BIOLOGICAL OVERVIEW

Grim encodes a protein required for programmed cell death in Drosophila. grim, maps between two previously identified cell death genes in this region: reaper (rpr) and head involution defective (hid). Expression of Grim RNA coincides with the onset of programmed cell death at all stages of embryonic development, whereas ectopic induction of grim triggers extensive apoptosis in both transgenic animals and in cell culture. Cell killing by Grim is blocked by coexpression of p35, a viral product that inactivates ICE-like proteases, and does not require the functions of rpr or hid. The predicted Grim protein shares an amino-terminal motif in common with Rpr. However, Grim is sufficient to elicit apoptosis in at least one context, where Rpr is not. The grim gene product might thus function in a parallel circuit of cell death signaling that ultimately activates a common set of downstream apoptotic effectors (Chen, 1996).

The Grim N-terminus induces apoptosis by disrupting IAP blockage of caspases; however, N-terminally-deleted Grim retains pro apoptotic activity. This study describes GH3, a 15 amino acid internal Grim domain absolutely required for its proapoptotic activity and sufficient to induce cell death when fused to heterologous carrier proteins. A GH3 homology region is present in the Drosophila proapoptotic proteins Reaper and Sickle. The GH3 domain and the homologous regions in Reaper and Sickle are predicted to be structured as amphipathic alpha-helixes. During apoptosis induction, Grim colocalizes with mitochondria and cytochrome c in a GH3-dependent but N-terminal- and caspase activity-independent manner. When Grim is overexpressed in vivo, both the N-terminal and the GH3 domains are equally necessary, and cooperate for apoptosis induction. The N-terminal and GH3 Grim domains thus activate independent apoptotic pathways that synergize to efficiently induce programmed cell death (Clavería, 2002).

Secondary structure prediction of the Grim protein has identified three regions with a very high probability of conforming to an a-helical structure. These regions have been termed GH1, GH2 and GH3, for Grim Helix 1, 2 and 3. The GH3 domain shows similarity to a region in Reaper and Sickle, both also predicted to conform as an a-helix. The homology region spans the 15 amino acids predicted to conform as an a-helix in Grim, with two 5 amino acid regions of high similarity flanking a 5 amino acid central region with lesser homology. Representation in a helical wheel projection of GH3 residues and of those in the Reaper and Sickle homology regions reveals the amphipathic nature of the predicted a-helices (Clavería, 2002).

Tests were performed to see whether GH3 is important for Grim proapoptotic function by assaying the cell killing ability of several Grim mutants altered in the GH3 domain in Drosophila SL2 cells. Wild-type (WT) Grim induces cell death when overexpressed in this assay. In contrast, a Grim mutant form with a 13 amino acid deletion that removes the GH3 residues most reliably predicted to form an a-helix only marginally induces apoptosis. A 5'-shifted 11 amino acid deletion, such that the 3' part of GH3 was respected, was less effective in eliminating the proapoptotic activity than the complete GH3 deletion. An internal deletion removing four amino acids commonly deleted in the two larger deletions (Delta89-92) also resulted in strong impairment of Grim killing ability, but to a lesser extent than the complete GH3 deletion. The relevance was tested of L89, a conserved residue within positions 89-92, whose hydrophobicity could be relevant for the amphipathic nature of the GH3 domain. A non-conservative replacement of L89 by glutamic acid (L89E) impaired GH3 killing ability nearly to the same level as the Delta89-92 mutant. Semi-conservative replacement of L89 by alanine (L89A) had a mild effect on Grim proapoptotic function. These results show that the GH3 domain is required for Grim proapoptotic function and that L89 is a functionally relevant residue in the domain (Clavería, 2002).

