HDAC4 mediates transcriptional repression by the acute ... - Nature

4 downloads 0 Views 318KB Size Report
Oct 4, 2004 - 1UPR 9079 CNRS-Ligue Nationale Contre le Cancer, Institut Andre´ Lwoff, 7 rue ... PLZF, the promyelocytic leukaemia zinc-finger protein, is.
Oncogene (2004) 23, 8777–8784

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $30.00 www.nature.com/onc

HDAC4 mediates transcriptional repression by the acute promyelocytic leukaemia-associated protein PLZF Anne Chauchereau1,4, Marion Mathieu1,4, Julie de Saintignon1, Roger Ferreira1, Linda L Pritchard1, Zohair Mishal2, Anne Dejean3 and Annick Harel-Bellan*,1 1 UPR 9079 CNRS-Ligue Nationale Contre le Cancer, Institut Andre´ Lwoff, 7 rue Guy Moˆquet, 94800 Villejuif, France; 2Laboratoire de Cytome´trie, Institut Andre´ Lwoff, 7 rue Guy Moˆquet, 94800 Villejuif, France; 3INSERM U579, Institut Pasteur, 28 rue du Dr Roux, 75015 Paris, France

PLZF, the promyelocytic leukaemia zinc-finger protein, is a transcriptional repressor essential to development. In some acute leukaemias, a chromosomal translocation fusing the PLZF gene to that encoding the retinoic acid receptor RARa gives rise to a fusion protein, PLZF– RARa, thought to be responsible for constitutive repression of differentiation-associated genes in these cells. Repression by both PLZF and PLZF–RARa is sensitive to the histone deacetylase inhibitor TSA, and PLZF was previously shown to interact physically with HDAC1, a class I histone deacetylase. We here asked whether class II histone deacetylases, known to be generally involved in differentiation processes, participate in the repression mediated by PLZF and PLZF–RARa, and found that PLZF interacts with HDAC4 in both GST-pull-down and co-immunoprecipitation assays. Furthermore, HDAC4 is indeed involved in PLZF and PLZF–RARa-mediated repression, since an enzymatically dead mutant of HDAC4 released the repression, as did an siRNA that blocks HDAC4 expression. Taken together, our data indicate that recruitment of HDAC4 is necessary for PLZF-mediated repression in both normal and leukaemic cells. Oncogene (2004) 23, 8777–8784. doi:10.1038/sj.onc.1208128 Published online 4 October 2004 Keywords: PLZF; HDAC; PLZF–RARa

Introduction The promyelocytic leukaemia zinc-finger (PLZF) gene was initially identified because of its rearrangement in acute promyelocytic leukaemia (APL) cells carrying the translocation t(11;17)(q23;q21) (Chen et al., 1993). This translocation fuses the PLZF gene to that encoding the transcription factor retinoic acid receptor a (RARa), *Correspondence: A Harel-Bellan, CNRS UPR 9079, Institut Andre´ Lwoff, 7 rue Guy Moˆquet, 94800 Villejuif, France; E-mail: [email protected] 4 These two authors contributed equally to this work Received 5 April 2004; revised 3 August 2004; accepted 4 August 2004; published online 4 October 2004

involved in myeloid differentiation (Zelent et al., 2001). The PLZFa fusion protein exerts a dominant-negative effect on RARa functions, provoking resistance to pharmacological doses of retinoic acid in this leukaemia (Guidez et al., 1998). More recently, PLZF has been shown to have pleiotropic functions (Nanba et al., 2003; Senbonmatsu et al., 2003) and to play an essential role in the development of nonhaematopoietic tissues (Barna et al., 2000). In the haematopoietic lineage, PLZF is expressed in undifferentiated cells and in myeloid progenitors, and is downregulated upon myeloid differentiation. PLZF is a transcriptional repressor that most likely affects cell proliferation by regulating, among others, cyclin A2 (Shaknovich et al., 1998) and c-myc (McConnell et al., 2003), and is considered to be a tumour suppressor (Zelent et al., 2001). PLZF is a member of the POK (BTB/POZ and Kru¨ppel-like zinc-finger) protein family. Its N-terminal POZ domain is a transcriptional repression domain mediating protein homo- and hetero-dimerization (Bardwell and Treisman, 1994), and its C-terminal region contains nine C2–H2 zinc-finger-like domains involved in both DNA binding and transcription repression. The repressive domains of PLZF function by recruiting N-CoR/Sin3A co-repressor complexes (David et al., 1998; Wong and Privalsky, 1998; Melnick et al., 2002), which in turn recruit histone deacetylases (HDACs). Moreover, repression by PLZF and PLZF– RARa is sensitive to the histone deacetylase inhibitor TSA (David et al., 1998; Grignani et al., 1998; Lin et al., 1998). HDACs participate in transcriptional repression by deacetylating histone tails that protrude from nucleosomes, resulting in local modification of chromatin structure (Wolffe and Guschin, 2000). Based on structural and biochemical criteria, HDAC proteins have been classified into three groups (Fischle et al., 2001). Class III HDACs such as the Sir2 protein are associated with silenced heterochromatin. Two of the Class I HDACs, HDAC1 and 2, are found in the NCoR/Sin3 complexes (Hassig et al., 1997; Heinzel et al., 1997), and are involved in regulating RARa activity (Heinzel et al., 1997) and cell proliferation (Magnaghi-Jaulin et al., 1998). HDAC1 has also been

