von Hippel-Lindau Partner Jade-1 Is a Transcriptional Co-activator Associated with Histone Acetyltransferase Activity*

Jade-1 was identified as a protein partner of the von Hippel-Lindau tumor suppressor pVHL. The interaction of Jade-1 and pVHL correlates with renal cancer risk. We have investigated the molecular function of Jade-1. Jade-1 has two zinc finger motifs called plant homeodomains (PHD). A line of evidence suggests that the PHD finger functions in chromatin remodeling and protein-protein interactions. We determined the cellular localization of Jade-1 and examined whether Jade-1 might have transcriptional and histone acetyltransferase (HAT) functions. Biochemical cell fractionation studies as well as confocal images of cells immunostained with a specific Jade-1 antibody revealed that endogenous Jade-1 is localized predominantly in the cell nucleus. Tethering of Gal4-Jade-1 fusion protein to Gal4-responsive promoters in co-transfection experiments activated transcription 5-6-fold, indicating that Jade-1 is a possible transcriptional activator. It was remarkable that overexpression of Jade-1 in cultured cells specifically increased levels of endogenous acetylated histone H4, but not histone H3, strongly suggesting that Jade-1 associates with HAT activity specific for histone H4. Deletion of the two PHD fingers completely abolished Jade-1 transcriptional and HAT activities, indicating that these domains are indispensable for Jade-1 nuclear functions. In addition, we demonstrated that TIP60, a known HAT with histone H4/H2A specificity, physically associates with Jade-1 and is able to augment Jade-1 HAT function in live cells, strongly suggesting that TIP60 might mediate Jade-1 HAT activity. Thus, Jade-1 is a novel candidate transcriptional co-activator associated with HAT activity and may play a key role in the pathogenesis of renal cancer and von Hippel-Lindau disease.

expressed in kidney and renal proximal tubule cells and may be involved in renal tubular epithelial cell differentiation, growth suppression, and apoptosis (1). Moreover, naturally occurring truncations and mutations of VHL altered its interaction with Jade-1, suggesting a correlation with renal cancer (2). Jade-1 has a strong possible PEST degradation domain (aa 5-28) and numerous possible sites for post-translational modification (1). The 509 amino acid Jade-1 protein contains one canonical plant homeodomain (PHD) finger (aa 203-253) followed by a noncanonical extended PHD (aa 257-371), both of which are zincbinding motifs. Mutational analysis studies demonstrated that both Jade-1 PHD zinc fingers were required for optimal interaction with wild-type 30-kDa VHL. Jade-1 has also been recently identified as a gene involved in anteroposterior axis development during mouse embryogenesis (3).
More than 300 gene products have been identified so far that contain one or more PHD-type zinc fingers (4 -6). Most characterized PHD proteins are found in the nucleus. PHD fingers are protein domains built around two zinc ions coordinated by cysteine residues and a histidine in a Cys 4 HisCis 3 motif. The properties of PHD zinc fingers make them good candidates for intracellular protein scaffolds because they are small, stable, and very diverse in sequence (7,8). There are several suggested functions for PHD fingers. A line of evidence strongly suggests that the PHD finger functions in chromatin remodeling. Thus, PHD motifs are found in transcriptional co-regulators and proteins of chromatin-modifying complexes, such as p300, CBP (6,9), ING1 (10), ING3, and TIF1 and Mi-2 family members (4,6). The extended PHD motif is found in several leukemia-associated proteins, such as AF10, AF17, and the mixed lineage leukemia protein MLL (11). The tandem of a canonical PHD followed by an extended PHD motif characteristic to Jade-1 is found in the closely related E9 (12), AF10, AF17, and BR140 (11). Several reports have provided evidence that PHD fingers may be protein-protein interaction domains. Thus, an extended PHD finger is responsible for oligomerization of the AF10 protein (13). It has recently been reported that the PHD finger of p300 binds isolated nucleosomes in vitro (4). In contrast to the bromodomain, which specifically binds acetylated nucleosomal histones, the PHD finger interacts with nucleosomal histones in an acetylation-independent manner. The PHD finger was also found to be an integral part of the CBP minimal acetyl transferase domain, which represents another function for the PHD finger (9,14). The PHD fingers of ING2 protein may function as phosphoinositide receptors (15). The PHD fingers in AIRE play a role in subnuclear targeting of this protein (16). PHD-type zinc fingers are targets of chromosomal translocations and mutations in several diseases, such as acute leukemias (MLL, AF10, AF17), ␣-thalassemia (ATRX), and autoimmune disease (AIRE-1) (17)(18)(19).
The correlation between the acetylation state of histones within chromatin and transcriptional regulation was proposed decades ago (20 -22). However, only over the last several years have proteins been identified that mediate histone acetylation (23,24). HAT enzymes acetylate the ⑀-amino groups of specific lysine residues on N-terminal tails of histone proteins that package DNA into chromatin. This packaging is mediated by nucleosome core particles, containing two copies of positively charged histones H2A, H2B, H3, and H4. HATs are recruited to specific promoters by DNA-bound transcriptional activators near specific histones. According to the histone code hypothesis, histone tails acetylated by these HATs serve as docking recognition sites for the binding of transcriptional cofactors and subsequent activation of general transcription (25,26). Most HATs are remarkably nonrandom and especially in vivo acetylate only certain lysine residues within specific histone tails. An important characteristic of all HATs is that they bind and functionally cooperate with other transcriptional regulators and HATs by assembling into multisubunit complexes, such as SAGA, ADA, PCAF, TIP60, NuA3, and NuA4 (reviewed in Ref. 24).
