Interferon Regulatory Factor-2 Regulates Cell Growth through Its Acetylation*

We have previously shown that interferon regulatory factor-2 (IRF-2) is acetylated by p300 and PCAF in vivo and in vitro . In this study we identified, by mass spectrometry, two lysine residues in the DNA binding domain (DBD), Lys-75 and Lys-78, to be the major acetylation sites in IRF-2. Although acetylation of IRF-2 did not alter DNA binding activity in vitro , mutation of Lys-75 diminished the IRF-2-dependent activation of histone H4 promoter activity. Acetylation of IRF-2 and IRF-2-stimulated H4 promoter activity were inhibited by the adenovirus E1A, indicating the involvement of p300/ CBP. Mutation of Lys-78, a residue conserved throughout the IRF family members, led to the abrogation of DNA binding activity independently of acetylation. H4 is transcribed only in rapidly growing cells and its promoter activity is dependent on cell growth. Consistent with a role for acetylated IRF-2 in cell growth control, IRF-2 was acetylated only in growing NIH 3T3 cells, but not in growth-arrested counterparts. Chromatin immunoprecipitation assays showed that IRF-2 interacted with p300 and bound to the endogenous H4 promoter only in growing antibodies and PBS-T. The antigen-antibody interaction was visualized by incubation in a chemiluminescent reagent (Amersham Biosciences) and exposure to x-ray incubated with at acetyltransferase reaction reaction mixture incubated with 10 units of thrombin (Amersham Biosciences) for 16 h to cleave the GST region and IRF-2DBD purified by reverse-phase high performance liquid chromatography (phenyl-5WRP). Purified GST-IRF-2 ( (cid:1) 4 (cid:1) g) was incubated with 100 ng of endopeptidase Lys-C (Roche Diagnostics) in a buffer containing 25 m M Tris-HCl, pH 8.5, 1 m M and 8 M urea for h 37 Acetylated peptides subjected to mass chromatogra- phy ESI-MS). Plasmid Constructs and Transient Transfection— GST-IRF-2DBD (amino acids 1–129) and GST-IRF-2 (cid:2) DBD (amino acids 86–349) were generated by cloning a PCR-amplified fragment into pGEX-4T (Amer- sham Biosciences). To make the IRF-2DBD mutant, a mutated oligonucleotide was produced by PCR and inserted into pGEX-4T (Amer- sham Biosciences) at the Bam HI and Sma I sites. GST-IRF-2DBD mutant proteins were synthesized in Escherichia coli and purified as previously described (39). To make full-length IRF-2 mutants, wild type IRF-2 was point mutated by changing to the oligonucleotide produced by PCR to produce an IRF-2KR

The IRF (interferon regulatory factor) family of transcription factors has been extensively studied in the context of host defense and oncogenesis. Members of this family have extensive homology within their DNA binding domain. By binding to the common target DNA, ISRE, and IRF proteins regulate transcription of interferon and interferon-inducible genes. Some IRF proteins interact with other transcription factors such as TFIIB, coactivators p300 and PCAF (29 -33). IRF-2 has generally been described as a transcriptional repressor, and it is thought to function by competing with the transcriptional activator IRF-1. However, IRF-2 can act as a positive regulator for ISRE-like sequences including the H4 promoter (34), VCAM promoter (35), and gp91 phox promoter (36). Hence, IRF-2 is similar to factors such as YY1, RAP-1, Dorsal, and Kruppel, which exhibit both transcriptional activating and repressing activities (37). Biologically, IRF-2 plays an important role in cell growth regulation, and has been shown to be a potential oncogene, as overexpression of IRF-2 causes anchorage-independent growth in NIH 3T3 cells and tumor formation in mice (38). We previously reported that IRF-2 in 12-O-tetradecanoylphorbol-13-acetate-treated U937 cells is complexed with p300, as well as PCAF, and is itself acetylated (39). Consistent with these results, recombinant IRF-2 was acetylated by p300, and to a lesser degree by PCAF, in vitro (39), providing the first example of acetylation in the family. Since then, additional IRF members, IRF-3 and IRF-7, were shown to be acetylated in vivo and in vitro (40,41). In a previous paper (39), we documented that acetylation of IRF-2 does not alter its DNA binding activity, although acetylation increases DNA binding activity and transcription for some transcription factors. However, acetylation of IRF-2 led to inhibition of histone acetylation by p300, suggesting a possible mechanism for transcriptional repression by IRF-2.
