Positive and negative modulation of the transcriptional activity of the ETS factor ESE-1 through interaction with p300, CREB-binding protein, and Ku 70/86.

Epithelium-specific ETS (ESE)-1 is a prototypic member of a novel subset of the ETS transcription factor family that is predominantly expressed in cells of epithelial origin but can also be induced in other cell types including vascular endothelial and smooth muscle cells in response to inflammatory stimuli. To further define the molecular mechanisms by which the transcriptional activity of ESE-1 is regulated, we have focused our attention on identifying proteins that interact with ESE-1. We have determined that Ku70, Ku86, p300, and CREB-binding protein (CBP) are ESE-1 interacting proteins. The Ku proteins have previously been shown to bind to breaks in DNA where they function to recruit additional proteins that promote DNA repair. Interestingly, Ku70 and Ku 86 negatively regulate the transcriptional activity of ESE-1. Using a series of deletion constructs, we have determined that the Ku proteins bind to the DNA-binding domain of ESE-1. The Ku proteins inhibit the ability of ESE-1 to bind to oligonucleotide probes in gel mobility shift assays. The finding that Ku proteins can interact with other transcription factors and block their function has not been previously demonstrated. In contrast, co-transfection of p300 and CBP with ESE-1 enhances the transcriptional activity of ESE-1. Moreover, the induction of ESE-1 in response to inflammatory cytokine interleukin-1 is associated with a parallel increase of the expression of p300 in vascular endothelial cells, suggesting that in the setting of inflammation, the transcriptional activity of ESE-1 is positively modulated by interaction with the transcriptional co-activator p300. In summary, our results demonstrated that the activity of ESE-1 is positively and negatively modulated by other interacting proteins including Ku70, Ku86, p300, and CBP.

ESE-1, also named ELF3, ESX, jen, and ERT, is the prototypic member of a novel subset of the ETS transcription factor family including ESE-2 and ESE-3 that under basal conditions is expressed exclusively in cells of epithelial origin (1)(2)(3)(4)(5)(6). ESE-1 is expressed in several tissues including kidney, pros-tate, small intestine, colon, ovary, pancreas, liver, and placenta, with particularly high expression in the gastrointestinal tract (1,2). Whereas ESE-2 expression is restricted to epithelial cells in the kidney and prostate, ESE3 is expressed in epithelial cells of salivary gland, prostate, and trachea, with a lower level of expression in colon, mammary gland, pancreas, lung, stomach, and appendix as well as fetal kidney and fetal lung (7,8).
Expression of ESE-1 is up-regulated during the differentiation of epithelial cells. For example, undifferentiated cultured keratinocytes express very little ESE-1. However, in the presence of calcium, the cells undergo a process of differentiation that is associated with the expression of several epithelial markers and the induction of ESE-1 (1). In situ hybridization also demonstrates increased expression of ESE-1 in the granular layers with little or no expression in the basal layers of the skin, suggesting a role for ESE-1 in the terminal differentiation of keratinocytes. The expression of ESE-1 in keratinocytes correlates with the expression of several markers of terminal differentiation in skin, including the small proline-rich proteins SPRR2A, SPRR1, transglutaminase 3, and profillagrin. The regulatory elements of each of these genes contain conserved binding sites for ETS factors that are critical for expression of during terminal differentiation of epithelial cells (9 -12). ESE-1 binds to the ETS sites within the SPRR2A, SPRR1, SPRR3, and transglutaminase genes and positively transactivates the corresponding promoters, further supporting the notion that ESE-1 is a critical regulator of the terminal differentiation of epithelial cells (1,5,11,12). Targeted disruption of the ESE-1 gene in the mouse leads to lethality of about 30% at embryonic day 11.5, whereas the surviving ESE-1 null mutant mice display severe alterations of tissue architecture in the small intestine (13), indicating that ESE-1 is required for the normal development of small intestine epithelial tissue.
We recently observed that ESE-1 expression could be induced in several nonepithelial cells such as vascular endothelial and smooth muscle cells in response to inflammatory stimuli (6,14). Induction of ESE-1 in these cells is dependent on the activation of NF-B (6). One of the main gene targets for ESE-1 in these cells is the inducible form of nitric-oxide synthase (NOS2). 1 ESE-1 acts synergistically with NF-B to induce NOS2 gene expression through the direct physical interaction of the p50 subunit (6). In a mouse model of endotoxemia, the expression of ESE-1 and NOS2 is rapidly induced in the smooth muscle and endothelial cells of the aorta after the administration of endotoxin (6).
