Membrane-spanning α-helical barrels as tractable protein-design targets

Abstract

The rational (de novo) design of membrane-spanning proteins lags behind that for water-soluble globular proteins. This is due to gaps in our knowledge of membrane-protein structure, and experimental difficulties in studying such proteins compared to water-soluble counterparts. One limiting factor is the small number of experimentally determined three-dimensional structures for transmembrane proteins. By contrast, many tens of thousands of globular protein structures provide a rich source of 'scaffolds' for protein design, and the means to garner sequence-to-structure relationships to guide the design process. The α-helical coiled coil is a protein-structure element found in both globular and membrane proteins, where it cements a variety of helix-helix interactions and helical bundles. Our deep understanding of coiled coils has enabled a large number of successful de novo designs. For one class, the α-helical barrels-that is, symmetric bundles of five or more helices with central accessible channels-there are both water-soluble and membrane-spanning examples. Recent computational designs of water-soluble α-helical barrels with five to seven helices have advanced the design field considerably. Here we identify and classify analogous and more complicated membrane-spanning α-helical barrels from the Protein Data Bank. These provide tantalizing but tractable targets for protein engineering and de novo protein design.This article is part of the themed issue 'Membrane pores: from structure and assembly, to medicine and technology'.

Keywords: coiled coil; de novo protein design; transmembrane proteins; α-helical barrel.

Figure 1.

Examples of three different families of natural membrane-spanning α-helical bundles. The transmembrane α-helical bundles are shown in red. (a) A heteromeric ABC transporter (3QF4) [13]. (b) A G-protein–coupled receptor, adenosine A_2A receptor (5G53) [11]. (c) A bacterial voltage-gated sodium channel (3RVY) [6].

Figure 2.

Structural models of ROCKER [20] (a), and cWza pore [19] (b).

Figure 3.

The 10 barrels in their four categories, with each row representing a class. The transmembrane region for each structure, as determined by TMDET, is shown in cartoon form. The coiled-coil regions are shown in rainbow colours, according to the assigned heptad register, with a residues in red and d residues in green. Row 1: symmetric, parallel, autonomous (2J58). Row 2: symmetric, parallel, buttressed (3UQ4, 4EV6, 4HKR, 4HW9, 4BEM). Row 3: symmetric, antiparallel, autonomous (4QND). Row 4: asymmetric, antiparallel, autonomous (4C9 J, 4IL3, 4CAD).