Tip60 and HDAC7 interact with the endothelin receptor a and may be involved in downstream signaling.

Endothelins exert their biological effects through G protein-coupled receptors. However, the precise mechanism of downstream signaling and trafficking of the receptors is largely unknown. Here we report that the histone acetyltransferase Tip60 and the histone deacetylase HDAC7 interact with one of the ET receptors, ETA, as determined by yeast two-hybrid analysis, glutathione S-transferase pull-down assays, and co-immunoprecipitation from transfected COS-7 cells. In the absence of ET-1, Tip60 and HDAC7 were localized mainly in the cell nucleus while ETA was predominantly confined to the plasma membrane. Stimulation with ET-1 resulted in the internalization of ETA to the perinuclear compartment and simultaneously in the efflux of Tip60 and HDAC7 from the nucleus to the same perinuclear compartment where each protein co-localized with the receptor. Upon co-transfection with ETA into COS-7 cells, Tip60 strongly increased ET-1-induced ERK1/2 phosphorylation, whereas HDAC7 had no significant effect. We thus suggest that protein acetylase and deacetylase interact with ETA in a ligand-dependent fashion and may participate in ET signal transduction.

Endothelins interact with the specific G protein-coupled receptors A (ETA) and B (ETB). The extracellular N terminus of these receptors is involved in ligand binding, whereas the intracellular C-terminal region is implicated in downstream signaling, receptor internalization, and desensitization (15)(16)(17)(18)(19). ET-1 interacts mainly with ETA, whereas the affinity of ET-2 and ET-3 to this receptor is much lower (20). In contrast, ETB has similar specificity for all three endothelin subtypes (21).
The diversity of endothelin action may be explained not only by the multiplicity of ligands and receptor heterogeneity but also by the ability of the receptor to activate different signaling pathways. Binding of ET-1 to ETA leads to the increase in intracellular calcium levels via activation of phospholipases A 2 , C, and D and Ca 2ϩ channels (22)(23)(24)(25)(26). It has also been documented that ET-1 activates the MAP kinase pathway (27)(28)(29). The sequence of biochemical events in which ETA and other heptahelical receptors activate this pathway is not clear and may depend on the cell type. Several possible mechanisms have been proposed. For example, Src and/or other upstream signaltransducing proteins may interact with these receptors directly or through arrestin (30 -32). Also, ET-1 and several other ligands may turn on ERK1/2 by "transactivation" of the receptortyrosine kinase; in particular, the epidermal growth factor receptor (33)(34)(35). Yet another mechanism involves calcium and tyrosine phosphorylation of PYK2 (36,37). It has also been shown that internalization of G protein-coupled receptors and/or transactivation of receptor-tyrosine kinases is crucial for the activation of the MAP kinase pathway (38,39).
Binding of ET-1 to ETA causes rapid receptor internalization via clathrin-coated pits and/or caveolae (40,41). Internalized ETA then traffics through the endosomal pathway to the pericentriolar recycling compartment that can also be marked by fluorescent-labeled transferrin (42). Eventually, a significant fraction of ETA recycles back to the plasma membrane (42,43).
To identify novel proteins that interact with ETA and may affect its biological functions, we have utilized the yeast twohybrid system using the ETA C-terminal region as bait. Two ETA-interacting proteins were identified: a histone acetyltransferase (HAT) Tip60 (44) and a human homolog of mouse HDAC7 (45) that represents a new member of the histone deacetylase (HDAC) family. We have further shown that HAT and HDAC proteins undergo ET-1-dependent translocation from nucleus to cytoplasm, where they co-localize and interact with ETA. Moreover, co-expression of Tip60 with ETA significantly potentiated phosphorylation of ERK1/2 in response to ET-1 stimulation. We propose that Tip60 and HDAC7 act as novel components of ETA-mediated signal transduction.
Cell Culture-Transformed African monkey kidney cell line COS-7 and rat smooth muscle cell line A10 were maintained in high glucose Dulbecco's modified Eagle's medium (DMEM, Life Technologies, Inc.) with 10% fetal bovine serum (Hyclone) in a 37°C, 5% CO 2 incubator.
Plasmids-The rat ETA cDNA (40) was subcloned into the HindIII site of pRC-CMV (Promega). The bait DNA, ETAC/pAS2-1, was constructed as described below. The ETA C-terminal region, which consists of amino acids 365-420 of rat ETA, was amplified from full-length ETA in pRC-CMV vector by PCR and cloned into the PstI site of pAS2-1 vector (CLONTECH).
The C-terminal region of ETA (amino acids 365-420) was amplified by PCR, conjugated in frame with glutathione S-transferase (GST) in pGEX-4T1 vector (Amersham Pharmacia Biotech), and was named ETACGST.
