Role of Histone Deacetylase in the Expression of CTP:Phosphocholine Cytidylyltransferase α*

Histone acetylation plays an important role in chromatin remodeling and gene expression. The molecular mechanisms involved in cell-specific expression of CTP:phosphocholine cytidylyltransferase α (CTα) are not fully understood. In this study, we investigated whether or not histone deacetylation is involved in repression of CTα expression in quiescent C3H10T1/2 mouse embryo fibroblasts. We have examined the contributions of the Sp1 and E2F binding sites in the repression of CTα gene expression. Immunoprecipitation experiments showed that histone deacetylase 1 (HDAC1) and HDAC activity are associated with Sp1 in serum-starved cells or during serum stimulation. However, HDAC1 association with E2F was only detected in serum-starved cells. By chromatin immunoprecipitation assays, we detected both direct and indirect association of HDAC1 with the CTα promoter. Treatment with the HDAC inhibitor trichostatin A induced CTα expression. Our data suggest that HDAC1 plays a critical role in CTα repression and that Sp1 and E2F may serve as key targets for HDAC1-mediated CTα repression in fibroblasts.

Transition through the mammalian cell cycle is accompanied by the periodic expression of genes that participate in cell cycle-dependent processes. The transcription of several genes involved in DNA replication and mitosis, the DNA polymerase ␣, cyclin-dependent kinase 2, and thymidine kinase genes (1,2), increases as cells transit through G 1 and enter S phase. Cell cycle progression is also sensitive to membrane phosphatidylcholine (PC) 3 content. Chinese hamster ovary cells harboring a temperature-sensitive mutant of CTP:phosphocholine cytidylyltransferase ␣ (CT␣) do not synthesize PC at 40°C, and they accumulate in G 1 (3). Choline deprivation in WI-38 fibroblasts, L6 myoblasts, or C3H10T1/2 fibroblasts results in decreased PC synthesis and mass with cell cycle arrest in G 1 (4,5). These cells undergo apoptosis unless rescued by the addition of PC or lyso-PC (3). PC biosynthesis occurs in all nucleated mammalian cells via the Kennedy (CDP-choline) pathway in which CT catalyzes the regulated and rate-limiting step (6 -8). CT␣ is ubiquitously expressed in nucleated cells (9), and its expression is tightly regulated by a post-translational reversible association with membrane lipids, which are required for CT activity (10 -12). It was reported that the wave of PC synthesis that accompanies the G 0 -G 1 transition is regulated by changes in CT activity, membrane affinity, and intracellular distribution (13). CT␣ expression increases when the cells reach S phase (14). We demonstrated that this activation is driven by the binding of Sp1 to the B site present in the CT␣ proximal promoter (15,16). However, the molecular mechanism involved in G 0 repression of CT␣ has not been fully elucidated. A recent study has demonstrated that Net acts as a repressor of CT␣ transcription (17).
The modification of core histones is important in alteration of chromatin structure and gene transcription. Acetylation of core histones unpacks the condensed chromatin and renders the target DNA accessible to transcriptional machinery, hence contributing to gene expression (18). Conversely, deacetylation of core histones by histone deacetylase (HDAC) increases chromatin condensation and precludes binding between DNA and transcription factors, leading to transcriptional silencing (19,20). Previous studies have identified different mechanisms for transcriptional repression. One mechanism involves a direct interaction of retinoblastoma protein (Rb) with the E2F transactivation domain, resulting in masking of this domain and blocking its ability to stimulate transcription (21). Another mechanism is based on the ability of Rb to recruit chromatin remodeling proteins, such as HDACs, and assemble transcription repression complexes at E2F-regulated promoters (21)(22)(23)(24). Other studies have demonstrated that Sp1 recruits HDAC to a promoter (21,25).
We have examined the role of HDAC on repression of the CT␣ promoter. Our data support a dynamic model in which Sp1⅐HDAC and E2F⅐Rb⅐HDAC complexes cooperate to establish stable repression of CT␣ gene expression in quiescent cells.

EXPERIMENTAL PROCEDURES
Materials-The luciferase vector, pGL3-Basic, which contains the cDNA for Photinus pyraris luciferase, and the dual luciferase reporter assay system were obtained from Promega (Madison, WI). Lipofectamine Plus reagent, Dulbecco's modified Eagle's medium (DMEM), and fetal bovine serum were from Invitrogen. Anti-Rb, anti-HDAC1, anti-Sp1, and anti-E2F (Santa Cruz) are commercially available. ECL immunoblotting reagents were purchased from Amersham Biosciences, and trichostatin (TSA) was from Sigma.
