| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Rheumatic and Autoimmune Diseases Division, and Center for Immunology, University of Minnesota Medical School, Minneapolis, MN 55455
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
AP-1 is a heterodimer containing a Fos and Jun protein that can both regulate the transcription of multiple lymphokine, chemokine, and cell surface receptor genes, as well as directly facilitate entry of T cells into cell cycle (8, 9, 10). Interestingly, the function of c-Jun can be enhanced by CD28 activation via JNK-dependent phosphorylation (11). c-Fos protein levels also vary greatly during the activation of T cells, depending on the strength of CD28 costimulatory signaling (4). Therefore, an increase in Fos gene expression may mediate at least part of the ability of CD28 costimulation to enhance AP-1-dependent transactivation. Nevertheless, the molecular mechanism involved in the control of Fos gene transcription in CD4+ T cells by CD28 remains only poorly understood.
Fos is considered an immediate-early gene, based on its very rapid induction of transcription in response to growth factors, physical stress, or other extracellular stimuli (and before new protein translation) (12). Transcription is regulated by several cis-acting DNA enhancer elements upstream of the transcriptional start site: a Sis-inducible element, a serum response element (SRE),5 and a cyclic AMP-responsive element (CRE) (13, 14). Using cloned CD4+ Th cells, we previously demonstrated that during stimulation with Ag, interactions between the B7 molecules CD80 and CD86 on the APC and the CD28 receptor are necessary for optimal transactivation by Elk-1, a NF that binds constitutively at the SRE (4). In this case, JNK appeared to mediate the costimulatory effects of CD28 on Elk-1 function. Hsueh et al. (16) have also shown that CD28 costimulation can promote transactivation by CREB protein, a factor that can be expected to bind constitutively at the Fos gene CRE in T cells (15, 16, 17). This augmentation of CREB function results from its more efficient binding to the nuclear coactivator CREB-binding protein (CBP) (18). Interestingly, this capacity of CD28 to augment CREB transactivation was mediated by p38 and calcium/calmodulin-dependent kinase IV protein kinase activities.
CBP and the related p300 protein are both large nuclear factors that share a capacity to bind to a group of DNA-specific transactivators, including CREB, Elk-1, NFAT, NF-
B, and AP-1 (19, 20, 21, 22, 23, 24). The p300/CBP proteins are essential to cellular function and appear to be limiting, as compound heterozygotes for Ep300 and Crebbp gene deletions are invariably associated with embryonic death (25). Using a cross-linked chromatin immunoprecipitation (X-ChIP) assay, p300 molecules have been observed to accumulate at the Fos promoter as well as other select genes during the stimulation of Jurkat T cells with ionomycin plus PMA (26). In general, these coactivators influence gene expression through a combination of acetyltransferase activities for all four histone proteins (histone acetyltransferase (HAT) activity) as well as other nearby nuclear factors (27). The hyperacetylation of histones can play a major role in the accessibility of nucleosomal DNA to the transcriptional apparatus, including RNA polymerase (Pol) II. In addition to acetyltransferase activities, p300/CBP proteins are also thought to promote transcription through their ability to bind multiple nuclear factors simultaneously and act as scaffolds that facilitate protein/protein interactions at the promoter with the Pol II holoenzyme (28). In this study, we investigated the hypothesis that increased p300/CBP activity at the Fos gene mediates the effects of CD28 costimulation on transcription. Our results not only demonstrate improved recruitment of p300/CBP to the Fos promoter, but also increased coactivator function by its C-terminal transactivation domain when CD28 signals are present. Taken together, the data suggest that p300/CBP molecules mediate the effects of CD28 costimulatory signals to hyperacetylate histone H4 and more effectively recruit and retain RNA Pol II at the Fos promoter.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Animal use has been approved by the University of Minnesota Institutional Animal Use and Care Committee and is in accordance with National Institutes of Health guidelines. The DO11 chicken OVA-specific and I-Ad-restricted murine CD4+ T cells were derived from the lymph node of DO11.10 TCR-transgenic mice (29) and maintained as previously described (4). The human T-leukemia cell line Jurkat (American Type Culture Collection (ATCC)) was also examined in some experiments, for comparison with the normal untransformed CD4+ T cells. APC were syngeneic BALB/c splenic adherent cells preactivated with 2.5–10 µg/ml LPS (Sigma-Aldrich) and were used as previously described (4).
