Human Telomerase Reverse Transcriptase (hTERT) Transcription Requires Sp1/Sp3 Binding to the Promoter and a Permissive Chromatin Environment*

Background: Regulation of telomerase gene (hTERT) by Sp1 family proteins remains contradictory in previous studies. Results: In chromatinized BAC reporters, Sp1/Sp3 binding to the promoter was essential for hTERT transcription, but did not affect its repression. Conclusion: Chromatin remodeling, not Sp1 binding to the promoter, was a limiting step in hTERT transcription. Significance: Chromatinized reporters are better models for studying telomerase regulation. The transcription of human telomerase gene hTERT is regulated by transcription factors (TFs), including Sp1 family proteins, and its chromatin environment. To understand its regulation in a relevant chromatin context, we employed bacterial artificial chromosome reporters containing 160 kb of human genomic sequence containing the hTERT gene. Upon chromosomal integration, the bacterial artificial chromosomes recapitulated endogenous hTERT expression, contrary to transient reporters. Sp1/Sp3 expression did not correlate with hTERT promoter activity, and these TFs bound to the hTERT promoters in both telomerase-positive and telomerase-negative cells. Mutation of the proximal GC-box resulted in a dramatic decrease of hTERT promoter activity, and mutations of all five GC-boxes eliminated its transcriptional activity. Neither mutations of GC-boxes nor knockdown of endogenous Sp1 impacted promoter binding by other TFs, including E-box-binding proteins, and histone acetylation and trimethylation of histone H3K9 at the hTERT promoter in telomerase-positive and -negative cells. The result indicated that promoter binding by Sp1/Sp3 was essential, but not a limiting step, for hTERT transcription. hTERT transcription required a permissive chromatin environment. Importantly, our data also revealed different functions of GC-boxes and E-boxes in hTERT regulation; although GC-boxes were essential for promoter activity, factors bound to the E-boxes functioned to de-repress hTERT promoter.

The transcription of human telomerase gene hTERT is regulated by transcription factors (TFs), including Sp1 family proteins, and its chromatin environment. To understand its regulation in a relevant chromatin context, we employed bacterial artificial chromosome reporters containing 160 kb of human genomic sequence containing the hTERT gene. Upon chromosomal integration, the bacterial artificial chromosomes recapitulated endogenous hTERT expression, contrary to transient reporters. Sp1/Sp3 expression did not correlate with hTERT promoter activity, and these TFs bound to the hTERT promoters in both telomerase-positive and telomerase-negative cells. Mutation of the proximal GCbox resulted in a dramatic decrease of hTERT promoter activity, and mutations of all five GC-boxes eliminated its transcriptional activity. Neither mutations of GC-boxes nor knockdown of endogenous Sp1 impacted promoter binding by other TFs, including E-box-binding proteins, and histone acetylation and trimethylation of histone H3K9 at the hTERT promoter in telomerase-positive and -negative cells. The result indicated that promoter binding by Sp1/Sp3 was essential, but not a limiting step, for hTERT transcription. hTERT transcription required a permissive chromatin environment. Importantly, our data also revealed different functions of GC-boxes and E-boxes in hTERT regulation; although GCboxes were essential for promoter activity, factors bound to the E-boxes functioned to de-repress hTERT promoter.
Chromosomal ends are capped by telomeres, the nucleoprotein structures containing simple DNA repeats of TTAGGG (1). In cancer cells, germ cells, and some stem cells, telomeres are maintained and/or elongated by telomerase (2), a ribonucleoprotein that contains an RNA template TERC, a reverse transcriptase TERT, and several associated proteins (3,4). In contrast, most human somatic cells express little or no telomerase (5). As a result, telomeres progressively shorten upon successive cell divisions, leading to replicative senescence or cell death and functioning as a biological aging clock in somatic cells (6).
