Designing proteins: Mimicking natural protein sequence heterogeneity

Area of Science:

• Computational biology
• Biochemistry
• Protein engineering

Background:

• Natural proteins display a limited use of the 20-letter amino acid alphabet, resulting in lower than theoretical permutation compositions.
• Understanding and replicating this natural sequence heterogeneity is crucial for advanced protein design.

Purpose of the Study:

• To develop an enhanced protein design algorithm that emulates natural protein sequence heterogeneity.
• To ensure designed proteins fold accurately into their native structures.
• To improve the prediction of protein contacts and interactions.

Main Methods:

• Developed a protein design algorithm allowing random reshuffling of amino acid composition, avoiding constraints.
• Utilized folding free energy computations to assess the refolding precision of designed sequences.
• Employed direct coupling analysis to evaluate the accuracy of predicted protein contacts.

Main Results:

• The algorithm generates sequences with equilibrium permutation compositions similar to natural proteins.
• Designed sequences demonstrated high-precision refolding to native structures, with exceptions for proteins containing large disordered regions.
• Direct coupling analysis revealed high accuracy (>82%) in predicting protein contacts, surpassing previous methods.

Conclusions:

• The enhanced algorithm successfully mimics natural protein heterogeneity and ensures correct folding.
• This advancement offers a more accurate representation of protein interactions, overcoming biases in prior designs.
• The study marks a significant step forward in protein design and engineering capabilities.