Acetylation-Dependent Interaction of SATB1 and CtBP1 Mediates Transcriptional Repression by SATB1

Abstract

Special AT-rich binding protein 1 (SATB1) acts as a global regulator of gene expression by recruiting various corepressor or coactivator complexes, thereby establishing a unique chromatin structure at its genomic targets in a context-dependent manner. Although SATB1 acts predominantly as a repressor via recruitment of histone deacetylase 1 (HDAC1) complexes, the precise mechanism of global repression is not clear. Here we report that SATB1 and C-terminal binding protein 1 (CtBP1) form a repressor complex in vivo. The interaction occurs via the CtBP1 interaction consensus motif PVPLS within the PDZ-like domain of SATB1. The acetylation of SATB1 upon LiCl and ionomycin treatments disrupts its association with CtBP1, resulting in enhanced target gene expression. Chromatin immunoprecipitation analysis indicated that the occupancy of CtBP1 and HDAC1 is gradually decreased and the occupancy of PCAF is elevated at the SATB1 binding sites within the human interleukin-2 and mouse c-Myc promoters. Moreover, gene expression profiling studies using cells in which expression of SATB1 and CtBP1 was silenced indicated commonly targeted genes that may be coordinately repressed by the SATB1-CtBP1 complex. Collectively, these results provide a mechanistic insight into the role of SATB1-CtBP1 interaction in the repression and derepression of SATB1 target genes during Wnt signaling in T cells.

The T-cell-enriched transcription factor special AT-rich binding protein 1 (SATB1) regulates the spatiotemporal expression of a large number of genes involved in T-cell development (1). SATB1 participates in the maintenance of chromatin architecture in a cell-type-specific manner by organizing higher-order chromatin loops into distinct domains via periodic anchoring of non-base-pairing regions to the nuclear matrix (9, 10, 25). In primary thymocytes, SATB1 exhibits a cage-like network distribution circumscribing heterochromatin domains and regulates distant genes in a coordinated manner (10). Implications of SATB1 domains in its functional regulation impart unique properties to this chromatin organizer. The C-terminal homeodomain (HD) acts in concert with the Cut repeat-containing domain (CD) and directs high-affinity binding of SATB1 to its targets in a sequence-specific manner (35). The N-terminal PDZ-like domain aids in the formation of a homodimer that is essential for the DNA binding activity of SATB1 (15, 35). The N-terminal region harboring the PDZ-like domain is a putative interface for its interaction with various cellular and viral proteins (25-27). SATB1 regulates gene expression in two distinct modes. Primary regulation is by specific binding of SATB1 to promoters and upstream regions, thereby directly influencing the promoter activity. SATB1 is known to directly regulate a number of genes, including those encoding globin, interleukin-2 (IL-2), and IL-2 receptor α (IL-2Rα), by recruiting either CBP (48) or histone deacetylase 1 (HDAC1) (26, 27). Secondly, context-specific regulation of SATB1 stems from its unique ability to bind to matrix attachment regions (MARs), thereby regulating a large number of genes in a coordinated manner by acting as a docking site for several chromatin remodeling complexes (50). Functional interaction of SATB1 and PML aids in the organization of the major histocompatibility complex (MHC) class I locus into a distinct higher-order chromatin loop architecture, thereby affecting the expression profiles of a subset of MHC class I genes (25). Recent evidence suggests that SATB1 regulates >10% of genes (27), presumably due to its unique ability to bind a large number of regulatory sequences and subsequent recruitment of its various interacting partners. Additionally, posttranslational modifications of SATB1 add another level of complexity to its role as a global regulator of gene expression. Phosphorylation and acetylation have contrasting effects on the ability of SATB1 to regulate transcription, as phosphorylated SATB1 associates with corepressors and acetylated SATB1 associates with coactivators (27).

Various corepressors target components of the basal transcription machinery and affect transcription by modulating chromatin structure. C-terminal binding protein (CtBP) is a nuclear protein that is phosphorylated in a cell cycle-dependent manner (6). CtBP was identified as a protein that specifically binds to the C-terminal region of the human adenovirus E1A proteins, and it is implicated in negative regulation of oncogenic transformation (6). The protein originally identified as the E1A interacting protein was termed CtBP1, and subsequently, a highly homologous human protein was identified by analysis of expressed sequence tag data bank sequences (23). CtBP represses transcription in an HDAC-dependent or -independent manner. Several transcription factors which recruit CtBP1 can also activate transcription in a context-dependent manner (11); however, the mechanism for such context-dependent effects is not clear. The HDAC-independent repressor activity of CtBP might be a manifestation of the recruitment of PcG complexes (12, 41). CtBP1 interacts with the Ku 70 subunit of DNA protein kinase (39, 47). CtBP1 has also been shown to bind to the carboxyl-terminal region of the breast cancer-associated tumor suppressor and transcription factor BRCA1 and may be involved in regulation of the p21 Waf1 and GADD45 genes by BRCA1 (30, 31). Human CtBP1 acts as a corepressor for the ZEB transcription factor that is involved in the regulation of lymphocytes and muscle differentiation (34). In addition to the repressor activity, a context-dependent weak transcriptional activation function of Drosophila CtBP has been reported (33). All of these cellular transcription factors and human adenovirus E1A proteins contain a conserved Pro-X-Asp-Leu-Ser (PXDLS) pentapeptide motif that is necessary for CtBP1 interaction (11, 40). CtBP also binds HDACs containing the PXDLS motif (11). CtBP1 acts as a homodimer that recognizes the consensus motif PXDLS in DNA binding proteins (45). Due to these characteristic properties, CtBP1 is likely to act as a protein bridge between HDACs and DNA-bound factors (45). CtBP1 also contains a PDZ domain binding motif (DLX) that is conserved in all vertebrates (11, 36). PDZ domains are protein interaction modules that are implicated in signal transduction and found to be present in membrane-associated proteins (21). A number of CtBP1-interacting DNA binding proteins possess a PXDLS-like motif (11). Many of these transcription factors undergo acetylation-mediated dissociation from CtBP1 (11, 47, 51).

SATB1 is the only chromatin organizer that harbors a PDZ-like domain and may act as a docking site for various PDZ domain interacting proteins. Since the DLX motif of CtBP1 is known to target PDZ domains, we investigated the association between SATB1 and CtBP1 and its consequences on the transcription regulatory function of SATB1. Here we demonstrate that SATB1 and CtBP1 interact in vitro and in vivo. The PDZ-like domain in SATB1 harbors the CtBP1 interaction consensus motif PVPLS. Interestingly, acetylation of the PDZ-like domain by PCAF disrupts its association with CtBP1. Furthermore, the PVPLS motif of SATB1 in conjunction with an adjoining lysine at amino acid (aa) 136 determines the interaction efficiency. The acetylation of K136 upon LiCl and ionomycin treatments disrupts the association of CtBP1 with SATB1, resulting in enhanced target gene expression and thereby demonstrating a different mode of regulation of gene expression by SATB1 depending upon its acetylation status.

MATERIALS AND METHODS

Cell culture and treatment.

Jurkat T cells and mouse thymocytes isolated from BALB/c mice were grown in RPMI 1640 supplemented with 10% fetal calf serum. Jurkat T cells (1 × 10⁶ cells/ml) were treated with 0.2 μM ionomycin, 10 mM KCl, or 10 mM LiCl for 0 to 12 h. Thymi were removed from 20-day-old mice, and thymocytes were suspended in RPMI 1640 medium. Mouse thymocytes (2 × 10⁶ cells/ml) were treated with 5 mM LiCl for 0 to 6 h. HEK-293 cells were grown in Dulbecco's modified Eagle's medium with 10% fetal calf serum and antibiotic. For reporter assays, fresh medium containing 10 mM LiCl or 10 mM KCl was added at 36 h posttransfection, and after further incubation for 6 h, the cells were processed.

Antibodies and reagents.

Antibodies to HDAC1, PCAF, CBP, p300, and CtBP1 were purchased from Santa Cruz Biotechnology. Anti-FLAG was procured from Sigma Chemical Company. Anti-acetyl-histone H3K9 and purified active catalytic domain of PCAF (14-309) were obtained from Upstate. Anti-SATB1 and anti-acetylated SATB1 have been described previously (27). CtBP1 small interfering RNA (siRNA) and SATB1 siRNA were obtained from Santa Cruz Biotechnology.

Plasmids and protein expression and purification.

T7CtBP1 and GST-CtBP1 were kindly gifted by G. Chinnadurai and Bonnie L. Firestein, respectively. Various deletion mutant DNA constructs of SATB1 used for in vitro transcription and translation have been reported before (15). K136R-SATB1 and P141A and PL-to-AS mutants in the CtBP1 interaction motif (PVPLS) of SATB1 were constructed using a QuikChange II site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. The 3XFLAG-CMV10-SATB1 construct, containing full-length human SATB1 cDNA, was used as a template for mutagenesis. The sequences of the oligonucleotides used for site-directed mutagenesis are listed in Table S1 in the supplemental material. Glutathione S-transferase (GST), GST-CtBP1, and GST-1-254, GST-1-204, GST-90-204, and GST-255-763 SATB1 were expressed in the XL1-Blue strain of Escherichia coli (Stratagene) and purified using glutathione-Sepharose affinity columns (GE Healthcare). A six-His-tagged N-terminal PDZ-containing region of SATB1 was expressed in the BL21(DE) strain of E. coli (Novagen) and purified using Ni-nitrilotriacetic acid fast-flow resin (Qiagen). The in vitro translation of various proteins was carried out using a TNT coupled in vitro transcription and translation kit (Promega) according to the manufacturer's instructions.

