The monocytic leukemia zinc finger protein MOZ is a histone ... - Nature

ã

Oncogene (2001) 20, 404 ± 409 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

The monocytic leukemia zinc ®nger protein MOZ is a histone acetyltransferase Nathalie Champagne1, Nadine Pelletier1 and Xiang-Jiao Yang*,1 1

Molecular Oncology Group, Department of Medicine, McGill University Health Centre, MontreÂal, QueÂbec H3A 1A1, Canada

The monocytic leukemia zinc ®nger protein (MOZ) gene is rearranged in t(8;16)(p11;p13), t(8;22)(p11;q13) and inv(8)(p11q13) associated with acute myeloid leukemia. The other fusion partners involved are CBP, p300 and TIF2, transcriptional coactivators with known or potential histone acetyltransferase (HAT) activity. MOZ itself is a 2004-residue protein containing a putative acetyl CoA-binding motif, so it was hypothesized that MOZ is a HAT. Here we present direct evidence that MOZ has intrinsic HAT activity. Moreover, MOZ possesses a transcriptional repression domain at its Nterminal part and an activation domain at its C-terminal part. The activation domain does not show sequence similarity to any yeast proteins, but when tethered, it is able to activate transcription in yeast. Therefore, MOZ is a HAT with characteristics of a transcriptional coregulator, supporting the hypothesis that aberrant acetylation by abnormal MOZ proteins leads to leukemogenesis. Oncogene (2001) 20, 404 ± 409. Keywords: histone acetyltransferase; MYST; MOZ; MORF; leukemia

Eukaryotic DNA is packaged into chromatin, a generally repressive structure that needs to be regulated for gene expression (Kornberg and Lorch, 1999). In the past few years, it has become clear that DNA-binding transcription factors recruit coactivators with histone acetyltransferase (HAT) activity for transcriptional upregulation (Kuo and Allis, 1998; Struhl, 1998; Wole et al., 1997; Workman and Kingston, 1998). Coactivators with HAT activity include GCN5 (Brownell et al., 1996), PCAF (p300/CBP-associated factor) (Yang et al., 1996), p300 (300 kDa proteins associated with the adenoviral oncoprotein E1A) (Ogryzko et al., 1996), CBP (CREB-binding protein) (Bannister and Kouzarides, 1996; Ogryzko et al., 1996), TAFII250 (Mizzen et al., 1996), SRC-1 (steroid receptor coactivator-1) (Spencer et al., 1997), and ACTR (activator of retinoid receptor) (Chen et al., 1997).

*Correspondence: X-J Yang Received 2 February 2000; revised 9 November 2000; accepted 13 November 2000

Moreover, PCAF, p300 and CBP acetylate nonhistone proteins (reviewed in Sterner and Berger, 2000). Therefore, HATs regulate transcription by modifying chromosomal proteins and/or transcription factors. HAT genes have been found to be ampli®ed or rearranged in several cancers. For example, the p300 and CBP genes are rearranged in myeloid leukemia (Borrow et al., 1996; Chaanet et al., 1999, 2000; Ida et al., 1997; Rowley et al., 1997; Satake et al., 1997; Sobulo et al., 1997; Taki et al., 1997). One fusion partner is the monocytic leukemia zinc ®nger protein (MOZ) (Borrow et al., 1996; Chaanet et al., 1999, 2000). The MOZ gene has also been found to be fused to that of TIF2 (transcriptional intermediary factor 2; a homolog of SRC-1 and ACTR) in leukemia patients with inv(8)(p11q13) (Carapeti et al., 1998, 1999; Jacobson and Pillus, 1999; Liang et al., 1998). MOZ is a founder of the MYST (MOZ, YBF2, SAS2 and TIP60) family of proteins (Borrow et al., 1996). Yeast MYST proteins include SAS2 (something about silencing 2) (Reifsnyder et al., 1996), SAS3 (also known as YBF2) (John et al., 2000; Reifsnyder et al., 1996; Takechi and Nakayama, 1999), and ESA1 (essential SAS2-related acetyltransferase 1) (Allard et al., 1999; Clarke et al., 1999; Smith et al., 1998). In Drosophila, MOF (male-absent on the first) has been characterized (Akhtar and Becker, 2000; Hil®ker et al., 1997). In humans, there are hMOF (Neal et al., 2000), TIP60 (HIV Tat-interacting 60 kDa protein) (Yamamoto and Horikoshi, 1997), HBO1 (HAT bound to ORC 1) (Iizuka and Stillman, 1999), MOZ (Borrow et al., 1996), and MORF (Champagne et al., 1999; Thomas et al., 2000). ESA1, SAS3, MOF, TIP60, HBO1 and MORF have been shown to possess intrinsic HAT activity (Akhtar and Becker, 2000; Champagne et al., 1999; Iizuka and Stillman, 1999; John et al., 2000; Neal et al., 2000; Smith et al., 1998, 1999; Takechi and Nakayama, 1999; Thomas et al., 2000; Yamamoto and Horikoshi, 1997). The function of this family of proteins is rather diverse, including roles in transcription, epigenetic control, DNA replication and repair (Akhtar and Becker, 2000; Galarneau et al., 2000; Iizuka and Stillman, 1999; Ikura et al., 2000; John et al., 2000; Reifsnyder et al., 1996; Smith et al., 1999). Among this family, MOZ is the only member whose aberrant forms have been directly linked to cancers. Because of lack of direct evidence, the HAT activity and function of MOZ remain largely speculative (John et al., 2000; Sterner and Berger,