The most studied role of histone acetyltransferases involves targeted chromatin transcription. However, it has been suggested that global acetylation of histones, specifically histone H4, is required in other types of DNA metabolism, including DNA repair, replication and recombination (27)(28)(29). In support of this notion, a novel role for HAT activity has been demonstrated for a member of the MYST family, the known transcriptional regulator and histone acetyltransferase TIP60 (HIV-1-Tat-interactive protein) (27). The first hint suggesting a novel function resulted from purification and identification of proteins composing the TIP60 complex. The TIP60 complex consists of at least 14 distinct subunits, three of which are homologs of known proteins involved in DNA remodeling. Indeed, the purified TIP60 complex possesses ATPase, DNA helicase, and structural DNA binding activities. Most importantly, the ectopic expression of mutated TIP60 lacking HAT activity causes defects in the ability of the cell to repair DNA and to trigger DNA damage-induced apoptosis. It has been suggested that, depending on the level of damage, TIP60, like p53, will initiate DNA repair or an apoptotic response. The mechanism of this dual-role stress response by TIP60 is unclear, but it is hypothesized that the TIP60 complex may interact with cell-cycle checkpoint proteins to activate an apoptotic pathway in response to DNA lesions (27). It is interesting that the TIP60 complex and its yeast homologue Esa1 have unique substrate specificity and, in vivo, specifically acetylate histone H4 and H2A (30).
In this study, we set out to determine the molecular function of Jade-1. Because PHDs are associated with nucleosomal histones and are found in some HATs, we examined the hypothesis that Jade-1 might be a transcription factor associated with HAT activities. We determined Jade-1 nuclear association by using monospecific polyclonal Jade-1 antibodies that we generated from Jade-1 anti-serum. We then examined transcriptional and HAT-associated activities of full-length Jade-1 and mutated Jade-1 lacking PHD zinc fingers. Our data strongly suggest that Jade-1 is a novel candidate transcription factor associated with HAT activity. Moreover, we identified the first HAT associated with Jade-1, TIP60. TIP60 might therefore mediate Jade-1 HAT activity in live cells.

EXPERIMENTAL PROCEDURES
Cell Lines and Transfection-293T17 human embryonic kidney cells and HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, glutamine, and penicillinstreptomycin (Invitrogen). Subconfluent cells grown in 35-, 60-, or 100-mm dishes were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.
Antibodies-Jade-1 antiserum has been described previously (1). Human VHL monoclonal and FLAG M5 monoclonal antibodies were from BD Pharmingen and Sigma, respectively. Hemagglutinin (HA) monoclonal antibody was from Santa Cruz Biotechnology. Acetylated histones acetyl-H4 and acetyl-H3 polyclonal IgG were from Upstate Biotechnology. Goat anti-mouse and anti-rabbit IgG-HRP conjugate were from Bio-Rad. Protein A/G agarose was from Santa Cruz Biotechnology.
Constructs-The vectors used as reporters in chloramphenicol acetyltransferase (CAT) assays contain five Gal4 binding sites cloned upstream of the E1B, SV40, E4, and AdML promoters, which were gifts from Dr. T. Kouzarides (Wellcome/CRC Institute). FLAG-TIP60 expression vector was a gift from Dr. K. Kandror (Boston University School of Medicine). FLAG-Jade-1 has been described previously (1). FLAG-Jade-1 dd lacks amino acids 202-253 and 312-371, while FLAG-Jade-1 d lacks FIG. 1. Endogenous Jade-1 is localized to the nucleus. A, Jade-1 protein is extracted by high ionic strength. Confluent monolayers of 293T17 or HeLa cells were washed with PBS and extracted in situ for 5 min by sequential addition of 150 l of lysis buffer (50 mM HEPES, pH 7.0, 5 mM MgCl 2 , 1 mM EDTA, and 0.2% Triton X-100) containing increasing concentrations of NaCl. Equal aliquots of each extract were analyzed by SDS-PAGE with anti-Jade-1 rabbit serum by Western blot. B, analysis of purified Jade-1 antibody. HeLa cells were transfected with empty vector (lanes 1 and 3) or untagged Jade-1 expression vector (lanes 2 and 4). Cells were extracted with radioimmunoprecipitation assay buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS). Cell lysates (50 g/lane) were separated by SDS-PAGE and transferred onto nitrocellulose. Single lanes were cut out and probed with either Jade-1 whole serum (lanes 1 and 2) or purified Jade-1 antibody (lanes 3 and 4). Note that three minor bands appearing in Jade-1 overexpressing cells are detected by Jade-1 whole serum and by purified Jade-1 antibody (lanes 2 and 4, bands are indicated by horizontal marks) and are probably related to partial degradation of overexpressed Jade-1. It is noteworthy that the whole Jade-1 serum and Jade-1 purified antibody recognize both endogenous Jade-1 (lanes 1 and 3) and recombinant Jade-1 (lanes 2 and 4), indicating antibody specificity. WB, Western blotting. C, immunohistochemistry of Jade-1 with purified antibody. HeLa cells were processed for conventional fluorescence or scanning confocal microscopy analysis (c and d, magnification 400ϫ, 0.2-m slice in z-dimension; image shown is taken through the cell nuclei), as described under "Experimental Procedures." b and d, enlarged selected fields of images a and c, respectively. amino acids 312-371. FLAG-Jade-1 dC has a C-terminal deletion of amino acids 418 -509. The FLAG-Jade-1 dd and FLAG-Jade-1 d were PCRamplified from HA-Jade-1 dd and HA-Jade-1 d templates and inserted using HindIII/BglII and BglII/XbaI restriction sites, respectively. FLAG-Jade-1 dC was PCR-amplified using full-length FLAG-Jade-1 as a template and inserted into pFLAG-CMV2 plasmid using HindIII/XbaI restriction sites. HA-Jade-1 and HA-Jade-1 dd have been described previously (1). Plasmid pSG424 encodes the DNA-binding domain of Gal4 driven by the SV40 early promoter/enhancer (31). The specified Jade-1 coding sequence was fused in frame C-terminal to Gal4 amino acids 1 to 147 yielding Gal4-Jade-1 or Gal4-Jade-1 dd fusion proteins.
Reporter Gene Assay-Cells seeded in 60-mm dishes were co-transfected with the indicated amount of either SV40-promoter driven Gal4 DNA binding domain or SV40-promoter driven Gal4-Jade-1 or Gal4-Jade-1 dd and 2 g of CAT-reporter plasmid (E1B, SV40, E4, and AdML, described above). After 36 h of transfection, cells were washed in phosphate-buffered saline (PBS), re-suspended in 150 l of 0.25 M Tris, pH 8.0, and lysed by freezing and thawing three times (liquid nitrogen/ 37°C). Supernatants were clarified by centrifugation (5 min; 12,000 ϫ g). CAT assays were performed as described previously (31,32), with modifications. In brief, 30 l of cell lysates were incubated for 1 h at 37°C in 180 l of reaction mixture, containing 125 l of 0.25 M Tris, pH 8.0, 5 l of [ 14 C]chloramphenicol (CAT assay grade; Amersham Biosciences) and 20 l of 4.0 mM acetyl-CoA (American Bioanalytical). Acetylated products of the CAT reaction were excised from thin layer chromatography plates and quantitated on a scintillation counter. Data presented in bar graphs are means of three experiments Ϯ S.E.
