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

Xu, Haiqi; Kong, Linzhen; Cheng, Jingfei; Al Moussawi, Khatoun; Chen, Xiufei; Iqbal, Aleema; Wing, Peter A. C.; Harris, James M.; Tsukuda, Senko; Embarc-Buh, Azman; Wei, Guifeng; Castello, Alfredo; Kriaucionis, Skirmantas; McKeating, Jane A.; Lu, Xin; Song, Chun-Xiao

doi:10.1038/s41592-024-02439-8

Download PDF

Article
Open access
Published: 30 September 2024

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

Nature Methods (2024)Cite this article

4005 Accesses
8 Altmetric
Metrics details

Subjects

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.

Quantitative profiling of pseudouridylation landscape in the human transcriptome

Article 30 March 2023

Direct RNA sequencing reveals m⁶A modifications on adenovirus RNA are necessary for efficient splicing

Article Open access 26 November 2020

Semi-quantitative detection of pseudouridine modifications and type I/II hypermodifications in human mRNAs using direct long-read sequencing

Article Open access 19 January 2023

Main

Ψ is one of the most abundant posttranscriptional modifications in cellular RNA^1,2. It is not only prevalent in various noncoding RNAs (ncRNAs), including ribosomal RNA (rRNA), small nuclear RNA (snRNA) and transfer RNA (tRNA)³, but also present in messenger RNA (mRNA)^4,5. Ψ plays important roles in splicing, translation, RNA stability and RNA–protein interactions⁶. In eukaryotes, Ψ is installed by pseudouridine synthases (PUSs)⁷, which have been shown to associate with many diseases including cancer⁶. Therefore, establishing an accurate and sensitive method to detect Ψ is highly desirable.

Traditionally, the detection of Ψ has been reliant on the N-cyclohexyl-N′-(2-morpholinoethyl)carbodiimide methyl-p-toluenesulfonate (CMC) chemistry⁸. Since the stable N³–CMC adduct of Ψ blocks the base pairing and would lead to reverse transcription (RT) truncations, CMC chemistry has been widely applied to transcriptome-wide mapping of Ψ, as shown in Pseudo-seq⁴, Ψ-seq⁵ and CeU-seq⁹. However, CMC-based methods have low labeling efficiency and selectivity for Ψ and lack stoichiometry information, making it intrinsically difficult to distinguish between true Ψ signals and background noises.

Recently, RBS-seq reexamined the bisulfite (BS)-mediated conversion of Ψ, and found that the Ψ–BS adduct could lead to deletion signatures during RT^10,11. BID-seq and similarly designed PRAISE further optimized BS treatment at near neutral pH to eliminate the side reaction on unmodified cytosine and enabled quantitative detection of Ψ^12,13. However, due to the deletion signature, BS-based methods cannot determine the exact position of Ψ in consecutive uridine sequences or consecutive Ψ sites and struggle to detect densely modified Ψ sites.

To overcome these limitations, we developed BACS for direct, quantitative and base-resolution sequencing of Ψ based on new bromoacrylamide cyclization chemistry. BACS induces Ψ-to-C mutation signatures rather than truncation or deletion signatures during RT, thereby providing higher resolution and more accurate quantification of Ψ stoichiometry compared with CMC-based and BS-based methods. We applied BACS to various types of ncRNAs and mRNA to build a comprehensive map of Ψ across the human transcriptome. Besides Ψ mapping, BACS delivers simultaneous detection of adenosine-to-inosine (A-to-I) editing sites in mRNA and N¹-methyladenosine (m¹A) in tRNA. We further utilized BACS to elucidate genuine Ψ targets and sequence motifs of three key PUS enzymes (TRUB1, PUS7 and PUS1) in HeLa cells. Finally, we applied BACS to various RNA and DNA viruses to investigate the presence of Ψ in viral RNAs.

Results

Development of BACS

The most distinct difference between Ψ and uridine is the free N¹ of Ψ, which is highly reactive toward Michael addition acceptors, such as acrylonitrile^14,15, acrylamide¹⁶ and other acrylic compounds¹⁷. Selective labeling of Ψ by acrylonitrile has been widely used to distinguish Ψ from uridine in mass spectrometry¹⁸. However, a simple N¹ adduct of Ψ with acrylic compounds would not induce mutation during RT. We envisioned that installing an α-halogen group would induce a tandem cyclization through intramolecular O²-alkylation and finally lead to Ψ-to-C mutation (Fig. 1a). We initially tested this chemistry on a short Ψ-containing oligonucleotide with 2-bromoacrylamide and analyzed the reaction product by matrix-assisted laser desorption/ionization mass spectrometry (MALDI). We found a 69-Da increase of mass values, indicating the formation of a cyclized product (carbamido-1, O²-ethano Ψ, nce^1,2Ψ; Fig. 1b and Supplementary Fig. 1a). This reaction was further confirmed by ultra-high-performance liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS; Supplementary Fig. 1b). To validate the mutation signature of nce^1,2Ψ, we applied BACS to a 72mer RNA containing two Ψ sites (Supplementary Table 1). Through RT and next-generation sequencing, we obtained 86.6% U-to-C mutation rates on these two sites, while U-to-R (R = A or G) mutation rates were lower than 1% (Supplementary Fig. 1c). Therefore, we confirm that the U-to-C mutation rate can serve as the conversion rate of BACS.

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

To understand the sequence preference of BACS, we built libraries with synthetic 30mer RNA spike-in containing NNΨNN and NNUNN (N = A, C, G or U), respectively (Fig. 1c). After BACS, we observed an 87.6% conversion rate of Ψ and a 0.75% false-positive rate of uridine when accumulating all motifs. Among all 256 motifs, 230 showed a conversion rate higher than 85% and 254 displayed a conversion rate higher than 80%, suggesting the high efficiency of BACS chemistry. We observed a low false-positive rate (<1%) in most motifs (213 of 256 motifs), while certain motifs displayed slightly higher false-positive rates (3–4%). Nevertheless, BACS clearly showed higher conversion rates than BID-seq both in general and in specific motifs¹². By mixing NNΨNN and NNUNN spike-in in different ratios, we generated excellent linear calibration curves for accurate quantification of Ψ modification levels (R² = 1.00; Fig. 1d and Supplementary Fig. 1d).

Validation of BACS on human rRNA

To evaluate the performance of BACS, we applied it to cytosolic rRNA (cy-rRNA) from cervical and nasopharyngeal cancer cell lines HeLa and C666-1, respectively (Fig. 2a). As expected, we observed high mutation rates only on Ψ, while other bases showed minimum mutation rates (Supplementary Fig. 2a,b). Using a 5% modification level cutoff, we detected 2, 40 and 62 Ψ sites in HeLa 5.8S, 18S and 28S rRNAs, respectively (Fig. 2b). Most of the sites displayed a high level of Ψ modification (>80%), consistent with reports that Ψ sites are highly modified in human cy-rRNA¹⁹ (Fig. 2c–e and Supplementary Fig. 2c–f). We also examined the raw signals of BACS, which revealed a strong correlation between two biological replicates (Pearson’s r = 1.00; Supplementary Fig. 2g). Compared with the reported SILNAS mass spectrometry (SILNAS MS) data, 103 of 105 known Ψ sites in human cy-rRNA (including one 2′-O-methylpseudouridine, Ψm site) were identified with high confidence¹⁹ (Fig. 2c–e and Supplementary Fig. 2c–f). However, Ψ₁₁₃₆ in 18S rRNA was not detected, possibly due to its low modification level in HeLa cells (4.5% by BACS; Supplementary Fig. 2c,d). Interestingly, we found a 16% U-to-C mutation rate of the known 18S rRNA Ψ₃₆ site in control libraries, although the mutation rate increased to 80% after BACS treatment (Fig. 2d and Supplementary Fig. 2h). Similar results were obtained from BID-seq control libraries¹², suggesting the presence of an uncharacterized single nucleotide polymorphism (SNP) site in HeLa cells. It is noteworthy that these two sites could be readily detected in C666-1 cells (Supplementary Fig. 2i). Therefore, BACS could detect all known Ψ sites in human cy-rRNAs (Fig. 2f). In addition, we detected a new Ψ₄₉₃₈ site in 28S rRNA from HeLa and C666-1 cells, located adjacent to the previously known Ψ₄₉₃₇ site (Supplementary Fig. 2e,f). The presence of Ψ₄₉₃₈ was supported by two public databases, both of which predicted that small nucleolar RNA (snoRNA) SNORA17B would be responsible for catalyzing this modification^20,21. We further identified four new Ψ sites (Ψ₃₁/Ψ₈₉₀/Ψ₈₉₉ in 18S rRNA and Ψ₁₆₇₄ in 28S rRNA) in cy-rRNAs from C666-1 cells (Supplementary Fig. 2i). It is important to note that while some Ψ or uridine modifications can induce intrinsic mutation signals (such as m¹acp³Ψ₁₂₄₈ in 18S rRNA and m³U₄₅₀₀ in 28S rRNA), these can be filtered out by comparing results of BACS libraries with control libraries (Supplementary Fig. 2h). In addition to cy-rRNA, we also applied BACS to mitochondrial rRNA (mt-rRNA) and detected eight and one Ψ sites in 12S and 16S rRNAs, respectively (Fig. 2b). Among them, four sites have also been detected by Pseudo-seq⁴. In general, the modification level of Ψ sites in mt-rRNA was substantially lower than that in their cytosolic counterparts (Supplementary Fig. 2j).

