Multifunctional reaper: sixty-five amino acids of fury

Introduction: The Core Cell Death Machinery

Apoptosis is a genetically driven form of cell death that results in systematic dismantling of the dying cell. Apoptotic cell death, a process critical for development and tissue homeostasis in multicellular organisms, is carried out by a family of proteases known as caspases. Caspases that lie at the start of apoptotic pathways are known as initiator or apical caspases. They in turn activate executioner caspases, which go on to cleave the various cellular substrates that package the cell for elimination. Apical caspases are first activated by adaptor proteins that are typically targeted by upstream apoptotic signals. Although the mechanism of adaptor protein activation is in some cases controversial, considerable characterization has been reported for the vertebrate Apaf-1 adaptor protein that is activated by cytochrome c. Cytosolic cytochrome c, released from the mitochondria during cell death binds to and induces oligomerization of Apaf-1 into heptamers, which in turn recruit and activate caspase 9.^{1, 2} In vertebrate cells, mitochondrial events of apoptosis, including cytochrome c release, are regulated by the Bcl-2 family. Antiapoptotic Bcl-2 family members (Bcl-2, Bcl-xL, Mcl-1) inhibit release of cytochrome c, whereas proapoptotic Bcl-2 family members (Bax, Bak, Bid, Bim Bad) promote cytochrome c release (reviewed by Cory et al.³ and Danial and Korsmeyer⁴).

In Drosophila melanogaster, the regulation of caspase activation is not fully understood. Although a Drosophila homolog of Apaf-1, known variously as Drosophila Apaf-1-related killer (Dark), homolog of Apaf-1 and Ced4 (Hac-1) or Drosophila Apaf-1 (dApaf-1) has been characterized and is known to promote activation of the initiator caspase Dronc,^{5, 6, 7} it is not known how or even if Dark is directly regulated.⁸ Reports from several laboratories have shown that Dark binds to cytochrome c^{5, 6} and that cytochrome c can induce the formation of a Dark apoptosome in vitro, although there have not yet been any reports documenting release of cytochrome c from Drosophila mitochondria in response to apoptotic stimuli.⁹ Additionally, a recent report suggests that cytochrome c is neither present within nor required for formation of Dark/Dronc complexes.¹⁰ In addition, although Drosophila cytochrome c isoforms can induce cell death upon ectopic expression in vertebrates, depletion of cytochrome c has no apparent effect on cell death in the fly.^{11, 12} Moreover, reconstituted apoptosis in fly cell extract does not require mitochondrial factors.¹³ However, there is evidence that caspase activation during spermatid differentiation is dependent on cytochrome c, but it is not known if this reflects an ability of cytochrome c to bind and promote Dark activation.^{14, 15} In addition to controversy over the role of the mitochondria in fly cell death, it is also unclear how the two Bcl-2 family members in Drosophila, Buffy (dBorg-2) and Debcl (Drob-1/dBorg-1/dBok) regulate cell death. Buffy, thought to behave as an antiapoptotic Bcl-2 family member, inhibits cell death during development as well as in response to γ irradiation.¹⁶ Debcl, a proapoptotic Bcl-2 family member mediates cell death during development, during UV irradiation, and when ectopically expressed.^{17, 18, 19, 20} A specific function for Buffy or Debcl has not been definitively shown and it is not clear if the Drosophila Bcl-2 family members act on mitochondria, like their vertebrate counterparts. Thus, although sequence homology can be informative, it may be misleading; it is not yet clear whether sequence relatedness is at all indicative of functional similarities between fly and vertebrate Bcl2 family members.

