Fragment‐based modeling of membrane protein loops : successes, failures, and prospects for the future

Abstract

Membrane proteins have become a major focus in structure prediction, due to their medical importance. There is, however, a lack of fast and reliable methods that specialise in the modelling of membrane protein loops. Often methods designed for soluble proteins are applied directly to membrane proteins. In this paper we investigate the validity of such an approach in the realm of fragment-based methods. We also examine the differences in membrane and soluble protein loops that might affect accuracy. We test our ability to predict soluble and membrane protein loops with the previously published method FREAD. We show that it is possible to predict accurately the structure of membrane protein loops using a database of membrane protein fragments (0.5-1°A median root mean square deviation). The presence of homologous proteins in the database helps prediction accuracy. However, even when homologues are removed better results are still achieved using fragments of membrane proteins (0.8-1.6°A) rather than soluble proteins (1-4°A) to model membrane protein loops. We find that many fragments of soluble proteins have shapes similar to their membrane protein counterparts but have very different sequences, however they do not appear to differ in their substitution patterns. Our findings may allow further improvements to fragment-based loop modelling algorithms for membrane proteins. The current version of our proof-of-concept loop modelling protocol produces high accuracy loop models for membrane proteins and is available as a web server at http://medeller.info/fread.