Avian Erythroleukemia: a Model for Corepressor Function in Cancer

Avian erythroleukemia: a model for corepressor function in cancer

Luc EG Rietveld1,2, Eric Caldenhoven1,2 and Hendrik G Stunnenberg*,1

1Department of Molecular Biology, NCMLS, Geert Grooteplein Zuid 26, PO Box 9101 6500 HB Nijmegen, The Netherlands

Transcriptional regulation at the level of chromatin plays of chromatin, such as inhibitors of methylation or crucial roles during eukaryotic development and dier- histone acetylation (Cameron et al., 1999). Thus, entiation. A plethora of studies revealed that the altering the state of chromatin may be an important acetylation status of histones is controlled by multi- tool in a cure for cancer. protein complexes containing (de)acetylase activities. In A link between (de)acetylation of histones and the current model, histone deacetylases and acetyltrans- transcriptional regulation was proposed several decades ferases are recruited to chromatin by DNA-bound ago (Allfrey et al., 1964). The identi®cation of proteins repressors and activators, respectively. Shifting the that can modulate the acetylation state of histones balance between deacetylation, i.e. repressive chromatin provided further evidence for the link between and acetylation, i.e. active chromatin can lead to chromatin structure and gene regulation. The extent aberrant gene transcription and cancer. In human acute of histone acetylation is determined by the opposing promyelocytic leukemia (APL) and avian erythroleuke- action of histone acetyltransferases (HATs) and histone mia (AEL), chromosomal translocations and/or muta- deacetylases (HDACs) (reviewed by Kouzarides, 1999). tions in nuclear hormone receptors, RARa [NR1B1] and In a simplistic view, these enzymes are recruited to TRa [NR1A1], yielded oncoproteins that deregulate chromatin by DNA-bound transcription factors there- transcription and alter chromatin structure. The onco- by (locally) modifying the histone tails and conse- genic receptors are locked in their `o' mode thereby quently the chromatin structure. Transcriptional constitutively repressing transcription of genes that are regulators that recruit histone deacetylases may control critical for dierentiation of hematopoietic cells. AEL tumour development at dierent stages and can grossly involves an oncogenic version of the chicken TRa,v- be divided into three dierent categories. Firstly, ErbA. Apart from repression by v-ErbA via recruitment transcriptional repressors that control cell-cycle and of corepressor complexes, other repressors and corepres- proliferation such as Mad and Rb. Secondly, methyl- sors appear to be involved in repression of v-ErbA target binding proteins that mediate transcriptional silencing genes, such as carbonic anhydrase II (CAII). Reactiva- of genes which may be critical in the formation of tion of repressed genes in APL and AEL by chromatin certain cancers. Thirdly, transcriptional regulators such modifying agents such as inhibitors of histone deacety- as TR and RAR that control dierentiation and lase or of methylation provides new therapeutic strategies development (reviewed in Cress and Seto, 2000). in the treatment of acute myeloid leukemia. Oncogene This review focuses on the role of nuclear hormone (2001) 20, 3100 ± 3109. receptors (NR) in cancer. Aberrant nuclear receptor function due to a hormonal imbalance or receptor Keywords: Erythroleukemia; corepressor; v-ErbA; his- mutations has been strongly correlated with disease tone deacetylase; carbonic anhydrase II; methylation and cancer (reviewed in Forrest et al., 1996; Goldhirsch and Gelber, 1996). Thyroid hormone receptor (TR) and retinoic acid receptor (RAR) are both class II Introduction nuclear receptors which are ligand-controlled transcription factors that govern transcription of a multitude of The mechanisms by which oncoproteins deregulate target genes by recruiting multicomponent cofactor transcription of speci®c target genes during oncogenic complexes. In the absence of ligand, nuclear hormone transformation are still poorly understood. Mutations, receptors recruit corepressor complexes containing deletions and fusions resulting from chromosomal HDAC activity, whereas in the presence of ligand they translocations in cellular or viral (proto-)oncogenes recruit coactivator proteins with HAT activity and/or have altered their cognate protein functions, causing multicomponent coactivator complexes. In human aberrant gene regulation and inducing cancer. Recent promyelocytic leukemia (APL), chromosomal translo- studies show that reactivation of genes silenced in cations lead to fusion proteins involving RARa and cancer can be accomplished by treatment of trans- dierent fusion partners, such as PML, PLZF, NPM formed cells with compounds that aect the structure and NUMA (reviewed in Redner et al., 1999). APL- patients carrying the PML-RARa fusion can be cured by treatment with the RARa ligand all-trans retinoic *Correspondence: HG Stunnenberg acid (ATRA) which induces terminal dierentiation of 2These authors contributed equally APL blasts (Warrell et al., 1991, 1993). PLZF ± RARa Corepressors in erythroleukemia LEG Rietveld et al 3101 patients are not responsive to ATRA, suggesting that v-ErbA and transcriptional repression dierent or additional mechanisms are involved in PLZF ± RARa- compared to PML ± RARa-induced The v-ErbA oncogene is a viral variant of chicken TRa APL. Treatment with histone deacetylase inhibitors and was fused to the viral gag gene during the genesis such as trichostatin A (TSA) plus ATRA overcomes of AEV. The way in which v-ErbA contributes to the the maturation blockade in APL induced by PLZF ± leukemic phenotype has long been elusive. The role of RARa (reviewed in Redner et al., 1999), suggesting a numerous mutations occurring throughout the receptor link between chromatin modi®cations and cancer moiety have been analysed in great detail (Figure 1, development. In chicks, the avian erythroblastosis virus reviewed in Beug et al., 1994). Due to the mutations in (AEV) causes erythroleukemia (AEL). AEV encodes an the Zn-®nger, the v-ErbA DNA-binding speci®city and oncogenic variant of the TRa, v-ErbA, and the anity have been aected. Furthermore, the ability of mutated EGF-receptor v-ErbB. Due to a number of v-ErbA to dimerise with RXR is impaired. Several mutations throughout the ligand-binding domain, v- mutations and a small N-terminal deletion also aect ErbA is unable to respond to ligand. Restoration of the ability of v-ErbA to bind ligand and to activate thyroid hormone responsiveness in erythroleukemic transcription (Munoz et al., 1988; Zenke et al., 1990; cells by overexpression of a ligand-responsive thyroid Barettino et al., 1993). hormone receptor partially restores the ability of these A decade ago, it was demonstrated that v-ErbA cells to dierentiate (Disela et al., 1991). The inability constitutively inhibits T3-mediated activation of cog- of oncogenic receptors to eectively respond to nate target genes (Sap et al., 1989; Damm et al., 1989; physiological concentrations of ligand provides a Pain et al., 1990; Zenke et al., 1990; Baniahmad et al., molecular basis of oncogenesis. Recently, a plethora 1992). From these and many other studies, v-ErbA was of molecular studies on the modulation of chromatin initially postulated to merely antagonise TR function by NR have been published that provide new and through constitutive binding to TR responsive elements sometimes contradictory explanations for the mechan- (TRE), occluding the endogenous receptor (the `bind- isms causing APL and AEL. In this review, we will ing site occlusion' model). Another step was the focus on chromatin modifying complexes and their role identi®cation of target genes that were directly in transcriptional repression in AEV-induced leukemia. regulated by v-ErbA: the erythrocyte anion transport protein gene (Band 3), the d-aminolevulinate synthase gene (ALA-S), the lysozyme gene and the carbonic AEV and leukemia anhydrase II gene (CAII) (Zenke et al., 1988; Pain et al., 1990; Baniahmad et al., 1990). Notwithstanding the AEV is an acute leukemogenic retrovirus that causes identi®cation of target genes, progress in elucidating fatal erythroleukemia in young chicks. The target cell the mechanisms by which v-ErbA silences transcription for AEV is the committed erythroid progenitor remained slow. A further piece of the puzzle was (Samarut and Gazzolo, 1982). In culture-transformed provided by Damm and Evans (1993). The authors erythroblast cells called HD3 are arrested in their showed that the transformation-defective AEV-mutant ability to terminally dierentiate and show an td359 encodes a v-ErbA variant that failed to suppress increased rate of proliferation (Graf and Beug, 1983). basal transcription and exhibited an impaired ability to AEV-transformed HD3 cells provide an excellent antagonise TR. One of two mutations speci®c for the model system to investigate the molecular basis of td359 variant (Figure 1, Pro144 Arg in the `hinge' the disease. Besides the v-ErbA oncoprotein, the AEV region), abolished repressive functions but not hor- virus encodes a second oncoprotein, v-ErbB, a mutated mone binding and activation properties when intro- constitutively active transmembrane receptor for epi- duced in the context of wild type TR. Thus, active dermal growth factor (EGF) (Downward et al., 1984; repression in addition to TR occlusion appeared to be Sap et al., 1986). v-ErbA has no transformation an important mechanism involved in AEV-induced capacity on its own, unless it is expressed at extremely oncogenesis. high levels (Casini and Graf, 1995). Co-expression of v- ErbA with v-ErbB does, however, lead to a much more severe leukemic phenotype compared to expression of Corepressor complexes v-ErbB alone (Miles and Robinson, 1985; Kahn et al., 1986; Forrest et al., 1990). The current view is that v- The recent observation that unliganded nuclear recep- ErbA cooperates with ligand-activated tyrosine kinases tors can `actively' repress transcription by recruiting such as stem cell factor (SCF)-activated c-Kit and multi-protein corepressor complexes has been a mile- TGF-loaded c-ErbB or constitutively active v-ErbB to stone (reviewed in Hu and Lazar, 2000; Burke and eectively arrest dierentiation and to induce leukemia Baniahmad, 2000). The repressive complexes are (reviewed in Beug et al., 1996). Recent studies suggest recruited to the DNA by unliganded nuclear hormone that v-ErbA mimics unliganded TRa in co-operating receptors and other speci®c DNA-binding proteins. with ligand-activated c-Kit to cause steroid-indepen- Yeast-two hybrid screens for proteins interacting with dent long-term proliferation and a block in dierentia- unliganded nuclear receptors led to the identi®cation of tion of primary erythroid progenitor cells (Bauer et al., the corepressor molecules NCoR (nuclear receptor 1998). corepressor) (Horlein et al., 1995; Kurokawa et al.,

