Caspase-1 and Caspase-8 Cleave and Inactivate Cellular Parkin*

Lesions in the parkin gene cause early onset Parkinson's disease by a loss of dopaminergic neurons, thus demonstrating a vital role for parkin in the survival of these neurons. Parkin is inactivated by caspase cleavage, and the major cleavage site is after Asp126. Caspases responsible for parkin cleavage were identified by several experimental paradigms. Transient coexpression of caspases and wild type parkin in HEK-293 cells identified caspase-1, -3, and -8 as efficient inducers of parkin cleavage whereas caspase-2, -7, -9, and -11 did not induce cleavage. A D126A parkin mutation abrogates cleavage induced by caspase-1 and -8, but not by caspase-3. In anti-Fas-treated Jurkat T cells, parkin cleavage was inhibited by caspase inhibitors hFlip and CrmA (but not by X-linked inhibitor of apoptosis (XIAP)), indicating that caspase-8 (but not caspase-3) is responsible for the parkin cleavage in this model. Moreover, induction of apoptosis in caspase-3-deficient MCF7 cells, either by caspase-1 or -8 overexpression or by tumor necrosis factor-α treatment, led to parkin cleavage. These results demonstrate that caspase-1 and -8 can directly cleave parkin and suggest that death receptor activation and inflammatory stress can cause loss of the ubiquitin ligase activity of parkin, thus causing accumulation of toxic parkin substrates and triggering dopaminergic cell death.

Parkinson's disease is characterized by a variety of motor dysfunctions including rigidity, tremor, slowness of movement, and disturbance of balance (1). The disease is a consequence of a reduced level of striatal dopamine, which follows the progressive degeneration of dopaminergic neurons in the substantia nigra pars compacta. Lesions in the parkin gene have been reported to be a frequent cause of autosomal recessive juvenile Parkinson's disease (2,3). The parkin gene encodes an intracellular parkin protein of 465 amino acids that contains an N-terminal ubiquitinlike domain and two C-terminal RING finger domains (2). RING finger domains are present in one class of ubiquitin protein ligases, and parkin has been shown to act as such (4 -9).
Protein ubiquitination is a post-translational protein modification that involves the ATP-dependent sequential action of a ubiquitin-activating enzyme (E1), 1 a ubiquitin-conjugating enzyme (E2), and a ubiquitin-protein ligase (E3) (10). The transfer of a ubiquitin molecule from an E2 onto a substrate is mediated by an E3 that specifies the substrate. Polyubiquitin chains provide a targeting signal for protein degradation by the 26 S proteasome (10). The target proteins range from misfolded proteins to highly controlled regulators of cell cycle progression, e.g. cyclins. Parkin expression is up-regulated during unfolded protein stress and has been shown to be a protecting factor in the response against this kind of stress (5,9,11). Several disease-causing parkin mutants do not exhibit ubiquitin protein ligase activity (4 -6), indicating that parkin-mediated protein ubiquitination plays a vital role for the survival of the dopaminergic neurons in substantia nigra.
Apoptosis is a cellular suicide program that is essential for correct development and homeostasis of multicellular organisms. Caspases comprise a family of intracellular cysteine proteases that are central initiators and executioners of apoptosis (12). Once activated, the caspases are responsible for cleavage of selective protein substrates that are cleaved after an aspartate residue. It has been reported that in neurodegenerative diseases, the affected neurons enter a degenerative phenotype that displays features associated with apoptosis, e.g. the presence of activated caspases (13). Apoptosis is generally considered a fast process, where activation of caspases rapidly is followed by cell fragmentation and phagocytosis. However, neurons show a relative resistance toward caspase activity that may allow a low level caspase activity to persist in long term neurodegenerative processes (14).

Miscellaneous-All
Cell Lines-HEK-293T cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 1 mM Lglutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. HEK-293T cells were transiently transfected using calcium phosphate precipitation in six-well culture dishes with the indicated DNA constructs. Twenty-four h after transfection, cells were harvested and washed with ice-cold phosphate-buffered saline, and their proteins were extracted for immunoblot analysis as indicated. In the experiments using different caspases, cells were co-transfected with 50 ng of pNLS-EGFP, 300 ng of pcDNA3.1-parkin (WT or D126A mutant) or empty vector, and 600 ng of caspase-containing plasmid or empty vector. In experiments with the caspase-inhibiting proteins, cells were co-transfected with 50 ng of pNLS-EGFP; 100 ng of pcDNA3.1-parkin (WT or its D126A mutant) or empty vector; 200 ng of caspase-containing plasmid or empty vector; and 1000 ng of pEF1A-CrmA, pEF1A-XIAP, pEF1A-hFlip, or empty vector.
