TRMT1L-catalyzed m22G27 on tyrosine tRNA is required for efficient mRNA translation and cell survival under oxidative stress

Abstract

tRNA modifications are critical for several aspects of their functions, including decoding, folding, and stability. Using a multifaceted approach encompassing eCLIP-seq and nanopore tRNA-seq, we show that the human tRNA methyltransferase TRMT1L interacts with the component of the Rix1 ribosome biogenesis complex and binds to the 28S rRNA as well as to a subset of tRNAs. Mechanistically, we demonstrate that TRMT1L is responsible for catalyzing N2,N2-dimethylguanosine (m²₂G) solely at position 27 of tRNA-Tyr-GUA. Surprisingly, TRMT1L depletion also impaired the deposition of 3-(3-amino-3-carboxypropyl) uridine (acp³U) and dihydrouridine on tRNA-Tyr-GUA, Cys-GCA, and Ala-CGC. TRMT1L knockout cells have a marked decrease in tRNA-Tyr-GUA levels, coinciding with a reduction in global translation rates and hypersensitivity to oxidative stress. Our results establish TRMT1L as the elusive methyltransferase catalyzing the m²₂G27 modification on tRNA Tyr, resolving a long-standing gap of knowledge and highlighting its potential role in a tRNA modification circuit crucial for translation regulation and stress response.

Keywords: CP: Molecular biology; N2,N2-dimethylguanosine; RNA methyltransferases; TRMT1L; eCLIP-seq; m(2)(2)G; nanopore tRNA-seq; oxidative stress response; rRNA; tRNA; translation.

Figure 1.. TRMT1L interacts with components of the Rix1 60S biogenesis complex in the nucleolus

(A) Mass spectrometry analysis of PELP1-associated complexes. (B and C) (B) Western blot analysis of endogenous PELP1 or (C) TRMT1L immunoprecipitated. Asterisk indicates the presence of the IgG cross-reactivity band. (D) Schematic representation of TRMT1L and TRMT1 proteins. (E) Immunofluorescence detection of endogenous TRMT1L. Scale bar, 20 μm (F) Inhibition of rRNA transcription by CX-5461 impairs TRMT1L nucleolar localization in U2OS cells. Scale bar, 20 μm. See also Figure S1.

Figure 2.. TRMT1L co-sediments with pre-60S ribosomes

(A) Pre-ribosomal particles from sucrose gradient fractionation. Proteins and RNA from each fraction were analyzed by western blot (top) or northern blot (bottom), respectively. (B) RNAi depletion of TRMT1L was confirmed by immunoblotting and densitometry quantification. Data are presented as mean ± SD. n = 3. (C) Nuclear pre-ribosomes from TRMT1L depleted cells were separated by sucrose gradient fractionation and analyzed by western blot. See also Figure S2.

Figure 3.. TRMT1L binds to tRNAs and rRNA

(A) eCLIP-sequencing analysis showing percentage of reads in SMInputs and TRMT1L IP samples mapped to CLIP peak regions for HCT116 cells. (B and C) (B) Log₂ fold enrichment of TRMT1L IP over SMInput with adjusted p values < 0.05, fold change ≥ 2, and basemean ≥ 1,000 from HCT116 or (C) HeLa duplicates for different tRNA iso-decoders. See also Figures S3 and S4; Table S1.

Figure 4.. Nano-tRNA-seq identifies potential tRNA modification changes in TRMT1L-depleted cells

A) Heatmap of summed base-calling errors in TRMT1L KO cells vs. WT, for each nucleotide and tRNA. Triangles mark positions with significantly decreased errors. (B) Snapshots of Integrative Genomics Viewer (IGV) tracks for tRNA alignments. (C) The percentage of deletions and mismatches at each position of tRNA iso-decoders were extracted and summarized from the .mpileup file. *p < 0.05. Data are presented as mean ± SD. n = 2. See also Figures S5 and S6; Table S2.

Figure 5.. TRMT1L catalyzes the deposition of m²₂G27 on tRNA-Tyr

(A) Primer extension analysis of tRNAs in WT, TRMT1L KO cells, and in TRMT1L KO cells re-expressing TRMT1L WT or TRMT1L D373V mutant. (B and C) Extracted ion chromatograms of tRNA-Tyr-GUA RNase A digestion products. (D) Extracted ion chromatograms of tRNA-Cys-GCA RNase T1 digestion products. (E) Sypro Ruby protein gel stain and western blot analysis of purified FLAG-TRMT1L WT and TRMT1L D373V. (F and G) In-vitro-transcribed tRNAs were incubated with purified FLAG-TRMT1L WT or D373V from (E) and analyzed by immuno-northern blotting with m²₂G antibody and densitometry quantification of the m²₂G signals. Data are presented as mean ± SD. n = 3. (H) The modifications on in-vitro-transcribed tRNA-Tyr WT, G26A, and G27A after incubating with FLAG-TRMT1L WT were analyzed by primer extension assay. Irrelevant lanes have been cropped. See also Figures S5 and S6.

Figure 6.. TRMT1L-depleted cells show a reduced abundance in tRNA-Tyr

(A and B) (A) Nano-tRNA-seq scatterplots of tRNA abundances of WT and (B) TRMT1L KO cells across biological replicates. The correlation strength is indicated by the Pearson correlation coefficient (r). (C) Volcano plot showing differential expression of tRNAs in TRMT1L KO relative to WT. The threshold of differential expression was set as p < 0.05 and absolute log₂ fold change > 0.6. (D, G, and J) Diagram showing the probes designed to cover different regions of tRNA-Tyr and tRNA-Cys. (E) Total RNAs from HCT116 WT and TRMT1L KO cells were analyzed by northern blot using the indicated Tyr-GUA and U6 probes. (F) Densitometry quantification of Tyr-GUA signal normalized to U6 probe signal from (E). Data are presented as mean ± SD. n = 3. (H, I, K, and L) Total RNAs from HCT116 WT and TRMT1L KO cells were analyzed by northern blot using the depicted tRNA-Tyr probes, tRNA-Cys probes, and U6 probe. Approximate sizes of fragments detected in nucleotides are indicated on the side of each panel. Diagram showing the speculative fragments generated from a cut at position 26/27. The blots in (K) and (L) were done on the same membrane. See also Figure S7.

Figure 7.. TRMT1L-depleted cells exhibit a reduced global translation rate and increased sensitivity to oxidative and ER stresses

(A) WT and TRMT1L KO HCT116 cells were analyzed by polysome profiling. (B) HCT116 cells were transfected with siRNAs and treated with puromycin for the indicated time and analyzed by western blot with the indicated antibodies. (C) WT and TRMT1L KO HCT116 cells were monitored for proliferation by cell counting. Relative cell counts were normalized to the cell counts on day 1. ns: p > 0.05. (D) WT and TRMT1L KO HCT116 cells were subjected to different stress conditions, and live cells were counted after 48 h of treatment. Data are presented as mean ± SD. n = 3.