Adenine Nucleotide Translocator Cooperates with Core Cell Death Machinery To Promote Apoptosis in Caenorhabditis elegans^{▿ ‡}

C. elegans strains and genetics.

C. elegans strains were provided by C. elegans genetic center (CGC) and worms were cultured and maintained by using standard procedures (5). The Bristol N2 strain was used as wild type. The deletion strains used in the present study are the wan-1(gk172), wan-2(gk294), wan-3(gk300), and wan-4(ok2209) strains. The wan-1(gk172) strain contains a deletion of 604 bp that removes most part of exon 1, the whole exons 2 and 3, and part of exon 4; the wan-2(gk294) strain carries a deletion region of 1,249 bp that removes the whole open reading frame and 324 bp of the promoter region of the wan-2 gene; the wan-3(gk300) strain contains a deletion of 471 bp that removes part of exons 1 and 2 and likely results in an early stop codon in the transcript; and the wan-4(ok2209) strain contains a deletion region of 892 bp including 646 bp of the promoter region and 246 bp of exons 1 and 2. Deletion mutants were outcrossed with N2 strain for at least four times. Double mutants were constructed by using standard protocol (5). Germ line transformation experiments were performed essentially as described previously (25).

Immunoprecipitation and MALDI-TOF analysis.

The yqIs1 transgenic strain expressing P_hsp ced-4flag was grown with liquid culture at 20°C. To induce the expression of CED-4Flag protein, worms were heat shocked at 33°C for 1 h and continued to grow at 20°C for another 3 h. Worms were then collected and broken in liquid nitrogen, and proteins were extracted in a lysis buffer (25 mM Tris [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, and 10% glycerol) to yield whole-worm lysate. For immunoprecipitation, whole-worm lysate was incubated with agarose beads conjugated with anti-Flag antibody (M2; Sigma) overnight at 4°C. Beads were extensively washed, and bound proteins were resolved on 12% sodium dodecyl sulfate (SDS) polyacrylamide gel and visualized with silver or Coomassie blue staining. Proteins of interest were subjected to matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) analysis. Briefly, gel slices were treated with 100 mM NH₄HCO₃ to remove Coomassie blue and dehydrated with 50% acetonitrile in 50 mM NH₄HCO₃. Gel pieces were then sequentially treated with 5 mM dithiothreitol for reduction and 0.5 M iodoacetamide for alkylation. After appropriate washing and dehydration, gel slices were soaked overnight in protein digestion buffer (0.02 μg of trypsin/μl in 25 mM NH₄HCO₃ [pH 8.0]) at 37°C. Reactions were quenched with 88% formic acid and sonicated to release protein peptides. The supernatant was further cleaned with zip-tip and applied to AutoFlex (Bruker) for mass spectrometric analysis.

RNAi.

The 3′-untranslated region (3′UTR) of wan-1 (125 bp) was in vitro synthesized into double-stranded RNA (dsRNA) and injected into gonads of young adult worms. Surviving progeny that developed normally were scored for embryonic cell corpses 48 h after injection. dsRNA of green fluorescent protein (GFP) was injected as control. Animals developed normally to L4 stage were scored for extra cells in the anterior pharynx. To examine wan-1 RNA interference (RNAi)-caused embryonic lethality, the dsRNA synthesized from the 3′UTR of wan-1 or the cDNA of GFP were injected as described above. At 48 h after injection, eggs were transferred to fresh plates, and hatched animals were counted 12 h later. To examine the wan-1 RNAi effect on germ line apoptosis, two approaches were used. First, a bacterial feeding assay was performed as described previously (36). Briefly, worms synchronized to L3 stage were fed with bacteria expressing either control dsRNA or full-length wan-1 dsRNA, and germ cell corpses were scored at different adult ages of the P₀ worms. Second, dsRNAs synthesized from either wan-1 3′UTR or GFP cDNA were injected into the body cavity of L4-stage animals as described by Mello and Fire (24), and germ cell corpses were scored at different time points after injection.

Quantification of cell corpses and extra cells.

Cell corpses and extra cells were scored by using Nomarski optics. For somatic cell corpses, only those in the head region of embryos were scored. Extra cells in the anterior pharyngeal region of L4 animals were scored. To quantify germ cell corpses, cell corpses in the germ line meiotic region of one gonad arm were scored 12, 24, 36, 48, and 60 h after the L4 stage.

Chemical treatment and measurement of ATP level.

