Underestimated Noncovalent Interactions in Protein Data Bank

Area of Science:

• Biochemistry
• Structural Biology
• Computational Chemistry

Background:

• Noncovalent interactions (NCIs) are fundamental to biomacromolecular structure and function.
• Hydrogen bonds (HBs), cation-π, π-π, and ionic interactions are key types of NCIs.
• Existing structural data may underestimate the prevalence of these interactions.

Purpose of the Study:

• To investigate the underestimation of noncovalent interactions in protein structures.
• To demonstrate the recoverability of overlooked hydrogen bonds.
• To assess the impact of structural resolution on NCI detection.

Main Methods:

• Analysis of hydrogen bond ratios (observed vs. pseudo HBs) in PDB structures.
• Quantum Mechanics/Molecular Mechanics (QM/MM) optimization of protein-ligand complexes.
• Systematic search of the Protein Data Bank (PDB) for NCI analysis.

Main Results:

• A significant number of hydrogen bonds are overlooked in protein-ligand and protein-protein interfaces within the PDB.
• QM/MM optimization successfully recovered these previously overlooked hydrogen bonds.
• The number of hydrogen bonds per residue (N_HB/R) decreases with lower structural resolution, indicating ongoing underestimation.
• Cation-π, π-π, and ionic interactions are also significantly underestimated.

Conclusions:

• There is a universal underestimation of various noncovalent interactions in structural databases.
• Recovering these NCIs is crucial for advancing drug design, protein-protein interaction studies, and understanding protein structure-function relationships.