We and others have demonstrated that Fas-mediated apoptosis is a potential therapeutic target for cholangiocarcinoma. Previously, we reported that CaM (calmodulin) antagonists induced apoptosis in cholangiocarcinoma cells through Fas-related mechanisms. Further, we identified a direct interaction between CaM and Fas with recruitment of CaM into the Fas-mediated DISC (death-inducing signalling complex), suggesting a novel role for CaM in Fas signalling. Therefore we characterized the interaction of CaM with proteins recruited into the Fas-mediated DISC, including FADD (Fas-associated death domain)-containing protein, caspase 8 and c-FLIP {cellular FLICE [FADD (Fas-associated death domain)-like interleukin 1β-converting enzyme]-like inhibitory protein}. A Ca2+-dependent direct interaction between CaM and FLIPL, but not FADD or caspase 8, was demonstrated. Furthermore, a 37.3±5.7% increase (n=6, P=0.001) in CaM–FLIP binding was observed at 30 min after Fas stimulation, which returned to the baseline after 60 min and correlated with a Fas-induced increase in intracellular Ca2+ that reached a peak at 30 min and decreased gradually over 60 min in cholangiocarcinoma cells. A CaM antagonist, TFP (trifluoperazine), inhibited the Fas-induced increase in CaM–FLIP binding concurrent with inhibition of ERK (extracellular-signal-regulated kinase) phosphorylation, a downstream signal of FLIP. Direct binding between CaM and FLIPL was demonstrated using recombinant proteins, and a CaM-binding region was identified in amino acids 197–213 of FLIPL. Compared with overexpression of wild-type FLIPL that resulted in decreased spontaneous as well as Fas-induced apoptosis, mutant FLIPL with deletion of the CaM-binding region resulted in increased spontaneous and Fas-induced apoptosis in cholangiocarcinoma cells. Understanding the biology of CaM–FLIP binding may provide new therapeutic targets for cholangiocarcinoma and possibly other cancers.

INTRODUCTION

Apoptosis or programmed cell death is a tightly regulated process by which abnormal cells are removed from the body without generating an inflammatory response. The hallmark of this process is the activation of specific proteases called caspases, to effect DNA fragmentation and ultimately cell death [1]. This process is crucial not only during embryonic development for proper organogenesis, but also throughout life to maintain cellular homoeostasis. Too much or too little apoptosis disturbs this homoeostasis, leading to various pathological conditions such as autoimmunity, neurodegeneration and cancer [2]. Induction of apoptosis has been used as a therapeutic modality in many cancers [36]. Previous studies from our group and others have demonstrated the importance of Fas-mediated apoptosis in tumorigenesis in cholangiocarcinoma, which is a fatal tumour arising from intra-hepatic or extra-hepatic biliary epithelium [711].

Fas (APO-1/CD-95) is a well-characterized member of the TNF (tumour necrosis factor) superfamily of death receptors [12]. In response to the cognate ligand, FasL, it undergoes oligomerization and recruits the adaptor protein FADD [Fas-associated DD (death domain)], which in turn recruits caspase 8 and/or c-FLIP {cellular FLICE [FADD (Fas-associated death domain)-like interleukin 1β-converting enzyme]-like inhibitory protein, to form the DISC (death-inducing signalling complex). Whether caspase 8 or its enzymatically inactive homologue, c-FLIP, is recruited into the DISC determines whether death or survival signals are transmitted through the DISC [1316]. c-FLIP has been described largely as an anti-apoptotic protein that interrupts apoptotic signalling by interfering with recruitment and/or activation of an initiator caspase, like caspase 8, into the DISC [17]. At the mRNA level, 11 distinct splice variants of c-FLIP have been reported [18]. At the protein level, however, two isoforms, namely a long isoform, FLIPL (55 kDa), and a short isoform, FLIPS (26 kDa), have been identified and extensively studied. FLIPL shares structural homology with caspase 8 and contains two DEDs (death effector domains) at its N-terminus and a long C-terminal tail. However, it lacks cysteine residues in the catalytic subunit that are essential for caspase activity [15,16]. Because of the structural similarity between these two proteins, it has been suggested that FLIP may have dual functions in apoptotic signalling [19]. On the one hand, it interferes with the recruitment and/or activation of caspase 8 in the DISC [17,20]. On the other hand, it may also facilitate activation of caspase 8 by forming a heterodimer with it, which helps achieve the initial cleavage step of procaspase 8 [19,21]. The short isoform, FLIPS, is believed to be exclusively anti-apoptotic. In situations where the expression of caspase 8 is stable, the expression levels of FLIPL are believed to be important in determining its role in apoptosis signalling [22,23]. However, the precise mechanism that enables FLIPL to switch these signals and modulate the sensitivity of cells to Fas-induced signalling is not clearly understood.

We have previously demonstrated that CaM (calmodulin) antagonists induce apoptosis through a Fas-related mechanism in a cholangiocarcinoma tumour model [7,10]. CaM is a 17 kDa, dumbbell-shaped, primarily cytoplasmic protein that binds Ca2+ through EF-hand motifs at each of its globular ends [24,25]. Ca2+ binding to CaM exposes a hydrophobic linker region between two globular domains of CaM that binds with a variety of target proteins and mediates their effects. Several studies have demonstrated the importance of CaM in apoptosis and the expression of a CaM gene was found to be increased during glucocorticoid- and Fas-mediated apoptosis [26,27]. CaM antagonists have been shown to induce apoptosis in several cancer cell lines [10,28,29]. However, the effects of CaM antagonists differ among cell types. In CD4+ T-cells from AIDS patients, the CaM antagonists TFP (trifluoperazine) and TMX (tamoxifen) have been reported to protect against apoptosis [30]. TFP, on the other hand, has been reported to induce apoptosis in some cell lines [7,10,29].

Previous studies from our group demonstrated a direct binding between CaM and the Fas receptor that was found to be regulated during Fas-induced apoptosis [31]. The CaM–Fas interaction is unique in that other members of the death receptor family such as TNF-R1 (TNF receptor 1) and the TRAIL (TNF-related apoptosis-inducing ligand) receptors TRAIL-R1 (TRAIL receptor 1)/DR4 and TRAIL-R2/DR5 were not found to interact with CaM. Furthermore, CaM was found to interact with the Fas DD through a classic Ca2+-dependent CaM-binding motif termed the 1-5-10 motif. CaM interacts with target proteins in both a Ca2+-dependent and a Ca2+-independent manner [32]. Specific characteristics, such as a net positive charge, moderate to high helical hydrophobic moment and moderate hydrophilicity, have been shown to predict the CaM-binding region in a given sequence [33]. Based on the structural analyses of various CaM-binding proteins, three classes of CaM-binding motifs have been described. These include two motifs for Ca2+-dependent binding, termed 1-8-14 and 1-5-10, with numbers indicating the conserved hydrophobic residues in a sequence and a consensus sequence for Ca2+-independent binding, called an IQ motif [34].

The binding between CaM and Fas that is regulated during Fas-mediated apoptosis prompted us to further study the interaction of CaM with other proteins in the Fas-induced DISC, since Fas-stimulated DISC formation is critical in Fas-mediated apoptosis. In the present paper we report that CaM specifically binds with FLIPL but not other members of Fas-induced DISC including FADD and caspase 8. CaM binds specifically to the long isoform of c-FLIP; the interaction is Ca2+-dependent and is regulated in response to activation of Fas. Furthermore, the CaM antagonist TFP inhibits CaM–FLIP binding with concurrent inhibition of ERK (extracellular-signal-regulated kinase) phosphorylation. In addition, CaM–FLIP binding is direct and the CaM-binding domain is localized to the FLIPL amino acids 197–213. Overexpression of FLIPL lacking this CaM-binding region increased both spontaneous and Fas-stimulated apoptosis in cholangiocarcinoma cells as compared with WT (wild-type) FLIPL-overexpressing cells, suggesting that CaM–FLIP binding is important in mediating survival signals by c-FLIP.

MATERIALS AND METHODS

Cells, antibodies and reagents

The human cholangiocarcinoma cell line Sk-ChA-1 was provided by Dr A. Knuth (Ludwig Institute for Cancer Research, London, U.K.). Cholangiocarcinoma cells were grown in RPMI 1640 medium with 2 mM L-glutamine, 5 units/ml penicillin, 5 μg/ml streptomycin and 10% (v/v) heat-inactivated fetal bovine serum. For the activation of the Fas pathway, cells were seeded at 70–80% confluence, followed by treatment with a Fas agonist antibody (CH-11) over 60 min. Antibodies include: monoclonal anti-FLIP antibody, NF-6 (Alexis, San Diego, CA, U.S.A.), polyclonal anti-FLIP and anti-FADD antibodies (BD Biosciences, San Jose, CA, U.S.A.), mouse monoclonal anti-caspase 8, rabbit polyclonal phospho-ERK and ERK (p44/42) antibodies (Cell Signaling Technology, Boston, MA, U.S.A.), Fas-activating antibody, clone CH-11 (Upstate Biotechnology, Lake Placid, NY, U.S.A.), Alexa Fluor® secondary antibodies [goat anti-mouse IgG (Alexa Flour® 488) and goat anti-rabbit IgG (Alexa Flour® 594)] (Molecular Probes, Invitrogen, Carlsbad, CA, U.S.A.), anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) antibody (Research Diagnostics, Concord, MA, U.S.A.), rabbit polyclonal anti-CaM antibody (Zymed Laboratories, San Francisco, CA, U.S.A.) and goat polyclonal anti-GST (glutathione transferase)–HRP (horseradish peroxidase) antibody (Amersham Biosciences, Piscataway, NJ, U.S.A.). The monoclonal antibody to CaM was developed as described in [35].