Endogenous Core Histone Extraction-Cell layers grown in 60-mm dishes were washed with PBS and lysed for 5 min in 0.5 ml of 10 mM Tris buffer, pH 8.0, containing 0.6% Nonidet P-40, 150 mM NaCl, and 1 mM EDTA, supplemented with protease inhibitor mixture (Roche Diagnostics). The nuclear fraction was isolated by centrifuging lysates at 1200 ϫ g for 5 min. Total histones were extracted by suspending nuclei pellets in 100 l of 0.4 N H 2 SO 4 and incubating on ice for 20 min. Histone extracts were cleared by centrifugation (13,000 ϫ g for 10 min) and precipitated by addition of 0.5 ml of ice cold 20% trichloroacetic acid after centrifugation (13,000 ϫ g for 10 min). Precipitated histones were washed twice with acetone and solubilized in 1ϫ SDS sample buffer.
Immunoprecipitation and Histone Acetyltransferase Assay-All procedures were done at 4°C, unless otherwise noted. Immunoprecipitation and HAT assays were done as described previously (27,33) with some modifications. Cultured cells grown in 60-mm dishes were lysed in 50 mM Tris, pH 7.8, 0.5% Nonidet P-40, 150 mM NaCl, 1 mM EDTA, and 5 mM MgCl 2 and protease inhibitor mixture. Lysates were cleared by centrifugation at 12,000 ϫ g for 10 min. Relevant antibodies were added to 1 ml of lysates and incubated overnight. Protein A-agarose/protein G-agarose (1:1 mix, 15 l total; Santa Cruz Biotechnology) was added and the mixture rotated slowly for 4 h. The immune complexes were pelleted by brief centrifugation and washed three times with 1 ml of lysis buffer. The immune complexes were mixed with 30 l of HAT reaction mix containing 50 mM Tris pH 7.8, 10% glycerol, 2 mM MgCl 2 , 0.5 mM EDTA, 15 M trichostatin A, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 0.8 mg/ml calf thymus core histones. The HAT reaction was initiated by addition of 1 l of [ 3 H]acetyl CoA (7.4 Gbq/mmol; PerkinElmer Life and Analytical Sciences). After 30 min of incubation, the reaction was stopped by addition of SDS-sample buffer. One-tenth aliquot of each immunoprecipitation sample was frozen at Ϫ20°C and later used to analyze proteins in immunocomplexes by SDS-PAGE/Western blot with corresponding antibodies. The rest of these samples were used to assess incorporation of [ 3 H]acetyl-CoA into core histones. Samples were separated on 15% SDS-PAGE, and core histones were visualized by staining with Coomassie Blue. Bands were excised with a razor blade, and histones were extracted from the polyacrylamide gels by incubating in Solvable reagent (PerkinElmer Life and Analytical Sciences) according to the manufacturer instructions. Ultima Gold scintillation mixture was added to the samples, and radioactivity was quantitated with a liquid scintillation analyzer (all from PerkinElmer Life and Analytical Sciences). Data presented in bar graphs are means of at least three experiments Ϯ S.E.
Purification of Monospecific Jade-1 Polyclonal Antibody-A highly specific fraction of Jade-1 polyclonal antibody was affinity-purified from Jade-1 whole rabbit antiserum using the 75-kDa FLAG-Jade-1 antigen and a scaled-up Western blotting approach. FLAG-Jade-1 includes additional 5Ј untranslated Jade-1 sequence that was present in the original library clone, increasing its mass to 75 kDa. This size difference was helpful in determining the specificity of the purified Jade-1 antibody (see below and Fig. 1B). 293T cells grown in 100-mm dishes were transiently transfected with the FLAG-Jade-1 expression vector. Cells were extracted sequentially with 900 l of 50, 150, and 420 mM NaCl in ice-cold solution containing 50 mM HEPES, pH 7.0, 5 mM MgCl 2 , 1 mM EDTA, and 0.2% Triton X-100. The 420 mM fraction of nuclear proteins enriched with FLAG-Jade-1 protein was separated by large scale preparative SDS-PAGE, transferred onto nitrocellulose membrane, and the FLAG-Jade-1 position was localized by Western blot. The membrane region corresponding to 75-kDa FLAG-Jade-1 antigen was cut out and used for affinity purification of antibody. The strip of nitrocellulose was blocked, incubated with Jade-1 antiserum, and washed according to a regular Western blot procedure. Jade-1-specific antibodies were eluted from the strips with 100 mM glycine, pH 2.5. pH was immediately neutralized with 1 M Tris, and antibody was dialyzed and A-C, 60-mm dishes of 293T17 cells were co-transfected with 1 g of AdML, E4, SV40, or E1b (data not shown) promoter-reporter constructs (see "Experimental Procedures") and with the indicated amounts of Gal4-Jade-1 fusion construct. The amount of DNA was adjusted to total 8 g by addition of empty Gal4 expression vector. Cells were harvested and CAT assays were performed as described under "Experimental Procedures." Gal4-Jade-1 expression in total cell lysates was visualized by Western blot with anti-Jade-1 serum (A and B, bottom). C, quantitation of Gal4-Jade-1 effects on transcriptional activities of AdML, E4, and SV40 promoters. The amount of product formed after CAT reaction was determined by liquid scintillation analyzer (PerkinElmer Life and Analytical Sciences). For each reporter, the activity derived from Gal4 expression vector was normalized to 1.0, and the activity of Gal4-Jade-1 was expressed relative to this.
concentrated with Centricon tubes. The specificity of purified Jade-1 antibody was then tested by Western blot against endogenous and untagged wild-type 64-kDa Jade-1 protein.
