Characterization of group I introns in generating circular RNAs as vaccines

Abstract

Circular RNAs are an increasingly important class of RNA molecules that can be engineered as RNA vaccines and therapeutics. Here, we screened eight different group I introns for their ability to circularize and delineated different features that are important for their function. First, we identified the Scytalidium dimidiatum group I intron as causing minimal innate immune activation inside cells, underscoring its potential to serve as an effective RNA vaccine without triggering unwanted reactogenicity. Additionally, mechanistic RNA structure analysis was used to identify the P9 domain as important for circularization, showing that swapping sequences can restore pairing to improve the circularization of poor circularizers. We also determined the diversity of sequence requirements for the exon 1 and exon 2 (E1 and E2) domains of different group I introns and engineered a S1 tag within the domains for positive purification of circular RNAs. In addition, this flexibility in E1 and E2 enables substitution with less immunostimulatory sequences to enhance protein production. Our work deepens the understanding of the properties of group I introns, expands the panel of introns that can be used, and improves the manufacturing process to generate circular RNAs for vaccines and therapeutics.

Graphical Abstract

Figure 1.

Identification and characterization of novel group I intron PIE for circular RNA production. (A) Schematics showing the engineering of group I intron PIE system to make circular RNA. (B)IVT RNA precursor was circularized at 55°C in buffers containing 25 mM NaCl, 15 mM MgCl₂, 25 mM HEPES (pH 7.5). Subsequently, these samples were purified by columns and then treated with RNase R. All these RNA samples were analysed by agarose gel electrophoresis. The black arrow (upper) indicates precursor RNA and red arrow (lower) indicates circularized RNA. (C) IVT RNA precursor was folded with indicated conditions and was analysed by agarose gel electrophoresis. Densitometry analysis was performed to quantify the lower band intensity using ImageJ. The black arrow (upper) indicates precursor RNA and red arrow (lower) indicates circularized RNA. Representative gels from three independent experiments are shown. (D) The capped linear RNA with 100% N¹-methylpseudouridine (linear m1ψ) that encodes Gluc and indicated circular RNA (circularized at 55°C: 25 mM NaCl, 15 mM MgCl₂, 25 mM HEPES, pH 7.5) that contains an internal ribosomal entry site CVB3 and Gluc was transfected into A549 cells. Six hours after transfection, total RNA was isolated and IFNβ and IFIT1 transcript abundance was measured by RT-qPCR. Data are mean ± SEM from two independent experiments. Statistical significance was determined using a two-tailed t-test: ∗P< .05; ∗∗∗P< .001. (E) Two doses of LNP-circSPIKE or LNP-Sd circSPIKE were injected into BALB/c mice for a dose-response study. Serum samples were collected at 3 weeks after each dose, and neutralizing activity against Delta SARS-CoV-2 was determined. Each dot represents one mouse. The dashed line indicates the cut-off that is being considered as positive for neutralizing activity.

Figure 2.

Incorporation of modified nucleotides in circular RNA reduces reactogenicity and IRES engineering increases protein production. (A) A549 cells were transfected with linearSPIKE-HiBit that has a cap and 100% N¹-methylpseudouridine and with circSPIKE-HiBit (circularized at 55°C: 15 mM MgCl₂, 50 mM Tris–HCl, pH 7.0, and 1 mM DTT) with or without various indicated modifications. Approximately 72 h post-transfection, cellular supernatants were harvested for luminescence measurement to assess protein expression. (B) Indicated RNA was transfected into A549 cells. About 6 h post-transfection, cells were harvested for RNA extraction, and IFNβ transcript abundance was determined by RT-qPCR. Data are presented as fold change relative to linearSPIKE-Hibit, and error bars indicate SEM. Each experiment was performed two or three times independently. Statistical significance was determined using a two-tailed t-test comparing circSPIKE-Hibit with and without modifications: ∗P< .05 and ∗∗P< .01. (C) 293T cells were transfected with circMNG containing CVB3-WT or indicated mutations. About 2 days post-transfection, cells were analyzed by flow cytometry, and data were presented as fold change relative to CVB3-WT. Statistical significance was determined using ANOVA: ∗P< .05 and ∗∗P< .01. (D) circSPIKE-Hibit containing CVB3-WT or CVB3-62G was transfected into 293T cells and luminescence was measured to assess protein expression. Each experiment was performed three times independently and data are presented as fold change relative to CVB3-WT and error bars indicate SEM. Statistical significance was determined using a two-tailed t-test: ∗P< .05.

Figure 3.

Structural mapping of GII and PIE RNA. (A) A scatter plot showing the correlation between PIE circularization and GII splicing from different group I intron sequences. (B) Comparison of Gv and Po in SHAPE-MaP correlation between GII and PIE. (C–E) Mapping of intramolecular interactions in Po PIE Gluc and Gv PIE Gluc. Data are presented in arc plots and bar graphs showing interaction types. H1 and H2 indicate homology 1 and homology 2, respectively. Differential analysis was performed using an independent student’s t-test: ∗P< .05. (F) Key interactions in Po PIE Gluc and base-pairing regions from RNAcofold analysis are highlighted in red (I) and green (II). (G, H) Indicated constructs were IVT and their IVT products were analysed by E-gel EX electrophoresis. The green circle indicates circular RNA. The blue line indicates IVT RNA precursor, and the red circle indicates nicked circular RNA. Each experiment was performed two times independently and representative gel images are shown.

Figure 4.

Characterization of E1 and E2 sequence requirement for PIE circularization. (A) A schematic showing WT and N60 constructs used in these experiments. N represents randomized nucleotide. (B) Indicated IVT RNA was circularized at 55°C in buffers containing 25 mM NaCl, 15 mM MgCl₂, and 25 mM HEPES, pH 7.5 and column purified. The resulting samples were analysed by agarose gel electrophoresis. The black arrow (upper) indicates precursor RNA and the red arrow (lower) indicates circularized RNA. (C) Junction sequences of circular T4, An, Sh, and Po PIE N60 Gluc RNA. (D) IVT products and folded RNA of T4, An, Sh, and Po PIE v4.2 Gluc were analysed by E-Gel EX electrophoresis. (E) A schematic showing an engineered S1 motif for circular RNA purification. (F, G) An PIE Gluc RNA with or without S1 was folded and mixed with streptavidin C1 beads. After washing and elution, bound fractions were analysed by E-Gel EX electrophoresis. Green circle: circular RNA. Blue line: IVT RNA precursor. Red circle: nicked circular RNA. Each experiment was performed two or three times independently and representative gel images are shown. (H) Schematics of T4 and T4SdEx circ hEPO-Hibit. (I–K) Circular T4 and T4SdEx RNAs encoding CVB3 and hEPO-Hibit were transfected into A549 cells. Six hours after transfection, total RNA was isolated, and IFNβ transcript abundance was measured by RT-qPCR. For assessing protein expression, supernatants from transfected A549 and 293T cells were collected for luminescence measurement 72 h post-transfection. Data are mean ± SEM from two independent experiments and are presented as fold change relative to T4. Statistical significance was determined using a two-tailed t-test: ∗P< .05; ∗∗P< .01; ∗∗∗P< .001.