Full text of DOI:10.1016/S1097-2765(00)80437-2

Molecular Cell, Vol. 5, 423–433, March, 2000, Copyright  2000 by Cell Press
The ced-8 Gene Controls the Timing
of Programmed Cell Deaths in C. elegans
Gillian M. Stanfield† and H. Robert Horvitz*
Howard Hughes Medical Institute
Department of Biology
Massachusetts Institute of Technology
Cambridge, Massachusetts 02139
by Metzstein et al., 1998). The activities of the genes
ced-3 (cell death abnormal), which encodes a caspase
(Yuan et al., 1993; Xue et al., 1996), and ced-4, which
encodes a protein similar to mammalian Apaf-1 (Yuan
and Horvitz, 1992; Zou et al., 1997), are required for cells
to undergo programmed cell death in C. elegans (Ellis
and Horvitz, 1986). Cells destined to survive are pro-
tected by the activity of the gene ced-9, which encodes
a member of the Bcl-2 family of proteins (Hengartner
and Horvitz, 1994b); in dying cells, ced-9 is negatively
regulated by the gene egl-1 (egg-laying defective), which
encodes a BH3 (Bcl-2 homology 3) domain–containing
protein (Conradt and Horvitz, 1998).
As a cell dies by programmed cell death, morphologi-
cal changes in many aspects of its structure are visible
(Kerr et al., 1972). These changes define the process of
apoptosis. The plasma membrane blebs, and the cell
becomes detached from its neighbors. The dying cell
may fragment into smaller membrane-bound bodies. As
observed using electron microscopy (EM), the cyto-
plasm condenses and becomes darkly staining (electron
dense), while organelles remain grossly intact. The DNA
becomes condensed at the nuclear periphery and is
degraded by endonuclease activity; a cell corpse can
be visualized using the TUNEL (terminal transferase-
mediated dUTP nick end labeling) technique, which la-
bels DNA nicks in situ (Gavrieli et al., 1992). The cell
corpse is eventually engulfed and degraded within the
engulfing cell.
Summary
Loss-of-function mutations in the gene ced-8 lead to
the late appearance of cell corpses during embryonic
development in C. elegans. ced-8 functions down-
stream of or in parallel to the regulatory cell death
gene ced-9 and may function as a cell death effector
downstream of the caspase encoded by the pro-
grammed cell death killer gene ced-3. In ced-8 mu-
tants, embryonic programmed cell death probably ini-
tiates normally but proceeds slowly. ced-8 encodes a
transmembrane protein that appears to be localized
to the plasma membrane. The CED-8 protein is similar
to human XK, a putative membrane transport protein
implicated in McLeod Syndrome, a form of hereditary
neuroacanthocytosis.
Introduction
It is critical for animals to be able to eliminate unwanted
cells in a regulated, timely manner in response to devel-
opmental and environmental cues. This elimination is
accomplished through a tightly controlled process of
programmed cell death. The functions of specific cell
death genes within cells that die are required for their
demise, and a number of the key molecules that control
this process have been identified and are conserved
among animals (Horvitz, 1999; Vaux and Korsmeyer,
By most morphological criteria, cell corpses in C. ele-
gans are similar to cells dying by apoptosis in mammals,
although dying cells in C. elegans do not fragment into
smaller bodies. As visualized by Nomarski microscopy,
C. elegans cell corpses are highly refractile and easily
distinguished from living cells (Sulston and Horvitz,
1
977). These refractile corpses probably correspond to
1
999). Failure to kill appropriate cells can lead to severe
the electron-dense stage seen by EM (Robertson and
Thomson, 1982). The DNA of C. elegans cell corpses is
degraded (Sulston, 1976), and C. elegans cell corpses
can be visualized using TUNEL (Wu et al., 2000).
developmental defects (Kuida et al., 1996, 1998; Hakem
et al., 1998; Varfolomeev et al., 1998).
In the Caenorhabditis elegans hermaphrodite, 131 of
the 1090 somatic cells generated during development
are destined to die (Sulston and Horvitz, 1977; Sulston
et al., 1983). The identities of the cells that undergo
programmed cell death in C. elegans are essentially
invariant from animal to animal. The majority of cell
deaths occur early in development during embryonic
stages coincident with a period of rapid cell division and
the initiation of morphogenesis. In genetic screens for
mutants abnormal in the pattern or appearance of cell
corpses, C. elegans mutants variously defective in the
specification, execution, engulfment, or degradation of
programmed cell deaths have been identified. Analyses
of these mutants have provided a framework for under-
standing how cells die by programmed cell death, not
only in C. elegans but also in other animals (reviewed
Although genes important for controlling the killing
step of programmed cell death have been well charac-
terized, it is not clear how downstream events, such as
the morphological changes that occur as cells die and
the engulfment and degradation of cell corpses, are
effected. Caspases are thought to be key mediators of
these downstream events (e.g., Liu et al., 1997; Janicke
et al., 1998), but while many in vitro targets of caspase
proteolysis have been identified, the functions of these
targets in cell death are in general unclear. DNA degra-
dation may be triggered by the caspase-dependent
cleavage of a protein that binds to and inhibits a cell
death endonuclease (Enari et al., 1998; Sakahira et al.,
1998). Several molecules implicated in the engulfment
process have been characterized recently (Franc et al.,
1
1
996; Liu and Hengartner, 1998; Wu and Horvitz, 1998a,
998b). Nonetheless, it is not known how the initiation
*
To whom correspondence should be addressed (e-mail: horvitz@
of programmed cell death leads to the engulfment of a
dying cell or whether the triggering of the engulfment
process is a direct or indirect result of caspase activity.
mit.edu).
†
Present address: Department of Developmental Biology, Stanford
University School of Medicine, Stanford, California 94305.

Molecular Cell
24
4
Figure 1. Delayed Appearance of Programmed
Cell Deaths in ced-8 Animals
(
(
(
(
A) N2.
B) ced-7(n1892); ced-5(n1812).
C) ced-8(n1891).
D) ced-7(n1892); ced-5(n1812); ced-8(n1891).
The y axis represents the average number
of corpses visible in the heads of embryos.
Stages of embryos examined: bean and
comma stage embryos (B/C); 1.5-fold and
2
-fold embryos (1.5/2ϫ); early 3-fold em-
bryos, which are 3.5 times the length of the
egg but do not yet have a well-developed
pharynx (3ϫE); late 3-fold embryos, which
have a well-developed pharynx and grinder
(3ϫL); and early L1 larvae with four cells in the
gonad (L1). Error bars indicate one standard
error of the mean. At least 14 embryos of each
stage were examined.
It is also unclear whether any downstream molecules in
the cell death process are activated by caspase-inde-
pendent mechanisms. The conservation among diverse
organisms of both the morphology of dying cells and
the molecules implicated in cell killing suggests that
aspects of cell killing downstream of the key regulatory
molecules are likely also to be conserved. For this rea-
son, genetic analysis of programmed cell death in C.
elegans, which has identified these regulatory mole-
cules, should also permit the characterization of the
events that occur downstream of cell death initiation
and of how these downstream events contribute to the
cell-killing process.
In this report, we demonstrate that the gene ced-8 is
involved in the execution step of programmed cell death
in C. elegans and that the function of ced-8 is important
for the proper kinetics of programmed cell death during
development. We cloned the ced-8 gene and found that
it is similar to XK, a putative membrane transport protein
from humans (Ho et al., 1994). We suggest that ced-8
functions in the downstream processes of cell killing.
from its living neighbors. The refractile disk-like appear-
ance of the dying cell disappears during the engulfment
process, which occurs within an hour of the initiation of
the cell’s death. The total number of cell corpses visible
at any time during development therefore represents
the difference between the number of cells that have
died up to that time and the number of cell corpses that
have been engulfed and degraded.
ced-8 was identified in a screen for C. elegans mutants
containing an increased number of cell corpses at the
late embryonic and early larval stages (Ellis et al., 1991).
