De novo design of protein binders targeting the human sweet taste receptor as potential sweet proteins

There is growing interest in alternative sweeteners that reduce dietary sugar intake while maintaining desirable sensory properties. Natural sweet proteins offer high sweetness potency but are limited by natural diversity and challenges in production, stability, and formulation. In this study, we developed an integrated computational framework for de novo design and prioritization of protein binders targeting the extracellular domain (ECD) of the human sweet taste receptor subunit TAS1R2. The workflow combines RFdiffusion-based backbone generation, ProteinMPNN sequence design, Boltz-1 structural evaluation, static GBSA energetic assessment, and molecular dynamics (MD) simulations. Designed binders targeted structurally relevant regions of the TAS1R2 Venus flytrap and cysteine-rich domains. Binder2 showed the strongest Boltz-1 structural confidence, whereas Binder1 exhibited the most favorable static GBSA estimate. Subsequent 200-ns MD simulations revealed distinct interaction profiles: Binder1 formed the largest and most interaction-rich interface, Binder2 maintained a persistent mixed polar/nonpolar network, and Binder3 showed greater conformational rearrangement and a more restricted interface. Advantame-bound TAS1R2 served as a structurally defined reference for trajectory-based analyses. These results demonstrate that receptor-guided de novo design can generate structurally diverse TAS1R2-binding proteins and support candidate prioritization for future experimental validation of receptor activation and sensory sweetness.