The mycotoxin penicillic acid inhibits Fas ligand-induced apoptosis by blocking self-processing of caspase-8 in death-inducing signaling complex.

Upon engagement with Fas ligand (FasL), Fas rapidly induces recruitment and self-processing of caspase-8 via the adaptor protein Fas-associated death domain (FADD), and activated caspase-8 cleaves downstream substrates such as caspase-3. We have found that penicillic acid (PCA) inhibits FasL-induced apoptosis and concomitant loss of cell viability in Burkitt's lymphoma Raji cells. PCA prevented activation of caspase-8 and caspase-3 upon treatment with FasL. However, PCA did not affect active caspase-3 in FasL-treated cells, suggesting that PCA primarily blocks early signaling events upstream of caspase-8 activation. FasL-induced processing of caspase-8 was severely impaired in the death-inducing signaling complex, although FasL-induced recruitment of FADD and caspase-8 occurred normally in PCA-treated cells. Although PCA inhibited the enzymatic activities of active recombinant caspase-3, caspase-8, and caspase-9 at similar concentrations, PCA exerted weak inhibitory effects on activation of caspase-9 and caspase-3 in staurosporine-treated cells but strongly inhibited caspase-8 activation in FasL-treated cells. Glutathione and cysteine neutralized an inhibitory effect of PCA on caspase-8, and PCA bound directly to the active center cysteine in the large subunit of caspase-8. Thus, our present results demonstrate that PCA inhibits FasL-induced apoptosis by targeting self-processing of caspase-8.

The death receptor Fas is widely expressed in diverse cell types and plays an essential role in the killing of harmful cells by cytotoxic T lymphocytes, as well as regulation of the immune system such as lymphocyte homeostasis (1)(2)(3)(4). Defective apoptotic signals in the Fas/FasL system result in accumulation of abnormal T cells in the peripheral lymphoid organs and develop autoimmune diseases in human and mice (1)(2)(3)(4). In addition to death-inducing function, the Fas/FasL system is also involved in T cell proliferation, as evidenced by reports from us and other groups (5)(6)(7) showing that Fas provides costimula-tory signals for human T cells, and caspase inhibitors block CD3-induced proliferation and interleukin-2 production by human T cells. Accelerated apoptotic signals via the Fas/FasL system are involved in the pathogenesis of fulminant hepatitis and autoimmune diseases such as experimental autoimmune encephalomyelitis, because virus-infected hepatocytes and oligodendrocytes are killed by cytotoxic T lymphocytes in a FasLdependent fashion (8 -10). Therefore, suppression of Fas-mediated signals might have therapeutic potential in the treatment of autoimmune diseases and fulminant hepatitis by blocking T cell-mediated killing and T cell proliferation.
Fas contains a cytoplasmic sequence known as the death domain that is utilized to interact with the adaptor protein Fas-associated death domain (FADD), 1 which contains both a death domain and a death effector domain (1)(2)(3)(4). Membranebound Fas ligand (FasL) induces Fas oligomerization that allows FADD recruitment to the death domain (1)(2)(3)(4). Through the interaction of their mutual death effector domains, FADD binds to caspase-8, which contains two death effector domains and a caspase domain, allowing the formation of death-inducing signaling complex (DISC) (11,12). In the DISC, caspase-8 zymogens are placed in proximity, which facilitates self-processing to generate the active form (13). Activated caspase-8 cleaves downstream substrates such as caspase-3, essential for apoptosis execution (14,15).
Screening System-Five thousand extracts of microbial culture broths and a hundred extracts of herbal medicines were screened to look for specific inhibitors that block DNA fragmentation induced by anti-Fas antibody Jo2 as monitored by [ 3 H]TdR-labeled A20 cells. The first screening led to an identification of 10 positive candidates. One sample did not suppress reduced cell viability of Jo2-treated A20 cells as analyzed by MTT assay. Two samples were excluded because of lack of reproducibility when prepared from new microbial culture broths. The bleb-forming assay using the human chronic myeloid leukemia K562 cells was utilized to detect activators of protein kinase C (21), because activators of protein kinase C block apoptosis induced by anti-Fas antibodies (22,23). Six samples induced a significant blebbing on K562 cells. Therefore, one crude extract derived from an unidentified fungus was further subjected to HPLC and UV spectrometric analysis. The active component was recovered as a single peak and identified as PCA. The anti-apoptotic activity of PCA was confirmed by the standard compound obtained from the commercial supplier.
