Phosphatidylinositol 3-Kinase/Protein Kinase C{zeta}-Induced Phosphorylation of Sp1 and p107 Repressor Release Have a Critical Role in Histone Deacetylase Inhibitor-Mediated Depression of Transcription of the Luteinizing Hormone Receptor Gene

We have demonstrated that silencing of luteinizing hormone receptor(LHR) gene transcription is mediated via a proximal Sp1 siteat its promoter. Trichostatin A (TSA) induced histone acetylationand gene activation in JAR cells that prevailed in the absenceof changes in Sp1/Sp3 expression, their binding activity, disassociationof the histone deacetylase/mSin3A complex from the Sp1 site,or demethylation of the promoter. This indicated a differentmechanism involved in TSA-induced derepression. The presentstudies have revealed that phosphatidylinositol 3-kinase/proteinkinase C ${zeta}$ (PI3K/PKC ${zeta}$ )-mediated Sp1 phosphorylation accounts forSp1 site-dependent LHR gene activation. TSA caused marked phosphorylationof Sp1 at serine 641 in JAR and MCF-7 cells. Blockade of PI3Kor PKC ${zeta}$ activity by specific inhibitors, kinase-deficient mutants,or small interfering RNA abolished the effect of TSA on theLHR gene and Sp1 phosphorylation. PKC ${zeta}$ was shown to associatewith Sp1, and this association was enhanced by TSA. Sp1 phosphorylationat serine 641 was required for the release of the pRb homologuep107 from the LHR gene promoter, while p107 acted as a repressorof the LHR gene. Inhibition of PKC ${zeta}$ activity blocked the dissociationof p107 from the LHR gene promoter and markedly reduced Sp1phosphorylation and transcription. These results have demonstratedthat phosphorylation of Sp1 by PI3K/PKC ${zeta}$ is critical for TSA-activatedLHR gene expression. These studies have revealed a novel mechanismof TSA action through derecruitment of a repressor from theLHR gene promoter in a PI3K/PKC ${zeta}$ -induced Sp1 phosphorylation-dependentmanner.

INTRODUCTION

Top

Introduction

The luteinizing hormone receptor (LHR) is a G-protein-coupledreceptor that is essential for sexual development and reproductivefunction in mammals. The LHR gene is expressed primarily ingonads but is also present in several nongonadal tissues andbreast and placenta cancer tissues and cells (21).

Characterization of the mechanism(s) involved in the regulationof LHR gene transcription has revealed that its basal promoteractivity is governed by two activating Sp1/Sp3 binding domainsand an inhibitory direct-repeat motif recognized by the nuclearorphan receptors EAR2 and EAR3/COUP-TFI (47-49, 56-58). We havefurther demonstrated that the status of histone modificationsand DNA methylation in the LHR gene promoter region operatecoordinately to elicit marked silencing or reactivation of thisgene's expression in human choriocarcinoma JAR cells and breasttumor MCF-7 cells (59, 60). Maximal derepression of the LHRgene in these cells is achieved upon hyperacetylation of histonesH3 and H4 within its promoter and complete demethylation ofthe CpG island encompassing the promoter (60). These findingshave illustrated the critical participation of an epigeneticprogram in the control of LHR gene transcriptional activity.The proximal Sp1 site of the LHR gene promoter was shown tobe an essential mediator of trichostatin A (TSA)-induced LHRgene activation. This site was also identified as an anchorto recruit the histone deacetylase (HDAC)/mSin3A complex tothe LHR gene promoter, leading to promoter-localized chromatincondensation through histone hypoacetylation (59). A TSA challenge,however, did not evoke a change in the patterns of DNA bindingat this site, the levels of the Sp1 and Sp3 proteins, or releaseof the HDAC/mSin3A complex from the LHR gene promoter in JARcells.

Sp1/Sp3 binding site-dependent gene activation via HDAC inhibitionhas also been observed for other genes in various cell linestreated with a variety of HDAC inhibitors (8, 20, 51, 55). Despitethe growing body of evidence in this regard, its mechanism isstill far from clear. It has also not yet been defined whetherSp1 and Sp3, which are both ubiquitously expressed and bindto GC-rich sequences with similar activities, contribute equallyto HDAC inhibitor-activated gene expression. The lack of changein Sp1/Sp3 binding properties and their protein expression statusin most of the cases studied imply that another mechanism(s)may be involved. Posttranslational modifications of Sp1 andSp3 are known to have a significant impact on Sp1 site-regulatedtarget gene expression in many physiological settings (5, 14,40). These include target genes involved in cell growth, apoptosis,angiogenesis, and tumorigenesis. In particular, phosphorylationand dephosphorylation of Sp1 in response to diverse exogenousstimuli and environmental cues can alter the transactivationactivity of Sp1 or its association with other transcriptionfactors or cofactors, resulting in a change in Sp1 site-controlledgene expression (4, 6, 15, 44). The present studies have demonstratedthat the participation of a signal transduction mechanism isessential for derepression of the LHR gene transcription inducedby the HDAC inhibitor TSA. Our findings have revealed a novelmechanism of TSA action through derecruitment of a repressorfrom the LHR gene promoter in a phosphatidylinositol 3-kinase/proteinkinase C ${zeta}$ (PI3K/PKC ${zeta}$ )-induced Sp1 phosphorylation-dependent manner.

MATERIALS AND METHODS

Top

Materials and Methods

Reagents, expression vectors, and antibodies. TSA, PD98059, H89, Gö6983, Gö6976, Rottlerin, a myristoylatedPKC ${zeta}$ pseudosubstrate peptide inhibitor (PS-PKC ${zeta}$ ), wortmannin,and LY294002 were obtained from Calbiochem (San Diego, CA).The human LHR promoter/luciferase reporter gene construct usedhas been described in our previous studies (57). The PKC ${zeta}$ kinase-inactivemutant construct was generously provided by Alex Toker (HarvardMedical School, Boston, MA). The PI3K p110 ${alpha}$ dominant negative(DN) mutant was a gift from Bai Lu (National Institute of ChildHealth and Human Development, NIH). The Flag-Sp1 expressionvector was kindly provided by Adrian Black (Roswell Park CancerInstitute, Buffalo, NY). The mutant Flag-Sp1 plasmid, in whichserine 641 was mutated to alanine, was constructed on the basisof wild-type (WT) Flag-Sp1 with the QuikChange II site-directedmutagenesis kit (Stratagene, La Jolla, CA) and verified by sequencinganalyses. The pCMV-Sp1, pCMV-pRb, and pCMV-p107 expression plasmidswere obtained from OriGene (Rockville, MD). The pcDNA1.1-Sp3expression vector was described previously (29).

