Different Subcellular Distribution of Caspase-3 and Caspase-7 following Fas-induced Apoptosis in Mouse Liver*

Caspases plays a key role in the execution phase of apoptosis. “Initiator” caspases, such as caspase-8, activate “effector” caspases, such as caspase-3 and -7, which subsequently cleave cellular substrates thereby precipitating the dramatic morphological changes of apoptosis. Following treatment of mice with an agonistic anti-Fas antibody to induce massive hepatocyte apoptosis, we now demonstrate a distinct subcellular localization of the effector caspases-3 and -7. Active caspase-3 is confined primarily to the cytosol, whereas active caspase-7 is associated almost exclusively with the mitochondrial and microsomal fractions. These data suggest that caspases-3 and -7 exert their primary functions in different cellular compartments and offer a possible explanation of the presence of caspase homologs with overlapping substrate specificities. Translocation and activation of caspase-7 to the endoplasmic reticulum correlates with the proteolytic cleavage of the endoplasmic reticular-specific substrate, sterol regulatory element-binding protein 1. Liver damage, induction of apoptosis, activation and translocation of caspase-7, and proteolysis of sterol regulatory element-binding protein 1 are all blocked by the caspase inhibitor, benzyloxycarbonyl-Val-Ala-Asp fluoromethyl ketone (Z-VAD.fmk). Our data demonstrate for the first time the differential subcellular compartmentalization of specific effector caspases following the induction of apoptosis in vivo.

Apoptosis is a crucial mechanism by which multicellular organisms control cell numbers and ensure the removal of damaged or potentially harmful cells (1). Administration of an agonistic anti-Fas antibody results in ligation of the Fas (CD95, APO-1) receptor, extensive hepatocyte apoptosis, and liver damage (2). The intracellular death domain of the Fas receptor binds to FADD/MORT1, which in turn recruits and activates caspase-8 (MACH/FLICE/Mch5) through its N-terminal death effector domain (3)(4)(5). Recombinant caspase-8 cleaves and activates all other known caspases and has been proposed to be at the apex of a hypothetical caspase cascade (3)(4)(5). Caspases are a family of aspartate-specific cysteine proteases, which preexist in the cytoplasm as single chain inactive zymogens (6,7). They are proteolytically processed to active heterodimeric enzymes during the execution phase of apoptosis. Caspases may be divided into "initiator" caspases with long prodomains (caspases-8, -9, and -10), which activate "effector" caspases with short prodomains (caspases-3, -6, and -7), which in turn cleave intracellular substrates, resulting in the dramatic morphological and biochemical changes of apoptosis (6 -8). Following Fas-induced apoptosis of cells in vitro, activation of a number of caspases, including caspases-3, -4, -6, -7, -8, and a caspase-1-like activity have all been reported (4, 9 -13).
To date, a family of at least 10 caspases have been identified, but it is not known precisely which of these caspase(s) are activated in vivo and which are responsible for the cleavage of particular substrates. Many of the caspases have overlapping substrate specificities, suggesting that there may be redundancy (6,7,14). Many cellular proteins are cleaved during the execution phase of apoptosis at a DXXD motif by the effector caspases-3 and -7 (reviewed in Refs. 6 and 7). Relatively little is known about the subcellular distribution of the caspases. Caspase-1 is found predominantly in the cytosol (15) although some has been localized to the external cell surface membrane (16). Other caspases have been considered to be cytosolic, a conclusion based largely on data with caspase-1 and on the isolation and purification of caspase-3 (17). In this study, we demonstrate for the first time the differential subcellular distribution of specific caspases during the induction of apoptosis in vivo. Following Fas-induced apoptosis in vivo, active caspase-3 is found primarily in the cytosol, whereas active caspase-7 is associated almost exclusively with the mitochondrial and microsomal fractions. Both the activation of caspase-7 in the endoplasmic reticulum and the cleavage of the endoplasmic reticular-specific substrate, sterol regulatory element-binding protein 1 (SREBP-1), 1 are blocked by the caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp fluoromethyl ketone (Z-VAD.fmk). These results support the hypothesis that during the execution phase of apoptosis, different caspase homologs cleave specific substrates in different cellular compartments.

EXPERIMENTAL PROCEDURES
Mice-In this study 6 -8-week-old (20 g) Balb/c males were used. All mice were bred in the Biomedical Sciences Department of the University of Leicester.