ESE-1 has a unique structural feature that distinguishes it from ESE2, ESE3, and other ETS proteins. ESE-1 contains a second DNA-binding domain: an A/T hook domain, which has been described in other nuclear factors and recognizes the minor groove of A/T-rich region of double-stranded DNA (1). Although the target DNA sequence of ESE-1 A/T hook domain has not been identified, the existence of such a domain implies that ESE-1 may have unusual function and physiological significance distinct from the traditional functions of other ETS factors.
Several aspects of how ESE-1 function is regulated in epithelial as well as nonepithelial cells remain unknown. We postulated that one of the major mechanisms by which the activity of ESE-1 could be modified is through protein-protein interactions. To identify ESE-1-interacting proteins, we employed the so-called "pull-down" assay. Using a GST-ESE-1 fusion protein we identify the transcriptional co-activators, p300 and CBP, and DNA damage repairing proteins Ku70 and Ku86, as ESE-1-interacting proteins. Whereas p300⅐CBP display co-activator activity and thereby up-regulate ESE-1 activity, Ku70 and Ku86 inhibit ESE-1 activity by blocking the DNA binding activity of ESE-1. The ability of the Ku proteins to interact with other transcription factors and block their function has not been previously demonstrated. Furthermore, we have also determined that the induction of ESE-1 in vascular endothelial cells in response to inflammatory cytokine IL-1␤ is associated with a parallel increase in the expression of p300, suggesting a role for p300 in synergistically promoting the activity of ESE-1 in the setting of inflammation.
Cell Culture, Transfection, and Reporter Assay-293T, HeLa, and THP-1 cells were routinely maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 units/ml of penicillin, and 100 g/ml of streptomycin (Invitrogen) at 37°C in 5% CO 2 . Human umbilical vein endothelial cell (HUVEC) and aortic smooth muscle cell were obtained from Clonetics and were cultured in endothelial cell growth medium (Clonetics) and smooth muscle cell growth medium (SmGM-2; Clonetics), respectively. Stimulation of the THP-1 and HUVECs with interleukin-1 (R & D Systems) was performed at a concentration of 0.1 ng/ml. 293T cells were transfected at about 70% confluency. DNA was introduced into the cells using the Polyfect transfection reagent (Qiagen) as recommended.
For the luciferase assay, 293T cells were grown in 24-well plates. The transfection mixture consisted of 1-2 g of DNA plus 3 l of Polyfect reagent. 0.05 or 0.01 g/well of CMV-␤-galactosidase (Clontech) was added as an internal control. Every transfection was performed in triplicate, and each experiment was repeated independently at least three times. After the transfection, the cells were cultured for 2 days and then collected in the passive lysis buffer (Promega). The luciferase activity was determined using the luciferase assay system (Promega) and the AutoLumat luminometer (EG&G Berthold) and normalized against the activity of ␤-galactosidase using the chromagenic substrate o-nitrophenyl-beta-glactoside (17).
GST Pull-down Assay-GST and GST fusion proteins were prepared using bacterial strain BL21-CodonPlus-RP cells (Stratagene) transfected with the expression vector pGEX-4T-2 or pGEX-4T-2-ESE-1 as previously described (18). The lysate used for the GST pull-down was The cell lysates were loaded for the GST pull-down assay using GST and GST-ESE-1 separately. The proteins isolated were separated by SDS-PAGE using a 12% polyacrylamide gel. B, GST-ESE-1 pull-down assay using cell lysates from the [ 35 S]Met/Cys-labeled THP-1 and HUVEC cells stimulated with and without the inflammatory cytokine IL-1␤. The proteins isolated were separated by SDS-PAGE using a 12% polyacrylamide gel. MW, molecular mass. prepared from 293T cells. 293T cells cultured in 10% fetal bovine serum Dulbecco's modified Eagle's medium were collected after low speed centrifugation and washed twice with phosphate-buffered saline. 