Multivalent coiled-coil interactions enable full-scale centrosome assembly and strength

Abstract

The outermost layer of centrosomes, called pericentriolar material (PCM), organizes microtubules for mitotic spindle assembly. The molecular interactions that enable PCM to assemble and resist external forces are poorly understood. Here, we use crosslinking mass spectrometry (XL-MS) to analyze PLK-1-potentiated multimerization of SPD-5, the main PCM scaffold protein in C. elegans. In the unassembled state, SPD-5 exhibits numerous intramolecular crosslinks that are eliminated after phosphorylation by PLK-1. Thus, phosphorylation induces a structural opening of SPD-5 that primes it for assembly. Multimerization of SPD-5 is driven by interactions between multiple dispersed coiled-coil domains. Structural analyses of a phosphorylated region (PReM) in SPD-5 revealed a helical hairpin that dimerizes to form a tetrameric coiled-coil. Mutations within this structure and other interacting regions cause PCM assembly defects that are partly rescued by eliminating microtubule-mediated forces, revealing that PCM assembly and strength are interdependent. We propose that PCM size and strength emerge from specific, multivalent coiled-coil interactions between SPD-5 proteins.

Figure 1.

XL-MS reveals SPD-5 interactions during multimerization. (A) Centrosome architecture in a C. elegans embryo. (B) Secondary structural analysis of SPD-5 featuring predicted coiled-coil domains (MARCOIL 50%), alpha helices (Alphafold, >70% confidence), and disorder scores (IUPRED2 score in purple and ANCHOR score in magenta). (C) Phosphorylation sites on SPD-5 (+PLK-1[CA]) identified by MS. All sites are listed in Data S1. (D) Flow chart of the DMTMM crosslinking procedure. Samples were split into two at step 4 to enrich for monomers (4a) or multimers (4b) (see Materials and methods). Right panel: A representative gel showing samples incubated with or without crosslinker. SPD-5 that has exclusively intramolecular crosslinks migrate as monomers on a gel (SDS-PAGE monomers). Crosslinked SPD-5 complexes migrate slower (SDS-PAGE multimers). (E) XL-MS analysis of SPD-5 SDS-PAGE monomers incubated with kinase-dead PLK-1 (n = 6 replicates). Left panel: Linear map of crosslinks. Center panel: Location and occurrence (indicated by bubble size) of crosslinked pairs across replicates. Orange bars indicate predicted coiled-coil domains. Right panel: Percentage of crosslinked pairs involving predicted coiled-coil domains (CC) or other regions (nonCC). All crosslinked pairs are listed in Data S2. (F) XL-MS analysis of SPD-5 SDS-PAGE monomers incubated with constitutively active PLK-1. (G) XL-MS analysis of SPD-5 SDS-PAGE multimers incubated with constitutively active PLK-1. (H) Model for SPD-5 assembly. Unphosphorylated SPD-5 folds back on itself in multiple configurations, which prevents intermolecular interactions. Phosphorylation blocks intramolecular interactions, allowing coiled-coil domains to interact in trans. Interactions can be between the same (homotypic) or different coiled-coil domains (heterotypic). Source data are available for this figure: SourceData F1.

Figure S1.

