English

• The therapeutic landscape is undergoing a major transformation, with mRNA therapy emerging as a key player [1]. However, mRNA's short cytoplasmic half-life remains a critical limitation. Circular RNA (circRNA) presents a promising alternative due to its superior stability and translation capacity. Unlike linear mRNA, circRNA forms a covalently closed loop through back-splicing, eliminating requirements for 5′ caps and 3′ polyA tails. This unique structure enables internal ribosome entry site (IRES)-mediated translation while conferring exonuclease resistance, resulting in prolonged stability-making circRNA an attractive therapeutic candidate [2]. However, achieving efficient and precise in vitro synthesis of circRNA remains a challenge. Moreover, preventing unwarranted immune reactions during the synthesis process is also a significant obstacle for the scientific community [3, 4].

  Research on in vitro circularization of RNA has spanned more than a decade, marking a significant paradigm shift from chemical synthesis to the current strategies [5, 6]. These strategies are broadly categorized into three distinct methods: chemical synthesis using cyanogen bromide or similar coupling agents, enzymatic approaches that employ DNA or RNA ligases, and ribozyme methods that leverage the self-splicing capabilities of introns [7]. Each method offers unique advantages and challenges in the pursuit of effective RNA circularization.

  Oligonucleotide chemical conjugation strategies have been evolving for decades, propelled by the advent of small nucleic acid drugs [8-10]. The industry has conducted extensive research on potential nucleic acid molecule reactions, such as the functionalization of the 5′ or 3′ termini with amines, thiols, azides, and cyanogen bromide for the purpose of circularization. Simple click chemistry, including azide-alkyne cycloaddition (ACC), is acknowledged for its efficiency in rapidly circularizing nucleic acids (Fig. 1a) [11, 12]. However, chemical ligation is seldom used for circRNA preparation compared to other well-established chemical synthesis methods for small nucleic acids. This is primarily due to the severe side reactions that can occur during the chemical ligation process, leading to the formation of 2′, 5′-phosphodiester bonds rather than the natural 3′, 5′-phosphate bonds [13]. In contrast, a variety of ligases originating from bacteriophage T4 are capable of catalyzing RNA ligation reactions, such as T4 DNA ligase (T4 Dnl), T4 RNA ligase 1 (T4 Rnl 1), and T4 RNA ligase 2 (T4 Rnl 2). This ligation process necessitates a 3′-OH group on the acceptor substrate and a 5′ monophosphate on the donor substrate (Fig. 1b) [13-18].

  Figure 1

  

  Figure 1.  In vitro circularization of RNA. (a) Chemical ligation of RNA molecules. (b) Enzymatic ligation of RNA molecules. Reproduced with permission [32]. Copyright 2022, Elsevier B.V.

  DownLoad: Full-Size Img PowerPoint

  Another important approach in RNA circularization involves ribozymes, a class of RNAs with enzymatic catalytic activities. Ribozymes play a crucial role in RNA circularization. One frequently used method is the permuted intron-exon (PIE) system, which incorporates type Ⅰ or type Ⅱ introns. This method leverages the self-splicing capability of these introns to perform splicing in an environment rich with magnesium ions and free GTP. This leads to the circularization of introns and the ligation of intermediate sequences, ultimately producing circRNA (Figs. 2a and b) [19-24]. Furthermore, hairpin ribozymes (HPR), originating from viroids and the hepatitis delta virus, engage a single-stranded DNA template to execute rolling circle transcription reactions. This process yields a lengthy, repetitive linear RNA precursor that undergoes self-cleavage, ultimately resulting in the formation of small circRNAs (Fig. 2c) [25-27].

  Figure 2

  

  Figure 2.  Ribozyme-mediated in vitro synthesis of circular RNA. (a) autonomous splicing of group Ⅰ introns in the group Ⅰ intron self-splicing system. (b) Self-splicing capability of group Ⅱ Introns in the group Ⅱ intron self-splicing system. (c) HPR conformation. Reproduced with permission [2]. Copyright 2021, Chen and Lu. Copied with permission [32]. Copyright 2022, Elsevier B.V.

