Acetylation inactivates the transcriptional repressor BCL6 - Nature

article

Acetylation inactivates the transcriptional repressor BCL6 Oksana R. Bereshchenko, Wei Gu & Riccardo Dalla-Favera

© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Published online 28 October 2002; doi:10.1038/ng1018 The proto-oncogene BCL6 encodes a BTB/POZ-zinc finger transcriptional repressor that is necessary for germinalcenter formation and has been implicated in the pathogenesis of B-cell lymphomas. Here we show that the coactivator p300 binds and acetylates BCL6 in vivo and inhibits its function. Acetylation disrupts the ability of BCL6 to recruit histone deacetylases (HDACs), thereby hindering its capacity to repress transcription and to induce cell transformation. BCL6 is acetylated under physiologic conditions in normal germinal-center B cells and in germinal center–derived B-cell tumors. Treatment with specific inhibitors shows that levels of acetylation of BCL6 are controlled by both HDAC-dependent and SIR2-dependent pathways. Pharmacological inhibition of these pathways leads to the accumulation of the inactive acetylated BCL6 and to cell-cycle arrest and apoptosis in B-cell lymphoma cells. These results identify a new mechanism of regulation of the proto-oncogene BCL6 with potential for therapeutic exploitation. Furthermore, these findings provide a new mechanism by which acetylation can promote transcription not only by modifying histones and activating transcriptional activators, but also by inhibiting transcriptional repressors.

Introduction The proto-oncogene BCL6 was identified by virtue of its involvement in chromosomal translocations associated with B cell–derived non-Hodgkin lymphoma (B-NHL)1–4. The product of BCL6 is a nuclear phosphoprotein belonging to the BTB/POZ (bric-à-brac, tramtrack, broad complex/Pox virus zinc fingers) zinc finger family of transcription factors5. BCL6 can repress transcription from promoters containing its DNA binding site6. This function requires the DNA-binding zinc-finger domain and two transcriptional repression domains: the amino-terminal POZ domain and a second separate domain located in the middle portion of the molecule6,7. In vitro evidence suggests that the transcriptional repressor function of BCL6 requires interaction of both of these domains with complexes containing HDACs and co-repressor molecules including SMRT and SIN3A8–10. In the B-cell lineage, BCL6 is selectively expressed in mature B cells within germinal centers11,12, where B cells undergo immunoglobulin gene hypermutation and isotype switching and are selected based on affinity maturation13. BCL6 is required for the formation of germinal centers14,15. Independent lines of evidence suggest that the function of BCL6 is to repress genes involved in the control of lymphocyte activation, differentiation and apoptosis within the germinal center16,17. BCL6 can bind the same DNA sequence that is recognized by the STAT6 transcriptional activator (the main nuclear effector of signaling induced by interleukin-4) and modulate expression

of those genes that are induced by interleukin-4 and mediated by STAT6 (refs 15,18). Expression of BCL6 is regulated by a number of signals that are critical for germinal-center development. At the protein level, engagement of B-cell receptors by antigens induces MAP kinase–mediated phosphorylation of BCL6, which, in turn, targets BCL6 for degradation by the ubiquitin proteasome pathway19. Signaling by the CD40 receptor, normally activated by T cells, has been shown to downregulate BCL6 expression at the transcriptional level20. The normal transcriptional regulation of BCL6 is disrupted by tumor-associated chromosomal translocations, which juxtapose heterologous promoters to the coding exons of BCL6, causing its deregulated expression by a mechanism called promoter substitution2. The 5′ non-coding region of BCL6 is also subjected to somatic hypermutation in normal germinal-center B cells and in germinal center–derived B-NHL21–23; in some diffuse large B-cell lymphomas (DLBCL), these mutations can deregulate BCL6 expression24). Overall, several lines of evidence suggest that downregulation of BCL6 is necessary for normal B cells to exit the germinal center, whereas BCL6 remains constitutively expressed in a substantial fraction of B-cell lymphomas11. Here we report that acetylation regulates the function of BCL6. Acetylation is known to stimulate transcription by modifying histones, leading to a transcriptionally active chromatin conformation, or by directly targeting transcriptional activators (for

Institute for Cancer Genetics and the Departments of Pathology and Genetics & Development, Columbia University, New York, New York 10032, USA. Correspondence should be addressed to R.D.-F. (e-mail: [email protected]).

606

nature genetics • volume 32 • december 2002

article

Results The transcriptional co-activator p300 acetylates BCL6 in vivo Noting that BCL6 contained motifs similar to those targeted by p300-mediated acetylation in p53 (ref. 27), we examined whether BCL6 could also be modified by transcriptional co-activators with histone acetyltransferase (HAT) activity28. Initial experiments indicated that BCL6 can be acetylated in vitro by p300, but not by the p300/CBP-associated acetyltransferase (data not shown). To test whether BCL6 is also acetylated in vivo, we co-transfected human embryonic kidney 293T cells with vectors expressing BCL6 and p300. The cell lysates were then analyzed by immunoprecipitation using an antibody against BCL6 followed by western blotting using an antibody previously shown to recognize acetylated lysines in various transcription factors29,30. Acetylated BCL6 was readily detected with co-expression of wildtype p300 but not with a HAT-deficient mutant of p300 (refs 31,32), indicating that acetylation of BCL6 required the enzymatic activity of p300. Acetylation of BCL6 was barely detectable in the absence of exogenous p300, possibly owing to the limiting amounts of endogenous p300 activity (Fig. 1b).

By co-transfecting 293T cells with vectors expressing p300 and various BCL6 mutants (Fig. 1a), we mapped the sequences susceptible to acetylation to the N-terminal region of BCL6 (Fig. 1b). This region contained 21 lysine residues that represented potential acetylation sites; however, we focused on the lysine cluster (KKYK) at position 376–379 because it resembles the site at which p53 is acetylated by p300 (ref. 27). We substituted all three lysines with arginines and tested the resulting BCL6 mutant (BCL6-RRYR, Fig. 1a) for acetylation on co-expression with p300 in 293T cells. The BCL6-RRYR mutant, BCL6-∆ZF-RRYR mutant lacking the zinc finger domain but having the RRYR substitutions and a deletion mutant lacking the PEST region that contains the KKYK motif (BCL6-∆PEST) were not acetylated by p300 (Fig. 1c), thereby indicating that the KKYK motif was indeed the primary site of BCL6 acetylation by p300. To identify the precise target(s) of acetylation among the lysines of the KKYK motif, we generated three point mutants, each of which had one of the lysines substituted with arginine, and tested them for acetylation by p300 in the immunoprecipitation and western blotting assays. Substitution of the carboxy-terminal lysine (residue 379) resulted in the complete abrogation of the acetylation signal, whereas substitution of the other two lysines did not have any effect on acetylation (Fig. 1d). These results indicate that BCL6 can be acetylated by p300 in vivo, and that Lys379 in the KKYK motif is the primary target for acetylation. BCL6 and p300 interact in vivo To investigate whether BCL6 can physically interact with p300, we co-transfected 293T cells with vectors expressing the various forms of BCL6 (Fig 1a) tagged with hemagglutinin and p300 tagged with FLAG, and analyzed the cell lysates by

