Interaction between p53 N terminus and core domain regulates specific and nonspecific DNA binding

Significance

The p53 tumor suppressor is a sequence-specific DNA binding protein that activates gene transcription to regulate cell survival and proliferation. This report investigates the role of the p53 N-terminal transactivation domain in regulating DNA binding affinity and specificity. The results suggest the N-terminal acidic transactivation region of p53 dynamically interacts with the DNA binding domain near residues that contact DNA, thus inhibiting DNA binding and increasing sequence specificity by nucleic acid mimicry or electrostatic shielding. Furthermore, p53 N-terminal–interacting proteins and posttranslational modifications may regulate DNA binding affinity and specificity partly by modulating the N terminus–DNA binding domain interaction.

Abstract

The p53 tumor suppressor is a sequence-specific DNA binding protein that activates gene transcription to regulate cell survival and proliferation. Dynamic control of p53 degradation and DNA binding in response to stress signals are critical for tumor suppression. The p53 N terminus (NT) contains two transactivation domains (TAD1 and TAD2), a proline-rich region (PRR), and multiple phosphorylation sites. Previous work revealed the p53 NT reduced DNA binding in vitro. Here, we show that TAD2 and the PRR inhibit DNA binding by directly interacting with the sequence-specific DNA binding domain (DBD). NMR spectroscopy revealed that TAD2 and the PRR interact with the DBD at or near the DNA binding surface, possibly acting as a nucleic acid mimetic to competitively block DNA binding. In vitro and in vivo DNA binding analyses showed that the NT reduced p53 DNA binding affinity but improved the ability of p53 to distinguish between specific and nonspecific sequences. MDMX inhibits p53 binding to specific target promoters but stimulates binding to nonspecific chromatin sites. The results suggest that the p53 NT regulates the affinity and specificity of DNA binding by the DBD. The p53 NT-interacting proteins and posttranslational modifications may regulate DNA binding, partly by modulating the NT–DBD interaction.

The p53 tumor suppressor protein plays a critical role in regulating proliferation, cell death, and differentiation. p53 is activated by a plethora of cellular stresses such as DNA damage, oncogene activation, replicative stress, and hypoxia (1). The activation process involves posttranslational modifications that suppress p53 degradation by MDM2 and increase p53 DNA binding affinity. Once activated, p53 tetramers bind to responsive elements in genomic DNA to stimulate the transcription of numerous target genes that orchestrate stress tolerance, DNA repair, cell cycle arrest, apoptosis, and tumor suppression (2). p53 is mutated in ∼50% of human tumors, with a higher frequency in specific tumor types and after relapse. Mutated p53 loses transcriptional activity and gains new functions that drive tumor progression (3).

The p53 polypeptide can be divided into distinct structural and functional domains. The N-terminal 90 residues [N terminus (NT)] contain two transactivation domains (TAD1: 1–40, TAD2: 40–60) and a proline-rich region (PRR: 60–90). TAD1 also contains the main binding site for MDM2 and MDMX. TAD2 is involved in transcription activation and protein interactions, such as with RPA. Mutations in both TAD1 and TAD2 are needed to eliminate the tumor suppressor function (4). TAD1 and TAD2 also contain multiple phosphorylation sites that regulate p53 degradation and activity during stress response. Mutations or deletions in the PRR compromise p53 degradation, transactivation, apoptosis, growth suppression, and tumor suppression (5–7). The central DNA binding core domain (DBD: 94–312) is required for sequence-specific DNA binding and is the target of most point mutations in cancer. The C terminus (CT: 312–393) contains an oligomerization domain (323–355) and a lysine-rich C-terminal tail (364–393) that harbors sites for acetylation, methylation, and phosphorylation.

Stress response by p53 is dependent on intricate mechanisms that regulate its level and activity. The MDM2/MDMX feedback loop plays a pivotal role in controlling the cellular level and transcriptional activity of p53 (8). MDM2 binds to the p53 NT to promote p53 ubiquitination and degradation in the absence of stress. Both MDM2 and MDMX have central acidic domains that engage in second-site interactions with the p53 DBD and inhibit DNA binding (9, 10). Additionally, the phosphorylation of p53 on NT residue T18 inhibits MDM2 binding (11, 12). DNA damage-induced phosphorylation of MDM2 inhibits dimerization and E3 ligase activity, which also contributes to accumulation of p53 (13).

Both the NT and CT of p53 have a high degree of intrinsic disorder. The p53 CT has well-documented effects in regulating DNA binding. The CT antibody Pab421 stimulates DNA binding by p53 in EMSAs. Acetylation of the CT by p300/CBP stimulates specific DNA binding in vitro (14). Analysis of p53 mutants suggested that the C-terminal tail was important for binding and activation of target genes with suboptimal binding sites (15). The C-terminal tail engages in intersubunit interaction with the DBD and stimulates p53 oligomerization (16). The CT is capable of nonspecific DNA binding and enabling p53 to slide on DNA, which is an important step in the search for specific binding sites (17). Mouse models that delete the C-terminal tail caused p53 stabilization and overactivation, resulting in developmental abnormalities (18, 19).

The p53 NT performs several important functions. TAD1 and TAD2 interact with basal transcription factors such as TBP, TFIID, and coactivator p300/CBP. Phosphorylation of the NT after DNA damage stimulates recruitment of p300/CBP that, in turn, acetylates the CT (12). The NT was also implicated in suppressing sequence-specific DNA binding by the DBD (20). Antibody binding to the NT changed the thermostability of the DBD, suggesting that the NT participates in DBD folding or is part of the final structure (21). Deletion of the PRR alone stimulated DNA binding in vitro (22). A single-molecule FRET study suggested the NT weakly interacts with the DBD (23). Intramolecular interaction between the DNA binding domain and adjacent regulatory domains was found to have autoinhibitory roles in other DNA binding proteins (24).