In this paper, we identified two lysine residues located in the IRF-2 DNA binding domain that are acetylated. We show that one residue, Lys-75 is important for the interaction with p300 and the stimulation of histone H4 promoter activity. Histone H4 is transcribed during cell replication, where IRF-2 functions as an activator (42). Providing further evidence of the biological * This work was supported by the Japan Society for Promotion of Sciences and the Ministry of Education, Science, Sports and Culture. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Immunoprecipitation/Western Blotting-Nuclear extracts were prepared by the methods of Dignam with modifications including the addition of a protease inhibitor mixture (Sigma) to buffers. Immunoprecipitation was performed using an anti-IRF-2 antibody as described previously (39). The precipitates or the whole cell lysates containing equal amounts of total proteins were suspended in the SDS sample buffer with boiling, separated on SDS-10% polyacrylamide gels, and transferred onto polyvinylidene difluoride membranes (Millipore) for 1 h. The membrane was blocked with 5% nonfat dry milk in a PBS-T buffer (containing PBS and 0.5% Tween 20) for 1 h, incubated with anti-IRF-2 (Santa Cruz), anti-p300 (Santa Cruz), anti-acetyllysine (New England Biology), anti-glutathione S-transferase (GST) (Santa Cruz Biotechnology Inc.), and anti-FLAG (Sigma), anti-histidine (Qiagen) antibodies for 1 h, and washed in PBS-T. The antigen-antibody interaction was visualized by incubation in a chemiluminescent reagent (Amersham Biosciences) and exposure to x-ray film.
Analysis of Acetylation Site-Ten micrograms of GST-IRF-2 DBD was incubated with 1 g of rp300 or rPCAF at 30°C for 2 h in 60 l of acetyltransferase reaction buffer (39). The reaction mixture was incubated with 10 units of thrombin (Amersham Biosciences) for 16 h to cleave the GST region and IRF-2DBD was purified by reverse-phase high performance liquid chromatography (phenyl-5WRP). Purified GST-IRF-2 (ϳ4 g) was incubated with 100 ng of endopeptidase Lys-C (Roche Diagnostics) in a buffer containing 25 mM Tris-HCl, pH 8.5, 1 mM EDTA, and 8 M urea for 16 h at 37°C. Acetylated peptides were subjected to mass spectrometry (high performance liquid chromatography ESI-MS).
Plasmid Constructs and Transient Transfection-GST-IRF-2DBD (amino acids 1-129) and GST-IRF-2⌬DBD (amino acids 86 -349) were generated by cloning a PCR-amplified fragment into pGEX-4T (Amersham Biosciences). To make the IRF-2DBD mutant, a mutated oligonucleotide was produced by PCR and inserted into pGEX-4T (Amersham Biosciences) at the BamHI and SmaI sites. GST-IRF-2DBD mutant proteins were synthesized in Escherichia coli and purified as previously described (39). To make full-length IRF-2 mutants, wild type IRF-2 was point mutated by changing to the oligonucleotide produced by PCR to produce an IRF-2KR pcDNA3.1 mutant. Adenovirus E1A and E1A⌬N plasmids were kindly provided from Dr. Y. Nakatani (Dana-Farber Cancer Institute).
Synthesis and Purification of Recombinant Proteins-FLAG-tagged human PCAF was purified in Sf9 cells by using a baculovirus expression system and purified as previously described (30). His 6 -tagged human p300 baculovirus were kindly gifted from Dr. Kraus and Dr. Ito (22). In vitro translatable products from plasmids were made using a rabbit reticulocyte TNT assay kit (Promega) as described previously (30).
In Vitro Acetylation Assays-Substrate proteins (GST-IRF-2DBD) (200 ng) were incubated with 100 ng of recombinant p300 or PCAF in 20 l of acetylation buffer containing 50 mM Tris, pH 8.0, 5% glycerol, 0.1 mM EDTA, 50 mM KCl, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 10 mM sodium butyrate with 10 M acetyl-CoA (Sigma) or [ 14 C]acetyl-CoA (Amersham Biosciences). Reaction mixtures were incubated at 30°C for 30 min, and resolved by SDS-PAGE. The reactions were analyzed first by Western blotting with an anti-acetyllysine antibody to evaluate acetylation activity, and the same gel was subse-quently subjected to immunoblot with an anti-GST antibody to verify the amounts of proteins used in each reaction. Radiolabeled reactions were analyzed by autoradiography after staining by Coomassie Brilliant Blue.