In vitro GST pull-down assays.

For pull-down assays using nuclear lysates, 200 μl glutathione bead-bound GST and GST-1-254 (PDZ) were incubated with 500 μg of HEK-293 nuclear lysate at 4°C for 8 h. The beads were washed four times with wash buffer (PBS containing 0.1% Triton X-100 and protease inhibitor cocktail). The interacting proteins were eluted with increasing concentrations of NaCl in wash buffer (0.3 to 1.0 M). Proteins were resolved by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (12% SDS-PAGE) and transferred to a polyvinyl difluoride (PVDF) membrane. The proteins transferred to the PVDF membrane were detected with anti-CtBP1. For pull-down assays using ³⁵S-labeled in vitro-translated proteins, 50-μl samples of glutathione bead-bound GST and GST-1-254, GST-1-204, GST-90-204, and GST-255-763 SATB1 were incubated with 50 μl of in vitro-translated ³⁵S-labeled CtBP1 at 4°C for 4 h, with rotation, in binding buffer (PBS containing 0.1% Triton X-100 and protease inhibitor cocktail [Roche]). Beads were washed six times with the binding buffer and eluted with 2× SDS-PAGE loading buffer. Proteins were resolved by 12% SDS-PAGE, dried under vacuum, and exposed to X-ray film for 24 h. For monitoring of acetylation-dependent interactions, glutathione bead-bound acetylated or unacetylated GST and GST-CtBP1 were incubated with in vitro-translated ³⁵S-labeled CtBP1. For delineation of the minimal region of SATB1 required for interaction with CtBP1, the full length (aa 1 to 763) or various truncations of SATB1 (113-763, 90-160:CD+HD, 215-763, CD+HD, and 1-204) were translated in vitro and incubated separately with GST or GST-CtBP1.

Coimmunoprecipitation.

Coimmunoprecipitation was carried out as described previously (27). Briefly, 100 μg of protein in nuclear extract was diluted by adding 2 volumes of IP-150 buffer and then precleared with rabbit immunoglobulin G (IgG; Sigma) and protein A/G Plus beads (Pierce). Precleared extract was then incubated with polyclonal anti-SATB1, rabbit IgG, mouse monoclonal anti-CtBP1, or mouse IgG. Protein-antibody complexes were analyzed by immunoblot analysis, using anti-SATB1 and anti-CtBP1. For delineation of the CtBP-interacting domain of SATB1, FLAG-tagged full-length SATB1 and N-terminal PDZ domain and C-terminal CD+HD mutants were overexpressed along with T7CtBP1 in HEK-293 cells. Complexes were immunoprecipitated with anti-FLAG antibody and detected by anti-CtBP1.

Immunofluorescence staining.

For immunostaining of thymocytes, approximately 1 × 10⁵ thymocytes in suspension were stained with anti-SATB1 and anti-CtBP1 after fixation and permeabilization, followed by incubation with conjugated secondary antibodies. The secondary antibodies used to detect staining of SATB1 and CtBP1 were Alexa Fluor 594 and Oregon green 488 (Invitrogen), respectively. DNA counterstaining was performed using DAPI (4′,6-diamidino-2-phenylindole). The cells were spun onto glass slides by use of Cytospin (Thermo Shandon), and after being mounted with coverslips, cells were visualized under an upright fluorescence microscope (AxioImager Z1; Carl Zeiss) and digital images were enhanced using the Apotome module (Carl Zeiss). z stacks of images were collected as 0.2-μm optical slices and were further processed offline using deconvolution software by applying the inverse theorem (Carl Zeiss).

ChIPs and real-time PCRs.

Chromatin immunoprecipitation assays (ChIPs) and real-time PCRs were performed essentially as described previously (26). Chromatin immunoprecipitated with anti-SATB1, anti-CtBP1, anti-HDAC1, anti-PCAF, and anti-H3K9Ac was amplified with the P1 IL-2-specific primer pair for human Jurkat T cells (26, 27) and the c-Myc SBS forward (CTAGGGTCTCTGCAGGCTCCCCAGATCTGG) and reverse (GTTGACTGGGCACATTCTTTCCAGAACGAC) primer set for amplification of the c-Myc SATB1 binding site (SBS) region from mouse thymocytes (9). Real-time quantitative PCR was carried out using an iCycler real-time PCR machine (Bio-Rad) and Sybr green premix (Bio-Rad). The changes in threshold cycle (C_T) values were calculated as follows: ΔC_T = C_{T target} − C_{T input}. Data are presented as values.

EMSA.

Electrophoretic mobility shift assays (EMSAs) were performed as described previously (35). Briefly, binding reactions were performed in a 10-μl total volume containing 1× EMSA buffer [10 mM HEPES (pH 7.9), 1 mM dithiothreitol, 2.5 mM MgCl₂, 10% glycerol, 1 μg of double-stranded poly(dI-dC), 10 μg of purified bovine serum albumin] and 100 ng of purified six-His-SATB1. Purified GST-CtBP1 alone (200 ng) was included as a control. To monitor the effect of CtBP1 interaction on the DNA binding activity of SATB1, 400 ng of GST-CtBP1 was included in the binding reaction mix. To perform an antibody-mediated supershift of the complex, 1 μg of anti-CtBP1 was added to the reaction mixture after incubation of the probe with SATB1 and CtBP1, and the same amount of mouse IgG was used as control.

Repression assay.

The PDZ and CD+HD domains, cloned separately into the pBIND vector (Gal4 DNA binding domain [DBD]-containing plasmid vector; Promega), were used to transfect cells along with the reporter vector in HEK-293 cells. Briefly, HEK-293 cells were seeded at 0.5 × 10⁶ cells per well in a six-well plate 24 h before transfection. Cells were transfected with pG5 Luc reporter vector (Promega), using Lipofectamine 2000 reagent (Invitrogen), along with either pBIND-PDZ, pBIND-CD+HD, or pBIND alone as a control. To monitor the effects of CtBP1, HDAC1, and PCAF on the repression function of the PDZ-like domain, the pBIND-PDZ construct was cotransfected with CtBP1, HDAC1, and PCAF constructs. The DNA amount was maintained at a constant level for each transfection. The cells were harvested at 48 h posttransfection for luciferase assay.

In vitro acetylation assay.

Purified histones, GST-PDZ, GST, and six-His-PDZ proteins were used for in vitro acetylation by PCAF as described previously (27). Briefly, equal amounts of proteins were incubated with PCAF (Upstate), using [³H]acetyl coenzyme A ([³H]acetyl-CoA) (ICN) as the donor of an acetyl group, per the procedure provided by the manufacturer. In parallel, these proteins were also acetylated using cold acetyl-CoA (ICN) as the donor.

Luciferase reporter assay.

Luciferase assays were performed using Luc Lite reagent (Perkin-Elmer), and luciferase activity was measured using Top Count (Packard). The IgH MAR-Luc reporter and IL-2 promoter-luciferase reporter constructs (26, 27) were used to score the effects of SATB1 and its complex with different corepressors/activators. All transfections were carried out using various combinations of constructs, and the final amount of DNA was normalized by including pCDNA3.1 DNA. HEK-293 cells were seeded at 0.2 × 10⁶ cells per well and transfected using Lipofectamine 2000 reagent (Invitrogen). Forty-eight hours after transfection, the culture medium was removed, and cells were washed with PBS and processed further. Cells were resuspended in 100 μl of reporter lysis buffer and stored frozen at −70°C. After being freeze-thawed twice, the lysate was clarified by spinning at 10,000 rpm for 15 min. For accurate quantitation of luciferase activity, equal amounts (100 μg) of the protein were assayed. Protein concentrations in the lysates were measured using Bradford reagent (Bio-Rad). The expression values were plotted after setting the luciferase activity of the control to 1.

RNA interference and RNA integrity.

HEK-293 cells were transfected with siSATB1 and siCtBP1 by use of siIMPORTER transfection reagent (Upstate Technologies) per the manufacturer's instructions. Subsequently, control and transfected cells were harvested, and RNAs were extracted using Tri reagent (Sigma) and used for reverse transcription-PCR (RT-PCR) analysis. The integrity of the RNA was analyzed by electrophoresis, using Bioanalyzer (Agilent), before proceeding to microarray hybridization. The RNA integrity numbers of samples were >9.0.

Microarray hybridizations and data analysis.

