Ab initio protein folding simulations using atomic burials as informational intermediates between sequence and...

Area of Science:

• Computational biology
• Structural biology
• Biophysics

Background:

• Protein three-dimensional structure is dictated by amino acid sequences, but the folding code remains largely undeciphered.
• Atomic burials, representing distances to the molecular center, are proposed as key information for protein structure determination from sequences.

Purpose of the Study:

• To provide direct evidence for the role of atomic burials in the protein folding code.
• To demonstrate that burial propensities, derived from local sequences, can be used for *ab initio* protein folding simulations.

Main Methods:

• Utilized a Hidden Markov Model (HMM) to classify protein heavy atoms into burial types based on sequence context.
• Employed molecular dynamics simulations with a potential function enforcing predicted burial levels and geometric constraints for covalent structure and hydrogen bonds.

Main Results:

• Successfully folded globular proteins from extended chains in *ab initio* simulations using predicted burial propensities.
• Achieved correct folded conformations for diverse protein structures across all three folding classes.
• Demonstrated high burial prediction quality, validating the approach.

Conclusions:

• Atomic burials serve as crucial informational intermediates linking protein sequence to structure.
• This finding offers a novel conceptual framework for enhancing protein structure prediction.
• Provides new insights into the fundamental principles governing protein folding.