Antibodies
Primary Abs used were anti-c-Fos (sc-253), anti-Elk-1 (sc-355), anti-CBP (sc-1211), anti-p300 (sc-585), anti-phosphorylated Elk-1 (serine 383; sc-8406), anti-RNA Pol II (sc-899), and anti-phosphorylated p38 (threonine 180, tyrosine 182; sc-535) were purchased from Santa Cruz Biotechnology. Anti-phosphorylated ERK1/2 (threonine 202, tyrosine 204; 9101), was obtained from Cell Signaling Technology. Finally, anti-acetyl lysine (06-933), anti-acetylated histone H3 (06-599), anti-acetylated histone H4 (06-866), and anti-p300 mAb (05-267) (a mixture of mAbs, RW109, RW105, and RW128), were obtained from Upstate Biotechnology. This last mAb mixture was raised against a GST fusion to the C terminus of human p300. The clone RW105 also detects CBP.
Stimulation conditions
For mAb stimulations, T cells were incubated with polystyrene bead (Interfacial Dynamics) immobilized hamster IgG anti-murine CD3 mAb 145-2C11 (30) and/or hamster anti-murine CD28 mAb 37.51 (31), as previously described (32). In addition, some T cells were activated with 1.5 µM ionomycin (Iono; Calbiochem) and/or 10–50 ng/ml PMA (Calbiochem) to bypass stimulation of the TCR and CD28. Finally, some T cells were preincubated with trichostatin A (TSA; Sigma-Aldrich) or sodium butyrate (Upstate Biotechnology) for 1 h before stimulation to inhibit histone deacetylases (HDAC). IL-2 production was monitored by ELISA.
Plasmids
The 5xGal4luc reporter gene plasmid called p5GT109luc (containing five tandem repeats of the Gal4 DNA cis-acting element linked to a minimal thymidine kinase promoter and luciferase cDNA), the Il2 luc reporter plasmid called pIL-2luc (containing Il2 gene promoter DNA bp –314 to +40), and expression plasmids pEF-Gal4 (empty vector) and pEF-Gal4-Elk-1 (EF-1
promoter followed by a cDNA encoding a fusion between the Gal4 DNA-binding domain and Elk-1 transactivation domain aa 307–428) were constructed using standard recombinant DNA techniques as previously described (4). A Fos luciferase reporter gene construct called c-fos-81TK-luc and containing two copies of the –357 to –276 promoter sequence (including the SRE but lacking the CRE) linked to a minimal (81 bp) thymidine kinase promoter was a gift from Dr. C. Lange (University of Minnesota, Minneapolis, MN) (33). The E1A and E1A-mutant (deleted for aa 2–36) expression vectors as well as CBP1–450, mouse CBP314–1100, and mouse CBP1191–2199 pGal4/CBP fusion protein expression constructs rely on an SV40 promoter and were provided by Dr. X.-H. Feng (Baylor College of Medicine, Houston, TX) (34). Transient transfection and luciferase reporter gene assays were conducted as previously described (4).
Immunoprecipitation and Western blots
T cells were solubilized in ice-cold lysis buffer containing 20 mM Tris-HCl (pH 7.4), 1% Triton X-100, 10% glycerol, 137 mM NaCl, 2 mM EDTA, 25 mM 
-glycerolphosphate, 1 mM NaVO4, 2 mM Na pyrophosphate, 1 mM PMSF, and 10 µg/ml leupeptin for 30 min. Nuclei were removed by high-speed microcentrifugation in the cold and samples were diluted 2/1 in 3x Laemmli sample buffer before gel loading. In some experiments, cell extracts were subjected to immunoprecipitations before Western blotting. Diluted lysates were mixed with 4 µg of the appropriate immunoprecipitating Ab (anti-Elk-1 or anti-p300/CBP) and then incubated overnight at 4°C by rotation. Immune complexes were collected by incubation with protein A-Sepharose beads for 1 h at 4°C, and then beads were collected, washed three times with wash buffer (20 mM HEPES (pH 7.6), 50 mM NaCl, 2.5 mM MgCl2, 0.1 mM EDTA, and 0.05% Triton X-100). Following separation by SDS-PAGE, proteins were transferred to nitrocellulose and probed with primary Abs. Detection was performed using a goat anti-rabbit Ig-HRP conjugate (Bio-Rad) and ECL chemiluminescence reagents (Amersham Biosciences).
