Neutral networks in protein space: a computational study based on knowledge-based potentials of mean force

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Many unrelated protein sequences can adopt similar three-dimensional folds.
• Sequences with the same fold form subsets within the broader sequence space.
• Understanding the topology of these subsets is crucial for protein evolution and de novo design.

Purpose of the Study:

• To investigate the topology and connectivity of protein sequence subsets that share a common fold.
• To explore the implications of sequence space distribution for protein evolution and design.

Main Methods:

• Utilized inverse folding techniques to analyze protein sequences and their corresponding structures.
• Examined the distribution and connectivity of sequences within sequence space for specific protein folds.
• Assessed the impact of restricted amino acid alphabets on sequence-structure relationships.

Main Results:

• Sequences adopting a specific fold are homogeneously distributed in sequence space, not clustered.
• No significant sequence homology was detected among sequences sharing the same fold.
• Despite lack of homology, sequences for a given fold are connected by neutral paths, allowing fold maintenance.
• Designability varies among native structures, with some folds being more amenable to sequence variation.

Conclusions:

• Protein sequence space exhibits extensive neutral networks, facilitating evolutionary exploration.
• The homogeneous distribution and connectivity of sequences enable de novo protein design with varied chemical properties.
• Findings suggest similarities between protein and nucleic acid sequence space features.