a

c

b

d


BCL6-∆PEST

BCL6-∆ZF-RRYR

BCL6-RRYR

BCL6-∆ZF-KKYR

BCL6-∆ZF

BCL6-ZF

BCL6-∆ZF-KRYK

BCL6-∆ZF-RRYR

BCL6-∆ZF-RKYK

BCL6

BCL6-∆ZF

BCL6-∆ZF

–

p300 p300-HAT

p300 p300-HAT

Fig. 1 p300-induced acetylation of KKYK p300 BCL6 and mapping of its acetylation sites. a, Schematic representation of BCL6 HA POZ PEST ZF the BCL6 protein and its derivatives used in transient transfection experiBCL6-∆ZF HA ments. POZ, protein–protein interacHAtion domain; PEST, region containing three putative PEST sequences; ZF, BCL6-ZF HA zinc finger DNA-binding domain. BCL6 b, p300 acetylates the N-terminal IP: α-HA BCL6-∆PEST ∆PEST HA region of BCL6. Top panel, vectors ∆ZF WB: α-AcLys RRYR expressing hemagglutinin-tagged ZF BCL6-∆ZF-RRYR HA full-length BCL6 (5 µg) or its N- (∆ZF, RRYR 2 µg) or C-terminal (ZF, 1 µg) portions BCL6 were transiently co-transfected with BCL6-RRYR ∆PEST HA vectors (7 µg) expressing FLAGIP: α-HA ∆ZF KKYR tagged wildtype p300 or a deletion WB: α-HA ZF mutant (p300-HAT) lacking acetylHA BCL6-KKYR transferase activity. Whole cell extracts were immunoprecipitated p300 (IP) using an hemagglutinin-specific ∆ZF HABCL6 ZF monoclonal antibody and analyzed for BCL6 acetylation by western blotting (WB) with an antibody specific for acetylated lysines (α-AcLys). BotHAtom panel, the same western-blot filter was subsequently probed with an antibody specific for hemagglutinin BCL6 IP: α-HA IP: α-HA to control for immunoprecipitation ∆ZF WB: α-AcLys ∆ZF WB: α-AcLys ZF and protein loading. c, BCL6 acetylation by p300 maps to its KKYK (376379) motif. Acetylation of various IP: α-HA BCL6 IP: α-HA BCL6 mutants (a) was analyzed by ∆ZF WB: α-BCL6 ∆ZF WB: α-HA immunoprecipitation and western ZF blotting as described in b. d, The Cterminal lysine (residue 379) of the KKYK motif is the target of p300-mediated acetylation. Acetylation of BCL6-∆ZF mutants with single lysine→arginine substitutions in the KKYK motif was analyzed by immunoprecipitation and western blotting as described in b. – p300 p300-HAT –


example, p53, GATA1, E2F), thereby contributing to their nuclear localization, association with co-activator complexes and perhaps increased DNA binding25,26. Our results show that acetylation can directly target and inactivate a transcriptional repressor such as BCL6. Acetylation of BCL6 occurs physiologically in germinal-center B cells and leads to its inactivation by preventing the recruitment of complexes containing HDACs. The results have biological and clinical implications for B-NHL and general implications for the role of acetylation in transcriptional regulation.

607

a

BCL6-RRYR

BCL6-ZF

BCL6

2% input

immunoprecipitation using an antibody against FLAG (for p300) followed by western blotting with an antibody against hemagglutinin (for the BCL6 mutants). Full-length BCL6 coimmunoprecipitated with p300 irrespective of the HAT activity of p300; both wildtype p300 and its HAT-deficient mutant bound BCL6 with comparable affinity (Fig. 2a). To map the region of BCL6 that interacts with p300, the same set of hemagglutinin-tagged BCL6 derivatives was co-expressed

BCL6-∆ZF

b

BCL6-∆PEST

Fig. 2 BCL6 and p300 interact in vivo. a, BCL6 can bind p300 independently of its HAT function. Top panel, Vectors HA-BCL6 – – + + + expressing FLAG-tagged p300 HAp300-FLAG – + – + – (5 µg) or p300-HAT (7 µg) were + – – – + p300-HAT-FLAG + – + – + – + – + – + – + p300-FLAG: co-transfected with a vector IP: α-FLAG IP: α-FLAG expressing full-length BCL6 p300-FLAG BCL6 WB: α-FLAG WB: α-BCL6 (5 µg) in 293T cells. Immunoprecipitations were done with IP: α-FLAG p300-FLAG antibody against FLAG BCL6 WB: α-FLAG p300-HAT-FLAG IP: α-FLAG ∆PEST agarose beads, and the pres∆ZF WB: α-HA ence of BCL6 in each immunoWB: α-BCL6 BCL6 ZF precipitate was evaluated by western blotting with antibody against BCL6. Middle panel, the same western-blot BCL6 WB: α-HA ∆PEST filter was probed with an anti∆ZF body against FLAG to control ZF for immunoprecipitation efficiency and p300-FLAG expression. Bottom panel, western blotting analysis (with antibody against BCL6) of the extracts prior to immunoprecipitation showed the expression levels of BCL6. b, p300 binds to the N-terminal domain of BCL6 and requires the BCL6 PEST region. In vivo mapping of p300 interaction with BCL6 was done with co-immunoprecipitation and western blotting assays. The set of BCL6 expression vectors (Fig. 1a) was transfected with (+) or without (–) the p300-FLAG expression vector in 293T cells. Middle panel, whole cell extracts were immunoprecipitated with M2 beads, and the presence of BCL6 derivatives in each immunoprecipitate was examined by western blotting using antibody against hemagglutinin. Top panel, the same filter was probed with an antibody against FLAG to control for immunoprecipitation efficiency and p300-FLAG expression. Bottom panel, western blotting analysis (with antibody against hemagglutinin) of the extracts prior to immunoprecipitation showed similar expression levels for all the BCL6 derivatives.