35 S-Labeled cell lysates were prepared using 293T, HeLa, THP-1, HUVEC, and aortic smooth muscle cell cells that were labeled with 2.0 mCi/ 10-cm plate of [ 35 S]Met/Cys mixture solution (Easytag; PerkinElmer Life Sciences) for 6 h after a 4-h starvation in Met/Cys-free Dulbecco's modified Eagle's medium (Invitrogen) plus 10% dialyzed fetal bovine serum (Invitrogen). For large scale GST pull-down, 50 g of the cell pellet was resuspended in 100 ml of Nonidet P-40 lysis buffer (50 mM HEPES, pH 7.4, 1% Nonidet P-40, 150 mM NaCl, 1.5 mM EDTA, 10% glycerol, 0.2 mM PefablocSC (Roche Applied Science), 1 g/ml leupeptin (Sigma), 1 g/ml pepstatin (Sigma), 1 mM dithiothreitol, 50 mM NaF, and 0.5 mM vanadate). The supernatant was collected after centrifugation at 14,000 ϫ g for 30 min twice at 4°C. The supernatant was added to 3 ml of GST-bound agarose beads for 6 h at 4°C to reduce nonspecific binding to GST. After centrifugation at 2,000 ϫ g for 10 min twice to get rid of the beads, 100 ml of the pre-cleared supernatant was split into two equal parts (i.e. 50 ml each). GST-bound agarose beads (about 400 g of GST protein) were added into one part as the negative control, whereas GST-ESE-1-bound agarose beads (about 200 g of protein) were added to the other part. The pull-down mixtures were then incubated at 4°C for 6 h with agitation. Finally, the beads were recovered by low speed spin (1,000 rpm) and washed with 50 ml of ice-cold Nonidet P-40 lysis buffer for five times. The beads were then stored at Ϫ70°C. For SDS-PAGE analysis, the beads were resuspended into equal amount of SDS sample buffer. To analyze the size of the GST bound proteins, the beads were resuspended into equal amount of SDS sample buffer and separated by SDS-PAGE on 8 and 12% polyacrylamide gels. Partial sequencing of proteins obtained from Coomassie Blue G-250stained gels was performed by tandem mass spectrometry using the service at the Taplin Biological Mass Spectrometry Facility at Harvard Medical School.
Co-immunoprecipitation-FLAG-and c-Myc-tagged target proteins were expressed in 293T cells transiently by co-transfecting the cells with the expression vectors using Polyfect reagent. Immunoprecipitation was then performed using anti-FLAG monoclonal antibody M2 (Sigma) plus protein A/G-agarose or the M2 antibody conjugated on agarose (Sigma). The mixture was incubated for 2 h at 4°C with agitation. Agarose beads were then recovered using low speed spin and washed with Nonidet P-40 lysis buffer. Finally, the beads were resuspended in SDS sample buffer, and the bound proteins were analyzed by Western blot analysis. The FLAG-tagged proteins were recognized using antibody M2, and the Myc-tagged proteins were recognized using the anti-c-Myc monoclonal antibody (9E10) (Santa Cruz Biotechnology). Ku86 was also analyzed using mouse anti-Ku86 antibody (BD Biosciences/Pharmingen). p300 and HAtagged CBP were analyzed using rabbit polyclonal anti-p300 antibody N15 and mouse anti-HA antibody or rabbit polyclonal anti-CBP C-20 (Santa Cruz Biotechnology), respectively.
In Vitro Transcription/Translation-ESE-1 and Myc-tagged Ku70/86 were also produced by the coupled in vitro transcription/translation reactions using the TNT reticulocyte lysate kit (Promega) as recommended.
Electrophoretic Mobility Shift Assay-DNA binding activity of ESE-1 was determined using the electrophoretic mobility shift assay as described previously (1,8,19). In brief, the 20-l binding reaction was made up of 2 l of ESE-1 prepared using an in vitro translation method with 32 P-labeled double-stranded probe (30,000 cpm) and 50 ng of cold mutant oligonucleotides to reduce the background. The samples were incubated at room temperature for 15 min and run on a 4% polyacrylamide gel (acrylamide-bisacrylamide, 29:1) containing 0.5ϫ TBE (45 mM Tris borate and 1 mM EDTA) buffer. The 32 P-labeled probe was an double-stranded oligonucleotide carrying encoding ESE-1-binding se-

FIG. 2. The ESE-1-interacting proteins identified by the pulldown assay using GST-ESE-1.
293T cells were co-transfected with pCDNA-FLAG-ESE-1 (0.4 g/well on a 24-well plate) and with NOS2or PSP-luciferase reporter (0.4 g/well on a 24-well plate). The empty expression vector pCDNA3.1-5FLAG was used as the negative control. After 48 h, the cell lysated were collected and assayed for luciferase activity (A). The activity was normalized for transfection efficiency using an internal control (CMV-␤-galactosidase). The error bars represent S.D. (n ϭ 3). 293T cell lysate was prepared from 50 g of 293T cell pellet. The cell lysate was incubated with GST and GST-ESE-1 beads and then washed. The bound proteins were analyzed by SDS-PAGE (B). GST-ESE-1 protein used for the pull-down assay was loaded as the control to indicate the GST-ESE-1 protein and its broken down products (lines g). GST pull-down products were considered as the "nonspecific" products (lines N). The Coomassie Blue-stained "specific" bands pulldown by GST-ESE-1 were denoted from top to bottom (lines A-E). The protein identification revealed that band A contains p300, and bands B and C represented Ku86 and Ku70, respectively. MW, molecular mass.