In-gel kinase assay
Ten percent SDS-polyacrylamide gels were poured with 100–200 µg/ml of either the purified recombinant GST-6His-Elk1308–428 transactivation domain (plasmid provided by Dr. R. A. Hipskind, Institut de Genetique Moleculaire de Montpellier, Montpellier, France) or GST alone as a control. CD4+ T cell protein lysates (prepared as described above) were electrophoresed, and then the gel was washed twice for 30 min with 100 ml of 20% isopropanol containing 50 mM HEPES (pH 7.6). After a third 30-min wash in buffer A (20 mM MgCl2 and 20 mM HEPES (pH 7.6)) the gel was incubated in 200 ml of 6 M urea in buffer A for 1 h at room temperature. Subsequently, the gel was subjected to serial washing steps with buffer A containing 0.05% Tween 20 and 3, 1.5, or 0.75 M urea. Finally, the gel was washed several times more with 100 ml of 0.05% Tween 20 in buffer A for 1 h at 4°C. To carry out the kinase reaction, the washed gel was incubated in kinase buffer (20 mM -glycerolphosphate, 20 mM p-nitrophenyl phosphate, 0.1 mM Na3VO4, 2 mM DTT, 50 mM HEPES (pH 7.6), and 5 mM 2-ME) containing 50 µM ATP and 5 µCi/ml 32P-labeled ATP at 37°C for 1 h. To quench the reaction, the gel was washed with 100 ml of 5% trichloroacetic acid plus 1% sodium pyrophosphate, and stained/destained with Coomassie brilliant blue. The gel was then dried and autoradiographed to examine for the presence of phosphorylated GST-Elk-1 protein. Control recombinant GST protein was not phosphorylated under these conditions (data not shown).
Preparation of mRNA and reverse-transcriptase quantitative real-time PCR
RNA was extracted using TRIzol (Invitrogen Life Technologies) and further purified by using the RNA Easy Mini kit (Qiagen). RNA samples were then used as a template for reverse-transcriptase quantitative real-time PCR (qRT-PCR) using a Superscript III Platinum SYBR Green Two-Step qRT-PCR kit (Invitrogen Life Technologies). qRT-PCR was conducted using a Cepheid Smart Thermocycler by adding SYBR Green (Molecular Probes) to the reaction mixture. Primers used for PCR were Fos exon-4 (forward, 5'-TAC TCC GGG CTG CAC TAC TTA CA-3'; reverse, 5'-GCT GCC TTG CCT TCT CTG ACT-3'); and Il2 (Forward, 5'-CCT GAG CAG GAT GGA GAA TTA CA-3'; reverse, 5'-TCC AGA ACA TGC CGC AGA G-3'). Primers spanning exon-7 and–8 regions of the Hprt1 housekeeping gene (forward, 5'-TGA AGA GCT ACT GTA ATG ATC AGT CA-3'; reverse, 5'-AGC AAG CTT GCA ACC TTA ACC A-3') were also used for PCR to normalize the Fos and Il2 mRNA data.
Chromatin immunoprecipitation analysis (X-ChIP)
Two to 8 x 106 T cells were treated with 1% formaldehyde (Fisher Scientific) for 10 min (35 min for anti-CBP or p300 Abs), washed, lysed, and sonicated using a Branson Sonifiers (Branson Ultasonics) or Fisher Scientific Ultrasonics Dismembrator (Fisher Scientific) to produce DNA fragments with an average length of 1,000 bp, and then microcentrifuged at 14,000 rpm in at 4°C for 10 min. Chromatin fragments were diluted 1/10 and then immunoprecipitated using polyclonal Abs and protein A-agarose beads (Pierce). After washing, immune complexes were eluted using an Upstate ChIP protocol (Upstate Biotechnology). Cross-links were reversed in the 500-µl samples by the addition of 10 µg of RNase and NaCl to a final concentration of 300 mM, followed by incubation at 65°C for 5–8.5 h. For X-ChIP/Western blots, the immune complexes were then processed for SDS-PAGE as described above. For the PCR analysis of immunoprecipitated chromatin, the sample DNA was precipitated overnight at –20°C in the presence of 20 µg of glycogen or 4 µg of yeast t-RNA, 100 mM sodium acetate, and two volumes of 100% ethanol, and then resuspended in 100 µl of water. The remaining protein in the DNA solution was digested with proteinase K for 1 h at 37–45°C. DNA was purified using QiaQuick DNA spin columns (Qiagen) according to the manufacturer’s instructions. In some experiments, the recovery of Fos, Il2, and Tnf genomic DNA in immune complexes was analyzed semiquantitatively using conventional PCR and analysis by 1% agarose gel electrophoresis. In other experiments, the X-ChIP DNA was amplified using a qRT-PCR (Cepheid Smart Thermocycler) procedure with SYBR Green in the reaction mixtures. The following primer pairs were used for the analysis of DNA: Fos promoter (forward, 5'-GGC GAG CTG TTC CCG TCA ATC C-3'; reverse, 5'-GCG GGC GCT CTG TCG TCA ACT CTA-3'); Fos exon-4 (forward, 5'-TAC TCC GGG CTG CAC TAC TTA CA-3'; reverse, 5'-GCT GCC TTG CCT TCT CTG ACT-3'); Il2 promoter (forward, 5'-TGT GTC TCC ACC CCA AAG AGG-3'; reverse, 5'-GGG GGT GTC ACG ATG TTT TAC-3'); Tnf promoter (forward, 5'-GGG GAC GAC GGG GAG GAG AT-3'; reverse, 5'-GGG AAG AGG GCG GGG AAA AG-3'). qRT-PCR cycle thresholds for both sample DNA recovered from immune complexes as well as input DNA were compared with plasmid DNA standard curves to determine the template copy number.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Previously, we reported that c-Fos protein induction in T cells is inhibited by the addition of neutralizing anti-CD80 and CD86 mAbs during Ag stimulation (4). This result indicated that B7/CD28 interactions at the immunological synapse induce a novel biochemical signal necessary for optimal c-Fos expression that cannot be mimicked by the TCR or other costimulatory receptors. In those experiments, CD28 ligation was also associated with increased Elk-1-dependent transactivation. Consequently, in this study, we investigated the hypothesis that CD28 costimulatory signals control the intensity of Elk-1-dependent Fos gene transcription. T cells were activated with mAbs directed against CD3 and/or CD28, and c-Fos protein expression was compared with steady-state Fos mRNA. Fos transcripts were found to accumulate to a greater level when a combination of CD3 and CD28 mAbs was present, as compared with either mAb alone (Fig. 1A).This increase in Fos steady-state mRNA during CD28 ligation correlated with both earlier and more intense accumulation of the c-Fos protein (Fig. 1B).
