STAT1-induced Apoptosis Is Mediated by Caspases 2, 3, and 7*

STAT1 (signal transducer and activator of transcription 1) has been implicated as a mediator of a variety of biological responses in response to stimulation by specific growth factors and cytokines. To understand better the role of STAT1 in the interferon- (cid:1) (IFN- (cid:1) )-induced phenotype, we generated an active form of STAT1 (STAT1C) by substituting Cys residues for both Arg-656 and Asn-658 within the C-terminal loop of the STAT1 SH2 domain. The IFN- (cid:1) activation site element was stimulated and bound efficiently by STAT1C without IFN- (cid:1) treatment. STAT1C was found to be tyrosine-phospho-rylated in the nucleus for more than 30 h after IFN- (cid:1) stimulation. STAT1-negative U3A cells reexpressing STAT1C showed retarded cell growth and underwent apoptosis when treated with IFN- (cid:1) . Further analysis demonstrated that apoptosis was preceded by proteo-lytic cleavage of caspases 2, 3, and 7, and wild type STAT1 also induced cleavage of caspase 7 when expressed in STAT1-negative U3A cells, indicating that STAT1C augments potential activity of wild type STAT1. Studies with cycloheximide treatment showed that protein synthesis induced in the first 24 h after IFN- (cid:1) treatment was required for apoptosis under these conditions. Finally, we found that STAT1C-induced apoptosis was, in part, the that IFN- -induced apoptosis. Our results using CHX to inhibit IFN- (cid:1) -induced protein synthesis revealed that de novo protein synthesis in the first 12 h after IFN- (cid:1) stimulation is necessary for apoptosis. These results suggest at least two possible mech-anisms: 1) the first 12 h, or 2) STAT1C target Our current studies did not identify

STAT1 (signal transducer and activator of transcription 1) has been implicated as a mediator of a variety of biological responses in response to stimulation by specific growth factors and cytokines. To understand better the role of STAT1 in the interferon-␥ (IFN-␥)-induced phenotype, we generated an active form of STAT1 (STAT1C) by substituting Cys residues for both Arg-656 and Asn-658 within the C-terminal loop of the STAT1 SH2 domain. The IFN-␥ activation site element was stimulated and bound efficiently by STAT1C without IFN-␥ treatment. STAT1C was found to be tyrosine-phosphorylated in the nucleus for more than 30 h after IFN-␥ stimulation. STAT1-negative U3A cells reexpressing STAT1C showed retarded cell growth and underwent apoptosis when treated with IFN-␥. Further analysis demonstrated that apoptosis was preceded by proteolytic cleavage of caspases 2, 3, and 7, and wild type STAT1 also induced cleavage of caspase 7 when expressed in STAT1-negative U3A cells, indicating that STAT1C augments potential activity of wild type STAT1. Studies with cycloheximide treatment showed that protein synthesis induced in the first 24 h after IFN-␥ treatment was required for apoptosis under these conditions. Finally, we found that STAT1C-induced apoptosis was, in part, mediated by caspase 2, 3, and 7 because benzyloxycarbonyl-valyl-aspartyl-valyl-alanyl-aspartic acid fluoromethyl ketone (Z-VDVAD-FMK) treatment partially blocked apoptosis. These results suggest that prolonged nuclear localization of activated STAT1 results in apoptosis involving specific regulation of caspase pathway.
In response to IFN-␥ stimulation, STAT1 becomes phosphorylated on Tyr-701 and forms a homodimer through reciprocal SH2-phosphotyrosine interactions. Phosphorylation on Tyr-701 is required for subsequent nuclear translocation, DNA binding, and gene activation (15)(16)(17). Tyrosine phosphorylation of STAT1 is regulated both by Jak1/Jak2 protein-tyrosine kinases and by protein-tyrosine phosphatase. Significantly, treatment with sodium vanadate, a protein-tyrosine phosphatase inhibitor, can sustain the phosphorylation of Tyr-701, preventing down-regulation of STAT1 transcriptional activity (18,19). Biochemical analysis has demonstrated that dephosphorylated STAT1 is rapidly exported back into the cytoplasm and takes part in subsequent activation-inactivation cycles and that loss of dephosphorylation leads to prolonged STAT1 activation in the nucleus (20 -27).
