Sp1 and Sp3 recruit histone deacetylase to repress transcription of human telomerase reverse transcriptase (hTERT) promoter in normal human somatic cells.

Activation of telomerase is crucial for cells to gain immortality. In human cells, telomerase activity is tightly regulated by the expression of its catalytic subunit, human telomerase reverse transcriptase (hTERT). In most normal human somatic cells, hTERT is not expressed, and its suppression acts as an important gatekeeper against tumorigenesis. Here we describe the systematic analyses of hTERT promoter to understand the transcriptional repression mechanism of the hTERT gene in normal human somatic cells. Through the serial deletion analysis of hTERT promoter in normal human fibroblasts, we identified a critical repressive element on the hTERT promoter. The repressive element formed DNA-protein complexes with Sp1 and Sp3 in nuclear extracts. Using formaldehyde cross-linked chromatin immunoprecipitation analysis, we found that Sp1 and Sp3 were associated with the endogenously repressed hTERT promoter in human fibroblasts. Furthermore, Sp1 and Sp3 interacted with histone deacetylase (HDAC) in these cells. Overexpression of dominant-negative mutants of Sp1 and Sp3, which contained mainly the HDAC2-binding domain, relieved the HDAC-mediated repression of the hTERT promoter. Taken together, these results suggest that Sp1 and Sp3 associate with the hTERT promoter, recruiting HDAC for the localized deacetylation of nucleosomal histones and transcriptional silencing of the hTERT gene in normal human somatic cells.

All tumor cells have functional telomere stabilization mechanisms, and most of them employ the enzyme telomerase to achieve this requirement (1)(2)(3). By catalyzing the addition of telomeric TTAGGG repeats onto the chromosome ends, telomerase stabilizes the telomere, and thus maintains the stability and integrity of linear chromosomes. In contrast, most normal human somatic cells lack the mechanisms to maintain telo-meric DNA, and the progressive loss of telomeres during DNA replication limits the life span of these cells both in vivo and in vitro (4 -6). Furthermore, abrogation of the telomerase activity in tumor cells has been reported to induce cell growth arrest and apoptosis (7).
Human telomerase is a complex composed of an RNA component that provides the template for the addition of new telomeric repeats, a catalytic subunit known as human telomerase reverse transcriptase (hTERT), 1 and some additional associated proteins (8 -10). The RNA component of human telomerase is expressed constitutively in most human tissues, whereas hTERT is not expressed in most normal human somatic tissues. Thus, in normal human somatic cells, telomerase activity is tightly regulated by the repression of the hTERT gene. In tumor cells, telomerase activity is closely correlated with the expression and derepression of hTERT. These findings indicate that derepression of the hTERT promoter might be an important mechanism leading to activation of the hTERT gene and telomerase enzyme in the cells during tumorigenesis.
Although hTERT gene expression has been well established as essential for the progression and maintenance of the human tumors, its activation/derepression mechanism in the course of tumorigenesis is not well understood. In the course of tumorigenesis many mutations occur, and some of these mutations are thought to contribute to the activation and derepression of hTERT. Several studies including ours have shown that one of the oncogenes deregulated in various human tumors, c-myc, can contribute to the transcriptional activation of the hTERT gene in tumor cells (11)(12)(13). In addition, amplification of the hTERT locus observed in some human tumors probably contributes to the misregulation of hTERT gene expression (14).
Genetic complementation approaches have suggested that the hTERT repression mechanism is dominant over the activation/derepression mechanism (15)(16)(17). Furthermore, hTERT expression in normal human somatic cells is tightly repressed even below the limit of detection by PCR. Thus, it is critical to understand the transcriptional repression mechanism of the hTERT gene in normal human somatic cells. Some of the transcriptional repressors of the hTERT gene have been identified, including WT1 and Mad from our studies (18 -21). However, these repressors seem to be either tissue-specific or relatively minor contributors. The understanding of the general mechanism of transcriptional repression of the hTERT gene in normal cells is still limited. Experiments in a wide variety of normal human somatic cells with trichostatin A (TSA), an inhibitor of histone deacetylase (HDAC), suggest that histone deacetylation is essential to the transcriptional repression mechanism of the hTERT gene (22)(23)(24). In addition, TSA treatment has been shown to significantly up-regulate hTERT promoter activity and the corresponding telomerase activity to a level comparable with that in human tumor cells. Thus, HDAC-mediated repression could be the major, universal transcriptional repression mechanism of the hTERT gene in normal human somatic cells.
