JBC Origene Your Gene Company

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M702885200 on July 10, 2007

J. Biol. Chem., Vol. 282, Issue 37, 26725-26739, September 14, 2007
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementa Data
Right arrow All Versions of this Article:
282/37/26725    most recent
M702885200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bu, Y.
Right arrow Articles by Gelman, I. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bu, Y.
Right arrow Articles by Gelman, I. H.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

v-Src-mediated Down-regulation of SSeCKS Metastasis Suppressor Gene Promoter by the Recruitment of HDAC1 into a USF1-Sp1-Sp3 Complex*Formula

Yahao Bu and Irwin H. Gelman1

From the Department of Cancer Genetics, Roswell Park Cancer Institute, Buffalo, New York 14263

Received for publication, April 4, 2007 , and in revised form, June 6, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
SSeCKS (Src-suppressed C kinase substrate), also called gravin/AKAP12, is a large scaffolding protein with metastasis suppressor activity. Two major isoforms of SSeCKS are expressed in most cell and tissue types under the control of two independent promoters, designated {alpha} and beta, separated by 68 kb. SSeCKS transcript and protein levels are severely decreased in Src- and Ras-transformed fibroblasts and in many epithelial tumors. By dissecting its promoters with progressive deletion analysis, we identified the sequence between –106 and –49 in the {alpha} proximal promoter as the minimal v-Src-responsive element, which contains E- and GC-boxes bound by USF1 and Sp1/Sp3, respectively. Both E- and GC-boxes are crucial for v-Src-responsive and basal promoter activities. v-Src does not alter USF1 binding levels at the E-box, but it increases Sp1/Sp3 binding to the GC-box despite no change in their cellular protein abundance. SSeCKS {alpha} and beta transcript levels in v-Src/3T3 cells can be restored by treatment with the histone deacetylase inhibitor, trichostatin A, but not with the DNA demethylation agent, 5-azacytidine. Chromatin changes are found only on the {alpha} promoter even though the beta proximal promoter contains a similar E- and GC-box arrangement. Recruitment of HDAC1 is necessary and sufficient to cause repression of {alpha} proximal promoter activity, and the addition of Sp1 and/or Sp3 potentiates the repression. Our data suggest that suppression of the beta promoter is facilitated by Src-induced changes in the {alpha} promoter chromatinization mediated by a USF1-Sp1-Sp3 complex.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
SSeCKS (Src-suppressed C kinase substrate) was originally identified in a screen for genes severely down-regulated by v-Src (1), and then characterized as a major in vitro and in vivo substrate of protein kinase C (2). SSeCKS is the rodent orthologue of human gravin, a kinase scaffold protein originally discovered as an autoantigen in some myasthenia gravis patients (3, 4). Based on their ability to bind protein kinase A RII isoforms (3), SSeCKS and gravin have been re-designated AKAP12 (A kinase anchoring protein 12). In addition to protein kinase A and protein kinase C, SSeCKS also binds with calmodulin, cyclin D1, beta-adrenergic receptor, beta-1,4 galactosyltransferase, and F-actin (5). With its ability to scaffold key signaling and cytoskeletal proteins, SSeCKS plays important roles in regulating G1 -> S phase transition and cytoskeletal organization (6, 7). Additionally, several lines of evidence suggest that SSeCKS/gravin/AKAP12 (heretofore called "SSeCKS") functions as a suppressor of tumorigenesis and metastasis. The expression of SSeCKS is down-regulated by several oncogenes and in various human epithelial tumors, including prostate, breast, ovarian, gastric, and lung (8–13). The locus of human GRAVIN gene in chromosome 6q24–25.2 is a deletion hotspot in advanced prostate, breast, and ovarian cancers (8, 14). Moreover, SSeCKS re-expression in Src-transformed fibroblasts can reverse Src-induced oncogenic growth by reducing anchorage-independent proliferation and inhibiting metastatic invasiveness through the suppression of podosome formation, most likely via the reestablishment of normal cytoskeletal architecture and the suppression of Rho family GTPase activity (15, 16). The re-expression of SSeCKS in rat Mat-LyLu prostate cancer cells slightly reduces the growth rate of primary subcutaneous tumors in nude mice but greatly suppresses the formation of lung metastases by decreasing angiogenesis through the inhibition of expression of pro-angiogenic factors such as vascular endothelial growth factor (17). The SSeCKS gene locus in human and rodents encodes three major transcripts under the control of three independent promoters, designated {alpha}, beta, and {gamma}, that are separated by 84 kb (5, 18). Two major protein isoforms of SSeCKS, {alpha} and beta, are expressed ubiquitously in most cell and tissue types (5), whereas the expression of isoform {gamma} is testes-restricted (19, 20). All proteins encoded by each transcript share >95% amino acid sequence identity encoded by a single large exon but differ only at their extreme N-terminal residues. For example, only the {alpha} isoform encodes a product that is myristoylated (2), and this modification facilitates {alpha} SSeCKS association with plasma membranes and vesicles of the endoplasmic reticulum (7, 18, 21), yet it is not sufficient for its plasma membrane targeting (22).

The mechanism by which SSeCKS is down-regulated in tumor cells has not been studied. The fact that SSeCKS is down-regulated by some oncogenes (Src, Ras, Myc, and Jun) but not by others (Raf, Mos, or Neu) suggests that this is not a generic effect in transformed cells but rather is controlled by specific mitogenic and oncogenic pathways (5). Given that down-regulation of SSeCKS is not a bystander effect during oncogenic transformation, understanding the molecular mechanism involved in SSeCKS gene silencing in the course of tumorigenesis will contribute to control tumor progression by means to reactivate SSeCKS expression.

In this study, we dissect the cis- and trans-factors responsible for the v-Src-induced repression of ssecks promoters. We find that the minimal v-Src-responsive element (VSRE)2 requires both E- and GC-boxes in the SSeCKS {alpha} proximal promoter (–106 to –49), which are bound by the transcription factors USF1 and Sp1/3, respectively. Our data indicate that v-Src-mediated transcriptional repression correlates with increased complex formation between USF1 and Sp1/3, increased binding activity of Sp1/3 to the {alpha} SSeCKS VSRE, and the recruitment of HDAC1. Moreover, although the mouse and human {alpha} and beta promoters share E- and GC-boxes in their proximal promoter, our data suggest that suppression of the beta promoter is facilitated by chromatin structure change in the {alpha} promoter 68 kb upstream.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Cell Culture—NIH3T3 mouse fibroblasts were maintained in Dulbecco's modified Eagle's medium containing 10% bovine serum. v-Src-transformed NIH3T3 cells (v-Src/3T3) were generated by retrovirus-mediated transduction. Briefly, pBabe-v-Src/puro DNA was transfected into the 293GPG packaging cell line (23) using Lipofectamine 2000 (Invitrogen). Virus was harvested 72 h after transfection and used to infect NIH3T3 cells (104/well in 12-well dishes) for 4 h. Cells were maintained in selection medium containing 2 µg/ml puromycin (Sigma). Single colonies with transformed cell morphology were picked and expanded, and then the expression of oncogenic v-Src was verified by Western blot with anti-v-Src (avian) mAb EC10, anti-Src[poY416] (BIOSOURCE), and anti-phosphotyrosine mAb-4G10 (Upstate, Charlottesville, VA) antibodies. v-Src/3T3 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% bovine serum and 2 µg/ml puromycin.

Immunofluorescence Staining—NIH3T3 or v-Src/3T3 cells grown to 60% confluence on 22-mm coverslips were fixed with 4% paraformaldehyde in PBS for 10 min and permeabilized with 0.1% Triton X-100 in PBS for another 10 min. After rinsing with PBS, cells were stained with rhodamine-conjugated phalloidin and 4',6-diamidino-2-phenylindole (1:500 dilution) (Invitrogen) for 1 h at room temperature. After three PBS washes, coverslips were mounted on glass slides in ProLong antifade reagent (Invitrogen). Fluorescent images were captured using a Nikon TE2000-E inverted microscope (Garden City, NY).

