Finding Reactive Configurations: A Machine Learning Approach for Estimating Energy Barriers Applied to Sirtuin 5

Area of Science:

• Biochemistry
• Computational Chemistry
• Enzymology

Background:

• Sirtuin 5 (SIRT5) is a class III histone deacetylase.
• SIRT5 primarily catalyzes desuccinylation and demanoylation, not deacetylation.
• Enzyme reaction pathway calculations are crucial for understanding biological mechanisms.

Purpose of the Study:

• To develop and test a novel machine learning approach for calculating enzyme reaction pathways.
• To identify suitable starting conformations for reaction path calculations of Sirtuin 5.
• To uncover structural features critical for the Sirtuin 5-catalyzed desuccinylation reaction.

Main Methods:

• Applied machine learning to molecular dynamics frames and QM/MM calculations.
• Utilized potential energy barriers to identify reactive configurations.
• Developed a novel approach to overcome limitations in traditional reaction path calculation schemes.

Main Results:

• Identified eleven key structural features governing Sirtuin 5 reactivity.
• Estimated reaction barriers with a mean absolute error of 3.6 kcal/mol.
• Successfully identified suitable start-conformations and reactive configurations for the desuccinylation reaction.

Conclusions:

• The novel machine learning approach enhances the accuracy of enzyme reaction pathway calculations.
• Understanding key structural features aids in elucidating enzyme mechanisms and active site interactions.
• This method provides a robust framework for studying large molecular systems and potential drug targets like SIRT5.