Ets-1 is a transcription factor that plays an important role in various physiological and pathological processes, such as development, angiogenesis, apoptosis and tumour invasion. In the present study, we have demonstrated that Ets-1 p51, but not the spliced variant Ets-1 p42, is processed in a caspase-dependent manner in Jurkat T-leukaemia cells undergoing apoptosis, resulting in three C-terminal fragments Cp20, Cp17 and Cp14 and a N-terminal fragment, Np36. In vitro cleavage of Ets-1 p51 by caspase 3 produces fragments consistent with those observed in cells undergoing apoptosis. These fragments are generated by cleavage at three sites located in the exon VII-encoded region of Ets-1 p51. This region is absent from the Ets-1 p42 isoform, which therefore cannot be cleaved by caspases. In Ets-1 p51, cleavage generates C-terminal fragments containing the DNA-binding domain, but lacking the transactivation domain. The Cp17 fragment, the major cleavage product generated during apoptosis, is devoid of transcriptional activity and inhibits Ets-1 p51-mediated transactivation of target genes by competing with Ets-1 p51 for binding to Ets-binding sites present in the target promoters. In the present study, we have demonstrated that caspase cleavage of Ets-1 within the exon VII-encoded region leads to specific down-regulation of the Ets-1 p51 isoform during apoptosis. Furthermore, our results establish that caspase cleavage generates a stable C-terminal fragment that acts as a natural dominant-negative form of the full-length Ets-1 p51 protein.
Apoptosis is a genetically controlled form of cell death that plays a critical role during development and tissue homoeostasis by ensuring the removal of damaged or unnecessary cells . Apoptotic stimulation leads to the cascade activation of cysteine proteases called caspases, which carry out the apoptotic programme by cleaving numerous substrates . Disruption in the regulation of apoptosis underlies the pathogenesis of many human diseases, including autoimmune diseases, neurogenerative disorders and cancers .
Since caspase-mediated proteolysis is crucial for the apoptotic process , identification of the molecular targets of caspases is necessary to better understand apoptotic signal transduction. Indeed, several classes of cellular proteins have been shown to be cleaved by caspases during apoptotic cell death . This usually results in either inactivation or activation of proteins that protect living cells or promote cell death. Among those substrates, several transcription factors are inactivated during apoptosis by caspase 3 including SRF (serum-response factor) , RARα (retinoid acid receptor α)  and NF-κB (nuclear factor κB) p65/RelA subunit . Likewise, in the Ets transcription factor family, Fli-1 has been shown to be cleaved in murine pre-B-leukaemic cells undergoing programmed cell death by a caspase-like activity and in vitro by caspase 3 .
Ets-1 is the founding member of the Ets family characterized by a well-conserved DBD (DNA-binding domain), called the ETS domain. It regulates gene expression by binding to specific DNA elements, called EBS (Ets-binding sites), found in the promoters of its target genes [10,11]. By binding to these elements, Ets-1 activates the transcription of various genes involved in numerous cellular mechanisms such as development, angiogenesis, proliferation and apoptosis . However, when overexpressed, Ets-1 is involved in the development of invasive pathologies such as rheumatoid arthritis , glomerulonephritis  and cancers .
In humans, two isoforms of Ets-1 have been described: p51, the predominant full-length isoform, and p42, a shorter alternatively spliced isoform that lacks the exon VII-encoded region . As a result of their respective structures, Ets-1 p51 and Ets-1 p42 isoforms have different DNA-binding and transcriptional properties [16–18]. Ets-1 p51 is autoinhibited for DNA binding owing to the presence of two inhibitory domains that flank and interact with its DBD . To counteract autoinhibition and improve DNA binding, Ets-1 p51 interacts with partners, enabling them to co-operatively bind to adjacent DNA elements . Ets-1 p51 can also counteract its autoinhibition without any interaction partners by binding to a particular palindromic arrangement of two EBS separated by 4 bp, as in the stromelysin-1 matrix metalloproteinase and the p53 oncosuppressor promoters [17,21,22]. Two molecules of Ets-1 p51 co-operatively bind to the EBS palindrome, thereby forming an Ets-1 p51–DNA–Ets-1 p51 ternary complex that is critical for full transactivation . In contrast, Ets-1 p42 only forms a binary Ets-1 p42–DNA complex, which is much less efficient in promoting transactivation [17,18]. These differences contribute to the Ets-1 isoform-specific modulation of particular target genes, illustrating that Ets-1 function can be regulated through the exon VII-encoded region [23,24].
Ets-1 p51 displays pro- and anti-apoptotic functions according to the biological context. Ets-1 p51 is required for the formation of a stable DNA–p53–CBP (CREB [cAMP-response-element-binding protein)-binding protein] complex to induce pro-apoptotic genes in the process of UV-induced apoptosis in embryonic stem cells . Furthermore, overexpression of Ets-1 p51 in human umbilical vein endothelial cells induces apoptosis under serum-deprived conditions by up-regulating pro-apoptotic genes . Anti-apoptotic functions have been ascribed to Ets-1 p51 in vascular smooth muscle cells protecting them from undergoing apoptosis by activating the transcription of p21waf1/cip1 . Moreover, Ets-1 p51 inhibits the expression of the pro-apoptotic gene bax by interacting with the transcription factor GFI-1 (growth factor independent-1) in the composite site present in the bax promoter . In contrast, Ets-1 p42 isoform expression seems to be essentially associated with apoptosis. The expression of Ets-1 p42, unlike that of Ets-1 p51, promotes Fas-mediated apoptosis by directly up-regulating ICE (interleukin-1β-converting enzyme)/caspase 1 gene expression in human colon cancer DLD-1 cells [24,29]. Ets-1 p42 expression can also induce apoptosis of these cells under low-serum conditions . Furthermore, Ets-1 p42-expressing DLD-1 cell xenografts in nude mice inhibit the tumorigenicity of the DLD-1 cancer cells by inducing apoptosis .
