Evolution and Structural Adaptation to Membranes of Single-Pass Transmembrane Proteins

Abstract

More than 6000 single-pass transmembrane (bitopic) proteins from six representatives of different kingdoms of life, Homo sapiens, Arabidopsis thaliana, Dictyostelium discoideum, Saccharomyces cerevisiae, Escherichia coli, and Methanococcus jannaschii (2545, 2115, 649, 517, 199 and 70 proteins, respectively), were identified, classified into 14 functional classes, 692 superfamilies and 1410 families, and collected in the Membranome database (at membranome.org). Three-dimensional (3D) models of transmembrane alpha-helices (TMH) were generated using an updated thermodynamic model of alpha-helix formation in membranes. Comparative analysis demonstrates that the repertoire of bitopic proteins is significantly enlarged in eukaryotic cells due to diversification of enzymes and emergence of proteins involved in vesicular traffic and biogenesis. A dramatic expansion of receptors, regulatory, and structural/adhesion proteins is seen in multicellular organisms. The relative abundance of receptors increases from prokaryotes ( 30%). Interestingly, the diversity of receptors, as estimated by the number of their superfamilies, is significantly smaller in A. thaliana than in humans. The calculated stabilities, lengths and tilt angles of TMHs are smallest for proteins from inner mitochondrial membranes, intermediate for prokaryotic proteins and proteins from endoplasmic reticulum/Golgi and chloroplast membranes and highest for proteins from eukaryotic plasma membrane. This reflects adaptation of TMHs to different membrane environments. The majority of bitopic proteins from multicellular organisms are located in plasma membrane with N-out topology, while prokaryotic cell membrane proteins mostly have N-in topology. The distributions of TMH lengths, hydrophobic thicknesses, and tilt angles are frequently bimodal (at ∼0-8° and ∼30°), which may be related to membrane heterogeneity or different modes of helix-helix association. The oligomerization propensity of bitopic proteins was investigated using computational docking, analysis of ∼450 experimental 3D structures of complexes and information from protein interaction databases.