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

As expected, BACS clearly outperformed BS-based methods in the following aspects^12,13. Firstly, the U-to-C mutation signature enabled BACS to determine the exact position and number of Ψ sites in consecutive uridine sequences (adjacent to one or more uridines, for instance, Ψ₈₀₁/Ψ₈₁₄/Ψ₈₁₅ in 18S rRNA, Ψ₁₈₄₇/Ψ₁₈₄₉ in 28S rRNA and Ψ₄₃₂₃/Ψ₄₃₃₁ in 28S rRNA; Supplementary Fig. 3a–c). The improved bioinformatics pipeline of BID-seq with realignment analysis (BID-pipe²²) could not resolve ambiguity in two consecutive uridine sequences and may introduce artifacts in some cases (for example, Ψ₆₈₁ in 18S rRNA, Ψ₁₀₄₅/Ψ₁₀₄₆ in 18S rRNA and Ψ₄₅₄₉ in 28S rRNA; Supplementary Fig. 3d–f). Secondly, more even conversion rates of Ψ sites were obtained over densely modified regions of rRNA using BACS compared with BS-based methods, indicating that BACS datasets would not be influenced by the density of pseudouridylation (Supplementary Fig. 3g,h). In particular, BACS successfully detected extremely dense Ψ sites in a narrow region (for example, Ψ₃₇₃₇/Ψ₃₇₄₁/Ψ₃₇₄₃/Ψ₃₇₄₇/Ψ₃₇₄₉ in 28S rRNA and Ψ₄₂₆₃/Ψ₄₂₆₆/Ψ₄₂₆₉ in 28S rRNA). Thirdly, when compared to SILNAS MS, BACS provided greatly improved accuracy in quantifying Ψ modification levels than BS-based methods (r = 0.90 for BACS, r = 0.37 for BID-seq, r = 0.48 for PRAISE; Supplementary Fig. 3i–k). These findings strongly indicate that quantifying Ψ using deletion signals, as done by BS-based methods, introduces inaccuracies into the analysis.

BACS identified dense Ψ sites in human spliceosomal snRNAs

We next applied BACS to spliceosomal snRNAs from HeLa, C666-1, Raji and Elijah cells. In major spliceosomal snRNA species, we detected 2, 14, 3, 4 and 4 Ψ sites in U1, U2, U4, U5 and U6 snRNAs from HeLa cells, respectively, which is highly consistent with the latest SILNAS MS results²³ (Fig. 2g). Only Ψ₅₉ in U4 snRNA was not detected by BACS, as it is likely to be lowly modified in HeLa cells. However, this position was modified to higher levels in C666-1 and Elijah cells (Supplementary Fig. 4a). Consequently, we detected all known Ψ sites in human major spliceosomal snRNAs (Fig. 2h). It is noteworthy that BACS successfully mapped all 14 Ψ sites in human U2 snRNA, which has not been realized by any other high-throughput sequencing methods, further highlighting the superiority of BACS in detecting dense and consecutive Ψ sites (Fig. 2i). In minor spliceosomal snRNA species, we detected 2, 2 and 1 Ψ sites in U12, U4atac and U6atac snRNAs from HeLa cells, respectively, while no Ψ site was detected in U11 snRNA (Fig. 2g). Notably, we confirmed that there are two consecutive Ψ sites (Ψ₁₁/Ψ₁₂) rather than one Ψ₁₂ site in U4atac snRNA²⁴ (Supplementary Fig. 4b,c). Additionally, we detected highly conserved Ψ₂₄₇ and Ψ₂₅₀ sites in 7SK RNA²⁵ and a differentially modified Ψ₂₁₁ site in 7SL RNA²³ (Supplementary Fig. 4a), while no high-confidence Ψ sites were detected in U7 snRNA, RNase P RNA, RNase MRP RNA, vault RNA and Y RNA (Supplementary Fig. 4d–h).

BACS revealed the Ψ profile of human snoRNA

The Ψ profile of human snoRNA remains relatively unexplored. Ψ-seq⁵ and BID-seq¹² only detected 11 and 39 Ψ sites in human snoRNA, respectively. In contrast, BACS detected 304 Ψ sites in snoRNA from HeLa cells, including 205, 67 and 32 Ψ sites in box C/D snoRNAs, box H/ACA snoRNAs and small Cajal body-specific RNAs, respectively, with a substantial number of highly modified sites (Supplementary Fig. 5a–c). SnoRNA Ψ sites detected through BACS largely covered those previously identified by Ψ-seq⁵ and BID-seq¹², demonstrating increased sensitivity of BACS (Supplementary Fig. 5d,e). Furthermore, we observed that Ψ sites in box C/D snoRNAs displayed enrichment in the 5′-upstream regions of box D′ and the 3′-downstream regions of box C′, while Ψ sites in box H/ACA snoRNAs were enriched in the 5′-upstream regions of box H and ACA, implying a potential role for Ψ in mediating interactions between snoRNAs and their targets (Supplementary Fig. 6a,b). Indeed, a subset of Ψ sites identified in box C/D and box H/ACA snoRNAs was also located in the predicted guide regions, which was in accordance with the Ψ-seq results⁵ (Supplementary Fig. 6c,d).

In addition, human telomerase RNA component (TERC) shares similar characteristics with snoRNAs. Upon BACS treatment, we detected seven Ψ sites in TERC from HeLa cells, four of which were putative Ψ sites previously discovered through a CMC-based primer extension approach²⁶ (Supplementary Fig. 5a,f). In particular, all three new Ψ sites (Ψ₃₈/Ψ₁₀₀/Ψ₁₅₅), together with the known Ψ₁₆₁ and Ψ₁₇₉ sites, were found in the core domain of TERC, which may contribute to the stabilization of TERC structure, similarly to the scenario for Ψ₃₀₆ and Ψ₃₀₇ within the P6.1 loop²⁷.

A comprehensive Ψ map of human tRNA

Ψ is one of the most fundamental and prevalent modifications in human tRNA²⁸. However, CMC-based and BS-based methods failed to map Ψ in human cytosolic tRNAs (cy-tRNAs)^13,29. We used BACS to generate the quantitative Ψ map of cy-tRNAs from HeLa cells with 609 high-confidence Ψ sites (Supplementary Fig. 7a). The number of Ψ sites identified per cy-tRNA varied among different isotypes (Fig. 3a). In cy-tRNAs, Ψ sites were predominantly located at highly conserved positions, including positions 13, 27–28, 38–40 and 55, while Ψ sites at other positions were limited to specific types of cy-tRNAs (Fig. 3b). An integrated view of the Ψ profile of human cy-tRNAs was summarized based on the canonical tRNA numbering system³⁰ (Supplementary Fig. 7b). Notably, position 55 emerged as the most frequently and highly modified Ψ site in cy-tRNAs (Fig. 3c). Moreover, position 13 also displayed a high level of Ψ modification. In contrast, the modification levels of position 27–28 and position 38–40 showed considerable variations.

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

We also detected 54 Ψ sites in HeLa mitochondrial tRNAs (mt-tRNAs; Fig. 3a and Supplementary Fig. 7a). Applying BACS to other human cell lines, we observed notable differential Ψ modification on mt-tRNAs. For example, Ψ₅₅ in mt-tRNA^Met was not characterized as a high-confidence site due to its low modification level in HeLa cells (3.4%), while it was readily detected in C666-1, Raji and Elijah cell lines (Supplementary Fig. 7c,d). In addition, C666-1 cell lines displayed an elevated Ψ level at position 66–68 compared to HeLa, Raji and Elijah cells (Supplementary Fig. 7d). We finally mapped a total of 65 Ψ sites in mt-tRNAs by merging the results of HeLa, C666-1, Raji and Elijah cells, which was highly consistent with the published dataset³¹ (Fig. 3d and Supplementary Table 13). Most of the BACS-only Ψ sites were located near the 3′-end of mt-tRNA, which could not be identified using CMC-based methods. Four Ψ sites in ref. ³¹ were not detected by BACS, including mt-tRNA^Gln Ψ₃₃/Ψ₄₀, mt-tRNA^His Ψ₃₅ and mt-tRNA^Pro Ψ₆₇. Overall, human mt-tRNAs were pseudouridylated to a lesser extent compared with cy-tRNAs (Fig. 3b,c and Supplementary Fig. 7c,d). In contrast, PRAISE detected only 34 Ψ sites in mt-tRNAs from HEK293T cells¹³ (Supplementary Fig. 8a). PRAISE encountered challenges in quantifying consecutive Ψ sites (8 of 34) and determining the precise position for a single Ψ site within multiple uridine contexts (13 of 34; Supplementary Fig. 8b–e). As a result, PRAISE achieved quantitative and single-base-resolution detection for only 13 Ψ sites in mt-tRNAs, revealing crucial limitations compared to BACS.

Profiling and quantification of Ψ in HeLa mRNA

After successfully applying BACS to various types of ncRNAs, we extended its usage to map and quantify Ψ modifications in HeLa mRNA. Given the low stoichiometry of Ψ modification in poly-A-tailed RNA, we applied in vitro transcribed poly-A-tailed RNA (IVT RNA) from HeLa cells as a modification-free control to help with Ψ calling³² (Supplementary Fig. 9a). We detected a total of 1,335 Ψ sites in HeLa poly-A-tailed RNA (Fig. 4a). In contrast to the aforementioned ncRNAs, the Ψ modification level in poly-A-tailed RNA was considerably lower (Supplementary Fig. 9b), with the majority exhibiting low modification levels (<20%) and only a limited number of Ψ sites displaying high levels of modification (>50%; Fig. 4a,b). Among the 1,335 Ψ sites, 1,294 and 41 were located in mRNA and ncRNA (excluding rRNA, snRNA, snoRNA and tRNA), respectively (Fig. 4c). Within mRNA, Ψ was enriched in the coding sequence and 3′-untranslated region (3′ UTR), consistent with previous findings^9,12,13 (Fig. 4c,d). The 1,335 Ψ sites were located across 1,103 poly-A-tailed RNA transcripts, with the majority carrying only one Ψ site (Fig. 4e). Gene Ontology (GO) analysis revealed that Ψ-modified mRNA was enriched in functions such as translation and regulation of apoptotic process (Supplementary Fig. 9c). Importantly, BACS could simultaneously provide mRNA expression levels while mapping Ψ, which showed strong correlation with control libraries (Pearson’s r = 0.99–1.00), suggesting minimal RNA degradation induced by BACS (Supplementary Fig. 9d).

