English

• In addition to the four canonical nucleosides of adenosine (A), guanosine (G), cytidine (C) and uridine (U), RNA molecules also contain diverse chemical modifications [1–3]. Since the first modification was characterized over 60 years ago, more than 150 chemical modifications have been identified in a variety of RNA species from three-domain of life [4–8]. RNA modifications occur in ribosomal RNA (rRNA), transfer RNA (tRNA), messenger RNA (mRNA), and non-coding RNAs (ncRNAs) [9–11]. Similar to epigenetic DNA methylation, RNA modifications have their own writers, erasers, and readers [12,13]. Now, the dynamic RNA modifications have been viewed as new posttranscriptional regulator in modulating gene expression as well as in a broad range of physiological processes [14–19].

  N⁶-methyladenosine (m⁶A) is the best-characterized internal modification in eukaryotic mRNA [20]. As an isomer of m⁶A, N¹-methyladenosine (m¹A) has also been found to be present in all the three kingdoms of life [21]. m¹A is a reversible modification occurring in tRNA, rRNA, mRNA, and mitochondrial RNA [22,23]. The methyl group in m¹A could disturb the formation of the Watson–Crick base pair (Fig. 1A), which would therefore affect the secondary structure of RNA and the interaction with its reader proteins. It has been reported that m¹A is involved in the folding of tRNA [24–26], maturation of ribosomes [27], and translation process [28,29].

  Figure 1

  

  Figure 1.  Schematic illustration of the principle of DA-m¹A method. (A) N1-methyl group in m¹A can disturb the formation of hydrogen bonding between m¹A and thymine, while adenosine and thymine can form classic Watson−Crick base pairing. (B) In DA-m¹A method, m¹A in RNA will cause RT stalling in m¹A, producing truncated cDNA. Removal of m¹A by AlkB will lead to the generation of full-length cDNA. Evaluation of the produced amounts of full-length cDNA by qPCR could realize site-specific detection and quantification of m¹A in RNA. (C, D) LC-MS/MS analysis of the demethylation efficiency toward m¹A by AlkB treatment. Error bars indicate mean ± SD with three replicates.

  DownLoad: Full-Size Img PowerPoint

  In recent years, mass spectrometry-based and sequencing-based techniques have been developed to discover, identify, and map nucleic acid modifications [30–37]. The presence of N1-methyl group in m¹A could induce stop or misincorporation feature in reverse transcription (RT), which was employed for the transcriptome-wide mapping of m¹A while in combination with high-throughput sequencing [28,29]. However, these high-throughput sequencing-based methods generally require complex sample pretreatment and sequencing library construction, which is time-consuming, costly and needs sophisticated bioinformatics analysis. Mass spectrometry-based analysis generally offers the identification and quantification of bulky RNA modifications [38–40]. However, the site information of RNA modifications typically couldn’t be obtained by mass spectrometry-based methods. Some studies aim to evaluate m¹A at certain critical sites with stoichiometry information. In this respect, a method that enables rapid, straightforward, site-specific detection and quantification of m¹A is highly desired. m¹A carries a methyl group in the N1 position of adenine, which can compromise the formation of hydrogen bonding between m¹A and thymine (Fig. 1A). By virtue of this property of m¹A, we developed a demethylase-assisted method (DA-m¹A) for the site-specific assessment of m¹A stoichiometry in RNA. The presence N1-methyl group in m¹A could cause the stalling of reverse transcription at m¹A site, thus producing the truncated cDNA. E. coli AlkB is a demethylase that acts as a repair protein and can demethylate m¹A [41,42]. Since m¹A can cause the stalling of reverse transcription and produce truncated cDNA in RT reaction, we envisioned that AlkB-assisted demethylation of m¹A to produce adenine in RNA would generate full-length cDNA. Therefore, evaluation of the produced amounts of full-length cDNA by qPCR could realize site-specific detection and quantification of m¹A in RNA (Fig. 1B).

