(Received for publication, May 2, 1997)
From the Laboratory of Molecular Endocrinology, Massachusetts General Hospital, Howard Hughes Medical Institute, and Harvard Medical School, Boston, Massachusetts 02114
We report the identification of the large subunit of the DNA replication factor, DSEB/RF-C140, as a new substrate for caspase-3 (CPP32/YAMA), or a very closely related protease activated during Fas-induced apoptosis in Jurkat T cells. DSEB/RF-C140 is a multifunctional DNA-binding protein with sequence homology to poly(ADP-ribose) polymerase (PARP). This similarity includes a consensus DEVD/G cleavage site for caspase-3. Cleavage of DSEB/RF-C140 is predicted to occurs between Asp706 and Gly707, generating 87-kDa and 53-kDa fragments. An antiserum raised against the amino-terminal domain of DSEB/RF-C140 detects a new 87-kDa protein in Jurkat T cells in which apoptosis is activated by a monoclonal antibody to Fas. This cleavage occurs shortly after PARP cleavage. In vitro translated DSEB/RF-C140 is specifically cleaved into the predicted fragments when incubated with a cytoplasmic extract from Fas antibody-treated cells. Proteolytic cleavage was prevented by substituting Asp706 by an alanine in the DEVD706/G caspase-3 cleavage site. The cleavage of DSEB/RF-C140 is prevented by iodoacetamide and the specific caspase-3 inhibitor, tetrapeptide aldehyde Ac-DEVD-CHO, but not by the specific ICE (interleukin-1-converting enzyme) inhibitors: CrmA and Ac-YVAD-CHO, indicating that the protease responsible for the cleavage of DSEB/RF-C140 during Fas-induced apoptosis in Jurkat cells is caspase-3, or a closely related protease. This conclusion is reinforced by the fact that recombinant caspase-3 but not caspase-1 reproduced the "in vivo" cleavage. Inasmuch as the cleavage of DSEB/RF-C140 separates its DNA binding from its association domain, required for replication complex formation, we propose that such a cleavage will impair DNA replication. Recent in vitro mutagenesis support this proposal (Uhlmann, F., Cai, J., Gibbs, E., O'Donnel, M., and Hurwitz, J. (1997) J. Biol. Chem. 272, 10058-10064).
Programmed cell death or apoptosis is well recognized as a physiological process essential for normal tissue homeostasis within multicellular organisms (1-4). Cells undergoing apoptosis are characterized by nuclear condensation, DNA fragmentation, contraction of the cytoplasm, and blebbing of the plasma membrane (5, 6). Some of the molecular processes responsible for the morphological changes observed during apoptosis in mammalian cells have been elucidated in the invertebrate Caenorhabditis elegans (7). Activation of a family of aspartate-specific cysteine proteases, initially defined by their homology to the C. elegans protein CED3, and homolog to the mammalian interleukin-1-converting enzyme (ICE),1 appear to be critical for apoptosis to occur (8-10). Due to a frenetic pace in the identification of new members of this protease family, a nomenclature committee has recently recommended the use of the term caspase to refer to those cysteine proteases that belong to the ICE/CED-3 family (49). Signals generated from inside the cells due to cellular damage or genetically determined space-time switches during development, and the presence or absence of cell-specific extracellular signals, can irreversibly trigger the biochemical cascade that results in programmed cell death. Extracellular signals such as the occupation of the Fas/APO-1 receptor by the Fas ligand or an antibody with agonist properties constitutes a well characterized model in which to study the events responsible for the initiation and execution of the apoptotic program (11). Fas-dependent apoptosis in mammalian cells resembles programmed cell death in C. elegans, inasmuch as both require the activation of ICE/CED-3 proteases (12) (caspases). The sequential activation of ICE (caspase-1)-like and CPP32/YAMA (caspase-3)-like proteases during apoptosis has been suggested (13, 16). Although current studies have focused on the elucidation of the biochemical pathways by which the occupation of the Fas-receptor results in the activation of the caspase (ICE/CED-3) protease cascade (14, 15) and the identification of the sequential activation of specific members of this protease family (16), less attention has been directed toward the identification of the cellular substrates for such proteases. Although several proteins have been identified as specific substrates for caspase proteases during apoptosis (10, 17-28), only the nuclear proteins poly(ADP-ribose) polymerase (PARP) (10), DNA-dependent protein kinase (DNA-PK) (25), the ribonucleoprotein U1-70 kDa (20), GTP dissociation inhibitor D4 (26), and huntingtin (48) have been shown to be cleaved by caspase-3 (CPP32/YAMA), one of the final known effectors of the protease cascade. It remains unknown whether the elimination of a particular protein or class of proteins essential for cell viability and integrity is the target of the caspase protease cascade, or the degradation of multiple substrates without a particular essential role, individually, could overcome the capacity for cells to repair themselves (29).