  DownLoad: Full-Size Img PowerPoint

  Given the immense therapeutic potential of circRNA, the exploration of efficient in vitro synthesis methods becomes crucial. Seong-Wook Lee's team has developed an innovative strategy for in vitro circRNA production, enhancing the design of self-targeting and splicing (STS) circularization vectors. These vectors, distinguished by the length of the P1 helix and the inclusion of the P10 helix, were initially engineered with extended P1 and P10 helices, internal guide sequences (IGS), and genes of interest (GOI). Subsequent refinements led to a streamlined design that incorporated solely the P1 helix and GOI. Research demonstrated that the P1 helix was capable of efficiently generating circRNA through the STS reaction in vitro, signifying a pivotal advancement in the simplification of circRNA preparation (Fig. 3) [28].

  Figure 3

  

  Figure 3.  End-to-end STS reaction for self-circularization and cellular transfection. (a) The IGS, ribozyme, GOI, and target site are in vitro transcribed and then circularized through an end-to-end STS reaction, which is catalyzed by the Tetrahymena group Ⅰ intron ribozyme. (b) The self-circularizing RNA structure, including the GOI sequence, is transcribed in vitro and circularized via the same STS reaction.

  DownLoad: Full-Size Img PowerPoint

  The STS circularization vector sequence, under the control of the CMV promoter, was initially integrated into a plasmid and transfected into 293A cells, with Luciferase expression monitored over a 12–48-h period. The circular junction was successfully verified by quantitative real-time PCR (qPCR), resulting in the generation of circFluc. The purity and yield of circRNA were further assessed using agarose gel and polyacrylamide gel electrophoresis, and subsequent confirmation was achieved post-purification by ion-pair reversed-phase high-performance liquid chromatography (IP-RP HPLC). These results validated the STS method's effectiveness and underscored its potential in cellular applications.

  Subsequent optimization studies evaluated the impact of antisense (AS) sequence length, antisense binding sequence (ABS) length, spacer sequences, and P1/P10/AS elements on STS circularization efficiency. AU-rich P1 helices significantly outperformed other sequences in terms of circularization efficiency. During the optimization process, AS/ABS sequence lengths of 50, 100, 150, 200, 250, and 300 nucleotides were individually incorporated into the DNA template. Agarose gel electrophoresis revealed that AS50, AS100, and AS150 sequences had significantly higher circularization efficiency than longer sequences. However, AS50 and AS100 sequences resulted in a significant reduction in vitro transcription (IVT) product yield. Therefore, the incorporation of AS150 into the initial STS circularization vector was deemed an optimized balance, enhancing both circularization efficiency and IVT yield.

  The highly structured elements of IRES and type Ⅰ intron ribozymes may impact their proper folding, thereby reducing circularization efficiency. To address this issue, spacer sequences are typically introduced between them to facilitate the circularization process. However, in the STS system, the addition of polyA sequences (A10/A30/A50) or other sequences did not significantly affect circularization efficiency or IVT yield. When examining the impact of P1/P10/AS elements on circularization efficiency, three unique STS circularization vectors were designed: one containing P1+P10+AS150, the second containing P1+P10, and the third containing only P1. Notably, the vector containing P1+P10+AS150 produced the highest circRNA yield, followed by the P1 vector, with the P1+P10 vector yielding the least. These results indicated that the P1 helix alone could efficiently generate circRNA via IVT without the need for P10 and AS elements.

  Building upon preliminary experimental data, further optimization was directed towards the base composition of the P1 helix. Agarose gel electrophoresis indicated that AU-rich P1 STS circularization vectors demonstrated significantly higher circularization efficiency compared to the initial P1+P10+AS/ABS vectors or other base compositions. When circular GFP (circGFP) RNA, derived from both PIE and STS circularization vectors, was transfected into 293A cells, the circRNA from the STS vector resulted in elevated GFP protein expression levels. Moreover, circRNA purified by IP-RP HPLC triggered minimal innate immune responses in A549 cells.

  It is worth highlighting that Rznomics, a reputable company originating from South Korea, has successfully developed an advanced self-targeting and self-splicing circularization vector system. This novel system leverages the ribozyme derived from the Tetrahymena type Ⅰ intron [7]. Table 1 provides a thorough overview of the unique features and wide-ranging applications of this groundbreaking innovation, enabling a precise comparison with the current methodologies for in vitro type Ⅰ intron circularization. This system enhances the efficiency of circRNA generation by reducing the reliance on P1 helices. It was discovered that AU-rich P1 helices bolster circularization efficiency. When the P1 STS circularization vector was designed to include only IGS, ribozyme, and GOI sequences, with an additional U appended to the 3′ end of GOI, the resulting circRNA contained solely the GOI and the extra U base. However, by designating an AU-rich region as the P1 helix target site within the GOI and positioning the 3′-GOI sequence following the target site after the ribozyme ΩG, the final circRNA contained only the GOI, excluding the extra U base. This STS circularization vector system broadens the scope for efficient circRNA production.

