The LIM-Only Protein FHL2 Is a Negative Regulator of E4F1

C Paul1, M Lacroix1, I Iankova1,4, E Julien1, BW Scha¨ fer2, C Labalette3, Y Wei3, A Le Cam1, L Le Cam1 and C Sardet1

1Institut de Ge´ne´tique Moleculaire, UMR 5535/IFR122, CNRS, Montpellier, France; 2Division of Clinical Chemistry and Biochemistry, Department of Pediatrics, University of Zu¨rich, Zu¨rich, Switzerland and 3Unite´ d’Oncogene`se et Virologie Mole´culaire, Institut Pasteur, Paris, France

The E1A-targeted transcription factor E4F1 is a key out by a GLI/Kruppel-related 50 kDa transcription player in the control of mammalian embryonic and factor (p50E4F1) generated, upon an E1A-activated somatic cell proliferation and survival. Mouse embryos proteolytic cleavage, from the ubiquitously expressed lacking E4F die at an early developmental stage, whereas 120 kDa protein (p120E4F1) (Fernandes et al., 1997) enforced expression of E4F1 in various cell lines inhibits previously identified in mouse as PhiAP3(Fognani cell cycle progression. E4F1-antiproliferative effects have et al., 1993). Although both p50E4F1 and p120E4F1 been shown to depend on its capacity to repress recognize the same DNA motives in vitro, they appear transcription and to interact with pRb and p53. Here we to differentially regulate gene expression in vivo. p50E4F1 show that full-length E4F1 protein (p120E4F1) but not its transactivates expression of the adenoviral E4 gene E1A-activated and truncated form (p50E4F1), interacts through multiple CRE-like elements in an E1A-depen- directly in vitro and in vivo with the LIM-only protein dent fashion (Fernandes et al., 1997). On the contrary, FHL2, the product of the p53-responsive gene FHL2/ p120E4F1, at least when overexpressed, is likely to DRAL (downregulated in rhabdomyosarcoma Lim pro- play a role of repressor for the adenoviral E4 and E1A tein). This E4F1-FHL2 association occurs in the nuclear promoters (Fognani et al., 1993; Fernandes et al., 1997) compartment and inhibits the capacity of E4F1 to block and for the cellular cyclin A2 gene (Fajas et al., 2001; cell proliferation. Consistent with this effect, ectopic Tessari et al., 2003; Ahmed-Choudhury et al., 2005), a expression of FHL2 inhibits E4F1 repressive effects on property that might rely on its direct interaction with transcription and correlates with a reduction of nuclear histone deacetylase 1 (Colombo et al., 2003). E4F1-p53 complexes. Overall, these results suggest that The precise function of E4F1 in normal cell physiol- FHL2/DRAL is an inhibitor of E4F1 activity. Finally, we ogy remains largely unknown. Nevertheless, several show that endogenous E4F1-FHL2 complexes form in recent observations demonstrate it is a key player in the U2OS cells upon UV-light-induced nuclear accumulation control of mammalian embryonic and somatic cell of FHL2. proliferation and survival. Indeed, embryos lacking Oncogene (2006) 25, 5475–5484. doi:10.1038/sj.onc.1209567; E4F1 die at the peri-implantation stage, while in vitro published online 1 May 2006 cultured E4FÀ/À blastocysts exhibit severe defects in mitotic progression, chromosomal missegregation and Keywords: FHL2; DRAL; E4F1; p53; proliferation; increased apoptosis (Le Cam et al., 2004). Conversely, repression enforced expression of p120E4F1 in various cell lines inhibits progression from G1 to S phase (Fernandes et al., 1998; Fajas et al., 2000, 2001; Sandy et al., 2000; Rizos et al., 2003; Fenton et al., 2004) and that of p50E4F1 enhances cell death in E1A/ras-transformed cells Introduction (Fernandes et al., 1999). Interestingly, the E4F1-induced G1 arrest appears to depend, at least in part, on pRB The mammalian transcriptional activity termed E4F was and p53pathways (Fajas et al., 2000; Sandy et al., 2000; first described as one of the cellular targets of the viral Rizos et al., 2003). Consistent with this, E4F1 was found oncoprotein E1A, which activation is required for full to interact directly in vivo with pRB (Fajas et al., 2000), expression of the viral E4 promoter during adenovirus p53(Sandy et al., 2000) and p19ARF (Rizos et al., 2003) infection (Raychaudhuri et al., 1987). This E4F activity and to induce the stabilization of the p21WAF1 protein was purified from human cells and shown to be carried (Fernandes et al., 1998). This suggests that E4F1 is somehow involved in both the pRB and p53pathways. Correspondence: Dr C Sardet, Institut de Ge´ ne´ tique Moleculaire, To address this question, we searched for novel E4F1- UMR 5535/IFR122, CNRS, 1919 Route de Mende, 34293, Montpel- interactors/regulators that might be involved in the lier cedex 5, France. pRB/p53pathways. Here, we show that p120 E4F1 E-mail: [email protected] 4Current address: INSERM U540, 34090 Montpellier, France. interacts with the product of the p53-responsive gene Received 3August 2005; revised 30January 2006; accepted 28 February downregulated in rhabdomyosarcoma Lim protein 2006; published online 1 May 2006 (DRAL)/FHL2. This gene was originally identified as LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5476 the product of the DRAL gene, which expression is E4F1 repressive effects on transcription. Association of downregulated in human rhabdomyosarcoma cells endogenous E4F1 and FHL2 proteins in U2OS is as compared with nonmalignant control myoblasts stimulated by UV, a signal that stimulates expression (Genini et al., 1997), suggesting a role in tumor and nuclear translocation of FHL2 in these cells. development. DARL/FHL2 is strongly expressed in cardiac and skeletal muscle cells and, at lower levels, in many other tissues, cell types and established cell lines (Chan et al., Results 1998; Chu et al., 2000a, b; Scholl et al., 2000; Tanahashi and Tabira, 2000; Kong et al., 2001; Muller et al., 2002; Identification of DARL/FHL2 as a p120E4F1 interacting Gabriel et al., 2004; Kang et al., 2003). It encodes a protein in vitro and in a cellular context member of the FHL protein family (FHL1-4, ACT), a To identify novel E4F1 interaction partners potentially subclass of the LIM-only protein family, that contains involved in the pRB/p53pathways, we performed a four and an half copies of the cysteine-rich LIM motif yeast two-hybrid screen using Gal4DB-p120E4F1 as a bait (CX2CX16-23HX2CX2CX2CX16-21CX2-3(C/H/D), a to probe a library of Gal4AD-cDNA preys generated widely represented protein domain which functions as from WI38 human primary fibroblasts (Sardet et al., an adapter or modifier in protein interactions (Retaux 1995). Several classes of clones were selected, one and Bachy, 2002). Although the cellular functions of corresponding to the p53-responsive gene DRAL/ these FHL proteins remain largely unknown, recent FHL2. This LIM-only protein was represented by five studies suggest that they may act as transcriptional independent positive clones containing Gal4-DRAL/ coregulators and mediators in signal transduction. Thus, FHL2 cDNAs (Figure 1a). of various lengths. The upon ectopic expression in reporter assays, FHL2 longest fusion corresponding to full-length FHL2 and behaves as a coactivator for several transcription the shortest one to a N-ter truncated version of the factors, including AP-1 (Morlon and Sassone-Corsi, protein that contained the last two LIM domains (data 2003), CREB (Fimia et al., 2000), androgen receptor not shown). (Muller et al., 2000), NF-kB (Stilo et al., 2002), WT1 To confirm this interaction in vitro, we performed (Du et al., 2002) and as a corepressor for the glutathione S-transferase (GST) pull-down assays by Promyelocytic leukemia zinc-finger protein (PLZF) mixing in vitro translated methionine-labeled p120E4F (McLoughlin et al., 2002). Interestingly, DRAL/FHL2 and either GST-FHL2 or the control fusion proteins, was also described to directly associates with proteins GST and GST-ELK. As shown in Figure 1b, labeled involved in signal transduction, such as IGFBP-5 E4F1 bound specifically to GST-FHL2 beads and not (Amaar et al., 2002), b-catenin (Martin et al., 2002; on control beads in this assay. Notably, dialysis of Wei et al., 2003; Labalette et al., 2004), a- and b- bacterial FHL2 in a buffer containing ZnCl2 was integrins (Wixler et al., 2000) and the map kinase erk2 required in order to get optimum E4F1-FHL2 interac- (Purcell et al., 2004), suggesting that it might function as tion (data not shown), suggesting that generation of a molecular transmitter linking various signaling path- GST-FHL2 in a zinc-finger LIM conformation is, at ways to transcriptional regulation. Consistent with this least in part, important for this interaction. scenario, DRAL/FHL2 cellular localization seems to be To document this association in a cellular context, we both nuclear and non-nuclear (Scholl et al., 2000). performed co-immunoprecipitation experiments on var- Moreover, it undergoes a Rho- and Crm1-dependent ious mammalian cell lines transiently co-transfected nuclear accumulation in response to various gowth, with expression vectors encoding full-length Flag-tagged morphogenic and apoptotic signals (Muller et al., 2002; FHL2 and E4F1 proteins. The two proteins were Morlon and Sassone-Corsi, 2003). The physiological efficiently co-immunoprecipitated from cellular extracts consequences of this DRAL/FHL2 nuclear accumula- prepared from transfected U2OS osteosarcoma cells tion remains unclear and might depend on the cellular using either an anti-Flag antibody (Figure 2b and d) or context, on the signals that initiates shuttling and on its an anti-E4F1 antibody (Figure 2b). The same results targets. Nevertheless, overexpression of DRAL/FHL2 were obtained when similar co-transfection experiments in several normal and tumor-derived cell lines efficiently were performed in NIH3T3 mouse fibroblasts or in induces an apoptotic program (Scholl et al., 2000). This human rhabdomyosarcoma cell line RDE (data not proapoptotic property might be physiologically relevant shown). To map the FHL2 and E4F1 protein regions since endogenous DRAL/FHL2 gene contains p53 involved in this interaction, we repeated these co- binding sites in its promoter and is strongly induced immunoprecipitation assays on cells transiently expres- by irradiation in a p53-dependent manner with kinetics sing similar levels of various truncated FHL2 (Figure 2a similar to p21WAF1 (Scholl et al., 2000). and b) and E4F1 (Figure 2c and d) proteins. In the present study, we provide evidence for another As shown in Figure 2b, removal of the first two and link between DRAL/FHL2 and the p53pathway. We an half N-terminal LIM domains did not significantly show that DRAL/FHL2 interacts directly with the p53- affect FHL2’s capacity to associate with E4F1 in this associated protein p120E4F1, but not p50E4F1, both in vitro assay, whereas deletion of the third LIM domain and in vivo. This E4F1-FHL2 association inhibits the completely abolished formation of the complex. These capacity of E4F1 to block the cell cycle and correlates results were consistent with the presence in the two- with a reduction of nuclear E4F1-p53complexes and of hybrid screen of a N-ter truncated version of FHL2 and