Immunocytochemistry and Confocal Microscopy-The intracellular localization of endogenous Jade-1 was studied by indirect immunofluorescence as described previously (34). HeLa cells were seeded onto glass coverslips placed in 12-well cluster plates and grown overnight. Cells were washed with PBS and then fixed with 4% paraformaldehyde for 20 min at room temperature, washed three times with PBS, permeabilized with 0.1% Triton X-100 in PBS for 15 min, and blocked with 1% bovine serum albumin in 0.1% Triton X-100 in PBS for 1 h. The cells were incubated with Jade-1 antiserum (1:200 diluted in blocking solution), preimmune rabbit serum, or with Jade-1 purified antibody in blocking solution for 1 h and washed with 0.1% Triton X-100 in PBS three times. Cells were then incubated with indocarbocyanine-conjugated secondary IgG (Jackson ImmunoResearch) in blocking solution and washed with 0.1% Triton X-100 in PBS three times. Coverslips were mounted onto slides with Gelvatol.
Indocarbocyanine signal was analyzed first by conventional fluorescence microscopy (Nikon OptiPhot). Control cells not incubated with primary antibody showed negative signal. Cells incubated with preimmune rabbit serum showed strong stress fiber fluorescence in the cytoplasm and showed no signal in cell nuclei. Cells incubated with the whole Jade-1 antiserum showed intense stress fiber and nuclear fluorescence. Cells incubated with Jade-1 purified antibody showed strong and clear nuclear signal and weak sparse extranuclear signal with no visible association to any cell organelle. Cells stained with Jade-1 purified antibody were further analyzed by confocal laser scanning microscopy (UltraVIEW LCI; PerkinElmer Life and Analytical Sciences).

Jade-1 Is Localized to Cell
Nuclei-Jade-1 protein has two zinc finger PHD motifs, which are found most commonly in transcription factors and proteins present in chromatin remodeling complexes (6). Because nuclear proteins associated with chromatin can be extracted from the nucleus with buffers of high ionic strength, we performed sequential salt extractions of Jade-1 with increasing concentrations of NaCl and low concentrations of Triton X-100. In both HeLa and 293T cells, Jade-1 was extracted predominantly with buffer containing 420 mM NaCl (Fig. 1A), suggesting an association with the nucleus and/or cytoskeleton. Jade-1 subcellular localization was then determined by Jade-1 immunofluorescence. For these studies, a highly specific fraction of Jade-1 polyclonal antibody was affinity-purified from Jade-1 whole rabbit antiserum using 75-kDa FLAG-Jade-1 antigen and a scaled-up Western blotting approach (see "Experimental Procedures"). The specificity of purified Jade-1 antibody was then tested by Western blot. HeLa cells were transiently transfected with empty vector or wild-type, untagged Jade-1 expression vector. Total cell lysate proteins were separated by SDS-PAGE and transferred to a membrane. Strips of nitrocellulose containing HeLa cell total lysates were probed with whole Jade-1 serum or with purified Jade-1 antibody. The whole serum recognized multiple protein bands in addition to endogenous Jade-1 in untransfected HeLa cell extracts (Fig. 1B, lane 1). In contrast, the purified antibody recognized one major band corresponding to 64-kDa endogenous Jade-1 (Fig. 1B, lane 3). As expected, the intensity of the Jade-1 band detected with Jade-1 whole serum was significantly increased in HeLa cells overexpressing Jade-1 (Fig. 1B,  lane 2). It is noteworthy that a proportional increase of the Jade-1 band intensity was detected in membrane strips probed with purified Jade-1 antibody (Fig. 1B, lane 4), demonstrating that this antibody recognizes both endogenous and recombinant Jade-1. This demonstrates that Jade-1 purified antibody is essentially mono-specific and is an appropriate tool for cell immunostaining. The affinity purified Jade-1 antibody was used on HeLa cells for conventional immunofluorescence (Fig.  1C, top) and laser-scanning confocal immunofluorescence imaging (Fig. 1C, bottom). It was striking that all images revealed Jade-1 to be predominantly intranuclear, exclusive of nucleoli. In addition, sparse cytoplasmic Jade-1 signal was also detectable. It has been previously reported that Jade-1 is localized to cytoplasm and nuclei in several types of cells (1). The current study revealed a significantly higher nuclei-to-cytoplasm ratio of Jade-1 signal, which is probably attributable to greater specificity of purified Jade-1 antibody as well as differences in the method of cell fixation (see "Experimental Procedures").
Jade-1 Tethered to Viral Promoters Activates Transcription-The nuclear localization and presence of PHD fingers in the Jade-1 molecule prompted us to examine Jade-1 transcrip- tional function. To assess Jade-1 transcriptional activity, we used an established promoter-reporter assay that was described previously (35). In this assay, we determined the ability of Jade-1 to activate transcription of specific viral promoters linked to a CAT reporter. We used a set of four vectors with viral promoters (E1B, SV40, E4, and AdML) containing five upstream Gal4 DNA binding sites (gift from Dr T. Kouzarides) and an expression vector that fuses the DNA binding domain of Gal4 transcription factor to the full-length Jade-1 protein. The Gal4 domain provides binding of the Gal4-Jade-1 fusion to the corresponding sites of the promoter and enables Jade-1 to transactivate. Gal4-Jade-1 fusion or control Gal4 expression vectors were transiently co-transfected with each of these specific promoters, and CAT assays were performed. Gal4-Jade-1 up-regulated transcription rates of the AdML and E4 promoters by 5.8-and 3.1-fold, respectively (Fig. 2, A-C). The upregulation occurred in a dose-dependent manner (Fig. 2, A-C) and correlated with levels of Gal4-Jade-1 protein expression, as assessed by Western blots with Jade-1 antibody (Fig. 2, A and  B, bottom). Note that the amounts of recombinant Gal4-Jade-1 protein synthesized by cells are modest and do not greatly exceed the amount of endogenous Jade-1 (Fig. 2, A and B,  bottom), indicating that the Gal4-Jade-1 dose response did not reach saturation and is still within its linear range. Thus, Gal4-Jade-1 potently activates transcription of at least two of four tested promoters. Others have shown in similar assays that the minimal acetyl transferase domain of CBP exhibits promoter selectivity, in that it transactivates the AdML and E4 but not the SV40 and E1B promoters (35). It is interesting that neither the SV40 (Fig. 2C) nor the E1b (not shown) viral promoters were affected by Gal4-Jade-1, which is in correlation with previously published data for the acetyltransferase domain of CBP (35). Together, these data indicate that nuclear protein Jade-1 might be a transcription factor.

