Baculovirus Caspase Inhibitors P49 and P35 Block Virus-Induced Apoptosis Downstream of Effector Caspase DrICE Activation in Drosophila melanogaster Cells

Abstract

Baculoviruses induce widespread apoptosis in invertebrates. To better understand the pathways by which these DNA viruses trigger apoptosis, we have used a combination of RNA silencing and overexpression of viral and host apoptotic regulators to identify cell death components in the model system of Drosophila melanogaster. Here we report that the principal effector caspase DrICE is required for baculovirus-induced apoptosis of Drosophila DL-1 cells as demonstrated by RNA silencing. proDrICE was proteolytically cleaved and activated during infection. Activation was blocked by overexpression of the cellular inhibitor-of-apoptosis proteins DIAP1 and SfIAP but not by the baculovirus caspase inhibitor P49 or P35. Rather, the substrate inhibitors P49 and P35 prevented virus-induced apoptosis by arresting active DrICE through formation of stable inhibitory complexes. Consistent with a two-step activation mechanism, proDrICE was cleaved at the large/small subunit junction TETD²³⁰-G by a DIAP1-inhibitable, P49/P35-resistant protease and then at the prodomain junction DHTD²⁸-A by a P49/P35-sensitive protease. Confirming that P49 targeted DrICE and not the initiator caspase DRONC, depletion of DrICE by RNA silencing suppressed virus-induced cleavage of P49. Collectively, our findings indicate that whereas DIAP1 functions upstream to block DrICE activation, P49 and P35 act downstream by inhibiting active DrICE. Given that P49 has the potential to inhibit both upstream initiator caspases and downstream effector caspases, we conclude that P49 is a broad-spectrum caspase inhibitor that likely provides a selective advantage to baculoviruses in different cellular backgrounds.

Baculoviruses are large, double-stranded DNA viruses that induce widespread apoptosis in insects. In permissive hosts, apoptosis likely represents a host antiviral defense that limits baculovirus multiplication and dissemination (7). As a consequence, baculoviruses have evolved distinct apoptotic suppressors that block this host response (6, 13). Many of the baculovirus apoptotic regulators function in diverse organisms, including Drosophila melanogaster (reviewed in references 6, 16, and 45), suggesting that the mechanisms by which these inhibitors prevent apoptosis are conserved. In Drosophila DL-1 cells, the baculovirus apoptotic suppressors P49 and P35 block apoptosis (56) induced by Autographa californica nucleopolyhedrovirus (AcMNPV), the best studied and prototype baculovirus (13). Although DL-1 cells are nonpermissive for AcMNPV (30), virus entry is efficient and early gene expression is sufficient to trigger apoptosis. Moreover, in the absence of P49 or P35, AcMNPV-induced apoptosis destroys nearly every cell within 24 h. Thus, DL-1 cells provide an exceptional opportunity to investigate the poorly understood mechanisms by which DNA viruses trigger and subsequently regulate apoptosis in an experimentally advantageous model system.

Due to the multiplicity of apoptotic pathways in Drosophila (16, 25), we have defined the pathway triggered by baculovirus infection. In multicellular organisms, apoptosis is a built-in, regulated process for elimination of unwanted cells, including those infected by viruses (1, 7, 14, 38). Characterized by membrane blebbing and cytolysis, cell destruction occurs through the action of the cysteine-dependent, aspartate-specific proteases known as caspases (reviewed in references 18, 39, 41, and 44). In Drosophila, seven different caspases have been described (16, 25). DRONC, DREDD, and Strica are classified as initiator caspases, whereas DrICE, DCP-1, Damm, and Decay are effector caspases. Upon signal-induced activation, the initiator caspases proteolytically process and thereby activate effector caspases from their zymogen proforms by a series of differentially regulated cleavage events (36, 39, 51). Once activated, effector caspases execute cell death. In Drosophila, the initiator caspase DRONC proteolytically cleaves and activates the caspase DrICE, an effector caspase that plays a critical role in development- and stress-induced apoptosis (16, 25). Both DrICE and DRONC are subject to negative regulation by the cellular inhibitor-of-apoptosis (IAP) protein DIAP1 (5, 29, 46, 52). Due to the prominence of DrICE in Drosophila apoptosis, we investigated its role during apoptosis induced by AcMNPV infection of cultured DL-1 cells.

In Drosophila, baculovirus P35 inhibits DrICE but not DRONC (15, 17, 29), thereby providing a sensitive probe for DrICE activity. P35 is a potent substrate inhibitor of effector caspases in which caspase-mediated cleavage at Asp87 within the P35 reactive-site loop (RSL) leads to formation of a stable, inhibited complex with the target caspase (2, 4, 9, 50, 54). P35 functions after proteolytic processing of effector caspases and directly inhibits their activity (26, 27, 43). Baculovirus P49 is a P35-related substrate inhibitor originally identified in the baculovirus Spodoptera littoralis nucleopolyhedrovirus (8). Like P35, cleavage of P49 within its predicted RSL is required for caspase inhibition (22, 35, 56). Unlike P35, P49 has the capacity to inhibit both effector and initiator caspases. When expressed together in permissive lepidopteran cells, P49 functions upstream of P35 by inhibiting an initiator caspase that is responsible for effector caspase activation (56). In lepidopteran cells, the baculovirus IAP protein OpIAP also functions upstream to block activation of effector caspases and thereby prevents apoptosis (26, 27, 43). Identified in the baculovirus Orgyia pseudotsugata nucleopolyhedrovirus (OpMNPV) (3), OpIAP was the first discovered member of the IAP family of viral and cellular apoptosis regulators (reviewed in references 40 and 48). Interestingly, OpIAP does not inhibit Drosophila caspases, nor does it block apoptosis in Drosophila (49, 56). Thus, OpIAP differs from DIAP1, which regulates initiator and effector caspases by distinct mechanisms in Drosophila (16, 25).

We have used viral and cellular regulators of apoptosis as tools to define the caspase-dependent pathways by which baculoviruses trigger apoptosis in Drosophila. We report here for the first time that DrICE is an important contributor to virus-induced apoptosis, a finding that broadens the role of this caspase in Drosophila. proDrICE was activated in AcMNPV-infected DL-1 cells by discrete proteolytic events that were differentially regulated by IAP proteins and the caspase inhibitors P49 and P35. Unexpectedly, we found that P49 blocked virus-induced apoptosis downstream from DrICE activation at a step also inhibited by P35. Thus, although the initiator caspase DRONC was active, P49 failed to affect DRONC and instead targeted active DrICE. Our study provides further evidence that baculovirus apoptotic suppressors involved in countering host defense mechanisms are highly diverse and inhibit a broad spectrum of caspases.