In this report, we provide evidence that the p53 NT interacts with the DNA binding site in the DBD using the TAD2 and PRR sequences. The interaction not only reduces overall DNA binding affinity; importantly, it also increases p53 selectivity for specific DNA in vitro and in vivo. The findings corroborate a recent in vitro analysis and provide insight for understanding p53 regulation by NT modifications and binding proteins (25).

Results

p53 NT Interacts with the DBD.

Previous work showed that recombinant p53 with deletion of the NT (∆1–90) had increased DNA binding in DNase I footprinting and EMSAs (20). It is also known that several antibodies against epitopes of the NT stimulated DNA binding (21). The p53 NT is an intrinsically disordered region that engages in numerous protein–protein interactions, some of which are kinases that regulate p53 through phosphorylation (26). These studies suggested the possibility of a direct interaction between the NT and DBD, but the absence of high-resolution structural studies to detect this interaction has left a number of open questions (23, 25). To test whether the NT interacts directly with the DBD, we used a proteolytic fragment release assay for detecting weak intramolecular and intermolecular domain interactions (9, 27). PreScission protease cleavage sites, followed by epitope tags, were inserted into p53 to generate the p53c2 and p53c1 constructs (Fig. 1 A and B). PreScission cleaves p53c2 into three fragments: FLAG-tagged 1–90, HA-tagged 91–299, and Myc-tagged 300–393. Cleavage of p53c1 produced FLAG-tagged 1–90 and untagged 91–393.

Fig. 1.

Detection of p53 domain binding. (A) Design of p53c1 and p53c2 constructs with Precision (PreS) cleavage sites and epitope tags (FLAG, HA, and Myc). (B) Diagram of proteolytic fragment release assay. (C and D) p53c2 expressed in H1299 was immobilized on FLAG or Myc beads and cleaved by PreScission for 30 min. The beads and supernatant were analyzed for the DBD fragment by HA Western blot (WB). Pairwise quantitation is shown below the blots. IP, immunoprecipitation. (E) p53c1 was immobilized on FLAG beads, cleaved by PreScission, and analyzed for DBD fragment release by WB using an antibody for full-length p53 (FL393). Pairwise quantitation is shown below the blots.

The p53c2 construct was expressed in H1299 cells, immobilized on beads, and cleaved with PreScission. Interactions between fragments were interrogated by immobilizing p53c2 using different antibodies and analyzing the retention/release ratio of each fragment by Western blot (Fig. 1B). Using this assay, the NT–DBD interaction (FLAG immobilization; Fig. 1C) was significantly stronger than the DBD–CT interaction (Myc immobilization; Fig. 1D). A similar analysis using the p53c1 construct with a single PreScission site also showed significant binding of the NT to the 91–393 fragment (Fig. 1E). The results demonstrate the p53 NT directly interacts with the DBD.

Intersubunit Interaction Between the NT and DBD.

p53 forms tetramers (dimer of dimer) mediated by the CT oligomerization domain (325–356). To determine whether the NT–DBD interaction diagrammed in Fig. 1 is intramolecular or intermolecular, point mutations were introduced into p53c1 to disrupt oligomerization (Fig. 2A). The ALAL mutant forms dimers but not tetramers, whereas the KEEK mutant is monomeric (28). Analysis of the ALAL and KEEK mutants using the fragment release assay showed the NT–DBD interaction was significantly reduced by the KEEK mutation but partially restored in the ALAL mutant (Fig. 2B), suggesting that when p53 forms dimers or tetramers, the NT of one subunit preferentially interacts with the DBD of another p53 within the same dimer or with p53 in a different dimer. The result did not rule out intramolecular foldback interaction, which is structurally permissible based on a crystallographic study (29).

Fig. 2.

Analysis of p53 domain binding. (A) Diagram of WT, ALAL, and KEEK p53c1 constructs. PreS, Precision cleavage sites. (B) WT (Wt), ALAL, and KEEK p53c1 expressed in H1299 were immobilized on M2 beads, cleaved by protease, and analyzed for DBD fragment release by Western blot using the p53 antibody FL393. (C) Lysate of H1299 transfected with FLAG-p53-90–393 was incubated with beads loaded with GST or GST-p53-1–90. Pull-down of FLAG-p53-90–393 was detected by FLAG Western blot. (D) Lysate of H1299 transfected with FLAG-p53-90–393 was incubated with glutathione beads loaded with GST, GST-p53-1–90, GST-p53-1–40 (TAD1), GST-p53-35–65 (TAD2), and GST-p53-65–90 (PRR). Pull-down of FLAG-p53-90–393 was detected by FLAG Western blot.

To confirm the interaction detected using the fragment release assay, GST-p53-1–90 produced in Escherichia coli was used to pull down p53-90–393 expressed in H1299 cells. GST-p53-1–90, but not GST, pulled down p53-90–393 (Fig. 2C). Further mapping performed using GST fusions of TAD1, TAD2, and the PRR showed that TAD2 and the PRR retained weak binding to p53-90–393, suggesting that both regions participate in binding to the DBD (Fig. 2D).

p53 NT Inhibits Binding to DNA.

Previous EMSAs and DNase footprinting assays showed that p53-90–393 produced in insect cells had increased DNA binding in vitro (20). To rule out effects from posttranslational modification caused by the NT deletion, we tested the immediate effect of removal of the NT by protease cleavage. The p53c1 and NT truncation mutants were purified from H1299 cells (Fig. 3A). The purified p53c1 was cleaved with PreScission to remove the NT before incubating with biotinylated oligonucleotide containing the p53 binding site. In this assay, cleavage of the NT stimulated DNA binding (Fig. 3B). The DNA binding activity of 90–393 was approximately fourfold higher than that of 1–393 in a titration analysis (SI Appendix, Fig. S1A). The p53c1 with deletion of 1–47 had modestly stronger DNA binding than that of 1–393 and was less significantly stimulated by cleavage of remaining residues 48–90. The p53c1 with deletion of 1–82 or 1–90 was not stimulated by cleavage (Fig. 3 B and C). The results showed that the NT inhibited DNA binding by the DBD.

