Acetylation of p65 at lysine 314 is important for late NF-κB-dependent gene expression

Background NF-κB regulates the expression of a large number of target genes involved in the immune and inflammatory response, apoptosis, cell proliferation, differentiation and survival. We have earlier reported that p65, a subunit of NF-κB, is acetylated in vitro and in vivo at three different lysines (K310, K314 and K315) by the histone acetyltransferase p300. Results In this study, we describe that site-specific mutation of p65 at lysines 314 and 315 enhances gene expression of a subset of NF-κB target genes including Mmp10 and Mmp13. Increased gene expression was mainly observed three hours after TNFα stimulation. Chromatin immunoprecipitation (ChIP) experiments with an antibody raised against acetylated lysine 314 revealed that chromatin-bound p65 is indeed acetylated at lysine 314. Conclusions Together, our results establish acetylation of K314 as an important regulatory modification of p65 and subsequently of NF-κB-dependent gene expression.


Background
The inducible transcription factor NF-B has an important function in regulating immune and inflammatory responses, apoptosis, cell proliferation and differentiation and tumorigenesis [1][2][3]. NF-B is induced in almost all cell types by different extracellular stimuli causing the activation of an enormous array of target genes [4]. The NF-B transcription factor family comprises NFKB1 (p50/p105), NFKB2 (p52/p100), p65 (RelA), c-Rel and RelB, which form homo-and heterodimers. The most abundant, best-studied and "classical" form of NF-B is a heterodimer consisting of the two subunits p50 and p65. In most unstimulated cells, NF-B is found as inactive transcription factor complex through its physical association with one of the several inhibitors of NF-B (IB) [5]. This family of IB's includes IBα, IBβ, IBγ, IBε (p105/p50, C-terminus), p100/p52 (C-terminus), IB-R and Bcl-3. Virtually all cell types show NF-B responses where the activity of NF-B is specifically regulated at multiple levels [1,3,6]: the level of protein synthesis, the existence of at least 12 different NF-B dimers, the interaction of these dimers with specific IBs and their subcellular localization, post-translational modification of these dimers in the cytoplasm and the nucleus, differential accessibility of B sites in various promoter and enhancer, differential binding to B's response elements due to different affinities, and cell type and stimuli specific interaction with a combination of cofactors.
NF-B is subject to a variety of post-translational modifications (e.g., phosphorylation [7], ubiquitination [8] or prolyl-isomerisation [9]) that modulate its activity. Phosphorylation of the p65 subunit by the PKAc, MSK1 and PKCζ kinases enhances its interaction with the coactivator p300/CBP and stimulates the NF-B transcriptional activity [7,[10][11][12], while dephosphorylation of p65 by the phosphatase WIP1 negatively affected the interaction with p300 [13]. It has recently been shown that p65 and p50 are reversibly acetylated by p300 and PCAF [14][15][16]. Chen et al. identified lysine residues (K) 218, 221 and 310 of p65 as acceptor sites for p300 acetylation. Kiernan et al. identified K122 and 123 in p65 as acetylation sites modified by both p300 and P/CAF. A recent report presented the TGF-β1-mediated acetylation of p65 at K221 in vitro and in vivo enhancing the induced activation of NF-B by bacteria [17]. We recently confirmed acetylation on K310 and provided further evidence for acetylation of p65 on K314 and 315, two novel acetylation sites [18]. Genetic complementation of p65 knockout (-/-) cells with wild type and acetylation-deficient mutants of p65 revealed that neither shuttling, DNA binding nor the induction of anti-apoptotic genes by TNFα was affected by acetylation on these residues. Micro array analysis of these cells treated with TNFα for only 45 minutes identified specific sets of genes differently regulated by wild type or acetylation-deficient mutants of p65 [18]. Specific genes were either stimulated or repressed by the acetylation-deficient mutants when compared to p65 wild type. These results support the hypothesis that site-specific p300-mediated acetylation of p65 regulates the specificity of NF-B dependent gene expression.
Here, we extended the gene expression analysis to three hours after TNFα stimulation and identified genes, which are higher expressed by mutating K314 and 315. ChIP experiments with antibodies directed against acetylated K314 revealed that this lysine is indeed acetylated when p65 is bound to chromatin. Together our results provide evidence that acetylation of K314 is important for the regulation of NF-B-dependent gene expression in vivo.

