AI-Guided Stability Tuning of a Heterodimeric Linker for Programmable Protein Tube Architectures

Abstract

Advances in artificial protein assembly design have enabled the construction of increasingly complex higher-order protein architectures. However, rationally tuning the morphology of such assemblies beyond their initially formed architectures remains challenging. Here, we show that the thermal responsiveness of a heterodimeric coiled-coil linker provides a molecular handle for tuning the formation and morphology of two-component protein tubes. Guided by computational stability predictions, we generated a panel of single-amino-acid variants in the M3L2/p66 linker and identified assembly-competent variants with distinct, experimentally accessible thermal transition behaviors. Incorporating two such variants into the tube-forming scaffold, together with the wild-type linker, enabled a focused comparison across three linker states, revealing stepwise shifts in tube-forming temperature windows and tube diameters. Notably, the variant with the lowest apparent thermal stability in this comparison uniquely accessed multi-walled tube architectures. Time-course electron microscopy of this variant revealed a sequential pathway from initially formed thin tubes to wall-thickened intermediates and ultimately to multi-walled states. Together, these findings establish linker thermal responsiveness as a design handle for tuning protein nanotube morphology and expanding the accessible structural states of designed protein tubes.