Laura Gherman

Talk: WormLink: A tool for engineering synthetic multivalent binding proteins

Condensates formed via liquid-liquid phase separation (LLPS) are a powerful mechanism for spatially organizing biochemical reactions. Pyrenoids are LLPS-based organelles in algae and hornworts that fix almost 30% of global CO2 and that assemble via interactions between Rubisco and multivalent “Linker” proteins. A potential route to improved crop yield and carbon fixation is the introduction of pyrenoid-like function into crop plants.

We now know many of the rules that underpin Linker protein architecture, and we can selectively modulate their Rubisco-binding and phase-separation properties through targeted engineering of their primary structure. However, modification of naturally occurring Linker proteins has limited potential, as algal Linker proteins do not effectively phase-separate Rubiscos from crop plants. Therefore, we are designing entirely synthetic Linker proteins from first principles. 

Linker proteins have the classic sticker-spacer architecture with multiple Rubisco-Binding Motifs (RBMs) connected by intrinsically disordered regions (IDRs). While tools exist for developing synthetic RBMs (e.g. phage display, computational design, etc), a major challenge lies in the design of IDRs that optimally link binding domains to generate a multivalent protein with the desired binding properties. Common approaches (e.g. glycine-serine repeats, transplanting an IDR from naturally occurring protein, etc) have severe practical limitations.

We have developed WormLink, a new computational tool to support the design of IDRs in multidomain binding proteins. The tool takes as input a 3D structure with binding motifs docked at their binding sites and calculates the optimal sequence lengths for IDRs connecting pairs of motifs. WormLink couples with tools that generate IDRs with appropriate ensemble properties to generate sequences for linking binding motifs. We will show how WormLink can support the design of multivalent binders using examples from different Rubisco-Linker complexes.