The mammalian DNA replication factor complex C, also called activation
complex A1, is a multimeric complex formed by five polypeptides with
apparent molecular masses of 140, 40, 38, 37, and 36 kDa (30, 31). This
complex is required for DNA replication in mammalian cells. It
recognizes and binds the primed DNA template in the replication fork
and recruits PCNA to allow the assembly of a replication complex with
the DNA polymerases (delta) and
(epsilon) (31, 32). We
fortuitously cloned the large subunit of this complex (designated DSEB)
while searching for a nuclear protein with capacity to bind a
differentiation-specific element (DSE) in the angiotensinogen gene
promoter (33). Binding of DSEB to the DSE enables irreversible
induction of angiotensinogen expression during adipocyte
differentiation (34). The large subunit of the replication factor C,
DSEB/RF-C140 is responsible for recognition of the primed DNA template
and the recruitment of the other four polypeptides of the RF-C complex
to the replication fork (31, 32). A large carboxyl-terminal region with
extensive homology to the other subunits of the RF-C is required for
their association (35). Domain B within the carboxyl-terminal region is
required for the direct interaction of DSEB/RF-C140 with proliferating cell nuclear antigen (PCNA) (36). We have mapped the DNA-binding domain
of DSEB/RF-C140 to an amino-terminal region of 126 amino acids
(residues 367-493) (33). This domain has extensive homology to the
prokaryotic DNA ligases and the eukaryotic enzyme, poly(ADP-ribose) polymerase, involved in the recognition of DNA breaks and the orchestration of a DNA repair response (37, 38). In its domain B,
DSEB/RF-C140 also shares with PARP a conserved region with the
consensus sequence DEVDG that is the cleavage site for caspase-3 during
apoptosis. This homology suggests that DSEB/RF-C140 could also be a
substrate for caspase-3 (CPP32/YAMA). Such a cleavage would inactivate
DSEB/RF-C140 by separating the DNA-binding domain from the association
domain required for its interaction with the other subunits of the
replication factor C (35), and would compromise its direct interaction
with PCNA (36). Our results confirm the predictions and indicate that
DSEB/RF-C140 is cleaved into two fragments during Fas-induced apoptosis
in T cells. By using site-directed mutagenesis, we identified the
Asp706 residue to be essential for the cleavage. This
residue is contained in the DEVD/G sequence with homology to the CPP32
cleavage site in PARP. Furthermore, by using specific protease
inhibitors, we identified caspase-3 (CPP32) or a very close homolog
protease, and not caspase-1 (ICE) as the cleaving protease. As in the
case of PARP and other substrates, for the CED3-like proteases, the exact relevance of their cleavages in apoptosis remains to be determined. However, the identification of DSEB/RF-C140 as a substrate cleaved by caspase-3 or a closely related protease may be an important clue for understanding the role that ICE/CED3-like proteases play during programmed cell death.
DNA modifying enzymes were obtained from New England Biolabs (Beverly, MA) or Boehringer Mannheim. Radioactive compounds were purchased from NEN Life Science Products. Nucleotides were obtained from Pharmacia Biotech Inc. Tissue culture media and reagents were from Life Technologies, Inc. Other reagents were purchased from Sigma. The cowpox-virus serpin CrmA was obtained from Kamiya Biomedical Co. (Tukwila, WA). ICE and CPP32 specific tetrapeptide aldehyde antagonists were obtained from Peptide Institute Inc. (Osaka, Japan).
