Caspases are well known for their role in the execution of apoptotic programs, in which they cleave specific target proteins, leading to the elimination of cells, and for their role in cytokine maturation. In this study, we identified a novel substrate, which, through cleavage by caspases, can regulate Drosophila neural precursor development. Shaggy (Sgg)46 protein, an isoform encoded by the sgg gene and essential for the negative regulation of Wingless signaling, is cleaved by the Dark‐dependent caspase. This cleavage converts it to an active kinase, which contributes to the formation of neural precursor (sensory organ precursor (SOP)) cells. Our evidence suggests that caspase regulation of the wingless pathway is not associated with apoptotic cell death. These results imply a novel role for caspases in modulating cell signaling pathways through substrate cleavage in neural precursor development.
Caspases, key mediators in cell‐killing pathways, are a class of cysteine proteases that includes mammalian interleukin‐1β‐converting enzyme (ICE) and the ced‐3 gene of Caenorhabditis elegans (C. elegans) (Miura and Yuan, 1996). A functional apoptotic program requires multiple caspases, each contributing to a different phase of apoptosis. During this program, activated caspases cleave a variety of substrates, including other caspases, which ultimately leads to morphological changes that indicate apoptosis. The requirement for ‘killer caspases’ in this process has been demonstrated through a variety of experimental approaches. Since several members of the caspase family process each other, resulting in the effective amplification of cell death signals (Thornberry and Lazebnik, 1998), studies on the activation mechanisms of these caspases and identification of caspase substrates are crucial for understanding the cell death machinery. However, although many caspase substrates have been identified, the biological functions of most of these substrates remain largely unknown. The identification of caspase targets and determination of the effects of caspase‐mediated cleavage on their function will be critical to understanding the mechanisms of apoptosis as well as other downstream events.
Recently, evidence for nonapoptotic functions of caspases in cellular processes has been accumulating. Caspase activity is required in some cells for the regulation of survival pathways, cell motility, receptor internalization, and may regulate cell‐cycle progression and proliferation (Algeciras‐Schimnich et al, 2002). In particular, recent studies implicate effector caspases in mouse lens development, highlighting a potential role for the caspase cleavage of α‐spectrin (Lee et al, 2001), and caspase activation is required for the terminal differentiation of human epidermal keratinocytes that are necessary for the loss of the nucleus (Weil et al, 1999). Moreover, a previous report implicated caspase‐8 as a negative regulator of erythroid development through its processing of GATA‐1 (De Maria et al, 1999), and a more recent study demonstrated that caspases are required for terminal erythroid differentiation through the cleavage of lamin B and acinus (Zermati et al, 2001). In Drosophila, caspase activity is required for the individualization of sperm differentiation (Arama et al, 2003; Huh et al, 2004b), and provided a model where apoptotic cells activate signaling cascades for compensatory proliferation (Huh et al, 2004a; Perez‐Garijo et al, 2004; Ryoo et al, 2004), but the precise mechanisms remain to be elucidated. Recently, a cell death‐independent role for Drosophila caspases in Rac‐mediated cell motility was reported (Geisbrecht and Montell, 2004), and loss of the Drosophila caspase DRONC leads to an unappreciated role in the maintenance of adult tissues including the eye, wing, and midgut (Chew et al, 2004; Daish et al, 2004). These findings point to the involvement of caspases in regulating cell death, survival, and differentiation.
In Drosophila, seven caspases (DCP‐1, drICE, Dredd, DRONC, Decay, DAMM, and STRICA) have been identified and shown to be required for cell death induced by various stimuli (Hay et al, 2004). These caspases are mediated, in part, by an Apaf‐1 homolog, Dark (Dapaf‐1/HAC‐1), and two Bcl‐2 family proteins (Drob‐1/Debcl/dBorg‐1/dBok and Buffy/dBorg‐2) (Baehrecke, 2002). In addition, caspase‐dependent cell death can be induced by the Drosophila killer proteins Reaper and Hid, and strongly inhibited by the overexpression of caspase inhibitory proteins, p35, DIAP1, and DIAP2 (McCall and Steller, 1997). The Drosophila caspases have distinct specificities for substrate recognition (Hawkins et al, 2000; Song et al, 2000), which could potentially be used to modulate nonapoptotic signaling pathways during development. In this study, we identified a novel substrate that, through cleavage by caspases, can be involved in the regulation of Drosophila neural precursor development. Here we reveal that the Shaggy (Sgg)46 protein, an isoform encoded by the sgg gene and essential for the negative regulation of Wingless signaling, is converted by caspase cleavage into an active kinase, which then might contribute to the formation of neural precursor (sensory organ precursor (SOP)) cells.