  We first expressed and purified the recombinant AlkB protein (Table S1, Figs. S1 and S2 in Supporting information). We employed LC-MS/MS to evaluate the demethylation efficiency of m¹A by AlkB protein (Table S2 in Supporting information). Since the bromodomain containing 2 (BRD2) mRNA was proved to contain m¹A at position 1674 [43], a 46-nt RNA (Oligo-m¹A, Table S3 in Supporting information) that has the same sequence context around the position 1674 of BRD2 mRNA was synthesized and used as the substrate for the evaluation. Oligo-m¹A was treated with AlkB followed by enzymatic digestion and LC-MS/MS analysis. The result showed that m¹A was almost undetectable after AlkB treatment (Fig. 1C). Quantification of the trace level of m¹A demonstrated that 99.95% demethylation efficiency toward m¹A was achieved upon AlkB treatment (Fig. 1D). The highly efficient demethylation of m¹A by AlkB indicates the high feasibility for the development of the DA-m¹A method.

  Two 46-nt RNAs of Oligo-m¹A and Oligo-A (Table S3) were used as the templates for the optimization reaction. A 78-nt DNA (Preamp oligo) was designed to hybridize with Oligo-m¹A and Oligo-A for the RT reaction. The RT stalling efficiencies were calculated using the equation of ΔCt = Ct_m1A − Ct_A (Figs. 2A and B).

  Figure 2

  

  Figure 2.  Optimization of the RT reaction in DA-m¹A method. (A) Schematic illustration showing that differential Ct values could be obtained from Oligo-m¹A (Ct_m1A) and Oligo-A (Ct_A) in DA-m¹A method. The ΔCt (Ct_m1A-Ct_A) values are used to evaluate the RT stalling efficiency. (B) Representative real-time fluorescence amplification curves with using Oligo-m¹A and Oligo-A templates. (C) Evaluation of the RT stalling efficiencies by different reverse transcriptases under different temperatures. (D) Evaluation of the RT stalling efficiencies by different reverse transcriptases with PAGE analysis. 22 cycles were carried out for PCR. (E−G) Optimization of the concentrations of Protoscript Ⅱ reverse transcriptase, dNTP, and Mg²⁺. Error bars indicate mean ± SD with three replicates.

  DownLoad: Full-Size Img PowerPoint

  Based on the ΔCt values, we first evaluated the RT stalling efficiency of m¹A by five commercial reverse transcriptases, including ProtoScript Ⅱ reverse transcriptase, Bst 3.0 DNA polymerase, SuperScript Ⅲ reverse transcriptase, M-MuLV reverse transcriptase, and AMV reverse transcriptase. The results showed that larger ΔCt values were obtained with using ProtoScript Ⅱ reverse transcriptase and Bst 3.0 DNA polymerase, indicating m¹A would lead to stronger RT stalling by these two transcriptases (Fig. 2C). We then optimized the reaction temperatures of these five reverse transcriptases from 37 ℃ to 70 ℃. The results showed that ProtoScript Ⅱ reverse transcriptase exhibited overall the best discrimination between m¹A and A under the temperature ranging from 37 ℃ to 52 ℃ (Fig. 2C). Similar to the qPCR results, the PAGE analysis also suggested that m¹A exhibited stronger RT stalling efficiency with using Protoscript Ⅱ reverse transcriptase (Fig. 2D). As expected, the band intensities from the AlkB-treated Oligo-m¹A template were comparable to the Oligo-A template, indicating that AlkB treatment would efficiently remove the methyl group of m¹A and therefore eliminate the RT stalling caused by m¹A (Fig. 2D).

  Since ProtoScript Ⅱ reverse transcriptase shows the best performance in discriminating m¹A and A under 42 ℃ (Fig. 2C), we chose this enzyme for the method development. To obtain the best discrimination, we further optimized the conditions of the RT reactions, including the concentrations of Protoscript Ⅱ reverse transcriptase, dNTP, and Mg²⁺. With the increased concentration of Protoscript Ⅱ reverse transcriptase, the ΔCt values increased and slightly decreased over 0.3 U/µL (Fig. 2E). The concentrations of dNTP and Mg²⁺also affected the discrimination between m¹A and A. The results showed that 0.5 mmol/L dNTP and 3 mmol/L Mg²⁺ offered the best discrimination (Figs. 2F and G). Collectively, the optimal RT reaction was conducted under 42 ℃ with 0.5 mmol/L dNTP, 3 mmol/L MgCl₂ and 0.3 U/µL Protoscript Ⅱ reverse transcriptase.