Protein pulldown

Protein pulldown with CaMS (CaM–Sepharose) 4B (Amersham Biosciences) or CS (control Sepharose) CL-4B beads (Sigma–Aldrich, St. Louis, MO, U.S.A.) was performed as described previously [31]. Briefly, cholangiocarcinoma cells were lysed in lysis buffer [20 mM Tris, 150 mM NaCl, 1% Triton X-100, 10% (v/v) glycerol, 1 mM PMSF, EDTA-free protease inhibitor cocktail (Roche), 1 mM sodium fluoride and 1 mM sodium orthovanadate], and 800 μg of extracted proteins was incubated with a 60 μl 1:1 slurry of beads overnight at 4 °C. The beads were washed five times with lysis buffer and proteins were eluted in 2×SDS buffer containing 10 mM EGTA.

Co-IP (co-immunoprecipitation)

A Seize™ primary mammalian immunoprecipitation kit (Pierce Biotechnology, Rockford, IL, U.S.A.) was used for covalent conjugation of an anti-CaM antibody to the activated agarose beads. Briefly, 50 μg of anti-CaM or mouse IgG1 control antibodies was used for conjugation with a 100 μl slurry of activated agarose beads in a column according to the manufacturer's protocol. Cell lysates were incubated overnight at 4 °C in these columns followed by washing and elution of the proteins from the columns as per the protocol. The beads were boiled in 2×SDS sample buffer.

Western blot analysis

Whole cell extracts (50 μg of proteins) or eluted proteins from pulldown or immunoprecipitation assays were separated by SDS/PAGE as described previously [22]. The proteins were transferred on to nitrocellulose membranes (Millipore, Bedford, MA, U.S.A.). For CaM, the membranes were fixed in 0.2% glutaraldehyde in Tris-buffered saline (150 mM NaCl/10 mM Tris, pH 7.6) for 30 min. The membranes were blocked in 3% nonfat dried skimmed milk in Tris-buffered saline with 0.05% Tween 20 for 1 h at room temperature (22 °C). Appropriate primary antibodies were added on to the membranes and incubated overnight at 4 °C followed by three washes of 10 min each. HRP-conjugated secondary antibodies in the same blocking buffer were incubated for 1 h at room temperature and washed three times. Signals were detected using an Immobilon Western chemiluminescent HRP substrate (Millipore, Billerica, MA, U.S.A.) detection kit. The band intensities were analysed by densitometry by using Adobe Photoshop Elements software (Adobe Systems, San Jose, CA, U.S.A.).

Apoptosis assay

Apoptosis was induced with Fas-activating antibody, CH-11, as described previously [22]. Briefly, 1×106 cells were seeded on to a 6-well plate and exposed to 500 ng/ml of CH-11 antibody over 24 h. Apoptosis was determined by staining the cells with annexin V and propidium iodide by using an annexin V–FITC apoptosis detection kit (BD Biosciences, Palo Alto, CA, U.S.A.) and analysed by flow cytometry according to the manufacturer's recommendations. The percentage of cells that were annexin V-positive and propidium iodide-negative were considered to be apoptotic cells.

Immunofluorescence staining

Cholangiocarcinoma cells (1×106) were seeded on to 18 mm×18 mm glass coverslips in a 6-well plate. Subsequently, cells were washed with ice-cold PBS and fixed with 3% (w/v) paraformaldehyde in PBS and permeabilized using 0.5% Triton X-100 in PBS. After blocking with 1% BSA in PBS, cells were incubated with primary antibody [a mouse monoclonal anti-CaM and a rabbit polyclonal anti-FLIP antibody (10 μg/ml) in blocking buffer] at room temperature for 1 h and then secondary antibody [Alexa Flour® 488 anti-mouse IgG or Alexa Flour® 594 anti-rabbit IgG (10 μg/ml) in blocking buffer] for 30 min. Cells incubated with secondary antibodies alone were used as a negative control. Microscopic images were taken at ×100 magnification using an Olympus IX 70 inverted microscope.

Expression and purification of GST fusion proteins in Escherichia coli

Expression and purification of GST–FLIP fusion proteins were performed as described previously [31]. The human c-FLIPL cDNA was amplified from a human umbilical vein endothelial cell library by PCR, cloned into pcDNA3 (Invitrogen) and confirmed by sequencing. To express recombinant FLIP proteins in vitro, FLIPL cDNA was cloned into a GST vector (pGEX-5X-3, Amersham Biosciences) using blunt-end ligation with the restriction enzyme SmaI. The mouse FLIPL construct was a gift from Dr R. Khosravi-Far (Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, U.S.A.). The deletion mutants with human FLIPL were generated using restriction enzyme (EcoRI, ApoI, ClaI, AflII, HindIII and BamHI)-mediated digestion of cDNA followed by cloning into a GST vector. ΔFLIP deletion mutant with deletion of amino acids 197–213 was generated using BamHI- and AflII-mediated digestion of FLIPL cDNA. All of the mutants were confirmed by sequencing before expression into the bacterial system. The fusion proteins were expressed in E. coli, strain DH-5α, by inducing with IPTG (isopropyl β-D-thiogalactoside) at 30 °C. The bacteria were lysed in GST lysis buffer (1×PBS, 50 mM EDTA, 10% glycerol, 0.5% aprotonin, 1 mM dithiothreitol and 1 mM PMSF) with lysozyme and purified with a GST expression and purification kit (Amersham Biosciences).

Assay for intracellular Ca2+

Intracellular Ca2+ was detected using FLUO-4 AM (acetoxymethyl ester) dye (Molecular Probes, Invitrogen) as described previously [36]. Briefly, 1×105 cholangiocarcinoma cells were seeded in a 35 mm dish with a 14 mm glass-bottom well. Cells were washed with HBSS (Hanks balanced salt solution) and incubated at room temperature with 5 μM FLUO-4 AM dye dissolved in DMSO followed by three washes of 5 min each with HBSS. After an initial image was recorded, Fas-activating antibody (CH-11; 500 ng/ml) was added to the sample followed by imaging over 60 min with one image recorded per minute. All images were taken at ×100 magnification using an Olympus IX 70 inverted microscope with epifluorescence optics. IPLab Spectrum and Ratio from Scanalytics (Fairfax, VA, U.S.A.) were used for image acquisition and data analysis.

Generation of cholangiocarcinoma cells stably overexpressing WT or mutant ΔFLIPL

Stable overexpression of FLIPL in cholangiocarcinoma cells was achieved using a lentiviral FLIPL expression vector (pLenti-FLIP) kindly provided by Dr Shi-Yong Sun (Winship Cancer Institute, Emory University, Atlanta, GA, U.S.A.) [37]. The lentiviral vector expressing a FLIPL deletion mutant lacking the CaM-binding region (ΔFLIP) was generated by using the restriction enzymes BamHI and AflII to delete amino acids 197–213 from full-length FLIPL cDNA. Infectious viral vector particles were generated by each expression vector into HEK-293T cells [HEK-293 cells (human embryonic kidney cells) expressing the large T-antigen of SV40 (simian virus 40)] together with the pΔ8.9 packaging construct and the Env pseudotyping construct pmD.G, as described previously [38]. Cholangiocarcinoma cells were transduced with each of the vectors for exogenous expression of the WT FLIP and ΔFLIP proteins respectively.

Statistical analysis

The differences between the two groups were analysed by a Student's t test. For multiple groups, one-way ANOVA and Student–Newman–Keuls tests were used to identify differences. Statistical significance was defined as P<0.05.