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5477 a E4F E4F Crm1-dependent nuclear-cytoplasmic shuttling (Muller GAL4 +FHL2 + CONT. et al., 2002; Morlon and Sassone-Corsi, 2003). In contrast, E4F1 was reported to be exclusively nuclear. To confirm that these two proteins associate in a cellular -L / -W context, we next tested whether they could modify the Y2H nuclear-cytoplasmic repartition of each other upon co- -L / -W / -H overexpression. 100AT As previously described, ectopically expressed E4F1 localized almost exclusively in the nucleus of exponen- b tially growing U2OS cells (Figure 3a) (Fajas et al., 2000, 2001). PM In contrast, in the absence of E4F1 overexpression, GST-ELK input GST GST-FHL2 both transfected Flag-tagged FHL2 (Figure 3a, b and c) 220 and endogenous FHL2 proteins (Figure 3d and e), were 120 E4F1 found predominantly in the cytoplasmic compartment, with less than 15% of the positive cells showing also a 72 clear nuclear staining in indirect immunofluorescence 46 (Figure 3b and e). Notably, co-expression of Flag- tagged FHL2 and E4F1 led to a threefold increase in cells showing nuclear FHL2 (Figure 3a and b). This 30 positive effect of E4F1 overexpression on FHL2 nuclear accumulation was confirmed by Western blot analysis of nuclear and cytoplasmic extracts prepared from transfected cells (Figure 3c). Importantly, this relocalization of FHL2 upon E4F1 overexpression likely depends on GST-FHL2 E4F1-FHL2 interaction, since it was not observed with GST-ELK the truncated forms of E4F1 and FHL2 that cannot interact, that is p50E4F1 (Figure 3b) and FHL2DL123 GST (data not shown). Figure 1 E4F1 associates with DRAF/FHL2 in vitro.(a) Altogether, these results suggest that E4F1 recruits Interaction of E4F1 with FHL2 in a two-hybrid assay (Y2H). FHL2 in the nucleus, and provide further evidence that Mav103yeast strain containing chromosomally integrated reporter genes whose expression is regulated by different Gal4 responsive these two factors interact in a cellular context. promoters was transfected with either intact Gal4 or combinations of Gal4DB-E4F1 (Bait, pPC97-p120E4F) with preys TA-FHL2 or TA-DP1 (control) in pPC-86. Yeast growing on permissive plates FHL2 antagonizes the p53-dependent antiproliferative (SC-Leu-Trp) were selected for reporter activation (GAL1:HIS3- dependent prototrophy) by replica plating on SC-Leu-Trp-His plus effects of E4F1 and impacts on E4F1-p53 complexes 100 mM 3-AT plates. (b) Glutathione S-transferase (GST) pull- We and others had previously demonstrated that E4F1 down assays. 35S-labeled in vitro translated full-length E4F1 overexpression leads to a strong reduction of DNA (p120E4F1) (input) was incubated with similar amounts of either synthesis, which results from a cell cycle arrest in G1 Sepharose-bound GST-FHL2, GST or GST-ELK1 (negative (Fernandes et al., 1998; Fajas et al., 2000, 2001; Sandy control). et al., 2000). To evaluate the impact of FHL2 on E4F- mediated growth arrest, we monitored DNA synthesis in asynchronously growing U2OS cells transiently co- suggest that the E4F1-binding region of FHL2 lies transfected with various combinations of E4F1 and within the third LIM domain. Flag-tagged FHL2, together with GFP expressing Notably, removal of the DNA-binding region of plasmids. At 16 h after transfection, cells were incubated E4F1 (DDBD-E4F1 and Gal4-DDBD-E4F1) did not with bromodeoxyuridine (BrdU) for 8 h. Transfected affect its capacity to associate with FHL2 in this assay cells, identified by GFP expression, were scored for (Figure 2d). Conversely, FHL2 was unable to associate BrdU incorporation by fluorescence microscopy with p50E4F1, the E1A-activated and C-terminal trun- (Figure 4a). FHL2 overexpression alone induced at best cated form of E4F1, which contains a functional DNA- a very moderate inhibition of BrdU incorporation. This binding domain. likely results from an increased proportion of apoptotic Altogether, these binding assays provide evidence that cells in FHL2-transfected cells, as shown by tunnel assay E4F1 and DRAL/FHL2 can interact directly in vitro (Figure 4a, left side of the figure). In contrast and in and in living cells. agreement with our previous observations, E4F1 overexpression in this cell line led to a strong reduction of the number of BrdU-positive cells (Figure 4a). Of note, p120E4F strongly stimulates the nuclear accumulation of this antiproliferative effect of E4F1 was almost com- downregulated in rhabdomyosarcoma Lim protein/FHL2 pletely eliminated upon co-expression of FHL2. Im- FHL2 was shown to localize in several cellular portantly, this antagonizing effect of FHL2 on E4F1- compartments and to undergo a continuous mediated growth arrest was not associated with

