![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Molecular and Cellular Biology, April 2007, p. 2919-2933, Vol. 27, No. 8
0270-7306/07/$08.00+0 doi:10.1128/MCB.00936-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Daniel G. Tenen,5 and
Trang Hoang1,2*
Institute of Research in Immunology and Cancer,1 Departments of Pharmacology, Biochemistry, and Molecular Biology, University of Montreal, Montréal, Québec, Canada H3C 3J7,2 Department of Hematology, Dokkyo University School of Medicine, Tochigi 321-0293, Japan,3 Department of Biology, York University, North York, Ontario, Canada M3J 1P3,4 Harvard Institutes of Medicine and Harvard Stem Cell institute, Harvard Medical School, Boston, Massachusetts 021155
Received 26 May 2006/ Returned for modification 25 July 2006/ Accepted 16 January 2007
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Macrophages are derived from bone marrow myeloid precursors and are activated by a large variety of stimuli including phorbol esters that evoke a stress-like response (57) and induce a repertoire of stress response genes that include interleukin-1ß (IL-1ß) (9, 70). Several transcription factors have been implicated in macrophage stress response. This includes c-Jun (4, 15, 35, 69), a basic leucine zipper transcription factor that can either homodimerize or heterodimerize with other members of the AP-1 family such as JunB, JunD, and c-Fos (reviewed in reference 18) and activate the expression of target genes through AP-1 binding sites (10). Although c-Jun homodimers are competent for transcription activation, it is not clear whether or not these homodimers have distinctive functions compared to heterodimers (10). The ETS transcription factor PU.1 drives the transcription of a large number of myelomonocytic genes (6, 32, 40, 42, 64) and plays essential roles in the development of myeloid and lymphoid cells (13, 28, 48, 52, 68). Finally, CCAAT/enhancer-binding protein ß (C/EBPß), a basic leucine zipper transcription factor of the C/EBP subfamily, is essential for macrophage activation and phagocytosis (75). c-Jun, PU.1, and C/EBPß have been shown to physically interact with each other (5, 27, 42) and enhance the transcription of monocyte-specific genes via binding to their respective sites on DNA (40, 50). Nonetheless, how PU.1, C/EBPß, and AP-1 govern macrophage activation and macrophage stress response remains to be documented.
The assembly of the preinitiation complex (PIC) on promoters is a rate-limiting step in transcription. Much effort has been dedicated to defining components of the PIC and their assembly on strong promoters containing multimerized high-affinity binding sites in vitro and in transfected cells. Furthermore, a large number of sequence-specific DNA binding activators have been identified, and their functional importance in gene expression and in specifying cell fate and/or homeostasis has been well documented. However, whether these activators influence the process of PIC assembly on endogenous promoters is largely unexplored in mammalian cells. Gene expression has also been linked with chromatin modification and remodeling, implicating, for example, histone H3 and H4 acetylation (46). In particular, histone tails acetylated at specific lysine residues can serve as docking sites for bromodomains and could facilitate the recruitment of bromodomain-containing proteins or complexes, e.g., the general transcription factor TFIID, to chromatin (1, 29). The formation of an enhanceosome at the interferon-ß promoter in response to viral infection provides one of the rare examples in which these mechanisms were addressed, and it was shown that the assembly of this particular multiprotein complex to the interferon-ß promoter leads to chromatin modifications, nucleosome remodeling, and PIC assembly (1, 2, 30, 44, 45). PU.1 has been shown to increase chromatin accessibility and transcription at target loci (56). However, facilitation of PIC assembly on promoters that are already poised for transcription has not been addressed. In the present study, we show that the interaction of DNA-bound PU.1 and C/EBPß recruits c-Jun as a coactivator and facilitates RNA polymerase II (Pol II) recruitment.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Tissue culture. The erythro-monocytic TF-1 cell line (31) was a kind gift from Kitamura (DNAX, Palo Alto, CA). The cells were maintained in Iscove's modified Dulbecco's medium (IMDM; Gibco/Invitrogen) supplemented with fetal calf serum (FCS; 10%) and granulocyte-macrophage colony-stimulating factor (200 pmol/liter) and were passaged every 48 h at a concentration of 1.5 x 105/ml in nonadherent culture dishes. The RAW 264.7 monocytic cell line was maintained in IMDM-10% inactivated FCS and passaged 1:4 to 1:5 to maintain 75 to 80% confluence in Falcon adherent tissue culture dishes. COS-7 cells and F9 embryonic carcinoma cells maintained in IMDM containing 10% FCS were passaged three times weekly (in gelatin-coated dishes for F9 cells).
TPA induction experiments. RAW or TF-1 cells were incubated for 4 h or 24 h, respectively, in the presence of TPA (22, 25). Granulocyte-macrophage colony-stimulating factor (200 pM) was also added to TF-1 cells. TF-1 cells were transferred to adherent tissue culture dishes during TPA treatment for selection of adherent differentiated TF-1 cells with macrophage-like phenotypes. After TPA treatment, TF-1 differentiated or RAW adherent cells were harvested and subjected to lysis according to the different experimental protocols.
Plasmid constructs.