|
These data appeared most compatible with a model in which Elk-1-mediated Fos gene transcription is optimally induced during TCR occupancy only when CD28 costimulatory signals are also generated. Consistent with this, a transiently transfected heterologous luciferase reporter gene that depended on the function of the Elk-1 transactivation domain was again found to be highly dependent on the coligation of CD28 for optimal expression (Fig. 1E) (4). Given that c-Fos is a component of the AP-1 complex that binds to the Il2 gene (7), the observed requirement for CD28 signaling to achieve optimal Il2 gene expression was also supportive (Fig. 1, F–H).
Phosphorylation of Elk-1 serine 383 is not dependent on costimulatory signals
In vitro, CD28 costimulatory signaling is necessary for optimal TCR/CD3-induced JNK and p38 activation in these Ag-experienced CD4+ T cells (4, 32). JNK and p38 are also synergistically activated by the combination of PMA and ionomycin (11). Previously, it was shown that an addition of the stress-activated protein kinase-specific inhibitor SB 202190 at doses sufficient to block both p38 and JNK activation would cause a decrease in Elk-1-dependent transactivation (4). These results, thus, led us to the hypothesis that in CD4+ T cells costimulation-dependent p38 or JNK activity was necessary for optimal phosphorylation and activation of Elk-1 at the Fos gene.
ERK, JNK, and p38 have all been shown capable of phosphorylating Elk-1 using in vitro kinase assay systems (36, 37, 38, 39). An in-gel-kinase assay was used here to identify those costimulation-dependent protein kinases within CD4+ T cells that had a capacity to phosphorylate the transactivation domain of Elk-1. T cells stimulated with either PMA or CD3 mAb alone contained 40- to 44-kDa proteins capable of phosphorylating Elk-1 within the gel matrix following enzyme renaturation, consistent with the mobility and function of ERK molecules (Fig. 2, A and B). The addition of CD28 mAb during the CD3 stimulation, or ionomycin during the PMA incubation did increase the intensity of phosphorylation in the region of these 40- to 44-kDa kinases as well as induce the appearance of an additional 
54-kDa kinase. The electrophoretic mobilities of these CD28-dependent Elk-1 kinases were consistent with the activation of p38 and JNK during the response. No other novel CD28-induced Elk-1 kinases were identified in these activated T cells. Therefore, p38 and JNK appeared to be the sole candidate regulators of Elk-1 phosphorylation and function during the recognition of costimulatory ligands.
|
Role for p300/CBP coactivator proteins in the transcription of CD28-dependent genes
Our inability to demonstrate an important CD28 costimulatory effect at the level of Elk-1 serine 383 phosphorylation raised the possibility that during T cell activation CD28 regulates the function of some other Elk-1-associated coactivator protein at the Fos gene, rather than acting to biochemically modify Elk-1 itself. The p300/CBP family of coactivator proteins has previously been shown to bind constitutively to Elk-1 in vitro (21). Therefore, we hypothesized that during T cell activation p300/CBP molecules cooperate with Elk-1 to induce the transcription of the Fos gene when CD28 costimulatory ligands are available. To assess the likelihood that CD3- and CD28-induced gene transcription depends on one or both of these coactivator proteins, we took advantage of the capacity of ectopically expressed adenoviral E1A protein to bind and inactivate both p300 and CBP (20, 42). Elk-1-mediated transactivation stimulated by either CD3 mAb alone, or by the combination of CD3 and CD28 mAbs, was inhibited in both Jurkat T cells and normal DO11.10 CD4+ T cells by the transfection of wild-type E1A (Fig. 3A and data not shown). This inhibitor effect of E1A relied on its ability to interact with p300/CBP, because a mutant E1A molecule lacking only the p300/CBP interaction site (amino acid residues 2–36) failed to block the response. Inhibition of 5' Fos and Il2 gene enhancer-dependent transcription was similarly achieved by coexpression of the wild-type but not mutant E1A protein in normal DO11.10 CD4+ T cells (Fig. 3, B and C). Therefore, CD4+ T cells depended on the function of p300/CBP coactivator proteins for CD3- plus CD28-induced Fos and Il2 gene activation.