MATERIALS AND METHODS

Cells lines and viruses.

Drosophila melanogaster DL-1 and S2 cell lines (42) were propagated at 27°C in Schneider's growth medium (Invitrogen) supplemented with 15% heat-inactivated fetal bovine serum (FBS) (HyClone). For infection, DL-1 monolayers (5 × 10⁶ cells/plate) were overlaid with TC100 growth medium (Invitrogen) plus 10% FBS containing the indicated PFU of virus per cell and rocked gently at room temperature. After 2 h, the inoculum was removed and replaced with supplemented Schneider's growth medium. Spodoptera frugiperda IPLB-SF21 cells (47) were propagated at 27°C in TC100 plus 10% heat-inactivated FBS. The AcMNPV recombinants vP49 (pIE1^prmp49/Δ35K/lacZ; p49⁺, p35⁻, iap⁻) and vP35 (pIE1^prmp35/Δ35K/lacZ; p35⁺, iap⁻) have been described (56). The recombinant viruses vDIAP1 (pIE1^prmdiap1^HA/Δ35K/lacZ/opiap^HA; diap1^HA, opiap^HA,p35⁻), vOpIAP (pIE1^prmopiap^HA/Δ35K/lacZ; Op-iap^HA, p35⁻), and vSfIAP (pIE1^prmsfiap^HA/Δ35K/lacZ; Sf-iap^HA, p35⁻) were created by allelic replacement in which the polyhedrin gene of the p35⁻ parent vΔ35K (19) was replaced with genes encoding DIAP1^HA, OpIAP1^HA, or SfIAP^HA fused to the AcMNPV ie-1 promoter (prm) and linked to lacZ. The hemagglutinin (HA) epitope was inserted at the N terminus of each iap gene. After plaque purification, all recombinant viruses were verified by PCR and sequence analysis.

Plasmids.

The gene encoding DrICE (a gift from S. Kumar) was inserted into the ie-1 promoter-based expression vector pIE1^prm/hr5/His6/PA (26) by PCR amplification to generate plasmid pIE1^prm/hr5/drice^T7-His6/PA, which encodes proDrICE with an N-terminal T7 tag (MASMTGGQQMGRDLMDATNN [single underline, T7 tag; italics, DrICE]) and a C-terminal His₆ tag. This plasmid served as the parent to generate plasmids encoding D28A-, C211A-, and D230A-mutated forms of proDrICE by standard PCR mutagenesis methods. pIE1^prm/hr5/PA-based expression vectors for wild-type AcMNPV P35 and D87A-mutated P35, wild-type SlNPV P49 and D94A-mutated P49, and OpMNPV OpIAP^HA have been described (2, 26, 28, 55, 56). pIE1^prmM4-SfIAP^HA/PA was created by PCR amplification of the SfIAP gene (a gift from J. Reed), insertion of the influenza virus HA epitope YPYDVPDYA at the N terminus, and subsequent placement into pIE1^prm/hr5/PA at the HincII and MfeI sites. pIE1^prmDIAP1^HA was created by placing an EcoRI-digested fragment of pGMR/DIAP1 (a gift from Bruce Hay) into plasmid pBS/KS⁺ (Stratagene), followed by insertion of the PstI-HindIII fragment into pIE1^prm/hr5/PA; an HA epitope tag was inserted at the N terminus of DIAP1. For expression in Escherichia coli, the wild-type and C211A-mutated DrICE genes were inserted into pET22b+ by PCR amplification. All plasmids and mutations thereof were verified by nucleotide sequencing.

Cell survival assays.

DL-1 monolayers (3.5 × 10⁶ cells/plate) were infected with the indicated baculovirus or irradiated with UV-B (BLAK-Lamp; Upland, CA) for 2 min at room temperature. At the indicated times, photographs were taken of three evenly distributed fields of view from each of three replicate plates by using computer-aided phase-contrast microscopy (Axiovert 135TV; Zeiss) as described previously (27). Intact cells were counted, and the percent survival is reported as the average ratio ± standard deviation of intact, nonapoptotic cells normalized to untreated cells, for which the value was set at 100% for each treatment.

Caspase assays.

DL-1 cells were harvested from triplicate plates and lysed on ice by using caspase activity buffer (100 mM HEPES [pH 7.0], 500 mM EDTA, 0.1% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}, and 100 mM dithiothreitol [DTT]). The extracts were clarified by centrifugation, and caspase activity was measured by using the substrate Ac-DEVD-7-amino-4-methylcoumarin (DEVD-AMC) in assays that were described previously (26). Values from triplicate assays are reported as the average fluorescence (RFU) ± standard deviation.

RNA and DNA transfections.

To generate double-stranded RNA (dsRNA), sense and antisense RNA was synthesized with in vitro transcription reactions (Ampliscribe T3 and T7 kits; Epicenter) using linearized pBS/KS⁺-based plasmids containing the complete genes for DrICE, DRONC, enhanced green fluorescent protein (EGFP), and nucleotides 2434 to 2894 of Drosophila Dark (ARK) (23, 37, 53). The complementary single-stranded RNAs were annealed by heating to 65°C, followed by cooling at 1°C per min to 20°C. DL-1 cells (2 × 10⁶ cells/ml) were transfected with dsRNA (20 μg) mixed with cationic liposomes (20 μl) consisting of DOTAP-DOPE {N-[1-(2,3 dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl sulfate-l-α-phosphatidylethanolamine, dioleoyl (C_18:1, [cis]-9)} in serum-free Schneider's growth medium. After a 4-h incubation at room temperature, the cells were suspended at 2 × 10⁶ cells per ml in serum-supplemented Schneider's medium. After 3 to 4 days of continuous shaking at 27°C, the cells were counted, plated, and treated as indicated.

For DNA transfections, DL-1 cell monolayers (3.5 × 10⁶ cells/plate) were mixed with CsCl-purified plasmid (2 μg) using DOTAP-DOPE (20 μl) as described previously (26, 34). SF21 monolayers (10⁶ cells/plate) were transfected with plasmid DNA (6 μg) in DOTAP-DOPE (10 μl). Cells were collected at 17 to 20 h after transfection and lysed in 1% sodium dodecyl sulfate (SDS)-1% β-mercaptoethanol (βME).

Antisera and immunoblots.