Dark‐dependent caspase activation is involved in macrochaete formation
Loss of maternal and zygotic Dark (Dapaf‐1/HAC‐1) function in Drosophila results in decreased caspase activity and reduced apoptosis in the embryonic and larval brain (Kanuka et al, 1999a; Rodriguez et al, 1999; Zhou et al, 1999). Macrochaetes are external sensory organs located on the notum and are typical structures of the Drosophila peripheral nervous system (PNS). Four successive asymmetric cell divisions from a common precursor cell called the SOP generate a sensory bristle consisting of four different non‐neuronal support cells and one neuron (Figure 1A). As shown in Figure 1B, four large macrochaetes are observed on the wild‐type scutellum. However, extra macrochaetes often appeared on the scutellum of dark mutant flies (48%) (Figure 1C; Kanuka et al, 1999a; Rodriguez et al, 1999). To determine if the appearance of these extra macrochaetes was caspase‐dependent, the caspase inhibitory protein p35 was expressed in the scutellum (GAL4/UAS system; Brand and Perrimon, 1993). Indeed, the inhibition of caspases in the scutellum by scabrous‐GAL4 (sca‐GAL4) resulted in the formation of ectopic macrochaetes similar to those observed in the dark mutants (Figure 1D and E; Kanuka et al, 1999a). This phenotype was not limited to p35 expression, given that both DIAP1 and DIAP2 expression also resulted in additional macrochaetes (Table I). Interestingly, blockage of the downstream Dark target, DRONC, either through the expression of a dominant‐negative form (DRONC DN) or ectopic expression of an inverted repeat (IR) dsRNA against DRONC (which abolishes protein expression, data not shown), induced the same phenotype (Table I). These results concerning DRONC function are consistent with a previous report that mild but not complete, loss of function of dronc also causes extra macrochaete formation (Chew et al, 2004). The more severe phenotype we observed likely stems from overexpression of DRONC DN, which may inhibit caspase function more effectively than in the partial dronc mutant. In contrast, the ectopic expression of p35 and DRONC DN specifically in SOP cells, through the use of neur‐GAL4, did not result in extra macrochaetes (Figure 1F–I, Table I). Taken together, the Dark/DRONC‐dependent caspase activation pathway appears to be involved in regulating the macrochaete number before SOP cell formation.
Dark‐dependent caspase activation negatively regulates SOP cell formation
We next examined whether extra macrochaetes are actually produced from extra SOP cells. Initially during the formation of SOP cells, a group of cells called a proneural cluster, characterized by the expression of the proneural genes achaete and scute, forms at the future site of the macrochaetes in the wing disc (Jan and Jan 1994). These proneural clusters also express the scabrous gene (Figure 1A and F), followed by the expression of achaete and senseless, in one or a few cells singled out from the proneural cluster (Figure 1A and 2A; Nolo et al, 2000). We performed immunostaining using antibodies against Achaete and Senseless on wing discs from the third‐instar larval stage. In wild‐type larvae, two sets of SOP cells were seen in scutellum region of one wing disc (Figure 2A). However, in both dark mutants and DRONC DN‐expressing flies, we observed one extra SOP cell expressing Achaete and Senseless (Figure 2B and C). This result indicates that caspase activation is involved in the control of SOP cell formation in the scutellum area.