Mutation of p65 K314/315 regulates TNFa-induced NF-B-dependent gene expression at 3 hours
We provided earlier evidence that acetylation of p65 at K310, 314 and 315 is important for the expression of a defined subset of genes [18]. These earlier studies provided a first glance of the functional relevance of p65 acetylation, since gene expression was measured only after 45 minutes of TNFα stimulation. In order to know if the requirement for site-specific acetylation is maintained for the same genes after longer exposure to TNFα, and to identify possible new genes regulated through p65 acetylation, we decided to extend our analysis to 3 hours of stimulation. For this, we used p65 (-/-) mouse embryonic fibroblasts (MEFs) complemented with acetylation-deficient mutants, where the target lysines for acetylation were mutated to arginines. These cells were described and extensively characterized previously [18]. p65(-/-) cells complemented either with wild type p65, an empty plasmid as control (pTV), the acetylation-deficient double mutant K314/315R or the triple mutant KTR (K310/314/315R), were stimulated by TNFα for 3 hours and total RNA was isolated in three independent replicates from these cells. RNA was amplified, labeled and hybridized to the Agilent Whole Mouse Genome Array. After statistical analysis of the expression profiles, differentially expressed genes were identified ( Fig. 1A-B, and Tables 1 and 2). We focused only on genes that required p65 for their proper induction, which were identified by comparing gene expression profiles from wild type and pTV cells. The majority of differentially expressed genes in KTR mutant were strongly downregulated compared to wild type cells, suggesting that acetylation of p65 at these residues is also an important modification for the expression of the extended NF-B-dependent gene expression (Fig. 1A). In contrast, experiments with the double K314/315R mutant revealed that the majority of genes were slightly upregulated after 3 hours compared to wild type cells (Fig. 1B).
Gene expression of Mmp10 and Mmp13 is enhanced when K314/315 are mutated We subsequently investigated the induction kinetics of several genes by real-time RT-PCR in the complemented cell lines stimulated by TNFα for different time points (between 20 and 360 minutes). Selection of these genes was based on their inducibility by TNFα, as well as their dependency on p65 and their regulation by acetylation of K314/315. Mmp13 and Mmp10 represented genes that were upregulated in the micro array experiments of K314/315R but not affected in KTR expressing cells. In contrast, Cfb and Mpa2l represented genes that were not affected in cells expressing K314/315R, but KTR, suggesting that acetylation of K310 is important for these genes. Overall, the absolute mRNA levels of Mmp10 and Mmp13 were strongly and significantly increased upon time (with a maximum value at 6 and 3 and hours, respectively) in cells expressing the K314/ 315R mutant ( Fig. 2A, B and Additional file 1: Supplemental Table S1), while the expression levels of Cfb and Mpa2l were not affected in the same cells (Fig. 2C, D and Additional file 1: Supplemental Table S1), corroborating our micro array results. Gene expression analysis in cells expressing the KTR mutant indeed confirmed that K310 is important for Cfb and Mpa2l, but counteracts the effect of K314/315R for Mmp10 or Mmp13 expression. Interestingly, the dependency on K310 acetylation for Cfb and Mpa2l was already observed when the basal expression levels were analyzed (0 hour time point).

