The emergence of novel versus known three-dimensional structures from random sequences

Area of Science:

• Computational biology
• Structural biology
• Molecular evolution

Background:

• Random protein sequences are generally unlikely to fold into stable globular structures.
• Repeated sequences may offer a pathway to stable protein folds, relevant to molecular evolution.

Purpose of the Study:

• To investigate the foldability of protein sequences composed of random repeats using structure prediction.
• To explore the potential for repeated sequences to form stable protein structures.

Main Methods:

• Utilized computational structure prediction methods.
• Analyzed approximately 120-residue sequences with 5- to 60-residue random repeats.
• Experimentally validated a novel alpha-helical screw structure using circular dichroism (CD) spectroscopy and X-ray crystallography.

Main Results:

• Sequences with repeats <30 residues frequently folded with high confidence (1-12%).
• Observed beta-solenoids (similar to natural proteins) and helical bundles.
• Discovered and experimentally confirmed a novel alpha-helical screw supersecondary structure.
• Foldability decreased beyond 40-residue repeat lengths.

Conclusions:

• Structure prediction tools can identify folds not well-represented in training data.
• Repeated sequences are a viable source of stably folded protein structures.
• Findings support the hypothesis that proteins evolved through the assembly of shorter peptide sequences.