FLAG-Pull Down Assay-Cell lysates from 293T transfectants were incubated with M2-agarose (Sigma) in a buffer containing 50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 5 mM EDTA, 0.1% Nonidet P-40, 0.1 mM phenylmethylsulfonyl fluoride, and 0.1 mg/ml aprotinin and washed three times. Bound materials were eluted in an SDS sample buffer, resolved on an SDS-12.5% PAGE, detected by Western blotting.
Affinity DNA Binding Assay and Electromobility Shift Assay-The DNA affinity binding assay was performed as described (30,39). Briefly, nuclear extracts (500 g of protein) were incubated with magnetic beads conjugated to biotinylated oligonucleotide from the H4 gene (34). Bound materials were immunoblotted with anti-IRF-2 and antiacetyllysine antibodies. Electrophoretic mobility shift assays were performed using a 32 P-labeled H4 promoter oligonucleotide (34) made by T4 kinase (Promaga). In vitro translated products were incubated with a reaction buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 5% glycerol, 1 g of poly(dI-dC)⅐poly(dI-dC) and 1 g of salmon sperm DNA) in the presence of a labeled probe. Protein-DNA complexes were separated from the free probe by gel electrophoresis on 5% polyacrylamide gels in a 0.5 ϫ TBE buffer. The gel was dried and analyzed by autoradiography.

Lys-75 and Lys-78 of IRF-2 Are Major
Sites of Acetylation-We previously reported that the DBD of recombinant IRF-2 is acetylated by p300 and PCAF in vitro (39). Here, we examined whether the DBD and C-terminal domain were acetylated to a similar degree. Bacterially expressed GST-IRF-2DBD and GST-IRF2 ⌬DBD containing only the C-terminal region were incubated with recombinant PCAF or p300 in the presence of acetyl-CoA, and acetylated peptides were detected by an anti-acetyllysine antibody. As shown in Fig. 1A, p300 strongly acetylated DBD. Although less strongly than p300, PCAF also acetylated DBD as expected (39). In contrast, ⌬DBD did not show detectable acetylation under these conditions. The control GST also did not show acetylation either. Predominant acetylation of DBD was also observed when acetylation was assessed by 14 C-acetyl-CoA uptake (data not shown).
We next sought to determine individual lysine residues within the IRF-2 DBD that are acetylated by p300 and PCAF. To this end, the GST-DBD peptides were acetylated by p300 and PCAF in vitro, digested by endopeptidase Lys-C, and the resultant peptides were subjected to mass spectrometry analysis to identify acetylated peptides. We found that peptide 1 and peptide 2 in Fig. 1B were clearly modified by acetylation among other peptides. Peptide 1 covers amino acids 26 -32, containing three lysines, Lys-29, Lys-31, Lys-32, whereas peptide 2 encompasses amino acids 65-78 containing Lys-71, Lys-75, and Lys-78. To further assess major acetylation sites within these regions, mutant DBDs were constructed in which each of the Lys residues was changed to arginine. These mutants were acetylated in vitro by p300 and PCAF, and the levels of acetylation were assessed by immunoblot with an anti-acetyllysine antibody. Wild type IRF-2DBD was intensely acetylated both by p300 and PCAF (Fig. 1C). Mutants K29R, K31R, K32R, and K71R were also acetylated, although slightly less than the wild type DBD. However, levels of acetylation were markedly reduced for K75R, observed both by PCAF and p300. In addition, acetylation by PCAF was significantly reduced for K78R, although the reduction was less significant when tested with p300 (Fig. 1C). These results indicate that Lys-75 is the major site of acetylation by p300 and PCAF, whereas Lys-78 is also acetylated by PCAF, albeit to a lesser degree. The lysine residue at 75 is shared between IRF-1 and IRF-2, but not in other IRF family members. Lys-78 is a residue conserved in all IRF family proteins, and is critical for ISRE binding activity (41). Thus, although mass spectrometry found other Lys residues (Lys-29, Lys-31, and Lys-32) to be acetylated, the overall acetylation level of these residues appear much less extensive.