Fig. 3.

p53 NT inhibits DNA binding in vitro. (A) Diagram of N-terminal–truncated p53c1 mutants with C-terminal FLAG tag. (B) FLAG-tagged p53c1 mutants were immunopurified from transfected H1299 cells, cleaved with protease, and incubated with beads loaded with biotinylated oligonucleotide containing the p53 binding site. The captured p53 was detected by Western blot using FL393. The level of p53 input was confirmed by Western blot. (C and D) Purified p53c1 mutants were cleaved with protease and incubated with immobilized biotinylated oligonucleotide in the presence of antibodies. The captured p53 was detected by Western blot using FL393. Pairwise quantitation is shown below the blots. (E) Purified FLAG-p53-90–393 was incubated with immobilized biotinylated oligonucleotide in the presence of NT peptides (200 μM). The captured p53 was detected by FLAG Western blot. Quantitation is shown below the blots.

A previous study showed that antibodies against epitopes of the NT stimulated DNA binding in EMSAs (20). In the oligonucleotide pull-down assay, Pab1801 (epitope: 46–55) and DO-1 (epitope: 11–25) also stimulated p53c1 DNA binding (Fig. 3C). Pab1801 did not stimulate DNA binding by the ∆1–47 mutant, suggesting that the antibody acts by neutralizing the inhibitory effect of the NT (Fig. 3D). Pab421 (epitope: 363–372) abrogated DNA binding in the oligonucleotide pull-down assay (Fig. 3D), similar to its effect in DNase footprinting (20).

The results above suggested that TAD2 and the PRR had inhibitory effects on DNA binding. Since the cleavage assay did not have the sensitivity to further dissect the TAD2/PRR region, synthetic peptides representing TAD2 (48–62) and the PRR (75–90) were tested in the DNA binding assay. The TAD2 peptide modestly inhibited DNA binding by 90–393, whereas two unrelated control peptides had no effect (Fig. 3E). Scrambling the negatively charged TAD2 peptide (SCR48–62) did not abrogate its activity, suggesting that charge was important for the peptide activity in this assay (Fig. 3E). The PRR peptide had significant activity that was abrogated by scrambling the sequence, suggesting that it interacted with the DBD in a sequence-dependent fashion (Fig. 3E). In control experiments, both peptides had no effect on the DNA binding by full-length p53 (SI Appendix, Fig. S1B). This was expected since full-length p53 was already in a self-inhibited state.

p53 NT Promotes Dissociation from DNA.

To further investigate how the NT inhibits DNA binding by the DBD, we used an in vitro DNA binding assay based on luciferase fragment complementation. The N-terminal domain of luciferase was fused to a high-affinity zinc finger protein (ZF-Nluc), and the C-terminal domain of luciferase was fused to p53 (p53-Cluc). Incubation of ZF-Nluc and p53-Cluc with an oligonucleotide containing both zinc finger and p53 binding sites will place Nluc and Cluc in close proximity to restore luciferase activity (Fig. 4A). After the luciferase activity reaches a sufficient level, adding excess oligo containing only the p53 binding site causes a loss of luciferase activity, presumably by sequestering p53-Cluc that dissociates from the dual-site oligo (Fig. 4B). Analysis using this assay showed that deleting the NT (90–393-Cluc, Fig. 4C) significantly reduced p53 dissociation from DNA (Fig. 4D). When the p53c1-Cluc construct was used (Fig. 4C), protease cleavage of the NT (producing 93–393-Cluc) reduced dissociation from DNA (Fig. 4E), whereas the protease had no effect on noncleavable p53-Cluc (Fig. 4E). The analysis revealed a role of the NT in promoting p53 dissociation from DNA.

Fig. 4.

p53 NT promotes dissociation from DNA. (A) Diagram of the in vitro luciferase fragment complementation assay for detecting p53 dissociation from DNA. (B) p53-Cluc was incubated with ZF-Nluc and ZPBS oligonucleotide for 30 min to assemble active complexes. Competitor oligonucleotide containing the p53 binding site was added, and luciferase activity was monitored for an additional 75 min. (C) Diagram of p53-Cluc fusion constructs with and without the protease cleavage site. (D) p53-Cluc and 90–393-Cluc were preincubated with ZF-Nluc and ZPBS oligonucleotide to assemble active complexes. Competitor oligonucleotide was added (0 min), and luciferase activity was monitored for 45 min. (E) Cleavable p53c1-Cluc and noncleavable p53-Cluc were preincubated with ZF-Nluc and ZPBS oligonucleotide in the presence and absence of protease to assemble active complexes. Competitor oligonucleotide was added (0 min), and luciferase activity was monitored for 45 min.

When equal amounts of p53-Cluc and 90–393-Cluc were tested in the complementation assay (SI Appendix, Fig. S2A), 90–393-Cluc reproducibly generated approximately twofold stronger luciferase activity than p53-Cluc (SI Appendix, Fig. S2B), corroborating the stronger DNA binding by 90–393 in the oligonucleotide pull-down assay (Fig. 3B). Despite having stronger DNA binding, 90–393-Cluc was more sensitive to competition by poly(deoxyinosinic-deoxycytidylic) acid [poly(dI-dC)] (SI Appendix, Fig. S2C), suggesting that it has lower specificity.

The NT Increases p53 DNA Binding Specificity in Vitro and in Vivo.

Isothermal titration calorimetry (ITC) was used to determine the effect of the p53 NT on DNA binding by titrating samples of either the 1-312 (ND) or 94-312 (DBD) with a 20-bp DNA fragment containing the consensus binding site or a scrambled binding site. The presence of the NT decreased the binding affinity of the DBD for consensus DNA by 26-fold (Fig. 5A). The NT also reduced the DBD affinity for scrambled DNA by 60-fold. One interesting result from the ITC experiments is the apparent difference in DNA binding specificity between the ND and DBD. This can be estimated by taking the ratio of K_d values for binding scrambled and consensus DNA. For the ND, this ratio is 8.2, and for the DBD, it is 3.5. Therefore, the presence of the NT decreases DNA binding affinity but increases binding specificity, which is essential for p53 to discriminate the promoter sites of target genes.

