Exploring one-state downhill protein folding in single molecules

Abstract

A one-state downhill protein folding process is barrierless at all conditions, resulting in gradual melting of native structure that permits resolving folding mechanisms step-by-step at atomic resolution. Experimental studies of one-state downhill folding have typically focused on the thermal denaturation of proteins that fold near the speed limit (ca. 10(6) s(-1)) at their unfolding temperature, thus being several orders of magnitude too fast for current single-molecule methods, such as single-molecule FRET. An important open question is whether one-state downhill folding kinetics can be slowed down to make them accessible to single-molecule approaches without turning the protein into a conventional activated folder. Here we address this question on the small helical protein BBL, a paradigm of one-state downhill thermal (un)folding. We decreased 200-fold the BBL folding-unfolding rate by combining chemical denaturation and low temperature, and carried out free-diffusion single-molecule FRET experiments with 50-μs resolution and maximal photoprotection using a recently developed Trolox-cysteamine cocktail. These experiments revealed a single conformational ensemble at all denaturing conditions. The chemical unfolding of BBL was then manifested by the gradual change of this unique ensemble, which shifts from high to low FRET efficiency and becomes broader at increasing denaturant. Furthermore, using detailed quantitative analysis, we could rule out the possibility that the BBL single-molecule data are produced by partly overlapping folded and unfolded peaks. Thus, our results demonstrate the one-state downhill folding regime at the single-molecule level and highlight that this folding scenario is not necessarily associated with ultrafast kinetics.

Fig. 1.

Cartoon that compares folding scenarios at the single-molecule level. Blue indicates native, green midpoint, and red unfolding conditions.

Fig. 2.

Equilibrium chemical denaturation of BBL measured by FRET. Blue (top scale), urea; green (bottom scale), GdmCl. The urea/GdmCl ratio is 2.75. Circles represent the bulk data and the curve the global fit to the two-state model. The two red lines signify the native (linear) and unfolded (curved) baselines used for the global two-state fit. The dashed lines are the baselines in the linear approximation. Inverted triangles show the position of the maximum on the single-molecule FEHs of Fig. 3.

Fig. 3.

Single-molecule FRET-efficiency histograms of BBL as a function of urea (blue) and GdmCl (green). The histograms were obtained from 50-μs bins with N_T = 40 (urea) or N_T = 30 (GdmCl). The red curves are fits to a lognormal distribution, and the black vertical lines signal the extreme FRET efficiency values. The red star signals the FEH closest to the denaturation midpoint.

Fig. 4.

The width of the FEH versus the position of its maximum. The width is represented by the variance (σ²) from fits of the histograms to a lognormal distribution (red curves shown in Fig. 3). The abscissa shows the position of the FRET efficiency maximum in the FEH. Blue, urea; green, GdmCl; cyan, shot-noise width for N_T = 40; red, simulation for a two-state model based on the fit to the bulk unfolding curve assuming N_T = 40 and no conformational exchange. For the data in GdmCl, the shot-noise variance is slightly higher (e.g., 0.008 at E = 0.4).

Fig. 5.

BBL conformational dynamics near the chemical-denaturation midpoint. Data are shown in blue and the fits to a single exponential decay in red. (A) Autocorrelation function of the FRET efficiency fluctuations undergone by individual free-diffusing BBL molecules (labeled with A488–A594) at 6.1 M urea and 279 K. (B) BBL relaxation decay measured in unlabeled BBL by infrared absorption at 1,649 cm^-1 after a nanosecond laser-induced T jump from 278 to 282 K in the presence of 2.2 M ¹³C-GdmCl (equivalent to ca. 6.1 M urea). Because BBL has maximum stability against cold-heat denaturation at approximately 280 K, the change in native signal upon this 4 K jump is only approximately 0.5% (measured by Fourier transform infrared spectroscopy and CD).

Fig. 6.

Global fitting of the FEH as function of chemical denaturant to the two-state model. The position of the native (green circles) and unfolded (red circles) states obtained from the global fit of the FEHs to the 3G Gopich–Szabo theory for a two-state system parameterized with the BBL rate data from Fig. S7 are shown. Filled circles are for urea (top scale) and open circles for GdmCl (bottom scale). The scales correspond to urea/GdmCl ratio of 2.75. The thin continuous lines are polynomial fits to guide the eye. The fit to the bulk data (gray curve) and the obtained from the FEH (blue small circles) are also shown for reference. (Inset) Changes in native probability as function of denaturant concentration obtained from the fit.

Fig. 7.

Quantitative analysis of the FEH near the denaturation midpoint. (Left) FEH of BBL measured at 6.1 M urea using T_b = 100 μs and N_T = 110. (Center) Fit of the FEH to a two-state model using the 3G Gopich–Szabo theory (dark blue). The 3G curves representing the native state (green), unfolded state (red), and conformational exchange (black) are also shown. (Right) Prediction for the one-state downhill folding scenario obtained from stochastic dynamic simulations (dark blue). (Center and Right) The experimental histogram is shown for reference as a shaded area in the background. The residuals from the fits are shown on top.