As shown in Figures 1A and 2D, one SOP cell will produce a set of one neuron and one thecogen cell in the pupal notum. These SOP daughter lineage cells can be characterized by specific marker proteins, Elav (neuron) and Prospero (thecogen) (Jan and Jan, 1998). To examine whether the extra SOP cells induced by caspase inhibition at the larval stage have the potential to develop into a complete set of extra sensory organs, the scutellum region of the pupal notum (24–30 h after pupal formation (APF)) was subjected to immunostaining. Indeed, we observed extra sets of both Elav‐ and Prospero‐positive cells in dark mutants and DRONC DN‐expressing flies (Figure 2F and G). These observations suggest that caspase activity negatively mediates the formation of scutellum SOP cells.
Caspase activation does not correlate with cell death in SOP cell development
To identify the exact moment at which caspase activity regulates macrochaete formation, we transiently expressed p35 using heat shock. Consistent with the SOP cell‐specific expression of p35 and DRONC DN, no significant increase in macrochaete number was seen as a result of caspase inhibition during SOP cell formation (i.e. late third‐instar, 6–12 h APF) or after it (24–30 h APF) (Figure 3A). At the stage prior to the appearance SOP cells, however, the transient expression of p35 had a strong effect on the production of extra bristles (Figure 3A). These data strongly support the idea that the inhibition of caspase activity leads to the formation of extra SOP cells rather than affecting differentiation at a later stage.
As caspases are well‐known executioners of apoptosis and cell death, which lead to the elimination of cells (Thornberry and Lazebnik, 1998), we investigated whether the increased number of SOP cells resulting from reduced caspase function was related to the inhibition of cell death. At first, we examined whether caspases are actually activated in the scabrous‐expressing proneural clusters that contain SOP cells. We previously reported the development of an indicator molecule for caspase activation (SCAT3) that uses fluorescence resonance energy transfer (FRET) between two types of fluorescent protein (ECFP and Venus), linked by a peptide containing the caspase‐3 cleavage sequence, DEVD (Takemoto et al, 2003). Using this probe, we can monitor caspase activation at the single‐cell level by observing the changes in the Venus/ECFP ratio. After performing confocal imaging of caspase activity in sca‐GAL4+UAS‐SCAT3 flies, a decrease in FRET (i.e. activation of caspase) was observed in scabrous‐expressing proneural clusters in wild‐type wing discs (Figure 3B). We also confirmed that this change of ratio was clearly recovered by coexpressing DRONC DN (Figure 3C), indicating that scabrous‐positive cells actually have a DEVDase‐like caspase activity that can be inhibited by the dominant‐negative form of DRONC.
A large amount of naturally occurring cell death is observed in the wing disc of second‐ and third‐instar larvae (Milan et al, 1997; Adachi‐Yamada et al, 1999). This cell death is programmed to eliminate excess cells produced by a high level of proliferation or differentiation (Adachi‐Yamada et al, 1999; Moreno et al, 2002). Consistent with this, we also observed massive cell death by TUNEL staining in the third‐instar larval wing disc (Figure 3D and E), allowing for the possibility that the extra SOP cells normally disappear via cell death in wild‐type flies. Surprisingly, whereas containing an amount of caspase activity (Figure 3B and C), careful examination of wing discs at early and late larval stages (n=21 and n=28, respectively), in which caspase inhibition appeared to be effective (Figure 3C), showed no dead cells in the scabrous‐expressing proneural clusters in wild‐type flies (Figure 3D and E). However, since TUNEL labeling is a transient event, it is difficult to rule out the possibility of inhibition of death, even in the case that these are no TUNEL‐positive cells in scabrous cluster. The mechanism to produce extra macrochaete by inhibition of caspase could arise through several possible mechanisms, and we also prefer the mechanism that the cell fate change could be induced by caspase inhibition. These results suggest that the increased number of SOP cells seen in the caspase‐inhibited wing discs was not a result of the artificial survival of cells that otherwise should have undergone apoptotic cell death, but rather was due to inhibition of an apoptosis‐independent caspase activity.