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3102

Figure 1 Comparison of the domain structure of the thyroid hormone receptor c-ErbA/TRa, with the oncogenic, mutated receptor v-ErbA and the transformation de®cient td359 variant. Highlighted are the DNA-binding and ligand binding domains and the activation function AF-2. v-ErbA contains an virus derived gag sequence, N- and C-terminal deletions and 13 point mutations (solid dots). The transformation-defective AEV-mutant td359 harbours a v-ErbA variant with two additional mutations, including the mutation Pro144 Arg in the `hinge' region

1995) and the highly related SMRT (Silencing Mediator transcriptional repressor, RPD3 (Vidal and Gaber, for Retinoic acid and Thyroid hormone receptors) 1991). The identi®cation of a human histone deacety- (Chen and Evans, 1995; Sande and Privalsky, 1996). lase with *60% identity to RPD3 linked histone Subsequent studies suggested that NCoR/SMRT func- deacetylases with Sin3A/Sin3B (Taunton et al., 1996). tions as corepressors not only for unliganded nuclear These and other eorts led to the identi®cation of a receptors, but for many other unrelated transcription multiprotein complex containing Sin3A/B and factors including, PLZF (Hong et al., 1997; Grignani et HDAC1/2, RbAP48 and 746, and two Sin-interacting al., 1998), BCL6 ± POZ (Dhordain et al., 1998) and protein SAP18 and 730 (Table 1, Hassig et al., 1997; MyoD (Bailey et al., 1999). While NCoR is a 270 kDa Kadosh and Struhl, 1997; Laherty et al., 1997; Zhang protein, SMRT was initially reported to be a 170 kDa et al., 1997). RbAp46/48, originally identi®ed to protein but recently a 270 kDa splice variant of SMRT associate with Rb, are of particular interest as they was identi®ed (Ordentlich et al., 1999, Park et al., 1999). can bind to helix 1 of histone H4 in vitro (Verreault et The 270 kDa SMRT protein has a high overall al., 1998). Concomitant with the identi®cation of the homology with NCoR; both possess a bipartite Sin3/HDAC complex, other studies reported on structure with at least three autonomous N-terminal protein-protein interactions between the Sin3/HDAC repressor domains and two independent C-terminal complex and the nuclear corepresssor NCoR/SMRT nuclear receptor interaction domains (RID) (reviewed (Heinzel et al., 1997; Nagy et al., 1997; Alland et al., by Hu and Lazar, 2000). NCoR/SMRT-mediated 1997). Finally, the Sin3/HDAC complex also partici- repression pathways have been suggested to play a pates in mediating repression by the methyl-CpG pathophysiological role in APL (Grignani et al., 1998; binding protein, MeCP2 (Nan et al., 1998; Jones et He et al., 1998), in AML (Gelmetti et al., 1998; al., 1998), suggesting a direct link between methylation Lutterbach et al., 1998) as well as in thyroid hormone and deacetylation. resistance syndrome (Safer et al., 1998). In addition to Recently, it was shown that the class II histone NCoR/SMRT, two other corepressors have been deacetylases HDAC4, HDAC5 and HDAC7 directly identi®ed that directly interact with unliganded nuclear interacted with NCoR/SMRT devoid of Sin3 or receptors. SUN-CoR is a highly basic 16 kDa nuclear HDAC1/2 proteins (Kao et al., 2000; Huang et al., protein that interacts with RevErb [NR1D1] as well as 2000). HDAC4, 75 and 77 share homology with the with TR in vitro (Zamir et al., 1996, 1997). Interestingly, yeast HDA1 protein (Fischle et al., 1999; Grozinger et SUN-CoR can also directly bind to NCoR/SMRT al., 1999; Verdel and Khochbin, 1999) and contain a suggesting a role for this protein in facilitating NCoR- non-catalytic N-terminal domain which may execute mediated repression. Recently, a TR-interacting protein additional functions. HDAC6 is unique because it called Alien has been identi®ed that is unrelated to contains two deacetylase domains that are less NCoR/SMRT (Dressel et al., 1999; Polly et al., 2000). conserved compared to other class II HDACs RAR has been reported not to interact with Alien, (Grozinger et al., 1999; Zhou et al., 2000; Kao et al., indicating that Alien displays a receptor speci®city 2000). The ®nding that class II HDACs are dieren- distinct from that of NCoR/SMRT. tially expressed suggests that they are not redundant New insights into the mechanism of transcriptional but rather have distinct physiological functions repression in mammalian cells came from the discovery (Grozinger et al., 1999; Verdel and Khochbin, 1999; that the mammalian homologues of yeast Sin3p, the Kao et al., 2000). Although it has been shown that Sin3A and Sin3B proteins, are recruited by the several HDACs can directly or indirectly interact with transcriptional repressor Mad (Ayer et al., 1995). In NCoR/SMRT in vitro, immunopuri®ed NCoR/SMRT yeast, the Sin3p repressor was reported to act as a complexes contained predominantly HDAC3 (Table 1, global repressor and genetic data indicated that it was Li et al., 2000; Guenther et al., 2000; Urnov et al., a component of a pathway that included the 2000; Wen et al., 2000). The NCoR/HDAC3 complex