Telomerase regulation plays important roles in human aging and cancer. Although TERC and other telomerase-associated proteins are ubiquitously expressed (7), hTERT 3 is the limiting subunit of telomerase in many cells. Studies have shown that hTERT transcription is the primary step of telomerase regulation (8). The regulation of hTERT transcription requires binding of sequence-specific transcription factors (TFs), such as those of Sp1, c-Myc, USF, and E2F families, to the hTERT promoter (9 -13). However, most of these TFs are widely expressed, and as such, their expression does not account for hTERT regulation during cell differentiation and tumorigenesis. Indeed, epigenetic mechanisms and chromatin environment also play critical roles in hTERT regulation (14). Our data showed that the hTERT gene was embedded in a condensed heterochromatin-like domain in many somatic cells (15). Inhibition of histone deacetylases (HDACs) led to an opening of this domain and activated hTERT transcription (14). It remains to be determined how TFs, such as Sp1, bind to the hTERT promoter in the context of this repressive chromatin environment (16).
Sp1 is a member of the specificity protein/Krüppel-like factor (SP/KLF) TF family, which also includes Sp2, Sp3, and Sp4 (17). Sp1 and Sp3 are widely expressed, but the expression of Sp2 and Sp4 is restricted to specific tissues. Sp1 family TFs bind to consensus sequence GGGGCGGGG, known as GC-boxes. Although Sp1 and Sp3 proteins are known to regulate transcription both positively and negatively, cumulative evidence also suggests that the chromatin context is an important factor in the differential loading of Sp1 and Sp3 to promoters (18).
The hTERT core promoter lacks a TATA box but contains an array of five GC-boxes surrounded by two E-boxes, in addition to an initiator element (9,19). This arrangement is reminiscent of the TATA-less promoters found in many "housekeeping" genes (20). GC-boxes are essential for promoter activity in many TATA-less promoters. However, studies on the role of Sp1 family proteins in hTERT regulation have yielded conflicting results thus far. Some studies indicated that Sp1 and GC-boxes played a positive role in hTERT transcription, because mutations of GC-boxes reduced hTERT promoter activity (21). Conversely, other studies concluded that Sp1 family proteins and their binding sites were critical for the repression of hTERT promoter (22,23). Such puzzling results might have resulted from the fact that most of these investigations were based on transient transfection of small plasmid reporters, because previous studies demonstrated the limitation of transient assays in deciphering the function of Sp1/Sp3 factors and their binding sites in promoters (24).
To understand the roles of Sp1 family TFs in hTERT regulation, we generated mutations of GC-boxes at the hTERT promoter in a bacterial artificial chromosome (BAC) reporter, H(wt), containing 160 kb of human genomic DNA encompassing the consecutive CRR9 (also called CLPTM1L gene), hTERT, and Xtrp2 (or SLC6A18) loci. H(wt) and its mutant derivatives were then integrated at an acceptor site in telomerase-positive 3C167b (Tel ϩ ) and -negative GM847 (Tel Ϫ ) cells, via Cre recombinase-mediated BAC targeting (RMBT) technique (25). Therefore, for the first time, the functions of GC-boxes and Sp1/Sp3 binding to these sites were examined in a relevant genomic and chromatin context. Our data indicated that the binding of Sp1 and Sp3 proteins to their cognate recognition sites was essential for the hTERT transcription in Tel ϩ cells. Interestingly, Sp1 and Sp3 bound to the hTERT promoter in both Tel ϩ and Tel Ϫ cells, indicating that the binding of Sp1/Sp3 TFs was required but not sufficient for hTERT transcription and hTERT activation required a permissive chromatin environment. Furthermore, in this chromatinized reporter system, mutations of GC-boxes did not affect the states of histone acetylation and trimethylation of histone H3K9 at the hTERT promoter or binding of other TFs to the promoter. The residual activities of mutant promoters were induced proportionally as H(wt) upon inhibition of HDACs by trichostatin A (TSA), indicating that the binding of Sp1/Sp3 to the hTERT promoter had no significant impact on the repressive states of hTERT promoter. This was in sharp contrast to the proximal E-box at the hTERT promoter: its binding by TFs de-repressed the hTERT promoter.

Experimental Procedures
BACs, Plasmids, and Transfection-hTERT promoter reporter plasmids were obtained from Dr. Horikawa (22). pBT-255 contained the 295-bp hTERT promoter fragment (Ϫ255 to ϩ40 nucleotides, relative to the transcription start site, and directly upstream of the hTERT initiation codon) in pGL3-Basic vector (Promega, WI). pBT-255gc, or hM12 (22), contained a point mutation that eliminated the proximal GC-box, Ϫ31 nucleotides relative to the transcription start site (see Fig. 1A). pLXSN-Sp1 was constructed by cloning a human Sp1 cDNA into retroviral vector pLXSN. In transient transfection assays with small plasmids, reporter and control plasmids were transfected into cells in triplicate using Lipofectamine 2000 (Life Technologies), and luciferase activities were measured 48 h later. Firefly luciferase (Fluc) activities were normalized to Renilla luciferase (Rluc) activities from co-transfected pRL-SV40 (Promega).
