Exploring the Sequence-based Prediction of Folding Initiation Sites in Proteins

Area of Science:

• Biochemistry
• Structural Biology
• Computational Biology

Background:

• Protein folding is essential for cellular function; misfolding is linked to various diseases.
• Early folding events, driven by local amino acid interactions, are poorly understood.
• Understanding initial folding stages is key to deciphering protein structure-function relationships.

Purpose of the Study:

• To develop a predictive method, EFoldMine, for identifying amino acids involved in early protein folding events.
• To leverage hydrogen deuterium exchange (HDX) data from NMR experiments for predicting early folding residues.
• To analyze the relationship between early folding residues and the final protein structure.

Main Methods:

• Utilized hydrogen deuterium exchange (HDX) data from NMR pulsed labeling experiments.
• Employed features including backbone and sidechain dynamics, and secondary structure propensities.
• Developed EFoldMine, a computational method for predicting early folding residues from primary amino acid sequences.

Main Results:

• EFoldMine successfully predicts amino acids involved in early folding stages.
• Predictions align with independent experimental observations, providing insights into folding pathways.
• On a proteome scale, predicted residues correlate with highly interacting residues in folded proteins and evolutionary covariation.

Conclusions:

• EFoldMine offers a novel approach to study protein folding dynamics.
• Early local interactions significantly influence the final protein structure and function.
• The findings connect early folding behavior to evolutionary patterns and disease-related misfolding.