Identification and characterization of a novel p300-mediated p53 acetylation site, Lysine 305

Identification of a novel p300-mediated p53 acetylation site


INTRODUCTION
The p53 tumor suppressor protein is a sequence-specific, DNA-binding transcription factor. In response to a wide variety of stress signals, it acts to regulate processes such as cell cycle, cell death, and DNA repair. Loss of p53 activity has been identified in 60% of the human cancers examined, and over 90% of the p53 missense mutations are clustered within the sequence-specific DNA-binding domain (11,20). The tumor suppressor functions of p53 are directly linked to its ability to mediate transcriptional activation (1).
The regulation of p53 transcriptional activity involves several mechanisms including post-translational modifications such as phosphorylation, acetylation and ubiquitination (2,4 apoptosis (23). Inhibition of p53 deacetylation also increases the p53 half-life, suggesting that acetylation may play important roles in the turnover of p53 protein (4).
Recently, we have employed a strategy based on LC/MS/MS for identifying residues bearing phosphorylation modifications (32). Based on this technique, moderately phosphorylated peptides could be identified by close examination of specific ion chromatograms (32). This selected ion tracing (SIT) method has a major advantage in analyzing the modified peptides of low abundance in a complex peptide mixture. More importantly, we have proposed that this method could be generally employed to identify peptides of various modifications. Here, we illustrate how this approach is adapted to identify acetylated peptides and to map the modified residues. While we successfully verified the previously reported acetylated residues of p53, we also uncovered two novel acetylation targets using this approach. We also present evidence that acetylation of p53 at Lys-305, like at Lys-382, is elevated by different stress signals. In addition, we found that the acetylation at Lys-305 is important for regulating p53 transcriptional activity. 6 clarified lysate on Ni-NTA agarose (Qiagen) according to the manufacturer's instructions and then dialyzed against the storage buffer (20 mM Tris-HCl [pH 8.0], 0.5 mM EDTA, 100 mM KCl, 20% glycerol, 0.5 mM DTT, and 0.5 mM PMSF) before storing at -80℃ in aliquots.

Site-directed mutagenesis
Mutants of FLAG-tagged p53 were constructed by site-directed mutagenesis by a two-step PCR technique as described (6,7). Briefly, mutants were prepared by introducing the desired mutation in overlapping oligonucleotide primers. In the first round of PCR, the mutagenic primer and an outside primer are used to amplify a partial fragment of the region where the mutation will be created. This mutagenized fragment is then purified and used as a "megaprimer," together with the second outside primer, in a second round of PCR to amplify the entire region to be subcloned. Conditions for the first round of PCR were 94℃ for 30 s, 47℃ for 30 s, and 72℃ for 1 min for 30 cycles.
Conditions for the second round of PCR was started with a 2 min denaturation step at 94 ℃, followed by 30 cycles of 94℃ for 30 s, 46.5℃ for 30 s, and 72℃ for 1 min 30 s. This PCR reaction was ended with an elongation step at 72℃ for 10 min.
Plasmid pcDNA3 containing FLAG-tagged p53 was used as the template in the PCR reaction. The outside primers used for all of the mutants in this study were as follows: were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (HyClone) at 37℃ in a 5% CO 2 -humidified atmosphere.
Approximately 1 × 10 6 cells were seeded in each 100-mm culture dish 12-24 h before transfection. Calcium phosphate-mediated DNA transfection was performed as described previously (8). DNA was prepared by the Clontech procedure and adjusted to 12 µg per transfection with pcDNA3 plasmid. Chloramphenicol acetyltransferase (CAT) assays were performed with two-thirds of total cell extract as described earlier (6,31).
The p50-2 that contains two copies of p53 binding motif in its promoter region was used 8 as p53 responsive CAT reporter plasmid (12). A 1 µg aliquot of a RSV-β-Gal vector was included as internal control in all transfections. Quantification of the acetylated [ 14 C] chloramphenicol was determined with a PhosphorImager (Molecular Dynamics). The β-galactosidase assays were performed according to protocols from the manufacturer (Promega). A total of 10 µg of the plasmid was used for each experiment.

Western blot and Dot blot analysis
To verify the antibody specificity, we used a dot-blots analysis as previously described (36). Equal amount of peptides (20 µg) were spotted on strips of Hybond-C membrane (Amersham Biosciences). After blocking by 5% nonfat milk in PBS and 0.1% Tween 20 buffer, membranes were probed with the purified rabbit polyclonal antibodies. Blots were then incubated with a HRP-conjugated goat anti-rabbit IgG 9 (Sigma). The blots were visualized using the enhanced chemiluminescence system (ECL; Amersham Biosciences).

