Aug 2026· Angewandte Chemie· pp.
e4052471
· 0 citations· 24 references
Medicine
TL;DR
This study illustrates how subtle residue-localized conformational bias can affect the overall protein conformational dynamics influencing protein-protein interactions that are important for cellular functions and related to diseases.
Abstract
Recently, the application of deep learning to structural data deposited in the Protein Data Bank has enabled the reliable and accurate prediction of 3D folded structures of proteins from their sequences. However, this approach is not applicable to highly dynamic proteins, where multiple structures interconvert. Furthermore, the mechanistic details of protein folding and unfolding remain challenging to study. Herein, we present a set of data highlighting these complexities. By chemical incorporation of stereoisomeric 4-fluoroproline residues at selected sites in the sequence of a folded, multi-conformational protein ("molten-globule"), we were able to modify its structural properties that propagated to highly distinct functional features, such as modulation of ligand binding affinities or misfolding and aggregation into amyloids. Application of NMR methods, notably 19F NMR spectroscopy, provided detailed molecular insights into the observed phenomena. This study illustrates how subtle residue-localized conformational bias can affect the overall protein conformational dynamics influencing protein-protein interactions that are important for cellular functions and related to diseases.
This work investigates how protein folding landscapes are altered inside condensates, using the protein α-helix as a model folded domain and develops a chemically specific, residue-resolution model for quantification of α-helical folding and applies it to characterize diverse helices within condensates of varying physicochemical properties.
Nathaniel Hess, Jerelle A. Joseph· Journal of the American Chem...· 0 citations
The advent of machine learning structural prediction tools has largely accelerated bottom-up structural biology and provides static Anfinsenian models with near-experimental accuracy, complementing X-ray crystallography and cryogenic electron microscopy. However, these techniques often overlook two fundamental aspects of biomolecular structure-function relationships: (i) conformational equilibria and (ii) transient noncovalent interactions. Indeed, these considerations are particularly important for integral membrane proteins (MPs), frequently serving mechanistic roles, such as conformational plasticity in ion channels facilitating heterogeneous state transitions, or dynamic binding partners modulating signal transduction in receptors and small-molecule transporters. In this vein, electron paramagnetic resonance (EPR) spectroscopy, an ensemble method, provides a powerful toolbox to guide high-resolution structural data by directly reporting on conformational subensembles and noncovalent binding events. Herein, we highlight recent EPR applications to integral MPs and emerging synergies with bottom-up atomistic approaches to enhance insights into structure-dynamics-function relationships.
Joshua L. Wort, Xinyu Liu, James Pullen et al.· Current Opinion in Structura...· 0 citations
It is proposed that the consistency principle is formulated by the quadratic relationship in the double logarithm plot of the residue-specific equilibrium and rate constants of a polypeptide chain.
The data reveal that both the mechanics and kinetics of peptide detachment are sensitive to the identity and sequence position of individual residues, highlighting the power of integrating CG MD and single-molecule force spectroscopy to unravel residue-specific, sequence-dependent factors underlying peptide–lipid interactions.
Ryan S. Smith, Krishna P Sigdel, D. R. Weaver et al.· Langmuir· 0 citations
Together, these studies establish how GAGs are organized in the CXCL8-bound complex and highlight the value of complementary low-resolution structural methods for characterizing GAG-protein complexes.
M. A. White, B. Mahler, P. R. B. Joseph et al.· Biochemical Journal· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduAug 27, 2026
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.
