Two major apoptotic signaling pathways have been defined in mammals, the extrinsic pathway, initiated by ligation of death receptors, and the intrinsic pathway, triggered by cytochrome c release from mitochondria. Here, we identified and characterized the Xenopus homologs of caspase-10 (xCaspase-10β), a novel initiator caspase, and Bid (xBid), a BH3-only molecule of the Bcl-2 family involved in both the extrinsic and intrinsic pathways. Exogenous expression of these molecules induced apoptosis of mammalian cells. By biochemical and cytological analyses, we clarified that xCaspase-10β and xBid exhibit structural and functional similarities to their mammalian orthologues. We also detected xCaspase-10β and xBid transcripts during embryogenesis by whole-mount in situ hybridization and RT-PCR analysis. Microinjection of mRNA encoding a protease-defect xCaspase-10β mutant into embryos resulted in irregular development. Enforced expression of active xBid induced cell death in developing embryos. Using transgenic frogs established to allow monitoring of caspase activation in vivo, we confirmed that this form of cell death is caspase-dependent apoptosis. Thus, we demonstrated that the machinery governing the extrinsic and intrinsic apoptotic pathways are already established in Xenopus embryos. Additionally, we propose that the functions of the initiator caspase and BH3-only molecule are evolutionarily conserved in vertebrates, functioning during embryonic development.
Apoptosis, or programmed cell death, removes unnecessary, superfluous, damaged or harmful cells in multicellular organisms. Apoptosis is important for tissue morphogenesis during development, for the maintenance of homeostasis in adulthood, and for proper immune responses (Raff 1992; Abbas 1996; Jacobson et al. 1997; Vaux & Korsmeyer 1999; Meier et al. 2000). A family of cysteine proteases, known as caspases, is activated to function as executors of apoptosis or regulators of the inflammatory response (Thornberry & Lazebnik 1998). Of the 15 caspases that have been previously identified in mammals, the caspases involved in apoptosis are divided into two categories, the initiator caspases, including caspases-2, -8, -9 and -10, and the effector caspases such as caspases-3, -6 and -7. The activation of caspases represents the convergence of two major signaling cascades, the extrinsic pathway, which is initiated by ligation of cell surface “death receptors,” such as Fas (APO-1/CD95), and the intrinsic pathway, which is triggered by cytochrome c release from mitochondria into the cytosol.
Caspase-8 and caspase-10 are involved in the extrinsic signaling pathway (Chen & Wang 2002). Activation of death receptors by ligand binding leads their association with the adaptor molecule FADD (Fas-associated death domain protein, also called MORT1). FADD then recruits caspase-8 (also known as FLICE/MACH1/Mch5) and caspase-10 (FLICE2/Mch4) via interactions between death effector domains (DED) (Boldin et al. 1996; Muzio et al. 1996; Vincenz & Dixit 1997). In the death-inducing signaling complex (DISC), these caspases execute autoprocessing to become active (Kischkel et al. 1995). Activated initiator caspases are released from the complex and cleave downstream effector caspases, leading to cell death (Stennicke et al. 1998). In contrast, the intrinsic signaling pathway is triggered by a wide variety of stimuli, including UV irradiation, steroids, chemotherapeutic drugs and growth factor deprivation. Molecules of the Bcl-2 family are the prime players in this pathway (Gross et al. 1999a). Of the members of the Bcl-2 family, the BH-3-only molecules such as Bid (BH3-interacting domain death agonist) are pro-apoptotic molecules that induce the release of cytochrome c from mitochondria by interacting with other classical Bcl-2 family members (Huang & Strasser 2000). Bid, discovered as a molecule that binds to Bcl-2 and Bax (Wang et al. 1996), is a target of caspase-8 (Li et al. 1998; Luo et al. 1998; Gross et al. 1999b). This cleavage event links the extrinsic and intrinsic apoptotic signaling pathways, allowing the processed form of Bid to translocate to mitochondria to induce intrinsic apoptotic signaling.
In amphibians, degeneration of the tadpole tail during metamorphosis is accompanied by cell death (Kerr et al. 1974; Tata 1994; Nishikawa et al. 1998). Recent reports suggest that apoptosis also occurs during the early development of Xenopus embryos (Hensey & Gautier 1998; Hensey & Gautier 1999; Charrier et al. 1999). Although these apoptotic events occur during different stages of development, several lines of evidence suggest that all apoptotic signaling pathways converge to a common final pathway, requiring a number of evolutionarily conserved genes. Eight caspase types have been identified in Xenopus (Yaoita & Nakajima 1997; Nakajima et al. 2000), each of which exhibits homology to its mammalian counterpart. Expression of the majority of these caspases increases in regressing organs during metamorphosis (Nakajima et al. 2000). Caspase-3 and caspase-9 have been directly implicated in cell death of the tail (Das et al. 2002; Rowe et al. 2005). Caspase-2 works as an executioner of stockpiled oocytes (Nutt et al. 2005). In addition, Bax, a pro-apoptotic molecule of the Bcl-2 family, plays a critical role in the regression of the tail during metamorphosis (Sachs et al. 2004). We previously identified Xenopus FADD as a protein that functions similarly to mammalian FADD, affecting embryonic development following ectopic over-expression (Sakamaki et al. 2004). Thus, the machinery involved in apoptotic signaling appears to be conserved between mammals and Xenopus. Numerous components required for apoptosis, including death ligands and BH-3-only molecules, have not yet been characterized in Xenopus. Further investigation is necessary to elucidate the common mechanisms controlling programmed cell death in vertebrates.
To understand the universal mechanisms that regulate cell death in vertebrates, we examined the apoptosis-inducing molecules in Xenopus that are essential for the extrinsic and intrinsic signaling pathways. Using mammalian cells and Xenopus embryos, we identified and characterized a novel initiator caspase, xCaspase-10β, and a BH3-only molecule, xBid. Our study clearly demonstrates that these molecules have both structural and functional similarities to their mammalian homologs, confirming the evolutionary conservation in vertebrates of apoptotic signaling mechanisms. In this study, we also present a novel system for the detection of apoptosis associated with caspase activation in vivo.