  Having optimized the conditions of the DA-m¹A method, we evaluated its quantitative ability toward m¹A at given sites in RNA. We first constructed a calibration curve using different concentrations of Oligo-A and Oligo-m¹A from 10⁻¹⁵ mol/L to 10⁻⁹ mol/L (Fig. S3 in Supporting information). The Ct values were plotted against the logarithm (lg) of the concentrations of Oligo-A or Oligo-m¹A. The resulting equations of Ct_A = −3.005lgC_Oligo-A – 14.681 and Ct_m1A = −3.028lgC_Ologo-m1A − 11.360 showed good linearity with R² being 0.997 (Figs. S4A and B in Supporting information). The obtained ΔCt (Ct_m1A − Ct_A) were similar under different concentrations of Oligo-A and Oligo-m¹A (from 10⁻¹⁵ mol/L to 10⁻⁹ mol/L, Fig. S4C in Supporting information), indicating the DA-m¹A method can be applied in the detection of m¹A with a wide range of concentrations of RNA substrates.

  We next quantitatively assessed the m¹A level in individual sites by the DA-m¹A method. In this respect, the mixtures of Oligo-A and Oligo-m¹A with various molar ratios of Oligo-m¹A/(Oligo-m¹A + Oligo-A) were prepared and subjected to AlkB treatment. The ΔCt (Ct_AlkB– – Ct_AlkB+) values between untreated mixtures (Ct_AlkB–) and AlkB-treated mixtures (Ct_AlkB–) were plotted against the theoretical percentages of m¹A fraction (Fig. 3A). A non-linear fitting curve (y = 1.156x/(100 – 0.675x)) was generated with R² being 0.998 (Fig. 3B). With this equation, we quantified the m¹A level at specific loci in RNA of HEK293T cells (Fig. S5 in Supporting information). We first examined a control site without m¹A modification (position 316) in 28S rRNA. As shown in Fig. S6A (Supporting information), the ΔCt (0.02) was extremely low, indicating low false positive rate of the DA-m¹A method. We then examined the previously well-characterized m¹A site (position 1322) in 28S rRNA [44]. The result showed a ΔCt of 3.49 was obtained, and the m¹A level was calculated to be 99.4% ± 1.2% (Fig. 3C and Fig. S6B in Supporting information), which was consistent with the previous study [44]. m¹A modification occurs in different RNA species, including mRNA, ncRNA, and mitochondrial RNA [22,23]. We also applied the DA-m¹A method to quantify m¹A level in individual sites of different RNA species. Some previously reported m¹A sites, including BRD2 mRNA position 1674 (BRD2: m¹A1674), CST telomere replication complex component 1 (CTC1) mRNA position 5643 (CTC1: m¹A5643), ncRNA prune exopolyphosphatase 1 (PRUNE1) position 58 (PRUNE: m¹A58), long intergenic non-protein coding RNA 324 (LINC00324) position 659 (LINC00324: m¹A659), mitochondrially encoded NADH: ubiquinone oxidoreductase core subunit 5 mRNA (MT-ND5: m¹A1374) and mitochondrially encoded cytochrome C oxidase Ⅱ mRNA (MT-CO2: m¹A297). The measured levels of m¹A at BRD2: m¹A1674, CTC1: m¹A5643, PRUNE: m¹A58, LINC00324: m¹A659, MT-ND5: m¹A1374, and MT-CO2: m¹A297 were 77.9% and 68.8%, 46.1%, 78.6%, 59.7%, and 66.3%, respectively (Fig. 3C and Figs. S6C-H in Supporting information). The quantification results of these m¹A sites are comparable to previously reported data (91% for BRD2: m¹A1674, 84% for CTC1: m¹A5643, 49% for PRUNE: m¹A58, 82% for LINC00324: m¹A659, 67% for MT-ND5: m¹A1374, and 76% for MT-CO2: m¹A297) [44–46]. Taken together, these results suggest that the DA-m¹A method is capable for site-specific assessment of m¹A stoichiometry in RNA.