In this study, we carried out systematic analyses of hTERT promoter to identify such repressive DNA elements and interacting regulatory factors. Through a serial deletion analysis of the hTERT promoter in cultured human fibroblasts, we identified a TSA-responsive repressive element. This element formed complexes with Sp1 and Sp3 proteins in nuclear extracts derived from normal human somatic cells. Moreover, the endogenous hTERT promoter in the cells was found to be tightly associated with Sp1 and Sp3. Sp1 and Sp3 were found to have the intrinsic ability to interact with HDAC, and through this interaction HDAC is recruited onto the hTERT promoter. Thus, we show that Sp1 and Sp3 are involved in the HDACmediated transcriptional repression of the hTERT gene in normal human somatic cells.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfection, and Luciferase Assay-IMR90 (normal human lung fibroblast), WI38 (normal human lung fibroblast), and HFF (normal human foreskin fibroblast) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 120 g/ml penicillin, and 200 g/ml streptomycin. Transfection was carried out with LipofectAMINE reagent (Invitrogen). Luciferase assays were performed 30 -48 h after transfection using the reagents from Promega according to the manufacturer's instructions. TSA (Sigma) was dissolved in Me 2 SO and added to the culture medium at a final concentration of 50 -200 nM. A corresponding volume of Me 2 SO was added to control cells not treated with TSA.
RNA Extraction and Quantitative Reverse Transcriptase-PCR-Total RNA was isolated with the Trizol reagent (Invitrogen) in accordance with the manufacturer's instructions. The cDNA synthesis, reverse transcription, and quantitative PCR were performed essentially as previously described (25). The amplified DNA was separated on 1% agarose gel and stained with ethidium bromide.
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assay (EMSA)-Nuclear extracts were prepared essentially as previously described (26) with minor modifications. Briefly, cultured cells were collected, washed with phosphate-buffered saline, and pelleted by centrifugation at 1500 ϫ g for 3 min. The pellet was resuspended in 5 packed pellet volumes of ice-cold buffer containing 10 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl 2 , 0.5 mM dithiothreitol, 1 mM sodium orthovanadate, and Boehringer Complete protease inhibitor mixture (Roche Molecular Biochemicals). The cells were allowed to swell on ice for 15 min, after which 1 ⁄3 packed pellet volume of a 10% Nonidet P-40 was added. The tube was vigorously vortexed for 10 s, and the homogenate was centrifuged at 20,000 ϫ g for 30 s. After removal of the supernatant, the nuclear pellet was resuspended in 2 packed pellet volumes of ice-cold buffer containing 20 mM HEPES (pH 7.9), 0.4 M NaCl, 10 mM KCl, 1.5 mM MgCl 2 , 25% glycerol, 0.5 mM dithiothreitol, 1 mM sodium orthovanadate, and Boehringer Complete protease inhibitor mixture. The tube was vigorously rocked at 4°C for 30 min on a shaking platform. The nuclear extract was centrifuged at 20,000 ϫ g for 5 min at 4°C and the supernatant was frozen in aliquots at Ϫ70°C.
Portions of nuclear extract (5 g) were incubated with 0.5 g of poly(dI⅐dC) in the absence or presence of specific antibodies at room temperature for 20 min in a 15-l reaction volume containing 10 mM Tris-HCl (pH 7.5), 4 mM HEPES (pH 7.9), 50 mM NaCl, 20 mM KCl, 1 mM MgCl 2 , 0.54 mM EDTA, 0.6 mM dithiothreitol, and 8% glycerol. Antibodies used for the supershift assay were anti-Sp1 (sc-59, Santa Cruz Biotechnology) and anti-Sp3 (07-107, Upstate Biotechnology). Following incubation, the 32 P-end-labeled double-stranded oligonucleotide probe (upper strand sequence: 5Ј-CCTTCCAGCTCCGCCTCCTC-CGCGCGGA-3Ј) was added and incubated at room temperature for an additional 20 min. These reaction mixtures were subjected to electrophoresis in a 6% polyacrylamide gel in 0.25ϫ TBE buffer, dried, and subjected to autoradiography.
