(Received for publication, February 20, 1997)
From the Hanson Centre for Cancer Research, Institute of Medical and Veterinary Science, Rundle Mall, Adelaide, SA 5000, Australia
The ICE/CED-3 family of proteases (caspases) play a central role in the execution phase of apoptosis. These proteases are synthesised as precursor molecules that require processing at specific aspartate residues to produce the two subunits that comprise the active enzyme. The activation of some of these proteases has been shown to occur during apoptosis. Here we show that Nedd2/ICH-1 (caspase-2) is activated during apoptosis induced by a variety of apoptotic stimuli. This activation occurs very early upon treatment of cells with apoptotic agents and appears to precede the activation of CPP32 (caspase-3). The activation of Nedd2 was not seen in cells that are resistant to apoptosis. These observations suggest that Nedd2 is an early effector in the pathway leading to cell death. Our observations also lend weight to the hypothesis that a group of caspases containing long prodomains are the first to be activated in response to apoptotic signals and that they lie upstream of a second class of caspases such as CPP32 containing short or no prodomains.
Nedd2 was initially identified as a developmentally
down-regulated gene in the mouse central nervous system (1), and its product was subsequently shown to be homologous to the
Caenorhabditis elegans CED-3 protein and the mammalian
interleukin-1-converting enzyme (ICE)1
(2). These enzymes are the prototypes of a growing family of
aspartate-specific cysteine proteases termed caspases (3) that play a
central role in the execution of apoptosis (4-6). Studies with ICE
suggest that caspases are synthesized as proenzymes that are cleaved at
specific aspartate residues to release two subunits of approximately 10 and 20 kDa, which heterodimerize (7) and possibly associate in a
tetramer to form the active enzyme (8, 9). The substrate specificity of
caspases requires an aspartate residue at the P1 cleavage
position, a property shared only by the cytotoxic T-lymphocyte serine
protease granzyme B (10).
The human caspases can be subdivided into three families based on sequence homology. The ICE-like protease family includes ICE (caspase-1) (7), TX/ICH-2/ICErel-II (caspase-4) (11-13), and TY/ICErel-III (caspase-5) (13, 14). The CED-3-like family includes CPP32/YAMA/apopain (caspase-3) (15-17), Mch2 (caspase-6) (18), Mch3/ICE-lap3/CMH-1/(caspase-7) (19-21), Mch4 (caspase-10) (22), MACH/FLICE/Mch5 (caspase-8) (22-24), and ICE-lap6/Mch6 (caspase-9) (25, 26). CPP32, ICE-lap3 and Mch2 have been shown to be activated in vivo in response to apoptotic stimuli (20, 27-29). Nedd2/ICH-1 (caspase-2) stands alone in the third caspase subfamily, and to date no systematic study has been carried out to evaluate its activation during apoptosis.
Once activated, the caspases cleave a range of cellular substrates. The
DNA repair enzyme poly(ADP-ribose)-polymerase (PARP) was one of the
first identified cellular substrates cleaved during apoptosis (30).
CPP32 was subsequently shown to cleave PARP with high efficiency (16,
17). Additional substrates that are specifically cleaved during
apoptosis by CPP32-like proteases include proteins such as
DNA-dependent protein kinase (31-33), U1-70 kDa
ribonucleoprotein (31), heteronuclear riboproteins C1 and C2 (34),
-fodrin (35), and nuclear lamins (29, 36). ICE has a substrate
specificity of Tyr-Val-His-Asp (YVHD), first identified as the site in
pro-interleukin-1
that is cleaved by ICE to release the mature
cytokine (7).