Synchronized young adult worms were incubated in Tris buffer (10 mM Tris [pH 7.5]) containing different concentration of Bongkrekic acid (BA; Sigma) or atractyloside potassium (Atr; Sigma) at 20°C for 14 h with supplement of OP50 bacteria. Chemicals were removed by transferring worms onto nematode growth medium plates. Germ cell corpses were scored at indicated time points after the removal of chemicals. To measure ATP levels, 500 animals treated with above chemicals were collected in 100 μl of buffer (100 mM Tris-HCl [pH 7.5], 40 mM EDTA). Worms were flash-frozen in liquid nitrogen and further boiled for 15 min, the resulting supernatant was collected, and the ATP concentration was measured by using an ENLITEN ATP assay kit (Promega). The luminescence was read on the Fluoroskan Ascent FL microplate reader (Thermo), and measurements were performed in triplicate.

Four-dimensional analysis of cell corpse duration.

The recording of both embryonic and germ line cell death events was performed as previously described (38).

Mammalian cell transfection and immunoprecipitation.

HEK293 cells were cultured in Dulbecco modified Eagle medium (HyClone) supplemented with 10% fetal bovine serum (HyClone), and transfections were performed by using the calcium phosphate method or Lipofectamine 2000 (Invitrogen) according to the supplier's instructions. Cells were harvested 36 h posttransfection and lysed in the lysis buffer (50 mM Tris [pH 8.0], 150 mM NaCl, 0.5% sodium deoxycholate, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride). Cell lysate was incubated with the indicated antibodies and protein A/G-Sepharose beads (Roche) for more than 2 h, and the beads were extensively washed. Bound proteins were resolved on SDS-polyacrylamide gel and detected by Western blotting. Immunoprecipitation with purified mitochondria was similarly performed. For immunostaining, transfected cells were fixed with 3.7% formaldehyde and stained with anti-Flag antibody.

Protein expression and GST pull-down.

Recombinant GST-CED-4 and GST-CED-9 proteins were expressed in bacterial BL21(DE3) cells and purified with glutathione-Sepharose beads (Amersham) according to the instructions provided by the supplier. GST pull-down was performed as described previously (36). Briefly, purified GST or GST fusion proteins were immobilized on glutathione-Sepharose beads and incubated with [³⁵S]methionine-labeled proteins at 4°C for more than 2 h. The beads were washed extensively, and bound proteins were separated on an SDS-polyacrylamide gel and viewed by autoradiography.

ANT interacts with CED-4 in C. elegans.

To identify additional factors that are involved in the central cell-killing process in C. elegans, we used a coimmunoprecipitation assay to search for proteins that are associated with the proapoptotic protein CED-4. We first expressed a Flag-tagged CED-4 (CED-4Flag) under the control of a heat shock promoter in ced-1(e1735); ced-4(n1162) double mutants. After heat shock, a large number of cell corpses were observed in transgenic embryos, indicating that the transgenic CED-4Flag was functional to rescue the cell death defects caused by ced-4 loss-of-function mutation (Fig. 1A and B). Western blot analysis further confirmed that CED-4Flag was well expressed in transgenic animals (Fig. 1C). We next prepared whole-worm lysate from the transgenic animals and performed immunoprecipitation of CED-4Flag with anti-Flag antibody. Three proteins with molecular masses of ∼90, ∼50, and ∼33 kDa were found to associate with CED-4 (Fig. 1D). Mass spectrometric analyses revealed that the peptide sequences of the 33-kDa protein corresponded to four proteins belonging to a family of highly homologous proteins, ANTs (Fig. 1D and see Fig. S1A in the supplemental material). Sequence comparison and structure modeling indicated that C. elegans ANTs share high homology with human ANT (see Fig. S1B and C in the supplemental material). We named them wan genes for “worm adenine nucleotide translocators.” wan-1 and wan-2 encode for proteins of 300 amino acids, while wan-3 and wan-4 encode for proteins of 313 amino acids, respectively (see Fig. S1B in the supplemental material).