Having identified Lys-75 and Lys-78 as the main sites of IRF-2 acetylation in vitro, it was important to determine whether these Lys residues were acetylated in vivo. FLAG-tagged wild type IRF-2 as well as FLAG-tagged, full-length IRF-2 mutants K75R and K78R were transfected into 293T cells. Extracts were precipitated by an anti-FLAG antibody and analyzed by immunoblot with an anti-acetyllysine antibody. As shown in Fig. 1D, the anti-acetyllysine antibody gave much stronger signals for wild type IRF-2 than the K75R and K78R point mutants and K75R,K78R double mutants. Acetylation of K75R,K78R double mutants was decreased further compared with each point mutant. Immunoblot with the anti-FLAG antibody verified that the four IRF-2 were expressed at comparable levels. These results verify that Lys-75 and Lys-78 are acetylated in vivo.
The Acetylation Mutant K75R Fails to Stimulate H4 Promoter Activity-We previously showed that acetylation of IRF-2 does not alter DNA binding activity (39). Here, we tested whether the substitution of Lys to Arg in amino acid positions 75 or 78 affected DNA binding activity. In vitro transcribedtranslated mutants K75R and K78R were tested in gel mobility shift assays using the ISRE of the histone H4 promoter as a probe (Fig. 2A). The K75R mutant bound to the ISRE as well as the wild type IRF-2. In contrast, the K78R mutant failed to bind to the DNA. Lysine at 78 is conserved throughout the IRF family and is directly involved in DNA binding (43), and the mutation of this residue likely impaired a structure/conformation required for DNA binding, because mutation of the homologous residue in several IRF family proteins abolishes DNA binding activity (41,44).   1, 4, and 7), rp300 (100 ng) (lanes 2, 5, and 8), and rPCAF (100 ng) (lanes 3, 6, and 9) as described under ''Materials and Methods.'' The reaction mixture was electrophoresed in an SDS-5-20% PAGE and immunoblotted using anti-acetyllysine and anti-GST antibodies. B, amino acids sequence of the IRF-2DBD. The IRF-2DBD peptide was cleaved by endopeptidase Lys-C and analyzed by mass spectrometry. Acetylated modification was observed in peptide 1 and peptide 2 and six-point mutants were constructed by changing lysine residues to arginine residues. Numbers indicate point-mutated amino acid numbers of IRF-2. C, PCAF and p300 highly acetylate lysine residues Lys-75 and Lys-78 in the IRF-2 DNA binding domain. Two  (34). To examine whether the mutation of Lys-75 or Lys-78 influences the ability of IRF-2 to activate H4 promoter activity, transfection assays were performed using vectors for K75R and K78R mutants and a H4luciferase reporter. Results obtained with NIH 3T3 cells are shown in Fig. 2B. As expected, transfection of wild type IRF-2 enhanced the H4 promoter activity in a dose-dependent manner. However, the K75R mutant only modestly enhanced promoter activity, reaching ϳ50% of the enhancement seen by wild type IRF-2. The K78R mutant that did not bind to the target DNA, did not enhance the H4 promoter. Thus, mutation of the acetylatable residue at Lys-75 diminishes activator function of IRF-2, supporting the role for acetylation of Lys-75 in the transcriptional activity of IRF-2.
Assuming that full activation of H4 promoter activity depends on acetylation of IRF-2, we wished to evaluate the contribution of p300/CBP or PCAF to H4 promoter activity. Because levels of PCAF expression are very low in NIH 3T3 cells, 2 p300/CBP is the likely acetylase affecting IRF-2 in these cells. It has been shown that p300/CBP binds to the adenovirus E1A, and that this binding inhibits p300-dependent transcription activation (45). However, the N-terminal mutant E1A (E1A⌬N) fails to bind to p300, and does not inhibit transcription (24 -28, 46). In experiments shown in Fig. 3, H4 promoter activity was examined following cotransfection of IRF-2 and E1A along with the H4-luciferase reporter. Cotransfection of wild type E1A strongly inhibited the H4 promoter activated by IRF-2. On the other hand, E1A⌬N did not inhibit the H4 promoter activation, suggesting that endogenous p300 takes part in IRF-2-dependent activation of H4 promoter activity.
To ascertain whether IRF-2 acetylation is inhibited by E1A, immunoblot analysis was performed for FLAG-tagged IRF-2 cotransfected with wild type E1A or E1A⌬N in 293T cells. As shown in Fig. 3B, coexpression of E1A, but not of the mutant E1A⌬N, strongly reduced acetylation of IRF-2 in vivo. These results suggest that p300/CBP acetylates IRF-2, thereby enhancing its transcriptional activity.