a

with p300-FLAG in 293T cells, and BCL6–p300 interactions were analyzed by immunoprecipitation and western blotting analysis. The BCL6-∆ZF mutant, but not the BCL6-ZF polypeptide, retained the ability to bind p300 (Fig. 2b), indicating that p300 binds to the N-terminal half of BCL6. The BCL6∆PEST polypeptide showed considerably weaker binding to p300, implicating the PEST region in the interaction with p300. But the BCL6-RRYR and BCL6-∆ZF-RRYR mutants still

b

608

0.02

HA-BCL6-QQYQ

HA-BCL6-KKYR

0.01

HA-BCL6

0.004

relative luciferase activity

Fig. 3 p300 inhibits the tranHA-BCL6 scriptional repression funcHA-BCL6 +p300 140 HA-BCL6-KKYR tion of BCL6. a, Expression of HA-BCL6-KKYR +p300 wildtype p300 but not its 120 HA-BCL6-QQYQ HAT-deficient mutant abro100 gates BCL6-mediated tran100 scriptional repression. Vectors 80 expressing BCL6, p300 and 60 pB6BS-TK-Luc were co-trans50 fected into 293T cells. The 40 pRL-TK-Luc vector (0.3 µg) 20 was included as an internal control for normalization. 0 Cells were collected 48 h postpB6BS-TK-Luc, 1 pmol pB6BS-TK-Luc, 1 pmol transfection, and luciferase – pMT2T-HA-BCL6, 1 pmol effector, pmol + + – – – + activity was measured as a pCI-p300-FLAG, 1 pmol – – – + – + function of BCL6-dependent + – + pCI-p300-HAT-FLAG, 1 pmol – – – reporter gene transcription6. The experiments were done WB: α-BCL6 in duplicate; representative data depict the mean ± s.d. of three independent experiWB: α-FLAG ments. Below the graph, controls are shown for the expression levels of BCL6, BCL6 IP: α-BCL6 p300-FLAG and p300-HATWB: α-AcLys IP: α- BCL6 FLAG by western blotting and for BCL6 acetylation by WB: α-BCL6 IP: α-BCL6 immunoprecipitation and WB: α-BCL6 western blotting (Fig. 1). b, The HA-BCL6-KKYR acety0.04 pmol lation-resistant BCL6 mutant can repress transcription, but its function cannot be inhibited by p300. The BCL6-QQYQ mutant (mimicking the acetylated state of BCL6) had weaker transcriptional repression activity. The pB6BS-TK-Luc reporter was co-transfected with the indicated amounts of the HA-BCL6, HA-BCL6-KKYR and HA-BCL6-QQYQ expression vectors (color-coded bars), and transcriptional repression by the three forms of BCL6 was compared in the luciferase reporter assay, with or without co-transfection of 1 pmol of the p300 expression vector. Below the graph, controls are shown for the expression levels of HA-BCL6, HA-BCL6-KKYR and HA-BCL6-QQYQ, as assessed by immunoprecipitation and western blotting analysis after transfection of 0.04 pmol of each vector. relative luciferase activity


BCL6-∆ZF-RRYR

article


article

a

relative luciferase activity

Acetylation inhibits the transcriptional repressor function of BCL6 To determine the functional consequences of BCL6 acetylation, we analyzed the effects of p300 on BCL6-mediated transcriptional repression. To this end, vectors expressing BCL6 and p300 were co-transfected into 293T cells along with a luciferase reporter gene (B6BS-TK-Luc) driven by a promoter linked to the BCL6 DNA binding site. The luciferase activity was then measured as a function of BCL6-dependent reporter gene transcription6. The potent transcriptional repression activity of BCL6 was markedly diminished by co-expression of p300 (Fig. 3a), but the basal reporter gene expression was not affected, indicating that the effect of p300 on BCL6 was specific and not due to a general enhancement of transcription by p300. The inhibitory effect on BCL6 was clearly dependent on the HAT activity of p300, as co-expression of the HAT-deficient mutant of p300 did not alter BCL6 function. These data also indicate that this effect was not due to recruitment of the

PolII machinery to the promoter by p300 (ref. 33), as the HATdeficient mutant of p300, which binds BCL6 (Fig. 2a) but does not acetylate it (Fig. 1b), had no effect on BCL6-mediated transcriptional repression. To confirm that acetylation was responsible for BCL6 inactivation, we studied the transcriptional activity of the acetylationresistant BCL6-KKYR mutant (see Fig. 1d) and a second mutant (BCL6-QQYQ) in which the three lysine residues of the KKYK motif were each substituted by glutamine, a residue that sterically mimics the effect of acetylated lysines29,34. The BCL6-KKYR mutant showed no difference in transcriptional repression activity compared with the wildtype BCL6 (Fig. 3b), but unlike wildtype BCL6, its activity was not inhibited by p300. This indicates that the inhibition of BCL6 function is dependent on p300mediated acetylation. Consistent with this, the acetylation-mimicking mutant (BCL6-QQYQ) showed considerably less transcriptionally repressive activity than wildtype BCL6, comparable to that of BCL6 after acetylation by p300. Co-expression of p300 did not impair transcriptional repression by BCL6-QQYQ further, suggesting that physical interaction with p300 does not contribute to BCL6 inactivation in the absence of acetylation (data not shown). Analogous results were obtained by transfecting the same vectors in H1299 small lung cell carcinoma cells (data not shown). Taken together, these results indicate that p300 negatively regulates BCL6 transcriptional repression function and that this regulatory activity requires BCL6 acetylation.

Fig. 4 Acetylation of BCL6 leads to its 120 dissociation from HDACs. a, BCL6mediated transcriptional repression is 100 inhibited by TSA. Luciferase reporter 80 assays were done as in Fig. 3 in the presence of the indicated concentraDMSO 60 tions of TSA (color-coded bars) or 200nM TSA dimethyl sulfoxide (DMSO), added 40 300nM TSA 12 h before cell harvesting. BCL6 protein expression levels (for 0.4 pmol of 20 500nM TSA transfected pMT2T-HA-BCL6) are 0 shown by western blotting analysis pMT2T-HA-BCL6 0 pmol 0.01 pmol 0.4 pmol 0.04 pmol below the graph. b, BCL6 acetylation and dose-dependent dissociation pB6BS-TK-Luc, 1 pmol from co-repressors. 293T cells were co-transfected with vectors expressTSA TSA TSA TSA ing FLAG-HDAC2 (7.5 µg), HA-BCL6 (1 µg) and GAL4-p300 (1, 2.5 and 12.5 µg). Cells were harvested 48 h WB: α-BCL6 BCL6 after transfection and divided in two parts for analysis of BCL6 interactions with HDAC2 and BCL6 acetylation status. For the analysis of BCL6–HDAC2 + + – – – – – – HA-BCL6 HA-BCL6 + – – + + + + + – interactions, whole-cell extracts were + + – – – – – – HA-BCL6-KKYR – + – – + + + + + FLAG-HDAC2 + + – – – – – HA-BCL6-QQYQ – immunoprecipitated with M2 beads + + + – – – – + + FLAG-HDAC2 + – – – GAL4-p300 (antibody against FLAG), and the IP: α-FLAG IP: α-FLAG amount of BCL6 in each immunopre1 BCL6 BCL6 WB: α-BCL6 WB: α-BCL6 cipitate was evaluated by western blotting (panel 1). The same westernIP: α-FLAG IP: α-BCL6 2 blot filter was then probed with an HDAC2 Ac-BCL6 WB: α-AcLys WB: α-FLAG antibody against FLAG to control for HDAC2 immunoprecipitation effiIP: α-BCL6 BCL6 3 ciency and protein loading (panel 5). WB: α-BCL6 WB: α-BCL6 BCL6 BCL6 acetylation was evaluated as described in Fig. 1 (panels 2,3). Cell p300 4 WB: α-GAL4 extracts prior to immunoprecipitation were analyzed for the expression levIP: α-FLAG els of GAL4-p300 (panel 4). c, A HDAC2 5 WB: α-FLAG mutant polypeptide (HA-BCL6-QQYQ) that mimics the acetylated isoform of BCL6 cannot bind HDAC2. Equal amounts (10 µg) of pMT2T-HA-BCL6, pMT2T-HA-BCL6-KKYR and pMT2T-HA-BCL6-QQYQ were co-transfected with an expression vector for FLAG-HDAC2 (7.5 µg) in 293T cells. Top panel, the HDAC2–BCL6 interactions were analyzed as in b by immunoprecipitation with M2 beads (antibody against FLAG) and western blotting with antibody against BCL6. Middle panel, the expression levels and immunoprecipitation efficiency for FLAG-HDAC2 were monitored by analyzing the same western-blot filter with antibody against FLAG. Bottom panel, western blotting (with antibody against BCL6) of the extracts prior to immunoprecipitation showed similar expression levels of the three BCL6 isoforms.