Characterization of antibodies raised against acetylated lysine 314 and 315 of p65
To further assess the functional relevance of p65 acetylation in vivo, we generated different antibodies raised against the acetylated K314 or K315. All raised antibodies recognized their specific p300-mediated acetylated residues on recombinant p65 acetylated in vitro (Fig. 3A) or on over expressed p65 acetylated in vivo (Fig. 3B).
To confirm that endogenous p65 is indeed acetylated at the indicated lysines, immunoprecipitation experiments with the antibodies raised against acetylated K314, K315 or against p65 were performed. The antibody anti-acK314 immunoprecipitated p65 only in cells complemented with wild type p65 in a TNFα-dependent manner (Fig. 3C). No p65 was immunoprecipitated in  (Fig. 3D), suggesting, that endogenous p65 is indeed acetylated at K314 upon stimulation with TNFα. Unfortunately, anti-acK315 was not specific enough to recognize endogenous acetylation at K315.
Chromatin-associated p65 is acetylated at lysine 314 From the above selected genes only Cfb was already described to contain a B site in its promoter [19]. We therefore searched for putative B sites within the DNA sequence 1 kB upstream of the transcription start site (TSS) of the selected genes. Bioinformatic analysis identified several putative B sites in the promoters of Mmp10 and Cfb, one site in the promoter of Mpa2l, but none for Mmp13 (Fig. 4A). Mmp13 was thus not further investigated. ChIP studies in p65(+/+) MEFs stimulated with TNFα for 20 and 180 minutes revealed that p65 was recruited to the promoters of Cfb and Mpa2l in a stimulus-dependent manner, while no enrichment was observed in p65(-/-) MEFs (Fig. 4B). These recruitments were promoter-specific, since p65 occupancy to promoter of glucagon, a negative control, was not induced upon TNFα stimulation. IP-10, a known NF-B target gene with very well characterized B sites at its promoter, served additionally as a positive control. Unfortunately, no p65 enrichment could be observed to the B sites of Mmp10 (data not shown), indicating that p65 would activate this gene through other B sites or other transcription factors.    Interestingly, ChIP experiments using the anti-acetyl K314 antibodies showed that chromatin-associated p65 is indeed acetylated at K314 on all analyzed genes (Cfb, Mpa2l and IP-10) (Fig. 4C), although the expression of these genes was not affected by mutating K314/315 ( Fig.  2C and 2D). Furthermore, the detected amount of recruited p65, which was acetylated at K314, increased over time (comparing values at 180 min and 20 min). Unfortunately, the antibody raised against acetyl K315 was not able to immunoprecipitate p65 under the tested conditions (data not shown). Together, these experiments identified Mpa2l as a novel NF-B target gene through the recruitment of p65 to its promoter and provide strong evidence that chromatin-bound p65 is indeed acetylated at K314.

