N(6)-methyladenosine Modulates Messenger RNA Translation Efficiency

Abstract

N(6)-methyladenosine (m(6)A) is the most abundant internal modification in mammalian mRNA. This modification is reversible and non-stoichiometric and adds another layer to the dynamic control of mRNA metabolism. The stability of m(6)A-modified mRNA is regulated by an m(6)A reader protein, human YTHDF2, which recognizes m(6)A and reduces the stability of target transcripts. Looking at additional functional roles for the modification, we find that another m(6)A reader protein, human YTHDF1, actively promotes protein synthesis by interacting with translation machinery. In a unified mechanism of m(6)A-based regulation in the cytoplasm, YTHDF2-mediated degradation controls the lifetime of target transcripts, whereas YTHDF1-mediated translation promotion increases translation efficiency, ensuring effective protein production from dynamic transcripts that are marked by m(6)A. Therefore, the m(6)A modification in mRNA endows gene expression with fast responses and controllable protein production through these mechanisms.

Figure 1. Transcriptome-wide Identification of YTHDF1 mRNA Targets

(A) YTHDF1-binding motifs identified by HOMER from PAR-CLIP peaks of two biological replicates. Motif length was restricted to 6–8 nucleotides. The motif with the lowest p value of replicate 1 (repl.1) was found in 53.8% of 24,753 sites (p = 1 × 10⁻⁷⁰³), and that of replicate 2 (repl.2) was found in 54.5% of 58,549 sites (p = 1 × 10⁻¹⁰⁹³). (B) The distribution of the distance from GGAC to T-to-C mutation sites. The optimum crosslinking sites are at the −3, +2, and +3 positions. (C) Overlap of YTHDF1 PAR-CLIP peaks and m⁶A-seq peaks in HeLa cells. (D) Overlap of target genes identified by PAR-CLIP and RIP-seq for YTHDF1. (E) Gene Ontology analysis of YTHDF1 mRNA targets. See also Figure S1 and Table S1.

Figure 2. Knockdown of YTHDF1 Leads to Reduced Translation of Its mRNA Targets

(A–C) Cumulative distribution log₂-fold changes of ribosome-bound fragments (A), mRNA input (B), and translation efficiency (C, ratio of ribosome-bound fragments and mRNA input) between siYTHDF1 and siControl for non-targets (gray), PAR-CLIP-only targets (blue), and common targets of PAR-CLIP and RIP (red). p values were calculated using a two-sided Mann-Whitney test. (D) The mRNA lifetime log₂-fold changes were further grouped and analyzed on the basis of the number of CLIP sites on each transcript. The extent of translation reduction caused by YTHDF1 knockdown correlates with the number of YTHDF1-binding sites for mRNA targets of YTHDF1. p values were calculated using a Kruskal-Wallis test. See also Figure S2 and Table S1.

Figure 3. YTHDF1 Enhances the Translation of m⁶A-Modified RNAs

(A) Knockdown of the m⁶A methyltransferase (METTL3) reduced the translation efficiency of YTHDF1 target transcripts. Cumulative distribution log₂-fold changes of the translation efficiency between siMETTL3 and siControl for non-targets (black) and YTHDF1 RNA targets (red). p = 0, two-sided Mann-Whitney test. (B) A diagram illustrating that YTHDF1 plays two potential roles in the translation of m⁶A-modified RNAs: in Role A, YTHDF1 shuttles more mRNAs to translation machinery; in Role B, YTHDF1 accelerates the translation initiation rate of methylated mRNAs. (C) Quantification of them⁶A/A ratio of total mRNA, the non-ribosome portion, 40S–80S, and polysome determined by LC-MS/MS for the YTHDF1 knockdown samples compared to controls after 48 hr. p values were determined using a two-sided Student’s t test for paired samples. Error bars represent mean ± SD. For total mRNA, n = 8 (four biological replicates × two technical replicates), p = 0.71. For the rest, n = 5 (two biological replicates, two technical replicates + three technical replicates), p = 0.083, 0.035, 0.049 for nonribosome, 40S–80S, and polysome fractions, respectively. See also Figure S3.

Figure 4. The N-Terminal Domain of YTHDF1 Promotes Protein Production in a Tethering Assay