STAT1 has been implicated in apoptosis in response to IFN-␥ stimulation. Mutant forms of STAT1 lacking the N-terminal 62 amino acid residues or carrying a single amino acid substitution of Ala-31 for Arg-31 inhibited STAT1 tyrosine dephosphorylation and greatly enhanced the antiproliferative activity of IFN-␥ when expressed in fibroblasts (28). It has also been shown that constitutive expression of caspases 1, 2, and 3 requires STAT1 without IFN-␥ stimulation and that STAT1-negative U3A cells reexpressing wild type and Y701F mutant forms of STAT1 restored apoptosis induced by tumor necrosis factor-␣ plus actinomycin D treatment (9,29). Although these results demonstrate that STAT1 is involved in negative regulation of cell growth, the mechanism of apoptosis induced by the IFN-␥/STAT1 pathway has not fully been described.
To dissect the role of STAT1 in the IFN-␥-induced phenotype, we generated an active form of STAT1 (STAT1C) by introducing Cys residues into Arg-656 and Asn-658 within the C-terminal loop of the STAT1 SH2 domain. Cysteine residues have been used successfully to induce dimerization in several signaling molecules, including Eyk, erythropoietin, c-Mpl, and ion channel proteins (30 -33). In particular, an engineered STAT3 protein carrying equivalent Cys mutations has been demonstrated to activate transcription and induce cell transformation without cytokine stimulation (34), suggesting that STAT1C may enhance a potential activity of STAT1.
Here we report that STAT1C alone can stimulate a promoter construct containing the GAS element. IFN-␥ treatment induced prolonged nuclear localization of STAT1C for more than 30 h, leading to apoptosis. Inhibition of caspases 2, 3, and 7 partially blocked apoptosis, suggesting that these caspases are, at least in part, involved in this pathway. More importantly, cleavage of caspase 3 and 7 was observed in cells expressing endogenous wild type STAT1 when treated with IFN-␥, demonstrating that STAT1C augments potential activity of STAT1. These results suggest a novel mechanism of apoptosis mediated by STAT1 involving specific cleavage of caspase.

EXPERIMENTAL PROCEDURES
Cell Lines, Transfection, and Luciferase Measurements-2fTGH (35), U3A (36), and HeLa cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Invitrogen). G8 cells (37) were grown in the same medium to which 100 g/ml G418 (Invitrogen) had been added. IFN-␥ was purchased from PeproTech, Inc. (Rocky Hill, NJ). Transfections were made with FuGENE 6 (Roche Applied Science). Luciferase assay basic procedures have been described (38). For each cell culture, 1 g of M67luc with STAT1wt, STAT1C, or empty vector was transfected. Cells were treated with IFN-␥ for 6 h before luciferase assay. ␤-Galactosidase activity was measured as an internal control by cotransfection of pcDNA3.1/His/LacZ (Invitrogen). Values shown represent a normalized average of three or four independent transfections.
Western Blot Analysis-Cells were lysed in EBC buffer (50 mM Tris-HCl, pH 8.0, 120 mM NaCl, 0.5% Nonidet P-40, 100 mM NaF, 200 M sodium orthovanadate, 10 g/ml phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin). Protein concentration in cell lysates was determined using the Bio-Rad protein assay kit. An equal amount of protein (20 or 100 g for caspase assay) for each sample was separated by 7.5-10 or 15% SDS-PAGE and transferred to Immobilon-P membrane (Millipore). The membranes were blocked with 5% skim milk or 3% bovine serum albumin in phosphate-buffered saline and sequentially incubated with primary antibodies and horseradish peroxidase-conjugated secondary antibodies (Jackson Laboratories) followed by ECL detection (Amersham Biosciences).
Electrophoretic Mobility Shift Assay-Total lysates or nuclear extracts were prepared from various type of cells both untreated and treated with IFN-␥ for various time periods, as described previously (39,40). Synthetic double strand oligonucleotide M67, 5Ј-CATTTCCCGTA-AATCAT-3Ј, was end labeled with [␥-32 P]ATP using T4 polynucleotide kinase. Extracts were incubated with the labeled probe in the presence of poly(dI-dC) in a binding buffer containing 20 mM HEPES at room temperature for 30 min. For the competition assay, 2-10 ng of unlabeled probe or 1-5 g of the indicated antibodies was added to the incubation mix. After resolving the DNA-protein complex in a 5% nondenaturing polyacrylamide gel, gels were dried and visualized by autoradiography.