Histone acetyltransferase MOZ N Champagne et al

2000). MOZ displays signi®cant sequence similarity to MORF (overall identity, 60%; similarity, 66%) (Champagne et al., 1999). This sequence similarity suggests, but does not prove, that MOZ is a HAT. In fact, SAS2 was found to be inactive in HAT assays (John et al., 2000). Here we present data demonstrating that MOZ is a HAT with multiple functional domains. To test if MOZ is really a HAT, we expressed the MYST domain of MOZ in E. coli as a protein fused to the maltose-binding protein domain (MBP; Figure 1a). The reason that the MBP tag was used is that no other anity-tags (e.g. GST and 66His) yielded soluble fusion proteins for the HAT domain of MORF (Champagne et al., 1999). The MBP fusion protein was anity-puri®ed on amylose resin (Figure 1b, lanes 1 and 2). To determine its HAT activity, the anity-puri®ed protein was subjected to Whatman P81 ®lter-binding HAT assays (Champagne et al., 1999; Mizzen et al., 1996; Yang et al., 1996). As shown in Figure 1c, MBP-MOZ-HAT eciently acetylated histones. As a positive control, the known HAT domain of MORF was similarly analysed in the same set of experiments (Figure 1a ± c). Comparison of the two fusion proteins indicated that MBP-MOZ-

HAT is almost as active as MBP-MORF-HAT. MORF is more active than PCAF (Champagne et al., 1999), so MOZ should be a relatively active HAT. To assess which histones were acetylated, HeLa core histones labeled with MBP-MOZ-HAT were separated by SDS ± PAGE and subjected to phosphoimaging analysis. For parallel comparison, MBP-MORF-HAT was also analysed. As reported previously (Champagne et al., 1999), MBP-MORF-HAT acetylated histones H3 and H4 (Figure 1d, lanes 1 and 2). Besides histones H3 and H4, MBP-MOZ-HAT also eciently acetylated histone H2A (Figure 1d, lane 3). MBP-MOZHAT was unable to acetylate nucleosomal histones (data not shown). Taken together, these results indicate that MOZ has intrinsic HAT activity and its substrate speci®city is slightly distinct from that of MORF. Next, we analysed the function of dierent domains of MOZ using reporter gene assays. The N-terminal part of MORF contains a repression domain (Champagne et al., 1999). The corresponding region of MOZ is highly conserved (identity, 73%; similarity, 78%), so we asked if the N-terminal domain of MOZ is able to repress transcription. For this, the MOZ fragment