The antibodies for various PKC isoforms, Sp3, actin, tubulin,HDAC1, HDAC2, mSin3A, pRb, and p107 were purchased from SantaCruz Biotechnology (Santa Cruz, CA). Whole-cell lysates of HeLa,3611-RF, Jurkat, and A431 cells obtained from Santa Cruz wereused as positive controls to verify the efficiency of the antibodiesfor individual PKC isoforms (HeLa cell lysates for PKC ${alpha}$ , ßI,ßII, and ${varepsilon}$ ; 3611-RF cell lysates for PKC ${gamma}$ and -µ;Jurkat cell lysates for PKC ${delta}$ , - ${zeta}$ , and - ${theta}$ ; A431 cell lysates forPKC ${iota}$ ). Lung tissue lysates (rat) obtained from Abcam (Cambridge,MA) were utilized as a positive control for PKC ${eta}$ antibody. Antibodiesfor phospho-PKC ${zeta}$ (Thr 410) and phospho-PKC ${delta}$ (Thr 505) were providedby Cell Signaling (Danvers, MA). Phosphoserine (p-Ser), phosphothreonine(p-Thr), or phosphotyrosine (p-Tyr) antibodies and Flag antibodywere purchased from Sigma (St. Louis, MO). The Sp1 antibodyfor immunoprecipitation (IP) was also from Sigma. Antibodiesfor the total and acetylated histones H3 and H4 were obtainedfrom Upstate (Charlottesville, VA).

Cell culture and transfection. Human choriocarcinoma JAR cells and MCF-7 cells (human mammarygland carcinoma cells) from the American Type Culture Collection(Manassas, VA) were maintained in RPMI 1640 medium supplementedwith 10% fetal bovine serum and 1% antibiotic-antimycotic solution(Invitrogen, Carlsbad, CA).

Transfections were carried out with Lipofectamine and Plus reagents(Invitrogen) according to the procedure recommended by the manufacture.For reporter gene analyses of cells transfected with expressionplasmids, the DNA amount used in each well was adjusted withempty vector plasmid so that every well contained the same amountof DNA. At 24 h posttransfection, cells were treated with theindicated inhibitors for 1 h, followed by TSA treatment foran additional 24 h. Luciferase activities were normalized tolight units per microgram of protein and expressed as means± standard errors. All experiments were performed atleast three times in triplicate wells.

RNA isolation, reverse transcription (RT)-PCR, and real-time PCR. Total RNA from treated or untreated cells was extracted withan RNeasy Kit (QIAGEN, Valencia, CA). RNA was pretreated withamplification grade DNase I (Invitrogen) at room temperaturefor 15 min before RT-PCR with a QIAGEN One Step RT-PCR kit.The primers for amplification of a 475-bp fragment encodinga C-terminal region of the human LHR gene (nucleotides 1557to 2031) were 5' GGAAACCACTCTCTCACAAGT 3' (forward) and 5' GGTGGATTGAGAAGGCTTATTTG3' (reverse). The primers for amplification of a 620-bp humanß-actin gene fragment were 5' CCTCGCCTTTGCCGATCC 3'(forward) and 5' GGATCTTCATGAGGTAGTCAGTC 3' (reverse).

For quantitative analyses of LHR gene mRNA levels, total RNAwas reverse transcribed with a SuperScript III kit (Invitrogen)and subjected to PCR analyses with SYBR green Master Mix andan ABI 7500 sequence detection system (Applied Biosystems, FosterCity, CA). The PCR conditions consist of a 10-min hot startat 95°C, followed by 40 cycles of 30 s at 95°C and 1min at 60°C. The primers used were 5' ATGGCAATCCTCATCTTCACCG3' (forward) and 5' GTGAATATTGCATACAGAAATG 3' (reverse). Eachsample was run in triplicate, and the results were normalizedto the level of ß-actin mRNA as described previously(60).

Whole-cell lysate and nuclear protein preparation. JAR whole-cell lysates were prepared with the M-PER mammalianprotein extraction reagent of Pierce (Rockford, IL). The cytosolicand nuclear fractions from treated and untreated cells wereisolated with the NucBuster protein extraction kit of Novagen(San Diego, CA) by following the protocol included. All of thepreparations were performed in the presence of 1x Complete proteaseinhibitor cocktail (Roche, Mannheim, Germany) and 1x Halt phosphataseinhibitor cocktail (Pierce).

Small interfering RNA (siRNA) analyses. siRNAs designed to knock down the endogenous expression of Sp1,Sp3, PKC ${zeta}$ , or PKC ${delta}$ and siRNA for the negative control (NTC) wereobtained from Ambion (Austin, TX). Transfections of siRNAs wereperformed by use of siPORTNeoFX reagent according to the proceduresrecommended by the manufacture (Ambion). Briefly, the siPORTNeoFXreagent was diluted in Opti-MEM I reduced-serum medium (Invitrogen)and incubated at room temperature for 10 min. This was thenmixed with the indicated siRNA (30 nM) diluted in Opti-MEM Ireduced-serum medium, followed by incubation at room temperaturefor 10 min. The complexes formed were dispensed into 24-wellor 6-well cell culture plates and overlaid with cells at a densityof 1 x 10⁵ cells/ml. At 24 h posttransfection, cells were replacedwith fresh medium and grown for an additional 48 h before harvest.For reporter gene analyses of LHR gene promoter activity inthe presence of siRNAs of Sp1, Sp3, PKC ${zeta}$ , PKC ${delta}$ , or NTC, the promoter/luciferaseconstruct was introduced by Lipofectamine and Plus reagentsat 24 h posttransfection with siRNA, and cells were terminated48 h later. TSA was added to the cells 24 h prior to termination.

IP analyses. For each IP assay, 250 µg nuclear proteins or 1 mg whole-celllysates in 1 ml IP buffer (20 mM Tris HCl [pH 8.0], 100 mM NaCl,1 mM EDTA, 0.5% NP-40, 1x protease inhibitor cocktail, 1x phosphataseinhibitor cocktail) was used. The proteins were initially preclearedwith 0.5 µg normal rabbit or mouse immunoglobulin G (IgG)and 40 µl protein A/G PLUS agarose beads (Santa Cruz)at 4°C for 20 min. The precleared supernatant was incubatedwith 1 µg of the specific antibody of interest at 4°Covernight with gentle shaking. This was followed by additionof 36 µl protein A/G agarose beads and incubation at 4°Cfor another 1 h. The protein-antibody-agarose complex was recoveredby brief centrifugation and washed three times with 1 ml IPbuffer for 5 min each time at 4°C. The complex was resuspendedin 30 µl 2x sodium dodecyl sulfate protein sample buffercontaining 2.5% ß-mercaptoethanol and boiled for 3min before being subjected to Western blot analyses.

DNA affinity precipitation assays (DAPA). DAPA were performed as described previously (59). Briefly, 50µg nuclear extracts was incubated with 0.4 µg biotin-labeledprobe harboring the sequences of the proximal Sp1 binding siteof the LHR gene promoter. The reaction was carried out on icefor 45 min, which was followed by addition of 30 µl oftetrameric avidin resin (Promega, Madison, WI) with anotherincubation at 4°C for 1 h. The DNA-protein complex was harvestedby brief centrifugation and analyzed by immunoblotting analyseswith the antibodies indicated.