Human possesses four ANT isoforms that are expressed in tissue- and development-specific manner. These ANTs function in mitochondria for the exchange of mitochondrial ATP with cytosolic ADP (14, 32). Importantly, ANT is a major component of mitochondrial PTPC that promotes the release of apoptotic factors from mitochondria during apoptosis (2, 16, 45). In addition, overexpression of ANT1 and ANT3 dominantly induced apoptosis (1, 44), while the knockdown of ANT2 sensitized cells to chemical-induced apoptosis (21). To determine whether C. elegans wan genes are important for programmed cell death, we analyzed the profiles of developmental cell deaths in wan-1(gk172), wan-2(gk294), wan-3(gk300), and wan-4(ok2209) deletion mutants which likely represent strong loss-of-function or null mutations of individual wan genes (Fig. 1E). Homozygous wan-1(gk172) worms are lethal. However, homozygous wan-2, -3, and -4 deletion mutants are viable and exhibit no obvious growth defects, allowing us to examine the cell death profiles in these mutants. A time course analysis of both embryonic and germ line cell deaths indicated that the wan-2(gk294), wan-3(gk300), and wan-4(ok2209) single mutants did not show pronounced defects in either somatic or germ line apoptosis compared to that of wild-type animals (see Fig. S2A in the supplemental material and data not shown). Furthermore, the wan-2(gk294); wan-3(gk300) and wan-2(gk294); wan-4(ok2209) double mutants, as well as the wan-2(gk294); wan-3(gk300) wan-4(RNAi) triple mutants did not show obvious cell death defects (see Fig. S2B in the supplemental material). In addition, we examined the cell death profiles of these wan gene mutants in the background of ced-1(e1735), which causes the accumulation of cell corpses owing to defects in cell corpse engulfment. In the double mutants between ced-1 and wan-2, -3, and -4, neither embryonic nor germ line cell death profiles are distinct from that of ced-1(e1735) single mutants alone (see Fig. S2C and D in the supplemental material). Taken together, these results indicate that wan-2, -3, and -4 are not essential for programmed cell death in C. elegans.

wan-1 functions in the cell-killing process.

Since wan-1(gk172) deletion mutants are embryonic lethal, we performed RNAi to inactivate wan-1 to determine whether it is involved in programmed cell death. For this purpose, we used the dsRNA derived from the wan-1 3′UTR, which shares no homology with other wan genes. wan-1 RNAi gave rise to an efficient inactivation of wan-1 (see Fig. S3A in the supplemental material) and caused ca. 60% embryonic lethality in wild-type animals, which was less severe than observed in wan-1(gk172) animals (see Table S1 in the supplemental material). Nevertheless, in animals expressing a WAN-1::GFP translation fusion protein under the control of the wan-1 promoter and the 3′UTR of the unc-54 gene (P_wan-1 wan-1::gfp::unc-54 3′UTR), the embryonic lethality caused by wan-1 RNAi was strongly suppressed (see Table S1 in the supplemental material), indicating that the 3′UTR-derived wan-1 RNAi specifically targeted wan-1 for inactivation. However, since fair number of wan-1(RNAi) embryos survived and developed normally to L4 stage, we were able to examine the cell death profiles in these escapers. In wan-1(RNAi) embryos, the number of cell corpses was significantly decreased at all examined embryonic stages compared to that in control RNAi-treated N2 animals (Fig. 2A). In contrast, in embryos expressing P_wan-1 wan-1::gfp::unc-54 3′UTR, wan-1 RNAi failed to cause obvious decrease of cell corpses (Fig. 2A), indicating that the expression of WAN-1::GFP rescued the embryonic cell death defects caused by the 3′UTR-derived wan-1 RNAi. wan-1 RNAi also significantly reduced the number of cell corpses in the triple mutants of wan-2, -3, and -4 (see Fig. S2B in the supplemental material). Moreover, we used four-dimensional microscopy analysis to determine the number of cell death events from the first cell division to 400 min during embryogenesis. We found that significantly fewer cell deaths occurred in wan-1(RNAi) embryos that developed normally than in control RNAi-treated animals (Fig. 2B). In addition, we examined whether wan-1 RNAi could enhance the cell death defects in the ced-3 and ced-4 weak loss-of-function mutants. In wild-type worms, 16 cells in the anterior pharynx normally undergo programmed cell death and barely extra “undead” cells can be observed. In the ced-3(n2438) and ced-4(n2273) weak loss-of-function mutants, however, cell death is partially blocked, and usually one to three extra cells can be seen. Importantly, wan-1 RNAi significantly enhanced the cell survival phenotype in both weak ced-3 and ced-4 mutants (Table 1) . Taken together, these findings suggest that wan-1 RNAi inhibited programmed cell death in C. elegans.

