Centriole and transition zone structures in photoreceptor cilia revealed by cryo-electron tomography

Abstract

Primary cilia mediate sensory signaling in multiple organisms and cell types but have structures adapted for specific roles. Structural defects in them lead to devastating diseases known as ciliopathies in humans. Key to their functions are structures at their base: the basal body, the transition zone, the "Y-shaped links," and the "ciliary necklace." We have used cryo-electron tomography with subtomogram averaging and conventional transmission electron microscopy to elucidate the structures associated with the basal region of the "connecting cilia" of rod outer segments in mouse retina. The longitudinal variations in microtubule (MT) structures and the lumenal scaffold complexes connecting them have been determined, as well as membrane-associated transition zone structures: Y-shaped links connecting MT to the membrane, and ciliary beads connected to them that protrude from the cell surface and form a necklace-like structure. These results represent a clearer structural scaffold onto which molecules identified by genetics, proteomics, and superresolution fluorescence can be placed in our emerging model of photoreceptor sensory cilia.

Figure 1.. Tomographic reconstruction and subtomogram averaging of three structural domains from mouse centrioles.

(A) Slice from a reconstructed tomogram showing mother and daughter centrioles (denoted as “M” and “D”). Centrioles were partitioned longitudinally into three regions, the proximal end (0–170 nm, light magenta) consisting of MT triplets, the mid-region (170–340 nm, blue) consisting of incomplete triplets, and the distal end (340–400 nm, light violet) consisting of MT doublets. The extending CC doublets are denoted in cyan. (B) Slices of a tomogram of a daughter centriole through the three centriole regions and the CC (from a different part of the tomogram). White boxes indicate examples of approximate regions used for initial stages of subtomogram averaging and refinement. (C) Four maps were generated by averaging tomogram subvolumes in each region of centrioles and of the CC. The A-, B-, and C-tubules, including non-tubulin densities, are shown, with protofilament numbering. Microtubule inner proteins are shown with green arrows. (D) Maps in C cut in half and rotated 90° toward the viewer to show microtubule inner proteins.

Figure 2.. Centriole model.

(A) Ninefold symmetrized cross sections of a map generated by rotational averaging of a daughter centriole at the proximal (left), middle (middle), and distal (right) regions, respectively, with maps generated by subtomogram averaging of triplet, incomplete triple, or doublet averages (those shown in Fig 1C) fitted onto the symmetrized map and superimposed on the right side of each corresponding map. (B) Fully refined, symmetrized, reconstructed/fitted models used for examination of extra-MT densities and inter-MT connections for microtubule triplets and microtubule doublets.

Figure 3.. Fit of MT subtomogram averages into raw tomogram and range of MT twist angles in the cross-sectional plane.

(A) Averaged subvolumes of triplets (cyan; the same subvolumes used for Fig 2) and doublets (magenta) were fitted back into the raw tomogram of an entire mother centriole–CC to generate a whole-cilium map. (A, B) Cross sections (top panel, model; bottom panel, raw tomogram) corresponding to the positions indicated by dashed white lines in (A) and denoted by A1–A5 in panel (B). (C, D, E) Twist angles β (defined in panel (C)) were measured and averaged for each cross section, and are plotted in (D) for the cilium map and (E) for the centriole map as a function of longitudinal position. (E) This procedure was repeated using a ninefold rotationally averaged centriole map (Fig S1) and are plotted as boxes in (E). Points and error bars represent means ± SD.

Figure S1.. Axial and cross-sectional twist angles of the centriole and CC.

(A) Projection views of the raw tomogram reconstruction of the centriole from the top (showing the distal end of the centriole) and side views. To the far right is a cutaway view. The side views illustrate the gradual twist of the entire centriole (highlighted with cyan [left] and green [right] dashed lines). (B) The 9-fold symmetrized map of the centriole displayed as for the raw map in (A), except for radial distance-coded color scheme and surface view. The diameters are indicated in (B). (C) Schematic diagram of centriole MTTs and how the ß-angles and α-angles are measured (see main Fig 3). (A, B) Note that for 9-fold symmetry, these two angles are related as follows: $\beta =70°-{\mathit{cos}}^{-1}(\frac{r_{ab}-r_{ma}\mathit{cos}(70°+\alpha )}{{{(r}_{ab}^2+r_{ma}^2 -2r_{ab}{r}_{ma}\mathit{cos}(70°+\alpha ))}^{1/2}})$,where r_ab is the distance between the centers of the (A and B) MT, r_ma is the distance from the center (middle) of the MT bundle to the center of the (A) MT. (D) Variation of the triplet and doublet twist angles along the entire cilium. Points and error bars represent means ± SD.

Figure 4.. Pinhead and A-C linker in the proximal and mid-regions of the centriole.

