A synergistic strategy for E2E+ESM2-driven protein a design and wet lab validation

Area of Science:

• Biochemistry
• Computational Biology
• Protein Engineering

Background:

• Protein A is crucial for antibody purification in biopharmaceuticals and biomolecule research.
• Current computational protein design methods struggle with predictive accuracy and efficient screening.

Purpose of the Study:

• To develop an improved computational strategy for designing functional Protein A variants.
• To address limitations in predictive accuracy and multidimensional screening of generative protein models.

Main Methods:

• A synergistic strategy combining E2E generative models and ESM2 for protein design.
• Introduction of 'feature distance' for multidimensional screening of synthetic Protein A sequences.
• Tertiary structure prediction using AlphaFold and functional validation via affinity testing.

Main Results:

• Generated synthetic Protein A variants with high functional consistency.
• Identified synthetic protein V2 with excellent binding kinetics (K_D of 3.81±0.17E-10 M).
• Demonstrated strong binding performance through balanced association and dissociation rates.

Conclusions:

• The proposed E2E+ESM2 strategy with feature distance significantly improves synthetic protein design.
• This method enhances functional consistency and application potential for bioengineered proteins.
• Offers a promising computational solution for advancing protein design in drug development.