Several lines of evidence suggest a role for Nedd2 in the apoptotic pathway. Overexpression of Nedd2 induces apoptosis in various cell types (2, 37, 38), whereas expression of antisense Nedd2 in FDC-P1 factor-dependent cells delays the onset of apoptosis induced by factor withdrawal (39). In a similar manner, an alternatively spliced form of Nedd2 that encodes a truncated protein has been shown to protect against cell death induced by serum withdrawal in Rat-1 and NIH-3T3 cells (37, 40). Up-regulation of Nedd2 mRNA has been observed in response to ischaemia-induced cell death in the Mongolian gerbil (41), whereas down-regulation of Nedd2 has been observed during gonadotropin-promoted follicular survival (42). Recently Nedd2 has been shown to associate via its pro-domain with RAIDD, a death adaptor molecule that is thought to be involved in death signaling through the TNF-R1 complex via association with the death domain proteins RIP and TRADD (43). An analogous association occurs between the pro-domain of MACH/FLICE/Mch5 and the death adaptor molecule MORT1/FADD, which recruits MACH/FLICE/Mch5 to the Fas/Apo1 signaling complex in response to Fas-induced apoptosis (23, 24). This suggests that MACH/FLICE/Mch5 activation is the most upstream enzymatic event in the Fas/Apo1 signaling pathway and that MACH/FLICE could subsequently activate other downstream caspases, resulting in apoptotic death. In a similar manner, Nedd2 activation may be an early, upstream event in TNF-R1-induced apoptosis.
In this report we show that proNedd2 is rapidly cleaved to generate active enzyme subunits in response to various apoptotic stimuli and that Nedd2 activation precedes the activation of CPP32. These observations demonstrate that Nedd2/ICH1 is a caspase that is activated early in the apoptotic cascade.
The human
megakaryoblastic cell line Mo7e was maintained in suspension culture at
37 °C, 5% CO2 in -minimal essential medium supplemented with 10% heat-inactivated fetal calf serum and
recombinant human interleukin-3 (IL-3) at a final concentration of 2 ng/ml (Amgen). The Burkitt's lymphoma cell lines BL30A and BL30K were maintained in suspension culture in RPMI 1640 medium containing either
20% (BL30A) or 10% (BL30K) fetal calf serum. Cells were induced to
undergo apoptosis by withdrawal of IL-3 and/or serum or by exposure to
either etoposide (40 µM) or 30 Gy of
-irradiation from
a 137Cs source.
Cytoplasmic extracts
were prepared essentially as described (44) with some modifications.
Cells were harvested and washed with ice-cold phosphate-buffered saline
and then resuspended (100 µl/107 cells) in ice-cold cell
extraction buffer (50 mM PIPES, pH 7.0, 50 mM
KCl, 5 mM EGTA, 2 mM MgCl2, 1 mM dithiothreitol, and 0.5 mM
phenylmethylsulfonyl fluoride). Cells were allowed to swell on ice for
20 min and then lysed by freeze/thawing twice. Cell lysis was confirmed
by trypan blue uptake. Lysates were centrifuged at 200 × g for 5 min at 4 °C, and supernatants were then removed and centrifuged at 9000 × g for 15 min at 4 °C. The
clear cytosol was removed and either used immediately or stored at
70 °C.
Cytoplasmic extracts (50 µg of total protein) were boiled in protein loading buffer (100 mM Tris-HCl, pH 6.8, 200 mM dithiothreitol, 20% glycerol, 4% SDS, 0.2% bromphenol blue) for 5 min and then centrifuged at 9000 × g for 5 min. The denatured proteins were electrophoresed on a 15% SDS-polyacrylamide gel and transferred to polyvinylidine difluoride membrane (DuPont NEN). Membranes were blocked in 5% skim milk in phosphate-buffered saline containing 0.05% Tween 20 at 4 °C overnight. Blots were probed with either an anti-ICH-1L polyclonal antibody (Santa Cruz Biotech., Inc.) at a 1:500 dilution for 4 h, anti-Mch3 polyclonal antiserum (kindly donated by Dr. G. Cohen) at a dilution of 1:2000 for 2 h, an anti-PARP polyclonal antiserum (Boehringer Mannheim) at a dilution of 1:2000 for 1 h, or an anti-CPP32 monoclonal antibody (Transduction Labs.) at a dilution of 1:1000 for 4 h at room temperature. This was followed by incubation with anti-rabbit (ICH-1L, Mch3, and PARP) or anti-mouse (CPP32) IgG conjugated with horseradish peroxidase (Amersham Corp.) for 1 h. Signals were detected using the ECL system (Amersham Corp.).