In vitro acetylation assay
Acetylation assays were carried out primarily according to the published method (11,13).

In-gel digestion
The gel piece containing p53 polypeptides was first reduced and pyridylethylated as described previously (32). Up to 0.2 µg of the enzyme, Arg-C (Roche) or Asp-N (Roche), was added to the dried gel. After overnight incubation, the supernatant was removed and the gel was extracted twice with adequate amount of 0.1% formic acid.
The supernatant and the extracts were combined together and dried in a Speed-Vac. The digests were kept at -20°C and suspended in 0.1% formic acid immediately before use.

LC/MS/MS analysis
Electrospray mass spectrometry was performed using a Finnigan Mat LCQ ion trap mass spectrometer interfaced with an ABI 140D HPLC (Perkin-Elmer). A 150 x 0.5 mm PE Brownlee C18 column (Perkin-Elmer) (5 mm particle diameter, 300 Å pore size) with mobile phases of A (0.1% formic acid in water) and B (0.085% formic acid in acetonitrile) were used. The peptides were eluted using the acetonitrile gradient and analyzed by "triple-play" experiment as described (32). The samples were analyzed by two runs of LC/MS/MS experiments. In the first analysis, the most abundant ion in an MS spectrum was selected for collision-induced dissociation (CID) experiment; in the later analysis, only the ions with m/z values corresponding to the potential acetylated peptides were selected for CID experiment. The acquired CID spectra were interpreted using a Finnigan Corporation software package, the SEQUEST Browser, which correlated the MS/MS spectrum with the amino acid sequence of human p53 protein. Enzyme was not specified in the search parameters, which increased the confidence of identification when the matched peptides had appropriate cleavage sites. A 105.14-Da mass tag was constitutively assigned to the Cys residue, which was modified with 4-vinylpyridine in experiments. Those MS/MS scans that matched the peptide sequences with proper cleavage sites were considered significant and also were subjected to manual evaluation to confirm the SEQUEST results.
The hypothetical m/z values of peptides used for generation of ion tracing chromatograms were derived using the PEPSTAT algorithm in the SEQUEST package.
The EXPLORE program (Finnigan) was used to plot the ion tracings.