Polyclonal DrICE antiserum (α-DrICE) was generated by immunization of New Zealand White rabbits (University of Wisconsin Medical School Polyclonal Antibody Service) with Ni⁺²-purified E. coli-produced C211A-mutated DrICE-His6 as antigen. For immunoblots, intact cells and apoptotic bodies were collected by centrifugation, lysed with 1% SDS-1% βME, and subjected to polyacrylamide gel electrophoresis and immunoblot analysis using the following antisera: α-P49 (1:1,000 dilution) (56), monoclonal α-P35 (1:5,000 dilution) (gift from Yuri Lazebnik), α-HA (1:1,000 dilution) (Covance), α-actin (1:1,000 dilution) (BD Transduction Laboratories), α-Sf-caspase-1 (1:1,000 dilution) (26), or α-DrICE (1:2,000 dilution). The membranes were subsequently treated with alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G or goat anti-mouse immunoglobulin G (Jackson ImmunoResearch Laboratories). The CDP-Star Chemiluminescent Detection System (Roche) was used for signal detection. Because polyclonal α-DrICE and α-P49 (56) were generated against His₆-tagged antigens, both antisera detect unrelated His₆-tagged proteins. Images were scanned at 300 dpi with an Epson TWAIN Pro scanner and prepared with Adobe Photoshop CS2 and Adobe Illustrator CS2.

Protein production.

Recombinant P49-His6, D94A-mutated P49-His6, P35-His6, D87A-mutated P35-His6, DrICE-His6, and C211A-mutated DrICE-His6 were isolated from E. coli strain BL-21(DE3) by Ni⁺² affinity chromatography (QIAGEN) as described previously (55). In brief, cells were induced with 0.1 to 1 mM isopropyl-β-d-thiogalactopyranoside for 3 h and lysed by sonication. Ni²⁺-nitrilotriacetic acid (NTA) beads (QIAGEN) were mixed with the soluble supernatant for 3 h at 4°C, washed with 20 mM Tris (pH 7.9)-500 mM NaCl-1 mM DTT, and incubated with 400 mM imidazole in 20 mM Tris (pH 7.9)-500 mM NaCl-1 mM DTT to elute bound proteins.

P49-His6 and P35-His6 pulldown assays.

DL-1 cells were suspended in 20 mM HEPES (pH 7.5), 10 mM KCl, 1.5 mM MgCl₂, 1 mM DTT, and 1× complete protease inhibitor (Roche Diagnostics) at 10⁸ cells per ml. After Dounce homogenization and centrifugation (10 min, 1,000 × g), the clarified supernatant was incubated for 7 h at 37°C to activate endogenous caspases. For pulldown assays, the activated cell extract was mixed with excess His₆-tagged recombinant proteins bound to Ni⁺²-NTA agarose beads (described above) for 15 min at 27°C. After the beads were washed, bound proteins were eluted with 400 mM imidazole in 20 mM Tris (pH 7.9)-500 mM NaCl-1 mM DTT and subjected to immunoblot analysis with α-DrICE.

RESULTS

DrICE is required for baculovirus-induced apoptosis.

Our previous studies indicated that baculovirus-induced apoptosis of Drosophila DL-1 cells is caspase mediated. Thus, to define the pathway by which baculoviruses trigger apoptosis, we have identified the caspases involved. DrICE is a principal effector caspase in Drosophila (11, 12, 16, 25). To determine the role of DrICE in baculovirus-induced apoptosis, we depleted endogenous DrICE by RNA silencing and subsequently assessed cellular sensitivity to apoptosis. Upon transfection with dsRNA specific to the heterologous EGFP gene as a control, DL-1 cells remained highly sensitive to apoptosis induced by UV irradiation or infection with an apoptosis-inducing baculovirus. Cell membrane blebbing and cytolysis were extensive (Fig. 1). Fewer than 10% of these cells survived for more than 24 h (Fig. 2A), a rate comparable to that of nontransfected cells. In contrast, DL-1 cells transfected with drice-specific dsRNA remained largely intact after UV irradiation or virus infection (Fig. 1). The 24-h survival of these cells after irradiation or infection was >50% (Fig. 2A).

FIG. 1.

Morphology of DrICE-silenced Drosophila cells after apoptotic signaling. DL-1 cells were transfected in suspension with dsRNA specific to genes encoding EGFP or DrICE. After 3 days, the cells were plated and either not treated (mock), UV irradiated (+ UV), or inoculated (+ virus) with the apoptosis-inducing baculovirus vOpIAP (multiplicity of infection, ≥20 PFU/cell). Cells were photographed (magnification, ×338) 12 h after irradiation or 20 h after infection.

FIG. 2.

Suppression of apoptosis upon DrICE silencing. (A) Cell survival. DL-1 cells transfected 3 days previously with EGFP-specific dsRNA, DrICE-specific dsRNA, or no dsRNA (none) were scored for viability by counting intact, nonapoptotic cells 12 h after UV irradiation (+UV) or 20 h after infection (+v). Values reported represent the average ratio ± standard deviation of the survival of irradiated or infected cells compared to that of untreated (−) cells, which was normalized to 100% for each condition. (B) Intracellular caspase activity. Extracts of DL-1 cells treated as described for panel A were assayed for caspase activity by using DEVD-AMC as a substrate. Values shown are reported as the average ± standard deviation of relative light units (RLU) measured from triplicate transfections. (C) Immunoblots. The DL-1 extracts of panel B were treated with SDS-βME and subjected to immunoblot analysis by using α-DrICE. For size markers, purified E. coli-produced wild-type (wt) DrICE large subunit and C211A-mutated proDrICE (pro) were included. Size markers (in kilodaltons) are indicated on the left. The nonspecific proteins denoted by arrows (left) indicated that comparable levels of protein were loaded in each lane. (D) Time course of DrICE processing. DL-1 or Schneider's S2 cells were lysed at the indicated time (in hours) after UV irradiation (hp UV) and analyzed by immunoblotting with α-DrICE or α-actin to verify comparable loading (10⁶ cell equivalents) of cell lysates. Untreated cells (−) were included.

To determine the effect of drice-specific dsRNA on caspase levels, we monitored caspase activity in UV-irradiated and virus-infected cells. Effector caspase activity was quantified in DL-1 extracts by cleavage of the substrate DEVD-AMC. Caspase activity was highest in untreated cells and EGFP-specific dsRNA-transfected cells upon UV irradiation (Fig. 2B). Baculovirus infection also induced high levels of caspase activity. In contrast, drice-specific dsRNA-transfected cells exhibited only basal levels of caspase activity after UV irradiation and infection (Fig. 2B). Thus, the decreased sensitivity to UV- and virus-induced apoptosis that was conferred by drice-specific dsRNA correlated with a reduction in caspase activity upon apoptotic signaling.