Tissue Preparation and Histopathological Examination-Livers were removed and fixed in 10% formaldehyde in buffered saline. Rep-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Fractionation of Liver-Following removal of the livers, excess hair and blood were removed by washing several times in buffer A (0.3 M mannitol, 5 mM MOPS, 1 mM EGTA, 4 mM KH 2 PO 4 ). The livers were then chopped up and homogenized using a dounce homogenizer in 5 ml of buffer A supplemented with protease inhibitors (20 g/ml leupeptin, 10 g/ml pepstatin A, 10 g/ml aprotinin, 2 mM phenylmethylsulfonyl fluoride). The crude homogenates were centrifuged at 650 ϫ g for 10 min at 4°C and the resultant supernatant centrifuged at 10,000 ϫ g for 15 min at 4°C to sediment the mitochondria. The mitochondria were washed in supplemented buffer A and pelleted. The microsomal and cytosolic fractions were obtained following centrifugation of the 10,000 ϫ g supernatant fraction at 100,000 ϫ g for 45 min at 4°C. Purity of the mitochondrial, cytosolic, and microsomal fractions was assessed by Western blotting using antibodies to cytochrome c oxidase subunit IV (Molecular Probes, Eugene, OR), glutathione S-transferase (18) (kindly provided by Dr. M. Manson, Medical Research Council Toxicology Unit) and SREBP-1. Cytochrome c oxidase subunit IV is located on the inner mitochondrial membrane (19), glutathione S-transferase is a cytosolic enzyme (20), and SREBP-1 is located in the endoplasmic reticulum (21). Densitometric analysis revealed that 77, 9, and 14% of total glutathione S-transferase and 5, 93, and 2% of total cytochrome c oxidase subunit IV were detected in the cytosolic, mitochondrial, and microsomal fractions, respectively. SREBP-1 was found almost exclusively in the endoplasmic reticulum ( Fig. 4 and data not shown).
Western Blotting-SDS-PAGE and Western blotting were carried out on liver fractions as described previously (22). The membranes were probed using a rabbit polyclonal antibody to the p17 subunit of caspase-3 (kindly provided by Merck Frosst, Quebec, Canada) (11, 23), a rabbit polyclonal antibody to the p17 fragment of caspase-7 (22), and a rabbit polyclonal antibody to amino acids 470 -479 of human SREBP-1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).

Fas-induced Apoptosis, Liver Damage, and Caspase 3/7-like Proteolytic Activity Are Blocked by Z-VAD.fmk-The agonistic
Fas receptor antibody JO2 induced extensive liver damage and hemorrhage in Balb/c mice, with Ͼ60% of hepatocytes showing apoptotic morphology after 4 h (Fig. 1B) (2). The caspase inhibitor, Z-VAD.fmk (50 mol/kg) blocked Fas-induced liver damage and hemorrhage and dramatically reduced hepatocyte apoptosis to Ͻ1% (Fig. 1C). Almost complete protection was still observed 24 h after exposure to the agonistic antibody (data not shown). The protection conferred by Z-VAD.fmk suggested a critical role for the activation of caspases in Fasinduced apoptosis in vivo, consistent with previous studies (24,25).
To date, very few studies on Fas-induced apoptosis have been carried out in vivo, most being in vitro (9 -13). In order to confirm the involvement of caspases in Fas-mediated apoptosis in vivo, their activation was assessed by measuring DEVDase in crude liver homogenates from control and treated mice. This activity is primarily a measure of caspase-3 and caspase-7 activities, although there may be a minor contribution from other caspases (14,17,26,27). Treatment with the agonistic Fas antibody resulted in a marked increase in total liver DEVDase (270 nmol/min) after 4 h in comparison with that in controls (80 nmol/min). These results demonstrated the activation of the effector caspase-3 and/or caspase-7 in Fas-induced apoptosis in vivo in agreement with in vitro studies, which have shown the activation of these caspases following treatment of cells with Fas or tumor necrosis factor (11,28,29). In order to further dissect the role of caspases in Fas-induced apoptosis in vivo, we examined DEVDase and caspase processing in subcellular liver fractions prepared from control and Fas-treated mice. Fas treatment induced 11-, 21-, and 23-fold increases in total DEVDase in cytosolic, microsomal, and mitochondrial fractions, respectively. In all cases, Z-VAD.fmk markedly inhibited the increases in DEVDase. Thus Z-VAD.fmk blocked apoptosis either by directly inhibiting the activity of caspases-3 and -7 or by inhibiting an upstream caspase, such as caspase-8.