RESULTS
Determination of p300-mediated in vitro acetylation sites on p53 by LC/MS/MS In our previous report, we established a selected ion tracing (SIT) method to identify phosphorylated peptides (32). Here, we set out to test whether the SIT method could be adapted to identify acetylated peptides. Previous studies have demonstrated that p53 could be acetylated by p300 in vitro (10). Therefore, we analyzed acetylated peptides derived from FLAG-tagged recombinant p53 protein that was acetylated in vitro by p300.
Baculovirus-expressed recombinant FLAG-tagged p53 was immobilized on M2-agrose beads and served as the substrate for the in vitro acetylation reactions. Various amounts of HAT domain from the p300 acetyltransferase (HAT-p300) were incubated with a fixed amount of FLAG-p53 in the presence of [ 14 C] acetyl-CoA (Fig. 1A). After in vitro acetylation, the reaction mixtures were analyzed by SDS-PAGE followed by autoradiographic analysis (Fig. 1B). The FLAG-tagged p53 was acetylated by the enzyme in a dose-dependent fashion. Quantitative determination of the autoradiogram showed that the acetylation was not saturated.
The acetylated p53 resolved in the gel was excised and divided into two portions.
One was treated with Arg-C and the other was digested with Asp-N.
An aliquot of each digested peptide mixture was analyzed by LC/MS/MS. The raw data were then interpreted by the SEQUEST software with optional addition of 44 daltons to the Lys residues. This analysis yielded fifteen Arg-C peptides and six Asp-N peptides (Table 1).
Subsequently, we used the SIT approach for examining all of these peptides and their putative acetylated analogues were chromatographed. We found that three Arg-C peptides and one Asp-N peptide were probably acetylated (Table 1).
In particular, we found that the triply charged, unmodified P291-306 with an m/z of 600.4, had a retention time of 35.66 min in our analysis (Fig. 2B). The +3 ion tracing for singly acetylated peptide has multiple peaks migrating closely with the unmodified peaks ( Fig. 2B, middle trace). In order to distinguish whether these contained the acetylated peptide ions, the +4 ion tracing was also chromatographed (Fig. 2B, lower trace). The result revealed that there were two major peaks, at 35.6-min and 39.8-min, which were commonly seen in two tracings. On the other hand, many signals found between 50~60 min in the +3 ion tracing were not observed in the +4 ion tracing. This is consistent with the notion that these peaks contain ions that may have different molecular masses but have similar m/z values as the triply charged P291-306. Also, based on their relative abundances, these lysine residues appeared to be modified at a lower efficiency (Fig.2B) The collision-induced dissociation (CID) spectra of the 615-m/z ion in the 35.6-min peak were acquired for identification of the acetylation sites (Fig. 2C). The presence of y 4 to y 5 ions argues against the possibility that Lys-305 is acetylated, while the ion masses of b 9 and b 10 indicates that the acetyl group was on the lysine residue at the N-terminal end. The observation of y 15 ion indicated that Lys-292 is the acetylated residue.
The CID spectra of the 615-m/z ion in the 39.8-min peak are shown in Fig. 2D.
Observation of b 7 to b 12 ions indicated that the acetyl group was on the C-terminal three amino acids. Lys-305 is presumably the acetylation target base on the molecular sizes of y 1 and y 2 ion. This assumption is further supported by the fact that it is the only acetylable residue in the region. Therefore, we identified two distinct acetylation sites, Subsequently, we used the SIT approach to analyze peptides P354-379 and P380-401.
Both peptides could take up as much as two acetyl groups. The three sets of Arg-C peptides were subjected to LC/MS/MS analysis that was carried out in a mass-specific mode such that the necessary MS/MS spectra were acquired. The MS/MS spectra of Detection of in vitro acetylated p53 by antibodies specific to acetyl-Lys-305 and acetyl-Lys-382 of p53 The rabbit antibodies raised against Lys-305 (Ab-165) and Lys-382 (Ab-166) were purified by peptide affinity chromatography. The specificity of the purified antibodies was verified by a dot-blots analysis (Fig. 3A). We found that Ab-165 specifically recognized acetylated Ac-P300-310 (Ac-K305), and it did not recognize acetylated Ac-P376-387 (Ac-K382) or an acetylated peptide from HMG-14 peptide (Ac-HMG) (Fig.   3A, upper panel). Likewise, Ab-166 only recognized Ac-K382 but neither Ac-K305 nor Ac-HMG (Fig. 3A, lower panel).
We next tested if these antibodies can recognize full-length p53 acetylated by p300 in vitro (Fig. 3B). After in vitro acetylation, recombinant FLAG-p53 was analyzed by immunoblotting using Ab-165 and Ab-166 antibodies. The amount of p53 used for each reaction was about the same (Fig. 3B lower panel). Ab-165 recognized p53 that was acetylated in the presence of p300 and acetyl-CoA. The acetylated signals increased proportionately to the p300 amounts in the reaction (Fig. 3B upper panel). Similarly, Ab-166 specifically recognized acetylated, but not the non-acetylated p53 (Fig. 3B middle   panel). Ab-165 and Ab-166 did not recognize p53 in the absence of acetyl CoA or p300 in the acetylation reactions (Fig. 3B). These results indicated that Ab-165 and Ab-166 could specifically recognize acetylated p53. Together with the fact that Ab-165 could specifically recognize acetylated Lys-305, these data further corroborated the conclusion that Lys-305 was indeed acetylated in vitro by p300 in a dose-dependent fashion.