Fig. 5.

p53 NT inhibits DNA binding in vivo. (A) DNA binding affinity of the p53 ND fragment (1–312) and DBD (94–312) to the consensus DNA binding site and scrambled control DNA as determined by ITC. (B) p53 and N-terminal deletion mutants were expressed in H1299 cells using a lentiviral vector and induced to comparable levels using doxycycline. (C) p53 mutant expression was induced by doxycycline for 24 h. p53 binding to the p21 promoter was determined by ChIP. (D) WT p53 and the 90–393 mutant were induced in H1299 by doxycycline for 24 h. p53 occupancy at specific target promoters (average of p21, MDM2, PUMA, and Fas1) and nonspecific repetitive sequences (average of LINE1, Alu, and α-satellite) was determined by ChIP. The occupancy by WT p53 was arbitrarily set as onefold.

To determine the effect of the NT on DNA binding in vivo, p53 and mutants 40–393, 70–393, and 90–393 were expressed in H1299 cells to similar levels using an inducible lentivirus vector (Fig. 5B). Chromatin immunoprecipitation (ChIP) showed that 70–393 and 90–393 bound to several target gene promoters (p21, MDM2, PUMA, and Fas1) with 10-fold and 20-fold higher efficiency, respectively, than WT p53 (Fig. 5C and SI Appendix, Fig. S3). DNA binding by 40–393 was similar to full-length p53. Therefore, TAD2 and the PRR strongly inhibited specific DNA binding in vivo.

ChIP was also used to detect p53 occupancy at repetitive elements LINE1, Alu, and pericentric α-satellite repeats as a measure of in vivo nonspecific DNA binding. The results showed that 90–393 binding to repetitive elements was, on average, ∼20-fold higher than WT p53 (Fig. 5D). In the same experiments, 90–393 binding to the specific targets p21, MDM2, and PUMA was, on average, ∼10-fold higher than WT p53. Therefore, removal of the NT stimulated p53 DNA binding, but resulted in an approximately twofold loss in specificity in vivo.

MDMX Regulates p53 DNA Binding Specificity.

Several important p53 regulators bind to TAD1 or TAD2, including MDM2/MDMX (TAD1), p300/CBP (TAD1 and TAD2), and RPA70 (TAD2). The effect of MDMX was examined since its p53 binding was more robust in our hands than p300 or RPA. In the in vitro DNA dissociation assay, GST-MDMX-1–121 (p53 binding domain) modestly inhibited the dissociation of p53-Cluc from DNA (Fig. 6A). Coexpression of MDMX-1–200 or full-length MDMX also inhibited the interaction of the p53 NT with the DBD in the fragment release assay (Fig. 6B). Therefore, MDMX binding may modulate the specific or nonspecific DNA binding by p53.

Fig. 6.

MDMX regulates p53 domain interaction and DNA binding. (A) p53-Cluc was preincubated with ZF-Nluc and ZPBS oligonucleotide to assemble active complexes in the presence of GST, GST-RPA-1–121, and GST-MDMX-1–120. Competitor oligonucleotide was added (0 min), and luciferase activity was detected at indicated time points. (B) MDMX disrupts p53 NT–DBD interaction. p53c1 was coexpressed with MDMX-1–200 and MDMX-1–490 in H1299 cells. p53c1 (and the associated MDMX) was immobilized on M2 beads, cleaved with protease for 40 min, and analyzed for NT–DBD interaction by Western blot detection of a released 91–393 fragment. Sup, supernatant. (C) U2OS cells stably infected with lentivirus expressing doxycycline-inducible MDMX were induced and treated with γ-radiation for 4 h. Expression of p53 pathway markers was determined by Western blot. IR, ionizing radiation. (D) U2OS cells expressing doxycycline-inducible MDMX were treated with doxycycline for 24 h and irradiated with 10 Gy IR. p53 binding to specific promoters (average of p21, MDM2, Fas1, and PUMA) and repetitive sequences (average of LINE1, Alu, and α-satellite) was analyzed by ChIP 4 h after irradiation. The occupancy by p53 in untreated cells was arbitrarily set as onefold. Con, control. *P < 0.05.

ChIP analysis of p53 in a U2OS cell line expressing ectopic MDMX (Fig. 6C) showed that MDMX suppressed specific DNA binding and stimulated nonspecific DNA binding, thus reducing the specificity of p53 in vivo (Fig. 6D). DNA damage (γ-radiation) preferentially enhanced p53 binding to specific sites, thus increasing the specificity of p53. In the presence of MDMX, DNA damage was less effective in stimulating specific binding (Fig. 6D). Therefore, MDMX reduced p53 specificity by blocking binding to specific sites and increasing binding to nonspecific DNA, consistent with its role as an important p53 inhibitor without promoting degradation.

NMR Reveals a Dynamic Interaction Between the p53 NT and DBD.

The structure and dynamics of the ND fragment (1–312) were examined by NMR. Fig. 7A shows an overlay of the ¹H-¹⁵N heteronuclear single quantum coherence (HSQC) spectra for the apo and DNA-bound ND. The ND fragment binds the consensus site to form a >100-kDa tetramer; thus, resonances for DBD residues are not visible. However, resonances for the disordered TAD1/2 and PRR residues are strong because they remain dynamic (30) (Fig. 7B). In the spectrum of the DNA-bound ND, several of the resonances for residues in TAD1, TAD2, and the PRR undergo chemical shift and intensity changes (Fig. 7 B and C). These chemical shifts are nearly identical to the ones we observe for a p53 NT fragment (1–89) without the DBD (SI Appendix, Fig. S4), suggesting that TAD1/2 and the PRR in the ND fragment are released from interactions with the DBD when it is bound to DNA.

Fig. 7.

