In vitro selection of ribozyme ligases that use prebiotically plausible 2-aminoimidazole–activated substrates

Significance Current models for the origin of life include an earlier period when prebiotic chemistry dictated the nonenzymatic copying of RNA polymers, followed by a period of ribozyme-catalyzed reactions using nucleoside-5′-triphosphate (NTP) substrates. Our study addresses the transition from nonenzymatic to ribozyme-catalyzed RNA template copying by determining whether ribozymes could have used substrates activated with 2-aminoimidazole, a prebiotically plausible leaving group, instead of NTPs. We identify ligase ribozymes that use RNA substrates activated with 2-aminoimidazole, thereby proving that ribozyme catalysis is compatible with the reactants of nonenzymatic template-directed ligation. This work suggests that the transition from nonenzymatic to ribozyme-catalyzed RNA replication could have involved ribozymes that utilize inherently reactive RNA substrates.

library using HiScribe TM T7 polymerase in a 2.4 mL reaction incubated for 3 h at 37 °C. The T7 reaction was then digested by DNase I, extracted with phenol chloroform, ethanol-precipitated, and purified by 10% PAGE. The PAGE-purified T7 product RNA was then digested using the restriction enzyme Ava-II, which cleaves DNA/RNA hybrid duplexes to generate defined 3′ ends on the RNA strand with free hydroxyls groups (2). Ava-II digests were incubated for 10 h at 37 °C and contained 3.3 µM PAGE-purified T7 product, 10 µM DNA template AD1, 1x buffer, and 5000 units/mL Ava-II enzyme. Digests were then treated with DNase I, extracted with phenol chloroform, ethanol precipitated, and PAGE purified.
To select for ligase activity, RNA pools were incubated at room temperature for 10-120 min. in reactions that contained 1 µM RNA library, 1.1 µM RNA template LT1, 1. Tween-20) for 15 min., twice with buffer C for 5 min., and three more times with buffer A. The magnetic beads were then suspended in 95% formamide containing 10 mM EDTA and heated for 6 min. at 65 °C to release the captured RNA. This sample was then diluted with 4.5 × volume of 300 mM sodium acetate with 100 ng / µL glycogen and ethanol-precipitated.
For rounds 1-3, ethanol-precipitated RNA was reverse transcribed in a reaction containing ProtoScript ® II reverse transcriptase and 10 µM DNA primer RT1 according to manufacturer recommendations. Reactions were incubated at 42 °C for 3 h, and then heat-denatured for 5 min. at 80 °C. The denatured reverse transcription reactions were then diluted 10 × in a PCR reaction containing Taq DNA polymerase, 0.5 µM DNA primer PCR1, and 0.5 µM DNA primer PCR2 using manufacturer recommendations. PCR reactions were first incubated at 95 °C for 2 min.; then underwent 16 cycles of 94 °C for 30 s, 60 °C for 1 min., and 72 °C for 1 min.; and finally, 72 °C for 10 min.
The reverse transcribed pool was then PCR amplified as described above for only 10 cycles, purified using a QIAquick spin column, and amplified another 8 cycles using PCR. These steps were undertaken to reduce PCR amplification by the reverse transcription primer RT2, which is contained in the heat-denatured reverse transcription reaction. QIAquick purified PCR reactions were then transcribed by T7 RNA polymerase, and the RNA digested by Ava-II as described above.
High throughput sequencing: To prepare the selected pools for sequencing, adaptor sequences were introduced onto the dsDNA pools resulting from each round of selection using several rounds of PCR with different primer sets. First, a 100 µL PCR reaction containing 25 ng dsDNA PCR product with 0.5 µM DNA primer PS1 and 0.5 µM DNA primer PS2, was amplified by Taq DNA polymerase for 4 cycles as described above. The PCR product was purified using QIAquick PCR Purification Kit (Qiagen), and then further amplified for 7 cycles in a second PCR reaction containing 40 ng dsDNA, 5 µM DNA primer PS3, and 5 µM DNA primer PS4. The purified product of the second PCR reaction was amplified once again in a 50 µL reaction containing 1 unit of Q5® Sequencing reads were filtered and counted using an R script. Reads were filtered based upon the presence of the predefined stem sequences and the quality (Q ≥ 20) of the sequence corresponding to the loop region of the RNA library. The loop region of each read was then counted to provide sequence abundance for each round of the selection. Loop sequences were then clustered by thresholding the score of global pairwise alignment using the Biostrings package. (NEB), 2.5 U/mL thermostable inorganic pyrophosphatase (TIPPase) (NEB), 4 mM each NTP, and 30-60 pmol/mL DNA template, for 2 h at 37°C. The RNA was subsequently treated with DNase I, phenol-chloroform extracted, and PAGE-purified. Our initial examination of ligase activity in Figure 2 and S4 uses RNA prepared by Ava-II digestion as described above.
Determination of phosphodiester bond regiospecificity Ligation reactions containing 10 pmol ribozyme, 10 pmol substrate oligo LS6, and 12 pmol of the DNA template LT2 were incubated for 2-6 h before DNase digestion to remove the template. Following phenolchloroform extraction, unreacted LS6 substrate was removed by 3-4 washes on a centrifugal 30 kDa molecular weight filter (EMD Millipore). The resulting concentrate was denatured by incubation at 95 °C for 6 minutes in the presence of 6 M urea and 1 mM EDTA, and then digested by RNase T1 for 24 h at 55 °C. These digestions were then extracted by phenol-chloroform and analyzed by 20% PAGE. To determine the sensitivity of this assay, a calibration curve was generated using a synthetic 17 nt RNA oligo standard to quantify the limit of detection in terms of pmol input RNA (Fig. S14). We then quantified the 16 nt band intensities (corresponding to the RNase T1 digestion product of the 3′-5′ linked ligation product) for each of the reactions for RS1-RS10 and used these values to calculate pmol input RNA based upon the calibration curve. The sensitivity of our assay is then given by dividing the pmol limit of detection by the calculated pmol input RNA of the 16 nt digestion products. Multiplying this fraction by 100 gives the reported percentage value corresponding to the maximum amount of 2′-5′ linked ligation product possible in the sample without being detected in our assay.

Secondary structure determination by SHAPE
The secondary structure of RS1 was determined by adapting previously published RNA SHAPE protocols         (Table S3). The template sequence was not changed in any of the reactions. The fraction of ligated product at 2 h was measured for the 16 different conditions and presented in a heat map. These results indicate that at least one intact base-pair is required for ligation; an AU or GU pair in the substrate-template duplex is sufficient to allow an AC mismatch in the primer-template duplex. Similarly, a GC pair in the primer-template duplex is sufficient to allow an CU (to some extent a UU) mismatch in the substrate-template duplex.
Other mismatch combinations at the ligation junction do not support ligation.  for ligases RS1-RS10 with templates strands of different sizes. The red dot indicates the length of template that reduces ligation yield by 50% relative to the standard 16 nt RNA template.