Coiled-coils: The long and short of it

Abstract

Coiled-coils are found in proteins throughout all three kingdoms of life. Coiled-coil domains of some proteins are almost invariant in sequence and length, betraying a structural and functional role for amino acids along the entire length of the coiled-coil. Other coiled-coils are divergent in sequence, but conserved in length, thereby functioning as molecular spacers. In this capacity, coiled-coil proteins influence the architecture of organelles such as centrioles and the Golgi, as well as permit the tethering of transport vesicles. Specialized coiled-coils, such as those found in motor proteins, are capable of propagating conformational changes along their length that regulate cargo binding and motor processivity. Coiled-coil domains have also been identified in enzymes, where they function as molecular rulers, positioning catalytic activities at fixed distances. Finally, while coiled-coils have been extensively discussed for their potential to nucleate and scaffold large macromolecular complexes, structural evidence to substantiate this claim is relatively scarce.

Keywords: allostery; coiled-coil; molecular ruler; molecular spacer; scaffold.

Figure 1

Coiled‐coil architecture. Coiled‐coil parameters. The basic parameters of the coiled‐coil are its pitch (periodicity of the supercoil), the associated pitch (or supercoil) angle (angle of the helix with respect to the supercoil axis), and the helix‐crossing angle (angle at which the two helices cross each other). Packing interactions. The canonical coiled‐coil is characterized by a heptad repeat in which hydrophobic residues are conserved at positions a and d. Undistorted α‐helices cannot pack together in a side‐by‐side arrangement due to the non‐integral periodicity of the helix (3.6). By supercoiling the helices, the periodicity is reduced to 3.5, leading to the 7/2 periodicity of a canonical, left‐handed coiled‐coil, with each heptad repeat measuring ∼1 nm along the coiled‐coil. Knobs‐into‐holes packing of two parallel, supercoiled helices results in a and d layers, while antiparallel helices exhibit mixed ad layers. Discontinuities. Insertions of 1, 2, 3, 4, or 6 amino acids give rise to discontinuities in the heptad repeat and local structural deformations in the coiled‐coil. Examples are shown for insertions of 1, 3, 4, and 2/6 amino acids in comparison to the canonical coiled‐coil of GCN4.

Figure 2

Coiled‐coils used as molecular spacers. Golgins. The Golgins are a family of tethering proteins with long coiled‐coils that capture vesicles on the surface of the Golgi. GMAP‐210 contains a curvature‐sensing ALPS motif in its C‐terminus, which allows it to selectively capture and tether vesicles in vitro and in vivo 116, 117, 118. This panel is based on Fig. 2 from Witkos and Lowe 33. Kinetochore. Ndc80 is a coiled‐coil component of the kinetochore that, in combination with the coiled‐coil proteins Spc24 and Spc25, links the centromere to the spindle microtubule. Ndc80 is highly conserved in length (Table 1). The Ska complex (Ska1/2/3) of coiled‐coil proteins forms a W‐shaped dimer of 35 nm that binds to kinetochore microtubules and is essential for mitosis. This panel is based on Fig. 1 from Matson and Stukenberg 119. Centrioles. Sas‐6 forms a coiled‐coil dimer. Dimerization of the head domains of Sas‐6 with neighboring Sas‐6 molecules leads to formation of the centriolar hub, a cartwheel with ninefold symmetry in which the coiled‐coil domains of Sas‐6 form the spokes. Bld10p, another coiled‐coil protein essential for centriole assembly, stabilizes the ninefold symmetric arrangement of the centriole via a yet to be determined mechanism. The ninefold symmetric cartwheel was recently visualized in a crystal structure of Leishmania Sas‐6 50, reflecting the structure of centrioles visualized previously by electron microscopy 51 and cryo‐electron tomography 53, 54 (reproduced with permission). Type I restriction‐modification enzymes. The DNA binding domains are held at a fixed distance apart by an intramolecular anti‐parallel coiled‐coil. The arrangement permits the recognition of two DNA sequences separated by a fixed spacer, and recruits a restriction enzyme that cuts randomly either side 55. Transcriptional regulators. Dimerization of CueR via an antiparallel coiled‐coil maintains the DNA binding domains (DBD) at a fixed distance of 55 Å that permits the recognition of successive major grooves of B‐DNA via a recognition helix in the DBD. Binding of heavy metals to a metal‐binding loop at the C‐terminus of CueR results in an allosteric change in the conformation of the DBDs with respect to the coiled‐coil. This conformational change distorts B‐DNA to a more open A‐DNA, converting CueR from a transcriptional repressor into an activator 57.