Primer Extension Analysis—Total RNA from NIH3T3 cells was extracted using TRIzol (Invitrogen) following the manufacturer's instructions. A 32P-end-labeled antisense primer (for SSeCKS {alpha},5'-AGGAGATGTGCGCCCAGGACCACAGG-3'; for SSeCKS beta,5'-CCTTCTCCTCTGTCTACTCCCGGCTAACC-3') corresponding to +66/+91 or +72/+101 relative to the transcriptional start site of SSeCKS {alpha} or beta, respectively, was hybridized at 58 °C overnight with 50 µg of total RNA (previously denatured for 3 min at 90 °C). The annealed primer was extended with SuperScript II reverse transcriptase (Invitrogen) at 42 °C for 1 h. The sizes of the extension products were determined by electrophoresis on 8% denaturing polyacrylamide gel containing 7 M urea. A 10-bp 32P-labeled ladder and sequencing reaction performed with the same primer on a genomic clone was used as a reference. The gel was dried, and the radioactive signals were identified by phosphorimaging (Storm-860, GE Healthcare).

Reporter Constructs—The SSeCKS {alpha} promoter sequence (–4920/+36) and beta promoter sequence (–4758/+119) were amplified from the BAC clone (RP11-27244) using the Triple-Master long run PCR system (Eppendorf, Westbury, NY) and then ligated into pCR-XL-TOPO vector (Invitrogen). The MluI/XhoI fragments from these plasmids were subcloned into the luciferase reporter vector pGL3-Basic (Promega, Madison WI) cut with MluI/XhoI. Progressive deletion mutants of SSeCKS promoter-luciferase constructs were created by inverse PCR with promoter-specific, MluI-flanked primers (Table 1), using the ~5-kb SSeCKS promoter/luciferase constructs as templates, followed by self-ligation. All constructs were validated by DNA sequencing.


View this table:
[in this window]
[in a new window]

  		TABLE 1
PCR primers

 
Promoter/reporter constructs with mutations in the E-box and/or GC-box were generated using the QuikChange site-directed mutagenesis kit (Stratagene, Cedar Creek, TX) according to manufacturer's instructions, using the –106/+36 {alpha} SSeCKS reporter construct as template. Nucleotide changes are indicated in italics for each primer (Table 1). All reporter mutations were confirmed by DNA sequencing.

For the SSeCKS-TK heterologous promoter constructs, a minimal TK promoter sequence was amplified from pRL-TK (Promega) with XhoI- or HindIII-flanked primers (Table 1). The PCR products were digested with XhoI/HindIII and cloned into pGL3-Basic vector cut with XhoI/HindIII to create the TKm-pGL3B luciferase reporter plasmid. The wild type SSeCKS {alpha} proximal promoter sequence between –106 and –49 and proximal promoters containing mutated E-Box and/or GC-Box were amplified with the KpnI- or XhoI-flanked primers (Table 1), digested with KpnI and XhoI, and inserted into TKm-pGL3B plasmid cut with KpnI/XhoI.

Reverse Transcription (RT)-PCR—Total RNA was isolated from NIH3T3 and v-Src/3T3 cells with or without treatment of fresh 5-azacytidine (5-aza-C) (500 nM for 72 h) and/or trichostatin A (TSA) (330 nM for 24 h) (Sigma) using TRIzol (Invitrogen). 1 µg of total RNA was subjected to reverse transcription using SuperScript first-strand synthesis system kit (Invitrogen) according to manufacturer's instructions. PCR was then performed using MJ Research PTC-200 DNA thermal cycler (Watertown, MA) with the optimized cycle numbers for each primer set (Table 1). The PCR products were separated by electrophoresis through a 1.5% agarose gel and digitally imaged using a Chemi-Genius2 Bio-Imager (Syngene, Frederick, MD).

mRNA Stability Analysis—NIH3T3 and v-Src/3T3 cells (2 x 105) seeded in 35-cm dishes were harvested at 2, 4, 8, 16, and 24 h after addition of actinomycin D (5 µg/ml) (Sigma). Total cellular RNA was extracted, and RT-PCR was performed as described above. Because the endogenous SSeCKS mRNA levels are suppressed in v-Src/3T3 cells, the PCR cycle numbers for SSeCKS {alpha} were 30 for NIH3T3 cells and 38 for v-Src/3T3 cells; for SSeCKS beta were 29 for NIH3T3 cells and 34 for v-Src/3T3 cells; and for the shared {alpha}/beta sequence were 28 for NIH3T3 cells and 32 for v-Src/3T3 cells.

Western Blot Analysis—Whole cell lysates were prepared in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate) with freshly added inhibitors (1 mM phenylmethylsulfonyl fluoride (PMSF), 10 mM NaF, 1 mM Na3VO4) and protease inhibitor mixture (Roche Applied Science). The protein concentration was determined by the Bradford protein assay (Bio-Rad). Proteins were separated on 10% SDS-polyacrylamide gels and transferred onto polyvinylidene difluoride membranes (PerkinElmer Life Sciences). The membranes were blocked with 5% nonfat milk in TBST (150 mM NaCl, 100 mM Tris-HCl, pH 7.4, 0.1% Tween 20), or with 5% bovine serum albumin in TBST for phosphoprotein blots. The following primary antibodies were used as indicated: anti-SSeCKS (2), anti-v-Src mAb EC10 (gift of Sarah Parsons, University of Virginia), anti-Src[poY416] (BIOSOURCE), anti-phosphotyrosine mAb-4G10 (Upstate), anti-FLAG and anti-actin clone AC40 (Sigma), anti-GAPDH, anti-lamin A/C, anti-USF1, anti-Sp1, anti-Sp3, anti-HDAC1, anti-HDAC2, anti-HDAC3, anti-PIAS1, and anti-HA tag (Santa Cruz Biotechnology, Santa Cruz, CA). After washing and incubating with horseradish peroxidase-labeled anti-rabbit or -mouse IgG secondary antibodies, the blots were washed, incubated with Lumi-Light chemiluminescence reagent (Roche Applied Science), and digitally imaged using a Chemi-Genius2 Bio-Imager.

Transient Transfection and Luciferase Assay—All transfections were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. For reporter assays, cells were seeded in 24-well plates at a density of 3 x 104 cells/well 24 h prior to transfection. Each transfection was performed in triplicate with 0.2 µg of promoter/reporter constructs along with 0.1 µgof Renilla luciferase reporter pRL-TK (Promega), used to normalize for transfection efficiency. Cells were harvested 48 h after transfection and lysed, and luciferase activities were measured using the dual luciferase assay kits (Promega). For overexpression experiments, a set amount of the SSeCKS {alpha} proximal promoter-reporter construct (–106/+36) was transfected in combination with various expression plasmids as indicated. The total amount of DNA transfected was normalized using the appropriate empty vectors. Data presented are representative of at least three independent experiments. The Sp1 expression vector (pFLAG-Sp1-HA) was a generous gift from Dr. Adrian Black (Roswell Park Cancer Institute). Expression vectors of pCMV-Sp3 and pN3-PIAS1, encoding a SUMO ligase, were kind gifts from Professor Guntram Suske (Institute für Molekularbiologie and Tumorforschung, Marburg, Germany). The expression vector for HA tagged SUMO-1 cDNA (HA-Sen1) was a gift kindly provided by Dr. Edward T. H. Yeh (University of Texas M. D. Anderson Cancer Center, Houston, TX). The HA-tagged HDAC1 expression plasmid (pCMV-2N3T-HDAC1) was a generous gift from Dr. Didier Trouche (Université Paul Sabatier, France).