Plasmid Constructs-STAT1wt FLAG-tagged in pcDNA3.1 was a gift from Dr. C. Horvath. The STAT1-C construct was made using sitedirected mutagenesis (QuikChange, Stratagene) with primer pairs 5Ј-AATTACAAAGTCATGGCTTGTGAGTGTATTCCTGAGAATCCC-3Ј. The construct was then transferred to a pRc/CMV vector (Invitrogen) and the cytomegalovirus promoter was replaced by EF1␣ promoter from pEF vector (Invitrogen).
Apoptosis Assay-Annexin V and propidium iodine staining was performed as described by the manufacturer (Roche Applied Science). After flow cytometry, cells were divided into four distinct populations using the control cells as a reference; the lower left quadrant of the dot plot represented viable cells, and the lower and upper right quadrants represented cells early and late in apoptosis, respectively. The cells in each quadrant were then gated, and the percentage of the total population was determined using CellQuest TM (BD Biosciences) software.

RESULTS
Generate Active Form of STAT1-It has been shown that the introduction of two Cys residues into Ala-662 and Asn-664 in the STAT3 SH2 domain produces a constitutively active form of STAT3 (34). We generated a mutant STAT1 (STAT1C) by introducing Cys residues into Ala-656 and Asn-658 in the SH2 domain, which are equivalent to Ala-662 and Asn-664 of STAT3 ( Fig. 1 A). Transcription activity of STAT1C was examined in the presence or absence of IFN-␥ using an M67luc reporter construct containing four copies of STAT1 binding sites upstream of the luciferase gene (41)(42)(43)(44). STAT1-positive 2fTGH cells and STAT1-negative U3A cells (36,37,45) were transfected with an empty, STAT1-, or STAT1C-expressing vector together with M67luc for 40 h and stimulated further by 20 ng/ml IFN-␥ for 6 h before measuring luciferase activity. As shown in Fig. 1B, STAT1 activated the reporter 46-fold after IFN-␥ treatment in U3A cells; an empty vector alone did not increase the activity because of a loss of endogenous STAT1 expression. Significantly, STAT1C alone activated the reporter plasmid 80-fold, and IFN-␥ treatment further enhanced the activity up to 147-fold. In STAT1-positive 2fTGH cells, luciferase activity was increased 2.5-fold when vector alone was transfected, and transfection of STAT1 enhanced the activity 27-fold when treated with IFN-␥. As shown in U3A cells, STAT1C alone activated the reporter plasmid 48-fold, an outcome enhanced by IFN-␥ treatment up to 88-fold. These results suggest that STAT1C alone can potentially activate a promoter containing STAT1 binding sites and that this activity is strongly enhanced by IFN-␥ stimulation.
STAT1C Shows Increased Tyrosine Phosphorylation-U3A cells were stably transfected with STAT1C, which has the FLAG tag at the C terminus, and G418-resistant clones were isolated for further analysis. Among the clones obtained, 1CC and 1C5 clones were found to express similar levels of STAT1 protein with G8 cells generated by stably expressing STAT1␣ in U3A cells (37). It has been shown that tyrosine phosphorylation of STAT1 Tyr-701 is essential for dimerization, nuclear translocation and DNA binding ability. To understand how Tyr-701 phosphorylation is involved in the strong transactivation activity of STAT1C, we examined the phosphorylation status of Tyr-701 and the induction of IRF1, a well character- ized IFN-␥ target gene, with or without treatment with the kinase inhibitor genistein, which is frequently used for the inhibition of tyrosine kinases. 100 M genistein was added to the cell culture 30 min after IFN-␥ stimulation, and nuclear extracts were prepared for immunoblot analysis at 30 min, 1 h, 6 h, 12 h, 24 h, and 30 h after IFN-␥ stimulation. The expression level of nuclear transcription factor Sp1 was unchanged, and it is shown as a loading control in Fig. 2. Although significant change in the nuclear STAT1 protein level was not observed in G8 cells, Tyr-701 phosphorylation was transiently induced (30 min) in these cells. Nuclear localization and Tyr-701 phosphorylation of STAT1 were undetectable in IFN-␥stimulated G8 cells when treated with genistein, serving as a positive control of Jak kinase inhibition by genistein. These results are consistent with the previous finding that Tyr-701 phosphorylation is required for nuclear localization of STAT1 (15)(16)(17). Interestingly, IRF1 was also induced in the presence of genistein in G8 cells, although induction was slightly weaker than untreated G8 cells. These results suggest that low levels of nuclear STAT1 are sufficient for induction of IRF1. In 1CC and 1C5 cells, STAT1C was more nuclear than in G8 even in the presence of genistein. Although Tyr-701 phosphorylation was slightly decreased by genistein in both 1CC and 1C5 cells, it was significantly sustained for 30 h after IFN-␥ stimulation even with genistein treatment, implying that prolonged STAT1C phosphorylation is a consequence not of kinase hyperactivation but of a decreased dephosphorylation rate. Consistent with the enhanced phosphorylation in Tyr-701, a remarkable induction of IRF1 was sustained for 30 h in 1CC and 1C5 cells after IFN-␥ stimulation. Interestingly, high expression of IRF1 was also detected in genistein-treated cells. These results suggest that IFN-␥-induced Tyr-701 phosphorylation is protected from nuclear phosphatase activity by the reinforced interaction of the SH2 domain through the Cys-Cys bond.