Previous studies have shown that the basal level of Ets-1 p51 decreases in cells undergoing apoptosis. For example, an apoptotic inducer, curcumin, decreases the amount of Ets-1 p51 protein in human endometrial carcinoma HEC-1-A cells in a time- and dose-dependent manner . Likewise, flow cytometry and confocal microscopy analysis have shown that Ets-1 expression is down-regulated in PBMC (peripheral blood mononuclear cells) and bursal cells induced to apoptosis .
In the present study, we have demonstrated that the induction of apoptosis in human acute T-cell leukaemia (Jurkat) cells causes a decrease in Ets-1 p51 levels that is associated with the generation of caspase-dependent fragments. Furthermore, we characterize the mechanisms and the functional consequences of Ets-1 p51 cleavage on the transcriptional modulation of Ets-1 target genes.
Cell culture and induction of apoptosis
HEK (human embryonic kidney)-293 cells were cultured in DMEM (Dulbecco's modified Eagle's medium) (Invitrogen) supplemented with 10% FBS (fetal bovine serum) and 50 μg/ml gentamycin. To induce apoptosis, HEK-293 cells were incubated 24 h after transfection in medium supplemented with 1 μM staurosporin (Calbiochem) for 8 h. Jurkat human T-leukaemia cells were cultured in RPMI 1640 medium (Invitrogen) supplemented with 10% FBS and 50 μg/ml gentamycin. To induce apoptosis, Jurkat cells (3×106) were incubated in medium supplemented with 1 μM staurosporin for 4 h. MDCK (Madin–Darby canine kidney) epithelial cells were cultured in DMEM supplemented with 10% FBS, 50 μg/ml gentamycin and 100 μg/ml penicillin/streptomycin. To induce apoptosis, MDCK cells were incubated 24 h after transfection in medium supplemented with 10 μM anisomycin (Calbiochem) for 8 h. To inhibit caspase activity, cells were pre-incubated for 30 min with 20 μM zVAD-FMK (benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone) (Calbiochem) before treatment with staurosporin. Rabbit synovial fibroblasts (HIG-82) cells (CRL-1832; A.T.C.C., Manassas, VA, U.S.A.) were cultured in F12 medium (Invitrogen), supplemented with 10% FBS and 50 μg/ml gentamycin, between passage 7 and 14.
Jurkat and HEK-293 cells were lysed by sonication (10 s, 16 μm amplitude) in 100 μl of buffer [50 mM Tris/HCl (pH 8), 150 mM NaCl, 0.1 mM EDTA and Complete™ protease inhibitor cocktail (Roche Diagnostics)]. MDCK cells were suspended in 200 μl of lysis buffer (25 mM Hepes, pH 7.5, 100 mM NaCl, 1.5 mM MgCl2, 0.5 mM EGTA, 0.25 mM EDTA, 0.1% Nonidet P40, 10 mM NaF, 20 mM β-glycerophosphate, 1 mM Na3VO4, 1 mM PMSF, 10 μg/ml leupeptin and 10 μg/ml aprotinin). Total protein concentration was measured by colorimetry (Bio-Rad Laboratories assay) before Western blot analysis.
Expression vector construction
The construction of a pcDNA3 vector expressing human Ets-1 p51 fused to a FLAG tag is described in . The pcDNA3 vector expressing human Ets-1 p42 fused to a FLAG tag was obtained by the same strategy as the FLAG–Ets-1 p51 pcDNA3 expression vector, using the human Ets-1 p42 pSG5 expression vector  as a template. To produce Ets-1 p51 mutants, aspartate residues were replaced by asparagine residues at position 262 for D1N, positions 262, 287 and 290 for D3N and positions 262, 287, 290 and 313 for D4N using the QuikChange® site-directed mutagenesis kit (Stratagene). Briefly, the pcDNA3 vector expressing human Ets-1 p51 fused to a FLAG tag was used as a template for PCR amplification, using the following primers in which mutated codons are underlined: 5′-TACGATAGTTGTAATCGCCTCACCCAG-3′ and 5′-CTGGGTGAGGCGATTACAACTATCGTA-3′ for the D262N mutation; 5′-GACAGCTTCAACTCAGAGAACTATCCGGCTGCC-3′ and 5′- GGCAGCCGGATAGTTCTCTGAGTTGAAGCTGTC-3′ for the D287N/D290N mutations; and 5′-CGGGACCGTGCTAACCTCAATAAGGACAAG-3′ and 5′-CTTGTCCTTATTGAGGTTAGCACGGTCCCG-3′ for the D313N mutation.
