Self-splicing introns make better circRNAs

 

Shen et al., Nature Commun, 2025

PIET & CIRC, two cleaner, faster ways to build therapeutic-grade circles

A new Nature Communications paper introduces two ribozyme-based workflows for in-vitro circRNA: PIET (Permuted Intron-Exon through Trans-splicing) and CIRC (Complete self-splicing Intron for RNA Circularization). CIRC uses intact group I or II introns to drive the second splicing step directly, yielding higher circularization, milder reaction conditions, fewer concatemers, options for scarless products, and workable purification with RNase R or oligo(dT). They even circularized a ~12 kb payload encoding full-length dystrophin (427 kDa) and showed protein expression.

---

Why this is a big deal (and how it differs from PIE)

Most industrial circRNAs are made with PIE (permuted group I intron), which needs specific split sites and often pushes Mg²⁺, temperature, and time—great for yield, rough on RNA integrity, and not always friendly to very large payloads or “scarless” designs. CIRC keeps the intron whole and bypasses the first splicing step, driving circularization via the second step alone. In head-to-head tests, CIRC delivered higher circularization, fewer concatemers, and works at lower Mg²⁺/pH with faster kinetics than classic PIE.

For historical context: PIE has been the workhorse since 2018 for long circRNAs with therapeutic expression, but it still demands careful engineering and can leave sequence “scars.” CIRC reduces that engineering burden and opens up more intron choices.

---

The two methods in a glance

• PIET: Split group I intron halves and mix them in trans with the payload; Mg²⁺ required, GTP not required (so you’re effectively skipping step 1), and tuning the 5′-intron:intermediate ratio boosts efficiency. It’s precise and controllable, but in the head-to-heads it didn’t beat PIE. 

• CIRC: Use an intact group I or II intron with minimal exons to circularize the payload directly via the second step; no GTP, lower Mg²⁺ and pH tolerated, rapid circularization, and less concatemer formation. Works with multiple natural introns (Tetrahymena, Azoarcus, more), not just Anabaena. 

---

Practical build tips they validated (good to know at the bench)

• Exons & homology: As little as ~16 nt per exon (or ~27 nt total split asymmetrically) is enough for fast CIRC; longer homology arms speed circularization but can hurt translation, so the authors cap at ~19–50 nt and recommend an AC30 spacer before an IRES for coding circles. 

• 5′ initiation G’s: T7 transcription likes ≥2 G at the 5′ end; this boosts yield without significantly hurting circularization. 

• 3′-end heterogeneity: T7 can produce ragged 3′ ends; using 2′-OMe-modified PCR primers (or mutant T7) improved circularization—because CIRC’s splice junction sits right at that 3′ end. 

• During IVT: For huge payloads, the intact intron can self-act prematurely. Adding cGMP and using custom buffers during IVT suppressed ribozyme activity so you get full-length precursors; circularize after. 

---

Size matters: expressing full-length dystrophin from circRNA

CIRC circularized payloads up to ~12 kb and demonstrated protein expression (EGFP, luciferases, base editor ABE8e, and full-length dystrophin). That breaks through the ~9 kb pain point reported for many platforms and makes true protein replacement (e.g., DMD) plausible for circRNA—not just minigenes. 

---

Scarless (or nearly) circRNAs and immunogenicity

• Scarless design: By exploiting minimal exons (e.g., CUU/AA or more generally NNUAA/NNUA motifs) and using internal payload structure (e.g., within CVB3 IRES) as the “homology,” the team made scarless or near-scarless circles without translation penalties. 

• Innate signaling: After phosphatase treatment, CIRC-made circRNAs showed low immunogenicity in cells, comparable to T4-ligase circRNA; scar length alone didn’t predict responses (sequence context likely matters). This aligns with prior work showing circRNAs can be less immunogenic than linear RNAs when purified well. 

---

Purification that doesn’t make you cry

• RNase R is easier with CIRC than PIE because CIRC precursors’ 3′ ends are more digestible; faster reactions also limit intron self-ligation, which otherwise resists RNase R. 

• Oligo(dT) trick: Inserting a poly(A) segment inside the intron lets you deplete byproducts on oligo(dT) beads/columns, leaving circRNA in the flow-through—simple, scalable, and not very costly.  

---

Limitations & open questions

• Generalizability: Many introns worked, but not all; sequence context still matters, and mass-scale GMP implementation needs testing. 

• Modifications: Certain base modifications can impair ribozymes and IRES activity. Making modified, translationally competent circRNAs at scale remains an active frontier. 

• In vivo circularization: CIRC is an in-vitro platform; whether “intact intron” strategies translate to in-cell circularization is a separate question. (Parallel efforts with group II/CP introns are emerging.) 

---

Where this lands in the therapeutic timeline

CircRNA is now in human trials (e.g., RXRG001 expressing AQP1 for radiation-induced xerostomia). Cleaner, scalable, large-payload methods like CIRC could widen the therapeutic scope to full-length protein replacement and multi-ORF constructs. 

---

If you want to try CIRC next week, start here

1. Pick your intron (Ana/Tet/Azo as strong baselines) and identify a NNUAA/NNUA motif for minimal exons. Keep exons ~16 nt each; homology ≤50 nt; add AC30 before IRES if coding. 

2. Control IVT (mutant T7 or 2′-OMe primers; add cGMP to suppress premature splicing for huge payloads). 

3. Circularize mild & fast (lower Mg²⁺/pH than PIE; shorter times). 

4. Purify smart (early RNase R or oligo(dT) flow-through if you embedded a poly(A) in the intron). Finish with phosphatase to cut immunostimulation. 

---

Source

Shen, Y., Li, B., Dong, L. et al. Self-splicing RNA circularization facilitated by intact group I and II introns. Nat Commun 16, 7376 (2025). https://doi.org/10.1038/s41467-025-62607-y

===

Keywords: circRNA, circular RNA, RNA circularization, self-splicing introns, group I intron, group II intron, ribozyme, PIE method, PIET method, CIRC method, trans-splicing, second transesterification step, exon ligation, minimal exons, homology arm, spacer sequence, IRES, CVB3 IRES, scarless circRNA, low immunogenicity, concatemer reduction, Mg2+ dependence, mild pH conditions, rapid circularization, T7 polymerase, 2′-OMe primers, 3′-end heterogeneity, cGMP inhibition, RNase R purification, oligo(dT) purification, poly(A) insertion, dystrophin expression, Duchenne muscular dystrophy, large payload circRNA, protein replacement therapy, in vitro transcription, translation efficiency, Anabaena intron, Tetrahymena intron, Azoarcus intron, Nature Communications 2025

Hashtags: #circRNA, #CircularRNA, #RNATherapeutics, #Ribozyme, #GroupIIntron, #GroupIIIntron, #SelfSplicing, #RNAEngineering, #RNACircularization, #PIE, #PIET, #CIRC, #TransSplicing, #ExonLigation, #IRES, #ScarlessDesign, #LowImmunogenicity, #RNaseR, #OligodT, #PolyA, #T7Polymerase, #cGMP, #LargePayload, #Dystrophin, #DMD, #ProteinReplacement, #GeneTherapy, #SyntheticBiology, #MolecularBiology, #NatureCommunications