Cells and Culture ConditionsJurkat cells were obtained from the ATCC (Rockville, MD) and grown in RPMI medium supplemented with 10% heat-inactivated fetal calf serum and antibiotics (penicillin, 100 units/ml; streptomycin, 100 µg/ml). To induce apoptosis, 1-ml aliquots containing 2 × 106 cells were exposed to the indicated concentrations (01-1 µg/ml) of the anti-human Fas monoclonal antibody (Upstate Biotechnologies, Lake Placid, NY) for the indicated periods of time.
Plasmid Construction and MutagenesisThe 3.6-kilobase pair
cDNA containing the entire protein coding sequence for DSEB/RF-C140
was inserted into the pCDNA-1 vector described previously (33). By
using site-directed mutagenesis, a mutated version of DSEB/RF-C140,
DSEB Asp706 Ala, was generated. In this mutated
protein, the normal Asp residue at position 706 was converted to an
alanine residue. Oligonucleotides ATGGATGAGGTCGCTGGCATGGCAGGCAATGAAGAC
and GCCTGCCATGCCAGCGACCTCATCCATGAGGGCGTG were used to create the
mutation using a polymerase chain reaction-based technique.
Full-length RF-C140/DSEB protein was synthesized in reticulocyte extracts by in vitro transcribing and translating the cDNA sequence contained in the pCDNA-DSEB plasmid (33), using T7 RNA polymerase in the TNT system (Promega, Madison, WI). The incorporation of [35S]methionine (>1000 Ci/mmol, NEN Life Science Products) into the protein and the use of an autoradiography enhancer (NEN Research Products) allowed its visualization after autoradiography. Prior to the cleavage assay, immunoprecipitation with a specific antiserum to DSEB/RF-C140 (33) was performed to eliminate the possible interference of other reticulocyte constituents with the protease cleavage assay.
In Vitro Proteolytic Cleavage AssayIncubation of in vitro translated DSEB/RF-C140 with a cytoplasmic extract from Fas antibody-treated or non-treated Jurkat cells, or a bacterial extract containing recombinant ICE or CPP-32, gift from Dr. J. Yuan (Harvard University, Boston, MA), was performed at 37 °C during 1 h in a final 20-µl volume. After addition of Laemmli buffer, samples were loaded on a 10% SDS-PAGE, and the gel was dried and autoradiographed. Cytosolic extracts were prepared by two different methods. No difference between their cleavage capacity was observed. The first method was a modification of a previously described method (13). Briefly, after incubations, cells were washed twice in cold phosphate-buffered saline and resuspended in 100 µl of extraction buffer containing 50 mM PIPES-NaOH, pH 7.0, 50 mM KCl, 5 mM EGTA, 2 mM MgCl2, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, then repeated cycles of freezing and thawing were applied. The second method used also has been described (39). Briefly, after washing in cold phosphate-buffered saline, cells were resuspended in 400 µl of buffer A containing: 10 mM Hepes, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride, then incubated on ice during 15 min and disrupted by vortexing after addition of 25 µl of 1% Nonidet P-40.
DNA Degradation AssayGenomic DNA was isolated by a method described previously (40). Aliquots of 2 × 106 cells were collected and incubated overnight at 37 °C in 100 mM Tris-HCl, pH 8, 5 mM EDTA, 200 mM NaCl, 0.2% w/v SDS, and 0.2 mg/ml proteinase K. Then phenol-chloroform extraction and ethanol precipitation was performed and RNA was digested with RNase A (0.1 mg/ml). After an additional phenol-chloroform extraction and ethanol precipitation, the DNA (5 µg) was resuspended in 20 µl of TE buffer and loaded in a 1.5% agarose gel. DNA was stained with Sybr Green I dye (Molecular Probes, Eugene, OR) and scanned with a FluorImager 575 (Molecular Dynamics, Sunnyvale, CA).
Western ImmunoblotAliquots of cells (2 × 106) processed in parallel with those for DNA degradation assay were washed twice in cold phosphate-buffered saline, and disrupted by the addition of Laemmli buffer prior separation in 10% SDS-PAGE. After overnight transfer to a nitrocellulose membrane, Western immunoblots were performed with a 1:10,000 dilution of the DSEB/RF-C140 primary antiserum raised against its amino-terminal domain (residues 1-510) (33). Horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (Bio-Rad) at 1:25,000 dilution was used as the secondary antibody. Immune complexes were detected with the ECL system (Amersham), according to manufacture's recommendations. PARP cleavage was detected on the same filter by stripping and reprobing with a commercially available PARP monoclonal antibody (PharMingen, San Diego, CA).