  Figure 3

  

  Figure 3.  Assessment of m¹A stoichiometry at given sites in RNA by DA-m¹A method. (A) Schematic illustration showing the assessment of m¹A stoichiometry at given sites in RNA by DA-m¹A method. The mixtures of Oligo-A and Oligo-m¹A with various molar ratios were prepared and subjected to AlkB treatment. (B) The ΔCt (Ct_AlkB– – Ct_AlkB+) values between untreated mixtures (Ct_AlkB–) and AlkB-treated mixtures (Ct_AlkB–) were plotted against the theoretical percentages of m¹A fraction. (C) Assessment of m¹A stoichiometry at specific loci in RNA of HEK293T cells by DA-m¹A method. Error bars indicate mean ± SD with three replicates.

  DownLoad: Full-Size Img PowerPoint

  Nucleic acid modifications are known to be involved in a variety of human cancer development [47–49]. It has been reported that m¹A in tRNA is elevated in human hepatocellular carcinoma (HCC) tissues and drives liver tumorigenesis [50]. However, site-specific quantification of the change of m¹A in RNA of HCC tissues is lacked.

  Here we explored the correlation of the levels of m¹A at specific sites in RNA with HCC. The total RNA from 5 pairs of human HCC tissues and their matched tumor-adjacent normal tissues were isolated and subjected to DA-m¹A analysis (Table S4 in Supporting information). An approval for the study was granted by the Wuhan University Ethics Committee and met the declaration of Helsinki. All experiments were conducted in accordance with Wuhan University Ethics Committee’s guidelines and regulations. The aforementioned six m¹A sites were evaluated (Tables S5 and S6 in Supporting information). The results showed that, compared to the tumor-adjacent normal tissues, m¹A level in HCC tissues was significantly decreased at BRD2: m¹A1674 (P = 0.013) and CTC1: m¹A5643 (P = 0.004) (Figs. 4A and B, Figs. S7A and B in Supporting information). m¹A level in the other 4 sites showed no significant change between HCC tissues and tumor-adjacent normal tissues (Figs. 4C–F and Figs. S7C–F in Supporting information). The results suggested that the two m¹A sites in mRNA (BRD2: m¹A1674 and CTC1: m¹A5643) may be involved in liver tumorigenesis.

  Figure 4

  

  Figure 4.  Site-specific quantification of m¹A in RNA of human HCC tissues by DA-m¹A method. Total RNA from five pairs of human HCC tissues and their matched tumor-adjacent normal tissues were isolated and subjected to DA-m¹A analysis. (A) BRD2: m¹A1674. (B) CTC1: m¹A5643. (C) PRUNE: m¹A58. (D) LINC00324: m¹A659. (E) MT-ND5: m¹A1374. (F) MT-CO2: m¹A297. All p values were evaluated by a two-sided paired t-test.

  DownLoad: Full-Size Img PowerPoint

  LC-MS/MS analysis only could provide the bulky level of m¹A, which is not able to accurately reflect the site-specific change of m¹A between samples. Indeed, the bulky m¹A level in the total RNA measured by LC-MS/MS showed no significant change between human HCC tissues and their matched tumor-adjacent normal tissues (P=0.930, Fig. S8 in Supporting information). Although the bulky m¹A level may not have obvious change, the levels of m¹A at certain sites may have significant changes between cancer tissues and normal tissues. Since the DA-m¹A method enables the site-specific quantification of m¹A with stoichiometry information, it allows the evaluation of precise alteration of m¹A at given sites. The DA-m¹A is a well complementary method to LC-MS/MS method. In addition, the DA-m¹A method is also suitable to be employed to confirm the m¹A sites determined by transcriptome-wide mapping methods. The limitation of the DA-m¹A method is that it relies on prior knowledge of sequence information flanking the target m¹A.

  In summary, we proposed the DA-m¹A method for the site-specific assessment of m¹A stoichiometry in RNA from biological and clinical samples. With the DA-m¹A method, we examined and successfully confirmed the previously well-characterized m¹A sites in 28S rRNA, mRNA, ncRNA, and mitochondrial RNA with low false positive rate. In addition, we found the level of m¹A was significantly decreased at BRD2 mRNA position 1674 and CTC1 mRNA position 5643 in human HCC tissues, suggesting that these two m¹A sites in mRNA might be involved in liver tumorigenesis. Collectively, the developed DA-m¹A method offers a straightforward, cost-effective, and site-specific detection of m¹A with stoichiometry. The DA-m¹A method is simple and rapid, which is especially suitable for the evaluation of the status of certain critical m¹A sites in RNA from a plenty number of biological or clinical samples.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  The work is supported by the National Key R&D Program of China (Nos. 2022YFC3400700 and 2022YFA0806600), the National Natural Science Foundation of China (Nos. 22277093, 22074110, 21721005 and 22207090), the Interdisciplinary Innovative Talents Foundation from Renmin Hospital of Wuhan University (No. JCRCGW-2022–008), and the Translational Medicine and Interdisciplinary Research Joint Fund of Zhongnan Hospital of Wuhan University (No. ZNJC202208).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108953.

