Figure 3. Effects of m⁶A modifications on RNA base-pairing and structure.

(A) Left panel: stable Watson–Crick base-pairing between m⁶A-U is only possible in the energetically less favorable anti-conformation of the m⁶A base. Right panel: two RNA strands containing an m⁶A-U instead of an A-U base-pair have a fourfold to ninefold decreased annealing rate constant but a dissociation rate constant which is not significantly changed by the methylation (Shi et al, 2019). (B) Left panel: Hoogsteen–Hoogsteen m⁶A-A base-pair is stabilized compared with Hoogsteen–Hoogsteen A–A (Roost et al, 2015). Right panel: base-pairing between pre-mRNA 5′splice site (5′SS) and modified (green) and unmodified (red) U6 snRNA (yellow) and U5 snRNA (orange). Successful base-pairing between modified U6 snRNA and 5′SS does not require a conserved AAG motif upstream of the 5′SS, but this conserved motif is required with unmodified U6 snRNA for stable base-pairing (Ishigami et al, 2021; Parker et al, 2022). (C) Left panel: depletion of m⁶A marks during METTL3 knock-out favors secondary structure formation of otherwise single-stranded endogenous RNAs, which triggers the recognition of these aberrant double-stranded RNAs by RIG-I and MDA5 resulting in an innate immune response (Gao et al, 2020). Right panel: double-stranded virus RNA from vesicular stomatitis virus gets methylated by METTL3, which leads to loss of structure and an impaired RLR recognition and innate immune response (Qiu et al, 2021). Structures used in this figure are as follows: METTL3/METTL14 (PDB-ID: 5IL1). ChimeraX version 1.6 was used for the visualization of experimental and predicted structures (Goddard et al, 2018; Pettersen et al, 2021). Panel (B) was adapted from Ishigami et al (2021) under the Creative Commons CC By license (license: https://creativecommons.org/licenses/by/4.0/).