To examine whether E1A affects the modest promoter activity of the IRF-2 mutant, E1A was cotransfected with IRF-2K75R mutant into NIH 3T3 cells. E1A transfection inhibited the modest activation of H4 promoter activity by mutant IRF-2 K75R as well as that of wild type IRF-2 (Fig. 3C).
IRF-2 Is Acetylated in Growing, but Not in Growth-arrested NIH 3T3 Cells-IRF-2 has an oncogenic property, as evidenced by anchorage-dependent cell growth when overexpressed in NIH 3T3 cells (38).
To assess whether acetylation of IRF-2 is relevant to cell growth control, we examined the status of IRF-2 acetylation in exponentially proliferating, or confluent, growth-arrested NIH 3T3 cells. Endogenous IRF-2 was precipitated by an anti-IRF2 antibody and tested for acetylation by immunoblot. As seen in Fig. 4, IRF-2 was expressed at a comparable level in growing and growth-arrested cells. However, acetylation of IRF-2 was seen only in growing cells, but not in growth-arrested cells.
These results indicate that acetylation of IRF-2 depends on the status of cell growth.
These results were interesting, because histone H4 is a cell cycle-regulated gene, expressed during the S phase, and is not expressed in growth-inhibited cells (42). Thus, we sought to determine whether acetylated IRF-2 binds to the H4 promoter in a cell growth-dependent manner. To this end, we first performed a DNA affinity binding assay with biotinylated H4 promoter oligonucleotides that had been conjugated to magnetic beads. Nuclear extracts from growing and growth-arrested cells were incubated with the bead-conjugated H4 promoter DNA, and bound IRF-2 was detected by immunoblot. Results are shown in Fig. 4B. Binding of IRF-2 to the H4 promoter was detected only with extracts from growing cells, but not from growth-arrested cells, despite that levels of total IRF-2 in the extracts being comparable, suggesting that only acetylated IRF-2 binds to the H4 promoter. We found that in addition to IRF-2, H4 promoter DNA bound to p300, again only 2 A. Masumi, unpublished data. from growing cells, but not growth-arrested cells. Like IRF-2, the total levels of p300 were similar between growing and arrested cells.
Finally, it was of importance to verify that IRF-2 binds to the promoter of the endogenous H4 gene in a cell growth-depend-ent manner. Thus, we performed chromatin immunoprecipitation assays for growing and growth-arrested NIH 3T3 cells. Cells were first treated with formaldehyde, and chromatin was sonicated to produce ϳ500-bp DNA fragments. Materials were precipitated by an anti-IRF-2 antibody, and precipitated DNA was subjected to PCR to amplify a ϳ150-bp H4 promoter fragment. As shown in Fig. 4D, the IRF-2 antibody precipitated the H4 promoter DNA from growing cells. However, the same antibody failed to precipitate DNA from growth-arrested cells. The anti-IRF-2 antibody did not precipitate glyceraldehyde-3phosphate dehydrogenase, run as a control. These results confirm that IRF-2 binds to the H4 promoter in growing NIH 3T3 cells, but not in their growth-arrested counterparts, supporting the role of IRF-2 acetylation in regulating H4 transcription and cell growth.  lanes 2 and 4) and growth-arrested (conf) (lanes 1 and 3) NIH 3T3 cells and were immunoprecipitated with control rabbit IgG (lanes 1 and 2) or an anti-IRF-2 antibody (lanes 3 and 4). Immunoprecipitates were immunoblotted with anti-acetyllysine antibody and then an anti-IRF-2 antibody after stripping with a stripping buffer (50 mM Tris-HCl, pH 7.8, 2% SDS, and 5% mercaptoethanol). B, magnetic beads conjugated to H4 promoter were incubated with nuclear extracts from confluent (lane 1) and growing (lane 2) NIH 3T3 cells. Bound materials were analyzed by immunoblot assay using anti-IRF-2, anti-acetyllysine, and anti-p300 antibodies. C, Western blot analysis of nuclear extracts from confluent and growing NIH 3T3 cells. Nuclear extracts from confluent (lane 1) and growing (lane 2) NIH 3T3 cells were electrophoresed and immunoblotted with anti-p300 antibody and anti-IRF-2 antibody. D, chromatin immunoprecipitation assay of the H4 promoter. Exponential growing (grow) and growth-arrested (conf) NIH 3T3 cells were cross-linked with 1% formaldehyde, chromatin was isolated as described under ''Materials and Methods'' and a chromatin immunoprecipitation assay of the H4 promoter and glyceraldehyde-3-phosphate dehydrogenase using an anti-IRF-2 antibody and IgG was performed. PCR quantitation was carried out as indicated under ''Materials and Methods.'' DISCUSSION A wide range of DNA