Deletion of the PHD Zinc Fingers Abolishes Transcriptional
Activity of Jade-1-PHD fingers participate in protein-protein interactions and are required for transcriptional function of some known transcriptional activators (9,14). Thus, a series of specific mutations, including a PHD finger deletion, simultaneously inactivated HAT and transcriptional function of CBP.
To determine the role of the PHD fingers in Jade-1 transcriptional function, we generated a Gal4-Jade-1 dd mutation lacking amino acids 203-253 and 312-371 and examined its transcriptional activity. An AdML promoter-reporter construct was cotransfected with either wild-type Gal4-Jade-1 wt or mutated Gal4-Jade-1 dd , and CAT assays were performed. As expected, the wild-type Gal4-Jade-1 wt reproducibly activated transcription driven by the AdML promoter. However, the mutated Gal4-Jade-1 dd lacking PHDs did not activate transcription (Fig. 3). It is noteworthy that both proteins were expressed at comparable levels (Fig. 3, bottom). These data further support the specificity of Jade-1 transcriptional activity and demonstrate that the PHD fingers are essential for Jade-1 transcriptional function.
Jade-1 Promotes Histone H4 Acetylation-PHD zinc fingers are found in nuclear proteins associated with histones, including those with intrinsic HAT activity, such as p300 and CBP. Moreover, it has been shown recently that PHD fingers can bind histones within the nucleosomal context (4). Thus far, we demonstrated that PHD-protein Jade-1 is localized to nuclei and can activate transcription when tethered to a viral promoter. We hypothesized that Jade-1 may be associated with HAT activity. To examine this possibility, we compared the levels of acetylated endogenous histones in cells transiently transfected with wild-type Jade-1, FLAG-Jade-1, or empty vector. Cells were harvested and the total histone fraction was purified from the nuclei by acid extraction (36). Purified total histone samples were analyzed for acetylated histone H3 and  Fig. 4. Histones were extracted (bottom), and levels of histone H4 acetylation were determined (see Fig. 4 and "Experimental Procedures" for details). H4 levels of expression by Western blot with corresponding antibodies. Jade-1 expression was monitored in parallel whole cell lysate samples with Jade-1 antibody (Fig. 4, top). It is remarkable that Jade-1 and FLAG-Jade-1 were able to dramatically increase levels of acetylated H4 in a dose-dependent manner (Fig. 4, second panel; see also Figs. 6, A and B, and 9B). In contrast, levels of acetylated H3 were unaffected by Jade-1 (Fig. 4, third panel). As expected, the general inhibitor of histone deacetylases trichostatin A, which was used as positive control, strongly up-regulated levels of both acetyl-H3 and acetyl-H4 (Fig. 4, lane 7). A vector expressing the DNA-binding domain of transcription factor Gal4 alone served as a negative control and did not affect acetylated H4 levels (data not shown). Thus, Jade-1 is a nuclear PHD protein and is associated with HAT activity specific for histone H4.
Jade-1-associated HAT Activity Requires PHD Zinc Finger-Because deletion of the PHD fingers knocked out Jade-1 transcriptional function, we examined whether the PHD fingers are essential for Jade-1-associated HAT activity. Hence, we generated three deletion mutations of Jade-1: Jade-1 dd missing both PHD fingers, Jade-1 d missing only the second PHD finger, and Jade-1 dC missing the C terminus beyond the second PHD finger (Fig. 5A). These constructs were examined for their ability to promote histone H4 acetylation in a whole-cell HAT assay. It was remarkable that although overexpression of full-length Jade-1 resulted in dramatic hyperacetylation of histone H4, the same levels of Jade-1 dd lacking both PHD fingers failed to promote H4 acetylation (Fig. 5B, lanes 2 and 3). Moreover, overexpression of Jade-1 d lacking only the second PHD finger resulted in reduced levels of H4 acetylation (Fig. 5B, lane 4), indicating that this specific domain composed of 61 aa residues is required for full Jade-1-associated HAT activity. It is interesting that overexpression of Jade-1 dC lacking C-terminal 92 aa residues also resulted in attenuated levels of H4 acetylation (Fig. 5B, lane 5), suggesting that this portion of Jade-1 polypeptide may be partly dispensable for its activity in this functional assay. These data also strongly support the specificity of Jade-1-associated HAT activity.
Jade-1 Lacking a Single PHD Finger Possesses a Dominantnegative Phenotype-Because PHD fingers are known to participate in protein-protein interactions, we hypothesized that deletion of this domain might result in a dominant-negative form of Jade-1. To investigate this possibility, we examined the effect of the Jade-1 d mutation on Jade-1 wt -mediated histone H4 hyperacetylation. We reasoned that if Jade-1 d could act in a dominant-negative manner, it should functionally compete with its wild-type counterpart and thus prevent Jade-1 wt -mediated H4 hyperacetylation. Cells were transfected with increasing amounts of Jade-1 wt DNA only or in combination with Jade-1 d . Jade-1 wt and Jade-1 d levels of expression were monitored by Western blot. As usual, expression of Jade-1 wt protein increased acetyl-H4 in a dose-dependent manner (Fig. 6A, lanes  1-4). In contrast, the presence of Jade-1 d significantly diminished histone H4 acetylation by Jade-1 wt protein (Fig. 6, A, compare acetyl-H4 levels in lanes 2-4 with those in lanes 5-7, and B, for quantitation), demonstrating that in a whole cell HAT functional assay Jade-1 d acted in a dominant-negative manner.
Histone Acetyltransferase TIP60 Interacts with Jade-1 and Enhances its HAT-Associated Function-Although our data clearly support an association of Jade-1 with HAT activity, the analysis of Jade-1 polypeptide using the Conserved Domain Architecture Retrieval Tool (CDART; www.ncbi.nlm.nih.gov) did not reveal the presence of an acetyltransferase domain. In addition, immunopurified and immobilized wild-type and tagged Jade-1 demonstrated reproducible but rather weak acetyltransferase activities against exogenously added core histones in a cell free HAT assay (Fig. 8A, bar graph, and data not shown) that could not explain the robust effects of Jade-1 on acetylation of endogenous histone H4 in a whole cell HAT assay (Figs. 4 -6). This suggested that Jade-1-associated HAT activity might be mediated by another protein with intrinsic acetyltransferase activity interacting with Jade-1. The dominant-negative phenotype of Jade-1 d also supported the notion that Jade-1 might physically interact with other proteins mediating Jade-1 HAT activity in vivo and in vitro.