b

c

+


bound p300 (Fig. 2b), indicating that binding and acetylation require distinct domains of BCL6, and that the inability of p300 to acetylate these mutants is due to the absence of BCL6 acetylation sites rather than to the lack of binding to p300. Taken together, these results indicate that BCL6 and p300 interact in vivo, and that this interaction requires the PEST region but not the KKYK acetylation motif of BCL6.


609

a

BCL6 to bind DNA or localize to the nucleus (data not shown). Taken together, these results suggest that p300-mediated acetylation regulates BCL6 function principally by interfering with its ability to bind HDAC-containing complexes. BCL6 acetylation occurs in B cells and is controlled by HDAC- and SIR2-dependent pathways To verify that acetylation and deacetylation of BCL6 occurs in native cells and does not result from protein overexpression typical of transient transfection assays, we analyzed the presence of acetylated BCL6 in normal germinal-center B cells isolated from human tonsils and in various B-cell lymphoma lines representing transformed counterparts of germinal-center B cells containing normal (Ramos), mutated (P3HR1 and Ly-1) or translocated (Val) copies of BCL6 (refs 2,22,23). Acetylated BCL6 was detected in both normal and transformed germinal-center B cells (Fig. 5a). Inhibition of HDAC activity by addition of TSA led to an increase in the levels of acetylated BCL6 (see Fig. 5b for representative results in Ramos cells), indicating that TSA-sensitive deacetylases are involved in deacetylating BCL6. An analogous effect was obtained after addition of niacinamide (NIA; Fig. 5b), which interferes with a different deacetylation pathway involving SIR2, recently shown to deacetylate p53 (refs 39,40). Inhibition of both the HDAC- and SIR2-dependent pathways by co-treatment with TSA and NIA led to an additive effect on the accumulation of acetylated BCL6 in both normal and transformed B cells (Fig. 5b). These results demonstrate that BCL6 acetylation and deacetylation processes occur physiologically in B cells and are controlled by both HDAC- and SIR2-dependent pathways.

c

IP: α-NOTCH

Acetylation of BCL6 leads to its dissociation from histone deacetylases To investigate the mechanism by which p300-mediated acetylation impairs BCL6 function, we tested whether acetylation interfered with the ability of BCL6 to recruit complexes containing HDACs, which contribute to transcriptional repression by deacetylating histones in the proximity of promoter regions35. To this end, we first confirmed that BCL6-mediated repression was dependent on HDAC function8,9,36,37 by showing that addition of trichostatin A (TSA), a potent and specific inhibitor of HDACs38, led to a dose-dependent inhibition of BCL6-mediated repression (Fig. 4a). On the basis of experiments indicating that BCL6 interacts preferentially with HDAC2 and, to a lesser extent, with HDAC1 (data not shown), we then examined the effect of BCL6 acetylation on the stability of the HDAC2–BCL6 interaction in 293T cells after transient co-transfection with vectors expressing BCL6, FLAG-tagged HDAC2 and p300 fused with GAL4. The results of co-immunoprecipitation and western blotting showed that BCL6 interacted with HDAC2 (Fig. 4b), and that this interaction was abolished when BCL6 was acetylated by p300. Consistent with its ability to functionally mimic acetylated BCL6 in the reporter assay (Fig. 3b), the BCL6-QQYQ mutant also showed much weaker binding to HDAC2 (Fig. 4c), indicating that the KKYK motif is required for BCL6 to bind HDAC complexes. Conversely, the KKYR mutant retained the ability to bind HDAC2, suggesting that acetylation (or glutamine substitution) has a specific effect in preventing HDAC2 binding. In analogous experiments, we observed that neither p300-mediated acetylation nor glutamine substitutions affected the ability of