Drosophila Inhibitors of Apoptosis

In both vertebrate and Drosophila cells, caspases can be inhibited by a family of proteins called inhibitor of apoptosis proteins (IAPs) (reviewed by Salvesen and Duckett²¹). IAPs contain one or more baculovirus IAP repeat (BIR) domains, which were first identified as regulators of survival in baculovirus-infected cells.²² BIR domains bind to activated caspases and can inhibit their catalytic activity.²¹ In contrast to mammalian cells, where the importance of IAPs has been most convincingly demonstrated in the regulation of cell death in postmitotic cells^{23, 24} and where no obvious role in developmental apoptosis has been observed,²⁵ IAPs appear to be the central locus of apoptotic regulation in the fly. Indeed, unlike oligomerization of the Apaf-1/capsase 9 apoptosome in vertebrates, formation of the fly Dark/Dronc apoptosome may be a relatively unregulated step, as processed Dronc appears to be generated and degraded continuously.⁸ Rather, the critical regulatory point of cell death in flies appears to be at the level of IAPs. In support of this, a decrease in the levels of the primary Drosophila IAP (DIAP-1) alone results in massive cell death,^{26, 27} indicating that there is constitutive caspase activity in Drosophila cells that is held in check by IAPs. DIAP-1 contains two BIR domains, responsible for caspase binding. Binding itself may not be sufficient for apoptotic inhibition, as there is growing evidence that caspases remain active when bound to DIAP-1,²⁸ suggesting that another activity is responsible for its antiapoptotic function. Two models (reviewed by Ditzel and Meier²⁹) have been proposed for this activity. The N-end rule model suggests that caspases are codegraded with DIAP-1 during cell death. This is accomplished by destabilization of DIAP-1 by caspase cleavage at its N-terminus, which induces its ubiquitylation, targeting it for degradation along with bound Dronc.³⁰ It has also been shown that DIAP-1 can induce the ubiquitylation and subsequent degradation of the apical caspase, Dronc^{31, 32, 33} as shown in Figure 1a. This occurs through DIAP-1's RING domain which is vital for its antiapoptotic function.^{32, 34} RING domain-containing proteins (E3 ligases) cooperate with E2 proteins in order to target proteins for degradation, suggesting that DIAP-1 may recruit an E2 ligase to induce the ubiquitylation of Dronc.

Figure 1

Reaper's regulation of caspase activation. (a) In addition to inhibiting the activity of active caspases, DIAP-1 induces ubiquitylation of the apical caspase Dronc through its RING domain. Ubiquitylation induced by DIAP-1 prevents the accumulation of active Dronc and apoptosis. (b) Reaper, Grim and Hid compete for the BIR domains of DIAP-1 with Dronc, allowing free Dronc to accumulate. When bound to Reaper, DIAP-1 is ubiquitinated, leading to its degradation, further preventing caspase inhibition. (c) DIAP-1 can also prevent its own inhibition by ubiquitinating Reaper. This increases Reaper turnover, preventing it from inducing cell death

Inhibition of DIAP-1 by IAP-Antagonists

Given the centrality of IAP proteins in the fly, it is perhaps not surprising that the key apoptotic inducers in this system, both in response to toxic insults and developmental cues, are a family of IAP antagonists. In both vertebrate and fly cells, IAPs can be inhibited by a family of antagonists that share a common IAP-binding motif (IBM). Although the IBM is well conserved both within and between species, there is little overall homology between vertebrate and fly IBM-containing IAP antagonists. Five potent apoptotic inducers in the fly, Reaper, Hid, Grim, Jafrac2 and Sickle, all appear to act through antagonism of IAP proteins. Embryos homozygous for the Df(3L)H99 deletion, which removes Reaper, Grim and Hid, fail to develop owing to a lack of normal programmed cell death. These embryos are also deficient in apoptosis induced by γ irradiation. In addition, Reaper and Hid have been shown to be transcriptionally induced during UV irradiation,^{35, 36} suggesting that these IAP antagonists mediate DNA stress-induced apoptosis. A number of studies have shown that targeted expression of IAP antagonists in the fly eye using the glass multimer reporter (GMR) enhancer disrupts eye patterning and morphology.^{37, 38, 39}

The IBM-containing proteins are known to bind to IAPs and displace them from caspases, thereby alleviating caspase inhibition.^{37, 38, 40, 41, 42, 43, 44} Whereas some IAP antagonists may simply release active caspases from IAPs, it is also apparent that some of these proteins can reduce total IAP levels in the cell.^{27, 45, 46} Ryoo et al.⁴⁶ demonstrated that Reaper binds to DIAP-1 and promotes auto-ubiquitylation (shown in Figure 1b) in a RING domain-dependent manner via the ubiquitin-conjugating enzyme UBCD1. Recent work has also shown that Grim can interact with UBCD1 and can accelerate ubiquitylation of DIAP-1 upon addition to fly embryonic extracts, although purified Grim did not accelerate this reaction using all recombinant components in vitro.⁴⁷ Similar findings were also obtained showing that Hid-induced DIAP-1 degradation is dependent on DIAP-1's E3 ligase activity.^{27, 34} There is some controversy over the importance of DIAP-1's RING domain in this process, as Reaper and Grim do not require the E3 ligase activity of DIAP-1 to reduce its levels in the cell.^{27, 34} There are also data showing that Reaper induces caspase cleavage of DIAP-1 at its N-terminus,³⁴ which may be critical in targeting it for degradation. As shown in Figure 1c, DIAP-1 can also inhibit Reaper-induced apoptotic signals by inducing Reaper ubiquitylation and degradation by the proteosome in a RING domain-dependent fashion. Moreover, a mutant of Reaper that cannot be ubiquitylated is a much more potent inducer of cell death.⁴⁸

