Mutant p53 proteins bind DNA in a DNA structure-selective mode

Abstract

Despite the loss of sequence-specific DNA binding, mutant p53 (mutp53) proteins can induce or repress transcription of mutp53-specific target genes. To date, the molecular basis for transcriptional modulation by mutp53 is not understood, but increasing evidence points to the possibility that specific interactions of mutp53 with DNA play an important role. So far, the lack of a common denominator for mutp53 DNA binding, i.e. the existence of common sequence elements, has hampered further characterization of mutp53 DNA binding. Emanating from our previous discovery that DNA structure is an important determinant of wild-type p53 (wtp53) DNA binding, we analyzed the binding of various mutp53 proteins to oligonucleotides mimicking non-B DNA structures. Using various DNA-binding assays we show that mutp53 proteins bind selectively and with high affinity to non-B DNA. In contrast to sequence-specific and DNA structure-dependent binding of wtp53, mutp53 DNA binding to non-B DNA is solely dependent on the stereo-specific configuration of the DNA, and not on DNA sequence. We propose that DNA structure-selective binding of mutp53 proteins is the basis for the well-documented interaction of mutp53 with MAR elements and for transcriptional activities mediates by mutp53.

Figure 1

Analysis of mutp53 DNA binding by EMSA. (A) Hot-spot p53 mutp53 proteins 273H, 245S and 248P were incubated with linear DNA_unspec lacking a p53-specific cognate motif (upper panel). Binding of mutp53 proteins 273H and 248P to a four-way junction structure (lower panel). Four-way junction structure was prepared as described in Materials and Methods by annealing four intercomplementary oligonucleotides b, h, r and x (the resulting structure is shown schematically), of which one was radioactively labeled. Arrowheads indicate p53 complexes formed with different types of DNA probes, as shown by the corresponding symbols in this and in other figures. (B) Binding of wtp53 and 245P proteins to a four-way junction structure.

Figure 2

(A) Binding of mutp53 proteins 273H, 245S and 248P to a stem–loop structure formed by unspecific DNA (stem–loop DNA_unspec). (B) Effects of p53-specific antibodies PAb421 and DO-1 on mutp53 interaction with non-linear DNA structures. An aliquot of 50 ng of p53 proteins were incubated with DNA in the presence or absence of 200 ng of the purified antibody.

Figure 3

Impact of the p53-specific sequence-motif and of mismatched bases on DNA binding of mutp53. (A) Stem–loop structures either containing (stem–loop DNA_spec) or lacking (stem–loop DNA_unspec) a p53-binding site from the p21 promoter were incubated with mutp53 proteins in the presence or the absence of PAb421. (B) Binding of 248P mutant to stem–loop structures lacking a p53-specific cognate motif and differing in the composition of non-matching bases within the stem (encircled bases). (C) Mutp53 and wtp53 bind differently to stem–loop structure lacking mismatched bases in the stem.

Figure 3

Impact of the p53-specific sequence-motif and of mismatched bases on DNA binding of mutp53. (A) Stem–loop structures either containing (stem–loop DNA_spec) or lacking (stem–loop DNA_unspec) a p53-binding site from the p21 promoter were incubated with mutp53 proteins in the presence or the absence of PAb421. (B) Binding of 248P mutant to stem–loop structures lacking a p53-specific cognate motif and differing in the composition of non-matching bases within the stem (encircled bases). (C) Mutp53 and wtp53 bind differently to stem–loop structure lacking mismatched bases in the stem.

Figure 4

Graphic representation showing typical results of mut p53 DNA-binding analyses using cFLA (shown for 273H and 248P). The results of all the cFLA experiments are summarized in Table 2.

Figure 5

Analyses of DNA binding by using mutp53 protein arrays. (A) Autoradiograph of radioactively labeled DNA_spec bound to p53 proteins. Mutp53 arrays were incubated with DNA_spec present either in linear or in stem–loop conformation, washed and exposed to X-ray film for autoradiographic detection, amplified with the aid of an intensifying screen at −70°C. Positions of wtp53 and mutp53 proteins 273H and 245S that were analyzed also by other assays are indicated by circles. Dotted lines indicate exclusive binding of stem–loop DNA_spec by those mutants that failed to bind linear DNA_spec. The results of binding experiments are summarized in Table 3, in which different mutp53 proteins have been grouped according to their ability to bind non-linear DNA. The apparent binding affinity was evaluated by densitometry whereby the intensity of the signal produced by p53-306X mutant lacking the oligomerization and the C-terminal domains was considered as background (designated as ‘−’ in Table 3). To score the binding to linear DNA, intensities that were equal to or consisted at least 50% of the value corresponding to wtp53 binding were designated as ‘++’. For stem–loop DNA binding, intensities higher than at least 10% of the value obtained for the same protein with linear DNA were designated as ‘+++’. (B) DNA competition assay. Arrays were first incubated with linear DNA_spec (upper panel) or with stem–loop DNA_spec (lower panel), and bound DNA was detected using autoradiography. The much weaker signal compared with images shown in (A) is due to more gentle conditions of autoradiography (shorter exposure time at room temperature under moist conditions) that were used for the sake of preserving protein–DNA complex. After the documentation of DNA binding, the p53–DNA complex were challenged by two sequential rounds of competition with a 1:1 and 1:5 molar excess of unlabeled linear DNA_spec. The lower panel shows a control array that was treated under identical conditions except that the competitor DNA was omitted.

Figure 6

Non-linear DNA binding protects mutp53 from Mdm2-mediated ubiquitination. In vitro ubiquitination of recombinant p53 proteins in the absence (lane 2) or in the presence of DNA_spec in stem–loop or in linear conformation as indicated. Lane 1 show control samples that were treated under identical conditions as samples in lane 2 except that ubiquitin was omitted from the reaction mixture.