To further prove that wan-1 is important for cell death, we performed wan-1 RNAi in animals defective in cell corpse engulfment. In C. elegans, the engulfment of cell corpses is mainly controlled by two parallel and partially redundant genetic pathways, with the genes ced-1, -6, and -7 and dyn-1 functioning in one pathway and the genes ced-2, -5, -10, and -12 and psr-1 functioning in the other pathway (29, 35, 43). In ced-1(e1735) and ced-7(n1892) animals, wan-1 RNAi significantly decreased the number of cell corpses at all examined embryonic stages (Fig. 2C and E). Similarly, wan-1 RNAi also caused strong reduction of cell corpses in ced-2(e1752) and ced-5(n1812) animals (Fig. 2D and F). The reduction of cell corpses in both engulfment pathways further proved that wan-1 RNAi indeed inhibited cell death, suggesting that wan-1 likely functions in the cell-killing process.

wan-1 is important for germ cell death.

Except for somatic cell death, C. elegans germ cells also undergo apoptosis, which is controlled by the same regulatory pathway as in somatic cells (15). To determine whether wan-1 is also important for germ cell death, we examined wan-1 RNAi-treated wild type as well as mutants defective in cell corpse engulfment. In the wild type, dying cells are swiftly engulfed and degraded so that only a small number of germ cell corpses can be seen at different adult ages (15), but even fewer cell corpses were observed after wan-1 RNAi treatment (Fig. 3A), suggesting that wan-1 RNAi likely inhibited germ cell death as well. Moreover, wan-1 RNAi treatment by either feeding with bacteria expressing dsRNA derived from the cDNA of wan-1 (Fig. 3B and C) or injection of dsRNA corresponding to the 3′UTR of wan-1 (see Fig. S4 in the supplemental material) strongly reduced the number of germ cell corpses to a similar level in all examined mutants defective in cell corpse engulfment. The reduction of germ cell corpses was more profound with the aging of animals (Fig. 3B and C and see Fig. S4 in the supplemental material); this was not likely due to defects in germ line development because the germ lines of ced-1(e1735); wan-1(RNAi) worms were morphologically indistinguishable from those of control RNAi-treated animals (data not shown). In addition, the number of nuclei in the germ line mitotic regions of both control RNAi- and wan-1 RNAi-treated ced-1(e1735) worms showed no significant differences (Fig. 3D). To determine whether the wan-1 RNAi-induced reduction of germ cell corpses had resulted from a suppression of apoptosis or an enhancement of cell corpse engulfment, we used four-dimensional microscopy analysis to measure the duration of germ cell corpse persistence and to monitor the occurrence of germ cell death. Our results indicated that germ cell corpses on average existed longer than 200 min in both control RNAi- and wan-1 RNAi-treated ced-1(e1735) worms, suggesting that wan-1 RNAi did not obviously affect the engulfment of cell corpses. However, wan-1 RNAi significantly reduced the number of cell death events. For example, we observed 5.6 ± 0.7 cells die within 200 min in the gonadal U-turn region in control RNAi-treated ced-1(e1735) animals with an adult age of 36 h post-L4 stage, but we only observed 1.1 ± 0.4 cells die within 200 min in the same gonadal region of wan-1 RNAi-treated ced-1(e1735) animals at the same adult age (Fig. 3E, region 1). In the oocyte region, we also found that wan-1 RNAi significantly reduced the number of cell death events compared to control RNAi treatment (Fig. 3E, region 2). These findings indicated that wan-1 RNAi suppressed germ cell death rather than enhanced cell corpse engulfment. Collectively, our results demonstrated that wan-1 regulates germ line apoptosis in C. elegans.

The ANT inhibitor BA suppresses apoptosis in C. elegans.