To produce the Cp17 fragment, the sequence coding amino acids 291–441 of Ets-1 p51 was obtained by PCR amplification, using the primers 5′-AATCACTAAGATCTTATCCGGCTGCCCTGCCCAACC-3′ and 5′-TATGCGGCCGCTCACTCGTCGGCATCTGGCTTGAC-3′. The PCR product was then cloned between the BamHI and NotI restriction sites in a modified pcDNA3 expression vector (kindly provided by Dr D. Monté, Institut de Biologie de Lille, Lille, France). This vector contained the coding sequence for a FLAG epitope followed by a BamHI site. To produce the Np36 fragment, the sequence coding amino acids 1–262 of Ets-1 p51 was obtained by PCR amplification, using the primers 5′-ATTTAAAGATCTAAGGCGGCCGTCGATCTC-3′ and 5′-AATCACTAAGATCTATCACAACTATCGTAGCTCTC-3′. The PCR product was then cloned in the BamHI site of the modified pcDNA3 expression vector. For all vectors, correct insertion was checked by sequencing. The construction of the −478/+4 WT (wild-type) and EBS mutant human stromelysin-1 promoter pGL3 reporter vector (Promega) and the construction of the −607/+51 TORU [PMA (‘TPA’) oncogene response unit] mutant human collagenase1 promoter pGL2 reporter vector are described in [17,33]. The −607/+51 WT human collagenase-1 promoter pGL2 reporter vector as well as the c-Jun- and the c-Fos-encoding pSG5 vectors were kindly provided by Dr M. Duterque-Coquillaud (Institut de Biologie de Lille, Lille, France).
The construction of the bacterial expression vector containing the cDNA sequence of human Ets-1 p51 is described in .
HEK-293, HIG-82 or MDCK cells were grown in six-well plates to reach 40–60% confluence at the time of transfection. Transfection of HEK-293 and HIG-82 cells was performed by incubating the ExGen 500 transfection reagent (Euromedex), according to the manufacturer's instructions, with 1 μg of expression vectors. Transfection of MDCK cells was performed using the lipofection method .
In vitro caspase cleavage reactions
After 48 h of transfection, HEK-293 cells and Jurkat cells were lysed in 80 μl of caspase buffer [20 mM Pipes (pH 7.2), 100 mM NaCl, 1% CHAPS, 10% sucrose, 5 mM DTT (dithiothreitol) and 0.05 mM EDTA]. Cell lysates were incubated for 4 h at 37 °C with 1 μl of purified caspase. Reactions were stopped by incubation at 70 °C for 1 min in denaturating buffer (125 mM Tris/HCl, pH 6.8, 4% SDS, 20% glycerol and 100 mM DTT). Total cell lysate proteins were then analysed by Western blotting. Purified active caspases were generously provided by Dr G.S. Salvesen (Burnham Institute, La Jolla, CA, U.S.A.). Titration of the active recombinant caspases were obtained using in vitro cleavage against synthetic substrates as described previously (caspase 3, 13.9 μM; caspase 6, 7.5 μM; caspase 7, 19.7 μM; caspase 8, 15 μM; caspase 9, 41.5 μM) .
Transfection and luciferase reporter gene assay
HEK-293 cells were grown in 12-well plates to reach 40–60% confluence at the time of transfection. Transfections were then performed by incubating the ExGen 500 transfection reagent, according to the manufacturer's instructions, with 500 ng of vectors. Transfections were carried out with increasing amounts of the Cp17 or the Np36 expression vector and human stromelysin-1 or collagenase-1 promoter reporter vector in the presence or absence of the Ets-1 p51 expression vector. Cells were harvested 48 h after transfection with 200 μl of passive lysis buffer (Promega), and each supernatant was sequentially tested for firefly and Renilla luciferase activity (Dual-Luciferase® Reporter Assay System, Promega) using a Centro LB 960 (Berthold) luminometer. For each expression vector combination and concentration, firefly luciferase activity (pGL-2/3 constructs) was normalized to the Renilla activity [pRL-null normalization vector (Promega)] to correct for variation in the number of transfected cells. Expression of proteins of interest was tested by Western blot analysis.
Western blot analysis
Total cell lysate proteins were boiled in Laemmli buffer (50 mM Tris, pH 6.8, 2% SDS, 5% 2-mercaptoethanol, 10% glycerol, 0.1% Bromophenol Blue) and resolved by SDS/PAGE. Proteins in gels were transferred on to a Hybond™-C Extra membrane (Amersham Biosciences) and blocked for 1 h at room temperature (25 °C) in 5% (w/v) non-fat dried skimmed milk powder in PBS. The membrane was then incubated with the primary antibody for 1 h at room temperature in blocking buffer. Ets-1 was detected with the C-20 polyclonal antibody directed against the C-terminal region or the N-276 polyclonal antibody directed against the N-terminal region (Santa Cruz Biotechnology), Erk-2 with the C-14 polyclonal antibody (Santa Cruz Biotechnology), cleaved caspase 3 with the 5A1 polyclonal antibody (Cell Signaling Technology), c-Jun with the H-79 polyclonal antibody (Santa Cruz Biotechnology) and c-Fos with the H-125 polyclonal antibody (Santa Cruz Biotechnology). The washed membrane was then incubated for 1 h at room temperature with HRP (horseradish peroxidase)-conjugated secondary antibody (Santa Cruz Biotechnology) in blocking buffer. Bound antibodies were visualized using the Western Lightning™ chemiluminescence detection system (PerkinElmer Life Sciences Biotechnology). For quantification of protein expression, luminescence was recorded using a digital camera with a cooled charge-coupled device (LAS 3000; Fuji) and quantified using the Multi Gauge V3.0 image analysis software (Fuji Film). Reprobing was performed using the Re-Blot Plus Strong Antibody Stripping Solution (Millipore) according to the manufacturer's instructions.
