Daxx has been shown to play an essential role in type I IFN-αβ-mediated suppression of B cell development and apoptosis. Recently, we demonstrated that Tyk2 is directly involved in IFN signaling for the induction and translocation of Daxx, which may result in growth arrest and/or apoptosis of B lymphocyte progenitors. To clarify how Daxx regulates B cell development, we examined Daxx interacting partners by yeast two-hybrid screening. DNA methyltransferase 1 (DNMT1)-associated protein (DMAP1) was identified and demonstrated to interact with Daxx. The interaction regions in both proteins were mapped, and the cellular localization of the interaction was examined. Both Daxx and DMAP1 formed a complex with DNMT1 and colocalized in the nucleus. DMAP1 enhanced Daxx-mediated repression of glucocorticoid receptor transcriptional activity. Furthermore, Daxx protected protein degradation of DMAP1 in vivo. These results provide the novel molecular link between Daxx and DNMT1, which establishes a repressive transcription complex in the nucleus.
Daxx was first identified as a Fas-binding protein by yeast two-hybrid screening and was known as a proapoptotic protein that can enhance Fas-mediated apoptosis through c-Jun N-terminal kinase (JNK)3 activation (1). However, disruption of Daxx in mice increased apoptosis during the embryonic development, suggesting that Daxx acts as an antiapoptotic protein in the embryo (2). Because of the diverse effects of Daxx between in vitro and in vivo experimental systems, the roles of Daxx in apoptotic signals still remain unclear. Although the interaction between Daxx and Fas indicated the importance of Daxx in cytoplasm, nuclear localization of Daxx was observed in various cell lines, and the interactions of Daxx with several nuclear proteins such as the centromeric protein C, DNA methyltransferase 1 (DNMT1), Pax3, Pax5, E26 avian leukemia oncogene ETS1, Ubc9, small ubiquitin-like modifier-1, and promyelocytic leukemia protein (PML) were reported (2, 3, 4, 5, 6, 7, 8, 9). Thus, Daxx is likely to play alternative roles in shuttling between nucleus and cytoplasm.
Our experiments using Tyk2-deficient mice revealed that Tyk2 is essential for the IFN-αβ-induced inhibition of colony formation of B lymphocyte progenitors in response to IL-7 as well as the up-regulation and nuclear translocation of Daxx (10). Because Daxx plays crucial roles in IFN-α-induced growth suppression of B lymphocyte progenitors (11), it is very informative to identify potential regulators of Daxx involved in the growth arrest and/or apoptosis in early B cell development.
Recent studies implied that Daxx might function as a transcriptional coregulator. Daxx has been shown to possess transcriptional repression activity by inhibiting several transcription factors such as Pax3, ETS1 and glucocorticoid receptor (GR) through direct protein-protein interactions (4, 6, 12). Daxx is also shown to act as a transcriptional coactivator or corepressor of Pax5 in different cell types (5). Although the exact mechanism accounting for these observations is still unclear, the recruitment of nuclear factors possessing either histone acetyltransferase or histone deacetylase (HDAC) activity by Daxx to modulate Pax5 transcriptional activity was proposed (5). In addition, the transcriptional repression effect of Daxx could be modulated by subnuclear compartmentalization through protein-protein interactions. Furthermore, PML has been shown to relieve the transrepression effect of Daxx on Pax3 or GR transcriptional activity through sequestering Daxx into the PML oncogenic domains (12, 13).
In an attempt to identify novel Daxx partners, we screened a mouse embryo cDNA library with a yeast two-hybrid system using the N-terminal (Daxx/N) domain or C-terminal (Daxx/C) domain of Daxx as bait. In this study, we identify DNMT1-associated protein (DMAP1) as a protein that interacts specifically with Daxx/N. DMAP1 was isolated as a DNMT1-interacting protein by yeast two-hybrid screening (14). DMAP1 interacts directly with the first 120 amino acids of DNMT1. The N-terminal noncatalytic domain of DNMT1 binds to HDAC2 and DMAP1 and can mediate transcriptional repression (14).
In this study, we have characterized both biochemical and functional interactions between Daxx and DMAP1. Daxx and DMAP1 with DNMT1 formed a complex and colocalized in the nucleus. DMAP1 enhanced Daxx-mediated repression of GR transcriptional activity. Furthermore, Daxx protected protein degradation of DMAP1 in vivo. These results provide an important linkage between Daxx and DNMT1, which establishes an efficient repressive transcription complex in the nucleus.