RESULTS

CaM interacts with FLIPL but not with FADD and caspase 8

We have previously characterized CaM binding to Fas, a member of the TNF superfamily of death receptors, and this binding is regulated during Fas-induced apoptosis [31]. The interaction of CaM and Fas is unique, because CaM does not bind to other members of the TNF family such as DR4, DR5 or TNF-R1. This exclusive interaction between CaM and Fas suggests a novel role for CaM in the Fas signalling pathway, which could be a mechanism to explain the variable effects of CaM antagonists in inducing apoptosis in various cell lines. To further delineate the role of CaM in Fas-induced signalling, we determined whether CaM interacted with other members of Fas-induced DISC. Protein pulldown assays were performed with proteins extracted from cholangiocarcinoma cells using CaMS beads. Western-blot analyses of proteins pulled down by CaMS demonstrated the specific interaction of CaM with c-FLIPL, but not with the other DISC-associated proteins, FADD and caspase 8 (Figure 1A). In addition, CaMS beads did not pulldown the short isoform of c-FLIP, FLIPS. The binding between CaM and FLIPL was further confirmed in cholangiocarcinoma cells by Co-IP analysis with an anti-CaM antibody (Figure 1B). Similarly, Co-IP of CaM and FLIPL was observed in Jurkat cells (results not shown), suggesting that CaM/FLIP binding could be a universal phenomenon. The CaM/FLIP binding was further visualized by immunofluorescence staining. As depicted in Figure 1(C), CaM (shown in green) and FLIP (shown in red) co-localized (seen as yellow in the merged image) in cholangiocarcinoma cells.

CaM binding to c-FLIPL

Figure 1
CaM binding to c-FLIPL

(A) CaM binds to c-FLIPL in cholangiocarcinoma cells. Whole cell lysates from cholangiocarcinoma cells were used for the protein pulldown assay using CaMS or CS beads. The beads were boiled in 2×SDS+10 mM EGTA buffer followed by separation with SDS/12% PAGE and immunoblotted for c-FLIP, FADD and caspase 8. The result shown here is representative of three independent experiments. (B) CaM and c-FLIPL were co-immunoprecipitated (IP) by an anti-CaM antibody. Cholangiocarcinoma cell lysates were used in Co-IP with an anti-CaM antibody or isotype-matched control mouse IgG1 antibody and immunoblotted for CaM and FLIP as shown in (B). α-CaM 1 and α-CaM 2 represent two consecutive samples eluted from an α-CaM antibody-conjugated column and control IgG1 represents the sample eluted from isotype-matched mouse (IgG1) antibody-conjugated columns. The asterisk in (B) represents anti-CaM antibody-conjugated beads used as a control for the heavy chain of the antibody. The results are representative of four independent experiments. IB, immunoblot. (C) Immunofluorescence staining showing co-localization of CaM and c-FLIP. Cholangiocarcinoma cells were fixed on the glass coverslips using 3% paraformaldehyde and stained with anti-CaM (green) and anti-FLIP (red) antibodies. Images were taken at ×100 magnification. The merge (yellow) represents the picture obtained by merging red and green fluorescent images. ‘Control’ in the right-hand panels represents Alexa Flour® 488 and Alexa Flour® 594 fluorescent secondary antibodies alone, used as negative controls to rule out non-specific binding. The results shown are representative of five independent experiments.

Figure 1
CaM binding to c-FLIPL

(A) CaM binds to c-FLIPL in cholangiocarcinoma cells. Whole cell lysates from cholangiocarcinoma cells were used for the protein pulldown assay using CaMS or CS beads. The beads were boiled in 2×SDS+10 mM EGTA buffer followed by separation with SDS/12% PAGE and immunoblotted for c-FLIP, FADD and caspase 8. The result shown here is representative of three independent experiments. (B) CaM and c-FLIPL were co-immunoprecipitated (IP) by an anti-CaM antibody. Cholangiocarcinoma cell lysates were used in Co-IP with an anti-CaM antibody or isotype-matched control mouse IgG1 antibody and immunoblotted for CaM and FLIP as shown in (B). α-CaM 1 and α-CaM 2 represent two consecutive samples eluted from an α-CaM antibody-conjugated column and control IgG1 represents the sample eluted from isotype-matched mouse (IgG1) antibody-conjugated columns. The asterisk in (B) represents anti-CaM antibody-conjugated beads used as a control for the heavy chain of the antibody. The results are representative of four independent experiments. IB, immunoblot. (C) Immunofluorescence staining showing co-localization of CaM and c-FLIP. Cholangiocarcinoma cells were fixed on the glass coverslips using 3% paraformaldehyde and stained with anti-CaM (green) and anti-FLIP (red) antibodies. Images were taken at ×100 magnification. The merge (yellow) represents the picture obtained by merging red and green fluorescent images. ‘Control’ in the right-hand panels represents Alexa Flour® 488 and Alexa Flour® 594 fluorescent secondary antibodies alone, used as negative controls to rule out non-specific binding. The results shown are representative of five independent experiments.

Interaction between CaM and FLIPL is Ca2+-dependent

Since CaM interacts with its target proteins in both a Ca2+-dependent and a Ca2+-independent manner, the interaction between CaM and FLIP was further characterized to determine whether the binding was affected by Ca2+. The effect of a Ca2+ chelator, EGTA, on CaM–FLIP binding was determined. EGTA (2 mM) inhibited the binding between CaM and FLIPL, as a decreased amount of FLIPL was pulled down by CaMS beads in protein pulldown assays (Figure 2A). FLIPS was not pulled down by CaMS, both in the presence and absence of EGTA. Similarly, EGTA decreased the amount of FLIPL co-precipitated with CaM in the immunoprecipitation using anti-CaM antibody, further confirming the Ca2+ dependence of the CaM–FLIPL interaction in vivo (Figure 2B).

Ca2+-dependent binding between CaM and c-FLIPL

Figure 2
Ca2+-dependent binding between CaM and c-FLIPL

(A) Whole cell lysates from cholangiocarcinoma cells were used in the protein pulldown assay with CaMS or CS beads in the presence or absence of 2 mM EGTA. The proteins pulled down were immunoblotted for c-FLIP and CaM. The results shown are representative of three independent experiments. (B) Cholangiocarcinoma cell lysates were used for Co-IP using anti-CaM antibody in the presence (+) or absence (–) of 2 mM EGTA, and Western blots were analysed for CaM and FLIP. The presence of 2 mM EGTA diminished the CaM–FLIPL binding as decreased FLIPL co-precipitated in the presence of EGTA. The results are representative of three independent experiments. IP, immunoprecipitate; IB, immunoblot.

Figure 2
Ca2+-dependent binding between CaM and c-FLIPL

(A) Whole cell lysates from cholangiocarcinoma cells were used in the protein pulldown assay with CaMS or CS beads in the presence or absence of 2 mM EGTA. The proteins pulled down were immunoblotted for c-FLIP and CaM. The results shown are representative of three independent experiments. (B) Cholangiocarcinoma cell lysates were used for Co-IP using anti-CaM antibody in the presence (+) or absence (–) of 2 mM EGTA, and Western blots were analysed for CaM and FLIP. The presence of 2 mM EGTA diminished the CaM–FLIPL binding as decreased FLIPL co-precipitated in the presence of EGTA. The results are representative of three independent experiments. IP, immunoprecipitate; IB, immunoblot.

CaM–FLIP binding is regulated in response to Fas activation

The biological significance of CaM–FLIPL binding in Fas-mediated apoptosis was characterized in cholangiocarcinoma cells. We determined that treatment with Fas-activating antibody (clone CH-11) over 24 h induced significant apoptosis (59±5%, n=8, P<0.05) in cholangiocarcinoma cells. We further determined whether Fas activation affected the CaM–FLIP interaction. Immunoprecipitation analysis of cholangiocarcinoma cells exposed to Fas-activating antibody for 30 and 60 min demonstrated that CaM–FLIP binding was increased significantly at 30 min and returned to the basal level at 60 min (Figure 3A). The band intensities for FLIP and CaM were measured using densitometry and the percentage change in the ratios of band intensities of FLIP with respect to the corresponding band intensities of CaM were used to determine the change in CaM and FLIP binding. As compared with untreated cells, a 37.3±5.7% increase (n=6, P=0.001) in CaM–FLIP binding was identified in cells exposed to Fas-activating antibody for 30 min and the binding was decreased to the basal level at 60 min (Figure 3B). Immunofluorescence staining for CaM (green) and FLIP (red) after stimulation of cholangiocarcinoma cells with Fas-activating antibody for 30 min showed that CaM and FLIP co-localized in the same apparent discrete subcellular location, as opposed to the diffused cytoplasmic staining seen in unstimulated cells, further confirming that the CaM–FLIP binding is regulated in response to Fas activation (Figure 3C). Fas and c-FLIP are known to move towards the membrane compartment of the cell to form the DISC during Fas-induced apoptosis. Thus it is likely that CaM and FLIP are moving towards the membrane compartment of the cell in response to Fas activation. These findings further underscore the involvement of CaM in Fas-induced apoptosis.

