The giant muscle protein titin known for its mechanical stability serves as an ideal model for developing protein-based biomaterials.However,practical applications are hindered by challenges such as low expression yie...The giant muscle protein titin known for its mechanical stability serves as an ideal model for developing protein-based biomaterials.However,practical applications are hindered by challenges such as low expression yields,poor solubility,and limited thermal stability.In this study,we combined artificial intelligence(AI)tools—ProteinMPNN and Alpha-Fold2—with human insights and steered molecular dynamics(SMD)simulations to redesign the titin Ig domain.This approach generated thousands of novel sequences,preserving structural features essential for mechanical stability.Six de novo proteins were experimentally validated,all demonstrating mechanical and kinetic stability comparable to the natural I27 domain at the 100 pN level.Notably,these proteins exhibited significantly improved physical properties,with up to a fivefold increase in expression yield and enhanced solubility.Circular dichroism and atomic force microscopy-based single molecule force spectroscopy(AFM-SMFS)confirmed proper folding and mechanical stability,while thermal stability was retained or improved in some designs.Control experiments with unguided random designs,which yielded negative results,underscored the critical role of integrating AI tools with domain-specific knowledge for functional outcomes.Our findings highlight the transformative potential of integrating AI-driven design with human expertise and computational simulations,enabling the development of mechanically robust proteins for applications in bioengineering and biomaterials science.展开更多
基金the funding support from the National Natural Science Foundation of China(grant nos.22222703 and 22477058)the Natural Science Foundation of Jiangsu Province,China(grant no.BK20202004)the Fundamental Research Funds for the Central Universities,China(grant no.020514380335).
文摘The giant muscle protein titin known for its mechanical stability serves as an ideal model for developing protein-based biomaterials.However,practical applications are hindered by challenges such as low expression yields,poor solubility,and limited thermal stability.In this study,we combined artificial intelligence(AI)tools—ProteinMPNN and Alpha-Fold2—with human insights and steered molecular dynamics(SMD)simulations to redesign the titin Ig domain.This approach generated thousands of novel sequences,preserving structural features essential for mechanical stability.Six de novo proteins were experimentally validated,all demonstrating mechanical and kinetic stability comparable to the natural I27 domain at the 100 pN level.Notably,these proteins exhibited significantly improved physical properties,with up to a fivefold increase in expression yield and enhanced solubility.Circular dichroism and atomic force microscopy-based single molecule force spectroscopy(AFM-SMFS)confirmed proper folding and mechanical stability,while thermal stability was retained or improved in some designs.Control experiments with unguided random designs,which yielded negative results,underscored the critical role of integrating AI tools with domain-specific knowledge for functional outcomes.Our findings highlight the transformative potential of integrating AI-driven design with human expertise and computational simulations,enabling the development of mechanically robust proteins for applications in bioengineering and biomaterials science.