PLZF and HDAC4 A Chauchereau et al

8778

shown to be present in repressive complexes associated with PLZF and PLZF–RARa (David et al., 1998; Lin et al., 1998; Melnick et al., 2002). Class II HDACs, HDAC4–7 and HDAC9–11, have an N-terminal domain of unknown function that participates in transcriptional repression (Verdel and Khochbin, 1999; de Ruijter et al., 2003), and they intervene in cell differentiation, particularly in muscle (Lu et al., 2000; McKinsey et al., 2000; Dressel et al., 2001). Interestingly, all known class II proteins can shuttle from the cytoplasm to the nucleus, whereas most of the class I HDACs are constitutively nuclear (de Ruijter et al., 2003). The POK protein Bcl-6 is also involved in human leukaemogenesis, for example, in non-Hodgkin’s lymphoma (Ahmed et al., 1993). Class II HDACs interact with Bcl-6, and preliminary evidence indicates that other POK proteins, including PLZF, also do so (Lemercier et al., 2002). Given the importance of class II HDACs in cell differentiation, we examined their role in PLZF function. Our data show that PLZF and PLZF–RARa physically interact with class II HDACs and cooperate with HDAC4 for gene repression, suggesting an essential role for HDAC4 in leukaemogenesis by PLZF–RARa.

Results PLZF interacts with class II HDACs in vitro and ex vivo We investigated physical interactions between PLZF and class I or class II HDACs in vitro by GST-pulldown, and observed an interaction between GST-PLZF (1–456) and the class I (HDAC-1, 2 and 3) or class II HDACs (HDAC4–6) (Figure 1b and c). Interaction of PLZF with HDAC4 tended to be slightly stronger than with HDAC1 (David et al., 1998), consistent with a higher affinity for HDAC4. Next, extracts of CV-1 cells transfected with expression vectors for Gal4–PLZF and for tagged HDACs were immunoprecipitated with anti-GAL4 antibody and analysed by Western blot. HDAC4 (Figure 1d), 5 and 6 (Figure 1e) were detected in the immunoprecipitates, indicating that these proteins interact with PLZF in live cells, whereas HDAC1 was not detected under the same conditions (not shown). In co-immunoprecipitation experiments, HDAC4 (Figure 1f), 5 and 6 (Figure 1g) also interacted with ectopically expressed untagged PLZF. HDAC4 and 5 are the most similar proteins among the class II HDACs (de Ruijter et al., 2003). They both interact with and repress the myogenic transcription factor MEF2 through their N-terminal part (aa 1–208 of HDAC4) (Wang et al., 1999; Chan et al., 2003). However, when tagged HDAC4 (1–208) was co-expressed with PLZF in CV-1 cells, we could not detect an interaction by co-immunoprecipitation (not shown), suggesting that the N-terminal region of HDAC4 is not the domain interacting with PLZF. Oncogene