Genetic identification of Sgg as a component of caspase‐regulated macrochaete development
To identify possible downstream targets of this apoptosis‐independent caspase activity, we performed a dominant‐modifier screen in which we monitored macrochaete numbers. The ectopic expression of DRONC DN using sca‐GAL4 gave rise to a dose‐dependent increase in the number of macrochaetes (Table I and data not shown). In our modifier screen, we identified 29 suppressors and three enhancers covering large chromosomal regions. We secondarily screened and identified small‐deficiency lines, and subsequently tested the ability of available pre‐existing mutants located in these regions to reproduce the phenotype exhibited by the sca‐GAL4+UAS‐DRONC DN flies. As a result, components of Notch signaling (numb, Su(H), Delta, and neur) and the Wingless pathway (wingless and shaggy) were identified (Table II and data not shown). Both of these signaling pathways are important for macrochaete development, especially in the regulation of positional patterning and number (Simpson et al, 1999). Thus, we further analyzed the Wingless pathway in relation to caspase‐mediated SOP cell formation.
In cells that receive Wingless signals, Armadillo (Arm)/β‐catenin accumulates and translocates to the nucleus to activate Wingless‐responsive genes. This signal‐induced accumulation of Arm/β‐catenin results from an interruption of its normal turnover mediated by phosphorylation and subsequent ubiquitination–proteasomal degradation (Wodarz and Nusse 1998). The kinase Sgg, the Drosophila GSK‐3β ortholog, is an antagonist of Wingless signal transduction and has been shown to be critical for this phosphorylation of Arm/β‐catenin (Dierick and Bejsovec 1999). In our modifier screen, we found a regulatory mutant of Wingless (wgSP−1) that was able to suppress the increase in extra macrochaetes, while a completely null mutant of Sgg (sggD127) enhanced the phenotype (Table II). This result genetically suggests that endogenous caspase activity antagonizes Wingless‐dependent macrochaete formation. In agreement with this, loss of Sgg function (achieved by an sggD127−/− mitotic clone) resulted in extra macrochaetes and microchaetes in the dorsal part and scutellum of the adult notum (Figure 4A–D), where Wingless is highly expressed (Figure 4E). Similarly, a previous study reported that sggD127 −/− adult flies produced by transient rescue of the sgg gene also exhibited severe hyperplasia of macrochaetes in the same region (Ruel et al, 1993a). In support of these data, sgg −/− cells in the wing discs significantly expressed both Senseless and Achaete, marker proteins for SOP cells (Figure 4F), indicating that these extra macrochaetes are mainly produced by ectopic differentiated SOP cells. In contrast, the ectopic expression of Sgg10, one of the major isoforms of the sgg gene transcripts, caused a clear loss of macrochaetes in the scutellum (Figure 4H, compared with 4G) and loss of SOP cells in the wing disc (Figure 6G). This phenotype depends on Sgg kinase activity because a kinase‐negative form of Sgg10 (Sgg10 A81T) did not affect macrochaete formation (Figure 4I). The same phenotypes were reported using another GAL4 driver (sc(455.2)‐GAL4) (Bourouis, 2002). Taken together, these data indicate that Sgg kinase activity negatively regulates SOP cell and macrochaete development.
Shaggy kinase isoform is a substrate for caspase
We next investigated how caspases are involved in the regulation of Sgg kinase activity. Some typical caspase families recognize and cleave motifs corresponding to DEVD (Asp‐Glu‐Val‐Asp). Interestingly, Sgg46, one of three major isoforms of Sgg (Ruel et al, 1993b), contains two such motifs, suggesting it may be a direct target for caspase cleavage (Figure 5A). Both of the caspase‐cleavage DEVD domains are found in the N‐terminal portion of Sgg46, before the kinase domain that is shared with Sgg10. To examine whether Sgg46 can be processed by caspases, Myc‐tagged Sgg46 was expressed along with various caspases in Drosophila S2 cells (Figure 5B). As a positive control, staurosporine, which is known to induce cell death via caspase activation in S2 cells, was used (Kanuka et al, 1999b). A single fragment (⩽39 kDa), which was likely processed from full‐length Sgg46 protein, was observed specifically in drICE‐ and DRONC‐expressing cells, in addition to staurosporine‐treated cells (Figure 5B). We did not detect this fragment when a putative noncleavable form of Sgg46, Sgg46 D300G, was expressed with drICE, indicating that caspases are responsible for cleaving Sgg46 at Asp300 (Figure 5B).