Cells (5 × 10⁶) were washed with PBS and harvested in Trizol (Invitrogen, Carlsbad, CA), and total RNA was isolated according to the manufacturer's protocol. Two micrograms of RNA was used to synthesize cDNA, using biotinylated and fluorescein-labeled deoxynucleoside triphosphates. The labeled cDNA was used to hybridize with human cDNA microarrays containing 8,000 single spots (University Health Network Microarray Center, Ontario Cancer Institute, Canada) according to the manufacturer's protocols. Posthybridization, the signal was amplified using a Micromax TSA signal amplification kit (Perkin-Elmer). Fluorescent images of hybridized microarrays were obtained using a ScanArray Express microarray scanner (Perkin-Elmer). The data were analyzed using GeneSpring GX software (Agilent Technologies).

RESULTS

NH₂-terminal PDZ-containing domain of SATB1 is required for repression via HDAC1 and CtBP1 interaction.

The various protein interactions of SATB1 involve its N-terminal PDZ-like domain, suggesting a regulatory role of the PDZ domain in mediating transcription of SATB1 target genes (25-27). To delineate the transcriptional function of the PDZ-like domain of SATB1, we designed an experiment to dissect the transcriptional functions of the two major domains in SATB1. The N-terminal PDZ-containing dimerization domain (aa 1 to 254) and the C-terminal CD+HD region (aa 255 to 763) were expressed as fusions with the Gal4 DBD. The heterologous luciferase reporter vector contains five Gal4-responsive elements next to the adenovirus MLP promoter (pG5 luc). Gal4 DBD fusion proteins bind very specifically to these elements and bring about either repression or activation of transcription, depending on the nature of the fusion protein. We assayed the activities of these two domains in regulating transcription of the heterologous MLP promoter in HEK-293 cells. The PDZ domain and CD+HD fused with the Gal4 DBD were transfected into HEK-293 cells along with the luciferase reporter vector. The PDZ domain fusion protein but not the CD+HD fusion protein repressed the transcription mediated by adenovirus MLP in a dose-dependent manner (Fig. 1A, compare bars 2 and 3 with bars 4 and 5). Thus, the N-terminal PDZ-containing region of SATB1 is essential for SATB1-mediated transcription repression. The HDAC1 complex has been found to be associated with this domain (26) and could be the component which mediates this repression activity.

FIG. 1.

PDZ-like domain is necessary and sufficient for SATB1-mediated repression via interaction with CtBP1 and HDAC1. (A) Repression assays were performed as described in Materials and Methods to examine the transcriptional activities of the N-terminal PDZ-like domain and the C-terminal CD+HD domain in a dose-dependent manner. Each transfection unit corresponds to 0.5 μg DNA (1×). The first bar indicates the luciferase activity of the control pBIND vector (Gal4-DBD) alone, which was arbitrarily set to 1. Error bars represent standard deviations calculated from triplicates. (B) Schematic representation of GST and GST-PDZ fusion proteins used for affinity chromatography. (C) Nuclear extract from HEK-293 cells (input; lane 1) was passed through GST and GST-PDZ affinity columns. Bound proteins were eluted with phosphate buffer containing 0.3 M (lanes 2 and 5), 0.6 M (lanes 3 and 6), or 1 M (lanes 4 and 7) NaCl. Eluted proteins were resolved by SDS-PAGE, transferred to a PVDF membrane, and immunoblotted with anti-CtBP1. (D) Coomassie brilliant blue-stained polyacrylamide gel showing GST (lanes 2 to 4) and GST-PDZ (lanes 5 to 7) eluted from glutathione-Sepharose beads. (E) Repression assays were performed as described in Materials and Methods. The indicated amounts (in μg) of CtBP1, HDAC1, and PCAF expression constructs were transfected into HEK-293 cells along with the pG5-luc reporter construct and with constructs expressing either the Gal4 DBD-PDZ fusion protein or the Gal4 DBD-CD+HD fusion protein. Bars 1 and 2 depict the luciferase reporter activities in the presence of the respective fusion proteins only.

Since the N-terminal PDZ-like domain of SATB1 alone is sufficient to act as a repressor, we reasoned that this repression must be mediated via interactions of the PDZ-like domain with repressor complexes. We performed pull-down assays using the PDZ-like domain to fish out interacting proteins from the pool of cellular proteins from HEK-293 cells. Nuclear extracts from HEK-293 cells were incubated with immobilized glutathione-bound GST-PDZ fusion protein or with GST alone. GST-PDZ and GST-associated proteins were eluted with increasing concentrations of NaCl (0.3 M to 1.0 M) (Fig. 1C). Western blots of pulled-down proteins revealed the presence of CtBP1 in the fractions eluted from GST-PDZ but not in the fractions containing GST alone (Fig. 1C). The interaction was strong, and protein remained bound in up to a 1 M salt concentration, as revealed by the elution profile (Fig. 1C). CtBP1 is a well-established non-DNA-binding corepressor protein which interacts with several DNA-binding transcription factors and represses transcription of target genes in an HDAC1-dependent and/or -independent manner (11). The PDZ domain of SATB1 acts as a repressor domain, so we wondered whether CtBP1-mediated interaction with the PDZ-like domain mediates repression in a dose-dependent manner. We performed a mammalian one-hybrid assay using the Gal4 DBD-PDZ fusion construct in combination with CtBP1 and HDAC1. We coexpressed CtBP1 along with the Gal4 DBD-PDZ fusion protein and monitored the effect of this interaction on MLP promoter-mediated transcription of the luciferase reporter gene (Fig. 1E). We also tested the effect of coexpression of HDAC1 or PCAF alone or in combination with CtBP1. HDAC1 and CtBP1, individually and in combination, could repress (Fig. 1E, bars 3, 5, 7, 9, and 11) and PCAF could activate the SV40 promoter activity via interaction with the PDZ-like domain of SATB1 (Fig. 1E, bar 13). There was no significant effect of coexpression of CtBP1, HDAC1, and PCAF alone or in combination when the Gal4 DBD-CD+HD fusion protein was used (bars 4, 6, 8, 10, 12, and 14). This result indicated that CtBP1 can bring about repression via association with the PDZ-like domain of SATB1 and that PCAF can abolish the repressor function of the PDZ-like domain of SATB1 in a heterologous system. Collectively, these results indicated that the PDZ-like domain is necessary and sufficient for SATB1-mediated repression via association with CtBP1 and HDAC1.

CtBP1 associates with SATB1 via a PVPLS motif contained in the PDZ-like domain.

To narrow down the minimal region of SATB1 required for the interaction with CtBP1, we prepared GST fusions with aa 1 to 254, 1 to 204, 90 to 204, and 255 to 763 truncations. We used GST and the above-mentioned GST fusion proteins for in vitro pulldown of ³⁵S-labeled CtBP1. CtBP1 was specifically pulled down with GST-1-254, GST-1-204, and GST-90-204 but not with GST or GST-255-763, indicating the importance of the aa 90 to 204 region of SATB1 for interaction with CtBP1 (Fig. 2A [summarized in Fig. 2B]). To further narrow down the CtBP1 interaction region within the PDZ-like domain of SATB1, we synthesized several ³⁵S-labeled deletion mutants of SATB1 and incubated them separately with GST and GST-CtBP1. The different deletion mutants of SATB1 used are represented schematically in Fig. 2F. GST-CtBP1 associated with and therefore could pull down the full-length (aa 1 to 763), 113-763, 90-160-CD+HD, 90-183-CD+HD, and 1-204 constructs but could not pull down the 215-763 and CD+HD (aa 330 to 763) constructs (Fig. 2D [summarized in Fig. 2F]). Thus, the aa 113 to 160 region of SATB1 is sufficient for interaction with CtBP1. Since it is known that CtBP1 interacts with a large number of proteins harboring a PXDLS-like motif (11, 47), we analyzed the N-terminal PDZ-like domain of SATB1 for the presence of a putative PXDLS-like motif. SATB1 possesses three putative CtBP1-binding PXDLS-like motifs within its PDZ-like domain (Fig. 2G). Of these, only the ¹³⁹PVPLS¹⁴³ motif lies within the aa 113 to 160 CtBP1-interacting region of SATB1. Interestingly, this motif is flanked by a PCAF acetylation site lysine (K136) within a three-residue distance (27). These results indicated that the CtBP1-interacting PDZ-like domain of SATB1 harbors a PVPLS motif, which is presumably required for this interaction.

FIG. 2.

A PVPLS motif-containing region in the PDZ-like domain of SATB1 is sufficient to interact with CtBP1. (A) The top panel shows the Coomassie brilliant blue (CBB)-stained SDS-polyacrylamide gel profiles of various GST fusion proteins used in the pull-down experiment with CtBP1. The bottom panel shows an autoradiogram of ³⁵S-labeled CtBP1 pulled down by various GST fusion domains of SATB1. (B) Schematic representation of various domains and truncations of SATB1 used in the above experiment and summary of interactions. (C) Autoradiogram indicating various ³⁵S-labeled domains of SATB1 used for pull-down assay with GST-fused CtBP1. (D) GST pull-down assay was performed by incubating various ³⁵S-labeled domains of SATB1 with GST or GST-CtBP1 as described in Materials and Methods. The panel depicts an autoradiogram of the SATB1 domains pulled by GST-CtBP1. For each of the proteins depicted in panel C, there are two lanes in panel D, to represent GST pulldown and GST-CtBP1 pulldown. (E) Coomassie brilliant blue-stained SDS-polyacrylamide gel profile of GST and GST-CtBP1 used for the pulldown. (F) Schematic representation of different domains of SATB1 used for the pull-down experiment along with a summary of interactions. (G) Analysis of the N-terminal PDZ-like domain of SATB1 for the presence of a putative PXDLS-like motif(s). There are three putative CtBP1-binding PXDLS-like motifs, indicated by vertical black lines in the block corresponding to the PDZ-like domain of SATB1, and they are expanded in blue below. The most N-terminal of the three motifs spanning ¹³⁹PVPLS¹⁴³ occurs in the vicinity of the lysine 136 residue that is acetylated by PCAF (27).