Both endogenous and ectopically expressed PLZF interact with HDAC4 Cytoplasmic HDAC4 is driven towards the nucleus by a variety of interacting proteins in several cell types (Borghi et al., 2001; Miska et al., 2001). To further document complex formation between PLZF and HDAC4, we transfected NIH 3T3 cells with expression vectors for Gal4–PLZF and myc-HDAC4, alone or in combination, and analysed the effect of PLZF on the subcellular localization of HDAC4 by immunofluorescence (Figure 2a). HDAC4 was mostly cytoplasmic when transfected alone, though a small proportion of transfected cells showed nuclear staining. This proportion increased significantly in the presence of Gal4– PLZF, indicating that PLZF affects HDAC4 localization, driving it to – or retaining it in – the nucleus. The 1–456 domain of PLZF, conserved in the leukaemogenic fusion protein PLZF–RARa, similarly increased the proportion of cells with nuclear HDAC4; identical results were observed in CV-1 cells when the full-length PLZF was co-transfected with myc-HDAC4 (Figure 2b). The nuclear re-localization was even more pronounced with the PLZF–RAR protein. Thus, both PLZF and PLZF–RARa can enforce a nuclear localization on HDAC4. Interaction between endogenous HDAC4 and PLZF was investigated in MDS myelodysplastic cells treated with calcium ionophore (CaI) to induce PLZF expression (Licht et al., 1996; Figure 3a). Clear colocalization of PLZF with HDAC4 was seen in the nucleus (Figure 3b). CaI, which did not modify the level of PLZF in individual cells (not shown), did not affect the colocalization profile, but simply increased the proportion of cells expressing PLZF under these conditions. In agreement with our in vitro analyses, HDAC1 also showed partial colocalization with PLZF in MDS cells, but to a lesser extent (Figure 3b). Colocalization of PLZF with HDAC5 or 6 was even less marked (not shown). Nuclear PLZF labelling was seen both in micro-speckles and in large dots resembling nuclear bodies, as described previously (Reid et al., 1995). Interestingly, PLZF–HDAC4 colocalization was generally not observed in the large dots (Figure 3c). Attempts to co-immunoprecipitate endogenous PLZF and HDAC4 from MDS cells under standard conditions did not succeed; perhaps the antibodies used interfere with complex formation. HDAC4 inhibition releases repression by PLZF and by PLZF–RARa Experiments using TSA have shown that HDAC activity is implicated in both PLZF- and PLZF– RARa-mediated transcriptional repression (David et al., 1998; Lin et al., 1998; and data not shown), but the enzyme involved remains to be identified. We addressed this issue using enzymatically dead HDAC mutants designed as in Hassig et al. (1998) for HDAC1, and Wang et al. (1999) for HDAC4, with essential histidines replaced (Figure 4a: histidines 140–141 for HDAC1, 802–803 for HDAC4). Though expressed at

PLZF and HDAC4 A Chauchereau et al

8779

Figure 1 PLZF interacts with HDAC4. (a) Schematic structure of RARa, PLZF–RARa and PLZF proteins. Grey boxes: RARaderived sequences. DBD: DNA-binding domain. HBD: hormone-binding domain. Z: zinc-finger motif. Black triangles: breakpoints for the t(11;17) translocation. (b, c) HDAC binding to PLZF1 456 in vitro. HDAC1–6 were analysed by GST-pull-down using GST- or GST-PLZF1 456-coated beads. Input: 10% of amount used for pull-down. Histogram: quantification of pull-down results, mean7s.d. of three experiments. (d, e) PLZF interacts with HDAC4, 5 and 6 in live cells. Whole-cell extracts (WCE) of CV-1 cells transfected as indicated were immunoprecipitated (IP) with anti-Gal4 antibody, then assayed for tagged HDACs by western blot (WB). Gal4-VP16 expression vector was used as control for Gal4–PLZF, and pcDNA3 for the HDACs. PLZF expression was monitored using antiGal4, HDAC4 with anti-Flag, and HDAC5 and 6 with anti-HA antibody. Input: 1% of amount used for IP. (f, g) CV-1 cells were transfected with either the pSG5 control vector or the full-length PLZF, together with tagged HDACs. WCE were immunoprecipitated with anti-PLZF antibody. HDACs were assayed as in (d, e). PLZF expression was monitored using the anti-PLZF antibody. Input: 2% of amount used for IP

levels similar to their wild-type counterparts, the mutants have no detectable enzymatic activity, directly measured after immunoprecipitation from transfected COS-7 cells (Figure 4b), but they do retain the ability to interact with PLZF (1–456) in GST pulldown (Figure 4c). Interaction between PLZF and HDAC4 KL was stronger than between PLZF and HDAC1 AA, mirroring results with the wild-type proteins (Figure 1b). Moreover, PLZF interacts strongly with the ectopically expressed HDAC4 KL mutant in live cells in co-immunoprecipitation experiments (not shown).