Chromatin IP (ChIP). ChIP experiments were performed with the ChIP assay kit fromUpstate Biotechnology (Lake Placid, NY). Briefly, 2 x 10⁷ treatedor untreated cells were fixed with 1% formaldehyde at 37°Cfor 10 min and lysed and sheared by sonication. The solublechromatin fraction was subjected to IP analyses with the antibodiesof interest as indicated. The precipitated complexes were washedsequentially with low-salt, high-salt, LiCl, and Tris-EDTA buffersand extracted twice with freshly prepared 1% sodium dodecylsulfate-0.1 M NaHCO₃. The cross-linking between DNA and proteinswas reversed, and DNA was purified by ethanol precipitation.The ChIP-precipitated DNA and input DNA were analyzed by real-timePCR with SYBR green Master Mix in an ABI 7500 sequence detectionsystem. The primers for the LHR gene promoter region were 5'ACTGGGCACTGTCGCAGGTC 3' (forward) and 5' CATGGCCGGCGAACTGGGCT3' (reverse), as described previously (60). The primers forthe human ß-actin promoter were 5' GGCCAACGCCAA AACTCTCCCTC3' (forward) and 5' GTCTCGGCGGTGGT GGCGCGT 3'. The amplified236-bp DNA fragment was cloned into the pCRZeroBlunt vector(Invitrogen), and the standard curve was created by a 10-folddilution of the cloned plasmid DNA for quantitative analysesof the binding of the indicated transcription factors to theß-actin promoter.

RESULTS

Top

Results

PKC ${zeta}$ has a critical role during TSA-induced LHR gene activation. Recent studies have demonstrated that treatment with an HDACinhibitor (e.g., sodium butyrate [NaB], trapoxin, or apicidin)caused increased activities of mitogen-activated protein kinases(MAPK) and protein kinase C (PKC) in several of the cell linesstudied (28, 39, 52). These findings indicated that these signaltransduction pathways might contribute to the stimulatory changesinduced by the HDAC inhibitors in the transcription of severalgenes. To investigate the putative participation of a proteinkinase(s) in TSA-activated LHR gene expression, reporter geneanalyses of LHR gene promoter activity were carried out initiallywith JAR cells that were pretreated with inhibitors of MAPK,PKA, or PKC prior to incubation with TSA (Fig. 1A). The PKCinhibition by Gö6983 (a broad-spectrum PKC inhibitor) diminishedTSA-induced activation of LHR gene promoter activity. In contrast,neither inhibition of MAPK by PD98059 nor that of PKA by H89had an impact on LHR gene promoter activity. This indicatedthe participation of a PKC family member(s) in the TSA-mediatedactivation of LHR gene expression.

View larger version (28K):
[in this window]
[in a new window]
FIG. 1. The PKC pathway is involved in TSA-induced LHR gene activation. (A) Reporter gene analyses of LHR gene promoter activity in JAR cells. Cells were pretreated with PD98059 (10 µM), H-89 (5 µM), or Gö6983 (2 µM) for 1 h prior to incubation with or without TSA (100 ng/ml) for 24 h. Relative promoter activities are indicated as n-fold induction over the activity in the absence of treatment (onefold). (B) Western blot analyses of the endogenous expression of PKC isoforms in JAR cells. (C and D) RT-PCR (C) and real-time PCR (D) analyses of human LHR (hLHR) gene expression in JAR cells that were treated with Rottlerin (5 µM) or PS-PKC ${zeta}$ (50 µM) for 1 h, followed by incubation with or without TSA (100 ng/ml) for 24 h. The relative mRNA levels from the quantitative analyses are indicated as n-fold induction over the level in the absence of treatment (onefold).

PKC is a family of serine/threonine kinases that includes atleast 11 isozymes (19, 46). These PKC isoforms are subgroupedinto three categories based on their regulatory properties,namely, conventional PKCs [ ${alpha}$ , ß(I), ß(II),and ${gamma}$ ], novel PKCs ( ${delta}$ , ${varepsilon}$ , ${eta}$ , ${theta}$ , and µ), and atypical PKCs ( ${zeta}$ , ${iota}$ / ${lambda}$ [human/mouse]). Western blot analyses of endogenous PKC isoenzymesin JAR cells revealed predominant expression of PKC ${delta}$ and - ${zeta}$ , whereasthe levels of other forms of PKC were undetectable, except forPKCß(II), which displayed a weak signal (Fig. 1B).The efficiency of the PKC antibodies used was confirmed by Westernblot analyses of lysates from cell lines expressing known PKCsubtypes (see Materials and Methods; data not shown). RT-PCRanalyses revealed that inhibition of PKC ${delta}$ activity by Rottlerindid not affect TSA-mediated induction at the LHR gene mRNA level(Fig. 1C and D). In addition, Gö6976, an inhibitor of conventionalPKC, did not have any effect on LHR gene expression endogenouslyor in reporter gene analyses (data not shown). These resultswere consistent with nonexpression or minimal expression ofthe conventional PKC isotypes in these cells (Fig. 1B). However,this induction was largely abolished upon blockade of the activityof the atypical PKC, PKC ${zeta}$ , by PS-PKC ${zeta}$ (Fig. 1C and D). Furthermore,TSA dose dependently induced the activation of PKC ${zeta}$ , as determinedby the phosphorylation of this enzyme at threonine 410 (Thr410) (Fig. 2A), which is recognized to be critical for its activity(2, 42). In contrast, phosphorylation of PKC ${delta}$ at Thr 505 (control)was not observed (Fig. 2B). Moreover, the critical participationof PKC ${zeta}$ in the up-regulation of LHR gene expression by TSA wasconfirmed by reporter gene analyses in which both the PKC ${zeta}$ inhibitorand a kinase-inactive mutant form of PKC ${zeta}$ (the DN mutant PKC ${zeta}$ DN) abrogated TSA-activated LHR gene promoter activity in adose-dependent manner (Fig. 2C and D). The specific role ofPKC ${zeta}$ was further demonstrated by siRNA analyses in which TSA-inducedLHR gene promoter activity was significantly abolished uponsiRNA suppression of PKC ${zeta}$ expression. In contrast, no such effectwas observed after knockdown of the expression of PKC ${delta}$ (Fig.2E). Collectively, it has been demonstrated that PKC ${zeta}$ is criticalfor TSA-induced LHR gene activation, since inhibition of itsactivity markedly diminished the inducible effect of TSA onLHR gene expression.

View larger version (27K):
[in this window]
[in a new window]
FIG. 2. PKC ${zeta}$ is critical for the LHR gene activation induced by TSA. (A) Levels of phospho-PKC ${zeta}$ (Thr 410) in JAR cells treated with increasing doses of TSA. Whole-cell lysates were immunoprecipitated with PKC ${zeta}$ antibody, followed by Western blotting with phospho-PKC ${zeta}$ (Thr 410) antibody. (B) Western blot analyses of phospho-PKC ${delta}$ (Thr 505) in JAR cells treated with the indicated doses of TSA. The total PKC ${zeta}$ and PKC ${delta}$ levels in these cells are also shown. (C and D) Reporter gene analyses of LHR gene promoter activity in JAR cells in the presence of increasing doses of a PKC ${zeta}$ inhibitor (C) or of a PKC ${zeta}$ DN mutant construct (D). (E, top) JAR cells were transfected with PKC ${zeta}$ or PKC ${delta}$ siRNA or a negative control siRNA (NTC), followed by Western blot analyses of the expression of PKC ${zeta}$ or PKC ${delta}$ protein. The expression of actin in these cells is also shown as a control. (E, bottom) Reporter gene analyses of LHR gene promoter activity in JAR cells cotransfected with PKC ${zeta}$ , PKC ${delta}$ siRNA, or NTC. Cells were treated with or without TSA for 24 h. Relative luciferase activities are indicated as n-fold induction of enzyme activity in the presence of TSA over enzyme activity in the absence of TSA.

