Making target sites in large structured RNAs accessible to RNA-cleaving DNAzymes through hybridization with synthetic DNA oligonucleotides

Abstract

The 10-23 DNAzyme is one of the most active DNA-based enzymes, and in theory, can be designed to target any purine-pyrimidine junction within an RNA sequence for cleavage. However, purine-pyrimidine junctions within a large, structured RNA (lsRNA) molecule of biological origin are not always accessible to 10-23, negating its general utility as an RNA-cutting molecular scissor. Herein, we report a generalizable strategy that allows 10-23 to access any purine-pyrimidine junction within an lsRNA. Using three large SARS-CoV-2 mRNA sequences of 566, 584 and 831 nucleotides in length as model systems, we show that the use of antisense DNA oligonucleotides (ASOs) that target the upstream and downstream regions flanking the cleavage site can restore the activity (kobs) of previously poorly active 10-23 DNAzyme systems by up to 2000-fold. We corroborated these findings mechanistically using in-line probing to demonstrate that ASOs reduced 10-23 DNAzyme target site structure within the lsRNA substrates. This approach represents a simple, efficient, cost-effective, and generalizable way to improve the accessibility of 10-23 to a chosen target site within an lsRNA molecule, especially where direct access to the genomic RNA target is necessary.

Graphical Abstract

Figure 1.

RNA cleavage activity comparison of a 10–23 DNAzyme, dZ-1, with a 584-nt RNA substrate, lsRNA-1 (A, B) and short 25-nt RNA substrate, sRNA-1 (C, D). The dZ-1 target site is shown in green for both lsRNA-1 (A) and sRNA (C) with A-U (diribonucleotides at cleavage site) shown in pink and dZ-1 is shown in yellow.

Figure 2.

Improvement of dZ-1 activity and accessibility against lsRNA-1 with rationally designed ASO-1A and B. ASO-1A (red sequence) and B (blue sequence), both 40-nt in length, hybridized to upstream and downstream sequence elements of the dZ-1 binding site (shown in green) within lsRNA-1. Addition of ASO-1A and B restored dZ-1 activity, enabling dZ-1 to cleave 82% of lsRNA-1 after 90 min with a k_obs of 0.078 min⁻¹ (A). Addition of ASO-1A and B to lsRNA-1 also increased the accessibility of dZ-1 by reducing RNA secondary structure of the dZ-1 binding site, as shown by in-line probing analysis (B). The in-line probing reaction was conducted using lsRNA-1 without ASO-1A or B (‘RNA in-line’) and with ASO-1A and B (‘RNA in-line + ASO-1A&B’).

Figure 3.

RNA cleavage activity comparison of a 10–23 DNAzyme, dZ-2 with the 831-nt lsRNA-2 (A) and short 25-nt sRNA-2 (B), followed by optimization with rationally designed ASOs (C). The dZ-2 target site is shown in green for both lsRNA-2 (A) and sRNA-2 (B) substrates with AU (diribonucleotides at the cleavage site) shown in pink. dZ-2 (shown in yellow) has limited activity with lsRNA-2 (k_obs = 1.6 × 10⁻⁴ min⁻¹) (A), but cleaves sRNA-2 with a k_obs of 0.047 min⁻¹ and Y₉₀ of 86% (B). ASO-2A and B (shown in red and blue, respectively), 40-nt in length, hybridized to upstream and downstream sequence elements of the dZ-2 binding site within lsRNA-2 and restored dZ-2 activity, cleaving 81% of lsRNA-2 after 90 min with a k_obs of 0.017 min⁻¹ (C). Addition of both ASO-2A and B to lsRNA-2 also increased the accessibility of the dZ-2 by reducing RNA secondary structure of the dZ-2 binding site as shown by in-line probing analysis (D).

Figure 4.

RNA cleavage activity comparison of a 10–23 DNAzyme, dZ-3 with 566-nt lsRNA-3 (A) and sRNA-3 (B), followed by optimization with rationally designed ASOs (C). The dZ-3 target site is shown in green for both lsRNA-3 (A) and sRNA-3 (B) with AU (the diribonucleotides at the cleavage site) shown in pink. dZ-3 (shown in yellow) is poorly active with lsRNA-3 (k_obs = 6.3 × 10⁻⁵ min⁻¹) (A), but cleaves sRNA-3 with a k_obs of 0.030 min⁻¹ and Y₉₀ of 76% (B). ASO-3A and B (shown in red and blue, respectively), 40-nt in length, hybridized to upstream and downstream sequence elements of the dZ-3 binding site and cleaved 18% of lsRNA-3 after 90 mins with a k_obs of 0.0027 min⁻¹ (C).

Figure 5.

Experimental ASO screen for dZ-3 with experimental screening ASO (esASO) binding sites within lsRNA-3. Each screening reaction contained equimolar amounts of dZ-3 (green) and 40-nt esASO (blue) starting with esASO-1 from the 3′-end of lsRNA-3, with 12 esASOs total needed to cover the entire sequence space of lsRNA-3 (A). After the 24 h reaction, esASO-11 was found to elicit the most significant enhancment in dZ-3 cleavage for lsRNA-3 out of the 12 esASOs tested, but was not more effective than ASO-3A and B (B). The same trend was observed after 90 min (C).

Figure 6.

Improvement of dZ-3 (yellow) activity and accessibility for lsRNA-3 with variant 1 of esASO-11 (ASO-V1, black), which was identified through an experimental ASO screen and possesses an additional 20-nt to its 3′ terminus (A). ASO-V1 increased the rate (k_obs) and fraction cleaved (Y₉₀) of lsRNA-3 by dZ-3 (black curve) compared to the same reaction with rationally designed ASO3-A and B (green curve). The addition of ASO-V1 also increased the accessibility of the dZ-3 binding site within lsRNA-3 by reducing RNA secondary structure of the binding site (green box) as shown by in-line probing analysis (B).