Changes in CaM–FLIP interaction in response to Fas activation

Figure 3
Changes in CaM–FLIP interaction in response to Fas activation

(A) Cholangiocarcinoma cells were treated with 500 ng/ml of Fas-activating antibody (CH-11) for 30 and 60 min. The whole cell lysates were used for co-immunoprecipitation (IP) using an anti-CaM antibody and immunoblotted (IB) for CaM and c-FLIP. (B) The band intensities for CaM and FLIP were measured by densitometry. The percentage change in the ratio of FLIP and CaM band intensities, shown in the Figure as means±S.E.M., was calculated for the indicated times (n=6, P=0.001). (C) Immunofluorescence staining for CaM (green) and FLIP (red) after the activation of cholangiocarcinoma cells with CH-11 for 30 min showed the movement of CaM and FLIP towards the same subcellular compartment, as shown in the lower panels in (C). The upper panels, shown as 0 min, represent untreated cholangiocarcinoma cells. The results shown are representative of three independent experiments.

Figure 3
Changes in CaM–FLIP interaction in response to Fas activation

(A) Cholangiocarcinoma cells were treated with 500 ng/ml of Fas-activating antibody (CH-11) for 30 and 60 min. The whole cell lysates were used for co-immunoprecipitation (IP) using an anti-CaM antibody and immunoblotted (IB) for CaM and c-FLIP. (B) The band intensities for CaM and FLIP were measured by densitometry. The percentage change in the ratio of FLIP and CaM band intensities, shown in the Figure as means±S.E.M., was calculated for the indicated times (n=6, P=0.001). (C) Immunofluorescence staining for CaM (green) and FLIP (red) after the activation of cholangiocarcinoma cells with CH-11 for 30 min showed the movement of CaM and FLIP towards the same subcellular compartment, as shown in the lower panels in (C). The upper panels, shown as 0 min, represent untreated cholangiocarcinoma cells. The results shown are representative of three independent experiments.

Increase in intracellular Ca2+ in response to Fas activation

Intracellular Ca2+ has been reported to increase in response to induction of apoptosis [39]. We investigated the response of Ca2+ in cholangiocarcinoma cells after Fas stimulation using FLUO-4 AM dye that stains intracellular Ca2+. The images were recorded and quantified as described in the Materials and methods section. An increase in intracellular Ca2+ that reached a peak at 30 min followed by a gradual decline over 60 min after Fas activation was observed in cholangiocarcinoma cells as shown in Figure 4. Thus the increased binding between CaM and FLIP at 30 min (Figure 3) following Fas stimulation paralleled the rise in intracellular Ca2+ that reached a peak at 30 min in response to Fas-activating antibody, suggesting that the Ca2+ dependence of CaM–FLIP binding is biologically significant.

Changes in intracellular Ca2+ with Fas activation in cholangiocarcinoma cells

Figure 4
Changes in intracellular Ca2+ with Fas activation in cholangiocarcinoma cells

The intracellular Ca2+ was labelled using FLUO-4 AM dye and, after recording a baseline image, Fas-activating antibody (CH-11; 500 ng/ml) was added to the sample and images were recorded at 1 min intervals over 60 min. Addition of isotype-matched mouse IgM antibody as a negative control showed no changes in intracellular Ca2+, whereas addition of 1 μM thapsigargin at the end of the experiment as a positive control showed a large increase in intracellular Ca2+ levels (results not shown). The levels of intracellular Ca2+ reached a peak at approx. 30 min and decreased gradually over the next 30 min. The results are representative of three independent experiments.

Figure 4
Changes in intracellular Ca2+ with Fas activation in cholangiocarcinoma cells

The intracellular Ca2+ was labelled using FLUO-4 AM dye and, after recording a baseline image, Fas-activating antibody (CH-11; 500 ng/ml) was added to the sample and images were recorded at 1 min intervals over 60 min. Addition of isotype-matched mouse IgM antibody as a negative control showed no changes in intracellular Ca2+, whereas addition of 1 μM thapsigargin at the end of the experiment as a positive control showed a large increase in intracellular Ca2+ levels (results not shown). The levels of intracellular Ca2+ reached a peak at approx. 30 min and decreased gradually over the next 30 min. The results are representative of three independent experiments.

CaM antagonist inhibits Fas-induced CaM–FLIP binding and ERK activity

TFP is a potent CaM antagonist and we have previously shown that it induces apoptosis in cholangiocarcinoma cells in a Fas-related mechanism [7,10]. To further characterize the biological relevance of CaM–FLIP binding in Fas-mediated signalling, we determined the effect of TFP (10 μM) on CaM–FLIP binding and the activity of the ERK pathway, a downstream signal that has been reported to mediate the survival pathway of FLIP in response to activation of death receptors [40]. Fas, extensively studied as a death receptor, is able to send survival signals as well through ERK activation and it is speculated that the final outcome of Fas stimulation depends on which signalling pathway, pro-apoptotic or pro-survival, predominates in a given system [41,42]. We found that pretreatment of cholangiocarcinoma cells with TFP (10 μM) for 30 min inhibited the Fas-stimulated increase in CaM–FLIP binding (Figure 5A). The basal level of CaM–FLIP binding was also decreased by TFP. Further, we determined whether inhibition of CaM–FLIP binding affected the phosphorylation of ERK signalling [p44/42 MAPK (mitogen-activated protein kinase)]. Exposure of cells to TFP for 30 and 60 min did not affect phosphorylation of ERK (pERK; Figure 5B, rightmost panel). However, pretreatment with TFP caused a decrease in phosphorylation of ERK after cells were exposed to Fas-activating antibody (CH-11), although CH-11 alone did not affect phosphorylation of ERK at 30 and 60 min. Such an inhibition of ERK phosphorylation by TFP paralleled its inhibition of CaM–FLIP binding. Thus increased CaM–FLIP binding after Fas stimulation appears to be important in maintaining ERK phosphorylation.

The CaM antagonist, TFP, interferes with CaM–FLIP binding with concurrent reduction in phosphorylation of ERK

Figure 5
The CaM antagonist, TFP, interferes with CaM–FLIP binding with concurrent reduction in phosphorylation of ERK

(A) Cholangiocarcinoma cells were treated with 10 μM TFP alone or Fas agonist antibody (CH-11) with or without pretreatment for 30 min with TFP. The whole cell lysates from these cells were used for Co-IP (IP) using anti-CaM antibody and immunoblotted (IB) for CaM and c-FLIP. (B) The whole cell lysates from these cells were separated by SDS/12% PAGE and immunoblotted for phosphorylated and total ERK (p44/42). The results are representative of three independent experiments.

Figure 5
The CaM antagonist, TFP, interferes with CaM–FLIP binding with concurrent reduction in phosphorylation of ERK

(A) Cholangiocarcinoma cells were treated with 10 μM TFP alone or Fas agonist antibody (CH-11) with or without pretreatment for 30 min with TFP. The whole cell lysates from these cells were used for Co-IP (IP) using anti-CaM antibody and immunoblotted (IB) for CaM and c-FLIP. (B) The whole cell lysates from these cells were separated by SDS/12% PAGE and immunoblotted for phosphorylated and total ERK (p44/42). The results are representative of three independent experiments.

Binding between CaM and FLIPL is direct

To determine whether CaM binds to FLIPL directly, purified recombinant FLIPL was used in the protein pull down assay with CaMS beads. Human FLIPL cDNA was cloned into a GST vector and expressed in E. coli and the proteins were purified using affinity purification. As shown in Figure 6(A), GST–FLIP protein bound to CaMS but not to CS beads, indicating a specific direct binding of CaM and FLIP. The binding between CaM and GST–FLIP was diminished in the presence of the Ca2+ chelator, EGTA (Figure 6B), confirming the Ca2+ dependence of the binding between purified proteins. To exclude the possibility of non-specific binding between CaM and GST, Factor Xa was used to cleave the GST tag from the GST–FLIP fusion protein. Recombinant FLIPL protein cleaved from the GST–FLIP fusion protein bound specifically to CaMS beads (Figure 6C). The direct binding between CaM and FLIPL was confirmed by using mouse FLIPL construct as well (results not shown).

CaM binds directly with recombinant FLIP protein

Figure 6
CaM binds directly with recombinant FLIP protein

(A) GST–FLIP fusion protein was used for protein pulldown using CaMS or CS beads. The proteins obtained from the binding reaction were separated by SDS/10% PAGE. Western blot using an anti-GST antibody showed that GST–FLIP fusion protein was pulled down specifically by CaMS beads. (B) Binding between purified GST–FLIP and CaM was diminished in the presence of EGTA as shown by Western-blot analysis for c-FLIP in (B). ‘Input’ indicates 1/30th of the lysate used in protein pulldown with CaMS. (C) GST tag was cleaved from GST–FLIP fusion protein using Factor Xa and this purified FLIP protein bound to CaMS beads specifically as shown in (C). The results shown are representative of three independent experiments. WB, Western blot.