Materials and Methods
Reagents and Abs
Dexamethasone and cycloheximide were purchased from Wako Pure Chemical (Osaka, Japan). MG132 was purchased from Peptide Institute (Osaka, Japan). A synthetic dimerizer, AP20187, was kindly provided by ARIAD Gene Therapeutics (Cambridge, MA) (15). Expression vectors, FLAG-tagged Daxx (15), murine mammary tumor virus-luciferase (MMTV-LUC) (16), and hemagglutinin (HA)-tagged Daxx (1) were kindly provided by Dr. H. Ariga (Hokkaido University, Sapporo, Japan), Dr. T. Taira (Hokkaido University), Dr. X. Yang (Massachusetts Institute of Technology, Cambridge, MA), and Dr. G. L. Hager (National Cancer Institute, National Institutes of Health, Bethesda, MD), respectively. Myc-tagged Daxx and DMAP1 mutants were generated by PCR methods and sequenced (primer sequences are available upon request). Myc-tagged mouse DNMT1 encoding amino acid residues 1–126 was also generated by PCR methods and sequenced (primer sequences are available upon request). Anti-HA and anti-Myc, anti-DNMT1 and anti-Daxx Abs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-FLAG M2 Ab was purchased from Sigma-Aldrich (St. Louis, MO). Anti-actin Ab was purchased from Chemicon International (Temecula, CA). Anti-DMAP1 Ab for Western blotting was purchased from Abcam (Cambridge, U.K.). We also generated anti-DMAP1 Ab for immunoprecipitation by immunizing GST-DMAP1 into a rabbit. JNK activation was determined by PhosphoPlus JNK Ab kit (Cell Signaling Technology, Beverly, MA) according to the manufacturer’s instructions.
Construction of fusion proteins with the Daxx/N or Daxx/C and the Gal4 DNA-binding domain
Full-length mouse Daxx cDNAs were obtained from Dr. X. Yang. To generate a bait construct with Daxx/N or Daxx/C, PCR was used to amplify the portion of the cDNA encoding amino acid residues 1–500 or residues 501–739, respectively (primer sequences are available upon request). The PCR product was digested with BamHI and SalI and inserted into pGBKT7 digested with BamHI and SalI (downstream of the Gal4 activation domain). All constructs were sequenced to verify integrity of the constructs.
Yeast two-hybrid screen
Gal4-Daxx/N was constructed by fusing Daxx/N coding sequence in-frame to the Gal4 DNA-binding domain in the pGBKT7 vector as previously described. Saccharomyces cerevisiae, strain AH109 cells transformed with pGal4-Daxx/N, followed by mating with a pretransformed mouse 11-day embryo Matchmaker cDNA library in Y187 cells (Clontech Laboratories, Palo Alto, CA), were plated onto medium that lacked tryptophan, leucine, and histidine and had been supplemented with 5 mM 3-amino-1, 2, 4-triazole (Sigma-Aldrich). Approximately 5 × 106 colonies were screened for growth in the absence of histidine. Plasmid DNAs derived from positive clones were extracted from yeasts, and sequenced Clones were reintroduced into yeast strain AH109 along with either empty pGBKT7, or pGBKT7-Daxx/N or pGBKT7-Daxx/C to verify the Daxx-clone interaction.
Cell culture, transfection, luciferase assays, and cell viability assays
An IL-3-dependent murine pro-B cell line, BAF3 was maintained in RPMI 1640 medium supplemented with 10% FCS and 10% conditioned medium from WEHI 3B cells as a source of IL-3 (15). A stable transforming cell line expressing FK506 binding protein (FKBP)-Fas, FLAG-tagged Daxx, and Myc-tagged DMAP1, BAF/FD7/DM1, was established as previously described (15) and maintained in the above medium in the presence of G418 (1 mg/ml). AP20187 stimulation was also performed in the presence of IL-3. Human embryonic kidney carcinoma cell line, 293T, was maintained in DMEM containing 10% FCS and transfected by the standard calcium precipitation protocol (17). The cells were harvested 48 h after transfection and lysed in 100 μl of PicaGene Reporter Lysis buffer (Toyo Ink, Tokyo, Japan) and assayed for luciferase and β-galactosidase activities according to the manufacturer’s instructions. Luciferase activities were normalized to the β-galactosidase activities. Three or more independent experiments were conducted for each assay. Cell viability was determined by Cell Counting kit 8 (Wako Chemicals, Tokyo, Japan) according to the manufacturer’s instructions.