MCF7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 1 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. The cells were transfected with 1 g of plasmid using FuGENE 6 transfection reagent according to the manufacturer's instructions (Roche). Stable transfected cell lines were selected in 100 g/ml Zeocin and tested for expression of parkin by immunoblotting. Apoptosis was induced by incubating cells with either 200 nM okadaic acid, 2 M staurosporine, 14 M camptothecine, or 30 ng/ml TNF-␣ and 10 g/ml cycloheximide (Chx) for 24 h. For transient co-transfection experiments, MCF7 cells were cotransfected with 400 ng of parkin or D126A construct, 400 ng of caspase construct, and 200 ng of pcDNA3.1 ␤-galactosidase construct and harvested for immunoblotting after 24 h.
Jurkat T cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, 1 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Cells were transfected with DMRIE-C reagent according to the manufacturer's instructions (Invitrogen). Cells were seeded at 5 ϫ 10 5 cells/ml on the day prior to transfection. Each transfection was performed in duplicate using 2 ϫ 10 6 cells. 2.5 g of pcDNA3-parkin (WT or D126A mutant) or empty vector were co-transfected with 5 g of pEF1A-CrmA, pEF1A-XIAP, pEF1A-hFlip or empty vector, and 1 g of pNLS-EGFP. Liposome-DNA complexes were allowed to form for 30 min at room temperature in 1 ml of Opti-MEM I serum-free medium (Invitrogen). Cells, washed and resuspended in the same medium, were added to the transfection solution and incubated for 4 h at 37°C in a CO 2 incubator, and the medium was reconstituted to normal culture conditions by adding 2 ml of 15% fetal bovine serum growth medium. After 24 h, cells transfected with the same DNA combinations were pooled together and re-divided into two wells. The day after, apoptosis was induced in one of each of the transfection sets by treatment with 250 ng/ml CH-11 anti-Fas antibody (BioCheck, Inc., Burlingame, CA). Cell death was monitored by trypan blue exclusion. Six h after the addition of anti-Fas antibody, cells were collected by centrifugation, washed with ice-cold phosphate-buffered saline, and lysed on ice with caspase lysis buffer.
Protein Extraction and Immunoblot Analysis-Cells were collected by centrifugation, washed with ice-cold phosphate-buffered saline, and lysed on ice with caspase lysis buffer containing 220 mM mannitol, 68 mM sucrose, 2 mM NaCl, 2.5 mM KH 2 PO 4, 10 mM Hepes, pH 7.4, 1 mM aprotinin, 1 mM leupeptin, and 100 M phenylmethylsulfonyl fluoride supplemented with 0.05% Nonidet P-40 and 1 mM oxidized glutathione. Protein concentrations were measured using Bio-Rad protein assay and 30 g from each sample were taken for immunoblot analysis.
Electrophoresis, immunoblotting, and the T160 antibody specific for parkin was prepared as described previously (15). The antibody recognizing poly(ADP-ribose) polymerase (PARP) was from BIOMOL Research Laboratories, Inc. (Plymouth Meeting, PA), and the antibody against caspase-8 was from BioSource Europe, S.A. (Nivelles, Belgium). Membranes were incubated with horseradish peroxidase-conjugated secondary antibodies to mouse and rabbit immunoglobulin (Amersham Biosciences). Proteins were visualized using a chemiluminescence substrate (DuPont), and blots were exposed to film (Amersham Biosciences).

Coexpression of Parkin with Caspase-1, Caspase-3, or
Caspase-8 Induces Parkin Cleavage-To identify caspase(s) responsible for cellular parkin cleavage, we co-transfected parkin cDNA with cDNAs encoding different caspases and a GFP construct in HEK-293T cells. To exclude the possibility that an absent induction of parkin cleavage was caused by inactive caspases, we analyzed the morphology of the GFP-transfected cells by microscopy. All cell cultures co-transfected with caspases showed a high degree of nonattached blebbing cells, whereas no cell death was seen in cells transfected with the inactive caspase-9s mutant or parkin alone (data not shown). A high degree of parkin cleavage was apparent when the protein was coexpressed with caspase-1 or caspase-8, whereas caspase-3 was less efficient (Fig. 1). In contrast, no parkin cleavage was observed when the protein was expressed alone or coexpressed with caspase-2, caspase-7, caspase-11, caspase-9, or the inactive caspase-9s mutant. We, therefore, conclude that overexpression of caspase-1, -3, or -8 induces parkin cleavage, whereas caspase-2, -7, -9, or -11 does not.
