Edit · 2026 01 11 221704 Nature S41594 025 01489 6

Saving needs a token. Log in or register to get one. Edits are public and licensed CC-BY-SA-4.0.

# Structures of a natural circularly permuted group II intron reveal mechanisms of branching and backsplicing | Nature Structural & Molecular Biology

**URL:** https://www.nature.com/articles/s41594-025-01489-6

**Saved:** 2026-01-11T06:17:09.676Z

---

Article
Published: 27 February 2025
Structures of a natural circularly permuted group II intron reveal mechanisms of branching and backsplicing
Xiaobin Ling, Yuqi Yao & Jinbiao Ma

Nature Structural & Molecular Biology 
volume
 32, 
pages
1101–1110 (2025)Cite this article

2130 Accesses

2 Citations

1 Altmetric

Metrics
details

Abstract

Circularly permuted (CP) group II introns, identified in various bacteria phyla, swap domains D5 and D6 near the 5′ end and have reversed splice sites (SSs), leading to backsplicing and circular RNA formation. In this study, we present multiple high-resolution cryo-electron microscopy structures of a natural CP group II intron from Comamonas testosteroni KF-1 (Cte 1), elucidating the molecular mechanisms of branching and backsplicing. During branching, the 5′ SS is positioned by an auxiliary sequence (AUX)-enhanced interaction between the exon-binding site and intron-binding site (IBS) and stacks on the branch-site adenosine within D6, allowing the attacking 2′-OH group to coordinate with a metal ion in the active center. In backsplicing, the 3′ SS is aligned with the branching step, leaving IBS in the active center, stabilized by base pairing with the AUX, which enables the free 3′-end hydroxyl group to directly attack the scissile phosphate of 3′ SS. Furthermore, a groove in Cte 1 may stabilize the circular RNA. These findings highlight a conserved catalytic mechanism for canonical group II introns, albeit facilitated by the versatile AUX, opening avenues for designing potent ribozymes producing circular RNAs.

You have full access to this article via your institution.

Download PDF
Similar content being viewed by others
Structural basis of circularly permuted group II intron self-splicing
Article 31 January 2025
Structural basis of branching during RNA splicing
Article 06 December 2023
Structural insights into the cross-exon to cross-intron spliceosome switch
Article Open access
22 May 2024
Main

Group II introns are categorized into three main groups, IIA, IIB and IIC, each characterized by conserved secondary structures with conserved domains and a similar organization around the ribozyme active site1. Group II introns have six structural domains, labeled D1-D2-D3-D4-D5-D6 from 5′ to 3′. Recently, a new class of group II introns, named circularly permuted (CP), was discovered through bioinformatics. Their structural domains are naturally inverted, labeled D5-D6-D1-D2-D3-D4 from 5′ to 3′, and unique short hairpin structures, P1–P2, are present before D5. According to conservation analysis of their secondary structure and biochemical mechanism, CP group II introns do not belong to the previously classified group II introns and represent a new class of naturally occurring group II introns that generate circular exons through backsplicing2. However, the structure of this new class of group II introns that generate circular RNAs (circRNAs) has not yet been reported.

The crystallographic analysis of the bacterium Oceanobacillus iheyensis group IIC intron demonstrated the presence of a well-defined catalytic core and highlighted the conservation of key structural elements essential for splicing3,4,5,6. Additionally, the Lactococcus lactis group IIA intron7 and the Pylaiella littoralis group IIB intron8,9 have provided insights into variations within specific regions while underscoring the maintenance of the essential six-domain organization. Structures such as the Eubacterium rectale group IIC intron and Thermosynechococcus elongatus group IIA intron have contributed to the growing body of evidence supporting a shared core architecture among diverse group II introns and D6 conformation flipping up and down in splicing or backsplicing processes10,11.

Typically, these introns initiate the first phosphoester transfer reaction at the 5′ splice site (SS) by requiring a ‘branch point (bp) adenosine’ in D6 (refs. 12,13). This branching pathway results in the formation of ligated exons and the excision of an intron with a 2′–5′ phosphodiester linkage. In addition to the branching pathway, group II introns use a hydrolysis pathway as the initial step of the self-splicing catalytic reaction14,15. Recently, the structures of the prebranching pathway stage during splicing of group IIC introns and the prehydrolysis pathway stage during transposition of the newly identified group II intron hydrolytic endonucleolytic ribozyme through bioinformatics were reported11,16. These structures provide the first structural insights into the mechanism of the branching stage. However, there are still no reports on the structure of the prebranching stage of backsplicing.

CircRNAs are stable and common RNA molecules in eukaryotic cells formed through backsplicing17,18. They have important roles in biology by functioning as microRNAs19,20 or protein inhibitors21, regulating protein activity or even being translated into proteins22,23. However, efficient generation of circRNAs either in vitro or in vivo presents technical challenges. Common circularization methods involve ligating linear RNAs with the T4 RNA ligase or using the permuted intron–exon (PIE) splicing strategy using self-catalyzed group I or group II introns24. However, direct ligation with RNA ligases often requires a ‘splint’ sequence to connect free ends, making large-scale production difficult because of concatemer formation. Introducing ligase proteins can also complicate product purification25. As early as 30 years ago, it was reported that rearranging the order of permuted group II domains using the PIE method can generate circRNAs but the PIE method introduces a foreign ‘scar’ sequence, potentially affecting circRNA structures and immune responses, making it less ideal for therapeutic use25,26. While scarless circRNAs produced by the ‘clean PIE’ method through engineered group I or group II introns have been reported for therapeutic applications, they involve artificial modifications and high-throughput screening27,28,29.

In this study, we successfully determined multiple sets of high-resolution cryo-electron microscopy (cryo-EM) structures of a natural CP group II intron from Comamonas testosteroni KF-1 (Cte 1) under varying metal ion conditions and with the aid of mutations. These findings elucidate the molecular mechanisms of backsplicing in prokaryotes and provide insights into the spatial requirements of the circular exon pocket. We propose that CP group II introns may more closely resemble the ancestors of U12-type spliceosomes or other unfound minor spliceosomes.

Results
CP group II intron Cte 1_∆D4 dynamic cryo-EM reconstruction

The CP group II intron Cte 1, which was well characterized previously, was truncated at the 3′ end by deleting D4 (Cte 1_∆D4) as this deletion had no effect on activity2 (Extended Data Fig. 1a,b). In the presence of cations (Mg2+ or Ca2+ combined with K+), the prebranching Cte 1 wild type (WT) folds properly, allowing the 2′-OH of the branch-site adenosine to engage in nucleophilic attacked on the scissile phosphate at the 5′ SS, which is located downstream of the branch site. This results in the formation of branched intermediate (BI) containing a 2′-5′-phosphodiester linkage in the first branching step. Then, the newly formed 3′ hydroxyl group attacks the 3′ SS, which occurs upstream in the permuted variant in the second backsplicing step, finally generating a branched product (BP) and a covalently ligated circRNA (Fig. 1a).