Identification of the genes for Xenopus caspase-10β and Bid
In our search for DED-containing proteins in the GenBank expressed sequence tag (EST) database, we identified two promising cDNA clones (XL070f19 and XL061i12) from the Xenopus tailbud cDNA library. Further sequence analysis revealed that both of these cDNA clones represent a single open reading frame encoding a 512 amino acid protein with a higher degree of homology to Xenopus caspase-10 (84% identity at the amino acid level) than that of caspase-8 (32% identical) (Fig. 1A,B). Alignment of the novel gene with caspase-10 demonstrated the presence of several differing amino acid and one gap. In addition, these molecules exhibit distinct expression profiles during early embryonic development (see below). Therefore, we concluded that this cDNA clone encodes a novel caspase that has diverged evolutionarily from caspase-10 in Xenopus. We refer to this molecule as a second caspase-10 (caspase-10β).
By searching for a Xenopus orthologue of mammalian Bid, we identified one EST clone (PBX0108H11) in the Xenopus egg cDNA library. Sequence analysis clarified that this clone encodes a 184 amino acid protein with homology (36% identical) to human Bid (Fig. 1C). Within this sequence was a typical LETD amino acid sequence that can be cleaved by initiator caspases. Thus, this molecule is a Xenopus homolog of mammalian Bid, designated xBid.
Characterization of Xenopus caspase-10β and Bid
In Xenopus, two initiator caspases, caspase-8 (xCaspase-8) and caspase-10 (xCaspase-10), have been identified (Nakajima et al. 2000). These molecules are the functional equivalents of their mammalian homologs, inducing apoptosis in mammalian cells upon exogenous expression (Sakamaki et al. 2004). Therefore, we investigated whether the exogenous expression of Xenopus caspase-10β (xCaspase-10β) in mammalian cell lines could induce apoptosis. We constructed egfp-xCasp10β, a fusion of enhanced green fluorescent protein (EGFP) and the full-length xCaspase-10β (Fig. 2A), and transfected this gene into human HeLa cells. The cell viability of transfectants was examined by monitoring EGFP-positive cells by microscopy. Ectopic expression of EGFP-xCasp10β molecules clearly induced cell death in HeLa cells, as evidenced by reduced numbers of EGFP-positive cells and by the appearance of apoptotic bodies compared with control cells expressing EGFP alone (Fig. 2B,C). In the presence of a pan-caspase inhibitor carbobenzoyl-Val-Ala-Asp-fluoromethylketone (z-VAD-fmk), however, an increased number of EGFP-positive cells were detected, exhibiting a normal morphology in a similar manner as the control cells (Fig. 2D). This result suggests an inhibition of apoptotic cell death induced by exogenous xCaspase-10β. Aggregates termed death-effector filaments (Siegel et al. 1998) were observed in surviving cells expressing EGFP-xCasp10β in the presence of z-VAD-fmk (Fig. 2E), suggesting that apoptotic induction by xCaspase-10β requires homophilic oligomerization in a manner similar to the mammalian DED-containing molecules.
We constructed egfp-txBid, a fusion of EGFP and truncated xBid (txBid), which lacks the 57 N-terminal amino acids (Fig. 2A). We then examined the killing activity of this gene product in transfected HeLa cells. EGFP-positive transfectants could be visualized in the presence, but not the absence, of z-VAD-fmk (Fig. 2F,G). These data suggest that exogenous expression of txBid induces cell death in a caspase-dependent manner. To assess the ability of txBid to activate the intrinsic mitochondrion-dependent pathway, we investigated the killing activity of txBid in the presence of the mouse anti-apoptotic molecule Bcl-XL. After co-transfection of plasmids carrying the egfp-txBid gene and the mouse bcl-xL cDNA, we examined EGFP-positive cells after 24 h of culture. We detected a small number of EGFP-positive cells (Fig. 2H), suggesting that Bcl-XL can partially inhibit xBid-induced apoptosis. Bcl-XL, however, also potently inhibited Fas-induced apoptosis, mediated by the extrinsic pathway, in transfectants, leading to significant increases in the number of observed EGFP-positive cells (Fig. 2K,L). To characterize the txBid-induced apoptotic signaling pathway, we also examined caspase-8-deficient and caspase-9-deficient mouse embryonic fibroblasts (MEFs) (Kuida et al. 1998; Sakamaki et al. 2002). Introduction of the egfp-txBid gene into these cells revealed that txBid is able to induce cell death in caspase-8-deficient, but not caspase-9-deficient, cells (Fig. 2I,J). These results indicate that both xCaspase-10β and xBid exhibit killing activity in mammalian cells. A truncated form of xBid activated the intrinsic pathway in a caspase-9-dependent manner, although this activation could not be completely suppressed by Bcl-XL.
Translocation of xBid to the mitochondria during apoptosis
In mammals, Bid is processed by activated caspase-8; truncated Bid translocates to the mitochondria (Gross et al. 1999b). To address the localization of the processed form of xBid to mitochondria, we generated a plasmid, pxBid-EGFP, encoding a fusion of full-length xBid and EGFP (Fig. 2A). Following co-transfection of HeLa cells with pDsRed2-Mito, the majority of the transfectants did not die (Fig. 3A). Then, we induced apoptosis in the transfectants by treatment with an agonistic anti-Fas antibody, CH11. The xBid-EGFP fusion clearly translocated to mitochondria in dying cells, as evidenced by the overlap of EGFP fluorescence with that of DsRed, a mitochondrial marker (Fig. 3B). We also generated an xBid mutant, xBid(D52A), in which the LETD sequence, the putative caspase recognition site, is replaced by LETA (Fig. 2A). This mutant failed to translocate to mitochondria, despite the induction of apoptosis in transfectants (Fig. 3C,D). These results demonstrate that the processed form of xBid translocates to mitochondria in a similar manner as observed for mammalian Bid. Immunoblotting analysis confirmed the cleavage of xBid-EGFP, but not the mutant xBid(D52A)-EGFP, after the induction of apoptosis. The processed form of xBid was only detectable in extracts of HEK293Fas cells transfected with the wild-type, but not the mutant, xBid after treatment with an anti-Fas antibody (Fig. 3E). Thus, in a manner similar to the mammalian orthologues, processed xBid translocated to mitochondria upon apoptotic stimulation.