Increased DNA Binding Ability of STAT1C-Next, DNA binding activity of STAT1C protein was studied using an electrophoretic mobility shift assay (Fig. 3A). Normally growing 2fTGH, G8, and 1CC cells were treated with IFN-␥ for 15, 30, and 120 min, and cell extracts were subjected to electrophoretic mobility shift assay using a 32 P-labeled STAT1-binding DNA element as a probe (41). Wild type STAT1 showed increased DNA binding after 15 min of treatment which gradually decreased in 120 min. The basal DNA binding ability of STAT1C was higher than that of wild type (lanes 1, 5, and 9) and was remarkably increased and sustained for more than 120 min after IFN-␥ stimulation (lanes 10 -12).
Specific binding of STAT1 and STAT1C to the probe was confirmed by a competition assay using an unlabeled DNA probe (Fig. 3B). Cell extracts were prepared from unstimulated or stimulated (15 min) cells, and an excess amount of unlabeled DNA probe was added to the extracts. Consistent with the results in Fig. 3A, STAT1C was found to bind DNA more efficiently than wild type STAT1 without IFN-␥ treatment (lanes 1, 7, and 13). Strong DNA binding ability of STAT1C was induced in 15 min, and it disappeared after administration of an excess DNA probe, confirming that STAT1C specifically recognizes the GAS element.
Specific binding of STAT1C to GAS element was confirmed further with antibodies recognizing STAT1C in gel shift assay (Fig. 3C). In 2fTGH cells, STAT1 binding to GAS element (lane 2) was supershifted with anti-STAT1 antibody (lanes 3 and 5) but not affected with anti-FLAG antibody (lanes 4 and 5). Enhanced DNA binding of STAT1C (lane 11) was supershifted with antibodies for FLAG, STAT1, or both (lanes 12-14) because STAT1C . Cell extracts were incubated with a ␥-32 P-labeled high affinity DNA binding site for STAT1 and resolved on a nondenaturing polyacrylamide gel. The gel was dried and exposed for autoradiography. The arrow indicates the position of the DNA⅐STAT1 complex. *N.S., nonspecific band. B, the same cell extracts as used in A (IFN-␥ (ϩ), 15 min) were preincubated with 2 or 10 ng of nonradiolabeled probe before incubation with 1 ng of radiolabeled M67. Samples were resolved on a nondenaturing polyacrylamide gel and analyzed by autoradiography. C, nuclear extract of 2fTGH (2f) and 1CC cells were used for gel mobility shift assay as described in B. DNAprotein complexes were coincubated with anti-FLAG and/or anti-STAT1 antibody. Unrelated DNA segment (2 ng (ϩ2) or 10 ng (ϩ10), 5Ј-GGCGGCCGCAAGCTTGCGGCCGCC-3Ј) was used for the competition assay. Dots show the positions of supershift (lanes 3 and 5).

Enhanced STAT1 Activation and Apoptosis
is FLAG-tagged at the C terminus. Consistent with results in Fig. 3B, DNA-binding signals were completed with excess M67 probe, not with unrelated cold probe (lanes 6 -9 and 15-18). These results demonstrate that STAT1C contains a DNA binding ability similar to that of wild type STAT1, but it is further enhanced by IFN-␥ treatment.