Formaldehyde Cross-linked Chromatin Immunoprecipitation (X-ChIP) Assay-X-ChIP assays were performed using the ChIP assay kit (Upstate Biotechnology) according to the manufacturer's instructions. Cells were fixed in normal culture medium with formaldehyde at a final concentration of 1% for 10 min at 37°C. Sonication was performed to achieve an average DNA length of 500 bp. The following antibodies were used for the immunoprecipitation of the cross-linked chromatin: anti-Mad (65396E, Pharmingen), anti-Sp1 (sc-59, Santa Cruz Biotechnology), anti-Sp3 (07-107, Upstate Biotechnology), and anti-HA (71-5500, Zymed Laboratories Inc.). A ϳ160-bp fragment in the hTERT proximal promoter was amplified using the primers 5Ј-TGCCCCT-TCACCTTCCAG-3Ј and 5Ј-CAGCGCTGCCTGAAACTC-3Ј. PCR was carried out as follows: 1 cycle at 94°C for 3 min; 34 cycles at 94°C for 30 s, 53°C for 30 s, 72°C for 20 s; and 1 cycle at 72°C for 1 min. The amplified DNA was separated on 2% agarose gel and visualized with ethidium bromide.
Co-immunoprecipitation-Co-immunoprecipitation was performed as described (27) with minor modifications. Whole cell extracts were prepared by resuspending IMR90 or HFF cells in lysis buffer containing 20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 1 mM sodium orthovanadate, 5 mM sodium fluoride, and Boehringer Complete protease inhibitor mixture, and incubated on ice for 30 min. After centrifugation of the lysates at 16,000 ϫ g for 5 min at 4°C, 1 mg of the supernatants (cell extracts) were precleared using 80 l of protein A-agarose (Santa Cruz Biotechnology) and incubated with the respective antibody with rotation for 2 h at 4°C. After addition of 60 l of protein A-agarose bead suspension (Santa Cruz Biotechnology), the mixture was further incubated with rotation for 4 h at 4°C. After three washes with extraction buffer, the beads were resuspended in 50 l of extraction buffer. For immunoprecipitation of the endogenous proteins, anti-Mad (65396E, Pharmingen), anti-Sp1 (sc-59, Santa Cruz Biotechnology), and anti-Sp3 (07-107, Upstate Biotechnology) antibodies were used. Anti-HA antibody (71-5500, Zymed Laboratories Inc.) was used as a negative control for the immunoprecipitation of the endogenous proteins. For immunoprecipitation of the HA-tagged proteins, anti-HA antibody (sc-7392, Santa Cruz Biotechnology) was used.
Western Blot-Proteins were separated by 10% SDS-PAGE and transferred to nitrocellulose membranes (Millipore). The membranes were blocked with 5% nonfat milk and probed with anti-HDAC2 antibody (51-5100, Zymed Laboratories Inc.). Membranes were then incubated with horseradish peroxidase-conjugated anti-rabbit IgG (sc-2054, Santa Cruz Biotechnology) and visualized using the ECL system (Amersham Biosciences).

RESULTS
Identification of the DNA Element Critical for the Transcriptional Repression of hTERT-To identify the DNA element within the hTERT promoter critical for its transcriptional repression, we generated a series of 5Ј-deletion mutant constructs from the 3396-bp hTERT promoter cloned into the firefly luciferase reporter plasmid (Fig. 1A). These constructs were transiently transfected into normal human lung fibroblasts (IMR90), where the expression of the hTERT gene was undetectable and tightly repressed. Subsequently, luciferase assays were performed. Serial deletions from Ϫ3396 up to Ϫ275 did not affect the hTERT promoter activity (Fig. 1A).

Sp1 and Sp3 Recruit HDAC to Repress hTERT Promoter
However, further deletion from Ϫ274 to Ϫ189 caused a slight increase in promoter activity, consistent with our previous report that the E-box within this region acts as the repressive element through interaction with the Mad-Max complex in normal human somatic cells (19). Interestingly, further deletion from Ϫ188 to Ϫ180 increased the promoter activity even more significantly. These findings suggest that a repressive regulatory element is within or extends into the region between Ϫ188 and Ϫ180 of the hTERT promoter.