Since N-terminally deleted Grim can still bind IAPs, GH3 domain function could be related to Grim's ability to bind IAPs and inhibit their protective function. Deletion of the Grim N-terminal domain was found to lead to a slight reduction in its ability to bind DIAP2, however, deletion of the GH3 domain, either alone or in combination with the N-terminal deletion, does not impair Grim's ability to bind DIAP2. These results suggest that GH3 activity is unrelated to the IAP inhibitory Grim activity (Clavería, 2002).

To determine whether the GH3 domain could function as a proapoptotic motif itself, a Grim fragment containing the GH3 domain was fused to green fluorescent protein (GFP) as a carrier protein (GH3-GFP) and the ability of this fusion protein to induce apoptosis in Drosophila SL2 cells was tested. GH3-GFP induces cell death with the same efficiency as does the complete Grim-GFP fusion protein. The cell death observed is specific to GH3 domain activity, since it is largely abolished by an L-to-E mutation in the residue equivalent to Grim L89. In all cases, cell death was rescued by coexpression with the baculoviral caspase inhibitor p35. The GH3 domain is therefore sufficient to trigger a specific proapoptotic route in SL2 cells (Clavería, 2002).

Immunocytochemistry was used to identify Grim subcellular localization in Drosophila SL2 cells. In phases previous to any obvious apoptotic phenotype, Grim generally shows a diffuse distribution in the cytoplasm, but also displays rings of stronger Grim staining. Grim rings are mitochondria-associated, as indicated by the presence of a mitochondrial matrix marker, mainly inside the rings, but also in colocalization with strong Grim staining. Grim colocalization with cytochrome c is similar to that observed with the mitochondrial matrix marker but, in addition, Grim and cytochrome c show extensive colocalization in larger dots. These larger cytochrome c dots are not observed in untransfected cells: they increase in size and abundance as apoptosis progresses, and do not colocalize with a mitochondrial marker. No substantial cytochrome c release to cytosol was found. Instead, Grim promotes redistribution of cytochrome c signal in large dots in colocalization with Grim itself (Clavería, 2002).

Grim subcellular localization is dependent on the presence of an intact GH3 domain. GH3-deficient Grim shows no obvious organization in rings associated with mitochondria and does not colocalize with either the mitochondrial marker or cytochrome c. In contrast, deletion of Grim 2-14 amino acids does not alter subcellular distribution of the Grim protein, nor the changes induced in cytochrome c display (Clavería, 2002).

The relevance of the GH3 and N-terminal Grim domains was examined by overexpressing the Grim mutants in transgenic flies using the Gal4-UAS system. To drive Grim expression, the GMR-Gal4 line, which targets expression to the eye disc, and the MS1096-Gal4 line, which directs expression to the wing imaginal disc, were used. Overexpression of WT Grim with either driver causes total or partial lethality in all transgenic lines. These results suggested that leaky expression from both promoters in vital tissues produces sufficient cell death to block development. Surviving adult flies overexpressing WT Grim display a considerable reduction in eye size with the GMR driver, and wing agenesis plus notum reduction and elimination of macro- and microchaetae with the MS1096 driver (Clavería, 2002).

In contrast to these results, overexpression of a GH3-deleted Grim (Delta86-98) results in very low lethality, as well as rescue of the eye, notum and wing phenotypes. Elimination of amino acids 89-92 results in a lesser impairment of Grim killing ability, showing that the 3' part of the GH3 domain is important for proapoptotic function. In correlation with these results, the Delta89-92 mutant rescues the eye phenotype induced by WT Grim, but only partially rescues the more sensitive wing and notum phenotypes. Elimination of amino acids 83-93, which extends the deletion N-terminal to the putative helical domain, does not increase the rescue observed with the 89-92 deletion, suggesting that residues 5' of leucine 89 might be less important for proapoptotic function. The relevance of leucine 89 was again shown by the significant rescue of viability and of eye, notum and wing phenotypes in flies overexpressing the non-conservative L89E substitution. In contrast, the semi-conservative substitution L89A results in mild, but significant, impairment of Grim death induction and targeted tissue deletion (Clavería, 2002).