|
|
The X-ChIP detection scheme relies in general on the stabilization of relatively weak protein-protein and protein-DNA interactions at the formaldehyde fixation step, and is facilitated by the fragmentation of chromatin into 1-kbp bits of genomic DNA. Accordingly, this immunoprecipitation assay allowed only for a determination that both phosphorylated Elk-1 and p300/CBP were present together at the same genetic locus 30 min after stimulation. The experiment could not determine whether their association had occurred at the Fos gene SRE. To test this, we purified the DNA fragments recovered along with the p300/CBP immunoprecipitates, and detected the presence of Fos gene promoter sequences (bp –200 to –23) using the PCR. As shown in Fig. 4B, the Fos promoter DNA sequence appeared to be weakly (or only infrequently) bound by p300/CBP molecules in resting T cells. However, stimulation with ionomycin and PMA led to a rapid up-regulation of p300/CBP binding at Fos promoter sequences that was sustained for at least 4 h. In contrast to the Fos gene, only a very modest amount of Il2 and Tnf promoter DNA template could be detected within the p300/CBP immune complexes of resting T cells. These two other promoters did demonstrate significant association with p300/CBP in response to T cell activation, but with kinetics different from Fos. The Tnf promoter quickly developed a strong but transient capacity to bind to p300/CBP, whereas the Il2 gene only recruited p300/CBP after >15 min of stimulation (Fig. 4B). Although the kinetics of p300/CBP binding to the Fos gene reported here for normal CD4+ T cells was somewhat faster than that previously observed using Jurkat T cells, this observation of inducible p300/CBP binding to all three gene promoters was consistent with the findings of Smith et al. (26).
To better evaluate the kinetics and dynamics of p300/CBP recruitment to these gene sequences, qRT-PCR was used to detect the presence of the X-ChIP immunoprecipitated genomic DNA. Using this technique, p300/CBP molecules were five times more likely to be bound to the Fos promoter in resting T cells than to Tnf (Fig. 4C). Binding to the Il2 promoter was found to be intermediate, as compared with the two other genes. The Fos gene binding was specific to the promoter region, because similar p300/CBP binding was never observed at downstream Fos exon 4 sequences, even though the exon could be efficiently amplified from total extracted (input) genomic DNA (Fig. 4, D and E, and data not shown). Within 5–7.5 min of activation with either CD3/CD28 mAb or ionomycin/PMA stimulation, p300/CBP molecules increased their binding to the Fos promoter (Fig. 4, C–E). No such enhancement of p300/CBP binding was observed at the Il2 promoter during this short incubation period; however, Tnf promoter binding was increased in this time frame.
Additional X-ChIP experiments suggested that p300/CBP can be induced to bind Fos promoter sequences in response to PMA treatment alone, consistent with the observed coimmunoprecipitation of CBP with phosphorylated Elk-1 (Fig. 4A and data not shown). Nevertheless, stimulation with the combination of CD3 and CD28 mAbs was necessary to achieve an optimal recruitment of these nuclear coactivators to the Fos (and Tnf) promoter sequences during T cell activation, as compared with CD3 mAb alone (Fig. 4, C and E). Therefore, CD28 signaling appeared to increase the opportunity for this nuclear coactivator to regulate Fos gene transcription during CD3 ligation.
In some systems, histone acetylation and nucleosomal modification have been shown to precede the transcription of immediate early genes such as Fos (43). Such alterations in chromatin structure are thought to increase the accessibility or binding of nuclear factors such as RNA Pol II to DNA sequences during the induction of the transcription response. Therefore, the HAT activities associated with p300/CBP may influence nucleosomal conformation in a way that promotes transcription initiation at the Fos gene (27). Transactivation domains within these coactivators may also facilitate the binding and function of RNA Pol II at the promoter (28). Consequently, we examined RNA Pol II at the Fos promoter. Stimulation with ionomycin and PMA led to increased promoter binding by RNA Pol II soon after p300/CBP molecules had bound the Fos gene (Fig. 4F), Interestingly, T cells stimulated with the combination of CD3 and CD28 mAbs were much more likely to have an RNA Pol II molecule bound to their Fos gene promoter, as compared with cells stimulated through CD3 alone (Fig. 4G). Thus, the recruitment of p300/CBP molecules to the Fos gene in the presence of CD28 costimulation correlated well with an increased loading of the promoter with RNA Pol II.
