DEVELOPMENTAL DYNAMICS 218:68 –79 (2000)

REVIEWS
A PEER REVIEWED FORUM

Regulation and Execution of Apoptosis During

ABSTRACT The development of the Dro- undergo apoptosis (for reviews, see Bergmann et al.,
sophila embryo into an adult fly is a process that 1998; Newton and Strasser, 1998; Vaux and Kors-
integrates cell proliferation and differentiation meyer, 1999). It is now clear that the components of the
with programmed cell death, or apoptosis. Apo- apoptotic machinery are constitutively expressed in
ptosis is an evolutionarily conserved process virtually all nucleated animal cells and that the acti-
that is controlled in the developing fly by the vation of this machinery is controlled by a set of intra-
products of the genes reaper, grim, and hid. We cellular regulatory proteins that transduce signals
discuss the role of programmed cell death in the from both inside and outside the cell. In most cases,
establishment and maintenance of correct pat- cells require specific survival signals from their local
terning in the embryo, and examine the coordi- environment to prevent the activation of apoptosis
nation of apoptosis with the hormonally con- (Raff, 1992). For example, when epithelial cells are
trolled degeneration of larval tissues during removed from the extracellular matrix and no longer
metamorphosis. Finally, we address the architec- have ligated integrin receptors, they respond by under-
ture of the adult eye as an example of how pro- going programmed cell death (reviewed in Frisch and
grammed cell death plays a key role in the devel- Ruoslahti, 1997). Cell death can also be induced by
opment of many adult structures. Dev Dyn 2000; specific signals from neighboring or nearby cells, as
218:68 –79. © 2000 Wiley-Liss, Inc. exemplified by the death ligands TNF and FasL (re-
viewed in Nagata, 1997).
Key words: apoptosis; Drosophila; development Apoptosis is mediated by the activation of a specific
class of proteases known as caspases. This name re-
INTRODUCTION flects the conserved cysteine in the active site and the
The fact that programmed cell death plays a signif- requirement for an aspartic acid at the cleavage site of
icant role in animal development has been known for a their target proteins. Caspases are constitutively
long time (Saunders and Fallon, 1967). Originally ob- present in cells as inactive zymogens with variable-
served in insect morphogenesis (Lockshin, 1969), cell length amino-terminal prodomains (Cryns and Yuan,
death has been recognized in both vertebrate and in- 1998). Caspases themselves have caspase cleavage
vertebrate development as necessary for the elimina- sites separating each domain, and activation of the
tion of superfluous cells, morphogenetic changes, and pro-caspases requires proteolysis at these sites.
the hollowing out of solid structures to form cavities or Caspases can be divided into two classes based on their
tubes (Coucouvanis and Martin, 1995). In many in- prodomains and roles in cell death. The upstream, or
stances, cell death is observed in developing tissues but initiator, caspases have longer prodomains that medi-
its function is not known, and it is unclear whether ate the transduction of death signals and the assembly
apoptosis is necessary for these processes to occur or if of activating complexes. When initiator caspases are
it is simply coincident with them (Glucksmann, 1951). brought into proximity with one another by assembly
The study of Drosophila and C. elegans mutants that into an activating complex, autoprocessing of the
completely lack developmental apoptosis has allowed caspases occurs through an intrinsic proteolytic activ-
this question to be addressed. These studies indicate ity (Salvesen and Dixit, 1999). The newly activated
that in these organisms apoptosis is not necessary for initiator caspases are then free to proteolytically acti-
some aspects of normal development (White et al.,
1994; Metzstein et al., 1998). However, as discussed
below, apoptosis provides developmental flexibility and Grant sponsor: National Institutes of Health; Grant sponsor: Shi-
is required for the maturation of the Drosophila ner- seido Company of Japan.
vous system and for metamorphosis. *Correspondence to: Kristin White, Cutaneous Biology Research Cen-
ter, Massachusetts General Hospital, Harvard Medical School, Charles-
The past few years have seen tremendous progress in town, MA02129. E-mail: [email protected]
the identification of the mechanisms by which cells Received 10 January 2000; Accepted 8 February 2000

© 2000 WILEY-LISS, INC.

vate the second class of caspases, the downstream or Regulators of embryonic cell death in Drosophila
effector caspases, which are characterized by shorter were initially identified by screening for mutants with
prodomains. The resulting protease cascade ends with disrupted patterns of cell death as assayed by AO
the cleavage of a multitude of cellular targets including staining. Three overlapping deletions mapping to the
structural proteins and enzymes involved in gene ex- left arm of chromosome 3 result in the complete abro-
pression, DNA replication and metabolic activities of gation of developmental cell death and lead to embry-
the cell. The dying cell eventually fragments into apo- onic lethality at a fairly late developmental stage
ptotic bodies which are subsequently engulfed by (White et al., 1994). Subsequent analysis of this region
neighboring cells or macrophages. has identified three novel genes, reaper (rpr), grim, and
Many of the proteins known to be important in apo- head involution defective (hid), whose gene products
ptosis were first identified genetically in C. elegans and are involved in the initiation of all embryonic cell death
human cancer mutations (reviewed in Hunter, 1997; in Drosophila (White et al., 1994; Grether et al., 1995;
Metzstein et al., 1998). Conservation across species is Chen et al., 1996). The apoptotic machinery is intact in
very high, indicating that the core apoptotic machinery embryos deleted for these genes, since apoptosis can be
and many regulatory features of apoptosis have been induced by X-irradiation, with an ultrastructural phe-
preserved through evolution. This has allowed the use notype that is indistinguishable from wild-type em-
of the genetic tools available in invertebrate systems to bryos. Therefore, rpr, grim, and hid must play signal-
identify many of the key proteins in apoptosis and to ing roles as opposed to participating in the actual
establish the pathways involved in its initiation and execution of the apoptotic mechanism.
execution. Sequence analyses of rpr, grim, and hid show them
The fruit fly Drosophila melanogaster has proven to to be unrelated to each other with the exception of
be a particularly convenient organism for the study of limited homology in the first 15 amino acids of each.
apoptosis. As will be discussed below, the genes re- When over-expressed in embryos via heat-shock induc-
quired for the induction of apoptosis in Drosophila ible transgenes, RPR, GRIM, and HID each induce
have been cloned, as have numerous other factors in- extensive cell death resulting in the demise of the
volved in the initiation and execution of the cell death embryo (Fig. 1). When expressed ectopically in the Dro-
pathway (reviewed by Abrams, 1999). The developing sophila eye, each of these genes is sufficient to induce
organism is relatively accessible to both observation massive cell death leading to ablation of the eye (Gre-
and dissection, and a veritable treasure chest of genetic ther et al., 1995; Hay et al., 1995; Chen et al., 1996;
tools allows for the manipulation of specific genes in White et al., 1996). Over-expression of any of these
specific tissues at precise timepoints. Additionally, the genes in cultured Drosophila cells or other insect cell
development of the fruit fly, from oogenesis through lines also results in the rapid induction of cell death
embryogenesis and metamorphosis into adulthood, has (Chen et al., 1996; Nordstrom et al., 1996; Vucic et al.,
been extremely well characterized and described 1998). Importantly, although vertebrate homologues of
through decades of work (for example, see Bate and rpr, grim, and hid have yet to be identified, each have
Martinez Arias, 1993). been shown to induce cell death in a caspase-dependent
manner when expressed in a number of vertebrate
REGULATION OF CELL DEATH IN systems (Evans et al., 1997; Claveria et al., 1998; Mc-
DROSOPHILA EMBRYOGENESIS Carthy and Dixit, 1998; Haining et al., 1999). In all
The vital dye acridine orange (AO) specifically stains cases of over-expression, killing by either of these
apoptotic cells and has been used to examine dying genes is independent of the other two.
cells in living Drosophila embryos (Abrams et al., In the embryo, rpr and grim expression appears to be
1993). Microscopic examination of AO stained embryos predictive of cells that are fated to die, and transcripts
shows that cell death is prominent and widespread of these genes can be detected in doomed cells approx-
during embryogenesis and that it occurs in a relatively imately 2 hr prior to the onset of morphologically rec-
predictable spatial and temporal pattern (Abrams et ognizable apoptosis (White et al., 1994; Chen et al.,
al., 1993). The first dying cells are invariably detected 1996; Robinow et al., 1997) (Fig. 2). In contrast, hid
in the dorsal region of the head approximately 7 hr expression is largely coincident with many cells which
after egg laying, which corresponds to the later part of undergo cell death, but this correlation is not absolute.
the fully extended germ band stage (stage 11). As de- For example, while there is considerable cell death
velopment proceeds, apoptotic cell death becomes more detected in the ventral nerve cord during late embryo-
prominent and widespread throughout the embryo, genesis, hid expression is not detected in these cells
and corpses are engulfed by circulating macrophages. (Abrams et al., 1993; Grether et al., 1995). In addition,
Time-lapse photography of AO-stained embryos shows hid mRNA can be detected throughout the optic lobe
that while cell death occurs in a consistent pattern, the primordium in stage 12 embryos, but only a fraction of
precise spatial and temporal aspects of this pattern are these cells die (Grether et al., 1995). The latter obser-
somewhat variable, indicating that there is a certain vation suggests that HID activity may be regulated
degree of plasticity in embryonic cell death (Abrams et post-translationally, and it has in fact been shown that
al., 1993; Pazdera et al., 1998). both hid expression and activity are modulated via the
70 BANGS AND WHITE

