Protein-protein interaction networks translate environmental inputs into specific physiological outputs. The signaling proteins in these networks require regulatory mechanisms to ensure proper molecular function. Two common regulatory features of signaling proteins are autoinhibition and ultrasensitivity. Autoinhibition locks the protein in an inactive state through cis interactions with a regulatory module until it is activated by a specific input signal. Ultrasensitivity, defined as steep activation after a threshold, allows cells to convert graded inputs into more switch-like outputs and can lead to complex decision making behaviors such as bistability. Although these mechanisms are common features of signaling proteins, their molecular origins are poorly understood. I used the Drosophila Pins protein, a regulator of spindle positioning in neuroblast cells, as a model to study the molecular origin and function of autoinhibition and ultrasensitivity.

Pins and its binding partners. Gαi and Mud, form a signaling pathway required for coordinating spindle positioning with cellular polarity in Drosophila neuroblasts. I found Pins switches from an autoinhibited to an activate state by modular allostery. Gαi binding to the third of three GoLoco (GL) domains allows Pins to interact with the microtubule binding protein Mud. The GL3 region is required for autoinhibitoon, as amino acids upstream and within GL3 constitute this regulatory behavior. This autoinhibitory module is conserved in LGN, the mammalian Pins orthologue.

I also demonstrated that Gαi activation of Pins is ultrasensitive. A Pins protein containing inactivating point mutations to GLs 1 and 2 exhibits non-ultrasensitive (graded) activation. Ultrasensitivity is required for Pins function in vivo as the graded Pins mutant fails to robustly orient the mitotic spindle. I considered two models for the source of ultrasensitivity in this pathway: cooperative or "decoy" Gai binding. I found ultrasensitivity arises from a decoy mechanism in which GLs 1 and 2 compete with the activating GL3 for the input, Gai. These findings suggest that molecular ultrasensitivity can be generated without cooperativity. This decoy mechanism is relatively simple, suggesting ultrasensitive responses can be evolved by the inclusion of domain repeats, a common feature observed in signaling proteins.

This dissertation includes previously published and unpublished co-authored material.