Cleavage of Fluorogenic Caspase SubstratesCytoplasmic extracts (10 µg of total protein) were incubated with 100 µM DEVD-7-amino-4-trifluoromethyl coumarin or YVAD-7-amino-4-trifluoromethyl coumarin (both from Enzyme Systems Inc.) at 37 °C for 30 min in a final volume of 20 µl in cleavage buffer (25 mM HEPES, pH 7.4, 10% sucrose, 5 mM dithiothreitol, 1 mM EDTA, 0.1% CHAPS). The tetrapeptide inhibitors DEVD-CHO and YVAD-CHO (Bachem) used at a final concentration of 1 µM were preincubated for 30 min at 37 °C with cytoplasmic extracts in a 10-µl volume prior to the addition of fluorogenic substrate. Fluorescence was quantified using an Aminco Bowman Luminescence Spectrophotometer (excitation, 400 nm; emission, 505 nm).
In Vitro Cleavage of PARPThe cDNA construct pBS hPARP
encompassing a truncated region of human PARP cDNA from nucleotide
93 to nucleotide 1156 (amino acids 1-339 containing the DEVD cleavage
site at amino acid residues 211-214) was generated by polymerase chain
reaction. The upstream primer 5-CGGAATTCTAGGTCGTGCGTCGG-3
was designed to contain an EcoRI site (underlined), and the
downstream primer used was 5
-GGAATATACGGTCCTGCT-3
. The pBS hPARP
construct was used as a template for the production of
[35S]methionine (ICN)-labeled PARP protein using the
Promega TNT T7 Coupled Reticulocyte Lysate System. 5 µl of labeled
product was incubated in proteolysis assays at 37 °C for 1 h
with cytoplasmic extracts from apoptotic and nonapoptotic cells (10 µg of total protein) in cleavage buffer in a total volume of 20 µl.
Cleavage products were resolved by 15% SDS-polyacrylamide gel
electrophoresis, transferred to polyvinylidine difluoride membrane
(DuPont NEN) using a semi-dry apparatus (Hoeffer), and visualized by
autoradiography.
To investigate the magnitude and specificity of protease
activity induced by various apoptotic stimuli, the tetrapeptide
fluorogenic substrates DEVD-AFC and YVAD-AFC were utilized. The
fluorogenic substrate DEVD-AFC mimics the cleavage site at which CPP32
and CPP32-like caspases cleave PARP, whereas the YVAD-AFC substrate is
specific for caspases closely related to ICE. Cleavage of these substrates by cellular cytoplasmic extracts is thus indicative of
in vivo proteolytic activation of these caspases in response to apoptotic signals. We initially studied caspase activation in the
factor-dependent hemopoietic cell line Mo7e, which
undergoes apoptosis in response to the removal of IL-3 and serum
from its growth medium. Following withdrawal of IL-3 and serum from
Mo7e, there was a significant, time-dependent increase in
the level of DEVD-AFC cleavage, indicating the progressive activation
of CPP32 and/or its close relatives in these cells. This activation reflected the progressive increase in the number of apoptotic cells as
measured by trypan blue exclusion and nuclear staining (not shown),
although an increase in DEVD-AFC cleavage activity was obvious prior to
any significant cell death (Fig. 1A).