TSA treatment elicited marked Sp1 phosphorylation at serine 641 dependent on the activity of PKC ${zeta}$ . The dependence on the proximal Sp1/Sp3 binding site in the processof TSA-activated LHR gene expression led to studies to determinewhether PKC ${zeta}$ exerts its impact through Sp1, Sp3, or both. IPof nuclear extracts from TSA-treated or untreated JAR cellsby an Sp1 or Sp3 antibody was performed, followed by Westernblot analyses with p-Ser, p-Thr, or p-Tyr antibodies. TSA treatmentcaused marked Sp1 phosphorylation at serine residues, whereasno phosphorylation of Sp1 at threonine or tyrosine residueswas detected (Fig. 3A). In contrast, no phosphorylation of Sp3was observed in the absence or presence of TSA. Furthermore,the TSA-induced serine phosphorylation of Sp1 was PKC ${zeta}$ dependentsince inhibition of the activity of this kinase by its inhibitor,its DN mutant form, and its siRNA significantly reduced Sp1phosphorylation (Fig. 3B and C). The TSA-induced Sp1 phosphorylationat a serine residue(s) was also in agreement with a previousreport showing that PKC ${zeta}$ -elicited Sp1 phosphorylation occurredsolely at a serine residue(s) (37).

View larger version (57K):
[in this window]
[in a new window]
FIG. 3. TSA causes marked Sp1 phosphorylation dependent on the activity of PKC ${zeta}$ . (A) Nuclear extracts isolated from TSA-treated (100 ng/ml, 24 h) or untreated JAR cells was immunoprecipitated (IP) with Sp1 or Sp3 antibody or normal rabbit IgG, followed by immunoblotting (IB) with p-Ser, p-Thr, or p-Tyr antibodies. Total Sp1 and Sp3 precipitates were also assayed with Sp1 or Sp3 antibody. (B) JAR cells pretreated with the PKC ${zeta}$ inhibitor PS-PKC ${zeta}$ or overexpressed of the PKC ${zeta}$ DN mutant form were incubated with TSA for 24 h. Nuclear extracts were immunoprecipitated with Sp1 antibody or normal IgG and then immunoblotted with p-Ser antibody. Immunoblotting with Sp1 antibody is also shown. (C) JAR cells transfected with PKC ${zeta}$ or NTC siRNA were treated with or without TSA for 24 h. Whole-cell lysates were immunoprecipitated with Sp1 antibody and then immunoblotted with p-Ser antibody. The total amount of Sp1 immunoprecipitated is also shown. (D) Cytosolic (Cyto) and nuclear (Nuc) fractions prepared from TSA-treated or untreated JAR cells were subject to immunoblotting analyses with antibodies to Sp1, PKC ${zeta}$ , phospho-PKC ${zeta}$ (Thr 410), and actin. HDAC1 and tubulin, used as nuclear and cytosolic markers, were also assessed by immunoblotting. (E) Nuclear extracts isolated from TSA-treated JAR cells were immunoprecipitated with Sp1 antibody and then immunoblotted with PKC ${zeta}$ antibody and vice versa.

PKC ${zeta}$ is distributed evenly in both the nuclear and cytosolicfractions of cells in the presence or absence of TSA (Fig. 3D).The activated form of PKC ${zeta}$ (phosphorylated at Thr 410; Fig. 2A)not only existed in the cytosol but was also present in thenucleus, strengthening the concept that Sp1 could be a nucleartarget of this kinase. Coimmunoprecipitation studies demonstratedthat endogenous PKC ${zeta}$ was in the immunocomplex precipitated byan Sp1 antibody and, similarly, endogenous Sp1 in the PKC ${zeta}$ immunocomplex(Fig. 3E). These observations have confirmed that Sp1and PKC ${zeta}$ coexisted in the same protein complex. Moreover, TSA treatmentcaused a significant increase in the association between Sp1and PKC ${zeta}$ in a TSA dose-dependent manner.

The differential effects of TSA on the phosphorylation of Sp1and Sp3 raised the possibility that Sp1 and Sp3 act differentlyduring TSA-activated LHR gene expression. In agreement withour previous findings (25), overexpression of Sp1 or Sp3 inJAR cells increased basal LHR gene promoter activity (Fig. 4A).On the other hand, Sp1 augmented the effect of TSA on LHR transcriptionalactivation, whereas no enhancement was observed with Sp3 (Fig.4B). Furthermore, transfection of cells with Sp1 siRNA, whichcaused complete knockdown of Sp1 protein expression, markedlyprevented TSA-induced LHR gene promoter activity (Fig. 4C and D).In contrast, suppression of Sp3 protein expression by its siRNAhad no effect on the TSA response of the LHR gene. Taken together,these results have illustrated that TSA treatment elicited markedSp1 but not Sp3 phosphorylation and that Sp1, but not Sp3, hasa significant role in TSA-induced LHR gene activation. Coincidentwith the requirement of PKC ${zeta}$ activation for the effect of TSAon LHR gene expression, TSA-induced Sp1 phosphorylation wasdependent on the activity of PKC ${zeta}$ .

View larger version (28K):
[in this window]
[in a new window]
FIG. 4. Sp1, but not Sp3, is required for TSA-activated LHR gene expression. (A and B) Reporter gene analyses of LHR gene promoter activity in JAR cells cotransfected with increasing doses of Sp1or Sp3 expression plasmid. Cells were treated with (B) or without (A) TSA for 24 h. (C) JAR cells were transfected with Sp1 or Sp3 siRNA or negative control siRNA (NTC), followed by immunoblotting analyses of the Sp1 or Sp3 protein level. The level of actin in these cells is shown as a control. (D) Reporter gene analyses of LHR gene promoter activity in JAR cells cotransfected with Sp1, Sp3 siRNA, or NTC. Cells were treated with or without TSA for 24 h. Relative luciferase activities are indicated as n-fold induction of enzyme activity in the presence of TSA over enzyme activity in the absence of TSA.

Computational analyses of the putative phosphorylation site(s)of Sp1 by PKC ${zeta}$ with the Scansite program (http://scansite.mit.edu/)at low stringency revealed that serine 641 and threonine 681in the Sp1 DNA binding domain were probably the candidate sites.This prediction is consistent with in vitro phosphorylationstudies of a series of truncated forms of the Sp1 protein, whichdemonstrated that PKC ${zeta}$ interacted with and phosphorylated theDNA binding domain of Sp1, whereas no interaction or phosphorylationwas detected for the rest of the Sp1 molecule (34). To delineatethe putative phosphorylation site(s) of Sp1 targeted by PKC ${zeta}$ during TSA-activated LHR gene expression, serine 641 of WT Flag-Sp1was mutated to alanine (S641A mutant) and expressed in JAR cells(Fig. 5A). Consistent with our previous results, TSA treatmentcaused significant phosphorylation of Sp1 precipitated witha Flag antibody (Fig. 5B). However, no phosphorylation of theS641A mutant Sp1 protein was detected. These results indicatedthat Ser 641 is the target site associated with the TSA-inducedresponses. In addition, overexpression of the Sp1 S641A mutantprotein dose dependently reduced the phosphorylation level ofSp1 in TSA-treated cells (Fig. 5C). Functionally, the Sp1 S641mutant antagonized TSA-activated LHR gene promoter activity,in contrast to the augmentation effect caused by WT Flag-Sp1(Fig. 5D). Taken together, these results have demonstrated thatphosphorylation of Sp1 at serine 641 has a critical role inTSA-induced LHR gene activation.