Preparation of Nuclear Extracts—Nuclear extracts were prepared essentially as described previously (24) with minor modifications. Briefly, cells growing in 10-cm dishes (80–90% confluence) were washed twice with cold PBS and scraped into 1.5 ml of PBS. Cells were collected by centrifuging at 2,000 rpm for 2 min and then resuspending in 1 ml of ice-cold buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl2, 1 mM dithiothreitol (DTT), 0.1% Nonidet P-40, 1 mM PMSF, and Roche Applied Science protease inhibitor mixture). After incubating on ice for 20 min, the cells were homogenized with a glass Dounce (type A) by applying 30 strokes. Nuclei were collected by centrifugation at 4,000 rpm for 15 min at 4 °C. The pellets were resuspended in 200 µl of buffer C (20 mM HEPES pH 7.9, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 25% glycerol, 1 mM DTT, 1mM PMSF, and Roche Applied Science protease inhibitor mixture) and incubated on ice for 30 min with occasional agitation. The nuclear extract was centrifuged at 13,000 rpm for 20 min at 4 °C, and the supernatant was stored in aliquots at –80 °C. Protein concentrations of nuclear extracts were determined using the Bradford protein assay (Bio-Rad).

Electrophoretic Mobility Shift Assay (EMSA)—All DNA oligonucleotides used for EMSA were synthesized by Integrated DNA Technologies (Coralville, IA). Oligonucleotides were annealed, end-labeled with [{gamma}-32P]ATP by T4 polynucleotide kinase, and purified by passage through Sephadex G-50 micro-columns (Amersham Biosciences). For each reaction, 5 µgof nuclear extract was preincubated with 1.5 µg of poly(dI-dC) (Sigma) for 20 min on ice in 10 µl of binding buffer (10 mM HEPES, pH 7.9, 5 mM KCl, 50 mM NaCl, 1 mM MgCl2, 0.5 mM EDTA, 0.5 mM DTT, 0.5 mM PMSF, 10% glycerol). 0.2 ng of 32P-labeled double-stranded oligonucleotides (~20,000 cpm) was added to the 10-µl reaction mixtures and incubated for 30 min at room temperature, and then electrophoresed on 4% nondenaturing polyacrylamide gels in 0.5x Tris borate/EDTA buffer run at 4 °C. The gels were sandwiched with Whatman 3M paper, dried, and then autoradiographed overnight with an intensifying screen. In competition or supershift assays, molar excess amounts of unlabeled DNA probe or 2 µgof antibody were added to the preincubation mixtures 20 min or 1 h, respectively, prior to the addition of 32P-labeled DNA oligonucleotides.

DNA Affinity Precipitation Assay—Nuclear extracts (100 µg) from NIH3T3 or v-Src/3T3 cells were incubated at room temperature for 15 min with 0.5 nmol of 5'-biotinylated double-stranded oligonucleotides (Integrated DNA Technologies), which were previously conjugated to streptavidin-agarose beads (Sigma), in 300 µl of low salt lysis buffer (1% Triton X-100, 0.1% sodium deoxycholate, 0.05 M Tris-HCl, pH 8.1, 5 mM EDTA, pH 8.0) containing 1 mM PMSF and Roche Applied Science protease inhibitor mixture. The beads were then washed six times with low salt lysis buffer containing 150 mM NaCl and boiled in 1x electrophoresis sample buffer to elute the bound proteins before running on an SDS-polyacrylamide gel. The separated proteins were transferred to polyvinylidene difluoride membrane and immunoblotted with antibodies against HDAC1 and USF1 (Santa Cruz Biotechnology).


Figure 1
View larger version (53K):
[in this window]
[in a new window]

  		FIGURE 1.
SSeCKS transcript and protein levels are severely down-regulated by v-Src. A, schematic diagram of mouse SSeCKS gene structure. Exons 1A1 and 1A2 encode a 103-amino acid myristoylated N-terminal domain of the {alpha} SSeCKS isoform driven by the {alpha} promoter, whereas exon1B encodes an 8-amino acid nonmyristoylated domain of the beta isoform driven by the beta promoter. Both are fused to a common exon2 sequence encoding roughly 1500 additional amino acids. B, phase-contrast images of NIH3T3 and v-Src/3T3 cells at confluence (upper panel) and immunofluorescence staining of F-actin with rhodamine-conjugated phalloidin and of nuclei with 4',6-diamidino-2-phenylindole (lower panel). C, cell lysates prepared from NIH3T3 and v-Src/3T3 cells were immunoblotted with antibodies specific for avian Src (mAb-EC10), SrcpoY416, total phosphotyrosine, SSeCKS, or actin (as a loading control). Both short and long exposures are presented to show that both SSeCKS isoforms are present in v-Src/3T3 lysates. D, steady-state SSeCKS mRNA levels in NIH3T3 and v-Src/3T3 cells were determined by semi-quantitative RT-PCR using isoform-specific or common {alpha}/beta primers. Actin-specific primers were used as a control.

 
Chromatin Immunoprecipitation (ChIP)—ChIP assays were performed following the protocol outlined by the manufacturer (Upstate) with minor modifications. Briefly, NIH3T3 and v-Src/3T3 cells growing in 10-cm dishes (80–90% confluence) were fixed in culture medium with formaldehyde (final concentration of 1%) for 10 min at 37 °C. After washing twice with cold PBS, cells were collected by centrifuging at 2,000 rpm for 2 min, resuspended in 0.8 ml of SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.1) containing 1 mM PMSF and Roche Applied Science protease inhibitor mixture, and incubated for 20 min on ice. DNA was sheared into ~200–1000-bp fragments by five 20-s sonications, followed by centrifugation for 10 min at 13,000 rpm at 4 °C to remove debris. Supernatant fractions were diluted 5-fold in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH 8.1, 167 mM NaCl) containing 1 mM PMSF and Roche Applied Science protease inhibitor mixture. Each chromatin fraction (1.2 ml per immunoprecipitation) was precleared with 60 µlof salmon sperm DNA-protein A-agarose beads (Upstate) for 1 h at 4 °C and then incubated at 4 °C overnight with 5 µgofthe following antibodies as indicated: anti-USF1, anti-Sp1, anti-Sp3, and normal rabbit IgG (Santa Cruz Biotechnology), anti-acetyl-histone H4 and anti-acetyl-histone H3 (Upstate). Immune complexes were isolated by binding to 40 µl of salmon sperm DNA-protein A-agarose beads for 1 h at 4 °C, and by washing sequentially with low salt buffer (50 mM Tris-HCl, pH 8.0, 0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet P-40, 1 mM EDTA, 150 mM NaCl), high salt buffer (50 mM Tris-HCl, pH 8.0, 0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet P-40, 1 mM EDTA, 500 mM NaCl), LiCl buffer (50 mM Tris-HCl, pH 8.0, 0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet P-40, 1 mM EDTA, 250 mM LiCl), and then twice with TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA). DNA-protein complexes were eluted twice with 250 µlof 1% SDS and 0.1 M NaHCO3 and incubated at 65 °C for 4 h to reverse the cross-linking. Proteins were digested with proteinase K, and DNA was recovered by phenol/chloroform extraction and ethanol precipitation with 20 µgof glycogen as carrier. Primers for PCR amplification of SSeCKS {alpha} proximal promoter sequence between –270 and +33 were 5'-TGCTGCTCCTGAACCTTCTG-3' and 5'-GATCCTGCTGAGAACACACC-3'. SSeCKS beta proximal promoter sequence between –248 and +43 were 5'-GTGCCAGGGATGAAGTCACC-3' and 5'-GAGCATCAAGGAAGCTCTCC-3'. PCR products were resolved on 2% agarose gels and stained with ethidium bromide, and the images were digitized with a Chemi-Genius2 Bio-Imager.

