Search

Genome-wide prediction of transcription factor binding sites is notoriously difficult. We have developed and applied a logistic regression approach for prediction of binding sites for the p53 transcription factor that incorporates sequence information and chromatin modification data. We tested this by comparison of predicted sites with known binding sites defined by chromatin immunoprecipitation (ChIP), by the location of predictions relative to genes, by the function of nearby genes and by analysis of gene expression data after p53 activation. We compared the predictions made by our novel model with predictions based only on matches to a sequence position weight matrix (PWM). In whole genome assays, the fraction of known sites identified by the two models was similar, suggesting that there was little to be gained from including chromatin modification data. In contrast, there were highly significant and biologically relevant differences between the two models in the location of the predicted binding sites relative to genes, in the function of nearby genes and in the responsiveness of nearby genes to p53 activation. We propose that these contradictory results can be explained by PWM and ChIP data reflecting primarily biophysical properties of protein–DNA interactions, whereas chromatin modification data capture biologically important functional information. Publisher PDF

Several reports have noted epidemiological differences in the prevalence or prognostic significance of p53 mutants with arginine (R) or proline (P) at the codon 72 polymorphism (R72/P72) in certain cancer types, but the biological significance of these variants is unclear. The ability of p53 mutants to interact with and inactivate the p53 homolog p73 was recently reported to depend on the conformational state of the p53 protein and the residue at codon 72. Since the conformation of p53 mutants may influence their ability to transdominantly inhibit wild-type p53, we tested whether there was a correlation between the amino acid at codon 72 and the transdominance of p53 alleles found in tumors. The transdominance test was performed using a simple yeast transcription assay, and the amino acid at codon 72 was determined by sequencing. A total of 100 p53 mutants were tested. Compared with the germline frequency (R:P = 427:297), an extreme bias in favor of the R72 allele was observed with recessive mutants (R:P = 50:7, P < 0.0002), whereas no selection for the R72 allele was seen with transdominant mutants (R:P = 23:20). p53 and p73 are known to transactivate overlapping sets of target genes. We interpret the R72 bias with recessive mutants as evidence that decreased activation of p53 target genes provides a selective growth advantage to tumor cells during the stage of tumorigenesis in which a wild-type and mutant p53 allele coexist. We suggest that transdominant p53 mutants achieve this by inactivation of the remaining wild-type p53 allele, whereas recessive p53 mutants achieve it through inactivation of p73.

Analysis of families with germline p53 mutations shows that the mutant p53 allele behaves as a dominant oncogene at the genetic level, although it behaves as a recessive oncogene at the cellular level, since tumours invariably show mutation or loss of both wild-type alleles. At the biochemical level it is possible that some clinically important mutant p53 proteins may be carcinogenic through a dominant mechanism. We show that p53 mutants can be readily classified according to their dominant potential using a simple yeast functional assay. Wild-type p53 is constitutively expressed from a TRP1 vector, p53 mutants are expressed from an otherwise identical LEU2 vector and net transcriptional activity is scored using an ADE2-based reporter. Twenty seven p53 mutants were tested: 19 were recessive, i.e. gave white colonies, and eight showed dominant activity, i.e. gave pink/red colonies. This simple assay should facilitate studies on p53 dominance.