A C-terminal transactivation domain of CBP responds to CD28 costimulatory signals
In addition to examining the recruitment of p300/CBP proteins to promoter DNA during T cell activation, we also investigated the sensitivity of CBP and its associated downstream effector molecules to CD28 costimulatory signals. Using Jurkat T cells cotransfected with plasmid DNA encoding a fusion protein of the Gal4 DNA-binding domain and CBP cDNA sequences together with a 5xGal4-luciferase report gene (Fig. 5A), we found that the transactivator function of a C-terminal CBP1191–2199 fragment could be strongly induced during T cell stimulation (Fig. 5B). Because the Gal4 DNA-binding domain of this chimeric CBP molecule was designed to associate with the reporter gene constitutively, the increased luciferase accumulation represented a change in the activity of the protein. The C-terminal portion of the protein contains intact HAT and transactivation domains that have inducible coactivator activity in Jurkat T cells (44). However, it lacks a bromodomain thought important for binding to chromatin (45). In contrast, a Gal4-CBP1–450 N-terminal fragment strongly transactivated in the absence of any stimulation (Fig. 5B). In normal CD4+ T cells, transactivation by Gal4-CBP1191–2199 was again enhanced during activation and was blocked in the presence of the wild-type E1A molecule, consistent with an intact E1A binding site (Fig. 5C). Remarkably, transactivation by this domain was significantly increased in the presence of CD28 costimulatory signals (Fig. 5, C and D). Finally, Gal4-CBP1191–2199 dependent transactivation was stronger in the presence of PMA plus ionomycin, as compared with treatment with PMA alone (Fig. 5D). Thus, the induction of CBP1191–2199 transactivation in normal CD4+ T cells was shown to mimic the regulation of Fos gene transcription itself.
|
Our observations of increased Fos gene accessibility to RNA Pol II during CD28 costimulation, and the finding that a HAT-containing C-terminal domain of CBP is sensitive to regulation by CD28, led to the hypothesis that CD28 enhances Fos gene transcription at least in part through increased histone acetylation. In normal CD4+ T cells, a modest fraction of the proteins present at the Fos promoter (but not exon 4) were found acetylated even at rest (Fig. 6). Nevertheless, with ionomycin and PMA activation the Fos promoter-associated proteins demonstrated significantly increased acetylation (Fig. 6A). This general protein hyperacetylation at the promoter was paralleled by the specific hyperacetylation of histone H4 (Fig. 6B). Remarkably, optimal hyperacetylation of Fos promoter-associated histone H4 was found to require stimulation with both ionomycin and PMA, or with a combination of CD3 and CD28 mAbs (Fig. 6, C and D), consistent with the pattern of recruitment of RNA Pol II that had been observed (Fig. 4G and data not shown). This CD28 costimulatory signal-dependent histone H4 hyperacetylation was significantly stronger than that observed with CD3 mAb alone (p = 0.037), and was unique to the Fos promoter, as H4 hyperacetylation was never observed at either the Il2 or Tnf genes (Fig. 6, E and F).
|
Because histone H4 hyperacetylation appeared unnecessary for p300/CBP to carryout its coactivator function at the Il2 promoter, we sought independent evidence that at the Fos promoter a hyperacetylation of histone H4 by p300/CBP could contribute to the increased binding of RNA Pol II and enhanced rate of transcription observed during CD28 costimulation. At some genes (including Fos), histone acetylation is countered by the biochemical actions of associated HDAC complexes (46). Therefore, to further investigate the importance of histone hyperacetylation at the Fos promoter, we made use of the HDAC inhibitors TSA and sodium butyrate. Although the mechanism of action of these agents remains uncertain, data suggest that they interfere with the binding of HDAC molecules at the gene. Our initial experiments using sodium butyrate confirmed a counterregulation of Fos gene expression by HDACs, as PMA-induced c-Fos protein accumulation was enhanced when butyrate was added (data not shown).
