Structures of SAS-6 coiled coil hold implications for the polarity of the centriolar cartwheel

Abstract

Centrioles are eukaryotic organelles that template the formation of cilia and flagella, as well as organize the microtubule network and the mitotic spindle in animal cells. Centrioles have proximal-distal polarity and a 9-fold radial symmetry imparted by a likewise symmetrical central scaffold, the cartwheel. The spindle assembly abnormal protein 6 (SAS-6) self-assembles into 9-fold radially symmetric ring-shaped oligomers that stack via an unknown mechanism to form the cartwheel. Here, we uncover a homo-oligomerization interaction mediated by the coiled-coil domain of SAS-6. Crystallographic structures of Chlamydomonas reinhardtii SAS-6 coiled-coil complexes suggest this interaction is asymmetric, thereby imparting polarity to the cartwheel. Using a cryoelectron microscopy (cryo-EM) reconstitution assay, we demonstrate that amino acid substitutions disrupting this asymmetric association also impair SAS-6 ring stacking. Our work raises the possibility that the asymmetric interaction inherent to SAS-6 coiled-coil provides a polar element for cartwheel assembly, which may assist the establishment of the centriolar proximal-distal axis.

Keywords: SAS-6; asymmetry; cartwheel; centriole; coiled-coil; complex; crystallography; electron microscopy; self-assembly; stacking.

Graphical abstract

Figure 1

Overview of SAS-6 oligomers in the centriolar cartwheel and of the coiled-coil domains interactions (A) Schematic of SAS-6 dimeric architecture from structural and biophysical studies. Red highlights the region that has been resolved previously with X-ray crystallography, gray a region with unknown structure, modeled here as the continuation of the coiled coil. (B) Nine SAS-6 dimers associate via interactions between their N-terminal head domain to form a SAS-6 ring. (C) Section of the cartwheel from C. reinhardtii centrioles resolved by cryo-ET (Klena et al., 2020), showing the central hub and spokes that progressively merge into larger bundles toward the organelle periphery. Red boxed region indicates the position of a single SAS-6 ring. Colored bars mark the regions that have resolved with X-ray crystallography in this study; green bar marks HsSAS-6_Middle, orange bar DrSAS-6_Middle, and blue bar CrSAS-6_Middle. Note that spokes, being initially discrete, have poor density toward the center of the cartwheel. (D–F) Schematic representation of the SAS-6 coiled-coil dimer structures resolved from HsSAS-6_CC_Middle (D), DrSAS-6_CC_Middle (E), and CrSAS-6_CC_Middle (F) crystals and ab initio phasing by SeMet incorporation. Lighter and darker colors distinguish the two protein chains in each dimer. The HsSAS-6_CC_Middle coiled coil is distorted compared with a canonical CC structure at a region N-terminal to the site of SeMet incorporation (indicated). See also Figures S1 and S2.

Figure 2

SAS-6 coiled-coil interactions in crystals (A and B) Schematic representation of how SAS-6 rings would position following an (A) parallel or (B) antiparallel association between coiled-coil domains. As shown, parallel association is the only mode compatible with stacking of SAS-6 rings, thereby giving rise to cartwheels. (C, D) Two adjacent asymmetric units from DrSAS-6_CC_Middle (C) and HsSAS-6_CC_Middle (D) crystals. Chains are colored from N-terminus (blue) to C-terminus (red). (E and F) Similar representation of a single asymmetric unit from CrSAS-6_CC_Middle crystal forms I (E) and II (F).