FIG. 3. Co-immunoprecipitation of ESE-1 and p300 and CBP.
Co-immunoprecipitation was performed using 293T cells expressing p300 or HA-CBP plus pCDNA3.1-5FLAG empty or FLAG-ESE-1. The expressions of FLAG-ESE-1, p300, and HA-CBP were confirmed in the total lysate (i.e. the input used for the immunoprecipitation) by Western blot using anti-FLAG antibody M2, anti-p300 antibody N-15, and anti-HA antibody, respectively (A). FLAG-ESE-1 was immunoprecipitated using anti-FLAG antibody M2-conjugated agarose beads. Co-immunoprecipitated products were separated by SDS-PAGE on a 6% polyacrylamide gel for the detection of p300 and HA-CBP using N-15 and anti-HA antibodies and on a 10% polyacrylamide gel for detection of FLAG-ESE-1 using antibody M2 (B). quence from the murine PSP promoter (5Ј-TCGACGAACATCCAG-GAAATAGGGCTC-3Ј) as reported previously (6).
Image Quantitation-The signal intensity of specific bands in the electrophoretic mobility shift assay experiments was determined using a PhosphorImager and the supporting software (Molecular Dynamic). The relative sizes of the bands on the Western blots were determined using Image-Pro Plus (Media Cybernetics, Inc.).

ESE-1 Interacts with Specific Proteins
-As an initial screen to identify whether there are specific proteins that interact with ESE-1, we used a GST pull-down assay with cell extracts derived from [ 35 S]Met/Cys-labeled cells. The proteins were separated by SDS-PAGE on a 12% polyacrylamide gel. Several proteins of varying molecular weight bind to GST-ESE-1 and not to GST alone, suggesting that they bind specifically to ESE-1 (Fig. 1A). The binding pattern of the proteins that bind to ESE-1 appears to be similar for epithelial (HeLa and 293T) as well as nonepithelial (THP-1, HUVEC, and aortic smooth muscle) cells.
We next examined whether any of the ESE-1-interacting proteins are differentially regulated in response to IL-1␤ in two cell types THP-1 and HUVECs in which ESE-1 has been previously shown to be induced in response to IL-1␤ (6). This induction occurs ϳ4 -6 h after stimulation and gradually diminishes over 24 h. There did not appear to be any change in the pattern of the molecular weight of these proteins after stimulation of these cells (Fig. 1B). These results indicate that the predominant ESE-1-binding proteins are constitutively ex-pressed in epithelial as well as nonepithelial cells.
Despite the absence of ESE-1 under basal conditions, the 293T cells appeared to contain several of the same ESE-1 binding proteins as other cells (Fig. 1A, tenth lane). To determine whether these cells could be useful for making a large scale preparation of cellular extracts, to isolate enough ESE-1 interacting proteins for micro-sequencing, we assessed the capacity of ESE-1 to function normally as a transcription factor in these cells. We tested the ability of ESE-1 to transactivate the promoters of two previously well characterized targets of ESE-1: the epithelial-specific PSP gene and the NOS2 gene (1,6). The co-transfection of the reporter with FLAG-tagged ESE-1 or its empty vector into 293T cells showed that ESE-1 strongly activated NOS2-and PSP-luciferase reporters by 3.9and 58-fold, respectively ( Fig. 2A).
Based on these data, we then performed a large scale pulldown assay using 293T cells (see "Materials and Methods" for details). The isolated proteins were separated by SDS-PAGE on a 12% polyacrylamide gel (Fig. 2B). To confirm the specificity of the GST-ESE-1-bound proteins, we included two controls: the GST pull-down product as the nonspecific control to identify the proteins bound on the beads and GST protein (lines N) and the GST-ESE-1 protein alone as the control to identify GST-ESE-1 protein and its subfragments, representing breakdown products (lines g). We observed five distinct additional bands (lines A-E) (Fig. 2B) in 35 S-labeled samples. By tandem mass spectrometry based peptide identification, we identified that To determine whether endogenous p300⅐CBP influences ESE-1 activity, E1A was expressed in 293T cells to sequester p300⅐CBP in the reporter assays (B). To further study the potential effects of interacting proteins on the ESE-1 activity, we determined the expressions of ESE-1 and its interacting proteins, p300, CBP, Ku70, and Ku86, in HUVEC challenged with inflammatory cytokine IL-1␤. The IL-1␤ induced ESE-1 expression was confirmed by reverse transcriptase (RT) PCR, whereas glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the control (C). The endogenous p300 and CBP protein levels were determined by Western blot analysis using anti-p300 (N15) and anti-CBP (C-20) antibodies. ␤-Actin was probed by anti-␤-actin antibody as the control to ensure the equal loading. The observed increase in p300 protein level was also analyzed using ImagePro-Plus analysis and normalized by ␤-actin signal intensities of the corresponding lanes (D). bands A, B, C, D, and E are p300, Ku86, Ku70, 60 S ribosomal protein l6, and NADPH-dependent carbonyl reductase 3, respectively. There is some evidence from previous studies that co-activator p300 and DNA repairing components Ku70 and K86 can interact with other transcription factors and could therefore be involved in regulating ESE-1 activity (20,21). However, there is no evidence that either 60 S ribosomal protein l6 or NADPH-dependent carbonyl reductase 3 interact with transcription factors. It is possible that 60 S ribosomal protein l6 and NADPH-dependent carbonyl reductase 3 are indirectly associated with Ku70 or Ku86 rather than with ESE-1.