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

Next, we analyzed the sequence contexts of Ψ in HeLa mRNA. First, the majority of Ψ sites (55.3%) were located in consecutive uridine sequences, which could not be precisely determined through BS-based methods^12,13 (Supplementary Fig. 10a). Benefiting from the high resolution of BACS, we found that Ψ was predominantly enriched in USΨAG (S = C or G) and GUΨCN (N = A, C, G or U) motifs, corresponding to the previously identified PUS7 and TRUB1 motifs, respectively³³ (Fig. 4f). In addition, we also observed that Ψ tends to be enriched in those motifs containing multiple consecutive uridines, such as CUΨUG, ACΨUU and UUΨUU. Among all motifs, GUΨCN exhibited a relatively high modification level (Fig. 4g). Furthermore, we analyzed the codon preference of Ψ in mRNA. As expected, Ψ was enriched in those codons containing consecutive uridines, such as UUY (Y = C or U), UUG, AUU and GUU, which encoded phenylalanine (Phe), leucine (Leu), isoleucine (Ile) and valine (Val), respectively (Supplementary Fig. 10b,c). Within codons, Ψ was mainly located in the second position (Supplementary Fig. 10d). Moreover, we observed one Ψ site positioned in the start codon (AUG), while two sites were found in the stop codon (UAG), which may promote stop codon readthrough according to previous research^12,34.

To further evaluate the performance of BACS, we compared our identified mRNA Ψ sites with published datasets. We first compared BACS with a recent dataset that consolidated three CMC-based methods³³. Remarkably, BACS accurately identified 61 of 70 Ψ sites (87.1%) listed in the ‘highest confidence’ category (Supplementary Fig. 11a). However, a strong overlap between BACS and the ‘high-confidence’ list was achieved only when considering Ψ sites consistently detected across multiple samples (>8; 177 of 320 Ψ sites, 55.3%), indicating the considerable variance between different CMC-based datasets (Supplementary Fig. 11b,c). We further compared BACS with two recently developed BS-based methods. Compared to CMC-based approaches, BACS demonstrated a better overlap with BID-seq results¹² (230 of 575 Ψ sites, 40.0%; Supplementary Fig. 11d). Most of the sites exclusive to the BID-seq dataset displayed low modification levels in our BACS libraries, possibly due to the inaccurate quantification of BID-seq (Supplementary Fig. 11e). When compared with PRAISE¹³, 651 of 1,995 Ψ sites (32.6%) showed an overlap with BACS results (Supplementary Fig. 11f). Similarly, the majority of PRAISE-only Ψ sites were lowly modified in our dataset (Supplementary Fig. 11g). Regarding the 7 Ψ sites identified in mitochondrial mRNAs (mt-mRNAs) by BACS, 4, 2 and 4 of them have also been detected by Pseudo-seq⁴, BID-seq¹² and PRAISE¹³, respectively.

Simultaneous profiling of A-to-I editing sites and m¹A

In addition to Ψ mapping, BACS enabled the simultaneous detection of A-to-I editing sites in a similar way with ICE-seq³⁵ (Supplementary Fig. 9a). We identified 1,121 A-to-I editing sites in HeLa poly-A-tailed RNA, with a mean modification level of 20% (Fig. 4a,b). In stark contrast to Ψ, most of A-to-I editing sites were resident in the Alu elements (Supplementary Fig. 9e). We further annotated 737 and 115 A-to-I editing sites to mRNA and ncRNAs, respectively (Fig. 4c). Within mRNA, the A-to-I editing sites were predominantly enriched in 3′ UTR, consistent with previous findings³⁶ (Fig. 4d). Interestingly, mRNA transcripts that carry A-to-I editing sites did not overlap with those possessing Ψ, suggesting distinct roles of A-to-I editing and pseudouridylation in mRNA processing (Fig. 4e).

Similarly to RBS-seq¹⁰, BACS induces Dimroth rearrangement of m¹A to N⁶-methyladenosine (m⁶A) and could potentially detect m¹A together with Ψ (ref. ³⁷; Supplementary Fig. 12a). As expected, we observed a marked reduction of m¹A mutation signals at tRNA position 58 (for cy-tRNAs) and 9 (for mt-tRNAs) after BACS treatment (Supplementary Fig. 12b,c). These results highlight the value of BACS to detect multiple modifications simultaneously.

BACS uncovered new PUS targets in HeLa cells

To elucidate the PUS-dependent Ψ profile in the HeLa transcriptome, we generated individual knockouts for three key PUS enzymes: TRUB1, PUS7 and PUS1 (Fig. 5a). In contrast to the conventional view that TRUB1 was the sole PUS enzyme responsible for cy-tRNA Ψ₅₅ (ref. ³⁸), depletion of TRUB1 did not eradicate Ψ₅₅ in human cy-tRNAs, suggesting redundancy of PUS enzymes for this position (Fig. 5b,c). Surprisingly, all Ψ₅₅ sites in mt-tRNA were eliminated upon TRUB1 depletion, challenging another belief that TRUB2 was solely responsible for mt-tRNA Ψ₅₅ (ref. ³⁹; Fig. 5b,c). We further discovered that the TRUB1 motif would extend beyond the recognized GUΨCNA (N = A, C, G or U) motif^5,33, since it could modify GUΨUAA in mt-tRNA^Asn, GUΨGUA in mt-tRNA^Glu and GUΨAAA in mt-tRNA^Pro with high efficiency (Fig. 5d and Supplementary Fig. 13a–c). We observed similar results in poly-A-tailed RNA, confirming that TRUB1 could edit GUΨANA, GUΨGNA and GUΨUNA motifs (Fig. 5d). Therefore, we demonstrate that the TRUB1 motif can be extended from GUΨCNA to GUΨNNA.

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

PUS7-knockout (KO) HeLa cells allowed us to reveal its role in the formation of Ψ_20B, Ψ₃₆ and Ψ₅₀ in cy-tRNAs, in addition to its previously known targets Ψ₁₃ and Ψ₃₅ (ref. ⁴⁰; Fig. 5e and Supplementary Fig. 13d). Moreover, our study uncovers new PUS7 activity in mitochondria by catalyzing the modification of Ψ₅₀ in mt-tRNA^Met (Fig. 5f). The majority of PUS7 targets in tRNA displayed the conserved UVΨAR (V = A, C or G; R = A or G) motif (Fig. 5g). However, PUS7 displayed comparable activity within the UGΨGG motif (cy-tRNA^Arg(CTT) Ψ₅₀) and relatively low activity within the UGΨUG motif (mt-tRNA^Met Ψ₅₀; Fig. 5f,g and Supplementary Fig. 13e). In poly-A-tailed RNA, PUS7 mainly catalyzed the pseudouridylation within the UBΨAG (B = C, G or U) motif (Fig. 5g). Collectively, the true PUS7 consensus motif would be UNΨAR (N = A, C, G or U; R = A or G) and UGΨKG (K = G or U), which was less strict than previously considered UGΨAR (R = A or G) motif^4,41.

Finally, PUS1 depletion resulted in the complete loss of Ψ_27/28 in human cy-tRNAs (Fig. 5h and Supplementary Fig. 13f). In mt-tRNAs, PUS1 not only catalyzed the modification of Ψ_27/28 but also induced the formation of Ψ_66/67/68, consistent with the findings on yeast and mouse PUS1 homologs⁴² (Fig. 5h and Supplementary Fig. 13f). We also observed noncanonical activity of PUS1, exemplified by Ψ₂₅ in mt-tRNA^Asn and Ψ₂₀ in mt-tRNA^Leu(UUR) (Supplementary Fig. 13g,h). Additionally, PUS1 was found to catalyze all seven Ψ sites identified in mt-mRNAs (Fig. 5i). These results show that PUS1 is the major PUS enzyme in mitochondria with diverse functions. PUS1-dependent Ψ sites did not show any sequence motifs, consistent with an earlier report that its activity is dependent on RNA structure⁴¹ (Fig. 5j).

Mimicry of tRNA has been considered as a general way for mRNA pseudouridylation^5,33. Although TRUB1, PUS7 and PUS1 could edit both tRNAs and poly-A-tailed RNA, the Ψ targets in poly-A-tailed RNA were substantially less modified than their counterparts in tRNAs (Supplementary Fig. 13i). These results suggest that mRNA may not be the primary substrate of these stand-alone PUS enzymes, consistent with our data showing that the majority of mRNA Ψ sites were modified to a low level.

Mapping of Ψ in viral RNAs

It has been widely accepted that Ψ-modified RNAs can suppress innate immune responses and may influence the mRNA vaccine design⁴³. Previous studies have reported the presence of Ψ in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by Nanopore sequencing^44,45; however, it has not been thoroughly confirmed with the latest sequencing technologies. Therefore, we applied BACS to five RNA viruses, including SARS-CoV-2, hepatitis C virus (HCV), Zika virus (ZIKV), hepatitis delta virus (HDV) and Sindbis virus (SINV). Surprisingly, we did not detect any high-confidence Ψ sites in the five RNA viruses (Fig. 6a and Supplementary Fig. 14). Importantly, we confirmed robust viral infection in our model systems, to ensure a high abundance of viral RNAs concomitant with high depth of coverages (Supplementary Table 3). These results suggest that Ψ is not directly involved in the modification of these RNA viruses.

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

In addition to these RNA viruses, we also applied BACS to human cell lines infected by Epstein–Barr virus (EBV), a DNA virus that encodes two highly expressed ncRNAs, EBER1 and EBER2. A previous study using HydraPsi-seq⁴⁶ and CMC-based primer extension methods reported one lowly modified Ψ₁₆₀ site in EBER2 (ref. ⁴⁷; Fig. 6b). However, we detected one highly modified Ψ₁₁₄ site in EBER2, while no Ψ site was identified in EBER1, indicating the previous results were likely caused by high backgrounds from hydrazine and CMC chemistry (Fig. 6b,c). The EBER2 Ψ₁₁₄ site was conserved across all EBV strains and host cell lines tested (Fig. 6c). To validate this Ψ site, we applied modified BID-seq (Methods) to EBV-encoded ncRNAs from Raji and Elijah cell lines. Indeed, we detected a potential Ψ site within the three consecutive uridines (U₁₁₂–U₁₁₄) in EBER2, while no Ψ site was found in EBER1 (Supplementary Fig. 15a). However, BS-based methods could not determine the exact position of this site due to the limitations of deletion signals, which further strengthens the advantages of BACS (Supplementary Fig. 15b). Taken together, these results suggest different ways of utilizing Ψ between virus families and further highlight BACS as a highly specific method.