Clonogenic Assay-Raji cells (10 6 cells/ml) were pretreated with various concentrations of PCA for the indicated times and washed with the RPMI 1640 medium to remove PCA. The cells were then treated with or without cross-linked FasL (500 ng/ml) for 2 h and washed with the medium to remove FasL. The cells were diluted with the medium (10 cells/ml) and cultured in 96-well microtiter plates (1 cell/well, 100 l) for 10 days. The number of colonies formed was counted.
FACS Analysis-Raji cells were treated with or without mouse antihuman Fas antibody DX2 (Calbiochem) for 1 h on ice. The cells were repeatedly washed and then stained with fluorescein isothiocyanateconjugated rabbit anti-mouse IgG (H ϩ L) (Jackson ImmunoResearch, West Grove, PA) for 45 min on ice. After washing, the stained cells were analyzed under flow cytometry (FACSCalibur; BD Biosciences).
DISC Analysis-Raji cells were treated with 2 g/ml FasL in the presence of 2 g/ml anti-FLAG antibody M2 (Sigma) for 15 min and quickly cooled down by adding 5 volumes of ice-cold PBS. Cells were washed twice with PBS and lysed with 0.2% Nonidet P-40, 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 2 mM sodium vanadate, 10% glycerol, and the protease inhibitor mixture. Postnuclear lysates were precleared with Sepharose 6B (Sigma) for 90 min and then incubated with protein A-Sepharose CL-4B (Amersham Biosciences) for 3 h. Sepharose beads were washed four times with the lysis buffer. Proteins were separated by SDS-PAGE and analyzed by Western blotting using anti-FADD and anti-caspase-8 antibodies.
Mass Spectrometry-Caspase-8 p20 (30 g) was treated with 100 M PCA in Tris-HCl (pH 8.5), 100 mM NaCl, 5% sucrose for 3 h. Caspase-8 p20 was precipitated by acetone and then digested with 2 g of lysyl endoprotease (Wako Pure Chemical Industries, Ltd., Osaka, Japan) in 25 mM Tris-HCl (pH 9.0), 1 mM DTT in the presence of 2 M urea at 37°C for 14 h. The resulting peptide mixture was subjected to liquid chromatography and tandem mass spectrometry (MS/MS). The digested peptides were separated by a reversed-phase column (Symmetry C18 5 m, 0.32 ϫ 150 mm, Waters Associates, Milford, MA) using a capillary HPLC system (CapLC system, Waters Associates). The column was eluted with a linear gradient of 5-50% acetonitrile in 0.1% formic acid at a flow rate of 5 l/min. The isolated peptides were identified by mass measurements with quadrupole time-of-flight mass spectrometer (Micromass Ltd., Altrincham, UK) equipped with an electrospray interface that was connected directly to the above HPLC system.

PCA Blocks Apoptosis Induced by Agonistic Anti-Fas Antibody or Cross-linked FasL-Fas-mediated apoptosis is induced by agonistic
Fas antibodies or physiological FasL. Murine B lymphoma A20 cells are highly sensitive to anti-Fas antibody Jo2 (19). To search for specific inhibitors that block Fas-mediated apoptosis, we screened microbial secondary metabolites and herbal medicines, and we identified PCA as an active component in the culture broth of an unidentified fungus. The rationale of the screening system was described under "Experimental Procedures." A20 cells underwent DNA fragmentation after treatment with Jo2 in a dose-dependent manner for 16 h (Fig. 1B). Four hours of incubation with 200 ng/ml Jo2 was sufficient to induce nearly maximal DNA fragmentation (data not shown). PCA inhibited DNA fragmentation induced by Jo2 in a dose-dependent manner, whereas PCA exhibited apoptotic effects marginally (Fig. 1C). PCA markedly reversed a reduction in cell viability of A20 cells treated with Jo2 (Fig. 1D). In addition to A20 cells, human Burkitt's lymphoma Raji cells were killed by cross-linked FasL in a dose-dependent manner, and most of the cells underwent apoptosis by 500 ng/ml FasL within 4 h (Fig. 1E). PCA also inhibited FasL-induced apoptosis in Raji cells at 100 -200 M (Fig. 1F). Although PCA alone slightly affected cell viability, PCA at 100 -200 M effectively blocked FasL-induced loss of cell viability (Fig. 1G). To determine the limits at which PCA does not affect long term viability, clonogenic assays were performed with a range of doses and incubation times with PCA ( Fig. 1, H and I). One hour of treatment with 200 M PCA reduced the long term viability by about 20% (Fig. 1H). However, cloning efficiency was markedly decreased when Raji cells were treated with 100 -200 M PCA for 2 h (Fig. 1I). Therefore, the 1-h treatment with 200 M PCA was used for assessment of its inhibitory effects on FasL-induced apoptosis by using clonogenic assay. Under this condition, PCA inhibited FasL-induced apoptosis in Raji cells and increased their survival rate slightly but significantly (Fig. 1J).