wash

5h

1.5 h

0.5 h

input

Ly-1

GC B cells

P3HR1

VaI

CB33

Ramos

Fig. 5 Protein interactions and acetylation of endogeTSA + NIA nous BCL6 in B cells. a, BCL6 is acetylated in normal human germinal-center B treatment: NT NT NT Ac-BCL6 cells and in germinal-center–derived lymphoma cells. IP: α-BCL6 HDAC2 1 Germinal center–centroblasts WB: α-HDAC2 BCL6 (CB) were isolated and pooled from human tonsils of four β-actin individuals. Top panel, cell BCL6 pellets from CB33 lymIP: α-BCL6 1 2 3 4 5 6 IgGH 2 phoblastoid cells (lacking WB: α-BCL6 BCL6 expression; negative IgGL control; lane 1), Burkitt lymphoma cells (Ramos, lane 2, IP: α-BCL6 (PGB6) Ac-BCL6 3 and P3HR1, lane 4), DLBCL WB: α-AcLys tonsillar cells (Val, lane 3, and Ly-1, Ramos Val P3HR1 Ly-1 B cells IP: α-BCL6 (PGB6) lane 5) and normal germinalBCL6 4 WB: α-BCL6 center B cells (GC, lane 6) were analyzed for BCL6 acetyWB: α-HDAC2 5 HDAC2 lation by immunoprecipitation and western blotting Ac-BCL6 6 using antibodies against BCL6 5 normalized HDAC2 binding and acetyl-lysine (AcLys). Mid4 3 dle panel, the same western2 BCL6 blot filters were subsequently 1 probed with antibody against 1 2 3 4 5 6 7 8 9 10 11 12 BCL6 to analyze the total levels of BCL6. Bottom panel, western blotting with antibody against β-actin was done for cell extracts used in immunoprecipitation. b, BCL6 acetylation in B cells is controlled by HDAC- and Sir2-dependent deacetylation pathways. Ramos, Val, P3HR1 and Ly-1 cells were grown in the presence of 1 µM TSA and 5 mM NIA, added together or individually, as indicated, for 3 h. Cell extracts were analyzed for BCL6 acetylation (top panel) and total BCL6 levels (lower panel) as in a. Human tonsillar lymphocytes were isolated and kept in culture for 6 h in the presence or absence of TSA and NIA. BCL6 acetylation status was evaluated as in a (lanes 11,12). c, BCL6 associates with endogenous HDAC2-containing complexes in Ramos cells, and BCL6 acetylation leads to its dissociation from HDAC2. Nuclear extracts from Ramos cells were immunoprecipitated with rabbit polyclonal antibody against BCL6. The antibody against Notch was used as a negative control in immunoprecipitation. Presence of HDAC2 in each immunoprecipitate was evaluated by western blotting with an antibody against HDAC2 (panel 1). The amounts of BCL6 in the immunoprecipitates were controlled by western blotting using an antibody against BCL6 (panel 2). Ramos cells were treated with 0.5 µM TSA and 5 mM NIA during the indicated time. Cells were divided in two aliquots: one for analysis of BCL6-HDAC2 interactions (panels 1,2) as in b, and one for analysis of BCL6 acetylation (panels 3,4) as in a and b. Panel 5 shows the expression levels of HDAC2 in the nuclear extracts used in the BCL6–HDAC2 interaction analysis, which were quantified by densitometry for normalization of the amount of HDAC2 bound to BCL6 in panel 1.

610

NT

TSA+ NIA

TSA+ NIA

NT

TSA+ NIA

NT

TSA+NIA

NT

TSA+NIA

NIA

TSA

b

NT


article


b


–

50

40

BCL6

number of colonies (>0.25 mm)/dish

a

BCL6-QQYQ

article

2,500 cells/dish 5,000 cells/dish

30

BCL6

10,000 cells/dish 20

10

0

pMT2T

pMT2T-HA-BCL6

pMT2T-HA-BCL6-QQYQ IP: α-BCL6 WB: α-BCL6

To establish whether acetylation prevents BCL6–HDAC interaction in native cells, we examined the association of endogenous BCL6, p300 and HDAC2 in Ramos cells. Indirect immunofluorescence staining experiments indicated that all three proteins resided in discrete nuclear subdomains outside of the nucleoli, consistent with the possibility of physical interaction (data not shown). Co-immunoprecipitation and western blotting experiments showed that BCL6 was associated with HDAC2 under basal conditions (Fig. 5c). To confirm that acetylation regulates the binding of BCL6 to HDAC2, we analyzed BCL6 acetylation and its binding to HDAC2 in parallel in Ramos cells treated with TSA and NIA. Drug-induced accumulation of acetylated BCL6 resulted in its progressive dissociation from HDAC2 (Fig. 5c). This phenomenon was reversible, as shown by the fact that drug removal by washing led to BCL6 deacetylation and restored binding to HDAC2. Taken together, these findings indicate that acetylation prevents stable association of BCL6 with HDACs, and that the function of BCL6 as a transcriptional repressor is regulated by the balance of the acetylation and deacetylation pathways in B cells. Acetylation sites are critical for BCL6 transforming activity To investigate whether acetylation affected the biological function of BCL6, we tested its effects on the ability of BCL6 to act as an oncogene. Although no assay is available to test its transforming potential in B cells, BCL6, like other oncogenes, can confer anchorage-independent growth to immortalized rodent fibroblasts, a classic trait of malignant transformation. Thus, we stably transfected Rat-1 cells with vectors expressing either wildtype BCL6 (pMT2T-HA-BCL6) or the mimetic of acetylated BCL6 (pMT2T-HA-BCL6-QQYQ). After antibiotic selection, transfected clones (>100) were pooled and compared for their ability to form colonies in soft agar41. The ability of the BCL6-QQYQ mutant to induce in vitro clonogenicity in Rat-1 cells was considerably impaired (Fig. 6). Both the number (Fig. 6a) and the size (data not shown) of the colonies induced by BCL6-QQYQ were markedly smaller than those induced by wildtype BCL6, despite the fact that both proteins were expressed at similar levels (Fig. 6b). These results indicate that the KKYK acetylation site is critical for BCL6-mediated transformation.

Discussion BCL6 is expressed in germinal-center centroblasts and centrocytes where it inhibits the expression of genes controlling cell cycle, apoptosis and differentiation16,17. Downregulation of BCL6 expression is required for B-cell differentiation towards memory and plasma cells17, and is achieved at the transcriptional nature genetics • volume 32 • december 2002

Fig. 6 BCL6 expression can transform Rat-1 fibroblasts, and this ability is impaired in the BCL6-QQYQ mutant, which mimics the BCL6 acetylated state. a, Anchorage-independent growth of Rat-1 cells expressing the BCL6 and BCL6-QQYQ proteins is reflected in the graph, which shows the number of colonies (scored when larger than 0.25 mm) formed by each Rat-1 transfectant at three seeding densities in soft agar. Each seeding was done in triplicates, and the error bars are shown. b, The same cells were analyzed for BCL6 expression by immunoprecipitation and western blotting using an antibody against BCL6.