In addition to acting on IAPs, several other biochemical activities (described further below) have been ascribed to IAP antagonists that likely contribute to the regulation of fly cell death. In particular, Reaper and Grim have been shown to inhibit protein synthesis.^{27, 45, 49} This may result in a loss of de novo synthesized prosurvival proteins or may simply shut down processes unnecessary for a dying cell. Reaper can also interact with and regulate the Hsp70/Hsc70 binding partner Scythe,^{50, 51} which appears to regulate apoptosis upstream of caspase activation.⁵² The multiple functions of the IAP antagonists suggest that a cell utilizes these death-inducing ligands in multiple ways to expedite cell death and to stop cellular processes that might interfere with the progression of apoptosis. It is possible then that increasing levels of IAP antagonists send a widely received apoptotic signal to the cell to prepare it for death.

Non-IAP-Dependent Functions of Reaper

Studies of Reaper and other IBM-containing proteins have indicated that these proteins are not simply IAP antagonists, but rather death ligands with diverse cellular targets. Although disruption of Reaper's IAP-binding ability markedly interferes with its proapoptotic activity, a mutant variant of Reaper missing its IBM (RPR16-65) has been shown to retain some ability to induce cell death in mammalian cultured cells⁵³ and cytochrome c release in Xenopus egg extracts,⁵⁴ indicating that Reaper can induce apoptosis in a non-IAP-related manner. IBM-independent death activities have also been observed in the fly, as overexpression of RPR16-65 in the fly eye using the GMR enhancer induces death.⁵⁵ Additionally, Claveria et al.⁵⁶ showed that Grim can also induce cell death without its IBM in tissue culture cells. Grim's IAP-independent ability to induce cell death is dependent on a hydrophobic region, known as the GH3 (Grim Homology 3) domain, which itself contains independent proapoptotic activity; this domain was shown to be indispensable for induction of apoptosis by full-length Grim. It is not clear how the GH3 domain induces cell death or why it is required, but it is apparent that the domain acts as a mitochondrial-targeting sequence. Reaper also has a similar hydrophobic sequence, critical for its ability to induce apoptosis and for its mitochondrial localization. Furthermore, Olson et al.⁵⁷ showed that deletion of the GH3 domain disrupts Reaper-induced DIAP-1 degradation. This loss of function could be rescued by restoring mitochondrial localization by attaching the C-terminal mitochondrial-targeting sequence of Bcl-xL. Hid also contains a mitochondrial localization sequence, but its role has not been established.⁵⁸ The dependence on mitochondrial localization for these IAP antagonists still remains a mystery. The available data suggest that degradation of DIAP-1 requires or is facilitated by its localization to the mitochondria.

Reaper or Reaper missing its IBM can induce mitochondrial cytochrome c release when added to Xenopus egg extract, suggesting that vertebrate systems are responsive to Reaper's IAP-independent proapoptotic activities. It is not known whether the ability of Reaper to permeabilize mitochondria is conserved in flies and whether such permeabilization is required for the mitochondrial contribution to Reaper-induced DIAP-1 degradation. In vertebrate cells, the mechanism of Reaper-induced cytochrome c release is not clear, but may well involve the Bcl-2 family. Addition of Bcl-2 to the Xenopus egg extract protects against Reaper-induced cytochrome c release.⁵⁴ Reaper addition may therefore lead to the activation of a proapoptotic Bcl-2 family member to initiate cytochrome c release. That said, it has been shown by one group that expression of Grim can induce GH3-dependent mitochondrial permeabilization even in Bax/Bak knockout mouse embryo fibroblasts, suggesting that Bcl-2-independent mechanisms may mediate GH3-dependent mitochondrial permeabilization.⁵⁹