Other mutants identified in this screen are defective in
the engulfment of cell corpses by neighboring cells. In
such engulfment-defective mutants, cell deaths occur
at the appropriate time during early embryogenesis, but
cell corpses persist in an unengulfed state (Hedgecock
et al., 1983; Ellis et al., 1991) (Figure 1B). By contrast,
we found that in ced-8 mutants the majority of cell
corpses did not appear until late in embryogenesis, and
the greatest number of cell corpses was visible in late
3-fold embryos (Figure 1C; Table 1). Furthermore, nearly
all of these cell corpses disappeared by the time of
hatching, suggesting that they are engulfed rapidly, as
usual. We observed a clear shift toward a later appear-
ance of cell corpses than in the wild type for all ced-8
alleles examined with the exception of n3113, which
was identified as a suppressor of the lethality conferred
by ced-4(n2273) ced-9(n1653) (Shaham and Horvitz,
1996a; E. K. Speliotes and H. R. H., unpublished data)
and was nearly indistinguishable from the wild type in
cell corpse assays (Table 1).
Results
The Appearance of Cell Corpses Is Delayed
in ced-8 Embryos
During the approximately 14 hr of C. elegans em-
bryogenesis, cell divisions and cell deaths together re-
sult in a total of 558 cells as the developing animal is
transformed from a single cell into a vermiform larva
capable of movement and feeding (Sulston et al., 1983).
Most somatic cells destined to die in C. elegans do so
soon after they are generated, and most cell corpses
are visible during the developmental period between the
bean and 2-fold stages when many cells are dividing
and the embryo begins to elongate (Sulston and Horvitz,
To confirm that the appearance of cell corpses is
delayed in ced-8 mutants, we examined the effect of a
ced-8 mutation on cell corpse number in a ced-7; ced-5
background, in which most cell corpses remain unen-
gulfed (Ellis et al., 1991). We found that most cell corpses
in ced-7; ced-5; ced-8 embryos were visible during the
3-fold stage of embryogenesis, just as in the ced-8 single
mutants (Figure 1D). We saw no more corpses in early
1
977; Sulston et al., 1983) (Figure 1A). As viewed by
differential interference contrast (Nomarski) micros-
copy, a dying cell becomes highly refractile and distinct

Programmed Cell Death Timing Gene ced-8
25
4
Table 1. ced-8 Alleles
Number of
Corpses Early
Number of
Corpses Late
Number of
Extra Cells
Range of
Extra Cells
a
b
c
d
Genotype
Wild type
ced-8(n1891)
ced-8(n1999)
ced-8(n2090)
ced-8(n2093)
ced-8(n3113)
ced-8(n3115)
ced-8(n3244)
ced-8(n3245)
ced-8(n3313)
lon-2(e678) ced-8(n1891)/yDf2
lon-2(e678)/yDf2
10.6 Ϯ 0.6
0.9 Ϯ 0.2
0.4 Ϯ 0.1
1.8 Ϯ 0.3
1.8 Ϯ 0.3
9.7 Ϯ 0.6
1.9 Ϯ 0.2
0.7 Ϯ 0.2
1.4 Ϯ 0.2
0.9 Ϯ 0.2
ND
2.3 Ϯ 0.5
19.5 Ϯ 1.1
12.0 Ϯ 0.5
19.0 Ϯ 1.2
19.5 Ϯ 1.5
2.1 Ϯ 0.4
16.7 Ϯ 0.8
19.5 Ϯ 1.1
17.1 Ϯ 1.0
17.5 Ϯ 0.7
ND
0.04 Ϯ 0.03
0.7 Ϯ 0.1
0.4 Ϯ 0.09
0.8 Ϯ 0.1
0.6 Ϯ 0.1
0.1 Ϯ 0.06
0.4 Ϯ 0.1
0.7 Ϯ 0.2
0.9 Ϯ 0.2
0.8 Ϯ 0.2
0.8 Ϯ 0.2
0.1 Ϯ 0.07
0–1
0–3
0–2
0–3
0–3
0–1
0–2
0–2
0–4
0–3
0–3
0–1
ND
ND
a
Average number of corpses in the head of bean and comma stage embryos Ϯ SEM. ND, not determined.
Average number of corpses in the head of 3-fold stage embryos Ϯ SEM.
Average number of extra cells in the anterior pharynx of L3 and L4 hermaphrodites Ϯ SEM.
b
c
d
Range in number of extra cells observed in the anterior pharynx. Number of animals examined: 50 for the wild type and 20–73 for other genotypes.
ced-7; ced-5; ced-8 embryos than in early ced-8
embryos, consistent with the interpretation that cell
corpses appear late, rather than persist, in animals mu-
tant for ced-8. Furthermore, the total number of cell
corpses that accumulated by the late 3-fold stage of
embryogenesis in ced-7; ced-5; ced-8 animals was simi-
lar to the maximum number observed in ced-7; ced-5
animals, consistent with the idea that a delay rather than
a reduction in cell killing is caused by mutations in
ced-8.
Meyer, personal communication). We found that there
was no significant difference in the number of extra
cells in ced-8(n1891)/yDf2 animals as compared to the
number in ced-8(n1891) homozygotes (Table 1). These
results are consistent with n1891s being a strong loss-
of-function or null allele of ced-8.
Although mutations in ced-8 alone did not result in
the survival of a significant number of cells that normally
die, it was possible that ced-8 might enhance the weak
cell survival phenotype that results from partial loss of
ced-3 or ced-4 function. We scored for the presence of
extra cells in the anterior pharynx of animals mutant
both for ced-8 and for one of the weak ced-3 mutations
n2427 or n2438. We found that a ced-8 mutation and a
weak ced-3 mutation had a synergistic effect on the
number of extra cells in the anterior pharynx (Table 2).
We also examined animals mutant for both ced-8 and
ced-4(n2273), a weak ced-4 mutation that results in the
survival of only a few cells that normally die. ced-
4(n2273); ced-8(n1891) animals also contained more ex-
tra cells than did either of the single mutants (Tables 1
and 2). Thus, mutations in ced-8 enhance the cell sur-
vival phenotype of all weak ced-3 and ced-4 mutants
tested. Double mutants between ced-8 and the strong
alleles ced-3(n1040), ced-3(n717), or ced-4(n1162) had
no more extra cells than did the ced-3 or ced-4 single
mutant animals (Table 2).
ced-8 Mutations Weakly Diminish Cell Killing
Strong loss-of-function (lf) mutations in ced-3 and ced-4
result in the survival of nearly all cells that normally die
in C. elegans (Ellis and Horvitz, 1986). Weak loss-of-
function mutations in ced-3 and ced-4 result in the sur-
vival of a variable and apparently random subset of
those cells that normally die (Hengartner and Horvitz,
1
994a; Shaham et al., 1999; M. O. Hengartner and S.
Shaham, personal communication) and in the delay of
some cell deaths (M. O. Hengartner and H. R. H., unpub-
lished data). Based on the number of corpses in em-
bryos, ced-8 mutations did not appear to cause many
cells to survive ectopically. However, since weak muta-
tions in ced-3 and ced-4 affect cell death kinetics and
also weakly affect the cell-killing process, we assessed
quantitatively whether ced-8 mutations block cell death.
Counts of extra cells in the anterior region of the phar-
ynx, the nematode feeding organ, are a convenient mea-
sure of the degree to which cell death is blocked in a ced
mutant (Hengartner et al., 1992). For instance, strong
mutations in ced-3 or ced-4 result in approximately 13
extra cells in this region of the animal (Table 2). We
found that very few, if any, extra cells were present in
ced-8 mutants as compared to strong ced-3 and ced-4
mutants (Tables 1 and 2).
One possible explanation for a weak ced-8 cell sur-
vival phenotype was that the known ced-8 mutations
might cause only a partial loss of function of this gene.
To test whether stronger ced-8(lf) mutations might cause
a more pronounced cell survival phenotype, we scored
for extra cells in the pharynges of ced-8(n1891)/yDf2
animals (yDf2 spans the ced-8 locus; L. Miller and B.