NIH 3T3 cells were transfected with a Gal4-driven luciferase reporter vector (Figure 4d), an expression vector for Gal4–PLZF, and increasing amounts of expression vectors for the HDAC-dead mutants. Neither mutant influenced Gal4–PLZF expression. Repression by PLZF was affected in a dose-dependent manner by expression of HDAC4 KL but not HDAC1 AA, suggesting that HDAC4 activity is required for repression by PLZF (Figure 4e). To confirm this hypothesis, HDACs were knocked down by RNA interference (Elbashir et al., 2001): cells were transfected with siRNAs directed against HDAC4 Oncogene

PLZF and HDAC4 A Chauchereau et al

8780

was observed in the absence of PLZF (not shown). A similar effect was previously observed in the presence of the HDAC inhibitor TSA (David et al., 1998) and is likely to be attributable to the PLZF homolog FAZF, which recognizes the same DNA target sequences (Hoatlin et al., 1999); the effectiveness of HDAC4 siRNA in counteracting Gal4–PLZF-mediated transcription confirms the implication of HDAC4 in the transcriptional repression mediated by PLZF. PLZF thus recruits class II HDACs. Since RAR are known to recruit class I HDACs (Heinzel et al., 1997; Nagy et al., 1997), we investigated the role of HDAC4 and HDAC1 in repression by PLZF–RARa, using a retinoic acid-responsive reporter gene, RARE-TK-CAT (Figure 5d). Repression by PLZF–RARa was strongly reduced by downregulating HDAC1 or 4. Interestingly, release was even better in the presence of HDAC siRNAs than with an siRNA directed against PLZF– RARa (P–R), most likely due to concomitant release of endogenous RAR-mediated repression.

Discussion

Figure 2 PLZF and PLZF–RARa influence subcellular localization of HDAC4. (a) NIH 3T3 cells were transfected with myctagged HDAC4 together with 0.5 mg of the indicated expression vectors 24 h prior to immunostaining. Merged images of myc (red) and Hoechst (blue) immunofluorescence. Proportion of cells containing nuclear myc-HDAC4 (either in the nucleus only, or in both nucleus and cytoplasm); mean7s.e. of three independent experiments, 100–150 cells per experiment. (b) CV-1 cells were transfected with myc-tagged HDAC4 (1 mg) together with a fivefold excess of the indicated expression vectors. HDAC4 was revealed after 48 h by immunostaining with an anti-myc antibody (red). Proportion of cells containing nuclear myc-HDAC4 was counted as in (a)

or HDAC1, or with a control siRNA. Effectiveness of the specific siRNAs was first documented by Western blot analysis (Figure 5a). Next, reporter assays based on Gal4–PLZF fusion protein (Figure 5b) or direct recognition of LexA sequence by full-length PLZF (Figure 5c) were used to monitor repression; in this last assay, repression was dependent on PLZF since it was released by siRNA-mediated knockdown of the protein (not shown); silencing of HDACs was verified in each experiment by Western blot (not shown). Knockdown of endogenous HDAC4 alleviated repression by PLZF in both assays, confirming the participation of HDAC4 in the repression process. In contrast, repression was only partially alleviated by HDAC1 knockdown. A low but reproducible effect of HD1 and HD4 siRNAs on the transcription release of the 80P-SV-Luc reporter gene Oncogene

Our GST-pull-down and co-immunoprecipitation data show that PLZF interacts with class II HDACs – most strongly with HDAC4 – but only HDAC4 colocalizes with endogenous PLZF in haematopoietic MDS cells. Though class I HDACs also interact with PLZF, the interaction with HDAC4 is stronger than with HDAC1 in vitro, and more readily detectable in live cells. Moreover, functional analyses using enzymatically dead mutants or RNA interference showed that HDAC4 is required for transcriptional repression by PLZF and PLZF–RARa. Inhibition of HDAC1 only mildly influences PLZF-mediated repression, but dramatically affects PLZF–RARa-mediated repression, most likely by interfering with repression mediated by the RAR moiety (see below). Taken together, our data suggest that the repressive activity of wild-type PLZF is mediated mainly by the class II histone deacetylase HDAC4. The POK protein family transcriptional repressors PLZF and Bcl-6 are both implicated in cell differentiation in the haematopoietic lineage, and both can recruit class II HDACs. Indeed, it has been speculated (Lemercier et al., 2002) that recruitment of class II HDACs to repress transcription could be a common mechanism for this family of proteins. Evidence is accumulating that class II HDACs help mediate the transcriptional repression of genes crucial for differentiation, as documented for the serum response factor (SRF) (Davis et al., 2003), MEF-2 in myogenesis (Lu et al., 2000; McKinsey et al., 2000; Dressel et al., 2001), or GATA-1 in erythroid differentiation (Watamoto et al., 2003), and our results suggest that HDAC4 could play a key role in PLZF-mediated myeloid differentiation. Interestingly, in contrast to MEF-2, PLZF does not seem to interact with the N-terminal region of HDAC4.