As a
consequence of screening cDNA expression libraries for proteins
that bind to a DSE in the angiotensinogen gene promoter (34),
previously we cloned a nuclear DNA-binding protein, DSEB (differentiation specific element-binding protein), the expression of
which is induced during the process of the differentiation of 3T3-L1
adipoblasts into adipocytes (33). Although initially no related
sequences were available in the GenBankTM data base, an identical
cDNA sequence was published that encoded the large subunit of the
DNA replication factor C (41, 42). We have mapped the DNA-binding
domain of DSEB to a region comprising residues 367-493 (33). This
region is structurally related to two major families of proteins:
bacterial DNA ligases and PARP (38). PARP is also a large nuclear
protein, which is able to recognize and bind damaged DNA. PARP is
important in the process of DNA repair because, once bound to DNA, it
becomes active and orchestrates the repair process by ADP-ribosylating
a number of nuclear proteins including itself. During the process of
programmed cell death, PARP is inactivated due to specific cleavage of
the protein into two fragments, one containing the DNA-binding domain
and the other containing the catalytic domain. The active form of
caspase-3 (CPP32/YAMA), an ICE/CED3-like protease, is responsible for
PARP cleavage through recognition of a specific site with the sequence DEVD/G (10). Fig. 1 shows the homology between the amino
acid sequences of DSEB/RF-C140 and PARP. Significant regions of
homology are found in the DNA-binding domain of DSEB/RF-C140 and two
domains of PARP with no ascribed function. A small region in
DSEB/RF-C140 with homology to the "DEAD box" family of RNA
helicases, and residing in the recently mapped domain B required for
PCNA interaction is identical to the consensus DEVD/G CPP32 cleavage
site of PARP.
In Vivo Cleavage of DSEB/RF-C140 during Fas-induced Apoptosis
To determine whether DSEB/RF-C140 is also a substrate
for ICE/CED-3 proteases during programmed cell death, we have used
Jurkat cells treated with anti-Fas-reactive antibodies, a well
established model for apoptosis. Jurkat T cells were incubated with
anti-Fas-reactive monoclonal antibodies, and the appearance of
apoptosis was measured by determining the pattern of migration of DNA
in agarose-gel electrophoresis (DNA laddering) (Fig.
2A). Incubation of Jurkat cells with
increasing concentrations of Fas antibodies causes a DNA laddering
degradation pattern characteristic of apoptosis. Western immunoblot
analyses were performed on the same cell extracts that were examined
for DNA laddering. The antiserum used for the immunoblot analysis was
raised against an amino-terminal fragment of DSEB/RF-C140 (33). The
antiserum detects the DSEB/RF-C140 in control Jurkat cells and, upon
the addition of Fas antibody to the cells, a new immunoreactive protein
with a molecular mass of 87 kDa appears (Fig. 2B). The size
of this new protein is consistent with a cleavage at the predicted
DEVD/G site, between Asp706 and Gly707. By
using the lower effective dose (0.1 µg/ml) of Fas antibody, we
determined whether the cleavage of DSEB/RF-C140 occurred earlier or
later than the Fas-dependent apoptosis as measured by DNA
degradation. Fig. 2C indicates that some DNA degradation is observed
after 1 h of Fas antibody treatment and becomes intense after
2 h. A Western immunoblot prepared from cells incubated in
parallel in the same experiment indicates that a marked cleavage of
DSEB/RF-C140 occurs after 2 h of treatment with Fas antibodies
(see Fig. 2D). After stripping and reprobing the same filter
with a PARP monoclonal antibody (Fig. 2E), we observed that
PARP cleavage occurs also in parallel with DNA degradation, starting
1 h after the addition of the Fas antibody. Although PARP cleavage
seems to occur earlier than DSEB/RF-C140 cleavage, an increased
concentration of PARP per cell or a more accessible pool could also
explain an earlier detection. This finding indicates that the cleavage
of DSEB/RF-C140 does not precede degradation of the DNA, but instead
occurs in conjunction with DNA degradation and the reported cleavage of the other known CPP32 substrates (25, 28).
