Designing bacterial signaling interactions with coevolutionary landscapes.

Area of Science:

• Biochemistry
• Computational Biology
• Protein Engineering

Background:

• Designing novel protein-protein interactions for specific functions, like catalysis, is complex.
• Current methods often involve time-consuming and expensive brute-force approaches.
• Predicting amino acid substitutions that alter protein interactions is a significant challenge.

Purpose of the Study:

• To develop a computational coevolutionary landscape for predicting amino acid substitutions that create or modify protein-protein interactions.
• To test the predictive power of this landscape for interspecies signaling partners.
• To demonstrate a more efficient approach to protein design compared to traditional methods.

Main Methods:

• Development of a computational coevolutionary landscape based on sequence analysis.
• Prediction of single amino acid substitutions to modulate phosphotransfer between Escherichia coli EnvZ and Bacillus subtilis Spo0F.
• Experimental validation of predicted mutations for their effect on kinase phosphotransfer activity.

Main Results:

• The coevolutionary landscape accurately predicted single amino acid substitutions that modulate phosphotransfer between non-cognate partners.
• The approach successfully identified mutations enhancing or suppressing interactions.
• Experimental results validated the theoretical predictions, demonstrating the landscape's efficacy.

Conclusions:

• A computational coevolutionary landscape, using limited structural data, significantly reduces the search space for predicting functional amino acid substitutions.
• This method offers a more efficient and predictive approach to protein engineering and design.
• The findings have implications for designing novel protein functions and understanding signaling pathways.