xCaspase-10β possesses protease activity and is able to cleave xBid at the specific site in vitro
To examine the protease activity of xCaspase-10β, we generated and purified recombinant xCaspase-10β and xCaspase-8 from E. coli. In an in vitro assay using a fluorescent substrate, recombinant xCaspase-10β exhibited a similar protease activity as recombinant xCaspase-8 (Fig. 4A). The large subunit of xCaspase-10β alone failed to cleave the substrate. Although the processing activity of caspase-8 for Bid has been recognized, it remains unclear if caspase-10 can process Bid. We next examined this possibility by using recombinant xCaspase-10β and xBid. We prepared both wild-type and mutant (D52A) xBid by the novel cell-free system (Sawasaki et al. 2002) and mixed them with recombinant xCaspase-10βin vitro. As shown in Fig. 4B, recombinant xCaspase-10β directly cleaved wild-type, but not mutant. In addition, xCaspase-10β protease activity was completely inhibited by the addition of z-VAD-fmk. These results suggest that xCaspase-10β has a potent protease activity, recognizing and directly cleaving xBid. Recently, human caspase-10 was reported to possess the potency to cleave Bid (Milhas et al. 2005), correlating well with our data.
xCaspase-10β specifically associates with Xenopus and mouse molecules by homophilic interactions of DEDs
We previously demonstrated that Xenopus caspase-10 interacts with the adaptor protein FADD (Sakamaki et al. 2004). In this study, we examined the association of xCaspase-10β with Xenopus and mouse FADD molecules. By immunoprecipitation assay, xCaspase-10β as well as mouse caspase-8 was detected in the complex with xFADD or mouse FADD upon co-expression in HEK293T cells, suggesting the interactions between xCaspase-10β and Xenopus and mouse adaptors (Fig. 5A). We further investigated whether xCaspase-10β forms homodimeric complexes or heterodimeric complexes with either xCaspase-10 or xCaspase-8 (Fig. 5B). xCaspase-10β clearly interacted with both xCaspase-10β and other xCaspase-8 and xCaspase-10 molecules. These data suggest that xCaspase-10β can interact with other Xenopus initiator caspases via the conformational activity of its DED.
The temporal and spatial expression of xCaspase-10β and xBid mRNA in Xenopus embryos and adult animals
To investigate the expression patterns of xCaspase-10β and xBid in Xenopus laevis, we detected their transcripts in a variety of adult and embryonic tissue by reverse transcription-polymerase chain reaction (RT-PCR). In previous studies, we demonstrated that caspase-10 is expressed in adult tissues (Sakamaki et al. 2004) and at very early stages of embryonic development (our unpublished data). To compare the expression patterns of xCaspase-10 and xCaspase-10β by RT-PCR, we designed new primers capable of distinguishing the PCR products derived from the two genes (see Fig. 1). Using this primer set, we could detect two distinct transcripts, representing xCaspase-10 and xCaspase-10β, in adult tissues, including the brain, heart, kidney, liver, lung, muscles, and spleen (Fig. 6A). The identities of these transcripts were confirmed by direct sequencing (our unpublished observation). In embryos, xCaspase-10 transcripts were expressed during stages 1-20, while xCaspase-10β increased beginning at stage 15/16, similar to the pattern seen for xCaspase-8 (Fig. 6B). Thus, the expression patterns of xCaspase-10β and xCaspase-10 transcripts were distinct during embryogenesis. In contrast, although the expression levels varied, we detected xBid transcripts in all samples of adults and embryos (Fig. 6A,B).
We examined the spatial patterns of xCaspase-10β expression in embryos by in situ hybridization. Whole-mount in situ hybridization using an xCaspase-10β cRNA probe demonstrated that xCaspase-10β transcripts were present ubiquitously in whole embryos during stages 8-32 (Supplementary Fig. S1A). As it is difficult to distinguish between xCaspase-10 and xCaspase-10β transcripts by hybridization, we hypothesize that the detected signal represents primarily xCaspase-10 mRNA at stages 8, 11 and 15 and xCaspase-10β mRNA at stages 22 and 32. We also examined the expression profile of xBid in embryos (Supplementary Fig. S1B). xBid mRNA was detected throughout the whole embryo at stage 12 and in the axial structure of embryos at stages 25 and 27.
Functional analysis of xCaspase-10β in embryonic development
To determine if xCaspase-10β is critical to embryogenesis, we examined the phenotype of embryos injected with an mRNA encoding a protease-deficient form of xCaspase-10β. This molecule is expected to act as a dominant-negative mutant by interacting with and sequestering endogenous xCaspase-10β. Indeed, HeLa cells expressing a protease-deficient xCaspase-10β, xCaspase-10β(C384S) exhibited resistance to Fas-mediated apoptotic stimulus compared with control cells, resulting in an increased number of EGFP-positive cells even in the presence of anti-Fas antibody (Fig. 7A). This result suggests that an xCaspase-10β(C384S) molecule is able to function as a dominant-negative molecule for initiator caspases. Following injection of the mRNA into four-cell-stage blastocysts, we analyzed the morphology of embryos after developing to stage 45. Injection into the equatorial areas of the two dorsal blastomeres resulted in embryos with shortened trunks in comparison with controls (Fig. 7B,a,c). We also observed edema in a number of embryos (Fig. 7B,b,c; shown by arrows). In contrast, abnormal embryos associated with scrape of the abdomen were observed after microinjection of the mRNA into the ventral area (Fig. 7B,d,e; shown by arrowheads). These phenotypes occurred in dose-dependent manner, but were not observed in embryos subjected to microinjection of control mRNA (our unpublished observation). These data (summarized in Fig. 7C) suggested that irregular development of embryos occurs at high efficiency following ectopic expression of protease-deficient xCaspase-10β. These results suggest that xCaspase-10β is involved in the early development of Xenopus embryos.