Transfection of DL-1 cells with drice-specific dsRNA also depleted the 38-kDa proform of DrICE to below the limits of immunoblot detection (Fig. 2C). In contrast, proDrICE was readily detected in cells transfected with egfp-specific dsRNA and in nontransfected cells. Indicative of proteolytic processing, appearance of the large subunit of DrICE coincided with loss of proDrICE in UV-irradiated and virus-infected cells (Fig. 2C). The large subunit of DrICE was not detected in DrICE-silenced cells. Transfection with drice-specific dsRNA also had no effect on endogenous levels of the closely related Drosophila effector caspase DCP-1 (data not shown), indicating that RNA silencing was selective for DrICE. The finding that drice-specific dsRNA eliminated endogenous DrICE, reduced caspase activity, and increased cell survival after apoptotic signaling indicated that DrICE is required for baculovirus- and UV-induced apoptosis.

Originally isolated in tandem with DL-1 cells (42), S2 cells are frequently used for studies of Drosophila apoptosis (21, 24, 31-33, 49). Comparison of S2 and DL-1 cells indicated that the kinetics of DrICE proteolytic processing upon UV irradiation (Fig. 2D) and virus infection (data not shown) are similar. In both cell lines, the appearance of large-subunit DrICE correlated with apoptosis that began ∼8 h after irradiation. Thus, the apoptotic responses of DL-1 and S2 appear to be similar.

DIAP1 and SfIAP, but not OpIAP, block DrICE activation during infection.

Due to the involvement of DrICE in virus-induced apoptosis, we predicted that apoptotic regulators would affect DrICE activation, activity, or both. To identify the targets of apoptotic regulation and define the pathway for DrICE activation, we constructed recombinant baculoviruses that overexpressed distinct apoptotic suppressors (Fig. 3A). By inoculating DL-1 cells with these viruses, each apoptotic regulator was synthesized in every cell that received an apoptotic signal, thereby allowing us to monitor the effect of each regulator on endogenous DrICE. Upon infection with recombinant vDIAP1, which expresses Drosophila DIAP1, little if any proteolytic processing of proDrICE was detected (Fig. 3B, lanes 2 to 5). DrICE remained in its proform and the pro-large and large subunits of DrICE were not detected even at late times after infection. When SfIAP, a related cellular IAP protein from the lepidopteran Spodoptera frugiperda (20), was expressed by recombinant vSfIAP, DrICE processing was detected (Fig. 3B, lanes 6 to 9) but at relatively low levels compared to that induced by vOpIAP, which expresses baculovirus OpIAP as its only apoptotic suppressor. The DrICE large subunit accumulated to relatively high levels in vOpIAP-infected cells (Fig. 3B, lanes 10 to 13). Thus, the level of DrICE processing paralleled the level of apoptosis induced by each virus, since vOpIAP caused widespread apoptosis (Fig. 1) whereas vDIAP1 and vSfIAP did not (data not shown). DIAP1, SfIAP, and OpIAP were each synthesized early in infection at or before DrICE processing (Fig. 3C). These findings indicated that in baculovirus-infected Drosophila cells, the cellular IAP proteins function at or upstream of effector caspase activation, but OpIAP does not.

FIG. 3.

Effect of caspase inhibitors on virus-induced DrICE activation. (A) Schematic of AcMNPV recombinants. Expression of the genes encoding the apoptotic regulators DIAP1, SfIAP, OpIAP, P49, and P35 was directed by the immediate-early ie-1 promoter inserted adjacent to a lacZ reporter at the polyhedrin locus of an AcMNPV p35 deletion mutant. An epitope tag (HA) was inserted at the N terminus of DIAP1, SfIAP, and OpIAP. (B to E) Immunoblot analyses. DL-1 cells were mock-infected (mi) or inoculated with the indicated recombinant viruses (multiplicity of infection, ≥20). Lysates prepared at the indicated time (in hours) after infection (hpi) were electrophoresed (10⁶ cells per lane) and immunoblotted with α-DrICE (B), α-HA (C), α-P49 (D), or α-P35 (E) antiserum. Protein size markers (in kilodaltons) are indicated on the left. Asterisks indicate the caspase-generated cleavage products of P49 and P35.

P49 fails to prevent proteolytic processing of DrICE.

In lepidopteran cells, baculovirus P49 functions upstream of P35 and inhibits proteolytic processing of the proform of the effector caspase Sf-caspase-1 (56). In DL-1 cells, however, proDrICE was proteolytically processed to its large subunit upon infection with P49-encoding vP49 (Fig. 3B, lanes 14 to 17); DrICE processing was comparable to that induced by vOpIAP (lanes 10 to 13). However, unlike OpIAP, P49 prevents vP49-induced apoptosis, as shown previously (56). Consistent with this apoptotic suppression, P49 was proteolytically cleaved to produce the fragments indicative of caspase inhibition (Fig. 3D). Upon infection of DL-1 cells with P35-encoding vP35, proDrICE was also processed to its large subunit (Fig. 3B, lanes 18 to 21). P35 was synthesized abundantly and was also cleaved to produce fragments indicative of caspase inhibition (Fig. 3E). Consequently, vP35-infected DL-1 cells failed to undergo apoptosis, as shown previously (56). We concluded that although both P49 and P35 block baculovirus-induced apoptosis, they do not inhibit the caspase(s) responsible for DrICE processing and activation.

P49 competes with P35 in Drosophila cells.

To determine whether P49 and P35 function at different steps, we tested whether either inhibitor was dominant. Previous competition assays using recombinant viruses in lepidopteran cells indicated that P49 is dominant and functions before P35, thereby blocking caspase-mediated cleavage of P35 (56). Our approach here involved coinfection with the recombinant viruses vP49 and vP35, in which both caspase inhibitors were produced in excess from the same ie-1 promoter in cells containing activated endogenous caspases due to infection. Thus, the steady-state levels of processed DrICE or that of the caspase cleavage products of P49 or P35 provided a measure of the effect of one inhibitor relative to the other in a single cell, a measurement not possible with plasmid transfections. Upon virus inoculation of DL-1 cells, proDrICE was processed to its pro-large and large subunits, as expected (Fig. 4A). When P49 alone was present, the pro-large and large subunits were comparable (Fig. 4A, lane 2). However, when P35 alone was present, the pro-large DrICE subunit predominated (Fig. 4A, lane 8). As vP35 was increased relative to vP49, the pro-large intermediate increased (lanes 3 to 7). In contrast, as vP49 increased relative to vP35, the level of the pro-large subunit decreased (lanes 9 to 13). Thus, the caspase inhibitor that was present in the greatest abundance dominated the other inhibitor's effect on DrICE processing, suggesting that P35 and P49 compete at the same step. The same conclusion is suggested by the effect of each inhibitor on caspase cleavage of the other. When the ratio of vP35 to vP49 increased, the level of P49 cleavage decreased (Fig. 4B, lanes 2 to 6) and P35 cleavage increased (Fig. 4C, lanes 3 to 7). Conversely, as the ratio of vP35 to vP49 decreased, the level of P49 cleavage increased (Fig. 4B, lanes 8 to 12) and P35 cleavage decreased (Fig. 4C, lanes 9 to 13). Since both inhibitors were present in excess, the simplest explanation for these findings is that P49 and P35 compete for the same effector caspase target(s) when simultaneously present. Confirming this conclusion, we next demonstrated that both P35 and P49 inhibit endogenous DrICE.