These data are consistent with a modulatory function
for ced-8 in cell killing. ced-8 appears to play a role in
cell death that is distinct from those of ced-3 and ced-4:
ced-8 is not absolutely required for cell killing, but it can
enhance the probability of dying for a cell poised near
the threshold of programmed cell death.
ced-8 Acts Downstream of or in Parallel to ced-9
The gene ced-9 is required for cell survival during devel-
opment and for fertility: ced-9(lf) mutations result in ec-
topic cell deaths leading to maternal-effect lethality and
in zygotic sterility (Hengartner et al., 1992). Mutations in
either ced-3 or ced-4 suppress the ced-9(lf) phenotype,
indicating that the lethality results from ectopic pro-
grammed cell deaths and that ced-9 normally functions
to negatively regulate ced-3 and ced-4 activity. We

Molecular Cell
26
4
Table 2. ced-8 Enhances Weak ced-3 and ced-4 Alleles
a
b
c
Genotype
Number of Extra Cells
Range
n
ced-3(n2427)
ced-8(n1891)
ced-8(n1999)
ced-3(n2427); ced-8(n1891)
ced-3(n2427); ced-8(n1999)
ced-3(n2438)
ced-3(n2438); ced-8(n1891)
ced-3(n2438); ced-8(n1999)
ced-3(n1040)
ced-3(n1040); ced-8(n1891)
ced-3(n1040); ced-8(n1999)
ced-3(n717)
ced-3(n717); ced-8(n1891)
ced-4(n2273)
ced-4(n2273); ced-8(n1891)
ced-4(n2273); ced-8(n1999)
ced-4(n1162)
1.0 Ϯ 0.1
0.7 Ϯ 0.1
0.4 Ϯ 0.1
5.2 Ϯ 0.3
4.8 Ϯ 0.3
2.0 Ϯ 0.3
6.5 Ϯ 0.3
5.0 Ϯ 0.3
11.5 Ϯ 0.3
10.3 Ϯ 0.5
11.1 Ϯ 0.4
12.8 Ϯ 0.4
12.3 Ϯ 0.3
3.9 Ϯ 0.3
7.1 Ϯ 0.4
7.0 Ϯ 0.4
13.1 Ϯ 0.3
13.3 Ϯ 0.4
0–4
0–3
0–2
3–10
2–8
0–7
3–11
3–8
8–14
6–16
8–15
8–18
9–15
3–6
42
73
50
30
29
28
32
26
33
23
24
30
30
10
10
10
31
28
d
e
d
5–9
5–8
10–16
9–16
d
ced-4(n1162); ced-8(n1891)
a
Average number of cells in the anterior pharynx of L3 and L4 hermaphrodites Ϯ SEM.
Range in number of extra cells observed in the anterior pharynx.
n, number of animals examined.
These strains contained the mutation lon-2(e678), which does not affect cell survival.
Although 16 cells die in the anterior region of the pharynx, 18 extra cells were observed in one animal of this genotype. These cells likely
b
c
d
e
originated in the posterior bulb of the pharynx, since the pharynx is separated from other cells by a basement membrane. Extra cells are
often slightly misplaced in strong ced mutants.
sought to determine whether ced-8 mutations similarly
could suppress the ced-9(lf) phenotype. We found that
the ced-8(n1891) and ced-8(n1999) mutations partially
suppressed the loss of fertility and developmental de-
fects of ced-9(n1950 n2161) mutants. For example, ced-
DNA Degradation Is Somewhat Delayed
in ced-8 Mutants
How might we distinguish whether ced-8 mutants initi-
ate programmed cell death late or execute the process
of programmed cell death slowly? If programmed cell
deaths were initiated late, then the appearance of all
markers of cell death should be delayed, but once killing
is triggered, the morphological changes that result in a
refractile corpse would presumably occur with normal
kinetics. However, if programmed cell deaths were initi-
ated at the normal time in ced-8 mutants, then some
cell death markers could appear at the appropriate time
even though the progression of the cell to refractility is
slowed.
Therefore, we sought additional biochemical markers
of cell corpses besides the refractility visible by Nomar-
ski microscopy. Degradation of the DNA of dying cells
is a hallmark of apoptosis and can be visualized in situ
using TUNEL (Gavrieli et al., 1992; Wu et al., 2000). In
wild-type C. elegans embryos, a small fraction of dying
cells are TUNEL reactive at any one time; by contrast,
embryos mutant for the gene nuc-1, which is important
for DNA degradation (Sulston, 1976; Hevelone and Hart-
man, 1988), contain many condensed TUNEL-reactive
nuclei that appear when cells die and persist (Figures
2A and 2B; Wu et al., 2000). Thus, DNA degradation is
apparently very rapid during programmed cell death in
9
(n1950 n2161) unc-69(e587); ced-8 homozygous
mothers laid more eggs than did ced-9(n1950 n2161)
unc-69(e587) mothers, and ced-9(n1950 n2161) unc-
6
9(e587); ced-8 embryos developed to later embryonic
stages than did ced-9(n1950 n2161) unc-69(e587) em-
bryos (data not shown). However, although ced-9(n1950
n2161) unc-69(e587); ced-8 animals occasionally sur-
vived until hatching, none progressed past the L1 larval
stage, so mutations in ced-8 improved the viability of
ced-9(lf) animals but did not completely suppress their
defects. To test whether ced-8 acts genetically up-
stream of ced-9, we determined whether ced-9 function
is required for the enhanced cell survival phenotype of a
ced-3(weak); ced-8 double mutant. We found that more
extra cells were present in ced-9(n2812); ced-3(n2427);
ced-8(n1891) mutant animals than were present in ced-
9
(n2812); ced-3(n2427) animals (p Ͻ 0.0001, paired t
test) (Table 3). ced-9(n2812) is likely to be null for ced-9
function (Hengartner and Horvitz, 1994a). Thus, muta-
tions in ced-8 probably can enhance cell survival even
in the absence of ced-9 activity, indicating that ced-8
acts downstream of or in parallel to ced-9.
Table 3. ced-8 Functions Downstream of or in Parallel to ced-9
Number of
Extra Cells
a
b
Genotype
n
ced-9(n2812); ced-3(n2427)
ced-9(n2812); ced-3(n2427); unc-10(e102) ced-8(n1891)
7.1 Ϯ 0.4
10.0 Ϯ 0.3
18
21
a
Average number of extra cells in the anterior pharynx in L3 and L4 hermaphrodites Ϯ SEM.
n, number of animals examined.
b

Programmed Cell Death Timing Gene ced-8
27
4
the greatest number of TUNEL-reactive corpses was
present in late-stage ced-8 nuc-1 embryos. However, in
early embryos, there were many more corpses in ced-8
nuc-1 embryos as visualized by TUNEL staining than
there were refractile corpses in ced-7; ced-5; ced-8 mu-
tant embryos as visualized by Nomarski optics (Figure
2
D versus Figure 1D). That so many TUNEL-positive
cells were present at this early stage suggests that al-
though morphological changes causing refractility are
delayed in most cell corpses until the 3-fold stage of
embryogenesis, DNA degradation is less delayed. This
result is consistent with the hypothesis that cell death
is triggered at the appropriate developmental time in
ced-8 mutants but proceeds to the refractile corpse
stage more slowly than in the wild type and that the
biochemical and cytological changes of dying cells that
occur nearly synchronously in the wild type are uncou-
pled from one another in ced-8 mutants.
Cloning of ced-8
ced-8 previously had been mapped to LGX to the right of
unc-10, near dpy-6 (Ellis et al., 1991). We more precisely
mapped ced-8 to the left of dpy-6, which lies to the left
of the polymorphism nP81 (Kornfeld et al., 1995). We
injected cosmids from the region to the left of nP81
(Figure 3A) into ced-8(n1999) hermaphrodites and found
that arrays containing either of the cosmids F54F4 or
F37F1 restored cell corpses to early embryos. We local-
ized the region required for rescue to a 4.2 kilobase (kb)
fragment from the region of overlap between F54F4 and
F37F1 (Figure 3B). We determined the sequence of this
DNA fragment and found that it contains a single gene
predicted by the program GENEFINDER (C. elegans Ge-
nome Sequencing Consortium, 1998).