CtBP1 associates with SATB1 and stabilizes the DNA-SATB1 complex.

To assess whether the SATB1 and CtBP1 interaction could be detected in vivo, nuclear extracts from HEK-293 and Jurkat T cells were incubated separately with anti-CtBP1 antibody and with mouse IgG. The antibody-protein complexes were precipitated with protein A/G beads and immunoblotted with anti-SATB1. SATB1 was coprecipitated from the two cell extracts by mouse anti-CtBP1 (Fig. 3A, lanes 4 and 6) but not by control mouse IgG (Fig. 3A, lanes 5 and 7). Rabbit anti-SATB1 was used to immunoprecipitate SATB1 as a positive control (Fig. 3A, lane 2), whereas rabbit IgG was used as a negative control (Fig. 3A, lane 3) for the pull-down assay. Conversely, CtBP1 was coprecipitated from the HEK-293 cell extract by anti-SATB1 (Fig. 3B, lane 2) but not by control rabbit IgG (Fig. 3B, lane 3). To unequivocally demonstrate the role of the PVPLS motif in the interaction with CtBP1, we generated mutants in which the residues ¹⁴¹Pro-Leu¹⁴³ were changed to Ala-Ser and in which only one residue, Pro at position 141, was changed to Ala (P141A). Coimmunoprecipitation analysis upon overexpression of FLAG-tagged SATB1 and its mutants in HEK-293 cells revealed the in vivo interaction of SATB1 (Fig. 3C, lane 1) but not of the PL-to-AS (lane 2) and P141A (lane 3) mutants with CtBP1, suggesting that these residues are critical for the interaction. Furthermore, to verify whether the PDZ-like domain of SATB1 can also interact with CtBP1 in vivo, we overexpressed CtBP1 separately with 3× FLAG-tagged full-length SATB1 and PDZ and CD+HD mutants and immunoprecipitated it with anti-FLAG antibody. CtBP1 immunoprecipitated with full-length SATB1 (Fig. 3D, lane 1) and its PDZ domain (Fig. 3D, lane 4) but not with the CD+HD region (Fig. 3D, lane 3).

FIG. 3.

SATB1 associates with CtBP1. (A) In vivo interaction of SATB1 with CtBP1 was monitored by coimmunoprecipitation analysis as described in Materials and Methods. Lane 1, input lysate; lane 2, immunoprecipitation (IP) with anti-SATB1; lane 3, IP with rabbit IgG; lanes 4 and 6, IP with mouse anti-CtBP1; lanes 5 and 7, IP with mouse IgG. The cells used for this assay were Jurkat J-6 (lanes 1, 6, and 7) and HEK-293 (lanes 2 to 5) cells. (B) Nuclear extract from HEK-293 cells (input; lane 1) was also used for coimmunoprecipitation using anti-SATB1 (lane 2) and rabbit IgG (lane 3), and complexes were immunoblotted with anti-CtBP1. (C) The CtBP1 interaction consensus motif PVPLS of SATB1 was mutated as described in Materials and Methods. FLAG-tagged wild-type SATB1 and its mutants were overexpressed in HEK-293 cells. In vivo interaction of wild-type SATB1 (lane 1) and the PL-to-AS (lane 2) and P141A (lane 3) mutants with CtBP1 was monitored by coimmunoprecipitation analysis as described in Materials and Methods. The lower panels show the expression levels of SATB1, CtBP1, and a loading control, using anti-FLAG, anti-CtBP1, and antitubulin antibodies, respectively. (D) The PDZ-like domain of SATB1 specifically interacts with CtBP1 in vivo. The nuclear extracts prepared from HEK-293 cells transfected with CtBP1 along with either FLAG-tagged full-length SATB1 (lane 1) or the CD+HD (lane 3) or PDZ (lane 4) mutant were used for coimmunoprecipitation as described in Materials and Methods, followed by immunoblotting with anti-CtBP1. The bands corresponding to CtBP1 and the IgG heavy chain are indicated by arrows. The expression of CtBP1 (middle panel) and various FLAG-tagged domains of SATB1 (lower panel) was monitored by immunoblotting using anti-CtBP1 and anti-FLAG, respectively. (E) EMSA using the heptamer of IgH MAR, showing stabilization of the DNA-SATB1 complex (lane 2) by GST-CtBP1 (lane 3). CtBP1 does not bind to DNA directly (lane 4). The complex is supershifted in the presence of anti-CtBP1 (lane 7) but not IgG (lane 6). The position of the SATB1-DNA complex is indicated by an arrow, and the supershifted complex is indicated by an asterisk. (F) Localization of endogenous SATB1 and CtBP1 in mouse thymocytes was monitored by immunostaining with respective antibodies.

Next, we reasoned that if CtBP1 interacts with SATB1, it must influence the DNA-SATB1 complex. To test this hypothesis, we performed EMSAs in which recombinant SATB1 was incubated with the IgH MAR probe in the presence of recombinant His-SATB1 (Fig. 3E, lane 2), GST-CtBP1 (lane 4), or both (lane 3). In the presence of CtBP1, the DNA-SATB1 complex appeared to be stabilized, but there was no appreciable change in the mobility of this complex. However, the addition of anti-CtBP1 supershifted this complex, confirming the presence of a ternary complex involving SATB1, CtBP1, and the DNA probe (Fig. 3E, lane 7). Anti-CtBP1 did not affect the SATB1-DNA complex (lane 5). We then monitored if SATB1 and CtBP1 colocalize inside the nucleus. In thymocytes, SATB1 is an important genome organizer and gene regulator which forms a distinguished cage-like structure demarcating the euchromatin domains (9, 16). Interestingly, immunostaining analysis of thymocytes with anti-CtBP1 revealed a cage-like structure resembling that of SATB1 (Fig. 3F). Moreover, the locations of SATB1 and CtBP1 overlapped considerably (Fig. 3F, merged image), suggesting that they may colocalize inside thymocyte nuclei by occupying a similar intranuclear space (see Movie S1 in the supplemental material). Collectively, the results show that CtBP1 associates with SATB1, stabilizes the DNA-SATB1 complex, and also forms a cage-like network distribution in thymocyte nuclei which, at least in part, colocalizes with that of SATB1.

Acetylation of SATB1 disrupts CtBP1-SATB1 complex in vitro and in vivo.

Acetylation of a Lys residue (K136) in the PDZ-like domain of SATB1 was shown to regulate its transcriptional activity via a loss of interaction with HDAC1 and an association with PCAF (27). PCAF-mediated acetylation of CtBP1-interacting proteins, such as E1A, Snail, and RIP 140, abolishes their interaction with CtBP1 (11, 47). Since the CtBP1-interacting region of SATB1 (¹³⁹PVPLS¹⁴³) is located adjacent to K136, which is known to be acetylated by PCAF (27), we tested the effect of acetylation of the PDZ-like domain of SATB1 by PCAF on the interaction with CtBP1. Tritium (³H)-labeled acetyl-CoA was used for PCAF-mediated acetylation of histones (Fig. 4A, lane 2), six-His-PDZ (lane 4), and GST-PDZ (lane 6). Acetylated GST and GST-PDZ were incubated separately with ³⁵S-labeled CtBP1. The interaction of CtBP1 with acetylated GST-PDZ was dramatically reduced (Fig. 4B, compare lane 5 with lane 4). CtBP1 did not interact with either GST or acetylated GST (Fig. 4B, lanes 2 and 3, respectively).

FIG. 4.