(A, B) Projection view and (B) surface view (with surface capping) of the section through the microtubule triplet (MTT) map obtained from subtomogram averaging of complete MTTs in the proximal region (0–170 nm) viewed down the long axis of the MTT. (B, C) Surface rendering from (B) rotated 90° to reveal the side facing the lumen of the MT bundle. The pinheads are colored with depth-coding (magenta closer to MT, cyan furthest from the MT surface). (C, D) is a magnified and rotated version of the rendering in (C), highlighting the PinF1 and PinF2 feet of the pinhead and their 8.4-nm spacing. (E) Surface rendering of the centriole model in the proximal region of the mother centriole (0–170 nm) (from Fig 2B). (E, F) Magnified view of the boxed area in (E). The A-C linker is the density that links PF A8 on the A MT at the bottom to PF C9 of the adjacent C MT of the MTT at the top. (G) Side view from the outside of the MT bundle highlighting the A-C linkers. (F, H) Cutaway view (the cutting plane is shown as a dotted line in (F)) from the lumen of the MT bundle with yellow arrows highlighting the 8.4-nm spacing of the A-C linkers. (I, J, K) Mid-region (170–340 nm) of the compete centriole model showing the incomplete MTTs; the proximal 100 nm of the boxed region is shown at a higher magnification in (J, K). (J) is a tilted version showing the progression of structural changes along the centriole axis. (J, K) shows cross-sectional views of three different subregions (K1–K3) extracted at different levels from the map as shown in (J). Red arrows and arrowheads indicate the variations in the connections between A MT and C MT along the axis.

Figure S2.. Comparison of pinhead architectures of complete triplets from different species.

(A) Cross-sectional views of microtubule triplets from mouse photoreceptor, CHO cells (Li et al, 2019), Trichonympha spp. (generic term for a group of termite gut protists), and Trichonympha agilis (Nazarov et al, 2020). All the pinheads (dashed box) connect to the A3 protofilament. (A, B) Longitudinal views of the pinheads (arrows in (A)). The pinheads have similar orientations, with the two pin feet moieties PinF1 and PinF2 extending from A3 PF. In T. spp, the average spacing between PinF1 and PinF2 of adjacent pinheads (7.9 nm) differs from that between PinF1 and PinF2 of the same pinhead (8.6 nm), whereas for the other species, these two spacings are the same: 8.4 nm for WT mouse, 8.2 nm for CHO, and 8.4 nm for T. agilis.

Figure 5.. Inner scaffold along the incomplete triplets in the mid-region of the basal body.

(A) Section from raw cryo-electron tomogram of the rod photoreceptor showing the basal body structure in a longitudinal view with labels for the proximal, mid-region, and distal positions along the BB. The red arrows around the mid-region indicate the visible inner scaffold densities attached to the MTs. A zoomed-in view of inner scaffold densities attached to microtubule triplets (as indicated with a dashed red box) is shown to the right. (B) Corresponding line intensity profile (right) of the inner scaffold showing an average periodicity in the mid-region of 8.5 nm. (B, C, D) Cross sections of the BB mid-region map from subtomogram averaging in a projection view (B) or a surface view with surface capping (D). These reveal the complete A- and B-, and partial C-tubules. In panel (D), the extending filamentous densities from the lumen side of PF A2 (Arm A) and PF B9 (Arm B) are highlighted in orange. (C, D, E, F) Longitudinal views of the same region as in (C, D), with the inner scaffold shown in orange bracketed by red arrows in (E). (E) View in (F) is of the map shown in (E) rotated 90° toward the viewer. (F, G) Line intensity profile along the direction indicated by the double-headed arrow in (F) is shown in panel (G) (periodicity of 8.6 nm). (H) Cross-sectional view of a section of a raw tomogram in the BB mid-region with red arrows indicating the presence of inner scaffold (top) and a slightly tilted view of the corresponding region of the model from Fig 2B highlighting the inner scaffold indicated in orange (bottom).

Figure S3.. Visualization of the connections between microtubule triplets (by a A-C linker) and microtubule doublets (by an inner scaffold) in raw tomograms.

(A) Section through a raw tomogram generated by cryo-ET of centrioles showing the entire structure in a longitudinal view. Dashed lines indicate the positions of the cross sections shown to the right. (A, B, C) Arrows indicate the visible inner scaffolding in microtubule doublets (red arrows) and (A, B, C) the linker connecting the microtubule triplet in the proximal region (blue arrows).

Figure S4.. Visualization of the inner scaffold through the centriole and CC.

(A, B) Longitudinal section of a raw cryo-electron tomogram. Dashed lines indicate the positions of the cross sections in (B). (B, C) Cross sections from the three regions along the centriole and CC length, (B) before and (C) after ninefold symmetry averaging. Arrows show the inner scaffold partially attached to the doublets (red arrows) and the A-C linker within adjacent triplets (blue arrows).

Figure S5.. Comparison of inner scaffolds and Arm A–Arm B scaffolds in centrioles from different species.