Fig. 1: Cryo-EM reconstruction of Cte 1_∆D4.

a, Cartoon of the Cte 1 backsplicing reaction. b, Secondary-structure cartoon and tertiary interactions of the CP group II intron. c,d, The SYBR gold-stained 6% urea–PAGE gels showing various conditions used to obtain samples for cryo-EM. Lane 1 is the size marker for RNA. Lanes 2 are the Cte 1_∆D4 backsplicing reactions showing migration of the prebranching (Pre) and BI products of WT Cte 1_∆D4 (top) and U537ins (bottom) in CaCl2 conditions. Lanes 3 are the Cte 1_∆D4 backsplicing reactions showing migration of the Pre, BI, BP and circRNA (C) products of WT Cte 1_∆D4 (top) and A695U (bottom) in MgCl2 conditions. An additional band labeled as BI* was observed between Pre and BI. The experiments shown in these panels were repeated more than three times with similar results. e–h, Cryo-EM map and molecular model of the Pre (\({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\), \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\)), BI (\({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\)) and BPwt in Cte 1 backsplicing.

Source data

Full size image

The backsplicing product of Cte 1_∆D4 was observed as early as 5 min in a time-course reaction (Extended Data Fig. 1c). Cte 1_∆D4 did not exhibit backsplicing activity at 4 °C but demonstrated this activity between 23 °C and 60 °C, indicating that Cte 1_∆D4 can function within a broad temperature range (Extended Data Fig. 1d). Previous studies showed that group II intron splicing activity relies on specific divalent and monovalent metal ions within the catalytic core as shown in the structural analysis of group II intron6,13,30,31,32,33. The backsplicing activity of Cte 1_∆D4 specifically depended on divalent metal ions Mg2+, while other divalent metal ions such as Ca2+, Mn2+ and Zn2+ did not work well (Extended Data Fig. 1e). Monovalent ions such as K+, \({{\rm{NH}}}_{4}^{+}\) and Cs+ all supported backsplicing activity, while Li+ and Na+ exhibited poor performance (Extended Data Fig. 1f).

Because the sequence and length of the D4 domain without the open reading frame are not conserved, we chose to use cryo-EM to directly analyze the structure of Cte 1_∆D4 in the presence of two different divalent cations, Mg2+ and Ca2+. Biochemical assays showed that backsplicing activity was optimal under Mg2+ conditions, while it was inhibited under Ca2+ conditions (Fig. 1c). Through this sample preparation, we were able to solve two cryo-EM structures of WT Cte 1_∆D4 with the two different divalent cations, at the resolution of 2.6 Å for the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) state and 2.78 Å for BPwt (Fig. 1e,f, Extended Data Figs. 2 and 3, Table 1 and Supplementary Video 1). The resulting near-atomic-resolution three-dimensional (3D) reconstruction was of sufficient quality to enable the modeling of individual nucleotides. This was achieved through a combination of secondary-structure prediction and validation on the basis of previous biochemical studies of group II introns. Additionally, we present a secondary-structure layout based on these models, which includes all known tertiary interactions of group II introns, as well as new tertiary interactions specific to CP group II introns (Fig. 1b and Extended Data Fig. 6a).

Table 1 Cryo-EM data collection, refinement and validation statistics
Full size table
Overall structural analysis of Cte 1 \({\mathbf{{Pre}}}_{\mathbf{{inactive}}}^{\mathbf{{wt}}}\) and BPwt conformations

In the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) conformation at the branching step, the D6 domain is positioned vertically on the D1 domain, achieved through a hook-like structure formed by A451 in D1 interacting with A113, A114 and G111 in D6 (\({{{\eta }}-{{\eta }}}_{{\rm{up}}}^{{\prime} }\) interaction), facilitating the involvement of the branch site in the catalytic core. Upon completion of the backsplicing process, the D6 domain flips down in the BPwt structure, allowing G111 to form a base pair with C657 in D2, referred to as the \({{{\eta }}-{{\eta }}}_{{\rm{down}}}^{{\prime} }\) interaction (Fig. 2a,g and Extended Data Fig. 6b). The varying conformations of D6 in the prebranching and BP states are consistent with the findings from previous studies10,11,34.

Fig. 2: Architecture of Cte 1_∆D4 catalytic core.

a, Different conformations of D6 (red) in the Pre and BP states. b, D5 (sky blue) steadily docks into the pocket formed by D1 (light gray) and D3 (light pink). c,d, D5 is docked in D1 through four non-Watson–Crick interactions named ξ–ξ′, ζ–ζ′, λ–λ′ and µ–µ′. e,f, D5 is held by D3 to dock in D1 through four non-Watson–Crick interactions named ρ–ρ′, κ–κ′, ω–ω′ and σ–σ′. g, D6 takes up and down conformations using anchor hook-like structures and the η–η′ interaction. h, Cte 1_∆D4 uses AUX with the EBS and IBS to form four triple interactions; EBS–AUX has different conformations in the Pre and BP states.

Full size image

Previous studies suggested that D1 functions as a scaffold that guides the folding of group II introns, with D5 docking into a pocket mainly formed by D1 (refs. 15,35,36). Our structures align with this model, as the sequence conservation analysis of CP group II introns indicates a high degree of conservation in the D5 region (Extended Data Fig. 7a) and further demonstrates that D5 docks into the pocket formed by D1 and D3, which is accomplished with the assistance of various important tertiary interactions (Fig. 2b and Supplementary Video 2). Firstly, the terminal loop of D5 and D1 forms the ξ–ξ′ and ζ–ζ′ tertiary interaction, comprising A74–U77 in D5 and A530, G600–G602 pairing with G476–C478 in D1 (Fig. 2c). Secondly, the 2-nt bulge of D5 is locked into D1 by λ–λ′ and µ–µ′ tertiary interactions, resulting in a triple interaction respectively formed by A424 in D1 with C69–G80 in D5 and A628 in D1 with U67–G84 in D5. Furthermore, U423 and G425 in D1 facilitate the anchoring of D5 (Fig. 2d). Subsequently, the kissing loop interaction of D1 and D3 causes U729 in D3 to adhere to the groove formed by C62-C63 in D5 and G89-U90, denoted as κ–κ′. A738–C740 in D3 respectively interact with G88-C87 in D5, forming the ρ–ρ′ interaction to promote the D5 docking (Fig. 2e). Finally, as the Cte 1_∆D4 stem sequence of D5 is longer than other group II introns2,15, the starting nucleotide of G55-C56 in D5 forms base pairs with G807-C808 in D3, leading to the interaction of A95 in D5 with the C56 and G807 base pair, resulting in triple interactions named ω–ω′ and σ–σ′, respectively (Fig. 2f).

A695-G696-A697 in J2/3 has an important structural support role and penetrates into the catalytic center. A695 has no density but its base interacts with the phosphate of U59 of D5 in the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) conformation (Extended Data Figs. 8a and 9b). In the BPwt conformation, A695 forms Watson–Crick pairing with U133 and the side chain of G221 wedges between them, forming a stable three-base triangular structure. Additionally, G220 connects with bpA127, forming a special form of 3′-phosphate linkage (Extended Data Fig. 8b). G696 and A697 interact with G61–U90 and A60–U91 in D5, anchoring the catalytic structure of D5 throughout J2/3 (Extended Data Fig. 8a).

Cte 1_∆D4 uses auxiliary sequence (AUX) to regulate backsplicing

Most group II introns modulate the choice of splicing pathways using the intron-encoded protein, such as maturase or reverse transcriptase, to facilitate exon recognition and D5 catalysis1,7,10,11,34,37. This protein-based assistance is also required for the ligation of linear RNAs in the postsplicing state7 or the reverse splicing prestate10,34. Notably, the unique sequence ‘CUGCG’ preceding the exon-binding site (EBS) of Cte 1_∆D4 forms a triple helix with the 5′ intron–exon SS. This unique sequence facilitates base pairing of the EBS with the intron-binding site (IBS), a feature not observed in other group II intron structures, and is referred to as AUX in this study (Figs. 1b and 2h). AUX forms four triple interactions, namely C545–G214–C535, U544–A215–U536, C542–G217–G537 and A541–U218–C538, to stabilize the IBS–EBS interactions. G537 and G539 in AUX separately form π–π stacking with mismatched A216 in the IBS and a Watson–Crick base pair with A540 in the EBS to enable it to form a twisted structure (Fig. 2h and Extended Data Fig. 8c,d). Interestingly, the EBS–AUX conformation exhibits a notable deviation in the prebranching and BP structures (Fig. 2h).