In accordance with the results observed in cultured cells, mutations of the GH3 domain impairs Grim proapoptotic activity in transgenic flies. Deletion of the N-terminal domain, in contrast to the results observed in cultured cells, results in highly significant elimination of Grim proapoptotic function in vivo. Both viability and appearance of the tissues targeted by the GMR and MS1096 drivers were rescued by the 2-14 deletion to a level similar to that observed for the GH3 deletion. Simultaneous deletion of the N-terminus and amino acids 89-92 of the GH3 domain results in even lower lethality and fewer alterations in the targeted tissues than those induced by each deletion in isolation (Clavería, 2002).

To determine whether the N-terminal and GH3 Grim domains can function independently of each other, tests were performed to see whether the simultaneous expression of independent 2-14- and GH3-deleted Grim proteins could induce apoptosis. Flies were generated carrying independent transgenes for 2-14 and 86-98 Grim deletion mutants driven by MS1096 expression. Whereas males carrying either protein alone showed little or no lethality and no alterations in wing development, double transgenic males simultaneously expressing Delta2-14 and Delta86-98 Grim mutants display severe lethality and reduced wings. Females did not display lethality in any situation, but frequently showed reduced wings in the double transgenics, although not in single transgenics. Functions of both the N-terminal and the GH3 domains are therefore essential for Grim activity in vivo and they independently activate specific death mechanisms that synergize to trigger apoptosis (Clavería, 2002).

Several lines of evidence point to the mitochondrial-cytochrome c pathway as the target of GH3 action. Grim associates with mitochondria in colocalization with cytochrome c and this activity resides in the GH3 domain. In vertebrate cells, Grim targets the mitochondria and induces cytochrome c release in a GH3-dependent and N-terminal-, caspase- and IAP-independent manner, suggesting functional conservation of the pathway. Interestingly, during apoptosis induction by Grim and Reaper in flies, changes in cytochrome c display are observed, rather than its free release to the cytosol as in vertebrates. Grim-expressing cells specifically show large cytoplasmic deposits of cytochrome c at sites where Grim itself is present, but other mitochondrial markers are not. The changes observed in the distribution of cytochrome c may result from its relocation from mitochondria to hypothetical specialized cytoplasmic structures involved in apoptosis induction. Alternatively, the apoptosome might be formed in the vicinity of the mitochondria, and cytochrome c deposits may constitute the remnants of damaged mitochondria, which have lost some of their constitutive components, but retain cytochrome c and Grim. The relevance of the cytochrome c proapoptotic pathway in Drosophila PCD is supported as well by the observation that elimination of Dark, a Drosophila homolog of Apaf-1 that mediates cytochrome c-primed apoptosis, impairs Reaper, Hid and Grim killing in flies. Even though no cytochrome c free release appears to take place in Drosophila cells, it is possible that a mechanism homologous to that of vertebrate cells is activated, but from different subcellular compartments (Clavería, 2002).

The involvement of the GH3 domain in a mitochondrial pathway and its predicted structure, an amphipathic a-helix, resemble the characteristics of the widespread proapoptotic BH3 domain. These similarities could be interpreted as functional homology between the two pathways; however, no association has been detected between Grim and either mammalian (Bcl-2 and Bcl-x_L) or insect (Debcl) Bcl-2-family members, as would be expected for a BH3-containing protein. Rather than representing homologous proapoptotic pathways, BH3 and GH3 domains may have converged during evolution to a similar proapoptotic mechanism. Since BH3-containing proteins coexist in Drosophila with GH3-containing proteins, the two pathways may operate in alternative routes, or even cooperate in apoptosis induction, not only in Drosophila, but perhaps also in other species (Clavería, 2002).