level by T cell–dependent CD40 signaling20 and at the protein level by antigen-dependent B-cell receptor activation19. Thus, acetylation-mediated downregulation of BCL6 activity represents a new pathway for the rapid and possibly signal-induced inhibition of BCL6 function. The signals and effectors inducing BCL6 acetylation in B cells are not known, but do not include DNA damage (data not shown), which has been shown to induce acetylation of p53 (ref. 42) or signaling from CD40 and B-cell receptors, which have been shown to regulate BCL6 (refs 19,20). Under experimental conditions, p300, but not p300/CBP-associated factor, efficiently acetylates BCL6, strongly suggesting that p300 is the physiologic effector of BCL6 acetylation in B cells. Identifying the pathways controlling BCL6 acetylation will be critical to understand the mechanisms that regulate germinalcenter formation and lymphomagenesis. Both acetylation and substitution of critical lysines with amino-acid residues that mimic acetylation markedly impair the ability of BCL6 to repress transcription. In both cases, loss of function is associated with loss of the ability of BCL6 to bind HDAC2, suggesting that recruitment of HDAC is an essential component of the transcription-repressing function of BCL6. The C-terminal lysine (residue 379) within the KKYK motif located in the central portion of BCL6 seems to be critical both for BCL6 function and for its interaction with HDACs. This observation explains previous findings that the 100-amino-acid region encompassing the KKYK motif is essential for full transcriptional repression by BCL6 (ref. 6) and is included in the region required for binding to the SIN3A co-repressor in experimental ‘two-hybrid’ systems8, and that a 17-amino-acid region encompassing the acetylation site is required for BCL6 transcriptional repression and can function as an autonomous repression module43. Overall, these observations indicate that acetylation may inhibit BCL6 function by preventing the recruitment of corepressor complexes containing HDAC. Our results (Fig. 5) indicate that BCL6 acetylation is controlled by two deacetylation pathways. On the basis of its inhibition by TSA and the observation that BCL6 binds HDACs, the first pathway conceivably involves HDACs belonging to conserved families of Class I and/or II histone deacetylases, including HDAC1 and HDAC2, but possibly involving other members44. The second pathway is inhibited by NIA (vitamin B3), strongly suggesting the involvement of the NAD-dependent TSA-resistant SIR2 deacetylase, which has been shown to deacetylate histones45, as well as a transcription factor such as p53 (refs 39,40). The involvement of SIR2 in inducing the non-acetylated active state of BCL6 is consistent with its established role as a transcriptional silencer46. On the other hand, the observation that the NADdependent deacetylase activity of SIR2 is closely linked to cellular 611


article metabolism in both yeast and mammalian cells46 suggests that BCL6 acetylation, and, therefore, the regulation of its function, may also be influenced by the cell metabolic rate. A notable aspect of our results is that the same two pathways (HDAC and SIR2) that regulate acetylation-mediated activation of p53 (refs 40,47) also control acetylation-mediated inactivation of BCL6. We have recently shown that BCL6 inhibits cell-cycle arrest and apoptosis and directly represses the transcription of a number of p53-induced genes, including the cell-cycle regulator WAF1/p21 and the putative anti-apoptotic protein PIG-7 (encoded by p53-induced gene 7; ref. 48 and Niu and R.D.-F., in preparation). Thus, it is possible that specific acetylation and deacetylation pathways may control cell-cycle arrest and apoptosis by activating p53 and inactivating BCL6 in germinal-center B cells. Consistent with this hypothesis, treatment with TSAor NIA, which induces p53 activation and BCL6 inactivation, induces apoptosis and expression of WAF1/p21 and PIG-7 in multiple B-NHL lines (data not shown). These responses are not entirely dependent on p53, because they can be observed in p53mutant cell lines49, consistent with a p53-independent, possibly BCL6-mediated, role of acetylation in controlling cell cycle and apoptosis in germinal-center B cells. The fact that induction of BCL6 acetylation, and therefore inhibition of its function, can be achieved pharmacologically by inhibiting HDAC- and SIR2-dependent deacetylation has important therapeutic implications. Combined treatment with TSA and NIA leads to complete cell death within 28–48 hours for B-NHL cell lines derived from Burkitt lymphoma with either normal or mutant p53 and DLBCL carrying either normal or translocated BCL6 genes (data not shown). As both HDAC-targeting deacetylase inhibitors such as TSA, SAHA and analogs38 and SIR2-targeting deacetylase inhibitors such as NIA represent pharmacologically well-characterized substances usable in humans, further research on their use for the therapy of B-NHL is warranted. This therapeutic approach would be generally applicable to most types of B-NHL (>90%) because all, except for mantle cell lymphoma, express BCL6 (ref. 11). On the basis of the data we obtained from the Ramos cell line, this approach could also be useful for p53-null chemotherapy-resistant cases50. Finally, these results have general implications for the mechanisms controlling gene expression in that they point to a new role for acetylation in transcriptional regulation. Multiple lines of evidence show that acetylation promotes gene transcription by rendering chromatin structure permissive for transcription through modification of histones and by directly targeting transcriptional activators so as to increase their transactivating activity, nuclear localization or stability25. The results herein indicate that acetylation can also promote transcription by inactivating transcriptional repressors.

Methods Plasmid constructs and antibodies. We used pMT2T-HA and pB6BS-TKLuc constructs as previously described6. The pMT2T-HA-based BCL6 mutants were generated by point mutagenesis using the QuickChange SiteDirected Mutagenesis Kit (Stratagene). The pCI-p300-FLAG and pCIp300-HAT-FLAG constructs were as previously described32. The pBJHDAC1-FLAG and pME18SFLAG-HDAC2 were obtained from R. Eisenman. We used antibody against acetyl-lysine (Cell Signaling Technology), antibody against FLAG (M2, Sigma) and antibody against hemagglutinin (a gift from J. Kitajewsky). Antibodies against BCL6 (N3 and C19), against GAL4 (DBD), against Notch2 (25-255) and against HDAC2 (H54) were purchased from Santa Cruz Biotechnology. PGB6 antibody was received from B. Falini. Transfections and immunoprecipitations. We transfected 293T cells using the calcium phosphate precipitation method as described6. For