(A, B, C) (Left) Cross-sectional slices from raw tomograms through the region at the beginning of the transition from complete to incomplete triplets in the (A) WT mouse photoreceptor (this work), (B) CHO cell (Greenan et al, 2018), and (C) Tetrahymena (Le Guennec et al, 2020). The regions outlined with dashed-line boxes are shown to the right rotated 90° to the right into the plane of the page, and show Arm A extending from the A/B junction. The corresponding intensity profiles (far right) display similar periodicity among the species (∼8.4 and 8.5 nm). (D, E) Different views (tilted cutaway views from the lumen of the MT bundle, and cross-sectional views of the models from mouse rod photoreceptors and P. tetraurelia (E)) are shown, with the incomplete microtubule triplet in blue (mouse rod, (D)) or violet (P. tetraurelia, (E)) and Arm A–Arm B in gold. (D) Stem density connecting the two is colored yellow in the P. tetraurelia map, but this feature is much weaker and not visualized at this contour level in the mouse rod (D). (D, E) Inside views from the centriole interior show the attached inner scaffold with spacing at ∼8.4 nm for mouse photoreceptor (D) and ∼8.5 nm for P. tetraurelia (E).

Figure 6.. Subtomogram averaging of the intramembranous beads/ciliary necklace along the ciliary membrane.

(A) Sections from raw tomograms from mouse rod photoreceptors displaying intramembranous beads periodically arranged on the ciliary plasma membrane (left displaying a slice from the outside view of the CC; right showing a slice through the center of the CC). The beads were categorized into two groups based on spacing, ciliary pocket region (orange arrows, ∼22-nm spacing), and CC region (purple arrows, ∼40 nm). (B) Spacing between beads for each of the two groups along the ciliary membrane (∼22 nm, n = 6 cells for ciliary pocket; and ∼40 nm, n = 7 cells for CC). The box-and-whisker plot shows the median (horizontal line), upper and lower quartile (box edges), and range of measurements (error bars), as well as individual points. (C) Conventional electron micrographs of mouse rod CC displaying the same intramembranous beads (average spacing 38.5 nm apart in this image, magenta arrows) along the CC. The dashed-line box in the image on the far left indicates the region shown at a higher magnification in the (middle) image immediately to its right. Dashed-line boxes on the image to the far right are the regions shown at a higher magnification below. (D) Map of the ciliary necklace beads obtained by subtomogram averaging was fit (in multiple copies) into a raw tomogram of a CC and the resulting map shown in a surface view from outside the CC (top, far left), in a projection view sliced through the CC membrane (immediately to the right of the surface view), and a magnified surface view (lower left). The series of images on the right show different high-magnification views of the average bead (cyan) and the portion of the CC membrane into which it was fit (gray). The longest dimension of the averaged structure is about 40 nm (magenta bars with inverted arrowheads), and the narrower dimension is about 10 nm (pink double-headed arrows). (E, F) Conventional transmission electron microscopy (TEM) images of CC showing the beads. (E, F) Dashed-line box in the upper image of (E) shows the region shown at higher magnifications below and in panel (F), to the right. The magenta bar with two inverted arrowheads represents 40 nm, and the cyan-outlined arrowheads indicate the ridge-like subdomains of the structure that yield the “bead” appearance. (D, G) is a similar view of a portion of a map from cryo-ET obtained as described in panel (D) above. (H) Plot of spacings measured for bead ridges in our TEM images and subtomogram averages, or in a previously published report of freeze-fracture/SEM studies. Whisker plots show medians, lower and upper quartiles, and range of the data, with n = 5 cells for both TEM and subtomogram average results.

Figure 7.. Y-link connections to microtubules.

(A) Left panel, a segmented version; and middle panel, a solid representation of the tomogram used to generate the template for the subtomogram average of doublets with attached bridges/links (right panel, magenta). BB, basal body, MTD, microtubule doublet. The MT-to-membrane bridges/links are indicated with orange arrows. The subtomogram average map on the right displays longitudinal spacings of 38 nm (double-headed arrow), also seen in the raw tomogram (orange arrows). The attached bridging density emerging from the A/B junction (A10/B1) is consistent with the observation in the CC longitudinal view (left and middle panels). (B) Rotationally averaged map of the CC shown in tilted (left) and cross-sectional (center) views; arrows point to Y-links. The averaged doublet + filamentous density map (magenta) and the averaged ciliary bead map (from Fig 6, cyan) were fit into the rotationally averaged map of the CC (gray mesh), providing a composite model of the MTD–Y-link–membrane complex (a square box in the center panel and an expanded view on the right). For comparison, in (C), a conventional transmission electron microscopy (TEM) image of a cross-sectional view of a detergent-extracted CC is shown at two different magnifications, with a magenta double-headed arrow displaying a distance of 41 nm. (B, D) MTD–Y-link ciliary model, aligned as in panel (B), is shown superimposed on the TEM image in (D). (B, E) View of the same map as in (B), with the cross section taken at a different position along the axis, and with the distance from the center color-coded. Three individual MTD–Y-link–membrane–bead complexes at different axial positions are shown to the right, with white arrows indicating the bridging/link density. (F) Different cross-sectional images from conventional TEM of CC, with red arrows pointing to the filamentous densities. (G) Schematic model of the MTD–Y-link–membrane–bead complexes in the CC cross section. (E, F) “Y” portion is meant to denote the approximate position, rather than the three-dimensional structure, given the heterogeneity demonstrated in (E, F).