Sequence-Dependent Conformational Landscapes of Intrinsically Disordered Proteins Reveal Asymmetric Chain Compaction

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Intrinsically disordered proteins (IDPs) lack stable 3D structures, exhibiting dynamic and heterogeneous conformational ensembles.
• Global properties of IDPs are well-studied, but fine-grained, sequence-specific variations crucial for function remain challenging to resolve.
• Understanding sequence-structure relationships in IDPs is key to deciphering their biological roles.

Purpose of the Study:

• To systematically investigate the relationship between sequence composition and conformational ensembles of IDPs.
• To develop novel descriptors that capture sequence-specific structural variations in IDPs.
• To provide a valuable resource for machine learning and coarse-grained force field development for IDPs.

Main Methods:

• Performed long-time scale atomistic simulations for 47 representative IDP sequences from the yeast proteome.
• Applied Uniform Manifold Approximation and Projection (UMAP) for nonlinear dimensionality reduction of high-dimensional structural data.
• Introduced and analyzed a novel metric, local compactness asymmetry (${\gamma }_{R_g}$), to quantify directional chain organization.

Main Results:

• UMAP effectively differentiated IDP conformational ensembles, revealing distinct structural patterns.
• Local compactness asymmetry (${\gamma }_{R_g}$) was identified as a descriptor orthogonal to global measures like radius of gyration.
• This metric correlates with sequence-level charge and hydropathy asymmetries, indicating preferential dynamics in more extended chain regions.

Conclusions:

• Local compactness asymmetry (${\gamma }_{R_g}$) provides a novel way to characterize IDP conformational ensembles based on sequence features.
• The findings highlight the importance of sequence-specific structural variations in governing IDP dynamics and function.
• The generated simulation data serves as a crucial resource for advancing computational methods for studying disordered proteins.