View larger version (38K):
[in this window]
[in a new window]
FIG. 5. Ser 641 of Sp1 is the target phosphorylation site during TSA-induced LHR gene activation. (A) WT or serine 641-mutated (to alanine, S641A) Flag-Sp1 plasmid was transfected into JAR cells. The expression of WT and mutant Sp1 proteins was assayed by immunoblotting (IB) analyses with Flag antibody. (B) JAR cells transfected with WT or S641A mutant Flag-Sp1 were treated with or without TSA for 24 h. Whole-cell lysates were immunoprecipitated with Flag antibody, followed by immunoblotting with p-Ser antibody. Total immunoprecipitated Flag-Sp1 proteins are also shown. (C) JAR cells were transfected with increasing doses of Flag-Sp1 S641A mutant construct, followed by treatment with TSA for 24 h. Whole-cell lysates were immunoprecipitated with Sp1 antibody, followed by immunoblotting with p-Ser, Flag, or Sp1 antibody. (D) Reporter gene analyses of LHR gene promoter activity in JAR cells transfected with a WT or S641A mutant Sp1 construct. Relative luciferase activity is shown as n-fold induction of enzyme activity in the presence of TSA over that in the absence of TSA.

Participation of PI3K during TSA-induced LHR gene activation. The activation of PKC ${zeta}$ by TSA indicated the participation ofanother kinase(s) as an upstream effector of PKC ${zeta}$ during TSA-inducedLHR gene activation. Recent studies have implied that PKC ${zeta}$ isa downstream target for PI3K in diverse physiological settings(13, 43, 45). Also, PI3K has been shown to be involved in theregulation of the gene for p21 by apicidin, an HDAC inhibitor(32). To address whether PI3K contributes to TSA-activated LHRgene expression, blockade of PI3K activity by the inhibitorwortmannin or LY294002 was tested in reporter gene analysesof LHR gene promoter activity. Both wortmannin and LY294002caused a dose-dependent reduction of promoter activity in thepresence of TSA (Fig. 6A and B). This reduction was specificto the TSA-mediated effect since no change in LHR gene basalpromoter activity was observed. We further confirmed the effectof PI3K on LHR gene promoter activity with a kinase-inactivemutant form of the PI3K p110 ${alpha}$ subunit. Similarly, the DN mutantPI3K p110 ${alpha}$ construct resulted in a significant reduction in theLHR gene promoter activity induced by TSA (Fig. 6C). Moreover,the PKC ${zeta}$ phosphorylation induced by TSA was abolished upon treatmentwith wortmannin (0.5 µM) or LY294002 (50 µM) (Fig.7A). Coincident with our previous studies showing a lack ofeffect of TSA on the binding activity or level of the Sp1 protein(59), no changes in these parameters were observed in the presenceof wortmannin or LY294002 or the PS-PKC ${zeta}$ mutant (Fig. 7B). However,inhibition of PI3K markedly suppressed TSA-induced Sp1 phosphorylation(Fig. 7C), demonstrating an important role for PI3K in TSA-activatedLHR gene expression. These results, in combination with theobservation that no nuclear location of PI3K was detected inthe absence or presence of TSA (data not shown), have demonstratedthat PI3K acts upstream of PKC ${zeta}$ , which is present in both thenuclear and cytoplasmic compartments, and that the activityof this kinase is necessary for PKC ${zeta}$ -elicited phosphorylationof Sp1.

View larger version (16K):
[in this window]
[in a new window]
FIG. 6. Participation of PI3K in TSA-induced LHR gene activation. (A and B) Reporter gene analyses of LHR gene promoter activity in JAR cells that were pretreated with wortmannin or LY294002 for 1 h, followed by incubation with or without TSA (100 ng/ml) for 24 h. (C) Reporter gene analyses of LHR gene promoter activity in JAR cells in the presence of a DN PI3K p110 ${alpha}$ mutant construct. Cells were treated with or without TSA for 24 h prior to analyses. Relative luciferase activities are indicated as n-fold induction over enzyme activity in the absence of drug treatment or the PI3K p110 ${alpha}$ DN mutant form.

View larger version (40K):
[in this window]
[in a new window]
FIG. 7. PI3K activity is required for PKC ${zeta}$ activation and Sp1 phosphorylation. (A) JAR cells were pretreated with or without wortmannin (0.5 µM) or LY294002 (50 µM) for 1 h, followed by incubation with or without TSA (100 ng/ml) for 24 h. Whole-cell lysates were immunoprecipitated with PKC ${zeta}$ antibody, followed by immunoblotting (IB) with phospho-PKC ${zeta}$ antibody (Thr 410). The total amounts of PKC ${zeta}$ precipitated from these cells are also shown as a control. (B) DAPA of Sp1 binding activity in JAR cells. Nuclear extracts isolated from JAR cells treated with or without TSA in the presence of wortmannin, LY294002, or PS-PKC ${zeta}$ were incubated with the biotin-labeled proximal Sp1 site of the LHR gene promoter, followed by immunoblotting with Sp1 antibody. The level of Sp1 protein is also shown. (C) IP of the nuclear extracts from panel A by Sp1 antibody was followed by immunoblotting with an antibody recognizing p-Ser and with an Sp1 antibody, respectively.

Release of p107 from the TSA-activated LHR gene promoter in a PKC ${zeta}$ -dependent manner. Promoter-specific histone hyperacetylation resulting from HDACinhibition has been recognized as crucial for the activationof target genes. To further elucidate the mechanism underlyingthe relevance of PKC ${zeta}$ in TSA-induced LHR gene expression, theputative impact of this kinase on histone acetylation statusin the LHR gene promoter region was next investigated. Blockadeof PKC ${zeta}$ activity did not affect the global histone H3 and H4hyperacetylation in TSA-treated cells (Fig. 8A). Also, as demonstratedby quantitative ChIP analyses, the presence of acetylated H3or H4 at the LHR gene promoter which was evoked by TSA treatmentremained unchanged in the presence of the PKC ${zeta}$ inhibitor (Fig.8B, left). These results thus indicate a lack of effect of PKC ${zeta}$ on LHR gene promoter-associated histone acetylation status.

View larger version (29K):
[in this window]
[in a new window]
FIG. 8. TSA causes release of p107 from the LHR gene promoter independently of histone acetylation status. (A) Immunoblotting analyses of the levels of acetylated histones H3 (AceH3) and H4 (AceH4) in TSA-treated or untreated JAR cells in the presence of different doses of the PKC ${zeta}$ inhibitor PS-PKC ${zeta}$ . The total H4 protein level is also shown. (B) Quantitative ChIP analyses of the association of acetylated H3 or H4 with the LHR gene promoter (left) and the human ß-actin promoter (right) in JAR cells treated with or without TSA, PS-PKC ${zeta}$ , or both. (C) Quantitative ChIP analyses of the association of Sp1, HDAC1, HDAC2, mSin3A, pRb, and p107 with the LHR gene promoter (left) and the human ß-actin promoter (right) in TSA-treated or untreated JAR cells. Results are expressed as percentages of the total input DNA.