EMSA (electrophoretic mobility-shift assay)
Double-stranded synthetic oligonucleotides corresponding to the WT and EBS mutant of the human stromelysin-1 (−223/−193) promoter region were end-labelled using T4 polynucleotide kinase and [γ-32P]ATP, and were subsequently purified by electrophoresis on a 20% polyacrylamide non-denaturing gel in TBE buffer (90 mM Tris/borate and 1 mM EDTA). Cell lysates prepared (30 μg) from HEK-293 cells transfected with the expression vector, empty or encoding the Cp17 fragment, or with recombinant Ets-1 p51 protein (1 μg) were incubated with 0.5 ng of labelled probe in 40 μl of binding reaction buffer [20 mM Tris/HCl (pH 7.9), 80 mM NaCl, 1 mM EDTA, 2 mM DTT, 40 μg/ml poly(dI-dC)·(dI-dC) (GE Healthcare) and 10% glycerol] for 30 min on ice. Complexes formed were resolved on a 5% polyacrylamide (29:1 acrylamide/bisacrylamide; Euromedex) non-denaturing gel in 0.25× TBE buffer at room temperature. Gels were dried and autoradiographed at −80 °C. For probe competition, 400 ng of non-labelled probe (800×) was added to the reaction mixture. For protein competition, cell lysates (30 μg) prepared from HEK-293 cells transfected with increasing amounts of the Cp17 expression vector were incubated with recombinant Ets-1 p51 protein (0.5 μg) with 0.5 ng of WT-labelled probe under the same conditions.
Expression and purification of Ets-1 p51 protein
Ets-1 p51 was expressed and purified using the T7-Impact™ System (New England Biolabs) as described previously .
RNA extraction and RT (reverse transcription)–PCR assays
RNA was extracted from HIG-82 cells using TRIzol® (Invitrogen), according to the manufacturer's instructions. RT was performed with random hexameric oligonucleotides with 1 μg of cDNA using the RevertAid™ First Strand cDNA kit (Fermentas). PCRs were performed with the High Fidelity PCR Master Mix (Roche) with 500 ng of reverse-transcribed RNAs and with 125 ng of forward and reverse primers for each of four genes: human Ets-1, 5′-ATGAAGGCGGCCGTCGATCTC-3′ and 5′-TTGGTCCACTGCCTGTGTAG-3′; GAPDH (glyceraldehyde-3-phosphate dehydrogenase), 5′-ATGAGGTCCACCACCCTGTT-3′ and 5′-ATCACTGCCACCCAGAAGAC-3′; gelatinase-B, 5′- CTGGGCAAGGGCGTCGTGGTC-3′ and 5′-CGTGGTGCAGGCGGTGTAGGAG-3′; and uPA (urokinase plasminogen activator), 5′-CCATCCCGGTCCATACAGACT-3′ and 5′- TCACAGCTTGTGCCAAAATTG-3′. PCRs were run for 35 cycles with 30 s of denaturation at 94 °C, 30 s of hybridization at 55 °C and 30 s of elongation at 72 °C. Amplified products were analysed on 2% agarose gels run in 0.5× TBE buffer and stained with ethidium bromide.
Ets-1 is a substrate for caspases
To investigate the fate of Ets-1 during programmed cell death, Jurkat T-leukaemia cells, which endogenously express Ets-1 p51 and Ets-1 p42 isoforms, were treated with the apoptotic inducer staurosporin (Figure 1A). This stress stimulus was found to activate caspase 3 with detection by Western blot analysis of the specific p19 and p17 caspase 3 fragments (Figure 1A, lane 2). This effect was associated with a decrease in Ets-1 levels and generation of three new Ets-1 fragments of approx. 14, 17 and 20 kDa in apparent molecular mass, designated as Cp14, Cp17 and Cp20, with Cp17 being the predominant fragment. These fragments were detected using an anti-Ets-1 antibody directed against its DBD located in the C-terminal region. Using an anti-Ets-1 antibody directed against its Pointed domain located in the N-terminal region, the same blot revealed a new Ets-1 fragment of approx. 36 kDa in apparent molecular mass, designated as Np36 (Figure 1A, lane 2). This suggests that Ets-1 factor was cleaved, leading to generation of three C-terminal and one N-terminal fragments. Furthermore, Ets-1 processing was abrogated by the addition of the potent general caspase inhibitor zVAD-FMK to Jurkat cells medium before staurosporin treatment (Figure 1A, lanes 3 and 4), demonstrating that caspases are involved in Ets-1 cleavage.
Caspase cleavage of Ets-1
To confirm that caspases are able to process the Ets-1 transcription factor, Jurkat cell extracts were exposed to recombinant active caspase 3 and analysed by Western blotting using the same antibodies as in the previous experiment (Figure 1B). Results showed that recombinant active caspase 3 reduced endogenous Ets-1 levels (Figure 1B, lanes 1 and 2) and produced the same fragments as those generated after staurosporin treatment (compare Figure 1A, lane 2 with Figure 1B, lane 2), i.e. the Cp14, Cp17 and Cp20 C-terminal and the Np36 N-terminal fragments. These results demonstrate unequivocally that the Ets-1 protein is processed by caspases during apoptosis.
Ets-1 p51, but not Ets-1 p42, is cleaved in vitro by recombinant active caspase 3
To determine whether one or both Ets-1 isoforms constitute caspase 3 substrates, Ets-1 p51 or Ets-1 p42 were exogenously expressed in HEK-293 cells, which do not encode Ets-1. Cell lysates were then exposed to recombinant active caspase 3 (Figure 2A). Western blot analysis revealed that exogenous Ets-1 p51 levels were drastically reduced after incubation with recombinant active caspase 3 (Figure 2A, lanes 3 and 4), whereas exogenous Ets-1 p42 levels were not affected (Figure 2A, lanes 5 and 6). Furthermore, the decrease in Ets-1 p51 intensities was associated with the production of the Cp14, Cp17, Cp20 and Np36 fragments (Figure 2A, lane 4). Thus Ets-1 p51, but not Ets-1 p42, is a substrate of recombinant active caspase 3 in vitro.