(A) Construct of the tethering reporter assay. The mRNA reporter consists of an inducible promoter, firefly luciferase as the coding region, and five Box B sequence at 3′ UTR (F-luc-5BoxB). The N-terminal domain of YTHDF1 (N_YTHDF1) was fused with λ peptide (N_YTHDF1_λ), which recognizes Box B RNA with a high affinity. R-luc lacks the inducible promoter and was used as an internal control to normalize the F-luc signal. (B) Under constant induction, the tethering of N_YTHDF1_λ to F-luc-5BoxB led to an on-average 42% increased translation in comparison with the control. The translation outcome was determined as a relative signal of F-luc divided by R-luc. Error bars, mean ± SD, p = 5.9 × 10⁻⁴ (two-sided Student’s t test for paired samples), n = 6 (three biological replicates × two technical replicates). (C) Under constant induction, the mRNA abundance decreased slightly in the N_YTHDF1_λ-tethered group compared with the control. The mRNA abundance was determined by qRT-PCR of F-luc and R-luc. Error bars, mean × SD, p = 0.031, n = 6. (D) The translation efficiency of the reporter mRNA increased by ~72% in the N_YTHDF1_λ-tethered group compared with the control. The translation efficiency is defined as the quotient of reporter protein production (F-luc/R-luc) divided by mRNA abundance. Error bars, mean ± SD, p = 3.2 × 10⁻⁵, n = 6. (E) F-luc-5BoxB was induced with a pulse expression for 2 hr. The mRNA reporter showed higher translation when tethered with N_YTHDF1_λ compared with the control. y axis, d(F-luc/R-luc)/dt, indicating the changing rate of protein production. Error bars, mean ± SD, n = 4. (F) After a 2 hr pulse expression and a 1 hr arsenite (1 mM) stress treatment, translation of the reporter protein was largely diminished. The translation recovery was assessed after the stress was released. The result showed that the N_YTHDF1_λ-tethered group exhibited faster translation recovery than the control group. Error bars, mean ± SD, n = 4.

Figure 5. YTHDF1 Associates with Translation Initiation Factors, Ribosome, and Stress Granule Marker

(A) Western blotting of Flag-tagged YTHDF1 on each fraction of 10%–50% sucrose gradient showing that YTHDF1 associates with ribosome. The fractions were grouped to non-ribosomal mRNPs, 40S–80S, and polysome. RPS6 is a protein subunit of 40S ribosome, and eIF3B is a component of the translation initiation complex (each lane is aligned to the corresponding fraction on the upper plot). (B) Western blotting showing that eIF3A, eIF3B, and G3BP1 co-immunoprecipitated with YTHDF1. The association of YTHDF1 with G3BP1 requires RNA, and those with eIF3A/eIF3B are independent of RNA. (C) Construct of the bicistronic IRES reporter assay. As the first coding region, R-luc reports cap-dependent translation of the mRNA and serves as control. The second coding region encodes F-luc whose translation is controlled by different types of upstream IRES elements. Five Box B sequence was inserted at the 3′ UTR as the tethering site for the N-terminal domain of YTHDF1. (D) Tethering N_YTHDF1_λ to the EMCV IRES reporter led to an on-average 27% increased translation of F-luc compared with R-luc, whereas N_YTHDF1_λ had no effect on the CrPV IRES reporter. Error bars, mean ± SD, p (EMCV) = 9.5 × 10⁻⁵, P (CrPV) = 0.050, two-sided Student’s t test for paired samples, n = 8 (biological replicates). See also Figures S4 and S5 and Table S2.

Figure 6. Translation Efficiency of the Common mRNA Targets of YTHDF1 and YTHDF2 is Affected by Both m⁶A Readers

(A and B) Evaluations of translation efficiency and mRNA half-lifetimes of shared targets or non-targets of YTHDF1 and YTHDF2 with and without perturbation. Ribosome profiling and mRNA-seq data were collected under YTHDF1 knockdown (A) or YTHDF2 knockdown (B) conditions. Transcripts were categorized into common targets of YTHDF1 and YTHDF2 (red) or non-targets (black) under the knockdown conditions, and each compared with their corresponding control (pink and gray, respectively). The solid diamonds and circles represent median of the translation efficiency and half-lifetime. The four periphery dots surrounding each median are data quartiles (25% and 75% of each variance) and connected by dashed lines. The blue arrows denote the directions of changes compared to the control. (C–E) Evaluations of both mRNA abundance and protein production of the reporter transcripts using the tethering assay. Cells transfected with inducible luciferase genes were tethered by YTHDF1 (C), YTHDF2 (D), or both (E). Protein production was calculated by normalized luciferase signal (F-luc/R-luc). mRNA abundance was quantified by qRTPCR with normalization. Data points representing the tethered group (red dots) were compared to those of the control group (tethered with control λ peptide, gray dots), and the directions of changes were shown by blue arrows. (F) Quantification of the 4-thiouridine (4SU) /U ratio of YTHDF1- and YTHDF2-bound RNAs. Nascent transcribed mRNAs were labeled by 4SU for 1 hr. RNAs bound by YTHDF1 or YTHDF2 were isolated at 2 or 4 hr post-labeling, then analyzed by LC-MS/ MS. Error bars represent mean ± SD, n = 2. The results indicate that YTHDF1 binds nascent RNA before YTHDF2. See also Figure S6.

Figure 7. A Proposed Model of Translation Promotion by YTHDF1

(A) YTHDF1 recruits m⁶A-modified transcripts to facilitate translation initiation. The association of YTHDF1 with translation initiation machinery may be dependent on the loop structure mediated by eIF4G and the interaction of YTHDF1 with eIF3. (B) The m⁶A-based regulation through binding of YTHDF1 and YTHDF2 shares similarities with that of the AU-rich element which is regulated by HuR/ HuD and AUF1.