Immunoprecipitation and immunoblotting
The immunoprecipitation and Western blot assays were performed as previously described (17). Cells were harvested and lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 0.15 M NaCl, containing 1% Nonidet P-40, 1 μM sodium orthovanadate, 1 μM PMSF, and 10 μg/ml each of aprotinin, pepstatin and leupeptin). The immunoprecipitates from cell lysates were resolved on SDS-PAGE and transferred to Immobilon filter (Millipore, Bedford, MA). The filters were then immunoblotted with each Ab. Immunoreactive proteins were visualized using an ECL detection system (Amersham Pharmacia Biotech, Piscataway, NJ).
Analysis of protein stability
To analyze the stability of DMAP1 in 293T cells, cells were transfected with Myc-tagged DMAP1 and full-length or Daxx mutant by the standard calcium precipitation protocol. After 36 h of incubation, the transfected cells were treated with or without 10 μM MG132 for 1 h and then with cycloheximide at a final concentration of 25 μg/ml for the indicated periods. The cell lysates were prepared and followed by Western blotting with an anti-Myc Ab or an anti-actin Ab.
Monkey COS7 or human HeLa cells were maintained in DMEM containing 10% FCS transfected with DMAP1 and/or Daxx and/or DNMT1 1–126 by the calcium phosphate precipitation protocol. After 48 h of transfection, cells were fixed with a solution containing 4% paraformaldehyde and reacted with an anti-FLAG Ab, an anti-Myc Ab, or an anti-Daxx Ab. The cells were then reacted with an FITC-conjugated anti-rabbit IgG or rhodamine-conjugated anti-mouse IgG (Chemicon International) and observed under a confocal laser fluorescent microscope. Images were obtained by using a Zeiss LSM 510 laser scanning microscope with an Apochromat ×63/1.4 oil immersion objective and ×4 zoom.
Identification of DMAP1 as an interaction partner of Daxx
To identify proteins that could be involved in the IFN/Daxx-mediated signaling, we screened a mouse 11-day embryo cDNA library using mouse Daxx/N Daxx/C as bait. Several Daxx-interacting proteins were identified from a screening of ∼5 × 106 transforming yeast. DNA sequencing analysis revealed that one of the positive clones that interacted specifically with Gal4 DNA-binding domain-fused Daxx/N was identical with a DMAP1 that contains a whole amino acid insertion (residue 1–467). To demonstrate specificity of binding, the plasmid was isolated from the positive two-hybrid clone and introduced back into S. cerevisiae along with either the Daxx/N or the Daxx/C fused to the DNA-binding domain of Gal4 or empty vector (Gal4 DNA-binding domain alone). Neither the Daxx/N nor the Daxx/C resulted in activation of the reporter genes (data not shown). After mating the indicated yeast, growth occurs only in the presence of the Daxx/N (Fig. 1,A), demonstrating that DMAP1 interacted with the Daxx/N in this assay. As previously described, Daxx had been reported to interact with a number of cellular proteins. Most of these factors have in common is that they require the Daxx/C for binding (see detailed listing in Fig. 1 B) (1, 2, 3, 4, 5, 6, 7, 8, 9, 18, 19, 20, 21, 22).