Mammalian ANT has two states, the cytosolic state (c-state) and the matrix state (m-state) (28). The PTP-inhibitory compounds Atr and BA can selectively bind to different regions of ANT (Fig. 4A). The binding of Atr to ANT fixes the latter to the c-state, which induces the swelling of mitochondria and facilitates apoptosis (23). In contrast, BA fixes ANT to the m-state and suppresses Atr- or Bax-induced apoptosis (23). Because WAN-1 highly resembles mammalian ANT (28) (see Fig. S1B and C in the supplemental material), we tested whether BA and Atr could affect programmed cell death in C. elegans by treating ced-1(e1735) animals with BA or Atr at concentrations comparable with those applied to mammalian cells for interfering apoptosis (1, 23). Although both BA and Atr caused significant reduction of ATP level in worms (Fig. 4B), only BA induced a strong suppression of germ line apoptosis in a concentration-dependent manner (Fig. 4C, D, and E). For example, 6 h after removal of BA, only a mean of 5.4 germ cell corpses were observed in ced-1(e1735) animals treated with 100 μM BA, compared to a mean of 31.2 germ cell corpses at the same time point in the control animals (Fig. 4D). In contrast, Atr did not cause obvious change of germ line apoptosis (Fig. 4E), indicating that the suppression of apoptosis by BA was not likely due to a general inhibition of ATP/ADP exchange. The suppression of germ line apoptosis by BA suggests that WAN-1 may function similarly to its mammalian equivalents in regulating apoptosis.

WAN-1 localizes to mitochondria.

To understand how WAN-1 may regulate programmed cell death, we examined the subcellular localization of the translational WAN-1::GFP fusion protein (P_wan-1 wan-1::gfp::unc-54 3′UTR). WAN-1::GFP expression was observed from embryonic to adult stages. In adult worms, WAN-1::GFP was seen in a wide range of tissues including the intestines, hypoderm, vulva, and some neurons (see Fig. S3B in the supplemental material). In embryos, WAN-1::GFP exhibited a punctate, cytoplasmic staining pattern which overlapped with that of the mitochondrion-specific fluorescent dye, Mitotracker Red (Fig. 5A), indicating that WAN-1 is localized to mitochondria. To confirm this, we expressed WAN-1Flag in HEK293 cells and found that it also displayed a punctate staining pattern that completely overlapped with that of Mitotracker Red (Fig. 5B). A further subcellular fractionation analysis revealed that WAN-1Flag existed in the same fraction with that of cytochrome c (see Fig. S3C in the supplemental material). These results indicate that WAN-1 indeed localizes to mitochondria.

Biochemical and structural studies have suggested that mammalian ANT functions as a dimer (28). Since WAN-1 highly resembles mammalian ANT, we tested whether WAN-1 could also dimerize. We coexpressed WAN-1 proteins with different tags in HEK293 cells and performed immunoprecipitation. Our results indicated that WAN-1Myc could nicely pull down WAN-1Flag and vice versa (Fig. 5C and D). Thus, WAN-1 likely dimerizes or oligomerizes to fulfill its normal function in living cells.

WAN-1 interplays with core regulators for cell death activation.

Previously, heterologous expression of C. elegans core cell death regulators in mammalian cultured cells has been used very successfully in dissecting the interaction among these proteins, which suggested a molecular mechanism underlying programmed cell death (12, 13, 37, 41). Since WAN-1 could localize to mitochondria in HEK293 cells as in worms (Fig. 5B), we decided to take advantage of mammalian cells to examine the potential protein interactions between WAN-1 and other core cell death regulators.

We first tested whether CED-4 and WAN-1 could interact in mammalian cells since WAN-1 was initially identified as a CED-4-interacting protein in C. elegans (Fig. 1). When WAN-1 and CED-4 labeled with different tags were coexpressed in HEK293 cells, CED-4 nicely pulled down WAN-1 as revealed by immunoprecipitation (Fig. 6A). In contrast, CED-4 failed to interact with an unrelated protein, the C. elegans p53 homolog CEP-1, when they were similarly coexpressed (Fig. 6A), indicating that CED-4 and WAN-1 interacted specifically in mammalian cells. Next, we examined whether CED-4 and WAN-1 could directly interact using the glutathione S-transferase (GST) fusion protein pull-down assay. Our data indicated that ³⁵S-labeled WAN-1 interacted with GST-CED-4 immobilized on glutathione-Sepharose beads but not with GST (Fig. 6B), further demonstrating that WAN-1 interacted with CED-4 directly and specifically.

In living cells, CED-9 forms complex with CED-4 to prevent the latter from promoting CED-3 processing. Since WAN-1 directly interacts with CED-4, we tested whether WAN-1 could also interact with CED-9. We found that CED-9 was associated with WAN-1 but not CEP-1 in HEK293 cells (Fig. 6C). In addition, the GST pull-down assay revealed that WAN-1 could directly interact with CED-9 in vitro (Fig. 6D). These results indicated that WAN-1 and CED-9 could also form complex.