We identified a cDNA (Figure 3C; see Experimental
Procedures) that rescued the Ced-8 phenotype when
expressed in ced-8(n1891) embryos under the control
of C. elegans heat-shock promoters (data not shown).
This transcript presumably corresponds to ced-8. To
determine whether we had identified the 5Ј end of this
rescuing transcript, we analyzed 5Ј RACE products from
C. elegans embryonic RNA and performed Northern
analysis. Putative full-length ced-8 transcripts identified
by RACE were 1576 bases in length and had 41 addi-
tional nucleotides of contiguous genomic sequence as
compared to the longest cDNAs isolated from phage
libraries. A Northern blot identified a single band migrat-
ing at 1.4–1.6 kb, the approximate size of the putative
full-length transcript, in RNA from embryos but not from
mixed stage animals (data not shown). Since the ced-8
transcript has two in-frame stop codons 5Ј to the puta-
tive transcriptional start, it is likely that we have identi-
fied the complete ced-8 coding sequence.
Figure 2. Delayed Onset of DNA Degradation in ced-8 Embryos
The genotypes indicated are for the ced-8 and nuc-1 loci. All four
strains also contained the mutation egl-15(484), which does not
affect TUNEL staining (data not shown), as indicated: (A) egl-
1
1
5(n484); (B) nuc-1(e1392) egl-15(n484); (C) ced-8(n1891) egl-
5(n484); (D) ced-8(n1891) nuc-1(e1392) egl-15(n484). The y axis
represents the number of TUNEL-positive cells present in embryos.
Stages examined: bean and comma stage embryos (B/C), 1.5-fold
embryos (1.5ϫ), and 3-fold embryos including both early and late
stages (3ϫ). Error bars indicate one standard error of the mean. At
least 15 embryos of each stage were examined.
wild-type C. elegans, and DNA degradation is slowed
in nuc-1 mutants relative to in the wild type. Furthermore,
engulfment per se is not necessary for the generation
of TUNEL-reactive DNA in cell corpses (Wu et al., 2000),
suggesting that TUNEL provides a cell-intrinsic marker
for cell deaths. We therefore examined whether muta-
tions in ced-8 affect the kinetics of DNA degradation as
visualized by TUNEL.
In ced-8 embryos, as in wild-type embryos, few TUNEL-
reactive corpses were present at any given develop-
mental stage (Figures 2A and 2C). However, in a ced-8
nuc-1 double mutant strain, many TUNEL-reactive
corpses were present late in embryogenesis (Figure 2D),
suggesting that the low number of TUNEL-reactive
corpses in the ced-8 single mutant was not caused by
a block in DNA degradation but rather was caused by
rapid progression through the DNA degradation process
(Figure 2D).
To compare the time course of the appearance of
TUNEL-reactive corpses with the time course of the
appearance of refractile cell corpses, we examined
TUNEL-stained ced-8 nuc-1 double mutant animals at
various developmental time points during embryogene-
sis and compared the number of TUNEL-positive cells
to the number of refractile corpses visible in a ced-7;
ced-5; ced-8 mutant strain at the same developmental
stage. The nuc-1 DNA degradation-defective and the
ced-7; ced-5 engulfment-defective backgrounds allow
a similar number of accumulated cell corpses to be
visualized at each embryonic stage (Figure 2B versus
Figure 1B), suggesting that these two methods of visual-
izing corpses have similar sensitivities. We found that
To confirm that we had identified the ced-8 gene, we
determined the sequences of the coding region from
ced-8 mutants and identified mutations for all nine al-
leles (Figure 3C). The nature of the lesions in the ced-8
alleles are consistent with their causing loss of ced-8
function. Eight of the ced-8 alleles cause an identical
phenotype that is consistent with their being strong or
complete loss of function. The n3113 mutation, which
eliminates the splice acceptor for intron 3 and is pre-
dicted to result in an in-frame insertion of 16 amino

Molecular Cell
28
4
Figure 3. ced-8 Cloning and cDNA Sequence
A) Cosmids assayed for transformation rescue. Rescuing cosmids are in bold.
B) Subclones assayed for transformation rescue; the fraction of assayed transgenic lines showing rescue of the Ced-8 phenotype is indicated
on the right. A, AgeI; H, HpaI; M, MluI; N, NheI.
C) A composite of the C. elegans ced-8 sequence derived from cDNA and 5Ј RACE analysis. Intron positions are indicated by vertical bars.
(
(
(
The predicted CED-8 protein sequence is shown beneath. Arrowheads designate base changes in the indicated ced-8 alleles. The ced-8(n3113)
allele (*) is predicted to eliminate the splice acceptor for intron 3, resulting in an in-frame insertion of the sequence VTHFLGSVRKTKSDLQ into
†
the CED-8 protein between amino acids 137 and 138. The ced-8(n2090) allele ( ) contains two nucleotide changes relative to the wild-type
sequence. Two in-frame stop codons upstream of the predicted initiation codon are underlined; the termination codon is also underlined.
acids, results in a weaker phenotype: very few cell
deaths are delayed until late stages of embryogenesis
in n3113 embryos (Table 1).
have been cleared from the animal by engulfment (Ellis
et al., 1991). That screen primarily identified mutants
with defects in cell corpse engulfment. ced-8 mutants,
like the engulfment-defective mutants identified in that
screen, contain visible cell corpses at these late stages,
and ced-8 was assumed initially to be defective in the
engulfment step of programmed cell death. However, by
examining earlier stages of development, we discovered
that in ced-8 embryos corpses were not visible at the
stages at which they normally appear both in the wild
type and in engulfment mutants; rather, corpses were
present at later stages than in the wild type. This obser-
vation revealed that the presence of cell corpses in
ced-8 L1 larvae reflects a delay in cell killing rather than
a defect in engulfment.
Our previous ultrastructural studies indicated that cell
corpses generated postembryonically in the ventral cord
of ced-8 mutants are engulfed (Ellis et al., 1991), consis-
tent with our current conclusion that the defect of ced-8
mutants is in cell killing rather than cell corpse en-
gulfment. However, we also reported previously that cell
corpses that are normally generated during early to mid
embryogenesis are not engulfed in recently hatched
ced-8 L1 larvae. We now understand that these corpses
are present at this stage simply because they have just
been generated. Specifically, in ced-8 mutants, embry-
onic deaths are delayed until very late embryogenesis
CED-8 Is Similar to XK
The ced-8 cDNA encodes a 458 amino acid protein that
is weakly similar to the product of the human XK gene
(Figure 4A) (Ho et al., 1994). CED-8 and XK share 19%
identity and 37% similarity overall and have similar hy-
dropathy plots (Figure 4B). Both proteins contain 10
hydrophobic predicted transmembrane-spanning seg-
ments, and most sequence similarity between the two
proteins lies within these regions. Consistent with this
structural prediction for CED-8, a ced-8:gfp (green fluo-
rescent protein) translational fusion, which includes
most of the ced-8 coding region and rescued the de-
layed death phenotype of ced-8(n1999) embryos (data
not shown), was localized to the plasma membrane (Fig-
ure 5). Two regions of similarity between CED-8 and XK
are shared with a predicted protein from the ascidian
Ciona intestinalis, and one of these regions is also
shared with a Drosophila EST (Figure 4C).
Discussion
ced-8 was discovered in a screen for animals containing
cell corpses in L1 larvae, a stage at which few pro-
grammed cell deaths occur and embryonic cell corpses

Programmed Cell Death Timing Gene ced-8
29
4
Figure 4. CED-8 Protein Sequence Comparisons
A) Alignment of the full-length proteins C. elegans CED-8, human XK (hXK) and mouse XK (mXK; Collec et al., 1999). Identities are shaded
black; similarities are shaded grey. Predicted transmembrane-spanning domains in CED-8 are indicated with brackets.