On the other hand, it was conceivable that PKC ${zeta}$ might exert itsaction through a change at the protein-protein interaction level.Thus, the effort was directed toward the identification of aprotein(s) whose associations with the LHR gene promoter werechanged by TSA treatment. Consistent with our previous findings(59), the binding of HDAC1/2 and the corepressor protein mSin3Awas unchanged in the absence or presence of TSA (Fig. 8C, left).This was also the case for histone acetyltransferase p300 andCBP corepressor proteins NcoR and SMRT (data not shown). Wefurther tested for retinoblastoma protein pRb and its closehomologue p107 since pRb is an HDAC-interacting protein andpRb and p107 have been shown to associate with the Sp1 bindingsite (9, 18, 36). TSA treatment profoundly decreased the associationof p107 with the LHR gene promoter (Fig. 8C, left). In contrast,the binding of pRb was not affected by TSA. The release of p107from the LHR gene promoter was found to be largely dependenton the activity of PKC ${zeta}$ , inhibition of which substantially preventedTSA-induced p107 dissociation (Fig. 9A). In addition, theseobserved changes were confirmed to be specifically associatedwith the LHR gene, whose promoter activity was markedly affectedby TSA, whereas no such effects were detected for the ß-actingene, whose expression was not influenced by TSA (Fig. 1, 8,and 9).

View larger version (32K):
[in this window]
[in a new window]
FIG. 9. Release of p107 from the LHR gene promoter in a PKC ${zeta}$ -dependent manner. (A) Quantitative ChIP analyses of p107 association with the LHR gene promoter and the human ß-actin promoter in JAR cells treated with or without TSA, PS-PKC ${zeta}$ , or both. (B) JAR cells were transfected with increasing doses of an S641A Flag-Sp1 mutant construct, followed by treatment with TSA for 24 h. Quantitative ChIP assays were carried out for analyses of p107 binding to the LHR gene promoter under these conditions. (C) DAPA of p107 association with the proximal Sp1 site of the LHR gene promoter. Nuclear extracts from cells treated as described for panel A were incubated with a biotin-labeled Sp1 probe, followed by immunoblotting analyses with a p107 antibody. The level of p107 protein is also shown (I.B.). (D) Reporter gene analyses of LHR gene promoter activity in JAR cells in the presence of a p107 or pRb expression plasmid. Relative luciferase activities are expressed as percentages of enzyme activity in the absence of the p107 or Rb construct (100%).

The dependency of p107 release on the phosphorylation statusof Sp1 was further analyzed by ChIP analyses of JAR cells, whereincreasing doses of the Flag-Sp1 S641A mutant construct wereexpressed (Fig. 9B). In agreement with the previous phosphorylationand functional analyses (Fig. 5), this phosphorylation-deficientSp1 mutant reduced TSA-elicited p107 release from the LHR genepromoter in a dose-dependent manner. Consistent with the resultsof ChIP assays, DAPA have shown that TSA caused significantdetachment of p107 from the proximal Sp1 site of the LHR genepromoter, whereas this effect was counteracted upon the blockadeof PKC ${zeta}$ activity (Fig. 9C). Furthermore, functional analyseshave demonstrated that p107 acted as a repressor of LHR geneexpression since coexpression of p107 but not pRb with the LHRgene promoter repressed promoter activity dose dependently (Fig.9D).

We had previously reported that TSA caused marked LHR gene derepressionin MCF-7 cells (60). The response elicited by TSA in this cellline was more prominent than in JAR cells (160-fold versus 40-fold),and in contrast to JAR cells, the LHR gene promoter is completelydemethylated in MCF-7 cells under basal conditions. This probablyfacilitated the dissociation of the HDAC/mSin3A complex fromthe LHR gene promoter observed in the presence of TSA, whilesuch a dissociation of the complex did not occurred in JAR cellstreated with TSA (59, 60). However, at that point we could notdetermine whether the properties observed, such as DNA methylationstatus and release of the HDAC/mSin3A inhibitory complex, inthese cell types would solely account for the differences observedin LHR gene promoter activities. Therefore, it was importantto learn whether the TSA-induced release of p107 from the LHRpromoter observed in JAR cells also occurred in MCF-7 cells.The activation of LHR gene expression by TSA in MCF-7 cellswas Sp1 site dependent, and the expression patterns of PKC isoformsin these cells were identical to those observed in JAR cells(data not shown). Reporter gene analyses revealed that the activationof LHR gene promoter activity by TSA was significantly repressedby inhibitors of PI3K (wortmannin, LY294002) and PKC ${zeta}$ (PS-PKC ${zeta}$ )and by PKC ${zeta}$ siRNA (Fig. 10A and B). Similar to JAR cells, TSAtreatment resulted in Sp1 serine phosphorylation in MCF-7 cellsand this phosphorylation was abolished in the presence of PS-PKC ${zeta}$ or wortmannin but not Gö6976 or Rottlerin (Fig. 10C). Furthermore,besides the dissociation of the HDAC/mSin3A complex, significantrelease of p107 from the LHR gene promoter was observed in TSA-treatedMCF-7 cells (Fig. 10D). In particular, release of p107 requiredthe activity of PKC ${zeta}$ because treatment of cells with PS-PKC ${zeta}$ incombination with TSA prevented p107 dissociation. In contrast,dissociation of the HDAC/mSin3A complex from the LHR gene promoterwas not affected by a blockade of PKC ${zeta}$ activity (Fig. 10D). Thiswas consistent with our previous finding demonstrating thatrecruitment or derecruitment of the HDAC/mSin3A complex is dependenton the methylation status of the promoter (60). Collectively,these results have demonstrated the participation of the PKC ${zeta}$ and PI3K signaling pathways in TSA-induced LHR gene activationin both JAR and MCF-7 cells. We have elucidated a novel mechanismby which PKC ${zeta}$ -dependent Sp1 phosphorylation and release of arepressor protein from the LHR gene promoter are critical forthe LHR gene activation elicited by TSA.

View larger version (27K):
[in this window]
[in a new window]
FIG. 10. Activation of LHR gene expression in MCF-7 cells by TSA involves PKC ${zeta}$ /PI3K, Sp1phosphorylation, and p107 release. (A) Reporter gene analyses of LHR gene promoter activity in MCF-7 cells which were treated with or without TSA (500 µg/ml, 24 h) in the presence or absence of PS-PKC ${zeta}$ , wortmannin, or LY294002. Relative luciferase activities are expressed as n-fold induction of enzyme activity in the presence of TSA over that observed in the absence of TSA. (B) Reporter gene analyses of LHR gene promoter activity in MCF-7 cells cotransfected with PKC ${zeta}$ siRNA or a negative control (NTC). Cells were treated with or without TSA for 24 h. The PKC ${zeta}$ and actin protein levels in these cells are also shown. (C) Sp1 phosphorylation status in MCF-7 cells. Nuclear extracts were isolated from untreated MCF-7 cells or cells treated with TSA in the absence or presence of Gö6976, Rottlerin, PS-PKC ${zeta}$ , or wortmannin. Nuclear extracts were immunoprecipitated with Sp1 antibody, followed by immunoblotting (IB) with p-Ser antibody. (D) Quantitative ChIP analyses of the association of indicated transcription factors to the LHR gene promoter in MCF-7 cells. Cells were treated with or without TSA in the presence of PS-PKC ${zeta}$ .