Reaper's regulation of premitochondrial events of apoptosis in vertebrates appears to involve a 150 kDa protein called Scythe (Bag6/Bat3), purified by Thress et al.^{51, 52} from the Xenopus egg extract.⁵¹ Xenopus Scythe interacts with Reaper and can inhibit apoptosis when added in excess to the extract. Scythe immuno-depletion also blocks cell death by co-depletion of a proapoptotic factor. Scythe contains a domain homologous to that found in the Bcl-2 interactor, Bag-1, which binds to and inhibits the Hsp70/Hsc70 chaperones and is critical for Scythe's regulation of cell death.⁵⁰ Biochemical characterization has indicated that Bag domain-containing proteins bind to Hsp70 and trap substrate proteins, thereby inhibiting protein folding.^{60, 61} Scythe's regulation of cell death likely depends on its ability to form a static complex with Hsp70 and a proapoptotic protein. As shown in Figure 2, Reaper, by binding to Scythe, can relieve its inhibition of Hsp70, thereby releasing a postulated trapped proapoptotic protein, which can induce cytochrome c release. Recently, Desmots et al.⁶² generated a Scythe knockout mouse, revealing a number of striking features of Scythe regulation of cell death in mammals. Scythe deficiency is lethal owing to a number of developmental defects in lungs, kidneys and brain. These developmental defects were due to dysregulation of cell death. Surprisingly, both increased and decreased apoptosis were seen in the Scythe-deficient mice in different tissue types. Although the data are not shown, Desmots et al.⁶² note that Scythe deficiency does not inhibit Reaper-induced cell death in cell culture. However, based on the ability of Scythe overexpression to inhibit Reaper-induced cell death in the Xenopus egg extract as well as biochemical evidence suggesting that Reaper can liberate a proapoptotic factor from Scythe sequestration, it is possible that genetic ablation of Scythe may sensitize cells to toxic insults. Although a true Reaper homolog has not been identified in vertebrate cells, the phenotypes seen in the Scythe-deficient mice suggest that a vertebrate Scythe-modulatory ligand must exist. Whether such a protein will bear primary sequence homology to Reaper is not known, although it is interesting to note that the known IAP antagonists in vertebrates (Smac and Omi) have no reported ability to induce apoptotic events upstream of mitochondria.

Figure 2

Reaper and Scythe's regulation of cell death. Scythe, through its Bag domain interacts with and inhibits the protein-folding activity of Hsp70. Bag domain containing proteins will trap Hsp70 substrates. In this case, Scythe traps a proapoptotic factor (factor X) bound to Hsp70. Reaper binds Scythe, releasing Hsp70 and its proapoptotic substrate. This released factor induces the release of cytochrome c from the mitochondria

IAP Antagonists and Translation Inhibition

In addition to activation of apoptotic pathways, there is growing evidence that Reaper can serve as an inhibitor of translation.^{27, 45, 49, 63} Clues that Reaper might possess such an activity came from two sources: first, as mentioned above, Reaper can downregulate DIAP-1 even when DIAP-1 lacks intrinsic E3 ligase activity.^{27, 34} Whereas this may reflect the involvement of another E3 ligase in DIAP-1 degradation, it was suggested that Reaper and Grim might inhibit also IAP synthesis. Second, in searching for sequence homologs of Reaper outside of leptidopterans, it was realized that Reaper bore some sequence homology to a group of small nonstructural proteins (NSs) encoded by members of the Bunyaviridae, a family of negative-sense arthropod-transmitted RNA viruses that can cause a number of illnesses including hemorrhagic fever and pediatric encephalitis.⁶⁴ The bunyaviral NSs is critical for viral pathogenesis as well as inhibition of translation during infection.⁶⁵ Given these observations, Reaper and related proteins were tested for their ability to inhibit protein translation. In pulse chase experiments in the Xenopus egg extract and in rabbit reticulocyte lysates, Reaper could clearly dampen protein synthesis. This effect was independent of IAP inhibition or caspase activation as it was not inhibitable by zVAD-FMK and still occurred when Reaper lacking the IBM was assayed.⁴⁵ Similar inhibition of translation in vitro was observed with Grim, but not Hid protein.²⁷ Interestingly, there are suggestive data that translational inhibition is critical for Reaper's ability to induce cell death in mammalian cells.⁴⁹ However, one study has shown that Reaper does not inhibit overall bulk protein translation in Drosophila tissue culture cells, although our recent studies suggest that Reaper may alter start codon selection, perhaps changing the population of proteins translated in Reaper-expressing cells.^{63, 66} Interestingly, we have recently found that Reaper bearing a mutation within the conserved motif (amino acids 19–27 of Reaper) shown to be important for translational inhibition in Reaper's bunyaviral homolog⁶⁷ has a reduced capacity to induce cell death when overexpressed in the fly eye (Figure 3a). It is clear from these data that this region of Reaper is important for cell death, which suggests a potential role for this translation inhibition domain in Reaper-induced apoptosis, although the possibility remains that this domain is also important for IAP binding or degradation. Although Reaper and Smac share no marked homology outside the IBM, a recent report has also demonstrated that Smac inhibits tumor growth in part by inhibiting global protein synthesis, demonstrating that this activity may be conserved in IAP antagonists.⁶⁸ With regard to the mechanism of translational inhibition, recent work by Colon-Ramos et al.⁶³ has shown that Reaper's ability to inhibit translation stems from direct binding to the 40S subunit of the ribosome. This binding compromises AUG start site recognition, which results in a decrease in CAP-dependent initiation. Moreover, this work demonstrates that Reaper does not affect translational initiation of a CAP-independent viral IRES. These findings suggest that Reaper may allow certain messages to be translated while suppressing others. This raises the possibility that proapoptotic or stress-related messages may continue to be expressed, whereas pro-growth or prosurvival genes are turned off. In this context, it is interesting to note that Yoo et al. found that the half-life of Dronc is significantly longer than that of DIAP-1. An inhibition of global translation would therefore lead to a rapid decrease in DIAP-1, potentially resulting in activation of the more stable Dronc.²⁷