B) Kyte-Doolittle hydropathy plots (Kyte and Doolittle, 1982) (scanning window, seven amino acids) of CED-8 and human XK. Bars indicate
predicted transmembrane-spanning domains.
C) Alignment of similar sequences of CED-8, human XK, mouse XK, Ciona intestinalis COS41.5, and Drosophila EST LD13628.
(
(
(
or even the early L1 larvae stage (Figures 1A and 1C);
furthermore, unlike in engulfment ced mutants, in ced-8
mutants cell corpses do not accumulate (Figures 1B and
required for programmed cell death nor sufficient to
induce cell killing, ced-8 enhances the efficiency of the
cell death program during development.
1
C), indicating that in ced-8 mutants engulfment is not
defective and probably proceeds at a normal rate once
corpses have been generated.
Function of ced-8 in Programmed Cell Death
Programmed cell death in C. elegans is believed to be
initiated molecularly by the activation of a core cell
death pathway, consisting of egl-1, ced-9, ced-4, and
ced-3 (reviewed by Metzstein et al., 1998). Morphological
changes and other downstream events are thought to
be dependent on cleavage by CED-3 of cell death sub-
strates, although it is not known precisely how the cell-
killing process results in the refractility or other morpho-
logical characteristics of a cell corpse. We have found
that ced-8, like ced-3 and ced-4, functions downstream
of or in parallel to ced-9. Thus, ced-8 could function
either to activate ced-3 and/or ced-4, that is, in the
initiation of cell killing, or to effect cell killing after cell
death has been initiated by ced-3 and ced-4.
We have shown that ced-8 is important to ensure
that cell corpses appear on time and that in sensitized
genetic backgrounds (with weak ced-3 and ced-4 mu-
tants) ced-8 is important for the cell-killing process.
While expression of ced-8 under the control of heat-
shock promoters was able to rescue the mutant pheno-
type of ced-8(n1891) animals, no additional abnormali-
ties were observed: e.g., there was no increase in the
number of cell deaths or any loss of viability associated
with overexpression of ced-8 (data not shown). By con-
trast, overexpression of ced-3 or ced-4 either under the
control of heat-shock promoters or under cell type–
specific promoters is able to induce programmed cell
death (Shaham and Horvitz, 1996b; S. Shaham and
H. R. H., unpublished data). Thus, while ced-8 is neither
In the former case, ced-8 might act directly on either
ced-3 or ced-4 to potentiate their activities and thereby

Molecular Cell
30
4
How does mutation of ced-8 in a sensitized ced-3 or
ced-4 genetic background result in the survival of cells
that normally die? It is possible that if a cell death is
delayed sufficiently then that cell can lose its compe-
tence to die, either because cell-killing protein(s) are no
longer present or because cell-protective function(s) are
activated. While such weak defects in cell death have
a very minor effect on C. elegans, which is viable even
if cell death is essentially completely blocked, it is possi-
ble that inefficient cell killing would cause more serious
defects in other organisms.
Evolutionary Conservation of CED-8
CED-8 shares similarity with human XK, the loss of which
is reported to cause McLeod Syndrome, a hereditary
disease defined by abnormalities in erythrocyte Kell anti-
gens (Ho et al., 1994). McLeod erythrocytes also have an
acanthocytic (spiculated) appearance, and the McLeod
phenotype comprises progressive myopathy and neuro-
degeneration. The cellular basis of these defects is not
understood.
Figure 5. ced-8:gfp Is Expressed in the Plasma Membrane
A) Embryo in which a ced-8:gfp fusion was expressed in most, if
not all, cells of early embryos and was localized to the plasma
membrane.
(
The XK protein is associated with the Kell protein in
the erythrocyte plasma membrane, and loss of XK leads
to the absence of Kell antigens at the cell surface (Kham-
lichi et al., 1995; Daniels et al., 1996; Russo et al., 1998).
The Kell protein is similar to neprilysin-like membrane
metallopeptidases, although its substrate is not known
(Turner and Tanzawa, 1997). XK has been postulated
to function as a transporter (Ho et al., 1994), so one
possibility is that XK transports either a substrate or a
cleavage product of Kell out of or into the cell. By anal-
ogy, it is possible that CED-8 functions in transporting
some substrate important for effecting the downstream
events of programmed cell death. However, although
the two proteins share some sequence similarity and
have similar hydropathy profiles, and both are appar-
ently localized to the plasma membrane, the similarity
between CED-8 and XK is not extensive. Functional
analysis of CED-8, XK, and the other related proteins
is needed to determine the extent of their functional
similarity.
It is possible that both CED-8 and XK act as transport-
ers that act to define the lipid content of the plasma
membrane. One proposed basis for acanthocytic mor-
phology, such as seen in McLeod erythrocytes, is asym-
metry between the inner and outer leaflets of the plasma
membrane (Zwaal et al., 1993; Handin et al., 1995); such
asymmetry could be a consequence of decreased phos-
pholipid transport by XK. CED-8 might similarly function
to modulate the distribution of lipids in the plasma mem-
brane; unlike living cells, dying cells display phosphati-
dylserine on the outer surface of their plasma mem-
branes (Fadok et al., 1992; Martin et al., 1995; Vermes et
al., 1995). Perhaps an abnormality in plasma membrane
lipid distribution leads to the slow progression of mem-
brane changes in dying cells, altering the kinetics of
programmed cell death in ced-8 mutants. We propose,
by analogy with the role of ced-8 in C. elegans, that XK
may function in apoptosis in humans and that the late
onset neurodegeneration associated with McLeod Syn-
drome is a consequence of a disruption in apoptosis.
Such an effect could be direct, that is, degenerating
cells may represent late occurring programmed cell
(
B) Nomarski image of the same embryo.
Scale bar, 10 ␮m.
initiate the cell death process. For instance, CED-8
might function as an adaptor protein between CED-3
and a protein required for the activation of CED-3 enzy-
matic activity. ced-4 appears to function genetically up-
stream of ced-3 (Shaham and Horvitz, 1996b), and bio-
chemical studies suggest that CED-4 helps induce the
conversion of pro-CED-3 to active CED-3 enzyme (Chin-
naiyan et al., 1997; Seshagiri and Miller, 1997; Wu et al.,
1
997a). When coexpressed with CED-9 in mammalian
cells, CED-4 is localized to intracellular membranes and
the perinuclear region (Wu et al., 1997b). However, we
observed that a ced-8:gfp fusion is localized to the
plasma membrane, suggesting that CED-8 does not
function in the CED-4-dependent activation of CED-3.
Alternatively, CED-8 might function to localize and
thereby activate other proteins that promote cell killing
or to sequester and thereby inactivate proteins that in-
hibit cell killing. A localization role for CED-8 is plausible
given the Ced-8 phenotype of delayed death. If ced-8
functions in a rate-enhancing role in the formation of a
cell-killing complex, it is reasonable that cell death oc-
curs eventually, albeit more slowly, in its absence.
Alternatively, if ced-8 functions in some aspect of cell
death execution downstream of ced-3 and ced-4, CED-8
might function as a substrate of CED-3 or by interacting
with such a substrate(s). We have found that the timing
of certain downstream cell death events, the onset of
refractility and DNA degradation, are uncoupled in ced-8
cell corpses. Defects in the timing and/or morphology
of cell killing may be a common result of compromising
events downstream of the central cell-killing pathway,
since the Ced-8 phenotype is reminiscent of the effect
of blocking the cleavage of lamins during apoptosis in
mammalian cells (Rao et al., 1996). When lamins with a
mutated caspase cleavage site were introduced into
mammalian cells in culture, a delay in death associated
with changes in the morphology of dying cells was ob-
served.

