Recognition of RNA by the p53 tumor suppressor protein in the yeast three-hybrid system

Abstract

The p53 tumor suppressor protein is a homotetrameric transcription factor whose gene is mutated in nearly half of all human cancers. In an unrelated screen of RNA/protein interactions using the yeast three-hybrid system, we inadvertently detected p53 interactions with several different RNAs. A literature review revealed previous reports of both sequence-specific and -nonspecific interactions between p53 and RNA. Using yeast three-hybrid selections to identify preferred RNA partners for p53, we failed to identify primary RNA sequences or obvious secondary structures required for p53 binding. The cationic p53 C-terminus was shown to be required for RNA binding in yeast. We show that while p53 strongly discriminates between certain RNAs in the yeast three-hybrid assay, the same RNAs are bound equally by p53 in vitro. We further show that the p53 RNA-binding preferences in yeast are mirrored almost exactly by a recombinant tetrameric form of the HIV-1 nucleocapsid (NC) protein thought to be a sequence-nonspecific RNA-binding protein. However, the possibility of specific RNA binding by p53 could not be ruled out because p53 and HIV-1 NC displayed certain differences in RNA-binding preference. We conclude that (1) p53 binds RNA in vivo, (2) RNA binding by p53 is largely sequence-nonspecific in the yeast nucleus, (3) some structure-specific RNA binding by p53 cannot be ruled out, and (4) caution is required when interpreting results of RNA screens in the yeast threehybrid system because sequence-dependent differences in RNA folding and display can masquerade as sequence-dependent differences in protein recognition.

Analysis of p53–RNA interactions in cultured human cells

“…Perhaps p53 posttranslational modifications prevent p53-RNA interactions in vivo. Together with other recent results [6,7], this work illuminates a historically mysterious aspect of p53 biology. The physiological significance of RNA binding by p53 remains obscure [4].…”

“…This RNA contamination can easily be misinterpreted as evidence for specific p53-RNA interactions. Whereas unmodified or partially post-translationally modified p53 proteins from E. coli [19], yeast [6], insect cells [7,13,[16][17][18], or reticulocyte lysate [16][17][18] are prone to nonspecific RNA binding, p53 protein is not complexed with RNA when purified from the tested human cells. Perhaps p53 posttranslational modifications prevent p53-RNA interactions in vivo.…”

“…1A; [4]). Although many RNAbinding proteins are characterized by RNA-recognition motifs (RRMs; [5]), the interaction between p53 and RNA involves the C-terminal 30 amino acids of p53 [6,7], which do not form a RRM but are disordered in solution [7,8] and do not reorganize upon RNA binding [7,9]. p53, like the HIV-1 nucleocapsid protein, has been termed an RNA chaperone [10] that may facilitate RNA folding [11] without ATP hydrolysis [11][12][13].…”

“…Similarly, Galy et al reported that recombinant p53 can bind the 5'-UTR of the FGF-2 mRNA [18]. In contrast, reports of p53 RNA-RNA annealing activity [13] and sequence-nonspecific RNA binding [4,6,7,19] suggested that p53 binds RNA promiscuously. Acetylation of four p53 C-terminal lysine residues eliminates detectable RNA binding in vitro [7], as it does for sequence-nonspecific DNA binding [8].…”

“…Our previous studies of RNA binding by p53 in yeast and in vitro demonstrated that p53 with incomplete post-translational modifications binds RNA without apparent sequence-or structure-specificity [4,6,7]. It is possible that a partially-modified form of p53 could exist in vivo with one or more specific RNA partners.…”