ABSTRACT N 6 ‐methyladenosine (m 6 A) modification constitutes a crucial layer of post‐transcriptional regulations, but the landscape of its downstream readout effects remains less comprehensively understood. Therefore, we systematically assess the readout effects of m 6 A on mRNA half‐life, translation efficiency, and alternative splicing across five cell lines (A549, HEK293T, HUVEC, JURKAT, and human embryonic stem cells (hESCs)) using actinomycin D‐disrupted temporal transcriptome, ribosome sequencing, and ultra‐high‐depth transcriptome sequencing, respectively. Our analysis, coupled with the integration of public and newly profiled m 6 A methylome data, reveals high cell type specificity in m 6 A readouts where m 6 A level alone is insufficient to predict m 6 A readouts. Nonetheless, machine learning models focusing on RNA‐binding protein (RBP) binding context can effectively predict the readouts and prioritize four novel m 6 A‐associated proteins (FUBP3, FXR2, L1TD1, and DDX6). Their m 6 A‐binding ability is validated by m 6 A RNA pull‐down, transcriptome‐wide binding site mapping, and electrophoretic mobility shift assay, while FUBP3 and L1TD1 are further suggested as m 6 A readers regulating mRNA stability based on half‐life profiling of knockout cells. Finally, FUBP3, FXR2, and L1TD1 are demonstrated to regulate hESC differentiation without affecting self‐renewal. Together, this study bridges the gap in understanding m 6 A functional readouts and lays the groundwork for future research on m 6 A‐mediated stem cell fate decisions.
Huang et al. (Tue,) studied this question.