Ets-1 p51, but not Ets-1 p42, is cleaved by caspase 3 in vitro
The same experiment was performed using other recombinant active caspases, in particular caspase 6, 7, 8 and 9 (Figure 2B). Results showed that among the tested proteases, the caspase 3 was the only one that cleaved exogenous Ets-1 p51 (Figure 2B, lanes 2–7), whereas Ets-1 p42 was not cleaved by caspase 3 or any other caspase (Figure 2B, lanes 8–13).
Ets-1 p51, but not Ets-1 p42, is cleaved in cells undergoing apoptosis
We then checked the consequences of apoptotic signals on the cleavage of Ets-1 p51 and Ets-1 p42 by caspases. Ets-1 p51 and Ets-1 p42 were expressed in HEK-293 cells, in the presence or absence of the apoptotic inducer staurosporin (Figure 3A). Western blot analysis of cell lysates showed that the three C-terminal cleavage fragments were generated in cells expressing Ets-1 p51 (Figure 3A, lanes 5 and 6). Furthermore, the incubation of cells with zVAD-FMK abrogated the production of the cleavage fragments (Figure 3A, lanes 7 and 8). In contrast, the exogenous expression of Ets-1 p42 in HEK-293 cells did not generate any proteolytic fragments (Figure 3A, lanes 9–12). Taken together, these results show that exogenous Ets-1 p51, but not Ets-1 p42, is cleaved by caspases in apoptotic HEK-293 cells.
Ets-1 p51, but not Ets-1 p42, is cleaved in HEK-293 and Jurkat cells undergoing apoptosis
To determine whether the differential caspase cleavage of the two Ets-1 isoforms could occur under endogenous conditions, apoptosis of Jurkat cells was induced by staurosporin treatment and cell lysates were analysed by Western blotting at several time intervals (Figure 3B). Upon induction of apoptosis, the three C-terminal fragments were generated as expected. The level of the Cp20 fragment, detected after 2 h, declined over the 8 h time course, suggesting that it may be degraded. In contrast, the Cp17 and Cp14 fragments were stable over time. Along with the generation of fragments, Ets-1 p51 levels simultaneously greatly decreased in a time-dependent manner, whereas Ets-1 p42 levels declined only weakly over the same period. The quantification of Ets-1 isoform levels in three independent experiments (Figure 3C) confirmed that the amount of Ets-1 p51 decreased approx. 60% during the 8 h time course experiment, whereas Ets-1 p42 level remained unchanged. Furthermore, the quantification of Cp17, the main fragment, showed that it was stable over time (Figure 3C). Incubating cells with zVAD-FMK before staurosporin treatment abrogated the reduction of Ets-1 p51 levels (Figure 3B, lane 5, and Figure 3C), as well as the formation of cleavage fragments (Figure 3B, lane 5). Our results demonstrate that, in Jurkat cells undergoing apoptosis, Ets-1 p51, unlike Ets-1 p42, is down-regulated. Taken together, this suggests that caspase cleavage of Ets-1 p51 during apoptosis contributes to the efficient degradation of Ets-1 and generates fragments of Ets-1.
Identification of caspase cleavage sites in Ets-1 p51
The cleavage motif for caspases is a sequence of four amino acids, numbered P4–P1, which is processed after the invariant P1 aspartate residue. Asp-Xaa-Xaa-Asp is a cleavage motif present in many caspase 3 substrates [36,37]. The amino acid sequence of human Ets-1 p51 has three putative cleavage sites for caspases in the exon VII-encoded region, DSCD262, a double site DSFD287SED290 and DRAD313 (Figure 4A, a). The position of these sites is consistent with the generation of three C-terminal fragments of 20, 17 and 14 kDa (Figure 4A, b). In addition, the location of these sites in the exon VII-encoded region is consistent with the resistance of Ets-1 p42 to caspase cleavage.
Identification of the caspase cleavage sites in Ets-1 p51 processing
To determine whether Ets-1 p51 is cleaved by caspases at these sites, site-directed mutagenesis was performed: (i) on the first putative caspase site in which Asp262 was replaced by asparagine (D1N); (ii) on the two first putative caspase sites in which Asp262 and Asp287/Asp290 were mutated to asparagine (D3N); and (iii) on all four of the caspase sites of the exon VII-encoded region (D4N) (Figure 4A, c).
After the exogenous expression of WT or mutant (D1N, D3N, D4N) Ets-1 p51 in HEK-293 cells, lysates were exposed to recombinant, active caspase 3 and analysed by Western blotting (Figure 4B). In contrast with WT Ets-1 p51 (Figure 4B, lane 2), all three mutants were resistant to caspase 3 cleavage (Figure 4B, lanes 4, 6 and 8). Thus DSCD262 was an essential site for in vitro caspase 3 cleavage because its mutation abrogates the processing at the other sites.
Secondly, we assessed the role of these sites for Ets-1 p51 processing in MDCK cells undergoing apoptosis. MDCK cells exogenously expressing WT or mutant Ets-1 p51 were incubated with or without anisomycin, and cell lysates were analysed by Western blotting (Figure 4C). As expected, WT Ets-1 p51 generated the C-terminal fragments with Cp17 predominating (Figure 4C, lane 2). It is noteworthy that Ets-1 fragments occurred in the absence of any apoptotic inducer; this may be due to stress caused by transfection conditions (Figure 4C, lane 3), as reported previously for MDCK cells [38,39]. The D1N mutant abrogated Cp20 generation, while Cp17 and Cp14 were still detectable (Figure 4C, lane 4), demonstrating that Cp20 was generated through cleavage on Asp262. The D3N mutant abrogated generation of all the fragments (Figure 4C, lane 6), demonstrating that cleavage on DSFD287SED290 is necessary for generation of the main Cp17 fragment and the secondary Cp14 fragment. Taken together, these results demonstrate that Ets-1 p51 was cleaved at caspase sites with aspartic acid located at positions 262, 290 and 313, allowing generation of the three C-terminal fragments Cp20, Cp17 and Cp14 respectively. In addition, DSCD262 and DSFDSED290 sites are critical for cleavage in vitro and during apoptosis respectively.