Two independent pathways may thus be triggered by Grim; an IAP inhibitory pathway activated by the N-terminal domain, and a mitochondrial-cytochrome c route activated by the GH3 domain. Either pathway could be alternatively or simultaneously promoted by Grim, and the relevance of each may depend on cellular context. The presence of a GH3 homology region in Reaper and Sickle suggests functional conservation of this domain in at least these other two Drosophila proapoptotic proteins. In this context, it is important to consider that Reaper promotes cytochrome c release in a cell-free Xenopus egg extract and does not require the N-terminal domain for this function (Clavería, 2002).

Although Reaper, Hid, Sickle and Grim induce specific apoptotic pathways in vertebrate cells, and in the fly participate in highly conserved routes, such as the p53 and Ras-MAPK pathways, no homolog for these proteins has been yet identified in any other organism. The vertebrate Smac/Diablo protein may, however, represent a functional homolog of the IAP inhibitory pathway. Smac/Diablo can bind to and block the protective effect of IAPs. However, it is unlikely that Smac/Diablo represent homologs of the mitochondrial-cytochrome c pathway. Database searches have failed to identify any protein with sequence similarity to the GH3 domain but, given the restricted sequence conservation among, for example, BH3 family members, this does not exclude conservation of this pathway. Whether vertebrate proapoptotic proteins exist that represent direct or functional homologs of the GH3 proapoptotic activity thus remains to be determined (Clavería, 2002).

GENE STRUCTURE

cDNA clone length - 1696 bases

Bases in 5' UTR - 315

Exons - 1

Bases in 3' UTR - 964

PROTEIN STRUCTURE

Amino Acids - 138

Structural Domains

In vivo binding studies demonstrate that both Grim and Reaper physically interacts with human IAPs through a 15-amino acid N-terminal segment. Deletion of this segment from either Reaper or Grim abolishes binding to cIAPs. In vitro binding experiments indicate that Reaper and Grim bind specifically to the BIR domain-containing region of cIAPs, since deletion of this region results in loss of binding. The physical interaction has been further confirmed by immunolocalization. When co-expressed, Reaper or Grim co-localize with cIAP1. However, deletion of the N-terminal 15 amino acids of Reaper or Grim abolishes co-localization with cIAP1, suggesting that this homologous region can serve as a protein-protein interacting domain in regulating cell death. Moreover, by virtue of this interaction, it has been demonstrated that cIAPs can regulate Reaper and Grim by abrogating their ability to activate caspases and thereby inhibit apoptosis (McCarthy, 1998).

N-terminally-deleted Grim retains pro apoptotic activity. GH3 is 15 amino acid internal Grim domain absolutely required for its proapoptotic activity and sufficient to induce cell death when fused to heterologous carrier proteins. A GH3 homology region is present in the Drosophila proapoptotic proteins Reaper and Sickle. The GH3 domain and the homologous regions in Reaper and Sickle are predicted to be structured as amphipathic alpha-helixes. During apoptosis induction, Grim colocalizes with mitochondria and cytochrome c in a GH3-dependent but N-terminal- and caspase activity-independent manner. When Grim is overexpressed in vivo, both the N-terminal and the GH3 domains are equally necessary, and cooperate for apoptosis induction. The N-terminal and GH3 Grim domains thus activate independent apoptotic pathways that synergize to efficiently induce programmed cell death (Clavería, 2002).

Secondary structure prediction of the Grim protein has identified three regions with a very high probability of conforming to an a-helical structure. These regions have been termed GH1, GH2 and GH3, for Grim Helix 1, 2 and 3. The GH3 domain shows similarity to a region in Reaper and Sickle, both also predicted to conform as an a-helix. The homology region spans the 15 amino acids predicted to conform as an a-helix in Grim, with two 5 amino acid regions of high similarity flanking a 5 amino acid central region with lesser homology. Representation in a helical wheel projection of GH3 residues and of those in the Reaper and Sickle homology regions reveals the amphipathic nature of the predicted a-helices (Clavería, 2002).

date revised: 15 November 2003

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.