Discussion

In this study, we report the development of BACS for quantitative and base-resolution sequencing of Ψ. BACS is based on new bromoacrylamide cyclization chemistry and induces Ψ-to-C mutation signatures rather than truncation or deletion signatures, allowing for accurate quantification of Ψ stoichiometry and sequencing of Ψ at absolute single-base resolution. Importantly, BACS overcomes the inherent limitations of BS-based methods in three crucial aspects: (1) it facilitates the precise determination of Ψ sites located adjacent to one or more uridines; (2) it enhances the detection of densely modified Ψ sites with higher accuracy and sensitivity; and (3) it offers much more accurate quantification of Ψ in all sequence contexts. These advances make BACS a valuable tool for studying Ψ modifications in cellular RNAs, as it can provide a more comprehensive and accurate picture of the Ψ landscape across various RNA species.

Using BACS, we successfully detected all known Ψ sites in human rRNA and spliceosomal snRNAs and generated the quantitative Ψ map of human snoRNA and tRNA. We further applied BACS to HeLa mRNA and revealed a rather low level of pseudouridylation. However, recent research has highlighted that Ψ could be much more abundant in pre-mRNA⁴⁸. In the future, BACS could be extended to investigate the pseudouridylation of nascent RNA. Moreover, by genetic depletion of PUS enzymes, we identified new targets of TRUB1, PUS7 and PUS1 in HeLa cells. The absolute single-base resolution of BACS enabled us to extend the sequence motifs of TRUB1 and PUS7. Indeed, BACS could serve as a valuable tool for studying PUS knockout cells to elucidate the properties and functions of the other PUS enzymes. Finally, we redefined the Ψ landscape of EBV-encoded ncRNAs EBER1 and EBER2, and several human RNA viruses (SARS-CoV-2, HCV, ZIKV, HDV and SINV), showing the importance of having a highly sensitive and specific method like BACS.

In light of these potential applications, we anticipate BACS to be widely adopted as the new standard to advance our understanding of Ψ modifications and their functional implications in diverse biological processes.

Methods

Preparation of model RNA

Regular and Ψ-labeled 10mer RNA oligonucleotides and 30mer spike-ins were purchased from Integrated DNA Technologies. The 72mer Ψ-containing model RNA used for mutation analysis and the 1.8-kb 10% Ψ-modified RNA used for UHPLC–MS/MS were prepared by T7 in vitro transcription using HiScribe T7 High Yield RNA Synthesis Kit (NEB) and Pseudo-UTP (Jena Bioscience) according to the manufacturer’s protocol. The DNA template was removed by adding 2 μl Turbo DNase (Invitrogen) to the reaction and incubating at 37 °C for 30 min. The products were finally purified with Monarch RNA Cleanup Kit (NEB). Sequences of RNA oligonucleotides can be found in Supplementary Table 1.

Mass spectrometry analysis of short oligonucleotides

MALDI was performed on a Voyager-DE Biospectrometry Workstation (Applied Biosystems) with 2′,4′,6′-trihydroxyacetophenone as matrix. All the oligonucleotides were analyzed in positive mode.

Quantification of Ψ level by UHPLC–MS/MS

The untreated and treated RNA were digested into nucleosides by Nucleoside Digestion Mix (NEB) in a 50 µl solution according to the manufacturer’s protocol. After filtering with Amicon Ultra-0.5 ml centrifugal filters (molecular weight cutoff of 3 kDa; Millipore), the digested samples were subjected to UHPLC–MS/MS analysis as described before⁴⁹. The 1290 Infinity LC System (Agilent) was equipped with a ZORBAX RRHD SB-C18 column (2.1 × 150 mm, 1.8 μm, Agilent) coupled to a 6495B Triple Quadrupole Mass Spectrometer (Agilent). The ions were monitored in positive mode with mass transitions of m/z 245 to 125 (Ψ+H) and m/z 245 to 113 (rU+H; Supplementary Table 2). Concentrations of nucleosides in RNA samples were deduced by fitting the signal peak areas into the standard curves.

Cell culture

HeLa cells (gifted from P. J. Ratcliffe, University of Oxford; originally obtained from the American Type Culture Collection (ATCC), CCL-2) were cultured in DMEM medium (Gibco) supplemented with 10% (vol/vol) FBS (Gibco) and 1% penicillin–streptomycin (Gibco) at 37 °C with 5% CO₂. For isolation of RNA, cells were harvested by centrifugation for 5 min at 1,000g at room temperature.

C666-1, Raji and Elijah cells were grown in RPMI 1640 medium (Thermo), complemented with 10% (vol/vol) FBS (Biosera), 2 mM l-glutamine (Thermo) and 100 units per ml penicillin and 100 µg ml⁻¹ streptomycin (Thermo). Cells were grown in a humidified incubator at 37 °C with 5% CO₂. The C666-1 cell line was gifted from C. Dawson (University of Warwick). Raji and Elijah cell lines were gifted from P. Farrell (Imperial College London). Cell lines were tested for mycoplasma monthly using MycoAlert Kit (Lonza) and were sent for authentication by Eurofins genomics.

Generation of CRISPR knockout cell lines

TRUB1-KO, PUS7-KO and PUS1-KO HeLa cells were generated using CRISPR–Cas9 technology (Supplementary Figs. 16 and 17). Briefly, single guide RNA (sgRNA) sequences were cloned into PX459 plasmids⁵⁰. Transfection was performed using Lipofectamine 3000 Transfection Reagent (Invitrogen) following the manufacturer’s protocol. Cells were then selected by 2 µg ml⁻¹ puromycin (Thermo). Serial dilution was performed to achieve clonal isolation. Finally, clones were expanded and picked for western blot validation with TRUB1 antibody (Proteintech, 12520-1-AP; 1:1,000 dilution), PUS7 antibody (Abcam, ab226257; 1:10,000 dilution) and PUS1 antibody (Proteintech, 11512-1-AP; 1:1,000 dilution). The sgRNA sequences were listed as follows:

TRUB1: 5′-CACGGCGAACACGCCGCTCAAGG-3′;

PUS7-sgRNA1: 5′-TTAATATTGAAACCCCGCTCTGG-3′;

PUS7-sgRNA2: 5′-TCGGAATGCAGTCTAACCAAAGG-3′;

PUS1: 5′-AATACAGCCTGACCGGACGAGGG-3′.

RNA isolation

Total RNA was isolated using TRIzol (Invitrogen) and Direct-zol RNA Miniprep Plus (Zymo Research) according to the manufacturer’s protocol. Ribo⁻ RNA was isolated using RiboMinus Eukaryote System v2 (Invitrogen) according to the manufacturer’s protocol. Poly-A⁺ RNA was isolated by two rounds of poly-A-tailed selection using Dynabeads mRNA DIRECT Purification Kit (Invitrogen) according to the manufacturer’s protocol. To remove genomic DNA contamination, RNA was then treated with Turbo DNase and purified by Zymo-IC Column with RNA binding buffer.

Viral infection and RNA isolation

SARS-CoV-2

Viral stocks were propagated as previously reported⁵¹. Briefly, Vero-TMPRSS2 cells (NIBSC, 100978) were infected with SARS-CoV-2 Victoria 02/20 strain at a multiplicity of infection (MOI) of 0.003 and incubated for 48–72 h until a visible cytopathic effect was observed. Viral titers were then determined by plaque assay from clarified supernatants. For sequencing, Calu-3 cells (gifted from N. Zitzmann, University of Oxford; originally obtained from the ATCC, HTB-55) were infected at an MOI of 1 for 1 h at 37 °C. The inoculum was then removed, cells washed thrice with PBS and incubated in Advanced DMEM with 10% FCS at 37 °C for 24 h. Total RNA was extracted using the RNeasy Mini Kit (Qiagen) and infection confirmed plaque assay and RT–qPCR using primers for the viral N gene: forward 5′-CACATTGGCACCCGCAATC-3′, reverse 5′-GAGGAACGAGAAGAGGCTTG-3′.

HCV and ZIKV

As previously reported⁵², the ZIKV MP1751 strain was propagated in Vero cells (ATCC, CCL-81) and concentrated with 8% polyethylene glycol (PEG) in NTE buffer. For infection, Huh-7.5 cells (gifted from C. Rice, Rockefeller University) were inoculated with either HCV or ZIKV (MOI 1) for 180 min, before extensive washing with PBS. The medium was replaced and infected cells were cultured for 72 h before lysing in RLT buffer. RNAs were extracted using RNeasy Mini Kit (Qiagen) and infection confirmed using RT–qPCR quantification of viral RNAs using specific primers pairs. HCV: forward 5′-TCCCGGGAGAGCCATAGTG-3′, reverse 5′-TCCAAGAAAGGACCCAGTC-3′; ZIKV: forward 5′-TCGTTGCCCAACACAAG-3′, reverse 5′-CCACTAATGTTCTTTTGCAGACAT-3′; and RPLP0: forward 5′-GCAATGTTGCCAGTGTCTG-3′, reverse 5′-GCCTTGACCTTTTCAGCAA-3′.

HDV

HDV inoculum was prepared by concentrating the culture supernatant of Huh-7 cells (gifted from A. Patel, University of Glasgow) transfected with pSVL(D3) and pT7HB2.7 plasmids as previously reported⁵³. HepG2-NTCP cells (gifted from S. Urban, University of Heidelberg) were differentiated with 2.5% dimethylsulfoxide (DMSO)-containing culture medium for 72 h before viral infection, before inoculation with HDV (MOI 50) in the presence of 4% PEG 8000 and 2.5% DMSO for 24 h. After 24 h, cells were washed with PBS and cultured for an additional 5 days in the presence of 2.5% DMSO. Cells were lysed in RLT buffer, and total cellular RNA was purified using RNeasy Mini Kit (Qiagen) and RNase-free DNase Set (Qiagen). Infection was confirmed by RT–qPCR using specific primers to detect HDV transcripts.