612

luciferase reporter assays, we seeded 7 × 105 cells per 60-mm dish 1 d before transfection. At 48 h post-transfection, we analyzed cell lysates for luciferase activity using the Dual-Luciferase Reporter Assay Kit (Promega) and for protein expression by western blotting as described40, using nitrocellulose membrane for immobilizing proteins and ECL reagent for detecting proteins (Amersham Pharmacia Biotech). For detection of protein–protein interactions in transfected 293T cells, we prepared whole-cell extracts in co-immunoprecipitation lysis buffer (50 mM Tris, 150 mM NaCl, 0.1% Nonidet P-40, 10% glycerol, 5 mM MgCl2) supplemented with 1 mM dithiothreitol, 10 mM β-glycerophosphate, 1 mM NaF, 1 mM sodium orthovanadate, 10 µM TSA (Sigma) and protease inhibitors cocktail (Calbiochem). Immunoprecipitations were done using the M2 beads (Sigma) for 6 h at 4 °C. We washed immobilized complexes five times with co-immunoprecipitation lysis buffer, eluted bound proteins from the beads using the FLAG peptide (Sigma) and analyzed eluates by western blotting using the indicated antibodies. Unless otherwise specified, western blots for BCL6 were done using the N3 antibody. For detection of BCL6-interacting proteins in Ramos cells, we prepared nuclear extracts from 1 × 108 cells and immunoprecipitated them with 2 µg of rabbit polyclonal antibody against BCL6 (C19) or with antibody against Notch as a negative control. Immune complexes were immobilized on Protein G beads (Amersham Pharmacia Biotech), washed five times with cold co-immunoprecipitation lysis buffer, and bound proteins were eluted from the beads using BC1000 buffer (20 mM Tris, pH 7.6, 1000 mM NaCl, 10% glycerol, 0.1 mM dithiothreitol, 10 µM TSA) to avoid contamination with immunoglobulin proteins. We analyzed eluates for the presence of HDAC2 and boiled and analyzed beads for BCL6 immunoprecipitation efficiency by western blotting using antibody against HDAC2 and antibody against BCL6, respectively. For detection of the acetylated endogenous BCL6, we washed 4 × 107 Ramos cells with cold phosphate-buffered saline containing 0.1 µM TSA and 1 mM NIA (Sigma) and lysed them in RIPA buffer (50 mM Tris, 250 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS) containing 50 µM TSA, 10 mM NIA and the protease inhibitors cocktail. Prior to immunoprecipitations, cell extracts were diluted 1:3 with co-immunoprecipitation buffer containing fresh TSA and NIA and cleared twice by centrifugation at 14,000 r.p.m. We immunoprecipitated whole-cell extracts using a mouse monoclonal antibody against BCL6 (PGB6) overnight at 4 °C. Immune complexes were immobilized on Protein G beads, washed four times with cold RIPA buffer containing TSA and NIA and eluted from the beads by boiling. Immunoprecipitated proteins were resolved by SDS–PAGE (4–15% gels, BIORAD) and analyzed by western blotting with antibody against acetyl-lysine. B-cell purification. Tonsils were obtained from routine tonsillectomies performed at the Babies and Children’s Hospital of Columbia–Presbyterian Medical Center. The specimens were kept on ice immediately after surgical removal. Separations were carried out in a cold room using ice cold phosphate-buffered saline with 0.5% bovine serum albumin. We isolated tonsillar mononuclear cells by mincing tonsillar tissue followed by FicollIsopaque density centrifugation. Germinal-center centroblasts (defined as CD77+CD38+ B cells) were purified from tonsils of four individuals by magnetic cell separation using the MidiMACS system (Miltenyi Biotech). We isolated centroblasts by staining tonsillar mononuclear cells with antibody against CD77 (Coulter/Immunotech), incubating with mouse antibody against rat IgM (Pharmingen) and then antibody against IgG1 MicroBeads (Miltenyi Biotech) and subsequently passing the stained cell suspension over an LS column (Miltenyi Biotech). The purity of the isolated centroblasts (95% in all cases) was determined on a Calibur fluorescence-activated cell sorter (Becton Dickinson) by staining for CD38, CD77 and CD3. Treatment with deacetylase inhibitors. We treated Ramos, Val and P3HR1 cells grown at 0.75 × 106 cells ml–1 in RPMI (Life Technologies) and Ly-1 B cells grown in Iscove’s modified Dulbecco’s medium supplemented with 10% fetal bovine serum (Life Technologies) with 1 µM TSA (dissolved in dimethyl sulfoxide) and 5 mM NIA (dissolved in water) for the indicated time. We placed normal tonsillar cells at a density of 5 × 106 cells ml–1 after isolation in RPMI medium supplemented with 20% fetal bovine serum and treated them the same way.


article Acknowledgments

We thank R. Baer and L. Pasqualucci for critical reading of the manuscript; U. Klein and G. Cattoretti for the isolation of human B cells; and J.C. Luo for the preliminary experiments on in vivo and in vitro interaction of p300 and BCL6. This work was supported by the US National Institutes of Health (to R.D.-F.). Competing interests statement

The authors declare that they have no competing financial interests.


Received 26 March; accepted 28 August 2002. 1. 2. 3.

4.

5.

6.

7. 8.

9. 10.

11. 12.

13. 14. 15.

16.

17.

18.

19.

20. 21.

Ye, B.H. et al. Alterations of a zinc finger-encoding gene, BCL-6, in diffuse largecell lymphoma. Science 262, 747–750 (1993). Ye, B.H. et al. Chromosomal translocations cause deregulated BCL6 expression by promoter substitution in B-cell lymphoma. EMBO J. 14, 6209–6217 (1995). Baron, B.W. et al. Identification of the gene associated with the recurring chromosomal translocations t(3;14)(q27;q32) and t(3;22)(q27;q11) in B-cell lymphomas. Proc. Natl Acad. Sci. USA 90, 5262–5266 (1993). Kerckaert, J.P. et al. LAZ3, a novel zinc-finger encoding gene, is disrupted by recurring chromosome 3q27 translocations in human lymphomas. Nature Genet. 5, 66–70 (1993). Zollman, S., Godt, D., Prive, G.G., Couderc, J.L. & Laski, F.A. The BTB domain, found primarily in zinc finger proteins, defines an evolutionarily conserved family that includes several developmentally regulated genes in Drosophila. Proc. Natl Acad. Sci. USA 91, 10717–10721 (1994). Chang, C.C., Ye, B.H., Chaganti, R.S. & Dalla-Favera, R. BCL-6, a POZ/zinc-finger protein, is a sequence-specific transcriptional repressor. Proc. Natl Acad. Sci. USA 93, 6947–6952 (1996). Seyfert, V.L., Allman, D., He, Y. & Staudt, L.M. Transcriptional repression by the proto-oncogene BCL-6. Oncogene 12, 2331–2342 (1996). Dhordain, P. et al. The LAZ3(BCL-6) oncoprotein recruits a SMRT/mSIN3A/histone deacetylase containing complex to mediate transcriptional repression. Nucleic Acids Res. 26, 4645–4651 (1998). Dhordain, P. et al. Corepressor SMRT binds the BTB/POZ repressing domain of the LAZ3/BCL6 oncoprotein. Proc. Natl Acad. Sci. USA 94, 10762–10767 (1997). Wong, C.W. & Privalsky, M.L. Components of the SMRT corepressor complex exhibit distinctive interactions with the POZ domain oncoproteins PLZF, PLZFRARα, and BCL-6. J. Biol. Chem. 273, 27695–27702 (1998). Cattoretti, G. et al. BCL-6 protein is expressed in germinal-center B cells. Blood 86, 45–53 (1995). Onizuka, T. et al. BCL-6 gene product, a 92- to 98-kD nuclear phosphoprotein, is highly expressed in germinal-center B cells and their neoplastic counterparts. Blood 86, 28–37 (1995). Rajewsky, K. Clonal selection and learning in the antibody system. Nature 381, 751–758 (1996). Ye, B.H. et al. The BCL-6 proto-oncogene controls germinal-center formation and Th2-type inflammation. Nature Genet. 16, 161–170 (1997). Dent, A.L., Shaffer, A.L., Yu, X., Allman, D. & Staudt, L.M. Control of inflammation, cytokine expression, and germinal-center formation by BCL-6. Science 276, 589–592 (1997). Shaffer, A.L. et al. BCL-6 represses genes that function in lymphocyte differentiation, inflammation, and cell-cycle control. Immunity 13, 199–212 (2000). Reljic, R., Wagner, S.D., Peakman, L.J. & Fearon, D.T. Suppression of signal transducer and activator of transcription 3-dependent B lymphocyte terminal differentiation by BCL-6. J. Exp. Med. 192, 1841–1848 (2000). Harris, M.B. et al. Transcriptional repression of Stat6-dependent interleukin-4induced genes by BCL-6: specific regulation of iepsilon transcription and immunoglobulin E switching. Mol. Cell. Biol. 19, 7264–7275 (1999). Niu, H., Ye, B.H. & Dalla-Favera, R. Antigen receptor signaling induces MAP kinase-mediated phosphorylation and degradation of the BCL-6 transcription factor. Genes Dev. 12, 1953–1961 (1998). Allman, D. et al. BCL-6 expression during B-cell activation. Blood 87, 5257–5268 (1996). Shen, H.M., Peters, A., Baron, B., Zhu, X. & Storb, U. Mutation of BCL-6 gene in normal B cells by the process of somatic hypermutation of Ig genes. Science 280,