STAT1C Shows Antiproliferative Activity-We studied the biological activity of STAT1C on cell growth by comparing its effects with those of wild type STAT1. Exponentially growing 2fTGH, U3A, G8, 1CC, and 1C5 cells were plated at 2 ϫ 10 5 cells/6-cm plate. After 24 h, IFN-␥ (20 ng/ml) was added to the medium, and the cells were counted at different times during a 5-day interval (Fig. 4A). No significant differences in growth rate were observed between untreated (solid line) or IFN-␥-treated 2fTGH, U3A, or G8 cells (dotted line). In contrast, 1CC and 1C5 cells showed striking growth retardation 24 -30 h after stimulation compared with untreated 1CC and 1C5 cells. This growth retardation was also observed when a lower concentration of IFN-␥ was used (5 ng/ml, data not shown).
Cell morphology was analyzed by phase-contrast microscopy of 2fTGH, U3A, G8, 1CC, and 1C5 cells treated with IFN-␥ for 48 h (Fig. 4B). Each cell line showed indistinguishable morphology before treatment (a-e), but 1CC and 1C5 cells detached from the plate after 48 h of IFN-␥ treatment and showed condensed nuclei, characteristic of cultured cells undergoing apoptosis (i and j). Although previous studies demonstrated that 2fTGH and STAT1-reconstituted U3A cells show the apoptosis phenotype in 72 h after IFN-␥ treatment (9), 2fTGH, U3A, and G8 cells did not show apoptosis after the same treatment within 48 h in our experiments. These results suggest that enhanced STAT1 activity results in cell death after IFN-␥ treatment.
STAT1C Sensitizes Cells to Apoptosis by IFN-␥-We further quantified apoptosis of STAT1C cells induced by IFN-␥ treatment. 2fTGH, U3A, G8, 1CC, and 1C5 cells were plated at 2 ϫ 10 6 cells/10-cm plate, respectively. After 24 h, cells were stimulated with 20 ng/ml IFN-␥ for 48 h. Collected cells were then stained with propidium iodide and fluorescein isothiocyanatelabeled annexin V, a marker of an early stage of apoptosis, and were analyzed by means of fluorescence-activated cell sorter (Fig. 5A). Apoptosis of 2fTGH, U3A, or G8 cells did not an increase by 48 h of IFN-␥ treatment. STAT1C-expressing 1CC and 1C5 cells showed slightly high levels of apoptosis without IFN-␥ treatment (11 and 13%, respectively), but increased cell death was observed at 24 -48 h after IFN-␥ treatment (33 and 43% at 48 h, respectively). As a positive control of apoptosis, hydroxyperoxide treatment of 1CC and 1C5 cells resulted in 66 and 69% frequency of apoptosis, respectively, in 24 h. These results confirm that STAT1C can sensitize cells to IFN-␥-induced apoptosis.

Enhanced STAT1 Activation and Apoptosis 4070
We next tested whether de novo protein synthesis is required for apoptosis by using the protein synthesis inhibitor cycloheximide (CHX). CHX was added to the 1CC cell cultures at 6, 12, and 24 h after IFN-␥ stimulation (Fig. 5B, top). At 48 h after IFN-␥ stimulation, cells were collected and divided into two aliquots, one for fluorescence-activated cell sorter analysis and the other for immunoblot of STAT1 and ␣-tubulin. When CHX was added to the cell culture after 0 -6 h of IFN-␥ treatment, 1CC cells did not show apoptosis (Fig. 5B, panels 1 and 2). However, when cells were exposed to IFN-␥ for 24 h, CHX did not prevent the cells from undergoing apoptosis (Fig. 5B, panels 3 and 4). Treatment with CHX alone resulted in the intermediate rate of apoptosis only at 48 h (35%, Fig. 5B, panel 6).
To avoid the possibility that this inhibition of apoptosis was the result of a decreased level of STAT1C as a result of CHX treatment, expression of STAT1C was confirmed by immunoblot with anti-STAT1 antibody (Fig. 5C). These results support the hypotheses that 1) STAT1C can transmit apoptosis signals; 2) apoptosis is not an immediate response to IFN-␥ treatment, but it is dependent on de novo protein synthesis during the first 6 -12 h after stimulation. Such protein synthesis is conceivably regulated by STAT1C.