Cleavage of DEVD-AFC by cytoplasmic extracts from factor-deprived Mo7e cells was completely inhibited by preincubation of the cell extracts with the tetrapeptide aldehyde inhibitor of CED-3-like caspases, DEVD-CHO. No cleavage of YVAD-AFC was observed at any time after IL-3
and serum withdrawal, indicating that ICE and its closest relatives are
not activated in response to this stimulus in these cells (data not
shown).
Exposure of Mo7e cells to 40 µM etoposide or 30-Gy
-irradiation also resulted in a progressive increase in CPP32-like
activity that correlated with the increase in the number of apoptotic
cells as measured by trypan blue exclusion and nuclear staining (Fig. 1, B and C). As previously observed, significant
substrate cleavage activity was evident before any morphological
appearance of apoptosis. DEVD-AFC cleavage activity of cell extracts
was inhibited by preincubation of the extracts with 1 µM
DEVD-CHO. Once again, there was no increase in caspase activity on the
YVAD-AFC substrate (data not shown).
In addition to the Mo7e
cell line, we studied the activation of caspases in two isogenic
Burkitt's lymphoma cell lines that are sensitive (BL30A) and resistant
(BL30K) to the induction of apoptosis by genotoxic agents (45). As for
Mo7e cells, the treatment of BL30A cells with 40 µM
etoposide induced a progressive increase in the magnitude of apoptotic
cells and CPP32-like activity. In contrast, BL30K cells that remain
viable following etoposide treatment did not show any increase in
CPP32-like activity, confirming that the activation of CPP32-like
caspases occurs during cell death (Fig. 2A).
In a similar manner, BL30A cells treated with 30-Gy -irradiation
exhibited a time-dependent increase in DEVD-AFC cleavage
activity, whereas BL30K cells were resistant to apoptosis and
cytoplasmic extracts prepared from these cells did not display any
increase in CPP32-like caspase activity (Fig. 2B).
We consistently observed a low level of DEVD-AFC cleavage in untreated Mo7e and BL30A cells. This may be a consequence of the low percentage of apoptotic cells in the normal cell population (approximately 3%).
Cleavage of PARP during Apoptosis in Mo7e and BL30A CellsTo
further investigate the proteolytic activity in apoptotic and
nonapoptotic extracts from Mo7e, BL30A, and BL30K cells, we examined
the cleavage of PARP both in vitro and in vivo.
It has been previously shown that CPP32, Mch3, and Mch2 cleave PARP efficiently (16-19), whereas ICE, TX, and Nedd2 do so poorly (46). Cytoplasmic extracts from Mo7e cells following factor withdrawal or
treatment with etoposide or -irradiation cleaved in vitro translated [35S]methionine-labeled PARP at the CPP32-like
caspase recognition sequence, liberating 24- and 14-kDa polypeptides
from the truncated 38-kDa PARP substrate (Fig. 1, A,
B, and C). Similarly, cytoplasmic extracts
prepared from BL30A cells following treatment with apoptosis inducing
stimuli contained PARP cleavage activity, whereas those prepared from
BL30K cells did not (Fig. 2, A and B). As
previously seen in fluorogenic substrate cleavage experiments, a
marginal base-line level of CPP32-like activity was observed in
untreated cells. These experiments confirm that activation of
CPP32-like proteases occurs in both Mo7e and BL30A cells early during
apoptosis.
Western analysis indicated that in Mo7e cells, PARP was rapidly cleaved
in vivo to 89- and 24-kDa products in response to IL-3
withdrawal, etoposide, and -irradiation, with the 115-kDa precursor
completely cleaved within the first 4 h following IL-3 withdrawal
and within the first 2 h following etoposide treatment or
-irradiation (Fig. 3). The complete proteolysis of
PARP significantly preceded the appearance of apoptotic cells,
suggesting that caspase activation occurs prior to the morphological
changes of apoptotic cell death.
