Design principles of human membrane protein topology.

Abstract

We have curated and annotated the topologic determinants for all human membrane proteins made at the endoplasmic reticulum. This census of 4,863 proteins allowed us to systematically analyze the physical properties of their 20,546 transmembrane domains (TMDs) and flanking soluble regions. Single-pass proteins house the majority of large exoplasmic and cytosolic domains, whereas multipass proteins overwhelmingly contain short loops and tails. All classes of TMDs have positively charged cytosolic flanks, but negatively charged exoplasmic flanks feature primarily on TMDs inserted by Oxa1 family insertases. The TMD pair, a topologic unit of two TMDs with a short exoplasmic loop, is the dominant building block of multipass proteins. TMD pairs accommodate high-hydrophilicity and charge-containing TMDs crucial for multipass protein functions. We interpret these context-dependent TMD features in light of current mechanistic models for membrane protein biogenesis and function. Our findings have implications for the evolution of membrane proteomes and for engineering new membrane proteins.