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3103 Table 1 Corepressor complexes Sin3/HDACa complex NCoR-1b complex NCoR-2c complex NCoR-3d complex Mi2/NuRDe complex MeCP2f complex

Sin3A NCoR NCoR NCoR Mi-2b MeCP2 SinB HDAC3 Sin3A/B HDAC3 MTA2 Sin3A/B HDAC1 BRG1 HDAC1 TBL-1 HDAC1 HDAC1 HDAC2 BAF170 HDAC2 HDAC2 HDAC2 RbAp48 BAF155 RbAp48 RbAp48 RbAp48 RbAp46 BAF47 RbAp46 RbAp46 RbAp46 SAP30 KAP SAP30 MBD3 SAP30 SAP18 SAP18 SAP18 aHassig et al., 1997; Zhang et al., 1997b; b and c Underhill et al., 2000; dLi et al., 2000 and Wen et al., 2000; eWade et al., 1999 and Zhang et al., 1999; fNan et al., 1998 and Jones et al., 1998 further contained TBL1 (Transducin b-like 1), a novel functions depending on the DNA-bound transcrip- WD40-repeat containing protein with analogy to the tional repressor. TUP1 and Groucho corepressors. TBL-1 has been reported to bind histone H3 suggesting that TBL-1 forms a bridge between NCoR/SMRT and chromatin Methylation and oncogenesis (Guenther et al., 2000). Independent evidence was obtained from Xenopus oocyte injections: unliganded v- Gene silencing by means of methylation has received a ErbA/TRa receptors were shown to interact with the lot of attention in recent years. Methylation of CpG NCoR/HDAC3 but not with the Sin3/HDAC complex residues has been tightly correlated with transcriptional (Urnov et al., 2000). Finally, two chromatographically repression and histone hypoacetylation (Siegfried et al., distinct NCoR complexes, NCoR-1 and NCoR-2 have 1999). Several studies have suggested that methylation- been identi®ed (Underhill et al., 2000). Whereas the dependent transcriptional silencing may be a critical NCoR-1 complex contained HDAC3 and the SWI- event in the etiology of cancer. A global-genome SNF related proteins BRG1, BAF170, BAF155, approach to analyse methylation of CpG islands in BAF47 and KAP, the NCoR-2 complex predominantly normal tissues and tumours suggested that methylation contained HDAC1 and HDAC2 and other components of CpG islands speci®cally occurs in tumours and bears typical of the Sin3/HDAC complex (Table 1). consequences for tumour formation (Costello et al., The ®rst compelling in vivo evidence for the role of 2000). It was estimated that in average 600 of the NCoR as a corepressor protein was obtained by roughly 45 000 CpG islands in the human genome disrupting the NCoR gene in mice (Jepsen et al., were aberrantly methylated in tumours. Other studies 2000). NCoR7/7 embryos exhibited defects in erythro- suggested that methylation of speci®c tumour suppres- cytes, thymocytes and neuronal development. NCoR7/7 sor genes, such as Rb and the cdk-inhibitors p15 and erythrocyte progenitors were impaired in de®nitive p16, may be a critical event in tumourigenesis erythropoiesis resulting in severe anemia in the mutant (reviewed by Jones and Laird, 1999). Although various embryos. Furthermore, transient transfections in mouse mechanisms may underlie methylation-dependent re- embryonic ®broblasts showed that NCoR is involved in pression, recent work has demonstrated the involve- repression mediated by speci®c nuclear receptors as ment of methyl-CpG binding proteins. MeCP2 was the well as other unrelated transcription factors. ®rst characterised methyl-CpG binding protein (Lewis HDAC1 and HDAC2 have also been detected in the et al., 1992; Meehan et al., 1992; Nan et al., 1998). Mi-2/NuRD multi-protein complex (Table 1), which MeCP2 can bind a single methylated CpG residue and contain two cancer related proteins Mi-2b and MTA2 was shown to recruit the Sin3/HDAC complex (Nan et (reviewed in Ayer, 1999; Knoep¯er and Eisenmann, al., 1998; Jones et al., 1998). MeCP2 is widely 1999). Mi-2b, originally identi®ed as a dermatomyosi- expressed and is thought to associate mainly with tis-speci®c auto-antigen, is a member of the SNF-2 dispersed methylated CpGs to prevent transcription, family of proteins and harbors ATP-dependent providing what has been referred to as a `transcrip- chromatin-remodeling activity (Boyer et al., 2000). tional noise reduction system' to the genome (Bird, Thus, Mi-2/NuRD complexes may not only be 1995). Mutations of MeCP2 lead to childhood involved in histone deacetylation but also in nucleo- neurodevelopmental disorders, the Rett-syndrome some remodeling, providing a link between both (Amir et al., 1999; Willard and Hendrich, 1999). Rett chromatin remodeling and histone acetylation. MTA2 syndrome mutations appearing in the methyl-binding is related to the metastasis-associated protein MTA1. domain of MeCP2 aect DNA-binding by these The NuRD complex also contains a 32 kDa subunit proteins and probably result in activation of normally corresponding to MBD3a/b, a member of a family of silenced genes (Ballestar et al., 2000). proteins containing methyl-CpG binding domains MeCP2 is the founding member of a family of (Wade et al., 1999; Zhang et al., 1999). The fact that methyl-binding proteins (MBD1-4) that are present in multiple distinct corepressor complexes co-exist indi- dierent repressor complexes and share a domain cates that they may perform tissue- or gene-speci®c homologous to MeCP2. MBD3 is a component of