In Vitro Proteolysis Assays Confirm That Fas-dependent Cleavage Occurs at Asp706
An in vitro
proteolysis system was established to study the cleavage pattern in
DSEB/RF-C140 induced by cytosolic extracts from Fas-treated Jurkat T
cells. DSEB/RF-C140 synthesized by coupled cell-free transcription and
translation in a reticulocyte lysate was isolated by
immunoprecipitation and incubated with a cytosolic extract from control
or Fas-stimulated cells. In the presence of Fas-treated but not
untreated cytosolic extracts, DSEB/RF-C140 was cleaved into two smaller
fragments (Fig. 3), the large fragment corresponding in
molecular weight to the amino-terminal fragment observed in the Western
immunoblot studies with an antibody that recognizes the amino-terminal
domain of DSEB/RF-C140. A new polypeptide with an estimated molecular
size of 53 kDa was also observed. Depending on the electrophoresis
conditions, the smaller protein fragment migrates as a duplex, perhaps
indicating the existence of two different conformations or a secondary
cleavage event. Although both the size of the cleaved fragments and the
immunoreactive pattern indicate that the predicted CPP32 consensus,
DEVDG(703-707) is the cleavage site during Fas-dependent
apoptosis, we confirmed the possibility by replacing the
Asp706 residue with an alanine by site-directed
mutagenesis. As was anticipated, this new mutant protein was not
cleaved by a Fas antibody-treated extract (Fig. 3). A similar Asp to
Ala substitution in another CPP32 substrate, steroid response
element-binding protein (SREBP), was also shown to prevent cleavage by
CPP32 (27).
DSEB/RF-C140 Cleavage Induced by Fas Treatment Is Due to Caspase-3 (CPP32) or a Closely Related Protease
We further characterized
the proteolytic activity responsible for DSEB/RF-C140 cleavage in
Fas-treated Jurkat cells. Two different classes of proteases have been
determined to be activated during Fas-dependent apoptosis
in T cells: 1) a very early initial ICE-like activity, inhibited
specifically by the tetrapeptide aldehyde Ac-YVAD-CHO; and 2) a later
appearance of a CPP32-like activity, inhibited by the tetrapeptide
aldehyde Ac-DEVD-CHO. By using the in vitro cleavage assay,
we examined the capacity of these different compounds to inhibit
DSEB/RF-C140 cleavage activity in cytosolic extracts prepared from
Jurkat cells treated with Fas antibodies. Fig.
4A shows that only the nonspecific
cysteine-protease inhibitor iodoacetamide and the CPP32-specific
inhibitor Ac-DEVD-CHO prevent the cleavage of DSEB/RF-C140 by the
cytosolic extracts prepared from cells undergoing apoptosis. A
non-cysteine protease inhibitor such as leupeptin and the ICE-specific
inhibitor AC-YVAD-CHO did not prevent cleavage. These experiments
indicate that CPP32, or a very close homolog of CPP32, which shares its
pattern of response to certain protease inhibitors, is responsible for
the cleavage of DSEB/RF-C140. The cowpox-virus CrmA serpin, a potent
inhibitor of ICE without a significant effect on caspase-3 activity,
does not prevent DSEB/RF-C140 cleavage, further indicating that ICE is
not the protease responsible for DSEB/RF-C140 cleavage in cytosolic extracts from Fas antibody-treated cells (Fig. 4B). It has
been shown that activation of ICE-like proteolytic activity is a very early event during Fas-induced apoptosis, whereas activation of CPP32
occurs later in parallel with the degradation of DNA (13). These data,
reported earlier, are in agreement with our observations.
Specific DSEB/RF-C140 Cleavage by Recombinant Caspase-3 (CPP-32)
Further confirmation that only caspase-3 (CPP-32) and not caspase-1 (ICE) is the proteolytic activity responsible for DSEB/RF-C140 cleavage was obtained by using a bacterial preparation of recombinant proteases (gift from Dr. J. Yuan). Fig. 4C shows that when "in vitro" translated DSEB/RF-C140 is incubated in the presence of recombinant CPP-32, cleavage occurs. This cleavage has the same pattern as the one induced by cytoplasmic extracts from Fas-treated Jurkat cells. On the other hand, recombinant ICE (caspase-1) has no cleavage activity on DSEB/RF-C140 as is predicted by the studies using specific inhibitors.