Next, we isolated an S2 cell line that stably expresses the cell death activator reaper under the control of the heat‐shock promoter. Expression of reaper remarkably induced the endogenous caspase activity (DEVDase) 60 min after heat‐shock treatment (Figure 5C) and the subsequent processing of drICE into the active form (Figure 5D). Correlated with this caspase activation, two small Myc‐tagged fragments and a large fragment recognized by an antibody against the cleaved form of Sgg46 (Asp300, Figure 5E) were observed (Figure 5F and G). These fragments could not be detected in S2 cells expressing Sgg46 D300G (upper band in Figure 5F and large band in 5G) and Sgg46 D235G (lower band of Figure 5F, data not shown), suggesting that Sgg46 can be cleaved by endogenous caspases at Asp235 and Asp300.
Caspase‐dependent cleavage alters Sgg46 function in SOP cell development
Previous genetic studies of sgg mutant flies showed the interesting observation that some phenotypes of sgg mutants can be rescued by the expression of sgg10 or sgg39 (the other sgg isoform similar to sgg10), but not sgg46 (Ruel et al, 1993a, 1993b), suggesting that Sgg46 might be an inactive form. The Sgg10 kinase phosphorylates the Arm protein and induces its degradation (Yanagawa et al, 1997; Ruel et al, 1999). We tested various forms of Sgg protein for this activity (Figure 6A). As previously reported, expression of Sgg10 induced Arm phosphorylation and degradation in a kinase‐dependent manner (Figure 6B, myc‐Sgg10 and myc‐Sgg10 K83R, a kinase‐negative form). In contrast, full‐length Sgg46 did not produce the same effects on the Arm protein (Figure 6B, myc‐Sgg46). Interestingly, expression of a putative cleaved form of Sgg46, containing the kinase domain (myc‐Sgg46 ΔN235 and myc‐Sgg46 ΔN300), led to Arm phosphorylation and degradation in a manner similar to that of Sgg10 (Figure 6B). These results suggest that full‐length Sgg46 is an inactive form that can be converted into an active kinase via caspase‐dependent cleavage.
We next examined whether these findings were applicable to macrochaete and SOP cell development in vivo by using transgenic flies expressing Sgg proteins. As shown above, the ectopic expression of Sgg10 by sca‐GAL4 caused the loss of macrochaetes and SOP cells (Figure 6D, compared with 6C). No apoptotic cells in the myc‐Sgg10 protein‐expressing region of the wing disc could be detected, indicating that this disappearance did not result from the death of SOP cells (Figure 6G). Consistent with the immunoblotting results, full‐length Sgg46 did not influence macrochaete and SOP cell formation (Figure 6E and H), whereas the cleaved form of Sgg46 (Sgg46 ΔN300) worked in a manner similar to that of Sgg10 (Figure 6F and I). After crossing sca‐GAL4+UAS‐DRONC DN to UAS‐sgg10, most F1 progeny showed a clear loss of macrochaetes in the scutellum (data not shown), indicating that Sgg kinase activation might be downstream of caspases. These observations suggest that the processing of Sgg46 by caspases leads to the formation of an active kinase that can negatively regulate SOP cell development.