Fig. 1. Cell killing by overexpression of rpr. The top row shows AO constitutively expressed lacZ. The cell on the right also carries a trans-
stained embryos at similar stages. The embryo on the right carries a gene in which rpr expression is driven by the metal inducible metallothio-
transgene that expresses rpr ubiquitously from the heat shock promoter. nein promoter. The cells have been induced with copper for 2 hr and
The embryos were heat shocked and stained 1.5 hr later. A substantial stained with lacz to identify transfected cells. When rpr is expressed, it
increase in apoptosis can be seen in the hsrpr embryo. The lower two rapidly results in blebbing and apoptosis of the cell.
panels show Drosophila S2 cells that have been transfected with a

Ras signal transduction pathway (Bergmann et al., ponents in the mechanism for cell killing in Drosophila
1998; Kurada and White, 1998). have been evolutionarily conserved (Table 1). The basic
Rpr, grim, and hid act cooperatively and combinato- apoptotic “machinery” was first characterized in C. el-
rially in what appears to be a cell lineage-specific man- egans following the identification of ced3, ced4, and
ner (Robinow et al., 1997; Zhou et al., 1997; Wing et al., ced9. ced3 encodes the caspase component of the C.
1998). For example, genetic analyses of deletions that elegans apoptotic program, and numerous caspases
remove hid alone, hid and grim but not rpr, or hid, have been identified in many different organisms in-
grim, and rpr indicate that all three death inducers are cluding Drosophila and humans (reviewed in Cryns
required for the normal pattern of cell death in the and Yuan, 1998). ced4 encodes a pro-apoptotic protein
embryonic central nervous system (CNS) midline that binds to, and is necessary for, the activation of the
(Zhou et al., 1997; Wing et al., 1998). Generally, each of caspase CED3. Apaf1, the vertebrate homologue of
these genes contributes to the overall level of apoptosis ced4, plays a similar role in that it interacts with and
(Fig. 3). Ectopic expression of either RPR or HID alone mediates the activation of caspase 9. Apaf1 differs from
can induce apoptosis. However, in the midline glia of CED4 in that Apaf1 contains a regulatory WD-repeat
the CNS expression of RPR or HID alone is insufficient domain and requires the binding of cytochrome C for
to induce apoptosis, whereas ectopic expression of both caspase activation. Recently, the Drosophila homo-
RPR and HID together results in the dramatic loss of logue of ced4/Apaf1 has been isolated and is identified
these cells, indicating that RPR and HID function syn- in the Drosophila database as Ark for apaf1 related
ergistically in the induction of programmed cell death. killer (Kanuka et al., 1999; Rodriguez et al., 1999; Zhou
Interestingly, the ectopic expression of GRIM alone is et al., 1999). ARK binds to the Drosophila caspases
sufficient to induce midline cell death, although GRIM DREDD and DRONC (Kanuka et al., 1999; Rodriguez
also acts synergistically with RPR and HID (Wing et et al., 1999), indicating that the proapoptotic activity of
al., 1998). ARK, like that of CED4 and Apaf1, is likely to mediate
Although vertebrate homologues for rpr, grim, and the activation of certain caspases. Like Apaf1, ARK
hid have not yet been identified, a number of key com- contains WD repeats and is positively regulated by
 APOPTOSIS IN DROSOPHILA DEVELOPMENT 71

Fig. 2. The patterns of rpr and hid expression in relation to the of apoptosis by 1 to 2 hr. Expression of hid expression corresponds
patterns of apoptosis. The pattern of apoptosis during various stages of generally to the locations of apoptosis, but it is more broadly expressed
embryonic development is shown in the upper row. In the second row, rpr in some regions. In addition, hid expression is not detected in the CNS at
expression, as assessed by in situs, can be seen to anticipate the pattern a time when significant numbers of these cells are dying.