SINV

SINV was produced from pT7-SVwt plasmid⁵⁴ that was first linearized with XhoI and purified to use it as a template for in vitro RNA transcription with HiScribe T7 ARCA mRNA kit (NEB). Transcribed viral RNA was transfected into BHK-21 cells (ATCC, CCL-10) using Lipofectamine 3000 reagent (Invitrogen) according to the manufacturer’s instruction. Viruses were collected from the supernatant 24 h later and cleared by centrifugation at 2,000 rpm for 5 min followed by filtration with 0.45-μm PVDF syringe filter units (Merck) and stored at −80 °C. Cleared supernatants were titrated by plaque assay.

Poly-A⁺ RNA purification was performed based on the previously described protocols^55,56, with the following alterations: A549 cells (ATCC, CCL-185) were seeded in two 10-cm dishes in DMEM 10% FBS 24 h before infection to reach 80% confluence. Next, cells were either mock-infected or infected using 0.1 MOI of SINV for 1 h in serum-free DMEM at 37 °C, followed by the replacement of the medium with DMEM supplemented with 2% FBS and incubated for 18 h. Cells were lysed with 1 ml of lysis buffer (20 mM Tris-HCl pH 7.5, 500 mM lithium chloride, 0.5% (wt/vol) lithium dodecyl sulfate, 1 mM EDTA, 0.1% IGEPAL (NP-40) and 5 mM dithiothreitol (DTT)). Lysates were homogenized by passing the lysate at high speed through a 5-ml syringe with a 27-gauge needle, repeating this process until the lysate was fully homogeneous. Then, the whole lysate was incubated with pre-equilibrated oligo(dT)25 magnetic beads (NEB) for 1 h at 4 °C with gentle rotation. Beads were collected in the magnet and washed twice with 2 ml of buffer 1 (20 mM Tris-HCl pH 7.5, 500 mM lithium chloride, 0.1% (wt/vol) lithium dodecyl sulfate, 1 mM EDTA, 0.1% IGEPAL and 5 mM DTT) for 5 min at 4 °C with gentle rotation, followed by two washes with buffer 2 (20 mM Tris-HCl pH 7.5, 500 mM lithium chloride, 1 mM EDTA, 0.01% IGEPAL and 5 mM DTT). Beads were then washed twice with 2 ml of buffer 3 (20 mM Tris-HCl pH 7.5, 200 mM lithium chloride, 1 mM EDTA and 5 mM DTT) at room temperature. Finally, beads were resuspended in 50 µl of elution buffer and incubated for 7 min at 55 °C with agitation. Eluates were stored at −80 °C. Approximately 5 µg of poly-A⁺ RNAs was used for further rRNA removal using the Ribo-Zero kit from the TruSeq Stranded Total RNA LT Kit (Illumina). Subsequently, RNAclean XP (Beckman Coulter) purification was conducted, and RNAs were finally eluted into 10–15 µl of nuclease-free water. The concentration and quality of the RNA were assessed using Qubit and RNA Bioanalyzer.

Preparation of IVT RNA control

A total of 100 ng poly-A⁺ RNA was annealed with 2 μl of 10 μM Oligo(dT)₃₀VN primer (5′-TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTVN-3′) and 2 μl of 10 mM dNTP mix (NEB) in 12 μl solution, incubated at 70 °C for 5 min and held at 4 °C. Next, 5 μl 4× Template Switching RT buffer (NEB), 1 µl of 75 μM T7-TSO (5′-/5Biosg/ACTCTAATACGACTCACTATAGGGAGAGGGCrGrGrG-3′), and 2 μl 10× Template Switching RT Enzyme Mix (NEB) were added to the mixture and the reaction was incubated at 42 °C for 90 min followed by 85 °C for 5 min. The second-strand synthesis was then performed by adding 100 μl Q5 Hot Start High Fidelity Master Mix (NEB), 10 μl RNase H (NEB) and 70 μl nuclease-free H₂O to the cDNA mixture before incubation at 37 °C for 15 min, 95 °C for 1 min and 65 °C for 10 min. The double-stranded cDNA product was purified with 0.8× AMPure XP beads (Beckman Coulter). The IVT RNA control was prepared by T7 in vitro transcription using the purified cDNA product and HiScribe T7 High Yield RNA Synthesis Kit according to the manufacturer’s protocol. To remove cDNA template, IVT RNA was then treated with Turbo DNase and purified by Zymo-IC Column with RNA binding buffer.

BACS for Ψ detection

Around 50–200 ng ribo⁻ or poly-A⁺ RNA was fragmented by NEBNext Magnesium RNA Fragmentation Module at 94 °C for 4 min according to the manufacturer’s protocol and purified by Zymo-IC Column with RNA binding buffer. The fragmented RNA was mixed with 5 μl 10× T4 PNK reaction buffer (NEB), 5 μl T4 PNK (NEB) and 2.5 μl SUPERase•In RNase Inhibitor (Invitrogen) in a 50 µl final solution and incubated at 37 °C for 1 h. The 3′-repaired RNA was purified by Zymo-IC Column with RNA binding buffer and eluted with 10 μl nuclease-free H₂O. The eluted RNA was then mixed with 1 μl synthetic 30mer spike-ins (2%) and 1 μl of 20 μM RNA adaptor (5′-/5rApp/AGATCGGAAGAGCGTCGTG/3SpC3/-3′), incubated at 70 °C for 2 min and immediately placed on ice. Next, 2.5 μl 10× T4 RNA Ligase reaction buffer (NEB), 1 μl SUPERase•In RNase Inhibitor, 7.5 μl 50% PEG 8000 (NEB) and 2 μl T4 RNA Ligase 2, truncated KQ (NEB) were added to the mixture and the reaction was incubated at 25 °C for 2 h followed by 16 °C for 14 h. To digest excess adaptors, the solution was further diluted to 47 μl with nuclease-free H₂O and treated with 2 μl 5′-deadenylase (NEB) at 30 °C for 1 h followed by adding 1 μl RecJ_f (NEB) and incubation at 37 °C for 1 h. The 3′-ligated RNA was purified by Zymo-IC Column with RNA binding buffer and eluted with 12 μl nuclease-free H₂O. A 10 μl aliquot was subjected to BACS library construction, while the remaining 2 μl was saved as control sample and diluted to 12.5 μl with nuclease-free H₂O. For BACS, 1 M 2-bromoacrylamide (Enamine) was prepared by dissolving the solid in DMSO. Next, 10 μl 3′-ligated RNA was added into a 20 μl solution containing 250 mM 2-bromoacrylamide and 625 mM phosphate buffer (pH 8.5) and incubated at 85 °C for 30 min. The treated RNA was double purified by Micro Bio-Spin P-6 Tris Column (Bio-Rad) and Zymo-IC Column with RNA binding buffer and finally eluted with 12.5 μl nuclease-free H₂O.

Both treated and control RNA samples were mixed with 1 μl of 2 μM RT primer (5′-ACACGACGCTCTTCCGATCT-3′) and 1 μl of 10 mM dNTP mix, incubated at 70 °C for 2 min and immediately placed on ice. Next, 4 μl 5× Maxima H⁻ RT buffer (Thermo), 0.5 μl RiboLock RNase Inhibitor (Thermo) and 1 μl Maxima H⁻ Reverse Transcriptase (Thermo) were added to the mixture and the reaction was incubated at 50 °C for 1 h. To digest excess RT primers, the solution was treated with 1 μl Exo I (NEB) and incubated at 37 °C for 30 min followed by adding 1 μl of 0.5 M EDTA (Sigma) to quench the reaction. To hydrolyze the RNA, 2.5 µl of 1 M sodium hydroxide (Sigma) was added and the solution was then incubated at 70 °C for 12 min followed by adding 2.5 µl of 1 M HCl (Sigma) to neutralize sodium hydroxide. The cDNA was finally purified with Dynabeads MyOne Silane (Invitrogen) and eluted with 13 µl nuclease-free H₂O. The eluted cDNA was then mixed with 2 µl of 25 μM cDNA adaptor (5′-/5Phos/NNNNNNAGATCGGAAGAGCACACGTCTG/3SpC3/-3′), incubated at 70 °C for 2 min and immediately placed on ice. Next, 5 μl 10× T4 RNA Ligase reaction buffer, 25 μl 50% PEG 8000, 0.5 μl of 100 mM ATP (NEB), 3.5 μl DMSO (Thermo) and 1 μl T4 RNA Ligase 1, high concentration (NEB) were added to the mixture and the reaction was incubated at 25 °C for 16 h. The ligated cDNA was purified with Dynabeads MyOne Silane and eluted with 15 µl nuclease-free H₂O. The eluted DNA was amplified with NEBNext Multiplex Oligos for Illumina (96 Unique Dual Index Primer Pairs) and NEBNext Ultra II Q5 Master Mix for 10–12 cycles according to the manufacturer’s protocol. The PCR products were purified with 0.8× AMPure XP beads and quantified with Qubit dsDNA HS Assay Kit (Thermo) according to the manufacturer’s protocol. BACS and control libraries were sequenced on a NextSeq 2000 (60-bp paired end reads) with no PhiX added.

Modified BID-seq

Library construction of modified BID-seq was conducted similarly with BACS except for the chemical conversion step. For BS treatment, revised BS and desulfonation reaction conditions of BID-seq were used²². Briefly, the 3′-ligated RNA was eluted in 10 μl nuclease-free H₂O. An 8.5 μl aliquot was subjected to modified BID-seq library construction, while the remaining 1.5 μl was saved as a control sample. The 8.5 μl aliquot was mixed with 45 μl freshly prepared BS reagent (2.4 M Na₂SO₃ and 0.36 M NaHSO₃, Sigma) and incubated at 70 °C for 3 h. The treated RNA was purified by Zymo-IC Column with RNA binding buffer. In-column desulfonation was performed using RNA desulfonation buffer (Zymo), with the column incubated at room temperature for 75 min.

