Waterless structures in the Protein Data Bank.

Area of Science:

• Structural Biology
• Biochemistry
• Data Science

Background:

• Water molecules are crucial for protein structure and function, expected to bind to proteins crystallized from aqueous solutions.
• The Protein Data Bank (PDB) is a primary repository for protein crystal structures, but its contents may not fully represent biological reality regarding solvent presence.

Purpose of the Study:

• To investigate the discrepancy between expected water molecule presence and their absence in deposited protein crystal structures.
• To identify potential issues in data collection, refinement, or deposition processes that lead to missing or inaccurate solvent modeling.

Main Methods:

• Analysis of protein crystal structure models in the Protein Data Bank (PDB).
• Examination of water molecule to amino-acid residue ratios across different resolutions.
• Detailed review of structures with missing or unusually modeled solvent molecules.

Main Results:

• A significant number of high-resolution protein crystal structures (800+ at ≥2.5 Å) in the PDB are deposited without any water molecules.
• Observed variations in water occupancy suggest potential problems with diffraction data, refinement, or deposition.
• The ratio of water molecules to residues varies significantly with resolution, from >1.5 at atomic resolution to 0.25 at 2.5 Å.

Conclusions:

• The absence or inaccurate modeling of water molecules in PDB structures can indicate underlying data or processing issues.
• The PDB should implement systems to flag and address problems related to solvent structure in deposited models.
• Depositors need to ensure accurate representation of solvent in protein crystal structures for improved biological interpretation.