The Interactions between ESE-1 and p300⅐CBP-The finding that p300 is a potential ESE-1 target was not a surprise because p300 and its related protein CBP had previously been reported to function as co-activators with several other transcription factors (20). Within the ETS family, it is known that p300⅐CBP forms a complex with Ets-1 (22), Ets-2 (23), ER81 (24), Spi-B (25), and GABP (26). To confirm our findings, 293T cells were transfected with mammalian expression plasmids encoding p300, HA-tagged CBP, and either FLAG-tagged ESE-1 or a control vector (Fig. 3A). Immunoprecipitation was then performed using the anti-FLAG antibody (M2). Western blot analysis of the immunoprecipitate with an anti-p300 polyclonal antibody demonstrated that ESE-1 interacts with p300 (Fig. 3B). Similarly when the lysates were probed with anti-HA antibody, ESE-1 appears to interact with HA-tagged CBP (Fig.  3B). Although the interactions were somewhat weak, this result confirms that ESE-1 can interact with p300 and its related protein CBP.
To explore the potential role that p300⅐CBP plays in modulating the transcriptional activity of ESE-1, we analyzed changes in ESE-1 transactivation of the PSP promoter in the presence or absence of p300⅐CBP. Our results demonstrate that the transcriptional activity of ESE-1 is increased in the presence of p300 or CBP in a dose-dependent manner (Fig. 4A). In contrast, p300 and CBP had little effect on the reporter activity in the absence of ESE-1 (Fig. 4A). Because it has previously been well defined that E1A blocks p300⅐CBP activity, we determined whether the addition of E1A would affect the activity of ESE-1. Transactivation of the PSP-luciferase reporter by ESE-1 was reduced in the presence of increasing amounts of E1A (Fig. 4B), suggesting that the endogenous p300⅐CBP may contribute to ESE-1 function in these cells. Together, our results suggested that p300 and CBP positively modulate ESE-1 function by acting as transcription co-activators.

Induction of ESE-1 in Response to Inflammatory Cytokine Stimulation Is Associated with a Parallel Increase in p300
Expression-ESE-1 is transiently induced in nonepithelial cells in response to inflammatory stimuli. To further define the role of p300 as an ESE-1-interacting protein in nonepithelial cells, we examined the expression in stimulated vascular endothelial cells. HUVECs were stimulated with the inflammatory cytokine IL-1␤; total protein samples were collected at 0, 6, 12, and 24 h; and RNA samples were collected at 0, 1, 2, 4, 6, 12, 24, and 36 h after stimulation as described previously (6). Consistent with previous reports, ESE-1 was induced in HUVEC and peaked at 4 -6 h (Fig. 4C). In these cells, the expression of p300 was induced at 6 h and persisted out to 24 h (Fig. 4D). The expression of CBP does not appear to change in response to IL-1␤. In addition to increasing p300 expression levels, IL-1␤ can also potentiate the activity of p300 through phosphorylation. One of the major IL-1␤ signaling pathways involves p38 mitogen-activated protein kinase and c-Jun N-terminal kinase (27)(28)(29). p300 activity is strongly augmented after phosphorylation by p38 (30 -32). These results suggest that IL-1␤ not only induces the expression of ESE-1 but also enhances ESE-1 activity by up-regulating the co-activators such as p300.
