Modelling the polypeptide backbone with 'spare parts' from known protein structures.

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Accurate protein structure prediction is crucial for understanding biological function.
• De novo protein backbone construction methods are essential for modeling and design.

Purpose of the Study:

• To develop an automated procedure for constructing protein polyalanine backbones.
• To assess the accuracy and robustness of the proposed method.
• To gain insights into sequence-3D structure relationships.

Main Methods:

• Utilizing C alpha positions and 'spare parts' from a database of 66 high-resolution protein structures.
• Constructing protein backbones from overlapping fragments of variable length.
• Analyzing the accuracy using root-mean-square deviation (r.m.s.d.) against crystal structures.

Main Results:

• Generated backbones show favorable comparison with refined X-ray structures (r.m.s.d. < 1Å).
• The method is insensitive to experimental errors in C alpha positions and database size.
• Identified beta alpha loops as structurally less common than alpha beta loops in a beta-barrel protein example.
• The 'spare parts' approach is useful for modeling local structural changes, yielding realistic backbones in ~2/3 of test cases.

Conclusions:

• The automated procedure effectively builds accurate protein polyalanine backbones.
• The method provides valuable insights into sequence-3D structure correlations.
• Incorporating sequence information, especially for residues like glycine, improves modeling accuracy for challenging cases.