PCA Inhibits Caspase-3 Activity, but Not Active Caspase-3, in FasL-treated Cells-Upon Fas oligomerization, caspase-8 is activated by self-cleavage, and active caspase-8 subsequently processes caspase-3 to an active form (1)(2)(3)(4). Consistent with this model, caspase-3 activity was undetectable during the first 30 min after FasL addition and increased rapidly thereafter in FasL-treated Raji cells (Fig. 2A). PCA inhibited caspase-3 ac-  (Fig. 2B). To examine whether PCA inhibits active caspase-3 directly in FasL-treated cells, PCA was added to Raji cells 0 or 60 min (sufficient time to induce active caspase-3) after treatment with FasL. In contrast to the caspase inhibitor Z-VAD-fmk, PCA did not prevent caspase-3 activity when PCA was added 60 min after FasL treatment (Fig. 2C). These results suggest that PCA does not inhibit active caspase-3 in FasL-treated cells.
PCA Blocks Self-processing of Caspase-8 in the DISC-Upon FasL engagement, Fas binds to FADD via the death domain interaction, and in turn recruits caspase-8 via the homotypic interaction of their death effector domain, allowing the formation of the DISC (1-4, 11, 12). Immediately after recruitment, caspase-8 molecules are placed in close proximity facilitating self-processing (13). PCA affected neither the cell-surface expression of Fas nor the cellular level of FADD and caspase-8 up to 5 h (Fig. 4, A and B). After treatment with FasL, PCA did not affect the recruitment of FADD and caspase-8 into the DISC (Fig. 4C). However, in cells treated with 100 -200 M PCA, only full-length caspase-8, but not processed caspase-8 (p43 and p41), was detected in the DISC (Fig. 4C). Caspase-8 processing in the DISC was significantly increased for 30 -60 min after treatment with FasL, and PCA still effectively blocked caspase-8 processing in the DISC (Fig. 4D). Similar to PCA, Z-VAD-fmk inhibited the processing of caspase-8 in the DISC without affecting the recruitment of FADD and caspase-8 (Fig. 4E).
Caspase-8 Is the Preferential Target of PCA in Living Cells-The above observations suggest the possibility that caspase-8 is a direct target of PCA. Consistent with this hypothesis, PCA inhibited the enzyme activity of active recombinant caspase-8 in a dose-dependent manner (Fig. 5A). However, the enzymatic activities of active recombinant caspase-3 and active recombinant caspase-9 were also inhibited by PCA at similar concentrations (Fig. 5, B and C). To address whether PCA inhibits caspase-8 preferentially at the enzyme level, competition assay between caspase-3 and caspase-8 was performed (Fig. 5D). Even in the presence of excess caspase-8 large subunit (p20) where the consensus active center QACXG is located (15,26), the inhibitory effect of PCA on caspase-3 activity was not suppressed (Fig. 5D). Thus, these observations suggest that PCA does not selectively inhibit active caspase-8 at the enzyme level. To determine whether PCA inhibits caspase-9 at the cellular level, Raji cells were pretreated with PCA for 1 h and then incubated for 6 h with staurosporine which induces apoptosis mediated by caspase-9 (Fig. 5E). Caspase-9 is processed into p35 subunit by self-cleavage, whereas active caspase-3 cleaves caspase-9 into p37 subunit (27). Caspase-9 processing was slightly inhibited by PCA in staurosporine-treated cells, although PCA profoundly inhibited caspase-8 processing in FasL-treated cells (Fig. 5E). Because, unlike caspase-3 and caspase-8, the cellular level of caspase-9 was markedly reduced when Raji cells were treated with 200 M PCA for over 7 h (Fig.  5F), a marginal reduction of caspase-9 processing might be due to the PCA-induced reduction of pro-caspase-9 rather than the inhibitory effect of PCA on caspse-9 activation. To avoid the extensive reduction of caspase-9 by PCA, Raji cells were first preincubated with staurosporine for 3 h and then treated with PCA or Z-VAD-fmk when caspase-9 activity was increasing (Fig. 5G). In contrast to Z-VAD-fmk, PCA exerted weak inhibitory effects on activation of caspase-9 and caspase-3 in staurosporine-treated cells. Moreover, a similar experiment was performed using FasL-treated Raji cells (Fig. 5H). Z-VAD-fmk prevented activities of caspase-8 and caspase-3 to the background level. By contrast, PCA strongly inhibited newly activated caspase-8 and caspase-3 but partially prevented already glutathione but not serine (Fig. 6A). Glutathione also suppressed the inhibitory effect of PCA on caspase-8 processing in the DISC (Fig. 6B). To address whether PCA binds to the active site cysteine, the recombinant caspase-8 p20 was mixed with PCA, and direct binding of PCA was assessed by liquid chromatography-MS/MS (Fig. 6, C and D). Compared with an unmodified peptide, the profiles of fragment ions of the peptide spanning the active center clearly revealed that PCA binds to the active center cysteine of caspase-8 catalytic domain. DISCUSSION Caspases play a central role in apoptotic cell death (14,15). Active caspases are modulated directly by cellular and viral proteins that interact with these proteases (14, 16, 17). Cell-permeable caspase inhibitors are important research tools for addressing the molecular basis of apoptosis, as well as being therapeutic agents for various diseases (28 -33). In contrast to cell-free screening for inhibitors of recombinant proteins such as caspases, we adopted cell-based models of apoptosis to overcome disadvantages such as the poor membrane permeability of inhibitors. We screened for specific inhibitors that block Fas-mediated apoptosis and identified PCA. In the present paper, we demonstrate that PCA primarily blocks self-processing of caspase-8 in the DISC by binding the active center cysteine.