Ectopic expression of truncated xBid induces apoptosis in Xenopus embryos
To investigate if xBid could induce apoptosis in Xenopus embryos, we enforce expression of a truncated xBid (txBid) by microinjection of mRNA into 4-cell-stage blastocysts. We detected cell death in a region of the embryo at stage 8 (Fig. 8A). Embryos showed a hollow inside due to the loss of some cells. The area involved in the induced cell death is likely derived from the blastomere into which the txBid mRNA was injected. To confirm the pro-apoptotic activity of truncated xBid in embryos, we examined the activation of caspase-3 by txBid in embryos derived from a transgenic frog. We have established a novel transgenic Xenopus line expressing a fluorescent indicator, SCAT3, throughout the animal (Supplementary Fig. S2B). SCAT3 is a fusion of super enhanced cyan fluorescent protein (seCFP) and Venus (a variant of yellow fluorescent protein YFP) joined by a peptide linker (Supplementary Fig. S2A). This substrate is specifically cleaved at the DEVD recognition sequence within the linker by activated caspase-3 (Takemoto et al. 2003). The cleaved form of SCAT3 in embryos treated with a txBid mRNA injection during development could clearly be detected (Fig. 8B). The ratio of cleaved to intact SCAT3 significantly increased in embryos at stage 10 over the levels seen in embryos at stage 7. In contrast, no cleaved products of SCAT3 could be detected in healthy control embryos during stages 5–10. These results suggest that enforced expression of active xBid induces cell death in embryos in a caspase-3-dependent manner.
In this study, we identified a novel initiator caspase, xCaspase-10β, and a BH-3-only molecule, xBid, in frogs. These molecules are structurally and functionally similar to their mammalian counterparts, caspase-10 and Bid. Our results revealed that xCaspase-10β is a pro-apoptotic protease activated by interactions with DISC components. This caspase is also able to cleave xBid directly. xBid has a potency to coordinate both the extrinsic and intrinsic apoptotic signaling pathways and is able to induce apoptosis. We verified xCaspase-10β and xBid expression in embryos and adults by whole-mount in situ hybridization and RT-PCR analysis. Based on experiments utilizing a dominant-negative mutant, we demonstrated that xCaspase-10β plays multiple critical functions in embryonic development. Furthermore, using embryos derived from a novel transgenic line, we demonstrated the association of xBid cytotoxic activity with caspase-3 activation. Thus, these lines of evidence indicate that the apoptotic machinery requiring for xCaspase-10β and xBid is evolutionarily conserved in vertebrates, and strongly suggest novel functions of these molecules involving in embryonic development.
Both caspase-8 and caspase-10 have been identified in humans, while only caspase-8 has been identified in rodents (Varfolomeev et al. 1998; Grenet et al. 1999). In amphibians, functional homologs of both caspase-8 and caspase-10 have been identified (Nakajima et al. 2000; Sakamaki et al. 2004). In this study, we identified yet another functional initiator caspase, xCaspase-10β. By RT-PCR analysis, we found that xCaspase-10β and xCaspase-10 exhibit distinct expression patterns in the early stages of embryogenesis (Fig. 6B). xCaspase-10β increased markedly beginning in the neurula stage, while the expression of xCaspase-10 was maternally derived in eggs and decreased during the neurula stage. In Xenopus laevis, two forms of Smad4 (Smadα and Smadβ) have been identified (Masuyama et al. 1999). The exchange of Smadβ for Smadα expression during embryogenesis is similar to that observed between xCaspase-10 and xCaspase-10β. xCaspase-10 and xCaspase-10β are likely transcribed from two distinct genes. Physical map analysis of the caspase-8 and caspase-10 genes in several species revealed that the genomes of chicken (Gallus gallus) and West African clawed frog (Xenopus tropicalis) include an additional caspase-8/10-like gene between these two genes (our unpublished observation). xCaspase-10β and xCaspase-10 may correspond to the caspase-10 and caspase-8/10-like genes observed in these animals, respectively. It is also possible that these are paralogues in tetraploid animals. Further genomic analysis will be necessary to clarify the conservation and diversification of initiator caspases throughout evolution. Importantly, the expression pattern of xCaspase-10β coincided with that of xCaspase-8, but not xCaspase-10, suggesting synchronized expression in embryos. Ectopic expression of a protease-deficient xCaspase-10β mutant molecule in developing embryos disrupted organogenesis by inhibiting the function of the endogenous xCaspase-10β (Fig. 7). In these abnormal phenotypes, edema is thought to be caused by defect in the blood vessels (Bergers & Song 2005), indirectly suggesting the involvement of caspase-10β in the development of the cardiovascular system. Caspase-8-deficient mouse embryos display malformations of cardiogenesis and angiogenesis (Varfolomeev et al. 1998; Sakamaki et al. 2002). Thus, our study suggests the involvement of caspase-10/caspase-10β in the early embryonic development of animals. In humans, caspase-10 is unable to substitute for caspase-8 function in death receptor-mediated apoptotic signaling (Sprick et al. 2002). Defects in caspase-10 cause autoimmune disease (Wang et al. 1999), suggesting that caspase-8 is unable to compensate for this deficiency. In Xenopus, however, the possibility remains that the xCaspase-10β mutant influenced the activity of xCaspase-8 and xCaspase-10 in embryos through homophilic DED interactions. Consequently, defining the distinct and complimentary roles during development of amphibian xCaspase-10β, xCaspase-10, and xCaspase-8 will be important in understanding the physiological roles of these initiator caspases in vertebrates.
