Multiscaled exploration of coupled folding and binding of an intrinsically disordered molecular recognition element in measles virus nucleoprotein

Abstract

Numerous relatively short regions within intrinsically disordered proteins (IDPs) serve as molecular recognition elements (MoREs). They fold into ordered structures upon binding to their partner molecules. Currently, there is still a lack of in-depth understanding of how coupled binding and folding occurs in MoREs. Here, we quantified the unbound ensembles of the α-MoRE within the intrinsically disordered C-terminal domain of the measles virus nucleoprotein. We developed a multiscaled approach by combining a physics-based and an atomic hybrid model to decipher the mechanism by which the α-MoRE interacts with the X domain of the measles virus phosphoprotein. Our multiscaled approach led to remarkable qualitative and quantitative agreements between the theoretical predictions and experimental results (e.g., chemical shifts). We found that the free α-MoRE rapidly interconverts between multiple discrete partially helical conformations and the unfolded state, in accordance with the experimental observations. We quantified the underlying global folding-binding landscape. This leads to a synergistic mechanism in which the recognition event proceeds via (minor) conformational selection, followed by (major) induced folding. We also provided evidence that the α-MoRE is a compact molten globule-like IDP and behaves as a downhill folder in the induced folding process. We further provided a theoretical explanation for the inherent connections between "downhill folding," "molten globule," and "intrinsic disorder" in IDP-related systems. Particularly, we proposed that binding and unbinding of IDPs proceed in a stepwise way through a "kinetic divide-and-conquer" strategy that confers them high specificity without high affinity.

Keywords: flexible binding; flexible recognition; free-energy surface; hybrid structure-based model; multiscale simulation.

Fig. 1.

Schematic representation of the coupled folding and binding that the α-MoRE of N_TAIL undergoes upon binding to XD. The picture on the Left corresponds to the free form of the α-MoRE (residues 484–504) of N_TAIL, which interconverts rapidly between unfolded conformations and discrete partially helical conformations. On the Right is shown the structure of the XD–N_TAIL complex as derived by X-ray crystallography [PDB ID code 1T6O (40)]. The α-MoRE within N_TAIL folds into a perfect α-helix upon interaction with XD.

Fig. 2.

Deciphering the heterogeneous disordered ensemble of the α-MoRE of N_TAIL by a clustering algorithm based on the REMD simulation at 298 K, and comparing the simulated chemical shifts with experimental data. (A) Structural representations of the top eight clusters. The structures are taken from the center of each cluster and colored by an index along the chain from blue (N terminus, Gly484) to red (C terminus, Ile504). RMSD refers to backbone atoms deviations from the bound form. Population distribution and helical content of all of the clusters are shown in SI Appendix, Fig. S3. The top eight clusters represent 70.5% of the population of the MD trajectory. (B) Helical propensity of the α-MoRE from REMD simulation in the presence of explicit solvent at 298 K. It shows the probability of forming an α-helix (or helicity) for each N_TAIL residue according to three different criteria (SI Appendix). (C) Comparison of simulated and experimental secondary chemical shifts for the α-MoRE. Green: experimental data of the free form. Red: simulated data of the free form from REMD simulation at 298 K. Blue: experimental data of the bound form at room temperature. Gray: estimated errors for the chemical shift prediction. (D) Correlation between simulated and experimental secondary chemical shifts with a correlation coefficient of 0.73. The gray points represent the residues, Asp487 and Asp493, which have relatively bad correlations.

Fig. 3.

Propensity of the free form of the α-MoRE of N_TAIL to sample folded conformations in a wide temperature range. Conformers located within the conformational region with R_COM less than 1.8 nm are considered as corresponding to the bound form. The graph shows the RMSD distribution of the free form of N_TAIL at various temperatures. The RMSD_N distribution in the bound state at 298 K is used as the baseline. The free form of N_TAIL can sample a wide range of conformations from 2.0 to 8.0 Å.

Fig. 4.

The free-energy surface supports a coupled folding and binding mechanism mainly according to induced folding rather than conformational selection despite the preexistence of folded conformational states before binding. (A) Free-energy surface as a function of R_COM (binding order parameter) and (folding order parameter). Free-energy surfaces as a function of other folding order parameters and binding order parameters are shown in SI Appendix, Fig. S12, indicating similar results. (B) Schematic free-energy surface showing that folding upon binding takes place according to an induced folding (labeled as IF) mechanism and not to a conformational selection (labeled as CS) mechanism despite the preexistence of unbound-folded state (UB-F). Note that, in the folding upon binding process, there are four possible states: unbound-unfolded (UB-UF), bound-unfolded (B-UF), unbound-folded (UB-F), and bound-folded (B-F). The schematic picture gives a good illustration that the preexistence of the UB-F state is a necessary but not a sufficient condition for a conformational selection mechanism (5).

Fig. 5.

Evidence that the free form of the α-MoRE behaves more like an MG than a random coil from both atomistic physics-based and structure-based models. (A) Distribution of the radius of gyration (R_g) of the α-MoRE of N_TAIL in both bound and unbound state (based on structure-based simulation). (B) Free-energy profile of the α-MoRE in the unbound state as a function of RMSD and R_g from the structure-based model (12 μs). (C) Free-energy profile as a function of RMSD and R_g from the physics-based model (REMD with explicit solvent). RMSD quantifies the similarity to a folded helical conformation and R_g quantifies the compactness. (D) In the protein trinity model, proteins can exist in any of the three forms: folded (having an ordered 3D structure), collapsed (MG-like), or extended (random coil-like), and function can arise from transitions among them.

Fig. 6.

Free-energy landscape illustrating that the coupled folding and binding process of the α-MoRE of N_TAIL occurs according to a mixed mechanism of conformational selection followed by induced folding. The α-MoRE of N_TAIL in the unbound state is a compact MG consisting of conformers rapidly fluctuating between disordered and ordered forms with various degree of helicity. As binding occurs, XD selects the partially ordered conformations of the α-MoRE of N_TAIL from the conformational ensemble, and subsequently induces downhill folding of the α-MoRE that adopts an α-helical conformation over its entire length. Therefore, the recognition process proceeds via a mechanism to which both conformational selection and induced folding contribute.

Fig. 7.

Kinetic divide-and-conquer strategy allows IDPs to rapid associate as well as fast dissociate from their partners. The binding of two proteins is generally understood by a two-step process: . Here, [A] is an ordered protein or an IDP, [B] is a rigid partner of [A], [E] is the encounter complex, and [C] is the final bound complex. There are three states: an unbound state, an encounter state, and a bound state. In such three-state model, [A] has to encounter [B] to form [E] before evolving to form [C]. The encountering between [A] and [B] is under the control of translational and rotational diffusion. The major differences in the binding mechanism between ordered proteins and IDPs lie in the evolution step. IDPs form and break interactions with their partners in a gradual way (red) rather than according to an all-or-none discrete mechanism as observed for the docking of ordered proteins (blue). In this way, the association/dissociation of IDPs is determined by the low-energy barrier transitions and becomes relatively fast.