  Table 1

  

  Table 1.  Comparison of different types of circularization mechanisms of type Ⅰ intron ribozymes.

  DownLoad: CSV

  Strategies for exogenous RNA circularization	Intron sequence	Intron secondary structure	Catalytic mechanism	Exon sequence	Cyclization condition
  Thymidylate synthase (td)	Thymidylate synthase (td) 1017 bp	P1 has 7 base pairs; P10 has either 2 or 3 base pairs	Esterification	E1, E2	GTP, Mg2+
  Anabaena pre-tRNA Leu	tRNA-Leu 249 bp	P1 has 3 base pairs; P10 has either 2 or 1 base pair	Esterification	E1, E2	GTP, Mg2+
  Tetrahymena	26S rRNA 413 bp	P1 has 6 base pairs; P10 has 7 base pairs	Esterification	–	GTP, Mg2+

  The STS method has significantly advanced the in vitro synthesis of circRNA, thereby propelling the field of biomedicine forward, particularly in the realms of gene therapy and RNA drug development. circRNA, generated with precision and efficiency via the STS method, is highly valued for its exceptional stability and low immunogenicity, rendering it an ideal candidate for gene therapy and effectively mitigating the risk of immune rejection. The approval of the world's first circRNA therapy, RXRG001, by the Food and Drug Administration (FDA) has ushered circRNA drugs into a new phase of clinical development. However, further optimization of the STS method is still required. Future research should focus on two primary directions: first, examining the impact of P1 helix variants on circularization efficiency; and second, enhancing the activity of the Tetrahymena type Ⅰ intron ribozyme through engineering approaches [29]. Additionally, the safety and immunogenicity of circRNA are critical research areas that necessitate investigation of their immunological effects across various cell types [30]. circRNA has demonstrated flexible roles in immune regulation. On one hand, certain circRNAs activate immune receptors (such as retinoic acid-inducible gene Ⅰ (RIG-Ⅰ) and protein kinase R (PKR)) through self-splicing, thereby enhancing innate immune responses and playing a crucial role in combating viruses and tumors. On the other hand, circRNAs designed with short RNA regions can inhibit PKR activity, thereby reducing the intensity of immune responses and showing promise in the treatment of autoimmune diseases and chronic inflammation. circRNA exhibits specific expression patterns in various diseases, including cancer. Its high stability in body fluids, such as blood, makes it a potential non-invasive disease biomarker. Studies have shown that the expression levels of specific circRNAs are closely related to cancer progression, treatment response, and prognosis. For example, highly expressed circRNAs can help identify and target diseased cells, thereby supporting personalized treatment plans and early detection technologies.

  Future development should prioritize interdisciplinary integration. Bioinformatics tools can revolutionize circRNA design through machine learning algorithms that predict optimal splicing sites and IRES configurations, thereby enabling high-throughput screening of therapeutic candidates. Meanwhile, innovations in nanotechnology are expected to improve delivery efficiency. For example, lipid nanoparticles with tissue-specific targeting ligands can enhance the biodistribution of circRNA. Material science approaches, such as programmable biomaterials, may enable stimulus-responsive circRNA release. Furthermore, combining ribozyme engineering with computational RNA folding simulations can optimize self-splicing efficiency while minimizing immunostimulatory motifs.

  In terms of clinical translation, circRNA therapies face numerous regulatory challenges. Regulatory bodies such as the FDA and European Medicines Agency (EMA) have issued a series of guidelines that clarify the specific requirements and considerations for the clinical application of circRNA therapies. For example, the FDA mandates that circRNA therapies provide detailed nonclinical research data before entering clinical trials, including the stability, biodistribution, metabolic pathways, and potential immunogenicity of the drug. The EMA also emphasizes strict regulation of circRNA therapies, requiring long-term follow-up of patients in clinical trials to assess the safety and effectiveness of the treatment [31].

