Figure 1. Reactivity of 2’, 3’ cyclic phosphate (>p) activated N₂₀ RNA pools.

(A) N₂₀>p pools are incubated in eutectic ice phases, spiked with 6% 5’FAM (Carboxyfluorescein)-labelled N₂₀>p to facilitate detection of ligation products (orange) by Urea-PAGE. (B) Scan of a denaturing gel for FAM-labelled products after 57 days of incubation either in eutectic ice (−9 °C) or at −80 °C, and +/- MgCl₂. Note that ligation of FAM-N₂₀>p is inhibited due to the blocked 5’-OH. (C) Densitogram of a SYBR-gold stained Urea-PAGE gel trace (-MgCl₂, −9 °C) (bottom panel) showing total ligation (both FAM-labelled and unlabelled ligation products (indicated by arrows)). The size distribution and size of the main ligation product (40mer) from deep sequencing indicates direct ligation of two eicosamers (inset). (D) Upper panel: Average nucleotide distribution profile of the 40mer ligation products. Lower panel: Changes in nucleotide composition compared to the unligated input N₂₀>p pool. Black arrows indicate the ligation site. (E) Upper panel: Average base-pairing frequencies of sequenced 40mers (as predicted by RNAfold (Lorenz et al., 2011): ‘Opening’ base pairs (obp, green; defined as hybridization to a downstream nucleotide) and ‘closing’ base pairs (cpb, hybridization to an upstream nucleotide, orange). Lower panel: Average difference of predicted base pairing between experimental pool and synthetic sequence pools (N = 3; 300,000 sequences each) generated in silico using experimental nucleotide frequencies shown in panel D. Grey indicates predicted unpaired bases (unp), black standard deviations.

Figure 1—figure supplement 1. N_-1pN₊₁ dinucleotide frequencies.

(A) Dinucleotide frequencies found in N₄₀ products from N₂₀>p ligation reactions (left panel). Differences between the expected dinucleotide frequencies calculated from N₂₀>pp starting material (using the frequencies from position 1 and 20 of the N₂₀>pp pool) are shown in the right panel. Equivalent graphs are shown for the semi-random sequence RNA pool ligation products of (B) C1₂₀-N₂₀>p × 5’OH-N₂₀-C2₂₀, (C) C3₂₀-N₂₀>p × 5’OH-N₂₀-C2₂₀, (D) C1₂₀-N₂₀>p × 5’OH-N₂₀-C4₂₀, and (E) the 60mer recombination product from C1₂₀-N₂₀-2’/3’p × 5’OH-N₂₀-C2₂₀ reactions shown in Figure 4D.

Figure 1—figure supplement 2. Evidence for intramolecular association and non-covalent complex formation in pools of random RNA oligonucleotides as judged by gel electrophoresis.

RNA oligonucleotide "FITC-N₂₀” (2.5 µM in 25 mM NaCl, 1 mM Tris•HCl pH 8.3, with or without 10 mM MgCl₂) incubated at room temperature or in ice, and analysed by native (non-denaturing) (top panel) or denaturing (bottom panel) PAGE. Under non-denaturing conditions, high molecular weight species can be observed following incubation at room temperature or in ice, which resolve to single bands under denaturing conditions indicating that these are non-covalently associated RNAs. A higher proportion of non-covalently associated RNA can be observed in the presence of MgCl₂. Single-sequence (i.e. non-random) 21mer and 30mer RNA oligonucleotides do not show formation of such intermolecular complexes and were included as molecular weight markers.