FIG. 4.

P49 competes with P35 in DL-1 cells. Monolayers were either mock-infected (mi), inoculated (multiplicity of infection, ≥10) with the recombinant virus vP49 alone (−) or vP35 alone (−), or coinoculated with an increasing ratio of both viruses (total multiplicity of infection, ≥10). Cell lysates (10⁶ cell equivalents) prepared 20 h later were subjected to immunoblot analysis using α-DrICE (A), α-P49 (B), or α-P35 (C).

P49 and P35 form inhibitory complexes with endogenous DrICE.

To determine whether DrICE is a target of P49 and P35, we tested the ability of both caspase inhibitors to form stable complexes with endogenous DrICE. To this end, we prepared cytosolic extracts of DL-1 cells that contained active DrICE. In this cell-free system, proDrICE was spontaneously processed to its large/small subunit complex (Fig. 5, lanes 2 and 3) with a concomitant increase in caspase (DEVD-ase) activity (data not shown). When E. coli-produced recombinant P35-His6 or P49-His6 was mixed in excess with the activated cell extracts and subjected to Ni²⁺ affinity chromatography, each substrate inhibitor selectively pulled down active DrICE, as indicated by the presence of the DrICE large subunit (Fig. 5, lanes 4 and 6). Thus, both P35 and P49 formed a stable complex with active DrICE. Confirming the requirement for P35 and P49 cleavage, cleavage-resistant D87A-mutated P35-His6 and D94A-mutated P49-His6 (2, 4, 35, 56) failed to form an inhibitory complex with DrICE (Fig. 5, lanes 5 and 7). Due to the presence of excess P35-His6 and P49-His6 in these assays, the uncleaved forms of both His₆-tagged proteins were also pulled down (lanes 4 to 7) and subsequently recognized by the His₆-specific reactivity of the polyclonal DrICE-His6 antiserum. Conversely, the caspase-generated cleavage fragments of P35-His6 and P49-His6 were below the limits of detection by α-DrICE. We concluded that exogenously added P49 and P35 inhibit endogenous DrICE, which suggests that both caspase inhibitors target active DrICE in the infected cell.

FIG. 5.

P49 and P35 target endogenous DrICE. Unactivated (unact) DL-1 cell extracts (lane 2) were incubated at 37°C until endogenous proDrICE was active (act) (lane 3). The activated extract was mixed with excess recombinant wild-type (wt) P35-His6 (lane 4), D87A-mutated P35-His6 (lane 5), wt P49-His6 (lane 6), or D94A-mutated P49-His6 (lane 7) bound to NTA beads and subjected to Ni⁺² affinity chromatography. The affinity-purified complexes were subjected to immunoblot analysis with α-DrICE. The His₆ tags of uncleaved P35-His6 (lanes 4 and 5) and P49-His6 (lanes 6 and 7) were also detected by polyclonal α-DrICE, which was generated against purified DrICE-His6. Processed DrICE from UV-irradiated DL-1 cells (lane 1) was included. Molecular mass standards are indicated on the left of each panel.

Depletion of DrICE by RNA silencing suppresses P49 cleavage.

Since P49 and P35 are substrate inhibitors in which their cleavage is required for caspase inhibition, the cleavage products of P49 and P35 denote the presence of a once-active caspase. To determine whether DrICE mediates cleavage of P49 and P35 and is thus targeted by these viral apoptotic inhibitors, we depleted endogenous DrICE by RNA silencing and then monitored the levels of P49 and P35 cleavage after infection. As expected, recombinant vP49 and vP35 triggered DrICE processing in nontransfected cells and cells transfected with control EGFP-specific dsRNA (Fig. 6A, lanes 2 to 3 and 7 to 8); DrICE was eliminated in drice-specific dsRNA-treated cells (lanes 4 and 9). In vP49-infected cells treated with EGFP-specific dsRNA, P49 cleavage was comparable to that in untreated cells (Fig. 6B, lanes 1 and 2). However, P49 cleavage was significantly reduced in vP49-infected cells treated with drice-specific dsRNA (Fig. 6B, lane 3). Thus, DrICE was required for P49 cleavage.

FIG. 6.

Cleavage of P49 or P35 in DrICE-silenced cells. DL-1 cells were transfected in suspension with the indicated dsRNAs, plated after 3 days, and mock-infected (mi) or inoculated with recombinant vP49 or vP35 (multiplicity of infection, ≥20). Lysates (10⁶ cell equivalents) prepared 20 h later were subjected to immunoblot analysis with α-DrICE (A), α-P49 (B), or α-P35 (C) antiserum. Proteins unrelated to P49 or P49 cleavage are indicated by the unlabeled arrowheads (left).

Because DrICE activation requires cleavage by the initiator caspase DRONC in association with the cofactor Dark (31), we next tested the effect of depleting DRONC and Dark on P49 cleavage. Confirming that RNA silencing effectively removed DRONC and Dark, virus-induced processing of proDrICE was blocked in DL-1 cells that were treated with dronc- or dark-specific dsRNA (Fig. 6A, lanes 5, 6, 10, and 11). In the DRONC- or Dark-depleted cells, little if any P49 cleavage was detected after vP49 infection (Fig. 6B, lanes 4 and 5). Thus, the removal of factors required for DrICE activation also suppressed P49 cleavage. These findings indicated that DrICE, not DRONC, mediates cleavage of P49 and is therefore the likely target of P49 in baculovirus-infected DL-1 cells.

Interestingly, RNA interference-mediated depletion of DrICE reduced but did not eliminate P35 cleavage in vP35-infected cells (Fig. 6C; compare lane 1 to lane 3). In addition, depletion of DRONC or Dark had only a minimal effect on P35 cleavage (Fig. 6C, lanes 4 and 5). Thus, P35 is cleaved by another caspase, possibly a constitutively active enzyme. These findings suggest that other, unidentified effector caspases are activated during virus infection.

ProDrICE is proteolytically processed upon overexpression.