Discussion
We have previously shown that p65 acetylation plays an important role in regulating NF-B-dependent transcription of a subset of genes. Our current results confirm those earlier findings by identifying new genes differentially regulated in the acetylation-deficient mutants compared to wild type cells. Remarkably, only a few genes identified after 3 hours TNFα stimulation overlapped with the genes identified in our previous study after 45 minutes TNFα stimulation (data not shown). We also observed that TNFα does not induce NF-B activity in the complemented MEFs as efficiently as in the parental p65(+/+) MEFs used for ChIP (data not shown). Therefore, although our complemented MEFs are useful to address the role of p65 acetylation in vivo, we might have identified only part of the genes influenced by p65 acetylation with the micro array screen.
Cfb was previously suggested to be a direct target of NF-B by EMSA experiments [19]. Our ChIP assays confirmed that p65 is recruited to Cfb promoter in a TNFα-dependent manner in vivo and further identified Mpa2l as a novel NF-B target gene. Although Mmp10 and Mmp13 were differentially expressed in acetylationdeficient mutant cells compared to wild type cells, we failed to detect p65 recruitment to those promoters. One explanation could be that p65 is recruited to the promoter region of these genes with a different kinetics than the one we investigated here. Alternatively, p65 could bind to a regulatory element located far away from the promoter to regulate transcription, as has been Figure 3 Characterization of specific antibodies against p65 acetylated at K314 or 315. (A) Purified recombinant p65 wild type and the acetylation-deficient mutants were incubated with recombinant p300 in the presence (+) or absence (-) of acetyl Co-A. Proteins were resolved on SDS-PAGE and analyzed by western blot using the indicated antibodies. (B) HEK 293T cells were transfected with p65 wild type or mutants, with (+) or without (-) p300 co-transfection. Acetylation of p65 at specific lysines was assessed by western blot using the specific antibodies. * indicates a non-specific band. (C and D) p65 was immunprecipitated from untreated or TNFα-stimulated cells using the indicated antibodies. Immunoprecipitated proteins were separated by 10% SDS-PAGE and subsequently analyzed by western blot using an anti-p65 antibody. * indicates IgG-band.
shown for several NF-B target genes [20][21][22]. NF-B is known to activate the expression of many transcription factors and their regulators [4]; consequently, a third possibility is that p65 directly induces the expression of a protein that regulates the expression of these genes. Along this line, NF-B has been shown to directly activate the expression of the transcription factor Elk-1, which in turn induces Mmp13 gene expression [23].
Interestingly, the stimulatory effect of mutating K314/ 315 was lost when additionally mutating K310, suggesting that K310 acetylation is not affected by the lack of K314/315 acetylation, and that the modification of K310 would counteract the effect of K314/315 at the particular tested genes. Furthermore, the effect of mutating K314/315 on gene expression was more obvious after 3 hours of stimulation compared to early time points (e.g. 20 minutes), suggesting that acetylation of these lysines might more likely affect termination of gene expression than its induction itself. Here, we provide for the first time evidence that endogenous, chromatin-bound p65 is indeed acetylated at K314 upon stimulation with TNFα. Unfortunately, acetylation of K315 could not be confirmed due to the lack of a specific acetylation-dependent antibody that recognizes endogenous proteins. Interestingly, we observed acetylation of p65 K314 at promoters of genes whose expression was not affected by mutating K314/315 (e.g. IP-10). Thus, it could well be that p65 is acetylated upon recruitment of p300 at promoters of many genes, but that the expression of only some of these genes is affected by p65 acetylation at distinct lysines.
The observation that mutation of K314/315 increases expression of certain genes (e.g. Mmp10 and Mmp13) suggests that acetylation of K314 and possibly K315 might represses gene expression. A possible molecular mechanism is that acetylated lysines at p65 create docking sites for bromodomain-containing proteins which need to be recruited to promoters to modulate transcription [24]. On the other hand, we can currently not exclude that lysines 314 and/or 315 are modified by another post-translation modification (e.g. methylation, [25]) and that mutating these residues would thus affect gene expression independently of acetylation. However, provided the observed acetylation of chromatin-bound p65 at K314, acetylation is more likely to compete with another lysine modification, possibly resulting in crosstalk between different modifications [25]. Indeed, a recent study reported that these same two lysines (K314/ 315) can be methylated by Set9 to induce ubiquitinylation of NF-B and subsequently terminate gene expression [26]. Thus, acetylation at K314 and possibly K315 could prevent methylation-mediated repression of target genes and thus avoid methylation-induced termination of p65-dependent gene expression. A direct evidence for methylated p65 at K314/315 bound to chromatin is however still missing. Whether the same genes are regulated by both post-translational modifications should be further addressed. Alternatively, both post-translational modifications might regulate distinct set of NF-Bdependent genes and thus not influence each other.

Conclusions
Together, our results establish acetylation of K314 as an important regulatory modification of p65 and subsequently of NF-B-dependent gene expression.

Tissue culture
Complemented p65(-/-) NIH 3T3 MEFs stably expressing p65 wild type or the acetylation-deficient mutants were generated by lentiviral complementation as previously described in [18]. Briefly, HEK 293T cells were transfected with the packaging plasmid, the envelope plasmid and pTV-myc-RelA/p65 wild type, RelA/p65 K/ R mutants or the control pTV vector. p65(-/-) MEFs were infected with the viral supernatant and split into selective medium (2.5 μg/ml Blasticidin) after thirty-six hours. Expression of recombinant proteins was confirmed by western blot analysis. Pools of cells were used for further analysis. They, as well as the p65(-/-) and (+/+) parental NIH 3T3 MEFs kindly provided by A. Beg [27], were maintained in DMEM supplemented with 10% FCS, 100 units/ml penicillin/streptomycin (Gibco) and non-essential amino acids (Gibco). HEK 293T cells were maintained in DMEM supplemented with 10% FCS and 100 units/ml penicillin/streptomycin.