PLZF and HDAC4 A Chauchereau et al

8781

Figure 3 Colocalization between PLZF and HDAC4 in haematopoietic cells. (a) MDS cells were treated (or not) with 400 nM of calcium ionophore (CaI) A23187 (Sigma) for 48 h to induce synthesis of PLZF protein, before double staining with anti-HDAC4 (green) and anti-PLZF (red) antibodies. Bars: 10 mm. The histogram shows the proportion of PLZF-containing cells (mean7s.d., n>100 cells). (b) MDS cells were double-labelled using polyclonal anti-HDAC1 or anti-HDAC4 (green) and monoclonal anti-PLZF (red) antibodies, and colocalization (yellow) determined by confocal microscopy. Bars, 10 mm. (c) MDS cells at higher magnification. Colocalization (white) between PLZF (red) and HDAC4 (green) was excluded from the large PLZF-containing nuclear dots

Although Bcl-6 and PLZF belong to the same POK protein family, Bcl-6 interacts preferentially with HDAC5; the interacting domain corresponds to the Cterminal zinc-finger domains of Bcl-6 (Lemercier et al., 2002), absent from some of the PLZF constructs used in our study. In fact, a strong interaction was observed in vitro between HDACs and the N-terminal portion of PLZF, consistent with mediation via its N-terminal repression domains. Both class I and class II HDACs interact with PLZF, suggesting that they could coexist in the same PLZF complex. Interactions between class I and class II HDACs have been reported previously, in particular between HDAC3 and 4 (Fischle et al., 2002). Bcl-6 has also been shown to recruit both class I and class II HDACs (Dhordain et al., 1997; David et al., 1998; Bereshchenko et al., 2002). The reasons for the

redundancy of HDAC activities associated with PLZF and Bcl-6 are unclear. Given that PLZF is a pleiotropic transcription factor, one may speculate that coexistence of a specific subset of HDACs allows PLZF to integrate signals from different cellular pathways, and to regulate distinct target genes. As for the PLZF–RARa fusion protein, which also recruits both HDAC1 and HDAC4, inhibiting HDAC1 markedly affected repression, probably due to the association of HDAC1 with SMRT/Sin3A complexes targeting the RAR moiety, as described previously (David et al., 1998). Thus, targeting HDAC1 with siRNA would disrupt repressor complexes interacting with both the PLZF and the RAR components of the fusion protein. Together, the data suggest that HDAC4 is most strongly associated with PLZF, whereas HDAC1 seems to be more important in the Oncogene

PLZF and HDAC4 A Chauchereau et al

8782

Figure 4 Release of PLZF-mediated repression by a dominantnegative mutant of HDAC4. (a) Schematic structure of wild-type and mutant HDAC1 and HDAC4 proteins. (b) HDAC1 AA and HDAC4 KL show no detectable enzymatic activity. COS-7 cells were transfected as indicated and lysed after 36 h. Tagged proteins were immunoprecipitated with an anti-Flag antibody, standardized by WB and tested for deacetylase activity. (c) HDAC1 AA and HDAC4 KL mutants interact with PLZF1 456 in vitro (GST-pulldown as in Figure 1). (d) Schematic representation of the Gal– PLZF system. (e) HDAC4 KL, but not HDAC1 AA, alleviates repression by PLZF. NIH 3T3 cells were transfected with LexA(Gal4)5X-Luc (250 ng) and CMV b-gal (5 ng) reporter vectors, together with expression vectors for LexA-VP16 (5 ng), Gal4DBD– PLZF (100 ng) and increasing amounts of HDAC1 AA or HDAC4 KL (0–400 ng). Repression by PLZF was calculated as the ratio of luciferase activity observed with GALDBD to that observed with GALDBD–PLZF, after normalization for b-galactosidase activity. Repression by Gal4–PLZF in the absence of HDAC mutant was assigned the value 1; mean7s.e. of three independent experiments

RAR-mediated repression. Alternatively, one might hypothesize that the PLZF–RARa fusion protein, due to its abnormal conformation, recruits an as yet Oncogene