binding transcription factors has been shown to be acetylated (6,10,14,17,18,21,26,39,41). Acetylation of transcription factors often affects DNA binding and transcriptional activity, thereby adding increased versatility to their regulatory functions (47). In the present paper we studied IRF-2 acetylation in detail in vivo and in vitro. We showed that IRF-2 is acetylated predominantly within the DBD, and that two residues, Lys-75 and Lys-78, within the DBD are major sites of acetylation. Lys-78 is a residue located within a region important for ISRE binding activity and is conserved throughout the IRF family (44). Interestingly, the equivalent residue in IRF-7 was shown to be acetylated in vivo and in vitro (41). Mutation of this residue in both IRF-2 and IRF-7 completely abolished DNA binding activity (Fig. 2). However, DNA binding activity is apparently not dependent on the acetylation of this residue, because an IRF-7 mutant in which this residue cannot be acetylated retains DNA binding activity (41). Further supporting this idea, the equivalent residue is present in ICSBP, another IRF member. Whereas ICSBP is not acetylated by p300/CBP or PCAF, a mutation of this residue eliminates ISRE binding (39,44). Similarly, acetylation of Lys-75 is not likely to affect ISRE binding activity under in vitro conditions, because acetylation of IRF-2 does not alter DNA binding activity in electrophoretic mobility shift assays (39).
Our results indicate that acetylation of Lys-75 is required for transcriptional activity of IRF-2, because a mutation of Lys-75 substantially diminished H4 promoter activity, coinciding with reduced levels of IRF-2 acetylation in vivo (Fig. 2). K75R mutant seems to be acetylated at the Lys-78 site; data in Fig. 1 shows that this residue is acetylated both in vivo and in vitro. The K78R mutant no longer binds to the ISRE, indicating that this residue is important for DNA binding as well as acetylation. Based on the in vitro binding assays and a report by Caillaud (41), DNA binding activity may not depend on the acetylation of this residue. In the absence of DNA binding activity by the K78R mutation, the significance of acetylation could be studied only with the K75R mutant. We infer that p300/CBP, rather than PCAF, is the main acetylase for IRF-2 in NIH 3T3 cells, because p300 is expressed at a high level in NIH 3T3 cells, whereas PCAF expression is very low (not shown). Because Lys-75 is acetylatable by both p300 and PCAF in vitro, PCAF may acetylate IRF-2 at this residue in other cell types. In agreement with the role of p300/CBP, E1A that competes with cellular factors for p300/CBP binding inhibited H4 promoter activity. H4 promoter activity was not inhibited by E1A⌬N, which does not bind to p300/CBP. Correlating with the E1A inhibition of promoter activity, IRF-2 acetylation in vivo was inhibited by E1A. These data indicate that p300/CBP interacts with IRF-2 and acetylates Lys-75 in NIH 3T3 cells, and that this event, susceptible to disruption by E1A, is critical for IRF-2 activation of the H4 promoter. Recently, Paulson et al. (48) demonstrated the ability of E1A to bind to GCN5 and inhibit its activity in interferon-stimulated transcription system. We cannot exclude the involvement of GCN5, which has histone acetyltransferase activity and binds to the N-terminal region of E1A similar to p300, in an IRF-2-mediated transactivation system. However, p300 acetylates IRF-2 much higher than that of PCAF in vitro, and a considerable amount of p300 associates with IRF-2 in NIH 3T3 cells. These finding suggest that at least p300 acetylase is required for IRF-2 transcriptional function in living cells, although acetylase such as GCN5 other than p300/CBP has to be examined in our system in the future. In line with these results, E1A has been reported to interfere with the acetylation of DNA binding factors, and hence transcription in several systems. Puri et al. (49) showed that MyoD forms a multimeric complex with PCAF and p300/ CBP leading to the acetylation of MyoD, which is critical for promotion of myogenic transcription. However, E1A inhibits this process by disrupting MyoD acetylation and the formation of the complex. In addition, Hung et al. (50) showed that CBP stimulates acetylation of GATA-1 in vivo in an E1A-sensitive manner, which is linked to transcriptional activity of GATA-1. Although acetylation is not required for GATA-1 to bind to DNA, the mutation that impairs acetylation reduces the ability of GATA-1 to induce cellular differentiation (50).