Chromatin Immunoprecipitation (ChIP) Assay-C3H10T1/2 fibroblasts were grown for 24 h in DMEM plus 10% FBS. After serum deprivation for 72 h, cells were cross-linked by the addition of 1% formaldehyde for 5 min at 37°C. Cells were collected, lysed, and sonicated three times for 10 s each time at 2.5% with an ultrasonic processor XL from Heat Systems and treated as recommended by the manufacturer. For PCR analysis, we used two set of primers: ChIPCT1 (5Ј-TTgCCCTCgCCTCTACTCCTgCTC)/ ChIPCT2 (5Ј-CTCCCgCCCgCCCTCTTgTC) and CB001 (5Ј-CCACA-CATCCggAATTCC)/CB002 (5Ј-CACCgACTAgCgAAgTC). PCRs were performed using 2.5 l of template DNA, 1.5 mM MgCl 2 , and 20 pmol of each primer for 30 cycles of 94°C for 1 min, 58°C for 1 min, and 72°C 1 min. For the assay performed on plasmids LUC.C7delE2F and LUC.C7delSp1, we used a set of primers, ChIPCT1 and GLP2, chosen specifically to detect the episomal CT␣ promoter but not the chromosomal one.
Nuclear Extract Preparations-Total nuclear extracts of C3H10T1/2 cells grown to G 0 or S phase of the cell cycle were prepared as described by Andrews and Faller (27).
Immunoprecipitation-Nuclear extracts were prepared as described above from cells collected after 20 h of cell cycle induction. 200 g of nuclear extract protein were incubated with 5 g of polyclonal anti-Sp1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or anti-E2F (Santa Cruz Biotechnology) in a 1-ml final volume of immunoprecipitation buffer (1% (v/v) Triton X-100, 150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM EGTA, 0.2 mM phenylmethylsulfonyl fluoride, 0.5% (v/v) Nonidet P-40). The reaction was incubated for 1 h at 4°C and then incubated for 30 min with 50 l of 10% Protein A-Sepharose (Staphylococus aureus, Cowan strain), and the complex was washed three times with immunoprecipitation buffer. The pellet was resuspended in 30 l of concentrated electrophoresis sample buffer and boiled, and the supernatant was loaded onto a SDS-polyacrylamide gel and electrophoresed. Proteins were transferred to polyvinylidene difluoride membranes and probed with appropriate antibodies.
Treatment with Trichostatin-Stably transfected LUC.C8 (Ϫ201/ϩ38), LUC.C7 (Ϫ1268/ϩ38), LUC.C7⌬E2F, and LUC.C7⌬Sp1C C3H10T1/2 fibroblasts were grown under normal conditions for 24 h and synchronized in G 0 phase with DMEM containing 0.5% FBS for 72 h. The cell cycle was induced by adding DMEM containing 10% FBS, and TSA was added at the concentrations indicated. Samples were collected at 20 h (S phase) after induction of the cell cycle, and luciferase activity was measured.
HDAC Activity-Nuclear extracts were immunoprecipitated using anti-Sp1 or anti-E2F antibodies, and HDAC activity was measured following the manufacturer's instructions.

RESULTS
Is the G 0 Repression of CT␣ Regulated by HDAC?-We have previously shown that in C3H10T1/2 mouse embryo fibroblasts, the CT␣ gene is transcriptionally regulated by Sp1. Binding of Sp1 to the consensus Sp1 binding site B is essential for CT␣ activation during the S phase of the cell cycle (15). The interaction of the DNA-binding proteins was direct and shown to be strongly enhanced during the S phase due to Sp1 phosphorylation by cyclin-dependent kinase-2 (16). Since CT␣ transcription is low during the G 0 phase of the cell cycle, we investigated whether HDAC is involved in such repression. To address this question, we used a LUC.C7 (Ϫ1268/ϩ38) (26) stable line of C3H10T1/2 cells. Cells were synchronized for 48 h by serum deprivation (G 0 phase), and after this time, duplicate samples were incubated for 20 h with the indicated concentrations of TSA or dimethyl sulfoxide as a control. TSA is a nonreversible inhibitor of HDAC (28). We analyzed the luciferase/ protein ratio as a measure of CT␣ expression. When cells were treated with TSA, a 2.8-fold increase in CT␣ expression was observed, consistent with a mechanism whereby HDAC causes repression through chromatin remodeling (Fig. 1).