We next investigated whether the abrupt increase in the hTERT promoter activity after deletion of the region from Ϫ188 to Ϫ180 can be observed in other normal human cells. The deletion constructs p-188 and p-179, which differ only by the "9 base pairs (bp)" between Ϫ188 and Ϫ180 of the hTERT promoter, were transfected into three different kinds of normal human fibroblasts, and then the promoter activities of the two hTERT promoter constructs were compared (Fig. 1B). Promoter activity of p-179 was significantly higher than that of p-188 in all of these cells. Thus, it is quite possible that the region of sequence between Ϫ188 and Ϫ180 may act as a repressive element in many types of normal human cells.
The Identified Repressive Element Responds to TSA, but Only in the Context of the hTERT Promoter-Because histone deacetylation has been proposed as one transcriptional repression mechanism for the hTERT gene (22), we examined whether the identified repressive element might involve this mechanism. To this end, we exploited the effects of TSA, which specifically inhibits histone deacetylases (30). As a preliminary experiment, we treated IMR90 cells with increasing amounts of TSA and examined the changes in the hTERT mRNA levels by quantitative reverse transcriptase-PCR with total RNA isolated from these cells 24 h after TSA treatment. TSA at 200 nM FIG. 1. Identification of the repressive DNA element within the hTERT promoter. A, deletion from Ϫ188 to Ϫ180 of the hTERT promoter results in a significant increase in the promoter activity in IMR90 cells. A schematic diagram of the firefly luciferase reporter plasmids with serial deletions of the hTERT promoter is presented. The A of the translation start codon (ATG) of the hTERT gene is designated as ϩ1, and the name of each reporter construct is assigned according to the 5Ј-end nucleotide number of the inserted promoter sequence. The locations of E-box and Sp-binding sites are presented as previously described (31). Each of these deletion constructs was transiently transfected into IMR90 cells. As a control for variations in transfection efficiency, a Renilla luciferase reporter plasmid under the control of the cytomegalovirus promoter was incorporated in these transfections. Luciferase activity of construct p-3396 was set to 100, and the relative luciferase activity of each of other deletion constructs is presented. B, deletion construct p-179 exhibits higher promoter activity than p-188 in WI38, HFF, and IMR90 cells. For each cell type, the luciferase activity of p-3396 was set to 100, and the relative luciferase activity of the two deletion constructs is presented. Results shown are the average of three experiments, and bars indicate standard deviations.

Sp1 and Sp3
Recruit HDAC to Repress hTERT Promoter resulted in maximal induction of hTERT mRNA without any significant effect on the ␤-actin mRNA level ( Fig. 2A, bottom panels). TSA at concentrations higher than 200 nM did not further increase the hTERT mRNA levels. 2 In parallel with this analysis, we transfected the 3396-bp hTERT promoterfirefly luciferase reporter plasmid into IMR90 cells, and these transfected cells were treated with increasing amounts of TSA. Similar to the results for endogenous hTERT mRNA, 200 nM TSA caused the maximal induction of activity from the transfected hTERT promoter ( Fig. 2A, top panel).
Next, we assayed the hTERT promoter activities of p-188 and p-179 after transfection into IMR90, WI38, and HFF cells and treatment with TSA (Fig. 2B). The promoter activities of p-188 in these normal human somatic cells were normalized to compare them with those of p-179 with or without TSA treatment. Interestingly, the promoter activity of p-188 was sharply up-regulated in comparison with that of p-179 after treatment with TSA. In contrast, p-188 had less activity than p-179 in the absence of TSA treatment (compare ϩ and Ϫ TSA in Fig. 2B; also see Fig. 1B). These results suggest that the region between Ϫ188 and Ϫ180 of the hTERT promoter may be engaged in the HDAC-mediated tight transcriptional repression of the hTERT gene in normal human somatic cells.
Given the repressive function of the region between Ϫ188 and Ϫ180, we investigated whether this region of the hTERT promoter was sufficient to function as a TSA-responsive repressive element within other promoters. To address this, the sequences around the region between Ϫ188 and Ϫ180 of the hTERT promoter (RE in Fig. 2C) were cloned upstream of the SV40 promoter. Then the activity of the SV40 promoter was measured by firefly luciferase reporter expression in IMR90, WI38, and HFF cells (Fig. 2C). The promoter activities of pGL3-Promoter in these normal human somatic cells were normalized to compare them with those of pGL3-Promoter-RE in the absence or presence of TSA treatment. In IMR90 cells, this introduced element decreased SV40 promoter activity in the absence of TSA treatment. In WI38 and HFF cells, this element did not significantly down-regulate the SV40 promoter activity. In addition, this isolated element supported the TSA-mediated induction of the SV40 promoter in IMR90 slightly, but not in WI38 and HFF cells. Taken together, these results imply that the repressive function of this element in normal human somatic cells requires the promoter context of the hTERT gene.