The Interactions between Ku70/86 and ESE-1-Next, we analyzed the interaction of ESE-1 with Ku70 and Ku86, which are two highly related and associated subunits of a heterodimer known as the Ku antigen. Because the amounts of Ku70 and Ku86 in the pull-down product were relatively high, we first repeated the pull-down assay using the cells transiently expressing Myc-tagged Ku70 and Ku86. GST-ESE-1 but not GST pulled down Myc-tagged Ku70, Ku86, or both Ku70 and Ku86 from the cells expressing Myc-tagged Ku70, Myc-tagged Ku86 or both Myc-tagged Ku70 and Ku86, respectively (Fig. 5A). These data confirm that GST-ESE-1 is able to pull-down Ku70 and Ku86 proteins. However, we could not rule out the possibility that ESE-1 actually pulled down the Ku70 -5Ku86 heterodimer by directly interacting with only one subunit, Ku70 or Ku86, because Myc-tagged Ku70/86 could form a heterodimer with endogenous Ku70/86. To determine whether this is the case, we probed the same Western blot filter using anti-human Ku86 antibody and found that endogenous Ku86 was indeed present in the pull-down products at comparable levels of Myctagged Ku86 (Fig. 5B). To identify which of the Ku subunits directly interacts with ESE-1, we next performed the pull-down assay using Myc-tagged Ku70 and K86 produced using in vitro These lysate were applied for the pull-down assay using GST and GST-ESE-1. The Ku proteins in the pull-down products were also determined by Western blot using anti-Myc antibody 9E10 (A). The endogenous Ku86 and Myc-tagged Ku86 in the lysate and pull-down products in A were determined using anti-human Ku86 antibody (B). In vitro translated Ku70 and Ku86 were verified by Western blot using anti-Myc antibody 9E10. The pCDNA empty vector was used as the negative control. In vitro translated Ku70 and Ku86 were incubated with GST and GST-ESE-1 beads, and the pull-down products were determined by Western blot using antibody 9E10. transcription/translation method because no significant amount of endogenous Ku86 was detected in the rabbit TNT system (Promega). These pull-down assays suggest that ESE-1 can interact directly with either Ku70 or Ku86 (Fig. 5C).
To further define the interaction between ESE-1 and the Ku proteins, we performed co-immunoprecipitation assays using the cells co-expressing Myc-tagged Ku70/86 and FLAG-tagged ESE-1 or its fragments. When Ku70 and/or Ku86 were coexpressed with the full-length ESE-1, complexes of Ku70/86 with ESE-1 were formed (Fig. 6A). Again, using anti-human Ku86 antibody, we observed the presence of the endogenous Ku86 in the co-immunoprecipitated products. When different ESE-1 deletion mutants were co-expressed with Ku70/86, we found that deletion of the ETS domain (ESE-1-BX and ESE-1-X as illustrated in Fig. 6B) abolished the interaction, but the deletion of the transactivation domain (ESE-1 dN1) had no effect on the interaction with Ku70/86 (Fig. 6, B and C), further supporting the notion that the ETS domain of ESE-1 mediates the protein-protein interactions with the Ku proteins.
To further define the specificity of the interaction of the Ku proteins with ESE-1, we generated GST fusion proteins with other ETS factor family members, including Ets-1, Ets-2, and the ESE-1-related ETS factor ESE-3. Furthermore, to define whether the ETS domain alone, is sufficient in mediating binding to the Ku proteins, we also generated a GST fusion protein with the ESE-1 ETS domain (GST-ESE1(ets)). As is shown in Fig. 7, the ETS domain of ESE-1 is sufficient to mediate binding of the Ku proteins. Interestingly, neither Ets-1 nor Ets-2 appear to interact with the Ku proteins. However, the ESE-1related ETS protein ESE-3 did interact with the Ku proteins. To confirm this interaction co-immunoprecipitation experiments were performed using FLAG-tagged ESE3 constructs and Myc-tagged Ku proteins. As is shown in Fig. 7B, the Ku proteins interact with the full-length ESE3 protein and its DNA-binding domain (FLAG-ESE3 (ets)) but not with a truncated form of ESE3 containing only the N terminus (ESE3N). These data suggest that the Ku proteins modulate the function of a subset of the ETS factor family and that the interaction occurs via the DNA-binding domain.