Lys-305 is acetylated by p300 in vivo
To address whether Lys-305 of p53 might be acetylated in vivo, we performed an in vivo acetylation assay that previously identified Lys-382 as an acetylated target. The H1299 cells were transfected with the pcDNA-p53 plasmid expressing wild-type p53 (Fig.   4). Without further treatment, no acetylated signal was detected on the Western blot using either antibody. Upon addition of sodium butyrate, an inhibitor of HDACs, acetylation at Lys-305 as well as at Lys-382 were detected (Fig. 4). Overexpression of HA-tagged p300 also enhanced acetylation at Lys-305, like Lys-382. This enhancement is correlated with the HA-p300 levels (Fig. 4). Synergism between sodium butyrate and p300 overexpression was also observed for the Lys-305 acetylation (Fig. 4). These data indicated that Lys-305, like Lys-382, was an acetylation target of p300.
Acetylation of p53 at Lys-305 is induced by various stress stimuli.
Previous studies have shown that various stress stimuli are capable of inducing p53 acetylation (3,15). To determine whether Lys-305 acetylation was induced in response to these stimuli we used Ab-165 and Ab-166 antibodies to monitor the acetylation levels of p53 in normal and stressed cells. First, we tested whether Lys-305 acetylation was induced by UV or ionization irradiation. MCF-7 cells were exposed to 50 J/m 2 UV-C or 20 Gy γrays (IR) and the cell lysates were prepared at various time points after treatment.
Before initiation of DNA damage, these cells were treated with sodium butyrate to reduce the rapid deacetylation by endogenous histone deacetylatase activity (27). Without any stimuli, the level of p53 in MCF-7 cells is low while the acetylation cannot be detected by either Ab-165 or Ab-166 ( Figure 5A, lane 1). Sodium butyrate treatment slightly increased the p53 level, which could be sustained for as long as 8 hr post treatment ( Figure 5A, lanes 2-4). After IR, acetylation at Lys-382 was detected as early as one hour and become prominent at 4 hr post irradiation. A similar course was also seen upon IR treatment for Lys-305 ( Figure 5A, lanes 5-7). Meanwhile, UV irradiation dramatically increased the acetylation for Lys-305 acetylation at 8 hr post treatment ( Figure 5A, lanes 8-10). The kinetics of irradiation-induced p53 acetylation for Lys-305 and Lys-382 were very similar. This suggests that Lys-305 and Lys-382 acetylation might be activated through a similar mechanism under these conditions.
The effects of H 2 O 2 and actinomycin D (ActD) on the p53 acetylation in MCF-7 cells were also investigated. Like irradiation-induced responses, actinomycine D induced the accumulation of p53 as well as the acetylation of Lys-305 (Fig. 5B). The H 2 O 2 -mediated oxidative stress, however, resulted in a slower accumulation of p53 than that elicited by actinomycine D. Intriguingly, it appeared that the kinetics of Lys-305 acetylation was quite different from that of Lys-382 (Fig. 5B). Lys-305 acetylation reached the highest level within 1 hr post treatment, while Lys-382 acetylation appeared to be higher at 4-8 hr post treatment (Fig. 5B)

Lys-305 is important for p53 transcriptional activity
To further determine the function of Lys-305, we tested whether it might affect p53-mediated transcriptional activation. We analyzed the effects of site-specific mutants of Lys-305 on transactivation function. Plasmids expressing various Lys-305 and Lys-382 mutants as well as wild-type proteins were introduced into the p53-null human lung cancer cell line H1299. These proteins were examined for their ability to activate a p53 responsive CAT reporter plasmid, p50-2, that contains two copies of p53 binding motif in its promoter region (12).
Overexpression of wild-type p53 generated a ~50-fold stimulation of the CAT activity (Fig. 6). It is similar to the previous reports that p53 is sufficient to stimulate this promoter activity of the reporter gene, p50-2 (29,35). When cells expressed p53 mutant K305R, the p50-2 promoter activity was about 2 folds higher than that induced by wild-type p53. In contrast, overexpression of the K305A mutant resulted in a 25-30% decrease of p50-2 promoter activity. These mutations have similar effects as seen with the Lys-382 mutants (Fig. 6A), further suggesting that Lys-305 may influence the transcriptional activity of p53 through a similar mechanism mediated by Lys-382 acetylation. Unexpectedly, overexpression of glutamine substituted p53 mutants, K305Q and K382Q, could not activate p50-2 promoter and the CAT activities were similar to mock control (Fig. 6A). These glutamine mutants seem to completely lose the intrinsic p53 transcriptional activity.
To further study the functional significance of Lys-305 acetylation, we tested whether it might affect p53-mediated transcriptional activation under stress stimuli (e.q. UV irradiation). After UV-irradiation, the wild-type p53-mediated transcriptional activity was elevated up to ~2-fold stimulation of the CAT activity (Fig. 6B). However, none of p53 mutants could respond to UV-irradiation to induce higher CAT activities.
The transcriptional activity of K305Q or K382Q was completely abolished whether the transfectants were treated with UV irradiation or not.