An IL-1ß promoter fragment linked to the chloramphenicol acetyltransferase gene and containing 4.4-kb upstream sequences of the human IL-1ß gene was generously provided by John Hiscott (Lady Davis Institute for Medical Research, Montreal, Canada) and was used as a template to generate luciferase reporter constructs. The IL-1ß131 construct contains sequences extending from –131 to +11 bp (–131/+11 fragment) of the IL-1ß promoter cloned in the promoterless luciferase reporter vector pXPII from a HindIII- and BglII-digested PCR fragment (sequence underlined) that was amplified with the oligonucleotides IL-1P-302 (5'-TGAAGCTTGGTACCTAACGTGGGAAAATCC-3') and IL-1P-+11 (5'-AAGCTTAGATCTAGAGGTTTGGTATCTG-3'). Mutagenesis of the PU.1 site at position –45 (relative to the underlined G in the core AGAA on the minus strand) and of the C/EBPß site at –90 (relative to the underlined T in TTGTGAAAT of the plus strand) of the –131/+11 promoter fragment created the m45 and m90 templates, respectively. These were generated by overlapping PCR mutagenesis using the following mutated oligonucleotides: for m45, 5'-TCAGCCTCCTACTTAGGCTTTTGAAAGCTA-3' and reverse; for m90, 5'-TAACTTGACCGTGAATTCAGGTATTCAACAG-3' and reverse. The underlined residues indicate the positions of the core ETS motif and C/EBP site at the –45 and –90 region, respectively, while residues in bold are mutations. The PU.1 –45 and C/EBP –90 double mutant (dm) form of IL-1ß131 was generated by overlapping PCR mutagenesis of IL-1ß131 mutated at the –45 PU.1 site with the mutated oligonucleotide for the –90 C/EBP site. All of the IL-1ß131 promoters cloned in pXPII were subcloned as BamHI/XbaI fragments (XbaI site in the 5' end of the luciferase reporter gene) into the BamHI and XbaI sites of pBluescript to generate biotinylated IL-1ß131 templates (see below). The murine PU.1 cDNA was PCR amplified with EcoRI-containing primers and subcloned as an EcoRI fragment in the murine stem cell virus (MSCV) expression vector (MSCV-PU.1). Murine PU.1 as an EcoRI fragment was subcloned into the EcoRI site of the pSG5-modified vector p513 (SV40-PU.1, where SV40 is simian virus 40). Human PU.1 cloned into the EcoRI/BamHI site of pGBT9 was transferred as a GAL4 DNA binding domain (DBD) fusion DNA fragment into the HindIII/BamHI sites of pcDNA3 (mammalian expression vector GAL4DBD-PU.1). The 5xGAL4UAS-tk109-luciferase reporter construct (five GAL4 binding sites upstream of 109 bp of the thymidine kinase promoter in pXP2-based luciferase reporter vector); the mammalian expression vector for ß-galactosidase (CMV-ßGal, where CMV is cytomegalovirus); and the mammalian SV40-driven expression vectors for murine c-Jun in its wild-type (wt) form, deleted from amino acids (aa) 251 to 276 (a deletion of the basic domain [
B]; aa 257 to 282 in human c-Jun according to reference 8), deleted from aa 6 to 194 (a deletion of the transactivation domain[TD]), or the S63A/S73A double point mutation mutant (c-JunAA) were obtained from Mona Nemer (IRCM, Montreal, Quebec, Canada). The murine C/EBPß expression vector (murine sarcoma virus C/EBPß) obtained from Alan D. Friedman (Johns Hopkins University, Baltimore, MD) was used to subclone C/EBPß as an EcoRI/XhoI fragment into the mammalian expression vector pCI (CMV-C/EBPß). Expression vectors driven by the SV40 enhancer for wt human c-Jun, c-Fos, JunB, and JunD cDNAs were gifts of Michael Karin (University of California, San Diego, CA). cDNAs for human c-Jun containing point mutations in the basic or leucine zipper domains were obtained from Dirk Bohmann (University of Rochester, Rochester, NY). Human wt c-Jun or point mutation mutants (as described in references 8 and 59) were HA tagged at their C terminus as described previously (76) before transfer into a human ubiquitin C promoter-driven expression vector (79) (these were named c-JunHA or mutHA). The c-JunHA vector was used to generate the point mutation mutants R270I/N271D (M13) or C278D/R279I (M14b) by PCR with the oligonucleotides GAGGAAGCGCATGATCGACCGCATCGCTGCCTCC (and reverse; mutated nucleotides are in bold) or CGCATCGCTGCCTCCAAGGACATAAAAAGGAAGCTGGAGAGAA (and reverse; mutated nucleotides are in bold), respectively, using the Pwo DNA polymerase as described previously (76). The M13 vector was then used to generate the double mutant R270I/N271D-C278D/R279I (M13-14b) as described above. Amino acid positions are as described by Bohmann and Tjian (8). All mutants were verified by DNA sequencing. The c-JunHA or mutHA fragments were also transferred in the pSG5-modified vector p514 that allows in vitro translation from a T7 promoter. The glutathione S-transferase (GST)-C/EBPß (aa 22 to 306) fusion construct was obtained from Wen-Hwa Lee (University of California, Irvine, CA), and the GST-murine PU.1 (aa 1 to 282) fusion construct was obtained from Tony Kouzarides (Wellcome Trust/Cancer Research UK Gurdon Institute, Cambridge, United Kingdom).