Figure 6
CaM binds directly with recombinant FLIP protein

(A) GST–FLIP fusion protein was used for protein pulldown using CaMS or CS beads. The proteins obtained from the binding reaction were separated by SDS/10% PAGE. Western blot using an anti-GST antibody showed that GST–FLIP fusion protein was pulled down specifically by CaMS beads. (B) Binding between purified GST–FLIP and CaM was diminished in the presence of EGTA as shown by Western-blot analysis for c-FLIP in (B). ‘Input’ indicates 1/30th of the lysate used in protein pulldown with CaMS. (C) GST tag was cleaved from GST–FLIP fusion protein using Factor Xa and this purified FLIP protein bound to CaMS beads specifically as shown in (C). The results shown are representative of three independent experiments. WB, Western blot.

Computer modelling of FLIPL structure and generation of GST–FLIP deletion mutants

Based on the structural characteristics of various CaM-binding proteins, the putative CaM-binding regions on a given sequence can be predicted using various criteria including the hydrophobic moment, net charge and propensity of the sequence to form an α-helical structure [33,34]. To date, only the short isoform of c-FLIP (FLIPS) has been crystallized [43] and hence the structure of the C-terminal tail of FLIPL is not known. Since FLIPL shares extensive sequence and probably structural homology with caspase 8, the structure of caspase 8 was used as a reference to determine the helical sequences in the three-dimensional model of FLIPL. Six α-helices, shown as coiled ribbons, were identified in the C-terminal FLIPL sequence as potential CaM-binding regions (Figure 7A) and used to direct the generation of FLIPL deletion mutants using restriction-enzyme-mediated digestion of human FLIP cDNA as shown in Figure 7(B). The FLIP cDNA fragments obtained were cloned into a GST vector to obtain six N-terminal, one C-terminal and a ΔFLIP deletion mutant with deletion of amino acids 197–213 from full-length FLIPL. After confirming the sequences of these GST–FLIP deletion mutants, proteins were expressed in E. coli using IPTG induction and confirmed by immunoblotting for GST and/or FLIP.

FLIPL structure and generation of deletion mutants

Figure 7
FLIPL structure and generation of deletion mutants

(A) Computer modelling of the FLIPL structure. The presence of helical structures in the FLIPL sequence was determined using the caspase 8 homology model. Based on the binding analysis using GST–FLIP deletion mutants, the CaM-binding region appears to lie in a region of c-FLIPL shown by a dotted line. The structure of FLIP DEDs is shown on the left side of this dotted line; on the right side of the dotted line is the C-terminal tail of FLIPL, modelled on the basis of sequence homology with caspase 8. (B) Generation of GST–FLIP deletion mutants. Schematic representation of various GST–FLIP mutants; ‘F (WT)’, full-length/WT c-FLIP; N383, N280, N257, N233, N213 and N196, mutants with corresponding number of N-terminal amino acids; C280, mutant with C-terminal 280 amino acids; ΔFLIP (delta FLIP), full-length FLIPL with deletion of amino acids 197–213.

Figure 7
FLIPL structure and generation of deletion mutants

(A) Computer modelling of the FLIPL structure. The presence of helical structures in the FLIPL sequence was determined using the caspase 8 homology model. Based on the binding analysis using GST–FLIP deletion mutants, the CaM-binding region appears to lie in a region of c-FLIPL shown by a dotted line. The structure of FLIP DEDs is shown on the left side of this dotted line; on the right side of the dotted line is the C-terminal tail of FLIPL, modelled on the basis of sequence homology with caspase 8. (B) Generation of GST–FLIP deletion mutants. Schematic representation of various GST–FLIP mutants; ‘F (WT)’, full-length/WT c-FLIP; N383, N280, N257, N233, N213 and N196, mutants with corresponding number of N-terminal amino acids; C280, mutant with C-terminal 280 amino acids; ΔFLIP (delta FLIP), full-length FLIPL with deletion of amino acids 197–213.

CaM binds to FLIPL between amino acids 197 and 213

The series of GST–FLIP deletion mutants (shown in Figure 7B) were used in protein pulldown assays to identify potential CaM-binding motifs on FLIP. We found that mutants N383, N280, N257, N233 N213 and C280 but not N196 bound to CaM (Figure 8). Thus the likely CaM-binding site was narrowed to amino acids 197–213 on FLIPL. Furthermore deletion of amino acids 197–213 (using the restriction enzymes BamHI and AflII) from FLIPL cDNA showed reduced binding of this mutant with CaM (Figure 8D), confirming that the region from amino acids 197–213 in FLIPL is important for binding with CaM. The short isoform, FLIPS, contains 221 amino acids, with the N-terminal 202 amino acids identical with those of FLIPL. Thus localization of the CaM-binding site on FLIPL to amino acids 197–213 explains the lack of CaM binding to FLIPS. Furthermore, the binding of fragment N213, the shortest FLIPL mutant to bind CaM, was Ca2+-dependent (Figure 8B). Interestingly, although the CaM–FLIPL interaction was found to be Ca2+-dependent, the classical CaM-binding motifs, 1-8-14 and 1-5-10, observed in many proteins binding with CaM in a Ca2+-dependent manner [34], were not identified in this FLIPL sequence. This may categorize FLIPL as a novel interacting partner in a family of CaM-binding proteins with a CaM-binding motif not previously identified.

Binding of GST–FLIP mutants with CaM

Figure 8
Binding of GST–FLIP mutants with CaM

Protein pulldown studies with a 20 μl 1:1 slurry of CaMS or CS beads were performed using various GST–FLIP mutants. Proteins bound to the CaMS beads were separated by SDS/12% PAGE, and Western-blot analyses were performed using anti-GST and/or anti-FLIP antibodies. (A) The upper panel in (A) shows inputs (1/30th) of various GST–FLIP mutants. The lower panel shows binding of the mutants to CaMS. Fragments N383, N280, N257, N233 and N213 bound to CaMS but fragment N196 did not bind to CaMS. (B) Effects of EGTA on the binding of the FLIP fragments N213 and N196. ‘Input’ indicates 1/30th of the GST–FLIP lysates used in the protein pulldown assay. (C) CaMS pulldown analysis of the FLIP fragment C280, which lacks the N-terminal region (196 amino acids) containing DEDs. (D) CaMS pulldown analysis of the ΔFLIP fragment with deletion of amino acids 197–213 from full-length FLIPL. Representative blots of three experiments are shown.

Figure 8
Binding of GST–FLIP mutants with CaM

Protein pulldown studies with a 20 μl 1:1 slurry of CaMS or CS beads were performed using various GST–FLIP mutants. Proteins bound to the CaMS beads were separated by SDS/12% PAGE, and Western-blot analyses were performed using anti-GST and/or anti-FLIP antibodies. (A) The upper panel in (A) shows inputs (1/30th) of various GST–FLIP mutants. The lower panel shows binding of the mutants to CaMS. Fragments N383, N280, N257, N233 and N213 bound to CaMS but fragment N196 did not bind to CaMS. (B) Effects of EGTA on the binding of the FLIP fragments N213 and N196. ‘Input’ indicates 1/30th of the GST–FLIP lysates used in the protein pulldown assay. (C) CaMS pulldown analysis of the FLIP fragment C280, which lacks the N-terminal region (196 amino acids) containing DEDs. (D) CaMS pulldown analysis of the ΔFLIP fragment with deletion of amino acids 197–213 from full-length FLIPL. Representative blots of three experiments are shown.

CaM–FLIP binding is important in mediating spontaneous as well as Fas-induced apoptosis by FLIPL

FLIPL has been ascribed with anti- as well as pro-apoptotic functions in Fas signalling [19]. To determine the function of FLIP in cholangiocarcinoma cell lines and the importance of CaM–FLIPL binding in mediating FLIPL function, we stably overexpressed WT FLIPL (WT FLIP) and FLIPL lacking the CaM-binding region (ΔFLIP) in cholangiocarcinoma cells using the lentiviral expression vectors (Figure 9A). Overexpression of WT FLIP reduced both basal apoptosis and the sensitivity to Fas-induced apoptosis as compared with control (LacZ) cells (Figure 9B). Even though the ΔFLIP-overexpressing cells were found to be resistant to Fas-induced apoptosis as compared with control (LacZ) cells, their sensitivity to Fas-induced apoptosis was higher than WT FLIP-overexpressing cells. Thus WT FLIP was found to be a more efficient inhibitor of Fas-stimulated apoptosis than ΔFLIP protein (Figure 9B, mean±S.E.M., n=8, P<0.05). Therefore CaM–FLIP binding seems to be important for the anti-apoptotic function of FLIPL in the Fas pathway. It is likely that the abrogation of CaM–FLIP binding may affect FLIPL such that there is inefficient recruitment and/or activity of FLIPL into the DISC. Interestingly, unlike WT FLIP that rendered cells more resistant to apoptosis under unstimulated conditions, basal apoptosis was increased significantly in cholangiocarcinoma cells that overexpressed the ΔFLIP protein, suggesting that CaM–FLIP binding is also an important determinant of spontaneous apoptosis under basal conditions.