Recognition of 5′ SS in prebranching

Previous reports and sequence conservation analysis identified ‘GGGCG’ as an extremely conserved 5′ SS sequence2,12,38 (Fig. 3b). This sequence differs from the canonical GT or AU motifs. Structural analysis reveals that during branching (Supplementary Videos 3 and 4), A424 and U423 in D1 interact with C81 and G80 in D5, as well as G217 and U218 of EBS–IBS through λ–λ′ interactions to stabilize the EBS (Fig. 3a,c). More importantly, G539 of AUX interacts with G220 of the 5′ SS to form a G–G noncanonical base pair, which is a new unique interaction for CP group II intron that we named the δ–δ′ interaction. This interaction leads to substantial distortion of G220 and U219, thereby exposing the bpA127 in the catalytic center (Fig. 3a,d,h). It is worth mentioning that, in the \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) conformation, the interaction between G220 and G539 begins to prepare to separate (Extended Data Fig. 8e). Additionally, G221 of the 5′ SS inserts into the minor groove of D6, stacking on bpA127 and, more importantly, specifically forming hydrogen bonds with the C102–U124 mismatched pair (Fig. 3a,e,h). Then, G222 and C223 of the 5′ SS form base pairs with C426-G425 of D1 through a highly conserved ε–ε′ interaction. G224 wedges into a 2-nt bulge and its base interacts with A628 of D1 while the phosphate group coordinates with a potassium ion (K3). Furthermore, the interaction between G696 and A697 in J2/3 with U59 and A60 in D5 also has a pivotal role in the catalytic core, as part of the catalytic triad (Fig. 3f and Extended Data Fig. 8a). Lastly, under the catalysis of D5 assisted by one K+ and two Mg2+ ions, bpA127 undergoes nucleophilic attack on the noncanonical 5′ SS in coordination with one of the Mg2+ ions (Fig. 3g, h).

Fig. 3: Molecular mechanism of the 5′ SS recognition for branching.

a, Schematic diagram of the overall model of highly conserved 5′ SS recognition by Cte 1_∆D4. b, Sequence alignment of the IBS–5′ SS and AUX–EBS regions of CP group II intron. c, The λ′–D5 interaction formed by A424 and U423 in D1 stabilizes the EBS–IBS interaction like a wedge. d, The 2′-OH of bpA127 attacking the G220 in the 5′ SS must be recognized by the G220 and G539 in AUX forming the new δ′–δ interaction. e, G221 in the 5′ SS sticks into the D6 spiral groove formed by G101-C102 and U124-C125, f, G222-C223 in the 5′ SS and G425-C426 in D1 form the ε–ε′ interaction; G224 in the 5′ SS sticks into the 2-nt bulge and interacts with A628 in D1 and one potassium ion. g,h, The highly conserved catalytic center is also formed by two divalent metal magnesium ions and one monovalent metal potassium ion; D5 catalyzes the bpA127 of D6 to attack the noncanonical 5′ SS between 5′ exon (forest green) and D1 (gray), while G696 and A697 in J2/3 (light pink) participate in the catalytic process.

Full size image
Structural analysis of the backsplicing mechanism

From the sequence conservation analysis in AUX–EBS of CP group II introns, Cte 1 is exactly missing a U between U536 and G537 (Fig. 3b). Furthermore, A695 within J2/3 is highly conserved (Extended Data Fig. 7b). On the basis of this analysis, we constructed two mutants, U537ins (with a U inserted between U536 and G537) and A695U, which has a pivotal role in defining the 3′ SS by γ–γ′ interaction. Fortunately, four more structures were obtained: U537ins in CaCl2 conditions and A695U in MgCl2 conditions in the prebranching (\({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\), 3.02 Å; \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\), 3.19 Å) and BI states (\({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\), 3.02 Å; \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\), 3.11 Å) (Fig. 1d,g,h, Extended Data Figs. 4 and 5 and Supplementary Videos 3 and 4).

The \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) conformation is particularly distinct, as evidenced by the increased density of exonic regions, which precisely align with the exterior of D6 and P1–P2. Analyzing portions of density amenable to atomic modeling, we discovered that U537 positioned at the site of mismatch A does not directly form triple interactions but rather disrupts the original triple interaction, thereby allowing for Watson–Crick base pairing between the IBS and EBS, except for the A–A mismatch, while AUX assists in forming non-Watson–Crick interactions with the IBS from positions 213 to 208. Notably, the conformation of G540 in \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) corresponds to G539 in the WT \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) conformation, where it no longer participates in crucial δ–δ′ interactions with the 5′ SS or 3′ exon upon extrusion from AUX–IBS–EBS. Instead, it engages in non-Watson–Crick interactions with the A209-C208 of the 5′ exon (Extended Data Fig. 8f–h). Additionally, U133 in the 3′ SS, along with the starting nucleotides C134-C135-A136 of the 3′ exon, pairs with G201-G200-G199-G198 of the 5′ exon, stabilizing the start of the exon (Extended Data Fig. 8i). In the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) conformation, we observe a more stable state compared to \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\).

Unique backsplicing catalytic center

By carefully analyzing the backsplicing catalytic core of these structures, we found that, although the formation mechanisms of the 2′–5′ and 3′–5′ phosphodiester bonds and the involved metal ions are similar to those in non-CP group II introns6,11,13, the activity center surrounding 5′ SS and 3′ SS recognition is very special. In the prebranching states, a flip occurs between the phosphate and base of G539 in AUX from \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) to \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\), leading to the recognition of G220 in the 5′ SS. G221 in the 5′ SS acts like a wedge anchoring into a groove formed by conserved bases C102-G101 and U124-C125. In the \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) state, G61 joins the active center and further pulls bpA127 closer to the 5′-phosphate of G220 with the help of a Mg ion (M1). The 2′-OH of bpA127 nucleophilically attacks G220 in the 5′ SS, facilitating backsplicing. Furthermore, in the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) conformation, after the recognition of G539 with G220, there are three Ca ions present in the active site. The two phosphate groups of C81 interact with two Ca ions (C2 and C3), exhibiting an inhibited state. However, upon transition to the \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) state, C81 coordinates only with one Mg ion (M2). At this point, G61 joins the active site and coordinates with the Mg ion (M1), followed by bpA127 further approaching to a distance of 3.12 Å to complete the coordination of of the 5′-phosphate of G220 with the Mg ion (M1). Subsequently, the 2′-OH of bpA127 nucleophilically attacks the G220 in the 5′ SS, releasing the 3′-OH of U219 and, thus, initiating backsplicing (Extended Data Fig. 9a).