Transfection protocols and luciferase assays.
TF-1 cells were transfected by electroporation. Cells were passaged 24 h before electroporation at a concentration of 3 x 105 cells/ml. Exponentially growing cells were then concentrated at 2.5 x 107 cells/ml and electroporated at 350 mV using a Bio-Rad electroporator with 12 µg of reporter DNA and 250 ng of CMV-ßGal used as an internal control for the experiment. The total amount of transfected DNA was 30 µg, with the balance made up of pGEM4 as a carrier DNA. Cell lysates were prepared 12 h after transfection and assayed for luciferase activity, which is expressed as relative light units after normalization for ß-Gal activity as described previously (38). For TPA induction, cells were treated immediately after electroporation with TPA or dimethyl sulfoxide alone for the indicated times. F9 cells were subjected to calcium phosphate transfection, as described previously (38). F9 cells were passaged 24 h before transfection at a concentration of 1 x 104 cells/22-mm dish. The cells were transfected with 1 µg of reporter DNA, 250 ng of CMV-ßGal, and different molar ratios of the following expression vectors: MSCV-PU.1 (1 µg), CMV-C/EBPß (250 ng), SV40-c-Jun (10 to 400 ng), ubiquitin C/c-JunHA or mutHA (10, 50, or 150 ng), SV40-c-JunAA (10 ng), and SV40-Fos (10 ng). The total amount of DNA transfected was balanced with pGEM4. Cell lysates were prepared 24 h after transfection, assayed for luciferase activity, and normalized from ß-Gal activities. Rous sarcoma virus-Luc was used as an external control for all the transfections, and pXPII was also transfected as a negative control. For mammalian one-hybrid assays, F9 cells were cotransfected with combinations of expression vectors for CMV-GAL4DBD-PU.1 (3 ng), SV40-c-Jun (300 ng), SV40-c-Jun B (300 ng), SV40-c-Jun TD (300 ng), SV40-JunB (300 ng), and SV40-JunD (300 ng). The reporter gene used was 5xGAL4UAS-tk109-luciferase (1 µg), and transfection efficiency was normalized with the ß-Gal expression vector (250 ng). The total amount of DNA transfected was balanced with pGEM4. All transfections were performed through calcium phosphate precipitation, and cells were lysed 24 h posttransfection for luciferase and ß-Gal assays. Data shown are the averages ± standard deviations of triplicate determinations and are typical of at least three independent experiments.
Electrophoretic mobility shift assay (EMSA). A 1-µl portion of in vitro translation reaction mixtures was incubated with 20 to 50 fmol of 32P-labeled IL-1ß promoter DNA fragments (–131 to +11 from a HindIII/BglII digestion of the IL-1ß131 vector) or double-stranded oligonucleotides containing on the plus strand the sequence TTCCGGCTGACTCATCAAGC (AP-1 site is underlined) in the presence of 0.1 µg/µl double-stranded poly(dI-dC) as described previously (38). The samples were then resolved through electrophoresis on a 4% polyacrylamide gel (bisacrylamide:acrylamide at a 1:19 ratio) in 0.25x TBE (45 mM Tris-borate, 1 mM EDTA) buffer at 150 V at 4°C. The retarded protein complexes were then exposed to a PhosphorImager screen for visualization.
Retroviral-mediated gene transfer. TF-1 cells were engineered to stably express a c-Jun antisense construct using the MSCV (23), as described previously (36). Briefly, the entire coding region of c-Jun was cloned in the antisense orientation in the MSCV retroviral vector. Amphotropic viruses were produced by transient transfection into a packaging cell line (36). For retroviral infection, 1 x 106 exponentially growing TF-1 cells were presensitized with polybrene at 2 µg/ml for 24 h and cocultured with virus-producing cells for another 24 h. Nonadherent TF-1 cells were separated from the infected fibroblasts. A polyclonal population was analyzed 7 days after selection in G418 at 1 mg/ml.
Immobilization of biotinylated DNA promoter templates on magnetic resin.
DNA promoter templates were biotinylated and immobilized on magnetic resin for the DNA pull-down assays (see below). The wt and mutant IL-1ß131 promoter fragments subcloned in pBluescript were cut at their 5' ends with BamHI or Xho, filled in with Klenow polymerase in the presence of biotin-14-dATP or biotin-14-dCTP, respectively, and complementing deoxynucleoside triphosphates (dNTPs). After biotinylation at the 5' end, the IL-1ß promoter templates were cut at their 3' ends with BglII or Xba and separated from vector fragments by agarose gel electrophoresis. The approximately 150-bp BamHI/BglII biotinylated IL-1ß fragment (IL-1ß131 short) or 300-bp Xho/Xba fragment (IL-1ß131 long) was cut from the agarose gel and purified with QN+-butanol (37). The resulting biotinylated IL-1ß131 fragments are designated IL-1ß131 wt, m45 (mutated fragment in the PU.1 –45 site), m90 (mutated fragment in the C/EBP –90 site), and dm (doubly mutated at both the –45 and –90 sites). Generation of a biotinylated TPA response element (TRE) was obtained by Klenow fill-in of the 5' overhang created after annealing of the sense AATTCTTCCGGCTGACTCATCAAGC (AP-1 site is underlined) and antisense GCTTGATGAGTCAGCCGGAAG oligonucleotides as described previously (16). Equal molar quantities of biotinylated DNA templates were than immobilized on magnetic resins conjugated with streptavidin according to the manufacturer's instructions (IL-1ß131 short and TRE; final concentration, 0.06 pmol/µl). IL-1ß131 long fragments were immobilized at 0.1x the concentration used for the short version and 32P-labeled at their 3' ends with [
-32P]dCTP with Klenow polymerase and complementing dNTPs. After labeling, the resin was washed twice with 10 mM Tris-HCl, pH 7.4, 2 M NaCl, and 20 mM EDTA and twice with 5 mM Tris-HCl, pH 7.4. Immobilized promoter templates were kept in 5 mM Tris-HCl buffer, pH 7.4, at 4°C for up to 3 months (IL-1ß131 or TRE) or 1 month (32P-labeled IL-1ß long).
