Interaction and Regulation of Subcellular Localization of CED-4 by CED-9

See allHide authors and affiliations

Science  21 Feb 1997:
Vol. 275, Issue 5303, pp. 1126-1129
DOI: 10.1126/science.275.5303.1126


The Caenorhabditis elegans survival gene ced-9 regulates ced-4 activity and inhibits cell death, but the mechanism by which this occurs is unknown. Through a genetic screen for CED-4-binding proteins, CED-9 was identified as an interacting partner of CED-4. CED-9, but not loss-of-function mutants, associated specifically with CED-4 in yeast or mammalian cells. The CED-9 protein localized primarily to intracellular membranes and the perinuclear region, whereas CED-4 was distributed in the cytosol. Expression of CED-9, but not a mutant lacking the carboxy-terminal hydrophobic domain, targeted CED-4 from the cytosol to intracellular membranes in mammalian cells. Thus, the actions of CED-4 and CED-9 are directly linked, which could provide the basis for the regulation of programmed cell death in C. elegans.

Programmed cell death (PCD) plays an essential role in animal development and homeostasis (1). Genetic studies in the nematode Caenorhabditis elegans have identified several components of the death pathway, some of which are conserved in vertebrates, including humans (2, 3). The nematode ced-9 gene was discovered by analysis of gain-of-function mutants, and it was subsequently shown that ced-9 protects cells that normally survive from undergoing PCD (4). In contrast, loss-of-function mutations in the ced-3 gene cause all 131 cells that normally die to survive, indicating that ced-3 is required for PCD (3). Both ced-9 and ced-3 have mammalian homologs. The ced-9 gene encodes a protein with significant sequence homology to the vertebrate Bcl-2 and Bcl-xL survival proteins, whereas the ced-3 product is homologous to the interleukin-1β-converting enzyme (ICE), which is a member of a growing family of ICE-like proteases, called caspases (5, 6). The mammalian bcl-2 gene can functionally substitute for ced-9 in C. elegans (5, 7), suggesting that Bcl-2 is a mammalian homolog of CED-9. Another C. elegans gene, ced-4, is required for developmental cell death in the worm, and its overexpression can cause cells that normally survive to undergo PCD (8, 9). Genetic experiments have indicated that ced-9 controls cell death by preventing the activation of the death genes ced-3 and ced-4 (4, 9) and have suggested that ced-3 acts downstream of ced-4 and that ced-9 acts upstream of ced-4 (4, 9). Consistent with this, Bcl-2 and Bcl-xL can inhibit the activation of ICE-like proteases and therefore appear to function upstream of the death proteases in the mammalian apoptotic pathway (10).

To search for proteins that bind to CED-4, we screened a C. elegans cDNA library for CED-4-interacting proteins by using GAL4-CED-4S as a “bait” in the yeast two-hybrid assay (11). In a screen of 2 × 106 library clones, four cDNAs were found to interact with the GAL4-CED-4 but not with control baits. One of the CED-4-interacting cDNAs encoded the entire CED-9 coding region fused in-frame to the GAL4 transcriptional activation domain. The association of CED-4 with CED-9 was specific in that CED-4 was unable to interact with empty vector or several plasmids encoding irrelevant baits (Fig. 1A) (12). To further characterize the CED-4-CED-9 interaction, we determined the ability of CED-4 to associate with three natural CED-9 mutants (Fig. 1A). In the yeast two-hybrid system, CED-4 interacted with the wild type and a gain-of-function CED-9 mutant but not with two loss-of-function mutants (Fig. 1B). These results indicate that the protective activity of CED-9 correlates with its ability to interact with CED-4 and strongly suggest that a physical interaction of CED-9 with CED-4 is critical for the regulation of PCD in C. elegans.

Fig. 1.

Interaction of CED-4 with wild-type CED-9 and CED-9 mutants in yeast. (A) Schematic representation of wild-type and mutant CED-9 proteins. BH1, BH2, and BH4 boxes depict conserved regions shared with Bcl-2 family members. TM represents the conserved COOH-terminal hydrophobic tail. The position of point mutations in n2077 and n2812 (loss-of-function) as well as n1950 (gain-of-function) mutations of ced-9 are shown (5). Q indicates Gln residues. G169E represents Gly169 → Glu169. ΔTM represents an engineered CED-9 mutation lacking residues 250 through 280. (B) Interaction of CED-4 with wild-type and mutant CED-9 proteins. A plasmid expressing CED-4 fused to the GAL4 DNA-binding domain was cotransfected with plasmids encoding wild-type CED-9, CED-9 mutants, wild-type CED-4, or empty vector sequences fused to a GAL4 transcriptional activation domain. Growth of yeast in the absence of leucine, tryptophan, and histidine is indicative of protein-protein interaction. Growth in the absence of leucine and tryptophan is shown as a control. The results are representative of three independent experiments.