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3104 the Mi-2/NuRD complex, involved in chromatin have a threefold higher DNMT1 expression than remodelling and histone deacetylation. The methyl- normal ®broblasts (Bakin and Curran, 1999). Inhibi- binding complex MeCP1 contains MBD2 along with tion of DNMT1 activity as well as inhibition of histone HDAC1 and 72 as well as the histone binding deacetylation resulted in reversion of fos-induced proteins RbAP46/48 (reviewed by Bird and Wole, transformation, suggesting that fos-mediated transfor- 1999). The tissue-speci®c distribution of these proteins mation requires alterations in histone deacetylation and suggests cell-type speci®c functions. DNA methylation. This in turn could suggest a CpG methylation patterns within the genome are mechanism in which oncogenes can transform cells by very stable and appear to be important for main- aecting the methylation- and/or acetylation machinery tenance of transcriptionally silenced chromatin by in the cell. methyl-binding proteins such as MeCP2. In contrast, DNA (de)methylation is remarkably dynamic during early mammalian development and tumourigenesis Repression of an erythroid enhancer (Wole et al., 1999). Alterations in the methylation status of speci®c genes can promote tumourigenesis. Unravelling repression of endogenous target genes has Little is still known about the enzymatic activities that become of utmost importance to understand the are involved in these alterations, but recent data have function of corepressor complexes in transcriptional shed some light on these processes. A DNA-demethy- repression in normal and transformed cells. Inhibitors lase has been identi®ed that is identical to the that can aect corepressor function such as AZAdC previously identi®ed methyl-binding protein MBD2b and TSA have been successfully used in clinical trials (Hendrich and Bird, 1998, Battacharya et al., 1999). to treat patients with leukemia and myelodysplastic This protein is widely expressed and may catalyse the syndromes (Lubbert, 2000; Limonta et al., 1993; Richel demethylation of a single methyl-CpG residue. De- et al., 1991). However, details on the molecular methylases antagonise the activity of another group of mechanisms underlying gene repression and the eects proteins that catalyse the addition of methyl-groups to of AZAdC and TSA on APL blasts are currently scars CpG residues, the methylases. Until recently, only one or missing. A severe drawback is the lack of well DNA-methyltransferase, DNMT1, had been cloned characterised cell systems and target genes to study the (Leonhardt et al., 1992). Disruption of the DNMT1 mechanisms of repression. AEL provides a good model gene in mice had severe eects on DNA methylation system to study corepressor function in vivo because and resulted in embryonic lethality (Li et al., 1992). the v-ErbA protein as well as v-ErbA target genes have Dnmt7/7 ES cells are, however, viable and still possess been characterised in great detail. the ability to methylate DNA de novo, suggesting the More than a decade ago, v-ErbA was postulated to presence of additional DNMT activities (Lei et al., act as a transcriptional repressor of erythroid-speci®c 1996). A second potential DNMT, DNMT2, was genes involved in dierentiation and to contribute to isolated by two dierent groups but hitherto has not leukemogenisis by eciently arresting terminal dier- been shown to possess DNA-methyltransferase activity entiation of erythroid precursor cells. Ensuing studies (Yoder and Bestor, 1998; Okano et al., 1998a). identi®ed potential v-ErbA target genes such as Band3, Recently, DNMT3a and DNMT3b were found in a ALA-S, lysozyme and the carbonic anhydrase II gene database search (Okano et al., 1998b) and shown to (CAII) (Zenke et al., 1988; Baniahmad et al., 1990; methylate hemimethylated and unmethylated DNA. Disela et al., 1991; Fuerstenberg et al., 1992). DNMT3a/b are expressed at high levels in undier- Repression of two of these potential target genes, entiated ES cells and are downregulated during ES cell band3 and CAII, was shown to be important for the v- dierentiation. Overexpression of DNMT1 and ErbA-induced malignant phenotype. DNMT3 has been reported in human tumours which Initial attempts to identify v-ErbA responsive might contribute to methylation abnormalities in elements in the CAII locus using transient transfection cancer cells. The mechanism by which DNMTs act is approaches yielded ambiguous or even con¯icting still unclear but recent data show that DNMT1 can results (Disela et al., 1991; Herman et al., 1993; Rascle participate in silencing of transcription in association et al., 1994). In an unbiased approach to map the with HDAC1 (Rountree et al., 2000; Fuks et al., 2000). regulatory regions of CAII and possibly a v-ErbA This suggests that DNMT1/HDAC1 proteins may be responsive element (VRE), two DNaseI hypersensitive required to assure that deacetylated histones are sites (HS1 and HS2) were identi®ed. Subsequent assembled on newly replicated DNA. DNMT1 would studies concentrated on HS2, that is located in the then generate methylated and hypo-acetylated chroma- second intron and harbours a VRE, a direct repeat tin, which might be maintained by recruiting methyl- with a spacer of four nucleotides, as well as three binding proteins such as MeCP2. Mutations in putative GATA-binding sites. In vivo footprint studies DNMTs, might disrupt its methyltransferase function showed that the VRE as well as two of the adjacent thereby creating demethylated chromatin resulting in GATA sites were protected in proliferating HD3 cells. histone hyperacetylation, aberrant gene activation and Transfection assays showed that the HS2 region acts as tumourigenesis. A direct correlation between oncogenic a silencer that can be turned into a potent enhancer in transformation, histone acetylation and DNMT1 v-ErbA expressing cells by mutation of the VRE, activity was found in fos-transformed ®broblasts that indicating that binding of v-ErbA to the VRE nulli®es