PCA inhibited the enzymatic activities of active recombinant caspase-3, caspase-8, and caspase-9 at similar concentrations. The non-selective inhibitory effects of PCA on active recombi- nant caspases at the enzyme level might be explained by the fact that PCA binds to cysteine residues of proteins via an ␣,␤-unsaturated lactone moiety. Nevertheless, PCA strongly inhibited self-activation of caspase-8 but weakly inhibited activation of caspase-3 and caspase-9, in cells treated with FasL or staurosporine. These observations suggest that caspase-8 should be the preferential target of PCA in intact cells. The overall structure of caspase-8 including the substrate-binding subsite (S 1 ) resembles the caspase-3 structure, with the exception of subsites (S 3 and S 4 ) essential for substrate specificity (34 -36). The similarity and difference in the binding cleft of caspase-3 and caspase-8 does not explain the preferential blockade of caspase-8 by PCA in intact cells. Although the three-dimensional structure of the caspase-8 zymogen is currently unknown, it seems likely that the substrate-binding pocket differs between the fully active caspase-8 heterodimer and the caspse-8 zymogen. As the likely explanation, we speculate that PCA might have a higher affinity to the caspase-8 zymogen and inactivate its intrinsic proteolytic activity.
In contrast to caspase-3 and caspase-8, the cellular amount of caspase-9 was selectively reduced in PCA-treated cells. The caspase-9 reduction by PCA was prevented by the proteasome inhibitor MG-132, 2 suggesting that caspase-9 is proteolytically degraded by the ubiquitin-proteasome system in PCA-treated cells. The molecular mechanism of caspase-9 degradation induced by PCA remains to be elucidated.
Among peptide-based caspase inhibitors so far developed, Z-VAD-fmk is an irreversible inhibitor that targets many (but not all) caspases (33,37) and is most frequently used to block apoptosis. It was reported that a series of non-peptide inhibitors of caspase-3 and caspase-7 (isatin sulfonamides) prevent apoptosis and maintain cell functionality (38). Unlike PCA, the isatin sulfonamides do not possess an ␣,␤-unsaturated lactone moiety that is capable of binding cysteine residues. Instead, these inhibitors primarily interact with the substrate binding cleft (S 2 ) unique to caspase-3 and caspase-7 (38). Most of the isatin sulfonamide derivatives do not inhibit caspase-8 (38). Thus, PCA seems to be the non-peptide inhibitor that blocks self-processing of caspase-8 in cell-based models of apoptosis.
Caspase-8 is the initiator caspase that functions as a key molecule critical for apoptosis induced by death receptors (1-4, 39, 40). Signaling by death receptors plays an essential role in homeostasis of various types of cells in the immune system and autoimmunity (41)(42)(43). Recently, our group and others (6,7) reported that caspase activation is required for T cell proliferation. In agreement with this notion, we have shown that the cellular caspase-8 inhibitor c-FLIP is essential for the regulation of T cell proliferation (44,45). Caspase inhibitors might thus be applicable to block T cell proliferation and prevent autoimmune diseases. Our present results demonstrate that PCA blocks self-processing of caspase-8 in the DISC. However, PCA exhibited a limited selectivity and induced unfavorable cytotoxic effects in the long term culture, possibly due to binding to sulfhydryl groups of other proteins. Thus, further design of synthetic analogues of PCA might lead to highly selective inhibitors for caspase-8.