized tissues, and equal amounts were assayed for caspase activity on the substrates VDVAD-amc or DEVD-amc. Error bars represent S.D. from three experiments. C, protein lysates from untreated and ecdysone-treated salivary glands were immunoblotted with anti-DRICE and anti-cytochrome c antibodies. Cytochrome c (Cyt c) levels were used as loading control.

DRICE Is Necessary for Maximal Ecdysone-induced Caspase
Activity-Because DRICE function is essential for the efficient execution of apoptosis in ecdysone-treated l(2)mbn cells we wanted to establish its contribution to caspase activity. Caspase assays were performed with lysates using DEVD-amc and VDVAD-amc substrates (Fig. 4, A and B, respectively). In the absence of DRICE, lysates displayed basal levels of DEV-Dase activity. Ablation of dcp-1 resulted in ϳ25% activity of controls. When both drice and dcp-1 were silenced DEVDase activity was virtually abolished (Fig. 4A). Previous studies have established that DCP-1 is a substrate for DRICE (26) and that DEVD-amc is a substrate for both DRICE and DCP-1. Because drice knockdown completely abolished DEVDase activity, whereas dcp-1 RNAi lysates still display DEVDase activity, these findings imply a necessity for DRICE in DCP-1 activation. Ablating dronc results in DEVDase activity that is ϳ60% of controls (Fig. 4A). This is consistent with DRONC being an upstream activator of the effector caspases. When caspase assays were performed using a VDVAD-amc substrate, caspase activity following ecdysone treatment of drice or dcp-1 knockdown cells was only ϳ20% of controls (Fig. 4B). Silencing both drice and dcp-1 resulted in basal VDVADase activity (Fig.  4B). In the absence of dronc expression ϳ60% of control VD-VADase activity was observed (Fig. 4B). Because VDVADase activity in part represents DRONC activity, these data suggest that DRICE and DCP-1 may affect DRONC activity. In addition, because activity on both DEVD-amc and VDVAD-amc substrates is still present in dronc knockdown cells, these data present evidence that DRICE activation may occur by means other than DRONC-mediated activation in l(2)mbn cells.
Evidence for a Caspase Activation Loop in l(2)mbn Cells-To further characterize the effect of ablating drice gene function we analyzed protein expression and processing of DRICE and DRONC. In ecdysone-treated control cells we observe up-regulation of DRICE precursor and accumulation of its processed form (Fig. 4C, DRICE proc.). The ablation of either dcp-1 or dronc did not affect DRICE precursor levels; however, we see a reduced processing of DRICE. dronc RNAi had a more profound affect on DRICE processing than dcp-1 RNAi, thus both DRONC and DCP-1 may be capable of activating DRICE. As discussed previously, ecdysone treatment of l(2)mbn cells leads to an accumulation of processed DRONC (Fig. 4C, control  lanes). We observe substantially reduced processed DRONC in the absence of DRICE (Fig. 4C). dcp-1 RNAi did not have any significant affect on processed DRONC levels. Because processed DRONC levels are significantly lower in drice-ablated ecdysone-treated cells, we propose that DRICE is required for amplification of DRONC processing following the initial activation of the upstream caspase.
Ecdysone-induced Factors E74A, E74B, E75A, E75B, E93, or ␤FTZ-F1 Are Not Required for drice Up-regulation-Because we established drice up-regulation and the importance of DRICE in ecdysone-mediated apoptosis in l(2)mbn cells, we could now use this system to investigate the role of known ecdysone-induced transcription factors in drice transcription. We therefore carried out the RNAi ablation of several ecdysoneresponsive transcription factors in l(2)mbn cells. The transcription factors E74A, E74B, E75A, E75B, E93, ␤FTZ-F1, and BR-C have been identified as important regulators of ecdysonemediated salivary gland death (14,15,27,28). In Fig. 5A we show the effect of ablating E74A, E74B, E75A, or E75B on ecdysone-induced l(2)mbn apoptosis. Compared with controls, induction of apoptosis by ecdysone treatment was not significantly affected by silencing any of these transcription factors. When we silence E93, ecdysone-induced apoptosis was reduced by 35% compared with controls (Fig. 5B). ␤FTZ-F1 RNAi did not affect apoptosis induction (Fig. 5B). Thus, of the transcription factors studied, only E93 was identified as being important for regulation of ecdysone-mediated apoptosis in l(2)mbn cells. We then analyzed drice transcription by RT-PCR. The ability of ecdysone to induce drice up-regulation was not affected by silencing E74A, E74B, E75A, E75B, E93, or ␤FTZ-F1 (Fig. 5, C  and D). Because E93 RNAi did not affect drice transcription, the rate of apoptosis in E93 knockdown cells is likely to be reduced as a result of reduced dronc expression, which has been previously reported (15).