siRNA Experiments—ON-TARGET plus SMART pool siRNA specific for murine HDAC1 was purchased from Dharmacon (Lafayette, CO). v-Src/3T3 cells plated in 6-well plates (1 x 105 cells/well) were transfected with 200 nM siRNA-HDAC1 using Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. After 72 h, cells were harvested for Western blot analysis and semi-quantitative RT-PCR.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
SSeCKS mRNA and Protein Levels Are Severely Down-regulated in v-Src-transformed NIH3T3 Cells—The mouse ssecks gene locus encodes two major transcript isoforms, {alpha} and beta (the testes-specific isoform {gamma} was not studied here), under the control of two independent promoters. As shown in Fig. 1A, exon 1A1 and exon 1A2 encode a 103-amino acid myristoylated N-terminal domain driven by the TATA-less {alpha} promoter. Exon1B encodes an 8-amino acid nonmyristoylated domain driven by the TATA-containing beta promoter. Both are fused to the common exon 2 encoding the remaining 1494 amino acids, and exon 3, which contains the 3'-untranslated region. The {alpha} and beta promoters are separated by 68 kb. Compared with untransformed NIH3T3 cells, NIH3T3 cells transduced with the v-Src oncogene (v-Src/3T3) are refractile, deficient in contact-inhibited growth, and lack F-actin stress fibers (Fig. 1B). Western blot analysis showed avian v-Src protein was specifically expressed in v-Src/3T3 cells (using mAb-EC10), resulting in dramatic increases in Src autophosphorylation (poY416) and total cellular tyrosine phosphorylation in v-Src/3T3 cells compared with the control NIH3T3 cells. This indicates that Src is constitutively activated in v-Src/3T3 cells (Fig. 1C). Consistent with our previous studies, the abundance of both SSeCKS {alpha} and beta protein isoforms was decreased markedly in v-Src/3T3 cells, although both SSeCKS isoforms could be detected in v-Src/3T3 lysates after longer exposures as shown in Fig. 1C. Semi-quantitative RT-PCR analysis with isoform-specific primers (Table 1) showed that the down-regulated levels of SSeCKS {alpha} and beta mRNAs (Fig. 1D) correlated with decreases in {alpha} and beta protein levels in v-Src/3T3 cells (Fig. 1C). Similar decreases in SSeCKS protein and mRNA levels by v-Src were found in at least three independent v-Src/3T3 clones and one v-Src-transformed mouse embryonic fibroblast (v-Src/MEF) cell line (data not shown). We previously showed that v-Src decreased SSeCKS transcript and protein levels to similar extents, based on Northern and Western blots (1, 2, 25). Thus, SSeCKS abundance is likely controlled by v-Src at the level of transcription. However, for Fig. 1D, we chose PCR cycle numbers that allowed the simultaneous visualization of SSeCKS isoform transcript levels in NIH3T3 and v-Src/3T3 cells, conditions that will allow us in the experiments below to gauge treatments that could derepress SSeCKS transcription in v-Src/3T3 cells.

Decreased SSeCKS mRNA Steady-state Levels in v-Src/3T3 Cells Are Not Mediated by Alteration in mRNA Stabilities—Because mRNA steady-state levels can be controlled by both mRNA synthesis rate and post-transcriptional mRNA stability (mRNA degradation rate), we examined whether the v-Src-induced down-regulation of SSeCKS was because of changes in mRNA stability. NIH3T3 and v-Src/3T3 cells were treated with actinomycin D to inhibit transcriptional initiation, and then SSeCKS mRNA levels were determined at various time points by semi-quantitative RT-PCR. The PCR products were resolved on 2% agarose gel, visualized by ethidium bromide staining, and quantified by densitometry analysis (Fig. 2A). The mRNA decay slopes reveal that there was no significant difference in the degradation rates of either SSeCKS mRNA isoform in NIH3T3 cells versus v-Src/3T3 cells (Fig. 2B). Therefore, v-Src-induced down-regulation of SSeCKS mRNA was not mediated by decreasing mRNA stability, suggesting that SSeCKS message abundance is controlled by repression of promoter activity.

Mapping of Transcriptional Start Sites (TSS) of SSeCKS—As a first step toward to studying SSeCKS promoter control mechanisms, we mapped the SSeCKS TSS using primer extension analysis. As shown in Fig. 3, specific extension bands were detected for SSeCKS {alpha} or beta isoforms, suggesting that each isoform has a single major TSS in NIH3T3 cells. The precise location of the TSS was obtained by running in parallel a sequencing reaction performed with the same primer used in the extension assay. The nucleotide C in the TTCAT motif was defined as the TSS (designated +1) of the SSeCKS {alpha} isoform (Fig. 3A), and the nucleotide C in the GTGCG motif was identified as the TSS of the beta isoform (Fig. 3B). The start sites identified here are very close to those identified by Streb et al. (18) in rat smooth muscle cells (3 bp downstream for {alpha} and 8 bp downstream for beta) using a5'-rapid amplification of cDNA ends technique.


Figure 2
View larger version (26K):
[in this window]
[in a new window]

  		FIGURE 2.
v-Src does not alter SSeCKS mRNA stability. A, total RNA was extracted from NIH3T3 or v-Src/3T3 cells following actinomycin D (5 µg/ml) treatment for the times indicated and analyzed for SSeCKS mRNA levels by RT-PCR. B, SSeCKS mRNA levels were determined by densitometric analysis and were plotted as decay curves. Because of the lower level of SSeCKS mRNA in v-Src/3T3 relative to NIH3T3 cells, more PCR cycles were used for v-Src/3T3 cells (see "Experimental Procedures") to get comparable amplifications. Calculated slopes were shown for comparison in the SSeCKS mRNA degradation rates between NIH3T3 and v-Src/3T3 cells.

 


Figure 3
View larger version (94K):
[in this window]
[in a new window]

  		FIGURE 3.
Mapping of the SSeCKS {alpha} and beta TSS in NIH3T3 cells by primer extension analysis. Total RNA (50 µg) isolated from NIH3T3 cells, or an equal volume of H2O as a control, was hybridized to a 32P-end-labeled antisense oligonucleotide primer specific to SSeCKS {alpha} (A) or beta (B) mRNA. The annealed primers were elongated with reverse transcriptase and then resolved on a denaturing 8% polyacrylamide gel along with a 10-bp DNA marker (M) and sequencing ladders. The annealed sequences of each primer used in the assay are underlined, and the nucleotide corresponding to the TSS is designated as +1. Arrows, major extension bands.

 
Identification of VSRE in SSeCKS Promoters—Roughly 5 kb of promoter region upstream of the TSS of each isoform (–4920/+36 for isoform {alpha} and –4758/+119 for isoform beta) was cloned into the pGL3-Basic luciferase reporter vector, and in order to locate the VSRE, we generated a series of progressive promoter deletion mutants. These full-length promoter and deletion constructs were then transiently co-transfected into NIH3T3 and v-Src/3T3 cells along with a Renilla luciferase reporter driven by the TK promoter, representing a normalization control. As shown in Fig. 4A, the –4920/+36 SSeCKS {alpha} promoter sequence induced robust luciferase activities in both cells and, more importantly, exhibited considerably lower relative promoter activity in v-Src/3T3 cells. Truncation of the sequence between –4920 and –2677 enhanced promoter activities in both cells but did not affect the v-Src responsiveness, suggesting that this region contains v-Src-independent repressors. Deletions of the sequence between –2677 and –106 reduced the promoter activities in both cell types but still had no effect on the v-Src-mediated repression. However, truncation of the sequence between –106 and –26 abolished both VSR and basal promoter activities. These findings were observed in at least two independently derived v-Src/3T3 clones and another v-Src/MEF cell line (data not shown). These results demonstrate that the SSeCKS {alpha} proximal promoter region between –106 and –26 encodes the minimal VSRE and also contains the minimal cis-acting sequences required for basal promoter activity. Interestingly, a luciferase-reporter construct containing roughly 5 kb of the human {alpha} SSeCKS (Gravin) promoter was down-regulated in LNCaP and C-42 prostate cancer cells compared with untransformed, immortalized P69SV40T human prostate epithelial cells (data not shown), suggesting that promoter sequences controlling down-regulation in cancer are also present in the human {alpha} SSeCKS allele.