The previous report showed that mouse truncated Bid is also able to induce apoptosis in a caspase-independent manner and Bcl-XL sequesters killing activity of this molecule (Cheng et al. 2001). In this study, however, we obtained contradicting results using xBid (Fig. 2). Mammalian Bid can interact with Bax and Bcl-XL, as well as with lipids, such as cardiolipin. These additional interactions confer a supplementary role in the induction of cell death (Esposti 2002). It is likely that xBid either cannot interact or has only weak interactions with Bcl-XL and lipids, but strongly promotes caspase-9-dependent intrinsic apoptotic signaling in mammalian cells. Although the genes encoding the pro-apoptotic molecules Bak and Bok and the anti-apoptotic protein Bcl-XL, all members of the Bcl-2 family, have been identified within the genome of the ascidian, Ciona (Dehal et al. 2002), we could not find a homolog of bid, as reported by Bates (2004). This evidence supports the possibility that Bid first appeared evolutionarily in vertebrates. A database search identified Bid from several species of fish and other vertebrates, with conservation of both the BH-3 domain and the initiator caspase recognition sequence in all animals examined (Supplementary Fig. S3). Phosphorylation of mouse Bid Ser61 and Ser64 by casein kinase attenuates caspase-8-mediated cleavage (Desagher et al. 2001). These amino acid residues are unlikely to be conserved in a subset of vertebrate species including Xenopus. Thus, the existence of Bid from fish to mammals suggests a common physiological role in vertebrates associating with diversified regulations in some animals such as rodents.
Fluorescence analysis of fusions of xCaspase-10β and xBid with EGFP enabled us to observe the subcellular localization of these molecules upon apoptotic stimulation. Death-effector filament formation occurs in cells undergoing apoptosis (Siegel et al. 1998); following induction of apoptosis with an anti-Fas antibody, we also observed the generation of death-effector filaments by EGFP-xCaspase10β in those cells surviving due to inhibition of cell death by z-VAD-fmk treatment (Fig. 2E). This result suggests that xCaspase-10β oligomerizes upon encountering death-receptor-mediated signals. Furthermore, we observed the translocation of an xBid-EGFP fusion into mitochondria by fluorescence microscopy. A previous study reported that, in mammalian cells, fusions of human Bid with YFP and CFP at the N- and C-termini is cleaved upon Fas-mediated apoptotic signaling, measured by the loss of fluorescent resonance energy transfer (FRET) (Onuki et al. 2002). Thus, the fluorescent labeling of apoptotic molecules involved in the extrinsic pathway permits the monitoring of the molecular dynamics associated apoptosis. In this study, we established a novel transgenic frog line that is useful for monitoring the activation of effector caspases in vivo. This transgenic line expresses the SCAT3 hybrid molecule, a fusion of CFP and Venus joined by a peptide linker, in the whole bodies of adults (Supplementary Fig. S2B) and in embryos at various developmental stages (our unpublished observation). This system enables caspase activation to be monitored in vivo throughout an animal's lifetime. In mammalian cell lines, as SCAT3 is specifically cleaved by activated caspase-3 and results in the loss of FRET, it enabled to detect cascapse-3 activation by measuring the dynamics of fluorescence (Takemoto et al. 2003). We demonstrated that enforced apoptosis induced SCAT3 cleavage, which was associated with cell death during development. We also examined the dynamics of FRET in the degenerating tail of transgenic animals. Detection of a ratio change in FRET intensity suggested cleavage of SCAT3 by caspase-3 (our unpublished observation). As the activation of caspase-3 in the degenerating tail has already been established (Das et al. 2002), our transgenic line may serve as a method to image caspase activation by monitoring changes in FRET in living animals. This system proves to be an extremely useful methodology for the biomonitoring of apoptosis in living organisms.
Oligomerization via interactions of DEDs is essential for the activation of initiator caspases. In this study, we observed the interactions of xCaspase-10β with both xCaspase-10 and xCaspase-8 via homophilic interactions of their DEDs (Fig. 5). Enforced dimerization of caspase-8 and caspase-10 by addition of an FKBP-binding domain fails to activate these proteases (Chen et al. 2002). As we have not yet examined the catalytic activity of complexes of initiator caspases in Xenopus, it is difficult to confirm the physiological roles of heterologous complex of xCaspase-10β, xCaspase-10, and xCaspase-8. It was recently reported that the DED-containing protein cFLIP interacts with and modulates caspase-8 activation (Micheau et al. 2002). The resultant heterodimer can cleave distinct substrates, such as RIP, implicating the existence of interactions between xCaspase-10β, xCaspase-10, and xCaspase-8 with differential activities and substrate specificities in vivo.
The Xenopus system is a powerful tool for the dissection of both the basic mechanisms and the complexities of apoptosis, especially those occurring during development. Our studies clarified that the components required for the extrinsic apoptotic signaling pathway, FADD and Bid, affect embryonic development. FADD and Bid, however, exhibit differential killing activities, despite similar expression of their transcripts at early stages of development. Exogenous expression of truncated xBid in embryos induced apoptosis at stage 7 by activating the intrinsic apoptotic pathway (Fig. 8). Over-expression of FADD, however, induced cell death at later stages (approximately stage 18) (our unpublished observation). As xCaspase-10β and xCaspase-8, but not xCaspase-10, were transcribed in stages 15-20 (Fig. 6B), these molecules may be required for the activity of FADD in vivo. The expression of both xCaspase-10 (and/or xCaspase-10β) and FADD is up-regulated in the degenerating tail during metamorphosis (Rowe et al. 2005). The mitochondria-dependent intrinsic apoptotic pathway, which is triggered by Bid, is involved in the removal of abnormally developing embryos during the early gastrula transition (stage 10) (Van Stry et al. 2004). In Xenopus, caspase-2 functions in the clearance of metabolically inactivated oocytes/eggs (Nutt et al. 2005). As Bid is essential for caspase-2-induced apoptosis (Gao et al. 2005), Bid may function as a downstream activator in this event. Our study also supports a role for Bid in these processes. Thus, the apoptosis-inducing machinery is already established early in embryogenesis. Consequently, the extrinsic and intrinsic apoptotic pathways may play distinct physiological roles during development.