Once branching is completed, the conformation of D6 flips down toward D2, pulling the 5′ SS along with bpA127 down through the 2′–5′ phosphodiester bond to G220, allowing the 3′ SS to dock into the active site and the 3′-OH of U219 in IBS to nucleophilically attack the 5′ phosphate of U133. During this process, the role of J2/3 is particularly special, with G696-A697 acting like rivets, stabilizing triple interactions with A60-G61 in D5, which facilitates swinging of A695 from the \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\) to BPWT conformation. Furthermore, a water molecule (W) in the active center of \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\) may regulate the BI transition from an inactive to \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\) state, in which A695 is inserted between G220 and G221, forming a π–π stacking interaction among the three, and interacts with U133 in the 3′ SS (Extended Data Fig. 8J). This ultimately accomplishes the backsplicing process, resulting in the formation of a BP and a circular exon (Extended Data Fig. 9b). Although the BP state resembles that of group IIC intron, the difference lies in the assistant role of AUX, making the circular exon easier to detach, while the linear exon assisted by the ribonucleoprotein (RNP) of group IIC intron remains (Extended Data Fig. 9c–e). Therefore, the exit of the linear exon may be a further step in the group II intron life cycle.

Recognition of the 3′ SS in backsplicing

Although previous studies defined AAU as the 3′ SS (ref. 2), A131-A132 are not completely conserved, with only U133 being fully conserved (Fig. 4d and Extended Data Fig. 7c). Through careful structural analysis, we obtained active and inactive BI structures from U537ins and A695U mutants (Fig. 4a,b). In the inactive BI (BI*) structure of the A695U mutant, C134 of the 3′ SS does not interact with G539 of AUX, while the scissile phosphate remains 10.5 Å away from the free 3′-OH of U219 of the 5′ exon, which coordinates with a water molecule in the catalytic center (Fig. 4a,e). In the active BI structure of U537ins, C134 of the 3′ SS pairs with G539 of AUX by δ–δ′ interaction, with the scissile phosphate occupying a position close to that of the water molecule in the inactive BI structure, which reduces the distance between the 3′-OH and scissile phosphate to 3.5 Å (Fig. 4b,e and Supplementary Videos 3 and 4). We also compared this structure to that of the pre-2F state of group IIC intron (Protein Data Bank (PDB) 8T2R), which may not be in the fully active conformation because the distance between the 3′-OH and scissile phosphate is 5.8 Å (Fig. 4c).

Fig. 4: Molecular mechanism of the 3′ SS recognition for backsplicing.

a,b, The 3′-OH of the 5′ exon attacking the U133 of the 3′ SS recognizes the C134 of 3′ exon (b) or not (a), which forms the δ′–δ interaction with G539 of AUX. In the whole-site view of the \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\) model (a, left), C135 is closer to G539 (a, left middle) and U219 and C135 coordinate with one water molecule (a, right middle), with the inactive attacking distance of 10.5 Å (a, right). In the whole-site view of the \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\) model (b, left), C134 interacts with G539 (b, left middle) and C134 coordinates with M2 (b, right middle), with an active attacking distance of 3.5 Å (b, right). c, The 3′-OH in the 5′ exon attacks U638 in the noncanonical 3′ SS of the group IIC intron (PDB 8T2R). In the whole-site view of the \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{IIC}}}\) model (left), the DBD of maturase engages in the δ′–δ interaction (left middle), U639 in the 3′ exon is not yet coordinated with M2 (right middle) and the inactive attacking distance is 5.8 Å (right). d, Sequence alignment of the bpA in D6, 3′ SS and the start of 3′ exon regions of CP group II introns. e, Secondary-structure schematics with annotated tertiary contacts of the dynamic branching and backsplicing by Cte 1_∆D4 in the Pre, BI and BP states.

Full size image
An intron groove provides the space for the circular exon

Through adjusting the contour level values to low-resolution density between the prebranching and BI conformations of the 3D classification, we observed that the density of exons in the prebranching state hangs outside through interactions with D6 and the exon itself, whereas, in the BI state, circular exons are generated, causing the exon to fall into the intron groove. Rotating 30° around the y axis from the front view, we can see that the intron groove of the CP group II intron resembles a pocket, allowing for the stable existence of circular exons, but this space may also limit the size of circular exons that CP group II introns can generate (Fig. 5a,b).

Fig. 5: A unified and unique model for backsplicing and noncanonical SS recognition in prokaryotes.

a, The full cryo-EM map structures of Cte 1_∆D4 in Pre and BI states show that the intron groove provides the exon backsplicing space without auxiliary proteins. Green and lime-green dashed lines indicate the circular exon; the black dashed line indicates the intron groove. b, The full cryo-EM map structures of Cte 1_∆D4 in the Pre and BI states with colored domains. Green represents the circular exon. c, Left, structures of the group IIC intron are shown in surface representation with D4, maturase and other regions. Structures of the pre-B complex in U2 type (middle) and the pre-B complex in U12 type (right) are shown in surface representation with U1, U2 and tri-snRNPs in different colors, illustrating that the intron groove space is essential for the backsplicing of exon; the space size may be a measuring stick for a long or short exon. The black dashed line indicates hypothetical path of exon backsplicing.

Full size image

Through comparison to the group IIC intron that generates linear exons7,10,11,34, we speculate that, if there is a short or long exon between the downstream 5′ SS and upstream 3′ SS of the group IIC intron, the space occupied by the maturase in the intron groove will hinder the generation of circular exons. This also suggests that, in prokaryotes, the presence of maturase and D4 ensures the correct generation of linear exons without frequent production of circular exons. Analysis of the cryo-EM structure of the spliceosome complex E in eukaryotes reveals that, for the major spliceosome, when the tri-small nuclear RNP (snRNP) occupies the original intron position and the downstream 5′ SS and upstream 3′ SS become exons, the exon must be long to complete the circularization process across the tri-snRNP39. Eukaryotes have not only long circular exons but also short circular exons. Therefore, with the recent resolution of the structure of the minor spliceosome, it was observed that the size of the tri-snRNP in the complex pre-B structure of the minor spliceosome and the distance between U11 and U12 are much smaller than those in the major spliceosome40. Therefore, it is reasonable to speculate that the U12-type spliceosome may be the pathway for generating short exons (Fig. 5c). Because CP group II introns are also speculated to be a pathway for generating short exons in prokaryotes on the basis of the size of the intron groove, CP group II introns may be the primitive form of U12-type spliceosomes or share a common ancestor. However, because of differences in their 5′ SS and 3′ SS sequences, there may be undiscovered specialized forms of CP group II introns in eukaryotes for generating short exons.

Discussion

For eukaryotic spliceosomes, recognition of the 5′ SS by the B* complex leads to the cleavage between the 5′ exon and the intron, which generates the C complex, completing the first splicing reaction. Then, the C complex recognizes the 3′ SS and connects the 5′ exon and the 3′ exon to generate the linear exon and P complex, completing the second reaction41 (Extended Data Fig. 10a). According to the identification of CP group II introns, we labeled the intron, exon, SS and bpA surface structures of the B* complex (PDB 6Q6J)42, C complex (PDB 7B9V)43 and P complex (PDB 6QDV)44 (Extended Data Fig. 10b). By enlarging and adjusting the positions, we can find that the active center of CP group II intron is very similar morphologically but the difference is that the spliceosome center requires the Prp8 protein to have an auxiliary role, while the CP group II intron relies on its own sequence AUX (Extended Data Fig. 10c,e).