The Asp 126 -Ser 127 peptide bond represents the major cellular caspase cleavage site in parkin, although some quantitatively minor but significant caspase-mediated degradation does occur at other locations (15). To test caspase-1, -3, and -8 for their specificity for the Asp 126 -Ser 127 peptide bond, we expressed these three caspases in HEK-293T cells either with wild type parkin or with the D126A parkin mutant. Wild type parkin was cleaved when it was coexpressed with any of the three caspases ( Fig. 2A). The D126A mutation abrogated parkin cleavage when coexpressed with caspase-1 and caspase-8, whereas some cleavage still occurred with caspase-3 (Fig. 2B).
Parkin Cleavage by Caspase-1 or -8 Occurs in MCF7 Cells in a Caspase-3-independent Manner-Caspase-8 can cleave and activate the effector caspase-3 (12). To assure that the cleavage induced by caspase-1 or -8 in HEK-293T cells (Fig. 2) is not because of a secondary activation of caspase-3, we coexpressed wild type and D126A mutant parkin with caspase-1, -3, and -8 in the caspase-3-deficient MCF7 cell line (24). As seen for HEK-293T cells (Fig. 2), all three tested caspases induced cleavage of parkin in MCF7 cells (Fig. 3A), and the D126A mutation significantly abrogated the cleavage induced by caspase-1 and -8 (Fig. 3B). This demonstrates that the caspase-1-and -8-induced cleavage of parkin after Asp 126 is not a cell-specific event and does not require activation of endogenous caspase-3. The role of caspase-1 and -8 was further corroborated by experiments where in vitro translated parkin was cleaved by purified recombinant caspase-1 and caspase-8 (data FIG. 1. Coexpression of parkin with different caspases identifies caspase-1, caspase-3, and caspase-8 as inducers of parkin cleavage. HEK-293T cells were co-transfected with parkin and either an empty vector (CTRL), caspase-2, -3, -7, -11, -1, -8, -9, or -9s, and parkin cleavage was determined after 24 h. Immunoblots were probed with the T160 parkin antibody. Positions of full-length parkin and the large parkin cleavage product are indicated by arrowheads to the right. Positions of molecular mass markers in kDa are indicated to the left. One of two similar experiments is presented. not shown), demonstrating that the cleavage of parkin induced by co-transfecting parkin with either caspase-1 or -8 can result from a direct action of the co-transfected caspases.
Activation of caspase-8 can be mediated through an extrinsic pathway where extracellular ligands bind to plasma membrane receptors, e.g. the TNF-␣ receptors (25), or through an intrinsic pathway where activated caspase-3 can cleave and activate caspase-8 (12). To test whether activation of endogenous caspase-8 can cause parkin cleavage independently of endogenous caspase-3, a MCF7 cell line stably transfected with parkin was generated. We have previously shown that stimuli that induce apoptosis through the intrinsic apoptosome-mediated cell death pathway, such as okadaic acid (26), staurosporine (27), and camptothecine (28), cause parkin cleavage in SH-SY5Y and Chinese hamster ovary cell lines with intact apoptotic machinery (15). However, okadaic acid, staurosporine, and camptothecine failed to induce cleavage of parkin in the caspase-3-deficient MCF7 cell line where caspase-8 cannot be activated by caspase-3 (Fig. 3C). In contrast, apoptosis initiated by TNF-␣ and Chx that leads to direct caspase-8 activation through the TNF-␣ receptors (25) caused cleavage of parkin (Fig. 3C). As a negative control, cells treated with Chx alone did not cause parkin cleavage. The caspase-8 cleavage was dependent on Asp 126 , as no parkin proteolysis was observed when similar apoptotic stimuli (including TNF-␣ and Chx) were applied to a stable transfected MCF7 cell line expressing the D126A parkin mutant (Fig. 3D). Induction of apoptosis in all of these experiments was verified by bisbenzimide staining for apoptosis-associated nuclear condensation and fragmentation (data not shown). Moreover, immunoblot analyses showed that pro-caspase-8 is processed specifically in the presence of TNF-␣ and Chx but not in the presence of any of the stimuli that induce apoptosis through the intrinsic apoptosome-mediated cell death pathway (Fig. 3, C and D). This demonstrates that in a caspase-3-deficient cell line, parkin cleavage can be induced specifically through activation of endogenous caspase-8 and occurs after Asp 126 .