Figure 3

Structures of symmetric and asymmetric complexes of CrSAS-6 coiled-coil domains (A) Surface representation of the CrSAS-6 symmetric complex of two coiled-coil domains, in green and blue, from crystal form I. The internal 2-fold symmetry axis is indicated. (B–D) Detailed views of three key contact areas between the coiled-coil domains. The side chains of amino acids participating in ionic (R312-E309′, E343-K344′, R377-D378′, and R386-E389′, where prime denotes residues of the second coiled coil in the complex) and hydrogen-bonding (Y283-E284′, E300-H294′, R304-S298′, E308-S305′, R312-E309′, S347-S347′, R377-D378′, and R386-E389′) interactions are represented as sticks, and those of a hydrophobic interaction (L339-L339′) as spheres. (E–H) Surface representation of CrSAS-6 asymmetric coiled-coil complex in yellow and red (E) and detailed views of three key contact areas (boxed) in (F–H). Shown in detailed views are amino acids forming key hydrogen-bonding (S290-D286′, H294-S293′, E309-S305′, R314-E308′, N332-N332′, E343-K337′, Q354-S347′, R358-D350′, R358-Q354′, R367-D361′, E370-R358′, Q371-Q368′, D375-R372′, R377-E370′, D378-C373, and D378-R372′), ionic (R314-E308′, E343-K337′, R356-D350′, R358-D350′, R367-D361′, E370-R358′, D375-R372′, R377-E370′, and D378-R372′), and hydrophobic (H294-G297ʹ/G301′, A340-A336′, V351-S347′ interactions. (I) Schematic representation of symmetric (green and blue) and asymmetric (orange and red) CrSAS-6 coiled-coil complexes from crystal form II. The side chains of leucine amino acids in the cores of coiled-coil dimers are shown as spheres for reference. In the asymmetric complex, one coiled-coil dimer is translated by 5 Å (approximately one helical turn) along the longitudinal axis. (J) Overlay of symmetric and asymmetric CrSAS-6 coiled-coil complexes, shown in perpendicular orientation compared with (E) and in a section around L331. One coiled-coil dimer from each complex was superimposed (green and orange); as seen, the second coiled-coil dimers (blue and red) are rotated by 35° relative to one another. See also Figure S2.

Figure 4

Analysis of CrSAS-6 oligomeric state in solution (A) ITC of CrSAS-6_CC WT, CrSAS-6_CC_LD, and CrSAS-6_CC_VD. Concentrated CrSAS-6_CC_Middle injected into sample buffer alone produced heat absorptions (ΔP) consistent with the dissociation of a weak protein complex. Substitutions of amino acids affecting either the symmetric (CrSAS-6_CC_LD) or the asymmetric (CrSAS-6_CC_VD) interaction mode reduces the observed heats of dilution compared with WT. (B) Analysis of the CrSAS-6_CC_Middle dilution heats plot in relation to the monomer concentration suggests that the K_d of CrSAS-6_CC_Middle coiled-coil dimers is in the low mM range. See also Figure S3.

Figure 5

Cartwheel reconstitution using CrSAS-6 WT and mutants (A) Exemplar cryoelectron micrographs of in vitro reconstituted cartwheels using WT CrSAS-6_NL; insets show magnification of field in dashed boxes. (B) Structures of symmetric and asymmetric CrSAS-6 coiled-coil complexes; black dashed boxes denote the areas of amino acid substitutions in (C, D). (C and D) Cryoelectron micrographs of in vitro reconstituted cartwheels using CrSAS-6_NL mutants designed to destabilize symmetric (C) or asymmetric (D) CrSAS-6 coiled-coil complexes. Structures to the right of micrographs highlight favorable interactions abolished by amino acid substitutions in the symmetric (C) and asymmetric (D) complexes. Red dashed lines in (C, D) indicate hydrogen bonds inferred from the structures. Scale bars 50 nm and 25 nm in the magnification inset. See also Figures S4–S6.

Figure 6

Model of the emergence of centriole polarity rooted in the cartwheel structure (A) Model of SAS-6 ring stacking viewed from the top. Two stacked rings (dark brown at the bottom, orange at the top) are shown schematically, the interacting section of SAS-6 coiled coils are represented as boxes. The rotational offset of the rings as observed in the cryo-ET maps (Klena et al., 2020; Nazarov et al., 2020) is in line with an offset at the coiled-coil merging uncovered here; blue box rotated is shown in (B), green box in (C). (B) Side view of a SAS-6 stack in the cartwheel, showing one-half of CrSAS-6 rings and spokes. The proximal-distal axis of centriole polarity is indicated on the left. Spacings from the centriole center are inferred from crystallographic structures of the CrSAS-6 N-terminal domain defining the hub diameter (Kitagawa et al., 2011), and our coiled-coil structures (this work). (C) Close-up view of the SAS-6 cartwheel model showing how ring offset and the asymmetric coiled coil are in agreement in the current stacking model.