To investigate the association of DMAP1 with Daxx in mammalian cells, 293T cells were transfected with FLAG-tagged Daxx together with or without Myc-tagged DMAP1. As shown in Fig. 1,C, Western blot analysis of the immunoprecipitates with an anti-FLAG Ab revealed that DMAP1 interacts with Daxx in 293T cells. To delineate the regions of Daxx that are involved in Daxx/DMAP1 interaction, various deletion constructs of Daxx were engineered (Fig. 1,B) and subjected to analyze in coimmunoprecipitation assays in 293T cells. As shown in Fig. 1,D, the N-terminal deletion mutant of Daxx (Daxx 493–740) failed to interact with DMAP1. In contrast, the C-terminal deletion mutant of Daxx (Daxx 1–240 or Daxx 241–492) was capable of interacting with DMAP1. This is consistent with results observed in the yeast two-hybrid assay. Together, these results implicated Daxx/N as necessary and sufficient for interaction with DMAP1. We then examined the interacting domains on DMAP1 with Daxx. To delineate the domains in the DAMP1 that mediate the protein-protein interaction with Daxx, coimmunoprecipitation experiments were performed with a series of mutant DMAP1 proteins (Fig. 2,A). Expression vectors encoding FLAG-tagged Daxx and a series of Myc-tagged DMAP1 mutants were transiently transfected into 293T cells. Cells were lysed and subjected to immunoprecipitation with an anti-FLAG Ab. Immunoprecipitates were then used in Western blot analysis with an anti-Myc Ab. As shown in Fig. 2,B, the C-terminal domain of DMAP1 (DMAP1/C; amino acid residues 235–467) strongly interacted with Daxx. Furthermore, the middle region of the C-terminal domain of DMAP1 (DMAP1/C2; amino acid residues 293–411) strongly mediated the protein-protein interaction between DMAP1 and Daxx. These results indicate that Daxx/N interacts with DMAP1/C. To examine the interactions between Daxx and DMAP1 under more physiological conditions through endogenous proteins, we first transfected FLAG-tagged DMAP1 alone into 293T cells. Cells were lysed and subjected to immunoprecipitation with an anti-FLAG Ab. Immunoprecipitates were then used in Western blot analysis with an anti-Daxx Ab. As shown in Fig. 2,C, FLAG-tagged DMAP1 interacted with endogenous Daxx. Furthermore, we performed coimmunoprecipitation experiments using untransfected 293T cells. As shown in Fig. 2 C, we found that the immunoprecipitate with anti-DMAP1 Ab contained Daxx proteins. These results strongly suggest that DMAP1 forms a complex with Daxx in 293T cells.
Association of Daxx with DNMT1 via DMAP1
DMAP1 was first identified as an interaction partner with DNMT1 through its N-terminal 120 amino acids without the C-terminal catalytic domain (14). The noncatalytic, N-terminal portion of DNMT1 is shown to function as a transcriptional repressor that directly interacts with, and is partially dependent on, the activity of HDAC2 (14). In a previous report, a direct interaction between Daxx and DNMT1 in yeast was described, although no data were presented (2).
We then examined whether Daxx forms the complex with DMAP1 and DNMT1 in mammalian cells. Expression vectors encoding HA-tagged Daxx and/or FLAG-tagged DMAP1, Myc-tagged N-terminal region of DNMT1 (amino acid residues 1–126) were transiently transfected into 293T cells. Cells were lysed and subjected to immunoprecipitation with an anti-HA Ab. Immunoprecipitates were then used in Western blot analysis with an anti-Myc Ab. As shown in Fig. 3 A, the immunoprecipitates with anti-HA Ab contained the N-terminal domain of DNMT1 only in the presence of DMAP1, suggesting that Daxx forms a complex with DNMT1 in the presence DMAP1.
To further examine whether the full-length DNMT1 forms the complex with Daxx, coimmunoprecipitation experiments were performed using endogenous DNMT1 proteins in 293T cells. Expression vectors encoding FLAG-tagged Daxx and/or Myc-tagged DMAP1 were transiently transfected into 293T cells. Cells were lysed and subjected to immunoprecipitation with control IgG or an anti-DNMT1 Ab. Immunoprecipitates were then used in Western blot analysis with an anti-FLAG or Myc Ab. As shown in Fig. 3 B, the immunoprecipitates with anti-DNMT1 Ab contained Daxx protein only in the presence of DMAP1. These results suggest that Daxx forms a complex with DMAP1 and DNMT1 in the nucleus.
Colocalization of Daxx with DMAP1 and DNMT1
To determine whether DMAP1 colocalizes with Daxx in the nucleus, expression vectors for Myc-tagged DMAP1 and FLAG-tagged Daxx were transfected into COS7 cells. After 48-h transfection, the cells were stained with anti-Myc and anti-FLAG Abs and were visualized with rhodamine and FITC-conjugated secondary Abs, respectively, under a confocal laser microscope (Fig. 4,A). In the previous study, DMAP1 was shown to colocalize with DNMT1 to small punctate structures characteristic of S phase replication in the nucleus (14). Our results showed that both DMAP1 and Daxx were located in the nucleus, and they were found to be colocalized after demonstration of the merged figure in which the red and green colors turned yellow (Fig. 4,A), suggesting that both DMAP1 and Daxx colocalize in the nucleus. We also examined whether both DMAP1 and Daxx colocalize with DNMT1, expression vectors for Myc-tagged DNMT1 (1–126), and FLAG-tagged DMAP1 and HA-tagged Daxx were transfected into COS7 cells. After transfection (48 h), the cells were stained with anti-Daxx and anti-Myc, anti-FLAG Abs, and were visualized with rhodamine and FITC-conjugated secondary Abs, respectively, under a confocal laser microscope (Fig. 4,B). As shown Fig. 4 B, they were found to be colocalized in the nucleus.
