(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Molecular architecture of the C. elegans centriole [1]

['Alexander Woglar', 'Swiss Institute For Experimental Cancer Research', 'Isrec', 'School Of Life Sciences', 'Swiss Federal Institute Of Technology Lausanne', 'Epfl', 'Lausanne', 'Marie Pierron', 'Fabian Zacharias Schneider', 'Keshav Jha']

Date: 2022-09

Uncovering organizing principles of organelle assembly is a fundamental pursuit in the life sciences. Caenorhabditis elegans was key in identifying evolutionary conserved components governing assembly of the centriole organelle. However, localizing these components with high precision has been hampered by the minute size of the worm centriole, thus impeding understanding of underlying assembly mechanisms. Here, we used Ultrastructure Expansion coupled with STimulated Emission Depletion (U-Ex-STED) microscopy, as well as electron microscopy (EM) and electron tomography (ET), to decipher the molecular architecture of the worm centriole. Achieving an effective lateral resolution of approximately 14 nm, we localize centriolar and PeriCentriolar Material (PCM) components in a comprehensive manner with utmost spatial precision. We found that all 12 components analysed exhibit a ring-like distribution with distinct diameters and often with a 9-fold radial symmetry. Moreover, we uncovered that the procentriole assembles at a location on the centriole margin where SPD-2 and ZYG-1 also accumulate. Moreover, SAS-6 and SAS-5 were found to be present in the nascent procentriole, with SAS-4 and microtubules recruited thereafter. We registered U-Ex-STED and EM data using the radial array of microtubules, thus allowing us to map each centriolar and PCM protein to a specific ultrastructural compartment. Importantly, we discovered that SAS-6 and SAS-4 exhibit a radial symmetry that is offset relative to microtubules, leading to a chiral centriole ensemble. Furthermore, we established that the centriole is surrounded by a region from which ribosomes are excluded and to which SAS-7 localizes. Overall, our work uncovers the molecular architecture of the C. elegans centriole in unprecedented detail and establishes a comprehensive framework for understanding mechanisms of organelle biogenesis and function.

Here, we set out to map in a comprehensive manner and with utmost precision the distribution of centriolar as well as PCM core component in the gonad of C. elegans. Considering the very small size of the worm centriole, we combined U-Ex and STED, reaching an effective lateral resolution of approximately 14 nm. Using mainly endogenously tagged components and validated antibodies, we could thus determine with exquisite precision the localization of 12 centriolar and PCM core proteins. Of particular interest, this revealed that SAS-6 and SAS-4 exhibit an angular offset with respect to the microtubules, resulting in a chiral arrangement in the organelle center. Moreover, we acquired a large corresponding EM data set, which we overlaid with the U-Ex-STED images to map each centriolar protein to a specific ultrastructural compartment of the organelle. Overall, we uncovered the molecular architecture of the C. elegans centriole and provide an unprecedented framework for a mechanistic dissection of centriole assembly and function.

The molecular architecture of the centrioles has been investigated using 3D-Structured Illumination Microscopy (SIM) or STED super-resolution microscopy in other systems where the organelle is larger than in C. elegans, including human cells and Drosophila [ 32 – 34 ]. Moreover, ultrastructure expansion (U-Ex) microscopy has been utilized to investigate the molecular architecture of centrioles from human cells [ 35 , 36 ]. In this method, the sample is embedded in a gel that is then expanded isotropically several fold, thus likewise expanding the effective resolution [ 37 ]. SIM, STED, and U-Ex have enabled placing in a more refined manner a subset of components in the centriole map in these systems. However, the resolution achieved with these approaches would be likely insufficient to resolve the molecular architecture of the minute worm centriole.

HYLS-1 and SAS-1 are 2 additional C. elegans centriolar proteins that are dispensable for procentriole assembly. However, HYLS-1 is needed for generating nonmotile cilia [ 17 ], whereas SAS-1 is critical for maintaining the integrity of the organelle once formed [ 18 ]. In addition, the Polo-like kinase PLK-1 is present at centrioles in the early worm embryo [ 19 ]. As in other systems, C. elegans centrioles recruit the PeriCentriolar Material (PCM), thus forming the centrosome, which acts as a microtubule organizing center (reviewed in [ 20 ]). Assembly of the C. elegans PCM core, which has been defined as the set of PCM proteins that are also present in interphase [ 21 ], relies on the interacting proteins SPD-2 [ 22 , 23 ] and SPD-5 [ 24 ], as well as on SAS-7 [ 8 ] and PCDM-1 [ 21 ]. Furthermore, the γ-tubulin protein TBG-1 [ 25 , 26 ], together with the γ-tubulin interacting proteins GIP-1 and GIP-2 [ 26 ], as well as the γ-tubulin partner MZT-1 [ 27 ] are present in the worm PCM core. Additional proteins, including PLK-1 and AIR-1 [ 28 , 29 ], as well as TAC-1 and ZYG-9 [ 30 , 31 ], are recruited to this PCM core when the centriole matures in mitosis in the embryo, leading to increased microtubule nucleation. Despite the probably near-comprehensive list of component parts of the centriole and the PCM core in C. elegans, the very small dimensions of the worm organelle have thus far prevented localizing with precision where each component resides, thus limiting understanding of how they function.

