Cooperative RNA Folding under Cellular Conditions Arises From Both Tertiary Structure Stabilization and Secondary Structure Destabilization

Abstract

RNA folding has been studied extensively in vitro, typically under dilute solution conditions and abiologically high salt concentrations of 1 M Na⁺ or 10 mM Mg²⁺. The cellular environment is very different, with 20-40% crowding and only 10-40 mM Na⁺, 140 mM K⁺, and 0.5-2.0 mM Mg²⁺. As such, RNA structures and functions can be radically altered under cellular conditions. We previously reported that tRNA^phe secondary and tertiary structures unfold together in a cooperative two-state fashion under crowded in vivo-like ionic conditions, but in a noncooperative multistate fashion under dilute in vitro ionic conditions unless in nonphysiologically high concentrations of Mg²⁺. The mechanistic basis behind these effects remains unclear, however. To address the mechanism that drives RNA folding cooperativity, we probe effects of cellular conditions on structures and stabilities of individual secondary structure fragments comprising the full-length RNA. We elucidate effects of a diverse set of crowders on tRNA secondary structural fragments and full-length tRNA at three levels: at the nucleotide level by temperature-dependent in-line probing, at the tertiary structure level by small-angle X-ray scattering, and at the global level by thermal denaturation. We conclude that cooperative RNA folding is induced by two overlapping mechanisms: increased stability and compaction of tertiary structure through effects of Mg²⁺, and decreased stability of certain secondary structure elements through the effects of molecular crowders. These findings reveal that despite having very different chemical makeups RNA and protein can both have weak secondary structures in vivo leading to cooperative folding.

Figure 1

Structures of FL yeast tRNA^phe and its model helical fragments (HF) with their predicted folds. (A) Three-dimensional folded structure of yeast FL tRNA^phe (PDB: 1ehz) and (B) FL tRNA secondary structure and the model HF derived from each stem. In panel (A) Mg²⁺ (cyan) and Mn²⁺ (black) ions associated with tRNA are shown as spheres. The colors of each structural element are the same in both panels. FL tRNA tertiary contacts are provided in Figure S1. The sequences of the model HF are the same as in FL tRNA^phe and are predicted to form the shown hairpins by the Mfold server for each model sequence.

Figure 2

First derivative curves of thermal denaturation experiments of FL tRNA and SSS under physiological ionic and crowded conditions. First derivative curves of thermal denaturation experiments in (A) 0% PEG200, (B) 20% PEG200 or PEG8000, and (C) 40% PEG200 or PEG8000 with increasing concentrations of Mg²⁺ from 0 to 0.5 to 2.0 mM (light to dark colors). The T_M’s are provided in Tables S1, S2 and S3. Solutions contain a background of 10 mM sodium cacdoylate (pH 7.0) and 140 mM KCl.

Figure 3

Difference in T_M of FL tRNA, SSS, and each of the four model HF in crowder compared to buffer. T_M in crowder minus T_M in buffer in (A) 0, (B) 0.5, and (C) 2.0 mM Mg²⁺. In all panels T_M differences in 20% PEG200, 40% PEG200, 20% PEG4000, 40% PEG4000, 20% PEG8000, 40% PEG8000, and 20% PEG20000 are in orange, red, light blue, dark blue, light green, dark green, and purple, respectively. The T_M’s are shown on the plot and in Tables S1, S2 and S3. Solutions contain a background of 10 mM sodium cacdoylate (pH 7.0) and 140 mM KCl.

Figure 4

Cooperative folding can be induced through crowder-driven secondary structure destabilization. First-derivative curves of thermal denaturation experiments of FL tRNA and SSS with increasing concentrations of (upper row) PEG200 and (lower row) PEG8000 in (A) 0, (B) 0.5 and (C) 2.0 mM Mg²⁺, with increasing concentrations of PEG200 or PEG8000 from 0% to 20% to 40% (w/v). The T_M’s are provided in Tables S1, S2 and S3. Solutions contain a background of 10 mM sodium cacdoylate (pH 7.0) and 140 mM KCl.

Figure 5

Nucleotide and helical stem fitting of temperature-dependent ILP data in buffer and 20% PEG200 with 0.5 mM Mg²⁺. Nucleotide fits (non-global fitting) and helical fits (globally fit for nucleotides shown with a bracket) were performed on buffer and 20% PEG200 samples, respectively, to obtain a T_M for unfolding of each nucleotide or each stem in (A) D SL, (B) AC SL, and (C) TΨC SL. The T_M values and residuals from the fits are provided in each figure and in Table 1. Fits could not be obtained in panel C in buffer; points are connected by lines to guide the eye.

Figure 6

Global fitting of temperature-dependent ILP data in buffer and 20% PEG200 with 0.5 mM Mg²⁺. The same ILP data Figure (5) were fit globally across each stem of the (A) D SL, (B) AC SL, and (C) TΨC SL for a single T_M of the RNA and to look for two-state behavior. The T_M values and residuals from the fits are provided in each figure and in Table 1.

Figure 7

Melting temperatures of FL tRNA and HF obtained by optical melting and ILP in buffer and crowded conditions. T_M values were obtained by (A) optical melting on FL tRNA or the model HF, or by (B) ILP on FL tRNA, data from which were globally fit for a T_M of either FL tRNA or the HF in FL tRNA. Blue closed and open circles overlapped in panel A left-hand side.

Figure 8

Physiological crowded and ionic conditions favor a compact folded state of FL tRNA. The tRNA^phe crystal structure (PDB ID: 1ehz) was aligned with FL tRNA SAXS bead models in (A) 0, (B) 0.5, and (C) 2.0 mM Mg²⁺, without (top) and with (bottom) 20% PEG8000. The bead models were made using DAMMIF and DAMAVER and overlayed with the crystal structure using SUPCOMB.