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5478 a Interaction c Interaction withE4F1 DNA binding with FHL2 Flag 1/2 LIM1 LIM2 LIM3 LIM4 784 279 E4F1 Flag-FHL2 ++ E4F1 / p120 + E4F1 L1 + p50 - L1/L2 + N-E4F1 + L1/L2/L3 - DBD-E4F1 + b GAL4-DBD-E4F1 GAL4DB +

L1 L1/L2/L3 L1/L2 d

E4F1 PM E4F1 +E4F1 FHL2 E4F1+ +E4F1 + DBD-E4F1 E4F1 Input 220 DBD-E4F1 N-E4F1 E4F1 E4F1 120

FHL2 + FHL2 +FHL2 +FHL2 p50 FHL2 + E4F1 FHL2 + GAL4- Input Flag-FHL2 72 FHL2 PM Input FlagFHL2 46 L1 FHL2 120 30 L1/L2 Input L1/L2/L3 72 E4F1 E4F 46

120 E4F1 120 IP IP -Flag 72 72 -Flag E4F1 IgG 46 46

72 IgG

IP 46 Flag-FHL2 -E4F1 L1 30 L1/L2

Figure 2 E4F1 associates with DRAL/FHL2 in U2OS cells. Mapping of the binding domains of FHL2 and E4F1. (a) Schematic representation of the various Flag-tagged FHL2 fragments used in (b) and summary of their capacity to bind E4F1. The LIM domains are shown as gray boxes and the Flag tag as an open circle. (b) Co-immunoprecipitation of full-length E4F1 and 3 Â Flagged-FHL2 deletion constructs (FHL2, DL1, DL1/L2 and DL1/L2/L3) ectopically expressed (CMV-driven expression vectors, PCiNeo) in U2OS cells. One-tenth of the input nuclear extracts were loaded in the two upper panels and probed for the presence of E4F1 and FHL2 as indicated by immunoblotting with an affinity purified rabbit polyclonal antibody (E4F-AD1) and anti-Flag monoclonal antibody (M2), respectively. Nuclear extracts were immunoprecipitated with a mixture of two anti-E4F1 antibodies (E4F-AD1 þ E4F-88) or with anti-Flag antibody coupled to agarose beads as indicated. Anti-Flag and anti-E4F1 precipitates were probed by immunoblotting for the presence of E4F1 and Flagged-FHL2, respectively (lower two panels). (c) Schematic representation of the various E4F1 expression constructs used in (d) and summary of their capacity to bind E4F1. The C2H2 Kru¨ ppel-like Zinc-fingers are shown as gray boxes. Gal4DB fusion is indicated as an open box. (d) Co-immunoprecipitation of full-length Flag-tagged FHL2 and pcDNA3-E4F1 deletion constructs (E4F1, p50E4F1, DN-E4F1, DDBD-E4F1 and Gal4-DDBD-E4F1) ectopically expressed in U2OS cells. One-tenth of the input nuclear extracts were probed as described for the presence of E4F1 and FHL2 by immunoblotting as described in (b) (upper two panels). Nuclear extracts were immunoprecipitated with anti-Flag antibody and precipitates were probed by immunoblotting for the presence of E4F1 deletions (lower panel).

increased apoptosis in cells co-transfected with FHL2 growth suppression nor induced apoptosis of U2OS in and E4F1 relative to FHL2 alone (Figure 4a, left side of the same assay (Figure 4a). the ﬁgure). Finally, overexpression of the carboxy This antagonizing effect of FHL2 on E4F1-mediated terminal-truncated construct FHL2DL123, which is growth arrest was observed in various other cell lines deﬁcient in binding to E4F1, neither rescued E4F1- (Paul and Sardet, unpublished data) including WT and