Through the splicing process of the spliceosome and previous speculations, we reasonably hypothesized the pathways for the B* complex, C complex and P complex during circRNA generation. From the cartoon schematic, it is evident that the spatial position of the U6 protein is very different because of the inversion of the SS sequence (Extended Data Fig. 10d) and we also simulated the inversion of the intron, exon, SS and bpA positions on the surface structures of the three spliceosomes. We found that this inversion actually affects the spatial interaction between U2, U5 and U6 and the intron and exon; the similarity of the active center is obtained by adjusting the observation position (Extended Data Fig. 10e). Thus, the spatial positions of the exon and intron lead to the difference in the active center of CP group II introns from that of the linear exon spliceosome. The structure of a spliceosome that generates circular exons has not been resolved.

Various group II introns use diverse strategies for 5′ SS and 3′ SS recognition, such as the δ element in group IIC, essential EBS–IBS 1 and 2 interactions between the 5′ exon and D1 in group IIA and those relying on the B2 domain in group IIB introns1,2,44. However, CP group II introns use different recognition mechanisms for 5′ SS and 3′ SS recognition. In the prebranching state, G539 of AUX forms a δ–δ′ interaction with G220 of the 5′ SS and G221 of the 5′ SS forms a π–π′ interaction with D6, facilitating bpA127 in nucleophilic attack on the 5′ exon. In the backsplicing state the 3′ SS is docked into the active site by flipping D6 down, enabling G539 to form a δ–δ′ interaction with C134 of the 3′ exon. Therefore, the CP group II intron represents a new class within group II introns.

Through cryo-EM, we resolved the conformations of the CP group II intron, represented by Cte 1_∆D4, from the inactive state of the prebranching and BI stages to the active state of the prebranching and BI stages and, finally, the release of circular exons and change to the BP stage. This is a comprehensive elucidation of the entire dynamic process of backsplicing in prokaryotes (Extended Data Fig. 10f). However, the ultimate fate of the BPs generated by CP group II introns in vivo and the function of circular exons remain unclear.

Group II introns offer advantages in generating diverse circRNAs without the need for additional guanosine triphosphate and high temperature (Extended Data Fig. 1). Although clean PIE technology offers improvements over Ana-PIE27,28, it has its limitations. However, engineering clean PIE group II introns poses challenges because of resource requirements, lack of known structures and purification issues. Cte 1_∆D4 shows promise for producing clean circRNAs. While Cte 1_∆D4 can be directly used, further rational design and improvements based on the elucidated structures are necessary to meet practical clinical production needs.

The ability of CP group II introns to produce circRNAs adds another dimension to the functional adaptability of these RNA molecules, prompting further investigation into their importance in cellular processes, molecular evolution and potential clinical applications.

Methods
RNA preparation

The Cte 1 group II intron sequence was derived from the National Center for Biotechnology Information (GenBank: NZ_AAUJ02000001.1). All sequences of RNA and primers can be found in the Supplementary Information. The WT and mutant RNAs were prepared as described previously45. In brief, the DNA templates in pUC-57 plasmid were amplified with the regular forward primer and reverse primer (Supplementary Table 2). In vitro transcription of all RNAs in this article were carried out with 0.3 µM DNA template, 5 mM NTPs and 1 U per µl pf RNase inhibitor in 1× transcription buffer containing 50 mM Tris-HCl pH 7.9, 0.01% TritonX-100, 20 mM MgCl2, 2 mM spermidine and 10 mM DTT, before incubating at 37 °C for 1 h (ref. 46). The transcription products were mixed with 2× denaturing gel loading buffer containing 95% formamide, 0.025% SDS, 10 mM EDTA, 0.025% xylene cyanol and 0.025% bromophenol blue and loaded on a 4% 19:1 acrylamide:bis 8 M urea polyacrylamide gel. The gel was run at 15 W for 3 h and then visualized briefly with a 254-nm ultraviolet lamp, held far from the gel to minimize RNA damage47. Then, the RNA was eluted from the gel overnight in elution buffer containing 30 mM sodium acetate pH 5.2 and 5 mM EDTA on an active rotator at 4 °C. The resulting gel slurry was then filtered through 0.45-μm filters (Minisart Syringe Filter, Sartorius). The resulting RNA was precipitated with equal volume of isopropanol to remove urea and salts and then washed with 75% ethanol 2–3 times. Air-dried RNA pellet was resuspended in RNase-free water. The RNA was quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific) and kept at −80 °C before use.

Backsplicing reactions

Backsplicing reactions of 0.05 μM WT and mutant Cte 1 RNAs were carried out in backsplicing buffer (40 mM HEPES pH 7.5 and 240 mM KCl) at 23 °C. The RNAs were heated to 80 °C for 1 min and cooled to 60 °C (0.5 °C s−1) in backsplicing buffer. After annealing, a final concentration of 20 mM or 100 mM MgCl2 or CaCl2 was added to start the backsplicing assay; samples were then incubated at 23 °C for 30 min.

For cryo-EM sample preparation, the backsplicing reaction was activated in a total volume of 0.5–1 ml as described above. The mixture was then rapidly chilled on ice and concentrated to approximately 15 μM by concentrator columns with a 100-kDa cutoff (Ultrafiltration Centrifugal Tube, Millipore) at 4 °C.

For subsequent biochemical analysis, the backsplicing assay was quenched by adding an equal volume of denaturing loading buffer containing 100 mM EDTA, 8 M urea, 0.1% SDS, 0.025% xylene cyanol and 0.025% bromophenol blue. Then, electrophoreses were performed on denaturing 8 M urea and 6% 19:1 acrylamide:bis gel, running at 12 W for 2.5 h. The gels were stained for 10 min in 1× SYBR Gold (Invitrogen) and visualized by Typhoon FLA 9500 (GE).

Cryo-EM sample preparation

First, 3 μl of the resulting backsplicing products (~15 μM) were applied onto glow-discharged (10 s) 200-mesh R1.2/1.3-2nmC Quantifoil Cu grids. The grids were blotted for 1.5 s in 100% humidity at 4 °C with no blotting offset and rapidly frozen in liquid ethane using a Vitrobot Mark IV (Thermo Fisher Scientific).

Cryo-EM data collection and image processing

Cryo-EM data were acquired using a Titan Krios microscope (Thermo Fisher Scientific) operating at 300 kV. The microscope was equipped with a K3 summit direct detector (Gatan) and a GIF quantum energy filter (Gatan) set to a slit width of 20 eV. Automated data collection was performed using the SerialEM software48 with the beam-image shift method49.

Videos of the WT–CaCl2, U537ins–CaCl2 and A695U–MgCl2 structures were captured in super-resolution mode at a nominal magnification of ×105,000, resulting in a physical pixel size of 0.832 Å. Videos of the WT–MgCl2 structure were recorded at a magnification of ×81,000, corresponding to a physical pixel size of 1.064 Å. Each video covered a defocus range from −1.2 μm to −2.2 μm. Video stacks were dose-fractionated into 40 frames, with a total exposure dose of approximately 52–54 e− per Å2.

For cryo-EM data processing, a standardized procedure was used throughout the analysis within cryoSPARC (version 4.4.1)50. The initial steps involved patch motion correction and patch contrast transfer function estimation to rectify stage shift, beam-induced motion and imaging distortions. Rigorous criteria were applied during the selection of high-quality images, taking into account factors such as ice thickness, defocus range and estimated resolution.