Figure 5 Release of PLZF- and PLZF–RARa-mediated repression by an siRNA directed against HDAC4. (a) Targeted inhibition of HDAC expression by siRNA. HeLa cells transfected with 2 mg of control (c), HDAC1 (HD1) or HDAC4 (HD4) siRNA were analysed by WB. (b) Effect of HDAC4 siRNA on Gal4–PLZFmediated repression. NIH 3T3 cells transfected with 1 mg of control (c) or HDAC4 (HD) siRNA were transfected again 24 h later with the reporter vector (0.5 mg), expression vectors for LexA-VP16 (5 ng) and increasing amounts of Gal4DBD–PLZF. (c) Effect of HDAC siRNA on repression by full-length PLZF. HeLa cells were transfected with the 8OP-SV-Luc reporter and an expression vector for PLZF, along with 2 mg of control (c), HDAC1 (HD1) or HDAC4 (HD4) siRNA. Mean7s.e. of five independent experiments. (d) Effect of HDAC siRNA on PLZF–RARa-mediated repression. HeLa cells were transfected with RARE-TK-CAT reporter plasmid, an expression vector for PLZF–RARa, and 2 mg of indicated siRNA(s): control (C), PLZF–RARa (P–R), HD1, HD4, or a 1 : 1 mixture of HD1 and HD4 siRNAs. In (c) and (d), repression by PLZF in presence of control was assigned the value 100%

uncharacterized repressor complex containing both class I and II HDAC activities. Unlike leukaemic patients whose cells express the PML–RARa fusion protein, those whose cells express PLZF–RARa do not respond to retinoic acid treatment. Our results show that functionally interfering with the HDAC4 component of this complex strongly compromises its repressive activity, and so provides proof of principle for designing novel antileukaemic therapies. Moreover, HDAC4 is a key determinant of the DNA damage response pathway (Kao et al., 2003), so that

PLZF and HDAC4 A Chauchereau et al

8783

blocking of HDAC4 function in combination with conventional therapy to maximize the killing of cancer cells could be a viable strategy for APL treatment.

Materials and methods Plasmids and oligonucleotidess pSG424-GAL41 147, pSG424pGEX-KG-PLZF1 456 aa, GAL4–PLZF1 456, pSG5-PLZF1 673-HA and 8OP-SV-Luc constructs were described previously (David et al., 1998), as were pSG5-PLZF–RARa (Hong et al., 1997) and RARE-TKCAT (Durand et al., 1992). XJ Yang provided Flag-tagged HDAC4 and HDAC4KL vectors (Wang et al., 1999), and S Khochbin, pcDNA3-HDAC5-HA and pcDNA3-HDAC6-HA vectors (Kao et al., 2001; Lemercier et al., 2002). pBJ5HDAC1 AA was derived from pBJ5-HDAC1 (Taunton et al., 1996) by PCR-mediated mutagenesis. The HD4 mutant xpHD4-(1–208) was provided by Z Wu (Chan et al., 2003). Construction details for pBJ5, pT7-HDAC1, pT7-HDAC1AA, pGEMT-HDAC2, pGEMT-HDAC3 and pcDNA3(LexA)4x-(Gal4)5x-LUC are available on request. SiRNA duplexes purchased from Genset Oligos (Proligo, France) were (sense strands): C, agggaggcaaacauugagactt; HDAC1, aagccucaccgaauccgcatt; HDAC4, aauguacgacgccaaagautt; PLZF– RARa, gcucauucagccauugagacctt. In vitro translation and GST pull-down GST and GST-PLZF proteins expressed in Escherichia coli were purified on glutathione-agarose beads (Sigma). 35Smethionine-labelled HDAC proteins were synthesized in vitro using the TNT system (Promega). GST-pull-down assays were performed as reported previously (Groisman et al., 1996). Bound proteins were extensively washed, fractionated on SDS–PAGE gels, visualized by autoradiography and quantified using NIH Image software. Transfections NIH 3T3, CV-1 and HeLa cells were maintained in DMEM (GIBCO) supplemented with 10% foetal calf serum (FCS, PAN), and MDS cells, in MEM alpha medium (GIBCO) with 10% FCS. For immunoprecipitation, CV-1 cells transfected with 5 mg of each of the appropriate expression vectors using PolyFect (Qiagen) were lysed after 36 h. CMV b-gal vector served as transfection control, and total DNA was adjusted using empty constructs. For luciferase assays, transfected NIH 3T3 cells were lysed after 36 h. Luciferase activity was assayed