Figure 3

SMC proteins – beyond simply a molecular spacer. Structural maintenance of chromosomes (SMC) proteins. SMC proteins are characterized by N‐ and C‐terminal Walker ATPase domains, separated by two stretches of sequence that fold back upon each other to form an antiparallel coiled‐coil. A hinge domain in the middle mediates dimerization of two SMC proteins in cohesin, condensin, and the DNA repair complex of SMC5/6. The SMC proteins are conserved in both length and sequence. Eukaryotic cohesin (SMC1/3), in particular, exhibits almost invariant sequence in vertebrates. This panel was adapted from Box 1 of Jeppsson et al. 120 with permission. Rad50. Closely related to the SMC proteins is the DNA damage response regulator Rad50, which contains an antiparallel coiled‐coil of 50 nm, the same length as SMCs. Rad50 dimerizes via a specialized zinc‐hook domain, the Rad50 counterpart of the SMC hinge domain, allowing it to bridge sister chromatids during meiotic recombination. Evolutionarily conserved in length, truncations in the coiled‐coil lead to defects in Rad50 function 60. This panel is based on Fig. 1a of Hohl et al. 60. Smc protein architectures. A number of Smc/Rad50 structures have been observed in vitro from the closed circle (mediated by dimerization of the hinge and Walker ATPase domains) to linear dimers and rod‐shaped dimers (in which the coiled‐coils pack together side‐by‐side).

Figure 4

Allosteric communication by coiled‐coils. Dynein. Dyneins are minus‐end directed motors highly conserved in both length and sequence (Table 1). The hexameric AAA ATPase motor domains are bridged to the microtubule‐binding domain by a coiled‐coil stalk, 15 nm in length. Conformational changes in the motor domain upon ATP hydrolysis in AAA1 are propagated to the stalk via the buttress. The buttress pulls on one helix of the coiled‐coil in the stalk, causing it to shift register by one full turn of the α‐helix. The sliding of the helices with respect to one another causes, in turn, a conformational change in the microtubule‐binding domain at the tip of the stalk, driving it into a low‐affinity conformation and promoting microtubule release. Microtubule re‐binding leads to a force‐producing swing of the linker (power stroke) back to its preferred straight conformation (post‐power stroke) and the release of ATP hydrolysis products resets the cycle. This figure was adapted from Fig. 4 of Schmidt et al. 69 with permission. Kinesin‐5. Kinesin‐5 is a homotetrameric bipolar plus‐end directed motor essential for spindle assembly and elongation during anaphase. Kinesin‐5, like kinesin‐1, is highly conserved in length (Table 1), suggesting that the distance at which microtubules are maintained is critical to their sliding in the elongating spindle. The recently determined structure of the central portion of kinesin‐5 revealed a 4‐helix bundle comprising the antiparallel coiled‐coils of each kinesin‐5 dimer 85. The structure suggests that the 90° offset of the coiled‐coils at either end of the kinesin‐5 mini‐filament might enable the motor domains at either end of the filament to step processively along each microtubule, thereby permitting or constraining outward sliding within the mitotic spindle. The coiled‐coil assembly is proposed to be mechanically stable, capable of resisting tensile and compressive forces during spindle elongation.

Figure 5

Molecular rulers. WbdD. O‐antigen synthesis in the pathogenic gram‐negative bacteria E. coli O9a is regulated by a polymerase, WbdA together with a capping enzyme, WbdD. Comprising N‐terminal methyltransferase and kinase domains followed by a C‐terminal coiled‐coil and transmembrane segment, WbdD assembles into trimers in which the catalytic domains are maintained at a distance of approximately 20 nm from the plasma membrane. Truncations in the coiled‐coil lead to a decrease in the modal length distribution of the O‐antigen polymannose chains 87. Source material for figure: G. Hagelueken and J.H. Naismith. ROCK2. Electron microscopy of full‐length human ROCK2 recently revealed it to be a constitutive dimer, 120 nm in length 90. N‐terminal kinase domains are separated from C‐terminal membrane‐binding domains by 107 nm of parallel coiled‐coil, which is evolutionarily conserved in length, but divergent in sequence (Table 1). ROCK2 is constitutively active in vitro and unaffected by membrane binding, RhoA, or activation loop phosphorylation. Truncations in the coiled‐coil, however, lead to defects in stress fiber formation in vivo, but do not affect catalytic activity. A new model proposes that the coiled‐coil of ROCK functions as a molecular ruler, positioning the kinase domains at a fixed distance from the plasma membrane, coincident with its substrates in the actin cytoskeleton.