DNA pull-down assay. The DNA pull-down assay was optimized from and is described in reference 20 as a DNA binding assay to immobilized promoters. Nuclear extracts (NE) from RAW cells treated with TPA or from COS-7 transfected cells were prepared as described previously (38). NE from COS cells overexpressing PU.1 (from SV40-PU.1), C/EBPß (from CMV-C/EBPß), c-Jun (from SV40-c-Jun), c-JunHA (from ubiquitin-c-JunHA or mutHA), or recombinant human c-Jun purified from E. coli were incubated with 0.3 pmol (IL-1ß131 short and TRE) or 0.015 pmol (32P-labeled IL-1ß131 long) of immobilized DNA templates in binding buffer (20 mM Tris [pH 8.0], 10% glycerol, 6.25 mM MgCl2, 5 mM dithiothreitol, 0.1 mM EDTA, 0.01% NP-40) in a final concentration of 100 mM NaCl and in the presence of 0.2 µg/µl of poly(dI-dC). Quantities of NE overexpressing C/EBPß, PU.1, c-Jun, or of purified recombinant c-Jun (150 ng) used in the assay were maintained below saturation of the DNA binding sites and optimized to allow for reproducible detection of PU.1, C/EBPß, and c-Jun assembled on immobilized IL-1ß templates by Western blotting. The concentration of COS NE was maintained constant at 0.3 µg/µl with NE obtained from untransfected COS cells in a minimum volume of 100 µl. After a mixing step, the samples were incubated by rotation for 60 to 90 min at 30°C. The resin coated with immobilized templates was then washed twice with 0.3 ml of binding buffer. The proteins bound to the immobilized templates were eluted from the templates by boiling of the beads for 5 min in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer (IL-1ß short and TRE) or subjected to DNase I treatment (32P-labeled IL-1ß131 long) as described below. Proteins in SDS-PAGE sample buffer were resolved on polyacrylamide gels and transferred to polyvinylidene difluoride membranes for Western blot analysis. Typically, 25% of each sample was used for the detection of C/EBPß and PU.1 while 50% was used for the detection of c-Jun.
ChIP. Chromatin immunoprecipitation (ChIP) assays were performed as described previously (38) with the use of 1 x 106 to 2 x 106 cross-linked TPA-treated RAW or TF-1 cells and 5 µg of anti-PU.1, anti-C/EBPß, anti-c-Jun, anti-c-Fos, anti-AcH3, or anti-RNA Pol II antibodies. Rabbit IgGs were used as a control for PU.1, C/EBPß, c-Jun, c-Fos, and acetylated-H3 antibodies while anti-HA antibodies in ascites fluid was used as a control for anti-RNA Pol II antibodies. SyberGreen real-time PCR was performed on an Mx3000 apparatus (Stratagene, La Jolla, CA) using a SYBR Green Stratagene PCR kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The following oligonucleotides were used for amplification: for the mouse IL-1ß promoter, 5'-CCACCCTTCAGTTTTGTTGT and 5'-CCCAGATGAGCCTATTAGGCCT; for the human IL-1ß promoter, 5'-AGGTCTATAGGTGGCTGCTTAACT and 5'-TGGCTTGTTTGAAAATGTGTTAGCA; for the mouse Mmp12 promoter, 5'-CTAATGGAGTTCTG and 5'-GCAGCTCATCAACCTGTTCC; for a region of inactive chromatin in the mouse c-Kit locus at position –4191, 5'-TGTGGGGGCTCCTGGTCTTA and 5'-TAGCGGCGCGCGACAG; and for a region in the human c-Kit locus at position +13532, 5'-AGGTCTATAGGTGGCTGCTTAACT and 5'-TGGCTTGTTTGAAAATGTGTTAGCA. Forty cycles of amplification were performed, followed by cycles of denaturation and annealing steps. Amplification plots and dissociation curves were analyzed with the Mx3000p (Stratagene, La Jolla, CA) and Excel (Microsoft, Redmond, CA) software programs.