405

Figure 1 HAT activity of MOZ. (a) Schematic representation of MOZ and the fusion proteins MBP-MOZ-HAT and MBPMORF-HAT. Domains are labeled as follows: PHD, plant homeodomain (also called leukemia-associated protein domain); MYST, putative HAT domain; ED, Glu/Asp-rich acidic region; S, Ser-rich module, P, Pro/Gln-rich stretch; M, Met-rich domain; and MBP, maltose-binding protein domain. (b) SDS ± PAGE analysis of puri®ed MBP fusion proteins. The fusion proteins were expressed in E. coli and anity-puri®ed on amylose agarose (New England Biolabs). (c) HAT activity of puri®ed MBP fusion proteins determined by Whatman P81 ®lter-binding assays. A typical reaction (20 ml) contained 75 nCi of [3H]acetyl coenzyme A (4.7 Ci/ mmol; Amersham) and 2 mg of calf thymus histones (type IIa; Sigma). The reaction mixture was incubated at 308C for 10 min and processed as described (Champagne et al., 1999; Mizzen et al., 1996; Yang et al., 1996). (d) Substrate speci®city of puri®ed MBP fusion proteins. One mg of HeLa core histones was incubated with no enzyme (lane 1), MBP-MOZ-HAT (0.13 mg; lane 3) or MBPMORF-HAT (0.12 mg; lane 2) in the presence of [14C]acetyl CoA as described (Champagne et al., 1999; Yang et al., 1996). After separation on 15% SDS ± PAGE gels, 14C-labeled histones were detected by phosphoimaging Oncogene


406

Figure 2 Mapping of the MOZ repression domain. (a,b) Schematic representation of the Gal4 fusion proteins (a) and the luciferase reporters Gal4-E4-Luc and 3TP-Lux (b). Gal4-E4-Luc was derived from the 3TP-Lux luciferase reporter by insertion of ®ve copies of Gal4-binding sites (Champagne et al., 1999). (c) Transcriptional potential of the N-terminal domain of MOZ. The luciferase reporter Gal4-E4-Luc (or 3TP-Lux) and the b-galactosidase reporter CMV-b-Gal were transfected into NIH3T3 cells with or without the mammalian vector expressing the Gal4 DNA-binding domain (residues 1 ± 147), Gal4-MOZ-N or Gal4-MORF-N as indicated. Luciferase activities were measured and normalized to internal b-galactosidase controls (the Luc activity without an eector was arbitrarily set to 1.0). Average values from 3 ± 5 independent experiments are shown with standard deviations. The transfection and reporter gene assays were performed as previously described (Champagne et al., 1999; Wang et al., 1999, 2000). (d) Western blot analysis of expressed Gal4 fusion proteins. Gal4 expression plasmids were transfected into 293T cells and nuclear extracts were prepared for immunoblotting with anti-Gal4 antibody (Santa Cruz Biotechnology, RK5C1) as described (Wang et al., 1999)

consisting of amino acids 1 ± 356 was expressed as a protein fused to the Gal4 DNA-binding domain in mammalian cells and subjected to reporter gene assays (Figure 2a,b). Unlike the Gal4 DNA-binding domain itself, Gal4-MOZ-N repressed Gal4-E4-Luc in a dosedependent manner (11- and 25-fold for 50 and 200 ng of the eector plasmid, respectively; Figure 2c). For comparison, we also analysed the repression domain of MORF side by side (Figure 2a). As reported previously (Champagne et al., 1999), Gal4-MORF-N repressed Gal4-E4-Luc in a dose-dependent manner (Figure 2c). Dierent from that reported previously (Champagne et al., 1999), Gal4-MORF-N activated 3TP-Lux when 200 ng of the eector plasmid was used (Figure 2c). Similarly, Gal4-MOZ-N activated 3TP-Lux (Figure 2c). 3TP-Lux does not possess any Gal4-binding sites (Figure 2b), so the nature of the activation and the apparent discrepancy between this and previous studies are presently unclear. Western blot analysis revealed that Gal4-MOZ-N and Gal4-MORF-N were expressed Oncogene

to approximately similar levels (Figure 2d). Therefore, the N-terminal part of MOZ constitutes a transcriptional repression domain. We also analysed the Ser/Met-rich domain of MOZ, which shows high sequence similarity to the corresponding region of MORF (identity, 72%; similarity, 75%) but possesses insertion of a PQ-stretch (Figure 3a) (Borrow et al., 1996; Champagne et al., 1999). For this, the MOZ fragment consisting of amino acids 1410 ± 2004 was expressed as a protein fused to the Gal4 DNA-binding domain in mammalian cells and subjected to reporter gene assays (Figure 3a). As shown in Figure 3b, Gal4-MOZ-SM activated the Gal4-E4-Luc reporter in a dose-dependent manner, whereas this fusion protein had minimal eects on the reporter 3TPLux. For comparison, Gal4-MORF-SM was analysed in the same set of experiments (Figure 3a). Compared to Gal4-MORF-SM, Gal4-MOZ-SM was less potent (10 ± 15-fold compared to 60 ± 170-fold; Figure 3b). Western analysis indicated that these two fusion proteins were