SAS-6 is the main building block of a scaffold referred to as the cartwheel, which is thought to contribute to imparting the 9-fold radial symmetry of the organelle [ 14 , 15 ]. Whereas SAS-6 proteins in other systems self-assemble into ring-containing polymers that stack to form the cartwheel, structural and biophysical evidence obtained with the C. elegans protein has led to the suggestion that SAS-6 forms a steep spiral [ 16 ]. However, whether this is the case in vivo has not been addressed.

As in other systems, starting approximately at the onset of S phase, the 2 resident centrioles in C. elegans each seed the assembly of a procentriole in their vicinity, such that 4 centriolar units are present during mitosis, 2 per spindle pole. Comprehensive genetic and functional genomic screens conducted in C. elegans led to the discovery of 6 components essential for procentriole formation (reviewed in [ 10 – 12 ]). Molecular epistasis experiments uncovered the order in which proteins essential for procentriole formation are recruited to the worm organelle [ 7 , 13 ]. These experiments established that SAS-7 and SPD-2 (Cep192 in humans) are first recruited to the resident centriole. Thereafter, the kinase ZYG-1 (Plk4 in humans) directs the interacting coiled-coil proteins SAS-6 (HsSAS-6 in humans) and SAS-5 (STIL in humans) to the procentriole assembly site. This is followed by SAS-4 (CPAP in human) recruitment to the procentriole, a protein thought to enable the addition of microtubules to the SAS-6/SAS-5 scaffold. Relatives of SPD-2, ZYG-1, SAS-6, SAS-5, and SAS-4 in other systems are recruited in a similar sequence and exert analogous functions in procentriole formation (reviewed in [ 10 – 12 ]).

There are variations in the architectural features of centrioles in some systems, which are usually correlated with the absence or reduction of ciliary and flagellar motility (reviewed in [ 5 ]). For instance, in the nematode Caenorhabditis elegans, motile cilia and flagella are absent, and the sperm moves in an amoeboid fashion. Perhaps in the absence of evolutionary pressure for ciliary and flagellar motility, centrioles are smaller (approximately 175 nm high and approximately 120 nm wide) in the embryo [ 6 – 8 ] and comprise a radial arrangement of 9 microtubule singlets instead of the usual triplets and doublets [ 9 ]. Electron microscopy (EM) of centrioles in the C. elegans embryo revealed ultrastructural compartments besides microtubules, including 9 peripheral paddlewheels, as well as the central tube and, more centrally still, the inner tube [ 6 – 8 ]. EM analysis of embryonic centrioles also led to the notion that each paddlewheel is offset with respect to its accompanying microtubule, with a clockwise twist when viewed from the distal end, resulting in a chiral ensemble [ 8 ]. Whether chirality of the C. elegans centriole is apparent more centrally in the organelle, where the assembly process is thought to initiate, is not known.

Centrioles are membrane-less organelles that were present in the last common ancestor of eukaryotes (reviewed in [ 1 ]). In cells with flagella or cilia, centrioles act as basal bodies that template the formation of these structures. Moreover, in animal cells, centrioles form the core of the centrosome, which organizes microtubules and is thereby critical for fundamental cellular processes, including polarity and division (reviewed in [ 2 ]). In most organisms, centrioles are cylindrical organelles approximately 500 nm high and approximately 250 nm wide, with a 9-fold radially symmetric distribution of microtubules (reviewed in [ 3 , 4 ]). These centriolar microtubules are organized in triplets in the proximal region of the organelle and in doublets more distally. Triplet and doublet microtubules are twisted in a clockwise direction with respect to the microtubules when viewed from the distal end of the centriole, resulting in the characteristic chiral geometry of the organelle. This 9-fold radially symmetric architecture is also imparted onto the ciliary and flagellar axoneme that stem from centriolar microtubules and might be evolutionarily conserved because it provides an optimal geometry for axonemal motility. Despite important progress in recent years, the detailed molecular architecture of the centriole, including the root of its characteristic chirality, remains incompletely understood.