DISCUSSION

Top

Discussion

These studies have identified a mechanism in which Sp1 phosphorylationand release of the p107 repressor from the LHR gene promotermediate TSA-evoked LHR gene activation in JAR and MCF-7 cells.PKC ${zeta}$ is responsible for the marked Sp1 phosphorylation inducedby TSA, and blockade of its activity largely abrogates Sp1 phosphorylation,p107 dissociation, and the TSA-induced increase in LHR geneexpression. Serine 641 of the Sp1 DNA binding domain has beenidentified as the functional phosphorylation site, and mutationof this site significantly prevented Sp1 phosphorylation, p107release, and TSA-mediated LHR gene activation. Moreover, PI3Kwas found to be an upstream effector of PKC ${zeta}$ 's action. Inhibitionof PI3K activity abolishes TSA-induced PKC ${zeta}$ activation, Sp1 phosphorylation,and the TSA-induced increase in LHR gene expression. Collectively,these findings have demonstrated that Sp1 phosphorylation bya PI3K/PKC ${zeta}$ kinase pathway is a critical component in HDAC inhibitor-elicitedderepression of a target gene such as the LHR gene.

Our recent studies revealed that the proximal Sp1/Sp3 bindingsite is essential for activation of LHR gene expression by theHDAC inhibitor TSA (59). The present observations have demonstratedthat although both Sp1and Sp3 function as activators of LHRgene expression under basal conditions, they have differentroles in TSA-induced LHR gene activation. Sp1 is a criticalmediator of the effect of TSA on the LHR gene, whereas Sp3 isnoncontributory (Fig. 4). Consistent with the differences inthe functional roles of Sp1 and Sp3 in TSA-mediated LHR geneactivation, TSA induced marked Sp1 phosphorylation with no changein Sp3. In cells exposed to exogenous stimuli, Sp1 phosphorylationis recognized as an important factor in Sp1 site-regulated geneexpression, particularly for genes involved in viral infectionand cell growth (5, 14). Although Sp1 phosphorylation was observedin Hep3B and SL2 cells treated with NaB (an HDAC inhibitor)(12), the functional impact of the signaling pathways and mechanismsinvolved in the activation of target gene transcription by HDACinhibitors have not been determined. Our studies have demonstratedthe direct relevance of Sp1 phosphorylation in HDAC inhibitor-regulatedgene expression and have revealed that phosphorylation of Sp1was attributed to the activity of the PI3K/PKC ${zeta}$ pathway.

Increased MAPK or PKC activity has been detected in cells treatedwith HDAC inhibitors. Inhibition of the MEK-ERK activity inK562 leukemia cells was shown to abrogate the G_i2 gene activationinduced by NaB and Helminthosporium carbonium toxin, but notTSA (54). Also, blockade of the MEK activity in CHP126 neuroepitheliomacells reduced the activation of the choline acetyltransferasegene by NaB, TSA, or trapoxin (23). On the other hand, the apicidin-mediatedp21 gene activation in HeLa cells required an Sp1 binding siteand the activity of PKC ${varepsilon}$ , while the Sp1 phosphorylation statuswas not evaluated (32). However, since apicidin caused significanttranslocation of PKC ${varepsilon}$ to the plasma membrane, it is unlikelythat PKC ${varepsilon}$ would exert an action in the nucleus. Our studies haveshown that the atypical PKC isoform PKC ${zeta}$ is required for TSA-inducedLHR gene expression, whereas the participation of conventionalor novel PKCs, MAPK, and PKA was excluded. The lack of conventionalPKC, novel PKC, and MAPK action was further confirmed by theuse of their DN or constitutively active mutant forms (datanot shown). This evidence indicates that different kinase pathwaysare utilized and associated with the effects of various HDACinhibitors and that they are subject to the nature of the inhibitorsand/or the genes studied. This notion is also supported by geneexpression profiling analyses of T24 bladder and MDA breastcarcinoma cells treated with three HDAC inhibitors (26). Itwas found that the set of genes influenced by the structurallyrelated compounds suberoylanilide hydroxamic acid and TSA wassubstantially different from those regulated by MS-27-275, areagent that is chemically unrelated to suberoylanilide hydroxamicacid and TSA.

The array of protein kinases that has been found to mediateSp1 phosphorylation includes DNA-PK, PKA, PKC ${zeta}$ , casein kinaseII, ERK, cyclin-dependent kinase, and others whose identitiesare not yet known (1, 6, 7, 16, 27, 30, 33). However, PKC ${zeta}$ -inducedSp1 phosphorylation has been exclusively associated with theregulation of growth factor genes (35, 37, 38). The presentstudy has identified the gene for LHR as a novel target of PKC ${zeta}$ and has linked the PKC ${zeta}$ pathway and the upstream activator PI3Kto the changes elicited by the HDAC inhibitor. TSA stronglyenhances PKC ${zeta}$ activation and its association with Sp1. It hasbeen further demonstrated that serine 641 in the Sp1 DNA bindingdomain is important for TSA-activated LHR gene expression. Mutationof this site simultaneously diminished the Sp1 phosphorylationand LHR gene induction elicited by TSA treatment, illustratinga causative role for Sp1 phosphorylation in this process. PI3Kis responsible for PKC ${zeta}$ activation and Sp1 phosphorylation duringTSA-activated LHR gene expression. The current literature onsignaling pathways and their integration in the regulation ofgene expression by HDAC inhibitors is limited, but some reportshave indicated a role for PI3K in this process (28). Specifically,inhibition of PI3K activity has been shown to repress the apicidin-inducedphosphorylation of PKC ${varepsilon}$ and activation of the p21 gene (32).However, whether the action of TSA impacts the PI3K pathway,directly or indirectly, is not yet clear. PI3K could indirectlyaffect the activity of PKC ${zeta}$ through phosphatidylinositol 3,4,5-trisphosphategenerated by activation of its lipid kinase activity througha known substrate (13). Alternatively, PI3K could influencedownstream effectors through its intrinsic protein kinase activity.In this study, activation of PI3K was indicated by AKT phosphorylationin MCF-7 cells (data not shown), in addition to the inhibitionof this enzyme's effect by its specific inhibitors wortmanninand LY294002 and by its DN mutant form (see above). However,inhibition of AKT by its DN mutant form did not affect TSA-inducedLHR gene activation, implying a lack of participation of AKTin PKC ${zeta}$ activation. It remains to be determined whether PKC ${zeta}$ isa target for PI3K protein kinase activity, as has been recognizedfor insulin receptor substrate docking protein and more recentlyfor tropomyosin (34, 41). Our preliminary results indicatedthat treatment of JAR and MCF-7 cells with insulin or platelet-derivedgrowth factor, which are known PI3K agonists, did not mimicthe induction of LHR gene expression by TSA (data not shown).This indicates that a mechanism(s) other than a direct stimulatoryeffect on PI3 kinase activity is probably involved. TSA, alternatively,may influence protein phosphatase activity or its interactionwith another protein(s), which in turn could change the phosphorylationstatus of a target protein. The finding that HDAC inhibitorsinduced dissociation of protein phosphatase 1 from HDAC6, causinga change in the AKT phosphorylation level in prostate cancercells, favors such a mode of action (11).