The TSA-responsive Repressive Element of the hTERT Promoter Forms Complexes with Sp1 and Sp3-To identify the nuclear protein(s) binding to the identified repressive element of the hTERT promoter, we performed an EMSA with nuclear extracts from IMR90 and HFF cells probed with the repressive element (Fig. 3). As a preliminary experiment, we performed the competition assays with unlabeled oligonucleotides containing either consensus or mutant-binding sites for Sp1, AP2,

Sp1 and Sp3 Recruit HDAC to Repress hTERT Promoter
Egr, and E2F. Because the DNA sequence within the identified repressive element is GC-rich (see Fig. 4A), it was possible that the putative transcription factors (Sp1, AP2, Egr, and E2F) could interact with the DNA element. We found that the upper three DNA-protein complexes (I, II, and III in Fig. 3) were specifically competed by the oligonucleotides containing consensus binding sequences for Sp1, but not by the oligonucleotides containing the binding sequences for AP2, Egr, or E2F. 2 Consistent with these results, when these DNA-protein complexes (I, II, and III) were challenged with anti-Sp1 and anti-Sp3 antibodies, we detected the presence of Sp1 in DNA-protein complexes I and II, and Sp3 in DNA-protein complexes I and III (Fig. 3). Under these conditions, we also found that antibodies against AP2, Egr, E2F1, and other Sp family members (e.g. Sp2 and Sp4) did not affect these specific band-shift patterns. 2 We next investigated any changes in these DNA-protein complexes after treatment with TSA. Nuclear extracts from IMR90 and HFF cells were treated with or without TSA. We then compared band-shift patterns with the radiolabeled probe encompassing the repressive element. No additional band(s) appeared after treatment with TSA (Fig. 3). Furthermore, we failed to detect any significant alteration in the band-shift patterns. Notably, the DNA binding activities of neither Sp1 nor Sp3 changed after TSA treatment. These results imply that the induced DNA binding activities of Sp1 and Sp3 may not be involved in the up-regulation of the hTERT promoter activity after treatment with TSA.
Mutation of the Sp Sites in the Proximal hTERT Promoter Increases Promoter Activity and Decreases TSA Responsiveness-Because the TSA-responsive repressive element (Ϫ188 to Ϫ180) can form complexes with Sp1 and Sp3 in nuclear extracts (Fig. 3), the Sp site within that region was assumed to be responsible for the HDAC-mediated repression of the hTERT promoter. To assess this, we mutated the core se-quences of the Sp site within this TSA-responsive repressive element (Sp site 1 in Fig. 4A; p-188-Sp-m1 in Fig. 4B). Then we examined whether the mutation produced the same effects (increased promoter activity and decreased TSA-responsiveness) as the deletion of the repressive element. The construct p-188-Sp-m1 showed enhanced promoter activity in comparison with its wild-type counterpart p-188 in all three cell types (Fig. 4B). This enhancement was similar to that observed by deletion of the repressive element (from Ϫ188 to Ϫ180) from p-188 in Fig. 1B. In addition, we investigated the TSA responsiveness of this Sp site by comparing the promoter activities of p-188 and p-188-Sp-m1 in these cells after treatment with TSA. The p-188-Sp-m1 showed less promoter activity in comparison with p-188 after treatment with TSA (Fig. 4B). Consistently, we found that a mutation in this Sp site produced similar results to those with the p-179 construct, relative to p-188 (compare Fig. 2B with Fig. 4B). Thus, the Sp site 1 is essential for the TSA-responsive repressive function of the hTERT promoter element between Ϫ188 and Ϫ180.