To add Myc and polyhistidine epitopes to the C-terminal end of ETA, the latter was subcloned into pcDNA3.1Myc-His(ϩ) vector. The stop codon at the end of ETA in pRCCMV vector was deleted, a XbaI site was introduced by the QuikChange site-directed mutagenesis kit (Stratagene), and this product was introduced into the pcDNA3.1Myc-His(ϩ) vector A by both HindIII and XbaI digestion. This construct was named ETAmychis.
Yeast Two-hybrid Screening-The cloned bait DNA, ETAC/pAS2-1, and human brain library DNA, in pACT10 vector, were co-transformed into yeast strain CG1945, and positive clones were selected. Positive clones were further selected for those that specifically interact with the ETA C-terminal region but not with other proteins (p53, lamin, and GAL4 DNA-binding domain alone) and the cDNAs were isolated from these clones and sequenced.
The Tip60 cDNA obtained by two-hybrid screening was subcloned into pFLAG-CMV2 vector (Kodak/Sigma) using the SalI and XbaIdigested vector. HDAC7 cDNA corresponding to the C-terminal fragment of the protein (amino acids 545-824) was cut with EcoRI and BglII and ligated into the pFLAG-CMV2 vector that was cut with EcoRI and XbaI. This vector is referred to as ctHDAC7/pFLAG-CMV2.
The full-length cDNA for Tip60 was obtained from the human brain library by PCR. The PCR product was then digested with KpnI and DrdI. The partial Tip60 cDNA in the pFLAG-CMV2 vector was cut with DrdI and XbaI. The pFLAGCMV2 vector was cut with KpnI and XbaI, and all three fragments were ligated together. This vector was named Tip60/pFLAGCMV2.
The full-length cDNA of HDAC7, DKFZp586J0917 (GenBank TM /EBI accession no. AL117455), was obtained from the human genome center (RZPD) in Heidelberg, Germany. The cDNA was cloned into the pFLAGCMV2 vector and this vector was named HDAC7/pFLAGCMV2.
Transfection-Plasmids (4 g each) were transfected into COS-7 cells using LipofectAMINE Plus reagent (Life Technologies, Inc.) and cells were harvested or stained after 72 h of transfection. Cells were washed twice with cold PBS and harvested in the extraction buffer (25 mM Tris, pH 7.4, 1% Triton X-100, 1 g/ml leupeptin, 1 g/ml aprotinin, 1 g/ml pepstatin, and 1 mM PMSF). For the analysis of ERK1/2 phosphorylation, 5 M Trichostatin A (Takara Inc.), 1 mM Na 3 VO 4 , and 20 mM NaF were added to the extraction buffer. After incubation on ice for 15 min, the extract was centrifuged at 14,000 rpm for 5 min to remove cell debris. The protein concentration in supernatant was determined by Pierce BCA protein assay.
GST Pull-down Assay-ETACGST (which comprises the entire sequence of the ETA C terminus) and GST alone were purified as described (47). Beads were blocked with 10 mg/ml BSA (Sigma) at RT for 30 min. The extracts from the Tip60-or HDAC7-transfected cells were added and incubated with the beads at RT for 1 h. Beads were washed twice (5 min each) with the binding buffer and then twice with the binding buffer containing 0.5 M NaCl. Equal volumes of 2ϫ sample buffer were added to washed beads for 10 min at 95°C, and the supernatant was analyzed by Western blot.
Co-immunoprecipitation Experiments-Transfected cells were harvested and anti-Myc antibody or nonspecific mouse IgG (1 l each) were added to the cell extracts (0.5-1 mg) and were incubated overnight at 4°C with rotation. Protein G-Sepharose beads (30 -50 l; Amersham Pharmacia Biotech), blocked with 2% BSA in PBS for 30 min at 4°C, were added to the mixture and incubated for 2 h at 4°C. The beads were washed once with the binding buffer (see the previous paragraph) at 4°C for 10 min and four times with the binding buffer containing 0.5% sodium deoxycholate. The beads were then rinsed once with 10 mM Tris, pH 7.4 and eluted with equal volumes of 2ϫ sample buffer at 65°C for 10 min. The eluates were analyzed by Western blot.
Immunofluorescence Experiments-Cells on coverslips were washed with cold PBS and fixed with 4% paraformaldehyde solution for 30 min at RT. Cells were solubilized with 0.1% Triton X-100 in PBS for 5 min at RT and then rinsed with PBS three times. Primary antibody in PBS with 5% BSA and 3% donkey serum was added for 30 min at RT. The cells were washed with PBS for 30 -60 min and Cy3-or FITC-conjugated secondary antibody (2-4 g/ml in the same solution as primary antibody) was added to the cells for another 30 min. The cells on coverslips were washed overnight in PBS at 4°C and mounted on slides. The fluorescent images were analyzed with the help of the LSM510 Zeiss confocal microscope.