BAC reporter constructs 117B23-cFtRvSVP, also named H(wt), and H(EboxD), which contained a mutation at the downstream E-box, were reported earlier (26). H(gc) and H(gc5) contained point mutations at the proximal GC-box (Ϫ31 nucleotides) and all five GC-boxes (Ϫ132, Ϫ112, Ϫ78, Ϫ58, and Ϫ31 nucleotides) at the hTERT promoter, respectively (see Fig. 1A). The mutations in the BAC reporters were generated using our two-step BAC recombineering method (27). In transient experiments, BACs were transfected into cells using FuGENE HD (Roche Applied Science), and luciferase assays were performed after 48 h.
Gene Expression Analyses-Total RNAs were extracted from cultured cells, and reverse transcription and quantitative PCR (qRT-PCR) were performed as described previously (28). Primers are shown in Table 1 or published previously (26). Western analyses were performed as described in Ref. 26. hTERT promoter. Our previous data showed that the hTERT gene was embedded in a nuclease-resistant chromatin domain in human somatic cells (15). The hTERT promoter was repressed in both telomerase-positive and telomerasenegative cells, and the repression was especially stringent in telomerase-negative cells. We set out to determine the roles of Sp1 family TFs in the regulation of hTERT transcription in this repressive chromatin environment. First, a set of human fibroblast cells was selected for this study (14): telomerasepositive 3C167b and GM639, telomerase-negative alternative lengthening of telomere (ALT) lines 3C166a and GM847, as well as normal human fibroblasts. Their endogenous hTERT mRNA expression was validated by qRT-PCR analysis (Fig. 1B). Although Sp1 and Sp3 were expressed in all the cells, their protein levels did not correlate with hTERT mRNA levels (Fig. 1C). Thus, the expression of Sp1 and Sp3 Horizontal arrows indicate directions of transcription. The 295-bp core promoter upstream of hTERT initiation codon is shown in the lower portion. BAC and plasmid constructs containing the mutations are listed on the right. B, hTERT mRNA expression in fibroblasts. mRNA levels were determined by qRT-PCR analyses and normalized to 18S ribosomal RNA. Normal human fibroblasts (NHF), GM847, and 3C166a were telomerase-negative cells, whereas GM639 and 3C167b were telomerase-positive lines. Error bars indicate means Ϯ S.E. C, expression of Sp1 and Sp3 TFs. Protein expression was determined by Western analyses using antibodies as indicated on the left. proteins did not account for the differential hTERT expression in these cells.
In most previous studies, small plasmid reporters were used to understand the hTERT promoter functions. Thus, we first used pBT-255, a firefly luciferase reporter plasmid containing a 295-bp hTERT promoter fragment (29), and its derivative pBT-255gc containing point mutations at the proximal GC-box (22). This promoter fragment contained all the TF binding sites shown in Fig. 1A, including five GC-boxes, two E-boxes, and three E2F sites. As shown in Fig. 2A, mutation of the proximal GC-box resulted in a 30% decrease in promoter activity in both telomerase-positive 3C167b and telomerase-negative GM847 cells, designated hereinafter as Tel ϩ and Tel Ϫ cells, respectively. In both cells, overexpression of Sp1 protein led to an ϳ2-fold increase of the hTERT promoter activity (Fig. 2, B and  C). Conversely, knockdown of Sp1 expression using lentiviral shRNAs resulted in a significant reduction of luciferase activity of pBT-255 (Fig. 2D). Therefore, in this transient plasmid reporter system, Sp1 factor and the GC-boxes mediated the activation of the hTERT promoter.