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5479 a b FHL2

70 FHL2 + p50 FHL2 60 FHL2 + E4F1 50 40 FHL2 + E4F1 30 20 DAPI α−Flag-FHL2 α−HA-E4F1 10 % of FlagFHL2-positive

cells with nuclear staining 0 c + E4F1 d e 70 + E4F1 . . . 60

Cyto. Nucl Cyto Nucl GFP 50 TBP 40 30 α−Flag-FHL2 20 GFP + α−Ha-E4F1 E4F1 10

% of GFP-positive cells 0

DAPI GFP α−FHL2 with FHL2 nuclear staining Figure 3 E4F1 stimulates nuclear accumulation of ectopically expressed and endogenous DRAL/FHL2 in U2OS cells. (a) U2OS cells were transfected with Flagged-FHL2 alone or together with Ha-tagged-E4F1 expression plasmid as indicated. At 16 h after transfection, cells were fixed, labeled with 40,60-diamidino-2-phenylindole (DAPI) and were assessed for ectopically expressed FHL2 and E4F1 proteins by immunoflurescence using mouse monoclonal anti-HA Ab and rabbit polyclonal anti-HA Ab, respectively. (b) Diagram showing the proportion of 3 Â Flagged-FHL2-positive cells with FHL2 nuclear staining. Average of two independent experiments (200 cells counted per experiment) performed as reported in a. U2OS cells were transfected with either Flagged-FHL2 alone or together with Ha-tagged-E4F1 or Ha-tagged-p50E4F1, as indicated. (c) Nuclear and cytoplasmic extracts were prepared from U2OS cells transfected as described in (a) and then probed for expression of Flagged-FHL2, Ha-tagged-E4F1 and TBP (control of fractionation) proteins by immunoblotting with anti-Flag, anti-Ha and anti-TBP Igs, respectively. (d) U2OS cells were transfected with GFP alone (control) or together with Ha-tagged-E4F1 expression plasmid as indicated. Fixed cells were processed as in (a) except that immunofluorescence was performed with Rabbit polyclonal anti-FHL2 Ab to reveal the cellular localization of endogenous FHL2 proteins. (e) Diagram showing the proportion of GFP-positive cells with endogenous FHL2 nuclear staining. Average of three experiments as reported in (d).

FHL2À/À mouse 3T3 fibroblasts (Figure 4b). Signifi- responsive promoters, cycA-Luc (Figure 5a) and E4- cantly, we found that E4F1 was more potent in inducing TK-Luc (Figure 5b), together with expression vectors growth-arrest in FHL2À/À cells than in control WT cells, encoding E4F1 and FHL2. In agreement with previous that is a larger fraction of E4F1-transfected cells were reports, E4F1 alone was found to strongly inhibit arrested in FHL2À/À fibroblasts than in control fibro- expression of these reporters, whereas FHL2 had little blasts, suggesting that endogenous FHL2 also contri- effect on their expression. Notably, this E4F1-mediated butes to antagonize the E4F1-antiproliferative effects repression was antagonized upon co-expression of (Figure 4b). FHL2 but not of FHL2DL123(Figure 5a and b). This These results, together with those showing that FHL2 result suggests that FHL2 is playing the role of an accumulates in the nucleus and associates with E4F1 inhibitor for E4F1 rather than acting as a coactivator or upon co-overexpression, suggest that FHL2 might a corepressor, a function it displays with several other participate to a pathway that interferes with E4F1- transcription factors. mediated growth controls. As previously described in mouse fibroblasts (Sandy et al., 2000; Rizos et al., 2003), the antiproliferative effect of E4F1 in U2OS also appears to depend on the FHL2 inhibits E4F1 repressive activity on transcription presence of Wt p53gene and of the direct interaction The antiproliferative effects of E4F1 has also been between the E4F1 and p53proteins. Accordingly, the shown to depend, at least in part, on its capacity to growth arrest induced by E4F1 overexpression in U2OS inhibit transcription of the cyclin A promoter. To cells is sensitive to short hairpin RNA (shRNA)- evaluate the impact of FHL2 on E4F1-mediated mediated depletion of endogenous p53(Figure 4a) transcriptional repression, U2OS cells were transfected (Brummelkamp et al., 2002). To investigate the potential with luciferase reporter constructs driven by E4F1- impact of FHL2 on E4F1-p53association in these cells,

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5480 a % of GFP-positive cells % of GFP-positive cells functional properties required for growth arrest, that is that are tunnel positive that are BrdU-positive its association with p53and its capacity to repress 30 20 1000 10 20 30 40 50 60 70 transcription. U2OS Control Identiﬁcation of an endogenous FHL2-E4F1 complex in E4F1 UV-irradiated U2OS cells FHL2 To gain information about the physiological function of this E4F1-FHL2 association, we looked for endogenous FHL2 + E4F1 E4F1-FHL2 complexes and asked whether their abun- FHL2123 dance could be regulated. Co-immunoprecipitation FHL2123 + E4F1 experiments were performed with an anti-FHL2 antibody or corresponding preimmune serum on nuclear E4F1 + ShRNA p53 extracts prepared from U2OS cells (Figure 6a), which expresses FHL2 and E4F1 (Fajas et al., 2000, 2001; b Control Amaar et al., 2002). As a negative control, we performed E4F1 the same experiment in RDE cells that express similar E4F1 + FHL2 levels of E4F1 (Figure 6b) but almost undetectable levels of FHL2 (Genini et al., 1997; Scholl et al., 2000). In WT 3T3 FL2-/- 3T3 exponentially growing cells, we repeatedly failed to 40 detect E4F1 in FHL2 precipitates (Figure 6a). Con- sidering the data shown in Figure 4e, we wondered 30 30 whether that this failure might reﬂect the fact that the pool of nuclear FHL2 was low in these growth 20 20 conditions. Therefore, we repeated this experiment after exposure of U2OS cells to UV (30 J/m2), a condition, 10 10 which had been described to stimulate the nuclear % of transfected - cells that are BrdU- positive

0 % of transfected - cells translocation of FHL2 in exponentially growing fibro- that are BrdU- positive blasts (Morlon and Sassone-Corsi, 2003). Consistent E4F1 E4F1+ FHL2 E4F1 E4F1+ FHL2 with this report, we found that FHL2 nuclear accumu- Figure 4 FHL2 antagonizes E4F1-mediated growth arrest. (a) lation was strongly stimulated 8 h after irradiation of U2OS cells were co-transfected with a CMV-GFP expression U2OS cells (Figure 6b and c). Under these conditions, plamid together with combinations of full-length FHL2 and E4F1 endogenous E4F1 was clearly and specifically detected expression vectors or with a short hairpin RNA (shRNA) directed in the anti-FHL2 immunoprecipitate (Figure 6a). Im- to human p53, as indicated. As a marker of cell proliferation, portantly, this association was not detected, in similar transfected (GFP positive) cells were assessed for BrdU uptake in their DNA by using an anti-BrdUrd-specific Ab (right). GFP- experiments carried out in RDE cells (Figure 6b) or positive cells were also assessed for apoptosis using a tunnel assay upon immunoprecipitation of irradiated-U2OS extracts (left). The diagram shows the average of two independent with preimmune serum (Figure 6a). These data show experiments. (b) FHL2À/ÀpR53 and WTpR53 3T3 mouse fibroblasts that genuine E4F1-FHL2 complexes exist in living cells were co-transfected with combinations of increasing amounts (100 ng, 300 ng) of full-length E4F1 expression vector alone or and suggest that they form upon UV-induced nuclear with 300 ng of full-length FHL2 expression vector, as indicated. As accumulation of FHL2. a marker of cell proliferation, transfected (E4F positive) cells were assessed for BrdU uptake in their DNA by using an anti-BrdUrd- specific Ab. The diagram shows the average of two independent experiments. Discussion