Cp17 binds to the EBS palindrome of the stromelysin-1 promoter and competes with Ets-1 p51 for binding
Caspase cleavage during apoptosis generated Cp17, the major C-terminal fragment that conserves the DBD surrounded by the two inhibitory domains (Figure 5A). This suggests that Cp17 would be able to bind DNA as Ets-1 p51 does. To test this hypothesis, the nucleotide sequence of Cp17, which codes for amino acids 291–441 of Ets-1 p51 (Figure 5A), was cloned into a eukaryotic expression vector and used in an EMSA.
Cp17 binds to the EBS palindrome of the stromelysin-1 promoter and competes with Ets-1 p51 for binding
Assays were carried out using: (i) lysates from HEK-293 cells transfected with empty vector; (ii) lysates from HEK293 cells transfected with Cp17 expression vector; or (iii) Ets-1 p51 recombinant protein. Proteins were incubated with 30-mer 32P-labelled DNA probes containing WT or mutant EBS palindromes of the stromelysin-1 promoter (Figure 5B). Results showed that Ets-1 p51 formed a complex with the WT promoter (Figure 5B, lane 5, arrow 3). This may correspond to a ternary complex with two protein molecules bound to the WT EBS palindrome, as reported previously . Two complexes were detected with lysates from cells transfected with Cp17. The slower migrating band was probably a ternary complex (Figure 5B, lane 9, arrow 3′) because we detected a smaller binding species, which may correspond to a binary complex formed by one protein molecule bound to the stromelysin-1 promoter (Figure 5B, lane 9, arrow 2′). Mutation of the EBS palindrome prevented both Ets-1 p51 and Cp17 from forming DNA–protein complexes (Figure 5B, lanes 6 and 10). These complexes were disrupted by an excess of non-labelled WT DNA probe (Figure 5B, lanes 7 and 11), but not by an excess of non-labelled double-mutant DNA probe (Figure 5B, lanes 8 and 12), confirming that these complexes were specific to the EBS palindrome.
Next, we assessed the ability of Cp17 to compete with Ets-1 p51 for DNA binding. EMSAs were carried out using a 30-mer 32P-labelled WT DNA probe and increasing concentrations of Cp17 exogenously expressed in HEK-293 cells, with or without recombinant Ets-1 p51 (Figure 5C). Results showed that the Ets-1 p51–DNA–Ets-1 p51 ternary complex (Figure 5C, lane 6, arrow 3) was disrupted in the presence of Cp17 (Figure 5C, lanes 7–9). This disruption was accompanied by the formation of intermediate complexes (indicated by arrow i) with lower molecular masses, which may be composed of both Ets-1 p51 and Cp17. Thus the Cp17 cleavage fragment competes with Ets-1 p51 for binding to the EBS palindrome in the stromelysin-1 promoter.
Cp17 cleavage fragment is dominant-negative for Ets-1 p51 transactivation
The caspase cleavage within the exon VII-encoded region separates the N-terminal TAD (transactivation domain) from the C-terminal DBD. This suggests that Cp17 would be dominant-negative for Ets-1 p51-mediated transcription. To test this hypothesis, luciferase transactivation assays were performed with the WT or EBS mutant stromelysin-1 promoter using increasing concentrations of Cp17 with or without Ets-1 p51 (Figure 6A). Cp17 was unable to substantially activate the stromelysin-1 promoter on its own (Figure 6A left-hand panel, lanes 1–4). Nevertheless, it repressed in a dose-dependent manner the Ets-1 p51-induced activation of the WT promoter (Figure 6A left-hand panel, lanes 5–8), with more than 60% suppression when present at a 1:1 ratio with Ets-1 p51 (Figure 6A left-hand panel, lane 7). This was mediated through the EBS palindrome because mutation of both EBSs abrogated any Cp17 effects (Figure 6A middle panel). Protein expression was confirmed by Western blot analysis for Ets-1 p51, Cp17 and Erk-2 as a loading control (Figure 6A blot). Thus Cp17 acts as dominant-negative for Ets-1 p51 transactivation of the stromelysin-1 promoter mediated through the EBS palindrome.
Cp17, but not Np36, acts as dominant-negative for transactivation of the stromelysin-1 and collagenase-1 promoters mediated by Ets-1 p51
Secondly, we examined the effect of Cp17 on the collagenase-1 matrix metalloproteinase promoter, another Ets-1 target gene, which is synergistically activated by Ets-1 p51 and the c-Jun–c-Fos [AP-1 (activator protein 1)] complex, through the TORU, composed of an EBS adjacent to an AP-1 site . Luciferase transactivation assays were performed with the collagenase-1 promoter (WT or EBS/AP-1 mutant) using increasing concentrations of Cp17 with or without Ets-1 p51 and the c-Jun–c-Fos complex (Figure 6B). Cp17 repressed, in a dose-dependent manner, the synergistic activation of the collagenase-1 promoter induced by Ets-1 p51 and the c-Jun–c-Fos complex (Figure 6B left-hand panel). This was mediated through the TORU, since its mutation abrogated any Cp17 effects (Figure 6B middle panel). Thus Cp17 acts also as dominant-negative for Ets-1 p51 transactivation of the collagenase-1 promoter mediated through the TORU.