Intramolecular interaction between TAD2 and the DBD. (A) Overlay of ¹H-¹⁵N HSQC spectra for the apo (blue) and DNA-bound (red) ND. (B) Expanded view showing most of the disordered TAD1/2 resonances. (C) TAD2 and PRR residues with the largest chemical shift changes. (D) Plot of the average amide ¹H and ¹⁵N chemical shift changes for TAD1/2 and PRR residues in the apo vs. DNA-bound ND. The gray line marks the digital resolution of the HSQC experiment. (E) Intensity ratios for TAD1/2 and PRR residues calculated as DNA-bound divided by the apo ND. The gray line shows the expected intensity ratio if there is no interaction. (F) Transverse R2s of TAD1/2 and PRR residues in the ND (blue) and the NT (1–89, black).

A plot of the averaged amide ¹H and ¹⁵N chemical shift changes between the apo and DNA-bound ND shows the largest differences localize to TAD2 (residues 40–60, also an RPA70 binding site), with a maximum shift observed for the consistently overlapped resonances of residues W53 and F54 (Fig. 7D). Resonances for the PRR also show significant chemical shifts upon DNA binding, consistent with previously published data where the PRR affects DNA binding (22, 31). The ¹H-¹⁵N HSQC spectra for a p53 1–73 fragment containing TAD1 and TAD2 (SI Appendix, Fig. S5) showed that the changes were not due to TAD1/2 interacting with DNA.

When comparing the intensity differences of the TAD1/2 region of the apo ND vs. the DNA-bound ND, there is an increase in the peak intensities of the residues located in TAD2 resulting in a maxima for the bound/apo intensity ratio that also overlaps the RPA70 binding site, consistent with the TAD2 no longer tumbling with the DBD. Transverse relaxation rates (R2s) were measured for the apo ND to further assess whether residues from TAD1/2 are tumbling more slowly than expected for an intrinsically disordered protein. This could indicate the presence of a direct interaction between TAD1/2 and the DBD. The R2 values for the apo ND and the disordered NT fragment (1–89) are compared in Fig. 7F. R2 values for residues in the RPA70 and MDM2 binding sites in the ND are elevated compared with the rest of the residues of the NT, consistent with a direct interaction between these regions and the DBD.

In further tests, the transinteraction between ¹⁵N-labeled 1–89 and the unlabeled DBD also identified similar interacting residues in TAD2 (SI Appendix, Fig. S6). TAD2 interaction with RPA70 results in α-helix formation (32, 33). However, chemical shift and d2D (determination of secondary structure populations from chemical shifts) analysis indicated that the interaction between TAD2 and the DBD results in little to no change in the secondary structure of TAD2 (SI Appendix, Supplemental Material and Fig. S7), presumably because the interaction is weak and dynamic. Overall, the results suggest that TAD2 interacts with the DBD and is displaced by DNA binding.

TAD2 Interacts with the DNA Binding Site on the DBD.

To locate the interaction site for TAD2 and the PRR on the DBD, we compared the ¹H-¹⁵N HSQC spectra for the apo ND (1–312) and DBD (94–312) to identify chemical shift differences for residues in the DBD region of the ND (Fig. 8A). The largest chemical shifts occur on residues K120, H168, H193, N210, and S240, as well as on a large cluster from R282 to E294 (Fig. 8B). DBD residues with the largest chemical shifts either make contact with DNA or are adjacent to contacting residues (red triangles in Fig. 8B). In addition, some residues with large shifts have been reported to interact with a portion of the NT; specifically, H168 and M169 form a pocket where the side chain of W91 binds and presumably stabilizes protein–protein interactions in the oligomeric form of p53 (29). Mapping residues with the largest chemical shifts to the DNA-bound structure of the DBD identifies a region localized to the periphery of the DNA binding cleft (Fig. 8C). The transinteraction between the ¹⁵N-labeled DBD and unlabeled 1–89 also identified similar interacting residues in the DBD (SI Appendix, Fig. S8). Overall, the NMR analyses suggest that the TAD2 and PRR regions interact with the DNA binding surface of the DBD to block binding to DNA.

Fig. 8.

TAD1/2 induces chemical shifts in the DBD. (A) Overlay of ¹H-¹⁵N HSQC spectra for the apo DBD (green) and ND (blue). (B) Plot of the average amide ¹H and ¹⁵N chemical shift changes for apo DBD and ND residues. Residues that contact DNA are indicated with red triangles, and red arrows show the positions of hotspot mutations. (C) Residues with the largest chemical shifts mapped onto the DBD tetramer (gray) bound to DNA (transparent cyan) (Protein Data Bank ID code 4hje).

Discussion

The p53 protein and key regulators MDM2 and MDMX contain extensive regions of disordered sequences that have important functions. In MDM2 and MDMX, the disordered central acidic regions form intramolecular contacts with the RING domain and p53 binding domain to activate E3 ligase function or inhibit p53 binding (27, 34, 35). The C-terminal tail of p53 interacts with the DBD of another p53 in the tetramer to stabilize oligomerization and DNA binding (16). In this report, we used several approaches to demonstrate that the p53 NT engages in intramolecular or intersubunit interaction with the DBD to regulate DNA binding. Our results provided mechanistic insight on how p53 NT deletion and binding of NT antibodies stimulate DNA binding. Furthermore, our in vitro and in vivo evidence showed that the NT differentially regulates specific and nonspecific DNA binding, suggesting a mechanism by which NT-interacting proteins regulate p53 DNA binding specificity in a physiological setting.