Here, we show that the kruppel-type zinc-finger we performed co-immunoprecipitation experiments on transcription factor E4F1 interacts in vitro and in vivo U2OS cells transiently co-transfected with combinations in the nucleus with DRAL/FHL2, a member of the four- of expression vectors encoding HA-tagged E4F1, full- and-a-half LIM domain protein family. Ectopic expres- length p53and flagged-FHL2. As in other cell systems sion of FHL2 impacts on E4F1-dependent cell cycle (Sandy et al., 2000; Rizos et al., 2003), ectopically arrest and repression of E4F1-responsive promoters, expressed E4F1 and p53proteins were efficiently co- and correlates with a strong reduction of nuclear E4F1- immunoprecipitated from U2OS nuclear extracts by an p53complexes. Conversely, E4F1 is more potent in anti-E4F1 antibody (Figure 5c). However, co-immuno- inducing growth-arrest in FHL2À/À fibroblasts than in precipitation of E4F1 and p53was reproducibly and control WT cells. This E4F1-FHL2 interaction occurs markedly reduced upon overexpression of FHL2 upon the translocation of FHL2 into the nucleus and (Figure 5c), suggesting that FHL2 interferes with the supports the notion that E4F1 biological activity might formation of E4F-p53nuclear complexes in cells. be inhibited by signals that control the nuclear Altogether, these data suggest that FHL2, upon accumulation of FHL2. Consistent with this hypothesis, its nuclear translocation, targets at least two E4F1 we found that UV irradiation of U2OS cells induces the

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5481 a Control b E4F-TK-Luc c FHL2 FHL2 - + - + E4F1 P53 + + + + CycA-Luc 45 E4F1 + + + + FHL2 + E4F1 40 70 L1/L2/L3 + E4F1

(RLU) 35

60 -3 (RLU)

30 -E4F1 -E4F1 -control -cont -3 50 25 α α α α IInput IP IInput IP IInput IP IInput IP 40 20 E4F1 30 15 20 10 10 p53 5 0 luciferase activity 10 luciferase activity 10 0 1X 5X 1X 5X FHL2 FHL2 FHL2 Figure 5 FHL2 inhibits E4F1 repression of transcription and alters E4F1-p53complexes. ( a) CycA promoter reporter assay. U2OS cells were transiently cotransfected with an E4F-responsive reporter gene, pCycA-mCCRE/CHR-Luc, alone (control) or together with combinations of expression vectors coding for E4F1, FHL2 or FHL2DL1/L2/L3mutant, as indicated. Luciferase activity (RLU) was normalized with b-galactosidase activity encoded by a CMV-b-Gal plasmid cotransfected in all assays. Standard deviations of triplicates are indicated. (b) Reporter assay performed as in (a) but with an other E4F-responsive reporter construct : E4F-TK-Luc. (c) Co-immunoprecipitation of nuclear p53and E4F1 in presence of FHL2. Nuclear extracts were prepared from U2OS cells transiently co-transfected with combinations of expression vectors encoding HA-tagged E4F1, full-length p53and Flagged-FHL2, as indicated. E4F1 was immunoprecipitated from these extract using affinity purified polyclonal anti-E4F1 Abs (E4F-AD1 þ E4F-88) or control antibody. These precipitates and one-tenth of the input nuclear extracts were probed for the presence of Ha-tagged-E4F1, Flag-tagged and p53by immunoblotting using monoclonal anti-Ha, anti-p53(DO1) and anti-Flag, respectively. nuclear translocation of endogenous FHL2 and the tion factors, including E4F1 and two other kruppel-type formation of nuclear E4F1-FHL2 complexes. zinc-finger transcription factors, the products of the Although expression of FHL2 is predominantly Wilms tumor suppressor gene Wt1 (Du et al., 2002) and detected in the heart (Chan et al., 1998), it has also the promyelocytic leukemia zinc-finger (plzf) protein been observed in several extracardiac tissues at certain (McLoughlin et al., 2002). In most cases, FHL2 stages of development and in many (if not all) different association with transcription factors was shown to cell lines in culture, including U2OS (Amaar et al., 2002) result in stimulation of the transcriptional activity of and fibroblastic cells used in this work. In these cell these factors, FHL2 acting either as a coactivator for lines, FHL2 has been reported to localize in several transactivators or a corepressor for repressors. Ex vivo cellular compartments, including focal adhesions (Scholl studies have demonstrated for E4F1 the potential to et al., 2000; Li et al., 2001; Gabriel et al., 2004), either repress or activate transcription depending if the cytoplasm and nucleus (Scholl et al., 2000; Muller et al., protein is full-length (p120) or proteolytically truncated 2002; Morlon and Sassone-Corsi, 2003). This cellular (p50), respectively (Fognani et al., 1993; Fajas et al., localization and/or expression of FHL2 in these cell 2000, 2001). Our mapping of E4F1-FHL2 interaction lines have been shown to be highly regulated and might shows that FHL2 does not associate with p50E4F2, be an essential parameter in modulating different whereas it strongly binds in vitro and in vivo to the cellular functions. Notably, signals, as different as repressor p120E4F1. Surprisingly, we found that p120E4F1 mitogenic restimulation of arrested cells, morphogenic transcriptional activity was inhibited by FHL2 rather stimulation or cell irradiation, strongly increase FHL2 than enhanced, suggesting that FHL2 acts as an expression and/or nuclear localization (Muller et al., inhibitor for E4F1 rather than as a corepressor. 2002; Morlon and Sassone-Corsi, 2003; Philippar et al., Although atypical, such inhibitory effect of FHL2 has 2004; Purcell et al., 2004). Consistent with this observa- already been reported on the transcription factors SRF tion, the FHL2 gene was shown to be a direct target of (Philippar et al., 2004), FOXO1 (Yang et al., 2005) and both p53(Scholl et al., 2000) and SRF (Philippar et al., Hand1 (Hill and Riley, 2004). In the case of SRF, FHL2 2004). Collectively, these data suggest that FHL2 was shown to compete for interaction of this factor with activation might be a general and early response to the coactivator MAL (Philippar et al., 2004), whereas it changes in cellular homeostasis. In some aspects it is inhibits FOXO1 transcription by recruiting the deace- similar to activation of NFKB or c-fos, which precedes tylase SIRT1 to FOXO1 complexes (Yang et al., 2005). and synergizes with more specific secondary responses The molecular mechanism by which FHL2 blocks E4F1 whose outcome is to direct cells either to apoptosis, transcriptional activity on cycA-Luc and E4-TK-Luc proliferation or differentiation. It is likely that FHL2 reporters remains unclear. Indeed, by using a number of exerts multiple functions upon its translocation into the assays we found no evidence for an interaction of FHL2 nucleus where it directly interacts with several transcrip- with the DNA binding domain of E4F1 (Figure 2), an

