LAUSR

Dear Editor, While the eukaryotic candidate mA methyltransferases belong to multiple distinct methylase lineages, the most widespread group belongs to the MT-A70 family exemplified by the yeast messenger RNA (mRNA) adenine methylase complex Ime4/Kar4. At the structural level, all of these enzymes are characterized by a 7-β-strand methyltransferase domain at their C terminus, fused to a predicted α-helical domain at their N terminus and require S-adenosyl-L-methionine (SAM) as a methyl donor. The catalytic motif, [DSH]PP[YFW], present in many members of this family, has shown to be critical for METTL3/METTL4-mediated mRNA mA methylation. The high degree of amino acid sequence conservation among the predicted N6-methyladenosine methyltransferases motivates further explorations into their potential functional conservation. METTL4 is a member of the MT-A70-like protein family, which is conserved during evolution (Fig. 1a). Previous studies suggested that METTL4 regulates DNA 6mA in vivo and therefore is a candidate DNA 6mA methyltransferase. However, the enzymatic activity of METTL4 in vitro has not been demonstrated. To identify the substrate(s) for METTL4, we purified His-tagged, wild-type (WT) as well as a catalytic mutant (DPPW mutated to NPPW) (Supplementary Fig. S1) of Drosophila melanogaster mettl4 from Escherichia coli strain BL21 (DE3). In order to unbiasedly identify potential substrates of mettl4, we performed in vitro enzymatic assays using various substrates, including both DNA and RNA with and without secondary structures. We used deuterated S-adenosyl methionine (SAM-d3) in the in vitro enzymatic assays in order to identify mA mediated by mettl4. Although we detected a weak enzymatic activity on DNA substrates composed of previously published sequence motifs, mettl4 appears to prefer RNA substrates potentially with secondary structures (Supplementary Fig. S2). We next performed enhanced crosslinking and immunoprecipitation (IP) followed by high-throughput sequencing (eCLIP-seq), which was originally developed to map binding sites of RNA-binding proteins on their target RNAs, to identify the RNA type that is targeted by mettl4 in vivo. Since there are no commercial antibodies available for fly mettl4, we generated a Drosophila Kc cell line with a FLAG-tagged mettl4 for the eCLIP-seq experiment. In total, we generated two biological replicates for IP samples, and their respective input samples, together with one IP-control and inputcontrol sample for the quality control and enrichment analysis. The two replicates showed a strong correlation with a Spearman’s correlation coefficient of 0.97, indicating great consistency between the replicates (Supplementary Fig. S3). Thus, we merged the two replicates to increase the sequencing depth and power for downstream analyses, which showed that mettl4 captured RNA molecules, mostly transfer RNA (tRNA) and small nuclear RNA (snRNA), including U2, U4, and U6atac (Fig. 1b, c). We next investigated whether the RNAs identified by the eCLIP experiments are indeed substrates of mettl4 by carrying out in vitro enzymatic assays. We synthesized oligonucleotides containing tRNA and snRNA sequences and various controls, including DNAs with the same sequences (Supplementary Table S1). The in vitro enzymatic activity of mettl4 on each candidate substrate and control