NMR analysis using the monomeric ND fragment showed that NT–DBD intramolecular interaction is permissible. However, the interaction is strengthened in p53 dimer or tetramer, suggesting that intersubunit interaction is favored in the tetramer. The NT-DBD binding causes small chemical shift changes and negligible secondary structure change in TAD2, suggesting it is a weak interaction (36). We propose that the NT may regulate the DBD DNA binding by nucleic acid mimicry or electrostatic screening. TAD2 can assume an α-helical structure and mimic single-stranded DNA when bound to RPA70 (32). The enhanced NT–DBD interaction in p53 oligomers may also favor a mimicry mechanism. Alternatively, acidic residues in the NT may dynamically interact with basic residues around the DNA binding surface to provide electrostatic screening. Without forming a stable structure, TAD2 may use multiple weak interactions with the DBD to cause a large effect on DNA binding affinity, as suggested in the polyvalency model (37, 38). Hydrophobic interactions may also play a role since W53 and F54 of TAD2 have the largest chemical shift changes despite remaining dynamic when they interact with the DBD. Interestingly, two DBD residues with the largest chemical shift changes due to the TAD2 interaction are cancer mutation hotspots important for conformation (G245) or DNA contact (R248).

Krois and coworkers (25) recently described an interaction between p53 TAD2 and the DBD using NMR-labeled 1–61 spliced to unlabeled 62–393. Our NMR analysis identified similar interactions as reported in the study by Krois et al. (25). The chemical shift changes and intensity ratios we observed for TAD2 were also similar but the intensity ratios for TAD1 were different suggesting a more complex interaction in the tetramer. Our analysis using a uniformly labeled 1–312 fragment enabled us to detect additional interactions between the PRR and DBD, and more clearly determine the effect of the NT on DNA binding specificity without interference from nonspecific DNA binding by the C-terminal tail. Taken together, our work and the study by Krois et al. (25) conclusively demonstrate the p53 NT is a regulator of DNA binding affinity and specificity in vitro and in vivo, leading to the intriguing possibility that NT phosphorylation or protein binding may alter the NT–DBD interaction and DNA binding.

Our in vitro and in vivo analyses provided corroborating evidence that the presence of the NT reduced the DBD DNA binding affinity by ∼10-fold but increased sequence specificity by twofold. These activities of the NT may be biologically relevant in ensuring p53 can rapidly bind and activate target promoters in the stress response. The nuclear level of p53 after stress-induced accumulation is on the order of 10⁴ molecules per cell (39). To specifically search and activate target genes in a timely fashion during the stress response, p53 may employ scanning and an interstrand transfer mechanism. The p53 CT is involved in nonspecific DNA binding and sliding that is critical for a rapid search of target sites (17). The DBD does not have the ability to slide on DNA; its high affinity for nonspecific DNA may interfere with interstrand transfer and CT-mediated sliding. By promoting DBD dissociation from DNA, the NT may facilitate rapid identification of target promoters. Given the large excess of nonspecific DNA sequences in the genome, the increase in specificity provided by the NT may have important roles in reducing trapping of p53 by irrelevant sequences.

The competition between the p53 NT and DNA for binding to the DBD provides a mechanism that explains certain effects of NT posttranslational modifications and binding proteins. Phosphorylation of T55 by the TAF1 coactivator causes p53 dissociation from DNA during recovery from the DNA damage response (40). Phosphorylation of T55 increases the negative charges on TAD2 and should further increase its inhibitory effect on the DBD. Phosphorylation of TAD1 stimulates interaction with p300/CBP, which should free the DBD for DNA binding (12). Recruitment of p300/CBP further acetylates the CT and stimulates specific DNA binding (14). Conversely, DNA binding by the DBD stimulates p53 acetylation by p300, possibly by freeing the NT and enabling interaction with p300 (41). A combination of these changes should cooperatively activate p53. Our results also revealed an activity of MDMX that stimulates p53 binding to nonspecific chromatin sites and inhibits specific DNA binding. MDMX engages in complex multisite interactions with p53 (9). The MDMX N-terminal domain binds to p53 TAD1 and stimulates DNA binding, whereas the central acidic region interacts with the p53 DBD in cooperation with CK1α and inhibits DNA binding. Both MDMX–p53 and MDMX–CK1α interactions are inhibited by DNA damage signaling, providing a mechanism to activate p53-specific DNA binding during the stress response (42).

Materials and Methods

Plasmids and Cell Lines.

The p53c1 construct contains the sequence LEVLFQGPG inserted before p53 residue 90, which allows removal of residues 1–89 by cleavage with PreScission protease. The p53c2 construct contains LEVLFQGPYPYDVPDYA and LEVLFQGPEEQKLISEEDL inserted before p53 residues 90 and 299, respectively. GST-PreScission protease fusion was purified from E. coli by a glutathione agarose column. Cell lines H1299 (p53-null), U2OS (p53 WT), H1299 with inducible p53, and U2OS with inducible MDMX were maintained in DMEM with 10% (vol/vol) FBS.

Proteolytic Fragment Release Assay.

H1299 cells were transiently transfected with p53c1 or p53c2 plasmid by standard polyethyleneimine (PEI) transfection (PEI 25000; Polysciences, Inc.). Cells were lysed with lysis buffer [150 mM NaCl, 50 mM Tris⋅HCl (pH 8.0), 0.5% Nonidet P-40, 5 mM EDTA, 0.5 mM DTT]. Cell lysate (1 mL) from ∼2 × 10⁶ cells (a 10-cm plate) was incubated with 20 μL of packed protein A beads with chemically cross-linked FLAG or Myc mouse monoclonal antibody for 18 h at 4 °C. The beads were washed two times with PreScission buffer [150 mM NaCl, 10 mM Hepes (pH 7.5), 0.05% Nonidet P-40, 0.5 mM DTT, 10% glycerol] and suspended in 200 μL of PreScission buffer. PreScission protease was added to a final concentration of 0.1 μg/μL, and the beads were incubated at 23 °C with shaking for 10–30 min. The digestion mixture was centrifuged for 10 s, and the beads (bound material) and supernatant (released material) were separated. The beads were washed once with PreScission buffer. The beads and supernatant were boiled in Laemmli sample buffer [4% (wt/vol) SDS, 20% glycerol, 200 mM DTT, 120 mM Tris (pH 6.8), 0.002% bromophenol blue] and analyzed by SDS/PAGE and Western blot using FL393 or HA antibody to determine the bound/released ratio of each fragment.