  
  
• 1. [1]

     D. Wiener, S. Schwartz, Nat. Rev. Genet. 22 (2021) 119–131.

  2. [2]

     S. Moshitch-Moshkovitz, D. Dominissini, G. Rechavi, Cell 185 (2022) 764–776. doi: 10.1016/j.cell.2022.02.007

  3. [3]

     Y. Zhao, X. Zuo, Q. Li, et al., Sci. China Chem. 64 (2021) 171–203. doi: 10.1007/s11426-020-9864-7

  4. [4]

     P. Boccaletto, F. Stefaniak, A. Ray, et al., Nucleic Acids Res. 50 (2022) D231–D235. doi: 10.1093/nar/gkab1083

  5. [5]

     M.Y. Chen, Z. Gui, K.K. Chen, et al., Chin. Chem. Lett. 33 (2022) 2086–2090. doi: 10.1016/j.cclet.2021.08.094

  6. [6]

     X.J. You, S. Zhang, J.J. Chen, et al., Nucleic Acids Res. 50 (2022) 9858–9872. doi: 10.1093/nar/gkac770

  7. [7]

     Y. Dai, C.B. Qi, Y. Feng, et al., Anal. Chem. 93 (2021) 6938–6946. doi: 10.1021/acs.analchem.0c04630

  8. [8]

     M.Y. Cheng, X.J. You, J.H. Ding, et al., Chem. Sci. 12 (2021) 8149–8156. doi: 10.1039/D1SC01972D

  9. [9]

     Y. Motorin, M. Helm, Wiley Interdiscip. Rev. RNA 13 (2022) e1691. doi: 10.1002/wrna.1691

  10. [10]

      J.H. Ding, M.Y. Chen, N.B. Xie, et al., Biosens. Bioelectron. 219 (2023) 114821. doi: 10.1016/j.bios.2022.114821

  11. [11]

      M.Y. Chen, C.B. Qi, X.M. Tang, et al., Chin. Chem. Lett. 33 (2022) 3772–3776. doi: 10.1016/j.cclet.2021.12.008

  12. [12]

      T.R. Fischer, L. Meidner, M. Schwickert, et al., Nucleic Acids Res. 50 (2022) 4216–4245. doi: 10.1093/nar/gkac224

  13. [13]

      F. Nie, P. Feng, X. Song, et al., Database 2020 (2020) baaa049. doi: 10.1093/database/baaa049

  14. [14]

      H.L. Huang, H.Y. Weng, X.L. Deng, J.J. Chen, Annu. Rev. Cancer Biol. 4 (2020) 221–240. doi: 10.1146/annurev-cancerbio-030419-033357

  15. [15]

      K. Chen, B.S. Zhao, C. He, Cell Chem. Biol. 23 (2016) 74–85. doi: 10.1016/j.chembiol.2015.11.007

  16. [16]

      H. Shi, P. Chai, R. Jia, X. Fan, Mol. Cancer 19 (2020) 78. doi: 10.1186/s12943-020-01194-6

  17. [17]

      M. Frye, B.T. Harada, M. Behm, C. He, Science 361 (2018) 1346–1349. doi: 10.1126/science.aau1646

  18. [18]

      Y.Y. Chen, Z. Gui, D. Hu, et al., Chin. Chem. Lett. 34 (2023) 108522.