Analysis of crosslinked residues in phosphorylated and unphosphorylated SPD-5. (A) Crosslinks were found in SPD-5 (+PLK-1[CA]) SDS-PAGE multimer samples and not in SPD-5(+PLK-1[CA]) SDS-PAGE monomer samples. (B) SPD-5 was incubated with kinase-dead PLK-1, then crosslinked using DMTMM. SPD-5 multimers were analyzed by mass spectrometry (n = 6 replicates). Left panel: Map of DMTMM-induced crosslinks. Center panel: Location and occurrence (indicated by bubble size) of crosslinked pairs across replicates. Orange bars indicate predicted coiled-coil domains. Right panel: Percentage of crosslinked pairs involving predicted coiled-coil domains (CC) or other regions (nonCC). (C) Quantification and location of crosslinked pairs in the pooled CA (red) or KD (blue) samples. Bubble size indicates number of replicates containing a given cross-linked pair. (D) The total number of times a residue was identified in a crosslinked pair was calculated for each condition. Data were normalized to account for overall differences in identified crosslinks. The difference between CA and KD samples is plotted. (E) The percentage of crosslinked pairs involving predicted coiled-coil domains (CC) or linker domains was determined for each condition. Shown is the difference in percentage points for each category between the CA and KD samples. (F) For each residue identified in a cross-linked pair, the number of unique partners was calculated. Data analyzed from multimer samples. (G) Histogram of sequence distances (in residues) between crosslinked pairs. Data were grouped into bins of 20 residues. Data analyzed from multimer samples.

Figure S2.

Biochemical and structural analysis of SPD-5(541–677). (A) Alphafold predicts that the central region of SPD-5 forms an alpha-helical hairpin. This motif contains Arginine 592 (blue) and PLK-1 phosphorylation sites (Serines 627, 653, and 658; magenta) that are critical for PCM assembly in C. elegans embryos. (B) Mass photometry of 50 nM SPD-5(541–677). (C) SPD-5(541–677) was incubated with PLK-1(KD) or PLK-1(CA) and analyzed by SDS-PAGE. The slower migrating bands (∼20 kDa) represent phosphorylated species of SPD-5. (D) Evaluation of 10,000 models built in Rosetta for SPD-5(541–677). (E) Predicted aligned error for the AlphaFold Multimer model of SPD-5(541–677). Source data are available for this figure: SourceData FS2.

Figure 2.

SPD-5 contains a central helical hairpin motif that homo-dimerizes. (A) SDS-PAGE gel showing purified SPD-5(a.a. 541–677). (B) Circular dichroism of SPD-5(541–677) reveals a strong alpha-helical signature. (C) 50 nM SPD-5 (541–677) was incubated with 20 nM kinase dead PLK-1 for 20 min and then analyzed by mass photometry. The inset lists the predicted monomeric masses of each protein. The calculated mass for each species is listed above the curves. (D) Mass photometry of SPD-5(541–677) incubated with active PLK-1. (E) Subsaturation crosslinking of SPD-5(541–677) reveals monomeric (M) and dimeric (D) species on an SDS-PAGE gel. (F) XL-MS of samples of SPD-5(541–677) incubated with PLK-1(KD). Crosslinks found in the monomers represent intramolecular interactions. Crosslinks unique to the dimeric state (red lines) are likely intermolecular. (G) XL-MS of samples of SPD-5(541–677) incubated with PLK-1(CA). (H) Best fit Rosetta model for dimeric SPD-5(541–677). PLK-1 phosphorylation sites are shown in magenta. Left, top view; right, en face view. (I) AlphaFold Multimer model for dimeric SPD-5(541–677). Left, top view; right, en face view. (J) Overlay of experimentally obtained crosslinks onto Rosetta and AlphaFold models. (K) Crosslink distances when mapped onto Rosetta and Alphafold models. Distance was measured between alpha carbons of linked residues (n = 18; mean ± 95% C.I.). UnP, links from the unphosphorylated sample (+PLK-1[KD]). Phos, links form the phosphorylated sample (+PLK-1[CA]). Source data are available for this figure: SourceData F2.

Figure 3.