Reverse transcription-PCR. Total RNA was prepared according to Chomczynski and Sacchi's protocol (11), using tRNA as carrier for ethanol precipitation. First-strand DNA synthesis was performed as described previously (24). A total of 2 µl of cDNA sample was added to the PCR mixture containing a 1 µM concentration of each specific 5' and 3' primer, a 1 mM concentration of each dNTP, 1.5 mM MgCl2, and 1 U of Taq DNA polymerase (Gibco). Primers corresponding to murine sequences are the following: S16 sense, 5'-AGGAGCGATTTGCTGGTGTGA-3'; S16 antisense, 5'-GCTACCAGGGCCTTTGAGATG-3'; S16 internal, 5' AAATTTATGCCATCCGACAGTC-3'; IL-1ß sense, 5'-CAAAATACCTGTGGCCTTGGG-3'; IL-1ß antisense, 5'-AAAACACAGGCTCTCTTTGAACAGA-3'; and IL-1ß internal, 5'-AGAAGTCAAGAGCAAAGTGG-3'. Thirty cycles of amplification were performed at a melting temperature of 55°C for S16 and 58°C for IL-1ß; 15 µl of each reaction mixture was loaded on 1% agarose gel, transferred on nylon membrane, and hybridized with the corresponding internal oligonucleotide probes. The hybridization signals were analyzed with an FLA-5000 apparatus (Fuji).
DNase I footprinting.
The DNase I footprinting procedure was adapted from Sandaltzopoulos and Becker (66), using immobilized DNA templates as described above. However, the concentration of 32P-labeled IL-1ß131 templates was 20-fold less in order to allow for near maximal occupancy of the DNA binding sites. In addition, maximal amounts of poly(dI-dC) were used (0.8 µg/µl) to minimize nonspecific interactions with immobilized IL-1ß templates. After incubation of C/EBPß (COS NE), PU.1 (COS NE), and c-Jun (purified recombinant; 300 ng) with immobilized 32P-labeled IL-1ß131 templates for 90 min at 30°C, beads were washed twice in binding buffer and kept in 25 µl of binding buffer. An equal volume of DNase I diluted in binding buffer supplemented with 0.1 mg/ml bovine serum albumin, 5 mM CaCl2, 10 mM MgCl2 was added to the beads for 1 min at 22°C. The optimal quantity of DNase I used was determined empirically on naked immobilized templates (4 x 10–4 U). The reaction was stopped, and the beads were washed as described previously (66) with stop and wash buffers supplemented with 0.01% NP-40, and the beads were kept in loading buffer. A Sequenase (USB) DNA sequencing ladder was obtained with the reverse oligonucleotide CTA GAG GAT AGA ATG GCG and [-32P]dCTP. DNase I-nicked fragments attached to the beads were eluted, denatured along with the sequencing ladder, and separated on a sequencing gel as described previously (66). The resolved 32P-labeled DNA fragments were then detected by exposure to a phosphor imager screen (FLA-5000; Fuji).
Pull-down assay. Pull-down assays were performed essentially as described previously (38). Briefly, [35S]methionine-labeled wt or mutant Jun proteins (1 to 2 µl) were incubated with 1 µg of GST, GST-PU.1, or GST-C/EBPß purified from E. coli in 250 µl of binding buffer (40 mM HEPES, pH 7.8, 50 mM KCl, 0.2 mM EDTA, 5 mM MgCl2, 0.1% Triton X-100, 10% glycerol, 1.5 mM dithiothreitol). The GST beads were preincubated with bovine serum albumin at 0.5 mg/ml for 1 h at 4°C. Bound proteins were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membranes before detection and quantification by exposure to a phosphor imager screen (FLA-5000; Fuji).
Northern blot analysis. RNA extracted from cells using the acid guanidinium-phenol method and resolved by electrophoresis in 1% formaldehyde-containing agarose was transferred and blotted with a 32P-labeled IL-1ß cDNA fragment as previously described (25).
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
PU.1, C/EBPß, and c-Jun from macrophages interact with immobilized IL-1ß promoter templates. To further address the recruitment of PU.1, C/EBPß, and c-Jun to the IL-1ß promoter and the mechanism through which these proteins assemble on DNA, we optimized the immobilized DNA template assay described previously (20). In this assay (DNA pull-down), DNA-interacting proteins are retained on immobilized DNA templates (Fig. 3A), eluted after elimination of unbound proteins, and revealed by Western blot analysis, thereby avoiding the potentially disruptive force of the electric field used in EMSA. The immobilized templates used for this assay (Fig. 3B) are the IL-1ß proximal promoter in its wt, m45, m90, or dm form and double-stranded oligonucleotides containing a consensus AP-1/TRE binding site. Binding assays were performed with NE from TPA-activated mouse RAW cells. As shown in Fig. 3C, Western blot analysis of proteins bound on IL-1ß templates indicates that C/EBPß, PU.1, and c-Jun interact with the IL-1ß promoter and that these interactions are disrupted when both the –45 and –90 sites are mutated. These observations indicate that c-Jun associates with the IL-1ß promoter in the absence of an AP-1 binding site.
|
|
The levels of PU.1 bound to DNA determined the extent of recruitment of c-Jun (Fig. 4E, compare lanes 1 and 2) without affecting that of C/EBPß. This suggests a role for PU.1 in tethering c-Jun to the IL-1ß promoter, which was further confirmed through the importance of the integrity of the –45 PU.1 binding site (Fig. 4E, compare lanes 3 and 4). Mutation of the C/EBPß binding site at –90 disrupted most of C/EBPß binding and reproducibly decreased PU.1 binding twofold (Fig. 4E, lane 5). Furthermore, c-Jun binding was decreased fourfold (lane 5) while mutation of both sites completely abrogated PU.1, C/EBPß, and c-Jun binding altogether (Fig. 4E, lane 6). Our binding assay therefore reveals that the presence of PU.1 and C/EBPß on their respective DNA sites allows for the recruitment of c-Jun to the IL-1ß promoter in the absence of an AP-1 site. Furthermore, both PU.1 and C/EBPß binding to the IL-1ß promoter appears to be reinforced by the presence of the other partner on DNA. This was most likely due to protein-protein interaction(s), as assessed by the residual binding of PU.1 or C/EBPß to promoter templates that harbor individual mutations, which was not observed when both sites were mutated (Fig. 4E, compare lane 4 or 5 with lane 6).