DISCUSSION
In this report, we first illustrated how partially acetylated peptides can be identified Here, our present data also indicate it can be applied to studying protein acetylation. Base on these observations, we surmise that this method should be applied to studies of other protein modifications, like methylation. Recently, we have indeed located the dimethylarginine residues on several nuclear proteins using similar approach (Tsay, unpublished observation).
The most intriguing finding in the present report is the identification of additional novel p53 acetylation sites, Lys-292 and Lys-305. Considering that they have not been reported in the literature, it is quite a surprise why acetylation at Lys-292 and Lys-305 was only detected in our system. There are several plausible explanations for our unique observation. The first is that the efficiency for acetylation at Lys-305 is probably much lower than that occurr at other sites (Fig 2B). , The ion counts of singly acetylated peptide for acetylation at Lys-381 and Lys-382 is about half of that for unmodified isoform (data not shown). In contrast, the amounts of the acetylated peptides at Lys-292 and Lys-305 are only about one thirtieth of the corresponding unmodified peptides.
Therefore, acetylation at Lys-292 and Lys-305 residues may have been neglected in previous experiments due to their suboptimal labeling under these conditions.
Another possibility is that we used a full-length p53 as the substrate, rather than the deletion mutants used by other investigators (10). The reductionistic approach has been frequently used in experiments involving identification of modified residues. However, an important pitfall is that construction of these deletion mutants, either randomly or deliberately, is associated with distortion of important local protein conformation. This can lead to introduction of false targets or obliteration of true ones. As these two sites are positioned over the region between DNA-binding and tetramerization domains, the context surrounding them may have been disrupted in mutants only consisting of individual domains.
Specific protein cleavage by endoproteinases is very important for SIT method.
One basic premise is that protein digestion is minimally or not affected by protein modification per se. If the moiety prevents the modified residue from being cleaved, it is conceivable that unmodified peptides derived from this region should have lengths different from those of modified peptides. Therefore, it becomes unlikely to identify the modified peptides using the unmodified counterparts as the references. Considering acetyl-lysine residues are very poor substrates for both trypsin and Lys-C, either enzyme is not suitable for our experiment. Therefore, we digested substrate proteins with enzymes whose cleavage is minimally affected by acetylation, for example, Asp-N and Previous studies have demonstrated that different post-translational modifications of p53 occurred under different stress-induced conditions. It may be activated through different signal transduction pathways (5,18). Our data demonstrated that the acetylation of p53 at Lys-305, like Lys-382, was responsive to all p53-activating agents.
Intriguingly, the acetylation profiles elicited by H 2 O 2 at Lys-305 and Lys-382 are quite different. The induction of Lys-382 acetylation is slower than that of Lys-305 (Fig. 5B).
It is likely that Lys-305 could be acetylated by a mechanism different from that of Lys-382. This interesting observation remains to be studied.
We have attempted to investigate the mechanism via which Lys-305 acetylation may modulate it transcriptional activity. Alteration of p53 subcellular distribution has been an attractive possibility since Lys-305 has been documented as part of bipartite nuclear localization signal (21). Nevertheless, we found that both K305R and K305A mutants and wild-type p53 have very similar immunofluorescence staining pattern (data not shown). The results are the same as those previously documented by other investigators (17). Therefore, it seems unlikely that alteration of subcellular/subnuclear distribution may fully account for the stimulatory/inhibitory effects of these mutants. In accordance with these findings, we also found that subcellullar/subnuclear localization pattern of the acetyl-Lys-305, as well as acetyl-Lys-382, is identical to the overall population of p53 molecules (data not shown). This also supports the idea that Lys-305 acetylation is not directly associated with subcellular/subnuclear targeting of p53 protein.
To further examine whether glutamine/arginine substitutions may mimic constitutively acetylated/nonacetylated lysine residues, cells were cotransfected with K305Q or K382Q mutants, the p50-2 promoter activity was markedly decreased in both cases (Fig. 6). and Lys-382. Meanwhile, these p53 mutants could not respond to UV-irradiation as efficiently and failed to produce optimal transcriptional activity that is intrinsic to wild-type p53. These results suggest that Lys-305, like Lys-382, played an important role in modulating the precise transcriptional activity of p53.
The Lys-292 of p53 was also identified as an in vitro p300 acetylated target by SIT approach (Fig. 2D). Unfortunately, we have not been able to generate the specific anti-acetyl-Lys-292 antibody to test the possibility that Lys-292 can also be acetylated in vivo. However, we have constructed K292 mutants and found that these mutants also We have also tested the possibility that Lys-305 acetylation might actually modulate the transcriptional activity of p53 through influencing the modification of other p53 residues. It has been documented that post-translational modification at certain sites can alter the properties of p53 such as protein stability (9,33), tetramerization (30), and DNA-binding capacity (4,16). In order to test this possibility, we investigated how modification states of other residues are changed in K305A and K305R mutants as compared to the wild-type protein. Our preliminary data appeared to show that K382 acetylation is suppressed in K305A mutant but enhanced in K305R mutant. It remains to be examined whether this is the primary mechanism responsible for the altered transcriptional activity of these mutants. It is also intriguing to examine whether the modification of other sites may also be affected by K305 acetylation.