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5482 a Mock U2OS + UV c DAPI FHL2

-FHL2 -FHL2 Mock PM input IP pre-im.IP input IP pre-im.IP

120 E4F1 72 IgG 46 + UV

b Mock RDE + UV d + UV 80

70 -FHL2 -FHL2 PM input IP pre-im.IP input IP pre-im.IP 60 50 40 E4F1 120 30

72 nuclear staining 20

IgG % of cells with FHL2 46 10 0 Figure 6 An endogenous FHL2–E4F1 complex is detected in UV-irradiated U2OS cells and correlates with the UV-stimulated nuclear accumulation of FHL2. (a) Co-immunoprecipitation of endogenous E4F1 and FHL2 from U2OS nuclear extracts. Nuclear extracts were prepared from U2OS cells either exponentially growing (Mock) or 8 h after UVC-irradiation (30 J/m2) and FHL2 was immunoprecipitated from these extracts using a rabbit polyclonal anti-FHL2 Ab (FHL2-CS1). These FHL2 precipitates (IP a-FHL2) or control immunoprecipitation realized with preimmune serum (IP pre-im) were probed for the presence of E4F1 by immunoblotting using affinity purified polyclonal anti-E4F1 Ab (E4F-AD1). Of the input nuclear extracts, 4%, were probed in each condition (input). (b) Co-immunoprecipitation assays performed as in (a) but on nuclear extracts prepared from the FHL2-negative rhabdomyosarcoma cell line RDE. (c) Cellular localization of endogenous FHL2 proteins in exponentially growing phase or 10 h after UVC irradiation (30 J/m2). Immunofluorescence was performed with Rabbit polyclonal anti-FHL2 as described in Figure 3d. (d) Diagram showing the proportion of GFP-positive cells with strong endogenous FHL2 nuclear staining in exponentially growing phase (open bar) or after UVC irradiation (filled bar). Average of three experiments as reported in (c).

in vitro effect of FHL2 on E4F-binding activity to DNA, growth arrest induced by E4F1 overexpression in these or direct binding of FHL2 to DNA (Paul and Sardet, cells is sensitive to ShRNA-mediated depletion of unpublished observations). Moreover, we also consid- endogenous p53(Figure 4). Significantly, ectopic ex- ered that FHL2 might interfere with the association of pression of FHL2 antagonizes the formation of these E4F1 with the corepressor histone deacetylase HDAC1 E4F1-p53nuclear complexes, suggesting that FHL2 also (Colombo et al., 2003), assuming that this would counteracts E4F1 effects on proliferation through the prevent nucleosome deacetylation and transcriptional disruption of this complex. Importantly, we did not repression at E4F-regulated promoters. However, we observe similar effects of FHL2 on E4F1-pRB com- were unable to test this hypothesis because, unlike plexes (Paul and Sardet, unpublished observation). others (Colombo et al., 2003), we did not detect E4F- These data suggest that FHL2, upon its nuclear HDAC1 complexes in our cell systems (Paul and translocation, targets at least two E4F1 functional Sardet., unpublished observations). A better knowledge properties required for growth arrest, that is its of E4F1 communication with the polymerase II association with p53and its capacity to repress holoenzyme and with the chromatin remodeling ma- transcription. Notably, others have shown that the chinery in U2OS will be required to further explore the FHL2 gene is a transcriptional target of p53with a mechanisms of this FHL2-mediated inhibition of E4F1 kinetics of induction similar to p21 (Scholl et al., 2000). transcriptional activity. Consistent with this observation, FHL2 expression is Beside this direct effect on transcription, evidence strongly induced in various cell lines (Scholl et al., 2000) indicates that E4F1-mediated growth arrest also de- and in primary human peripheral blood lymphocytes pends on its association with p53and pRb (Sandy et al., (PBLs) (Kang et al., 2003) exposed to irradiation, that is 2000; Fajas et al., 2000; Rizos et al., 2003). As previously under conditions where p53is stabilized and activated. shown in mouse fibroblasts (Sandy et al., 2000), we As shown here in U2OS, and by others in fibroblasts found that ectopically expressed E4F1 and p53form a (Morlon and Sassone-Corsi, 2003), this response of nuclear complex in Human U2OS cells and that the FHL2 to irradiation also occurs at the level of its

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5483 nuclear accumulation. Thus, UV irradiation rapidly Two-hybrid screening induces an accumulation of FHL2 in the nucleus with a Yeast strain MAV103expressing pPC97-p120 E4F1 was trans- kinetics that suggests this translocation precedes the fected with a WI38 human fibroblasts cDNA pPC86 preys stimulation of its expression by p53(Scholl et al., 2000; library and interactors were selected essentially as previous Morlon and Sassone-Corsi, 2003). Interestingly, we described (Sardet et al., 1995; Fajas et al., 2000). observed that this UV-induced nuclear relocalization Cell culture, transfections and luciferase assays correlates with the formation of native E4F-FHL2 U2OS osteosarcoma cells line and RDE human embryonal complexes in U2OS cells. While, these data do not rhabdomyosarcoma cells line were grown in Dulbecco definitively address the physiological function of this modified Eagle medium (Sigma, Saint Quentin Fallavier, targeting it is tempting to speculate that FHL2 could be France) supplemented with 10% FBS. All transfections were the effector of a feed back loop controlling of the p53 carried out using calcium phosphate procedure. For E4F1/ pathway, whose functions would be to target p53-E4F1 FHL2 Luciferase assays, cells were co-transfected with 0.2 mg nuclear complexes and E4F1 transcriptional program in of pCH110b-galactosidase expression vector and 2 mg of E4- TK-LUC or pCycA-DCCRE/CHR-Luc alone or with various response to irradiation. In addition, while there are E4F1 evidence that E4F1, p53and FHL2 participate indivi- combinations of 100 ng of pcDNA-p120 , pCI3XF-FHL2 or pCI3XF-FHL2DL1/L2/L3. b-Galactosidase and luciferase dually in cellular growth control and apoptosis, the role activity were measured 36 h later as previously described of genuine p53-E4F1 and E4F1-FHL2 complexes in (Fajas et al., 2000, 2001). Luciferase activity values were these processes remain to be investigated and will normalized to the b-galactosidase activity to account for constitute the subject of our future studies. variations in transfection efficiencies. FHL2À/ÀpR53 and WTpR53 3T3 mouse fibroblasts were derived from FHL2À/À and control WT 3T3 cells (Labalette and Wei, Materials and methods 2004) to contain a retrovirally transduced, functional p53protein.