We next examined whether Daxx and DMAP1 colocalized in the nucleus under more physiological conditions. Expression vectors for FLAG-tagged DMAP1 alone were transfected into HeLa cells. After 48-h transfection, the cells were stained with anti-Daxx and anti-FLAG Abs, and they were visualized with rhodamine and FITC-conjugated secondary Abs, respectively, under a confocal laser microscope (Fig. 4,C). As shown in Fig. 4 B, endogenous Daxx and FLAG-tagged DMAP1 were found to be colocalized in the nucleus. These results indicate that Daxx and DMAP1 with DNMT1 colocalize in the nucleus.
Effect of DMAP1 on Daxx-mediated apoptosis in murine pro-B cells
In our previous study, we established a murine pro-B cell line, BAF/FD7, which expresses the fusion protein composed of FKBP and membrane-anchored intracellular domain of Fas (FKBP-Fas) and FLAG-tagged Daxx after transfecting these expression constructs into original BAF3 cells (15). Aggregation of FKBP-Fas proteins by the addition of a bivalent FKBP ligand, AP20187, strongly induced cell death. However, original BAF3 as well as BAF3 cells expressing FKBP-Fas alone did not die after the treatment with AP20187 (15). Furthermore, AP28187 treatment of BAF/FD7 cells induced the marked JNK activation, when we monitored by immunoblotting using an Ab directed to phospho-JNK Thr183/Tyr185 (15).
To examine whether DMAP1 has an effect on Fas/Daxx-mediated cell death, we transfected Myc-tagged DMAP1 into BAF/FD7 cells and established a stable transformation with BAF/FD7DM1, which expresses both Daxx and DMAP1. We first examined the interaction between Daxx and DMAP1 in BAF/FD7/DM1 cells. Coimmunoprecipitation experiments were performed using cell lysates obtained from original BAF3, BAF/FD7, or BAF/FD7DM1 cells that were untreated with AP20187. Cell lysates were subjected to immunoprecipitation with an anti-FLAG Ab. Immunoprecipitates were then used in Western blot analysis with an anti-Myc Ab. As shown in Fig. 5 A, we found that FLAG-tagged Daxx immunoprecipitated from BAF/FD7/DM1 cells resulted in a complex with Myc-tagged DMAP1 and this interaction was independent on the presence of AP20187, suggesting that Daxx constitutively associates with DMAP1 in murine pro-B cells.
We next examined whether DMAP1 has an effect on Fas/Daxx-mediated cell death using BAF/FD7DM1 cells. As previously described, BAF/FD7 cells underwent cell death with the addition of AP20187 as shown in Fig. 5,B. Similarly, BAF/FD7DM1 cells underwent cell death by AP20187 in a dose-dependent manner (Fig. 5,B). Upon AP20187 stimulation, the strong-sustained phosphorylation of JNK was induced in both BAF/FD7 and BAF/FD7DM1 cells as shown in Fig. 5 C. These results suggest that DMAP1 expression and association with Daxx has no effect on Fas/Daxx-mediated cell death and JNK activation when we overexpressed both of them in murine pro-B cells.