We set out to determine what protein might mediate Jade-1-associated HAT activity and reasoned that a potential Jade-1 partner should be a HAT with histone H4 specificity. Although  5-7). In lane 8, cells were transfected with 4 g of Jade-1 d only. The total amount of DNA in each transfection sample was adjusted to 4 g by filling in with FLAG-empty vector plasmid. FLAG-Jade-1 and mutated protein expression levels in total cell extracts were monitored by Western blot with Jade-1 serum (top). Histones were extracted (bottom), and levels of histone H4 acetylation were determined by Western blot (middle; see legend to Fig. 5 and "Experimental Procedures" for details). Note that FLAG-Jade-1 d attenuated FLAG-Jade-1 wt -dependent dose response of histone H4 acetylation (compare acetyl H4 levels in lanes 2 and 5, 3 and 6, and 4 and 7). B, quantitation of FLAG-Jade-1 d effects on FLAG-Jade-1 wt -associated HAT activity. Quantitation of Western blot in Fig. 7A was done by densitometry. Signal intensities for acetyl-H4 were normalized to the intensities of the total histones and plotted as multiples of control (0 g of FLAG-Jade-1 wt ) for both FLAG-Jade-1 wt (OE) and FLAG-Jade-1 wt /FLAG-Jade-1 d (f) dose responses.
numerous HATs acetylate nucleosomal histones H3 and H4, so far, few known HATs have unique specificity for histone H4 (24). TIP60 has been previously characterized as a HAT with histone H4 specificity and seemed to be a good candidate (27,30). To examine whether TIP60 might mediate Jade-1 effects, we first determined whether Jade-1 directly interacted with TIP60 using a co-immunoprecipitation approach. Cells were transiently transfected with wild-type Jade-1, HA-Jade-1, or Gal4-Jade-1, and/or FLAG-TIP60. The Jade-1 complex was immunoprecipitated with Jade-1 antiserum or corresponding tag antibody, and the presence of FLAG-TIP60 in Jade-1 immunoprecipitates was assessed by Western blot with FLAG antibody (Fig. 7, A). In a reverse experiment, FLAG-TIP60 was immunoprecipitated with FLAG antibody, and the presence of Jade-1 was assessed by Western blot with Jade-1 antiserum (Fig. 7B). Jade-1, HA-Jade-1, and Gal-4Jade-1 were able to pull down TIP60 (Fig. 7A, top) and TIP60 was able to pull down all three species of Jade-1 (Fig. 7B, top), demonstrating that the two transfected proteins strongly interact. Moreover, endogenous Jade-1 immunoprecipitated with Jade-1 specific polyclonal antiserum efficiently pulled down a significant amount of FLAG-TIP60 (Fig. 7A, lane 1), whereas immunoprecipitated FLAG-TIP60 pulled down endogenous Jade-1 (Fig. 7B, lane 1), supporting the specificity of the Jade-1-TIP60 interaction. Therefore, both membranes were stripped and reprobed with Jade-1 and FLAG antibodies (Fig. 7, A and B, bottom). This allowed evaluation of the amount of the interacting partners in the complex and supported the efficiency of Jade-1-TIP60 binding. We then examined whether deletion of the PHD zinc fingers from Jade-1 molecule will affect these interactions. We were surprised to find that full-length Jade-1 and Jade-1 dd pulled down TIP60 with similar efficiency (Fig. 8A), indicating that the PHD zinc fingers are not involved in the Jade-1-TIP60 physical interaction. Thus, these results demonstrate that Jade-1 physically interacts with TIP60 and suggest that TIP60 might mediate Jade-1-associated HAT activity.
To further explore this possibility, we examined whether TIP60 co-immunoprecipitated with Jade-1 or Jade-1 dd can enhance [ 3 H]acetyl incorporation into core histones in a cell free immunoprecipitation-HAT assay. An aliquot of each immunoprecipitation sample presented in Fig. 8A, top two panels, was subjected to an IP-HAT assay. It is clear that the HAT activities in HA-Jade-1 and HA-Jade-1 dd immunoprecipitates were increased up to 3-fold with FLAG-TIP60 (Fig. 8A, bar graph). The increase in HAT activity was proportional to the amount of coimmunoprecipitated FLAG-TIP60; hence, similar amounts of FLAG-TIP60 immunoprecipitated directly without Jade-1 had comparable levels of activity (data not shown). In a reverse experiment, FLAG-TIP60 was immunoprecipitated alone or in the complex with HA-Jade-1, and HAT activities were determined (Fig. 8B, bar graph). As expected, TIP60 alone potently acetylated core histones. TIP60 bound to Jade-1 also acetylated histones with similar efficiency, indicating that the TIP60-Jade-1 complex is capable of acetylating core histones in vitro in a cell free assay. In sum, these data strongly suggest that TIP60 might mediate Jade-1 effects on levels of acetylated endogenous histone H4 in an intact cell.