HDAC1 Binds the CT␣ Promoter-The association between HDAC1 and the CT␣ promoter was monitored by ChIP assay using chromatin prepared from C3H10T1/2 fibroblasts that were grown to different phases of the cell cycle (determined by [ 3 H]thymidine incorporation into DNA; data not shown). Target sequences were detected by PCR using two sets of primers as described under "Experimental Procedures" and in Fig. 2A. A portion of the CT␣ promoter was found in association with anti-HDAC1 immunoprecipitate in chromatin extracted from G 0and S-phase cells (Fig. 2B). As a control, chromatin was immunoprecipitated in the absence of specific antibodies, and in these samples, no significant quantity of CT␣ promoter was detected. Although the PCR is not quantitative, clearly, the amount of CT␣ promoter associated with HDAC1 decreased as cells progressed through the G 0 -S transition.
Interactions among HDAC1, Sp1, and E2F-HDAC1-Sp1 and HDAC1-E2F interactions were observed by immunoprecipitation experiments. Nuclear extracts obtained from resting and S-phase C3H10T1/2 mouse embryo fibroblasts were immunoprecipitated with anti-Sp1 or anti-E2F antibodies and the co-immunoprecipitated HDAC1 was detected by immunoblot. As Fig. 3 shows, Sp1 directly or indirectly binds HDAC1 in both phases of the cell cycle. Whereas HDAC1 was expressed at a higher level in replicating cells, the affinity to Sp1 seemed to be the same. In the case of immunoprecipitation of E2F, HDAC1 was only immunoprecipitated during the G 0 phase as expected from previous models where the progression of the cell cycle induced a release of HDAC from E2F (29).
HDAC activity associated with Sp1 and E2F was also analyzed in the same nuclear extracts as described above. In agreement with the immunoblot experiments, we observed that Sp1 is associated with significant HDAC activity in both G 0 and S phases, whereas E2F is associated with HDAC activity in G 0 phase (Fig. 4), with much less activity found in the S phase. HDAC activity was also measured in nuclear extracts treated with TSA to assess the effectiveness of the assay, and HDAC activity was inhibited by 50% (data not shown).
CT␣ Expression Is Regulated by E2F and Sp1-E2F regulates several families of genes whose products are required for cell progression. In G 0 or G 1 cells, E2F sites in the promoters of these genes are generally occupied by complexes that include E2F proteins and Rb or other members of the pocket family. Experimental data indicate that Rb-E2F repressive complexes function in association with HDAC (24,30,31) and that E2F appears to play a role in both repression and activation (32).
Investigation of the DNA sequence of the proximal promoter for the CT␣ gene revealed a putative E2F binding site at the Ϫ212 position upstream of the transcriptional start sites (see Fig. 2A). Using the luciferase reporter construct LUC.C7 (Ϫ1268/ϩ38), we constructed a plasmid (LUC.C7⌬E2F) harboring a point mutation in the putative E2F binding site (see "Experimental Procedures"). C3H10T1/2 fibroblasts were transfected with the wild type LUC.C7 or mutated LUC.C7⌬E2F plus pSV-␤-galactosidase as a control for transfection efficiency. Samples were collected at G 0 phase and S phase. We also analyzed the activity of a reporter construct LUC.C7⌬ Sp1(C) harboring a mutation in the previously identified Sp1 binding site C that was shown to be a negative regulator of CT␣ expression (15). The activity of the LUC.D3 (Ϫ52/ϩ38) construct, which excludes both the E2F and the Sp1 binding sites (see Fig. 2A), was also analyzed. As Fig. 5 shows, mutation of the E2F binding site abruptly decreased CT␣ expression in both phases analyzed. Like LUC.D3, the mutation in E2F not only decreased the basal activity in G 0 but also affected S-phase activation. In contrast, mutation of the Sp1 binding site C resulted in enhanced expression of the luciferase reporter (Fig. 5). From these results and a previous report (32), E2F appears to play a role in both repression and activation.
E2F, Rb, and HDAC Associate with the CT␣ Promoter-To determine whether E2F plays a role in CT␣ regulation during G 0 phase, we utilized a ChIP assay to investigate the association of E2F, Rb, and HDAC1 with the CT␣ promoter. Fig. 6 shows that significant quantities of the CT␣ promoter associate with each of the proteins assayed,     whereas no significant amount was detected with a nonspecific antibody, anti-IgG. Although the PCR used was not quantitative, there were differences in the intensities of the bands that may reflect a direct or indirect interaction between each protein with DNA (33).