Activation of Nedd2/ICH-1 during Apoptosis
Activation of
Nedd2 requires the cleavage of the 51-kDa precursor molecule into
subunits of 19 and 12 kDa (47, 48). Having established that the
cytoplasmic extracts from cells treated with apoptotic stimuli
contained CPP32-like caspase activity, we investigated whether proNedd2
is cleaved to generate the p19 and p12 subunits in response to
apoptotic stimuli. We examined the activation of Nedd2 by Western
blotting with an antiserum against the p12 subunit of human Nedd2 that
recognizes both p12 and p51 proNedd2. In Mo7e cells, IL-3/serum
withdrawal and etoposide or -irradiation treatment resulted in the
cleavage of p51 and the appearance of p12 (Fig. 3, A,
B, and C). From time course studies, the
activation of Nedd2 appears to be rapid, detectable as early as 4 h after IL-3 and serum withdrawal and as early as 2 h following
exposure to etoposide and
-irradiation. This correlates well with
the first significant increase in fluorogenic substrate cleavage and
PARP cleavage, both in vitro and in vivo (Figs.
1, 2, 3). In BL30A cells, but not in apoptosis-resistant BL30K cells,
progressive proteolysis of p51 proNedd2 over time was observed in
response to treatment with both etoposide and
-irradiation (Fig.
4, A and B).
Although the majority of activity on DEVD-AFC observed in our experiments is probably due to CPP32 and caspases closely related to it, we and others have observed the ability of Nedd2 to cleave the DEVD substrate (49).2 Therefore, it is possible that some DEVD-AFC cleavage may be attributed to activated Nedd2, the optimal substrate of which is thus far unknown.
Activation of Nedd2 Occurs Prior to That of CPP32 during ApoptosisTo investigate at which stage in the apoptotic pathway
Nedd2 activation occurs in relation to that of other caspases, we
assessed the activation of CPP32 by Western blotting using an
anti-CPP32 monoclonal antibody that detects both proCPP32 and the p17
subunit. We did not observe significant cleavage of CPP32 in Mo7e cells subjected to IL-3/serum withdrawal or treated with either etoposide or
-irradiation at time points where proNedd2 was completely cleaved
(Fig. 3, A, B, and C). In later time
points some cleavage of proCPP32 was evident (data not shown). These
results indicate that in the apoptotic pathway, Nedd2 activation occurs
upstream of CPP32 activation in Mo7e cells.
Activation of CPP32 was clearly observed in BL30A cells following
treatment with both etoposide and -irradiation (Fig. 4, A
and B). The appearance of the CPP32 p17 subunit was visible approximately 3 h post treatment with etoposide, once again a later stage than the initial activation of Nedd2, visible 2 h post
treatment (Fig. 4A). An analogous situation was observed in
response to
-irradiation (Fig. 4B). In contrast, no
cleavage of CPP32 precursor was seen in extracts from the nonresponsive BL30K cells treated with etoposide or
-irradiation.
Because no early activation of CPP32 was observed in Mo7e cells, we investigated the activation of Mch3, the caspase most homologous to CPP32. Western blotting with an anti-Mch3 polyclonal antiserum revealed that proMch3 is rapidly processed, as evident by the disappearance of the precursor, in response to all apoptotic stimuli (Fig. 3, A-C). In our experiments, the activation of Mch3 appears to occur concurrently with that of Nedd2 and correlates with the cleavage of PARP we observed in vitro and in vivo.
Previous studies have demonstrated the activation of CPP32, Mch2, and ICE-LAP3 caspases in response to various apoptotic stimuli (20, 27-29). We sought to investigate whether the Nedd2/ICH1 caspase is activated in response to various apoptosis inducing stimuli and at what stage in the protease cascade pathway this occurs. Here we show that Nedd2 is activated early in the passage of events that leads to the execution of cell death, preceding the activation of CPP32. In the megakaryoblastic cell line Mo7e, the activation of Nedd2 occurred much before the activation of CPP32; however, in the Burkitt's lymphoma cell line BL30A, CPP32 activation followed Nedd2 activation much more rapidly.