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3105 its intrinsic enhancer activity (Ciana et al., 1998; proliferating HD3 cells (manuscript in preparation). Braliou et al., 2000). The fact that the HS2 is Similarly, Xenopus oocyte injection studies showed transcriptionally silent notwithstanding the binding of recruitment of corepressor proteins by v-ErbA. Injected GATA-transactivators, suggested that v-ErbA inter- v-ErbA and unliganded TR recruited NCoR/HDAC3 fered with the positive activity of the erythroid GATA- to an arti®cial v-ErbA responsive promoter resulting in 1 factors bound to the enhancer. Thus, v-ErbA repression (Urnov et al., 2000). Interestingly, neither in converted an enhancer into a silencer. the oocyte system nor in HD3 cells the initially How in molecular terms v-ErbA quenches the reported interaction between NCoR and the Sin3/ activity of GATA-factors is yet unclear but several HDAC complex (Nagy et al., 1997; Heinzel et al., not necessarily mutually exclusive models can be 1997) could not be detected (Urnov et al., 2000, envisioned. The close proximity of the GATA-sites to manuscript in preparation). The use of dierent cell the VRE could imply that binding of a multi-protein systems and experimental set-up might explain the corepressor complex to v-ErbA might sterically inter- dierences in corepressor components recruited by v- fere with binding of coactivator components to GATA- ErbA and NCoR/SMRT. The observation that the 1, such as p300/CBP or FOG thereby preventing signal association of corepressor components with unliganded transduction to or recruitment of the basal machinery. nuclear receptors may be under control of speci®c A second mechanism could involve deacetylation of signalling pathways shed new light on this ambiguous GATA-1. It was shown by Boyes and coworkers that issue. Hong and Privalsky (2000) showed that the acetylation of GATA1 potentiated its transcriptional ability of TR to mediate repression via NCoR/SMRT activity and was essential for erythroid dierentiation was strongly inhibited by activation of tyrosine kinase (Boyes et al., 1998; Hung et al., 1999). The association signaling pathways, such as that originating from v- of an HDAC-containing corepressor complex with v- ErbB. The authors reached the intriguing conclusion ErbA could cause deacetylation of GATA-1 and render that SMRT function was potently inhibited by a it inactive in proliferating, undierentiated HD3 cells. mitogen-activated protein kinase (MAPK) cascade that Thirdly, the corepressor complex recruited by v-ErbA operates downstream of the constitutively active may inhibit GATA-1 or GATA-1-associated coactiva- growth factor receptor v-ErbB. SMRT itself appeared tors through direct physical contact. A similar to be a substrate for phosphorylation by protein mechanism has been postulated for inhibition of kinases operating in this MAPK pathway. Phosphor- GATA-1 function by the glucocorticoid receptor ylation of SMRT by MEKK-1 and, to a lesser extent, (Chang et al., 1993). MEK-1 inhibited the ability of SMRT to bind The emerging model for the role of v-ErbA in CAII unliganded nuclear receptors in transient transfection silencing is that of a nuclear receptor locked in its experiments. The ®nding that phosphorylation trig- repressive `o' mode by constitutive recruitment of an gered by v-ErbB abolishes the interaction of SMRT HDAC-containing corepressor complex. Supporting interaction with unliganded TR/v-ErbA is in apparent evidence came from experiments with the HDAC- contradiction with the postulated role of v-ErbA in inhibitor Trichostatin A (TSA), which caused an leukemia. In AEV-transformed HD3 cells, where v- activation of CAII transcription (Ciana et al., 1998). ErbB is constitutively active, v-ErbA is able to repress Unpublished observations from our lab showed that v- CAII transcription and to interact with the corepressor ErbA indeed recruits HDAC-activity to the silencer in protein SMRT/NCoR (EC and LR, manuscript in

Figure 2 Schematic representation of the CAII genomic locus. Only exons 1 ± 4 are indicated in black boxes. DNaseI hypersensitive sites HS1, HS2 and promoter HS (prHS) are indicated with arrow heads, the size is indicative of relative DNaseI hypersensitivity. GATA sites I and II and the VRE that are protected in vivo footprinting are indicated, as well as the distribution of CpG residues in the promoter region between 7800 and +200

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3106 preparation). The apparent discrepancies may be explained if v-ErbB activates distinct signalling pathways in dierent cell systems. In our lab, we have obtained compelling data showing that an additional pathway is in place to repress CAII transcription in HD3 cells. Treatment of HD3 cells with the methylation inhibitor AZAdC caused an activation of CAII transcription (manuscript in preparation). Inspection of the CAII locus revealed that the CAII promoter region is G/C-rich and contains a hypermethylated CpG-island (Figure 2). Preliminary data suggest that the prototypical member of the methyl-CpG binding protein family, MeCP2, may bind to the hypermethylated promoter and recruit a multi-protein corepressor complex that includes the corepressor Sin3A and the histone deacetylases HDAC1 and HDAC2 (Table 1) (Nan et al., 1998; Jones et al., 1998). The observation that both MeCP2 and nuclear hormone receptors recruit corepressor complexes is suggestive of cross-talk and synergy between the repressive complexes bound to the CAII locus (Figure 3). Protein-protein interactions between subunits in the distinct corepressor complexes ± bound at the enhancer and at the promoter ± might result in the formation of a higher-order complex, the `repressosome'. The repressosome might be stabilised by interaction between unliganded TR or v-ErbA with components of the general transcriptional machinery, such as TFIIB and TBP. Although this may apply to unliganded TR, an interaction between v-ErbA and TFIIB seems less likely because of its reported lower anity for the basal transcription factor (Urnov et al., 2000). Interactions between the corepressor proteins SMRT/NCoR, Sin3 and MeCP2 with TFIIB have also been observed (Wong and Privalsky, 1998; Kaludov and Wole, 2000; Muscat et al., 1998). Interactions of repressosome components with basal transcription factors may inhibit the activity or assembly of a transcription competent preinitiation complex, to `actively' repress transcription (Baniahmad et al., 1993; Fondell et al., 1993, 1996; Urnov et al., 2000). The multitude of possible interactions suggests that transcriptional silencing by nuclear hormone receptors involves a higher-order network of concordant physical Figure 3 Models for CAII gene repression (a) Binding of two interactions between the nuclear receptor, the core- similar corepressor complexes, a MeCP2/Sin3/HDAC complex pressor complex and the general transcriptional binding to a CpG island in the promoter and a v-ErbA/NCoR/ machinery. Sin3/HDAC complex binding to the silencer, (b) Two distinct corepressor complexes, a MeCP2/Sin3/HDAC complex and a v- Parallels between the regulation of transcription by ErbA/NCoR/HDAC3/TBL-1 complex, (c) Formation of a higher v-ErbA/TR at the chicken lysozyme as well as CAII order repressosome through physical interactions between subunits gene are striking. A silencer element 2.4 kb upstream in the corepressor complexes bound to the promoter as well as the of the lysozyme transcription start site has been silencer thereby facilitating promoter-silencer communication identi®ed. It is composed of two elements F1 and F2; F2 is a TR/v-ErbA responsive element, whereas the F1 element harbours a binding site for the CTCF protein al., 2000b). Collectively the data suggest that coopera- (reviewed in Lutz et al., 2000a). v-ErbA or unliganded tive recruitment of multiple corepressor complexes by TR cooperate with CTCF in repressing lysozyme v-ErbA and CTCF or v-ErbA and MeCP2 at the transcription. The molecular mechanism underlying lysozyme and CAII loci, respectively, may lead to this synergism is unclear, but cooperation between ecient repression. transcription corepressor complexes recruited by v- The cooperation of multiple corepressor complexes ErbA and CTCF might be involved, because also in mediating repression of genes in cancer might be a CTCF can recruit the Sin3/HDAC complex (Lutz et recurring theme and may have important implications