The particle processing pipeline began with initial particle picking using blob picking and template picking, followed by reference-free two-dimensional classification and removal of duplicated particles. Subsequently, a subset of high-quality particles underwent ab initio reconstruction, yielding initial references, and heterorefinement was applied to acquire particles most similar to the initial reference particles without binning. After multiround 3D classification, classes containing well-defined D6 domains were re-extracted. Through homorefinement, nonuniform refinement, reference-based motion correction and local/global refinement, high-resolution maps were obtained. Subsequently, local resolution estimation was conducted. The final refined maps underwent sharpening by EMReady (beta version 2.0)51. The resulting sharpened maps, featuring enhanced density, were used for subsequent structural analyses. The reported resolutions above are based on the gold-standard Fourier shell correlation (FSC) 0.143 criterion. All visualization and evaluation of 3D density maps were performed with UCSF Chimera52 and ChimeraX53. Detailed cryo-EM data processing workflows are outlined in Extended Data Figs. 2, 3, 4 and 5.

Cryo-EM model building and refinement

We de novo reconstructed the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) model on the basis of the information of IBS–EBS according to the conserved second structure as the start point of model. The other regions with no high-resolution regions were rebuilt according to RNA 3D structure prediction54,55 and then manually adjusted and rebuilt with Coot as needed56. These models were refined with phenix.real_space_refine57, yielding an averaged model–map correlation coefficient of 0.84, 0.83, 0.92, 0.91, 0.92 and 0.82, respectively, for \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\), \({{\rm{BP}}}^{{\rm{wt}}}\), \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\) and \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\). Final models were validated by MolProbity58. All figures of models and maps in the article were drawn using PyMol, Chimera and ChimeraX. Secondary-structure diagrams were manually drawn with Adobe Illustrator.

RNA conservation analysis

Sequences corresponding to CP group II intron sequences were obtained from the supplementary data of a previous study2. Sequences were aligned using ClustalOmega and alignments were visualized and analyzed with JalView.

Reporting summary

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

Data availability

All data are available in the main text and Supplementary Information. The cryo-EM maps and associated atomic coordinate models of \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\), \({{\rm{BP}}}^{{\rm{wt}}}\), \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\) and \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) were deposited to the wwPDB OneDrop System under EM Data Bank accession codes EMD-38173, EMD-38175, EMD-38174, EMD-38177, EMD-38176 and EMD-38182 and PDB accession codes 8X9K, 8X9M, 8X9L, 8X9O, 8X9N and 8X9Q, respectively. Source data are provided with this paper.

References

Zimmerly, S. & Semper, C. Evolution of group II introns. Mobile DNA 6, 7 (2015).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Roth, A., Weinberg, Z., Vanderschuren, K., Murdock, M. H. & Breaker, R. R. Natural circularly permuted group II introns in bacteria produce RNA circles. iScience 24, 103431 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Toor, N. et al. Tertiary architecture of the Oceanobacillus iheyensis group II intron. Rna 16, 57–69 (2010).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Pyle, A. M. The tertiary structure of group II introns: implications for biological function and evolution. Crit. Rev. Biochem. Mol. Biol. 45, 215–232 (2010).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Toor, N., Keating, K. S., Taylor, S. D. & Pyle, A. M. Crystal structure of a self-spliced group II intron. Science 320, 77–82 (2008).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Marcia, M. & Pyle, A. M. Visualizing group II intron catalysis through the stages of splicing. Cell 151, 497–507 (2012).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Qu, G. et al. Structure of a group II intron in complex with its reverse transcriptase. Nat. Struct. Mol. Biol. 23, 549–557 (2016).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Robart, A. R., Chan, R. T., Peters, J. K., Rajashankar, K. R. & Toor, N. Crystal structure of a eukaryotic group II intron lariat. Nature 514, 193–197 (2014).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Chan, R. T. et al. Structural basis for the second step of group II intron splicing. Nat. Commun. 9, 4676 (2018).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Haack, D. B. et al. Cryo-EM structures of a group II intron reverse splicing into DNA. Cell 178, 612–623 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Xu, L., Liu, T., Chung, K. & Pyle, A. M. Structural insights into intron catalysis and dynamics during splicing. Nature 624, 682–688 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

de Lencastre, A. & Pyle, A. M. Three essential and conserved regions of the group II intron are proximal to the 5′-splice site. RNA 14, 11–24 (2008).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Zhao, C. & Pyle, A. M. Structural insights into the mechanism of group II intron splicing. Trends Biochem. Sci. 42, 470–482 (2017).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Daniels, D. L., Michels, W. J. Jr. & Pyle, A. M. Two competing pathways for self-splicing by group II introns: a quantitative analysis of in vitro reaction rates and products. J. Mol. Biol. 256, 31–49 (1996).

Article
 
CAS
 
PubMed
 
Google Scholar

Pyle, A. M. Group II intron self-splicing. Annu. Rev. Biophys. 45, 183–205 (2016).

Article
 
CAS
 
PubMed
 
Google Scholar

Liu, Z.-X. et al. Hydrolytic endonucleolytic ribozyme (HYER) is programmable for sequence-specific DNA cleavage. Science 383, eadh4859 (2024).

Article
 
CAS
 
PubMed
 
Google Scholar

Chen, L., Huang, C., Wang, X. & Shan, G. Circular RNAs in eukaryotic cells. Curr. Genomics 16, 312–318 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Liu, C. X. & Chen, L. L. Circular RNAs: characterization, cellular roles, and applications. Cell 185, 2016–2034 (2022).

Article
 
CAS
 
PubMed
 
Google Scholar

Panda, A. C. Circular RNAs act as miRNA sponges. Adv. Exp. Med. Biol. 1087, 67–79 (2018).

Article
 
CAS
 
PubMed
 
Google Scholar

Hansen, T. B. et al. Natural RNA circles function as efficient microRNA sponges. Nature 495, 384–388 (2013).

Article
 
CAS
 
PubMed
 
Google Scholar

Zhou, W.-Y. et al. Circular RNA: metabolism, functions and interactions with proteins. Mol. Cancer 19, 172 (2020).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Pamudurti, N. R. et al. Translation of circRNAs. Mol. Cell 66, 9–21 (2017).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Wang, Y. et al. Expanding uncapped translation and emerging function of circular RNA in carcinomas and noncarcinomas. Mol. Cancer 21, 13 (2022).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Petkovic, S. & Müller, S. RNA circularization strategies in vivo and in vitro. Nucleic Acids Res. 43, 2454–2465 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Obi, P. & Chen, Y. G. The design and synthesis of circular RNAs. Methods 196, 85–103 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Jarrell, K. A. Inverse splicing of a group II intron. Proc. Natl Acad. Sci. USA 90, 8624–8627 (1993).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Chen, C. et al. A flexible, efficient, and scalable platform to produce circular RNAs as new therapeutics. Preprint at bioRxiv https://doi.org/10.1101/2022.05.31.494115 (2022).

Qiu, Z. et al. Clean-PIE: a novel strategy for efficiently constructing precise circRNA with thoroughly minimized immunogenicity to direct potent and durable protein expression. Preprint at bioRxiv https://doi.org/10.1101/2022.06.20.496777 (2022).

Lee, K. H. et al. Efficient circular RNA engineering by end-to-end self-targeting and splicing reaction using Tetrahymena group I intron ribozyme. Mol. Ther. Nucleic Acids 33, 587–598 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Sontheimer, E. J., Gordon, P. M. & Piccirilli, J. A. Metal ion catalysis during group II intron self-splicing: parallels with the spliceosome. Genes Dev. 13, 1729–1741 (1999).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Zhang, X., Li, S., Pintilie, G., Palo, M. Z. & Zhang, K. Snapshots of the first-step self-splicing of Tetrahymena ribozyme revealed by cryo-EM. Nucleic Acids Res. 51, 1317–1325 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Su, Z. et al. Cryo-EM structures of full-length Tetrahymena ribozyme at 3.1 Å resolution. Nature 596, 603–607 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Luo, B. et al. Cryo-EM reveals dynamics of Tetrahymena group I intron self-splicing. Nat. Catal. 6, 298–309 (2023).

