Identification of new p53 target microRNAs by bioinformatics and functional analysis

Abstract

Background: The tumor suppressor p53 is a sequence-specific transcription factor that regulates an extensive network of coding genes, long non-coding RNAs and microRNAs, that establish intricate gene regulatory circuits influencing many cellular responses beyond the prototypical control of cell cycle, apoptosis and DNA repair.

Methods: Using bioinformatic approaches, we identified an additional group of candidate microRNAs (miRs) under direct p53 transcriptional control. To validate p53 family-mediated responsiveness of the newly predicted target miRs we first evaluated the potential for wild type p53, p63β and p73β to transactivate from p53 response elements (REs) mapped in the miR promoters, using an established yeast-based assay.

Results: The REs found in miR-10b, -23b, -106a, -151a, -191, -198, -202, -221, -320, -1204, -1206 promoters were responsive to p53 and 8 of them were also responsive to p63β or p73β. The potential for germline p53 mutations to drive transactivation at selected miR-associated REs was also examined. Chromatin Immuno-Precipitation (ChIP) assays conducted in doxorubicin-treated MCF7 cells and HCT116 p53+/+ revealed moderate induction of p53 occupancy at the miR-202, -1204, -1206, -10b RE-containing sites, while weak occupancy was observed for the miR-23b-associated RE only in MCF7 cells. RT-qPCR analyses cells showed modest doxorubicin- and/or Nutlin-dependent induction of the levels of mature miR-10b, -23b, -151a in HCT116 p53+/+ and MCF7 cells. The long noncoding RNA PVT1 comprising miR-1204 and -1206 was weakly induced only in HCT116 p53+/+ cells, but the mature miRs were not detected. miR-202 expression was not influenced by p53-activating stimuli in our cell systems.

Conclusions: Our study reveals additional miRs, particularly miR-10b and miR-151a, that could be directly regulated by the p53-family of transcription factors and contribute to the tuning of p53-induced responses.

Figure 1

p53 family members can transactivate p53-REs found in miR-associated promoter regions. A-B) Transactivation potential of p53 protein tested on a panel of 15 putative p53-dependent miR-REs and miR-34a-RE (a positive control) using the yeast functional assay. The expression of p53 was modulated by increasing concentrations of galactose in the culture medium (A, 0.008%: a moderate p53 expression; B, 0.128%: high p53 expression). C-D) The same panel of p53 miR-REs was tested using the other members of the p53 family p63 and p73 (gray and black bars respectively). The expression of p63β and p73β isoforms was induced using two different concentrations of galactose (C, 0.008%: moderate expression; D, 1%: maximal induction). Results are presented as fold of induction calculated over an empty expression vector. Bars plot the averages and standard deviations of at least three independent biological repeats. (E) Western Blot establishing galactose-dependent expression of p53, p63 or p73 in yeast. PGK1 was used as reference.

Figure 2

p53-REs found in miR promoters can be used to classify p53 germline alleles associated with Li-Fraumeni Syndrome. Five p53 missense mutants representative of partial function (A138S, C141Y, R337C) or loss-of-function (A138P, R175H) germline p53 alleles were tested for transactivation from 5 p53 miR-REs using the yeast functional assay. Yeast transformants were cultured over-night in selective medium and the luciferase activity was measured using the miniaturized assay format [43]. Bars represent the averages and standard deviations of at least three independent biological repeats.

Figure 3

p53 can bind chromatin region surrounding the identified p53 REs in miR genes. A) ChIP assays were performed in HCT116 p53^+/+(gray bars) and HCT116 p53^−/− (black bars) upon doxorubicin treatment for 24 hours. The results of Real-Time qPCR are presented as fold of mock treatment, normalized with respect to the signal obtained with Input DNA. The results from three control locations corresponding to promoter regions of Actin, GAPDH and exon 9 of CCNB1 genes were averaged and are also presented in panel A (Neg. Ctrls). P21 and miR-34a occupancy were measured as positive controls. Bars represent average and standard deviations of three independent experiments. C) ChIP assays of MCF7 cells treated with doxorubicin for 24 hours. Results obtained after ChIP with an antibody against IgG were included as a negative control for the p53 miR-REs. Examples of agarose gel analysis of standard ChIP-PCR are given in panels B and D. Specifically, panel B shows experiments performed in HCT116 p53^+/+ and p53^−/− cells, while panel D presents results from MCF7 cells. The DO-1 p53 antibody was used for immunoprecipitation; NTC, no template control. Regions surrounding the established P21 and miR-34a p53 REs were examined as positive controls.

Figure 4

p53-induced expression of mature and pre-miR genes. RT-qPCR were performed in HCT116 p53^+/+(black bars), HCT116 p53^−/− (white bars) and MCF7 (gray bars) cells upon doxorubicin (A, B) or Nutlin (C) for 24 hours. The expression of the processed mature miR (A) or of the pre-miR RNA (B, C) was tested. p21 mRNA expression was measured as p53-dependent positive control. PVT1 non-coding RNA expression levels were measured as additional evidence of p53-dependent expression of miR-1204 and miR-1206. Results are presented as fold of induction with respect to the mock condition. Bars plot average and standard deviations of three independent experiments. (D) Western Blot establishing stabilization of p53 protein in doxorubicin and Nutlin treated cells and the induction of the p53 target gene p21. GAPDH was used [6] as reference.