cytochrome C. Mutations in Ark greatly suppress kill- death as well as killing by ectopically expressed RPR,
ing by ectopically expressed RPR, GRIM, and HID GRIM, and HID in the eye (Hay et al., 1995). dIAP1
(Kanuka et al., 1999; Rodriguez et al., 1999; Zhou et al., binds to a number of Drosophila caspases and inhibits
1999). However, Ark does not appear to be required for their activity (Kaiser et al., 1998; Hawkins et al., 1999;
viability, as mutants develop into adult flies, albeit Wang et al., 1999), and null mutations of dIAP1 are
with a number of morphologic abnormalities and a lethal very early in embryogenesis (Wang et al., 1999;
significant level of sterility in males (Rodriguez et al., Lisi et al., 2000), leading to the suggestion that contin-
1999). Ark is expressed in many, but not all, cells that uous levels of dIAP1 are necessary to inactivate consti-
die during development (Zhou et al., 1999), and it may tutively expressed caspases. RPR, GRIM, and HID
be that expression of Ark sensitizes cells to death sig- physically interact with dIAP1 (Vucic et al., 1997,
nals but is not absolutely required for death to occur. 1998), and binding by HID (presumably RPR and
Apoptosis in Drosophila is negatively regulated by a GRIM as well) negatively regulates the capacity of
class of evolutionarily conserved molecules known as dIAP1 to inhibit caspase activity (Wang et al., 1999). A
inhibitors of apoptosis, or IAPs. Originally identified in model for the induction of apoptosis by RPR, GRIM,
baculoviruses, IAPs have now been characterized from and HID includes binding of dIAP1 and thereby block-
a wide variety of organisms including Drosophila, ing the ability of dIAP1 to inhibit caspases, although
mice, and humans (reviewed in Deveraux and Reed, this may not be the only mechanism by which RPR,
1999). IAPs are characterized by a novel domain of GRIM, and HID initiate apoptosis. In addition, RPR
approximately 70 amino acids called the baculoviral has been shown to bind to and inactivate certain volt-
IAP repeat (BIR). A conserved arrangement of cys- age-gated K⫹ potassium channels and may initiate ap-
teines and histidines in the BIR binds a zinc ion to form optosis in some instances by inducing membrane depo-
a zinc finger-like motif (Hinds et al., 1999). IAPs can larization (Avdonin et al., 1998).
have up to three tandemly repeated BIR domains, and Five caspases have thus far been identified and char-
many of them also have another structural motif, the acterized in Drosophila (Fraser and Evan, 1997; Song
RING domain, located near the carboxy-terminus. et al., 1997; Chen et al., 1998; Dorstyn et al., 1999a,b).
Two IAPs, dIAP1 and dIAP2, have been identified in DREDD and DRONC are related to the initiator class
Drosophila (Hay et al., 1995). Overexpression of either of caspases, in that they have relatively long prodo-
dIAP1 or dIAP2 suppresses normal developmental cell mains and share substrate specificities with, and ho-
72 BANGS AND WHITE

Fig. 3. Apoptosis in embryos mutant for one or more of the proapop- grim are deleted the number of apoptotic cells is decreased further. In
totic regulators in 75C1,2. Embryos of similar stages were stained with the absence of hid, grim, and rpr, there is no apoptosis during embryo-
AO to visualize apoptosis. When embryos are null for hid, but wild type for genesis.
grim and rpr, there is a slight reduction in apoptosis. When both hid and

mology to, several of the mammalian initiator caspases gonads, and seem to have fragile tracheae. In addition,
(Chen et al., 1998; Dorstyn et al., 1999a). DCP-1, dcp-1 homozygous larvae have prominent melanotic
drICE, and DECAY on the other hand, have the shorter tumors indicative either of over-proliferation of blood
prodomains of the effector caspases and show a closer cells (which does not appear to be the case here; Song et
homology to this class (Fraser and Evan, 1997; Song et al., 1997) or of immune responses toward abnormal
al., 1997; Dorstyn et al., 1999b). DREDD is unique in cells or tissues within the larva (Watson et al., 1991).
that rather than the typical constitutive expression dcp-1 is also necessary for proper oogenesis, as germ-
observed for other caspases, its expression appears to line clones homozygous for dcp-1 result in sterile fe-
be upregulated in dying cells. This upregulation is males with arrested oogenesis phenotypes. Analysis of
tightly linked to signaling by rpr, grim, and hid (Chen the ovaries from these animals showed that nurse cell
et al., 1998).
dumping, the process by which nurse cells transfer the
To date, only one loss of function mutation has been
contents of their cytoplasm into the developing oocyte,
characterized for Drosophila caspases. dcp-1 mutants
was severely curtailed (McCall and Steller, 1998).
appear to have normal embryonic patterns of cell death
Dumping may be a modified form of apoptosis that
suggesting that zygotically derived dcp-1 is unneces-
sary for programmed cell death in the developing em- requires caspase activation.
bryo. This may be because of the abundant maternally A reduction in the dosage of dredd, achieved in flies
loaded dcp-1 or because other caspases are responsible heterozygous for a small genomic deletion that over-
for embryonic deaths (Song et al., 1997). dcp-1 alleles laps the gene, results in the suppression of killing
do cause larval lethality however, with most dcp-1 ho- induced by ectopic expression of either RPR or GRIM in
mozygotes dying before the third instar larval stage. the eye (Chen et al., 1998). This suggests that dredd is
Within those larvae that do reach the third instar a functional effector of the cell death pathway initiated
larval stage, several abnormalities are apparent. Al- by RPR and GRIM. Interestingly, reducing the dosage
though their central nervous systems appear to be nor- of dredd does not appear to effect the level of killing
mal, dcp-1 mutant larvae lack imaginal discs, have no induced by the overexpression of HID in the eye. This
 APOPTOSIS IN DROSOPHILA DEVELOPMENT 73