Results

Combining nuclei spreading and U-Ex microscopy for improved resolution of centrioles We set out to analyze the molecular architecture of the C. elegans centriole with utmost spatial resolution, using the adult hermaphrodite gonad as an experimental system (Fig 1A). The distal part of the syncytial gonad (the “mitotic zone” from here on) constitutes a stem cell pool where nuclei undergo cell cycles characterized by short G1 and M phases, with merely approximately 2% of nuclei being in one of these 2 phases combined [38]. Once nuclei have traveled far enough from the distal end of the gonad, they undergo premeiotic S phase and enter meiotic prophase I, a prolonged G2 phase during which meiotic recombination occurs. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Combining gonad nuclei spreading and U-Ex microscopy to analyze worm gonad centrioles. (A) Widefield imaging of ethanol-fixed worm expressing GFP::SAS-7. One layer of nuclei of the gonad is max intensity Z-projected (in this case, a height of 6.25 μm). White, grey, and red bold dashed lines indicate progression through the gonad from the mitotic zone to early and then late meiotic prophase; other white dashed line outlines the gonad. Grey box is magnified in (B). Yellow dashed regions mark 3 oocytes, purple dashed region the spermatheca. Note that centrioles are eliminated in oocytes, prior to fertilization. (B) (Left) Magnification of grey box region from (A). (Middle) Schematic representation of a single nucleus shown in the left and right panels. (Right) Early prophase region of a spread gonad from a worm expressing GFP::SAS-7. Note that spread nuclei are flattened and thus occupy a larger area compared to not spread nuclei. Note also that at this stage, centrioles do not act as microtubule organizing centers [55,56]. (C) Widefield imaging of centrioles in the early prophase region of the gonad from worms expressing GFP::SAS-7 before (left) and after (right) gel expansion. Grey mesh in the background represents the gel matrix. https://doi.org/10.1371/journal.pbio.3001784.g001 The gonad can be easily extruded from the animal and contains hundreds of nuclei, which are almost all in S or G2 phases of the cell cycle. Since procentriole formation begins in early S phase, most gonad nuclei harbor 2 pairs of centriole/procentriole, which are in close proximity to one another and cannot be resolved by immunofluorescence (IF) in widefield microscopy, usually appearing instead as a single focus (Fig 1B, left). We took 2 steps to improve the spatial resolution for our analysis. First, nuclei from extracted gonads were adhered as a single layer to a coverslip using mild chromatin spreading [39], resulting in superior detection by IF since the specimen is closer to the coverslip. Moreover, the pool of cytoplasmic proteins, which would otherwise contribute to poor signal-to-noise ratio of the centriolar signal, is largely washed out in this manner (Fig 1B, right). Second, we adapted previously validated ultrastructure gel expansion methods (U-Ex) [35,37], reaching approximately 5-fold isotropic expansion of the specimens (Fig 1C, Materials and methods). Combination of spreading with U-ExM enabled us to distinguish centriole and procentriole with widefield microscopy (S1A Fig), as well as to localize components to distinct regions within the C. elegans centriole (Fig 1C, right).