Is HDAC Recruited to the CT␣ Promoter by either Sp1 or an E2F-Rb Complex?-Luciferase reporter constructs LUC.C7⌬E2F and LUC. C7⌬Sp1(C), with mutations in the E2F and Sp1(C) binding sites, respectively, were transfected into C3H10T1/2 fibroblasts. After synchronization in G 0 phase, the cells were analyzed by ChIP assay using antibodies against Sp1, Rb, HDAC, and E2F (Fig. 7). The fragment of the CT␣ promoter harboring the corresponding mutation in each case was analyzed by designing primers ChIPCT1 and GLP2 (see Fig. 2A), which specifically recognize the plasmids. As a negative control, we used a sample incubated with a nonrelated antibody anti-IgG. Mutation in the E2F binding site does not affect the association between the CT␣ promoter with Sp1, HDAC1, or E2F-Rb. In contrast, a mutation in Sp1(C) decreases the association with HDAC1 but has no effect on E2F and Rb association. Because the primers amplified a fragment of the CT␣ promoter that includes three Sp1 binding sites, we did not evaluate the effect of a mutation in the Sp1 binding site C on Sp1 association. Thus, the interaction of HDAC1 with the CT␣ promoter depends at least in part on the Sp1(C) binding site.
We also analyzed whether or not inhibition of HDAC activity by TSA would alter luciferase expression in LUC.C8 (Ϫ201/ϩ38), LUC.C7⌬E2F, and LUC.C7⌬Sp1(C) stable cell lines. Cells were synchronized for 48 h by serum deprivation, after which duplicate samples were incubated with the indicated concentration of TSA or dimethyl sulfoxide as a control. We analyzed the luciferase/protein ratio as a measure of CT␣ expression. Mutation in the Sp1(C) binding site did not abolish the induction of luciferase activity caused by TSA (Fig. 8A). However, mutation of E2F decreased the induction by TSA, but we still observed a 1.7-fold induction of luciferase activity after treatment. In agreement with our previous results, analysis of LUC.C8, which excludes the E2F binding site, revealed an increase in CT␣ expression in both phases of the cell cycle after TSA treatment (Fig. 8B).

DISCUSSION
We provide evidence for the first time that HDAC1 is involved in the repression of murine CT␣ expression in quiescent cells. The repression of CT␣ expression in C3H10T1/2 mouse embryo fibroblasts arrested in G 0 phase could be partially mitigated by inhibition of HDAC activity. TSA treatment of the LUC.C7 cell line (Fig. 1), but not in cells transiently transfected with the same construct (data not shown), restores FIGURE 7. Mutation of the Sp1(C), but not the E2F, site decreases the binding of HDAC1 to the CT␣ promoter. ChIP assays were performed with anti-Rb, -E2F, or -HDAC1 antibodies using cells transfected with the indicated mutated LUC.C7 plasmids. A nonrelated antibody, anti-IgG, was used as a control. The pairs of primers (ChiPCT1 and GLP2) used to amplify the DNA isolated from the ChIP assay were designed to specifically detect the episomal DNA (see "Experimental Procedures"). MW, molecular weight markers. FIGURE 8. Differential sensitivity of the E2F and Sp1 binding sites on the CT␣ promoter to TSA. A, stable cell lines expressing the nonmutated LUC.C7 luciferase constructs or constructs with mutations in the designated promoter binding sites were synchronized to G 0 and then incubated with the indicated concentration of TSA. Luciferase activity/protein ratio between treated and untreated cells was used to calculate the induction caused by TSA treatment. Values represent the mean Ϯ S.D. of three independent experiments. B, the LUC.C8 stable cell line was synchronized, and the cell cycle was induced by adding 10% FBS. Cells from G 0 and S phase were treated with the indicated TSA concentration for 20 h, and the luciferase activity/protein ratio was calculated. Values represent the mean Ϯ S.D. of three independent experiments. CT␣ expression. After the episomal CT␣ promoter, which is methylation-free, is transfected into fibroblasts, it will form minichromatin but not in a repressive chromatin structure as does the endogenous CT␣ promoter (34). We also demonstrated by ChIP assays that HDAC associates in vivo with the CT␣ proximal promoter. Based on this result, it would be logical to ask how CT␣ is repressed in quiescent C3H10T1/2 fibroblasts by HDAC1.