Figure 3

(a) Reaper's translation inhibition motif (amino acids 19–27) is important for its ability to induce cell death. One copy of GMR-RPR (a) induces cell death in the fly eye to a greater extent than one (b) or two (c) copies of GMR-RPRΔ19–27. GMR-RPRΔ19–27 has an intact IBM and GH3 domain, suggesting that translation inhibition is a critical step in Reaper toxicity. (b) Reaper inhibits normal cap-dependent translation of mRNA by compromising AUG recognition, allowing the ribosome to initiate translation at alternative AUGs. Reaper will therefore prevent the production of the normal poly-peptide and may generate a unique protein from a secondary start site

Reaper as a Multifunctional Killer

Reaper appears to be an efficient death inducer with multiple distinct functions. Deletion of the IBM, the GH3 or the translational inhibitory motif can reduce the capacity of Reaper to induce cell death, suggesting that their distinct functions all contribute to Reaper activity. It is interesting that the relative importance of each Reaper function may vary depending upon where Reaper is expressed. When expressed heterologously in vertebrates, where cytochrome c release is a pivotal event in apoptosis, the GH3 domain is of clear functional importance for mitochondrial permeabilization. Although no vertebrate GH3-containing proteins have been reported, the sequence can be found in the NSs proteins from a number of Bunyaviridae, which are mammalian pathogens. Consistent with a role for this domain in apoptosis, the NSs proteins were shown to have mitochondrial cytochrome c-releasing activity dependent upon this domain. Moreover, the NSs proteins appear to act through Scythe to induce cytochrome c release.⁶⁷ Therefore, this may be an evolutionarily conserved function important for death induction in vertebrates, but of secondary importance in fly cells. Conversely, Reaper lacking an IBM or NSs proteins, which naturally lack an IBM are as potent as full-length Reaper in inducing apoptosis in vertebrate systems, whereas the IBM clearly makes the largest contribution to Reaper's biological activity in fly cells. In both settings (vertebrate and invertebrate), Reaper/NSs regulation of translation may be important for inhibiting the production of pro-survival proteins or perhaps to simply shut down the cell in preparation for cell death. Indeed, regulation of translation is a common response to cellular stress.⁶⁹ Whether or not a vertebrate death regulator containing all of the regulatory features of Drosophila Reaper exists is not yet clear. However, the machinery necessary to recognize these features of Reaper is clearly conserved, suggesting that the regulatory motifs must exist, even if they are separated on distinct molecules in vertebrate cells. Drosophila Reaper is a remarkably efficient machine, packing multiple death-inducing functions into only 65 amino acids. Such economy may represent a rare twist of evolution.

In Memory of Stan Korsmeyer

Although Stan Korsmeyer never worked on Reaper himself, we had a number of opportunities to discuss this work and he was always supportive, thoughtful and insightful. Stan had an ability to appreciate a diversity of experimental systems and approaches, viewing new data with a broad open mind. Thus, when I first met him as a starting investigator and we were studying the biochemical activities of heterologous proteins in Xenopus egg extracts, Stan was enthusiastic about the idea that we might learn something novel using this system. I found his excitement about science and encouraging words inspiring. I also found him to be generous and a gentleman, helping with reagents, letters of support and constructive criticism. He set the standard for how colleagues should behave and the cell death field and science in general are sadder without him. –SK.