Topological constraints are major determinants of tRNA tertiary structure and dynamics and provide basis for tertiary folding cooperativity

Abstract

Recent studies have shown that basic steric and connectivity constraints encoded at the secondary structure level are key determinants of 3D structure and dynamics in simple two-way RNA junctions. However, the role of these topological constraints in higher order RNA junctions remains poorly understood. Here, we use a specialized coarse-grained molecular dynamics model to directly probe the thermodynamic contributions of topological constraints in defining the 3D architecture and dynamics of transfer RNA (tRNA). Topological constraints alone restrict tRNA's allowed conformational space by over an order of magnitude and strongly discriminate against formation of non-native tertiary contacts, providing a sequence independent source of folding specificity. Topological constraints also give rise to long-range correlations between the relative orientation of tRNA's helices, which in turn provides a mechanism for encoding thermodynamic cooperativity between distinct tertiary interactions. These aspects of topological constraints make it such that only several tertiary interactions are needed to confine tRNA to its native global structure and specify functionally important 3D dynamics. We further show that topological constraints are conserved across tRNA's different naturally occurring secondary structures. Taken together, our results emphasize the central role of secondary-structure-encoded topological constraints in defining RNA 3D structure, dynamics and folding.

Figure 1.

Secondary structure cartoon (A) and TOPRNA implementation (B) of a 2-nt bulge two-way junction. Filled and open circles indicate paired and single-stranded nucleotides, respectively.

Figure 2.

Secondary structure limits the set of global conformations accessible to WT tRNA. (A) Secondary structure and labeling scheme of tRNA^Phe. Loop residues are bolded and cut locations marked. (B) The Euler angle convention used to describe the relative 3D orientation of RNA helices. Shown is a representative TOPRNA snapshot of the AC- and T-stems, colored as in (A), with the A- and D-stems and connecting loops not shown for clarity. (C) Fraction of possible (α_h, β_h, γ_h) angles sampled between pairs of tRNA helices by the WT (black), cut A/D-loop (red) and cut V-loop (blue) simulations. (D) The mutual information (MI) between pairs of interhelical Euler angles measured with respect to a common reference helix. The two helices whose orientations are being correlated are bolded.

Figure 3.

Secondary structure prevents tRNA from forming non-native tertiary contacts. (A) Free energy cost of forming different interloop residue–residue contacts in WT tRNA. Contacts observed in the crystal structure are outlined in black. (B) The free energy cost of forming different contacts upon cutting the A/D-loop. The ΔG^topo is shown in upper left triangle using the same color scale as (A). In the lower triangle, the ΔΔG^topo between the cut A/D-loop relative to WT tRNA is shown, with the color scale to the right. (C) Entropies and all P-bead RMSDs of the 500 best-packed conformers sampled by WT tRNA. Conformations that possess only native-consistent contacts and have D-T loop–loop contacts are colored black, those that possess only native-consistent contacts but lack D-T loop–loop contacts colored red, and those that possess native-inconsistent contacts are colored blue (see methods). Note that high entropies indicate conformers that are thermodynamically favored. (D) Superposition of the crystal structure (blue) and the five highest entropy best-packed conformers from the WT tRNA simulation (red). (E) Entropies and all P-bead RMSDs of the 500 best-packed conformers sampled by cut A/D-loop tRNA. The color scheme is the same as in (C).

Figure 4.

Tertiary interactions confine WT tRNA to native-like conformations. (A) Diagram of conserved tRNA residues and tertiary interactions. Residues conserved in <90% of tRNA species are indicated by circles (38). Conserved tertiary interactions are labeled and semiconserved base triples are drawn as gray lines. (B) RMSD distributions of simulations of unrestrained WT tRNA (black), WT tRNA restrained by all nine conserved tertiary interactions (tRNA_9R; light gray) and WT tRNA restrained by the four non-redundant interactions (tRNA_4R; dark gray). (C) The average structure of tRNA_9R. (D) Three representative structures from the tRNA_9R simulation illustrating the orientations sampled between the D- and AC-stems. Structures are superimposed by the AC stem. Residues of the A- and T-stems and connecting loops are not colored for clarity. (E) 2D projections of the (α_h, β_h, γ_h) angles sampled between the AC- and D-stems. Angles only sampled by unrestrained tRNA are shown in black; angles sampled by both unrestrained and tRNA_4R are shown in dark gray; and angles sampled by tRNA_9R, tRNA_4R and unrestrained WT tRNA are shown in light gray. Red points correspond to angles measured from 109 different tRNA crystal structures. A reference cartoon of the three angles is shown at left. Note that as discussed in the text, examining only one pairwise set of (α_h, β_h, γ_h) angles provides an incomplete picture of the extent to which topological constraints confine tRNA conformation.

Figure 5.

tRNA's tertiary interaction network is cooperative. (A) The mean cooperativity of jointly forming n number of loop–loop or stacking contacts, averaged over all combinations of native contacts (solid line, circles) or combinations containing at least one non-native contact (dashed line, triangles). Results are shown for the WT (black), cut A/D-loop (red) and cut V-loop (blue) simulations. Combinations that were observed ≤10 times were excluded from the averages. (B) The thermodynamic cooperativity among tRNA's tertiary interactions, computed with Equation (6). Restrained tertiary interactions are numbered on the x-axis according to the shown key. Loop–loop, loop–stem and stacking contacts for which cooperativities were measured are shown along the y-axis. Cooperativities are not computed for loops that have an active restraint placed between them.

Figure 6.

Naturally occurring tRNA secondary structures conserve topological constraints. (A) Class I tRNAs. The A/D and V-loop are shown as blue and red lines, respectively, with tested length variations labeled. Naturally observed lengths are bolded. Full sequences are shown in Supplementary Figure S9. (B) Example Class II tRNA. Inserted V-stem is shown in red, and G26·U44 pair shown by a dashed line. Note the additional D-loop nt (gray) and 3-bp D-stem. (C) Relative fraction of 3×(α_h, β_h, γ_h) interhelical conformations sampled by different tRNAs compared to WT. (D) Mutual information between different pairs of interhelical Euler angles relative to WT tRNA. The helices whose orientations are being correlated are bolded in the key. (E) The probability and cooperativity of jointly forming loop–loop contacts between the D and T loops and both native interhelical stacks. (F) The fraction of 500 best-packed conformers that possess native-like folds, weighted by entropy as described in methods. The gray background in (C–F) is used to highlight natural tRNA variants.