Procaspase-3 and Active Caspase-3 Are in the Cytosol-Caspase-3 is generally present in control cells as an inactive p32 zymogen (11,22,29). On induction of apoptosis, it is initially processed at Asp-175 between the large and small subunits, yielding a p20 subunit, which is further processed at Asp-9 and Asp-28 to yield p19 and p17 large subunits, respectively (17,30). In control mice, procaspase-3 was present in the cytosolic fraction (Fig. 2, lane 3) with none detected in the mitochondrial or microsomal fractions (Fig. 2, lanes 1 and 6). Following treatment with the Fas antibody, complete processing of procaspase-3 together with the appearance of its catalytically active p17 subunit was observed in the cytosolic fraction (Fig. 2, lane 4). In addition, immunologically reactive fragments of ϳ29 kDa (Fig. 2, lanes 3-7) and ϳ25 kDa (Fig. 2,  lane 4) were also observed. The ϳp29 fragment was observed in all the cytosolic and microsomal fractions (Fig. 2, lanes 3-7). While the identity of this fragment is not known, it has been proposed to be due to processing of caspase-3 following cleavage at Asp-28, yielding a zymogen, which is not further processed to active caspase-3 (23). Treatment with Z-VAD.fmk resulted in almost complete inhibition of the processing of procaspase-3, formation of the p17 large subunit, and the p25 fragment (Fig.  2, lane 5). Inhibition of caspase-3 processing was also accompanied by the appearance of a very small amount of an ϳp19 fragment (Fig. 2, lane 5), which may be attributed either to irreversible binding of Z-VAD.fmk to the p17 subunit or to partial blocking of the processing of the large subunit (22,31). The p17 subunit was detected primarily in the cytosolic fraction of livers from Fas-treated mice (Fig. 2, lane 4) with little if any being present in any other fraction (Fig. 2). These data demonstrate that following Fas induction of apoptosis, active caspase-3 is located primarily in the cytosol.
Active Caspase-7 Is Primarily in the Microsomal and Mitochondrial Fractions-Caspase-7 exists as an inactive p35 zymogen in control cells (22,28,32). On induction of apoptosis it is activated by initial processing at Asp 198 between the large and small subunits followed by cleavage at Asp-23 to yield the catalytically active p19 large subunit (22,27,28). In livers from control mice, caspase-7 was present as an unprocessed p35 proform in both the cytosolic and microsomal fractions with no detectable p19 subunit (Fig. 3, lanes 4 and 7). While caspase-7 has previously been detected in the cytoplasm of Jurkat cells (28), this is the first time it has also been recognized to have a microsomal location. No detectable procaspase-7 or p19 subunit was present in the mitochondrial fraction from control mouse livers (Fig. 3, lane 1).
Following treatment with the Fas antibody, complete processing of procaspase-7 was observed in both the cytosolic and microsomal fractions (Fig. 3, lanes 5 and 8). Although complete processing of caspase-7 was observed in the cytosolic fraction, little if any p19 subunit was detected (Fig. 3, lane 5). However, the p19 catalytically active large subunit of caspase-7 was clearly detected in both the mitochondrial and microsomal fractions (Fig. 3, lanes 2 and 8). It was very unlikely that the p19 fragment in the mitochondrial fraction was due to microsomal contamination, because the endoplasmic reticular protein SREBP-1 was located exclusively in the microsomal fraction ( Fig. 4 and data not shown). In addition, the uncharacterized p29 band, detected using the caspase-3 antibody, in the microsomal fraction from control or Fas-treated livers (Fig. 2, lanes 6  and 7) was not present in the mitochondrial fraction (Fig. 2,   lanes 1 and 2). The amount of the large p19 subunit of caspase-7 in the microsomal fraction following Fas-induced apoptosis was greater than the amount of procaspase-7 in control liver microsomes (Fig. 3, compare lanes 7 and 8). These results suggested that caspase-7 was translocated from the cytosol to the microsomes following its catalytic activation by an initiator caspase. The data clearly demonstrate that following Fas induction of apoptosis in mouse liver, caspase-7 is completely processed to its catalytically active p19 subunit, which is found primarily in the mitochondrial and microsomal fractions with little if any remaining in the cytosol.
Z-VAD.fmk completely inhibited the Fas-induced cleavage of procaspase-7 as well as the formation of the p19 subunit in all subcellular fractions (Fig. 3, lanes 3, 6, and 9). Z-VAD.fmk also blocked the appearance of an uncharacterized ϳp32 fragment in the microsomal fraction (Fig. 3, lane 9). Taken together with the caspase-3 results, our data suggest that Z-VAD.fmk blocks the processing of both caspase-3 and caspase-7. Although it is not known precisely which caspase activates caspase-3 and caspase-7 during Fas-induced apoptosis, caspase-8 has been considered the most likely candidate (3)(4)(5). Recent studies have demonstrated that procaspase-9 binds to Apaf-1 (apoptotic protease-activating factor 1) in a cytochrome c-and dATP-dependent manner (33,34). This complex results in the activation of caspase-9, which in turn cleaves and activates caspase-3 (34). The mechanism of activation of procaspase-7 is not known, it may be due to activation by caspase-8 (5) or to a mechanism involving Apaf-1 and procaspase-9 (34), but it does not appear to be due to a direct activation by caspase-3 (35). Although activation of caspase-8 is clearly a very early event following Fas-induced apoptosis, it is not yet clear how this is related to the activation of caspases following mitochondrial damage with the subsequent release of cytochrome c and the activation of procaspase-9. Therefore, Z-VAD.fmk may exert its action by blocking the caspase cascade initiated by both caspase-8 and caspase-9.