Plasmids
Plasmids for the mammalian overexpression of human p65 wild type and mutants K310R, K314/315R and KTR in HEK 293T cells were described elsewhere [18]. Briefly, pph-CMV-Km-RelA/p65 wild type, previously described in [28], was used as template vector for the generation of the p65 mutants by site-directed mutagenesis according to the QuickChange protocol (Stratagene). All introduced mutations were confirmed by sequencing.

Generation of recombinant proteins
The recombinant proteins were expressed by baculovirus in Sf21 cells using either the FastBac or the Bac-PAK systems (Clontech). His-tagged proteins were purified over Ni 2+ -beads (ProBond, Invitrogen).

Micro array
Complemented cell lines p65 wild type or acetylationdeficient mutants were starved overnight before either left untreated or stimulated with 30 ng/ml TNFα for 3 hours. Total RNA from three biological replicates per sample was isolated at different days using the 'Total RNA isolation mini kit' (Agilent Technologies). RNA quality was measured on the 2100 Bioanalyzer (Agilent Technologies). Microarray experiments were performed using 'Whole Mouse Genome (4 × 44 K) Oligo Microarray Kit' (Agilent Technologies) and 'One-color micro array-based gene expression analysis' (Agilent Technologies) following the manufacturer's protocol. Cy3-labeled cRNA was purified with the RNeasy kit (Qiagen). Dye incorporation was assessed with the ND-1000 Spectrophotometer (NanoDrop Technologies). Per sample, 1.65 μg cRNA from each of the three biological replicates was hybridized to independent arrays according to the manufacturer's protocol. Hybridized slides were scanned with the Agilent DNA Microarray scanner and quantified using the Agilent Feature Extraction software. The data analysis was performed using 'Rosetta Resolver® Gene Expression Data Management and Analysis System' (Rosetta Biosoftware). Briefly, data was processed and normalized with default settings. Then, lowsignal genes with signal intensities < 0.1 were filtered out. Differential expression between two conditions was assessed based on the average ratio and significance. All genes with expression ratios < 0.556 and >1.8, and a pvalue < 0.05 were selected to generate the tables of significantly regulated genes. These sequence data have been submitted to the GEO database http://www.ncbi. nlm.nih.gov/geo under accession number GSE15196.

Gene expression by quantitative RT-PCR
Complemented MEFs were starved overnight before treatment with 30 ng/ml recombinant TNFα for different time points. Total RNA was isolated from at least two biological samples at different days with the 'Total RNA isolation mini kit' (Agilent Technologies). RNA quantity was assessed with the ND-1000 Spectrophotometer (NanoDrop Technologies). 2 μg RNA was subsequently retro-transcribed using the 'High-capacity cDNA reverse transcription kit' (Applied Biosystems) following the manufacturer's protocol. Real-time PCR was performed using the Rotor-Gene 3000 (Corbett Life Science, now Qiagen) and TaqMan assays from Applied Biosystems for the following genes: Cfb, Mpa2l, Mmp10, Mmp13, Rps6 and CanX, according to Applied Biosystems' protocol. The last two genes were used as internal controls to normalize for RNA input. RNA from at least two biological replicates per sample was measured and analyzed with REST [29]. Each experiment was run three independent times, the mean value and ± SD was calculated and blotted into graphs. A 2-tails, paired t-Test was performed with the log of the values to additionally calculate the p-values.

Acetylation assays in cells
HEK 293T cells were transfected with expression plasmids for p300 and either myc-tagged p65 wild type, the acetylation-deficient mutants or an empty vector, using the calcium phosphate precipitation method. After 23 hours, cells were treated with 10 ng/ml human TNFα for 30 minutes. Then, whole cell extracts were prepared and 40 μg protein was analyzed by SDS-PAGE and western blot.
Additional file 1: Gene expression analysis of Mmp10, Mmp13, Cfb and Mpa2l with quantitative RT-PCR. The mean value of three independent runs relative to wild type unstimulated is shown, as well as ± SD and p-values. P-values were calculated comparing data of each cell line with wild type. Click here for file [ http://www.biomedcentral.com/content/supplementary/1471-2164-11-22-S1.DOC ]