Full Protein Sequence Redesign with an MMGBSA Energy Function.

Area of Science:

• Computational biology
• Protein engineering
• Biophysics

Background:

• Accurate energy functions are crucial for computational protein design.
• Existing models face challenges in predicting novel protein properties.
• The MMGBSA approach combines molecular mechanics with a generalized Born and surface area solvent model.

Purpose of the Study:

• To evaluate the performance of the MMGBSA energy function for computational protein design.
• To assess the contributions of individual energy terms within the MMGBSA model.
• To optimize the MMGBSA protocol for designing realistic protein sequences and conformations.

Main Methods:

• Implementation of the MMGBSA energy function within the Proteus software.
• Redesigning protein sequences one position at a time and simultaneously designing all positions.
• Evaluating individual energy terms (GB and SA) and their parametrizations.
• Comparing MMGBSA performance against simple Coulomb electrostatics.

Main Results:

• The generalized Born (GB) term significantly improves design accuracy over simple electrostatics.
• Pairwise decomposition errors affect the GB term when all positions are designed concurrently.
• The surface area (SA) term, with distinct coefficients for nonpolar and polar atoms, is vital for realistic sequences and can partially offset GB term limitations.
• The optimized GBSA protocol yields native-like protein cores and successful superfamily recognition.

Conclusions:

• The MMGBSA energy function, particularly with optimized SA terms, is effective for computational protein design.
• The study highlights the importance of accurate solvation and surface area modeling for protein design.
• The Proteus software and the developed GBSA protocol offer a robust tool for creating novel proteins with desired properties.