Enzyme-mediated methylation and alkynylation enables transcriptome-wide identification of pseudouridine modifications

Pseudouridine (?) is one of the most abundant chemical modifications and plays important roles in RNA function. Advances in our understanding of ? have been hindered by a limit of robust methods to precisely and sensitively map their distributions in cellular RNAs. Here, we present ELAP-seq (Enzymatic Labeling and Pull-down for Sequencing) for ? detection, which leverages a naturally occurring N1-methyl pseudouridine methyltransferase from Methanocaldococcus jannaschii (Mj1640). This enzyme promiscuously converts ? to N1-methyl-? (m1?) or installs a propargyl group at the same location in vitro under a mild condition, exhibiting high sensitivity and specificity, and is also functional inside cells. ELAP-seq enriches ?-containing RNA fragments and enables single-nucleotide-resolution ? detection with markedly enhanced signal-to-noise ratio and reduced sequencing and computational demands. Using ELAP-seq, we identify thousands of candidate ? sites in human HeLa and HEK 293T transcriptomes, validating many previously identified sites as well as reporting new ones. This versatile enzymatic platform expands the toolkit for sensitive labeling and detection of ?, advancing the study of RNA modification biology. Overall design: We developed a method named ELAP-seq to detect pseudouridine modification in the whole transcriptome. In this method, poly-A RNA is isolated and fragmented, after which pseudouridine is specifically labeled by methyltransferase Mj1640 with a propargyl group. Further clicking on a biotin molecule enables enrichment of RNA fragments containing pseudouridine for sequencing. Pseudouridine causes stop signature during reverse transcription thus can be detected at single base resolution. Input libraries are built without enzymatic labeling and is for decreasing false positives which also generate stop signatures in the input samples. The input samples are also used for semi-quantitative evaluation of pull-down enrichment levels for each modification site. Two different reverse transcriptases, Superscript IV (corresponing to samples with letter "IV" in their names) and Superscript III (corresponding to samples with letter "III" in their names), were used when constructing the sequencing libraries. We first applied ELAP-seq to synthesic oligos to evaluate the enzyme's sequence preference and performed two replicates using only superscript IV (samples with names starting with 'oligo' mean synthetic probes containing NN?NN motif and samples with names starting with "spikein" means synthetic probes containing NNUNN motif). We further performed three biological replicates for HeLa and HEK293T cell lines (HeLa-rep1 to 3 and HEK rep1 to 3) to identify candidate modification sites in each cell line. To identify sites that are responsive to DKC1 knockdown, we treated HEK293T cells with negative control siRNA (sictrl) or DKC1 siRNA (siDKC1) and prepared two biological replicates for each condition using only superscript IV. We also employed a recently established in vitro transcribed RNA (IVT RNA) library from HEK293T transcriptome as a negative control and performed two replicates (sample names start with 'IVT').