TABLE 1. Conserved Elements of the Apoptotic division patterns of the different mitotic domains con-
Machinery in Nematodes, Mammalian Systems, and tribute to the transformation of the monolayered blas-
Drosophila
toderm into the multi-layered gastrula and the devel-
C. elegans Mammals Drosophila opment of larval and adult structures.
Ced 3 Initiator caspases (CP)
Although cell death does not normally occur during
DED domain: CP-8, 10 Dredd these early stages of development (Abrams et al.,
1993), the stage is set at this point for apoptosis to play
CARD domain: CP-2, 9 Dronc
a vital role later. A general model for animal develop-
ment includes the accumulation of excess cells required
Effector caspases DrICE
CP-3, 6, 7 Dcp-1
for the formation of a viable embryo. Extraneous cells
Decay are then removed at later points by programmed cell
death (reviewed in Jacobson et al., 1997). Development
Ced 4 Apaf-1 Ark in Drosophila is consistent with this model, since early
divisions produce more cells than are needed for the
Proapoptotic bcl-2 Drob-1 subsequent development of specific structures. These
family proteins Others? extra cells are removed apoptotically at precise times
Ced 9 Antiapoptotic bcl-2 ?
along the developmental pathway. For instance, when
family proteins cell death is blocked in the absence of rpr, grim, and
hid, the embryonic nervous system contains a large
Bir-1 cIAP-1 dIAP1 excess of cells at the end of embryogenesis (White et al.,
Bir-2 cIAP-2 dIAP2 1994). Acridine orange staining shows that cell death
NAIP occurs in a consistent spatial and temporal pattern
XIAP (Abrams et al., 1993; Pazdera et al., 1998). However,
survivin
the precise locations and numbers of dying cells varies,
indicating that the number of cells being eliminated
? ? Hid
Grim depends on the developmental circumstances for a
Rpr given embryo.
The elimination of excess cells appears to be at least
in part related to intercellular signaling via the seg-
may be an indication that the cell death pathway ini- ment polarity genes. For instance, cell death observed
tiated by HID does not rely on dredd activity. in the epidermal segments at stage 12–14 occurs
among cells expressing the segment polarity gene en-
CELL DEATH DURING grailed or are immediately adjacent to cells that ex-
DROSOPHILA DEVELOPMENT press engrailed (Pazdera et al., 1998). When cell sig-
Although cell death has been recognized as a compo- naling is disrupted by mutations in the wingless
nent of Drosophila development for many years (Lock- signaling pathway, cell death among the ENGRAILED-
shin, 1969; Campos-Ortega and Hartenstein, 1985), expressing cells increases approximately 5-fold. This
specific roles for apoptosis have only recently been rec- increased death is seen in cells located approximately
ognized. As described below, programmed cell death is six rows away from WINGLESS-secreting cells but
a necessary feature of all phases of development, from does not occur in the cells that secrete WINGLESS or
early patterning of the embryo to the removal of obso- those adjacent to them. Mutants in armadillo, an in-
lete larval tissues and the final sculpting of adult tis- tracellular transducer of WINGLESS signaling, also
sues. show increased cell death in the same stripe of cells
seen with wingless mutants. A similar situation seems
Cell Death and Embryonic Pattern Formation to exist in the embryonic brain. When the wingless
Drosophila development begins in the fertilized egg gene is deleted in null mutants, the protocerebrum
with a series of 13 syncytial divisions, followed by three develops initially but is then largely deleted by pro-
rounds of cellular divisions for most cells. Cells are grammed cell death in later embryonic stages (Richter
allocated and positioned along dorsal-ventral and an- et al., 1998). Conversely, when WINGLESS is over-
terior-posterior axes according to maternally estab- expressed, a dramatic enlargement of the CNS is ob-
lished morphogen gradients. The fate of these cells served. These results indicate that one of the roles for
depends on their positions relative to the axes of the the segment polarity genes is to establish cell survival
egg. This results in the compartmentalization of the through a precise signaling mechanism. It may be that
developing embryo into specific domains of cells that in normal development excess cells are eliminated af-
will further develop into precise regions of the larva ter pattern formation because they are not in the ideal
and adult (reviewed in Lawrence and Struhl, 1996). position to receive the appropriate survival signals at
The generation of defined regions within the embryo the appropriate time.
includes the establishment of synchronously dividing Successful development requires the tight regulation
mitotic zones (Foe, 1989). The varied morphologies and of cell number to assure correct patterning and cell
74 BANGS AND WHITE

Fig. 4. Apoptosis can contribute to pattern repair. Multiple copies of extra copies of bicoid into a background where hid, grim, and rpr are
the anterior morphogen bicoid lead to an enlarged head in young em- deleted (6⫻ bcd; H99/H99, D), then the head is greatly enlarged at the
bryos. Increased apoptosis (arrow) can be seen in the head of these end of embryogenesis (6⫻ bcd; H99/H99, F) when compared to the
embryos (6⫻ bcd, B), when compared to the wild-type levels seen in A. phenotype of the deletion alone (H99/H99, E). Brackets denote the head
This contributes to the normal appearance and viability of the embryos by region.
the end of embryogenesis (6⫻ bcd, C). If apoptosis is blocked by putting

signaling. When the normal regulation of cell number terns in embryos with inappropriate dosages of bicoid
is disrupted, cell death increases to compensate for the shows that excess cells generated by an expansion of
extra cells. This has been examined by using ectopic the head domain are deleted by apoptosis (Namba et
expression of CYCLIN E to extend the normal cycles of al., 1997) (Fig. 4). These extra cell deaths occur on the
cell division by one extra round (Li et al., 1999). When same schedule as seen for wild-type embryos and ap-
this is done, the cell density in the embryo is nearly pear to be by the same mechanism, since increased rpr
doubled within 1 hr following the expression of expression is also observed in this domain (Namba et
CYCLIN E. Despite the huge excess of cells, the striped al., 1997). Interestingly, the compressed posterior do-
patterns of ENGRAILED and WINGLESS staining are main is repaired by a down-regulation of cell death,
unaltered when compared to wild-type embryos, and rather than by an increase in cell proliferation. This
most of the embryos develop and hatch into larvae, use of cell death regulation to repair both the expan-
albeit at a slower rate of development. Examination of sion and compression of fate maps reflects the elegance
cell death in these embryos shows that hyperplasia of the system.
induced by CYCLIN E expression is largely compen- There are limits to the extent of expansion and com-
sated for by increased apoptosis in the first 4 hr follow- pression that the embryo can tolerate. Tumorous struc-
ing the induction of CYCLIN E. The excess cells are tures and breaches in the epithelium are observed in
removed in a pattern-specific fashion rather than ran- domains that have expanded to the point where cell
domly, and the dying cells express rpr prior to their death can no longer completely eliminate the excess.
demise. Whether this is due to difficulties with signaling within
The plasticity of embryogenesis and the role of apo- an overly large cell mass or simply a matter of time
ptosis in this plasticity are apparent when examining running out before all of the extra deaths can be initi-
pattern repair in the Drosophila embryo. It has been ated remains unclear. Conversely, if the compression of
shown that the mis-expression of the anterior morpho- a developmental domain drops the cell number below a
gen bicoid results in the mispatterning of a number of certain threshold, structural defects occur or particular
anterior morphological markers. Eggs containing inap- organs fail to develop (Namba et al., 1997).
propriate doses of bicoid have a number of early devel-
opmental abnormalities. These include the shifting of Cell Death During Metamorphosis
the cephalic furrow either anteriorly, in embryos with The transition from larval to adult life in Drosophila
reduced bicoid levels, or posteriorly in embryos with is regulated by the steroid hormone 20-hydroxyecdys-
increased bicoid (Driever and Nüsslein-Volhard, 1988). one (ecdysone). The onset of metamorphosis is signaled
This latter shift results in the enlargement of head by a sharp rise in the ecdysone titer that initiates
structures in the early embryo (Fig. 4). Despite the puparium formation. During the next several hours,
rather significant departure from the normal pattern of most larval tissues undergo programmed cell death,
development, these embryos frequently develop into adult tissues begin to differentiate from imaginal pre-
healthy adults, indicating the embryo’s plasticity and cursor cells, and the ecdysone titer returns to basal
capacity for pattern repair. Analysis of cell death pat- levels (reviewed in Thummel, 1996). A second major
 APOPTOSIS IN DROSOPHILA DEVELOPMENT 75

