Comprehensive profiling of N ⁶ -methyladnosine (m ⁶ A) readouts reveals novel m ⁶ A readers that regulate human embryonic stem cell differentiation

N ⁶ -methyladenosine (m ⁶ A) methylation has emerged as a prevalent RNA modification that extensively impacts various physiological and pathological processes via various post-transcriptional readout effects in mammals. High-throughput methylome profiling has outlined the landscape of m ⁶ A modification sites, but their downstream readouts require comprehensive investigation. To this end, we systematically assessed the 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 sequencing, ribosome sequencing, and ultra-high-depth transcriptome sequencing, respectively. Our analysis, coupled with the integration of public and re-profiled m ⁶ A methylome data, revealed high cell type specificity in m ⁶ A readouts where m ⁶ A level alone is insufficient to predict m ⁶ A readouts. Machine learning models focused on the RNA binding protein (RBP) binding context of m ⁶ A sites demonstrated substantial predictive ability of m ⁶ A readouts while prioritizing putative m ⁶ A readers from their informative RBP features. Four novel m ⁶ A readers (DDX6, FUBP3, FXR2, and L1TD1) were identified and validated through m ⁶ A RNA pull-down assays and transcriptome-wide RBP binding site mapping. Notably, FUBP3, FXR2 and L1TD1 were found to regulate hESC differentiation without impairing self-renewal, underscoring their critical roles in stem cell biology. 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.