To determine the potential roles of distal regulatory elements in hTERT regulation, we introduced H(wt), a BAC reporter that contains a 160-kb human genomic DNA encompassing the CRR9, hTERT, and Xtrp2 genes, as shown in Fig. 1A. A Renilla (Rluc) and a firefly luciferase (Fluc) cassette were inserted into the initiation codons of hTERT and CRR9 genes, respectively. Because CRR9 was ubiquitously expressed in all cells and tissues examined (30), Fluc was used as an internal control, and the ratio of Rluc to Fluc activities was a direct measurement of hTERT promoter activity. As shown in Fig. 3A, upon transient transfection of H(wt), Rluc/Fluc ratios varied in different cell lines, but did not correlate with endogenous hTERT expression (Fig. 2B). Sp1 overexpression resulted in increases from 40% to 1-fold in the hTERT promoter activity in H(wt) (Fig. 3B), consistent with an activator role of the Sp1 protein.
To test the functions of Sp1 in the context of BAC reporters, mutations were introduced into GC-boxes in H(wt). H(gc) contained a point mutation in the proximal GC-box, whereas all five GC-boxes in the promoter region were mutated in H(gc5) hTERT Regulation by Sp1 Family Transcription Factors (Fig . 1A). In both Tel ϩ and Tel Ϫ cells, hTERT promoter activity was reduced by 20 -30% in H(gc) and by 40 -50% in H(gc5) in transient transfection of BAC reporters (Fig. 3C). The impact of GC-box mutations on hTERT promoter in the BAC reporter was similar to that in plasmid reporter pBT-255. Overexpression of Sp1 activated the hTERT promoter in wild-type H(wt) by about 2-fold, and its effect on H(gc) and H(gc5) was greatly diminished (Fig. 3D), indicating that Sp1 activated hTERT promoter primarily via these GC-boxes.
Chromatin is an essential component of hTERT regulation (16). To determine the roles of GC-boxes and their binding proteins in the hTERT regulation in a relevant chromatin context, we used our previously reported RMBT method to integrate single-copy BAC reporters into chromosomal acceptor sites in Tel ϩ and Tel Ϫ cells (Fig. 4A) (25). As we previously reported, the hTERT promoter in H(wt) was over 20-fold stronger in Tel ϩ cells than in Tel Ϫ cells (Fig. 4B). HDAC inhibition by TSA dramatically induced hTERT promoter activity in both cells, indicating that the chromatinized H(wt) recapitulated the repression of endogenous hTERT locus.
To determine the binding of Sp1 protein to the hTERT promoters, ChIP experiments were performed. As shown in Fig.  4C, although the association of E-box-binding proteins Max and USF1/2 with the endogenous and transgenic hTERT promoters was higher in Tel ϩ cells than in Tel Ϫ cells, Sp1 bound to the hTERT promoters similarly in these two cells. As a control, trimethylation of H3K4 was significantly higher in Tel ϩ cells than in Tel Ϫ cells, consistent with hTERT transcription in Tel ϩ cells. Thus, Sp1 binding to hTERT promoter occurred in both Tel ϩ and Tel Ϫ cells and did not correlate with hTERT activa-tion. However, when Sp1 was overexpressed, the chromatinized hTERT promoters were up-regulated by about 2-fold in both Tel ϩ and Tel Ϫ cells (Fig. 4D).
To decipher the role of GC-boxes in a relevant chromatin context, H(gc) and H(gc5) were inserted into the same chromosomal acceptor sites as H(wt) in Tel ϩ and Tel Ϫ cells via RMBT (Figs. 1A and 4A). As shown in Fig. 5A, mutation of the proximal GC-box in H(gc) resulted in over 80% loss of hTERT promoter activity, and mutations of all five GC-boxes virtually abolished promoter function. When HDACs were inhibited by TSA, Rluc expression from H(gc) was proportionally induced, as compared with those from H(wt) in Tel ϩ and Tel Ϫ cells. TSA failed to induce Rluc expression from H(gc5), consistent with the fact that the hTERT promoter function was abolished in this reporter. This was in sharp contrast to H(EboxD), in which the downstream E-box was mutated. The hTERT promoter activity in H(EboxD) was reduced by 5-fold as compared with that of H(wt). However, upon TSA treatment, the mutant promoter was induced to a level similar to those of wild-type promoter in H(wt) in both Tel ϩ and Tel Ϫ cells. This result suggested that GC-boxes and E-boxes played different roles in hTERT regulation. While proteins that bind to E-boxes might recruit histone acetyltransferases and de-repress the hTERT promoter, the binding of Sp1 family TFs to GC-boxes was likely essential for its promoter function.