with Promega reagents, and b-galactosidase with a kit from Tropix (Bedford, MA, USA). For RNA interference, NIH 3T3 cells were transfected with siRNA oligonucleotides using Lipofectamine (Invitrogen). HeLa cells were transfected with 330 ml of calcium-phosphate precipitate containing reporter construct 8OP-SV-Luc (10 mg/ml) or RARE-TK-CAT (5 mg/ ml), PLZF (0.5 mg/ml) or PLZF–RARa (1 mg/ml) expression vector, siRNA (as indicated) and CMV b-gal (100 ng/ml), adjusted with herring sperm DNA to a total concentration of 20 mg/ml. Extracts were prepared 48 h post-transfection and processed for CAT (CAT Elisa kit, Roche) or luciferase activity, normalized on b-galactosidase activity. Antibodies Primary antibodies were: M2 anti-Flag (F-3165, Sigma), antiGal4DBD (RK5C1, Santa Cruz), anti-myc (9E10, Santa Cruz), anti-HA (12CA5, Boehringer), anti-tubulin (Sigma), antiHDAC1 (Cell Signaling), anti-HDAC4 (A Dejean for Western blot; Santa Cruz H-92 for immunofluorescence), anti-RARa (provided by P Chambon) and anti-PLZF (Oncogene). Deacetylase assays These were performed in duplicate as described previously (Magnaghi-Jaulin et al., 1998). Immunocytochemistry NIH 3T3 cells were transfected with the appropriate expression vectors, fixed in 4% formaldehyde and permeabilized in 0.5% Triton X-100. For colocalization experiments, MDS cells were grown in fibronectin-precoated (5 mg/ml) Labteck chambers, washed with PBS and fixed as above, then incubated overnight at 41C with primary antibodies in the presence of 1% goat serum. Primary antibodies were detected by Alexa 488- and/or Cy3-labelled secondary antibodies (1/400); nuclei were counterstained with Hoechst 33258. Images acquired with a Leica TCS laser-scanning confocal microscope were analysed using ImageJ software (NIH, USA). Acknowledgements We thank XJ Wang for providing HDAC4 expression vectors, S Khochbin for HDAC5 and 6 expression vectors, W-M Yang for HDAC2 and 3 expression vectors, Z Wu for xp-HD4-(1– 208) expression vector and P Chambon for RARE-TK-CAT and anti-RARa antibodies. This work was supported by grants from ARC and the European 5th FP (grant QLG1-199900866). MM was supported by the CNRS, RF by the Comite´ du Val d’Oise de la Ligue contre le Cancer.

References Ahmed N, Williams JF and Weidemann MJ. (1993). Biochem. Mol. Biol. Int., 29, 1055–1067. Bardwell VJ and Treisman R. (1994). Genes Dev., 8, 1664–1677. Barna M, Hawe N, Niswander L and Pandolfi PP. (2000). Nat. Genet., 25, 166–172. Bereshchenko OR, Gu W and Dalla-Favera R. (2002). Nat. Genet., 32, 606–613. Borghi S, Molinari S, Razzini G, Parise F, Battini R and Ferrari S. (2001). J. Cell Sci., 114, 4477–4483. Chan JK, Sun L, Yang XJ, Zhu G and Wu Z. (2003). J. Biol. Chem., 278, 23515–23521.

Chen SJ, Zelent A, Tong JH, Yu HQ, Wang ZY, Derre J, Berger R, Waxman S and Chen Z. (1993). J. Clin. Invest., 91, 2260–2267. David G, Alland L, Hong SH, Wong CW, Depinho RA and Dejean A. (1998). Oncogene, 16, 2549–2556. Davis FJ, Gupta M, Camoretti-Mercado B, Schwartz RJ and Gupta MP. (2003). J. Biol. Chem., 278, 20047–20058. de Ruijter AJ, van Gennip AH, Caron HN, Kemp S and van Kuilenburg AB. (2003). Biochem. J., 370, 737–749. Dhordain P, Albagli O, Lin RJ, Ansieau S, Quief S, Leutz A, Kerckaert JP, Evans RM and Leprince D. (1997). Proc. Natl. Acad. Sci. USA, 94, 10762–10767. Oncogene