Figure 4
View larger version (38K):
[in this window]
[in a new window]

  		FIGURE 4.
Mapping of VSRE in SSeCKS{alpha} andbeta promoters. A and B, NIH3T3 and v-Src/3T3 cells were transiently co-transfected with a deletion series of the SSeCKS {alpha} (A) andbeta promoter (B) sequences cloned into pGL3B-luciferase vectors along with pRL-TK (as a normalization control for transfection efficiency and cell proliferation rates). Cells lysates were produced after 48 h and then assayed for luciferase activity. The normalized luciferase value for each assay is shown as the mean of three replicates ± S.D. The representative result of three independent experiments is shown. C, DNA sequence alignment of the SSeCKS{alpha} proximal promoters from mouse, human, chimp, rat, and dog. Asterisks indicate the conserved nucleotides among all five species. Consensus E- and GC-boxes are highlighted.

 
To our surprise, neither the –4758/+119 promoter sequence of SSeCKS beta isoform nor any of its deletion constructs exhibited v-Src-mediated repression in the luciferase assay, although the region proximal to the TSS (–157/+119) was sufficient to encode adequate promoter activity (Fig. 4B). As with the {alpha} promoter, the distal beta promoter region from –4758 to –2482 seems to harbor v-Src-independent control sequences that repress the basal activity.

Sequence conservation across different species, especially in promoter regions, strongly suggests common regulatory mechanisms. We carried out a cross-species comparison of SSeCKS {alpha} proximal promoter sequences shown in Fig. 4A to be sufficient for VSR and basal promoter activity. As shown in Fig. 4C, the {alpha} proximal promoter sequence was highly conserved between mouse, human, chimp, rat, and dog, and moreover, there was equal E- and GC-box spacing relative to the TSS, further strengthening its functional importance in regulating gene expression in the context of the {alpha} promoter.

E- and GC-boxes in the SSeCKS {alpha} Proximal Promoter, Bound by USF1 and Sp1/3, Respectively, Are Crucial for the v-Src Responsiveness—To identify the precise regulatory elements responsible for VSR activity, we generated fine deletion constructs with the SSeCKS {alpha} proximal promoter and then performed luciferase assays. As shown in Fig. 5A, deletion of the region between –106 and –89, which contains the consensus E-box, markedly reduced promoter activities in both NIH3T3 and v-Src/3T3 cells and abrogated v-Src responsiveness. Further deletion of the region between –89 and –67, which contains a GC-box site, resulted in a similar decrease in the basal promoter activities. These results indicate that the E-box in the proximal promoter is critical for the v-Src responsiveness.

We investigated the transcriptional factors binding to the {alpha} proximal promoter by EMSA with three overlapping synthetic DNA oligonucleotides, spanning the sequence between –106 and –26 (Fig. 5B). Equal amounts of total nuclear extract (5 µg) prepared from NIH3T3 and v-Src/3T3 cells were used in each assay. As shown in Fig. 5C, oligo-1 (–106/–71) facilitated the formation of two DNA-protein complexes, C1 and C2, and two major binding complexes, C2 and C3, were formed to oligo-2 (–85/–47), but no significant gel shifts were detected using oligo-3 (–63/–26). All three DNA-protein complexes (C1, C2, and C3) were specific as demonstrated by competition experiments with molar excesses of specific and nonspecific unlabeled oligonucleotides (Fig. 5D). The binding activity of protein complex C1 to oligo-1 was comparable between the NIH3T3 and v-Src/3T3 cells (Fig. 5C, lanes 1 and 2). In contrast, the binding complexes of C2 and C3 were formed preferentially in v-Src/3T3 cells compared with NIH3T3 cells (Fig. 5C, lanes 3 and 4), suggesting that these two complexes correlate with v-Src responsiveness. These findings were also observed using nuclear extract prepared from another independent v-Src/3T3 clone (data not shown). The presence of C1 complex solely in oligo-1 but not in oligo-2 strongly suggested that the binding site of C1 was in the nonoverlapping region of oligo-1 (e.g. the E-box). Given that C2 co-migrated with either oligo-1 or -2, but that it showed stronger binding with oligo-2, it is possible that the C2 binding is at the overlap of oligo-1 and -2 (e.g. the GC-box). To test whether the binding site of C1 is the E-box, and C2 is the GC-box, we performed EMSAs with oligo-4 and oligo-4m (Fig. 6A), versions of oligo-1 lacking the GC-box site at its 3'-end or encoding a mutated E-box site. As shown in Fig. 6B, oligo-4 failed to induce the formation of complex C2, although it had no effect on the formation of complex C1 (lanes 3 and 4), strongly suggesting that the binding site of C2 induced by oligo-1 was at the GC-box. Mutation of the E-box in oligo-4m completely abrogated the formation of complex C1 (Fig. 6B, lanes 5 and 6), strongly suggesting that the protein-DNA interaction site of C1 was at the E-box. In addition, mutation of the GC box in oligo-2 (oligo-2m) abolished the formation of both complex C2 and C3 (Fig. 6B, lanes 9 and 10), showing that the DNA-protein interaction site of complex C2 or C3 with oligo-2 was likely at the GC-box.


Figure 5
View larger version (48K):
[in this window]
[in a new window]

  		FIGURE 5.
VSRE are located in the {alpha} SSeCKS proximal promoter. A, NIH3T3 or v-Src/3T3 cells were transfected with various SSeCKS {alpha} proximal promoter-luciferase reporter constructs along with pRL-TK, and lysates isolated after 48 h were assayed for luciferase activity. The normalized luciferase value for each assay is shown as the mean of three replicates ± S.D. The representative result of three independent experiments is shown. B, DNA sequence of the SSeCKS {alpha} proximal promoter from nucleotides –106 to –26. The three overlapping oligonucleotides spanning this region used in the EMSA are indicated by solid or dotted lines. C, EMSA analysis was carried out with 32P-labeled oligonucleotides and nuclear extracts (NE)(5 µg/lane) prepared from NIH3T3 or v-Src/3T3 cells. The two DNA-protein complexes formed with oligo-1 were designated as C1 and C2, and the two major binding complexes formed with oligo-2 were designated as C2 and C3. D, nuclear extracts prepared from v-Src/3T3 cells were used in competition assays with molar excess of unlabeled nonspecific (NS) or specific oligonucleotides as indicated.

 
The c-Myc- and c-Myc-related family of proteins, such as USF1 and USF2, is among the helix-loop-helix transcriptional factors known to bind E-box sequences (26, 27). Moreover, it has been reported that USF1 binds to the E-box on {alpha} SSeCKS promoter in Rat-2 fibroblasts (18). To determine whether c-Myc or USF1 was the E-box-binding protein that forms the C1 complex, we performed supershift EMSAs with antibodies (Ab) specific for c-Myc or USF1. Addition of the anti-USF1 Ab supershifted the C1 complex (Fig. 6B, ss), whereas preimmune IgG control had no effect on the migration of C1 (Fig. 6B, lanes 13 and 14). In contrast, Ab against c-Myc failed to either compete or supershift the C1 complex (data not shown). These data indicate that USF1, but not c-Myc, binds in vitro to the SSeCKS {alpha} proximal promoter at the E-box.

