Structure Prediction of Alternate Frame Folding Systems with AlphaFold3.

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• AlphaFold2 accurately predicts static protein structures.
• Protein conformational flexibility is crucial for function.
• Tuning sequence alignments can reveal alternative protein states with AlphaFold2.

Purpose of the Study:

• Evaluate AlphaFold3's accuracy in modeling alternative protein conformations.
• Investigate if AlphaFold3 can natively model conformational diversity.
• Determine the influence of disordered regions on AlphaFold3 predictions.

Main Methods:

• Utilized a family of engineered green fluorescent proteins.
• Employed alternate frame folding techniques.
• Assessed AlphaFold3's prediction accuracy for conformational states.

Main Results:

• AlphaFold3 demonstrates potential for modeling alternative protein conformations.
• Disordered protein regions unexpectedly influence conformational preferences in AlphaFold3 models.
• Tuning multiple sequence alignments may be necessary for capturing full conformational landscapes.

Conclusions:

• AlphaFold3 advances protein structure prediction, including dynamic aspects.
• Disordered regions are critical determinants of protein conformational states.
• Further research is needed to fully leverage AlphaFold3 for conformational dynamics.