Data preprocessing

Raw sequencing reads were processed by Cutadapt (v.4.2)⁵⁷ to remove low-quality bases (--q 20) and short reads (--m 18), as well as to trim adaptors. 6mer unique molecular identifiers (UMIs) were extracted by UMI-tools extract (v.1.0.1)⁵⁸ and used for deduplication. Paired reads were then merged into single reads using fastp (v.1.0.1)⁵⁹.

Read alignment

Cleaned reads were first mapped to synthetic spike-ins and rRNA references using Bowtie 2 (v.2.4.4)⁶⁰. The key parameters are as follows: bowtie2 -p 2 --no-unal --local -L 16 -N 1 --mp 4. Human rRNA sequences (NR_023363.1, NR_003285.3, NR_003286.4 and NR_003287.4) were downloaded from the National Center for Biotechnology Information (NCBI). The unaligned reads were subsequently mapped to snoRNA references and then to tRNA references, using the same parameters. Human snoRNA sequences that belong to the HUGO Gene Nomenclature Committee (HGNC) ‘small nucleolar RNAs’ gene group (https://www.genenames.org/) were downloaded from RefSeq (https://www.ncbi.nlm.nih.gov/refseq/). Duplicate snoRNA sequences were removed. High-confidence human tRNA sequences (hg38) were downloaded from GtRNAdb⁶¹. Only nonredundant tRNA sequences were kept and appended with a 3′-CCA end. Finally, unmapped reads were aligned to human genome (hg38) with GENCODE v.43 annotation by STAR (v.2.7.9a)⁶².

For RNA viruses, the following reference genomes were used: SARS-CoV-2 isolate Wuhan-Hu-1 (NCBI, NC_045512.2), recombinant HCV J6 (5′ UTR-NS2)/JFH1 (NCBI, JF343782.1), Zika virus isolate ZIKV/H. sapiens/Brazil/Natal/2015 (NCBI, NC_035889.1), HDV sequence from the pSVL(D3) plasmid⁶³ (Addgene plasmid, 29335; https://www.addgene.org/29335/) and SINV (NCBI, NC_001547.1). For EBV samples, reads were aligned to the EBV genome, strain B95-8 (NCBI, V01555.2).

The aligned reads were then filtered and sorted using SAMtools (v.1.16.1)⁶⁴. For synthetic spike-ins and rRNA, only reads with MAPQ ≥ 10 were kept. For snoRNA and tRNA, only reads with MAPQ ≥ 1 were kept. For mRNA, only uniquely mapped reads (-q 30) with a maximum of three mutation counts were kept. Deduplication was performed using UMI-tools dedup (v.1.0.1)⁵⁸. Finally, mutations were counted by SAMtools mpileup (v.1.16.1)⁶⁴ and cpup (v.0.1.0; https://github.com/y9c/cpup/). The sequencing metrics and sample information can be found in Supplementary Table 3.

Calling Ψ sites

BACS raw conversion rates were calculated as C/(T + C). The Ψ modification levels were calculated using the linear equation: Ψ modification level = (R–F)/(C–F), where R, F and C indicated raw conversion rates, motif-specific false-positive rates (from NNUNN spike-in for ncRNAs or IVT-BACS library for poly-A-tailed RNA) and motif-specific conversion rates (from NNΨNN spike-in), respectively. The following criteria were used to call Ψ sites in ncRNAs: (1) coverage ≥ 20 in both BACS and control libraries; (2) background conversion rate ≤ 0.01 or T-to-C mutation counts ≤ 2 in control libraries; (3) background T-to-R (R = A or G) mutation ratio ≤ 0.10 in control libraries; (4) Ψ modification level ≥ 0.05; (5) a P value was calculated for each site using the motif-specific false-positive rates and then adjusted following the Benjamini–Hochberg procedure; the adjusted P value is required to be <0.001; (6) consistently detected in all replicates. For calling cy-tRNA Ψ sites, criteria (4) and (6) were modified to require a Ψ modification level ≥ 0.10 in at least one of two replicates. Only Ψ sites identified in expressed cy-tRNA isodecoders were reported. The following criteria were used to call Ψ sites in poly-A-tailed RNA: (1) coverage ≥ 20 in BACS, control and IVT-BACS libraries; (2) background conversion rate ≤ 0.01 or T-to-C mutation counts ≤ 2 in control libraries; conversion rate in BACS libraries higher than that in IVT-BACS libraries; (3) background T-to-R (R = A or G) mutation ratio ≤ 0.10 in control libraries; (4) Ψ modification level ≥ 0.05; (5) a contingency table test was performed for each site between BACS and IVT-BACS libraries; the P value is required to be <0.01. Statistical analyses were performed in R (v.4.0.3). The full list of Ψ sites identified in the HeLa transcriptome can be found in Supplementary Tables 4–9.

Calling A-to-I editing sites

A-to-I editing sites were called based on the ICE-seq protocol⁶⁵, with minor modifications. Candidate A-to-I editing sites are required to meet the following criteria: (1) coverage ≥ 10 in control libraries; (2) A-to-G mutation ratio ≥ 0.05 in control libraries; (3) A-to-Y (Y = C or T) mutation ratio < 0.05 in control libraries. Supporting reads for the candidate sites are required to meet the following criteria: (1) supporting reads containing an indel call within 5 bp upstream or downstream to the candidate sites were filtered out; (2) supporting reads containing mismatches other than the A-to-G mutation were filtered out; (3) total counts of supporting reads ≥ 3. A ΔG score for each candidate site was calculated as follows:

$$\Delta {\rm{G}}={\log }_{2}\left(\frac{{{\rm{N}}}_{{\rm{G}}}\left({\rm{BACS}}\right)+1}{{{\rm{N}}}_{{\rm{G}}}\left({\rm{control}}\right)+1}\cdot \frac{{\rm{N}}\left({\rm{control}}\right)}{{\rm{N}}\left({\rm{BACS}}\right)}\right)$$

where N_G and N denote the counts of A-to-G mutation and total mapped bases, respectively. The ΔG score is required to be ≤−1. Finally, candidate A-to-I editing sites registered as common SNPs in the dbSNP⁶⁶ were removed. The full list of A-to-I editing sites identified in the HeLa transcriptome can be found in Supplementary Table 10.

RNA structure visualization

The RNA–RNA interactions were visualized using r2r (v.1.0.6)⁶⁷. The snoRNA–rRNA interactions were adapted from snoRNA Atlas²¹.

Downstream analysis

The snoRNA box and guide sequences were downloaded from snoDB 2.0 (ref. ⁶⁸). In the metagene analysis, snoRNA sequences that displayed considerable similarity were streamlined, retaining only one representative snoRNA. The annotation of Ψ sites identified in poly-A-tailed RNA was performed using bedtools intersect (v.2.30.0)⁶⁹ with GENCODE v.43 annotation. Read counts obtained from featureCounts (v.1.6.4)⁷⁰ were normalized based on sequencing depth and gene length using the transcripts per million method. GO analysis was performed with mRNA Ψ sites using enrichR (v.3.2)⁷¹. Sequence logos were generated using ggseqlogo (v.0.1)⁷² in R (v.4.3.1).

Calling PUS-dependent Ψ sites

The following criteria were used to call PUS-dependent Ψ sites: (1) coverage ≥ 20 in both WT and KO samples; (2) Ψ modification level ≥ 0.05 in WT samples; (3) T-to-C mutation counts ≥ 10 in WT samples; (4) Ψ modification level ≤ 0.01 in KO samples. For downregulated Ψ sites, we required the reduction in Ψ modification level to be ≥0.20.

Published data

Related published data were downloaded from the Gene Expression Omnibus (GEO) database: BID-seq for HeLa cells (GSE179798)¹². BID-seq data were processed using the original pipeline (BID-seq¹²) and updated pipeline (BID-pipe²²), respectively.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All sequencing data are available on the GEO database under accession GSE241849. Published data were downloaded from the GEO database: BID-seq for HeLa cells (GSE179798)¹². All relevant additional data have been published with the manuscript, either as part of the main text or in Supplementary Information. Source data are provided with this paper.

Code availability

The analysis scripts are available at https://github.com/lkong888/bacs/.