We then checked the transcriptional response of the N-terminal Np36 fragment. The nucleotide sequence, which codes for amino acids 1–262 of Ets-1 p51, was cloned in a eukaryotic expression vector and used in luciferase transactivation assays. Results show that Np36 does not inhibit the Ets-1 p51-induced transcriptional activation of the stromelysin-1 and collagenase-1 promoters, owing to the absence of the DBD (Figure 6D). Thus, in contrast with Cp17 fragment, the Np36 fragment is not able to inhibit the transcriptional activity of Ets-1 p51.
Cp17 decreases the Ets-1 p51 target genes expression
Given the dominant-negative function of Cp17, we explored its effect on the expression of endogenous Ets-1 p51 target genes. To this aim, we performed RT–PCR analysis on HIG-82 rabbit synovial fibroblasts, which endogenously express the Ets-1 protein and represent a good model to study the transcriptional regulation of Ets-1 target genes . We transfected these cells with an empty vector, with a human Ets-1 p51 expression vector or with both human Ets-1 p51 and Cp17 expression vectors. Results showed that the overexpression of human Ets-1 p51 alone induced an increase in the mRNA levels of well-known Ets-1 target genes, such as gelatinase-B and uPA (Figure 7; compare lanes 2 and 3), indicating that Ets-1 p51 up-regulated the expression of these genes. However, the overexpression of both Ets-1 p51 and Cp17 decreased the mRNA levels of these target genes (Figure 7, lane 4). This shows that Cp17 inhibits the expression of endogenous Ets-1 target genes, confirming its dominant-negative function on Ets-1 p51.
Cp17 decreases Ets-1 p51 target gene expression
Although previous studies have described the down-regulation of the Ets-1 p51 protein in cells undergoing apoptosis [31,32], the molecular mechanism underlying this effect was not known. In the present study, we demonstrated that the down-regulation of Ets-1 p51 protein in Jurkat T-cells undergoing apoptosis is a consequence of its caspase-dependent cleavage. Caspase cleavage generated three C-terminal fragments, Cp14, Cp17 and Cp20, and one detectable N-terminal fragment, Np36. Screening the amino acid sequence of Ets-1 for putative caspase cleavage sites made it possible to identify three consensus sites (DXXD) located in the exon VII-encoded region: DSCD262, a double site DSFD287SED290 and DRAD313. Mutational analysis indicated that Ets-1 p51 processing occurs at these sites, thus generating the Cp20, Cp17 and Cp14 fragments respectively. Cp17 was the major fragment generated during apoptosis, suggesting that DSFD287SED290 is the main cleavage site. The N-terminal Np36 fragment detected during apoptosis may be the N-terminal counterpart of the C-terminal Cp20 fragment generated through cleavage at the DSCD262 site. Consistent with the location of the caspase sites in the alternative exon VII-encoded region, Ets-1 p42 was not processed by caspases in in vitro cleavage experiments, nor during apoptosis.
The exon VII-encoded region of Ets-1 is a regulatory domain described as (i) a regulator of DNA-binding [16,17,41], (ii) a mediator of protein–protein interactions , and (iii) a target of calcium-mediated phosphorylation [43,44]. The cleavage of Ets-1 p51 and not Ets-1 p42 allocates a new function to the exon VII-encoded region, thus expanding on the functional differences between the two Ets-1 isoforms.
Cleavage of Ets-1 p51 in the exon VII-encoded region separate the TAD, crucial for transcriptional activity, from the rest of the molecule, but leave the DBD intact. Therefore C-terminal Ets-1 fragments containing the DBD should be able to bind to Ets-1 p51 target promoters without activating transcription, suggesting that they have a dominant-negative function. Our results confirmed this hypothesis for the major cleavage product, Cp17. Its DNA-binding and transcriptional properties were assessed on the EBS palindrome of the stromelysin-1 promoter, which has been shown to be an important target gene for Ets-1 and is involved in the process of tumour invasion and metastasis . Our results demonstrated that Cp17 competes with Ets-1 p51 for binding to the EBS palindrome of the stromelysin-1 promoter, demonstrating its dominant-negative effect on Ets-1 p51 binding. This competition involves the formation of an intermediate complex, suggesting that both proteins interact with each other on the EBS palindrome of the stromelysin-1 promoter. Cp17 forms also a binary complex with the DNA, which may represent a transient complex facilitating the binding of a second Cp17 molecule to form a ternary complex. This is consistent with previous work that showed that deleting the N-terminal region of Ets-1 p51 spanning amino acids 1–300 does not alter the binding stoichiometry for the EBS palindrome found in the stromelysin-1 promoter [17,18]. It is noteworthy that we did not detect any binary complexes involving Ets-1 p51 because full-length Ets-1 p51 more readily forms ternary complexes. Furthermore, Cp17 inhibited Ets-1 p51 transactivation of the stromelysin-1 promoter. This effect may be the result of Cp17 binding to the palindromic EBS, which then becomes inaccessible to Ets-1 p51. Cp17 also repressed the Ets-1 p51-mediated transactivation of the collagenase-1 promoter. This may result from the ability of Cp17 to replace Ets-1 p51 in the interaction with the c-Fos–c-Jun complex, required for combinatorial regulation. This is supported by the fact that the interaction of Ets-1 p51 with the c-Fos–c-Jun complex is mediated through the DBD, which is conserved in Cp17 . The N-terminal Np36 fragment did not have any effect on the Ets-1 p51-induced transcriptional activation of either the stromelysin-1 or the collagenase-1 promoters. This is consistent with the inability of Np36 to bind DNA because it lacks the DBD. This confirms that the dominant-negative effect of Cp17 effectively results from its binding to the Ets-1 p51 target promoters. Finally, Cp17 decreased Ets-1-induced expression of uPA and gelatinase-B, two well-known target genes of Ets-1, confirming the dominant-negative function of Cp17 on Ets-1 p51.