Because DrICE is a principal target of the baculovirus antiapoptotic factors, we further investigated the steps of DrICE activation affected by these and other apoptotic regulators. To this end, we generated alanine substitutions within proDrICE that were predicted to block its proteolytic processing (Fig. 7A), including Asp230 at the large/small subunit junction (TETD²³⁰↓G), Asp28 at the prodomain/large subunit junction (DHTD²⁸↓A), and the cysteine-containing active site (Cys211). For overexpression, plasmids encoding wild-type or mutated DrICE under control of the strong baculovirus ie-1 promoter were transfected into DL-1 cells. Ectopically expressed DrICE was distinguished from endogenous DrICE by the presence of its T7 and His₆ tags.

FIG. 7.

Effects of IAP proteins on DrICE processing. (A) DrICE mutations. The locations of alanine substitutions are indicated along with the boundaries of pro-large, large, and small subunits of DrICE. The N-terminal T7 and C-terminal His₆ tags are shown. (B and C) Immunoblot analysis. DL-1 cells were transfected with ie-1 promoter expression plasmids encoding T7- and His₆-tagged wild-type (wt) DrICE or the indicated mutations either with or without plasmids encoding epitope-tagged DIAP1^HA, SfIAP^HA, or OpIAP^HA. SDS lysates of cells harvested 20 h after transfection were subjected to immunoblot analysis (10⁶ cell equivalents) by using α-DrICE (B) or α-HA (C). Lysates of cells transfected without (−) DrICE-encoding plasmids were included. A nonspecific protein detected by α-HA is indicated by the unlabeled arrowhead.

Upon overexpression in DL-1 cells, T7-proDrICE-His6 was proteolytically processed. The DrICE large subunit and the large subunit missing the link domain [large (−link) subunit], which lacks residues 218 to 230, were both produced (Fig. 7B, lane 1). Substitution of Asp28 within D28A-mutated DrICE eliminated cleavage at the DHTD²⁸↓A junction of the prodomain and produced three larger fragments. The appearance of these products suggested that alternative cleavages occurred within the prodomain upon loss of the normal processing site; indeed, the T7-tagged prodomain contains two other potential Asp-containing cleavage sites, including one with the consensus sequence DLMD↓A. Substitution of Asp230 within D230A-mutated DrICE eliminated cleavage at the large/small subunit junction TETD²³⁰↓G and yielded a fragment consistent with the size of a prodomainless DrICE, Δpro (Fig. 7, lane 4). Lastly, loss-of-function C211A-mutated DrICE was also cleaved to produce the large subunit (lane 2), albeit at lower levels. Since active-site C211A-mutated DrICE lacks enzymatic activity, this cleavage was most likely mediated by a constitutively active caspase(s). We concluded that proteolytic processing of proDrICE involves caspase-mediated cleavages at DHTD²⁸↓A and TETD²³⁰↓G in a pattern that resembles that of the related effector caspase Sf-caspase-1 of lepidopteran cells (26).

Proteolytic processing of proDrICE is differentially affected by IAP proteins.

Upon coexpression of Drosophila DIAP1, processing of T7-proDrICE-His6 was blocked (Fig. 7B, lane 5 to 8). Thus, a DIAP1-sensitive caspase was responsible for the Asp28 and Asp230 cleavages of overexpressed proDrICE. In contrast, lepidopteran SfIAP failed to block processing of overexpressed DrICE (Fig. 7B, lanes 9 to 12) and thus differed from when it was expressed during infection (Fig. 3). As expected, baculovirus OpIAP failed to block processing (lanes 13 to 16). Interestingly, both SfIAP and OpIAP blocked the Asp28 and Asp230 cleavages of C211A-mutated DrICE (Fig. 7B; compare lane 2 with lanes 10 and 14), suggesting that the endogenous caspase responsible for these events is inhibited by both heterologous IAP proteins. All three IAP proteins were readily detected in cotransfected cells (Fig. 7C).

Neither P49 nor P35 blocks proteolytic processing of overexpressed proDrICE.

The effects of P49 and P35 on proteolytic processing of overexpressed DrICE were determined by cotransfection of plasmids encoding T7-proDrICE-His6 and P49 or P35. The patterns of processing of wild-type and mutated DrICE in the presence of wild-type P35 were nearly identical to that in the presence of loss-of-function D87A-mutated P35 (Fig. 8B, lanes 1 to 8). Likewise, DrICE processing in the presence of wild-type P49 and loss-of-function D94A-mutated P49 were nearly identical (Fig. 8B, lanes 9 to 16). The only effect of P49 and P35 was to suppress accumulation of the large subunit from C211A-mutated DrICE (lanes 6 and 10), which indicated that the endogenous caspase(s) responsible for this cleavage is P35 and P49 sensitive. Thus, similar to that in the context of infected cells (Fig. 3B), processing of DrICE occurred in the presence of P49 and P35.

FIG. 8.

Effects of P49 and P35 on DrICE processing. (A) DrICE mutations. Alanine substitutions relative to the boundaries of pro-large, large, and small subunits of DrICE are indicated. (B, C, and D) Immunoblots. DL-1 cells were transfected with ie-1 promoter expression plasmids encoding T7- and His₆-tagged wild-type (wt) DrICE or the indicated mutations either with or without plasmids encoding wild-type P35, loss-of-function (lof) D87A-mutated P35, wild-type P49, or lof D94A-mutated P49. Lysates of cells harvested 20 h after transfection were subjected to immunoblot analysis (10⁶ cell equivalents) with α-DrICE (B), α-P35 (C), or α-P49 (D). Lysates of cells transfected without (−) plasmids were included. A nonspecific protein detected by α-P35 is indicated the unlabeled arrowhead (right). In panel D, polyclonal α-P49 (generated against P49-His6) also detected the His₆ epitope of T7-proDrICE-His6 (arrows).

Upon DrICE overexpression, P35 and P49 were cleaved vigorously (Fig. 8C, lane 2, and Fig. 8D, lane 7). This cleavage was due to the exogenously expressed T7-proDrICE-His6, as neither caspase substrate was cleaved in the absence of DrICE-encoding plasmid or in the presence of C211A-mutated DrICE plasmid (Fig. 8C and D). The nearly complete disappearance of P35 and P49 (Fig. 8C, lanes 2 and 7, and Fig. 8D, lanes 2, 3, and 7) suggested that both substrate inhibitors were fully titrated by an excess of DrICE and thus were unable to block processing of the pro-large intermediate to the mature large subunit, which was the most abundant product detected (Fig. 8B, lanes 5 and 9). As expected, D87A-mutated P35 and D94A-mutated P49 were not cleaved (Fig. 8C, lanes 6 to 10, and Fig. 8D, lanes 1 to 5). P35 and P49 were cleaved upon overexpression of D28A- and D230A-mutated DrICE, indicating that these processing-defective forms of DrICE were at least partially active. Consistent with these findings was the observation that DL-1 cells transfected with plasmids encoding functional DrICE displayed signs of cytolysis and apoptosis, whereas those cells expressing nonfunctional C211A-mutated DrICE did not (data not shown).