Article
 
CAS
 
Google Scholar

Chung, K. et al. Structures of a mobile intron retroelement poised to attack its structured DNA target. Science 378, 627–634 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Zhao, C., Rajashankar, K. R., Marcia, M. & Pyle, A. M. Crystal structure of group II intron domain 1 reveals a template for RNA assembly. Nat. Chem. Biol. 11, 967–972 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Pyle, A. M., Fedorova, O. & Waldsich, C. Folding of group II introns: a model system for large, multidomain RNAs?. Trends Biochem. Sci. 32, 138–145 (2007).

Article
 
CAS
 
PubMed
 
Google Scholar

Zhao, C. & Pyle, A. M. The group II intron maturase: a reverse transcriptase and splicing factor go hand in hand. Curr. Opin. Struct. Biol. 47, 30–39 (2017).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Toor, N., Robart, A. R., Christianson, J. & Zimmerly, S. Self-splicing of a group IIC intron: 5′ exon recognition and alternative 5′ splicing events implicate the stem–loop motif of a transcriptional terminator. Nucleic Acids Res. 34, 6461–6471 (2006).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Li, X. et al. A unified mechanism for intron and exon definition and back-splicing. Nature 573, 375–380 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Bai, R. et al. Structural basis of U12-type intron engagement by the fully assembled human minor spliceosome. Science 383, 1245–1252 (2024).

Article
 
CAS
 
PubMed
 
Google Scholar

Shi, Y. Mechanistic insights into precursor messenger RNA splicing by the spliceosome. Nat. Rev. Mol. Cell Biol. 18, 655–670 (2017).

Article
 
CAS
 
PubMed
 
Google Scholar

Haufroid, M., Mirgaux, M., Leherte, L. & Wouters, J. Crystal structures and snapshots along the reaction pathway of human phosphoserine phosphatase. Acta Crystallogr. D Struct. Biol. 75, 592–604 (2019).

Article
 
CAS
 
PubMed
 
Google Scholar

Wilkinson, M. E., Fica, S. M., Galej, W. P. & Nagai, K. Structural basis for conformational equilibrium of the catalytic spliceosome. Mol. Cell 81, 1439–1452 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Fica, S. M., Oubridge, C., Wilkinson, M. E., Newman, A. J. & Nagai, K. A human postcatalytic spliceosome structure reveals essential roles of metazoan factors for exon ligation. Science 363, 710–714 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Woodson, S. A. & Cech, T. R. Alternative secondary structures in the 5′ exon affect both forward and reverse self-splicing of the Tetrahymena intervening sequence RNA. Biochemistry 30, 2042–2050 (1991).

Article
 
CAS
 
PubMed
 
Google Scholar

Inoue, T., Sullivan, F. X. & Cech, T. R. New reactions of the ribosomal RNA precursor of Tetrahymena and the mechanism of self-splicing. J. Mol. Biol. 189, 143–165 (1986).

Article
 
CAS
 
PubMed
 
Google Scholar

Kladwang, W., Hum, J. & Das, R. Ultraviolet shadowing of RNA can cause significant chemical damage in seconds. Sci. Rep. 2, 517 (2012).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Mastronarde, D. N. Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol. 152, 36–51 (2005).

Article
 
PubMed
 
Google Scholar

Wu, C., Huang, X., Cheng, J., Zhu, D. & Zhang, X. High-quality, high-throughput cryo-electron microscopy data collection via beam tilt and astigmatism-free beam-image shift. J. Struct. Biol. 208, 107396 (2019).

Article
 
CAS
 
PubMed
 
Google Scholar

Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017).

Article
 
CAS
 
PubMed
 
Google Scholar

He, J., Li, T. & Huang, S.-Y. Improvement of cryo-EM maps by simultaneous local and non-local deep learning. Nat. Commun. 14, 3217 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Pettersen, E. F. et al. UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci. 30, 70–82 (2021).

Article
 
CAS
 
PubMed
 
Google Scholar

Kappel, K. et al. De novo computational RNA modeling into cryo-EM maps of large ribonucleoprotein complexes. Nat. Methods 15, 947–954 (2018).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Smith, L. G., Zhao, J., Mathews, D. H. & Turner, D. H. Physics-based all-atom modeling of RNA energetics and structure. Wiley Interdiscip. Rev. RNA 8, e1422 (2017).

Article
 
Google Scholar

Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. D Struct. Biol. 74, 531–544 (2018).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Chen, V. B. et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12–21 (2010).

Article
 
CAS
 
PubMed
 
Google Scholar

Download references

Acknowledgements

Cryo-EM data were collected in the Electron Microscope Center of Fudan University at the Shanghai Medical School and at School of Life Sciences, Fudan University. The data were processed by High-performance Computing Clusters in the Electron Microscope Center of Fudan University at the Shanghai Medical School and at School of Life Sciences, Fudan University. We thank Z. Chen and A. Dong from the Center of Cryo-EM at Shanghai Medical School, Fudan University and the students from Z. Chen’s lab for helping 300-kV cryo-EM data collection and processing. This work was supported by the Natural Science Foundation of China (32471347 and 31971130 to J.M.).

Author information
Author notes

These authors contributed equally: Xiaobin Ling, Yuqi Yao.

Authors and Affiliations

State Key Laboratory of Genetic Engineering, Collaborative Innovation Center of Genetics and Development, Department of Biochemistry and Biophysics, School of Life Sciences, Fudan University, Shanghai, China

Xiaobin Ling, Yuqi Yao & Jinbiao Ma

RNA Therapeutics Institute, UMass Chan Medical School, Worcester, MA, USA

Xiaobin Ling

Contributions

X.L. conceptualized the project. J.M. supervised the project. X.L. and Y.Y. prepared the cryo-EM samples, performed the electrophoresis and analyzed the gel electrophoresis results. X.L. collected and processed the cryo-EM data. X.L. and J.M. built, refined and validated the atomic coordinate models. X.L. wrote the first version of the manuscript. All authors contributed to the preparation of the manuscript.

Corresponding authors

Correspondence to Xiaobin Ling or Jinbiao Ma.

Ethics declarations
Competing interests

Fudan University has successfully applied for a patent (no. 202410378541.1) with X.L., J.M. and Y.Y. as inventors. All other authors declare no competing interests.

Peer review
Peer review information

Nature Structural & Molecular Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editors: Sara Osman, Dimitris Typas, and Florian Ullrich, in collaboration with the Nature Structural & Molecular Biology team.

Additional information

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

Extended data
Extended Data Fig. 1 Biochemical analysis Cte 1.

a. Schematic diagram of the full-length Cte 1 and Cte 1_∆D4. b, The back-splicing activity of Cte 1_∆D4 is higher than that of the full-length Cte 1. c, Time-course analysis of the back-splicing activity of Cte 1_∆D4. d, The back-splicing activity of Cte 1_∆D4 over a wide range of temperatures. e, Dependance of back-splicing activity of Cte 1_∆D4 on divalent ions. f, Dependance of back-splicing activity of Cte 1_∆D4 on monovalent potassium, ions. The experiments shown in these panels were repeated more than three times with similar results.