ecdysone pulse initiates the start of pupation, during subsequent deaths of the neurons, can be suppressed
which the pupal cuticle degenerates and the adult cu- by treatment with ecdysone (Robinow et al., 1997).
ticle forms. As the pupal stage ends, the ecdysone titer These observations demonstrate the complexity of the
again returns to basal levels and the adult fly emerges. cell death response as regulated by steroid hormone
Larval tissue degeneration occurs in a stage- and levels. On the one hand, certain tissues respond to
tissue-specific manner that follows each ecdysone increasing levels of ecdysone by undergoing apoptosis,
pulse. For instance, the degeneration of the anterior as seen with the larval midgut and salivary glands.
larval muscles and the larval midgut follows the first Other tissues, for instance some larval neurons, re-
pupal ecdysone pulse, while histolysis of some abdom- spond in the opposite manner and die due to decreasing
inal muscles and the larval salivary glands occurs levels of ecdysone. These differential responses are at
shortly after the second major ecdysone pulse during least in part due to the ratio of ecdysone receptor iso-
pupation. Acridine orange staining and TUNEL assays forms expressed in these cells. Those cells that die
show that cells in these tissues initiate apoptosis just when the ecdysone titer increases express mainly the
prior to the first morphological indications of histolysis EcR-B isoforms, whereas those cells dying in response
(Jiang et al., 1997). Expression of the caspase inhibitor to a drop in ecdysone levels express primarily the
p35 blocks histolysis in these tissues, further confirm- EcR-A isoform.
ing that the loss of larval tissue is through apoptosis
(Jiang et al., 1997). CELL DEATH IN DEVELOPING
The response of tissues to ecdysone is mediated ADULT STRUCTURES
through the ecdysone receptor which is encoded by the Adult structures in Drosophila develop from imagi-
EcR gene (Koelle et al., 1991). Like steroid hormone nal discs of primordial cells. As in the embryo, cell fates
receptors in other organisms, EcR is a nuclear receptor in the imaginal discs are determined by positional pa-
that must heterodimerize with a retinoid X receptor rameters that relate to the numbers of cells in partic-
(RXR) in order to exert its regulatory functions (Oro et ular domains and cell signaling within and between
al., 1990; Yao et al., 1992; Hall and Thummel, 1998). domains. The development of a properly proportioned
The EcR gene encodes at least three protein isoforms structure requires the strict regulation of cell numbers,
(EcR-A, EcR-B1, and EcR-B2), which differ in their and programmed cell death plays a key role. In the
N-terminal sequences (Talbot et al., 1993). The obser- developing wing, a small number of epithelial cells are
vation that cells expressing different forms of the EcR undergoing apoptosis at any given time during the
respond differently to ecdysone has led to the hypoth- second and third larval instars and during pupal de-
esis that the ratio of distinct EcR isoforms within cells velopment (Milán et al., 1997). Cell death during this
dictates the specificity of the response to ecdysone period is seen in isolated single cells or in small clus-
(Robinow et al., 1993). This model is supported by ters of synchronously dying cells. This pattern is con-
experiments that show that the metamorphic re- sistent with the pattern of rpr expression (Nordstrom
sponses of various tissues are differentially effected in et al., 1996; Milán et al., 1997) and indicates that the
larvae mutant for the receptor EcR-B1 (Schubiger et cells die in a coordinated, non-random way. When the
al., 1998). Cells in which EcR-B1 is normally the pre- normal regulation of cell death is disrupted locally, the
dominant form fail to respond to ecdysone pulses, while result is an altered cell death pattern in other parts of
cells that typically express larger amounts of EcR-A the wing to compensate for disproportionate cell num-
undergo the morphogenetic changes expected of them bers. For example, when the expression of a ricin toxin
(Bender et al., 1997; Schubiger et al., 1998). is targeted to the posterior compartment of the devel-
Programmed cell death in larval tissues appears to oping wing, cell death in this compartment is increased
be transcriptionally regulated in response to ecdysone during the period of ricin expression. Cell death in the
levels. In both the larval midgut and the larval salivary posterior decreases following the inactivation of ricin,
gland, ecdysone pulses are followed promptly by the but an increase in cell death is then observed in the
expression of the cell-death genes rpr and hid and the anterior compartment (Milán et al., 1997). This ectopic
subsequent destruction of these tissues (Jiang et al., cell death is not due to diffusible pro-apoptotic factors
1997). Ecdysone treatment of isolated midgut and sal- released from dying cells but instead seems to be in
ivary gland tissue has also been shown to induce the response to the change in positional values caused by
transcription of the caspase dronc (Dorstyn et al., the abnormal loss of cells in the posterior domain.
1999). Furthermore, the anti-apoptotic diap 2 gene is Direct disruption of positional values by inappropriate
transcriptionally down-regulated following the steroid levels of morphogens can also result in apoptosis (for
pulses (Jiang et al., 1997). Following adult eclosion, example Bryant, 1988; Adachi-Yamada et al., 1999).
and a corresponding drop in ecdysone levels, certain Imaginal discs also accommodate for localized in-
larval neurons degenerate coincidentally with the eclo- creases in cell death by modulating cell proliferation.
sion-specific muscles they innervate (Robinow et al., Thus, an increase in cell proliferation follows ricin-
1993). These neurons accumulate rpr and grim tran- induced cell death in the posterior domain, while a
scripts in response to the drop in ecdysone levels, and decrease in normal cell proliferation is observed con-
the accumulation of these transcripts, as well as the comitantly with increased cell death in the anterior
76 BANGS AND WHITE

Fig. 5. Inhibition of EGF receptor signaling results in massive apo- red. In the wild-type disc very few cells undergo apoptosis during devel-
ptosis in the developing eye. Third instar eye discs were labeled with an opment. If a dominant negative version of the EGFR is expressed in
antibody to a nuclear neuronal antigen (elav), in green. The nuclear these cells, massive apoptosis is detected, and the regular array of
material of apoptotic cells is labeled with the TUNEL technique, shown in developing photoreceptors is disrupted.