GC-boxes are known to be bound by the Sp and Krüppel-like factor families of transcription factors (28). We then examined whether Sp1 or Sp3 were present in the binding complexes, C2 and C3. Upon more careful resolution, the thick band of C2 formed with oligo-2 in EMSA (Fig. 5C, lanes 3 and 4) can be separated into several complexes as follows: a doublet, designated C2a and C2b, and a faster migration C2c (Fig. 6B, lanes 15 and 16). Addition of anti-Sp1 Ab led to a supershift of C2a but had little effect on C2b, C2c, or C3 (Fig. 6B, lane 17), whereas anti-Sp3 Ab resulted in supershift of C2b and of the entire C3 (Fig. 6B, lane 18). Neither Ab affected the migration of C2c. Addition of both Abs supershifted C2a, C2b, and C3 but still had little effect on C2c (Fig. 6B, lane 19). These results strongly suggest that the DNA-protein complexes formed at the GC-box primarily contain Sp1 and Sp3; whether C2c represents another Sp family member or an additional non-Sp complex or nonspecific binding is still unclear. It should be noted that the C2c complex is also formed preferentially in v-Src/3T3 cells. In addition, a ChIP assay confirmed the EMSA findings, namely that binding to the SSeCKS {alpha} proximal promoter by Sp1 and Sp3, but not by USF1, is enhanced in v-Src/3T3 cells compared with NIH3T3 cells (Fig. 6C). Importantly, the increased binding of Sp1 and Sp3 to the GC box in v-Src/3T3 cells was not because of either increased total protein expression or increased nuclear localization of these proteins in v-Src/3T3 cells (Fig. 6D). The lamin A/C protein level is used merely as a marker of nuclear preparation; it cannot be used as a loading control because its relative abundance is altered by v-Src (as is true for many other typical loading control proteins).3


Figure 6
View larger version (63K):
[in this window]
[in a new window]

  		FIGURE 6.
E- and GC-boxes in the SSeCKS {alpha} proximal promoter, bound by USF1 and Sp1/3, respectively, are crucial for VSR activity. A, sequences of the five oligonucleotides used in the EMSA, containing wild type E- or GC-boxes (oligo 1, -2, and -4) or mutated E- or GC-boxes (oligo-2m, -4m; only the mutated nucleotides are shown). B, EMSA analysis using 32P-labeled oligonucleotides shown in A. Supershift experiments were carried out by adding antibodies specific for USF1, Sp1, and/or Sp3 (or IgG controls), as indicated. SS, supershift. C, ChIP assay. NIH3T3 or v-Src/3T3 cells were treated with formaldehyde, lysed, and subjected to immunoprecipitation with antibodies specific for USF1, Sp1, Sp3, or IgG control. SSeCKS {alpha} proximal promoter sequences (–270 to +33) were amplified by PCR from the immunoprecipitates. D, whole cell lysates or nuclear extracts prepared from NIH3T3 or v-Src/3T3 cells were immunoblotted with antibodies specific for USF1, Sp1, or Sp3. Actin and GAPDH blots are shown as loading controls for whole cell lysates; lamin A/C is a marker of nuclear preparation. Note that v-Src typically decreases actin 2-fold, increases GAPDH 2-fold, and increases lamin A/C 2–3-fold (Y. Bu and I. H. Gelman, unpublished observations); and thus, protein-loading normalization usually requires analyzing several proteins. E, NIH3T3 or v-Src/3T3 cells were co-transfected with the wild type –106/+36 SSeCKS {alpha} promoter/luciferase construct or similar reporter constructs containing mutated E- or GC-boxes, or both (top panel), along with pRL-TK. Cell lysates were harvested after 48 h and assayed for luciferase activity (bottom panel). The normalized luciferase value for each assay is shown as the mean of three replicates ± S.D. The representative result of three independent experiments is shown.

 
We next investigated the functional importance of E- and GC-boxes in mediating v-Src responsiveness. Thus, proximal {alpha} promoter-luciferase constructs were produced that incorporate the mutations in oligo-4m (loss of USF1 binding), oligo-2m (loss of Sp1/3 binding), or both (Fig. 6E). Loss of either USF1 or Sp1/3 binding ablated both VSR and basal promoter activities, and mutation to both boxes (DM) decreased these activities roughly 50% more (Fig. 6E). Taken together with the increased binding of Sp1/3 to the GC-box in v-Src/3T3 cells (Fig. 6B), these data strongly suggest that Sp1/3 encode dual regulatory roles for this promoter as follows: as inducers of basal promoter activity in both untransformed and v-Src-transformed cells and as repressors of promoter activity in v-Src/3T3 cells.

Although the beta proximal promoter also contains E- and GC-boxes spaced similarly upstream of the TSS as in the {alpha} promoter, these boxes seem not to be sufficient for the VSR activity in the beta promoter. This suggests that VSR activity is governed by E-/GC-box spacing constraints and/or by the E-/GC-boxes plus other sequences found only in the {alpha} proximal promoter.

The SSeCKS {alpha} Proximal Promoter Is Sufficient to Confer VSR to a Heterologous Promoter—To test whether the SSeCKS {alpha} proximal sequence between –106 and –49, containing E- and GC-boxes, is sufficient to confer VSR activity, this sequence was spliced immediately upstream of a minimal TK promoter driving firefly luciferase. As shown in Fig. 7, this 58-bp sequence was sufficient to induce VSR to the heterologous TK promoter. Moreover, mutation of either the E- or GC-box in this sequence abrogated its ability to confer VSR. Collectively, therefore, the E- and GC-boxes in the context of the SSeCKS {alpha} proximal promoter were both necessary and sufficient for VSR activity.


Figure 7
View larger version (20K):
[in this window]
[in a new window]

  		FIGURE 7.
The SSeCKS {alpha} proximal promoter is sufficient to confer VSR activity to a heterologous TK promoter. Either the wild type SSeCKS {alpha} proximal promoter (–106 to –49), or those containing E- and/or GC-box mutations (top panel) were fused upstream of the minimal TK promoter driving a luciferase reporter and used to co-transfect NIH3T3 or v-Src/3T3 cells along with pRL-TK control plasmid. Cell lysates were prepared after 48 h and assayed for luciferase activity. The normalized luciferase value for each assay is shown as the mean of three replicates ± S.D. Data shown are representative of three independent experiments.

 
Effects of Sp1 and/or Sp3 Overexpression on SSeCKS {alpha} Proximal Promoter Activity in NIH3T3 and v-Src/3T3 Cells—Given that the GC-box supports VSR activity and that increased Sp1/Sp3 binding to the GC-box was observed in EMSA using v-Src/3T3 lysates, we examined whether overexpression of Sp1 and/or Sp3 was sufficient for transcriptional repression in NIH3T3 and v-Src/3T3 cells. Thus, we co-transfected increasing amounts of Sp1 or Sp3 expression plasmid (from 0.01 to 0.4 µg) with a set amount of the –106/+36 {alpha} promoter construct. As shown in Fig. 8A, overexpression of Sp1 led to a dose-dependent increase in promoter activity in both cells. In contrast, only high overexpression levels of Sp3 increased promoter activity in NIH3T3 cells but had little effect in the v-Src/3T3 cells (Fig. 8B). Because Sp3 can exert transcriptional inhibition by competitively antagonizing the action of Sp1 (29), and sumoylation of Sp3 potentiates its repressive activity (30), we reasoned that the failure of Sp3 to repress the SSeCKS {alpha} promoter activity could be because of limited sumoylation of the exogenous Sp3. However, inclusion of expression vectors for SUMO and for the PIAS1 ubiquitin-protein isopeptide ligase-3, required to conjugate SUMO to Sp3 (31), led to increased promoter activity in both cells, albeit to a much lesser extent in v-Src/3T3 cells (Fig. 8C). In contrast, increasing levels of Sp3 (0.1–0.4 µg) co-transfected with a set level of Sp1 (0.1 µg) resulted in the reduction of the Sp1-mediated transactivation in a dose-dependent manner in v-Src/3T3 cells but not in NIH3T3 cells (Fig. 8D). The exogenous expression of the Sp1, Sp3, PIAS1, and SUMO proteins was confirmed by immunoblotting (supplemental Fig. 1). Taken together, these results indicate that Sp1 alone is a strong activator of the SSeCKS {alpha} promoter in both NIH3T3 and v-Src/3T3 cells, whereas Sp3 is a weak activator alone in both cells, but in the presence of both Sp1 and Sp3, Sp3 can antagonize the Sp1-mediated transactivation of {alpha} promoter activity only in v-Src/3T3 cells. This suggests that v-Src converts the Sp1/3 complex from an activator to a repressor, possible through post-translational modifications or through the induction of co-repressors.