DMAP1 and Daxx cooperatively repress GR-mediated transcriptional activity
Daxx has been reported to function as a transcriptional modulator in the nucleus (4, 6, 9, 23, 24). It has been also demonstrated that overexpression of Daxx suppresses GR-mediated activation of the MMTV promoter in a human embryonic kidney carcinoma cell line, 293T cells (12). To examine the functional relevance of the Daxx-DMAP1 interaction in the context of GR signaling pathway, we performed the transient transfection assay using 293T. The GR-mediated transcriptional responses were measured by MMTV-LUC, which is one of the standard reporters for assessing GR activity (16). When 293T cells were transfected with MMTV-LUC together with an expression vector for GR and treated with dexamethasone, luciferase expression was increased by 70-fold (Fig. 6,A). As shown in Fig. 6,A, overexpression of Daxx suppressed GR-mediated transactivation in a dose-dependent manner. DMAP1 was shown to act as a corepressor with DNMT1 independent of HDACs (14). We then examined the effect of overexpression of DMAP1 on GR-mediated transactivation in 293T cells. When 293T cells were transfected with an expression vector for DMAP1, GR, and MMTV-LUC, overexpression of DMAP1 showed a moderate suppression in GR-induced MMTV-LUC activity (Fig. 6,B). Interestingly, in the presence of Daxx, overexpression of DMAP1 effectively suppressed GR-induced MMTV-LUC activity, although the effect was additive. The previous study demonstrated that the suppression of GR activity was mediated through the C-terminal region of Daxx (16). We then tested whether DMAP1 has any effect on the suppression of GR activity by Daxx 493–740, which did not interact with DMAP1. As shown in Fig. 6 C, the DAMP1-mediated suppression of GR activity was enhanced by Daxx 493–740. These results suggest that Daxx and DMAP1 can independently act as a corepressor for GR-mediated transcription.
Daxx protects DMAP1 from the proteasomal degradation
During our experiments on the Daxx-DMAP1 interaction, we found that DMAP1 was a short-lived protein in 293T cells in the absence of Daxx. As shown in Fig. 7,A, coexpression of Daxx protein in 293T cells resulted in the accumulation of DMAP-1 protein in a dose-dependent manner. We also examined whether the proteasome-specific inhibitor MG132 can block DMAP1 degradation in 293T cells. Treatment of 293T cells with MG132 also resulted in the accumulation of DMAP1 in a dose-dependent manner (Fig. 7,B). To further determine whether Daxx or the proteasome inhibitor, MG132, directly affects degradation of DMAP-1 protein in 293T cells, we examined the stability of DMAP1 protein under conditions in which protein synthesis had been blocked by cycloheximide. Western blot analysis revealed that the DMAP-1 protein was stabilized when cells were transfected with Daxx but not Daxx/C-Daxx 493–740, or in the presence of MG132 (Fig. 7 C). These results suggest that Daxx but not Daxx 493–740 protects DMAP1 protein from proteasomal degradation in vivo.
In this study, two-hybrid screening has led to our identification of DNMT1 associated protein, DMAP1, as a Daxx-interacting protein. The interaction involves Daxx/N and DMAP1/C2. Daxx formed a complex with DNMT1 via DMAP1 in the nucleus. Furthermore, Daxx and DMAP1 additively repressed GR-mediated transcription, although DMAP1 had no effect on Fas/Daxx-mediated apoptosis in a murine pro-B cell line. Interestingly, Daxx protected DMAP1 protein from proteasomal degradation in vivo.
Daxx was reported to be involved in the Fas- and TGF-β-mediated apoptotic signaling pathway (1, 25). Daxx was originally cloned as a Fas-associated protein and binds specifically to the death domain of Fas, although Daxx by itself lacks a death domain (1). There are two independent signaling pathways downstream of Fas, involving the adapter protein Fas-associated death domain and Daxx (26). The activation of Fas-associated death domain induces a protease cascade (27), whereas that of Daxx enhances JNK activation, leading to apoptosis (1, 27). Overexpression of Daxx enhances Fas-induced apoptosis (1, 15, 24), and the targeted disruption of the Daxx gene in mice results in embryonic lethality (2). Daxx is also involved in coupling of type II TGF-β receptor signaling with components of the apoptotic machinery (25). Daxx associates with the cytoplasmic domain of the type II TGF-β receptor and transduces apoptotic signals by TGF-β (25). Recently, Daxx was also reported to be essential for IFN-induced suppression of B cell development (11). IFN-α enhances Daxx expression, with concomitant increases in Daxx protein levels and nuclear body translocation. Moreover, Daxx antisense oligonucleotides rescue IFN-α-treated pro-B cells from growth arrest and apoptosis. We also demonstrated that Tyk2 is essential for the transduction of IFN-α-induced suppression of B cell development through the activation of some signaling molecules other than STAT1, followed by the up-regulation and nuclear translocation of Daxx (10). The apoptotic signaling pathway downstream of Daxx still remains unknown. One candidate of Daxx targets is a JNK signaling pathway. Alternatively, the localization of Daxx is also important because the localization of Daxx in either the cytoplasmic or nuclear compartment was reported to be dependent upon the cell type and/or its functional status (28, 29, 30). In this study, we have tested whether a novel Daxx-interacting partner, DMAP1 involves in Daxx-mediated apoptotic signaling by using our conditional Fas/Daxx suicide within a murine pro-B cell line. However, DMAP1 had no effect on its signaling including JNK activation, although DMAP1 showed a constitutive association with Daxx in cells.
