Absolute quantitative and base-resolution sequencing reveals comprehensive landscape of pseudouridine across the human transcriptome

Abstract

Pseudouridine (Ψ) is one of the most abundant modifications in cellular RNA. However, its function remains elusive, mainly due to the lack of highly sensitive and accurate detection methods. Here, we introduced 2-bromoacrylamide-assisted cyclization sequencing (BACS), which enables Ψ-to-C transitions, for quantitative profiling of Ψ at single-base resolution. BACS allowed the precise identification of Ψ positions, especially in densely modified Ψ regions and consecutive uridine sequences. BACS detected all known Ψ sites in human rRNA and spliceosomal small nuclear RNAs and generated the quantitative Ψ map of human small nucleolar RNA and tRNA. Furthermore, BACS simultaneously detected adenosine-to-inosine editing sites and N¹-methyladenosine. Depletion of pseudouridine synthases TRUB1, PUS7 and PUS1 elucidated their targets and sequence motifs. We further identified a highly abundant Ψ₁₁₄ site in Epstein-Barr virus-encoded small RNA EBER2. Surprisingly, applying BACS to a panel of RNA viruses demonstrated the absence of Ψ in their viral transcripts or genomes, shedding light on differences in pseudouridylation across virus families.

Fig. 1. BACS achieved quantitative detection of Ψ through cyclization chemistry.

a, Schematic overview of BACS reaction. b, MALDI characterization of BACS labeling of a 10mer Ψ-containing RNA oligonucleotide. Calculated mass is shown in black. Observed mass is shown in red. Data are representative of two independent experiments. c, Cumulative (left) and motif-dependent (middle and right) results of BACS conversion rates and false-positive rates on synthetic 30mer NNΨNN and NNUNN spike-in. Data are shown as means ± s.d. of six independent experiments (n = 6). d, BACS calibration curve for quantification of Ψ stoichiometry in NNUNN motif. Data are representative of two independent experiments. Source data

Fig. 2. BACS detected known Ψ sites in human rRNA and spliceosomal snRNAs.

a, Flowchart of BACS library construction. b, Numbers of Ψ sites identified in HeLa cy-rRNAs and mt-rRNAs. c–e, Conversion rates of BACS (pink) and control (gray) samples in HeLa 28S rRNA (c), 18S rRNA (d) and 5.8S rRNA (e), respectively. Data are presented as means of two independent experiments. f, Venn diagram illustrating the overlap of Ψ sites detected in human cy-rRNAs between BACS and SILNAS MS. g, Numbers of Ψ sites identified in HeLa spliceosomal snRNAs. h, Venn diagram illustrating the overlap of Ψ sites detected in human spliceosomal snRNAs between BACS and SILNAS MS. i, Conversion rates of BACS (pink) and control (gray) samples in HeLa U2 snRNA. Data are presented as means of two independent experiments. Source data

Fig. 3. BACS unveiled the comprehensive Ψ profile of human tRNA.

a, Median numbers of Ψ sites identified per tRNA in each cy-tRNA and mt-tRNA isotype from HeLa cells. NA, not applicable. b, Heat map showing the modification levels of high-confidence Ψ sites in HeLa cy-tRNAs. Only one representative tRNA isodecoder was presented for each isoacceptor family. c, Comparison of the modification levels of Ψ sites at selected positions of HeLa cy-tRNAs. Box plots visualize all Ψ sites at each position; boxes represent the 25th to 75th percentiles with a line at the median; whiskers correspond to 1.5 times the interquartile range (tRNA position: 13, n = 43; 20B, n = 12; 27, n = 81; 28, n = 53; 32, n = 18; 38, n = 29; 39, n = 86; 40, n = 13; e12, n = 17; 54, n = 32; 55, n = 185). d. Venn diagram illustrating the overlap of Ψ sites in human mt-tRNAs reported by BACS and a previously published dataset. Source data

Fig. 4. Simultaneous characterization of Ψ and A-to-I editing sites in HeLa mRNA.

