Apolar chemical environments compact unfolded RNAs and can promote folding

Abstract

It is well documented that the structure, and thus function, of nucleic acids depends on the chemical environment surrounding them, which often includes potential proteinaceous binding partners. The nonpolar amino acid side chains of these proteins will invariably alter the polarity of the local chemical environment around the nucleic acid. However, we are only beginning to understand how environmental polarity generally influences the structural and energetic properties of RNA folding. Here, we use a series of aqueous-organic cosolvent mixtures to systematically modulate the solvent polarity around two different RNA folding constructs that can form either secondary or tertiary structural elements. Using single-molecule Förster resonance energy transfer spectroscopy to simultaneously monitor the structural and energetic properties of these RNAs, we show that the unfolded conformations of both model RNAs become more compact in apolar environments characterized by dielectric constants less than that of pure water. In the case of tertiary structure formation, this compaction also gives rise to more energetically favorable folding. We propose that these physical changes arise from an enhanced accumulation of counterions in the low dielectric environment surrounding the unfolded RNA.

Figure 1

Formation of RNA secondary and tertiary structures in apolar environments. The relative dielectric constant (ε_r) is used to quantify solvent polarity. (A) Reported values of ε_r for several binary aqueous-organic solvent mixtures at 25°C. Solvent polarity monotonically decreases with increasing amounts of organic solvent. Values for water-MeOH, water-EtOH, and water-DMK were taken from (30). References (31) and (32) were used to derive the dielectric constants of water-IPA and water-ACN mixtures, respectively. (B) Color-coded sequence diagrams of the two model RNA constructs used in this study. Folding of the HP-RNA (left) results in the formation of five canonical basepairs, which hold together the hairpin secondary structure with a large poly-rA loop. Folding of the TL-RNA (right) involves docking the GAAA tetraloop into its cognate receptor. This common tertiary structural motif is held together by several noncanonical interactions (33). For the single-molecule fluorescence studies, both RNAs were modified with Cy3B and CF660R, which function as the FRET donor and acceptor fluorophores, respectively. For both RNAs, the efficiency of energy transfer from the donor to the acceptor is highest in the folded conformation.

Figure 2

Experimental design and analysis. (A) Schematic overview of a single-molecule confocal fluorescence microscope. The alternating (20 kHz) output from two diode lasers (515 and 642 nm) is directed into the back aperture of the water immersion objective using a dual-band dichroic mirror resulting in a diffraction-limited focal spot. The stokes-shifted fluorescence emitted from the sample is collected by the same objective and directed through a 100 μm pinhole to remove out-of-plane light. A polarizing beam splitter (PBS) cube and a set of long-pass (LP) dichroic mirrors spatially separates the emission into four photon streams based on polarization and color. The emitted photons in each of the four streams are detected by four separate single-photon avalanche photodiodes. (B) Cartoon depicting the confocal volume of an inverted confocal microscope (not to scale). The donor- and acceptor-labeled RNA molecules that stochastically diffuse through the confocal volume get excited by the focused laser beams producing a burst of photons. (C) Representative 2.5 s segment of a 900 s fluorescence time trace with 500 μs time bins depicting numerous bursts of photons resulting from alternating 515 and 642 nm excitation of the FRET-labeled RNAs (100 pM RNA, 25 mM HEPES, 12.5 mM NaOH, 250 mM KCl, 0.01% v/v Tween 20). (D) Insets on the left and right are representative 12.5 ms segments of the data in (C), each depicting a burst of photons. Transfer efficiency (E) and stoichiometry (S) values are calculated for each burst of photons recorded during the measurement using Eq. 1 and Eq. 2, respectively. The values of E arising from those bursts with more than 50 total corrected photons emitted by FRET-labeled molecules (i.e., 0.25 < S < 0.75) are then compiled into a transfer efficiency histogram. Histograms are fitted to a sum of two Gaussian functions to quantify the mean transfer efficiency, ⟨E⟩, and fractional abundance, Θ, of the folded and unfolded subpopulations, with typical experimental uncertainties of ⟨E⟩ ± 0.02 and Θ ± 0.03.

Figure 3

Effect of MeOH on RNA folding. Single-molecule transfer efficiency histograms for the HP-RNA (A) and TL-RNA (B) constructs in various mixtures of H₂O-MeOH are shown. Vertical dashed lines show how the high- and low-transfer-efficiency subpopulations change relative to the pure aqueous conditions (top). For the HP-RNA construct, the addition of MeOH as a cosolvent greatly increases the mean transfer efficiency of the unfolded subpopulation while slightly modulating the fractional abundance of the folded subpopulation. For the TL-RNA construct, MeOH increases the mean transfer efficiency of both subpopulations in addition to systematically increasing the fractional abundance of the folded subpopulation (100 pM RNA, 25 mM HEPES, 12.5 mM NaOH, 250 mM KCl, 0.01% v/v Tween 20, and the specified amount MeOH).

Figure 4

Apolar chemical environments compact unfolded RNAs. The mean transfer efficiencies of the folded and unfolded subpopulations of the (A) HP-RNA and (B) TL-RNA constructs plotted against the relative dielectric constant, ε_r, of various binary aqueous-organic solvent mixtures. Apolar conditions, characterized by values of ε_r that are less than pure water (≈ 78.5), strongly affect the mean transfer efficiency of the unfolded conformations of both RNAs, with almost no dependence on cosolvent identity. Notably, the mean transfer efficiency of folded HP-RNA, ^HP⟨E⟩_f, and the folded TL-RNA, ^TL⟨E⟩_f, were much less sensitive to changes in ε_r. The differential free energy change for RNA folding (ΔΔG°_fold) in apolar environments is shown for two RNA constructs designed to probe the formation of (C) secondary structure (HP-RNA) and (D) tertiary structure (TL-RNA). Based on our definition of ΔΔG°_fold (see Materials and methods), negative values represent conditions that stabilize the folded conformation (100 pM RNA, 25 mM HEPES, 12.5 mM NaOH, 250 mM KCl, 0.01% v/v Tween 20, and an amount of organic solvent to achieve the specified value of ε_r. The corresponding weight percentages for each measurement are shown in Fig. S3).

Figure 5

Effect of KCl concentration on the conformational dimensions and folding equilibria of RNA secondary and tertiary structure formation. Transfer efficiency histograms show compaction of the unfolded conformations of the (A) HP-RNA and (B) TL-RNA at elevated KCl concentrations. Conversely, the concentration of KCl has almost no effect on the mean transfer efficiency associated with folded conformations of the HP-RNA and TL-RNA, as shown by the dashed line at ^HP⟨E⟩_f ≈ 0.95 and ^TL⟨E⟩_f ≈ 0.8, respectively. However, in both cases, the two RNAs tend to favor the folded conformation at high concentrations of KCl (100 pM RNA, 25 mM HEPES, 12.5 mM NaOH, 0.01% Tween 20, and, unlike the previous experiments, no organic solvent but instead variable concentrations of KCl). Mean transfer efficiencies of the folded and unfolded subpopulations of the (C) HP-RNA and (D) TL-RNA constructs are plotted against KCl concentration. The differential free energy change for folding (ΔΔG°_fold), relative to 250 mM KCl, associated with the HP-RNA and (E) TL-RNA (F) is plotted against KCl concentration.