compartment. If the expression of ricin is temporary, Evidence for this comes from the observation that ex-
the wing eventually develops into a correctly propor- pression of a dominant-negative form of EGFR in the
tioned adult-sized wing. Interestingly, when increased developing eye results in massive apoptosis (Freeman,
cell death is chronically maintained by continuous ricin 1996) (Fig. 5), as does the ectopic expression of argos, a
expression, the wing develops with up to 40% fewer diffusible inhibitor of the EGFR (Sawamoto et al.,
cells, but with the correct wing shape, proportions, and 1998). Furthermore, when EGFR signaling is main-
vein patterning (Milán et al., 1997). Correct propor- tained by ectopic expression of a constitutively acti-
tions in developing adult structures are thereby main- vated form of the receptor, cell death is dramatically
tained through a combination of cell proliferation and reduced in the eye, resulting in excess cells in the
cell death. As in embryos, if the number of cells in a ommatidial lattice (Miller and Cagan, 1998).
developing disc is reduced below a threshold, for exam- The signal mediated by EGFR activation is trans-
ple, by a cell cycle block, the ability of the tissue to duced through the Ras/Raf/MAPK pathway (Diaz-Ben-
compensate is overwhelmed, and defects in adult struc- jumea and Hafen, 1994; Freeman, 1997), suggesting
tures occur (de Nooij and Hariharan, 1995). that the anti-apoptotic activity of EGFR is dependent
on Ras signaling. This has been confirmed by the dem-
PATTERNING IN THE EYE AND onstration that the eye ablation phenotype resulting
EGFR SIGNALING from the ectopic expression of the death-inducing genes
The compound eye of the adult fly is a highly ordered rpr and hid is enhanced by a reduction in Ras activity,
array of approximately 750 ommatidia. Each ommatid- and suppressed by constitutively activated Ras or mu-
ium is composed of an invariant arrangement of cells tations that increase the signaling strength of the Ras
consisting of eight photoreceptor neurons, four lens- pathway (Bergmann et al., 1998; Kurada and White,
secreting cone cells, and two primary pigment cells. 1998). Ras signaling is thought to result in the phos-
Ommatidia are arrayed within a hexagonal lattice phorylation of HID, which apparently downregulates
composed of secondary and tertiary pigment cells. The its cell killing activity (Bergmann et al., 1998). A sig-
components of the ommatidia are specified in a stereo- nificant reduction in the transcription of hid is also
typical sequential order by the recruitment of undiffer- observed following increased activity of the Ras path-
entiated cells via signaling from their differentiated way (Kurada and White, 1998). Thus, cell survival in
neighbors. Thus the photoreceptors are specified in the the developing eye is mediated by signaling through
order R8, then pairwise R2 and R5, R3 and R4, and R1 the EGFR, resulting in the activation of the Ras/Raf/
and R6, followed at the end by the addition of R7. The MAPK pathway and a downregulation in both the lev-
cone cells are recruited next and then the primary, els and activity of the death-inducing HID.
secondary, and tertiary pigment cells in that order. One model for the development of the adult eye in-
Excess cells are subsequently removed by apoptosis, vokes the progressive recruitment and differentiation
resulting in the precise architecture of the adult eye of the various cell types of the ommatidia by reiterative
(reviewed in Wolff and Ready, 1993). signaling by the EGFR, modulated by secretion of the
Cell survival in the developing eye is regulated by activating and inactivating ligands SPITZ and AR-
cell-cell signaling through EGFR, the Drosophila ho- GOS, respectively (Freeman, 1997). Multiple functions
mologue of the epidermal growth factor (EGF) receptor. for the EGFR are suggested for this model including
 APOPTOSIS IN DROSOPHILA DEVELOPMENT 77

roles in cell proliferation, cell survival and cell differ- degeneration of obsolete larval tissues coincidentally
entiation (Dominguez et al., 1998). However, it is pos- with the development of adult structures. This work
sible that the role of EGFR signaling may be to specify provides an excellent opportunity to understand differ-
which of the cells in the developing retina survive, and ential sensitivity to death-inducing signals.
thus differentiate, and which cells are not needed, and Finally, it is likely that the regulation of apoptosis
thus die. While differentiation is aborted in the absence in a wide variety of developmental processes is more
of EGFR, it may be that the failure to differentiate complex and important than has been previously rec-
reflects the fact that the cells are dying rather than the ognized. Recent work has suggested that survival sig-
converse view that the cells are dying because they fail naling may play a significant role throughout develop-
to differentiate. This can only be addressed by exam- ment. For example, reiterative signaling through the
ining the consequences of the loss of EGFR in the EGF receptor integrates differentiation with cell death
absence of cell death. Considering the fact that EGFR to finely sculpt the compound eye of the fly. The con-
signaling protects against cell death by downregulat- tribution of apoptosis must be considered to fully un-
ing hid, it is likely that this question can be addressed derstand the phenotypes of many genes important in
by eliminating EGFR function in a hid null back- development.
ground. Our understanding of apoptosis during animal devel-
Cell death regulation by the EGFR may occur in a opment is still in its infancy. It is apparent that cell
number of developing tissues. For instance, mutations death occurs throughout development, and rudimen-
in the EGFR ligand spitz disrupt embryonic muscle tary maps detailing the spatial and temporal occur-
patterning, and targeted expression of a dominant neg- rence of cell death allow us to examine its conse-
ative EGFR in developing mesoderm results in the loss quences, and thus gain insight about specific functions.
of approximately half of the normal complement of The identification of many of the central genes required
myofibrils. Conversely, upregulation of EGFR activity to execute apoptosis provide the tools for the explora-
in the developing mesoderm leads to an increase in the tion of how this process is integrated with signaling
number of myogenic founder cells and the duplication pathways that specify proliferation or differentiation
of some adult muscle types (Buff et al., 1998). These during development.
observations suggest that, as has been proposed for
the developing eye, cell numbers are regulated by ACKNOWLEDGMENTS
EGFR signaling and that this regulation is necessary We thank Christian Peterson for the data shown in
for normal muscle development. In fact, the regulation Figure 3 and Phani Kurada for the data in Figure 5. PB
of cell number may prove to be a general role of EGFR is sponsored by an NIH training grant to MGH. KW’s
signaling. research is sponsored by a grant from the NIH and a
grant from the Shiseido Company of Japan to Massa-
CONCLUSIONS AND FUTURE DIRECTIONS chusetts General Hospital/Harvard Medical School.
Drosophila provides a powerful model system to in-
REFERENCES
vestigate both the mechanisms and roles of apoptosis
during animal development. Genetic manipulation al- Abrams J. 1999. An emerging blueprint for apoptosis in Drosophila.
lows the study of apoptotic regulators and effectors in Trends Cell Biol 9:435– 440.
Abrams JM, White K, Fessler LI, Steller H. 1993. Programmed cell
the context of the whole animal. In addition this work death during Drosophila embryogenesis. Development 117:29 – 43.
takes place in the context of a huge body of knowledge Adachi-Yamada T, Fujimura-Kamada K, Nishida Y, Matsumoto K.
on other aspects of Drosophila development. This pro- 1999. Distortion of proximodistal information causes JNK-depen-
vides the opportunity to understand how different cel- dent apoptosis in Drosophila wing. Nature 400:166 –169.
Avdonin V, Kasuya J, Ciorba M, Kaplan B, Hoshi T, Iverson L. 1998.
lular processes are coordinately regulated to form the
Apoptotic proteins Reaper and Grim induce stable inactivation in
developed organism. For example, apoptosis is neces- voltage-gated K⫹ channels. Proc Natl Acad Sci USA 95:11703–
sary to establish and maintain the correct cell densities 11708.
required for efficient and precise patterning mecha- Bate M, Martinez Arias A. 1993. The development of Drosophila
nisms. Coupled with cell proliferation, cell death en- melanogaster. New York: Cold Spring Harbor Press.
Bender M, Imam F, Talbot W, Ganetzky B, Hogness D. 1997. Dro-
sures a plasticity in developmental processes that al- sophila ecdysone receptor mutations reveal functional differences
lows an embryo to develop dynamically according to among receptor isoforms. Cell 91:777–788.
circumstances. The investigation of how cell division Bergmann A, Agapite J, McCall K, Steller H. 1998. The Drosophila
and apoptosis are coupled is likely to reveal fundamen- gene hid is a direct molecular target of Ras-dependent survival
signaling. Cell 95:331–341.
tal mechanisms of organogenesis.
Bergmann A, Agapite J, Steller H. 1998. Mechanisms and control of
The study of apoptosis during metamorphosis pro- programmed cell death in invertebrates. Oncogene 17:3215–3223.
vides a fertile area for the investigation of how differ- Bryant P. 1988. Localized cell death caused by mutations in a Dro-
ent tissues respond differently to the same hormonal sophila gene coding for a transforming growth factor-␤ homolog.
signal. The use of alternate isoforms of the ecdysone Developmental Biol. 128:386 –395.
Buff E, Carmena A, Gisselbrecht S, Jimenez F, Michelson A. 1998.
receptor coordinates the transcriptional regulation of Signaling by the Drosophila epidermal growth factor receptor is
apoptotic effectors in some cells with proliferation and required for the specification and diversification of embryonic mus-
differentiation in others. This results in the timely cle progenitors. Development 125:2075–2086.
78 BANGS AND WHITE

