YTHDF3 facilitates translation and decay of N6-methyladenosine-modified RNA

Abstract

N⁶-methyladenosine (m⁶A) is the most abundant internal modification in eukaryotic messenger RNAs (mRNAs), and plays important roles in cell differentiation and tissue development. It regulates multiple steps throughout the RNA life cycle including RNA processing, translation, and decay, via the recognition by selective binding proteins. In the cytoplasm, m⁶A binding protein YTHDF1 facilitates translation of m⁶A-modified mRNAs, and YTHDF2 accelerates the decay of m⁶A-modified transcripts. The biological function of YTHDF3, another cytoplasmic m⁶A binder of the YTH (YT521-B homology) domain family, remains unknown. Here, we report that YTHDF3 promotes protein synthesis in synergy with YTHDF1, and affects methylated mRNA decay mediated through YTHDF2. Cells deficient in all three YTHDF proteins experience the most dramatic accumulation of m⁶A-modified transcripts. These results indicate that together with YTHDF1 and YTHDF2, YTHDF3 plays critical roles to accelerate metabolism of m⁶A-modified mRNAs in the cytoplasm. All three YTHDF proteins may act in an integrated and cooperative manner to impact fundamental biological processes related to m⁶A RNA methylation.

Figure 1

YTHDF3 selectively binds m⁶A in cells. (A) LC-MS/MS quantification showing m⁶A is enriched in RNAs pulled down with YTHDF3 from the cell lysate. Error bars, mean ± sd, n = 2, technical replicates. (B) Distribution of common PAR-CLIP peaks identified in biological triplicates along the length of mRNA. (C) YTHDF3-binding motif identified by HOMER with common PAR-CLIP peaks identified in triplicates. P = 1× 10⁻⁴⁹⁴, the motif is identified in 67% of PAR-CLIP peaks. (D) Overlap of common YTHDF3 PAR-CLIP peaks and m⁶A-seq peaks from HeLa cells. PAR-CLIP peak numbers, blue; m6A peak numbers, red. (E) Overlap of target transcripts identified in PAR-CLIP triplicates and RIP replicates. (F) Overlap of target transcripts among YTHDF1-3.

Figure 2

YTHDF3 promotes translation efficiency of its mRNA targets and facilitates the function of YTHDF1. (A-B, D) Cumulative distribution of log₂-fold changes of translation efficiency (ratio between ribosome-bound fragments and input RNA) between siControl and siYTHDF3, biological replicate 1 (A), siYTHDF3, biological replicate 2 (B), and siMETTL3 (D). Three groups of genes were plotted: non-targets (neither targets of YTHDF1 nor targets of YTHDF3, black), YTHDF3 CLIP+IP (high-confidence YTHDF3 targets, red), and YTHDF1 unique (YTHDF1 targets that are not targets of YTHDF3, blue). Number of genes in each group was indicated in parentheses. P values were calculated from a two-sided Mann-Whitney test compared to non-targets. (C) Redistribution of representative targets in non-ribosome and polysome portions of mRNPs upon depletion of YTHDF3 measured by RT-qPCR. APC, a YTHDF3 target; TSC1 and DST, YTHDF1 unique targets; and RPL30, a non-target. Error bars, mean ± sd, n = 2, technical replicates. (E) Construct of the double tethering assay. A sequence of two BoxB followed by two MS2 stem loops was inserted at the 3′UTR of F-Luc (firefly luciferase) mRNA. The C-terminal YTH domains of YTHDF1 and YTHDF3 were respectively replaced with λ peptide (binding BoxB motif) and MS2 binding protein (binding MS2 motif). R-Luc (renilla luciferase) mRNA was used as an internal control for normalizing luciferase signals from different samples. (F) Translation efficiency of F-Luc normalized with R-Luc 4-hour post F-Luc induction, with the expression of effectors indicated at x-axis. The ratio between YTHDF1N-λ (yellow) and the corresponding control sample (grey) was calculated. Error bars, mean ± sd, n = 3. P = 0.05 (paired two-sided Student's t-test).

Figure 3

Protein partner analyses showing that YTHDF3 may interact with translation machineries through YTHDF1. (A) Polysome profiling of HeLa cells stably overexpressing Flag-HA tagged YTHDF3, and western blotting of Flag, eIF3A, and eIF3B in each fraction. (B) Comparison of components of protein complex co-immunoprecipitated with Flag-HA-YTHDF3 and Flag-HA-YTHDF1. Scores of each protein partner obtained from mass spectrometry were used in the x-y scatter plot. Subunits of 40S were labelled with red diamond; subunit of 60S, dark blue diamond; EIF components, green triangle; HNPNP proteins, light blue rectangle. See also Supplementary information, Table S2. (C) In vitro binding assay showing that YTHDFs co-immunoprecipitated with one another. Purified GST tagged YTHDF1-3 were incubated with cell lysates from Flag-HA (FH) tagged YTHDF1-3. FH tagged proteins were western blotted in the eluent after GST affinity pull down. FH-Ctrl, Flag-HA expression control cell line.

Figure 4

YTHDF proteins form an interconnected network in the cytosol. (A) Total RNAs bound by YTHDF1 and YTHDF2 quantified with PAR-CLIP followed by 5′-³²P labelling in the control HeLa cells and cells depleted of YTHDF3 using two different siYTHDF3 oligos. Knockdown efficiency of the two siYTHDF3 oligoes was indicated, respectively. Samples loaded in the radioactivity gel were normalized with immunostaining of the Flag-tagged protein. (B-D) Genome-wide analysis of target affinity of YTHDF1 and YTHDF2 with or without YTHDF3. Box plot of RIP enrichment of different groups of YTHDF targets in siControl samples (B), and that of YTHDF1 targets (C) or of YTHDF2 targets (D) in siControl and siYTHDF3 samples. Box, 25%-75%; “−”, max and min; “×”, 1% and 99%; “□”, median. P values were calculated from a two-sided Mann-Whitney test. (E) LC-MS/MS quantification of m⁶A levels of HeLa cells treated with siControl, siYTHDF1, siYTHDF2, siYTHDF3, and combinations of those oligoes. Error bars, mean ± sd, n = 4 (two biological replicates × two technical replicates). (F) LC-MS/MS quantification of 4SU (4-thio-uridine) level in mRNAs pulled down with YTHDF1-3 2-hour and 4-hour post a 1-hour 4SU labeling of nascent RNAs. Error bars, mean ± sd, n = 3∼4. P values were calculated using paired two-sided Student's t-test. ^*P < 0.05; ^**P < 0.005; ^***P < 0.0005; (G) A proposed model for an integrated partition network for m⁶A-modified transcripts mediated by YTHDFs in the cytosol. While YTHDF1 functions in translation regulation and YTHDF2 dominates in accelerating mRNA decay, YTHDF3 could serve as a hub for fine-tuning the RNA accessibility of YTHDF1-2. These three mRNA pools controlled by YTHDF1-3 could be interchangeable and highly dynamic, resulting in an interconnected and dynamic mRNA modulation through m⁶A. YTHDF3 might also interact with other protein partners (grey) to negatively impact translation.