Predicting fragment binding poses using a combined MCSS MM-GBSA approach.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate prediction of small molecule binding poses is essential for fragment-based drug discovery, especially when high-resolution structures are unobtainable.
• Fragment screening campaigns identify low molecular weight compounds that require reliable pose prediction to guide subsequent chemical modifications.

Purpose of the Study:

• To evaluate the utility of implicit solvent methods for assessing fragment binding poses generated by the Multiple Copy Simultaneous Search (MCSS) method.
• To investigate the impact of utilizing multiple receptor structures to account for receptor flexibility.
• To compare the performance of MCSS and GOLD docking programs when combined with MM-GBSA rescoring.

Main Methods:

• Fragment binding poses were generated using the Multiple Copy Simultaneous Search (MCSS) method in CHARMm.
• Poses were rescored using a Molecular Mechanics Generalized Born approach with molecular volume integration and Surface Area model (MM-GBSA).
• The docking program GOLD was also employed, followed by MM-GBSA rescoring, and its performance was compared to MCSS.
• The effect of using multiple receptor structures, exemplified by Hsp90, was assessed.

Main Results:

• The original MCSS performance of 50% accurate predictions increased to 67% when poses were rescored with MM-GBSA.
• Similar performance improvements were observed using the GOLD docking program followed by MM-GBSA rescoring.
• Combining results from both MCSS and GOLD docking methods led to a higher success rate in identifying correct binding poses.
• Using multiple receptor structures for Hsp90 showed variable performance improvements.

Conclusions:

• Implicit solvent methods, specifically MM-GBSA, significantly enhance the accuracy of fragment binding pose prediction when applied to MCSS or GOLD docking results.
• Comparing different docking methodologies can improve the identification of correct binding poses.
• These enhanced docking predictions can confidently guide chemical optimization efforts, particularly for targets with relatively rigid structures or when multiple diverse crystal structures are available.