Campos-Ortega JA, Hartenstein V. 1985. The embryonic development Jiang C, Baehrecke E, Thummel C. 1997. Steroid regulated pro-
of Drosophila melanogaster. New York: Springer Verlag. grammed cell death during Drosophila metamorphosis. Develop-
Chen P, Nordstrom W, Gish B, Abrams JM. 1996. grim, a novel cell ment 124:4673– 4683.
death gene in Drosophila. Genes Dev. 10:1773–1782. Kaiser W, Vucic D, Miller L. 1998. The Drosophila inhibitor of apo-
Chen P, Rodriguez A, Erskine R, Thach T, Abrams J. 1998. Dredd, a ptosis D-IAP1 suppresses cell death induced by the caspase drICE.
novel effector of the apoptosis activators reaper, grim, and hid in FEBS Lett 440:243–248.
Drosophila. Dev Biol 201:202–216. Kanuka H, Sawamoto K, Inohara N, Matsuno K, Okano H, Miura M.
Claveria C, Albar J, Serrano A, Buesa J, Barbero J, Martinez-A C, 1999. Control of the cell death pathway by Dapaf-1, a Drosophila
Torres M. 1998. Drosophila grim induces apoptosis in mammalian Apaf-1/CED-4-Related Caspase Activator. Mol Cell 4:757–769.
cells. EMBO J 17:7199 –7208. Koelle M, Talbot W, Segraves W, Bender M, Hogness D. 1991. The
Coucouvanis E, Martin GR. 1995. Signals for death and survival: a Drosophila EcR gene encodes an ecdysone receptor, a new member
two step mechanism for cavitation in the vertebrate embryo. Cell of the steroid receptor superfamily. Cell 67:59 –77.
83:279 –287. Kurada P, White K. 1998. Ras promotes cell survival in Drosophila by
Cryns V, Yuan J. 1998. Proteases to die for. Genes Dev 12:1551–1570. downregulating hid expression. Cell 95:319 –329.
de Nooij JC, Hariharan IK. 1995. Uncoupling cell fate determination Lawrence P, Struhl G. 1996. Morphogens, compartments and pattern:
from patterned cell division in the Drosophila eye. Science. 270: Lessons from Drosophila? Cell 85:951–961.
983–985. Li Q, Pazdera T, Minden J. 1999. Drosophila embryonic pattern
Deveraux QL, Reed JL. 1999. IAP family proteins-suppressors of repair: how embryos respond to cyclin E-induced ectopic division.
apoptosis. Genes Dev 13:239 –252. Development 126:2299 –307.
Diaz-Benjumea FJ, Hafen E. 1994. The sevenless signaling cassette Lisi S, Mazzon I, White K. 2000. Diverse domains of DIAP1/THREAD
mediates Drosophila EGF receptor function during epidermal de- are required to inhibit apoptosis induced by REAPER and HID in
velopment. Development 120:569 –578. Drosophila. Genetics 154:669 – 678.
Dominguez M, Wasserman JD, Freeman M. 1998. Multiple functions Lockshin R. 1969. Programmed cell death. Activation of lysis by a
of the EGF receptor in Drosophila eye development. Curr Biol mechanism involving the synthesis of a protein. J. Insect Physiol
8:1039 –1048. 15:1505–1516.
Dorstyn L, Colussi P, Quinn L, Richardson H, Kumar S. 1999a. McCall K, Steller H. 1998. Requirement for DCP-1 caspase during
DRONC, an ecdysone-inducible Drosophila caspase. PNAS Drosophila oogenesis. Science 279:230 – 4.
96:4307– 4312. McCarthy J, Dixit V. 1998. Apoptosis induced by Drosophila Reaper
Dorstyn L, Read S, Quinn L, Richardson H, Kumar S. 1999b. DECAY, and Grim in a human system. J Biol Chem 273:24009 –24015.
a novel Drosophila caspase related to mammalian caspase-3 and Metzstein M, Stanfield G, Horvitz H. 1998. Genetics of programmed
caspase-7. J Biol Chem 274:30778 –3083. cell death in C. elegans: past, present and future. Trends Genet
14:410 – 416.
Driever W, Nüsslein-Volhard C. 1988. The BICOID protein deter-
Milán M, Campuzano S, Garcı́a-Bellido A. 1997. Developmental pa-
mines position in the Drosophila embryo in a concentration-depen-
rameters of cell death in the wing disc of Drosophila. Proc Natl Acad
dent manner. Cell 54:95–104.
Sci USA 94:5691–5696.
Evans E, Kuwana T, Strum S, Smith J, Newmeyer D, Kornbluth S.
Miller DT, Cagan RL. 1998. Local induction of patterning and pro-
1997. Reaper-induced apoptosis in a vertebrate system. EMBO J
grammed cell death in the developing Drosophila retina. Develop-
16:7372–7381.
ment 125:2327–2335.
Foe V. 1989. Mitotic domains reveal early commitment of cells in
Nagata S. 1997. Apoptosis by death factor. Cell 88:355–365.
Drosophila embryos. Development 107:1–22.
Namba R, Pazdera T, Cerrone R, Minden J. 1997. Drosophila embry-
Fraser A, Evan G. 1997. Identification of a Drosophila melanogaster
onic pattern repair: how embryos respond to bicoid dosage alter-
ICE/CED-3-related protease, drICE. EMBO J 16:2805–2813.
ation. Development 124:1393–1403.
Freeman M. 1996. Reiterative use of the EGF receptor triggers dif-
Newton K, Strasser A. 1998. The bcl-2 family and cell death regula-
ferentiation of all cell types in the Drosophila eye. Cell 87:651– 660.
tion. Curr Opin Genet Dev 8:68 –75.
Freeman M. 1997. Cell determination strategies in the Drosophila Nordstrom W, Chen P, Steller H, Abrams JM. 1996. Activation of the
eye. Development 124:261–270. reaper gene during ectopic cell killing in Drosophila. Dev Biol
Frisch S, Ruoslahti E. 1997. Intergrins and anoikis. Curr Opin Cell 180:213–226.
Biol 9:701–706. Oro A, McKeown M, Evans R. 1990. Relationship between the product
Glucksmann A. 1951. Cell deaths in normal vertebrate ontogeny. Biol of the Drosophila ultraspiracle locus and the vertebrate retiniod X
Rev 26:59 – 86. receptor. Nature 347:298 –301.
Grether ME, Abrams JM, Agapite J, White K, Steller H. 1995. The Pazdera T, Janardhan P, Minden J. 1998. Patterned epidermal cell
head involution defective gene of Drosophila melanogaster func- death in wild-type and segment polarity mutant Drosophila em-
tions in programmed cell death. Genes Dev 9:1694 –1708. bryos. Development 125:3427–3436.
Haining W, Carboy-Newcomb C, Wei C, Steller H. 1999. The proapop- Raff M. 1992. Social controls on cell survival and cell death. Nature
totic function of Drosophila Hid is conserved in mammalian cells. 356:397– 400.
Proc Natl Acad Sci USA 96:4936 – 4941. Richter S, Hartman B, Reichert H. 1998. The wingless gene is re-
Hall B, Thummel C. 1998. The RXR homolog ultraspiracle is an quired for early embryonic brain development. Dev Genes Evol
essential component of the Drosophila ecdysone receptor. Develop- 208:37– 45.
ment. 125:4709 –17. Robinow S, Draizen TA, Truman JW. 1997. Genes that induce apo-
Hawkins C, Wang S, Hay B. 1999. A cloning method to identify ptosis: transcriptional regulation in identified, doomed neurons of
caspases and their regulators in yeast: identification of Drosophila the Drosophila CNS. Dev Biol 190:206 –213.
IAP1 as an inhibitor of the Drosophila caspase DCP-1. Proc Natl Robinow S, Talbot WS, S. HD, Truman JW. 1993. Programmed cell
Acad Sci USA 96:2885–2890. death in the Drosophila CNS is ecdysone-regulated and coupled
Hay BA, Wassarman DA, Rubin GM. 1995. Drosophila homologs of with a specific receptor isoform. Development 119:1251–1259.
baculovirus inhibitor of apoptosis proteins function to block cell Rodriguez A, Oliver H, Zou H, Chen P, Wang X, Abrams J. 1999.
death. Cell 83:1253–1262. DARK, a Drosophila homolog of Apaf-1/ced-4, functions in an evo-
Hinds M, Norton R, Vaux D, Day C. 1999. Solution structure of a lutionarily conserved death pathway. Nature Cell Biol 1:272–279.
baculoviral inhibitor of apoptosis repeat. Nat Struct Biol 6:648 – Salvesen G, Dixit V. 1999. Caspase activation: the induced-proximity
651. model. Proc Natl Acad Sci USA 96:10964 –10967.
Hunter T. 1997. Oncoprotein networks. Cell 88:333–346. Saunders JW, Fallon JF. 1967. Cell death in morphogenesis. In: M.
Jacobson MD, Weil M, Raff MC. 1997. Programmed cell death in Locke, editor. Major problems in development biology. New York:
animal development. Cell 88:347–354. Academic Press. p 289 –314.
 APOPTOSIS IN DROSOPHILA DEVELOPMENT 79