ChIP experiments revealed that the binding of Sp1 and Sp3 to the transgenic promoter was significantly reduced in mutant BACs H(gc) and H(gc5) (Fig. 5B). As a control, Sp1/Sp3 binding to the endogenous hTERT promoter was unaffected in the same cells. Mutations of the GC-boxes were accompanied by a  DECEMBER 11, 2015 • VOLUME 290 • NUMBER 50 reduction of di-and trimethylation of histone H3 lysine 4 (H3K4me2 and H3K4me3) (Fig. 5C), consistent with reduced transcriptional activities of the mutant promoters. However, the recruitment of E-box factors, Max and USF1/2, to the mutant promoters was not affected. Therefore, these results indicated that the binding of Sp1 family proteins to GC-boxes was critical for the promoter activity, but did not affect TF binding to neighboring E-boxes, which activated hTERT transcription by de-repressing the promoter (26). Consistently, these GC-box mutations did not affect acetylation status of histones H3 and H4 (H3Ac and H4Ac), trimethylation of histone H3 lysine 9 (H3K9me3), and the association of histone H1 at the hTERT promoters (Fig. 5C), supporting the notion that Sp1/ Sp3 binding did not influence the repressive chromatin state of hTERT promoter.

hTERT Regulation by Sp1 Family Transcription Factors
To further determine the roles of Sp1 family TFs in regulating hTERT promoter activity, the endogenous expression of Sp1 and Sp3 was knocked down by lentiviral shRNAs. As shown in Fig. 6A, two shRNAs against Sp1, shSp1a and shSp1b, reduced its protein level by 90 and 70%, respectively, and two shRNAs, shSp3a and shSp3b, inhibited Sp3 expression by 70 and 30%, respectively. KD of either Sp1 or Sp3 resulted in over 50% decreases of both endogenous hTERT mRNA levels (Fig.  6B) and transcription from the transgenic hTERT promoter in chromatinized H(wt) (Fig. 6C). Co-transductions of cells with lentiviral shRNAs against Sp1 and Sp3 further reduced the expression from endogenous and transgenic hTERT promoters, indicating that both Sp1 and Sp3 contributed to hTERT transcriptional activation.
Finally, ChIP experiments were performed to determine how Sp1 KD affected the chromatin structure of the hTERT promoters. Sp1 KD reduced the association of Sp1 with both endogenous and transgenic hTERT promoters, but had no overall effects on the binding of other TFs, such as E-box binding factors and E2F family proteins (Fig. 7A). Moreover, H3Ac, H4Ac, H3K4me2, and H3K4me3 were not consistently affected by Sp1 KD (Fig. 7B).

Discussion
Sp1 family proteins, Sp1 and Sp3, are near ubiquitous TFs (17). They can function as activators or repressors depending on their binding sites and interacting cofactors (31). Here, we studied the roles of these TFs at the hTERT promoter in the context of a chromatinized transgenic BAC reporter. Our data showed that, although the expression of Sp1 and Sp3 did not correlate with hTERT transcription, they both contributed to hTERT transcriptional activation, and their binding sites, the five GC-boxes, were crucial for the hTERT promoter function. Although this conclusion was not a surprise given that some of the previous publications using transient plasmid reporters also showed that Sp1 family TFs and the GC-boxes were involved in hTERT activation, our current study using chromatinized BAC reporters demonstrated that these factors were not involved in hTERT repression, in contrast to several earlier studies (22,23). Moreover, the repressive states of hTERT promoter in Tel ϩ and Tel Ϫ cells did not seem to affect Sp1 binding to the promoter, nor did Sp1/Sp3 binding to GC-boxes significantly change the states of histone acetylation and trimethylation of H3K9, a repressive chromatin mark, at the hTERT promoter. In Tel Ϫ cells, the hTERT promoter was silenced while Sp1 bound to the promoter, indicating that the limitation for hTERT transcription occurred at steps following Sp1 binding. Chromatin fragments from Tel ϩ cells containing BAC reporters were precipitated using antibodies against TFs (B) or specific histone modifications (C), followed by qPCR analyses. Shown are precipitated fragments as the percentages of input chromatin. H3Ac and H4Ac refer to acetylated histone H3 and H4, respectively. H3K4me2 and H3K4me3 are di-and trimethylated H3K4. Error bars indicate means Ϯ S.E. DECEMBER 11, 2015 • VOLUME 290 • NUMBER 50