It is becoming clear that in addition to the division of the caspases into subfamilies by virtue of their relatedness, caspases can be divided into two classes based on the length of their amino-terminal pro-domains and their position in the caspase heirachy. Proteases containing a long pro-domain include CED-3, Nedd2/ICH-1, Mch4, MACH/FLICE/Mch5, ICE, TX, and ICErel-III, whereas those with very short or absent pro-domains include CPP32, Mch2, Mch3, and Mch6. The pro-domains of Nedd2 and MACH/FLICE/Mch5 have been shown to contain protein motifs that mediate their association with similar motifs present in the amino termini of RAIDD and MORT1/FADD, respectively (23, 24, 43). The carboxyl termini of the RAIDD and MORT1/FADD molecules harbor a "death domain," a motif first identified in TNF-R1 (50) and Fas/Apo1 receptors (51) that serves to mediate homo- and heterotypic protein associations necessary for cell death. A death domain also mediates the association of MORT1/FADD with TRADD, an adaptor protein involved in the TNF-R1 cell death pathway (52), suggesting that these two cell death pathways may be linked. Mch4 also contains a long pro-domain that harbors death effector domains (22). Mch4 and Mch5 have both been shown to activate all known caspases in vitro, although they differ with respect to their inhibition by CrmA (53), suggesting that several parallel pathways to cell death may exist that utilize the same effector machinery but are initialized by different upstream caspases in response to different stimuli.
From our data obtained in Mo7e, it appears that CPP32 may not be involved in the apoptotic pathway in these cells. Instead, the function of CPP32 may be replaced by Mch3, which we observed to be rapidly cleaved following apoptotic stimuli. Mch3 may be the caspase responsible for PARP cleavage in these cells. It is not known whether Nedd2 is responsible for Mch3 cleavage in Mo7e cells or whether another caspase member such as Mch4 or Mch5 is activated in response to these stimuli and proceeds to activate Mch3. We have previously shown that Nedd2 is unable to cleave CPP32 in vitro (47), which suggests that another protease that is activated early in the caspase cascade such as Mch4 or Mch5 may be responsible for the cleavage of CPP32 we observed in BL30A cells. Alternatively, Nedd2 may activate an unidentified caspase, which in turn activates CPP32. These results lend weight to the classification of caspases in a hierarchy based on their sequence of activation in response to apoptotic stimuli. It appears that caspases containing long pro-domains that mediate their direct physical coupling to signal transduction machinery constitute the first class of proteases to be activated in response to apoptotic signals and that they in turn activate the second class of caspases, which are the cell death effectors.
Our results show that Nedd2/ICH-1 is activated in response to various
apoptotic stimuli and that this activation occurs at an earlier stage
in the caspase hierarchy than that of CPP32, suggesting that Nedd2
belongs to the first group of caspases activated in response to
apoptotic stimuli. The interaction of Nedd2 with RAIDD suggests that
Nedd2 might be activated in response to TNF or Fas-mediated apoptosis,
but our results demonstrate that Nedd2 activation also occurs in
response to other apoptosis inducing stimuli. Nedd2 may therefore also
comprise part of the signaling complex that transduces death signals
from stimuli such as factor withdrawal, drug treatment, and
-irradiation, although the identity of the other components of this
complex and their location in the cell is currently not known. In
future studies it will be interesting to examine if Nedd2, once
activated, can further process other members of the caspase family,
thereby initiating the proteolytic cascade characteristic of the
execution phase of apoptosis. It will also be interesting to see
whether the Nedd2 prodomain can interact with other, as yet
unidentified, adaptor molecules.
We thank Martin Lavin and Dianne Watters for providing the BL30A and BL30K cells, Gerry Cohen and Marion MacFarlane for kindly donating the Mch3 antiserum and for helpful discussions, Gayathri Parasivam for technical assistance, and Andreas Strasser and David Vaux for useful comments.