Effect of disrupted CaM–FLIP binding on Fas-mediated apoptosis

Figure 9
Effect of disrupted CaM–FLIP binding on Fas-mediated apoptosis

(A) Western-blot analysis for FLIP expression using cell lysates extracted from control cholangiocarcinoma cells (LacZ) or cells overexpressing WT and mutant ΔFLIP (with deletion of CaM-binding region). Native FLIPL is marked by a bold arrow, whereas overexpressed FLIP is marked by a dotted grey arrow. (B) Apoptosis was determined by annexin V and propidium iodide staining in cholangiocarcinoma cells overexpressing WT FLIP, ΔFLIP or LacZ as a control. Spontaneous apoptosis in WT FLIP-overexpressing cells was compared with control vector (LacZ)-transduced cells, and that of ΔFLIP-overexpressing cells was compared with LacZ- as well as WT FLIP-overexpressing cholangiocarcinoma cells. Fas-induced apoptosis in each cell line was found to be statistically significantly different compared with control conditions in the respective cell lines (statistical significance not marked). Results shown are means±S.E.M. (n=8). Statistical significance (*) represents P<0.05 for comparisons shown by curly brackets.

Figure 9
Effect of disrupted CaM–FLIP binding on Fas-mediated apoptosis

(A) Western-blot analysis for FLIP expression using cell lysates extracted from control cholangiocarcinoma cells (LacZ) or cells overexpressing WT and mutant ΔFLIP (with deletion of CaM-binding region). Native FLIPL is marked by a bold arrow, whereas overexpressed FLIP is marked by a dotted grey arrow. (B) Apoptosis was determined by annexin V and propidium iodide staining in cholangiocarcinoma cells overexpressing WT FLIP, ΔFLIP or LacZ as a control. Spontaneous apoptosis in WT FLIP-overexpressing cells was compared with control vector (LacZ)-transduced cells, and that of ΔFLIP-overexpressing cells was compared with LacZ- as well as WT FLIP-overexpressing cholangiocarcinoma cells. Fas-induced apoptosis in each cell line was found to be statistically significantly different compared with control conditions in the respective cell lines (statistical significance not marked). Results shown are means±S.E.M. (n=8). Statistical significance (*) represents P<0.05 for comparisons shown by curly brackets.

DISCUSSION

CaM interacts with a variety of proteins such as calcineurin, CaM kinases, myosin light-chain kinase, nitric oxide synthase and neuromodulin [34] and affects numerous signalling pathways such as inflammation, memory, muscle contraction, the immune response and ion channel functioning [44]. Previously, we reported a direct and dynamic interaction between CaM and Fas suggesting a novel role of CaM in Fas-mediated signalling [31]. CaM binding to Fas is mediated through a classic Ca2+-dependent CaM-binding region, termed the 1-5-10 motif, present in the cytoplasmic DD of Fas [31]. The point mutation (V254N) in the cytoplasmic DD of Fas in lprcg mice, whose phenotype is analogous to that of the human autoimmune lymphoproliferative syndrome, results in the inability of Fas to form the DISC [45]. This same mutation also causes reduced binding between CaM and Fas as seen in osteoclasts from lprcg mice [46]. Considering these findings and the multitude of signalling pathways CaM is involved in, it is likely that CaM plays an important but yet undetermined role in Fas-mediated signalling. Consistently, we found that CaM is recruited into the Fas-mediated DISC [47]. Therefore we studied the interaction of CaM with other proteins involved in Fas-induced DISC. In the present paper we report that c-FLIPL, a mediator in the Fas-stimulated DISC, is a novel binding partner of CaM. We confirmed the direct binding between CaM and c-FLIPL by various methods including CaMS protein pulldown, immunoprecipitation, epifluorescence and in vitro binding assays using recombinant proteins (Figures 1, 2 and 6). The CaM–FLIP binding was increased in response to Fas activation and this increase in binding correlated with a rise in intracellular Ca2+ in response to Fas activation, further corroborating a likely functional significance of this binding in cholangiocarcinoma cells (Figures 3A, 3B and 4). Interestingly, CaM does not interact with other homologous proteins of the DISC, namely FADD and caspase 8, which gives rise to the interesting possibility that CaM–Fas and CaM–FLIP binding may act as counter mechanisms for Fas–FADD–caspase 8 assembly during DISC formation.

The CaM-binding region was localized to amino acids 197–213 on FLIPL, which also explains the lack of binding of CaM with c-FLIPS (short isoform of c-FLIP). FLIPS is a 26 kDa protein with 221 amino acids whose sequence differs from the long isoform, FLIPL, from amino acid 203 to the C-terminus. Also this region lies outside the DEDs of c-FLIP, further explaining the lack of CaM binding with the other homologous proteins, FADD and caspase 8, which share DEDs as the homologous domain in the family. In addition, the CaM-binding site on FLIPL does not appear to share the characteristics of any of the classical Ca2+-dependent CaM-binding motifs described so far and thus FLIPL may be a novel target in a family of CaM-binding peptides.

The precise role of the CaM–FLIP interaction in regulating Fas-mediated apoptosis is yet to be fully elucidated. In the present study we showed that overexpression of WT FLIP decreased spontaneous as well as Fas-induced apoptosis in cholangiocarcinoma cells. However, disruption of CaM–FLIP binding with ΔFLIP rendered cholangiocarcinoma cells more sensitive to spontaneous and Fas-induced apoptosis as compared with WT FLIP-overexpressing cells. Thus CaM–FLIP binding appears to be important not only for mediating anti-apoptotic effects of FLIPL in the Fas pathway, but also for maintaining survival signals under basal conditions. It is possible that the disruption of CaM–FLIP binding causes a conformational change in FLIP, which affects its recruitment into the DISC, thus affecting its anti-apoptotic role in cholangiocarcinoma cells. Further investigations are warranted to define the mechanisms by which CaM–FLIP modulates cell survival and apoptosis under basal conditions and upon Fas stimulation.

We showed in the present study that CaM–FLIP binding increased after 30 min of Fas activation in cholangiocarcinoma cells (Figures 3A and 3B). Furthermore, CaM and FLIP were shown to co-localize in the same apparent discrete subcellular compartment after Fas activation. In combination, these observations demonstrated that the CaM–FLIP binding is regulated upon Fas stimulation, which might be important for mediating anti-apoptotic and/or pro-survival signalling by FLIP in response to Fas stimulation. Consistently, we showed here that the CaM antagonist, TFP, interfered with Fas-induced increase in CaM–FLIP binding with a concurrent decrease in phosphorylation of ERK, a pro-survival signal downstream of FLIP (Figure 5). It is possible that the functions of CaM are cell-type-specific, possibly paralleling the cell-specific apoptosis-inducing abilities of CaM antagonists [10,2830].

Both CaM–Fas and CaM–FLIP interactions are Ca2+-dependent. Therefore there is a possibility that these interactions, by engaging CaM, affect the Ca2+ flux in the vicinity of the membrane compartment of the cell. As opposed to Jurkat cells in which a rise in intracellular Ca2+ is comparatively a late phenomenon [39], cholangiocarcinoma cells showed a relatively rapid rise in intracellular Ca2+ in response to Fas stimulation (Figure 4). Jurkat cells are known to be type II cells in which only minimal formation of the DISC is required to initiate apoptosis that proceeds mainly through the mitochondrial pathway [48]. Type I cells, however, are characterized by the formation of the DISC (to recruit caspase 8), which is required for the activation of downstream caspases. Overexpression of FLIPL in cholangiocarcinoma cells interfered with caspase 8 activation and significantly reduced their sensitivity to Fas-mediated apoptosis, which is consistent with previous reports attributing an inhibitory role to c-FLIP in Fas-mediated apoptosis in cholangiocarcinoma cells [9]. Thus the cholangiocarcinoma cells that we used in the present studies appear to be type I cells in which modulation (with increased recruitment of c-FLIP) in the formation of the DISC abrogates apoptotic cell death. Therefore it is very likely that the CaM–FLIP interaction is an important modulator of Fas-induced DISC, which is the essential pathway for apoptotic cell death in cholangiocarcinoma cells.

Although Fas is well characterized as a death receptor in the apoptotic machinery, it can also activate non-apoptotic signalling events [45]. Activation of Fas has been shown to induce cell proliferation and tissue regeneration as well [49] and c-FLIP is touted as one of the important determinants of downstream signalling initiated after Fas activation. Fas-induced activation of the ERK pathway has been shown to be independent of caspase activation such that caspase inhibition did not affect Fas-induced ERK activation [42]. This supports the notion that the Fas–CaM–FLIP arm of the DISC might be important in modulating survival signals in response to Fas activation, independent of caspase involvement. It will be interesting to study the CaM–FLIP interaction with respect to other death receptors such as DR4, DR5 and TNF-R1, which may provide further insights into the non-apoptotic signalling events mediated by these receptors. Further characterizing the nature of CaM–FLIP interaction may potentially provide new therapeutic targets for cancer therapy.

