Generative design of programmable asymmetric β-barrel nanopores

Area of Science:

• Biophysics
• Materials Science
• Artificial Intelligence

Background:

• Native homo-oligomeric transmembrane beta-barrels (TMBs) are useful but have limited spatial resolution due to uniform lumens.
• Previous monomeric TMB design methods required extensive expert input, hindering scalability and functionality.

Purpose of the Study:

• To develop a rapid, scalable generative AI method for designing functional monomeric TMBs.
• To create TMBs with tailored properties for specific applications like sensing, sequencing, and separation.

Main Methods:

• A diffusion-based generative AI approach was used for TMB backbone generation, conditioned on structural features.
• TMB-optimized sequence design was integrated with backbone generation.
• 48 TMB designs were characterized, and two were structurally determined via crystallography.

Main Results:

• The AI method successfully designed 48 TMBs with measurable conductances, corresponding to pore diameters of 0.7-1.5 nm.
• Crystal structures confirmed atomic-level agreement between designed models and actual TMBs.
• Designed nanopores demonstrated capabilities for selective ion sensing (copper binding), DNA translocation, and ion transport in hybrid membranes.

Conclusions:

• The generative AI method offers a rapid and scalable approach for designing custom TMBs.
• This AI-driven design enables versatile nanopore engineering for diverse applications in molecular sensing and separation.
• The designed TMBs show promise for advanced applications in biomolecular analysis and materials science.