Based on the importance of the Sp site in the repressive function of our identified element, we examined the possible involvement of other Sp sites within the proximal hTERT promoter. Kyo et al. (31) described the existence of four other Sp sites within the region from Ϫ188 to the translation start site of the hTERT promoter (see Fig. 4A). We mutated these remaining four Sp sites within the promoter region of the construct p-188 (Fig. 4A), and transfected these mutant constructs (p-188-Sp-m2, p-188-Sp-m3, p-188-Sp-m4, and p-188-Sp-m5) into IMR90, WI38, and HFF cells. Mutation of Sp sites 2 and 5, but not Sp sites 3 and 4, increased the promoter activity in these cells, to a similar extent as mutation of the Sp site 1 (Fig.  4B). Furthermore, the mutations introduced in Sp sites 2 and 5 decreased the induction of the hTERT promoter after TSA treatment (Fig. 4B). Taken together, these results suggest that multiple Sp sites within the proximal hTERT promoter may be engaged in the HDAC-mediated transcriptional repression of the hTERT gene in normal human somatic cells.
The Endogenous hTERT Promoter Is Associated with Sp1 and Sp3-To directly address whether Sp1 and Sp3 are associated with the repressed hTERT promoter, we used formaldehyde X-ChIP analysis (Fig. 5A). After formaldehyde cross-linking of the IMR90, WI38, and HFF cells, chromatin immunoprecipitation was performed with antibodies directed against Mad1, Sp1, Sp3, and HA. The precipitated DNA was subjected to PCR with the use of specific primers for the hTERT proximal promoter region. As expected from previous reports (19,32), tight association of the hTERT promoter with Mad1 was observed in all three cell types (Fig. 5A). In addition, both Sp1 and Sp3 were found to be associated with the endogenous hTERT promoter. The tight association of Sp1 and Sp3 with the repressed hTERT promoter strongly suggested that Sp1 and Sp3 may play a functional role in transcriptional repression of the hTERT promoter. This idea is also consistent with our results in Fig. 4B showing that mutation of Sp sites increased the hTERT promoter activity in normal human somatic cells.
Examination of the DNA binding ability of Sp1 and Sp3 after treatment with TSA showed that there is no significant alteration in the DNA-protein complex formation with the repressive element of the hTERT promoter in the EMSA (Fig. 3). We further examined these observations in normal human somatic cells with X-ChIP analysis after treatment with TSA (Fig. 5B). The cross-linked chromatin immunoprecipitated with anti-Sp1, anti-Sp3, or anti-HA antibody was analyzed by quantitative PCR with specific primers for the proximal hTERT promoter. Consistent with our EMSA results, the occupancy of the hTERT promoter by Sp1 and Sp3 within IMR90, WI38, and HFF cells did not change after treatment with TSA (Fig. 5B). These results further indicate that the TSA-mediated increase in the hTERT promoter activity in normal human somatic cells is not because of the increased recruitment of Sp1 or Sp3 onto the hTERT promoter.
Endogenous Sp1 and Sp3 Are Associated with HDAC in Human Normal Somatic Cells-Based on the association of Sp1 and Sp3 with the repressed hTERT promoter, we examined the physical interaction of Sp1 and Sp3 with HDAC through co-immunoprecipitation (Fig. 6). As a preliminary experiment, we performed a Western blot with IMR90 and HFF cell extracts using anti-HDAC1, anti-HDAC2, anti-HDAC3, anti-HDAC4, anti-HDAC5, and anti-HDAC6 antibodies to examine which HDAC family member is highly expressed in these cells. Because HDAC2 is the predominant form in these cells, 2 we immunoprecipitated Sp1 and Sp3 protein complexes from IMR90 and HFF cell extracts with anti-Sp1 and anti-Sp3 antibody, respectively, and then we performed Western blot with anti-HDAC2 antibody (Fig. 6). As a positive control, we used pRB, which is known to interact with HDAC2 (33). Both Sp1 and Sp3 formed complexes with HDAC2 (Fig. 6). Interestingly, we observed a reproducibly stronger interaction of Sp3 with HDAC2 under these conditions (Fig. 6). 2 The association of Sp1 and Sp3 with HDAC further supports our conclusion that Sp1 and Sp3 play an important role in transcriptional repression of the hTERT promoter through recruitment of HDAC in normal human somatic cells.