Finally, we investigated Sgg46 contributes significantly to macrochaete and SOP cell formation in vivo. The ectopic expression of Sgg46 D235G/D300G by sca‐GAL4 significantly induced extra macrochaetes and SOP cells (Figure 7A–C). Since Sgg46 D235G/D300G could not be cleaved by caspases (Figure 5B, F and G), this noncleaved Sgg46 might act as dominant‐negative form against endogenous Sgg function. Furthermore, an ectopic knockdown of Sgg protein expression by dsRNA‐expressing constructs (Figure 7D) revealed that the specific reduction of the Sgg46 protein induced extra macrochaetes (Figure 7E–G). However, inhibition of Sgg46 is less effective at producing extra macrochaetes than inhibiting Dark or DRONC, suggesting that modulation of Sgg kinase activity may not be the only mechanism contributing to SOP formation. It still remains to be examined whether or not Sgg46 is actually cleaved and converted into an active form in proneural clusters, and will require further examination in vivo. Based on our findings that loss of Sgg function or inhibition of caspase activity resulted in extra macrochaetes mainly in the scutellum of the adult notum (pSC and aSC), where Wingless is highly expressed (Figure 4A–E), and that caspases are activated in scabrous‐expressing cluster (Figure 3B and C), we could consider that scabrous‐expressing SOP cells that will produce pSC and aSC macrochaetes are located in specific region, where precise formation of each set of macrochaetes might require both (1) Wg expression (to increase bristle) and (2) caspase activation (to decrease bristle). Thus, it appears that Dark‐dependent caspase signaling mediates the total Sgg kinase activity by processing Sgg46 into an active form, thereby negatively regulating Wingless‐sensitive macrochaete development (Figure 7H).
Signal transduction mediated by the caspase family
Although the contribution of caspases to development through apoptosis has been well documented, caspase downstream targets and potential nonapoptotic roles have remained elusive. In this study, we observed that the Sgg46 protein, an isoform encoded by the sgg gene and essential for the negative regulation of Wingless signaling, could be cleaved by caspases and thereby converted into an active kinase. This cleaved form of Sgg46 was then able to contribute to the formation of neural precursor (SOP) cells (Figure 7H). However, whereas a plethora of homologous Sgg/GSK‐3β proteins in many species including yeast and mammals have been identified, we could not find any isoforms similar to Sgg46, even in the mosquito Anopheles genome (data not shown). This indicates that the additional Sgg46 domain attached to the conserved kinase domain of Sgg/GSK‐3β, which is able to be processed by caspases, might be an evolutionarily unique phenomenon in Drosophila.
Even if the Sgg46 isoform is unique, similar forms of caspase‐dependent regulation may occur in other signaling pathways. The tumor suppressor APC (adenomatous polyposis coli) is implicated in the Wingless/Wnt signaling pathway that is involved in early embryonic development and tumorigenesis in vertebrates. Browne et al (1998) and Webb et al (1999) reported that mammalian APC is processed to a 90‐kDa amino‐terminal fragment by caspase‐3. This cleavage at the DNID site of APC may be biologically relevant, given that the 90‐kDa fragment consists of a sequence that is highly conserved in the human, rat, mouse, Xenopus, and Drosophila APC (Webb et al, 1999). Further circumstantial evidence for caspase‐mediated regulation of the Wnt pathway comes from the observation that caspases also induce the cleavage of Arm/β‐catenin in mammals, a protein known to accumulate in cells depleted of functional APC. This regulation appears to link cell–cell signaling to changes in transcription and cell movement (Brancolini et al, 1997). Although the precise functions of the cleaved APC and Arm/β‐catenin products remain unknown, these findings strongly suggest a conservation of the caspase‐dependent regulation of Wingless/Wnt signaling between Drosophila and mammalian systems. Considering the number of caspases present in the different systems (three in C. elegans, seven in Drosophila, and 14 in mammals) (Baehrecke, 2002), a broad and manifold evolution of caspase families may allow for a diversity of specific substrates with various nonapoptotic roles.