Programmed Cell Death Timing Gene ced-8
31
4
deaths, or indirect, that is, degenerating cells may be a
subsequent consequence of earlier defects in pro-
grammed cell death.
for cell corpses; embryos with more than six corpses in the head
at this stage were scored as rescued.
We injected heat-shock constructs at 50 ng/␮l each into ced-
8
(n1891) hermaphrodites. For heat-shock experiments, gravid Rol
hermaphrodites were allowed to lay eggs for 1.5–2 hr, the adults
were removed from the plates, and the plates were incubated at
Experimental Procedures
3
3ЊC for 1 hr. The number of corpses was counted when embryos
Nematodes
reached either the bean and comma stages or the 3-fold stage of
development. Non-heat-shocked control animals were studied in
parallel and treated identically, except that the plates were placed
at 20ЊC rather than 33ЊC for 1 hr.
C. elegans was cultured at 20ЊC on NGM agar seeded with E. coli
strain OP50 as described by Brenner (1974). Worms used for RNA
preparations were grown in liquid culture essentially as described
in Wood et al. (1988). All mutant strains were derived from the wild-
type Bristol strain N2. Mutations are described by Riddle et al. (1997)
unless otherwise noted. A list of the mutations used follows.
LG III: ced-7(n1892); ced-9(n2812) (Hengartner and Horvitz, 1994a);
qC1[dpy-19(e1259ts) glp-1(q339)] (Epstein and Shakes, 1995); ced-
Molecular Biology
Standard molecular biological procedures were followed (Sambrook
et al., 1989). The sequences of primers used for determining DNA
sequences, for PCR amplification from ced-8 mutants, and for
5Ј-RACE are available upon request. We determined sequence from
the minimal rescuing fragment pGS23 by a shotgun protocol using
an ABI 373A sequencer. CED-8 corresponds to the predicted gene
F08F1.5 (C. elegans Genome Sequencing Consortium, 1998), except
that the splice donor predicted by GENEFINDER for exon five of
F08F1.5 is incorrect. We isolated RNA from nematodes by standard
methods and performed poly(A) selection using Fast Track (In-
vitrogen). 5Ј-RACE reactions were performed on embryonic RNA
using the 5Ј RACE system (GIBCO).
4
(n1162, n2273). LG IV: ced-3(n717); ced-3(n2427, n2438) (Hen-
gartner and Horvitz, 1994a); ced-3(n2452) (Shaham et al., 1999);
ced-5(n1812). LGX: lon-2(e678); unc-10(e102); xol-1(y9); dpy-6(e14);
nuc-1(e1392); egl-15(n484); yDf2 (L. Miller and B. Meyer, personal
communication); szT1[lon-2(e678)]. The ced-8 alleles n1891, n1999,
n2090, and n2093 are described by Ellis et al. (1991). ced-8(n3113)
and ced-8(n3115) were isolated as suppressors of the lethality and
sterility caused by ced-4(n2273) ced-9(n1653) (Shaham and Horvitz,
1
996a; E. K. Speliotes and H. R. H., unpublished data), ced-8(n3244)
and ced-8(n3245) were isolated as containing late cell corpses (Z.
Zhou and H. R. H., unpublished results), and ced-8(n3313) was
isolated as an enhancer of the extra cell phenotype of ced-3(n2427)
animals (P. W. Reddien and H. R. H., unpublished data).
cDNA Analysis
To identify ced-8 cDNAs, a 3.5 kb HpaI-SphI fragment of pGS23
was used to probe an embryonic stage cDNA library in ␭gt11 (Ok-
kema and Fire, 1994) and a mixed stage cDNA library in ␭ZAP
Genetic Mapping
From the progeny of unc-10(e102) xol-1(y9) dpy-6(e14)/ced-
(
Barstead and Waterston, 1989). We found two classes of alterna-
tively spliced cDNA clones in addition to a clone containing introns
and 3, which were spliced out of all other isolated cDNAs. We
8
(n1891) heterozygotes, we selected 56 Unc non-Dpy recombinants
and found that 53 segregated Ced-8 progeny, indicating that ced-8
lies between unc-10 and dpy-6, very close to dpy-6 and therefore
left of the physical marker nP81 (Kornfeld et al., 1995).
1
refer to these two classes of alternatively spliced cDNA clones as
type I and type II, and they differed in two ways: the splice donor
for the second intron differed between the two clones, and the type
II cDNA was shorter at the 5Ј end. The type I splicing pattern was
present in five independent clones, while the type II splicing pattern
was present in only a single independent clone. The type I cDNA is
predicted to encode a 458 amino acid protein. The longest ORF
from the type II cDNA can encode a protein corresponding to the
Cell Death Assays
The number of extra cells in the anterior pharynx of L4 animals was
scored using Nomarski microscopy as described by Hengartner et
al. (1992). TUNEL experiments were performed as described by Wu
et al. (2000). DAPI (4Ј-6-diamidino-2-phenylindole dihydrochloride)
was included at 1 ␮g/ml in the final wash, and animals were staged
using the pattern of DAPI-stained nuclei to identify developmental
landmarks.
3
63 C-terminal amino acids of the type I ORF. Plasmids pGS35 (type
I) and pGS33 (type II) were constructed by digesting phage DNA
with BsiWI and ligating the insert fragment into pSL1190 (Brosius,
1
989) digested with BsiWI. To determine whether either or both of
ced-8 Dosage Experiments
the two alternatively spliced transcripts encode(s) functional CED-8
protein, we placed the longest ORF from each of the two transcripts
under the control of C. elegans heat-shock promoters (Fire et al.,
We crossed lon-2 or lon-2 ced-8 males with szT1/ϩ; unc-32(e189);
yDf2/szT1[lon-2(e678)] hermaphrodites. Non-Lon cross progeny
were scored for extra cells in the pharynx. To assay viability, several
of these non-Lon hermaphrodites were picked to a plate and allowed
to lay eggs for 2 to 3 hr. The number of eggs was counted immedi-
ately after removal of the adult worms and then counted again after
at least 16 hr, and the number of worms reaching at least the L3
stage was counted 2 days later. There was no decrease in viability
of embryos from lon-2(e678) ced-8(n1891)/yDf2 mothers as com-
pared to embryos from lon-2(e678)/yDf2 mothers; in both cases,
approximately a quarter of embryos failed to hatch (data not shown),
as expected if only the yDf2 homozygous class of progeny was
inviable.
1
990) and assayed the ability of these heat shock–cDNA constructs
to rescue the Ced-8 phenotype of ced-8(n1891) embryos. Only the
type I ORF rescued, suggesting that the type II transcript does not
encode a functional ced-8 gene.
To assess further the possibility that the type II transcript might
be relevant for either the function or regulation of ced-8, we cloned
and determined the sequence of the ced-8 region from the closely
related nematode C. briggsae. We obtained clones containing geno-
mic C. briggsae ced-8 by probing a ␭Charon4 library (T. Snutch and
D. Baillie, personal communication) and a fosmid library (Genome
Systems; fosmid clones were provided by the C. elegans Genome
Sequencing Consortium). We found that the type I splice donor
sequence is conserved from C. elegans to C. briggsae, but the type
II splice donor site is absent in C. briggsae. We also found that the
majority of the predicted amino acid sequence of C. briggsae CED-8
is well conserved as compared to that of C. elegans type I, and in
particular, the coding potential in the region spliced out of the ced-8
message in the event of a type II splice donor choice is identical
between C. elegans and C. briggsae. Our findings that the type II
long ORF does not appear to encode a functional protein and that
the splice site necessary to generate a type II transcript are not
conserved between C. elegans and C. briggsae are consistent with
the idea that the type II cDNA identified in the library is not functional.
This cDNA may represent an aberrant transcript that is not normally
Rescue Experiments
We performed injections as described by Mello et al. (1991) using
the coinjection markers pRF4 at 50–75 ng/␮l and egl-5:gfp (A. Chis-
holm and H. R. H., unpublished data) at 50–75 ng/␮l, and we selected
transgenic lines using the Roller phenotype.