JOURNAL OF BIOLOGICAL CHEMISTRY 30199
There were several advantages using chromatinized BAC reporters to study hTERT regulation. First, the expression from transgenic hTERT promoter in chromosomally integrated H(wt) recapitulated endogenous hTERT transcription in host cells (25), whereas the hTERT promoter in small plasmid reporters or even the transient BAC reporter H(wt) bore no resemblance to the host gene, indicating that some of regulatory components were missing in transient reporter systems. It was likely that the repressive chromatin environment crucial for hTERT regulation was absent in transiently transfected DNA constructs (16). Second, the expression of luciferase reporter in chromatinized H(wt) depended much more on the GC-boxes at the hTERT promoter than those of transiently transfected hTERT reporters. Mutation of the proximal GCbox reduced transcription from the hTERT promoter on pBT-255 by 40 -60% ( Fig. 2A) and that from the transiently transfected H(wt) by 20 -30% (Fig. 3C). In contrast, the same mutation reduced hTERT promoter in chromatinized H(gc) by 7-fold (Fig. 5A). Mutation of all five GC-boxes in H(gc5) decreased hTERT promoter activity in Tel ϩ cells by 300-fold and effectively eliminated hTERT promoter activity even in the presence of HDAC inhibitor TSA in both Tel ϩ and Tel Ϫ cells. The result indicated that it was easier for RNA Pol II and general transcription factors to be loaded onto the transiently transfected DNA constructs than onto chromatinized reporters or endogenous genes, likely due to the lack of an appropriately assembled nucleosomal array in these newly transfected DNA constructs (32,33). In chromatinized BAC reporters, transcriptional initiation depended strictly on the hTERT promoter, making them a more reliable model for studying hTERT regulation.
Repressive chromatin environment, signified by histone deacetylation and other histone modifications, such as H3K9me3, was a key component of hTERT regulation (14). In a previous study, we showed that the hTERT locus was embedded in a condensed heterochromatin-like domain in many somatic cells (15). This domain was present in both 3C167b (Tel ϩ ) and GM847 (Tel Ϫ ) cells used in this study. How hTERT transcription occurs in such a repressive chromatin is currently unknown and a focus of our research. hTERT transcription in Tel ϩ cells correlated with the appearance of a major DNase I hypersensitive site (DHS) at the hTERT promoter (14,34). Because Sp1 binding induced an asymmetric bend in DNA (35), we previously speculated that Sp1 binding might have caused DNA bending and nucleosome sliding, resulting in the formation of a DHS (14). In this study, we showed that Sp1 bound to the hTERT promoter not only in Tel ϩ cells, but also in Tel Ϫ cells without this DHS (14). As such, Sp1 binding was insufficient to initiate hTERT transcription (Fig. 8, top). Because E-box-binding proteins, Myc family proteins, and USF1/2 were more enriched on hTERT promoter in Tel ϩ cells than in Tel Ϫ cells, these factors might be directly involved in local de-repression/activation of the hTERT promoter in Tel ϩ cells. However, because these TFs were also abundant in Tel Ϫ cells (26), their binding to the hTERT promoter in Tel ϩ cells was likely a result of chromatin remodeling, possibly the establishment of DHS at hTERT promoter (Fig. 8, middle). The binding of these TFs may further recruit additional histone-modifying factors and remodel the ϩ1 nucleosome immediately downstream of the hTERT transcription start site, leading to its transcription in a permissive chromatin environment (Fig. 8, bottom). The current data, together with our earlier study on temporal repression of the hTERT gene during HL60 cell differentiation (36), suggested that local chromatin remodeling to create a nucleosome-free region (i.e. DHS) at the hTERT promoter, instead of TF binding, was a limiting step in hTERT transcription. This model is consistent with the data that transiently transfected hTERT promoters were active in both Tel ϩ and Tel Ϫ cells. Identification of the events that trigger chromatin remodeling is likely a key to understand hTERT transcriptional activation.