To examine the functional interactions of Jade-1 and TIP60 in an intact cell, we studied the effects of TIP60 on Jade-1associated induction of endogenous histone H4 acetylation. Cells were transfected with increasing amounts of Jade-1 without or with TIP60, and levels of endogenous acetylated histone H4 and histone H3 were assessed by Western blot. As expected, Jade-1 alone increased acetylation of histone H4 in a dose-dependent manner (Fig. 9, A, lanes 1, 3, 5, and 7, and B), whereas Jade-1 dd failed to exert any effects (Fig. 9, A, lane 9, and B). Overexpression of TIP60 alone, a potent HAT with histone H4 specificity, failed to modulate levels of endogenous acetylated histone H4  1). In addition, note that the amount of Jade-1 in each sample is proportional to the amount of pulled down TIP60, indicating that interactions are efficient and specific. B, FLAG-TIP60 pulls down endogenous and transfected Jade-1. Transfection and immunoprecipitations were done as described above (A), except that FLAG monoclonal antibody was used to precipitate FLAG-TIP60. Co-immunoprecipitated endogenous (top, lane 1) and transfected Jade-1 (top, lanes 2-4) were visualized with Jade-1 antiserum. Lane 5 is a negative control. To visualize immunoprecipitated FLAG-TIP60, membranes were stripped and re-probed with FLAG antibody (bottom). (Fig. 9A, lane 2, and B). It is striking that co-expression of TIP60 with Jade-1 further augmented levels of acetyl-H4, demonstrating that Jade-1 co-operates with HAT TIP60 in an intact cell (Fig.  9, A, lanes 4, 6, and 8, and B). In contrast, co-expression of TIP60 with Jade-1 dd failed to augment levels of acetyl-H4 (Fig. 9A, lane  10). Note that the levels of acetylated histone H3 were not changed in any samples, supporting the specificity of observed effects (Fig. 9A, middle). These data strongly suggest that TIP60 might mediate Jade-1-associated effects on levels of endogenous acetylated histone H4.
Thus far, we have demonstrated that in vitro TIP60 promotes histone acetylation of core histones whether in the complex with Jade-1 or Jade-1 dd . In contrast, in intact cells, TIP60 failed to co-operate with mutated Jade-1 dd in promoting hyperacetylation of endogenous histone H4, indicating a requirement of intact PHD zinc fingers for an in vivo functional interaction of TIP60 and Jade-1. A recent study demonstrated that the PHD zinc finger of CBP binds nucleosomal histones in vitro (4). We speculate that PHD fingers are required for a nuclear Jade-1 complex to bind to the nucleosome, to provide a docking site for TIP60 and thereby promote acetylation of nucleosomal histone H4 (Fig. 10). DISCUSSION Jade-1 is a novel PHD zinc finger family protein that interacts with the VHL tumor suppressor (1, 2). Jade-1 is highly FIG. 8. Jade-1 and Jade-1 dd lacking PHD zinc fingers precipitate TIP60 and HAT activity. Cells were co-transfected with HA-Jade-1 or HA-Jade-1 dd with or without FLAG-TIP60 and lysed and processed for immunoprecipitation (IP) with HA antibody. See detailed description in the legend to Fig. 7. After washing, the immobilized immunocomplexes were used for IP-HAT assay with core histones as described under "Experimental Procedures." The HAT reaction was stopped by addition of SDS-sample buffer. One-tenth aliquot of each IP-HAT sample was analyzed by Western blot (WB) with FLAG antibody to visualize FLAG-TIP60 pulled down by HA-Jade-1 or HA-Jade-1 dd (I). Note that HA-Jade-1 and HA-Jade-1 dd both efficiently pulled down FLAG-TIP60, indicating an interaction. Membranes were stripped and re-probed with HA antibody to visualize HA-Jade-1 and HA-Jade-1 dd in precipitates (II). In III and IV, Western blots of the total cell lysates (10% of input) were probed with FLAG (III) and HA (IV) antibody to visualize FLAG-TIP60, HA-Jade-1 and HA-Jade-1 dd . The rest of each of IP-HAT sample was subjected to 15% SDS-PAGE to separate and excise core histone substrates. [ 3 H]Acetyl incorporation into histones was determined as described under "Experimental Procedures" (V). B, FLAG-TIP60 in complex with Jade-1 potently acetylates core histones. Cells were transfected with FLAG-TIP60 with or without HA-Jade-1. Cells were lysed and processed for IP-HAT assay as described in A and under "Experimental Procedures," except FLAG antibody was used. One tenth aliquot of IP-HAT reaction was analyzed by Western blot with HA (I) and FLAG antibody (II) to visualize co-immunoprecipitated HA-Jade-1 and FLAG-TIP60. The rest of each sample was used to determine HAT activity of the FLAG-TIP60 and FLAG-TIP60-Jade-1 complex (V). Note that because of FLAG antibody efficiency and concentration, the FLAG-TIP60 direct precipitation with FLAG antibody yielded 8-fold more FLAG-TIP60 (as assessed by densitometry) than co-precipitation of FLAG-TIP60 via HA-Jade-1 (compare FLAG-TIP60 bands intensities in A, I, and B, II). As a result, the assessed HAT activities are proportionally higher in these samples (compare V in A and B). expressed in kidney and renal proximal tubule cells (1,2), which are renal cancer precursors. Jade-1 may be involved in renal tubular epithelial cell differentiation, growth control and apoptosis (1). 2 The molecular function of Jade-1 is undefined and was the specific focus of the current study.
We have made the following major findings: 1) Jade-1 is localized predominantly to the nucleus and is capable of activating transcription when tethered to viral promoters; 2) Jade-1 is associated with HAT activity specific for histone H4; 3) Jade-1 physically and functionally interacts with TIP60, a powerful HAT with histone H4 specificity; and 4) both transcriptional and HAT-associated functions of Jade-1 are fully dependent on its PHD zinc fingers. The presented set of experimental observations indicates that Jade-1 is a candidate transcription factor and is associated with acetyltransferase activity specific for histone H4. Transcriptional and HAT-associated Jade-1 activities require intact PHD zinc fingers within the Jade-1 polypeptide. Our data strongly suggest that TIP60 is at least one of the candidates that might mediate Jade-1-associated in vivo HAT activities.