  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.

  CRediT authorship contribution statement

  Huiping Shi: Writing – original draft, Writing – review & editing. Shaojun Peng: Validation. Minghui Yang: Validation. Yuanyu Huang: Writing – review & editing, Validation, Supervision, Funding acquisition.

  Acknowledgments

  This work was supported by the National Key Research & Development Program of China (Nos. 2021YFC2302400, 2021YFA1201000, 2023YFC2606004), the Fundamental Research Funds for the Central Universities (No. 2022CX01013).

  
  
• 1. [1]

     M.J. Zhang, A. Hussain, B. Hu, et al., Nat. Commun. 15 (2024) 6463. doi: 10.1038/s41467-024-50752-9

  2. [2]

     X.J. Chen, Y. Lu, Front. Bioeng. Biotechnol. 9 (2021) 787881. doi: 10.3389/fbioe.2021.787881

  3. [3]

     C.X. Liu, S.K. Guo, F. Nan, et al., Mol. Cell. 82 (2022) 420–434 e426. doi: 10.1016/j.molcel.2021.11.019

  4. [4]

     Z.Y. Ye, L.S. Li, L. Yang, et al., Exploration 4 (2024) 20230138. doi: 10.1002/EXP.20230138

  5. [5]

     S. Khalid, M. Bilal, N. Rasool, M. Imran, Chin. Chem. Lett. 35 (2024) 109498. doi: 10.1016/j.cclet.2024.109498

  6. [6]

     Y.J. Li, Y.N. Wang, Y.G. Luo, et al., Chin. Chem. Lett. 35 (2024) 109576. doi: 10.1016/j.cclet.2024.109576

  7. [7]

     S. Petkovic, S. Müller, Nucleic Acids Res. 43 (2015) 2454–2465. doi: 10.1093/nar/gkv045

  8. [8]

     Y.J. Dang, Y.R. Feng, X. Chen, et al., Chin. Chem. Lett. 35 (2024) 109660. doi: 10.1016/j.cclet.2024.109660

  9. [9]

     B. Hu, B. Li, K. Li, et al., Sci. Adv. 8 (2022) eabm1418. doi: 10.1126/sciadv.abm1418

  10. [10]

      Y.J. Li, J.Y. Li, Y.Q. Chang, Y.F. Lin, L. Sui, Chin. Chem. Lett. 35 (2024) 109414. doi: 10.1016/j.cclet.2023.109414

  11. [11]

      N.G. Dolinnaya, N.I. Sokolova, D.T. Ashirbekova, Z.A. Shabarova, Nucleic Acids Res. 19 (1991) 3067–3072. doi: 10.1093/nar/19.11.3067

  12. [12]

      N.I. Sokolova, D.T. Ashirbekova, N.G. Dolinnaya, Z.A. Shabarova, FEBS Lett. 232 (1988) 153–155. doi: 10.1016/0014-5793(88)80406-X

  13. [13]

      N.G. Dolinnaya, M. Blumenfeld, I.N. Merenkova, et al., Nucleic Acids Res. 21 (1993) 5403–5407. doi: 10.1093/nar/21.23.5403

  14. [14]

      T.E. England, O.C. Uhlenbeck, Biochemstry 17 (1978) 2069–2076. doi: 10.1021/bi00604a008

  15. [15]

      C.K. Ho, S. Shuman, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 12709–12714. doi: 10.1073/pnas.192184699

  16. [16]

      G. Kaufmann, T. Klein, U.Z. Littauer, FEBS Lett. 46 (1974) 271–275. doi: 10.1016/0014-5793(74)80385-6

  17. [17]

      R. Silber, V.G. Malathi, J. Hurwitz, Proc. Natl. Acad. Sci. U. S. A. 69 (1972) 3009–3013. doi: 10.1073/pnas.69.10.3009

  18. [18]

      S. Yin, C.K. Ho, E.S. Miller, S. Shuman, Virology 319 (2004) 141–151. doi: 10.1016/j.virol.2003.10.037

  19. [19]

      E. Ford, J.M. Ares, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 3117–3121. doi: 10.1073/pnas.91.8.3117

  20. [20]

      K.A. Jarrell, Proc. Natl. Acad. Sci. U. S. A. 90 (1993) 8624–8627. doi: 10.1073/pnas.90.18.8624