1750–1752 (1998). 22. Migliazza, A. et al. Frequent somatic hypermutation of the 5′ noncoding region of the BCL6 gene in B-cell lymphoma. Proc. Natl Acad. Sci. USA 92, 12520–12524 (1995). 23. Pasqualucci, L. et al. BCL-6 mutations in normal germinal center B cells: evidence of somatic hypermutation acting outside Ig loci. Proc. Natl Acad. Sci. USA 95, 11816–11821 (1998). 24. Artiga, M.J. et al. A short mutational hot spot in the first intron of BCL-6 is associated with increased BCL-6 expression and with longer overall survival in large B-cell lymphomas. Am. J. Pathol. 160, 1371–1380 (2002). 25. Sterner, D.E. & Berger, S.L. Acetylation of histones and transcription-related factors. Microbiol. Mol. Biol. Rev. 64, 435–459 (2000). 26. Barlev, N.A. et al. Acetylation of p53 activates transcription through recruitment of coactivators/histone acetyltransferases. Mol. Cell 8, 1243–1254 (2001). 27. Gu, W. & Roeder, R.G. Activation of p53 sequence-specific DNA binding by acetylation of the p53 C-terminal domain. Cell 90, 595–606 (1997). 28. Roth, S.Y., Denu, J.M. & Allis, C.D. Histone acetyltransferases. Annu. Rev. Biochem. 70, 81–120 (2001). 29. Chan, H.M., Krstic-Demonacos, M., Smith, L., Demonacos, C. & La Thangue, N.B. Acetylation control of the retinoblastoma tumour-suppressor protein. Nature Cell Biol. 3, 667–674 (2001). 30. Martinez-Balbas, M.A., Bauer, U.M., Nielsen, S.J., Brehm, A. & Kouzarides, T. Regulation of E2F1 activity by acetylation. EMBO J. 19, 662–671 (2000). 31. Ogryzko, V.V., Schiltz, R.L., Russanova, V., Howard, B.H. & Nakatani, Y. The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell 87, 953–959 (1996). 32. Boyes, J., Byfield, P., Nakatani, Y. & Ogryzko, V. Regulation of activity of the transcription factor GATA-1 by acetylation. Nature 396, 594–598 (1998). 33. Chan, H.M. & La Thangue, N.B. p300/CBP proteins: HATs for transcriptional bridges and scaffolds. J. Cell Sci. 114, 2363–2373 (2001). 34. Zhang, W., Bone, J.R., Edmondson, D.G., Turner, B.M. & Roth, S.Y. Essential and redundant functions of histone acetylation revealed by mutation of target lysines and loss of the Gcn5p acetyltransferase. EMBO J. 17, 3155–3167 (1998). 35. Ng, H.H. & Bird, A. Histone deacetylases: silencers for hire. Trends Biochem. Sci. 25, 121–126 (2000). 36. Albagli, O. et al. Multiple domains participate in distance-independent LAZ3/BCL6-mediated transcriptional repression. Biochem. Biophys. Res. Commun. 220, 911–915 (1996). 37. Huynh, K.D. & Bardwell, V.J. The BCL-6 POZ domain and other POZ domains interact with the co-repressors N-CoR and SMRT. Oncogene 17, 2473–2484 (1998). 38. Marks, P.A., Richon, V.M. & Rifkind, R.A. Histone deacetylase inhibitors: inducers of differentiation or apoptosis of transformed cells. J. Natl Cancer Inst. 92, 1210–1216 (2000). 39. Vaziri, H. et al. hSIR2(SIRT1) functions as an NAD-dependent p53 deacetylase. Cell 107, 149–159 (2001). 40. Luo, J. et al. Negative control of p53 by Sir2α promotes cell survival under stress. Cell 107, 137–148 (2001). 41. Iida, S. et al. Deregulation of MUM1/IRF4 by chromosomal translocation in multiple myeloma. Nature Genet. 17, 226–230 (1997). 42. Sakaguchi, K. et al. DNA damage activates p53 through a phosphorylationacetylation cascade. Genes Dev. 12, 2831–2841 (1998). 43. Zhang, H., Okada, S., Hatano, M., Okabe, S. & Tokuhisa, T. A new functional domain of Bcl6 family that recruits histone deacetylases. Biochim. Biophys. Acta 1540, 188–200 (2001). 44. Cress, W.D. & Seto, E. Histone deacetylases, transcriptional control, and cancer. J. Cell Physiol. 184, 1–16 (2000). 45. Imai, S., Armstrong, C.M., Kaeberlein, M. & Guarente, L. Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase. Nature 403, 795–800 (2000). 46. Guarente, L. Sir2 links chromatin silencing, metabolism, and aging. Genes Dev. 14, 1021–1026 (2000). 47. Luo, J., Su, F., Chen, D., Shiloh, A. & Gu, W. Deacetylation of p53 modulates its effect on cell growth and apoptosis. Nature 408, 377–381 (2000). 48. Polyak, K., Xia, Y., Zweier, J.L., Kinzler, K.W. & Vogelstein, B. A model for p53induced apoptosis. Nature 389, 300–305 (1997). 49. Gaidano, G. et al. p53 mutations in human lymphoid malignancies: association with Burkitt lymphoma and chronic lymphocytic leukemia. Proc. Natl Acad. Sci. USA 88, 5413–5417 (1991). 50. Moller, M.B., Gerdes, A.M., Skjodt, K., Mortensen, L.S. & Pedersen, N.T. Disrupted p53 function as predictor of treatment failure and poor prognosis in B- and T-cell non-Hodgkin’s lymphoma. Clin. Cancer Res. 5, 1085–1091 (1999).

613