Cleavage of Caspase 2, 3, and 7 in IFN-␥-treated STAT1C Cells-Genetic studies in Caenorhabditis elegans defined an evolutionarily conserved apoptosis pathway including activation of cysteine proteases called caspases. Once activated, these proteases cleave specific intracellular substrates, leading to cell death. We sought to determine whether specific activation of caspases is involved in IFN-␥-induced apoptosis of 1CC and 1C5 cells. 2fTGH, U3A, G8, 1CC, and 1C5 cells were treated with IFN-␥ for 24 or 48 h. Total cell lysates were studied by immunoblot analysis using anti-caspase 1, 2, 3, 6, and 7 antibodies that recognize the full-length and/or cleaved forms of caspases (Fig.  6A). A strong increase in caspase 1 protein was observed in IFN-␥-treated 1CC and 1C5 cells, although previous studies had showed that the mRNA level of caspase 1 is increased in a STAT1-dependent manner in response to IFN-␥ stimulation (9). Cleaved forms of caspases 2, 3, and 7 were specifically detected in 1CC and 1C5 cells from 24 to 48 h of IFN-␥ treatment, whereas no cleaved forms of caspases 1 and 6 were detected in 2fTGH, U3A, G8, 1CC, or 1C5 cells.
To rule out the possibility that this caspase activation and apoptosis are induced only in STAT1C cells, we studied also HeLa cells expressing wild type STAT1, because it has been reported that IFN-␥ treatment can induce caspase 1 expression and apoptosis particularly maintained in 1% serum medium (9). As shown in Fig. 6B, IFN-␥ treatment induced full-length caspase 1 in 1% serum and caspase 3 in both 10 and 1% serum medium when HeLa cells were treated with IFN-␥ for 72 h. Significantly, cleaved forms of both caspase 3 and 7 were also immunodetected in IFN-␥-treated HeLa cells. IRF1 induction showed that the IFN-␥ pathway is intact in HeLa cells. We further examined whether U3A cells that are transiently transfected with an empty vector, STAT1wt (WT) or STAT1C (1C) expression vectors can activate caspase pathway when treated with IFN-␥ (Fig. 6C). Induction of caspase 1 by WT and 1C and cleavage of caspase 3 by WT were not detected in this transfection assay perhaps because of low transfection efficiency, because stable transfectants of 1CC and 1C5 cells showed strong induction and cleavage of both caspases (Fig. 6A). Nevertheless, cleavage of caspase 7 was detected in transient transfection of WT, and cleavage of caspase 3 and 7 was detected in transfection of 1C. These results indicate that activation of caspase pathway by IFN-␥ is not specific in STAT1C cells but is generally observed in IFN-␥-treated cells.
Partial Inhibition of Apoptosis by Z-VDVAD-FMK-On the basis of the observation described above, we attempted to determine whether STAT1C-mediated apoptosis specifically involves the caspase pathway. As shown in Fig. 7, 1CC cells were grown in normal medium with IFN-␥ alone and/or Z-VDVAD-FMK, which specifically inhibits caspase 2 and, to a lesser degree, caspase 3 and 7 (46,47). Cell morphology and annexin V staining were studied after 48 h. Apoptosis of 1CC cells by IFN-␥ (75%) was partially inhibited by coincubation with Z-VDVAD-FMK (49%). These results demonstrate that caspase 2 stimulated with IFN-␥, and samples were taken at different times for detailed analysis. In control experiments 1CC and 1C5 cells were treated with hydroxyperoxide (H 2 O 2 ) for 24 h. B, top, experimental procedures to study whether de novo protein synthesis is necessary to induce apoptosis. 1CC cells were treated with cycloheximide (CHX) at 0, 6, 12, and 24 h after IFN-␥ stimulation (panels 2-5). At 48 h after IFN-␥ stimulation, samples were collected and divided into two aliquots: one for annexin V analysis (B) and the other for immunoblot analysis of STAT1 and tubulin (C). As a control, CHX-dependent apoptosis was measured by incubating 1CC cells with CHX alone for 48 h (panel 6). Apoptotic cells are indicated. and, perhaps, caspase 3 and 7 are partially involved in STAT1C-mediated apoptosis. DISCUSSION IFN-␥ inhibits cell growth and regulates apoptosis, two activities involved in suppression of cancer development and modulation of the immune response. The sensitivity of cells of different types to IFN-␥ antiproliferative activity is extremely variable. For instance, IFN-␥ has been shown either to promote or to inhibit the apoptosis in murine pre-B cells or B chronic lymphocytic leukemia cells, respectively (48 -50). Thus, the involvement of IFN-␥ in apoptosis is determined by collaboration with other modulators of apoptosis and/or the enzymes they regulate. STAT1 is a key molecule in signal transmission after IFN-␥ stimulation and has been implicated in the regu-lation of cell proliferation-associated genes. Recent investigations have demonstrated that STAT1 is essential for regulation of the constitutive mRNA level of c-myc and caspases 1, 2, and 3, suggesting that STAT1 is involved in apoptosis (29,51). Significantly, it has been shown that STAT1 has promoter selectivity through the interaction with breast cancer tumor suppressor protein BRCA1, which is frequently mutated in breast cancer patients, suggesting that the IFN-␥-induced phenotype can be determined by cytogenetic background (8). These results reinforce the hypothesis that when cells are stimulated with IFN-␥, STAT1 plays a crucial role in determining their fate.