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3107 for the treatment of cancer. Combinatorial treatments corepressor function might provide new and powerful of cancer cells with agents that target dierent tools in a cure for cancer. corepressor complexes might proof to have important therapeutic value. Combinatorial treatment has already been successfully used in APL-patients carrying the PLZF ± RARa fusion. The patients are not responsive Acknowledgments to ATRA alone, but the addition of histone deacetylase The authors wish to thank C Logie for critical reading of inhibitors such as trichostatin A (TSA) plus ATRA can the manuscript. E Caldenhoven was supported by a KWF- cure patients (reviewed in Redner et al., 1999). Thus, a fellowship. This work was supported by a EU Biomed II `chromatin therapy' focussing on the inhibition of Network grant.

References

Alland L, Muhle R, Hou Jr H, Potes J, Chin L, Schreiber- Ciana P, Braliou GG, Demay FG, von Lindern M, Barettino Agus N and DePinho RA. (1997). Nature, 387, 49 ± 55. D, Beug H and Stunnenberg HG. (1998). EMBO J., 17, Allfrey VG, Faulkner R and Mirsky AE. (1964). Proc. Natl. 7382 ± 7394. Acad. Sci. USA., 51, 786 ± 794. Costello JF, Fruhwald MC, Smiraglia DJ, Rush LJ, AmirRE,VandenVeyverIB,WanM,TranCQ,FranckeU Robertson GP, Gao X, Wright FA, Feramisco JD, and Zoghbi HY. (1999). Nat. Genet., 23, 185 ± 188. Peltomaki P, Lang JC, Schuller DE, Yu L, Bloom®eld Ayer DE, Lawrence QA and Eisenman RN. (1995). Cell, 10, CD, Caligiuri MA, Yates A, Nishikawa R, Su Huang H, 767 ± 776. Petrelli NJ, Zhang X, O'Dorisio MS, Held WA, Cavenee Ayer DE. (1999). Trends Cell. Biol., 9, 193 ± 198. WK and Plass C. (2000). Nat. Genet., 24, 132 ± 138. Bailey P, Downes M, Lau P, Harris J, Chen SL, Hamamori Cress WD and Seto E. (2000). J. Cell. Physiol., 184, 1 ± 16. Y, Sartorelli V and Muscat GE. (1999). Mol. Endocrinol., Damm K and Evans RM. (1993). Proc. Natl. Acad. Sci. USA, 13, 1155 ± 1168. 90, 10668 ± 10672. Bakin AV and Curran T. (1999). Science, 283, 387 ± 390. Damm K, Thompson CC and Evans RM. (1989). Nature, Ballestar E, Yusufzai TM and Wole AP. (2000). Biochem- 339, 593 ± 597. istry, 39, 7100 ± 7106. DiselaC,GlineurC,BuggeT,SapJ,StenglG,DodgsonJ, BaniahmadA,HaI,ReinbergD,TsaiS,TsaiMJand Stunnenberg H, Beug H and Zenke M. (1991). Genes Dev., O'Malley BW. (1993). Proc. Natl. Acad. Sci. USA, 90, 5, 2033 ± 2047. 8832 ± 8836. Dhordain P, Lin RJ, Quief S, Lantoine D, Kerckaert JP, Baniahmad A, Steiner C, Kohne AC and Renkawitz R. Evans RM and Albagli O. (1998). Nucl. Acids Res., 26, (1990). Cell, 61, 505 ± 514. 4645 ± 4651. Baniahmad A, Kohne AC and Renkawitz R. (1992). EMBO Downward J, Yarden Y, Mayes E, Scrace G, Totty N, J., 11, 1015 ± 1023. Stockwell P, Ullrich A, Schlessinger J and Water®eld MD. BarettinoD,BuggeTH,BartunekP,VivancoRuizMD, (1984). Nature, 307, 521 ± 527. Sonntag-Buck V, Beug H, Zenke M and Stunnenberg HG. Dressel U, Thormeyer D, Altincicek B, Paululat A, Eggert (1993). EMBO J., 12, 1343 ± 1354. M, Schneider S, Tenbaum SP, Renkawitz R and BauerA,MikulitsW,LaggerG,StenglG,BroschGand Baniahmad A. (1999). Mol. Cell. Biol., 19, 3383 ± 3394. Beug H. (1998). EMBO J., 17, 4291 ± 4303. Fischle W, Emiliani S, Hendzel MJ, Nagase T, Nomura N, Beug H, Bauer A, Dolznig H, von Lindern M, Lobmayer L, Voelter W and Verdin E. (1999). J. Biol. Chem., 274, Mellitzer G, Steinlein P, Wessely O and Mullner E. (1996). 11713 ± 11720. Biochim. Biophys. Acta, 1288, M35 ± M47. Fondell JD, Brunel F, Hisatake K and Roeder RG. (1996). Beug H, Mullner EW and Hayman MJ. (1994). Curr. Opin. Mol. Cell. Biol., 16, 281 ± 287. Cell. Biol., 6, 816 ± 824. Fondell JD, Roy AL and Roeder RG. (1993). Genes Dev., 7, Bhattacharya SK, Ramchandani S, Cervoni N and Szyf M. 1400 ± 1410. (1999). Nature, 397, 579 ± 583. Forrest D, Sjoberg M and Vennstrom B. (1990). EMBO J., 9, Bird AP and Wole AP. (1999). Cell, 99, 451 ± 454. 1519 ± 1528. Bird AP. (1995). Trends Genet., 11, 94 ± 100. Forrest D, Hanebuth E, Smeyne RJ, Everds N, Stewart CL, Boyer LA, Logie C, Bonte E, Becker PB, Wade PA, Wole Wehner JM and Curran T. (1996). EMBO J., 15, 3006 ± AP, Wu C, Imbalzano AN and Peterson CL. (2000). J. 3015. Biol. Chem., 275, 18864 ± 18870. Fuerstenberg S, Leitner I, Schroeder C, Schwarz H, Boyes J, By®eld P, Nakatani Y and Ogryzko V. (1998). Vennstrom B and Beug H. (1992). EMBO J., 11, 3355 ± Nature, 396, 594 ± 598. 3365. Braliou GG, Ciana P, Klaassen W and Stunnenberg HG. Fuks F, Burgers WA, Brehm A, Hughes-Davies L and (2000). Oncogene, in press. Kouzarides T. (2000). Nat. Genet., 24, 88 ± 91. Burke LJ and Baniahmad A. (2000). FASEB J., 14, 1876 ± Gelmetti V, Zhang J, Fanelli M, Minucci S, Pelicci PG and 1888. Lazar MA. (1998). Mol. Cell. Biol., 18, 7185 ± 7191. Cameron EE, Bachman KE, Myohanen S, Herman JG and Goldhirsch A and Gelber RD. (1996). Semin. Oncol., 23, Baylin SB. (1999). Nat. Genet., 21, 103 ± 107. 494 ± 505. Casini T and Graf T. (1995). Oncogene, 11, 1019 ± 1026. Graf T and Beug H. (1983). Cell, 34, 7±9. Chang TJ, Scher BM, Waxman S and Scher W. (1993). Mol. Endocrinol., 7, 528 ± 542. Chen JD and Evans RM. (1995). Nature, 377, 454 ± 457.