Ku70 and Ku86 Down-regulated the ESE Activity by Blocking the DNA Binding Activity of the ETS Domain-To determine whether the interaction with Ku70/86 can modulate the transcriptional activity of ESE-1, we performed co-transfection experiments with ESE-1 and the PSP-luciferase reporter in the presence or absence of Ku70/86 (Fig. 8). The addition of Ku70 or Ku86 individually had a modest effect on down-regulation of ESE-1 transactivation (Fig. 8, A and B). Thus, the addition of Ku70 together with Ku86 down-regulated more potently the ESE-1-dependent PSP-luciferase activity in a dose-dependent manner (Fig. 8C). The down-regulation of the transcriptional activity of ESE-1 by Ku70 and Ku86 through the interactions with the ETS could occur by interfering with the DNA binding activity of ESE-1. However, other interacting proteins such as p300⅐CBP also target the ETS domain and do not interfere with DNA binding (33). To examine whether the Ku proteins inhibit DNA binding of ESE-1, we performed gel mobility shift assays using an oligonucleotide probe encoding the ESE-1-binding sequence from the PSP promoter in the presence and absence of Ku70 and Ku86 (Fig. 8D) (6, 19). The addition of both Ku proteins appeared to reduce ESE-1 binding to DNA. Whereas the addition of Ku70 (lanes 3,4) had a relatively modest effect on blocking ESE-1 binding, the addition of Ku86 appeared to block ESE-1 binding in a dose-dependent manner, similar to what was observed with the transactivation studies. These results suggest that the principal mechanism by which Ku proteins down-regulate ESE-1 transcriptional activity is by blocking DNA binding.
To confirm that the down-regulation of ESE-1 transcriptional activity is solely mediated via disruption of the DNA binding of ESE-1 and not by additional effects upon transactivation, we made a chimeric constructs encoding the GAL4 DB domain and ESE-1 without its DB domain and evaluated the ability of the Ku proteins to block transactivation of 5ϫ GAL4 site-luciferase reporter by this fusion protein. The GAL4 DB-ESE-1was able to potently transactivate the 5ϫ GAL4 siteluciferase reporter compared with ESE-1 and the GAL4 DNAbinding domain, which were essentially inactive (Fig. 9A). We next examined the ability of the Ku70/86 proteins to block GAL4 DB-ESE-1-mediated transactivation of the 5XGAL4 siteluciferase reporter. The addition of the Ku proteins had little effect on GAL4DB-ESE-1-mediated transactivation of the GAL4-site reporter (Fig. 9B). These results suggest that Ku70 and Ku86 interact with ESE-1 and lead to the decrease in the ESE-1 activity solely by interfering with ESE-1 DNA binding activity.

DISCUSSION
The overall goal of this study was to identify additional proteins that interact with the ETS transcription factor ESE-1, because protein-protein interactions are one of the principal mechanisms by which the activity of transcription factors can be modulated. We have identified proteins that interact with ESE-1 to positively and negatively affect its function. Our results regarding the ability of p300⅐CBP to promote ESE-1 activity are consistent with the previously defined role of p300⅐CBP as a co-activator. p300 and its related protein CBP are known to coordinate multiple signal-dependent biological processes such as cell growth, transformation, differentiation, apoptosis, and development with the transcription apparatus of multiple genes (20). Several ETS factors have previously been shown to interact with p300⅐CBP. Transactivation of the c-Myb-dependent CD13 aminopeptidase N by Ets-1 is potentiated by p300⅐CBP. Ets-1 binds to two cysteine-histidine rich regions of CBP (22). In addition to acting as a co-activator of Ets-1, the p300⅐CBP⅐Ets-1 complex exhibits functional histone acetyl transferase activity that promotes chromatin remodeling and modification of transcription factors and adaptor proteins (22). The ability of p300⅐CBP to function as a co-activator with selected Ets factors depends on the specific target gene. For example, the addition of p300⅐CBP promotes the transactivation of the human stromelysin promoter by Ets-2 but not by other ETS factors such as PEA-3 and Erg-2 that are also involved in the regulation of this gene (23). The ETS⅐p300⅐CBP complex may bind to additional proteins to form a complex known as the enhanceosome. Interactions of p300⅐CBP have also been reported for the ETS factors ER81 (24), Spi-B (25), and GABP (26). p300⅐CBP binds to the ETS domain of each of these transcription factors, thereby promoting the transactivation of the corresponding target genes.
In response to LPS induction, the TNF-␣ gene is activated by a specific enhanceosome in T-lymphocytes and monocytic cells Other ETS factors (ETS1, ETS2, and ESE3) were also loaded in the pull-down assay. B, the interaction between Ku proteins and ESE3 was also analyzed by the co-immunoprecipitation. Briefly, the Ku proteins were co-expressed with ESE3 or its truncated forms (ESE3N, containing the N-terminal part of ESE3 protein; and ESE3 (ets), containing the ETS domain of ESE3). The anti-FLAG antibody immunoprecipitated products were analyzed by anti-FLAG antibody to confirm the FLAGtagged ESE3 proteins and by anti-Myc antibody to determine the Myctagged Ku proteins.
consisting of p300⅐CBP, NFAT, Ap-1 members, Sp1, and ETS-1 (34). Previous investigations have demonstrated the co-activator role of p300⅐CBP for NF-B in the regulation of gene expression (35,36) and its essential role for NF-B in the setting of inflammation (37,38). The results of our study demonstrate that p300 expression is up-regulated in endothelial cells in response to IL-1 and promotes ESE-1 transactivation of its target genes such as NOS2 that are similarly known to be activated by NF-B, further supporting an important role for p300⅐CBP as a mediator of inflammation.