VSR Activity Correlates with Changes in Chromatin Structure—The transient luciferase assay (Fig. 4A) only partly recapitulated the repression of SSeCKS transcripts at the endogenous level (Fig. 1D). This suggests that VSR activity of the endogenous {alpha} promoter may also be controlled by epigenetic mechanisms, such as DNA methylation or changes in chromatin structure, that are not manifest on exogenous reporter plasmids during a transient expression assay. Therefore, we examined whether v-Src-mediated down-regulation of SSeCKS transcription was controlled by epigenetic mechanisms. Indeed, previous reports indicated that AKAP12/Gravin is inactivated by promoter hypermethylation in gastric and colon cancer (12, 32). RT-PCR analysis showed that treatment of v-Src/3T3 cells with the DNA methyltransferase inhibitor, 5-aza-C, failed to restore the steady mRNA levels of either isoform of SSeCKS (Fig. 9A, lanes 3 and 4), indicating that DNA methylation was not involved in the v-Src-mediated down-regulation of SSeCKS. Moreover, we examined the methylation status of CpG islands in the SSeCKS proximal promoter sequences of the {alpha} isoform (see Table 1 for primers) by a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry approach (33) (Sequenom MassARRAY, RPCI Microarray and Genomics Core Facility), and we found no significant differences between control NIH3T3 and v-Src/3T3 cells (data not shown). In contrast, inhibition of histone deacetylase (HDAC) activities by treatment with TSA partly restored the steady mRNA level of the {alpha} isoform and fully restored beta isoform in v-Src/3T3 cells (Fig. 9A, lanes 5 and 6), whereas combining with 5-aza-C treatment had little additive effect (Fig. 9A, lanes 7 and 8). Similar data were also found in another independent v-Src/3T3 clone and a v-Src/MEF cell line (data not shown). These data suggest that histone deacetylation, but not DNA methylation, plays a role in v-Src-mediated down-regulation of SSeCKS. This finding agrees with Rombouts et al. (34) who showed that TSA could derepress SSeCKS in a model of hepatic injury.

Because TSA treatment is known to affect the expression of a wide range of genes, the derepression of SSeCKS in v-Src/3T3 cells induced by TSA could be indirect, i.e. not because of a direct increase in acetylated histones on the SSeCKS promoter. To test this possibility, we performed ChIP assays to examine the acetylation status of histone H3 and histone H4 markers, often associated with a more transcriptionally active chromatin structure (35, 36), on the SSeCKS proximal promoters in NIH3T3 and v-Src/3T3 cells. As shown in Fig. 9B, the degree of Ac-H3 and Ac-H4 binding to the SSeCKS {alpha} promoter was lower in v-Src/3T3 than that in the control NIH3T3 cells. Moreover, TSA treatment increased the binding of Ac-H3 and Ac-H4 in the v-Src/3T3 cells, showing that the derepression of SSeCKS {alpha} isoform by TSA in v-Src transformed cells is controlled by an increase in histone acetylation levels. In contrast, there was no difference in Ac-H3 and Ac-H4 binding levels on the proximal beta promoter between the two cells, and correspondingly, TSA treatment did not alter Ac-H3 and Ac-H4 binding levels in v-Src/3T3 cells. This lack of change to the chromatinization of the beta promoter contrasts with the finding in Fig. 9A that TSA strongly derepressed the transcription of beta-SSeCKS in v-Src/3T3 cells, yet it agrees with our finding that the exogenously expressed beta promoter failed to be down-regulated in v-Src/3T3 cells (Fig. 4B). These data suggest that the repression of beta isoform may not be controlled by its own promoter but rather through a coordinate regulation of chromatin structure at the {alpha} promoter 68 kb upstream. Indeed, Streb and Miano (37) show evidence that the serum responsiveness of the beta promoter could be controlled by the CArG box found in the {alpha} promoter. It cannot be ruled out, however, that the TSA-induced derepression of the beta isoform was caused by the induction of transcription activators specific for the beta promoter.


Figure 8
View larger version (37K):
[in this window]
[in a new window]

  		FIGURE 8.
Effects of overexpression of Sp1 and/or Sp3 on SSeCKS{alpha} proximal promoter activity in NIH3T3 and v-Src/3T3 cells. A and B, effect of Sp1 (A) or Sp3 (B) on SSeCKS {alpha} proximal promoter activity, as determined by luciferase assay. NIH3T3 (white bars) or v-Src/3T3 cells (black bars) were transfected with the SSeCKS {alpha} proximal promoter-luciferase alone or with varying amounts of Sp1 (A) or Sp3 (B) expression plasmids as indicated. Cell lysates were prepared after 48 h and assayed for luciferase activity. C, effect of Sp3 sumoylation on the {alpha} proximal promoter. The SSeCKS {alpha} proximal promoter-luciferase construct was transfected into NIH3T3 or v-Src/3T3 cells alone or in combination with expression plasmids for Sp3, the PIAS1 ubiquitin-protein isopeptide ligase-3, and SUMO-1, as indicated. D, effect of Sp3 on the Sp1-mediated activation of the {alpha} proximal promoter. The SSeCKS {alpha} proximal promoter-luciferase construct was co-transfected into NIH3T3 or v-Src/3T3 cells with 0.1 µg of Sp1 expression plasmid plus varying amounts of Sp3 expression plasmid, as indicated. For all assays, the luciferase activity values are expressed as fold induction of the reporter alone, arbitrarily assigned a value of 1. Each bar is the mean of three replicates ± S.D. Data shown are representative of three independent experiments.

 
Given that hypo-acetylation of histones contributed to the VSR activity of the {alpha} promoter in v-Src/3T3 cells, we examined whether this could be due to changes in the abundance of HDAC1, -2, or -3. As shown in Fig. 9C, HDAC1 protein levels were increased in either whole cell lysates or nuclear extracts from v-Src/3T3 cells, whereas no significant differences in HDAC2 and HDAC3 levels were observed. Given that Sp1 and Sp3 can directly interact with HDAC1 (38–40) and that more Sp1/Sp3 bound to the GC-box on the SSeCKS {alpha} proximal promoter in v-Src/3T3 cells (Fig. 6B), we examined whether there is increased recruitment of HDAC1 to the {alpha} proximal promoter in v-Src/3T3 cells by performing a DNA affinity precipitation assay. As predicted, the oligonucleotides containing both E- and GC-boxes (–106/–47) and the one containing only the nonmutated GC-box (–85/–47 WT) pulled down an increased amount of HDAC1 from v-Src/3T3 lysates compared with lysates from NIH3T3 cells (Fig. 9D). Moreover, mutation in the GC-box significantly reduced its ability to interact with HDAC1 in both cells, indicating the HDAC1 binding to the SSeCKS {alpha} proximal promoter is dependent on the GC-box. We attempted to perform a ChIP assay to confirm these findings from the DNA affinity precipitation assay, but we were not able to detect any significant chromatin precipitation after normalization with the IgG control using the available HDAC1 antibody. However, the RNA interference-mediated knockdown of HDAC1 expression in v-Src/3T3 cells resulted in significant increases in SSeCKS mRNA levels of both isoforms (Fig. 9E), indicating that the co-repressor HDAC1 is involved in the transcriptional repression of SSeCKS in v-Src transformed cells. Taken together, these results suggest that VSR activity is mediated by the increased recruitment of the HDAC1 co-repressor via the enhanced binding of Sp1/Sp3 to the GC-box on the SSeCKS {alpha} proximal promoter.