  21. [21]

      S. Mikheeva, M. Hakim Zargar, D. Carlson, K. Jarrell, Nucleic Acids Res. 25 (1997) 5085–5094. doi: 10.1093/nar/25.24.5085

  22. [22]

      M. Puttaraju, M.D. Been, Nucleic Acids Res. 20 (1992) 5357–5364. doi: 10.1093/nar/20.20.5357

  23. [23]

      R.A. Wesselhoeft, P.S. Kowalski, D.G. Anderson, Nat. Commun. 9 (2018) 2629. doi: 10.1038/s41467-018-05096-6

  24. [24]

      R.A. Wesselhoeft, P.S. Kowalski, F.C. Parker-Hale, et al., Mol. Cell. 74 (2019) 508–520. e504. doi: 10.1016/j.molcel.2019.02.015

  25. [25]

      A. Dallas, S.V. Balatskaya, T.C. Kuo, et al., Nucleic Acids Res. 36 (2008) 6752–6766. doi: 10.1093/nar/gkn637

  26. [26]

      P.A. Feldstein, G. Bruening, Nucleic Acids Res. 21 (1993) 1991–1998. doi: 10.1093/nar/21.8.1991

  27. [27]

      S.A. Kazakov, S.V. Balatskaya, B.H. Johnston, RNA 12 (2006) 446–456. doi: 10.1261/rna.2123506

  28. [28]

      K.H. Lee, S. Kim, J. Song, et al., Mol. Ther. Nucleic Acids. 33 (2023) 587–598. doi: 10.1016/j.omtn.2023.07.034

  29. [29]

      J.M. Burke, M. Belfort, T.R. Cech, et al., Nucleic Acids Res. 15 (1987) 7217–7221. doi: 10.1093/nar/15.18.7217

  30. [30]

      L.X. Xu, Z.X. Shao, X. Fang, et al., Exploration 5 (2024) 20231065.

  31. [31]

      P.P. De Deyn, R. D'Hooge, L.F.M. van Zutphen, Neurosci. Res. Commun. 26 (2000) 141–148. doi: 10.1002/1520-6769(200005/06)26:3<141::AID-NRC2>3.0.CO;2-V

  32. [32]

      X. Liu, Y. Zhang, S.R. Zhou, et al., J. Control. Release 348 (2022) 84–94. doi: 10.1016/j.jconrel.2022.05.043

• 

  Figure 1  In vitro circularization of RNA. (a) Chemical ligation of RNA molecules. (b) Enzymatic ligation of RNA molecules. Reproduced with permission [32]. Copyright 2022, Elsevier B.V.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Ribozyme-mediated in vitro synthesis of circular RNA. (a) autonomous splicing of group Ⅰ introns in the group Ⅰ intron self-splicing system. (b) Self-splicing capability of group Ⅱ Introns in the group Ⅱ intron self-splicing system. (c) HPR conformation. Reproduced with permission [2]. Copyright 2021, Chen and Lu. Copied with permission [32]. Copyright 2022, Elsevier B.V.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  End-to-end STS reaction for self-circularization and cellular transfection. (a) The IGS, ribozyme, GOI, and target site are in vitro transcribed and then circularized through an end-to-end STS reaction, which is catalyzed by the Tetrahymena group Ⅰ intron ribozyme. (b) The self-circularizing RNA structure, including the GOI sequence, is transcribed in vitro and circularized via the same STS reaction.

  下载: 全尺寸图片 幻灯片

  Table 1.  Comparison of different types of circularization mechanisms of type Ⅰ intron ribozymes.

  Strategies for exogenous RNA circularization	Intron sequence	Intron secondary structure	Catalytic mechanism	Exon sequence	Cyclization condition
  Thymidylate synthase (td)	Thymidylate synthase (td) 1017 bp	P1 has 7 base pairs; P10 has either 2 or 3 base pairs	Esterification	E1, E2	GTP, Mg2+
  Anabaena pre-tRNA Leu	tRNA-Leu 249 bp	P1 has 3 base pairs; P10 has either 2 or 1 base pair	Esterification	E1, E2	GTP, Mg2+
  Tetrahymena	26S rRNA 413 bp	P1 has 6 base pairs; P10 has 7 base pairs	Esterification	–	GTP, Mg2+

  下载: 导出CSV
•