Using X-ChIP to monitor the level of histone H4 acetylation at both promoter and exon 4 Fos gene DNA sequences, TSA treatment alone was shown to cause a large increase in H4 acetylation, predominantly at the promoter (Fig. 7A). Despite a TSA-induced hyperacetylation of promoter nucleosomes that approached the level observed during activation with the combination of CD3 and CD28 mAb, this hyperacetylation response alone proved to be a poor stimulus for RNA Pol II recruitment to the promoter and was insufficient to induce any accumulation of Fos mRNA (Fig. 7, B and C). Nevertheless, treatment of T cells with TSA during the course of activation with the CD3 mAb alone led to similarly strong histone H4 hyperacetylation that was now accompanied by an enhanced, albeit suboptimal Fos mRNA accumulation. This result was in contrast to the relatively weak Fos histone H4 acetylation and mRNA production seen during anti-CD3 stimulation in the absence of TSA. Therefore, it did appear that T cell stimulation through the TCR/CD3 complex alone is a suboptimal stimulus for Fos transcription at least in part because of its inability to promote hyperacetylation of nucleosomal histone H4 proteins at the promoter in the face of HDAC counterregulation. Interestingly, CD3-induced RNA Pol II recruitment to the Fos promoter was only modestly increased by the addition of TSA and was still only a fraction of that observed following stimulation of T cells with the combination of CD3 and CD28 mAbs (Fig. 7B). Thus, the ability of CD28 costimulatory signaling to increase the binding of RNA Pol II at the promoter appeared in fact to be largely independent of any role in the hyperacetylation of histone H4. Taken together, these data provided strong support for a model in which Fos gene transcription is controlled by CD28 through the actions of p300/CBP, both at the level of histone H4 hyperacetylation and through an independent effect on RNA Pol II recruitment. In addition, these inhibitor studies revealed a role for HDAC proteins in the counterregulation of CD3-induced Fos responses whenever CD28 costimulatory signals were limiting.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
(PKC), and Akt, resulting in more potent NF-KB-dependent transcription and improved cell survival (50, 51, 52). CD28 also has the capacity to synergize with TCR signals in memory cell phenotype CD4+ T cells for the activation of JNK, leading to the down-stream activation of nuclear effectors such as Elk-1 and the stabilization of Il2 mRNA (4, 11, 53).
We now provide evidence that CD28 costimulatory signaling pathways have the capacity to regulate gene transcription by influencing the location and function of the nuclear coactivator proteins p300 and CBP. At the Fos promoter in intact resting CD4+ T cells, p300/CBP molecules were found to only inefficiently bind to the chromatin. Previous studies have indicated a capacity of the SRE-specific transactivators Elk-1 and serum response factor (SRF) to bind N-terminal transactivation, bromo, acetyltransferase, or C-terminal transactivation domains of p300/CBP, even in the absence of protein phosphorylation (21, 54, 55, 56). However, wild-type Elk-1 (but not mutated Elk-1 protein lacking serines at aa 383 and 389) demonstrates significantly enhanced in vitro binding to the N-terminal p3001514–1922 protein sequence following phosphorylation by ERK2 (56). Therefore, strong ERK-dependent phosphorylation of Elk-1 may be expected to stabilize the interaction between Elk-1 and p300/CBP at the SRE of the Fos promoter. In fact, an increased p300/CBP binding at the Fos promoter was elicited by our treatment of CD4+ T cells with the ERK-activating phorbol ester PMA (data not shown). Nevertheless, an intense ligation of CD3 in the absence of CD28 costimulatory signals failed to cause a reliable increase in the association of p300/CBP with the Fos promoter, despite its ability to activate ERK and phosphorylate Elk-1. This difference may reflect the fact that PMA treatment in T cells is also sufficient to activate PKC (a putative CREB activator), whereas CD3 ligation requires CD28 costimulation to optimally increase the activity of this protein kinase (57). CREB is constitutively bound to Fos DNA sequences and is optimally activated only in the presence of CD28 signals (16). Therefore, CREB likely also contributes to p300/CBP recruitment to the gene during stimulation with the combination of CD3 and CD28 mAbs. Taken together, our results are most consistent with a model in which p300/CBP remains only loosely bound to the Elk-1/SRF complex at the Fos promoter in resting CD4+ T cells and following TCR ligation alone, but achieves a tighter association to the promoter region during TCR ligation in the presence of CD28 costimulatory signals, perhaps as a consequence of the phosphorylation of CREB by p38 (Fig. 8).
|
What then are the direct consequences of strong p300/CBP coactivation in the presence of CD28 costimulatory signals? Soon after the recruitment of fully activated p300/CBP to the Fos promoter, histone H4 hyperacetylation develops at nearby nucleosomes (Fig. 8). This predicts an important role for p300/CBP in the acetylation of H4. Hyperacetylation of Fos nucleosomal histone proteins is thought to induce a conformational change in the chromatin structure that increases the accessibility of the promoter to the transcriptional apparatus (59, 60). The finding of a CD28 dependence for this histone H4 hyperacetylation taken together with the ability of HDAC inhibitors to independently raise the level of H4 acetylation and enhance CD3-induced Fos gene expression predicts that CD28-activated p300/CBP molecules promote nearby histone hyperacetylation, thus facilitating an increased rate of Fos gene transcription. The CD28-responsive C-terminal CBP1191–2199 fragment examined in our reporter gene system contains an intact acetyltransferase domain; therefore, it is conceivable that CD28 signals simply augment the HAT activity of these coactivator proteins. Against this, we have been unable to demonstrate increased HAT enzymatic activity in CBP immunoprecipitates from activated T cells (W. Li, unpublished observation). Furthermore, the reporter gene system used here is an extrachromosomal template not thought to rely on histone hyperacetylation for function (59). We also observed at the endogenous Fos gene that the hyperacetylation of histone H4 with HDAC inhibitors was insufficient to recruit additional RNA Pol II molecules to the Fos promoter. Finally, p300/CBP molecules had the capacity to positively regulate the Il2 gene in the absence of observed histone hyperacetylation. Taken together with previous observations that p300 and CBP protein sequences lacking the HAT domain can still coactivate at 5' Il2 gene NFAT and CD28RE/TPA-responsive element enhancer sequences (22, 23, 61), these data suggest that CD28 costimulatory signals may also act to regulate a functional activity of the C-terminal CBP domain that is independent of histone hyperacetylation but necessary for RNA Pol II recruitment to the Fos promoter.