References

Cohn, W. E. & Volkin, E. Nucleoside-5′-phosphates from ribonucleic acid. Nature 167, 483–484 (1951).
Article CAS Google Scholar
Li, X., Ma, S. & Yi, C. Pseudouridine: the fifth RNA nucleotide with renewed interests. Curr. Opin. Chem. Biol. 33, 108–116 (2016).
Article CAS PubMed Google Scholar
Ge, J. & Yu, Y. T. RNA pseudouridylation: new insights into an old modification. Trends Biochem. Sci. 38, 210–218 (2013).
Article CAS PubMed PubMed Central Google Scholar
Carlile, T. M. et al. Pseudouridine profiling reveals regulated mRNA pseudouridylation in yeast and human cells. Nature 515, 143–146 (2014).
Article CAS PubMed PubMed Central Google Scholar
Schwartz, S. et al. Transcriptome-wide mapping reveals widespread dynamic-regulated pseudouridylation of ncRNA and mRNA. Cell 159, 148–162 (2014).
Article CAS PubMed PubMed Central Google Scholar
Cerneckis, J., Cui, Q., He, C., Yi, C. & Shi, Y. Decoding pseudouridine: an emerging target for therapeutic development. Trends Pharmacol. Sci. 43, 522–535 (2022).
Article CAS PubMed Google Scholar
Borchardt, E. K., Martinez, N. M. & Gilbert, W. V. Regulation and function of RNA pseudouridylation in human cells. Annu. Rev. Genet. 54, 309–336 (2020).
Article CAS PubMed PubMed Central Google Scholar
Bakin, A. V. & Ofengand, J. Mapping of pseudouridine residues in RNA to nucleotide resolution. Methods Mol. Biol. 77, 297–309 (1998).
CAS PubMed Google Scholar
Li, X. et al. Chemical pulldown reveals dynamic pseudouridylation of the mammalian transcriptome. Nat. Chem. Biol. 11, 592–597 (2015).
Article CAS PubMed Google Scholar
Khoddami, V. et al. Transcriptome-wide profiling of multiple RNA modifications simultaneously at single-base resolution. Proc. Natl Acad. Sci. USA 116, 6784–6789 (2019).
Article CAS PubMed PubMed Central Google Scholar
Fleming, A. M. et al. Structural elucidation of bisulfite adducts to pseudouridine that result in deletion signatures during reverse transcription of RNA. J. Am. Chem. Soc. 141, 16450–16460 (2019).
Article CAS PubMed PubMed Central Google Scholar
Dai, Q. et al. Quantitative sequencing using BID-seq uncovers abundant pseudouridines in mammalian mRNA at base resolution. Nat. Biotechnol. 41, 344–354 (2023).
Article CAS PubMed Google Scholar
Zhang, M. et al. Quantitative profiling of pseudouridylation landscape in the human transcriptome. Nat. Chem. Biol. 19, 1185–1195 (2023).
Article CAS PubMed Google Scholar
Chambers, R. W., Kurkov, V. & Shapiro, R. The chemistry of pseudouridine. Synthesis of pseudouridine-5′-diphosphate. Biochemistry 2, 1192–1203 (1963).
Article CAS PubMed Google Scholar
Chambers, R. W. The chemistry of pseudouridine IV. Cyanoethylation. Biochemistry 4, 219–226 (1965).
Article CAS Google Scholar
Knutson, S. D., Ayele, T. M. & Heemstra, J. M. Chemical labeling and affinity capture of inosine-containing RNAs using acrylamidofluorescein. Bioconjug. Chem. 29, 2899–2903 (2018).
Article CAS PubMed PubMed Central Google Scholar
Emmerechts, G., Herdewijn, P. & Rozenski, J. Pseudouridine detection improvement by derivatization with methyl vinyl sulfone and capillary HPLC–mass spectrometry. J. Chromatogr. B 825, 233–238 (2005).
Article CAS Google Scholar
Mengel-Jorgensen, J. & Kirpekar, F. Detection of pseudouridine and other modifications in tRNA by cyanoethylation and MALDI mass spectrometry. Nucleic Acids Res. 30, e135 (2002).
Article PubMed PubMed Central Google Scholar
Taoka, M. et al. Landscape of the complete RNA chemical modifications in the human 80S ribosome. Nucleic Acids Res. 46, 9289–9298 (2018).
Article CAS PubMed PubMed Central Google Scholar
Lestrade, L. & Weber, M. J. snoRNA-LBME-db, a comprehensive database of human H/ACA and C/D box snoRNAs. Nucleic Acids Res. 34, D158–D162 (2006).
Article CAS PubMed Google Scholar
Jorjani, H. et al. An updated human snoRNAome. Nucleic Acids Res. 44, 5068–5082 (2016).
Article CAS PubMed PubMed Central Google Scholar
Zhang, L.-S. et al. BID-seq for transcriptome-wide quantitative sequencing of mRNA pseudouridine at base resolution. Nat. Protoc. 19, 517–538 (2024).
Article CAS PubMed Google Scholar
Yamaki, Y. et al. Direct determination of pseudouridine in RNA by mass spectrometry coupled with stable isotope labeling. Anal. Chem. 92, 11349–11356 (2020).
Article CAS PubMed Google Scholar
Massenet, S. & Branlant, C. A limited number of pseudouridine residues in the human atac spliceosomal UsnRNAs as compared to human major spliceosomal UsnRNAs. RNA 5, 1495–1503 (1999).
Article CAS PubMed PubMed Central Google Scholar
Zhao, Y., Karijolich, J., Glaunsinger, B. & Zhou, Q. Pseudouridylation of 7SK snRNA promotes 7SK snRNP formation to suppress HIV-1 transcription and escape from latency. EMBO Rep. 17, 1441–1451 (2016).
Article CAS PubMed PubMed Central Google Scholar
Kim, N. K., Theimer, C. A., Mitchell, J. R., Collins, K. & Feigon, J. Effect of pseudouridylation on the structure and activity of the catalytically essential P6.1 hairpin in human telomerase RNA. Nucleic Acids Res. 38, 6746–6756 (2010).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Q., Kim, N. K. & Feigon, J. Architecture of human telomerase RNA. Proc. Natl Acad. Sci. USA 108, 20325–20332 (2011).
Article CAS PubMed PubMed Central Google Scholar
Suzuki, T. The expanding world of tRNA modifications and their disease relevance. Nat. Rev. Mol. Cell Biol. 22, 375–392 (2021).
Article CAS PubMed Google Scholar
Song, J. et al. Differential roles of human PUS10 in miRNA processing and tRNA pseudouridylation. Nat. Chem. Biol. 16, 160–169 (2020).
Article CAS PubMed Google Scholar
Sprinzl, M., Horn, C., Brown, M., Ioudovitch, A. & Steinberg, S. Compilation of tRNA sequences and sequences of tRNA genes. Nucleic Acids Res. 26, 148–153 (1998).
Article CAS PubMed PubMed Central Google Scholar
Suzuki, T. et al. Complete chemical structures of human mitochondrial tRNAs. Nat. Commun. 11, 4269 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Z. et al. Systematic calibration of epitranscriptomic maps using a synthetic modification-free RNA library. Nat. Methods 18, 1213–1222 (2021).
Article CAS PubMed Google Scholar
Safra, M., Nir, R., Farouq, D., Slutzkin, I. V. & Schwartz, S. TRUB1 is the predominant pseudouridine synthase acting on mammalian mRNA via a predictable and conserved code. Genome Res. 27, 393–406 (2017).
Article CAS PubMed PubMed Central Google Scholar
Karijolich, J. & Yu, Y. T. Converting nonsense codons into sense codons by targeted pseudouridylation. Nature 474, 395–398 (2011).
Article CAS PubMed PubMed Central Google Scholar
Sakurai, M., Yano, T., Kawabata, H., Ueda, H. & Suzuki, T. Inosine cyanoethylation identifies A-to-I RNA editing sites in the human transcriptome. Nat. Chem. Biol. 6, 733–740 (2010).
Article CAS PubMed Google Scholar
Sakurai, M. et al. A biochemical landscape of A-to-I RNA editing in the human brain transcriptome. Genome Res. 24, 522–534 (2014).
Article CAS PubMed PubMed Central Google Scholar
Safra, M. et al. The m¹A landscape on cytosolic and mitochondrial mRNA at single-base resolution. Nature 551, 251–255 (2017).
Article CAS PubMed Google Scholar
Becker, H. F., Motorin, Y., Planta, R. J. & Grosjean, H. The yeast gene YNL292w encodes a pseudouridine synthase (Pus4) catalyzing the formation of Ψ55 in both mitochondrial and cytoplasmic tRNAs. Nucleic Acids Res. 25, 4493–4499 (1997).
Article CAS PubMed PubMed Central Google Scholar
Mukhopadhyay, S., Deogharia, M. & Gupta, R. Mammalian nuclear TRUB1, mitochondrial TRUB2, and cytoplasmic PUS10 produce conserved pseudouridine 55 in different sets of tRNA. RNA 27, 66–79 (2021).
Article CAS PubMed PubMed Central Google Scholar
Behm-Ansmant, I. et al. The Saccharomyces cerevisiae U2 snRNA: pseudouridine-synthase Pus7p is a novel multisite-multisubstrate RNA: Ψ-synthase also acting on tRNAs. RNA 9, 1371–1382 (2003).
Article CAS PubMed PubMed Central Google Scholar
Carlile, T. M. et al. mRNA structure determines modification by pseudouridine synthase 1. Nat. Chem. Biol. 15, 966–974 (2019).
Article CAS PubMed PubMed Central Google Scholar
Behm-Ansmant, I. et al. A previously unidentified activity of yeast and mouse RNA: pseudouridine synthases 1 (Pus1p) on tRNAs. RNA 12, 1583–1593 (2006).
Article CAS PubMed PubMed Central Google Scholar
Karikó, K., Buckstein, M., Ni, H. P. & Weissman, D. Suppression of RNA recognition by Toll-like receptors: the impact of nucleoside modification and the evolutionary origin of RNA. Immunity 23, 165–175 (2005).
Article PubMed Google Scholar
Fleming, A. M., Mathewson, N. J., Manage, S. A. H. & Burrows, C. J. Nanopore dwell time analysis permits sequencing and conformational assignment of pseudouridine in SARS-CoV-2. ACS Cent. Sci. 7, 1707–1717 (2021).
Article CAS PubMed PubMed Central Google Scholar
Giambruno, R. et al. Unveiling the role of PUS7-mediated pseudouridylation in host protein interactions specific for the SARS-CoV-2 RNA genome. Mol. Ther. Nucleic Acids 34, 102052 (2023).
Article CAS PubMed PubMed Central Google Scholar
Marchand, V. et al. HydraPsiSeq: a method for systematic and quantitative mapping of pseudouridines in RNA. Nucleic Acids Res. 48, e110 (2020).
Article CAS PubMed PubMed Central Google Scholar
Henry, B. A. et al. Pseudouridylation of Epstein–Barr virus noncoding RNA EBER2 facilitates lytic replication. RNA 28, 1542–1552 (2022).
Article CAS PubMed PubMed Central Google Scholar
Martinez, N. M. et al. Pseudouridine synthases modify human pre-mRNA co-transcriptionally and affect pre-mRNA processing. Mol. Cell 82, 645–659 (2022).
Article CAS PubMed PubMed Central Google Scholar
Muller, C. A. et al. Capturing the dynamics of genome replication on individual ultra-long nanopore sequence reads. Nat. Methods 16, 429–436 (2019).
Article CAS PubMed Google Scholar
Ran, F. A. et al. Genome engineering using the CRISPR–Cas9 system. Nat. Protoc. 8, 2281–2308 (2013).
Article CAS PubMed PubMed Central Google Scholar
Wing, P. A. C. et al. Hypoxic and pharmacological activation of HIF inhibits SARS-CoV-2 infection of lung epithelial cells. Cell Rep. 35, 109020 (2021).
Article CAS PubMed PubMed Central Google Scholar
Zhuang, X. D. et al. The circadian clock components BMAL1 and REV-ERBα regulate flavivirus replication. Nat. Commun. 10, 377 (2019).
Article CAS PubMed PubMed Central Google Scholar
Sureau, C. The use of hepatocytes to investigate HDV infection: the HDV/HepaRG model. Methods Mol. Biol. 640, 463–473 (2010).
Article CAS PubMed Google Scholar
Sanz, M. A. & Carrasco, L. Sindbis virus variant with a deletion in the 6K gene shows defects in glycoprotein processing and trafficking: lack of complementation by a wild-type 6K gene in trans. J. Virol. 75, 7778–7784 (2001).
Article CAS PubMed PubMed Central Google Scholar
Castello, A. et al. Insights into RNA biology from an atlas of mammalian mRNA-binding proteins. Cell 149, 1393–1406 (2012).
Article CAS PubMed Google Scholar
Castello, A. et al. System-wide identification of RNA-binding proteins by interactome capture. Nat. Protoc. 8, 491–500 (2013).
Article CAS PubMed Google Scholar
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 10–12 (2011).
Article Google Scholar
Smith, T., Heger, A. & Sudbery, I. UMI-tools: modeling sequencing errors in unique molecular identifiers to improve quantification accuracy. Genome Res. 27, 491–499 (2017).
Article CAS PubMed PubMed Central Google Scholar
Chen, S. F., Zhou, Y. Q., Chen, Y. R. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, 884–890 (2018).
Article Google Scholar
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Article CAS PubMed PubMed Central Google Scholar
Chan, P. P. & Lowe, T. M. GtRNAdb 2.0: an expanded database of transfer RNA genes identified in complete and draft genomes. Nucleic Acids Res. 44, D184–D189 (2016).
Article CAS PubMed Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article CAS PubMed Google Scholar
Kuo, M. Y. P., Chao, M. & Taylor, J. Initiation of replication of the human hepatitis delta virus genome from cloned DNA: role of delta antigen. J. Virol. 63, 1945–1950 (1989).
Article CAS PubMed PubMed Central Google Scholar
Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Article PubMed PubMed Central Google Scholar
Suzuki, T., Ueda, H., Okada, S. & Sakurai, M. Transcriptome-wide identification of adenosine-to-inosine editing using the ICE-seq method. Nat. Protoc. 10, 715–732 (2015).
Article CAS PubMed Google Scholar
Sherry, S. T. et al. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 29, 308–311 (2001).
Article CAS PubMed PubMed Central Google Scholar
Weinberg, Z. & Breaker, R. R. R2R—software to speed the depiction of aesthetic consensus RNA secondary structures. BMC Bioinformatics 12, 3 (2011).
Article CAS PubMed PubMed Central Google Scholar
Bergeron, D. et al. snoDB 2.0: an enhanced interactive database, specializing in human snoRNAs. Nucleic Acids Res. 51, D291–D296 (2022).
Article PubMed Central Google Scholar
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
Article CAS PubMed PubMed Central Google Scholar
Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930 (2014).
Article CAS PubMed Google Scholar
Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 44, W90–W97 (2016).
Article CAS PubMed PubMed Central Google Scholar
Wagih, O. ggseqlogo: a versatile R package for drawing sequence logos. Bioinformatics 33, 3645–3647 (2017).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

