br In corroboration with the above
In corroboration with the above-described body of results, an in-verse correlation of pp53 expression to m6A modification with METTL3 was comparably observed in TP53-Dox cells upon immunofluorescence staining. Increased levels of pp53 displayed prominently in cell nuclei (green, pp53-Alexa Fluor 488) of TP53-Dox cells treated with PDMP (GCS inhibition), with decreased METTL3 protein levels (red, METTL3-Qdot 605), and likewise when these cells were alternatively treated with NPC (inhibition of m6A methylation), in both cases comparing to TP53-Dox cells treated with vehicle (Fig. 3C). Furthermore, we identi-fied p53 protein after immunoprecipitation (IP) with anti-p53 or anti-pp53 antibodies. Western blotting after IP showed that levels of wt p53 (pp53) were suppressed in TP53-Dox cells as compared to SW48-Dox, and PDMP treatments restored the pp53 levels in TP53-Dox cells (Fig. 3D). Based on two peptides (amino acids 102–110: TYQGSYGFR; 5-Azacytidine 307–319: ALPNNTSSPQPK), tandem mass spectrometry (MS/MS) analysis of proteins resolved by SDS-PAGE identified p53 protein expressed in SW48-Dox and TP53-Dox cells treated with PDMP, but not TP53-Dox in the absence of PDMP treatment (Fig. 3D, E).
In human cells, METTL3 is an essential methyltransferase for the methylation of adenosines to generate m6As in RNA [29,30]. As an-ticipated, siRNA silencing of METTL3 expression reduced METTL3 protein levels to 20% vs. vehicle treatment in TP53-Dox cells, which, interestingly, concomitantly enhanced the pp53 levels 5-fold over the levels seen for vehicle control treatment (Fig. 4A). These results clearly Biochemical Pharmacology 160 (2019) 134–145
implicate m6A methylation catalyzed by METTL3 (Fig. 4B) as being involved in regulating the expression of mut-p53 vs. wt-p53 protein in cells heterozygously carrying the p53 R273H mutation. Perhaps more noteworthy, however, was that PDMP treatments also reduced METTL3 protein levels, to 20% vs. vehicle, once again concomitantly enhancing pp53 protein levels by more than 5-fold (Fig. 4A).
To directly identify the presence of m6A in R273H pre-mRNA, we precipitated RNA transcripts with anti-m6A antibody, and then ana-lyzed a region of p53 mRNA that included the segment coding for R273 (p53-c273, 95 bp, codons 249–280) using quantitative RT-PCR (MeRIP qRT-PCR). The levels of m6A p53-c273 in TP53-Dox and in WiDr cells were 16-fold (330 vs. 20 pg/µg RNA) and 19-fold (376 vs. 20 pg/µg RNA) higher, respectively, than in SW48-Dox cells (Fig. 4C, D). These results directly evidenced that the adenosine transited from guanosine at codon 273 of the mutant pre-mRNA (-CAU-), due to DNA point mutation (Fig. 1A), was extensively converted to m6A by METTL3 (Fig. 4B). NPC treatments significantly decreased m6A p53-c273 levels, by 2-fold (140 vs. 329 pg/µg RNA), in TP53-Dox cells, as compared to vehicle control (Fig. 4C, D). Perhaps more interestingly, PDMP treat-ments also decreased m6A p53-c273 levels, and to a greater degree (by 9-fold: 36 vs. 329 pg/µg RNA), in TP53-Dox cells, as compared to ve-hicle. Altogether, current work directly pinpoints that m6A production, which only appears in the adenosine-transited codon of the R273H-mutant p53 pre-mRNA, positively regulates missense protein expression in cancer cells carrying p53 R273H mutations. This result also implies that inhibition of METTL3-catalyzed RNA methylation of R273H-mu-tant might be an effective approach for restoring wt p53 expression in cells heterozygously carrying R273H.
3.3. Glycosphingolipid Gb3 mediates signaling of cSrc and β-catenin to increase METTL3 expression and RNA m6A methylation To explore how GSLs modulate m6A methylation and protein ex-pression of the p53 R273H mutant, we assessed and characterized the effects of ceramide, Gb3, cSrc and β-catenin on METTL3 expression in TP53-Dox cells. GCS, a rate-limiting enzyme in GSL synthesis, converts ceramide to glucosylceramide (GlcCer), which serves as a precursor for production of further-elaborated GSLs, including globo-series GSLs and gangliosides, upon serial glycosylation in cells. Previous works indicate that certain globo-series GSLs (in particular, globotriaosylceramide Gb3) play important roles in modulating transcription of gene expres-sion via cSrc and β-catenin signaling pathways [36,37]. It has been elucidated that the active proto-oncogene protein cSrc elevates the le-vels of β-catenin, which in turn upregulates the expression of cyclin D1, c-Myc, MDR1 and fibroblast growth factor 2 (FGF2) [36,37,44]. PDMP treatment, which inhibits GCS activity and ceramide glycosylation , reduced levels of phosphorylated cSrc (pcSrc), β-catenin and METTL3, and thence restored pp53 levels (Fig. 5A, B); however, fumonisin B1 (FB1), an inhibitor of ceramide synthase [27,45], could not affect the levels of these proteins (inclusive of pp53) in TP53-Dox cells (Fig. 5A, B). These results indicate that GSLs, rather than ceramides, modulate METTL3 expression.