The association of Jade-1 with HAT activity is, in part, similar to the HAT association of the mammalian ING (10) family and their yeast orthologs, the Yng (37) family of proteins. The ING1 candidate tumor suppressor gene expresses a family of PHD proteins that localize to the nucleus and are growth inhibitory. p33 ING1b is the most characterized isoform of Ing1 and is involved in apoptosis and cell cycle regulation, presumably via association with p53, PCNA, and HAT complexes (10,37,38). Ectopic expression of p33 ING1b in human fibroblasts resulted in elevated levels of acetylated histones H4 and H3 (38). We find that overexpression of Jade-1 in 293T cells resulted in a robust increase of endogenous acetylated histone H4, but not histone H3. Yeast protein Yng2 is associated with the yeast homolog of TIP60, Esa1 (37), whereas another ING family member, Ing3 protein, was identified as a specific subunit of the human TIP60 complex (39). Likewise, we find a physical and functional interaction between Jade-1 and TIP60. The most important observation supporting in vivo interaction of TIP60 with Jade-1 is the ability of TIP60 to enhance Jade-1-mediated acetylation of endogenous histone H4. It is interesting that others have shown that TIP60 by itself in vitro can only acetylate individual histones and is unable to acetylate isolated nucleosomal chromatin substrates. Two other proteins were required for the TIP60 to gain affinity toward nucleosomal histones. One of the required proteins of the minimal triple core complex conferring nucleosomal affinity to TIP60 was Ing3 (39). It is conceivable that the whole cell HAT assay in our study is an indirect measurement of Jade-1or TIP60-mediated effects on acetylation of native nucleosomal histones. We demonstrate that although expression of Jade-1 alone promoted endogenous histone H4 acetylation, expression of TIP60 alone was insufficient to do so. However, co-expression of TIP60 further enhanced H4 acetylation induced by Jade-1, demonstrating the cooperative relationship of these two proteins. We suggest that Jade-1 targets TIP60 to nucleosomal histones and enables TIP60 to acetylate histone H4 lysine residues in vivo. It is possible that, like Ing3, Jade-1 might play a role in conferring TIP60 affinity to nucleosomal histones (Fig.  10). More direct examination of Jade-1's ability to cooperate with TIP60 in acetylating nucleosomal histones in vitro is required to provide evidence for this proposed mechanism.
PHDs are often found in proteins that function in the formation, maintenance, or regulation of chromatin structure and are thought to function as protein interaction domains. According to a recent study, the isolated PHD finger of p300 was capable of binding to nucleosomal histones as determined by an in vitro EMSA assay (4). We show that, in intact cells, deletion of the two PHD fingers abrogated Jade-1-associated HAT activity. It is possible that this mutation impaired Jade-1's ability to associate with nucleosome and target endogenous HATs, such as TIP60 to their substrate. It is not excluded that other, as-yet unidentified proteins might also be required for this function of Jade-1. Like Yng2 (37), TIP60 physical interaction  1, 3, 5, 6, and 9) or the presence (lanes 2, 4, 6, 8, and 10) of TIP60. Thirty-six hours after transfection, cells were harvested, histones extracted as described under "Experimental Procedures," and levels of acetylated histone H4 (top) and H3 (middle) were determined by Western blot. Jade-1 and TIP60 levels of expression were monitored by Western blot (data not shown). Higher levels of acetyl histone H4 in sample 9 reflect uneven loading (see the total histone image, bottom). Note that TIP60 did not affect acetylation of histones (lane 1), but enhanced levels of acetyl-H4 when co-transfected with 1 g (lanes 3 and 4), 2 g (lanes 5 and 6), and 3 g (lanes 7 and 8) of Jade-1. Jade-1 dd failed to induce histone acetylation (lane 9), and TIP60 failed to rescue Jade-1 dd negative phenotype (lane 10). B, densitometry of results from A. with Jade-1 did not require PHD fingers, in that TIP60 and its HAT activity were precipitated by mutated Jade-1 dd .
The TIP60 complex has been previously isolated from stably transfected HeLa cells (27). Fourteen individual specific proteins were co-purified with TIP60. Several proteins of the TIP60 complex turned out to be well known proteins; at least three of them are involved in chromatin and DNA remodeling, such as ruvB-like DNA helicases and SWI2/SNF2-like ATPasecontaining p400. Since then, it has been reported that thyroid receptor co-activating protein (TRCp120), DNA methyltransferase 1-associated protein (DMAP1), and two other related proteins are previously unrecognized components of the TRRAP/TIP60 HAT complex (40). A most recent study characterized the TIP60 complex from mammalian and yeast cells by a similar biochemical approach and identified at least three more new members of native stable TIP60 complex (39). Whether Jade-1 might be among the unidentified members of a stable TIP60 complex is unclear. On the other hand, it is possible that Jade-1 may not be a constitutive member of the TIP60 complex. Interaction of the TIP60 complex with endogenous Jade-1 in vivo might be an inducible event, perhaps depending on other factors and/or stimuli. Functional studies demonstrate that the TIP60 complex is involved in apoptosis and chromatin remodeling/DNA repair processes induced by ␥-irradiation and that these functions are dependent on TIP60 acetyltransferase activity (27). Jade-1 has growth inhibitory properties, promotes apoptosis, and may be a tumor suppressor. This and other available information suggest that together with TIP60, Jade-1 might be involved in similar biological activities triggered by a cellular stress, such as DNA damage.
In this study, we demonstrated that Jade-1 can activate transcription when tethered to two viral promoters. The magnitude of activation was modest but very reproducible and specific. The specificity of this effect is supported by the fact that, similar to CBP, Jade-1 activated only two of four tested promoters and that the deletion of the two PHD fingers completely inactivated Jade-1 transcriptional function. It is unlikely that the deletion of PHD fingers resulted in impaired nuclear transport and therefore inhibited Jade-1 transcriptional and HAT activities. Thus, genetic manipulations aimed at identifying developmentally regulated genes in the mouse led to Jade-1 gene disruption, resulting in the expression of only a 47-amino acid fragment of Jade-1 (3). This Jade-1 fragment lacking PHD fingers was efficiently targeted to the nucleus in mouse embryo in a developmentally regulated manner. Taking into consideration the current study, as well as data reported by others, we hypothesize that the PHD fingers might be important for Jade-1 association with nucleosomal histone and for targeting other required regulatory proteins to specific promoter regions.
In sum, the current study indicates that VHL-interacting protein Jade-1 is a strong candidate transcription factor and is associated with acetyltransferase activity specific for histone H4. Our data support the idea that TIP60 is at least one candidate that might mediate Jade-1-associated HAT activities in vivo. Jade-1 interaction with TIP60 implies a potential role in DNAdamage stress response, cell cycle and apoptosis, cellular processes directly related to cancer development. Previous studies demonstrated that naturally occurring mutations of VHL altered its interactions with Jade-1, suggesting a renal cancer relationship. It would be critical to investigate whether the VHL tumor suppressor, an established partner of Jade-1, might be involved in Jade-1 novel nuclear activities, including its interactions with TIP60. Such studies might provide a link between TIP60mediated Jade-1 nuclear function and VHL-related renal cancer.