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3108 Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti Lubbert M. (2000). Curr. Top. Microbiol. Immunol., 249, V, Cioce M, Fanelli M, Ruthardt M, Ferrara FF, Zamir I, 135 ± 164. Seiser C, Grignani F, Lazar MA, Minucci S and Pelicci Lutterbach B, Westendorf JJ, Linggi B, Patten A, Moniwa PG. (1998). Nature, 391, 815 ± 818. M, Davie JR, Huynh KD, Bardwell VJ, Lavinsky RM, Grozinger CM, Hassig CA and Schreiber SL. (1999). Proc. Rosenfeld MG, Glass C, Seto E and Hiebert SW. (1998). Natl. Acad. Sci. USA, 96, 4868 ± 4873. Mol. Cell. Biol., 18, 7176 ± 7184. GuentherMG,LaneWS,FischleW,VerdinE,LazarMA Lutz M, Baniahmad A and Renkawitz R. (2000a). Biochem. and Shiekhattar R. (2000). Genes Dev., 14, 1048 ± 1057. Soc. Trans., 28, 386 ± 389. Hassig CA, Fleischer TC, Billin AN, Schreiber SL and Ayer Lutz M, Burke LJ, Barreto G, Goeman F, Greb H, Arnold DE. (1997). Cell, 89, 341 ± 347. R, Schultheiss H, Brehm A, Kouzarides T, Lobanenkov V He LZ, Guidez F, Triboli C, Peruzzi D, Ruthardt M, Zelent and Renkawitz R. (2000b). Nucl. Acids Res., 28, 1707 ± A and Randolphi PP. (1998). Nat. Genet., 18, 126 ± 135. 1713. Heinzel T, Lavinsky RM, Mullen TM, Soderstrom M, Meehan RR, Lewis JD and Bird AP. (1992). Nucl. Acids LahertyCD,TorchiaJ,YangWM,BrardG,NgoSD, Res., 20, 5085 ± 5092. Davie JR, Seto E, Eisenman RN, Rose DW, Glass CK and Miles BD and Robinson HL. (1985). J. Virol., 54, 295 ± 303. Rosenfeld MG. (1997). Nature, 387, 43 ± 48. Munoz A, Zenke M, Gehring U, Sap J, Beug H and Hendrich B and Bird A. (1998). Mol. Cell. Biol., 18, 6538 ± Vennstrom B. (1988). EMBO J., 7, 155 ± 159. 6547. Muscat GE, Burke LJ and Downes M. (1998). Nucl. Acids Hermann T, Homann B, Piedra®ta FJ, Zhang XK and Res., 26, 2899 ± 2907. Pfahl M. (1993). Oncogene, 8, 55 ± 65. Nagy L, Kao HY, Chakravarti D, Lin RJ, Hassig CA, Ayer Hong SH and Privalsky ML. (2000). Mol. Cell. Biol., 20, DE, Schreiber SL and Evans RM. (1997). Cell, 89, 373 ± 6612 ± 6625. 380. Hong SH, David G, Wong CW, Dejean A and Privalsky ML. Nan X, Ng HH, Johnson CA, Laherty CD, Turner BM, (1997). Proc. Natl. Acad. Sci. USA, 94, 9028 ± 9033. Eisenman RN and Bird A. (1998). Nature, 393, 386 ± 389. Horlein AJ, Naar AM, Heinzel T, Torchia J, Gloss B, Okano M, Xie S and Li E. (1998a). Nucl. Acids Res., 26, Kurokawa R, Ryan A, Kamei Y, Soderstrom M, Glass 2536 ± 2540. CK and Rosenfelt MG. (1995). Nature, 377, 397 ± 404. Okano M, Xie S and Li E. (1998b). Nat. Genet., 19, 219 ± 220. HuIandLazarMA.(2000).Trends Endocrinol. Metab., 11, Ordentlich P, Downes M, Xie W, Genin A, Spinner NB and 6±10. Evans RM. (1999). Proc. Natl. Acad. Sci. USA, 96, 2639 ± Huang EY, Zhang J, Miska EA, Guenther MG, Kouzarides 2644. T and Lazar MA. (2000). Genes Dev., 14, 45 ± 54. Pain B, Melet F, Jurdic P and Samarut J. (1990). New Biol., Hung HL, Lau J, Kim AY, Weiss MJ and Blobel GA. (1999). 2, 284 ± 294. Mol. Cell. Biol., 19, 3496 ± 3505. ParkEJ,SchroenDJ,YangM,LiH,LiLandChenJD. Jepsen K, Hermanson O, Onami TM, Gleiberman AS, (1999). Proc. Natl. Acad. Sci. USA, 96, 3519 ± 3524. Lunyak V, McEvilly RJ, Kurokawa R, Kumar V, Liu F, Polly P, Herdick M, Moehren U, Baniahmad A, Heinzel T Seto E, Hedrick SM, Mandel G, Glass CK, Rose DW and and Carlberg C. (2000). FASEB J., 14, 1455 ± 1463. Rosenfeld MG. (2000). Cell, 102, 753 ± 763. Rascle A, Ghysdael J and Samarut J. (1994). Oncogene, 9, Jones PL, Veenstra GJ, Wade PA, Vermaak D, Kass SU, 2853 ± 2867. Landsberger N, Strouboulis J and Wole AP. (1998). Nat, Redner RL, Wang J and Liu JM. (1999). Blood, 94, 417 ± 428. Genet., 19, 187 ± 191. Richel DJ, Colly LP, Kluin-Nelemans JC and Willemze R. Jones PA and Laird PW. (1999). Nat. Genet., 21, 163 ± 167. (1991). Br. J. Cancer, 64, 144 ± 148. Kadosh D and Struhl K. (1997). Cell, 89, 365 ± 371. Rountree MR, Bachman KE and Baylin SB. (2000). Nat. Kahn P, Frykberg L, Brady C, Stanley I, Beug H, Vennstrom Genet., 25, 269 ± 277. B and Graf T. (1986). Cell, 45, 349 ± 356. Safer JD, Cohen RN, Hollenberg AN and Wondisford FE. Kaludov NK and Wole AP. (2000). Nucl. Acids Res., 28, (1998). J. Biol. Chem., 273, 30175 ± 30182. 1921 ± 1928. Samarut J and Gazzolo L. (1982). Cell, 28, 921 ± 929. Kao HY, Downes M, Ordentlich P and Evans RM. (2000). Sande S and Privalsky ML. (1996). Mol. Endocrinol., 10, Genes Dev., 14, 55 ± 66. 813 ± 825. Knoep¯er PS and Eisenman RN. (1999). Cell, 99, 447 ± 450. Sap J, Munoz A, Damm K, Goldberg Y, Ghysdael J, Leutz Kouzarides T. (1999). Curr. Opin. Genet. Dev., 9, 40 ± 48. A, Beug H and Vennstrom B. (1986). Nature, 324, 635 ± Kurokawa R, Soderstrom M, Horlein A, Halachmi S, Brown 640. M, Rosenfeld MG and Glass CK. (1995). Nature, 377, Sap J, Munoz A, Schmitt J, Stunnenberg H and Vennstrom 451 ± 454. B. (1989). Nature, 340, 242 ± 244. Laherty CD, Yang WM, Sun JM, Davie JR, Seto E and Siegfried Z, Eden S, Mendelsohn M, Feng X, Tsuberi BZ and Eisenman RN. (1997). Cell, 89, 349 ± 356. Cedar H. (1999). Nat. Genet., 22, 203 ± 206. LeiH,OhSP,OkanoM,JuttermannR,GossKA,Jaenisch Taunton J, Hassig CA and Schreiber SL. (1996). Science, R and Li E. (1996). Development, 122, 3195 ± 3205. 272, 408 ± 411. Leonhardt H, Page AW, Weier HU and Bestor TH. (1992). Underhill C, Qutob MS, Yee SP and Torchia J. (2000). J. Cell, 71, 865 ± 873. Biol. Chem., 275, 40463 ± 40470. Lewis JD, Meehan RR, Henzel WJ, Maurer-Fogy I, Urnov FD, Yee J, Sachs L, Collingwood TN, Bauer A, Beug Jeppesen P, Klein F and Bird A. (1992). Cell, 69, 905 ± 914. H, Shi YB and Wole AP. (2000). EMBO J., 19, 4074 ± Li E, Bestor TH and Jaenisch R. (1992). Cell, 69, 915 ± 926. 4090. Li J, Wang J, Nawaz Z, Liu JM, Qin J and Wong J. (2000). Verdel A and Khochbin S. (1999). J. Biol. Chem., 274, 2440 ± EMBO J., 19, 4342 ± 4350. 2445. Limonta M, Colombo T, Damia G, Catapano CV, Conter V, Verreault A, Kaufman PD, Kobayashi R and Stillman B. GervasoniM,MaseraG,LisoV,SpecchiaGandGiudici (1998). Curr. Biol., 8, 96 ± 108. G. (1993). Leuk. Res., 17, 977 ± 982. Vidal M and Gaber. (1991). Mol. Cell. Biol., 11, 6317 ± 6327.