PCAF-mediated acetylation of SATB1 abolishes its interaction with CtBP1 in vitro and in vivo. (A) In vitro acetylation autoradiogram of histones (lanes 1 and 2), His-PDZ (lanes 3 and 4), and GST-PDZ (lanes 5 and 6) in the presence (lanes 2, 4, and 6) or absence (lanes 1, 3, and 5) of PCAF, using [³H]acetyl-CoA as described in Materials and Methods. (B) Autoradiogram of pulldown of ³⁵S-labeled CtBP1, using unmodified and acetylated proteins, as indicated on top of the lanes. Lane 1 represents the ³⁵S-labeled in vitro-transcribed and -translated CtBP1 preparation used for the pulldown. (C) Coomassie brilliant blue (CBB)-stained SDS-polyacrylamide gel depicting the proteins used in the pull-down assay. The lower panel depicts an immunoblot of a duplicate gel, created using anti-AcSATB1. (D) Jurkat cells were treated with 10 mM LiCl as described in Materials and Methods. At the indicated time points, cells were harvested and processed for immunoblot analysis to monitor the levels of acetylated SATB1 (AcSATB1; top panel), PCAF (middle panel), and SATB1 (lower panel). (E) Real-time RT-PCR analysis of PCAF transcript upon LiCl or ionomycin treatment of Jurkat T cells and LiCl treatment of mouse thymocytes. (F) Coimmunoprecipitation analysis of SATB1 and CtBP1 interaction upon LiCl treatment of Jurkat cells. Cells were harvested at the indicated time points, and the SATB1 and CtBp1 interaction was monitored by coimmunoprecipitation upon LiCl treatment (upper panel). Immunoblot analysis was performed to monitor the levels of SATB1 (middle panel) and CtBP1 (lower panel) during the indicated time course. (G) Effects of ionomycin (lane 5), LiCl (lane 4), and TSA (lane 7) treatments of Jurkat T cells on SATB1 and CtBP1 interaction were monitored by coimmunoprecipitation (upper panel). Immunoblot analysis was performed to monitor the levels of FLAG-SATB1 (middle panel) and CtBP1 (lower panel) in the nuclear extracts. The graph below represents densitometric quantitation of bands corresponding to SATB1 expression in the indicated lanes. (H) Overexpression of PCAF but not the p300 acetyltransferase affects the interaction between SATB1 and PCAF. HEK-293 cells were transfected with FLAG-SATB1, p300, and PCAF expression constructs as indicated or treated with LiCl. Lane 1, coprecipitation of SATB1 with anti-CtBP1 upon LiCl treatment; lane 2, coprecipitation of SATB1 with anti-CtBP1 when p300 is overexpressed; lane 3, coprecipitation of SATB1 with anti-CtBP1, with no treatment or overexpression; lane 4, coprecipitation of SATB1 with anti-CtBP1 when PCAF is overexpressed; lane 5, mouse IgG as a control for immunoprecipitation. The middle and lower panels depict the expression levels of FLAG-SATB1 and CtBP1, respectively. The graph below represents densitometric quantitation of bands corresponding to FLAG-SATB1 expression in the indicated lanes. (I) Overexpression of PCAF disrupts association of SATB1 with CtBP1. HEK-293 cells were transfected with FLAG-SATB1 and processed for coimmunoprecipitation after LiCl treatment (lane 1), PCAF cotransfection (lane 2), and p300 cotransfection (lane 3) as described in Materials and Methods. Arrows depict positions of the IgG heavy chain and CtBP1. Immunoblot analysis was performed to monitor the levels of FLAG-SATB1 (middle panel) and CtBP1 (lower panel) in the extracts from transfected cells. (J) Interaction of CtBP1 and FLAG-SATB1-K136R remains unaffected by PCAF overexpression or LiCl treatment. HEK-293 cells were transfected with FLAG-SATB1-K136R and processed for coimmunoprecipitation after LiCl treatment (lane 1), PCAF cotransfection (lane 2), and p300 cotransfection (lane 3) as described in Materials and Methods. Arrows depict positions of FLAG-SATB1-K136R and the IgG heavy chain. Immunoblot analysis was performed to monitor the levels of FLAG-SATB1 (middle panel) and CtBP1 (lower panel) in the extracts from transfected cells.

Several CtBP1 target genes are also targets of the Wnt signaling pathway (8, 11, 42, 46). The transcription factor and CtBP1 complexes are bound to promoters to execute repression of transcription. The molecular mechanism which relieves this repression is usually signal dependent. Acetylation of a transcription factor flanking the CtBP1 binding motif leads to dissociation of CtBP1, which culminates in the derepression of target genes (11). Since acetylation of SATB1 has a negative influence on its association with CtBP1 in vitro, we tested whether the same effect is observed in vivo. SATB1 is acetylated upon treatment with ionomycin, in a time-dependent manner (27); however, its acetylation status upon initiation of Wnt signaling is not known. LiCl treatment mimics Wnt signaling by inhibiting glycogen synthase kinase 3β and thereby stabilizing β-catenin (24). A polyclonal antibody against K136-acetylated SATB1 (27) was used to detect the acetylation pattern of SATB1 in Jurkat T cells upon treatment with 10 mM LiCl in a time course experiment. An increase in acetylation was observed for up to 6 h, followed by decreases at later time points (Fig. 4D, top panel). We also detected increased expression of PCAF under these conditions (Fig. 4D, middle panel), whereas there was no change in the expression of SATB1 (Fig. 4D, bottom panel). Furthermore, we also monitored the transcript level of PCAF acetyltransferase and found that it was upregulated for up to 6 h in Jurkat T cells as well as in mouse thymocytes (Fig. 4E). Ionomycin treatment also led to an increase in expression of PCAF acetyltransferase in Jurkat T cells (Fig. 4E, bottom graph). Interaction of SATB1 with CtBP1 was reduced significantly after 6 h of LiCl treatment of Jurkat cells (Fig. 4F). Trichostatin A (TSA) is an HDAC inhibitor which enhances acetylation of histones and also several chromatin-associated proteins. We found that acetylation of SATB1 was increased upon TSA treatment (data not shown). Next, we monitored the interaction of CtBP1 with SATB1 following these treatments. Interaction of SATB1 with CtBP1 was reduced when Jurkat cells were treated with TSA (Fig. 4G, lane 7), ionomycin (Fig. 4G, lane 5), and LiCl (Fig. 4G, lane 4) for 6 h, whereas there was no change in interaction of SATB1 with CtBP1 when cells were treated with dimethyl sulfoxide (Fig. 4G, lane 6). Next, we overexpressed PCAF and p300 along with SATB1 and CtBP1 and performed coimmunoprecipitation to monitor the interaction between SATB1 and CtBP1. When proteins were pulled down with anti-CtBP1 and immunoblotted with anti-SATB1, a decreased association of SATB1 with CtBP1 was observed only when PCAF was overexpressed (Fig. 4H, lane 4), but a relatively less inhibitory effect was observed when p300 was overexpressed (Fig. 4H, lane 2). LiCl treatment also resulted in decreased interaction between SATB1 and CtBP1 (Fig. 4H, lane 1). Similarly, when SATB1 was immunoprecipitated with anti-FLAG antibody, we found decreased coimmunoprecipitation of associated CtBP1 when PCAF was overexpressed (Fig. 4I, lane 2) and upon LiCl treatment (Fig. 4I, lane 1) but not when p300 was overexpressed (Fig. 4I, lane 3). Next, we mutated the lysine 136 residue of SATB1 to arginine to abolish the effects of acetylation on its interaction with CtBP1. Strikingly, coimmunoprecipitation analysis revealed that unlike the case with wild-type SATB1, the interaction of K136R-SATB1 with CtBP1 is not sensitive to LiCl (Fig. 4J, lane 1) or PCAF (lane 2), indicating that the acetylation status of K136 of SATB1 is an important determinant of its association with CtBP1. Altogether, these results indicate that SATB1 is acetylated early during the onset of Wnt signaling, presumably by PCAF acetyltransferase, which eventually leads to dissociation of the CtBP1 complex from SATB1.

CtBP1 cooperates with SATB1 to repress transcription.

IgH MAR-Luc is a well-characterized MAR reporter containing seven concatemerized SATB1-binding core MAR elements in a pGL3-based luciferase reporter system (26, 27), and we employed this reporter to score for the effect of CtBP1 on the ability of SATB1 to repress transcription. Towards this end, we overexpressed SATB1 and/or CtBP1 in HEK-293 cells that express SATB1 endogenously (28) and cotransfected them with the reporter construct. As expected, SATB1 repressed the reporter activity in a dose-dependent fashion (Fig. 5A, bars 2 and 3). Interestingly, we found that CtBP1 alone could also repress the MAR reporter activity (Fig. 5A, bars 4 and 5), indicating that CtBP1 must cooperate with endogenous SATB1 to repress the reporter activity. HDAC1 is known to repress transcription mediated by SATB1 (26, 27), and similarly, we observed HDAC1-mediated repression in a dose-dependent manner (Fig. 5A, bars 6 and 7). Next, we tested the activity of CtBP1 on reporter expression in the presence of exogenous SATB1 and found that CtBP1 repressed the reporter more in combination with SATB1 than either SATB1 or CtBP1 alone did (Fig. 5A, compare bar 9 with bars 2 and 4). A similar effect was also seen when HDAC1 was used along with SATB1 (Fig. 5A, compare bar 10 with bars 2 and 6). Repression mediated by CtBP1 and HDAC1 was similar to that with SATB1 and CtBP1 (Fig. 5A, compare bar 8 with bar 9), suggesting that CtBP1-mediated repression may be HDAC1 dependent and that CtBP1 collaborates with SATB1 to repress transcription. However, since the levels of repression were not affected substantially, factors other than HDAC1 may also be involved in the repression.

FIG. 5.