Ets-1 functions have largely been studied through the use of artificial dominant-negative constructs made up of the DBD of the molecule [23,46]. These constructs have helped to demonstrate, for instance, the involvement of Ets-1 in tumorigenesis and angiogenesis [46–48]. The Cp17 fragment represents a natural dominant-negative form generated by proteolytic cleavage, able to inhibit the transactivation response induced by the full-length protein. However, in contrast with artificial DBD constructs, Cp17 possesses the two inhibitory domains flanking the DBD. Cp17 is thereby more specific, particularly for promoters that have either palindromic EBS or adjacent DNA elements and that can be activated only by overriding Ets-1 autoinhibition.
Along the same lines, we demonstrated recently that a novel Ets-1 isoform, p27, which lacks the Pointed domain and the TADs, but conserves the DBD flanked by the two inhibitory domains, also acts dominant-negatively for Ets-1 p51 . The overexpression of this novel isoform in invasive mammary carcinoma cells represses proliferation, transformation and invasion, and reduces the growth of tumour xenografts in nude mice. The splice variant Ets-1 p27 and the Ets-1 p51 cleavage product Cp17, both of which possess the DBD and the inhibitory domains, represent natural dominant-negative forms of the full-length protein generated by two different mechanisms.
The caspase-mediated generation of dominant-negative fragments has also been described for several other transcription factors. For instance, NF-κB p65/RelA subunit  and SRF  provide two examples of transcription factors in which caspase cleavage leads to separation of the TAD from the DBD. These truncated proteins are transcriptionally inactive and also act as dominant-negative inhibitors [6,8].
Since some studies have shown that Ets-1 plays a role in apoptosis through the control of gene expression critical to this process, we can assume that down-regulation of Ets-1 p51 during apoptosis could prevent this transcriptional programme. Following this line of reasoning, the dominant-negative function of Cp17 may provide an additional mechanism to amplify the cell death process by inhibiting the Ets-1 p51-mediated expression of anti-apoptotic proteins. This would be consistent with a previous study demonstrating that artificial Ets dominant-negative forms suppress Ets transcriptional activity and induce programmed cell death, mediated by lower levels of c-Myc expression, in thyroid carcinoma cell lines . Although the anti-apoptotic target genes of Ets-1 p51 and their promoters are currently not well characterized, it would interesting to assess the dominant-negative properties of Cp17 on these target genes. In addition, we demonstrated that caspase cleavage targets Ets-1 p51 without affecting the Ets-1 p42 isoform. Consequently, during apoptosis, the ratio between the two isoforms is modified in favour of the Ets-1 p42 isoform. Interestingly, it has been shown that, compared with Ets-1 p51, Ets-1 p42 is preferentially associated with the induction of apoptosis. Therefore the modification of the ratio between the two isoforms may be a mechanism favouring cell death through the specific preservation of the pro-apoptotic isoform.
Taken together, our data suggest a novel mechanism of Ets-1 p51 regulation through caspase-mediated cleavage and generation of a dominant-negative fragment, which may play an active role during apoptosis.
activator protein 1
Dulbecco's modified Eagle's medium
electrophoretic mobility-shift assay
fetal bovine serum
human embryonic kidney
Madin–Darby canine kidney
nuclear factor κB
PMA (‘TPA’) oncogene response unit
urokinase plasminogen activator
Souhaila Choul-li contributed to the apoptosis induction of Jurkat cells, the luciferase reporter gene assays, the EMSAs, the RNA extraction, the RT–PCR assays and the construction of Cp17 and Np36 expression vectors. Catherine Leroy contributed to the in vitro caspase cleavage reactions and the apoptosis induction of transfected cells. Gabriel Leprivier contributed to the construction of Ets-1 p51 mutant expression vectors and experiments performed with Ets-1 p51 mutants. Clélia Laitem contributed to cell culture, preparation of cell lysates and expression and purification of recombinant Ets-1 p51 protein. David Tulasne and Marc Aumercier contributed to the design of the study, the interpretation of the results and the writing of the paper.
We warmly thank Dr M. Duterque-Coquillaud for generously providing the human collagenase-1 promoter as well as the pSG5-c-Fos, and pSG5-c-Jun vectors and Dr D. Monté for providing the pcDNA3-Flag vector. We are grateful to I. Roland for technical assistance.
This work was supported by the Centre National de la Recherche Scientifique (CNRS) and by grants from the Ligue contre le Cancer-Comité Pas-de-Calais, the Fondation pour la Recherche Médicale-Comité Nord-Pas-de-Calais (to M.A. and D.T.), the Association pour la Recherche sur le Cancer and Agence Nationale de la Recherche-Young Investigator Program (to D.T.). A Ph.D. fellowship was provided to S.C. by the CNRS and the Conseil Régional Nord-Pas-de-Calais (BDI: Bourse de Docteur-Ingénieur), to G.L. by the Ligue Nationale contre le Cancer and to C.L. by the French Ministère de la Recherche et de l'Enseignement Supérieur.
Present address: Department of Molecular Oncology, British Columbia Cancer Research Centre, Vancouver, British Columbia, Canada, V5Z 1L3
Present address: Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, U.K.