  19. [19]

      J.J. Chen, X.J. You, L. Li, et al., Anal. Chem. 94 (2022) 8740–8747. doi: 10.1021/acs.analchem.2c01226

  20. [20]

      S. Murakami, S.R. Jaffrey, Mol. Cell. 82 (2022) 2236–2251. doi: 10.1016/j.molcel.2022.05.029

  21. [21]

      J. Li, H. Zhang, H. Wang, Comput. Struct. Biotechnol. J. 20 (2022) 6578–6585. doi: 10.1016/j.csbj.2022.11.045

  22. [22]

      C. Zhang, G. Jia, Genom. Proteom. Bioinform. 16 (2018) 155–161. doi: 10.1016/j.gpb.2018.03.003

  23. [23]

      X. Xiong, X. Li, C. Yi, Curr. Opin. Chem. Biol. 45 (2018) 179–186. doi: 10.1016/j.cbpa.2018.06.017

  24. [24]

      S. Ozanick, A. Krecic, J. Andersland, J.T. Anderson, RNA 11 (2005) 1281–1290. doi: 10.1261/rna.5040605

  25. [25]

      E. Vilardo, C. Nachbagauer, A. Buzet, et al., Nucleic Acids Res. 40 (2012) 11583–11593. doi: 10.1093/nar/gks910

  26. [26]

      T. Chujo, T. Suzuki, RNA 18 (2012) 2269–2276. doi: 10.1261/rna.035600.112

  27. [27]

      T. Waku, Y. Nakajima, W. Yokoyama, et al., J. Cell Sci. 129 (2016) 2382–2393.

  28. [28]

      X. Li, X. Xiong, K. Wang, et al., Nat. Chem. Biol. 12 (2016) 311–316. doi: 10.1038/nchembio.2040

  29. [29]

      D. Dominissini, S. Nachtergaele, S. Moshitch-Moshkovitz, et al., Nature 530 (2016) 441–446. doi: 10.1038/nature16998

  30. [30]

      Y. Zhang, L. Lu, X. Li, Exp. Mol. Med. 54 (2022) 1601–1616. doi: 10.1038/s12276-022-00821-0

  31. [31]

      H. Liu, Y. Wang, X. Zhou, RSC Chem. Biol. 3 (2022) 994–1007. doi: 10.1039/D2CB00087C

  32. [32]

      A.S. Felix, A.L. Quillin, S. Mousavi, J.M. Heemstra, Acc. Chem. Res. 55 (2022) 2271–2279. doi: 10.1021/acs.accounts.2c00287

  33. [33]

      D. Bartee, S. Thalalla Gamage, C.N. Link, J.L. Meier, Chem. Soc. Rev. 50 (2021) 9482–9502. doi: 10.1039/D1CS00214G

  34. [34]

      J.D. Alfonzo, J.A. Brown, P.H. Byers, et al., Nat. Genet. 53 (2021) 1113–1116. doi: 10.1038/s41588-021-00903-1

  35. [35]

      X.J. You, L. Li, T.T. Ji, et al., Chin. Chem. Lett. 34 (2023) 107181. doi: 10.1016/j.cclet.2022.01.074

  36. [36]

      Q. Wang, J.H. Ding, J. Xiong, et al., Chin. Chem. Lett. 32 (2021) 3426–3430. doi: 10.1016/j.cclet.2021.05.020

  37. [37]

      Q.F. Zhang, H.M. Xiao, J.T. Zhan, B.F. Yuan, Y.Q. Feng, Chin. Chem. Lett. 33 (2022) 4746–4749. doi: 10.1016/j.cclet.2022.01.004

  38. [38]

      B. Chen, B.F. Yuan, Y.Q. Feng, Anal. Chem. 91 (2019) 743–756. doi: 10.1021/acs.analchem.8b04078

  39. [39]

      X.M. Tang, T.T. Ye, X.J. You, et al., Chin. Chem. Lett. 34 (2023) 107531. doi: 10.1016/j.cclet.2022.05.045

  40. [40]

      W.B. Tao, N.B. Xie, Q.Y. Cheng, Y.Q. Feng, B.F. Yuan, Chin. Chem. Lett. 34 (2023) 108243. doi: 10.1016/j.cclet.2023.108243

  41. [41]

      S.C. Trewick, T.F. Henshaw, R.P. Hausinger, T. Lindahl, B. Sedgwick, Nature 419 (2002) 174–178. doi: 10.1038/nature00908

  42. [42]

      P.O. Falnes, R.F. Johansen, E. Seeberg, Nature 419 (2002) 178–182. doi: 10.1038/nature01048

  43. [43]