a, Numbers of Ψ and A-to-I editing sites with high (50–100%; red), medium (20–50%; yellow) and low (5–20%; green) modification levels identified in HeLa poly-A-tailed RNA. b, Modification level distribution of Ψ and A-to-I editing sites in HeLa poly-A-tailed RNA. c, Distribution of Ψ and A-to-I editing sites within different features of HeLa mRNA and ncRNA. d, Metagene profile of Ψ and A-to-I editing sites in HeLa mRNA. e, Venn diagram illustrating the overlap of transcripts possessing Ψ and A-to-I editing sites. f, Motif frequency of Ψ sites in HeLa mRNA. g, Modification level distributions of HeLa mRNA Ψ sites within selected motifs, with means indicated in each plot by a horizontal line. Motif: UGΨAG, n = 84; UCΨAG, n = 16; GUΨCA, n = 35; GUΨCC, n = 39; GUΨCG, n = 20; GUΨCU, n = 27; GUΨAA, n = 22; AAΨUG, n = 19; ACΨUU, n = 26; CAΨUU, n = 18; CUΨUG, n = 26; GUΨUG, n = 20; UUΨUU, n = 16. CDS, coding sequence. Source data

Fig. 5. PUS-dependent Ψ landscape across the HeLa transcriptome.

a, Integrated view of the PUS-dependent Ψ profiles of HeLa cy-tRNAs and mt-tRNAs. Ψ₅₅ in HeLa cy-tRNAs is partially dependent on TRUB1, which is labeled by the dashed line. b, Comparison of the modification levels of Ψ₅₅ in HeLa cy-tRNAs and mt-tRNAs upon TRUB1 depletion. Box plots visualize all Ψ sites at each position; boxes represent the 25th to 75th percentiles with a line at the median; whiskers correspond to 1.5 times the interquartile range (cy-tRNA Ψ₅₅, n = 180; mt-tRNA Ψ₅₅, n = 6). c, Scatterplot illustrating all TRUB1-dependant Ψ sites across the HeLa transcriptome. d, Sequence motifs of TRUB1-dependent Ψ sites in mt-tRNAs and poly-A-tailed RNA. e, Comparison of the modification levels of Ψ sites at selected positions of HeLa cy-tRNAs upon PUS7 depletion. Box plots visualize all Ψ sites at each position; boxes represent the 25th to 75th percentiles with a line at the median; whiskers correspond to 1.5 times the interquartile range (cy-tRNA: Ψ₁₃, n = 43; Ψ_20B, n = 12; Ψ₃₅, n = 7; Ψ₃₆, n = 4; Ψ₅₀, n = 3). f, Comparison of the modification levels of Ψ₅₀ in mt-tRNA^Met between wild-type (WT) and PUS7-KO cell lines. g, Sequence motifs of PUS7-dependent Ψ sites in tRNAs and poly-A-tailed RNA. h, Comparison of the modification levels of Ψ sites at selected positions of HeLa cy-tRNAs and mt-tRNAs upon PUS1 depletion. Box plots visualize all Ψ sites at each position; boxes represent the 25th to 75th percentiles with a line at the median; whiskers correspond to 1.5 times the interquartile range (cy-tRNA: Ψ_27/28, n = 130; mt-tRNA: Ψ_27/28, n = 21; Ψ_66/67/68, n = 4). i, Scatterplot illustrating all PUS1-dependant Ψ sites in HeLa mt-mRNAs. j, Sequence motifs of PUS1-dependent Ψ sites in cy-tRNAs and mt-tRNAs. Source data

Fig. 6. Investigation of Ψ modification in viral RNAs.

a, Ψ modification levels in SARS-CoV-2 viral RNA. b, Canonical EBER2 structure, with the Ψ₁₁₄ site labeled accordingly. c, Ψ modification levels in EBER1 and EBER2 from C666-1, Raji and Elijah cell lines. Source data