We injected cosmids at 20 ng/␮l each into ced-8(n1999) hermaph-
rodites. We examined 3-fold embryos that expressed the egl-5:gfp
marker for the presence of cell corpses in the head region; embryos
with fewer than three corpses in the head were scored as rescued.
We confirmed that the absence of cell corpses was a consequence
of rescue of the Ced-8 phenotype rather than a block in cell death
by examining bean and comma stage embryos from candidate lines

Molecular Cell
32
4
produced. We consider the type I cDNA, referred to in the text as
the ced-8 cDNA, to be the functional ced-8 product.
Hakem, R., Hakem, A., Duncan, G.S., Henderson, J.T., Woo, M.,
Soengas, M.S., Elia, A., de la Pompa, J.L., Kagi, D., Khoo, W., et al.
(1998). Differential requirement for Caspase 9 in apoptotic pathways
Acknowledgments
in vivo. Cell 94, 339–352.
Handin, R.I., Lux, S.E., and Stossel, T.P. (1995). Blood: Principles
and Practice of Hematology (Philadelphia, PA: J.B. Lippincott).
We thank B. Castor for determining DNA sequences; A. Coulson,
J. Schein, and the C. elegans Genome Sequencing Consortium for
DNA clones; B. Barstead and P. Okkema for C. elegans cDNA librar-
ies; T. Snutch and D. Baillie for the C. briggsae genomic library; A.
Chisholm for egl-5:gfp; and E. Speliotes, Z. Zhou, and P. Reddien
for ced-8 alleles. We also thank B. Hersh, M. Metzstein, and R.
Ranganathan for helpful comments concerning the manuscript.
G. M. S. was supported by a Howard Hughes Medical Institute
Predoctoral Fellowship. H. R. H. is an Investigator of the Howard
Hughes Medical Institute.
Hedgecock, E.M., Sulston, J.E., and Thomson, J.N. (1983). Mutations
affecting programmed cell deaths in the nematode Caenorhabditis
elegans. Science 220, 1277–1279.
Hengartner, M.O., and Horvitz, H.R. (1994a). Activation of C. elegans
cell death protein CED-9 by an amino-acid substitution in a domain
conserved in Bcl-2. Nature 369, 318–320.
Hengartner, M.O., and Horvitz, H.R. (1994b). C. elegans cell survival
gene ced-9 encodes a functional homolog of the mammalian proto-
oncogene bcl-2. Cell 76, 665–676.
Received August 2, 1999; revised December 29, 1999.
Hengartner, M.O., Ellis, R.E., and Horvitz, H.R. (1992). Caenorhab-
ditis elegans gene ced-9 protects cells from programmed cell death.
Nature 356, 494–499.
References
Hevelone, J., and Hartman, P.S. (1988). An endonuclease from
Caenorhabditis elegans: partial purification and characterization.
Biochem. Genet. 26, 447–461.
Barstead, R.J., and Waterston, R.H. (1989). The basal component
of the nematode dense-body is vinculin. J. Biol. Chem. 264, 10177–
1
0185.
Brenner, S. (1974). The genetics of Caenorhabditis elegans. Genetics
7, 71–94.
Ho, M., Chelly, J., Carter, N., Danek, A., Crocker, P., and Monaco,
A.P. (1994). Isolation of the gene for McLeod syndrome that encodes
a novel membrane transport protein. Cell 77, 869–880.
7
Brosius, J. (1989). Superpolylinkers in cloning and expression vec-
Horvitz, H.R. (1999). Genetic control of programmed cell death in the
tors. DNA 8, 759–777.
nematode Caenorhabditis elegans. Cancer Res. 59, 1701s–1706s.
C. elegans Genome Sequencing Consortium (1998). Genome se-
quence of the nematode C. elegans: a platform for investigating
biology. Science 282, 2012–2018.
Janicke, R.U., Sprengart, M.L., Wati, M.R., and Porter, A.G. (1998).
Caspase-3 is required for DNA fragmentation and morphological
changes associated with apoptosis. J. Biol. Chem. 273, 9357–9360.
Chinnaiyan, A.M., Chaudhary, D., O’Rourke, K., Koonin, E.V., and
Dixit, V.M. (1997). Role of CED-4 in the activation of CED-3. Nature
Kerr, J.F.R., Wyllie, A.H., and Currie, A.R. (1972). Apoptosis: a basic
biological phenomenon with wide-ranging implications in tissue ki-
netics. Br. J. Cancer. 26, 239–257.
3
88, 728–729.
Collec, E., Colin, Y., Carbonnet, F., Hattab, C., Bertrand, O., Cartron,
J.P., and Kim, C.L. (1999). Structure and expression of the mouse
homologue of the XK gene. Immunogenetics 50, 16–21.
Khamlichi, S., Bailly, P., Blanchard, D., Goossens, D., Cartron, J.P.,
and Bertrand, O. (1995). Purification and partial characterization of
the erythrocyte Kx protein deficient in McLeod patients. Eur. J.
Biochem. 228, 931–934.
Conradt, B., and Horvitz, H.R. (1998). The C. elegans protein EGL-1
is required for programmed cell death and interacts with the Bcl-
Kornfeld, K., Hom, D.B., and Horvitz, H.R. (1995). The ksr-1 gene
encodes a novel protein kinase involved in Ras-mediated signaling
in C. elegans. Cell 83, 903–913.
2
-like protein CED-9. Cell 93, 519–529.
Daniels, G.L., Weinauer, F., Stone, C., Ho, M., Green, C.A., Jahn-
Jochem, H., Offner, R., and Monaco, A.P. (1996). A combination of
the effects of rare genotypes at the XK and KEL blood group loci
results in absence of Kell system antigens from the red blood cells.
Blood 88, 4045–4050.
Kuida, K., Zheng, T.S., Na, S., Kuan, C., Yang, D., Karasuyama, H.,
Rakic, P., and Flavell, R.A. (1996). Decreased apoptosis in the brain
and premature lethality in CPP32-deficient mice. Nature 384,
368–372.
Ellis, H.M., and Horvitz, H.R. (1986). Genetic control of programmed
cell death in the nematode C. elegans. Cell 44, 817–829.
Kuida, K., Haydar, T.F., Kuan, C.Y., Gu, Y., Taya, C., Karasuyama,
H., Su, M.S.-S., Rakic, P., and Flavell, R.A. (1998). Reduced apopto-
sis and cytochrome c–mediated caspase activation in mice lacking
Caspase 9. Cell 94, 325–337.
Ellis, R.E., Jacobson, D.M., and Horvitz, H.R. (1991). Genes required
for the engulfment of cell corpses during programmed cell death in
Caenorhabditis elegans. Genetics 129, 79–94.
Kyte, J., and Doolittle, R.F. (1982). A simple method for displaying
the hydropathic character of a protein. J. Mol. Biol. 157, 105–132.
Enari, M., Sakahira, H., Yokoyama, H., Okawa, K., Iwamatsu, A., and
Nagata, S. (1998). A caspase-activated DNase that degrades DNA
during apoptosis, and its inhibitor ICAD. Nature 391, 43–50.
Liu, Q.A., and Hengartner, M.O. (1998). Candidate adaptor protein
CED-6 promotes the engulfment of apoptotic cells in C. elegans.
Cell 93, 961–972.
Epstein, H.F., and Shakes, D.C. (1995). Caenorhabditis elegans:
modern biological analysis of an organism. In Methods in Cell Biol-
ogy, L. Wilson and P. Matsudaira, eds. (San Diego, California: Aca-
demic Press, Inc.).
Liu, X., Zou, H., Slaughter, C., and Wang, X. (1997). DFF, a heterodi-
meric protein that functions downstream of caspase-3 to trigger
DNA fragmentation during apoptosis. Cell 89, 175–184.
Fadok, V.A., Voelker, D.R., Campbell, P.A., Cohen, J.J., Bratton,
D.L., and Henson, P.M. (1992). Exposure of phosphatidylserine on
the surface of apoptotic lymphocytes triggers specific recognition
and removal by macrophages. J. Immunol. 148, 2207–2216.
