A neural m6A/YTHDF pathway is required for learning and memory in Drosophila

The roles of epitranscriptomic modifications in mRNA regulation have recently received substantial attention, with appreciation growing for their phenotypically selective impacts within the animal. We adopted Drosophila melanogaster as a model system to study m6A, the most abundant internal modification of mRNA. Here, we report proteomic and functional analyses of fly m6A-binding proteins, confirming nuclear (YTHDC) and cytoplasmic (YTHDF) YTH domain proteins as the major m6A binders. Since all core m6A pathway mutants are viable, we assessed in vivo requirements of the m6A pathway in cognitive processes. Assays of short term memory revealed an age-dependent requirement of m6A writers working via YTHDF, but not YTHDC, comprising the first phenotypes assigned to Drosophila mutants of the cytoplasmic m6A reader. These factors promote memory via neural-autonomous activities, and are required in the mushroom body, the center for associative learning. To inform their basis, we mapped m6A from wild-type and mettl3 null mutant heads, allowing robust discrimination of Mettl3-dependent m6A sites. In contrast to mammalian m6A, which is predominant in 3’ UTRs, Drosophila m6A is highly enriched in 5’ UTRs and occurs in an adenosine-rich context. Genomic analyses demonstrate that Drosophila m6A does not directionally affect RNA stability, but is preferentially deposited on genes with low translational efficiency. However, functional tests indicate a role for m6A in translational activation, since we observe reduced nascent protein synthesis in mettl3-KO cells. Finally, we show that ectopic YTHDF can increase m6A target reporter output in an m6A-binding dependent manner, and that this activity is required for in vivo neural function of YTHDF in memory. Altogether, we provide the first tissue-specific m6A maps in this model organism and reveal selective behavioral and translational defects for m6A/YTHDF mutants.