Ectopic Expression of HDAC-binding Domain of Sp1 and Sp3 Increases hTERT Promoter Activity and Decreases Its TSA Responsiveness--Lastly, we addressed whether the derepression of the hTERT promoter by mutations (Figs. 1, 2B, and 4B) FIG. 5. Association of Sp1 and Sp3 with the hTERT proximal promoter. A, the endogenous hTERT promoter is associated with Sp1 and Sp3. A formaldehyde cross-linked X-ChIP was performed with IMR90, WI38, and HFF cells using the antibodies indicated. Genomic DNA (input) and anti-HA antibody served as positive and negative controls, respectively. B, TSA treatment does not increase the recruitment of Sp1 or Sp3 onto the hTERT promoter. IMR90, WI38, and HFF cells were treated with or without 200 nM TSA for 24 h and analyzed by X-ChIP assay with anti-Sp1, anti-Sp3, and anti-HA (an irrelevant, negative control antibody) antibodies. No bands corresponding to hTERT promoter were detected using anti-HA antibody 2 and these negative control lanes were omitted for simplicity. can also be obtained by blocking the protein-protein interactions between the HDAC and transcription factors Sp1 and Sp3. Our results from the analyses of TSA responsiveness (Figs. 2B and 4B) and HDAC interaction (Fig. 6) indicated that Sp1 and Sp3 may play a role in the transcriptional repression of the hTERT gene through interaction with HDAC and its recruitment onto the promoter. Thus, we designed a dominantnegative approach in which we overexpressed the deletion mutants of Sp1 and Sp3, which retain the region important for the HDAC interaction but lack the DNA-binding domain (zinc finger regions indicated in Fig. 7A).
We first examined the interaction of HDAC2 with the Nterminal and C-terminal domains of Sp1 and Sp3 in normal human somatic cells. To this end, HA-tagged Sp1, Sp3, and their truncated mutants were expressed in IMR90 cells (Fig.  7A). After transient expression in IMR90 cells, protein was immunoprecipitated with anti-HA antibody and assayed for the presence of HDAC2 by Western blot with anti-HDAC2 antibody (Fig. 7B). The expression of each HA-tagged Sp protein was normalized in transfected cells. 2 Under these conditions, full-length Sp1 and Sp3 interacted with HDAC2. Interestingly, the N-terminal domain of Sp1 and Sp3 seemed to interact with HADC2 more efficiently than the C-terminal DNA-binding domain. These results were reproducible, 2 and suggest that the N-terminal domain of Sp1 and Sp3 can interact with HDAC2 in normal human somatic cells.
We next investigated the intrinsic TSA responsiveness of the N-terminal regions of Sp1 and Sp3. To this end, we took advantage of a GAL4 system in which the N-terminal domains of Sp1 and Sp3 were fused with the GAL4 DNA-binding domain (Fig. 7C). These GAL4 fusion constructs were co-transfected with pG5-luciferase reporter plasmids containing five consensus GAL4-binding sites into IMR90, WI38, and HFF cells in the absence or presence of TSA. The GAL4 DNA-binding domain itself supported TSA-mediated transcriptional activation slightly (by 4 -6-fold). Under these conditions, the N-terminal 621-amino acid residues of Sp1 and Sp3 exhibited a dramatic increase in TSA responsiveness (Fig. 7C). Interestingly, consistent with stronger HDAC2 interaction of the N-terminal domain of Sp1, compared with full-length Sp1 (Fig. 7B) mutant that specifically abrogates the HDAC-mediated repression of the hTERT promoter. In transient transfection experiments, these deletion mutants of Sp1 and Sp3 up-regulated the hTERT promoter activity in a dose-dependent manner in IMR90 cells (Fig. 7D). These results suggest that hTERT promoter activity may be derepressed through blocking the Sp1/ Sp3 interaction with HDAC. In addition, we investigated the TSA responsiveness with overexpression of these deletion mutants after treatment with TSA (Fig. 7E). Consistent with the results obtained from the mutation of the repressive element ( Fig. 2B and 4B), overexpression of the mutants inhibited the induction of the hTERT promoter activity in the presence of TSA (Fig. 7E). These results are consistent with the idea that Sp1 and Sp3 play a role in transcriptional repression of the hTERT promoter through recruitment of HDAC in normal human somatic cells. DISCUSSION Previous studies have implicated histone deacetylation in the transcriptional repression of the hTERT gene in normal human somatic cells (22)(23)(24). However, little is known about the mechanisms, including the identity of transcription factor(s) that recruit HDAC to the hTERT promoter. Our present studies provided direct evidence that the hTERT promoter is tightly associated with Sp1 and Sp3 in normal human somatic cells, and these transcription factors in turn recruit HDAC, resulting in transcriptional silencing of the hTERT gene.
