Connectable Components for Protein Design

Protein design requires reusable, trustworthy, and connectable parts in order to scale to complex challenges. The recent explosion of protein structures stored within the Protein Data Bank provides a wealth of small motifs we can harvest, but we still lack tools to combine them into larger proteins. Here I explore two approaches for connecting reusable protein components on two different length scales. On the atomic scale, I build an interactive search engine for connecting chemical fragments together. Protein fragments built using this search engine recapitulate native-like protein assemblies that can be integrated into existing protein scaffolds using backbone search engines such as MaDCaT. On the protein domain scale, I quantitatively dissect structural variations in two-component systems in order to extract general principles for engineering interfacial flexibility between modular four-helix bundles. These bundles exhibit large scissoring motions where helices move towards or away from the bundle axis and these motions propagate across domain boundaries. Together, these two approaches form the beginnings of a multiscale methodology for connecting reusable protein fragments where there is a constant interplay and feedback between design of atomic structure, secondary structure, and tertiary structure. Rapid iteration, visualization, and search glue these diverse length scales together into a cohesive whole.