As WAN-1, CED-4, and CED-9 directly interacted with one another, we examined whether they could form a tripartite complex by performing immunoprecipitation of WAN-1Flag, CED-4Myc, and CED-9Myc, which were coexpressed in HEK293 cells. As shown in Fig. 6E, both CED-4 and CED-9 were coimmunoprecipitated with WAN-1, indicating that they formed a complex. Next, we investigated whether these interactions occurred on mitochondria. We isolated the mitochondrial fraction from HEK293 cells coexpressing WAN-1 and CED-4 or WAN-1 and CED-9 and found that WAN-1 interacted with both CED-4 and CED-9 in purified mitochondria (see Fig. S3D and E in the supplemental material). Nevertheless, these interactions seemed not to be affected by either BA or Atr (see Fig. S3D and E in the supplemental material), implying that BA may not function by directly interfering with the protein interaction between WAN-1 and these two cell death regulators.

The binding of the cell death initiator EGL-1 to CED-9 to disrupt CED-9-CED-4 complex is the key to the onset of apoptosis in C. elegans (6, 39). Because WAN-1 interacts with CED-9, we tested whether EGL-1 could disrupt the complex of these two proteins. In HEK293 cells, WAN-1 was associated with CED-9 as shown by coimmunoprecipitation (Fig. 6C and E). However, when EGL-1 was coexpressed with WAN-1 and CED-9, the interaction between WAN-1 and CED-9 was strongly reduced (Fig. 6F). Moreover, we found that EGL-1 efficiently pulled down CED-9 but not WAN-1 when these three proteins were coexpressed in HEK293 cells, indicating that EGL-1 and WAN-1 did not interact with each other (Fig. 6G). Thus, the strong reduction of WAN-1-CED-9 interaction was likely caused by an EGL-1-induced conformational change of CED-9 (40). The disruption of CED-9-WAN-1 interaction by EGL-1 may release the proapoptotic activity of WAN-1 which together with CED-4 promotes the activation of cell death in C. elegans.

wan-1 overexpression induces ectopic cell death in C. elegans.

Because WAN-1 interacts with core cell death regulators and plays an essential role in programmed cell death, we further examined whether WAN-1 could induce cell death by ectopically expressing WAN-1 in touch receptor neurons under the control of the promoter of the mec-7 gene (P_mec-7 wan-1) (30). We then scored the percentage of posterior lateral microtubule (PLM) touch receptor neurons that were killed. Our data revealed that overexpression of WAN-1 resulted in >50% death of PLM neurons (51 to 56% PLM death in three independent transgenic lines), which was similar to that caused by CED-4 overexpression (P_mec-7 ced-4) (48 to 56% PLM death) (Table 2). These results indicate that WAN-1 on its own can induce ectopic cell death. To determine whether WAN-1-induced cell death requires the core cell death machinery, we crossed transgenic lines expressing WAN-1 into the ced-3(n717) and ced-4(n1162) strong loss-of-function mutants, as well as into the ced-9(n1950) gain-of-function mutants. We found that PLM killing caused by WAN-1 overexpression was almost completely abolished in these mutants, indicating that WAN-1 induced cell killing through the core cell death pathway (Table 2).

To investigate further how WAN-1 may act to induce PLM death, we coexpressed WAN-1 and CED-4 in touch receptor neurons and found that it caused a higher percentage of PLM death (56 to 68%) than that caused by either WAN-1 or CED-4 alone (Table 2). Previously, it was shown that CED-4-induced ectopic cell killing was enhanced by ced-9 loss-of-function mutation (31), which suggests that endogenous CED-9 protects against cell killing by CED-4 overexpression. Thus, we speculated that the increased PLM death observed in animals coexpressing WAN-1 and CED-4 was due to an antagonistic effect of WAN-1 on endogenous CED-9. To test this possibility, we crossed transgenic lines expressing CED-4 alone and lines coexpressing WAN-1 and CED-4 into the ced-9(n1950) gain-of-function mutants. Consistent with previous report (31), the ced-9(n1950) mutation did not obviously inhibit the PLM death induced by CED-4. In contrast, the ectopic PLM death induced by coexpressing WAN-1 and CED-4 was partially suppressed by the ced-9(n1950) gain-of-function mutation (Table 2), which resulted in PLM death similarly to that caused by overexpressed CED-4 in the ced-9(n1950) mutants (Table 2). Altogether, these data suggest that WAN-1 likely promotes cell killing by antagonizing the antiapoptotic role of CED-9.