Discovery of multiple hidden allosteric sites by combining Markov state models and experiments

Area of Science:

• Biochemistry
• Computational Biology
• Pharmacology

Background:

• Allosteric drug design targets protein pockets not typically seen in static structures.
• Hidden allosteric sites offer therapeutic potential but are challenging to identify and target.
• Current methods often rely on serendipitous drug discovery rather than rational design.

Purpose of the Study:

• To develop a method for identifying hidden allosteric sites independent of drug discovery.
• To demonstrate the utility of Markov state modeling in uncovering novel allosteric pockets.
• To validate the identified sites through experimental methods.

Main Methods:

• Utilized Markov state modeling to analyze microsecond- to millisecond-timescale protein structural fluctuations.
• Developed a visualization technique to identify potential hidden allosteric sites from dynamic structures.
• Employed thiol labeling experiments to experimentally validate predicted hidden allosteric sites.

Main Results:

• Successfully identified multiple hidden allosteric sites in TEM-1 β-lactamase, a key antibiotic target.
• Demonstrated that Markov state models can predict the location of previously unknown allosteric pockets.
• Validated the computational predictions through experimental thiol labeling.

Conclusions:

• The study presents a novel computational and experimental approach to discover hidden allosteric sites.
• This methodology facilitates the rational design of drugs targeting previously inaccessible allosteric pockets.
• The findings suggest a vast, untapped potential for hidden allosteric sites in pharmaceutical development.