Fig 2: RNA methyltransferase METTL16 regulates erythroid differentiation in vivo.a Mettl16 mRNA expression levels in each hematopoietic compartment (n = 3 technical replicates). HSC hematopoietic stem cells, MPP multipotent progenitors, CMP common myeloid progenitors, GMP granulocyte-monocyte progenitors, MEP megakaryocyte-erythrocyte progenitors. b Gene expression patterns of Mettl16 mRNA during erythroid differentiation generated from the single-cell RNA-seq analysis48. c Survival rates of mice with each genotype. d Fetuses on E12.5 of control and Mettl16fl/flEpor-Cre+ mice. e–g Flow cytometric analysis of the E10.5 peripheral blood from Mettl16fl/flEpor-Cre+ mice (e). The histogram of CD71 intensity (f) and the relative CD71 mean fluorescence intensity (MFI) in the gated population (g, n = 4–5 mice). Data were pooled from two independent experiments. h Flow cytometric analysis of the E11.5 FL from Mettl16fl/flEpor-Cre+ mice (n = 4–5 mice). i May-Grünwald-Giemsa stain of cytospin slides of E11.5 FL cells from Mettl16fl/flEpor-Cre+ mice. j The histogram of FSC intensity in each population in h. k Flow cytometric analysis of the E12.5 FL from Mettl16fl/flEpor-Cre+ mice. Data are expressed as mean ± SD (a, g, h). The p-values were calculated using two-tailed Student’s t-test (g, h). Images are representative of samples obtained from at least three mice (d and i–k). Source data are provided as a Source Data file.

Fig 3: Binding of the CFIm complex is sufficient to promote MAT2A splicing.(A) Diagram of the tethering assay. The 2XMS2 ß-globin reporter consists of MAT2A exon 8, the detained intron, exon 9, and the full-length 3´UTR with two bacteriophage MS2-coat protein binding sites inserted 3´ of A4G mutant hp1 (asterisk). A matched reporter lacking the MS2 site (‘No MS2’) is used to measure background. All MS2 fusion proteins have an N-terminal MS2 coat protein, an SV40 nuclear localization signal, and flag tag (MS2-NLS-fl). The MS2-NLS-fl alone is expressed as negative control. Diagram not to scale. (B) Northern analysis of ß-globin reporter RNA after tethering CFIm25 variants. MS2-NLS-fl fusions to METTL16 (M16), CFIm25 variants (wt, m1, m2, m3), or MS2-NLS-fl alone (-) and were expressed with the indicated reporters. Statistical analysis is relative to matched MS2-vector control. Orange, 2XMS2 reporter. Gray, No MS2 reporter. n = 3. (C) Coimmunoprecipitation of CFIm68 and CFIm59 with flag-CFIm25. Flag-tagged wild-type CFIm25 (wt) or mutants (m1, m2, m3), flag-tagged METTL16 (M16), or flag-vector (-) were expressed in HEK293 cells before immunoprecipitation with anti-flag beads. The immunoprecipitates were then probed for endogenous CFIm68 and CFIm59, flag, or ß-actin. Input is 10% of the lysate volume applied to flag beads. n = 3. (D) Diagrams of CFIm25, CFIm68, CFIm68?RS, CFIm59, and CFIm59?RS proteins; diagrams to scale. (E) Northern analysis of ß-globin reporter RNA after tethering CFIm68 or CFIm59. MS2-NLS-flag-tagged METTL16, CFIm68, CFIm68?RS, CFIm59, CFIm59?RS, or vector (-) were co-transfected with the 2XMS2 ß-globin reporter or no MS2 reporter control. Statistical analysis is relative to matched MS2-vector control. F.L., full length; ?, ?RS domain. n = 3.

Fig 4: A CRISPR screen identifies CFIm25 as a candidate MAT2A splicing factor.(A) Diagram of endogenous MAT2A gene (top) and GFP-MAT2A reporter (bottom). DI, detained intron; hp, hairpins; diagram is not to scale. (B) Representative images of reporter cells (phase) and GFP 5 days after transduction with lentivirus expressing Cas9 and sgRNA targeting MAT2A or non-targeting (sgNT) control. Cells were maintained in methionine-rich media. (C) Flow cytometry results monitoring GFP production in the reporter cell line after 24 hr conditioning in methionine-rich (blue) or methionine-free (red) media. Displayed as percent of maximum cell count for a given GFP intensity. (D) Northern blot analysis of GFP and endogenous MAT2A RNAs produced from the reporter line after 12 hr in media with the indicated methionine concentrations. In all figures, ‘DI’ and ‘m’ mark the MAT2A detained intron and mRNA isoforms, respectively. (E) Northern blot analysis of GFP and endogenous MAT2A RNA expression after four hours ± methionine depletion in the reporter line. Cells were treated with non-targeting control (siNT) or METTL16 (siMETTL16) siRNAs. (F) CRISPR screen timeline. (G) CRISPR screen results. CRISPR screen was performed in biological triplicate before analysis by MAGeCK. The -log10(FDR) of the analysis is plotted on the y-axis and genes are organized alphabetically by chromosome number on the x-axis. Non-targeting (NT) guides are also included. Genes above the dotted line have an FDR < 0.01.

Fig 5: Co-depletion of CFIm68 and CFIm59 reduces induction of splicing of the MAT2A DI.(A-B) Representative northern blot and quantification of MAT2A intron detention after CFIm68 and CFIm59 depletion. Two independent siRNAs for each factor were used (labeled #1 and #2). 293A-TOA cells were conditioned with methionine-rich or methionine-free media for 4 hr prior to harvesting. n = 3. (C-D) Representative northern blot and quantification of MAT2A expression after methionine titration after depletion of the indicated factors. 293A-TOA cells were conditioned with media containing the specified methionine concentration for 12 hr prior to harvesting. Non-targeting control, siNT, gray. METTL16, siM16, purple. siCFIm25, orange. siCFIm68 and siCFIm59 co-depletion, green. n = 3.