      M. Safra, A. Sas-Chen, R. Nir, et al., Nature 551 (2017) 251–255. doi: 10.1038/nature24456

  44. [44]

      H. Zhou, S. Rauch, Q. Dai, et al., Nat. Methods 16 (2019) 1281–1288. doi: 10.1038/s41592-019-0550-4

  45. [45]

      X. Li, X. Xiong, M. Zhang, et al., Mol. Cell. 68 (2017) 993–1005.e9. doi: 10.1016/j.molcel.2017.10.019

  46. [46]

      J.H. Ding, C.J. Ma, M.Y. Chen, et al., Anal. Chem. 92 (2020) 2612–2619. doi: 10.1021/acs.analchem.9b04454

  47. [47]

      I. Barbieri, T. Kouzarides, Nat. Rev. Cancer 20 (2020) 303–322. doi: 10.1038/s41568-020-0253-2

  48. [48]

      J. Xiong, K.K. Chen, N.B. Xie, et al., Anal. Chem. 94 (2022) 15489–15498. doi: 10.1021/acs.analchem.2c03808

  49. [49]

      N.B. Xie, M. Wang, T.T. Ji, et al., Chem. Sci. 13 (2022) 7046–7056. doi: 10.1039/D2SC01052F

  50. [50]

      Y. Wang, J. Wang, X. Li, et al., Nat. Commun. 12 (2021) 6314. doi: 10.1038/s41467-021-26718-6

• 

  Figure 1  Schematic illustration of the principle of DA-m¹A method. (A) N1-methyl group in m¹A can disturb the formation of hydrogen bonding between m¹A and thymine, while adenosine and thymine can form classic Watson−Crick base pairing. (B) In DA-m¹A method, m¹A in RNA will cause RT stalling in m¹A, producing truncated cDNA. Removal of m¹A by AlkB will lead to the generation of full-length cDNA. Evaluation of the produced amounts of full-length cDNA by qPCR could realize site-specific detection and quantification of m¹A in RNA. (C, D) LC-MS/MS analysis of the demethylation efficiency toward m¹A by AlkB treatment. Error bars indicate mean ± SD with three replicates.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Optimization of the RT reaction in DA-m¹A method. (A) Schematic illustration showing that differential Ct values could be obtained from Oligo-m¹A (Ct_m1A) and Oligo-A (Ct_A) in DA-m¹A method. The ΔCt (Ct_m1A-Ct_A) values are used to evaluate the RT stalling efficiency. (B) Representative real-time fluorescence amplification curves with using Oligo-m¹A and Oligo-A templates. (C) Evaluation of the RT stalling efficiencies by different reverse transcriptases under different temperatures. (D) Evaluation of the RT stalling efficiencies by different reverse transcriptases with PAGE analysis. 22 cycles were carried out for PCR. (E−G) Optimization of the concentrations of Protoscript Ⅱ reverse transcriptase, dNTP, and Mg²⁺. Error bars indicate mean ± SD with three replicates.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Assessment of m¹A stoichiometry at given sites in RNA by DA-m¹A method. (A) Schematic illustration showing the assessment of m¹A stoichiometry at given sites in RNA by DA-m¹A method. The mixtures of Oligo-A and Oligo-m¹A with various molar ratios were prepared and subjected to AlkB treatment. (B) The ΔCt (Ct_AlkB– – Ct_AlkB+) values between untreated mixtures (Ct_AlkB–) and AlkB-treated mixtures (Ct_AlkB–) were plotted against the theoretical percentages of m¹A fraction. (C) Assessment of m¹A stoichiometry at specific loci in RNA of HEK293T cells by DA-m¹A method. Error bars indicate mean ± SD with three replicates.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Site-specific quantification of m¹A in RNA of human HCC tissues by DA-m¹A method. Total RNA from five pairs of human HCC tissues and their matched tumor-adjacent normal tissues were isolated and subjected to DA-m¹A analysis. (A) BRD2: m¹A1674. (B) CTC1: m¹A5643. (C) PRUNE: m¹A58. (D) LINC00324: m¹A659. (E) MT-ND5: m¹A1374. (F) MT-CO2: m¹A297. All p values were evaluated by a two-sided paired t-test.

  下载: 全尺寸图片 幻灯片
•