Fitness landscape of a dynamic RNA structure

Abstract

RNA structures are dynamic. As a consequence, mutational effects can be hard to rationalize with reference to a single static native structure. We reasoned that deep mutational scanning experiments, which couple molecular function to fitness, should capture mutational effects across multiple conformational states simultaneously. Here, we provide a proof-of-principle that this is indeed the case, using the self-splicing group I intron from Tetrahymena thermophila as a model system. We comprehensively mutagenized two 4-bp segments of the intron. These segments first come together to form the P1 extension (P1ex) helix at the 5' splice site. Following cleavage at the 5' splice site, the two halves of the helix dissociate to allow formation of an alternative helix (P10) at the 3' splice site. Using an in vivo reporter system that couples splicing activity to fitness in E. coli, we demonstrate that fitness is driven jointly by constraints on P1ex and P10 formation. We further show that patterns of epistasis can be used to infer the presence of intramolecular pleiotropy. Using a machine learning approach that allows quantification of mutational effects in a genotype-specific manner, we demonstrate that the fitness landscape can be deconvoluted to implicate P1ex or P10 as the effective genetic background in which molecular fitness is compromised or enhanced. Our results highlight deep mutational scanning as a tool to study alternative conformational states, with the capacity to provide critical insights into the structure, evolution and evolvability of RNAs as dynamic ensembles. Our findings also suggest that, in the future, deep mutational scanning approaches might help reverse-engineer multiple alternative or successive conformations from a single fitness landscape.

Fig 1. Determining the fitness landscape of a dynamic RNA structure.

(A) The sequence and secondary structure of the Tet-119(C20A) group I intron with its 5’ and 3’ exonic context. Secondary structure conformations during sequential formation of P1ex and P10 are highlighted in the blow-ups. The two sub-regions that were subjected to mutagenesis (N₂..N₅ and N₁₈..N₂₁) are shaded grey. (B) Schematic representation of the knt-intron construct, library generation, and selection protocol. (C) In the presence of kanamycin, self-splicing activity (molecular fitness) of the group I is coupled to organismal fitness as intron removal is required for reconstitution of the knt open reading frame.

Fig 2. Fitness across intron genotypes.

(A) Distribution of fitness effects at 30°C and 37°C. (B) Shannon diversity of intron genotype pools under different conditions. (C) Similarity in genotype pool composition across all replicates and conditions measured as Bray-Curtis (BC) dissimilarity, where BC = 1 indicates maximum dissimilarity between samples. (D) Fitness of intron genotypes at 30°C as a function of Hamming distance (i.e. the number of mutational steps away from the master sequence).

Fig 3. Causes and correlates of variable fitness across intron genotypes.

(A) Fitness weakly correlates with predicted minimum free energy of the intron. For orientation, note that the predicted minimum free energy (ΔG) of the master sequence is -362.8 (B) Fitness varies according to the number of guanosines or cytosines (#GC) in the N₂..N₅ and N₁₈..N₂₁ regions. (C) Fitness varies as a function of the number of strong or weak base-pairs that could be formed in P1ex assuming that base-pairing follows the established master/wildtype pattern (see Fig 1A). (D) Distribution of pairwise epistasis values for genotypes that are two mutations away from the master sequence (Hamming distance = 2). ε>0 indicates positive epistasis, ε<0 indicates negative epistasis. (E) Pairwise epistasis for genotypes in (D) by position and mutation. Diagrams on the left highlight the N₃/N₂₀ couple, where mutations that are predicted to lead to base-pairing are associated with positive epistasis.

Fig 4. Assessing the contribution of individual nucleotide identities to fitness across multiple structural conformations.

(A) Contribution to XGBoost-predicted relative fitness across all intron genotypes, as measured by Shapley’s additive explanation (SHAP) scores, of three example site/nucleotide features. More positive SHAP scores are associated with higher fitness. (B) The average contribution across all genotypes of all individual site/nucleotide features, measured as ΔSHAP = SHAP_present—SHAP_absent, where SHAP_present and SHAP_absent correspond to the mean SHAP score of all genotypes where a given nucleotide at a given site is present and absent, respectively. (C) Fitness landscape at 30°C as a function of RNA stability of P1ex and P10 across all genotypes assuming bases are aligned to pair as in the master/wildtype structure (see Fig 1A, Materials and methods). There are 211 unique energy values across all 4⁸ P1ex genotypes. These were consolidated into ten bins of increasing stability for visualization purposes. The 21 unique energy values across 4⁴ P10 genotypes are shown in full as 21 bins of increasing stability. Bar heights correspond to the median fitness in each bin. (D) Fitness as a function of N₂/N₂₁ genotype, with a focus on cytosines. (E) Minor groove width associated with different N₂/N₂₁ genotypes as determined using molecular dynamics simulations (see Materials and methods). (F) Three overlaid representative conformations of the P1/P1ex helix (randomly sampled from the final 50 ns of each simulation) for the master sequence and the C₂/C₂₁ genotype.

Fig 5. Asymmetric fitness effects across the N₂..N₅ and N₁₈..N₂₁ sub-regions.

(A) Proportion of gains in the model (see main text) contributed by site/nucleotide identity features at N₂..N₅ and N₁₈..N₂₁. The solid line corresponds to the mean contribution made by a sub-region across 100 random samples, where individual gains are randomly shuffled across site/nucleotide identity features. Dashed lines correspond to 95% confidence intervals. (B) Pairwise epistasis for double mutants where both mutations are located in N₁₈..N₂₁ (orange), both mutations are located in N₂..N₅ (brown), or N₂..N₅ and N₁₈..N₂₁ carry one mutation each (grey). (C) The correlation of fitness effects (γ) of intron mutants at various mutational distances from the master sequence.