CtBP1 cooperates with SATB1 to repress transcription. (A) Luciferase (Luc) reporter assay performed upon cotransfection of 1 to 2 μg of various expression constructs, as indicated, along with 1 μg of IgH MAR-luciferase reporter construct into HEK-293 cells. Cells were harvested 40 h after transfection, and luciferase activity was measured using 100 μg of protein from each cell lysate. Luciferase activity is expressed as the level of expression compared to that of the control, which was set to 1 (lane 1). (B to D) RT-PCR analysis of human IL-2, mMyc, and mouse IL-2R transcripts indicating that genes repressed by SATB1 are derepressed during LiCl treatment or CtBP1 knockdown. Cell treatments and knockdowns were performed as described in Materials and Methods. Panels below the graphs depict gel profiles of the respective PCR products. (B) The first and second graphs represent the relative expression of IL-2 transcript during ionomycin and LiCl treatment, respectively, in Jurkat T cells. (C) The first and second graphs indicate the relative expression of mMyc and mouse IL-2R transcripts, respectively, during LiCl treatment of mouse thymocytes. (D) CtBP1 knockdown elevates expression of the IL-2 gene transcript (compare bar 3 with bar 1), similar to what is observed when SATB1 is silenced (compare bar 2 with bar 1). Lanes 5 and 7 depict the relative knockdown of CtBP1 and SATB1, respectively. (E) Knockdown of SATB1 (lanes 2 and 5) and CtBP1 (lanes 3 and 6) was monitored by RT-PCR (lanes 1 to 3) and immunoblot analysis (lanes 4 to 6). Samples from cells transfected with the pSuper vector were used as controls (lanes 1 and 4). siCtBP1 does not affect the expression of CtBP2 (panel 3).

Human IL-2, mouse IL-2R, and c-Myc genes are known SATB1 targets (9, 26, 50). Ionomycin treatment has been shown to increase the acetylation of SATB1, which leads to derepression of IL-2 gene transcription in a time-dependent manner (27). Since we observed similar kinetics of acetylation of SATB1 upon LiCl treatment, we tested whether SATB1 target genes are affected by LiCl treatment. Treatment of Jurkat T cells with either ionomycin or LiCl for up to 6 h elevated the transcript level of IL-2 in a time-dependent manner (Fig. 5B). Since the SATB1-CtBP1 complex forms a characteristic cage-like structure in mouse thymocytes, we reasoned that CtBP1 and SATB1 may also share their targets and therefore tested whether SATB1 target genes in mouse T cells are also affected by Wnt signaling. We observed elevated expression of the mouse c-Myc and IL-2R genes when mouse thymocytes were treated with LiCl for up to 6 h (Fig. 5C). To know whether CtBP1 can directly affect the expression of SATB1 target genes, we knocked down CtBP1 and monitored the expression of IL-2. IL-2 expression was elevated at least 2.5-fold upon 2-fold knockdown of CtBP1 in Jurkat T cells (Fig. 5D, first graph, compare bar 3 with bar 1). This is similar to the upregulation of IL-2 observed upon knockdown of SATB1 (Fig. 5D, bar 2). These results suggest that the genes repressed by SATB1 are derepressed upon LiCl treatment and therefore could be activated upon Wnt signaling.

Acetylation of SATB1 at K136 alleviates repression mediated by SATB1 and CtBP1.

We further tested the effect of LiCl in a luciferase reporter assay using the heptameric IgH MAR and the IL-2 promoter cloned into the pGL3 basic vector, termed IgH MAR-luc and IL-2P-luc, respectively (26, 27). LiCl treatment of Jurkat T cells transfected with the IgH MAR-reporter or IL-2 promoter-reporter construct led to increased luciferase activity without (Fig. 6A, bars 2 versus bars 1) as well as with exogenously expressed SATB1 (Fig. 6A, bars 6 versus bars 5), CtBP1 (Fig. 6A, bars 4 versus bars 3), or both together (Fig. 6A, bars 8 versus bars 7) compared to that in cells treated with KCl. In contrast, the K136A acetylation-defective mutant of SATB1 showed comparatively less enhancement in the effect of LiCl treatment on IgH MAR-reporter activity and virtually no effect on IL-2 promoter-reporter activity (Fig. 6A, bars 10 versus bars 9), even in the presence of CtBP1 (Fig. 6A, bars 12 versus bars 11), suggesting that the SATB1 acetylation site lysine 136 is required for the LiCl/Wnt response. This result also revealed that the repression mediated by SATB1 is abolished upon LiCl treatment. This derepression could be attributed to the enhanced acetylation at K136 of SATB1. We therefore tested the effect of the acetyltransferases PCAF and p300 on the transcriptional activity of SATB1 in these reporter assays. Overexpression of PCAF abolished the repression mediated by SATB1 (Fig. 6B, bars 3 versus bars 2) and the SATB1-CtBP1 complex (Fig. 6B, bars 6 versus bars 5), indicating that PCAF activity is required for the derepression mediated by the SATB1-CtBP1 complex. In contrast, overexpression of p300 did not significantly affect the repression mediated by SATB1 (Fig. 6B, bars 4 versus bars 2) and the SATB1-CtBP1 complex (Fig. 6B, bars 7 versus bars 5). Overexpression of K136A-SATB1 resulted in repression (Fig. 6B, bars 8) comparable with that of wild-type SATB1 (Fig. 6B, bars 2). Strikingly, the repression mediated by this acetylation-defective mutant of SATB1 and its complex with CtBP1 was not affected significantly by PCAF overexpression (Fig. 6B, bars 9 and 12, respectively). Similar to that of wild-type SATB1, p300 overexpression did not affect repression mediated by K136A-SATB1 and its complex with CtBP1 (Fig. 6B, bars 7 and 13, respectively). These results are indicative of the requirement of acetylation of K136 by PCAF for the dissociation of the CtBP1 complex from SATB1, which happens during Wnt signaling for the derepression of SATB1 target genes.

FIG. 6.

Repression mediated by SATB1 and CtBP1 is abolished upon acetylation of SATB1. A luciferase (Luc) reporter assay was performed upon cotransfection of 1 μg of various expression constructs, as indicated, along with 1 μg of either IgH MAR-luciferase reporter construct (left graphs) or IL-2 promoter-reporter construct (right graphs) into HEK-293 cells. Cells were harvested 40 h after transfection, and luciferase activity was measured as described in Materials and Methods. Luciferase activity is expressed as activity compared to that of the control, which was arbitrarily set to 1 (lane 1). The reporter activity was checked essentially to monitor the effect of the K136A acetylation-defective mutant of SATB1 either in the presence of LiCl or KCl treatment (A) or upon overexpression of PCAF, p300, and CtBP1 individually or in the indicated combinations (B).

CtBP1 dissociates from SATB1 bound to the human IL-2 promoter and mouse c-Myc promoter upon acetylation of SATB1 following LiCl treatment.

SATB1 regulates the transcription of the IL-2 gene by binding to its distal promoter region (26). The HDAC1 complex associated with SATB1 on the IL-2 promoter is displaced by the human immunodeficiency virus type 1 (HIV-1) Tat protein during HIV-1 infection, leading to the upregulation of the IL-2 transcript (26). Posttranslational modifications of SATB1 are involved in the transcriptional regulation of IL-2 expression (27). The 1.5 kb upstream of the mouse c-Myc transcription start site harbors an SBS and is bound by SATB1, thereby establishing a unique chromatin structure to regulate the transcription of c-Myc in mouse thymocytes (9). Thymocytes from SATB1 knockout mice revealed ectopic expression of the c-Myc gene (9). We therefore performed ChIP analysis of these two SBSs to monitor the occupancy of SATB1, HDAC1, and CtBP1 in Jurkat cells (Fig. 7A, top panel, lanes 1, 2, and 3, respectively). SATB1 knockdown in Jurkat T cells led to a reduced occupancy of HDAC1 (Fig. 7A, compare lane 2 in the second panel with lane 2 in the top panel) and CtBP1 (Fig. 7A, compare lane 3 in the second panel with lane 3 in the top panel) on the IL-2 promoter. Upon CtBP1 knockdown, there was no appreciable change in association of SATB1 with the IL-2 promoter (Fig. 7A, compare lane 1 in the third panel with lane 1 in the top panel), but there was a reduction in HDAC1 occupancy (Fig. 7A, compare lane 2 in the third panel with lane 2 in the top panel). CtBP1 overexpression led to enhanced occupancy of CtBP1 at the IL-2 promoter region (Fig. 7A, lane 3 in the bottom panel). Overexpression of CtBP1 also led to increased occupancy of HDAC1 (Fig. 7A, lane 2 in the bottom panel) at this region, whereas there was no significant change in the occupancy of SATB1 at this promoter (Fig. 7A, lanes 1 of the top and bottom panels). These results indicate that SATB1 is associated with the IL-2 promoter region, along with CtBP1 and HDAC1, in vivo. Next, we monitored the occupancy of SATB1 and CtBP1 during ionomycin and LiCl treatments at this promoter and the mouse c-Myc SBS region. A distant non-SATB1-binding upstream region was used to normalize the specific PCR products of the Myc SBS. Although no appreciable change in SATB1 occupancy was observed at these SBSs for up to 6 h of treatment (Fig. 7B and D, top panels), the occupancy of HDAC1 and CtBP1 at these promoter regions decreased drastically at the mouse Myc (mMyc) SBS (Fig. 7D, panels 2 and 3) and significantly at the IL-2 promoter (Fig. 7B, second and third panels) upon LiCl treatment. Concomitantly, occupancy of PCAF increased at these regions for up to 6 h of ionomycin and LiCl treatment (Fig. 7B, fourth panel, for the IL-2 promoter and Fig. 7D, fourth panel, for the mMyc SBS). As a result of this recruitment, the acetylation of lysine 9 of histone H3 at the IL-2 promoter and the mMyc SBS was enriched during these treatments (Fig. 7B, bottom panel, for the IL-2 promoter and Fig. 7D, bottom panel, for mMyc SBS), revealing that the IL-2 promoter- and mMyc SBS-associated chromatin becomes transcriptionally activated. Taken together, these results indicate that SATB1 is associated with the CtBP1 and HDAC1 complex at the IL-2 promoter region in human T cells and at the c-Myc SBS region in mouse thymocytes. During ionomycin and LiCl treatments, this association is disrupted and PCAF is simultaneously recruited, which presumably causes acetylation of histones at the two SBSs.