Artificially engineered STAT3 (STAT3C) protein carrying the Cys substitutions in the SH2 domain has provided a model of the constitutively active form of STATs, which stimulate their own targets in a cytokine-independent manner (34). In the current studies, a similar strategy was followed to produce constitutively active STAT1 (STAT1C) for analysis of the potential activity of STAT1 in the regulation of cell proliferation. Intermolecular STAT1 homodimerization through phosphorylated Tyr-701 and the SH2 domain is required for DNA binding and transcription of the target genes upon IFN-␥ simulation (15)(16)(17). We found that reporter constructs containing the high affinity STAT1 binding sequence were activated by STAT1C without IFN-␥, demonstrating that the engineered STAT1C forms a homodimer and binds DNA, consistent with our results showing that STAT1C is more nuclear than wild type STAT1 without IFN-␥. However, IFN-␥ treatment further enhanced STAT1C transcription activity, indicating that the IFN-␥-induced modification of STAT1C is required to achieve maximal activity. In this context, it is likely that Ser-727 in the transactivation domain is involved in this activation because it has been shown that Ser-727 phosphorylation is crucial for the interaction with transcription coactivators, including BRCA1 and MCM5, which leads to maximal STAT1 activity (8,52). Further biochemical studies are needed to understand the modification of STAT1C protein after IFN-␥ stimulation.
Genistein treatment did not significantly change Tyr-701 phosphorylation induced by IFN-␥ in 1CC and 1C5 cells, suggesting that prolonged Tyr-701 phosphorylation is not the result of the constitutive activation of tyrosine kinase(s). Recent studies indicate that tyrosine phosphatase TC45 and SHP-2 are involved in dephosphorylation of Tyr-701 (24,53). Although it is unclear that the activities of these phosphatases are regulated by STAT1C, it is possible that phosphorylated Tyr-701 is strongly protected from dephosphorylation by phosphatase through the interaction with the SH2 domain, which is intensified by the introduced Cys-Cys bond.
Our present results support the hypothesis that sustained nuclear localization of the active form of STAT1 is crucial for IFN-␥-induced apoptosis. Our results using CHX to inhibit IFN-␥-induced protein synthesis revealed that de novo protein synthesis in the first 12 h after IFN-␥ stimulation is necessary for apoptosis. These results suggest at least two possible mechanisms: 1) STAT1C continuously produces a proapoptotic protein(s) during the first 12 h, or 2) STAT1C target protein subsequently induces the secondary target genes responsible for apoptosis. Our current studies did not identify the mechanisms of how STAT1C directly or indirectly induces apoptosis. Further investigation is required to identify the gene(s) mediating apoptosis of STAT1C cells.
Although our results demonstrate that STAT1C induces cleavage of specific sets of caspases, a detailed mechanism of caspase activation is largely unclear. It has been shown previously that IFN-␥ can potentially suppress Bcl-xL promoter in STAT1-dependent manner (54), and our preliminary results show that Bcl-xL protein is decreased in STAT1C cells in response to IFN-␥ treatment. 2 Given that the cytochrome c level is regulated by balanced expression of Bcl-2 family of proteins (55), it is possible that decreased level of Bcl-xL results in activation of cytochrome c-Apaf/caspase 9 pathway, leading to caspase 3 cleavage.
Our results demonstrate that IFN-␥-and wild type STAT1dependent cleavage of caspase and apoptosis were confirmed in HeLa cells and U3A cells that are transiently transfected with wild type STAT1. These studies indicate that STAT1C enhances potential activities of wild type STAT1, cleavage of caspase, and subsequent apoptosis upon IFN-␥ stimulation.
These studies provide a model in which STAT1 activity is critical in determining the balance between cell survival and apoptosis, and enhanced STAT1 activity tips the balance toward apoptosis through activation of the caspase pathway.