Martin, S.J., Reutelingsperger, C.P.M., McGahon, A.J., Rader, J.A.,
van Schie, R.C.A.A., LaFace, D.M., and Green, D.R. (1995). Early
redistribution of plasma membrane phosphatidylserine is a general
feature of apoptosis regardless of the initiating stimulus: inhibition
by overexpression of Bcl-2 and Abl. J. Exp. Med. 182, 1545–1556.
Fire, A., Harrison, S.W., and Dixon, D. (1990). A modular set of
lacZ fusion vectors for studying gene expression in Caenorhabditis
elegans. Gene 93, 189–198.
Mello, C.C., Kramer, J.M., Stinchcomb, D., and Ambros, V. (1991).
Efficient gene transfer in C. elegans: extrachromosomal mainte-
nance and integration of transforming sequences. EMBO J. 10,
Franc, N.C., Dimarcq, J.L., Lagueux, M., Hoffmann, J., and Ezeko-
witz, R.A. (1996). Croquemort, a novel Drosophila hemocyte/macro-
phage receptor that recognizes apoptotic cells. Immunity 4,
3
959–3970.
Metzstein, M.M., Stanfield, G.M., and Horvitz, H.R. (1998). Genetics
of programmed cell death in C. elegans: past, present and future.
Trends Genet. 14, 410–416.
4
31–443.
Gavrieli, Y., Sherman, Y., and Ben-Sasson, S.A. (1992). Identification
of programmed cell death in situ via specific labeling of nuclear
DNA fragmentation. J. Cell Biol. 119, 493–501.
Okkema, P.G., and Fire, A. (1994). The Caenorhabditis elegans NK-2
class homeoprotein CEH-22 is involved in combinatorial activation

Programmed Cell Death Timing Gene ced-8
33
4
of gene expression in pharyngeal muscle. Development 120, 2175–
186.
Xue, D., Shaham, S., and Horvitz, H.R. (1996). The Caenorhabditis
elegans cell-death protein CED-3 is a cysteine protease with sub-
strate specificities similar to those of the human CPP32 protease.
Genes Dev. 10, 1073–1083.
2
Rao, L., Perez, D., and White, E. (1996). Lamin proteolysis facilitates
nuclear events during apoptosis. J. Cell Biol. 135, 1441–1455.
Yuan, J., and Horvitz, H.R. (1992). The Caenorhabditis elegans cell
death gene ced-4 encodes a novel protein and is expressed during
the period of extensive programmed cell death. Development 116,
Riddle, D.L., Blumenthal, T., Meyer, B.J., and Priess, J.R. (1997).
C. elegans II (Cold Spring Harbor, New York: Cold Spring Harbor
Laboratory Press).
3
09–320.
Robertson, A.M.G., and Thomson, J.N. (1982). Morphology of pro-
grammed cell death in the ventral nerve cord of Caenorhabditis
elegans larvae. J. Embryol. Exp. Morphol. 67, 89–100.
Yuan, J., Shaham, S., Ledoux, S., Ellis, H.M., and Horvitz, H.R. (1993).
The C. elegans cell death gene ced-3 encodes a protein similar to
mammalian interleukin-1␤-converting enzyme. Cell 75, 641–652.
Russo, D., Redman, C., and Lee, S. (1998). Association of XK and
Kell blood group proteins. J. Biol. Chem. 273, 13950–13956.
Zou, H., Henzel, W.J., Liu, X., Lutschg, A., and Wang, X. (1997).
Apaf-1, a human protein homologous to C. elegans CED-4, partici-
pates in cytochrome c-dependent activation of caspase-3. Cell 90,
Sakahira, H., Enari, M., and Nagata, S. (1998). Cleavage of CAD
inhibitor in CAD activation and DNA degradation during apoptosis.
Nature 391, 96–99.
4
05–413.
Zwaal, R.F., Comfurius, P., and Bevers, E.M. (1993). Mechanism
and function of changes in membrane-phospholipid asymmetry in
platelets and erythrocytes. Biochem. Soc. Trans. 21, 248–253.
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Clon-
ing: A Laboratory Manual (Cold Spring Harbor, New York: Cold
Spring Harbor Laboratory Press).
Seshagiri, S., and Miller, L.K. (1997). Caenorhabditis elegans CED-4
stimulates CED-3 processing and CED-3-induced apoptosis. Curr.
Biol. 7, 455–460.
Shaham, S., and Horvitz, H.R. (1996a). An alternatively spliced C.
elegans ced-4 RNA encodes a novel cell death inhibitor. Cell 86,
2
01–208.
Shaham, S., and Horvitz, H.R. (1996b). Developing Caenorhabditis
elegans neurons may contain both cell-death protective and killer
activities. Genes Dev. 10, 578–591.
Shaham, S., Reddien, P.W., Davies, B., and Horvitz, H.R. (1999).
Mutational analysis of the Caenorhabditis elegans cell-death gene
ced-3. Genetics 153, 1655–1671.
Sulston, J.E. (1976). Post-embryonic development in the ventral cord
of Caenorhabditis elegans. Philos. Trans. R. Soc. Lond. B Biol. Sci.
2
75, 287–297.
Sulston, J.E., and Horvitz, H.R. (1977). Post-embryonic cell lineages
of the nematode Caenorhabditis elegans. Dev. Biol. 56, 110–156.
Sulston, J.E., Schierenberg, E., White, J.G., and Thomson, J.N.
(1983). The embryonic cell lineage of the nematode Caenorhabditis
elegans. Dev. Biol. 100, 64–119.
Turner, A.J., and Tanzawa, K. (1997). Mammalian membrane metal-
lopeptidases: NEP, ECE, KELL, and PEX. FASEB J. 11, 355–364.
Varfolomeev, E.E., Schuchmann, M., Luria, V., Chiannilkulchai, N.,
Beckmann, J.S., Mett, I.L., Rebrikov, D., Brodianski, V.M., Kemper,
O.C., Kollet, O., Lapidot, T., et al. (1998). Targeted disruption of the
mouse caspase 8 gene ablates cell death induction by the TNF
receptors, Fas/Apo1, and DR3 and is lethal prenatally. Immunity 9,
2
67–276.
Vaux, D.L., and Korsmeyer, S.J. (1999). Cell death in development.
Cell 96, 245–254.
Vermes, I., Haanen, C., Steffens-Nakken, H., and Reutelingsperger,
C. (1995). A novel assay for apoptosis. Flow cytometric detection
of phosphatidylserine expression on early apoptotic cells using fluo-
rescein labelled Annexin V. J. Immunol. Meth. 184, 39–51.
Wood, W.B., and the Community of C. elegans Researchers. (1988).
The Nematode Caenorhabditis elegans (Cold Spring Harbor, New
York: Cold Spring Harbor Laboratory Press).
Wu, Y.-C., and Horvitz, H.R. (1998a). The C. elegans cell corpse
engulfment gene ced-7 encodes a protein similar to ABC transport-
ers. Cell 93, 951–960.
Wu, Y.-C., and Horvitz, H.R. (1998b). C. elegans phagocytosis and
cell-migration protein CED-5 is similar to human DOCK180. Nature
3
92, 501–504.
Wu, D., Wallen, H.D., Inohara, N., and Nunez, G. (1997a). Interaction
and regulation of the Caenorhabditis elegans death protease CED-3
by CED-4 and CED-9. J. Biol. Chem. 272, 21449–21454.
Wu, D., Wallen, H.D., and Nunez, G. (1997b). Interaction and regula-
tion of subcellular localization of CED-4 by CED-9. Science 275,
1
126–1129.
Wu, Y.-C., Stanfield, G.M., and Horvitz, H.R. (2000). NUC-1, a C.
elegans DNase II homolog, functions in an intermediate step of DNA
degradation during apoptosis. Genes Dev., in press.