Although phosphorylation of Sp1 by PKC ${zeta}$ increases its DNA bindingactivity in the upregulation of platelet-derived growth factorB-chain gene expression by nogalamycin (37), activation of vascularendothelial growth factor/vascular permeability factor geneexpression by hepatocyte growth factor, which is also dependenton PKC ${zeta}$ /Sp1 phosphorylation, does not involve a change in Sp1binding activity (38). Sp1 is ubiquitously expressed, and itsinteraction with a number of transcription factors and cofactorsis influenced by its phosphorylation status (14, 31). Modulationof such protein interactions appears to be a critical componentof the specificity of Sp1 site-dependent gene regulation. Ourstudies have shown that TSA causes prominent release of p107but not pRb from the LHR gene promoter and that p107 acts asa repressor of LHR gene promoter activity. Although pRb andp107 were originally recognized as potent repressors of E2F-regulatedgene expression during cell cycle progression (9, 22, 24), theyare also involved in nuclear receptor (Nur, SF-1, HNF4, andothers)-, TGF-ß1-, and Sp1-dependent gene regulationvia association with SRC/p160 coactivators, Smad, and Sp1 proteins(3, 10, 50, 53). Similar functional roles of pRb and p107 wereobserved for several genes, but no redundant or substitute functionsof these two proteins have been demonstrated (17). Associationof p107 and Sp1 has been observed previously in multiple celllines (HeLa, COS, 293, NIH 3T3, and mouse L cells), and p107had an Sp1 site-dependent silencing effect on the simian virus40 early promoter and the transforming growth factor ß1promoter in Saos-2 cells (18). However, it has not been definedin these studies whether the association of p107 and Sp1 issubject to regulation by the phosphorylation status of Sp1.In our study, the observations that p107 release occurred ina PKC ${zeta}$ -dependent manner and that the Sp1S641A phosphorylation-deficientmutant significantly antagonized this p107 dissociation haveindicated a clear involvement of the kinase activity, whereSp1 phosphorylation at serine 641 would elicit a conformationalchange to disrupt the balance of the p107/Sp1 complex.

Our results and a previous report have shown that HDAC inhibitor-mediatedPKC activation ( ${zeta}$ / ${varepsilon}$ ) does not affect the histone acetylation levelin the target gene promoter regions (32). TSA-induced LHR promoteractivity is significantly reduced (60 to 70%) during blockadeof Sp1 expression by its siRNA by inhibition of the activitiesof PKC ${zeta}$ and PI3K. The residual promoter activity (30 to 40%)indicates the presence of a mechanism involving histone acetylationindependently of Sp1 phosphorylation or could be related tothe residual Sp1 phosphorylation observed during repressionof the activities of PKC ${zeta}$ and PI3K (Fig. 3 and 7). On the otherhand, since marked TSA-induced histone hyperacetylation in theLHR gene promoter region is concurrent with Sp1 phosphorylation,the possibility cannot be excluded that the changes in chromatinstructure might provide a permissive state of the LHR gene promoterfavoring the phosphorylation of Sp1 and/or the release of p107.Similarly, dissociation of the HDAC/mSi3A complex from the LHRgene promoter in MCF-7 cells was independent of the effect ofPI3K/PKC ${zeta}$ and Sp1 phosphorylation. These findings indicate thatDNA methylation and histone acetylation levels in the LHR genepromoter region and dissociation of the HDAC/mSin3A complexfrom the LHR gene promoter, singly or combined, are not sufficientto elicit derepression of the LHR promoter to attain maximalactivation of LHR gene expression. Multiple layers of regulation,including PKC ${zeta}$ /PI3K-induced Sp1 phosphorylation, histone acetylation,DNA demethylation of the LHR gene promoter, and release of theinhibitory HDAC/mSin3A complex and p107, all account for TSA-inducedLHR gene activation and derepression (a model is shown in Fig.11).

View larger version (26K):
[in this window]
[in a new window]
FIG. 11. Model of transcriptional derepression and activation of LHR gene expression by TSA. The LHR gene promoter is hypermethylated in JAR cells (top) but demethylated in MCF-7 cells (bottom) (60). In the absence of TSA, the LHR gene promoter is repressed in both cell types and the HDAC/mSin3A complex and p107 are both anchored by Sp1 (left). No Sp1 phosphorylation was observed under these repressive conditions. TSA treatment of JAR cells induced LHR gene promoter-localized histone hyperacetylation without an effect on its DNA methylation status. TSA caused (i) Sp1 phosphorylation dependent on the activity of PKC ${zeta}$ and PI3K and (ii) release of p107 but not of the HDAC/mSin3A complex from the LHR gene promoter. This causes partial LHR gene derepression with a 40-fold induction of its promoter activity (top right). Upon promoter demethylation and HDAC/mSin3A dissociation by TSA-5-azacytidine treatment, maximal derepression of LHR gene expression was observed (60). In MCF-7 cells, TSA induced maximal derepression of LHR gene expression (160-fold). This was caused by histone hyperacetylation and DNA demethylation at its promoter (60) and release of the HDAC/mSin3A complex, more critically, triggered by the dissociation of p107 which resulted from Sp1 phosphorylation by PKC ${zeta}$ and PI3K (bottom right). Me, 13 methylated CpG sites within the promoter of the LHR (60); Ac, acetylated histones; P, phosphorylated; De-Me, promoter DNA demethylation.

Taken together, the present studies have demonstrated that derecruitmentof repressor protein p107 from the LHR gene promoter in a PI3K/PKC ${zeta}$ -inducedSp1 phosphorylation-dependent manner is critically importantfor TSA-induced activation of the LHR gene.

ACKNOWLEDGMENTS

We thank Alex Toker (Department of Pathology, Beth Israel DeaconessMedical Center, Harvard Medical School, Boston, MA), Bai Lu(Laboratory of Cellular and Synaptic Neurophysiology, NationalInstitute of Child Health and Human Development, NIH, Bethesda,MD), and Adrian Black (Department of Pharmacology and Therapeutics,Roswell Park Cancer Institute, Buffalo, NY) for kindly providingthe dominant negative mutant PKC ${zeta}$ , dominant negative mutant PI3Kp110 ${alpha}$ ,and Flag-Sp1 construct, respectively.

Our research is supported by the Intramural Research Programof the National Institute of Child Health and Human Developmentof the National Institutes of Health.

FOOTNOTES

* Corresponding author. Mailing address: Bldg. 49, Rm. 6A-36, 49 Convent Dr. MSC 4510, National Institutes of Health, Bethesda, MD 20892-4510. Phone: (301) 496-2021. Fax: (301) 480-8010. E-mail: dufaum{at}mail.nih.gov.

REFERENCES

Top