Recent studies demonstrated that Daxx might function as a transcriptional coregulator for several transcription factors such as Pax3, ETS1, and GR through direct protein-protein interactions (4, 6, 12). Daxx recruits nuclear factors possessing either histone acetyltransferase or HDAC to modulate transcriptional activity (5, 20). In fact, Daxx is shown to bind to HDAC2 (20). In addition, the transcriptional repression effect of Daxx could be modulated by subnuclear compartmentalization through protein-protein interactions. For example, PML has been shown to relieve the transrepression effect of Daxx through sequestering Daxx into the PML oncogenic domain (9). A nucleolar 58-kDa microspherule protein is also shown to relieve the transcriptional repression by Daxx through protein-protein interactions (21). DMAP1 is shown to have an intrinsic transcriptionally repressive activity (14). We report that Daxx and DMAP1 cooperatively repressed GR-mediated transcription in 293T cells. To our interest, Daxx protected DMAP1 protein degradation through proteasomal signaling pathway. These results strongly suggested that Daxx and DMAP1 form an effective repression complex in the nucleus.
Furthermore, our experiments revealed that another DMAP1-interacting partner also joined and formed a complex with DMAP1 and Daxx in the nucleus. DNMT1 is a major enzyme that maintains mammalian DNA methylation. DNA methylation is known to contribute to transcriptional silencing through several transcriptionally repressive complexes, which include methyl-CpG binding domain proteins and HDACs (31, 32, 33). The N-terminal noncatalytic domain of DNMT1 binds to both HDAC2 and DMAP1 (14), and can mediate transcriptional repression. DMAP1 is also shown to directly interact with tumor susceptibility gene 101 (TSG101) (14), a protein recently demonstrated to be a transcriptional corepressor (34), although we do not have any evidence yet whether TSG101 also interacts with Daxx. It has been also demonstrated that DMAP1 is targeted to replication foci through interaction with the N-terminal domain of DNMT1 throughout S phase, whereas HDAC2 associates with DNMT1 and DMAP1 only during late S phase following DNA replication (14). Thus, DMAP1 may mediate to form more effective repression complex in the nucleus by linking between Daxx and DNMT1, although DMAP1 by itself has a transcriptionally repressive activity. At the present time, we do not know whether the Daxx-DMAP1 interaction has any effect on DNA methylation through methyltransferase activity by DNMT1. This interaction may modify DNMT1 enzymatic activity to carry out an effective repression in the nucleus. Further work will be required to assess this possibility.
The present report describes both physical and functional interactions between Daxx and DMAP1. Daxx and DMAP1 with DNMT1 formed a complex and colocalized in the nucleus. DMAP1 enhanced Daxx-mediated repression of a GR transcriptional activity. Furthermore, Daxx protected protein degradation of DMAP1 in vivo. These results provide an important linkage between Daxx and DNMT1, which forms an efficient transcription repression complex in the nucleus.
We thank Dr. H. Ariga, Dr. T. Taira, Dr. X. Yang, and Dr. G. L. Hager, for their kind gifts of reagents. We also thank Dr. S. Matsuzawa for critical comments and Dr. J. Akiyama for encouraging our work.
This work was supported by grants-in-aid for Scientific Research from the Ministry of Education, Science, Sports and Culture in Japan, the Osaka Foundation for Promotion of Clinical Immunology, the Akiyama Foundation, the Suhara Memorial Foundation, Mochida Memorial Foundation for Medical and Pharmceutical Reasearch, and the Uehara Memorial Foundation.
Abbreviations used in this paper: JNK, c-Jun N-terminal kinase; DNMT, DNA methyltransferase; DMAP, DNMT1 associated protein; PML, promyelocytic leukemia protein; ETS, E26 avian leukemia oncogene 1; FKBP, FK506 binding protein; MMTV, murine mammary tumor virus; LUC, luciferase; HA, hemagglutinin; GR, glucocorticoid receptor, HDAC, histone deacetylase.