Perturbation of ACE2 Structural Ensembles by SARS-CoV-2 Spike Protein Binding

Area of Science:

• Biochemistry
• Structural Biology
• Computational Biology

Background:

• The human Angiotensin-Converting Enzyme 2 (ACE2) is a key receptor for coronaviruses, including SARS-CoV-2.
• ACE2's interaction with the SARS-CoV-2 spike protein's S1 subunit is crucial for viral entry.
• X-ray crystallography has characterized this interaction but hasn't revealed significant ACE2 structural changes upon binding.

Purpose of the Study:

• To investigate persistent structural differences in ACE2 upon SARS-CoV-2 S1 protein binding.
• To utilize machine learning to identify and validate these structural changes.
• To predict the compatibility of ACE2-binding compounds with viral S1 protein binding and identify allosteric effects.

Main Methods:

• All-atom molecular dynamics simulations of ACE2.
• Linear Discriminant Analysis (LDA) machine learning for structural classification.
• Validation using independent datasets and long simulation trajectories (Anton 2 supercomputer).
• Projection of 78 ACE2-ligand complex trajectories onto the LDA classification vector.

Main Results:

• Persistent differences in ACE2 structure upon S1 protein binding were identified.
• LDA successfully classified ACE2 structures as 'apo-like' or 'complex-like'.
• Ligand-bound ACE2 structures were analyzed for compatibility with S1 binding, revealing potential allosteric effects.

Conclusions:

• ACE2 undergoes significant structural changes upon SARS-CoV-2 S1 binding, not apparent in static crystallography.
• Machine learning effectively identifies these dynamic structural alterations.
• This approach can predict drug efficacy by assessing their influence on ACE2 structure and viral binding compatibility.