Molecular docking for substrate identification: the short-chain dehydrogenases/reductases

Area of Science:

• Biochemistry
• Computational Biology
• Enzymology

Background:

• Protein ligand docking is a computational method used for predicting protein function and identifying substrates.
• Identifying enzyme substrates, especially within diverse protein families like short-chain dehydrogenases/reductases (SDRs), presents a significant computational challenge.
• Cross-docking large numbers of enzymes with numerous potential ligands requires efficient and accurate methodologies.

Purpose of the Study:

• To evaluate the efficacy of protein ligand docking protocols for identifying substrates of short-chain dehydrogenases/reductases (SDRs).
• To assess the ability of docking to accurately predict substrate binding modes and rankings within a large metabolite dataset.
• To explore a reduced computational cost protocol using representative structures (medoids) for characterizing SDR binding site ligand preferences.

Main Methods:

• Docking of over 900 human metabolites against 27 SDR proteins with known functions.
• Comparison of two distinct docking methods and two scoring functions.
• Evaluation of a novel protocol using representative structures (medoids) for binding site characterization.
• Analysis of docking success in reproducing known substrate binding and ranking.

Main Results:

• Docking successfully reproduced viable binding modes for known substrates of SDRs.
• The ability of docking to highly rank known substrates among other metabolites varied depending on the protocol used.
• The medoid-based protocol showed correlation between representative rank and cluster mean rank, but simple clustering was insufficient for accurate substrate identification.
• Some clusters contained ligands with diverse affinities, indicating potential for missing key substrates when using single representatives.

Conclusions:

• Protein ligand docking is a valuable tool for predicting enzyme-substrate relationships within specific protein families like SDRs.
• Careful selection of docking protocols and scoring functions is crucial for accurate substrate identification.
• A simplified approach using representative structures may not be sufficient for comprehensive substrate discovery due to ligand binding heterogeneity.