Animals, embryos and cell lines
Adult wild-type Xenopus laevis were purchased from Hamamatsu Seibutsukyozai Co. (Shizuoka, Japan). In vitro fertilization of Xenopus eggs was performed as described (Suzuki et al. 1994). Fertilized embryos were dejellied in 3% cysteine hydrochloride, washed several times in water, and used for mRNA microinjection, whole-mount in situ hybridization, and RT-PCR analysis. Developing embryo stages were determined according to the scheme described by Nieuwkoop & Faber (1967). Mouse embryonic fibroblasts (MEF cells) isolated from the wild-type, caspase-8-deficient (Sakamaki et al. 2002), and caspase-9-deficient (Kuida et al. 1998) embryos, human cervical carcinoma HeLa-S3 (HeLa) cells, and two variants of human embryonal kidney 293 (HEK293) cells, HEK293Fas and HEK293T cells, that express high levels of Fas and large T antigen, respectively, were cultured in Dulbecco's Modified Eagle's medium supplemented with 10% fetal calf serum.
Database searches and DNA sequencing
To identify Xenopus homologs of DED-containing molecules and the BH3-only molecule Bid in the GenBank database, we utilized Blast software (Altschul et al. 1990). Two Xenopus EST clones (GenBank accession nos. BJ063173 and BJ0590094) were identified using an amino acid sequence corresponding to the DED of human FADD as a probe. A single EST clone (GenBank accession no. AW641666) was homologous to human Bid. The nucleotide sequences of both strands of these clones were confirmed by DyeDeoxyterminator Cycle sequencing (Applied Biosystems Inc., Foster City, CA, USA) using automated DNA sequencers (PRISM™ 310, Applied Biosystems Inc. and LI-COR 4000, LI-COR Biosciences, Lincoln, NE, USA). The complete sequences of these Xenopus EST clones have been deposited as the GenBank accession nos. AY519262 and AY518731, respectively. We compared the aligned amino acid sequences of the Xenopus and human molecules using CLUSTAL W software (Higgins & Sharp 1988).
Sequence alignment of Bid
To align the vertebrate Bid sequences, we cited the amino acid sequences previously published for human (Homo sapiens), mouse (Mus musculus), rat (Rattus norvegicus), and chicken (Gallus gallus) Bid (GenBank accession nos. NP_001187, NP_031570, NP_073175, and NP_989883, respectively). We also predicted the Bid amino acid sequences of macaque monkey (Macaca mulatta) (GenBank accession nos. CB554325 and CB554668), cows (Bos taurus) (GenBank accession no. CK831238), pigs (Sus scrofa) (GenBank accession no. CK460098), salamanders (Math Ambystoma mexicanum) (GenBank accession no. CN036937), frogs (Xenopus tropicalis) (GenBank accession no. AL774229), and five fish species, including Danio rerio (GenBank accession nos. CO359216 and AL924614), Ictalurus punctatus (GenBank accession no. CB937949), Oryzias latipes (GenBank accession no. BJ707272), Oncorhynchus mykiss (GenBank accession no. CA350756), and Salmo salar (GenBank accession no. CB513070) from EST databases. To compare the aligned amino acid sequences, we utilized CLUSTAL W software (Higgins & Sharp 1988).
Construction and transfection of expression vectors
To express a fusion protein of Xenopus caspase-10β (xCaspase-10β) and EGFP in mammalian cell lines, the DNA fragment containing the PCR-amplified caspase-10β cDNA was inserted into pEGFP-C1 (BD-Clontech, Palo Alto, CA, USA). To express Flag-tagged xCaspase-10β, caspase-10β cDNA was cloned into pME18S or pCAGGS (Niwa et al. 1991; Sakamaki et al. 1992) with the Flag-tag DNA sequence. A protease-deficient xCaspase-10β mutant, xCaspase-10β(C384S), was generated by replacing the cysteine residue (amino acid 384) in the QACQG active site of the protease region with serine by exchange of the PCR-amplified DNA fragment containing the mutant sequence. The resulting mutant xCaspase-10β(C384S) cDNA was then inserted into the pCSII vector. To express a fusion of the truncated form of Xenopus Bid (xBid) with EGFP, we cloned the BglII-SalI DNA fragment of Xenopus bid cDNA into pEGFP-C1. This truncated form of xBid (txBid) was also inserted into the pCSII vector. To monitor mitochondrial translocation and processing of xBid, we generated a plasmid construct, pxBid-EGFP, by inserting the Xenopus bid-coding region into pEGFP-N1 (BD-Clontech). A mutant form of xBid, xBid(D52A), was generated by replacing the putative site (IETD) recognized by initiator caspases, with the sequence IETA and subcloned into pEGFP-N1 to create pxBid(D52A)-EGFP. The GFP-mCasp8 DNA fragment, encoding a mouse caspase-8 cDNA (Sakamaki et al. 1998) fusion with the gfp gene, the Flag-tagged mouse Fadd cDNA (Hsu et al. 1996), the kind gift of Dr D. Goeddel (Amgen), and the mouse bcl-xL cDNA were cloned into pME18S. The pCX-p35 plasmid, carrying the baculovirus p35 gene, has been described (Sakamaki et al. 2004).
Transfection of the constructed plasmid DNAs into cells was performed using LipofectAMINE PLUS Reagent (Invitrogen Corp., Carlsbad, CA, USA), according to the manufacturer's protocol.