Sawamoto A, Taguchi A, Hirota Y, Yamada C, Jin M, Okano H. 1998. Watson K, Johnson T, Denell R. 1991. Lethal(1) aberrant immune
Argos induces programmed cell death in the developing Drosophila response mutations leading to melanotic tumor formation in Dro-
eye by inhibition of the ras pathway. Cell Death Diff 5:262–270. sophila melanogaster. Dev Genet 12:173–187.
Schubiger M, Wade A, Carney G, Truman J, Bender M. 1998. Dro- White K, Grether ME, Abrams JM, Young L, Farrell K, Steller H.
sophila EcR-B ecdysone receptor isoforms are required for larval 1994. Genetic control of programmed cell death in Drosophila.
molting and for neuron remodeling during metamorphosis. Devel- Science 264:677– 683.
opment 125:2053–2062. White K, Tahaoglu E, Steller H. 1996. Cell killing by the Drosophila
Song Z, McCall K, Steller H. 1997. DCP-1, a Drosophila cell death gene reaper. Science 271:805– 807.
protease essential for development. Science 275:536 –540. Wing J, Zhou L, Schwartz L, Nambu J. 1998. Distinct killing proper-
Talbot W, Swyryd E, Hogness D. 1993. Drosophila tissues with dif- ties of the Drosophila reaper, head involution defective, and grim
ferent metamorphic responses to ecdysone express different recep- genes. Cell Death Diff 5:930 –939.
tor isoforms. Cell 73:1323–1337. Wolff T, Ready D. 1993. Pattern formation in the Drosophila retina.
Thummel C. 1996. Flies on steroids-Drosophila metamorphosis and In: Bate M, Martinez Arias A, editors. The development of Drosoph-
the mechanisms of steroid hormone actions. Trends Genet 12:306 – ila melanogaster. New York: Cold Spring Harbor Laboratory Press.
310. p 1277–1325.
Vaux D, Korsmeyer S. 1999. Cell death in development. Cell 96:245–254. Yao T, Segraves W, Oro A, McKeown M, Evans R. 1992. Drosophila
Vucic D, Kaiser WJ, Harvey AJ, Miller LK. 1997. Inhibition of Reaper- ultraspiracle modulates ecdysone receptor function via hetero
induced apoptosis by interaction with inhibitor of apoptosis proteins dimer formation. Cell 71:63–72.
(IAPs). Proc Natl Acad Sci USA 94:10183–10188. Zhou L, Schnitzler A, Agapite J, Schwartz LM, Steller H, Nambu JR.
Vucic D, Kaiser WJ, Miller LK. 1998. Inhibitor of apoptosis proteins 1997. Cooperative functions of the reaper and head involution defective
physically interact with and block apoptosis induced by Drosophila genes in the programmed cell death of Drosophila central nervous
proteins HID and GRIM. Mol Cell Biol 18:3300 –3309. system midline cells. Proc Natl Acad Sci USA 94:5131–5136.
Wang S, Hawkins C, Yoo S, Müller H-A, Hay B. 1999. The Drosophila Zhou L, Song Z, Tittel J, Steller H. 1999. HAC-1, a Drosophila ho-
caspase inhibitor DIAP1 is essential for cell survival and is nega- molog of APAF-1 and CED-4, functions in developmental and radi-
tively regulated by HID. Cell. 98:453– 463. ation-induced apoptosis. Mol Cell 4:745–755.