Figure 3 Mapping of the MOZ activation domain. (a) Schematic representation of the Gal4 fusion proteins. (b) Transcriptional activation of the luciferase reporter Gal4-E4-Luc by the SM domain of MOZ. The luciferase reporter (Gal4-E4-Luc or 3TP-Lux) and the CMV-b-Gal galactosidase reporter were transfected into NIH3T3 cells with a mammalian vector expressing the Gal4 DNA-binding domain (residues 1 ± 147) or the indicated Gal4 fusion protein. Luciferase activities were measured and normalized as in Figure 2c. (c) Western blot analysis of expressed Gal4 fusion proteins. The fusion proteins were expressed and analysed as in Figure 2d

expressed to approximately similar levels (Figure 3c). Therefore, the SM domain of MOZ constitutes an activation domain that is weaker than that of MORF.

To analyse the activation domain further, we asked if it is functional in yeast. For this, the MOZ domain was expressed as a protein fused to the bacterial LexA DNA binding domain and tested for the ability to activate the b-galactosidase reporters pSH18 ± 34 and pLR1D 1 (Figure 4a,b). As shown in Figure 4c, LexAMOZ-SM activated the pSH18 ± 34 reporter by *30fold. We also tested LexA-MORF-SM (Figure 4a). LexA-MORF-SM activated the pSH18 ± 34 reporter by *160-fold (Figure 4c). In contrast, LexA-MOZ-SM and LexA-MORF-SM had minimal eects on the reporter pLR1D1, which lacks any LexA-binding sites (Figure 4c). Western blot analysis showed that LexAMOZ-SM and LexA-MORF-SM were expressed to approximately similar levels (Figure 4d). Therefore, the SM domains of MOZ and MORF function as transcriptional activation domains in yeast. The results described above demonstrate that MOZ is a HAT with characteristics of a transcriptional coregulator (Figure 4e). This is reminiscent of MORF (Champagne et al., 1999). However, there are subtle dierences between MOZ and MORF. First, the HAT domain of MOZ acetylated histone H2A more eciently than that of MORF (Figure 1d). Second, as an activation domain, the SM domain of MOZ was weaker than that of MORF (Figures 3 and 4), raising the interesting possibility that insertion of a PQ-stretch in the SM domain of MOZ reduces its activation potential (Figure 3a). For other MYST proteins, no evidence indicates that they possess transcriptional activation domains besides their HAT domains (Iizuka and Stillman, 1999; Smith et al., 1998, 1999; Yamamoto and Horikoshi, 1997). To the contrary, it was recently reported that HBO1 and TIP60 possess repression domains (Sharma et al., 2000). As for other HATs, p300 and CBP contain transcriptional activation domains in addition to their HAT domains (Bannister and Kouzarides, 1996; Ogryzko et al., 1996; Swope et al., 1996). The possession of multiple functional domains by MOZ and MORF suggests that they may regulate transcription in both acetylationdependent and -independent manners. The t(8;16)(8p11;p13), t(8;22)(p11;q13) and inv(8)(p11q13) chromosome translocations create MOZ fusion proteins lacking its activation domain (Borrow et al., 1996; Carapeti et al., 1998, 1999, Chaanet et al., 2000; Jacobson and Pillus, 1999; Liang et al., 1998). Various models have been proposed to explain how these translocations cause the development of leukemia (Borrow et al., 1996; Carapeti et al., 1998; Jacobson and Pillus, 1999; Liang et al., 1998). Our data suggest that these translocations either deregulate expression of MOZ-dependent genes due to loss of its activation domain and acquisition of a HAT domain from CBP, p300 or TIF2; or deregulate expression of CBP- p300-, or TIF2-dependent genes due to acquisition of the HAT domain of MOZ by CBP, p300 or TIF2. In summary, this study demonstrates that MOZ is a HAT with multiple functional domains and thus provides new insight into how aberrant MOZ proteins lead to leukemogenesis.

