Protein-free catalysis of DNA hydrolysis and self-integration by a ribozyme

Abstract

Group II introns are ancient self-splicing ribozymes and retrotransposons. Though long speculated to have originated before translation, their dependence on intron-encoded proteins for splicing and mobility has cast doubt on this hypothesis. While some group II introns are known to retain part of their catalytic repertoire in the absence of protein cofactors, protein-free complete reverse splicing of a group II intron into a DNA target has never been demonstrated. Here, we demonstrate the complete independence of a group II intron from protein cofactors in all intron-catalyzed reactions. The ribozyme is capable of fully reverse splicing into single-stranded DNA targets in vitro, readily hydrolyzes DNA substrates and is even able to unwind and react with stably duplexed DNA. Our findings make a protein-free origin for group II introns plausible by expanding their known catalytic capabilities beyond what would be needed to survive the transition from RNA to DNA genomes. Furthermore, the intron's capacity to react with both single and double-stranded DNA in conjunction with its expanded sequence recognition may represent a promising starting point for the development of protein-free genomic editing tools.

Graphical Abstract

Figure 1.

(A) Schematic of the proposed reverse-splicing reaction network including all known sub-reactions. 1: The lariat base pairs with the DNA substrate. 2: The 3′-hydroxyl of the lariat performs a nucleophilic attack on the DNA substrate at the insertion site. This results in the transfer of the 3′-exon to the 3′-end of the lariat, producing the lariat-3′-exon intermediate. 3: The 3′-hydroxyl of the 5′-exon performs a nucleophilic attack on the 2′-5′ bond of the branchpoint adenosine, transferring the 5′-exon to the 5′-end of the now linearized intron. 4: By using water as a nucleophile, the intron irreversibly hydrolyzes the DNA at the insertion site. 5: The lariat base pairs to a free 5′-exon that is produced in Steps 2 and 4. 6: Same as Step 3 but without the 3′-exon attached to the lariat. 7: The 5′-exon can be released by the lariat-3′-exon intermediate, which can bind the lariat in Step 5. (B) Secondary structure diagram of the modified G2I P.li.LSU.I2. The sequence is based on the crystallization construct designed by Robart et al. (12). Letters in black font indicate mutations introduced in this study. Roman numerals mark the functional domains of the ribozyme. The motifs EBS1, EBS2 and EBS3 interact with the substrate DNA. (C) The insertion site of the intron on the DNA substrate. Recognition of the insertion site is mediated by base pairing between the exon binding sites (EBSs) in the intron and their complementary intron binding sites (IBSs) in the DNA. The intron inserts itself between the 5′-exon (left) and the 3′-exon (right), at the site denoted by the triangle.

Figure 2.

Biochemical characterization of protein-free reverse splicing into ssDNA. (A) Percentage of substrate converted to RSP after 1 h of incubation at 25°C of 100 nM intron lariat and 200 nM DNA substrate, in a buffer containing 40 mM Tris–HCl pH 7.5, 0.5 M NH₄Cl, 0.001% PEG 8000 and varying concentrations of MgCl₂. Values were corrected for the 600 mM MgCl₂ sample (see ‘Data analysis’ section in ‘Materials and methods’ section). (B) Example denaturing PAGE of the reverse-splicing reaction time course. Conditions are the same as in panel (A), except the reaction was carried out at 30°C and in 50 mM MgCl₂. Bands are labeled with a drawing of the corresponding species. (C), (D), (E) and (F) are kinetic traces of RSP, free 5′-exon, debranching product and the DNA substrate, respectively. Conditions as in panel (B), except that those reactions represented by triangles contained 200 nM DNA substrate, and those with circles contained 50 nM DNA substrate. Datapoints are the averages of triplicates. Error bars represent ± standard deviations (n = 3).

Figure 3.

(A) Reaction scheme of P.li.LSU.I2 reverse splicing into a dsDNA. The intron unwinds the dsDNA, base pairs with the IBSs and catalyzes reverse splicing and hydrolysis of DNA. The reaction is expected to produce a mixture of the two products. (B) For analysis, the reaction products were reverse-transcribed with a primer specific to the intron. The hydrolysis/debranching product (front, solid) and the RSP (back, faded) produce indistinguishable cDNA. Reverse transcription was followed by a PCR with a reverse primer specific to the intron, and a forward primer specific to the plasmid DNA. (C) Two per cent agarose TAE gel of RT-PCRs in absence or presence of lariat (LAR), and plasmid substrate (SUB) either with (pS) or without (pN) the intron insertion site. (D) Section of the Sanger sequencing chromatogram of the RT-PCR product from panel (C) showing that the intron is sequence-specifically inserted at the expected 5′-splice-site.

Figure 4.

(A) Illustration of intron lariat reacting with the dsDNA hairpin substrate carrying a 5′-Cy5 label in the 5′-exon, and an internal FAM modification in the loop of the 3′-exon. The lariat unwinds the hairpin DNA at the IBSs (Step 1) and catalyzes the reversible reverse-splicing (Steps 2 and 3) or quasi-irreversible hydrolysis reaction followed by debranching (Steps 4 and 6). (B) PAGE analysis of reverse splicing into an ssDNA (ss) and the hairpin (hp) dsDNA substrate. The contrast was increased in the region bordered with the dotted line. Reactions were conducted at 25°C for 20 h in 200 mM MgCl₂, 0.5 M NH₄Cl, 0.001% PEG 8000 and 40 mM Tris–HCl pH 7.5 using 1.5 μM intron lariat and 100 nM of either DNA substrate. The band labeled with * in the ssDNA sample is inferred to be the remaining RSP (see high-contrast gel image, Supplementary Figure S7).