Subcellular Fractionation of A10 Cells-Rat vascular smooth muscle cells A10 were incubated in the absence or presence of ET-1 (10 Ϫ8 M) for 30 min and separated into nuclei and cytoplasmic fractions as described (48).

RESULTS
To identify proteins interacting with ETA, the human brain cDNA library (ϳ8 ϫ 10 5 clones) was screened with the Cterminal region of ETA (amino acids 365-420) as a bait using the yeast two-hybrid system. The final selection process yielded two strong positive clones. A BLAST search revealed that the first clone was identical to Tat-interactive protein 60 or Tip60 (44), which belongs to the subfamily of histone acetyltransferases called MYST (49). The second clone encoded a novel protein with a region of strong homology to the conserved catalytic domain of the class II histone deacetylases, which includes HDAC4, HDAC5, and HDAC6 (50). According to a recent report (45), this protein represents a human homolog of mouse HDAC7.
To confirm interaction of Tip60 and HDAC7 with ETA, GSTconjugated peptide corresponding to the C terminus of ETA was used in GST pull-down experiments. Fig. 1a demonstrates that Tip60 and ctHDAC7 bind to ETACGST but not to GST alone. These results confirm the direct interaction between the ETA C terminus and Tip60 or HDAC7 in vitro.
The interactions were further verified by co-immunoprecipitation experiments using transfected COS-7 cells. ETA with Myc epitope and poly(His) tag at the C terminus (ETAmychis) was expressed in COS-7 cells alone or co-expressed with either Tip60 or ctHDAC7 (both in the pFLAG-CMV2 vector) and immunoprecipitated with anti-Myc antibody. Western blot with anti-FLAG antibody demonstrates that both Tip60 and ctHDAC7 can interact with ETA in vivo (Fig. 1b). In control experiments, when cells were transfected with Tip60 or ctHDAC7 alone, no specific signal was detected in the immunoprecipitated material.
To study the intracellular localization of Tip60, HDAC7, and FIG. 1. Interaction of Tip60 and HDAC7 with ETA in vitro and in vivo. a, purified GST proteins (2 g each) were incubated with extracts (10 -20 g) of Tip60 or ctHDAC7 expressing COS-7 cells. Proteins were isolated with glutathione beads (25 l of packed beads) and were analyzed by Western blot with anti-FLAG antibody. A representative result from at least three independent experiments is shown. b, co-immunoprecipitation of Tip60 and ctHDAC7 with ETA. ETAmychis was transfected into COS-7 cells alone or together with Tip60 or ctHDAC7 (both in the pFLAG-CMV2 vector). Cells were homogenized, and ETAmychis was immunoprecipitated from cell lysates (500 g) with anti-Myc antibody (1 g) and protein G-Sepharose (25 l of packed beads). Immunopurified proteins were analyzed by Western blot with anti-FLAG antibody. A representative result from at least three independent experiments is shown.

Acetyltransferase and Deacetylase in Signaling 16598
ETA, transfected COS-7 cells were incubated in the presence or in the absence of ET-1. These cells were immunostained with rabbit polyclonal anti-ETA antibody, which specifically recognizes the N terminus of ETA (46) and with monoclonal anti-FLAG antibody for the detection of Tip60 and HDAC7. Confocal images showed that in the absence of ET-1 ETA was expressed mainly on the cell surface (Fig. 2). Tip60 (Fig. 2a) and HDAC7 (Fig. 2b) were detected mainly in the cell nucleus similar to all other known histone acetyltransferases and deacetylases (50,51).
Treatment of cells with ET-1 changed the intracellular localization of all three proteins. As expected, ETA was internalized from the cell surface into the perinuclear region (Fig. 2, a and  b). ET-1 stimulation also caused a dramatic redistribution of both Tip60 and HDAC7 from the nucleus into the same perinuclear region where these proteins co-localized with ETA. Note, that Fig. 2b shows two cells (bottom middle panel). One cell (transfected with both HDAC7 and ETA) demonstrates the redistribution of HDAC7 (green) from the nucleus to the ETAcontaining perinuclear region upon ET-1 administration. Another cell (transfected with HDAC7 only) does not show this effect. Thus, expression of ETA is required for the ET-1-dependent nuclear efflux of HDAC7.
ET-1-induced efflux of Tip60 from the nucleus was confirmed in biochemical experiments with non-transfected smooth muscle cells. A10 cells treated and not treated with ET-1 were separated into the nuclear and the cytosolic fractions, and the presence of endogenous Tip60 in these samples was analyzed by Western blot. It is evident in Fig. 3 that Tip60 is localized in the nucleus of basal cells and is partially relocated to the cytosol upon ET-1 stimulation.