Further support for the hypothesis that different effector caspases are responsible for the enzymic activity in different subcellular compartments was provided by comparing the Western blot data with DEVDase. DEVDase in cells undergoing apoptosis is believed to be primarily due to activation of caspase-3 and caspase-7 (14,17,26,27). Although the total DEVDase loaded onto the polyacrylamide gel from the cytosolic fraction (20 pmol/min) (Fig. 3, lane 5) was greater than that from either the mitochondrial or microsomal fractions (9 and 15 pmol/min) (Fig. 3, lanes 2 and 8, respectively), the antibody to caspase-7 only detected the p19 large subunit in the mitochon- drial and microsomal fractions (Fig. 3, lanes 2 and 8). This suggested that a caspase other than caspase-7 was primarily responsible for DEVDase in the cytosolic fraction. Most probably this was caspase-3, based on the data demonstrating that the p17 catalytically active large subunit of caspase-3 was primarily located in the cytosol (Fig. 2). Thus our results strongly suggest that the major DEVDase in the microsomal and mitochondrial fractions is due to caspase-7, while in the cytosolic fraction it is due to caspase-3. Taken together, our results suggest that following its activation, caspase-7 is translocated to the microsomal and mitochondrial fractions, where it is responsible for the cleavage of specific substrates in these distinct subcellular compartments. A recent study using an affinity label also noted differences in the pattern of active caspases in the nuclei and cytosol between two cell lines (36). Their results together with the present study raise the question about how different active caspases may be targeted to different subcellular localizations.
Fate of Microsomal SREBP-1 following Fas-induced Apoptosis in Mouse Liver--Many previous studies have highlighted overlapping substrate specificities of caspases-3 and -7. For example, combinatorial studies using tetrapeptide substrates assigned virtually indistinguishable substrate specificities to caspases-3 and -7 (14), and both enzymes effectively cleave poly(ADP-ribose) polymerase (17,26,27). It has often been suggested that there is a redundancy for certain caspases, which may be due to the important biological function of this system in removing damaged or unwanted cells (6,7). However results with caspase-3 knockout mice, which exhibit normal apoptosis in most tissues except neuronal cells, indicate that some caspases may function in a tissue selective manner (37). Alternatively our data suggest that, at least in some tissues, caspases with overlapping substrate specificity may exert their functions in different cellular compartments, where they catalyze the cleavage of specific substrates. To explore this possibility, we examined the fate of SREBP-1, one of the few endoplasmic reticulum-associated proteins known to be cleaved in apoptosis (21,38,39). SREBPs belong to the basic-helix-loophelix-leucine zipper family of transcription factors and are involved in the regulation of sterol metabolism (21). On the induction of apoptosis both SREBP-1 and SREBP-2 are cleaved by the hamster homologs of caspases-3 and -7 (38,39).
In livers from control mice, SREBP-1 was exclusively associated with the microsomal fraction with none being detectable in the cytosolic or mitochondrial fractions (Fig. 4, compare lanes 1 and 2, and data not shown). Complete loss of the ϳ125-kDa SREBP-1 was observed in liver microsomes obtained from mice treated 4 h earlier with the agonistic Fas antibody (Fig. 4, lane 3). The antibody to SREBP-1 detected the intact but not the cleaved molecule. The Fas-induced cleavage of SREBP-1 was largely prevented by Z-VAD.fmk (Fig. 4, lane  4). As active caspase-7 and SREBP-1 share the same subcellular localization, it is possible that in vivo caspase-7 is responsible for the Fas-induced cleavage of SREBP-1. However based on our present data and the contiguous nature of the cytosol and endoplasmic reticulum, we cannot totally exclude the possibility that SREBP-1 may also be cleaved, at least in part, by caspase-3.
We have clearly shown that following Fas-induced apoptosis in vivo, active caspase-3 was located primarily in the cytosol, whereas active caspase-7 was associated with both the mitochondrial and microsomal fractions. Our data represent the first example of the differential subcellular distribution of specific caspases in an in vivo model of apoptosis.