Caspase‐dependent modification of kinase activity
The constitutive activation of kinases can lead to catastrophic effects in a cell. Here, we clearly showed that full‐length Sgg46 is an inactive form that can be converted into an active kinase via caspase‐dependent cleavage. While some substrates are functionally inactivated upon caspase‐mediated cleavage, other proteins and enzymes can be activated, mostly via the removal of an inhibitory or regulatory domain. The physiological consequence of this gain‐of‐function cleavage in apoptosis remains unclear. Several members of the PKC family and MAP kinase pathway are constitutively activated by the separation of N‐terminal regulatory from C‐terminal catalytic domains (e.g. p21‐activated kinase PAK2 and ROCK‐1), and this activation is important for cytoskeletal reorganization and plasma membrane blebbing (Rudel and Bokoch 1997; Coleman et al, 2001; Sebbagh et al, 2001). Interestingly, a DEVD motif‐specific caspase that cleaves MEKK‐1 is specifically activated when cells lose matrix contact and this cleavage is required for the MEKK‐1 kinase activity (Cardone et al, 1997). In another example, MST1, mammalian STE20‐like kinase 1, is cleaved and activated by caspases during apoptosis (Graves et al, 1998, Lee et al, 1998; Ura et al, 2001). This raises the possibility that the N‐terminal portion of Sgg46 might inhibit its own kinase domain, thereby imposing an additional form of regulation on its kinase activity.
Cellular differentiation via caspase activation
Little is known about the proteins cleaved by caspases during the differentiation process, and only a limited number of distinct substrates have been identified. For instance, in erythroblasts, PARP, lamin B, and acinus are cleaved, while GATA‐1, a transcription factor essential for erythrocyte formation, and ICAD, remains intact. Interestingly, MST1 kinase was identified as a crucial caspase‐3 effector in myoblast differentiation (Fernando et al, 2002): the cleaved and activated MST1 enhances the activity of downstream MAP kinases to promote skeletal muscle differentiation. In this report, we observed caspase activation in clusters of neural precursor cells that was able to induce slight but significant changes in SOP cell differentiation through the cleavage of Sgg46 rather than by cell death. It is still unclear why Dark‐dependent caspase activation in the scabrous‐expressing regions of wing discs does not induce cell death. Similar observations of caspase‐dependent cellular differentiation have been made, for example, in the lens, keratinocytes, erythrocytes, and sperm (De Maria et al, 1999; Weil et al, 1999; Lee et al, 2001; Arama et al, 2003; Huh et al, 2004b). In these cases, caspase activation does not lead to any apoptotic or cell death‐related events. Furthermore, differentiation‐related caspase activation must be tightly regulated to prevent cells from dying through apoptosis. During cellular differentiation, caspase activation is apparently very limited, transient, or localized. For instance, during megakaryocyte differentiation, the limited caspase activation is confined to dot‐like structures (De Botton et al, 2002). When senescent megakaryocytes die, however, the site of caspase activation changes from a localized to a diffused and greatly enlarged cytosolic area. Thus, restricting or lowering the levels of caspase activation in specific cellular spaces might enable Sgg46 processing without concurrent cell death.
We showed that the caspase family antagonizes the Wingless pathway during macrochaete development via Sgg in a kinase‐dependent manner. This process is achieved through the caspase‐dependent cleavage of Sgg46, which converts it to an active kinase. Our study clearly indicates an apoptosis‐independent mechanism in which caspase activation modulates a signaling pathway via substrate cleavage. Such caspase cleavage of intracellular target proteins may strongly depend on the cellular context, including the differentiation status. Clearly, much remains to be learned about the potential dual role of caspases in apoptosis and cellular differentiation. Characterization of the molecules that regulate this limited caspase activation and the relevant substrates will provide exciting new insights into processes that, beyond cell death, might link caspase cleavage to important nonapoptotic biological processes.
Materials and methods
Flies were raised on standard Drosophila medium at 25°C. The fly strains used in this study were described in Supplementary data. The ectopic expression of various kinds of Sgg proteins and sgg dsRNA in the fly was achieved using the GAL4/UAS system. All pUAST constructs were injected into w1118; Dr/TMS, Sb P[ry+, Δ2–3] embryos to produce transgenic flies as described (Kanuka et al, 1999b). At least three independent transformed lines were obtained for each transgenic construct. Drosophila crosses were carried out by standard procedures at 25°C.