Endogenous Caspase-8, but Not Endogenous Caspase-3, Cleaves Parkin in Anti-Fas-mediated Apoptosis in Jurkat T Cells-We used Jurkat T cells expressing the Fas receptor to
test how the presence of endogenous caspase-3 and caspase-8 affects parkin cleavage in Fas-mediated apoptosis. Incubation of these cells with an agonistic anti-Fas antibody leads to activation of caspase-8 that cleaves and activates caspase-3 (12). When Jurkat T cells were transfected with a parkinexpressing plasmid followed by incubation with anti-Fas antibody, parkin cleavage appeared (Fig. 4A). In contrast, no cleav-age products were observed in anti-Fas-treated cells transfected with the D126A plasmid or an empty plasmid (Fig.  4A). Cell death monitored by trypan blue exclusion and typical apoptotic morphology was observed in all of the cell populations incubated with the anti-Fas antibody (data not shown). Furthermore, immunoblot analyses with an antibody directed against PARP, a typical caspase-3 substrate (29), showed that the protein was cleaved, and thus caspase-3 was activated in cells that were incubated with anti-Fas (Fig. 4A).
CrmA, XIAP, and hFlip are selective protein inhibitors of specific caspases. XIAP inhibits the activity of caspase-3. CrmA and hFlip (but not XIAP) inhibit the activity of caspase-8, whereas hFlip does not inhibit the activation of either caspase-1 or caspase-3 (12). To verify that caspase-8 (and not caspase-3) cleaves parkin in this system, we co-transfected parkin with either CrmA, XIAP, and hFlip and induced apoptosis with anti-Fas. Cleavage of parkin was inhibited by coexpression with hFlip and CrmA (but not with XIAP) demonstrating that in Fas-mediated apoptosis in Jurkat T cells endogenous caspase-8 (and not endogenous caspase-3) is responsible for parkin cleavage (Fig. 4B). We coexpressed the three caspase inhibitors with parkin and either caspase-1, -3, or -8 in HEK-293T cells to confirm their inhibition selectivity. CrmA and hFlip inhibited parkin cleavage when coexpressed with caspase-8 (Fig. 4C). XIAP inhibited the cleavage of parkin by caspase-3 but failed to prevent cleavage by caspase-8 demonstrating that the inhibitor is functional (Fig. 4C). These results demonstrate that in anti-Fas-induced apoptosis in Jurkat T cells, it is activated endogenous caspase-8 (and not activated caspase-3) that cleaves parkin. DISCUSSION The recessive heritability of Parkinson's disease caused by parkin gene lesions underscores the importance of parkin function for the survival of the dopaminergic neurons of substantia nigra pars compacta and suggests that its dysfunction can contribute to the disease progression in sporadic Parkinson's disease. This is corroborated by the detection of cleaved parkin fragments in substantia nigral tissue affected by Parkinson's disease and in Lewy bodies isolated from diseased tissue (30,31). Moreover, proteolytic inactivation of parkin may play a broader role in neurodegenerative settings as recently demonstrated by parkin's rapid catabolism upon hypoxia (32). Growing evidence suggests that inappropriate activation of apoptotic processes is involved in the pathology of degenerative diseases, including sporadic Parkinson's disease (13,33), and caspase activation could represent a link to parkin dysfunction in sporadic Parkinson's disease. Indeed, we recently demonstrated that parkin is cleaved by caspases after Asp 126 (15). Here we show that cellular parkin can be cleaved by caspase-1 and -8. Parkin is cleaved upon coexpression with caspase-1, -3, and -8, but only caspase-1 and -8 cleave parkin selectively after Asp 126 ( Figs. 1 and 2). Furthermore, purified recombinant caspase-1 and -8 cleaved in vitro translated parkin (data not shown). The caspase-3-deficient MCF7 cell line allowed us to study parkin cleavage without interference from secondary effects caused by caspase-3. In this model, parkin was cleaved upon overexpression of either caspase-1 or -8 and upon activation of endogenous caspase-8 via the TNF receptor pathway (Fig. 3). Agents that activate the intrinsic cell death pathway through the Apaf-1 apoptosome complex and activation of caspase-9 did not lead to proteolysis of parkin in the MCF7 cell line. Furthermore, the prevention of parkin cleavage by the caspase inhibitors CrmA or hFlip demonstrated that endogenous caspase-8 cleaves parkin in the presence of activated caspase-3 in Fas-mediated apoptosis in Jurkat T cells (Fig. 4). This is further corroborated by overexpression of XIAP, which did not affect parkin proteolysis, thus excluding the involvement of caspase-3, -7, and -9 in parkin cleavage in anti-Fas-induced apoptosis. It is unclear how parkin is processed when caspase-3 is overexpressed in cells, as caspase-3 is not specific for the Asp 126 (Figs. 2 and 3), but it could be because of secondary activation of downstream caspases.