U-Ex-STED reveals consecutive ring-like distribution of C. elegans centriolar proteins We proceeded to comprehensively uncover the precise distribution of centriolar and PCM core proteins using U-Ex-STED. We used top views of centrioles to determine the radial distribution of these components (Fig 3A). Remarkably, except for ZYG-1 (see S1B Fig), such top views revealed that all components exhibit a ring-like distribution, with distinct diameters. To analyze the position of each component with respect to the others, we determined the diameter of each ring relative to that of α-tubulin, which was used as an invariant reference in this analysis (Fig 3B). To verify the validity of this approach, we costained α-tubulin with 2 different antibodies, finding that the 2 signals colocalize and that the corresponding rings hence exhibit the same diameter (without correction for the expansion factor: C-terminus 458 ± 38 nm, N-terminus 455 ± 35 nm, p = 0.81, N = 19; Fig 3A). Furthermore, an antibody raised against the middle portion of SAS-4 likewise had the same perimeter as α-tubulin, in line with the fact that the SAS-4-relative CPAP is a microtubule binding protein (Fig 3A and 3C, #9) [40–42]. Thus, α-tubulin and SAS-4 can be used interchangeably as invariant references in this analysis. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Relative position of centriolar components within the centriole. (A) U-Ex-STED of centrioles from early meiotic prophase stained for the indicated proteins. Each component (green) was imaged together with either α-tubulin (visualized with an antibody recognizing the C-terminus of the protein) or SAS-4 (visualized with an antibody raised against amino acid 350–517 of the protein) (both magenta). (B) Examples of fitted rings on fluorescent signal used to calculate the diameter of each component relative to α-tubulin or SAS-4 standards. In each image, the diameter of the centriolar component (in these cases RFP::SPD-2 (top) and SPD-5 (middle)) and that of the α-tubulin signal were measured along the dashed lines. The perimeter of the centriolar component was then divided by that of the α-tubulin signal. To obtain the theoretical diameter of the component before expansion, this value was normalized by the diameter of microtubules in EM images of centrioles (see Fig 5). (C) Calculated diameter of each centriolar component as determined in (B), arranged from the smallest to the largest. Magenta box highlights α-tubulin (#8) and SAS-4 (#9). Numbers in the graph indicate the identity of the component. Colors indicate whether the diameter is significantly different from 0 (red), 1 (blue), or 2 (green) neighboring values (Student two-tailed t test, significance p < 0.005). The middle lines of the boxplots correspond to the median, the cross represents the mean, the box includes 50% of values (IQR), and the whiskers show the range of values within 1.5*IQR. N = Flag::SAS-6: 21, SAS-5: 25, SAS-6: 22, SAS-1::FLAG: 10, SAS-4::GFP: 21, SAS-6::GFP: 15, FLAG::SAS-1 15, α-tubulin (N-ter): 19, SAS-4: 22, mCherrry::HYLS-1: 20, PCDM-1::GFP: 24, SPD-5: 20, GFP::PCDM-1: 21, SPD-2::GFP: 25, RFP::SPD-2: 20, GFP::SAS-7: 15, GFP::MZT-1: 20 and TBG-1: 20. Data underlying the graphs shown in the figure can be found in S1 Data. EM, electron microscopy; U-Ex-STED, Ultrastructure Expansion coupled with STimulated Emission Depletion. https://doi.org/10.1371/journal.pbio.3001784.g003 To estimate the ring diameter of each component in nonexpanded samples, we determined the diameter of the ring formed by the 9 microtubules in a novel EM data set of early meiotic prophase centrioles to be 87.9 ± 5.7 nm (N = 44; see below) and compared this value to the α-tubulin signal diameter determined with U-Ex-STED (Fig 3B). Moreover, we found that the α-tubulin diameter determined with U-Ex-STED following correction of the expansion factor (5.2) is similar to that measured for microtubules by EM (88 ± 8 nm; N = 38). This standardization method enabled us to estimate the actual diameter of the ring distribution of each protein, going from the smallest one, SAS-6[N], to the largest ones, SAS-7[N], MZT-1, and TBG-1 (Fig 3C). This analysis established that most components that were shown previously through biochemical and cell biological assays to physically interact are indeed located in close vicinity to one another. This is the case for SAS-6 and SAS-5 [43], SAS-4 and HYLS-1 [17], SAS-7 and SPD-2 [8,44], SPD-2 and SPD-5 [45], as well as PCMD-1 and SAS-4 or SPD-5 [46]. Overall, U-Ex-STED enabled us to localize in a comprehensive manner centriolar and PCM core component with unprecedented spatial precision.