DNA plasmids Immunofluorescence Details of all constructs are available upon request. Bait for U2OS and RDE cells were grown on coverslips for 24 h and the two-hybrid assays in pPC97 is as follow: pPC97-p120E4F transfected for 10 h with indicated plasmids and further grown (E4F cDNA nt 980-3325 Sal1/BamH1, ( þ )50-ACGCGTC for 20 h before different treatment of cells. For BrdU staining, GACGGGCGAGATGGCAGTGCG-30,(À)50-ACGCGGA cells were incubated for 5 h with BrdU. After formalin fixation TCCTAGACGATGACCGTCTGCAC-30; pPC97Sal/Bgl2). for 20 min at 41C followed by a 5 min methanol permeabiliza- Mammalian E4F1 expression constructs are in pcDNA3 tion, cells were treated with 2 N HCl for 15 min at room (Invitrogen, Cergy-Pontoise, France) as follows: pcDNA- temperature and incubated with anti-BrdU mouse monoclonal p120E4F1 (full-length E4F1), pcDNA-p50E4F1 (amino-acid antibody (Immunotech-Beckman Coulter, Luminy, France). (aa)1-357 of E4F1), DDBD-E4F1 (aa357-784 of E4F1) and Apoptosis analysis was performed with In Situ Cell Death Gal4-DDBD-E4F1 were previously described in Fajas et al. Detection Kit, TMR red following the manufacturer’s (2000, 2001). Mammalian FHL2 expression constructs are as instructions (Roche, Penzberg, Germany). Detection of follows: pCI-FHL2, a 50 end Not1/Sal1 fragment from Y2H transfected FHL2 was performed with anti-FLAG primary prey clone pPC86-FHL2 encoding full-length FHL2 was monoclonal anti-body M2 (Sigma-Aldrich, Saint Quentin cloned Not1/Sal1 in pCINeo. 3 Â Flagged FHL3constructs, Fallavier, France) and that of endogenous FHL2 with a goat pCI3XF-FHL2, was constructed by cloning first a single copy anti-FHL2 (Santa Cruz via tebu-bio, le Perray en Yvelines, of the double-stranded oligonucleotide ( þ )50-GGCCGC France), followed by incubation with suitable Texas-red- TAATACGACTCACTATAGGGATGGATTATAAAGAT labeled secondary antibodies (Jackson ImmunoResearch GATGACGATAAGGATTACAAAGACGATGATGACAA Laboratories, West Grove, USA). Detection of E4F1 was as GGACTATAAAGACGACGATGATAAAG-30/(À)50-AAT described previously (Fajas et al., 2000, 2001). TCTTTATCATCGTCGTCTTTATAGTCCTTGTCATCATC Glutathione S-transferase pull-down assay, protein extracts, GTCTTTGTCATCCTTATCGTCATCATCTTTATAATCC immunoprecipitations. Glutathione S-transferase pull-down ATCCCTATAGTGAGTCGTATTAGC-30 into digested assays were carried out with similar amounts of GST, GST- Nhe1/Not1 pCI-FHL2 and by replacing the Not1/Xho1 ELK and GST-FHL2 proteins bound to Gluthatione-sephar- fragment by a fragment generated by polymerase chain ose Beads (Amersham-GE HealthCare Biosciences, Orsay, reaction (PCR) ( þ )50-CTGACGCGGCCGCTGACTGAGC France) combined with 35S-labeled in vitro translated full- 0 GCTTTGACTGCCAC-3 and with the commun minus strand length E4F1 in PBST (PBS with 0.1% Tween, 20 mM ZnSO4). (À)50-CTGACCTCGAGTCAGATGTCTTTCCCACAGT For co-immunoprecipitation of transfected p53-E4F1, FHL2- CG-30. FHL2 deletion constructs were built using similar E4F1 complexes, U2OS cells were transfected with indicated strategy: pCI3XF-FHL2DL1, ( þ )50-CTGACGCGGCCGCT expression plasmids (pCMV-p53, pCI3XF-FHL2, pCDNA3- GTTTGAGACCCTGTTCGCCAACAC-30and the commun E4F1, pCI3XF-FHL2 or FHL2 and E4F1 deletions) and used minus strand; pCI3XF-FHL2DL1/L2, ( þ )50-CTGACGCG to prepare nuclear extracts as previously described (Fajas GCCGCTGTATTCCAACGAGTACTCATCCAAG-30 and et al., 2000, 2001). Immunoprecipitations were carried on 1 mg minus commun strand; pCI3XF-FHL2DL1/L2/L3, ( þ )50- of nuclear extracts with either a mixture of rabbit anti-E4F1 CTGACGCGGCCGCTGTATGAGAAACAACATGCCATG polyclonal antibodies (AD1 and 88-2, Fajas et al., 2000, 2001), CAG-30 and minus commun strand. Reporter constructs, E4- anti-Flag (Sigma-Aldrich) or anti-HA Igs.ProtA/G-sepharose TK-LUC, pCycA-DCCRE/CHR-Luc and CMV-GFP were as precipitates (Amersham) were eluted and separated on 10% previously described (Fajas et al., 2000, 2001). FHL2 GST acrylamide sodium dodecyl sulfate/polyacrylamide gel electro- fusion (pGKG-FHL2) was constructed by cloning a fragment phoresis (SDS/PAGE) gels, blotted and probed with either Bam1/Xho1 from pCINeoFHL2 in pGEX-KG (BamH1/ anti-E4F1 polyclonal antibody (p120lyon, Fajas et al., 2000, Xho1). Plasmid that direct the synthesis of shRNA directed 2001), Anti-p53(DO1), Anti-FLAG M2 (Sigma-Aldrich) or to human p53is as described in Brummelkamp et al. (2002). Anti-TBP (Dantonel et al., 2000) as indicated in figures. For

Oncogene LIM-only protein FHL2 is a negative regulator of E4F1 C Paul et al 5484 detection of endogenous E4F1-FHL2 complexes in RDE and Acknowledgements U2OS cells, nuclear extracts were prepared 4 h after UVC irradiation (30 J/m2 Amersham), immunoprecipitated with a We are grateful to R Bernards and R Agami for shRNA puriﬁed rabbit polyclonal anti-FHL2 developed in the lab (J81, reagents and C Gauthier-Rouvierre for RDE cells. We thank generated against GST-FHL2) or corresponding preimmune R Hipskind and JM Blanchard for critical reading of the serum, and precipitates were probed by Western blot with anti- manuscript. M Lacroix is supported by a fellowship from the E4F1. For FHL2 localization, 20 mg of nuclear and cytosolic French Association pour la recherche contre le Cancer (ARC). extract prepared as described (Fajas et al., 2000, 2001) and This work was supported by La ligue Contre Le Cancer (C probed by Western blotting with indicated Igs. Sardet and JM Blanchard, e´ quipe labellise´ e 2003).