DSEB/RF-C140 Cleavage Inactivates Its Functional Role in DNA ReplicationCaspase-3 cleavage of DSEB/RF-C140 at
Asp706 during apoptosis results in disappearance of the
normal full-length protein and the generation of two new fragments.
These two newly formed proteins still contain clusters of functional
domains of the wild-type DSEB/RF-C140 that theoretically could support
some DNA replication activity. However, during the preparation of this
manuscript, an extensive study of the replication function of the
RF-C140 was published (50). In this study two deletion mutants,
p140C555 (residues 1-555) and p140C976 (residues 1-976), which
comprise the amino-terminal fragment (residues 1-706) generated by
caspase-3 during programmed cell death, lack the capacity to form the
replication complex and therefore are inactive in DNA replication (Fig.
5). Two additional deletion mutants, also lacking DNA
replication activity, comprise the carboxyl-terminal fragment (residues
707-1131) generated by caspase-3 cleavage; these are p140N687
(residues 687-1131) and p140N604 (residues 604-1131) (see Fig. 5 and
Ref. 50). Because the reported deletion mutants deficient in DNA replication comprise the same clusters of functional domains as the two
caspase-3 generated fragments, it is concluded that both fragments are
inactive and do not support DNA replication.
An Emerging Common Pattern of CPP32 Substrate Cleavage
Activation of CPP32, as defined by inhibition by the
tetrapeptide Ac-DEVD-CHO, is an early event during programmed cell
death and its exact role in apoptosis is unknown. Activated CPP32 has been shown to cleave only a selected number of nuclear substrates: PARP
(10), DNA-PK (25, 28), and the U1-70 kDa subunit of the small nuclear
ribonucleoprotein complex (20). Our study adds a fourth protein,
DSEB/RF-C140, to this relatively short list of specific nuclear
substrates. In all these instances, cleavage occurs between an
anchoring domain (DNA- or RNA-binding domains) and a functional or
catalytic domain (Fig. 6). Cleavage of these substrates
occurs in parallel with nucleosomal DNA degradation. This circumstance
may indicate that cleavage of substrates and DNA degradation are not
causally linked and that they occur independently from one another.
This concept has recently received experimental support (43). Because
both PARP and DNA-PK are involved in DNA repair processes, it has been
proposed that their cleavages, if not essential, at least would
facilitate the completion of the apoptotic process by blocking the DNA
repair capacity of those cells committed to the programmed cell death
(44). Because RF-C is required for PCNA loading onto DNA, and PCNA is
not only required for DNA replication but also for the DNA
excision-repair mechanisms (45), it has been suggested that the RF-C
could also participate in certain types of DNA repair processes where
PCNA is involved. If this were the case, cleavage of the DSEB/RF-C140
by disruption of its B domain required for PCNA interaction (36) would
almost certainly impair its role in DNA repair, contributing to the
completion of DNA degradation. However, another and perhaps more
interesting possibility is that suppression of DNA replication,
especially the elongation phase of DNA replication performed by
polymerase and
, is a necessary or facilitating step for the
completion of DNA degradation and the apoptotic process. It seems
unquestioned that both suppression of DNA synthesis, as well as
suppression of DNA repair mechanisms are logical events that might
occur in a cell that is degrading its DNA as part of its commitment to a predetermined and organized suicide program. Another feature that the
known CPP32 substrates share among themselves is that they are abundant
ubiquitously distributed proteins (25). DSEB/RF-C140 itself is present
in at least 10,000 molecules/cell (42). They also share the capacity to
consume ATP as a source of energy. Therefore, the inactivation of the
CPP32 substrate proteins during apoptosis would save a considerable
amount of energy that can be applied to processes of endonucleolytic
cleavages during apoptosis (46). If such a circumstance does exist, it
is expected that the rate of DNA degradation would increase, resulting
in a shortening of the critical time required for the completion of the
apoptotic process.
We thank Marion Matzilevich and Edward Maytin for helpful discussions, Linda Fucci for expert technical assistance, and Townley Budde for help in the preparation of the manuscript. We also thank Dr. J. Yuan for providing us with recombinant ICE and CPP-32 proteins.