Another new finding in the current study was that Sp1 family TFs and E-box-binding proteins played different roles in hTERT transcription. Mutation of the downstream E-box significantly reduced hTERT promoter activity, but this reduction disappeared upon inhibition of HDACs by TSA. Therefore, it is likely that E-box-binding proteins recruit histone acetyltransferases to the promoter, de-repressing the hTERT promoter and thereby activating hTERT transcription. On the other hand, binding of Sp1 family TFs to GC-boxes was essential for hTERT promoter function. Mutations of GC-boxes or KD of Sp1 protein reduced promoter activity, but had no effect on the histone acetylation and the recruitment of other TFs. The functions of Sp1 at hTERT promoter might include recruiting Pol II containing pre-initiation complex (PIC). We attempted to determine the effect of Sp1 KD on Pol II recruitment, but were unable to obtain conclusive results due to low Pol II ChIP signals (Figs. 5B and 7A). This is not surprising because hTERT promoter is under strong HDAC-mediated repression in Tel ϩ cells, and a previous study showed that hTERT gene is tran-scribed at a low level even in cancer cells, which contained no more than a few copies of mRNA per cell (14,37).
Although induction of hTERT expression by TSA suggested that repression required histone deacetylation, the possibility of its indirect effects on the hTERT promoter could not be ignored. However, we and others previously showed that the TSA-induced hTERT expression and alteration of chromatin configuration were not blocked by the protein synthesis inhibitor cycloheximide (14,37), indicating that TSA-induced hTERT transcription did not require new protein synthesis and thus was likely a result of HDAC inhibition at the hTERT promoter. In the current study, we further tested several potential hTERT activators, c-Myc/Max, Sp1, and p53. As shown in Fig.  9, only the c-Myc mRNA level was elevated in Tel ϩ cells upon TSA treatment, but its protein level was not increased in these cells. Therefore, the regulation of these TFs could not account for the dramatic activation of hTERT transcription following HDAC inhibition by TSA.
It was reported that Sp1 and Sp3 were subjected to a variety of post-translational modifications, including phosphorylation, ubiquitination, acetylation, sumoylation, glycosylation, and hTERT Regulation by Sp1 Family Transcription Factors DECEMBER 11, 2015 • VOLUME 290 • NUMBER 50 poly(ADP-ribosyl)ation (reviewed in Ref. 38). Although some of these modifications may impact their DNA binding activities, other modifications may affect the abilities of Sp1 family TFs to interact with other TFs and chromatin factors. It remains to be determined whether Sp1 and Sp3 proteins are subjected to differential modifications in Tel ϩ and Tel Ϫ cells.
In this study, we used immortal Tel ϩ and Tel Ϫ cell lines that were derived from SV40 large T antigen-transformed human diploid fibroblasts (14). First, these cell lines were chosen for the current study because the RMBT method involved two rounds of clonal expansion of the host cells, the insertion of a retroviral acceptor site into a host cell chromosome, and Cremediated integration of the BAC reporters. This precluded the use of most primary human cells due to their limited lifespan in culture. Second, because Tel ϩ and Tel Ϫ cells are both SV40transformed IMR90 cells, their genetic backgrounds are more related than most cancer cell lines and especially suited for studying hTERT activation during immortalization of human fibroblasts following M2 crisis (39). One caveat is that inactivation of p53 and pRb pathways by T antigens alters cell cycle control and potentially impacts on hTERT regulation in these cells. However, mutations of these two tumor suppressor pathways occur in the majority of cancers, making these cells relevant models for studying telomerase activation during tumorigenesis. In addition, the chromatin state of the hTERT promoter is likely very complex. This study examined only a limited set of histone and epigenetic modifications. Further analyses of additional epigenetic marks, such H3K27 methylation and DNA methylation, may shed more light on how the hTERT promoter is regulated by Sp1 TFs in these cells.
In summary, we focused on the role of Sp1 family TFs in hTERT regulation in a relevant genomic and chromatin context in this study. Our data showed that the binding of these TFs to the GC-boxes was essential, but not sufficient, for hTERT acti-vation. hTERT transcription occurred only in a permissive chromatin environment, and Sp1 family TFs played no roles in this repression.