We acknowledge T. McMahon for helping with UHPLC–MS/MS experiments. We thank C. Rice (Rockefeller University), S. Urban (University of Heidelberg) and N. Zitzmann (University of Oxford) for the generous provision of Huh-7.5, HepG2-NTCP and Calu-3 cell lines, respectively. Furthermore, we acknowledge C. Rice (Rockefeller University), A. Kohl (CVR, University of Glasgow) and W. James (University of Oxford), for the kind provision of HCV J6/JFH1, ZIKV and SARS-CoV-2 stocks, as well as S. Camille (Université de Tours) for the pT7HB2.7 plasmid. This work was funded by the Ludwig Institute for Cancer Research (to C.-X.S., X.L. and S.K.). The laboratory of C.-X.S. is also supported by National Institute for Health Research (NIHR) Oxford Biomedical Research Centre. The laboratory of J.A.M. is funded by a Wellcome Investigator Award (200838/Z/16/Z), a Wellcome Discovery Award (225198/Z/22/Z) and the Chinese Academy of Medical Sciences Innovation Fund for Medical Science, China (2018-I2M-2-002). A.C. is funded by the European Research Council (ERC) Consolidator Grant ‘vRNP-capture’ 101001634 and the UK Medical Research Council grants MR/R021562/1 and MC_UU_00034/2. H.X. and L.K. are supported by China Scholarship Council. A.E.-B. is funded by Fundación Ramón Areces postdoctoral fellowship program. The views expressed are those of the author(s) and not necessarily those of the National Health Service, the National Institute for Health Research or the Department of Health. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Author information

These authors contributed equally: Haiqi Xu, Linzhen Kong.

Authors and Affiliations

Ludwig Institute for Cancer Research, Nuffield Department of Medicine, University of Oxford, Oxford, UK
Haiqi Xu, Linzhen Kong, Jingfei Cheng, Khatoun Al Moussawi, Xiufei Chen, Aleema Iqbal, Skirmantas Kriaucionis, Xin Lu & Chun-Xiao Song
Target Discovery Institute, Nuffield Department of Medicine, University of Oxford, Oxford, UK
Haiqi Xu, Linzhen Kong, Jingfei Cheng, Xiufei Chen, Aleema Iqbal & Chun-Xiao Song
Chinese Academy of Medical Sciences Oxford Institute, Nuffield Department of Medicine, University of Oxford, Oxford, UK
Peter A. C. Wing
Nuffield Department of Medicine, University of Oxford, Oxford, UK
James M. Harris, Senko Tsukuda & Jane A. McKeating
MRC University of Glasgow Centre for Virus Research, Glasgow, UK
Azman Embarc-Buh & Alfredo Castello
Department of Biochemistry, University of Oxford, Oxford, UK
Guifeng Wei

Authors

Haiqi Xu
View author publications
You can also search for this author in PubMed Google Scholar
Linzhen Kong
View author publications
You can also search for this author in PubMed Google Scholar
Jingfei Cheng
View author publications
You can also search for this author in PubMed Google Scholar
Khatoun Al Moussawi
View author publications
You can also search for this author in PubMed Google Scholar
Xiufei Chen
View author publications
You can also search for this author in PubMed Google Scholar
Aleema Iqbal
View author publications
You can also search for this author in PubMed Google Scholar
Peter A. C. Wing
View author publications
You can also search for this author in PubMed Google Scholar
James M. Harris
View author publications
You can also search for this author in PubMed Google Scholar
Senko Tsukuda
View author publications
You can also search for this author in PubMed Google Scholar
Azman Embarc-Buh
View author publications
You can also search for this author in PubMed Google Scholar
Guifeng Wei
View author publications
You can also search for this author in PubMed Google Scholar
Alfredo Castello
View author publications
You can also search for this author in PubMed Google Scholar
Skirmantas Kriaucionis
View author publications
You can also search for this author in PubMed Google Scholar
Jane A. McKeating
View author publications
You can also search for this author in PubMed Google Scholar
Xin Lu
View author publications
You can also search for this author in PubMed Google Scholar
Chun-Xiao Song
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

H.X. and C.-X.S. conceived and designed the study. H.X. performed the experiments with the help from L.K., K.A.M., X.C., A.I., S.K. and X.L. P.A.C.W., J.M.H., S.T. and J.A.M. performed isolation of SARS-CoV-2, HCV, ZIKV and HDV RNA. A.E.-B., G.W. and A.C. performed isolation of SINV RNA. L.K. performed the computational analysis with the help from J.C. H.X., L.K. and C.-X.S. wrote the paper.

Corresponding author

Correspondence to Chun-Xiao Song.

Ethics declarations

Competing interests

C.-X.S. and H.X. are named as inventors on pending patent applications filed by the Ludwig Institute for Cancer Research for the technologies described here. The other authors declare no competing interests.

Peer review

Peer review information

Nature Methods thanks Tsutomu Suzuki and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: Lei Tang, in collaboration with the Nature Methods team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Figs. 1–17 and Supplementary Tables 1 and 2.

Reporting Summary

Supplementary Table 3–13

Source data

Source Data Fig. 1

Statistical source data.

Source Data Fig. 2

Statistical source data.

Source Data Fig. 3

Statistical source data.

Source Data Fig. 4

Statistical source data.

Source Data Fig. 5

Statistical source data.

Source Data Fig. 6

Statistical source data.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Xu, H., Kong, L., Cheng, J. et al. Absolute quantitative and base-resolution sequencing reveals comprehensive landscape of pseudouridine across the human transcriptome. Nat Methods (2024). https://doi.org/10.1038/s41592-024-02439-8

Download citation

Received: 13 September 2023
Accepted: 04 September 2024
Published: 30 September 2024
DOI: https://doi.org/10.1038/s41592-024-02439-8

Subjects

Abstract

Similar content being viewed by others

Main

Results

Development of BACS

Validation of BACS on human rRNA

BACS identified dense Ψ sites in human spliceosomal snRNAs

BACS revealed the Ψ profile of human snoRNA

A comprehensive Ψ map of human tRNA

Profiling and quantification of Ψ in HeLa mRNA

Simultaneous profiling of A-to-I editing sites and m1A

BACS uncovered new PUS targets in HeLa cells

Mapping of Ψ in viral RNAs

Discussion

Methods

Preparation of model RNA

Mass spectrometry analysis of short oligonucleotides

Quantification of Ψ level by UHPLC–MS/MS

Cell culture

Generation of CRISPR knockout cell lines

RNA isolation

Viral infection and RNA isolation

SARS-CoV-2

HCV and ZIKV

HDV

SINV

Preparation of IVT RNA control

BACS for Ψ detection

Modified BID-seq

Data preprocessing

Read alignment

Calling Ψ sites

Calling A-to-I editing sites

RNA structure visualization

Downstream analysis

Calling PUS-dependent Ψ sites

Published data

Reporting summary

Data availability

Code availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

Search

Quick links

Simultaneous profiling of A-to-I editing sites and m¹A