Histology, immunohistochemistry, and TUNEL
Flies were prepared for scanning electron microscopy as described (Kanuka et al, 1999b). For light microscopic images of the adult notum, flies were anesthetized and examined with a Nikon SMZ1000 microscope (Nikon) equipped with an AxioCam digital camera (Carl Zeiss). Immunohistochemistry for wing discs and pupal nota were described previously (Tsuneizumi et al, 1997; Sato et al, 1999). TUNEL labeling of the wing discs was carried out as described (Hay et al, 1994) with some modifications; in the secondary antibody incubation step, the wing discs were also incubated with Cy2‐ or Cy3‐conjugated streptavidin (Jackson). All fluorescently labeled samples were examined with an LSM510 confocal microscope (Carl Zeiss). The antibodies used for immunostaining in this study were described in Supplementary data.
Confocal imaging analysis of caspase activity in vivo
Confocal imaging analysis of caspase activation was performed based on our previous report with several modifications (Takemoto et al, 2003). Briefly, wing discs dissected from third‐instar larvae were prefixed in 4% PFA at 4°C for 5 min, and then fixed in fresh 4% PFA at room temperature for 20 min. After several washes by PBS, confocal FRET images were acquired on the Aquacosmos/Ashura system (Hamamatsu Photonics) using an IX‐71 equipped with a UPlanApo 10 × 0.40 NA objective (Olympus), a spinning disk‐type confocal unit (CSU10, Yokogawa, Tokyo), a diode‐pumped solid‐state laser (430 nm) (Melles Griot), and a three CCD (charge‐coupled device) color camera (C7780‐22; Hamamatsu Photonics). Image acquisition and analysis were performed using Aquacosmos/Ashura software (Hamamatsu Photonics).
Cell culture, transfections, immunoblotting, and caspase activity
Drosophila S2 cells were cultured as described previously (Kanuka et al, 1999a). To generate stable transformants of the reaper expression plasmid, S2 cells were transfected with pCasepeR‐hs‐reaper and pAct5cp‐Neomycin, and stable S2 cell lines were selected and established as described previously (Kondo et al, 1997). For the detection of proteins, S2 cells were cultured in six‐well plates (5 × 105 cells/well) and transfected using CellFectin (GIBCO) with various amounts of the indicated plasmids plus pWAGAL4, which expresses the GAL4 protein under the control of the actin5C promoter. After transfection, the cells were lysed in SDS sample buffer. All samples were separated by 10% SDS–PAGE and subjected to immunoblotting. The antibodies used for immunoblotting in this study were described in Supplementary data. The signals were visualized using ECL plus (Amersham). Specific antibodies that recognize cleaved forms of Sgg46 and drICE were raised in rabbits using a synthetic peptide corresponding to the N‐terminus of Sgg46 (AKPKNR) and the C‐terminus of drICE (SQTETD) conjugated with keyhole limpet hemocyanin as the immunogen. These specific antibodies were purified by sequential protein affinity purification and used for immunoblotting (1:100 dilution). The assay for caspase activity in S2 cells was performed as described (Kanuka et al, 1999a).
Supplementary data are available at The EMBO Journal Online.
We are grateful to R Akai for technical support, M Sato, K Tsuneizumi, Y Hiromi, B Hay, S Yanagawa, P Simpson, M Bourouis, H Bellen, R Ueda, H Richardson, R Phillips, R Carthew, V Rodrigues, and H Steller for fly strains and materials, the Developmental Studies Hybridoma Bank for antibodies, and the Bloomington Stock Center for fly stocks. We are also grateful to B Nelson, M Sato, R Niwa, and K Nakao for valuable discussions. This work was supported in part by grants from the Japanese Ministry of Education, Science, Sports, Culture, and Technology to HK, EK, HO, and MM. This work was also supported in part to HO by Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Corporation, and to MM by a RIKEN Bioarchitect Research Grant, TAKEDA Science Foundation, and TORAY Science Foundation. EK and KT were a research fellow of the Junior Research Associate Program, RIKEN. HK was a research fellow of the Special Postdoctoral Researchers Program, RIKEN.
- Copyright © 2005 European Molecular Biology Organization