Coiled‐coils are structurally conserved
In 1952, Francis Crick described a new type of protein structure, which he called the coiled‐coil 2. Hypothetical at the time, the coiled‐coil was proposed as a solution to the X‐ray fiber diffraction patterns of α‐keratin, which showed a strong reflection at 5.1 Å compared to the reflection at 5.4 Å observed for a canonical α‐helix. Crick correctly predicted that by supercoiling an α‐helix, one could obtain a repeating structure in which two or more adjacent α‐helices could pack together in a knobs‐into‐holes arrangement. Such a structure is impossible to obtain from two straight helices arranged side by side due to the non‐integral nature of the helix. Crick proposed that the energy required to deform the helices could be compensated for by the closer packing of the side chains. The structure of influenza virus hemagglutinin 3 later confirmed the basic parameters of the coiled‐coil 4, 5. Indeed, Crick’s parametric equations, which describe coiled‐coils using just four parameters, are able to reproduce the majority of subsequently experimentally determined coiled‐coil structures and are at the heart of CCCP and CC‐builder, two web‐based tools for designing, building, and assessing coiled‐coil assemblies 6, 7.
The canonical coiled‐coil is formed from a heptad repeat, labeled abcdefg, in which hydrophobic amino acids at positions a and d are conserved (Fig. (Fig.1).1). The conservation of hydrophobic residues alternately three and four residues apart (average 3.5) is close to the 3.6 amino acids per turn periodicity of a regular α‐helix. Consequently, helices deriving from such repeating sequences exhibit distinct amphipathic character, with both hydrophobic and polar faces. The association of two helices via their hydrophobic faces drives coiled‐coil formation. However, in order to pack two helices together and maintain hydrophobic contacts, the knobs‐into‐holes packing of side chains requires that these residues occupy equivalent positions, turn after turn. By supercoiling the helices around each other, the periodicity is effectively reduced from 3.6 to 3.5 with respect to the supercoil axis. This allows the positions of side chains to repeat after two turns, or seven residues, instead of drifting continuously on the helical surface 8.
Coiled‐coil domains have since been shown to be ubiquitous repetitive peptide motifs present in all the domains of life 9. In total, coiled‐coil forming sequences comprise up to 10% of an organism’s proteome 9, 10, 11. Naturally occurring coiled‐coils comprise between two and six helices arranged either parallel or antiparallel to each other. They may be homo‐ or hetero‐oligomers, and may be formed from separate chains, or from consecutive helices of the same chain. While variations on the canonical heptad repeat are known, these non‐canonical motifs are also based on repeating hydrophobic and polar residues spaced three or four residues apart. These coiled‐coils exhibit slight variations in the knobs‐into‐holes packing and different degrees of supercoiling 8, 12. Discontinuities caused by the insertion of one, three, or four residues result in local deformations in the coiled‐coil assembly (Fig. (Fig.1),1), while insertions of two or six residues strain the supercoil to breaking point, leading to the local formation of β‐strands that move the path of the chain by 120° around the trimer axis. The resulting structure is a so‐called α‐β coiled‐coil, in which the β‐strands associate to form a triangular structure called a β‐layer (Fig. (Fig.1).1). Hartmann and co‐woorkers show that β‐layers are found in naturally occurring proteins in which they form fibers with repeating α and β structure 13. Clear sequence to structure relationships 14 coupled with their intrinsic properties of periodically repeating structure, capacity for self‐assembly, and mechanical strength have seen coiled‐coils feature prominently in recent protein engineering efforts 15, 16, 17, 18. While the applications for de novo coiled‐coil design and prospects for synthetic biology fall outside the scope of this review, the potential to inform our understanding of coiled‐coil structure, folding, and function as well as for innovation in biotechnology, material science, and medicine is manifold 19.
Based on the underlying sequence repeats that govern their assembly, coiled‐coils can be reliably predicted from primary sequence 20. However, a more recent extension of the SUPERFAMILY database for coiled‐coil domains has improved these predictions by incorporating existing genomic and structural information 11. Examples of structurally validated coiled‐coils of different architectures may be found in the periodic table of coiled‐coil structures 21, while the CC+ database 22 of coiled‐coil structures (http://coiledcoils.chm.bris.ac.uk/ccplus/search/dynamic_interface) is a useful resource for structural and cell biologists alike. The rest of this review will focus on the biological roles of coiled‐coil domains.