Computational Prediction and Design for Creating Iteratively Larger Heterospecific Coiled Coil Sets

Area of Science:

• Biochemistry
• Computational Biology
• Protein Engineering

Background:

• Mimicking nature's specific protein-protein interactions (PPIs) is a key biochemical engineering goal.
• Previous methods identified limited sets of specific interacting peptides.
• Developing faster, more versatile tools for PPI engineering is needed.

Purpose of the Study:

• To develop new software for rapidly designing large sets of specific heterospecific protein-protein interactions.
• To improve the prediction accuracy and utility of computational approaches for protein interaction engineering.

Main Methods:

• Developed new software (qCIPA) to predict dimer stability (Tm) based on core, electrostatic, and helical propensity components.
• Used computational interactome screening to derive large heterospecific peptide sets targeting antiparallel off-targets.
• Experimentally validated predictions using circular dichroism and size exclusion chromatography.

Main Results:

• The qCIPA software accurately predicts dimer specificity by weighing core and electrostatic components.
• An expanded dataset revealed sequence context rules, enhancing designed coiled coil (CC) stability while reducing off-target interactions.
• Achieved increased prediction accuracy and speed for designing specific PPIs.

Conclusions:

• The new computational approach significantly expands the toolkit for engineering specific protein-protein interactions.
• The software enables the rapid design of large heterospecific peptide sets with high specificity.
• This method has broad applications in synthetic biology and for imposing specificity on diverse PPIs.