The importance of the PU.1 and C/EBP binding sites in recruiting the complex was further confirmed by the capacity of double-stranded oligonucleotides covering these sites to specifically displace DNA binding in our assay (data not shown). Surprisingly, a large excess of free double-stranded oligonucleotides containing an AP-1/TRE binding site that specifically displaces c-Jun binding to an immobilized TRE (Fig. 4D, lane 4) also prevented c-Jun recruitment to the IL-1ß promoter without affecting C/EBPß and PU.1 binding (Fig. 4E, lane 7).
Together, our observations suggest that the recruitment of c-Jun to DNA may be mediated by protein-protein interactions with DNA-bound PU.1 and C/EBPß and that the DNA-binding domain of c-Jun may be involved in these interactions, possibilities that will be further tested.
To test the possibility of cooperative DNA binding between C/EBPß and PU.1, we assessed the impact of C/EBPß on PU.1 binding at a lower concentration (fourfold less COS NE containing PU.1). Under these conditions, PU.1 binding was increased by C/EBPß, and this increase depended on the integrity of the –45 and –90 DNA sites (Fig. 5A, lanes 1 to 4) and on the concentration of C/EBPß (data not shown). To further assess the mechanism of C/EBPß and PU.1 DNA binding, we used a 3' extended and 32P-labeled version of IL-1ß131 templates immobilized on magnetic beads (Fig. 5B) for DNase I footprinting. In this assay, C/EBPß alone specifically protects a region encompassing the core –90 C/EBPß binding site (Fig. 5B, lane 3). In contrast, no significant protection of the –45 PU.1 binding site is observed in the presence of PU.1 alone (Fig. 5B, lane 4). This site, together with immediate upstream sequences, was nonetheless protected when both proteins were used together (Fig. 5B, lane 5), indicating that C/EBPß can help PU.1 bind to the IL-1ß promoter when the concentration of PU.1 is limiting, as observed in Fig. 5A. Together, our DNA binding assays suggest that DNA-bound C/EBPß can facilitate the binding of PU.1 to its DNA binding site by a mechanism that potentially involves protein-protein interaction(s).
|
Together, our DNA binding data indicate that C/EBPß, PU.1, and c-Jun form a complex that can assemble cooperatively on the IL-1ß promoter via C/EBPß and PU.1, directly tethered to their respective DNA binding sites. We suggest that this anchoring then allows recruitment of c-Jun to the IL-1ß promoter via protein-protein interactions. Our observations also suggest that the assembly of a competent multiprotein complex on specific regulatory sequences is a limiting step in c-Jun-dependent transcriptional synergy.
PU.1, c-Jun, and C/EBPß interact in vitro and in vivo. The association of proteins with DNA is often strengthened by direct protein-protein interactions. c-Jun associates with PU.1 or C/EBPß in vitro (5, 27), suggesting that c-Jun can be recruited to DNA by PU.1 and C/EBPß in the absence of a canonical AP-1 binding site. Pull-down assays with GST-PU.1 and GST-C/EBPß confirm that both proteins interact with c-Jun in vitro (Fig. 6A), as expected (5, 27). Moreover, the interaction was also observed with JunB and JunD, two family members that share 86% and 95% homology with c-Jun basic leucine zipper domains, as illustrated in Fig. 6B. Deletion of the basic domain of c-Jun disrupted its interaction with C/EBPß and PU.1 (Fig. 6A). Together, our results indicate that c-Jun interacts through its basic domain with PU.1, as reported previously (5), and with C/EBPß. The involvement of the basic domain in protein-protein interaction is consistent with the observation that a TRE compromised the recruitment of c-Jun to immobilized IL-1ß promoter templates by DNA bound PU.1 and C/EBPß (Fig. 4E). Finally, we observed that PU.1 interacts through its ETS-containing C-terminal domain with C/EBPß (data not shown), consistent with the effect of C/EBPß on PU.1 binding to the IL-1ß promoter (Fig. 5A and C).
|
These same observations were made using the IL-1ß 131-luciferase reporter (compare Fig. 6C and D). Together, our results indicate that the N terminal TD of c-Jun confers functional specificity in transcriptional cooperativity with PU.1, while the basic domain is important for its interaction with both C/EBPß and PU.1.
