A microscopic view of miniprotein folding: enhanced folding efficiency through formation of an intermediate

Abstract

The role of polypeptide collapse and formation of intermediates in protein folding is still under debate. Miniproteins, small globular peptide structures, serve as ideal model systems to study the basic principles that govern folding. Experimental investigations of folding dynamics of such small systems, however, turn out to be challenging, because requirements for high temporal and spatial resolution have to be met simultaneously. Here, we demonstrate how selective quenching of an extrinsic fluorescent label by the amino acid tryptophan (Trp) can be used to probe folding dynamics of Trp-cage (TC), the smallest protein known to date. Using fluorescence correlation spectroscopy, we monitor folding transitions as well as conformational flexibility in the denatured state of the 20-residue protein under thermodynamic equilibrium conditions with nanosecond time resolution. Besides microsecond folding kinetics, we reveal hierarchical folding of TC, hidden to previous experimental studies. We show that specific collapse of the peptide to a molten globule-like intermediate enhances folding efficiency considerably. A single point mutation destabilizes the intermediate, switching the protein to two-state folding behavior and slowing down the folding process. Our results underscore the importance of preformed structure in the denatured state for folding of even the smallest globular structures. A unique method emerges for monitoring conformational dynamics and ultrafast folding events of polypeptides at the nanometer scale.

Fig. 1.

Thermal unfolding of modified TC peptides monitored by fluorescence intensity measurements. (A) Engineered NMR structure (Protein Data Bank ID code 1L2Y) of fluorescently modified TC*. The fluorophore MR121 labeled to the ε-amino group of lysine (K8) and the Trp residue (W6) are shown in red and blue, respectively. (B) Relative quantum yields (Φ_rel) and corresponding folded fractions (F_F) measured from the MR121 fluorescence of labeled (red) and the Trp fluorescence of unlabeled (blue) TC (circles). Data recorded from the mutant TC(I4G) are shown as squares. Gray lines are sigmoidal data fits. Note that, for better comparison, Φ_rel for unlabeled TC is shown as 1 – Φ_Trp (Φ_Trp being a measure of the relative Trp fluorescence), because, in contrast to MR121 in TC*, the fluorescence intensity of Trp in TC increases upon unfolding. Φ_rel values measured from MR121 in labeled control peptides TC*F1 and TC*F2(W6F) are shown as black and green circles, respectively. Melting curves for TC* (circles) and TC*(I4G) (squares), calculated from normalized FCS amplitudes (see Materials and Methods), are shown in magenta.

Fig. 2.

Folding of TC* monitored by FCS. (A) Normalized autocorrelation function [Norm. G(t)] measured from the TC fragments TC*F1 (black), TC*F2(W6F) (green), and the unstructured control peptide p53*1 (red) at 20°C. (Inset) Nanosecond relaxation kinetics of TC*F1 and corresponding exponential fit. (B) Temperature-dependent autocorrelation functions of TC* measured in 5-K temperature intervals (only samples spaced by 10 K are shown; 5–55°C). (Inset) Folding (red) and unfolding (black) rates, k, calculated from the relaxation times and corresponding normalized amplitudes, follow Arrhenius folding behavior at moderate temperatures. Gray lines are linear data fits. (C) Autocorrelation functions of fast relaxation kinetics of TC* measured at various GdmCl concentrations. Additionally, corresponding data recorded from the fragment TC*F1 at 6 M GdmCl are shown. (D) Fraction of the denatured (F_D) and the unfolded (F_U) ensemble of TC measured as a function of the GdmCl concentration, monitored by Trp fluorescence of TC (black) and by the amplitude of the nanosecond relaxation kinetics (K₂) of TC* (red). The amplitude K₂ of the fragment TC*F1 measured as a function of the GdmCl concentration is shown in blue. (E) Illustration of the TC folding ensemble as revealed by FCS. The unfolded state (U) is characterized by extended random-coil conformations giving rise to fluorescence quenching, with fluorophore (red)–Trp (blue) interaction kinetics on nanosecond time scales. Under mildly denaturing conditions, the denatured ensemble (D) is dominated by a collapsed intermediate with Trp being solvent-exposed. The labeled fluorophore is within a short interaction distance to Trp, leading to efficient fluorescence quenching with interaction kinetics beyond the time window of FCS. In the folded state (F), the fluorophore is well shielded from quenching interactions with Trp, being buried in the hydrophobic pocket.

Fig. 3.

Temperature-dependent folding dynamics of the mutant TC*(I4G). (A) Expanded view of autocorrelation functions recorded from TC*(I4G) at various temperatures on the nanosecond to microsecond time scale. (B) Correlation plot of the denatured fraction (F_D) versus the unfolded fraction (F_U) of TC* (filled green bars) and TC*(I4G) (filled red bars). Open green bars show corresponding data recorded from TC* at pH 3.0. F_D values are monitored by steady-state MR121 fluorescence intensities, whereas F_U values correspond to the normalized amplitudes of nanosecond relaxation kinetics (K₂) extracted from FCS data. The theoretical blue line represents 100% correlation.

Fig. 4.

Folding dynamics of TC* monitored under acidic conditions. Autocorrelation functions recorded from TC* under acidic conditions at 5°C (black), 35°C (green), and 55°C (red). Gray lines are data fitting curves obtained by using the model described in Materials and Methods.