Structural Insights of Inter-domain Interactions in Multi-domain Proteins

Abstract

Embargo up to August 11, 2026. Proteins are essential biomolecules whose diverse functions stem from their modular architecture built from independently folding units called domains. This thesis investigates the nature and significance of domain-domain interactions in multi-domain proteins, focusing primarily on intra-protein interactions. Chapter 1 briefly explains the fundamentals used in this thesis. In Chapter 2, the study classifies domain-domain interactions within monomeric two domain proteins as either permanent or transient—a concept traditionally applied to interprotein interactions. It reveals that around 40% of these proteins exhibit permanent domain interactions, while 60% are transient. Permanent interactions are characterised by stronger bonding due to larger interfacial areas and a predominance of hydrophobic contacts, offering structural rigidity. In contrast, transient interactions rely more on hydrogen bonding, providing specificity and flexibility. Dynamic correlation analysis shows stronger coordinated motions in permanent interactions, contributing to greater structural coherence, whereas transient interactions allow functional flexibility. Interestingly, both interaction types have comparable interfacial residue conservation, suggesting evolutionary independence. Furthermore, permanent domains preferentially interact with structurally similar partners and are biased toward specific protein folds, while transient domains show greater structural diversity. Chapter 3 shifts focus to remote homologous domains, domains with conserved structural frameworks despite low sequence similarity. Using residue-level network comparisons derived from crystal structures, the study explores three types of domain comparisons: within a single protein architecture, across identical architectures, and across different ones. The greatest structural variability is observed in remote homologs from different domain architectures, where changes in local residue clusters and side-chain orientations lead to significant global structural differences. These variations highlight how network organisation adapts based on domain pairing, affecting structural integrity and interaction modes. Chapter 4 explores structural conformations of the SARS-CoV2 spike protein using computational methods to automate the classification of its open and closed forms. Traditional approaches like RMSD and radius of gyration had limitations due to structural heterogeneity, but Principal Component Analysis (PCA) proved more effective in distinguishing conformers, revealing previously unnoticed structural patterns across viral variants. This method offers a refined tool for identifying conformational states in dynamic, multi-domain proteins. Chapter 5 consolidates these insights, underlining the broader applications of domain-domain interaction analysis in protein folding, structural prediction, and drug design. By redefining homology through structural network comparison and highlighting domain-specific interaction dynamics, the thesis contributes to an improved understanding of protein architecture and paves the way for advancements in computational biology, protein engineering, and functional annotation.