FIG. 7.

IL-2 promoter-bound SATB1 is associated with CtBP1 and dissociates when SATB1 is acetylated during ionomycin or LiCl treatment. (A) Chromatin from Jurkat cells was isolated, subjected to immunoprecipitation, and processed further as described in Materials and Methods. PCR amplification of the purified DNAs from such reactions revealed that SATB1, CtBP1, and HDAC1 are associated with the IL-2 promoter region (panel 1). ChIP was also performed with Jurkat cells in which SATB1 or CtBP1 was knocked down (siSATB1 and siCtBP1) or CtBP1 was overexpressed (bottom panel). Antibodies used for immunoprecipitation are indicated at the top. (B) ChIP analysis of occupancy of SATB1, HDAC1, CtBP1, PCAF, and histone H3 lysine 9 acetylation (H3K9Ac) at the IL-2 promoter upon ionomycin (lanes 1 to 3) and LiCl (lanes 4 to 6) treatment of Jurkat cells. The time course of treatments is indicated at the top of the lanes. (C) Controls for the ChIP experiment. Lanes 1 and 2 depict a lack of PCR amplification from mouse IgG- and rabbit IgG-precipitated chromatin. Lanes 3 to 6 represent PCR products of chromatin inputs prepared from cells transfected with control pCDNA vector, siSATB1, siCtBP1, and CtBP1 expression constructs, respectively. (D) Mouse thymocytes were treated with ionomycin and LiCl for the indicated time points to analyze the occupancy of the indicated factors at the SBS within the mMyc locus (9). Antibodies used for the ChIP are indicated on the right. Graphs represent the quantitation of the ChIP PCRs by real-time analysis, and changes in occupancy are represented as enrichment with respect to the occupancy at 0 h. Black bars represent occupancy upon ionomycin treatment, whereas gray bars represent occupancy upon LiCl treatment. The gel profile of PCR products from each of these reactions is depicted below each graph.

Gene expression profiling identifies several common targets of SATB1 and CtBP1.

SATB1 regulates the transcription of a large number of genes, either negatively or positively (20, 27). There is only one report documenting the repertoire of genes regulated by CtBP1 in a mouse model (18). Since both of these proteins form a complex and regulate transcription, we asked how many genes are regulated by these two proteins in a similar manner in a specific cell type. We adopted the microarray profiling approach to analyze the genes which are dysregulated upon siRNA-mediated silencing of SATB1 and CtBP1 in HEK-293 cells. Labeled cDNAs prepared from RNAs isolated from siSATB1- or siCtBP1-transfected cells and control scrambled RNA-transfected cells were used to hybridize with an 8,000-cDNA array. Genes whose expression was affected by knockdown of SATB1 and CtBP1 were identified and clustered. Such analysis revealed that about 300 SATB1-regulated genes are also regulated by CtBP1 in a similar manner (see Fig. S1 in the supplemental material). About 100 genes were upregulated upon knockdown of SATB1 and CtBP1, implying that these genes represent a subset of genes that presumably undergo acetylation-dependent transcriptional regulation by the SATB1-CtBP1 complex in response to cellular signals (see Table S2 in the supplemental material). These effects are gene specific, since CtBP1 regulates more than 800 genes independently of SATB1 (see Table S3 in the supplemental material). Many of the genes that are positively regulated by SATB1 are also positively regulated by CtBP1 and therefore were repressed upon knockdown of these two proteins (see Fig. S1 in the supplemental material). Furthermore, the coregulated genes could be classified into multiple pathways of biological significance, such as focal adhesion, mitogen-activated protein kinase (MAPK) signaling, regulation of the actin cytoskeleton, long-term potentiation, ribosome biogenesis, glycolysis, cell cycle, and extracellular matrix-receptor interactions, etc. (Table 1). Thus, the transcriptional targets of SATB1 and CtBP1 demonstrate a significant overlap pointing toward their common regulatory mechanism(s).

TABLE 1.

List of various pathways commonly regulated by SATB1 and CtBP1^a

Pathway	No. of common genes	GeneList vs Pathway random overlap P value
Ribosome	5	0.00
Focal adhesion	4	0.03
Pyruvate metabolism	3	0.00
Glycolysis and gluconeogenesis	3	0.00
Cell cycle	3	0.03
Regulation of actin cytoskeleton	3	0.07
Neurodegenerative disorders	2	0.01
Propanoate metabolism	2	0.02
Butanoate metabolism	2	0.03
Tyrosine metabolism	2	0.04
Extracellular matric-receptor interaction	2	0.05
MAPK signaling pathway	7	0.05
Glioma	4	0.01
Chronic myeloid leukemia	4	0.02
Epithelial cell signaling in Helicobacter pylori infection	4	0.02
Tryptophan metabolism	3	0.06
Long-term potentiation	3	0.08
Vitamin B6 metabolism	2	0.01
Cholera infection	2	0.08

^aGene expression profiling was performed using RNAs isolated from cells in which expression of SATB1 and CtBP1 was silenced using respective siRNAs as described in Materials and Methods. Pathway analysis was performed using GeneSpring GX software, version 7.3. Only those pathways that involved two or more genes coregulated by SATB1 and CtBP1 are included, and the P values are <0.08.

DISCUSSION

The mechanism of transcription repression or activation by a transcription factor involves direct or indirect recruitment of various corepressor or coactivator complexes to the specific binding sites on DNA, thereby modulating chromatin structure in a manner that renders it either resistant or conducive for the execution of transcription (4, 17, 19, 29, 44). Large numbers of DNA binding proteins that are posttranscriptionally modified are required not only for cognate site binding but also for the recruitment of corepressor or coactivator complexes. Specific extracellular signals trigger transcription repression or activation in two distinct ways. Certain signals induce phosphorylation-mediated translocation of specific transcription factors, such as NF-κB, into the nucleus, where they bind to target promoters and alter the local chromatin structure to activate transcription (5, 19, 22, 37, 44). In contrast, a large number of DNA-binding transcription factors could also be bound to their genomic targets in the absence of any signal to appropriately regulate transcription. Once a specific signal is received by the cell, the DNA-bound transcription factor undergoes a change in its posttranslational modification, which either changes the association with the corepressor/coactivator complex or modulates the affinity with DNA via conformational changes (2, 3, 7, 13, 27, 38, 49, 52). Both types of alterations in properties of a transcription factor may affect the transcription of target genes. SATB1 serves as an organizer of chromatin higher order and also as a transcription factor (16). The posttranslational modification status of SATB1 determines whether it will act as an activator or as a repressor (27). The homeodomain of SATB1 recognizes its specific binding targets in concert with the Cut domain, in a dimerization-dependent manner. SATB1 mediates repression of a large number of genes (9, 26, 27, 50) and seems to act predominantly as a repressor via recruitment of HDAC1 (26, 27, 50). However, the precise mechanism of repression is not known. We therefore designed an experiment to delineate the functions of different functional domains of SATB1. A luciferase reporter-based assay was devised, which is similar to the repression assay used for delineating the transcriptional functions of various domains of protein (32, 47). This assay revealed that the N-terminal PDZ-like domain of SATB1 is involved in mediating repression. Recently, it was shown that the PDZ-like domain is required for the dimerization of SATB1 (15, 35) and is also involved in interaction with several chromatin-modifying complexes that modulate chromatin structure for the regulation of transcription (26, 27, 50). The identification of a limited number of components of various histone-modifying enzymes and some of the specialized substructures in the nucleus indicates that SATB1 function is largely dependent on its associated complexes. In this report, identification of CtBP1 as a SATB1 interacting partner to contribute in repression provides a clue toward how a chromatin organizer may function in a signal-dependent manner. CtBP1 is known to interact with various DNA-binding transcription factors to repress the transcription of their target genes (11). CtBP1 mediates transcriptional repression by recruiting HDAC1 and the Groucho and/or Polycomb group of proteins to the transcription factor binding regions (11). The N-terminal PDZ-like domain of SATB1 harboring a PVPLS motif interacts with CtBP1. A specialized cage-like distribution of CtBP1 in mouse thymocyte nuclei similar to what is formed by SATB1 signifies that it could be a component of a complex formed by SATB1 in organizing the higher-order chromatin structure. This peculiar distribution of CtBP1 in thymocytes is also the first report of its expression in T cells, although the presence of CtBP1 but not CtBP2 transcripts has been shown by Northern blot analysis (41). Such specialized higher-order organization of chromatin seems to be critical for the proper transcription of genes (10, 16, 25), as disruption in this kind of a structure may lead to altered gene expression, as seen during tumorigenesis (20). The DNA-SATB1 complex seems to be stabilized in the presence of CtBP1 in vitro, and presumably a similar mechanism may operate in vivo wherein SATB1-CtBP1 complexes associated at multiple genomic loci are bound stronger than is SATB1 alone. SATB1 makes a very-high-molecular-weight complex with the IgH MAR DNA, and presumably, therefore, a migration difference is not observed due to association of CtBP1 in the EMSA experiment. CtBP1 seems to stabilize the SATB1-DNA complex by an unknown mechanism. The interaction of CtBP1 with SATB1 also provides a repressor function similar to what is observed with HDAC1, in a dose-dependent manner. Repression by CtBP1 is enhanced in the presence of HDAC1, indicating that CtBP1 presumably acts to repress the transcription mediated by SATB1 in an HDAC1-dependent manner. SATB1 is known to interact directly with HDAC1 and to repress IL-2 and IL-2Rα gene expression (26). Thus, SATB1 may repress the reporter activity by associating either directly with HDAC1 or indirectly via the CtBP1 complex. CtBP1 is known to mediate repression in two modes, i.e., HDAC1-dependent and HDAC1-independent modes (11). We have not tested the HDAC1-independent mechanism of repression by CtBP1 associated with SATB1. Nevertheless, our results prompted us to postulate two mechanisms for SATB1-mediated transcriptional repression activity (Fig. 8). The primary mechanism may involve direct recruitment of HDAC1, and the secondary mechanism may involve indirect recruitment of HDAC1 by corepressors such as CtBP1 to SATB1's genomic targets, although recruitment of additional repressor complexes, such as the Polycomb group of proteins, and inhibition of CBP/p300 activity by CtBP1 at SATB1 target sites could also be possible. Results using the HDAC inhibitor TSA also suggested a possible role for acetylation in the dysregulation of SATB1-CtBP-mediated repression activity. SATB1 is known to act as a repressor upon binding to its consensus sites, and the extent of repression is proportional to SATB1's binding affinity for these sequences (35). Thus, this may explain the context-specific effects on transcription. We propose that the SATB1-CtBP1 complex may also be subjected to this kind of regulation, depending upon the genomic context.