Construction of the transgene and generation of transgenic frogs
To construct a plasmid carrying the CAG-SCAT3 transgene (Takemoto et al. 2003), we inserted the DNA fragment encoding a hybrid protein composed of seCFP and Venus joined by a linker peptide followed by the rabbit β-globin polyadenylation (polyA) signal sequence into the pCAGGS expression vector (Niwa et al. 1991). SCAT3 is a fluorescent indicator allowing the detection of cell death. This substrate is cleaved by activated caspase-3 in mammalian cells. To generate transgenic frogs expressing SCAT3, we utilized the established method (Kroll & Amaya 1996; Amaya & Kroll 1999), with the modifications detailed previously (Sakamaki et al. 2004). We established a transgenic line expressing SCAT3 throughout the whole body of adult animals and confirmed that SCAT3 is also expressed in eggs (our unpublished data). In this study, we used F1 generation females from the established transgenic frog lines for experiments.
Pro-apoptotic activity assays of xCaspase10β and xBid in cell cultures
Cells were transfected either with plasmids carrying EGFP-xCaspase10β or EGFP-txBid in the absence or presence of 100 µm carbobenzoyl-Val-Ala-Asp-fluoromethylketone (z-VAD-fmk) (Kamiya Biomedical Co., Seattle, WA, USA). A subset of cells was transfected with txBid and a plasmid encoding mouse Bcl-XL. Twenty-four hours after transfection, cells were fixed in PBS containing 3.7% formaldehyde. The number of EGFP-positive cells was assessed in photographs taken under the microscope (DMIRE2, Leica Microsystems, Wetzlar, Germany). MEF cells isolated from caspase-8-deficient and caspase-9-deficient mice were transfected with EGFP-txBid, cultured for 24 h, and examined by microscopy. To examine the ability of a protease-deficient xCaspase-10β mutant, xCaspase-10β(C384S) as a dominant-negative molecule, HeLa cells were co-transfected with pCSII-xCaspase-10β(C384S) and pEGFP-C1. After a 24 h culture period, cells were treated for an additional 8 h with 200 ng/mL of an agonistic anti-human Fas antibody, CH11 (MBL, Nagoya, Japan) in the presence of 5 µg/mL cycloheximide (CHX). Similarly, to examine the anti-apoptotic activity of Bcl-XL in HeLa cells, cells were co-transfected with Bcl-XL and pEGFP-C1. After a 24 h culture period, cells were treated with 200 ng/mL of CH11 and 5 µg/mL of CHX for 24 h.
Cytological analysis by confocal laser scanning microscopy
After co-transfection of HeLa cells with pDesRed2-Mito (BD-Clontech) with either pxBid-EGFP or pxBid(D52A)-EGFP, apoptosis was induced with 200 ng/mL of an anti-Fas antibody and 5 µg/mL of CHX. The subcellular localization of xBid was examined by confocal laser scanning microscopy (TCS SP2, Leica Microsystems).
Synthesis of active xCaspase-10β recombinant protein and protease activity assay in vitro
To generate recombinant xCasp10β and xCasp8, DNA sequences encoding the protease domains were amplified by PCR. The resulting fragments were then cloned into the pMAL-c2E plasmid (New England Biolabs, Beverly, MA, USA). As a negative control, we also generated a construct encoding xCasp10β(p21), which only contained the large, catalytically inactive subunit. After transformation of E. coli Rosetta (DE3) (Novagen, Madison, WI, USA) with the appropriate constructs, maltose-binding protein (MBP)-tagged recombinants were produced following induction of protein expression with IPTG. Recombinant proteins were purified with amylose resin (New England Biolabs), according to the manufacturer's instructions. To detect protease activity, assays examined cleavage of an acetyl-Ile-Glu-Thr-Asp-α-(4-ethylcoumaryl-7-amide) (Ac-IETD-MCA) substrate (50 µm) (Peptide Institute Inc., Osaka, Japan) in the presence of 1 µg recombinant enzyme in 200 µL of protease assay buffer (50 mm PIPES (pH 7.2), 10% sucrose, 100 mm NaCl, 1 mm EDTA, 0.1% CHAPS, and 10 mm dithiothreitol) at 37 °C. The production of 7-amino-4-methylcoumarin (AMC) from the substrate was measured at the indicated times in a plate reader (ARVO-SX, PerkinElmer Life Science-Wallac Oy, Turku, Finland) using excitation and emission wavelengths of 355 nm and 460 nm, respectively. All values were assessed in triplicate. To insure for any effect of the MBP modification, we also produced a recombinant xCasp10β-His, containing only the protease domain fused to His-tags at both termini, in the transformed DE3 cells. Recombinant proteins were purified using nickel-NTA column chromatography according to the manufacturer's recommendations (Qiagen, Valencia, CA, USA).
To analyze the processing activity of xCasp10β protein against xBid in vitro, we synthesized both wild-type and mutant forms of xBid fused to EGFP in a cell-free system as described (Sawasaki et al. 2002). Briefly, xBid-EGFP or xBid(D52A)-EGFP were transcribed from the SP6 promoter of the pEU expression vector; the resulting transcripts were translated in vitro within wheat embryo extracts. An in vitro cleavage assay mixed active xCasp10β-His with xBid-EGFP or xBid(D52A)-EGFP in conjunction without or with 200 µm z-VAD-fmk in the protease assay buffer at 37 °C for 3 h. Following resolution by SDS-PAGE, the processed products were detected with an imaging analyzer (LAS-3000, FUJIFILM, Tokyo, Japan). For estimation of the molecular size of products, FITC-labeled molecular weight markers (SP-0130, APRO Life Science Institute, Tokushima, Japan) were used.
Immunoblotting and immunoprecipitation
To examine xBid cleavage following apoptotic stimulation, xBid and the xBid(D52A) mutant were transiently transfected into HEK293Fas cells. Two days later, transfectants were incubated in the presence or absence of 200 ng/mL anti-Fas antibody for 6 h. Samples were then suspended in lysis buffer containing 50 mm Tris-HCl (pH 7.5), 150 mm NaCl, 5 mm EDTA, 1% TritonX-100, and 0.5% sodium deoxycholate. Cell extracts were resolved by SDS-PAGE and examined by immunoblotting with an anti-GFP antibody (IE4) (MBL) as described (Sakamaki et al. 2004).