Purification of p53 and DNA Affinity Immunoblotting.

H1299 cells were transiently transfected with FLAG-tagged p53 using the PEI method. Cells from a 10-cm plate were lysed in 1 mL of lysis buffer and centrifuged for 10 min at 14,000 × g, and the insoluble debris was discarded. The lysate was incubated with a 30-μL slurry of anti-FLAG M2-agarose beads (Sigma) for 18 h at 4 °C. The beads were washed with lysis buffer, and FLAG-p53 was eluted with 200 μL of lysis buffer containing 200 μg/mL FLAG epitope peptide (Sigma) for 2 h at 4 °C. An aliquot of the eluted proteins was analyzed for expression levels by Western blot. Lysate containing an equal level of p53 was added to a 200-μL DNA binding reaction mixture and incubated at 4 °C for 30 min. The DNA binding reaction contains 25 nM (0.01 nmol) double-stranded biotinylated oligonucleotide DNA containing a consensus p53 binding site (biotin-5′TCGAGAGGCATGTCTAGGCATGTCTC annealed to 5′GAGACATGCCTAGACATGCCTCTCGA), 2 μg of poly(dI-dC), 5 mM DTT, 150 mM NaCl, 20 mM Tris⋅HCl (pH 7.2), and 4% glycerol. Mutant control oligonucleotide contains biotin-5′TCGAGAGGTCGCTCTAGGTCGCTCTC annealed to 5′GAGAGCGACCTAGAGCGACCTCTCGA. The DNA/protein complexes were captured with 0.1 mg of magnetic streptavidin beads (Promega) at 4 °C for 30 min. The beads were collected using a magnet and washed two times with DNA binding buffer. The bound p53 was eluted by boiling in Laemmli sample buffer and analyzed by SDS/PAGE and Western blot using FL393 antibody. The following peptides were tested in the p53 DNA affinity immunoblotting: 48–62 (DDIEQWFTEDPGPDE), SCR48–62 (PIDPDWETDGEQDFE), 75–89 (PAPAAPTPAAPAPAP), SCR75–89 (PPPPPPPTAAAAAAA), control1 (KPLQSTANNTPK), and control2 (HQNKSNLSSGLM).

ChIP.

ChIP was carried out using a standard procedure with Pab421 p53 monoclonal antibody. The samples were subjected to SYBR Green real-time PCR analysis using the following primers: p21 promoter (5′AGGAAGGGGATGGTAGGAGA and 5′ACACAAGCACACATGCATCA), MDM2 promoter (5′CGGGAGTTCAGGGTAAAGGT and 5′CCTTTTACTGCAGTTTCG), PUMA promoter (5′CTGTGGCCTTGTGTCTGTGAGTAC and 5′CCTAGCCCAAGGCAAGGAGGAC), Fas promoter (5′ACAGGAATTGAAGCGGAAGTCT and 5′GAGTTCCGCTCCTCTCTCCAA), LINE1 (5′CAGAATCTCTGGGACGCATT and 5′ATTGTGATGTTCGGGTGTCA), Alu (5′ACGAGGTCAGGAGATCGAGA and 5′CTCAGCCTCCCAAGTAGCTG), and α-satellite (5′TAGACAGAAATATTCTCACAATCGT and 5′GCCCTCAAAGCGCTCCAAG).

Western Blot.

Cells were lysed in lysis buffer and centrifuged for 10 min at 14,000 × g to remove the insoluble debris. Cell lysate (10–50 μg of protein) was boiled in Laemmli sample buffer, fractionated by SDS/PAGE, and transferred to Immobilon P filters (Millipore). The filter was blocked for 1 h with PBS containing 5% (wt/vol) nonfat dry milk and 0.1% Tween 20, incubated with primary and secondary antibodies, and developed using Supersignal reagent (Thermo Fisher Scientific). The following antibodies were used: FL393 (Santa Cruz Biotechnology) and DO-1 (BD Pharmingen) for p53, FLAG antibody (Sigma), and P21 antibody (BD Pharmingen).

Luciferase Fragment Complementation p53 DNA Binding Assay.

To construct ZF-Nluc, luciferase 1–437 was fused to the CT of a high-affinity six-finger ZF and cloned into pET28 vector (43). To construct p53-Cluc, luciferase 398–550 was fused to the CT of p53 and cloned into pET28. Zinc finger/p53 binding site (ZPBS) double-stranded oligonucleotide (5′CCGATGTAGGGAAAAGCCCGGGAACATGTCCCAACATGTTGAGC) contains a zinc finger binding site (5′ATGTAGGGAAAAGCCCGG, K_d = 50 pM) and p53 binding site from the p21 promoter (5′GAACATGTCCCAACATGTTG, K_d = 5 nM). BL21DE3 cells transformed with p53-Cluc or ZF-Nluc were cultured to OD₆₀₀ = 0.6 at 37 °C, supplemented with 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and 150 μM ZnCl₂, and cultured for 20 h at 16 °C. Pelleted E. coli cells were sonicated in lysis buffer [50 mM Hepes (pH 7.5), 150 mM NaCl, 0.1% Nonidet P-40, 5% glycerol, 10 μM ZnCl₂, 1 mM DTT] and centrifuged at 10,000 × g for 10 min at 4 °C. The lysate was diluted in dilution buffer [4.25% (vol/vol) 0.2 M NaH₂PO₄, 45.75% (vol/vol) 0.2 M Na₂HPO₄, 5% glycerol, 1 μM ZnCl₂, 1 mM DTT] to ∼10 ng/μL total protein. The diluted BL21DE3 extract (∼200 ng) containing ZF-Nluc was mixed with 20 ng of ZPBS oligo and 1% FBS and incubated at 23 °C for 15 min to allow ZF-Nluc prebinding. BL21DE3 extract (∼200 ng) containing p53-Cluc was added to the mixture and incubated for 45 min at 23 °C to assemble active complexes. Competitor oligonucleotide [5′CCGACGTCGGTAACAGTCCAGGAACATGTCCCAACATGTTGAGC, ZPBS with a mutated ZF binding site (2 µg)] was then added and incubated at 23 °C. Luciferase substrates were added at different time points after the addition of competitor for luciferase activity readout.