To determine whether or not Tip60 and/or HDAC7 are involved in the downstream signaling of ETA, we analyzed the activation of p44/42 MAP kinase (Erk1/2) in response to ET-1 stimulation. As described previously (52,53) and is also shown in Fig. 4, stimulation with ET-1 causes transient activation of this pathway in COS-7 cells transfected with ETAmychis. How-ever, in cells transfected with ETAmychis together with Tip60, the level of ET-1-induced ERK1/2 phosphorylation is dramatically increased, which suggests that Tip60 may be directly involved in the acute downstream signaling of ETA. Deletion of the C-terminal region of ETA abolishes the effect of Tip60, indicating that direct interaction is required for the elevation of ERK1/2 phosphorylation (data not shown).
The effect of exogenously expressed HDAC7 on the ERK1/2 pathway was inconsistent. In some cases, ERK1/2 phosphorylation was attenuated, whereas other experiments showed no effect (data not shown). However, we were never able to detect an increase in ERK1/2 phosphorylation in ETAmychis-and HDAC7-transfected cells. DISCUSSION Acetylation/deacetylation of histones is crucial for the regulation of transcription of many genes. Known substrates for HATs, however, are not limited to histones and include several transcription factors, importin, and ␣-tubulin (54 -59). Acetylation affects DNA-binding activity, protein stability, and protein-protein interaction (51). Therefore, acetylation and deacetylation may represent general post-translational modifications that take place in both nuclear and cytosolic compartments of the cell (reviewed in Ref. 51). In fact, we have shown that Tip60, a HAT family protein, can potentiate the effect of ET-1 on ERK1/2 phosphorylation, which suggests that the acetylation event may also regulate receptor-mediated signaling pathways.
Our studies on the localization of Tip60 and HDAC7 revealed that under basal conditions the major pools of Tip60 and HDAC7, like other HATs and HDACs, reside in the cell nucleus. ET-1 stimulation led to the redistribution of Tip60 and HDAC7 from the nucleus into the perinuclear region where each protein co-localized with the internalized ETA. Efflux of Tip60 and HDAC7 from the nucleus may have dual effects. First, it may result in changes in the transcription of several ET-1 responsive genes, such as c-myc, c-jun, c-fos, etc. Second, interaction of Tip60 and HDAC7 with ETA or ETA-associated proteins may affect cytoplasmic signaling pathways, such as phosphorylation of ERK.
It was recently shown that other class II histone deacety- Co-localization of Tip60 and HDAC7 with ETA by confocal laser-scanning microscopy. COS-7 cells expressing ETA and Tip60 (a) or ETA and HDAC (b) were incubated in the absence or presence of ET-1 (10 nM) for 10 min at 37°C. Fixed cells were stained with rabbit polyclonal anti-ETA antibody and mouse monoclonal anti-FLAG antibody followed by Cy3-conjugated anti-rabbit IgG for the detection of ETA (red) and FITC-conjugated anti-mouse IgG for the detection of Tip60 and HDAC7 (green). A representative result from at least three independent experiments is shown. Acetyltransferase and Deacetylase in Signaling 16599 lases, HDAC4 and 5, shuttle between the nucleus and the cytosol (60,61). The physiological importance of the nucleocytoplasmic distribution of HDAC4 and 5 was emphasized in a recent report demonstrating that muscle differentiation can be controlled by nuclear export of HDAC4 and 5 in response to calcium/calmodulin-dependent protein kinase signaling (62). On the other hand, constitutive activation of the MAP kinase pathway in Ras-or MEK1-transfected cells results in the increased nuclear localization of HDAC4 (63). ERK1/2 was shown to interact with and phosphorylate HDAC4, although it is not clear to what extent phosphorylation of HDAC4 may regulate its intracellular localization (63). Our findings provide the first specific example of ligandinduced translocation of HAT and HDAC. However, the molecular mechanism(s) triggering the nuclear export of Tip60 and HDAC7 in response to ET-1 is not known. It is possible that interaction with internalized ETA may anchor Tip60 and HDAC7 in the cytosolic compartment and induce their efflux from the nucleus. There is evidence that protein 14-3-3 may play an important role in cytoplasmic retention of HDAC4 and 5 (61,64). With the experimental system established in our study, future work should be able to ascertain the role of 14-3-3 or other candidate proteins in ET-1-induced nucleo-cytoplasmic transport of Tip60 and HDAC7.
In addition to the MAP kinase pathway, Tip60 and HDAC7 may be involved in other regulatory events. For example, we have noticed that transfection of cells with Tip60 increases (or with HDAC7 decreases) the level of ETA (data not shown). This suggests that Tip60 may protect ETA from degradation by re-routing the receptor from lysosomes to the recycling pathway. We are now trying to follow up on these experiments to uncover the potential role of acetylation/deacetylation in the regulation of receptor trafficking and stability.