Structural studies of sorting nexin proteins from two distinct subfamilies

Abstract

Sorting nexin (SNX) proteins are a large family of proteins with critical roles in endocytosis, membrane trafficking and intracellular signalling. Each SNX protein contains a Phox homology (PX) domain that typically recognizes phosphoinositide lipids to enable their anchoring to defined organelles. SNXs also contain additional domains other than the PX domain thus allowing them to participate in a wide range of functions such as protein-protein interactions, membrane remodeling, lipid metabolism and other functions that are yet to be explored. My research focuses on the structural and functional characterization of two different sub-families of SNX proteins; those that have both PX and BAR (bin/amphiphysin/rvs) domains (SNX5, SNX6 and SNX32), and the poorly characterized SNX-RGS sub-family that contain a regulator of G-protein signaling (RGS) domain (SNX13, SNX14, SNX19 and SNX25). These protein families are required for membrane trafficking and lipid droplet formation, and are implicated in diseases including pathogen invasion and cerebellar ataxia respectively.I first describe the structural mechanism for how a Chlamydial effector protein called IncE hijacks and recruits SNX5 related proteins to the bacterial inclusion membrane during Chlamydial infection (Chapter 2). Using X-ray crystallography, I demonstrate that the C-terminal region of IncE forms a β-hairpin structure and binds to a unique α-helical insertion on the PX domain of SNX5, which is a separate site from the canonical PX lipid-binding site. Using isothermal titration calorimetry, I also investigate the binding mechanisms of the SNX5 homologs, SNX6 and SNX32 to IncE, and demonstrate that these three proteins share a common binding mechanism. These results suggest that the IncE protein mimics other SNX5-related proteins, where SNX5 PX might be functioning as a protein-binding scaffold that could potentially orchestrate the trafficking of certain transmembrane receptors.The research presented in Chapter 3 extends the work described in Chapter 2 confirming the IncE binding site in the SNX32 PX domain using X-ray crystallography. The crystal structure of SNX32 PX-IncE complex not only confirmed the biophysical data from chapter 2, but also revealed the binding mechanism to be identical to the SNX5PX-IncE binding. In addition, this is the first reported structure of the SNX32 protein in either an apo or ligand bound state. This result further strengthens the idea that SNX-BAR proteins can act as receptor recyclers in the cell.In Chapter 4 using biophysical and structural studies, I next investigated the potential functional role of the SNX-BAR proteins in transmembrane cargo trafficking. My biophysical experiments directly confirmed the recent indirect proteomic studies that suggested cargos including CIMPR, IGF1R and SEMAC can bind to SNX-BAR proteins, and my first low-resolution crystal structure of the SNX5 PX domain with a transmembrane cargo, CIMPR, further confirms my hypothesis that endogenous human transmembrane cargos binds SNX5 related proteins in an analogous manner to IncE. Even though CIMPR does not possess sequence similarity with IncE, the identified hydrophobic patch in SNX5 is required and sufficient for the binding to the CIMPR peptide.In Chapter 5, I explore the idea of synthesizing cyclic peptides based on my structure of the linear IncE peptide, which when bound to SNX-BAR proteins I showed forms a β-hairpin. The N- and C-termini of IncE were linked through various macrocyclisation approaches, and this yielded a cyclic IncE derived peptide that binds to SNX5 related SNX-BAR proteins with a nanomolar binding affinity that is 30x higher than the standard IncE peptide, and ~300x higher than cellular cargo molecules. I further demonstrate the ability of this cyclic peptide to inhibit the binding of a transmembrane cargo, CIMPR to SNX5 in vitro.In Chapter 6, I describe my initial studies regarding whether the RGS domains of SNX-RGS proteins possess GTPase activating protein (GAP) activity for the small GTPase Gαs. My studies reveal that RGS domains of SNX13, SNX14 and SNX25 binds to Gαs, but do not possess GAP activity. My results lead to me to speculate that since SNX-RGS proteins bind to Gαs but do not regulate its GDP/GTP-bound activity states, these proteins may still act as negative regulators of Gαs signaling through ways that are yet to be explored.In Chapter 7 I provide a summary of my attempts to successfully express and purify the structurally uncharacterized PXA domain from the SNX-RGS subfamily, which is directly implicated in lipid droplet biogenesis. My trials have yielded soluble PXA protein and initial crystal hits, which provides a solid foundation for further structural and functional studies of this domain.Finally, in Chapter 8 I provide an overall summary of my work, both outcomes and future directions. Overall, my studies have revealed the crucial role of PX domains not only in membrane interactions but in direct protein-protein interactions required for cargo sorting, provides insights into how SNX proteins contribute to pathogen invasion and other diseases, and how structural information can be used to target these proteins with peptide tools in functional studies and in potential therapeutic approaches.