IRF-2 has a dual activity and can act as a repressor for many interferon-inducible genes (51). In this context, we previously reported that acetylation of IRF-2 reduces p300-mediated acetylation of core histones, suggesting that acetylated IRF-2 can negatively affect transcription in some cases (39). Whether acetylated IRF-2 can act as an activator or repressor may be dependent on specific promoters.
We have extended our study to address the role of acetylated IRF-2 in cell growth, and showed that IRF-2 is acetylated only in growing cells, but not growth-arrested cells. By DNA affinity binding, as well as a chromatin immunoprecipitation assay, we found that IRF-2 binds to the H4 promoter only in growing cells, and that H4-bound IRF-2 was acetylated. That IRF-2 was expressed in growth-arrested cells at levels comparable with rapidly growing cells, but was not acetylated or bound to the H4 promoter in growth-arrested cells, suggests that only acetylated IRF-2 binds to the H4 promoter and activates H4 promoter activity. Our data indicate that acetylated IRF-2 has an increased affinity for the H4 promoter in vivo, although acetylation did not alter the DNA binding activity of recombinant IRF-2 in electrophoretic mobility shift assays (39). Different from a recombinant protein-DNA complex, endogenous IRF-2 is likely to bind to the H4 promoter by associating with multiple factors. Acetylated IRF-2 could form a complex distinct from that of non-acetylated IRF-2, and bind to the promoter more efficiently than non-acetylated IRF-2 in vivo. Thus, the effect of acetylation may not have been observed with electrophoretic mobility shift assays where DNA binding is detected without complex formation. In this context, Barlev et al. (52) showed that acetylated p53 binds more tightly to the transcriptional cofactors TRRAP and CBP than nonacetylated p53, although acetylated and nonacetylated p53 bind to the p21 promoter in the same manner. Interestingly, Suhara et al. (40) reported that IRF-3 was acetylated when incorporated into the large holocomplex that includes p300, and that a p300 lacking histone acetylase activity failed to confer IRF-3 DNA binding activity, indicating that acetylation of IRF-3 is important for the DNA binding activity of the IRF-3 holocomplex. These results support the notion that acetylation influences the DNA binding activity of IRF family members, albeit under varying conditions.
Our results indicate that IRF-2 is acetylated by p300/CBP in a cell growth-regulated manner. The temporal control of IRF-2 acetylation and consequently transcriptional activity may provide a mechanistic basis for cell cycle-dependent transcription of the H4 gene (34). It has been shown that although p300/CBP is expressed throughout the cell cycle, levels of phosphorylation and coactivator function change during the cell cycle (53). Importantly, CBP is phosphorylated by cyclin E-cdk2 at the G 1 /S phase, increasing the histone acetylase activity during this phase of the cell cycle (54). p300/CBP is also phosphorylated by another kinase, causing a different functional outcome (55). Thus, acetylation of IRF-2 may be attributed to cell cycle-dependent phosphorylation/acetylase activity of p300/CBP. In addition, some nuclear cofactors may be preferentially recruited to acetylated IRF-2, rather than unacetylated IRF-2, resulting in more efficient transcription. IRF-2 may interact with other proteins, but not with p300 in confluent NIH 3T3 cells and it cannot bind to the H4 promoter efficiently. In contrast, in growing NIH 3T3 cells IRF-2 associates with p300 and is acetylated. IRF-2 binds to the H4 promoter by recruiting some factors that access cell growth. Adenovirus E1A inhibits acetylation of IRF-2 and IRF-2-induced H4 gene activation by dissociation of IRF-2 and p300. In growth-arrested cells, IRF-2 may act only for those promoters where it can bind in an unacetylated state.
In summary, we identified Lys-75 as a major site of IRF-2 acetylation. The acetylation of this residue is required for full activation of histone H4 promoter. Furthermore, being acetylated in a cell growth-dependent manner by p300, IRF-2 regulates cell growth-dependent transcription of the H4 gene.