This work was supported by a Veterans Affairs Merit Review Award to J. M. M. We thank Albert Tousson [Imaging Facility, UAB (University of Alabama at Birmingham)] for help with live cell imaging, Dr Shi-Yong Sun (Winship Cancer Institute, Emory University, Atlanta, GA, U.S.A.) for providing us with the Lenti-FLIPL construct. J. C. K. is supported by a VA merit award. We thank the Genetically Defined Microbe and Expression Core of the UAB mucosal HIV and Immunology Center (R24DK64400) for technical assistance.

Abbreviations

     
  • AM

    acetoxymethyl ester

  •  
  • CaM

    calmodulin

  •  
  • CaMS

    CaM–Sepharose

  •  
  • c-FLIP

    cellular FLICE-like inhibitory protein

  •  
  • Co-IP

    co-immunoprecipitation

  •  
  • CS

    control Sepharose

  •  
  • DD

    death domain

  •  
  • DED

    death effector domain

  •  
  • DISC

    death-inducing signalling complex

  •  
  • ERK

    extracellular-signal-regulated kinase

  •  
  • FADD

    Fas-associated DD

  •  
  • FLICE

    FADD-like interleukin 1β-converting enzyme

  •  
  • GST

    glutathione transferase

  •  
  • HBSS

    Hanks balanced salt solution

  •  
  • HRP

    horseradish peroxidase

  •  
  • IPTG

    isopropyl β-D-thiogalactoside

  •  
  • TFP

    trifluoperazine

  •  
  • TNF

    tumour necrosis factor

  •  
  • TNF-R1

    TNF receptor 1

  •  
  • TRAIL

    TNF-related apoptosis-inducing ligand

  •  
  • TRAIL-R1

    TRAIL receptor 1

  •  
  • WT

    wild-type

References

References
1
Kerr
 
J. F.
Wyllie
 
A. H.
Currie
 
A. R.
 
Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics
Br. J. Cancer
1972
, vol. 
26
 (pg. 
239
-
257
)
2
Thompson
 
C. B.
 
Apoptosis in the pathogenesis and treatment of disease
Science
1995
, vol. 
267
 (pg. 
1456
-
1462
)
3
Gerl
 
R.
Vaux
 
D. L.
 
Apoptosis in the development and treatment of cancer
Carcinogenesis
2005
, vol. 
26
 (pg. 
263
-
270
)
4
Stahnke
 
K.
Fulda
 
S.
Friesen
 
C.
Strauss
 
G.
Debatin
 
K. M.
 
Activation of apoptosis pathways in peripheral blood lymphocytes by in vivo chemotherapy
Blood
2001
, vol. 
98
 (pg. 
3066
-
3073
)
5
Muller
 
M.
Strand
 
S.
Hug
 
H.
Heinemann
 
E. M.
Walczak
 
H.
Hofmann
 
W. J.
Stremmel
 
W.
Krammer
 
P. H.
Galle
 
P. R.
 
Drug-induced apoptosis in hepatoma cells is mediated by the CD95 (APO-1/Fas) receptor/ligand system and involves activation of wild-type p53
J. Clin. Invest.
1997
, vol. 
99
 (pg. 
403
-
413
)
6
Searle
 
J.
Lawson
 
T. A.
Abbott
 
P. J.
Harmon
 
B.
Kerr
 
J. F.
 
An electron-microscope study of the mode of cell death induced by cancer-chemotherapeutic agents in populations of proliferating normal and neoplastic cells
J. Pathol.
1975
, vol. 
116
 (pg. 
129
-
138
)
7
Ahn
 
E. Y.
Pan
 
G.
Oh
 
J. H.
Tytler
 
E. M.
McDonald
 
J. M.
 
The combination of calmodulin antagonists and interferon-γ induces apoptosis through caspase-dependent and -independent pathways in cholangiocarcinoma cells
Am. J. Pathol.
2003
, vol. 
163
 (pg. 
2053
-
2063
)
8
Jhala
 
N. C.
Vickers
 
S. M.
Argani
 
P.
McDonald
 
J. M.
 
Regulators of apoptosis in cholangiocarcinoma
Arch. Pathol. Lab. Med.
2005
, vol. 
129
 (pg. 
481
-
486
)
9
Que
 
F. G.
Phan
 
V. A.
Phan
 
V. H.
Celli
 
A.
Batts
 
K.
LaRusso
 
N. F.
Gores
 
G. J.
 
Cholangiocarcinomas express Fas ligand and disable the Fas receptor
Hepatology
1999
, vol. 
30
 (pg. 
1398
-
1404
)
10
Pan
 
G.
Vickers
 
S. M.
Pickens
 
A.
Phillips
 
J. O.
Ying
 
W.
Thompson
 
J. A.
Siegal
 
G. P.
McDonald
 
J. M.
 
Apoptosis and tumorigenesis in human cholangiocarcinoma cells. Involvement of Fas/APO-1 (CD95) and calmodulin
Am. J. Pathol.
1999
, vol. 
155
 (pg. 
193
-
203
)
11
Shimonishi
 
T.
Isse
 
K.
Shibata
 
F.
Aburatani
 
I.
Tsuneyama
 
K.
Sabit
 
H.
Harada
 
K.
Miyazaki
 
K.
Nakanuma
 
Y.
 
Up-regulation of Fas ligand at early stages and down-regulation of Fas at progressed stages of intrahepatic cholangiocarcinoma reflect evasion from immune surveillance
Hepatology
2000
, vol. 
32
 (pg. 
761
-
769
)
12
Peter
 
M. E.
Scaffidi
 
C.
Medema
 
J. P.
Kischkel
 
F.
Krammer
 
P. H.
 
The death receptors
Results Probl. Cell Differ.
1999
, vol. 
23
 (pg. 
25
-
63
)
13
Baumler
 
C.
Duan
 
F.
Onel
 
K.
Rapaport
 
B.
Jhanwar
 
S.
Offit
 
K.
Elkon
 
K. B.
 
Differential recruitment of caspase 8 to cFlip confers sensitivity or resistance to Fas-mediated apoptosis in a subset of familial lymphoma patients
Leuk. Res.
2003
, vol. 
27
 (pg. 
841
-
851
)
14
Irmler
 
M.
Thome
 
M.
Hahne
 
M.
Schneider
 
P.
Hofmann
 
K.
Steiner
 
V.
Bodmer
 
J. L.
Schroter
 
M.
Burns
 
K.
Mattmann
 
C.
, et al 
Inhibition of death receptor signals by cellular FLIP
Nature
1997
, vol. 
388
 (pg. 
190
-
195
)
15
Thome
 
M.
Schneider
 
P.
Hofmann
 
K.
Fickenscher
 
H.
Meinl
 
E.
Neipel
 
F.
Mattmann
 
C.
Burns
 
K.
Bodmer
 
J. L.
Schroter
 
M.
, et al 
Viral FLICE-inhibitory proteins (FLIPs) prevent apoptosis induced by death receptors
Nature
1997
, vol. 
386
 (pg. 
517
-
521
)
16
Shu
 
H. B.
Halpin
 
D. R.
Goeddel
 
D. V.
 
Casper is a FADD- and caspase-related inducer of apoptosis
Immunity
1997
, vol. 
6
 (pg. 
751
-
763
)
17
Scaffidi
 
C.
Schmitz
 
I.
Krammer
 
P. H.
Peter
 
M. E.
 
The role of c-FLIP in modulation of CD95-induced apoptosis
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
1541
-
1548
)
18
Djerbi
 
M.
Darreh-Shori
 
T.
Zhivotovsky
 
B.
Grandien
 
A.
 
Characterization of the human FLICE-inhibitory protein locus and comparison of the anti-apoptotic activity of four different flip isoforms
Scand. J. Immunol.
2001
, vol. 
54
 (pg. 
180
-
189
)
19
Chang
 
D. W.
Xing
 
Z.
Pan
 
Y.
Algeciras-Schimnich
 
A.
Barnhart
 
B. C.
Yaish-Ohad
 
S.
Peter
 
M. E.
Yang
 
X.
 
c-FLIP(L) is a dual function regulator for caspase-8 activation and CD95-mediated apoptosis
EMBO J.
2002
, vol. 
21
 (pg. 
3704
-
3714
)
20
Thome
 
M.
Tschopp
 
J.
 
