Structure Determination of Challenging Protein-Peptide Complexes Combining NMR Chemical Shift Data and Molecular...

Area of Science:

• Biochemistry and Structural Biology
• Computational Biology
• Molecular Dynamics

Background:

• Intrinsically disordered proteins (IDPs) play crucial roles in cellular signaling, often mediating protein-protein interactions.
• Predicting the structures of complexes involving IDPs is challenging due to their inherent flexibility and disorder-to-order transitions upon binding.
• Existing modeling tools struggle with the dynamic and plastic nature of these interactions.

Purpose of the Study:

• To develop and validate a computational approach for determining the three-dimensional structures of peptide-protein complexes, particularly those involving intrinsically disordered regions.
• To assess the performance of the MELD (Modeling Employing Limited Data) technique combined with NMR chemical shift data for modeling these challenging complexes.
• To compare the capabilities of the MELD+NMR approach with AlphaFold for predicting complex structures and estimating binding affinities.

Main Methods:

• Utilized the MELD (Modeling Employing Limited Data) technique integrated with Nuclear Magnetic Resonance (NMR) chemical shift data.
• Applied the hybrid approach to model complexes formed between the extraterminal (ET) domain of bromo and extraterminal domain (BET) proteins and various polypeptide ligands.
• Conducted blind studies to evaluate the accuracy of predicted structures against experimentally determined conformations.

Main Results:

• The MELD+NMR approach successfully modeled bound-state conformations and binding poses for moderately tight binders (K_d ~100 nM) using only protein backbone chemical shift data.
• Accurate structural predictions were achieved in excellent agreement with experimental data.
• For weaker binders (K_d ~250 μM), additional peptide ligand chemical shift data improved the accuracy of the MELD+NMR method.
• While AlphaFold could qualitatively rank peptides, MELD directly estimated relative binding affinities.

Conclusions:

• The hybrid MELD+NMR approach provides a powerful new tool for the structural analysis of protein-polypeptide complexes, especially those involving disorder-to-order transitions.
• This method accurately predicts both the 3D structures and relative binding affinities of such complexes.
• It overcomes limitations of other prediction methods for intrinsically disordered protein interactions.