Oncogene Corepressors in erythroleukemia LEG Rietveld et al 3109 Wade PA, Gegonne A, Jones PL, Ballestar E, Aubry F and ZamirI,HardingHP,AtkinsGB,HorleinA,GlassCK, Wole AP. (1999). Nat. Genet., 23, 62 ± 66. Rosenfeld MG and Lazar MA. (1996). Mol. Cell. Biol., 16, Warrell Jr RP, Frankel SR, Miller Jr WH, Scheinberg DA, 5458 ± 5465. ItriLM,HittelmanWN,VyasR,AndreeM,TafuriA Zamir I, Dawson J, Lavinsky RM, Glass CK, Rosenfeld MG and Jakubowski A. (1991). N. Engl. J. Med., 324, 1385 ± and Lazar MA. (1997). Proc.Natl.Acad.Sci.USA,94, 1393. 14400 ± 14405. Warrell Jr RP, de The H, Wang ZY and Degos L. (1993). N. ZenkeM,KahnP,DiselaC,VennstromB,LeutzA,Keegan Engl. J. Med., 329, 177 ± 189. K,HaymanMJ,CoiHR,YewN,EngelJDandBeugH. Wen YD, Perissi V, Staszewski LM, Yang WM, Krones A, (1988). Cell, 52, 107 ± 119. Glass CK, Rosenfeld MG and Seto E. (2000). Proc. Natl. Zenke M, Munoz A, Sap J, Vennstrom B and Beug H. (1990). Acad. Sci. USA, 97, 7202 ± 7207. Cell, 61, 1035 ± 1049. Willard HF and Hendrich BD. (1999). Nat. Genet., 23, 127 ± Zhang Y, Iratni R, Erdjument-Bromage H, Tempst P & 128. Reinberg D. (1997). Cell, 89, 357 ± 364. Wole AP, Jones PL and Wade PA. (1999). Proc. Natl. Acad. Zhang Y, Ng HH, Erdjument-Bromage H, Tempst P, Bird A Sci. USA, 96, 5894 ± 5896. and Reinberg D. (1999). Genes Dev., 13, 1924 ± 1935. Wong CW and Privalsky ML. (1998). Mol. Cell. Biol., 18, Zhou X, Richon VM, Rifkind RA and Marks PA. (2000). 5500 ± 5510. Proc. Natl. Acad. Sci. USA, 97, 1056 ± 1061. Yoder JA and Bestor TH. (1998). Hum. Mol. Genet., 7, 2792 ± 2784.

Oncogene