Mounting evidence associates activation of caspase-8, which relays death signals from the death domain receptors for TNF-␣ and Fas ligand and the proinflammatory caspase-1, with the pathogenesis of several neurodegenerative disorders including Parkinson's disease (13,33,34). Elevated levels of the proinflammatory cytokines interleukin-1␤ and TNF-␣ are present in the nigro-striatal dopaminergic system as well as in the cerebrospinal fluid of patients with Parkinson's disease (34 -36). This demonstrates, first, an increase in the activation of the pro-interleukin-1␤ converting enzyme (caspase-1) and, second, an increased potential for stimulating the caspase-8 activating TNF-␣ receptors on the dopaminergic neurons. This is corroborated by the increased activity of caspase-1 in substantia nigra from patients with Parkinson's disease as compared with substantia nigra from control patients (37) and by the demonstration of increased levels of activated caspase-8 in dopaminergic neurons of the substantia nigra in patients with Parkinson's disease compared with control patients (38,39). Caspase-8 is also activated in the MPTP mouse model of Parkinson's disease (38,39).
Caspase-1 and -8 are activated as initial events in the process of apoptotic cell death (12,40,41). Because parkin is a direct substrate of caspase-1 and -8, parkin cleavage might be important for the degeneration of dopaminergic neurons.

FIG. 4. Endogenous caspase-8 (and not endogenous caspase-3) cleaves parkin in anti-Fas-mediated apoptosis in Jurkat T cells.
A, Jurkat T cells were transfected with wild type parkin, D126A mutant parkin, or an empty vector. Positions of processed and full-length parkin and PARP are indicated by arrowheads to the left. B, Jurkat T cells were co-transfected with a parkin plasmid and either CrmA, XIAP, hFlip, or an empty vector. Cells were cultured in the presence of (ϩ) or the absence of (Ϫ) the anti-Fas antibody. Immunoblots were probed with the T160 parkin antibody (top panel in A and B) or an antibody recognizing PARP (lower panels in A and B). Positions of processed and full-length parkin and PARP are indicated by arrowheads to the left. Note that parkin cleavage is inhibited in the presence of CrmA and hFlip, but not XIAP. C, HEK-293T cells were co-transfected with parkin, caspase-1 (first panel), caspase-3 (second panel), or caspase-8 (third panel) and either CrmA, XIAP, or hFlip. Controls from cells co-transfected with parkin and caspase-1, -3, -8, or an empty vector (CTRL) without caspase inhibitors from the same experiment are shown in the fourth panel. Nontransfected cells (Mock) are also shown. Immunoblots were probed with the T160 antibody. Positions of full-length and cleaved parkin are indicated by arrows to the left. The molecular mass markers in kDa are indicated to the right for all panels. One of two similar experiments is presented.
Caspase-1 and caspase-8 have also been reported to regulate cellular processes other than apoptosis and are activated without causing apoptotic cell death in macrophages and T cells, respectively (42,43). If similar activation of these caspases occurs in dopaminergic neurons, it may cause parkin cleavage and affect the viability of these cells. We, therefore, hypothesize that caspase-1-and caspase-8-dependent parkin cleavage in sporadic Parkinson's disease may play an important role in the degenerative process by initiating a vicious circle that leads to the accumulation of toxic parkin substrates, e.g. ␣-synuclein (44). The failure to remove these substrates may lead directly to cell death or lower the cellular threshold to further stressful insults. The process may be long lasting, as neurons are relatively resistant to caspase activity (14) and thus may remain functionally crippled but viable. Therefore, inhibition of parkin cleavage could represent a potential target for neuroprotective strategies against Parkinson's disease aiming to delay the functional decline of dopaminergic neurons. Such an approach may have a wider applicability when considering that active forms of caspase-1 or -8 were also detected in diseases like Huntington's disease (45), amyotrophic lateral sclerosis (46), and ischemia (47).