407

Oncogene


408

Figure 4 Function of the MOZ activation domain in yeast. (a) Schematic illustration of LexA-fusion proteins. (b) Schematic representation of the b-galactosidase reporters pSH18-34 (OriGene Technologies, Inc) and pLR1D1. pLR1D1 contains the bgalactosidase gene driven by the minimal Gal1 promoter (West et al., 1984). pSH18-34 is the same as pLR1D1 except eight LexAbinding sites inserted upstream from the Gal1 minimal promoter. (c) Activation by the C-terminal domain of MOZ. The yeast expression plasmid for the LexA DNA-binding domain or the indicated LexA fusion protein was transformed into the strain EGY48 harboring pSH18-34 or pLR1D1. b-Galactosidase activities of the yeast extracts were determined using chemiluminescent substrates (Tropix), normalized to cell density and used to calculate relative activation potential (the b-Gal activity without an eector was arbitrarily set to 1.0). Average values from three independent experiments are shown with standard deviations. (d) Western blot analysis of LexA and LexA fusion proteins. Nuclear extracts were prepared from EGY48 harboring the LexA expression plasmids, and subjected to Western blot analysis with anti-LexA antibody. (e) Functional domain organization of MOZ. Domains are labeled as in Figure 1a

Acknowledgments We thank M Ptashne and E Moore for the reporter plasmid pLR1D 1 and J CoÃteÂ for HeLa histones. This work was supported by a fellowship from Fonds de la recherche

en santeÂ du QueÂbec (FRSQ) (to N Champagne), grants from the National Cancer Institute of Canada and a scholarship from the Medical Research Council of Canada (to X-J Yang).

References Akhtar A and Becker PB. (2000). Mol. Cell, 5, 367 ± 375. Allard S, Utley RT, Savard J, Clarke A, Grant P, Brandl CJ, Pillus L, Workman JL and Cote J. (1999). EMBO J., 18, 5108 ± 5119. Oncogene

Bannister AJ and Kouzarides T. (1996). Nature, 384, 641 ± 643.


Borrow J, Stanton Jr. VP, Andresen JM, Becher R, Behm FG, Chaganti RS, Civin CI, Disteche C, Dube I, Frischauf AM, Horsman D, Mitelman F, Volinia S, Watmore AE and Housman DE. (1996). Nat. Genet., 14, 33 ± 41. Brownell JE, Zhou J, Ranalli T, Kobayashi R, Edmondson DG, Roth SY and Allis CD. (1996). Cell, 84, 843 ± 851. Carapeti M, Aguiar RC, Goldman JM and Cross NC. (1998). Blood, 91, 3127 ± 3133. Carapeti M, Aguiar RC, Watmore AE, Goldman JM and Cross NC. (1999). Cancer Genet. Cytogenet., 113, 70 ± 72. Chaanet M, Gressin L, Preudhomme C, Soenen-Cornu V, Birnbaum D and Pebusque MJ. (2000). Genes Chrom. Cancer, 28, 138 ± 144. Chaanet M, Mozziconacci MJ, Fernandez F, Sainty D, Lafage-Pochitalo M, Birnbaum D and Pebusque MJ. (1999). Genes Chromosomes Cancer, 26, 161 ± 165. Champagne N, Bertos NR, Pelletier N, Wang AH, Vezmar M, Yang Y, Heng HH and Yang XJ. (1999). J. Biol. Chem., 274, 28528 ± 28536. Chen H, Lin RJ, Schiltz RL, Chakravati D, Nash A, Nagy L, Privalsky ML, Nakatani Y and Evans RM. (1997). Cell, 90, 569 ± 580. Clarke AS, Lowell JE, Jacobson SJ and Pillus L. (1999). Mol. Cell. Biol., 19, 2515 ± 2526. Galarneau L, Nourani A, Boudreault AA, Zhang Y, Heliot L, Allard S, Savard J, Lane WS, Stillman D and Cote J. (2000). Mol. Cell, 5, 927 ± 937. Hil®ker A, Hil®ker-Kleiner D, Pannuti A and Lucchesi JC. (1997). EMBO J., 16, 2054 ± 2060. Ida K, Kitabayashi I, Taki T, Taniwaki M, Noro K, Yamamoto M, Ohki M and Hayashi Y. (1997). Blood, 90, 4699 ± 4704. Iizuka M and Stillman B. (1999). J. Biol. Chem., 274, 23027 ± 23034. Ikura T, Ogryzko VV, Grigoriev M, Groisman R, Wang J, Horikoshi M, Scully R, Qin J and Nakatani Y. (2000). Cell, 102, 463 ± 473. Jacobson S and Pillus L. (1999). Curr. Opin. Gent. Dev., 9, 175 ± 184. John S, Howe L, Tafrov ST, Grant PA, Sternglanz R and Workman JL. (2000). Genes Dev., 14, 1196 ± 1208. Kornberg RD and Lorch Y. (1999). Cell, 98, 285 ± 294. Kuo M-H and Allis CD. (1998). Bio. Essays, 20, 615 ± 626. Liang J, Prouty L, Williams BJ, Dayton MA and Blanchard KL. (1998). Blood, 92, 2118 ± 2122. Mizzen CA, Yang XJ, Kokubo T, Brownell JE, Bannister AJ, Owen-Hughes T, Workman J, Wang L, Berger SL, Kouzarides T, Nakatani Y and Allis CD. (1996). Cell, 87, 1261 ± 1270. Neal KC, Pannuti A, Smith ER and Lucchesi JC. (2000). Biochim. Biophys. Acta., 1490, 170 ± 174. Ogryzko VV, Schiltz RL, Russanova V, Howard BH and Nakatani Y. (1996). Cell, 87, 953 ± 959.