The observation that ESE-1 can interact with the Ku complex interaction came as somewhat of a surprise because Ku70 and Ku86 are abundant nuclear proteins with essential functions in the repair of DNA double-stranded breaks, V(D)J re-combination, and telomere maintenance (21,39). Ku70 and Ku86 are two tightly associated subunits of the so-called Ku autoantigen or complex that displays prodigious DNA binding and recognizes DNA ends, discontinuity in double-stranded DNA, structural transitions in DNA, and specific DNA sequences (39). Once bound to DNA, the Ku complex recruits the 470-kDa DNA-dependent protein kinase catalytic subunit, forms an active protein kinase complex, and initiates cooperation with other factors to carry out repair of damaged DNA.
More recently it has been observed that the Ku complex may influence gene transcription via several different mechanisms. First, Ku proteins can regulate transcription by binding to putative Ku-specific binding sites. Ku antigenbinding sites have been identified in several cellular genes. For example, the erythroid cell-specific glycoprotein glycophorin B, which is exclusively activated by GATA1 in erythroid cells, is repressed in nonerythroid cells by the replacement of GATA1 with Ku70-Ku80 heterodimer on the GATA site (40,41). Several other cellular genes including c-Myc (42), transferrin receptor (43), collagen III (44), and the U1 snRNA have similar Ku-binding sites (45). Ku antigen-binding sites also exist in several viral genes, especially the long terminal repeat sequences of HTLV-1, mouse mammary tumor virus, and HIV-1 (46 -49). In these long terminal repeats, a DNA motif was identified as the negative regulation element 1, which is directly targeted by Ku proteins and leads to the repression on the viral transcription (48). Thus, it would appear that binding of the Ku antigen to binding sites provides a potent mechanism of silencing gene transcription.
A second mechanism by which Ku proteins can alter gene transcription is through protein-protein interactions with other transcription factors. The Ku complex can interact with the glucocorticoid receptor enhancing factors 1 and 2 (43), purine box-binding proteins NF45 and NF90 (50), progesterone receptor (51), Tbdn100 (52), HoxC4, and Oct-1 (53). Such interactions result in enhanced gene transcription.
A third mechanism by which the Ku antigen contributes to gene transcription is through a direct interaction of Ku86 with subunits of the RNA polymerase II holoenzyme (54). Functional disruption of the Ku-polymerase II interaction by a dominant negative form of Ku86 suggests that Ku is required for gene transcription. The transcriptional activity of several different promoters (i.e. AdMLP, Hsp70, or HTLV-I) is significantly reduced in cells lacking Ku as well as DNA-PK (55,56). It has recently been shown that different domains of Ku proteins mediate the effects on DNA repair and transcription (15).
The ability of the Ku antigen to block binding of ESE-1 to DNA would appear to be different from the mechanisms described above. The Ku antigen does not appear to bind to the ESE-1-binding site on the DNA because no additional bands were observed in the gel mobility shift assay. Instead, the Ku antigen appears to block ESE-1 binding to DNA by attaching to the ESE-1 Ets domain.
In summary, our results confirm the importance of proteinprotein interactions in modulating ESE-1 function. In particular we have identified the transcription co-activators p300⅐CBP and Ku antigen as ESE-1-interacting factors that mediate opposing effects upon the transcriptional activity of ESE-1. Interaction with p300⅐CBP enhances ESE-1-mediated transcription. The results of our study demonstrate that p300 expression is up-regulated in endothelial cells in response to IL-1 and promotes ESE-1 transactivation of its target genes such as NOS2 that are similarly known to be activated by NF-B, further supporting an important role for p300⅐CBP as a mediator of inflammation. Down-regulation of ESE-1 activity by the Ku antigen is mediated by a direct physical interaction of the Ku antigen with the DNA binding activity of ESE-1, thereby inhibiting the ability of ESE-1 to bind to DNA. The demonstration of this interaction provides a novel mechanism by which Ku antigens modulate the activity of transcription factors. Future studies will be directed at determining whether other ETS family members are similarly targeted by Ku antigen and identification of which of the ESE-1-binding proteins interact with ESE-1 in vitro as well as in vivo depending on the cell type and whether cells are in a quiescent or inflammatory environment. The ultimate goal of these studies is to further define the molecular mechanisms by which ESE-1 contributes to the regulation of its target genes in epithelial as well as nonepithelial cells.