DrICE is proteolytically activated upon overexpression in heterologous cells.

To determine if the pattern of DrICE processing is cell line dependent, we monitored the fate of T7-proDrICE-His6 after transfection of cultured SF21 cells, a cell line derived from the baculovirus-permissive lepidopteran host Spodoptera frugiperda. Upon overexpression, T7-proDrICE-His6 was processed in part to its large and large (−link) subunits (Fig. 9A, lane 1). Because nonfunctional C211A-mutated DrICE was not cleaved (lane 10), the processing of wild-type DrICE was attributed to autocleavage. D28A-mutated and D230A-mutated DrICE also yielded processing products (lanes 9 and 11) comparable to those observed in DL-1 cells. Interestingly, coexpression of P35 (lane 2), P49 (lane 4), and DIAP1 (lane 6), but not SfIAP, OpIAP, loss-of-function mutated P35, or loss-of-function mutated P49, reduced proDrICE processing to the large subunit. Consistent with their mechanism of substrate inhibition, both P35 and P49 were cleaved upon coexpression with wild-type DrICE and D28A- and D230A-mutated DrICE (Fig. 9B and C). The finding that neither P35 nor P49 was cleaved when expressed alone or in the presence of loss-of-function C211A-mutated DrICE suggested that the P35/P49 cleavage activity was DrICE mediated.

FIG. 9.

Effects of apoptotic inhibitors on DrICE processing in heterologous cells. SF21 cells were transfected with plasmids encoding wild-type (wt) T7-proDrICE-His6 or the indicated mutations either with or without plasmids encoding wild-type or loss-of-function (lof) D87A-mutated P35, wild-type or lof D94A-mutated P49, DIAP1^HA, SfIAP^HA, or OpIAP^HA. SDS lysates of cells harvested 20 h after transfection were subjected to immunoblot analysis (10⁶ cell equivalents) by using α-DrICE (A), α-P35 (B), α-P49 (C), or α-Sf-caspase-1 (D). Lysates of SF21 cells transfected without (−) plasmids or with plasmids encoding only OpIAP^HA were included. α-DrICE failed to detect endogenous Sf-caspase-1 (panel A, lane 1). However, α-Sf-caspase-1 detected overexpressed DrICE (panel D, lanes 9 to 19). The constant levels of pro-Sf-caspase-1 in cells transfected with wild-type DrICE and inactive C211A-mutated DrICE-encoding plasmids indicated that comparable levels of protein were loaded in each lane.

Because proDrICE itself is a caspase substrate, we determined whether endogenous Spodoptera caspases contributed to DrICE processing in SF21 cells. Immunoblot analyses indicated that Sf-caspase-1 was proteolytically processed to its active large subunit but only at low levels, as suggested by the predominance of pro-Sf-caspase-1 present in transfected cells (Fig. 9D, lanes 9 to 19). Sf-caspase-1 processing was reduced upon coexpression of P35 and P49 but not of OpIAP (compare lane 9 to lanes 10, 12, and 16). Because OpIAP does not inhibit DrICE (Fig. 3 and 7) (49), we concluded that the overexpressed DrICE was responsible for the low level of Sf-caspase-1 processing. Thus, DrICE itself, and not Sf-caspase-1, was the principal mediator of DrICE activation in transfected SF21 cells.

DISCUSSION

Baculoviruses are well known for the capacity to induce rapid and widespread apoptosis in lepidopteran cells (7, 13). Virus-mediated suppression of apoptosis blocks premature cell death and enhances baculovirus multiplication, which suggests that apoptosis is an antiviral defense by the host. Although AcMNPV infection of Drosophila DL-1 cells is nonproductive (30), virus entry is efficient and early gene expression is sufficient to trigger apoptosis in every cell. Here we report that AcMNPV induces apoptosis in nonpermissive Drosophila cells by activating a prototypical initiator → effector caspase pathway that requires the Drosophila effector caspase DrICE. Moreover, baculovirus P49 and P35 prevented apoptosis by targeting active DrICE and thus functioned at a step different from that in permissive lepidopteran cells. Thus, our studies indicate an unusual versatility of the mechanisms by which baculovirus apoptotic suppressors function in different cells.

DrICE-mediated apoptosis during baculovirus infection.

DrICE is a principal effector caspase in Drosophila and is an important contributor to apoptosis induced by developmental cues and external signals (16, 25). By depleting endogenous DrICE through RNA silencing, we have shown that DrICE is a major contributor to baculovirus-induced apoptosis, a finding that broadens the role of this caspase in Drosophila. DrICE depletion increased virus-infected cell survival by more than fivefold compared to control-silenced cells (Fig. 2A) and reduced intracellular DEVD-ase activity to background levels (Fig. 2B). Interestingly, despite the decrease in DrICE to below the limits of immunoblot detection (Fig. 2C and 6A), not all virus- or UV radiation-induced apoptosis was blocked by drice-specific RNA silencing (Fig. 2A). Thus, other Drosophila caspases may contribute to virus-induced apoptosis. Indeed, the closely related effector caspase DCP-1 is proteolytically processed from its proform during infection (E. Lannan and P. Friesen, unpublished data), suggesting a possible role for DCP-1. Both DrICE and DCP-1 are activated through proteolytic cleavage by the Drosophila initiator caspase DRONC and its cofactor Dark (16, 25). Consistent with this initiator role for DRONC in DL-1 cells, RNA interference-mediated depletion of DRONC or Dark blocked virus-induced processing of DrICE (Fig. 6A). These findings suggested that baculovirus infection initiates a host pathway that activates the initiator caspase DRONC, which in turn activates DrICE and other effector caspases. Thus, our studies indicate that activation of DrICE follows a path that is common to invertebrate effector caspases (Fig. 10), including that of the closely related effector caspase Sf-caspase-1 from Spodoptera frugiperda (10, 26, 27, 43, 56).

FIG. 10.

Model for regulation of the Drosophila effector caspase DrICE. Upon baculovirus-induced apoptotic signaling, proDrICE is activated by cleavage at the large/small subunit junction (TETD²³⁰↓G) by a protease (DRONC) that is inhibited by overexpressed cellular IAP protein DIAP1 (and SfIAP) but not by the baculovirus caspase inhibitor P49 or P35. Subsequently, the DrICE prodomain is removed from the pro-large subunit at DHTD²⁸↓A by a protease (possibly DrICE) that is sensitive to P49 and P35 inhibition. Active DrICE, composed of mature large and small subunits, is inhibitable by P35, P49, and DIAP1 but not by baculovirus OpIAP.

IAP regulation of DrICE activation.

