ABSTRACT Cholesterol (CLR) is an integral component of mammalian membranes. It has been shown to modulate membrane dynamics and alter integral membrane protein (IMP) function. However, understanding the molecular mechanisms of these processes is complicated by limited and conflicting structural data: Specifically, in co-crystal structures of CLR-IMP complexes it is difficult to distinguish specific and biologically relevant CLR-IMP interactions from a nonspecific association captured by the crystallization process. The only widely recognized search algorithm for CLR-IMP interaction sites is sequence-based, i.e. searching for the so-called ‘CRAC’ or ‘CARC’ motifs. While these motifs are present in numerous IMPs, there is inconclusive evidence to support their necessity or sufficiency for CLR binding. Here we leverage the increasing number of experimental CLR-IMP structures to systematically analyze putative interaction sites based on their spatial arrangement and evolutionary conservation. From this analysis we create three-dimensional representations of general CLR interaction sites that form clusters across multiple IMP classes and classify them as being either specific or nonspecific. Information gleaned from our characterization will eventually enable a structure-based approach for prediction and design of CLR-IMP interaction sites. SIGNIFICANCE CLR plays an important role in composition and function of membranes and often surrounds and interacts with IMPs. It is a daunting challenge to disentangle CLRs dual roles as a direct modulator of IMP function through binding or indirect actor as a modulator of membrane plasticity. Only recently studies have delved into characterizing specific CLR-IMP interactions. We build on this previous work by using a combination of structural and evolutionary characteristics to distinguish specific from nonspecific CLR interaction sites. Understanding how CLR interacts with IMPs will underpin future development towards detecting and engineering CLR-IMP interaction sites.

Structural Determinants of Cholesterol Recognition in Helical Integral Membrane Proteins