An Effective Pipeline to Develop Modification-Specific Intracellular Antibodies Using AI Tools

Posttranslational protein modifications (PTMs) are often associated with protein functions. To detect PTMs in living cells, genetically encoded antibody-based probes, termed modification-specific intracellular antibodies (mintbodies), have been developed. A mintbody consists of a single-chain variable fragment (scFv) of a modification-specific antibody and a fluorescent protein. By expressing mintbodies in cells or organisms, the localization and changes in the specifically modified form of a protein can be tracked in real time. In particular, mintbodies against histone and RNA polymerase II PTMs have been developed and are available. However, most mouse monoclonal antibodies fail to remain functional when expressed in the reduced environment inside eukaryotic cells due to unstable folding. This results in an inability to bind antigen, instability, aggregation, and/or degradation. Whereas there have been no general methods to develop functional intracellular antibodies, we have recently developed a convenient pipeline to efficiently convert unstable scFvs into stable and functional probes using AI-assisted tools. We here describe a practical, end-to-end protocol for converting existing antibody sequences into functional intracellular antibodies using an AI-assisted design pipeline combined with live-cell screening. The method integrates antibody sequence annotation, structure prediction, deep-learning-based framework optimization, and experimental validation in mammalian cells.