Regulation of lymphocyte proliferation and death by FLIP
Nat. Rev. Immunol.
2001
, vol. 
1
 (pg. 
50
-
58
)
21
Dohrman
 
A.
Russell
 
J. Q.
Cuenin
 
S.
Fortner
 
K.
Tschopp
 
J.
Budd
 
R. C.
 
Cellular FLIP long form augments caspase activity and death of T cells through heterodimerization with and activation of caspase-8
J. Immunol.
2005
, vol. 
175
 (pg. 
311
-
318
)
22
Chen
 
Y.
Xu
 
J.
Jhala
 
N.
Pawar
 
P.
Zhu
 
Z. B.
Ma
 
L.
Byon
 
C. H.
McDonald
 
J. M.
 
Fas-mediated apoptosis in cholangiocarcinoma cells is enhanced by 3,3′-diindolylmethane through inhibition of AKT signaling and FLICE-like inhibitory protein
Am. J. Pathol.
2006
, vol. 
169
 (pg. 
1833
-
1842
)
23
Kiener
 
P. A.
Davis
 
P. M.
Starling
 
G. C.
Mehlin
 
C.
Klebanoff
 
S. J.
Ledbetter
 
J. A.
Liles
 
W. C.
 
Differential induction of apoptosis by Fas–Fas ligand interactions in human monocytes and macrophages
J. Exp. Med.
1997
, vol. 
185
 (pg. 
1511
-
1516
)
24
Babu
 
Y. S.
Sack
 
J. S.
Greenhough
 
T. J.
Bugg
 
C. E.
Means
 
A. R.
Cook
 
W. J.
 
Three-dimensional structure of calmodulin
Nature
1985
, vol. 
315
 (pg. 
37
-
40
)
25
Hoeflich
 
K. P.
Ikura
 
M.
 
Calmodulin in action: diversity in target recognition and activation mechanisms
Cell
2002
, vol. 
108
 (pg. 
739
-
742
)
26
Dowd
 
D. R.
MacDonald
 
P. N.
Komm
 
B. S.
Haussler
 
M. R.
Miesfeld
 
R.
 
Evidence for early induction of calmodulin gene expression in lymphocytes undergoing glucocorticoid-mediated apoptosis
J. Biol. Chem.
1991
, vol. 
266
 (pg. 
18423
-
18426
)
27
Gerner
 
C.
Frohwein
 
U.
Gotzmann
 
J.
Bayer
 
E.
Gelbmann
 
D.
Bursch
 
W.
Schulte-Hermann
 
R.
 
The Fas-induced apoptosis analyzed by high throughput proteome analysis
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
39018
-
39026
)
28
Jan
 
C. R.
Lu
 
C. H.
Chen
 
Y. C.
Cheng
 
J. S.
Tseng
 
L. L.
Jun-Wen
 
W.
 
Ca2+ mobilization induced by W-7 in MG63 human osteosarcoma cells
Pharmacol. Res.
2000
, vol. 
42
 (pg. 
323
-
327
)
29
Kang
 
U. G.
Kim
 
M. J.
Suh
 
P. G.
Ryu
 
S. H.
Park
 
J. B.
Kim
 
J. H.
Kim
 
Y. S.
Lee
 
Y. H.
 
Inhibition of trifluoperazine-induced DNA fragmentation by cyclic AMP mediated signaling
Mol. Cells
1999
, vol. 
9
 (pg. 
596
-
602
)
30
Pan
 
G.
Zhou
 
T.
Radding
 
W.
Saag
 
M. S.
Mountz
 
J. D.
McDonald
 
J. M.
 
Calmodulin antagonists inhibit apoptosis of CD4+ T-cells from patients with AIDS
Immunopharmacology
1998
, vol. 
40
 (pg. 
91
-
103
)
31
Ahn
 
E. Y.
Lim
 
S. T.
Cook
 
W. J.
McDonald
 
J. M.
 
Calmodulin binding to the Fas death domain. Regulation by Fas activation
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
5661
-
5666
)
32
Jurado
 
L. A.
Chockalingam
 
P. S.
Jarrett
 
H. W.
 
Apocalmodulin
Physiol. Rev.
1999
, vol. 
79
 (pg. 
661
-
682
)
33
O'Neil
 
K. T.
DeGrado
 
W. F.
 
How calmodulin binds its targets: sequence independent recognition of amphiphilic α-helices
Trends Biochem. Sci.
1990
, vol. 
15
 (pg. 
59
-
64
)
34
Rhoads
 
A. R.
Friedberg
 
F.
 
Sequence motifs for calmodulin recognition
FASEB J.
1997
, vol. 
11
 (pg. 
331
-
340
)
35
Sacks
 
D. B.
Porter
 
S. E.
Ladenson
 
J. H.
McDonald
 
J. M.
 
Monoclonal antibody to calmodulin: development, characterization, and comparison with polyclonal anti-calmodulin antibodies
Anal. Biochem.
1991
, vol. 
194
 (pg. 
369
-
377
)
36
Schneider
 
E. M.
Menzl
 
I.
Weber
 
O.
Hug
 
H.
 
Differential calcium response in HeLa and HeLa-Fas cells by cytotoxic T lymphocytes
Biochem. Biophys. Res. Commun.
2003
, vol. 
301
 (pg. 
159
-
166
)
37
Liu
 
X.
Yue
 
P.
Schonthal
 
A. H.
Khuri
 
F. R.
Sun
 
S. Y.
 
Cellular FLICE-inhibitory protein down-regulation contributes to celecoxib-induced apoptosis in human lung cancer cells
Cancer Res.
2006
, vol. 
66
 (pg. 
11115
-
11119
)
38
Zufferey
 
R.
Nagy
 
D.
Mandel
 
R. J.
Naldini
 
L.
Trono
 
D.
 
Multiply attenuated lentiviral vector achieves efficient gene delivery in vivo
Nat. Biotechnol.
1997
, vol. 
15
 (pg. 
871
-
875
)
39
Sen
 
C. K.
Sashwati
 
R.
Packer
 
L.
 
Fas mediated apoptosis of human Jurkat T-cells: intracellular events and potentiation by redox-active α-lipoic acid
Cell Death Differ.
1999
, vol. 
6
 (pg. 
481
-
491
)
40
Kataoka
 
T.
Budd
 
R. C.
Holler
 
N.
Thome
 
M.
Martinon
 
F.
Irmler
 
M.
Burns
 
K.
Hahne
 
M.
Kennedy
 
N.
Kovacsovics
 
M.
Tschopp
 
J.
 
The caspase-8 inhibitor FLIP promotes activation of NF-κB and Erk signaling pathways
Curr. Biol.
2000
, vol. 
10
 (pg. 
640
-
648
)
41
Peter
 
M. E.
Budd
 
R. C.
Desbarats
 
J.
Hedrick
 
S. M.
Hueber
 
A. O.
Newell
 
M. K.
Owen
 
L. B.
Pope
 
R. M.
Tschopp
 
J.
Wajant
 
H.
, et al 
The CD95 receptor: apoptosis revisited
Cell
2007
, vol. 
129
 (pg. 
447
-
450
)
42
Tran
 
S. E.
Holmstrom
 
T. H.
Ahonen
 
M.
Kahari
 
V. M.
Eriksson
 
J. E.
 
MAPK/ERK overrides the apoptotic signaling from Fas, TNF, and TRAIL receptors
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
16484
-
16490
)
43
Li
 
F. Y.
Jeffrey
 
P. D.
Yu
 
J. W.
Shi
 
Y.
 
Crystal structure of a viral FLIP: insights into FLIP-mediated inhibition of death receptor signaling
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
2960
-
2968
)
44
O'Day
 
D. H.
 
CaMBOT: profiling and characterizing calmodulin-binding proteins
Cell. Signalling
2003
, vol. 
15
 (pg. 
347
-
354
)
45
Legembre
 
P.
Barnhart
 
B. C.
Zheng
 
L.
Vijayan
 
S.
Straus
 
S. E.
Puck
 
J.
Dale
 
J. K.
Lenardo
 
M.
Peter
 
M. E.
 
Induction of apoptosis and activation of NF-kappaB by CD95 require different signalling thresholds
EMBO Rep.
2004
, vol. 
5
 (pg. 
1084
-
1089
)
46
Wu
 
X.
Ahn
 
E. Y.
McKenna
 
M. A.
Yeo
 
H.
McDonald
 
J. M.
 
Fas binding to calmodulin regulates apoptosis in osteoclasts
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
29964
-
29970
)
47
Chen
 
Y.
Pawar
 
P.
Pan
 
G.
Ma
 
L.
Liu
 
H.
McDonald
 
J. M.
 
Calmodulin binding to the Fas-mediated death-inducing signaling complex in cholangiocarcinoma cells
J. Cell. Biochem.
2008
, vol. 
103
 (pg. 
788
-
799
)
48
Scaffidi
 
C.
Schmitz
 
I.
Zha
 
J.
Korsmeyer
 
S. J.
Krammer
 
P. H.
Peter
 
M. E.
 
Differential modulation of apoptosis sensitivity in CD95 type I and type II cells
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
22532
-
22538
)
49
Desbarats
 
J.
Newell
 
M. K.
 
Fas engagement accelerates liver regeneration after partial hepatectomy
Nat. Med.
2000
, vol. 
6
 (pg. 
920
-
923
)