BR-C Regulates Ecdysone-induced drice Transcription-The ecdysone-induced transcription factor BR-C plays a key role in salivary gland death and is a regulator of dronc transcription (14,19). There are four isoforms of BR-C protein, Z1-Z4 (29). The Z2 isoform is constitutively present in l(2)mbn cells, and following ecdysone treatment the Z1, Z3, and Z4 isoforms are expressed (19). To test whether BR-C regulates ecdysone-induced drice transcription we carried out RNAi ablation of BR-C in l(2)mbn cells. BR-C RNAi significantly inhibited the rate of ecdysone-mediated apoptosis indicating that BR-C is an important regulator of ecdysone-induced apoptosis in l(2)mbn cells (Fig. 7A). Analysis of transcription by RT-PCR using total RNA shows that BR-C RNAi reduces ecdysone-induced transcription FIG. 4. DRICE is required for maximal caspase activity in l(2)mbn cells following ecdysone treatment. l(2)mbn cells were exposed to 40 nM of dsRNA for drice, dcp-1, drice and dcp-1, dronc, or a negative control for 3 days. Cells were then harvested or treated with 10 M ecdysone. Equal amounts of cell lysate were assayed for caspase activity using DEVD-amc (A) or VDVAD-amc (B) substrate. Activity was represented as a percentage of no RNAi ϩ ecdysone and is expressed in relative fluorescence units (RFU). C, caspase expression and processing (proc.) was assessed by immunoblotting with anti-DRICE or anti-DRONC antibody. Cytochrome c (Cyt c) levels were used as loading control. Samples were derived from the experiments analyzed in Fig. 3. of drice (Fig. 7B). Activity on both DEVD-amc and VDVAD-amc substrates was reduced to ϳ30% of control activity in l(2)mbn cells treated with BR-C dsRNA (Fig. 7, C and D) correlating with the effects seen on apoptosis induction. Immunoblot analysis indicated that processing of both DRICE and DRONC was markedly reduced upon knockdown of BR-C (Fig. 7E). Additionally, ecdysone-induction of full-length DRICE and DRONC was significantly reduced (Fig. 7E). Our results therefore demonstrate that BR-C is required for ecdysone-mediated regulation of drice transcription.

DISCUSSION
The data presented here demonstrate that drice expression is up-regulated in response to ecdysone in both larval salivary glands and Drosophila l(2)mbn cells. The observations are consistent with a large scale gene expression analysis during salivary gland cell death, which shows that drice, among many other genes, is up-regulated in dying salivary glands (30). When drice expression was ablated by RNAi, we observed low levels of caspase activity and a dramatic reduction in the rate of apoptosis following ecdysone treatment. In addition, our data suggest that DRICE is required for DRONC processing in ecdysone-treated l(2)mbn cells. Recent studies have shown that effector caspase activation in some tissues may occur in the absence of DRONC (13,18). Data in this paper complement these findings and supports the observation that some DRICE activation may occur in the absence of DRONC.
Expression of the cell death genes rpr, hid, crq, dark, and dronc in salivary glands and midgut has been shown to be regulated by ecdysone-responsive transcription factors BR-C and E93 (14 -16). The knockdown of BR-C in l(2)mbn cells reduced ecdysone-induced drice transcription and resulted in reduced DRICE expression, thus demonstrating that BR-C regulates drice transcription in l(2)mbn cells. E93 knockdown did not affect drice regulation following ecdysone treatment of l(2)mbn cells. As E93 regulates dronc expression (15,21,31), the decrease in apoptosis levels that we observed is likely to be a result of reduced dronc transcription and protein expression. Although BR-C and E93 are required in both salivary gland and midgut apoptosis, their functions differ between the tissues. For example, dronc expression is regulated in salivary glands by both BR-C and E93 but only by E93 in the midgut (15,33). It would be interesting to test whether like dronc, BR-C regulation of drice is tissuespecific. The competence factor ␤FTZ-F1 is needed for the precise developmental re-expression of transcription factors including BR-C and E93 (32). We investigated its involvement in ecdysone-induced l(2)mbn apoptosis. Our findings demonstrate no clear role for ␤FTZ-F1 in regulating the apoptosis of l(2)mbn cells. We speculate that in l(2)mbn cells, FIG. 6. E93 knockdown reduced caspase activity but did not affect DRICE levels in ecdysone-treated l(2)mbn cells. We further analyzed the RNAi experiments carried out in Fig. 5. A and B, DEVDamc was used as a substrate to analyze DRICE-like caspase activity. Error bars represent S.D.; RFU, relative fluorescence units. C and D, caspase activity was assessed using VDVAD-amc substrate. Error bars represent S.D. E, DRICE expression and processing (proc.) was analyzed by immunoblot analysis using anti-DRICE polyclonal antibody. Cytochrome c (Cyt c) levels were used as loading control. FIG. 7. BR-C regulates drice expression in ecdysone-treated l(2)mbn cells. Drosophila l(2)mbn cells were treated with 40 nM dsRNA for BR-C or a negative control. Cells were harvested on day 3 or treated with 10 M ecdysone. A, apoptosis was calculated as the percentile of cells with condensed nuclei. At least 300 cells were scored for each treatment. Error bars represent S.D. from four experiments. B, RT-PCR was used to analyze drice expression and knockdown of BR-C. rp49 acts as loading control. DEVD-amc (C) and VDVAD-amc (D) substrates were used to assess caspase activity in equal amounts of lysate. E, immunoblot analysis of DRICE and DRONC expression and the accumulation of their processed bands (proc.). Cytochrome c (Cyt c) levels were used as loading control.
␤FTZ-F1 is not required, because l(2)mbn cells do not undergo developmental transitions.
Studies in this paper suggest that ecdysone-regulated DRICE is an essential effector of apoptosis execution in l(2)mbn cells. Based on the data presented here, we propose a model for DRICE-mediated amplification of the caspase cascade in ecdysone-treated cells (Fig. 8). Given that drice mutants are currently unavailable, specific ablation of this caspase in salivary glands and midgut may be required to further investigate the role of DRICE in ecdysone-induced programmed cell death in vivo. In l(2)mbn cells treated with ecdysone dronc and drice are up-regulated, resulting in the activation of these caspases. Initial activation of DRONC probably occurs autocatalytically. Active DRONC then mediates DRICE processing and activation. We proposed that as active DRICE accumulates there is positive feedback (shown by broken lines) and DRICE processes DRONC. The processing of DRONC by DRICE further potentiates activation of the effector caspases, enabling efficient execution of apoptosis. The caspase cascade is further amplified by a positive feedback loop between DRICE and DCP-1.