The interaction of CED-4 with CED-9 could not be assessed in C. elegans cells in culture, because no such cells are currently available. To verify that CED-4 associates with CED-9 in vivo, we transiently cotransfected a 293T human kidney cell line with expression plasmids producing a Myc epitope- tagged CED-4 and hemagglutinin (HA)- tagged CED-9 protein (13). Immunoprecipitates were prepared with rabbit antibody to Myc (anti-Myc) and subjected to SDS-polyacrylamide gel electrophoresis (PAGE). Immunoblotting with a monoclonal antibody to HA (anti-HA) revealed that CED-9 was coimmunoprecipitated specifically with CED-4 (Fig. 2A). Control immunoblotting with anti-HA and anti-Myc confirmed that CED-9 and CED-4 proteins were expressed in total lysates of cells transfected with the corresponding plasmids but not in cells transfected with control plasmids (Fig. 2A). To further verify the interaction, we performed reciprocal experiments with anti-HA to immunoprecipitate CED-9, followed by immunoblotting with rabbit anti-Myc. In agreement with the reverse experiment, CED-4 coimmunoprecipitated specifically with CED-9 (Fig. 2B). The CED-9 protein contains a conserved COOH-terminal hydrophobic tail that may serve as an anchoring domain to intracellular membranes (5). We engineered a mutant form of CED-9 with a deletion of the COOH hydrophobic domain (CED-9ΔTM) to determine if the putative membrane anchoring region is required for CED-9 to localize to intracellular membranes in mammalian cells (13). Immunoprecipitation analysis revealed that the CED-9ΔTM mutant protein retained its ability to associate with CED-4, indicating that the COOH hydrophobic domain of CED-9 is not required for the interaction with CED-4 (Fig. 2B). In addition, CED-4 was specifically coimmunoprecipitated with Bcl-xL, a mammalian homolog of CED-9 (Fig. 2C) but not with Bcl-xS, a mutant form of Bcl-xL lacking an internal region of 62 amino acids that contain the conserved Bcl-2 homology region 1 (BH1) and BH2 domains (Fig. 2D) (13).

Fig. 2.

Interaction of CED-4 with CED-9, CED-9ΔTM mutant, and human Bcl-x proteins in mammalian cells. (A) 293T cells (5 × 106 per 100-mm plate) were transiently transfected with 2.5 μg of the indicated Myc- or HA-tagged expression plasmids or empty vector. In the case of transfection with a single plasmid, cells were cotransfected with 2.5 μg of empty vector so that the total amount of transfected plasmid DNA was always 5 μg. Lysates were immunoprecipitated with rabbit anti-Myc or normal rabbit serum. Immunoprecipitates were immunoblotted with mouse anti-HA to detect the HA-CED-9 protein. In the lower two panels, aliquots of total protein lysates were immunoblotted with anti-Myc or anti-HA. Molecular size markers are indicated in kilodaltons. (B) Anti-HA or isotype control immunoprecipitates were immunoblotted with anti-Myc to detect Myc-tagged CED-4. The amount of plasmid DNA and number of transfected cells were identical to that described for (A). (C) Anti-Myc or control immunoprecipitates were blotted with anti-FLAG. In the lower two panels, aliquots of total protein lysates were immunoblotted with anti-FLAG or anti-Myc(D) Anti-FLAG, anti-HA, or control immunoprecipitates were blotted with anti-Myc. In the lower panel, aliquots of total protein lysates were immunoblotted with anti-FLAG (to detect Bcl-xL) or anti-HA (to detect Bcl-xS or CED-9). The results are representative of three independent experiments.

To assess the subcellular localization of wild-type CED-9, we transfected an HA-ced-9 expression plasmid into 293T cells and determined the labeling pattern of CED-9 by immunostaining and confocal microscopy (15). Analysis of labeled cells revealed that HA-CED-9 displayed a compact, granular, and extranuclear staining pattern consistent with a localization confined to membranes of intracellular organelles and the perinuclear region (Fig. 3A). In contrast, the labeling pattern of the CED-9ΔTM mutant was diffuse and cytoplasmic, which is consistent with a cytosolic localization (Fig. 3B). Analysis of cells transfected with a Myc-CED-4 construct with anti-Myc revealed a cytoplasmic labeling pattern that was distinct from wild-type CED-9 but similar to that of the CED-9ΔTM mutant (Fig. 3C). Transfection with a control GATA-1 expression plasmid revealed a nuclear labeling pattern as expected for a nuclear factor (Fig. 3F). Significantly, coexpression of CED-9 with CED-4 strongly altered the distribution of CED-4 to a compact granular, perinuclear pattern similar to that observed for CED-9 (compare panels C and D in Fig. 3). The change in subcellular localization of CED-4 induced by coexpression with CED-9 was specific in that it was not observed by cotransfection with an expression plasmid that encodes the nuclear GATA-1 (16). The labeling pattern of CED-4 was not affected by coexpression of the CED-9ΔTM mutant (Fig. 3E), indicating that the membrane-anchoring domain is required for CED-9 to alter the subcellular localization of CED-4.