p300/CBP has been shown to act as a bridging molecule to recruit additional gene regulators to a promoter together with RNA Pol II (62, 63). Both RNA helicase A and p68 RNA helicase have been shown to directly interact with the C terminus of p300/CBP and RNA Pol II, and they can influence the capacity of p300/CBP to transactivate (64, 65). The finding of increased RNA Pol II association with the Fos promoter during activation using a combination of CD3 and CD28 mAbs is consistent with this notion. Thus, our results now indicate that CD28 costimulatory signals, acting in parallel with the TCR to ensure an optimal activation of JNK and p38, lead to both 1) enhanced recruitment of p300/CBP to the Fos promoter, thus allowing for increased histone H4 acetylation and chromatin accessibility, and 2) enhanced capacity of p300/CBP to stabilize the association of RNA Pol II and other transcription factors with the Fos promoter DNA (Fig. 8).
This model is consistent with the previous findings of Alberts et al. (59) that demonstrated a requirement for two signals in the induction of Fos gene transcription in NIH3T3 fibroblasts. In their system, the activation of the small GTPase RhoA by serum was necessary, but not sufficient, for an induction of either the endogenous Fos gene or a stably integrated (chromosomal) Fos SRE-driven reporter gene. In addition to RhoA, the activation of a Rac1 or Cdc42 down-stream signaling cascade leading to JNK appeared to be essential for the activation of the Fos SRE. Remarkably, both histone de-acetylase inhibitors and stress conditions sufficient to induce the activation of JNK demonstrated a capacity to induce a hyperacetylation of histone H4 and to cooperate with RhoA for the induction of Fos. Therefore, their results are indicative of a role for stress-activated protein kinases such as JNK and p38 in causing a hyperacetylation of H4 that can facilitate the induction of transcription.
These data may shed new light on the mechanisms used by downstream CD28 signaling pathways to regulate new gene transcription during the initiation of CD4+ T cell clonal expansion response. Although the work shown here uses the paradigm of Fos gene regulation, the capacity of CD28 to enhance the function of CBP in a heterologous Gal4 reporter gene system predicts that the actions of CD28 on p300/CBP will be felt at other genes that normally recruit these coactivator proteins during the course of T cell activation (24, 26). Even though some of these genes (e.g., Il2, Tnf) may not rely on local changes in histone H4 acetylation for their activation, they still require the recruitment of activated p300/CBP molecules for optimal enhancer function. Therefore, we propose that the control of p300/CBP C-terminal coactivator function by CD28 represents a "master switch" that ultimately shapes the transcriptional response from one that promotes an induction of clonal anergy to one that stimulates a cell cycle progression away from G0 phase and allows for a productive and aggressive clonal expansion response.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by National Institutes of Health Grant R01 GM54706 (to D.L.M.). 
2 S.L.N. and W.L. contributed equally to this work.
3 Current address: Department of Microbiology and Immunology, Indiana University School of Medicine, Indianapolis, IN 46202.
4 Address correspondence and reprint requests to Dr. Daniel L. Mueller, Center for Immunology, University of Minnesota Medical School, Mayo Mail Code 334, 6-120 Nils Hassilmo Hall, 312 Church Street Southeast, Minneapolis, MN 55455. E-mail address: muell002{at}umn.edu
5 Abbreviations used in this paper: SRE, serum response element; CRE, cyclic AMP-responsive element; CBP, CREB-binding protein; X-ChIP, cross-linked chromatin immunoprecipitation; HAT, histone acetyltransferase; Pol, polymerase; TSA, trichostatin A; HDAC, histone deacetylase; PKC, protein kinase C ; qRT-PCR, quantitative real-time PCR; SRF, serum response factor.
Received for publication December 28, 2005. Accepted for publication April 17, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
) or with (
) a combination of CD3 and CD28 mAbs. Proteins were then extracted and assayed for luciferase activity. Error bars represent the SEM for replicate samples.
), Il2 (