Reifsnyder C, Lowell J, Clarke A and Pillus L. (1996). Nat. Genet., 14, 42 ± 49. Rowley JD, Reshmi S, Sobulo O, Musvee T, Anastasi J, Raimondi S, Schneider NR, Barredo JC, Cantu ES, Schlegelberger B, Behm F, Doggett NA, Borrow J and Zeleznik-Le N. (1997). Blood, 90, 535 ± 541. Satake N, Ishida Y, Hinohara S, Kobayashi H, Sakashita A, Maseki N and Kaneko Y. (1997). Genes Chrom. Cancer, 20, 60 ± 63. Sharma M, Zarnegar M, Li X, Lim B and Sun J. (2000). J. Biol. Chem., 275, 35200 ± 35208. Smith ER, Eisen A, Gu W, Sattah M, Pannuti A, Zhou J, Cook RG, Lucchesi JC and Allis CD. (1998). Proc. Natl. Acad. Sci. USA, 95, 3561 ± 3565. Smith ER, Pannuti A, Gu W, Steurnagel A, Cook RG, Allis CD and Lucchesi JC. (1999). Mol. Cell. Biol., 20, 312 ± 318. Sobulo OM, Borrow J, Tomek R, Reshmi S, Harden A, Schlegeberger B, Housman D, Doggett NA, Rowley JD and Zeleznikle NJ. (1997). Proc. Natl. Acad. Sci. USA, 94, 8732 ± 8737. Spencer TE, Jenster G, Burcin MM, Allis CD, Zhou J, Mizzen CA, Mckenna NJ, Onate SA, Tsai SY, Tsai M-J and O'Malley BW. (1997). Nature, 389, 194 ± 197. Sterner DE and Berger SL. (2000). Microbiol. Mol. Biol. Rev., 64, 435 ± 459. Struhl K. (1998). Genes Dev., 12, 599 ± 606. Swope DL, Mueller CL and Chrivia JC. (1996). J. Biol. Chem., 271, 28138 ± 28145. Takechi S and Nakayama T. (1999). Biochem. Biophys. Res. Commun., 266, 405 ± 410. Taki T, Sako M, Tsuchida M and Hayashi Y. (1997). Blood, 89, 3945 ± 3950. Thomas T, Voss AK, Chowdhury K and Gruss P. (2000). Development, 127, 2537 ± 2548. 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. Wang AH, Kruhlak MJ, Wu J, Bertos NR, Vezmar M, Posner BI, Bazett-Jones DP and Yang XJ. (2000). Mol. Cell. Biol., 20, 6904 ± 6912. West RW, Yocum RR and Ptashne M. (1984). Mol. Cell. Biol., 4, 2467 ± 2478. Wole AP, Wong J and Pruss D. (1997). Genes Cells, 2, 291 ± 302. Workman JL and Kingston RE. (1998). Annu. Rev. Biochem., 67, 545 ± 579. Yamamoto T and Horikoshi M. (1997). J. Biol. Chem., 272, 30595 ± 30598. Yang XJ, Ogryzko VV, Nishikawa J, Howard BH and Nakatani Y. (1996). Nature, 382, 319 ± 324.

409

Oncogene