Fig. 3.

Subcellular localization of CED-9 and CED-4 proteins in 293T cells. Cells were transiently transfected with expression plasmids encoding HA-CED-9 (A), HA-CED-9ΔTM mutant (B), Myc-CED-4 (C), HA-CED-9 plus Myc-CED-4 (D), Myc-CED-4 plus HA-CED-9ΔTM mutant (E), or FLAG-GATA-1 as a control (F). Shown are confocal images after labeling with anti-HA (A, B), anti-Myc (C, D, E) or anti-FLAG (F). Arrowheads, perinuclear region. N, nucleus. Scale bar, 10 μm. Labeling with control IgG was negative in all cases. The results are representative of at least two independent experiments.

We performed subcellular fractionation of protein lysates in which cytosolic and membrane fractions were prepared and analyzed by immunoblotting for the presence of CED-4 and CED-9 proteins (17). Analysis of cellular lysates revealed that the majority of CED-9 was contained within the membrane fraction, whereas the CED-9ΔTM mutant protein was located mainly in the cytosolic fraction (Fig. 4A). Similar subcellular localizations have been reported, respectively, for the CED-9 mammalian homolog Bcl-2 and its ΔTM mutant (18). In contrast to CED-9, the CED-4 protein resided mostly in the cytosolic fraction (Fig. 4B). Significantly, a large fraction of the CED-4 protein was directed to the membrane fraction in 293T cells cotransfected with ced-4 and ced-9 expression plasmids (Fig. 4B). However, the cytosolic distribution of CED-4 was largely unaffected in cells that expressed the ΔTM mutant form of CED-9 (Fig. 4C), which is consistent with the confocal analysis shown in Fig. 3. These results indicate that CED-9 requires its conserved COOH-terminal hydrophobic region for both attachment to intracellular membranes and targeting of CED-4 to intracellular membranes in mammalian cells.

Fig. 4.

Subcellular fractionation of 293T cells transfected with ced-9 and ced-4 expression plasmids. Cytosolic (C) and membrane (M) protein fractions were analyzed for CED-9 or CED-4 protein expression by immunoblotting using monoclonal anti-HA or polyclonal anti-Myc. Protein lysates from an equal number of cells were loaded in the C and M lanes. (A) Cells were transfected with an HA-ced-9 plasmid (top panel) or an HA-ced-9 mutant plasmid (ΔTM). Molecular size markers are indicated in kilodaltons. (B) Cells were transfected with empty vector (control), Myc-ced-4, or Myc-ced-4 plus HA-ced-9. (C) Cells were transfected with Myc-ced-4 plus HA-ced-9ΔTM or control plasmids. The relative ratio of CED-4 in the cytosolic or membrane fractions was quantitated by densitometry analysis of the corresponding bands. The results are representative of three independent experiments.

Genetic studies have shown that ced-9 controls cell death by preventing the activation of the death genes ced-3 and ced-4 (4, 9). Two forms of ced-4ced-4S and ced-4L—that exhibit different functions have been identified (19). Here, we studied ced-4S, the most abundant form of ced-4 expressed in C. elegans cells (19). Our studies showed that CED-4 and CED-9 proteins interact in yeast and mammalian cells and presumably in C. elegans cells. Thus, we propose that CED-9 regulates cell death at least in part by binding and inactivating CED-4. This could be accomplished by sequestering CED-4 from its otherwise cytosolic distribution to specific intracellular membrane sites. The removal of cytosolic CED-4 could result in the reduced ability of CED-4 to activate downstream effector molecules such as CED-3, thereby promoting cell survival. This model is most consistent with genetic analysis in C. elegans, where it has been shown that ced-9 inhibits ced-3 activity at least in part through ced-4 (9). We have shown that CED-4 interacts with Bcl-xL, a mammalian CED-9 homolog. These results would predict that the mammalian CED-9 homologs Bcl-2 and Bcl-xL will regulate cell death at least in part by interacting with a mammalian CED-4 counterpart. The membrane localization of CED-9 may be important for the regulation of CED-4 activity, because Bcl-2 mutants lacking the COOH-terminal membrane-anchoring tail exhibit a greatly reduced anti-apoptotic activity when compared with wild-type Bcl-2 (20). Bcl-2 associates with and targets the cytosolic protein kinase Raf-1 to mitochondrial membranes (21). Unlike the pro-apoptotic CED-4, Raf-1 functions to improve Bcl-2-mediated resistance to apoptosis, perhaps indicating that the Bcl-2 family may play a more general role in the targeting or sequestration of specific apoptotic regulatory proteins to intracellular membranes.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
View Abstract

Stay Connected to Science

Navigate This Article