Sp1 Recruits HDAC to the CT␣ Promoter and Regulates Its Expression in Quiescent Cells-CT␣ expression during cell cycle progression is mainly regulated by Sp1. Three Sp1 binding sites have been identified in the CT␣ proximal promoter (see Fig. 2A); Sp1 site B drives CT␣ induction in G 1 -S phase, whereas site A acts as a transcriptional enhancer. Conversely, the Sp1 binding site C appears to be a repressor of CT␣ expression (15). HDAC1 was readily recovered in immunoprecipitates of Sp1, and this interaction persisted in serum-starved cells, supporting the notion that Sp1 plays a role in repression of CT␣. Doetzlhofer et al. (25) suggest that Sp1 can serve as a target for HDAC1-mediated transcriptional repression, although Sp1 is usually known as a positive transcriptional factor. Moreover, in different cell lines, gene repression is regulated by the association of HDAC1 or -2 and Sp1 or Sp3 (35,36). Given that Sp1 associates with HDAC1 and Sp1 binds and regulates CT␣ expression, it is likely that Sp1 may serve as a target for HDAC1involved CT␣ suppression. To address this hypothesis, we analyzed by ChIP assay the ability of a mutated Sp1 binding site to recruit HDAC1 (Fig. 7). A limited quantity of CT␣ promoter harboring a mutation in the Sp1 binding site C was immunoprecipitated with anti-HDAC antibody. However, this mutation did not abolish the effect of the HDACspecific inhibitor TSA on CT␣ expression; the LUC.C7⌬Sp1(C) stable cell line showed a 3.5-fold increase in luciferase reporter activity after treatment (Fig. 8A). Taken together, these results suggest that HDACmediated regulation of CT␣ expression in quiescent cells is in part due to a Sp1⅐HDAC complex. However, because we still observed regulation in the absence of Sp1 binding site C, we propose that additional protein(s) could be involved in HDAC recruitment to the promoter.
E2F Cooperates with Sp1 in Recruiting HDAC-Binding sites for E2F are found in the promoters of genes whose expression occurs at G 1 /S transition, and repression of promoters containing E2F-binding sites has been shown to involve binding of E2F/Rb pocket protein complexes to these sites (37). One of the mechanisms of transcriptional repression is based on the ability of Rb to recruit chromatin remodeling proteins, such as HDACs, and assemble transcriptional repression complexes at E2F-regulated promoters (22)(23)(24). With this information in mind, we evaluated the presence of E2F⅐HDAC complexes by immunoblot, using nuclear extracts obtained during G 0 and S phases (Fig. 3). We only detected the complex in extracts obtained in G 0 phase, in agreement with a previous report that Rb is released after phosphorylation by cyclin D-cyclin-dependent kinase 4 during G 1 -S phase (38). To further investigate the role of E2F in CT␣ regulation, we analyzed the activity of a reporter construct having a point mutation in the putative E2F binding site found in the Ϫ210 position with respect to the transcriptional start point (Fig. 5). Surprisingly, this mutation decreased CT␣ expression to a similar extent as LUC.D3 (Ϫ54/ϩ38), in which the promoter is almost entirely deleted. This result indicates that, as with other cell cycle-dependent genes, CT␣ expression is regulated by E2F.
E2F, Rb, and HDAC1 were detected on the CT␣ promoter by ChIP assay (Fig. 6). The different intensities of the PCR band may reflect the type of interaction between the factors and the DNA, with E2F acting as the anchor protein to recruit Rb⅐HDAC. The role of E2F in HDAC-dependent repression of CT␣ was examined by measuring the ability of a mutant E2F binding site to allow for the CT␣ promoter to be immuno-precipitated by various antibodies (Fig. 7). Because mutation of the E2F binding site did not affect the binding of E2F, HDAC, and Rb to the promoter, we can speculate that these and probably other proteins such as Sp1 form a multimolecular complex that regulates CT␣ expression (Fig. 9). The cooperation between E2F and Sp1 is supported by the observation that members of these two families interact with one another (39 -41). In concurrence with this result, a stable cell line with a mutated E2F binding site is still sensitive to TSA, showing a 1.7-fold increase in luciferase activity. (Fig. 8A). On the basis of these results, we propose that an interaction between E2F and Sp1 or other proteins not yet identified as part of the complex cooperates in recruiting HDAC to the CT␣ promoter.
In conclusion, this work describes a model for the control of the murine CT␣ promoter in quiescent cells (Fig. 9). During growth arrest, E2F binds Rb⅐HDAC1, and the nearby Sp1 recruits HDAC1. Thus, both E2F-Rb and Sp1 bind HDAC1, thereby causing repression of the expression of the CT␣ promoter and, thereafter, a decrease in PC biosynthesis, which should impact on membrane biogenesis. The model is in agreement with reports on regulation of the dihydrofolate reductase promoter (42,43) and thymidine kinase promoter (25), which have shown that Sp1 in addition to E2F plays an active role in the growth control of transcription. Further experiments will be necessary to clarify the type of interaction between each of these proteins.