To analyze the interactions between xCaspase-10β and Xenopus and human apoptotic signaling molecules, HEK293T cells were transiently transfected with an expression vector encoding either EGFP-xCaspase10β or GFP-mouse caspase-8 without or with Flag-xFADD or Flag-mouse FADD. After 24 h of cultivation, transfectants were suspended in lysis buffer; proteins were immunoprecipitated from cell lysates with an anti-Flag antibody (M2) (Sigma Chemicals Co., St. Louis, MO, USA) and Protein-A Sepharose (Amersham Biosciences, Arlington Heights, IL, USA). Immunoprecipitates were analyzed by SDS-PAGE and immunoblotting with anti-Flag and anti-GFP antibodies. To examine the formation of homodimeric complexes of xCaspase10β or heterodimeric complexes of xCaspase-10β with xCaspase-10 or xCaspase-8, we co-transfected a plasmid encoding EGFP-xCaspase10β into HEK293T cells without or with plasmids encoding Flag-xCaspase10β, Myc-xCaspase-10, or HA-xCaspase-8 (Sakamaki et al. 2004). After 24 h cultivation period, transfectants were suspended in lysis buffer. Proteins were immunoprecipitated from cell lysates with an anti-GFP antibody. Immunoprecipitates were then analyzed by SDS-PAGE and immunoblotting with anti-GFP, anti-Flag, anti-Myc (9E10) (ATCC, Manassas, VA, USA), and anti-HA (ab-hatag) (InvivoGen, San Diego, CA, USA) antibodies. Bound antibody was visualized with Mouse IgG TrueBlot™ (eBioscience, San Diego, CA, USA) or horseradish peroxidase-conjugated anti-mouse IgG antibody (Cell Signaling Technology Inc., Beverly, MA, USA). For preventing apoptosis of transfectants, the p35 gene derived from baculovirus, which is well known as a strong inhibitor of caspases (Zhou et al. 1998), was introduced into all examined samples with an expression vector (Sakamaki et al. 2004).
Total RNAs were isolated from several tissues of adult animals and from embryos at various stages with ISOGEN reagent (Nippongene, Toyama, Japan), according to the manufacturer's instructions. First strand cDNAs were synthesized from each RNA sample (1 µg) using random hexameric primers on Ready-To-Go RT-PCR beads (Pharmacia), according to the manufacturer's protocol. These cDNAs were amplified with several sets of primers to detect both xCaspase-10β and xCaspase-10 (forward: 5′-CAAGACCTTTCTGGATGTGTTGTG-3′, reverse: 5′-TTTGCTTGAAACCTGGATGGAG-3′), xCaspase-8 (forward: 5′-CGCTTCTATACTGGAAATATTCTT-3′, reverse: 5′-TTGTACTGAAATCTTCTCAAAT-3′), and xBid (forward: 5′-GGAAACGTCCAATTAAGATCT-3′, reverse: 5′-CCTCTGTCTGGCGAGACGCTC-3′). As positive controls in embryos and adult tissues, we examined transcription of EF1α and ornithine decarboxylase using two sets of primers (forward: 5′-CAGATTGGTGCTGGATATGC-3′, reverse: 5′-ACTGCCTTGATGACTCCTAG-3′) and (forward: 5′-CAGCTAGCTGTGGTGTGG-3′, reverse: 5′-CAACATGGAAACTCACACC-3′), respectively. PCR was performed as previously described (Sakamaki et al. 2004). Amplified PCR products were analyzed by 2.5% agarose gel electrophoresis.
Microinjection of xCaspase-10β(C384S) and txBid mRNAs
Capped mRNAs encoding the protease-deficient xCaspase-10β(C384S) mutant or the truncated xBid were synthesized from pCSII plasmid construct templates using an mMESSAGE mMACHINE SP6 kit (Ambion, Austin, TX, USA). mRNAs were purified using a Sephadex G-50 column (Amersham Biosciences) as described (Yamamoto et al. 2001). Synthetic mRNAs were microinjected at either 20, 50 or 200 pg into the equatorial areas of two dorsal or two ventral blastomeres at the four-cell-stage. The injected embryos were cultured in 3% Ficoll and 0.1× Steinberg's solution (Yamamoto et al. 2001). Photographs were acquired at the appropriate stages of development.
Fluorescent analyses of cell extracts of transgenic animals
After mRNA microinjection, SCAT3-expressing embryos developed to the indicated stages and were lyzed in lysis buffer in the presence of a protease inhibitor cocktail (Nacalai tesque, Kyoto). Whole cell lysates were then obtained by centrifugation. After resolution of the extracts by SDS-PAGE without heat denaturation, the fluorescence was analyzed using an imaging analyzer (Typhoon 9410, Amersham Biosciences). Then, gels were stained with 0.2% Coomassie Brilliant Blue-R250, 40% methanol, and 10% acetic acid and destained in a mixture of 40% methanol and 10% acetic acid.
Whole-mount in situ hybridization
Embryos were fixed in 4% paraformaldehyde in 0.1 m potassium phosphate buffer (pH 7.8), washed with PBS. Digoxigenin-labeled RNA probes were prepared according to the manufacturer's recommendations (Roche Diagnostics, Mannheim, Germany). Whole-mount in situ hybridization of various stages of embryos was performed as previously described (Harland 1991). Hybridized RNAs were detected with alkaline-phosphatase-conjugated anti-DIG antibody (Roche) and developed using BM purple (Roche).
The authors are grateful to Drs D. Goeddel (Amgen), Y. Tsujimoto (Osaka University) and M. Miura (University of Tokyo) for providing mouse FADD and Bcl-XL cDNAs and a SCAT3 construct, and to Drs T.S. Yamamoto (National Institute for Basic Biology) and Y. Satou and S. Yonehara (Kyoto University) for their technical assistance and advice and experimental support. A part of this work was supported by a Grant-in-Aid for Creative Scientific Research to N.M and K.S. from the Japan Society for the Promotion of Science (13GS0008).