PLZF and HDAC4 A Chauchereau et al

8784 Dressel U, Bailey PJ, Wang SC, Downes M, Evans RM and Muscat GE. (2001). J. Biol. Chem., 276, 17007–17013. Durand B, Saunders M, Leroy P, Leid M and Chambon P. (1992). Cell, 71, 73–85. Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K and Tuschl T. (2001). Nature, 411, 494–498. Fischle W, Dequiedt F, Hendzel MJ, Guenther MG, Lazar MA, Voelter W and Verdin E. (2002). Mol. Cell, 9, 45–57. Fischle W, Kiermer V, Dequiedt F and Verdin E. (2001). Biochem. Cell Biol., 79, 337–348. Grignani F, Dematteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M, Fanelli M, Ruthardt M, Ferrara FF, Zamir I, Seiser C, Grignani F, Lazar MA, Minucci S and Pelicci PG. (1998). Nature, 391, 815–818. Groisman R, Masutani H, Leibovitch MP, Robin P, Soudant I, Trouche D and Harel-Bellan A. (1996). J. Biol. Chem., 271, 5258–5264. Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and Zelent A. (1998). Blood, 91, 2634–2642. Hassig CA, Fleischer TC, Billin AN, Schreiber SL and Ayer DE. (1997). Cell, 89, 341–347. Hassig CA, Tong JK, Fleischer TC, Owa T, Grable PG, Ayer DE and Schreiber SL. (1998). Proc. Natl. Acad. Sci. USA, 95, 3519–3524. Heinzel T, Lavinsky RM, Mullen TM, Soderstrom M, Laherty CD, Torchia J, Yang WM, Brard G, Ngo SD, Davie JR, Seto E, Eisenman RN, Rose DW, Glass CK and Rosenfeld MG. (1997). Nature, 387, 43–48. Hoatlin ME, Zhi Y, Ball H, Silvey K, Melnick A, Stone S, Arai S, Hawe N, Owen G, Zelent A and Licht JD. (1999). Blood, 94, 3737–3747. Hong SH, David G, Wong CW, Dejean A and Privalsky ML. (1997). Proc. Natl. Acad. Sci. USA, 94, 9028–9033. Kao GD, McKenna WG, Guenther MG, Muschel RJ, Lazar MA and Yen TJ. (2003). J. Cell. Biol., 160, 1017–1027. Kao HY, Verdel A, Tsai CC, Simon C, Juguilon H and Khochbin S. (2001). J. Biol. Chem., 276, 47496–47507. Lemercier C, Brocard MP, Puvion-Dutilleul F, Kao HY, Albagli O and Khochbin S. (2002). J. Biol. Chem., 277, 22045–22052. Licht JD, Shaknovich R, English MA, Melnick A, Li JY, Reddy JC, Dong S, Chen SJ, Zelent A and Waxman S. (1996). Oncogene, 12, 323–336. Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH and Evans RM. (1998). Nature, 391, 811–814.

Oncogene

Lu J, McKinsey TA, Zhang CL and Olson EN. (2000). Mol. Cells, 6, 233–244. Magnaghi-Jaulin L, Groisman R, Naguibneva I, Robin P, Lorain S, Le Villain JP, Troalen F, Trouche D and HarelBellan A. (1998). Nature, 391, 601–605. McConnell MJ, Chevallier N, Berkofsky-Fessler W, Giltnane JM, Malani RB, Staudt LM and Licht JD. (2003). Mol. Cell. Biol., 23, 9375–9388. McKinsey TA, Zhang CL, Lu J and Olson EN. (2000). Nature, 408, 106–111. Melnick A, Carlile G, Ahmad KF, Kiang CL, Corcoran C, Bardwell V, Prive GG and Licht JD. (2002). Mol. Cell. Biol., 22, 1804–1818. Miska EA, Langley E, Wolf D, Karlsson C, Pines J and Kouzarides T. (2001). Nucleic Acids Res., 29, 3439–3447. Nagy L, Kao HY, Chakravarti D, Lin RJ, Hassig CA, Ayer DE, Schreiber SL and Evans RM. (1997). Cell, 89, 373–380. Nanba D, Mammoto A, Hashimoto K and Higashiyama S. (2003). J. Cell Biol., 163, 489–502. Reid A, Gould A, Brand N, Cook M, Strutt P, Li J, Licht J, Waxman S, Krumlauf R and Zelent A. (1995). Blood, 86, 4544–4552. Senbonmatsu T, Saito T, Landon EJ, Watanabe O, Price Jr E, Roberts RL, Imboden H, Fitzgerald TG, Gaffney FA and Inagami T. (2003). EMBO J., 22, 6471–6482. Shaknovich R, Yeyati PL, Ivins S, Melnick A, Lempert C, Waxman S, Zelent A and Licht JD. (1998). Mol. Cell. Biol., 18, 5533–5545. Taunton J, Hassig CA and Schreiber SL. (1996). Science, 272, 408–411. Verdel A and Khochbin S. (1999). J. Biol. Chem., 274, 2440– 2445. Wang AH, Bertos NR, Vezmar M, Pelletier N, Crosato M, Heng HH, Th’ng J, Han J and Yang XJ. (1999). Mol. Cell. Biol., 19, 7816–7827. Watamoto K, Towatari M, Ozawa Y, Miyata Y, Okamoto M, Abe A, Naoe T and Saito H. (2003). Oncogene, 22, 9176–9184. Wolffe AP and Guschin D. (2000). J. Struct. Biol., 129, 102– 122. Wong CW and Privalsky ML. (1998). J. Biol. Chem., 273, 27695–27702. Zelent A, Guidez F, Melnick A, Waxman S and Licht JD. (2001). Oncogene, 20, 7186–7203.