p53, transcriptional repression and drug sensitivity

Upon exposure to multiple different types of cellular stress, including DNA, damage, hypoxia, and oncogene activation, the p53 tumor suppressor protein becomes activated as a transcription factor. It then functions as either a transcriptional activator or a transcriptional repressor of gene expression in order to induce apoptosis, growth arrest, or senescence. To transactivate gene expression p53 binds in the upstream regulatory sequences of a gene to a consensus DNA binding motif of two repeats of the 10 base pair element 5′-PuPuPuC(A/T)(T/A)GpyPyPy-3′, where Pu is a purine residue and Py is pyrimidine; these repeats can be separated by 0–13 nucleotides.¹

That p53 also functions as a transcriptional repressor has been known for close to 20 y.^2,3 However, because many powerful transcription factors can commandeer components of the basal transcriptional machinery (so-called transcriptional ‘squelching’), they also possess potent non-specific transcriptional repression activity. This makes it notoriously difficult to map necessary and sufficient p53 binding sites on repressed promoters. An added complication is the fact that p53 induces cell cycle arrest and apoptosis, making it difficult to separate these outcomes from transcriptional repression.

In the present study by Meek and colleagues, the authors perform a meticulous study documenting the sequence-specific transcriptional repression by p53 of the gene encoding Polo-like kinase 1 (PLK1).⁴ PLK1 is a member of a family of serine-threonine kinases that control mitotic progression and the DNA damage-induced G₂/M checkpoint. Because PLK1 belongs to a class of genes that are repressed in the G₁ phase of the cell cycle by virtue of an element in their promoters termed a CDE/CHR (cell cycle dependent element/cell cycle genes homology region), the authors use several means to eliminate p53’s ability to cell cycle arrest from its ability to repress PLK1. Notable among these is their finding that p53 induction can still repress PLK1 in cells in which the p21 gene, a critical mediator of p53-mediated cell cycle arrest, is silenced. The authors then use chromatin immunoprecipitation to identify the sites for p53 binding, and they map these to two canonical p53 consensus elements located approximately 800, and 200, nucleotides upstream of the transcription start site.

The study by Meek and colleagues echoes some findings reported previously for p53-mediated repression of the Cdc25c gene.⁵ Like PLK1, Cdc25c is a critical mitotic checkpoint gene that possesses a CDE/CHR element in its promoter. Like Cdc25c, p53 binds to a consensus element in the PLK1 promoter that resembles a canonical p53 response element—two copies of the 10 base pair element 5′-PuPuPuC(A/T)(T/A)GpyPyPy-3′. How p53 represses transcription from this element remains a burgeoning question in the field. One answer may come from the studies of Meek and Manfredi; in both cases the authors discovered that the p53 consensus element overlapped with a binding site for another transcription factor: one that played a major role in the activated expression of these genes.^4,5 Therefore, p53 may repress the transcription of these genes by interfering with the activity of another factor.

The future of p53, repression and the G₂/M checkpoint seems clear. Which other genes that function in the G₂/M checkpoint and contain p53 consensus binding sites belong to this ‘class’ of genes? One intriguing possibility is Aurora Kinase, which like Cdc25c and PLK1 contains a CDE/CHR element, and is expressed at abnormally high levels in cells with inactive p53.⁶ Another question relates to the differences between p53 activating elements and p53 repressing elements; are they identical or are there subtle differences that we are unaware of? Computational analyses suggest that p53 activating elements typically have spacers of 0–1 nucleotides between the dimer binding sites, while repressed genes more often have spacers of five nucleotides or greater;¹ this might be predicted to place p53 dimers on different faces of the DNA helix, and thereby alter their protein-protein interaction partners. Finally, the clinical relevance of this line of research is important. Inhibitors of Aurora kinase, as well as PLK1, are actively in use in clinical trials for cancer. These genes are typically repressed in non-cycling (normal) cells, and overexpressed in tumor cells with mutant p53; therefore, identifying other G₂/M kinases that might be repressed by p53, and using cocktails of such inhibitors of such on tumors with mutant p53, is likely to represent a productive area for cancer research.

McKenzie L, King S, Marcar L, Nicol S, Dias SS, Schumm K, et al. p53-dependent repression of polo-like kinase-1 (PLK1) Cell Cycle. 2010;9:4200–12. doi: 10.4161/cc.9.20.13532.