Because c-Jun can form homodimers or heterodimers with c-Fos, we addressed the role of c-Fos in the context of the IL-1ß promoter (Fig. 6E). Our results show that c-Fos was functionally inactive in potentiating PU.1 and C/EBPß. Furthermore, the activity of c-Jun is not enhanced by c-Fos and is even reduced, suggesting that heterodimerization with c-Fos is not required (Fig. 6E). We also addressed the role of the Ser63 and Ser73 Jun N-terminal protein kinase (JNK) phosphoacceptor sites (14) in synergy with PU.1. For that purpose we used a c-Jun mutant that harbors an Ala instead of a Ser at these positions (Fig. 7, c -JunAA). These mutations did not affect synergy with PU.1 (Fig. 6F), indicating that this activity is independent of phosphorylation by JNK. Together, these observations indicate that c-Jun interacts with PU.1 and C/EBPß to increase transcriptional output through a mechanism which is c-Fos and JNK independent.
|
We next tested the activity of these mutants in transcriptional synergy with PU.1 and C/EBPß (Fig. 8A). At low levels, the mutant with a reduced capacity to interact with PU.1 only (M14) exhibits decreased synergy with PU.1 alone (not shown) or in combination with C/EBPß (Fig. 8A, left panel), correlating with its decreased capacity to be recruited to immobilized IL-1ß templates (Fig. 8B, lane 2). However, at high levels, when wt c-Jun followed a bell curve shape, the M14 mutant did not follow this trend and remained active. In contrast, the mutants that show reduced (M13 and M14b) or no (M22-23) capacity to interact with C/EBPß are almost inactive (Fig. 8, middle panel) or totally inactive (Fig. 8A, right panel) in synergy with C/EBPß and PU.1. This correlated with a severe incapacity for these mutants (M13, M14b, and M22-23) to be recruited to immobilized IL-1ß templates (Fig. 8B, lanes 3, 7 and 8; most if not all of the signal observed in lanes 7 to 9 corresponds to nonspecific interaction as observed in lane 10 and described in the figure legend). The mutant incapable of interaction with both C/EBPß and PU.1 (M13-14b) shows both an absence of synergy (Fig. 8A, middle panel) and a severe incapacity to be recruited to immobilized templates (Fig. 8B, lane 9). Differences in binding to immobilized templates were not due to differing protein levels, as revealed by comparable input levels in each binding samples (Fig. 8B, input panels). Together, these results indicate that residues R270/N271 and K277/C278/R279 in the basic domain of c-Jun are essential for interactions and transcriptional synergy with PU.1 and C/EBPß. Furthermore, the structural integrity of the leucine zipper domain is essential in transcriptional synergy, despite the fact that it is dispensable for interaction with PU.1. We therefore tested the mutant that exhibits increased homodimer stability and impaired heterodimer stability with Fos (M17). This mutant mediated a twofold more robust synergy with C/EBPß and PU.1 compared to the wt protein (Fig. 8A, right panel), consistent with the functional importance of c-Jun homodimers. Furthermore, increased transcriptional activity does not correlate with increased recruitment to immobilized IL-1ß promoter (Fig. 8B, lane 5), suggesting that this higher activity was determined by a step that occurs after complex assembly.
|
|
Combined, our in vitro and in vivo results suggest that the complex formed on the IL-1ß promoter between PU.1, C/EBPß, and c-Jun directly facilitates the assembly of the RNA Pol II machinery, thereby enhancing transcriptional output.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Consequences of the assembly of c-Jun and DNA-bound PU.1 and C/EBPß in a complex on Pol II recruitment. Transcription activation by PU.1 or C/EBPß is enhanced through association with multiple protein partners (26, 53, 60, 61). However, subsequent events that follow their binding to cis-regulatory sequences remain to be ascertained. In the present study, we show that PU.1 and C/EBPß mutually facilitate their association with DNA via specific binding to cis-acting sequences. This allows for c-Jun recruitment despite the absence of an AP-1 site and increased RNA Pol II occupation of a target promoter.
The assembly of the RNA Pol II machinery into transcriptional PICs is a very limiting step in transcription activation that can be facilitated by physical interactions with promoter-bound activators (12, 39, 62, 63). One consequence of the assembly of PU.1, C/EBPß, and c-Jun on the IL-1ß promoter observed here is to allow for maximal recruitment of the RNA Pol II to this promoter. PU.1, C/EBPß, or c-Jun have all been reported separately to interact with basal Pol II transcription factors, suggesting that these factors may together facilitate PIC assembly. Indeed, PU.1 was shown to interact with CBP (80), C/EBPß with the CRSP130/Sur2 protein present in Mediator complexes or p300 (49, 67), and c-Jun with TFIIE, TFIIF, or TAF1 (43, 47). Together with our observations of a more efficient recruitment of RNA Pol II, we propose that maximal assembly of PIC on the IL-1ß promoter is achieved by multiple interactions and recruitment of TFIID and RNA Pol II holoenzymes containing Mediator complexes, CBP/p300, and some general transcription factors by the PU.1-C/EBPß-c-Jun complex.