Source data

Extended Data Fig. 2 Cryo-EM workflow of the wild type Cte 1_∆D4 in CaCl2 condition.

a, Cryo-EM data processing workflow of the wild type Cte 1_∆D4 at the CaCl2 condition. b-e, Particle distribution of full micrograph b, local resolution map c, FSC curves with the gold standard threshold of 0.143 for the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) conformation map d and eular angular distribution e.

Extended Data Fig. 3 Cryo-EM workflow of the wild type Cte 1_∆D4 in MgCl2 condition.

a, Cryo-EM data processing workflow of the wild type Cte 1_∆D4 at the MgCl2 condition. b-e, Particle distribution of full micrograph b, local resolution map c, FSC curves with the gold standard threshold of 0.143 for the BPWT conformation map d and eular angular distribution e.

Extended Data Fig. 4 Cryo-EM workflow of the mutant U537ins of Cte 1_∆D4 in CaCl2 condition.

a, Cryo-EM data processing workflow of the mutant U537ins at the CaCl2 condition. b-e, Particle distribution of full micrograph b, local resolution map for \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) and \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\) conformation c, FSC curves with the gold standard threshold of 0.143 for the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) and eular angular distribution d and FSC curves with the gold standard threshold of 0.143 for the \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\) and eular angular distribution e.

Extended Data Fig. 5 Cryo-EM workflow of the mutant A695U of Cte 1_∆D4 in MgCl2 condition.

a, Cryo-EM data processing workflow of the mutant A695U at the MgCl2 condition. b-e, Particle distribution of full micrograph b, local resolution map for \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) and \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\) conformation c, FSC curves with the gold standard threshold of 0.143 for the \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\) and eular angular distribution d and FSC curves with the gold standard threshold of 0.143 for the \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) and eular angular distribution e.

Extended Data Fig. 6 The secondary structure schematic diagram of the Cte 1_∆D4.

a, Secondary structure with sequence information and long-range interactions of the Cte 1_∆D4, with domains labeled D1–D6. Tertiary interactions are labeled with Greek letters. Exon binding sequences, intron binding sequences and Auxiliary Exon binding sequences are labeled EBS, IBS and AUX, respectively. b, Revised secondary structure reflecting the changes from the Pre conformation to the BP conformation of wild type Cte 1_∆D4. Domains are colored as in a.

Extended Data Fig. 7 Sequence conservation analysis of Cte 1_∆D4.

a–c, Sequence conservation analysis of D5, J2/3 and D6 in Cte 1_∆D4.

Extended Data Fig. 8 Tertiary interactions.

a, The A695 in J2/3 has no density in the cryo-EM map when Cte 1_∆D4 go through branching. The G696 and A697 in J2/3 form a triple interaction with D5 to stabilize the catalytic center. b, The 2’-OH of branch point A127 is ligated with the 5’-P of G220 forming 2’-5’ phosphodiester bond. c, AUX forms four triple interactions with IBS-EBS. Green represents IBS, purple represents EBS and yellow represents AUX. d, G537 in AUX stacks with mismatched A216. G539 kicked out the IBS-EBS interaction. e, The G539 in AUX does not interact with G220 in the 5′ SS when Cte 1_∆D4 is in the branching active state. f, The overall AUX-IBS-EBS interaction in the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) structure. g, AUX interacts with other 5’-exon and does not form triple interaction with IBS-EBS in the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) structure. h, IBS and EBS forms canonical Watson-Crick base pair. i, More 5’-exon interaction with 3’-exon and 3′ SS in the \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\) structure. j, P1 stabilizes the 5’-exon by interaction with A195. K, G221 of the 5′ SS interacts with U133 of 3′ SS when Cte 1_∆D4 is in the branching intermediate inactive state.

Extended Data Fig. 9 Analysis of the catalytic core in different states.

a, Zoomed-in view of the organization of catalytic elements in \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{U}}537{\rm{ins}}}\), \({{\rm{Pre}}}_{{\rm{inactive}}}^{{\rm{wt}}}\) and \({{\rm{Pre}}}_{{\rm{active}}}^{{\rm{A}}695{\rm{U}}}\) states of Cte-1_∆D4, showing noncanonical 5′ SS recognition by G220 in 5′ SS interaction with G539 in AUX and G221 in 5′ SS sticking into the D6 groove. b, Zoomed-in view of the organization of catalytic elements in \({{\rm{BI}}}_{{\rm{inactive}}}^{{\rm{A}}695{\rm{U}}}\), \({{\rm{BI}}}_{{\rm{active}}}^{{\rm{U}}537{\rm{ins}}}\) and BPwt states of Cte 1_∆D4, showing noncanonical 3′ SS recognition. c, The circular exon of the BPwt state in Cte 1_∆D4 is released. d, the linear exon of post-2f state in group IIC intron is retained, assisted by the maturase. e, Zoomed-in view of c and d.

Extended Data Fig. 10 Difference analysis and presumed evolutionary relationships between Cte 1_∆D4 and spliceosomes.

a, Cartoon diagram of the spliceosome from B* compex, C complex to P complex. b, Surface structure of linear exon generation process from spliceosome B* complex, C complex to P complex. Lime green represents the 3’-exon, forest green represents the 5’-exon, blue represents 5′ SS, red represents 3′ SS, orange represents the branch point A, dim grey represents the intron and dashed line represents the invisible intron corresponding to Cte 1_∆D4. The position of U6/U5/U2 as show in the picture. c, Zoomed-in view of b shows the spatial position of Prp8 is involved in splicing catalysis and anchoring exons. d, According to the process of linear-exon splicing by the spliceosome, a cartoon schematic diagram of the possible back-splicing process of circular exon in eukaryotes is speculated based on the existing structure. e, The back-splicing catalytic core of Cte 1_∆D4 shows that the function of AUX in prokaryotes is similar to that of the Prp8 protein in eukaryotes (up), and the speculated back-splicing process of circular exon by the spliceosome appears to be different because of their spatial positions. f, The back-splicing model of Cte 1_∆D4 showing the recognition of the noncanonical splice site and circular-exon production.

Supplementary information
Supplementary Information

Supplementary Tables 1 and 2.

Reporting Summary
Peer Review File
Supplementary Video 1

Dynamic conformational changes from prebranching to BI states.

Supplementary Video 2

D5 anchoring and teriary interactions.

Supplementary Video 3

The 5′ SS and 3′ SS recognition and important teriary interaction in these two processes.

Supplementary Video 4

The 2′-5′ and 3′-5′ phosphodiester bonds forming at catalytical center.

Source data
Source Data Fig. 1

Unprocessed gels.

Source Data Extended Data Fig. 1

Unprocessed gels.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article
Cite this article

Ling, X., Yao, Y. & Ma, J. Structures of a natural circularly permuted group II intron reveal mechanisms of branching and backsplicing. Nat Struct Mol Biol 32, 1101–1110 (2025). https://doi.org/10.1038/s41594-025-01489-6

Download citation

Received
10 June 2024

Accepted
09 January 2025

Published
27 February 2025

Version of record
27 February 2025

Issue date
June 2025

DOI
https://doi.org/10.1038/s41594-025-01489-6

Share this article

Anyone you share the following link with will be able to read this content:

Get shareable link

Provided by the Springer Nature SharedIt content-sharing initiative

Subjects
Cryoelectron microscopy
Enzyme mechanisms

Cancel

Preview · live