Protein Purification for NMR and ITC.

p53 NT 1–89 (72R) was expressed as previously described, except using pET47 and PreScission cleavage in place of pET28 and thrombin (33). The longer p53 ND (1–312) and p53 DBD (94–312) were expressed in BL21 cells grown in M9 media. The cells were cultured at 37 °C until OD₆₀₀ = 0.5, ZnCl₂ was added to 29 μM, and the cultures were cooled to 15 °C for 30 min. IPTG (1 mM) was added to induce expression for 18–20 h at 15 °C. Cells were pelleted and frozen at −80 °C until lysis. Cells were lysed in Tris-buffered saline (50 mM Tris⋅HCl, 500 mM NaCl, 1 mM DTT, 20 μM ZnCl₂) using a French press. The lysate was centrifuged, and the supernatant was applied to a GE GST-fast flow column and eluted with reduced glutathione. The eluate was cleaved with PreScission for 18 h, and the glutathione was removed by dialysis. The DBD sample was concentrated and run on a GE Superdex-75 column for final purification. The p53 ND construct was run through an anion exchange column using standard Tris buffer containing 20 μM ZnCl₂ before the final size exclusion chromatography step.

NMR Data Collection and Analysis.

NMR experiments were carried out at 25 °C on a Varian VNMRS 800-MHz spectrometer equipped with a triple-resonance pulse field z-axis gradient cold probe. Sensitivity-enhanced ¹H-¹⁵N HSQ; ¹H-¹⁵N heteronuclear single-quantum coherence-transverse relaxation optimized spectroscopy (HSQC-TROSY); and 3D HNCACB, HNCA-TROSY, and HNCO experiments were performed on the uniformly ¹⁵N- and ¹³C-labeled samples in 90% H₂O/10% D₂O. Unless otherwise stated, the buffer was 10 mM phosphate buffer with 66 mM NaCl, 4 mM DTT, and 0.02% NaN₃ at pH 6.8. For the HNCACB experiment, data were acquired in ¹H, ¹³C, and ¹⁵N dimensions using 9,689.9-Hz (t₃) × 14,074.9-Hz (t₂) × 2,754.4-Hz (t₁) sweep widths and 1,024 (t₃) × 128 (t₂) × 32 (t₁) complex data points. For the HNCO experiment, the sweep widths were 9,689.9 Hz (t₃) × 2,010.4 Hz (t₂) × 2,754.4 Hz (t₁), and the complex data points were 1,024 (t₃) × 128 (t₂) × 32 (t₁). The HNCA-TROSY was collected with the following sweep widths, 9,689.9-Hz (t₃) × 6,031.9-Hz (t₂) × 2,754.4-Hz (t₁), and complex points, 1,024 (t₃) × 128 (t₂) × 32 (t₁) (44–47). The sweep widths and complex points of the HSQC experiment were 9,689.9 Hz (t₂) × 2,754.4 Hz (t₁) and 1,024 (t₂) × 128 (t₁), respectively. Relaxation experiments were collected using at least 10 time points from 10 ms up to 250 ms and fit to an exponential decay function using NMRViewJ software (48).

All 3D experiments were conducted with 140–200 μM samples. The ¹⁵N-labeled p53 NT relaxation data were collected at 250 μM, and the p53 NT with 4× p53 DBD experiment was conducted with 50 μM labeled NT titrated with 200 μM unlabeled p53 DBD. The ¹⁵N-labeled p53 DBD to unlabeled p53 NT data were collected with 111 μM p53 DBD titrated with up to 666 μM NT. The relaxation data for the p53 ND sample were collected at 200 μM labeled ND. For the p53 ND HNCACB and HNCO experiments, only residues in the disordered p53 NT and toward the CT were assignable, resulting in 83 Ca, 82 Cb, 66 CO, and 67 HN and N assignments to the first 86 residues. The p53 DBD HSQC-TROSY and HNCA-TROSY assignments were made by comparing chemical shift values with those published by Kriwacki and coworkers (49).

All NMR spectra were processed with the NMRFx program and analyzed using NMRViewJ software (50). Apodization was achieved in the ¹H, ¹³C, and ¹⁵N dimensions using a squared sine bell function shifted by 70°. Apodization was followed by zero-filling to double the number of real data points, and linear prediction was used in the ¹⁵N dimension. The ¹H carrier frequency was set on the water peak, and 4.753 ppm was used as the reference frequency. Secondary chemical shift values were calculated by subtracting the residue-specific random coil chemical shifts in the neighbor-corrected intrinsically disordered protein (IDP) chemical shift library (ncIDP) from the measured chemical shifts (51). Secondary structure populations were calculated with d2D using the measured proton, nitrogen, α- and β-carbon, and carbonyl carbon chemical shifts (52).

ITC.

p53 polypeptides and DNA were codialyzed into 10 mM NaPO₄, 66 mM NaCl, 0.02% sodium azide, and 8 mM beta mecaptoethanol at pH 6.8. For the p53 DBD constructs, three replicate titrations were conducted with 5 μM p53 DBD in the cell and 12.5 μM DNA in the syringe using 10-μL injections utilizing a MicroCal-VP-ITC system at 25 °C. For the p53 ND constructs, two to three titrations were conducted using a MicroCal-ITC 200 system with 50 μM ND in the cell and 125 μM DNA in the syringe using 15-μL injections at 25 °C. The corrected heat values were fit using Microcal Origin software (7.0) in a nonlinear least square curve-fitting algorithm yielding the stoichiometry, enthalpy, and affinity constants reported for a single binding site. The following double-stranded DNA oligonucleotides were used: consensus 5′AGACATGCCTAGGACATGCCT and scrambled 5′TGCCGATCAAAACCGATTCG (53).