In contrast to the demonstrated role of Sp3 as a transcriptional repressor (34 -38), Sp1 has most often been described as a transcriptional activator. The ectopic overexpression of Sp1 was reported to induce transcription of the hTERT gene in tumor cells (21,31). We also observed that transient transfection of Sp1 expression plasmids induced the hTERT promoter in normal human somatic cells. 2 However, as demonstrated in many previous studies, these overexpression experiments could misidentify the normal function of the transcription factor in cells. Indeed, we have obtained a series of results indicating that Sp1, like Sp3, plays a role in the transcriptional repression of the hTERT promoter. Most importantly, our X-ChIP analysis demonstrates that the endogenous hTERT promoter was associated with Sp1 in the absence of ectopic Sp1 expression (Fig. 5). Furthermore, a dominant-negative deletion mutant of Sp1, which contains the N-terminal 621-amino acid residue encompassing the activation domains, was found to relieve the repression of the hTERT promoter (Fig. 7D). This dominant-negative experiment suggested that endogenous Sp1 may repress the hTERT promoter in normal human somatic cells. We would expect the down-regulation of hTERT promoter activity in this dominant-negative experiment, if endogenous Sp1 is an activator. In fact, in certain promoter contexts, Sp1 has been reported to be engaged in transcriptional repression (39 -44). Moreover, several studies have indicated the potential interaction of Sp1 with HDAC (45)(46)(47)(48). Given our consistent results, how can we explain the activation of the hTERT promoter upon the overexpression of Sp1? Sp1 has the ability to interact with other Sp1 proteins to form multimeric complexes, which in turn activate transcription synergistically (49). Thus, excessive Sp1 protein could shift to form higher order Sp1 homomultimers, which might be nucleated by the Sp1 proteins already occupying the hTERT promoter. Related to this possibility, Sp3, which cannot form these multimeric complexes, failed to activate hTERT promoter upon its ectopic overexpression in normal human somatic cells. Taken together, our studies revealed the repressive function of Sp1 (and Sp3) in the hTERT promoter activity in normal human somatic cells.
TSA-mediated induction of the hTERT transcription was observed previously in various normal human somatic cells.
Based on these results, an HDAC-mediated mechanism was proposed as the universal transcriptional repression mechanism of the hTERT gene (22)(23)(24). Our results demonstrated the involvement of Sp1 and Sp3 in the TSA-mediated induction of the hTERT promoter activity through recruitment of HDAC. Furthermore, we found that Sp1 and Sp3, the transcription factors binding to the identified TSA-responsive element, were not altered by TSA treatment; TSA did not alter their abundance in the nucleus, their DNA binding activities, or their occupation of the hTERT promoter ( Figs. 2A and 3B). 2 What then is responsible for the Sp1/Sp3-mediated activation/derepression of the hTERT promoter in response to TSA? We observed that Sp1 and Sp3 fused to the GAL4 DNA-binding domain were able to dramatically induce transcription after treatment with TSA. These results suggest that TSA may convert the Sp1 and Sp3 on the hTERT promoter from repressors into activators by the abrogation of the associated HDAC activity.
Activation of the hTERT gene is a crucial step during the immortalization and malignant transformation of human cells. However, it is poorly understood how the hTERT gene is activated. Multiple mechanisms are probably involved in transcriptional activation of the hTERT gene during tumorigenesis. For example, deregulated c-Myc overexpression might contribute to the up-regulation of the hTERT promoter in the course of tumorigenesis. The amplification of the hTERT locus during tumorigenesis probably contributes to the misregulation of hTERT transcription in some kinds of human tumors (14). In addition to these previously proposed mechanisms, our studies indicate that abrogation of the regulatory pathway(s) governing the functional interaction of Sp1 or Sp3 with HDAC could be an important mechanism for inducing hTERT gene expression during tumorigenesis. This newly proposed mechanism for the transcriptional derepression of the hTERT gene could provide the platform for the detailed analyses of the transcriptional regulation of the hTERT gene, and help to unveil possible therapeutic targets for the development of therapeutic drugs against cancer.