De novo protein design. II. Plasticity in sequence space

Area of Science:

• Protein engineering and bioinformatics.
• Computational biology and structural biology.

Background:

• Multiple protein sequences can fold into similar three-dimensional structures.
• Protein design should focus on accessible sequence space for a given fold, not just single sequences.

Purpose of the Study:

• To develop and evaluate a novel computational approach for protein sequence design.
• To assess the ability of designed sequences to maintain a specific protein fold.
• To explore the application of designed sequence alignments in protein fold recognition.

Main Methods:

• Optimizing complete protein sequences using backbone structure, amino acid composition, and a physical energy function.
• Imposing sequence specificity by maintaining fixed amino acid composition.
• Generating multiple sequence alignments from designed sequences sharing a common fold.

Main Results:

• The protein design procedure converges within sequence space but does not yield the native sequence.
• Polar residues are generally conserved, while surface non-polar residues may be replaced by polar ones.
• A derived profile matrix for chicken triose phosphate isomerase (TIM) effectively recognizes native and related sequences.

Conclusions:

• The developed protein design method generates diverse sequences that maintain a target fold.
• Designed sequence alignments can be used to create profile matrices for protein fold recognition.
• This approach shows potential for improving protein fold recognition and understanding sequence-structure relationships.