FIG. 8.

Model depicting the mechanism of derepression by SATB1 during Wnt signaling. The mechanism of transcription regulation without (top) and with (bottom) LiCl treatment of cells is depicted. Derepression involves acetylation of SATB1, dissociation of the CtBP1 and HDAC1 corepressor complex, and recruitment of an activator complex via SATB1 to its binding target. See the text for details.

The loss of interaction due to acetylation by PCAF at K136, which is adjacent to the CtBP1-interacting pentapeptide motif in SATB1, revealed a unique feature of SATB1 in changing the interacting partner depending upon its specific posttranslational modification. This is in concert with a number of other transcription factors which have been shown to possess a PXDLX motif for interaction with CtBP1 (11, 47). The binding of CtBP1 to transcription factors is largely dependent on the acetylation of a lysine adjacent to the pentapeptide (11, 47). In SATB1, the acetylation site K136 is 3 aa upstream of the PVPLS motif, and this may constitute a distinct class of CtBP-interacting proteins, since the lysines in all other CtBP-interacting proteins occur 2 or 3 aa downstream of the PXDLX motif (11). Interestingly, SATB2, a closely related homolog of SATB1 (14), lacks this lysine residue in the proximity of the PXDLX motif. Recently, it was reported that ionomycin-mediated activation of T cells causes enhanced acetylation of SATB1 (27). However, the signals/stimuli which bring about change in the acetylation status of SATB1 are not fully characterized. Since CtBP1 is known to maintain the repressed status of Wnt target genes (8, 11, 42, 46), we therefore monitored the effect of Wnt signaling on the activity of SATB1 target genes. LiCl is a pharmacological agent that acts via inhibition of glycogen synthase kinase 3β, which phosphorylates β-catenin to mediate its degradation and therefore is widely used to mimic the Wnt signaling pathway (24, 42, 43). Once the kinase activity is inhibited, β-catenin is stabilized, accumulates in the cytosol, and subsequently translocates to the nucleus. In the nucleus, β-catenin associates with the TCF/LEF transcription factors and activates the transcription of target genes. The Wnt activation of Jurkat human T cells and mouse thymocytes activated SATB1 target genes in a time-dependent manner, indicating that SATB1-mediated transcriptional repression is changed when Wnt signaling is on. Intriguingly, ionomycin activation also showed a similar pattern of activation of target genes. Thus, these two pathways may overlap downstream, leading to similar effects on SATB1 modification and function. The PCAF transcript and protein are elevated during these treatments, providing an additional clue toward how acetylation of SATB1 might be regulated. Upon Wnt signaling, acetylation of SATB1 is enriched during early time points, resulting in dissociation of CtBP1 as well as HDAC1 and enrichment of PCAF at SATB1 binding sites. This series of molecular events constitutes an important mechanism of switching of SATB1 from repressor mode to activator mode. The time-dependent acetylation and deacetylation of SATB1, as observed during the early and late time points during Wnt signaling, may also constitute another molecular switch dictating the fate of SATB1-regulated genes that are also common targets of this signaling pathway. The insensitivity of acetylation-defective K136A and K136R mutants toward the Wnt stimuli and toward PCAF acetyltransferase strengthens the involvement and importance of K136 acetylation in Wnt-mediated activation of SATB1 target genes. The increased occupancy of the coactivator complex upon Wnt “on” conditions also elevates the acetylation of histones at the SATB1-binding target gene promoters, such as the human IL-2 promoter and the mMyc SBS, resulting in their activation.

Gene expression profiling upon knockdown of SATB1 and CtBP1 identified a large number of common genes targeted by SATB1 and CtBP1. These commonly targeted genes may be coordinately repressed by the SATB1 and CtBP1 complex. Many of these genes may change their transcription level during Wnt or other specific signaling. Gene expression profiling of CtBP1 knockout versus CtBP1-rescued mouse embryo fibroblasts revealed that CtBP1 regulates many epithelium-specific and proapoptotic genes (18). Interestingly, SATB1 also regulates multiple proapoptotic genes (1). Not surprisingly, both SATB1-deficient and CtBP1-deficient cells are hypersensitive to apoptosis (1, 18). Comparison of CtBP1-regulated genes from our study with those in the study of Grooteclaes et al. (18) revealed 84 genes in common within the two studies. This number is quite significant considering that the two microarray platforms used are very different and that their study was done with mouse cells, whereas we used human cell lines. Intriguingly, a significant number of genes were also regulated by SATB1 (see Table S4 in the supplemental material). Pathway analysis of genes that are regulated similarly by these two proteins revealed that they constitute important components of various signaling pathways. Further investigation of the requirement of the collaboration of these factors at the level of transcription of these genes could potentially reveal an important candidate regulator of the pathway, which can be targeted during a disease condition or a physiological disorder.

In summary, this study identified for the first time a large number of common SATB1- and CtBP1-regulated genes which may undergo changes in expression upon acetylation-mediated loss of SATB1-CtBP1 interaction induced by Wnt signaling. A number of genes are commonly downregulated, whereas relatively fewer genes are upregulated, upon knockdown of either CtBP1 or SATB1. We propose that the genes which are repressed directly by the SATB1-CtBP1 complex are upregulated upon knockdown of either of these proteins. Since a large number of genes are also downregulated upon knockdown of either of the proteins, the positive regulation by these proteins could be an indirect effect. Additionally, our results also point toward a mechanism by which CtBP1 may contribute toward the repression function of SATB1 in an HDAC1-dependent manner, at least in part. Signal-induced acetylation of SATB1 by PCAF results in dissociation of the SATB1-CtBP1 complex at human IL-2 and mouse c-Myc promoters to derepress the respective genes. We propose that this mechanism may be required to overcome the CtBP1-mediated repression of SATB1 target genes involved in various signal transduction pathways. Elucidation of the mechanism by which repressor complexes associated with transcription factors such as SATB1 are replaced by activator complexes would provide a complete understanding of the mechanism of transcriptional activation following a signal. The dissociation of repressor complexes and association of activator complexes with the same transcription factor depend on the posttranslational modification status of the transcription factor. For SATB1, it appears that acetylation of K136 by PCAF is an important switch which regulates its association with the CtBP1 corepressor complex to execute appropriate transcription following a specific signal. These studies argue for a role of acetylation of SATB1 in the Wnt signaling pathway and also provide novel insights into the regulation of transcription by SATB1 by virtue of its acetylation-dependent association with the CtBP1 repressor complex.