Figure 9
View larger version (58K):
[in this window]
[in a new window]

  		FIGURE 9.
Histone deacetylation is involved in the repression of SSeCKS transcription in v-Src/3T3 cells. A, total RNA was isolated from NIH3T3 or v-Src/3T3 cells following treatment with vehicle (DMSO) or the fresh DNA methylation inhibitor 5-aza-C (500 nM for 72 h), and/or the histone deacetylase inhibitor TSA (330 nM for 24 h) as indicated and analyzed for SSeCKS mRNA levels by semi-quantitative RT-PCR. B, ChIP assay. After treating NIH3T3 or v-Src/3T3 cells with vehicle (DMSO) or TSA (330 nM for 24 h), chromatin was cross-linked by formaldehyde and then subjected to immunoprecipitation with antibodies specific for Ac-H3 or Ac-H4 as described under "Experimental Procedures." SSeCKS proximal promoter sequences (–270 to +33 for {alpha}, –248 to +43 for beta) were amplified by PCR from the immunoprecipitates. C, whole cell lysates or nuclear extracts prepared from NIH3T3 and v-Src/3T3 cells were immunoblotted with antibodies against HDAC1, HDAC2, and HDAC3. Actin and GAPDH blots are shown as loading controls for whole cell lysates; lamin A/C is a marker of nuclear preparation. D, nuclear extracts (100 µg) from NIH3T3 or v-Src/3T3 cells were incubated with streptavidin-agarose beads coupled with or without biotin-labeled oligonucleotides containing SSeCKS {alpha} proximal promoter sequences (WT –106 to –47, WT –85 to –47, or –85 to –47 with a mutated GC-box). The precipitated protein complexes were immunoblotted with antibodies specific for HDAC1 or USF1. E,v-Src/3T3 cells were transfected with siRNA-HDAC1 (200 nM) or untransfected (mock) for 72 h. HDAC1 protein level was determined using Western blot analysis and GAPDH was the loading control. SSeCKS mRNA levels were analyzed by semi-quantitative RT-PCR using isoform-specific or common {alpha}/beta primers.

 
Sp1/3 Potentiates the Repressive Activity of HDAC1 on the {alpha} Promoter—To support the notion that the recruitment of HDAC1 by Sp1/3 to the SSeCKS {alpha} promoter is involved in VSR activity, we co-transfected the –106/+36 reporter construct (0.2 µg) with increasing amounts of an HDAC1 expression plasmid (0.1–0.4 µg) alone or in combination with Sp1 and/or Sp3 expression plasmids (0.05 µg) into NIH3T3 or v-Src/3T3 cells. As shown in Fig. 10, the overexpression of HDAC1 alone led to a dose-dependent repression of the {alpha} proximal promoter activity in both NIH3T3 and v-Src/3T3 cells. The potency of this repression was ~2-fold higher in v-Src/3T3 cells, possibly because the enhanced binding of Sp1/Sp3 facilitated increased recruitment of HDAC1 to the {alpha} proximal promoter in the v-Src/3T3 cells. Moreover, inclusion of Sp1 or Sp3 potentiated the HDAC1-mediated repression in both cells; the combination of Sp1 and Sp3 had an effect, which was additive at best. Additionally, Sp3 consistently was a more potent co-repressor, especially at the higher concentrations of HDAC1. The exogenous expression of the HDAC1 protein was confirmed by immunoblotting (supplemental Fig. 1). These data strongly suggest that HDAC1 participates in VSR activity of the SSeCKS {alpha} promoter via an enhanced Sp1/Sp3 binding to the proximal promoter.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
A wide range of genes have been shown to be up-regulated (41–45) or down-regulated (1, 46–51) in cells oncogenically transformed by v-Src. Some of these v-Src down-regulated genes, including ssecks, are recognized as tumor suppressor genes (1, 46, 52, 53). With the exception of studies on the QR1 gene (50, 54), little is known regarding the mechanism by which v-Src down-regulates gene transcription. In this study, we used the ssecks gene as a prototype in order to explore the mechanisms involved in v-Src-mediated gene repression. We localize an essential VSRE to the E- and GC-boxes sites in the SSeCKS {alpha} proximal promoter. However, although these same elements (and similar relative spacing) are found in the SSeCKS beta proximal promoter, neither the proximal nor 5-kb beta promoter exhibited VSR activity. This could be due to the following: (i) the VSRE being >5 kb upstream in beta promoter, (ii) the inactivity of the beta promoter VSRE in transient expression assays, or (iii) a coordinated control of the beta promoter by v-Src through VSRE found in the {alpha} promoter 68 kb upstream. The possibility that the transient expression assay fails to fully reflect VSR activity is borne out by our finding that even the v-Src-mediated down-regulation of the exogenous {alpha} promoter in the luciferase assays is roughly 2-fold lower than that of the endogenous {alpha} transcript levels (compare Fig. 1D to Fig. 4A). The notion that the upstream {alpha} promoter controls the VSR of the beta promoter is supported by our findings that the beta proximal promoter had similar levels of associated Ac-H3 and Ac-H4 in NIH3T3 and v-Src/3T3 cells, yet TSA treatment derepressed both {alpha} and beta transcript levels in v-Src/3T3 cells. Thus, the ability of TSA to increase Ac-H3/Ac-H4 binding to the {alpha} proximal promoter strengthens our notion that v-Src affects chromatin structure at the upstream {alpha} promoter only.


Figure 10
View larger version (31K):
[in this window]
[in a new window]

  		FIGURE 10.
HDAC1 combines with Sp1 and Sp3 to repress SSeCKS {alpha} proximal promoter activity in both NIH3T3 and v-Src/3T3 cells. NIH3T3 cells (A) or v-Src/3T3 cells (B) were transfected with the SSeCKS {alpha} proximal promoter-luciferase construct alone or in combination with varying amounts of expression plasmid encoding HDAC1, with or without Sp1 and/or Sp3 expression plasmids as indicated. Cell lysates were prepared after 48 h and assayed for luciferase activity. The luciferase activity values for each assay are expressed as a percentage of those from the reporter alone, arbitrarily set at 100%. Each bar is the mean of three replicates ± S.D. Data shown are representative of three independent experiments.

 
Transcriptional repression can be mediated either by recruiting repressors or by releasing activators from the gene promoter. Our observation that increased Sp1/Sp3 binding to the GC-box in the {alpha} promoter in v-Src/3T3 cells, coupled with the fact that Sp1 can behave as either activator or repressor depending on the gene promoter context, strongly suggests that the repression of {alpha}-SSeCKS by v-Src is mediated by either the recruitment of repressors to the promoter or the post-translational modification of Sp1/Sp3. Indeed, our data indicate that Sp1 plays dual roles as follows: activating basal promoter activity in both NIH3T3 and v-Src/3T3 cells and repressing promoter activity in v-Src/3T3 cells. In addition, our overexpression experiments suggest that the increased Sp3 binding to the GC-box in {alpha} proximal promoter in v-Src/3T3 cells can antagonize Sp1-mediated transactivation in v-Src/3T3 cells. Indeed, Sp3 seems able to convert an apparent Sp1-containing complex into a repressor of the {alpha} proximal promoter in v-Src/3T3 but not in NIH3T3 cells.

Several studies have demonstrated post-translational modifications to Sp1/Sp3 as well as the induction of several Sp1/Sp3 partners in cancer cells. For example, the activation of the ERK MAPK by growth factors is known to induce Sp1 phosphorylation, correlating with increased DNA binding activity (55, 56). Other modifications such as acetylation occur in cancer cells, (reviewed in Ref. 57), although no specific modifications induced by Src have been described to date. Interestingly, Kuo et al. (58) showed that v-Src induced higher Sp1 binding activity to a GC-rich box in the proximal promoter of MMP-2 via an ERK-dependent pathway, but this correlated with induced MMP-2 expression.

In addition to the GC-box, our mutation assays indicate that the E-box is also crucial for VSR. Although USF1 was first identified as a transcriptional activator for the adenovirus late promoter (59), recent studies demonstrate that USF1 is also involved in transcriptional repression of certain genes (60–62). Moreover, Ge et al. (63) showed physical interaction between USF1 and Sp1, and at low Sp1 concentrations, Sp1 and USF1 could cooperatively transactivate the deoxycytidine kinase promoter, whereas at higher levels of Sp1, USF1 helps form a repressor complex. Therefore, although v-Src does not alter USF1 binding to the E-box on the {alpha} promoter, it is likely that the enhanced Sp1 binding to the adjacent GC-box in v-Src/3T3 cells converts the USF1 from a transcriptional activator to a repressor. Interestingly, we find that the USF1-Sp1-Sp3 complex seems more stable in v-Src/3T3 than in NIH3T3 cells because an oligonucleotide missing the proximal {alpha} promoter E-box (–87 to –47) can pull down USF1 in a GC-box-dependent manner (i.e. requiring Sp1/Sp3 binding) in v-Src/3T3 cells only (Fig. 9D). Taken together, the cooperat