Establishing the molecular architecture of the C. elegans centriole: Beyond microtubules To better understand the cellular context in which the centriole resides, we conducted tomographic analysis of the EM sections (ET), which revealed a ribosome free area approximately 262 ± 26 nm in diameter extending beyond the paddlewheels (Fig 6A; N = 3). This diameter is approximately 60 nm larger than that of the largest ring-like distribution observed in this work (see Fig 3C), raising the possibility that other proteins may be present in this area. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Overlay of EM and U-Ex images. (A) (Left) Max intensity Z-projection of ET of an early meiotic prophase centriole and surrounding region. (Middle) Magnification of the black box in the image on the left. (Right) Manually annotated ribosomes are shown in magenta and paddlewheel structures with dark-yellow outlines. Note that the ribosome-free area extends beyond the paddlewheels. Purple inset shows a magnified ribosome from the same ET image. (B, C, F, G) Overlay of U-Ex-STED and EM images (inverted grey levels) of centrioles from early meiotic prophase. Circularized images (left two panels), corresponding 9-fold symmetrized versions (next two panels), and magnification of the insets highlighted by the white box (very right). (B) Note that SAS-7 extends beyond the paddlewheel. (C) Overlay of paddlewheel components. (F) Overlay of components around microtubules. (G) Overlay of SAS-6 (N- and C-ter) and microtubules. (D, E) (Left) Magnification of a 9-fold symmetrized centriole imaged by EM (D) and highest populated class from class-averaging of particles containing microtubules and paddlewheels from individual ET tilt series of four centrioles (E) (see S4 Fig). Images are colorized with the LUT “Fire” (low intensities in blue, high intensities in magenta and red). Light green arrowheads point to the small density next to the paddlewheel (IPD), filled white arrowheads to that spanning from the central tube to 1 side of the microtubule (SCD). (Right) Intensity profiles were obtained along the indicated dashed lines (10 pixels wide). Microtubules display consistently more density on the side located under the paddlewheel (marked with an A) compared to the other side (marked with a B). EM, electron microscopy; ET, electron tomography; IPD, Inter Paddlewheel Density; SCD, SAS-6/4/1 Containing Density; U-Ex-STED, Ultrastructure Expansion coupled with STimulated Emission Depletion. https://doi.org/10.1371/journal.pbio.3001784.g006 We set out to determine the identity of the centriolar and PCM core proteins that correspond to given ultrastructural compartment of the organelle. To this end, we devised a method that relies on overlaying U-Ex-STED and EM images, using microtubules as a joint registration standard. In brief, we circularized, rotated, and size-adjusted jointly the 2 U-Ex-STED channel signals, aligning the α-tubulin signal with the microtubules in the EM images (S4 Fig). We applied this method initially on the symmetrized images and then likewise adjusted the raw data (S4 Fig). We report the results of this analysis hereafter, starting with the outside of the organelle. Overlaying the U-Ex-STED and EM data revealed that SAS-7[N] localizes just outside the paddlewheel, partially filling the region devoid of ribosomes surrounding the centriole (Fig 6B). Four components were found to localize to the paddlewheel: HYLS-1[N], SPD-2, SPD-5, and PCMD-1. SPD-5 and PCMD-1[C] are on the same angular axis as microtubules in the U-Ex-STED data set (see Fig 4D and 4E), and we indeed find PCMD-1[C] just outwards of microtubules in the overlay, constituting the base of the paddlewheel (Fig 6C). HYLS-1[N] also localizes to the base of the paddlewheel, but in contrast to SPD-5 and PCMD-1[C], it does so with an offset with respect to the microtubules (Fig 6C). SPD-2 is the outermost component of the paddlewheel with the 2 ends showing distinct distributions: SPD-2[C] appears as foci positioned just outside of microtubules, with an angular offset with respect to them (Fig 6C; see also Fig 4F), whereas SPD-2[N] localizes slightly further to the outside as an epitrochoid with 9 lobes extending left and right over the paddlewheel (Fig 6C). Interestingly, we detected a previously unnoticed small electron-dense region in the EM and ET data sets (see below) located between neighboring paddlewheels (Fig 6D and 6E, green arrowheads), which can be partially matched with the position of SPD-2[N] in these overlays. We name this density Inter Paddlewheel Density (IPD). Overall, this analysis reveals in exquisite detail the molecular architecture of components located outside the centriolar microtubules.

Molecular architecture at the level of the microtubules We next report the analysis of components located more centrally. Upon careful analysis of the symmetrized EM data set, we noticed another novel density, which starts from the central tube (Fig 6D, dashed arrowhead), extends towards and along each microtubule, rendering 1 side of the microtubule more pronounced than the other (Fig 6D, white arrowhead). This density displays the same angular offset with respect to the microtubules as the paddlewheel. Since microtubules are not always perfectly perpendicular to the plane of sectioning, we performed ET to obtain bona fide top views of microtubules and thus better analyze this novel density. From individual tilt series of 4 centrioles, we picked 628 particles containing microtubules and paddlewheels; class-averaging resulted in 3 well-defined classes containing 92% of input particles (S5A Fig). In all 3 classes, the novel density is present on the side of the microtubule above which the paddlewheel is located (Figs 6E and S5B). Given that SAS-6, SAS-4, and SAS-1 all display the same angular offset direction with respect to microtubules as the paddlewheel component SPD-2[C] (see Fig 4F and 4H–4L), we propose that these 3 proteins together could compose this novel offset density. Therefore, we name this novel density “SAS-6/4/1 Containing Density” (SCD). Overlays of the corresponding U-Ex-STED and EM images indeed revealed perfect alignment of SAS-4, SAS-6[C], and SAS-1[N] with the SCD, below 1 side of the microtubule (Figs 6F, 6G, and S5C). Moreover, SAS-4[C] overlaps almost perfectly with the SCD at the level of the central tube, whereas SAS-1[N] has an indistinguishable diameter from SAS-6[C] (see Fig 3C). Taken together, our data suggest that SAS-4, SAS-6, and SAS-1 form the newly described chiral SCD, with SAS-4 potentially bridging it to HYLS-1.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001784

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.