References

Ahmed-Choudhury J, Agathanggelou A, Fenton SL, Ricketts Kong Y, Shelton JM, Rothermel B, Li X, Richardson JA, C, Clark GJ, Maher ER et al. (2005). Cancer Res 65: 2690– Bassel-Duby R et al. (2001). Circulation 103: 2731–2738. 2697. Labalette C, Renard CA, Christine Neuveut C, Buendia MA, Amaar YG, Thompson GR, Linkhart TA, Chen ST, Baylink Wei Y. (2004). Mol Cell Biol 24: 10689–10702. DJ, Mohan S. (2002). J Biol Chem 277: 12053–12060. Le Cam L, Lacroix M, Ciemerych MA, Sardet C, Sicinski P. Brummelkamp TR, Bernards R, Agami R. (2002). Science 296: (2004). Mol Cell Biol 24: 6467–6475. 550–553. Li HY, Kotaka M, Kostin S, Lee SM, Kok LD, Chan KK Chan KK, Tsui SK, Lee SM, Luk SC, Liew CC, Fung KP et al. (2001). Cell Motil Cytoskeleton 48: 11–23. et al. (1998). Gene 210: 345–350. McLoughlin P, Ehler E, Carlile G, Licht JD, Schafer BW. Chu PH, Bardwell WM, Gu Y, Ross Jr J, Chen J. (2000a). Mol (2002). J Biol Chem 277: 37045–37053. Cell Biol 20: 7460–7462. Martin B, Schneider R, Janetzky S, Waibler Z, Pandur P, Kuhl Chu PH, Ruiz-Lozano P, Zhou Q, Cai C, Chen J. (2000b). M et al. (2002). J Cell Biol 59: 113–122. J Mech Dev 95: 265. Morlon A, Sassone-Corsi P. (2003). Proc Natl Acad Sci USA Colombo R, Draetta GF, Chiocca S. (2003). Oncogene 22: 100: 3977–3982. 2541–2547. Muller JM, Isele U, Metzger E, Rempel A, Moser M, Pscherer Dantonel JC, Quintin S, Lakatos L, Labouesse M, Tora L. A et al. (2000). EMBO J 19: 359–369. (2000). Mol Cell 6: 715–722. Muller JM, Metzger E, Greschik H, Bosserhoff AK, Mercep Du X, Hublitz P, Gunther T, Wilhelm D, Englert C, Schule R. L, Buettner R et al. (2002). EMBO J 21: 736–748. (2002). Biochim Biophys Acta 577: 93–101. Philippar U, Schratt G, Dieterich C, Muller JM, Galgoczy P, Fajas L, Paul C, Zugasti O, Le Cam L, Polanowska J, Engel FB et al. (2004). Mol Cell 16: 867–880. Fabbrizio E et al. (2000). Proc Natl Acad Sci USA 97: Purcell NH, Darwis D, Bueno OF, Muller JM, Schule R, 7738–7743. Molkentin JD. (2004). Mol Cell Biol 24: 1081–1095. Fajas L, Paul C, Vie A, Estrach S, Medema R, Blanchard JM Retaux S, Bachy I. (2002). Mol Neurobiol 26: 269–281. et al. (2001). Mol Cell Biol 21: 2956–2966. Raychaudhuri P, Rooney R, Nevins JR. (1987). EMBO J 6: Fenton SL, Dallol A, Agathanggelou A, Hesson L, Ahmed- 4073–4081. Choudhury J, Baksh S et al. (2004). Cancer Res 64: 102–107. Rizos H, Diefenbach E, Badhwar P, Woodruff S, Fimia GM, De Cesare D, Sassone-Corsi P. (2000). Mol Cell Becker TM, Rooney RJ et al. (2003). J Biol Chem 278: Biol 20: 8613–8622. 4981–4989. Fernandes ER, Rooney RJ. (1997). Mol Cell Biol 17: Sandy P, Gostissa M, Fogal V, Cecco LD, Szalay K, Rooney 1890–1903. RJ et al. (2000). Oncogene 19: 188–199. Fernandes ER, Zhang JY, Rooney RJ. (1998). Mol Cell Biol Sardet C, Vidal M, Cobinik D, Geng Y, Onufrik C, Chen A 18: 459–467. et al. (1995). Proc Natl Acad Sci USA 92: 2403–2407. Fernandes ER, Rooney RJ. (1999). Mol Cell Biol 19: Scholl FA, McLoughlin P, Ehler E, de Giovanni C, Schafer 4739–4749. BW. (2000). J Cell Biol 151: 495–506. Fognani C, Della Valle G, Babiss LE. (1993). EMBO J 12: Stilo R, Leonardi A, Formisano L, Di Jeso B, Vito P, Liguoro 4985–4992. D. (2002). FEBS Lett 521: 165–169. Gabriel B, Mildenberger S, Weisser CW, Metzger E, Gitsch G, Tanahashi H, Tabira T. (2000). Hum Mol Genet 9: 2281–2289. Schule R et al. (2004). Anticancer Res 24: 921–927. Tessari MA, Gostissa M, Altamura S, Sgarra R, Rustighi A, Genini M, Schwalbe P, Scholl FA, Remppis A, Mattei MG, Salvagno C et al. (2003). Mol Cell Biol 23: 9104–9116. Schafer BW. (1997). DNA Cell Biol 16: 433–442. Wei Y, Renard CA, Labalette C, Wu Y, Levy L, Neuveut C Hill AA, Riley PR. (2004). Mol Cell Biol 24: 9835–9847. et al. (2003). J Biol Chem 278: 5188–5194. Differential Regulation of Hand1 Homodimer and Hand1- Wixler V, Geerts D, Laplantine E, Westhoff D, Smyth N, E12 Heterodimer Activity by the Cofactor FHL2. Aumailley M et al. (2000). J Biol Chem 275: 33669–33678. Kang CM, Park KP, Song JE, Jeoung DI, Cho CK, Kim TH Yang Y, Hou H, Haller EM, Nicosia SV, Bai W. (2005). et al. (2003). Radiat Res 159: 312–319. EMBO J 24: 1021–1032.

Oncogene