When overexpressed by a recombinant virus, DIAP1 blocked virus-induced processing of proDrICE (Fig. 3B). Thus, DIAP1 functioned upstream of effector caspase activation, which is consistent with one of its documented antiapoptotic functions in Drosophila (16, 25). Suggesting a conservation of function between the lepidopteran and dipteran IAP homologs, overexpression of SfIAP also blocked DrICE processing (Fig. 3B). We concluded that the path by which baculoviruses trigger apoptosis is regulated by endogenous host IAP proteins that are likely inactivated by virus-induced signaling. In contrast, baculovirus OpIAP failed to block DrICE activation. This finding is consistent with the inability of OpIAP to prevent apoptosis in Drosophila or to affect caspase activity (28, 43, 49, 56). Thus, the antiapoptotic mechanisms of viral and cellular IAP proteins differ significantly.

P49 and P35 function as effector caspase inhibitors in Drosophila cells.

In Spodoptera cells (Lepidoptera; Noctuidae), baculovirus P49 functions upstream of P35 and inhibits proteolytic processing and activation of the effector caspase Sf-caspase-1 (56). Thus, an unexpected finding of this study is that overexpressed P49 failed to block processing of the effector caspase DrICE (Fig. 3B), even though it prevented virus-induced apoptosis. Instead, P49 inhibited DrICE after its activation by forming a stable inhibitory complex through a P35-like mechanism that requires the cleavage site residue Asp94 (Fig. 5). Upon depletion of DrICE by RNA silencing, cleavage of P49 was blocked (Fig. 6). This finding suggested that P49 is cleaved by DrICE and thus targets processed DrICE during infection. Like P49, P35 also failed to prevent DrICE activation but formed an inhibitory complex with active DrICE. Since it is expected that DRONC mediates DrICE activation (15, 29), then DRONC must be resistant to inhibition by P49 and P35 during infection. It has been shown previously that P49, but not P35, blocks DRONC-mediated cell death in yeast viability assays (22). However, direct tests of P49 inhibition of DRONC in biochemical assays have not been reported. The molecular mechanism by which DRONC escapes inhibition by both baculovirus caspase inhibitors remains to be determined.

Whereas P49 is a potent inhibitor of effector caspases from diverse organisms, including insects and humans, it is also an inhibitor of initiator caspases, including human caspase 9 and the initiator caspase that activates Sf-caspase-1 (22, 35, 56). This unique feature is not shared with P35, which is more selective for effector caspases (2, 15, 17, 29, 54). P49's versatility for blocking apoptosis at different steps, depending upon the infected cell, likely provides a selective advantage to P49-encoding baculoviruses. The molecular mechanism for this broad-spectrum inhibition is under investigation.

Although P49 and P35 blocked virus-induced apoptosis by inactivation of active DrICE, our study suggests that these caspase inhibitors are not equivalent in Drosophila. The pronounced accumulation of pro-large DrICE in the presence of P35 compared to P49 (Fig. 4A) suggested that P35 is more efficient than P49 at inhibiting the protease that removes the DrICE prodomain (see below). Moreover, when DrICE was depleted by RNA silencing, residual caspase-mediated cleavage of P35 was detected, whereas cleavage of P49 was not (Fig. 6). This finding raises the possibility that in addition to DrICE, P35 targets other active Drosophila caspases, but P49 is more selective for DrICE. Additional studies are required to examine these interesting differences.

Proteolytic steps in DrICE activation.

Due to the importance of DrICE in the execution of baculovirus-induced apoptosis, we have also defined the steps by which proDrICE is proteolytically processed and activated. Our studies involving the overexpression of cleavage site substitutions of proDrICE indicated that processing occurs at the large/small subunit junction TETD²³⁰↓G and at the prodomain/large subunit junction DHTD²⁸↓A (Fig. 10). When endogenous DrICE was activated in the presence of P49 and P35, the pro-large subunit accumulated to high levels compared to that in the absence of these inhibitors (Fig. 2D and 3B). This finding suggested that the pro-large fragment is a processing intermediate from which the fully cleaved large subunit is generated (Fig. 10). The removal of the prodomain from the pro-large fragment by cleavage at DHTD²⁸↓A is mediated by a P49- and P35-sensitive protease, whereas cleavage at TETD²³⁰↓G is mediated by a P49-, P35-insensitive, DIAP1-inhibitable protease (Fig. 3, 7, and 8). All of these forms of processed DrICE were also observed in our caspase-activated cell extracts of DL-1 cells (R. Vandergaast and P. Friesen, unpublished data). Collectively, these data are consistent with a model (Fig. 10) in which virus-induced activation of DrICE begins by cleavage at the large/small subunit junction by a DIAP1-sensitive caspase (DRONC), followed by removal of the prodomain by a DXXD-A/G effector caspase (possibly DrICE). Thus, this model resembles that for the activation of the insect effector caspase Sf-caspase-1, in which processing occurs at similar sequences, TETD↓G and DEGD↓A, respectively (26, 27). The pro-large subunit of DCP-1 was also apparent in virus-infected DL-1 cells (E. Lannan and P. Friesen, unpublished data), suggesting that DCP-1 follows a similar activation path. An alternative pathway in which activation of proDrICE begins with removal of the prodomain has been proposed upon DrICE overexpression in Drosophila S2 cells (46).

When overexpressed in Drosophila cells by plasmid transfection, a significant fraction of the proDrICE was processed to its fully mature subunits (Fig. 7 and 8). The observation that overexpressed DIAP1 but not P49 or P35 blocked the activation cleavages (Fig. 7 and 8) argues that endogenous DRONC was involved. Indeed, constitutively active DRONC has been detected in cultured Drosophila cells (33). Inactive C211A-mutated proDrICE was also cleaved by a P35-, P49-, and DIAP1-inhibitable caspase (Fig. 8), indicating that a low-level, constitutively active effector caspase may also contribute to initial processing. However, we expect that exogenous DrICE mediated most of the processing, since plasmid-transfected DL-1 cells exhibited high levels of caspase activity, as judged by vigorous cleavage of P35 and P49 (Fig. 8). Thus, when DrICE is overexpressed, autoactivation and feedback amplification may occur. Consistent with this possibility, ΔproDrICE that was generated by a P35- and P49-sensitive cleavage at the DXXD-A effector caspase site was abundant upon overexpression of D230A-mutated DrICE (Fig. 8). It is noteworthy that each of the DrICE processing mutants studied here, except nonfunctional C211A-mutated DrICE, exhibited high levels of caspase activity. Thus, complete processing of proDrICE may not be required for its activity in situations where DrICE is overexpressed. Consequently, caution must be exercised when caspase regulation is studied in systems requiring caspase overexpression.