The presence of this complex could also facilitate other limiting steps in mRNA production following initial PIC assembly such as transcription reinitiation or promoter clearance (17, 63, 78, 81). The mechanism described here is, nonetheless, distinct from the cooperation between the long form of C/EBPß and Myb at the mim-1 promoter that probably necessitates a derepression step involving the recruitment of the SWI/SNF complex and chromatin remodeling before activation per se (34). Indeed, the IL-1ß gene is already transcriptionally active in macrophages and macrophage progenitors. Furthermore, the C/EBPß isoform detected in the two progenitor/macrophage cell lines studied here does not appear to contain the extra N-terminal 22 amino acids that characterize the long form of C/EBPß. Our data are consistent with an important role for the PU.1/C/EBPß/c-Jun complex in RNA Pol II recruitment and transcription activation.
c-Jun homodimer as a coactivator. AP-1 transcriptional activity is determined by the nature of the dimerization partner. Hence, c-Jun-c-Fos heterodimers bound to DNA via AP-1 sites exhibit the highest transcriptional activity (reviewed in reference 10). In contrast, we show here that when c-Jun is in a complex with PU.1 and C/EBPß, c-Jun activity is highest as homodimers, whereas heterodimerization with c-Fos decreases transcriptional synergy. Furthermore, a c-Jun mutant with increased homodimer stability has a twofold-impaired capacity to interact with PU.1 and C/EBPß yet exhibits increased transcriptional synergy with PU.1 and C/EBPß. These observations suggest that a postrecruitment step which is limiting for transcriptional output may be facilitated by c-Jun homodimers, possibly via additional protein-protein interaction(s) with the PIC, as discussed below. Our observations also suggest that in this context, c-Jun does not function as a heterodimer with C/EBPß, consistent with published results (54, 73, 77). Indeed, a potential c-Jun/C/EBPß heterodimer is unlikely to bind an AP-1 site, while we show here that a TRE oligonucleotide displaces all of c-Jun recruitment to the IL-1ß promoter without affecting C/EBPß binding.
Basic residues are often involved in specific electrostatic bonds between two proteins. Here, we show that the basic domain of c-Jun serves as a protein interaction interface with PU.1 and C/EBPß, specifically, residues R270/N271 and K277/C278/R279 (amino acid positions are according to reference 8) that are otherwise involved in DNA binding. These observations suggest that c-Jun DNA binding and protein-protein interactions are mutually exclusive, a hypothesis which is further supported by additional lines of evidence. Initially, we observed that c-Jun recruitment to DNA-bound PU.1 and C/EBPß is out-competed by an excess of the TRE double-stranded oligonucleotide. However, c-Jun does not directly bind IL-1ß proximal promoter sequences either by gel shift assays, an optimized DNA pull-down assay, or by footprinting. Furthermore, the dispensability of direct DNA binding by c-Jun is further supported by the M14 mutation that disrupts TRE binding but does not abrogate functional and physical interaction with PU.1 and C/EBPß. Finally, these five residues within the basic domain of c-Jun are conserved in evolution, suggesting that the duality of DNA and protein-protein interactions is also conserved. Structural studies of basic leucine zipper proteins in solution or in crystals suggest that the basic domain of these proteins is disorganized when they are free in solution while it adopts an -helical structure when bound to DNA, probably via an induced fit (19, 58, 72). We hypothesize that the c-Jun basic domain, including the DNA binding interface, can adopt in the context of homodimers a non--helical conformation which is stabilized upon binding to C/EBPß and PU.1, thereby simultaneously preventing specific interactions with DNA.
c-Jun can mediate transcriptional stress response via two different mechanisms. c-Jun has been implicated in macrophage differentiation (74) and stress response (4, 15, 35, 69). Our study unravels two different mechanisms through which c-Jun can mediate transcriptional activation of stress-responsive genes. c-Jun commonly drives gene expression via binding to an AP-1 site, and, in this context, heterodimers are favored over homodimers. This is exemplified from studies with the Mmp12 promoter (57) where TPA induction allows (i) recruitment of c-Jun/c-Fos heterodimers to an AP-1 site, (ii) displacement of the NcoR corepressor through a mechanism that requires c-Jun phosphorylation at Ser63/Ser73, and (iii) as shown here, a sharp increase in histone H3 acetylation. In contrast to the Mmp12 promoter, induction from the IL-1ß promoter (i) occurs in the absence of an AP-1 site, (ii) is mediated by c-Jun homodimers acting as coactivators that interact with promoter-bound PU.1 and C/EBPß, (iii) is independent of c-Fos and c-Jun Ser63/Ser73 phosphorylation, and (iv) is associated with a modest increase in histone H3 acetylation. It is therefore possible that the promoter architecture and transcriptional ground state of c-Jun target genes will select between c-Jun/c-Fos and c-Jun/c-Jun dimer usage. A repressed state prior to activation, as exemplified with the Mmp12 promoter, requiring derepression would implicate heterodimer binding to an AP-1 site, allowing for displacement of a corepressor, possibly triggered by the phosphorylation of Ser63/Ser73 (57). In contrast, an active basal state as exemplified by the IL-1ß promoter would implicate homodimers in the absence of DNA binding (Fig. 10) and does not require Ser63/Ser73 phosphorylation. In summary, our observations are consistent with the view that c-Jun heterodimers shown elsewhere (3, 65) to bind AP-1 sites with highest affinity may function as classical transcription factors whereas c-Jun homodimers may preferentially act as coactivators. Our data indicate that these two activities are mutually exclusive. Although we have not addressed what governs this switch, we propose that these two activities are determined by the architecture of the promoter or the ratio of heterodimers to homodimers and that this duality could represent a novel mechanism of fine-tuning gene expression profiles.
|