R592 in SPD-5 is essential for full-scale PCM assembly and strength. (A) Wild-type (N2) or spd-5(or213 ts) worms were grown at 25°C for 16 h, then embryos were extracted and exposed to DMSO or 20 μM nocodazole for 2 min, then fixed. Embryos were permeabilized with perm-1(RNAi). SPD-5 was detected using immunofluorescence. Insets show zoomed-in images of centrosomes. (B) Quantification of fluorescence integrated density of SPD-5 signal localized at PCM. Mean ± 95% C.I.; N = 12–31 centrosomes; P value from a Kruskal–Wallis test. (C) MosSCI-generated transgenes (GFP::SPD-5, WT or R592K) were expressed from chromosome II and are re-encoded to be resistant to RNAi knockdown of endogenous spd-5. Worms were treated with RNAi against endogenous SPD-5 for 24 h, then embryos were excised and imaged by fluorescence confocal microscopy. Embryos were permeabilized using light pressure. Insets show zoomed in images of centrosomes. (D) Quantification of fluorescence integrated density of SPD-5 signal localized at PCM in C. Mean ± 95% C.I.; N = 36–40 centrosomes; P value from a Kruskal–Wallis test. (E) Surface-rendered model of the SPD-5(541–677) dimer. Key residues for one side are labeled. Inset: Diagram of SPD-5 structure surrounding R592; dashed yellow lines represent hydrogen bonds. (F) 50 nM of wild-type (WT) or mutant (R592K) SPD-5(541–677) was analyzed by mass photometry. The calculated molecular weights are indicated. (G) CD spectroscopy of WT and R592K SPD-5(541–677). (H) Thermal denaturation of WT and R592K SPD-5(541–677) was measured by CD spectroscopy at 222 nm. (I) 500 nM of purified full-length SPD-5(WT)::GFP or SPD-5(R592K)::GFP were incubated in 150 mM KCl + 7.5% PEG (MW: 3,300) for 10 min to form condensates, then imaged with fluorescence confocal microscopy. (J) Quantification of total condensate mass per field of view in I. Mean ± 95% C.I.; N = 11 images; P value from Mann–Whitney test. (K) 1,000 nM of SPD-5(WT)::RFP was assembled into condensates for 2 min, then 10 nM of the client (SPD-5(WT)::GFP or SPD-5(R592K)::GFP) was added, incubated for 8 min, and then imaged. (L) Quantification of client partitioning into preassembled WT condensates in K. Mean ± 95% C.I. (N = 4,702 (WT) and 1122 (R592K) condensates; P value from Welch’s t test.

Figure S3.

Extended analysis of SPD-5(R592K). (A) Western blots depicting expression levels of indicated proteins. Alpha tubulin was detected as a loading control. Worms were grown at 23°C. (B) Western blots showing depletion of endogenous SPD-5, but not transgenic GFP::SPD-5, following RNAi. Alpha tubulin was used as a loading control. (C) Viability of offspring in worms expressing RNA-resistant GFP::SPD-5 transgenes after depletion of endogenous SPD-5. Mean ± 95% C.I. (n = 11 worms). (D) The energetics of 10,000 ab initio models were calculated and then compared with the same model containing the R592K substitution. Each dot represents one model. Data on the diagonal indicate that the mutation does not change the energetics of folding in a particular model. (E) SPD-5(541–677) (WT or R592K) was incubated with different amounts of DMTMM crosslinker (0–10 mM) for 45 min, then analyzed by SDS-PAGE. Source data are available for this figure: SourceData FS3.

Figure 4.

Residues within the helical hairpin are essential for PCM assembly and strength. (A) Endogenous RFP-tagged SPD-5 was modified by CRISPR to delete residues 610–640. Embryos expressing the unmodified (WT) and mutant (Δ610–640) RFP::SPD-5 were visualized with fluorescence confocal microscopy. Embryos were permeabilized with perm-1(RNAi) to allow entry of DMSO or 20 μM nocodazole. (B) Quantification of fluorescence integrated density of RFP signal localized at PCM. Mean ± 95% C.I.; N = 18–22 embryos; P value from a Kruskal–Wallis test. (C) Quantification of PCM circularity. Mean ± 95% C.I.; N = 18–22 embryos; P value from a Kruskal–Wallis test. (D) PCM fragmentation from nuclear envelope breakdown onward in one-cell embryos. Mean ± 95% C.I.; N = 9–11 embryos.

Figure S4.

Extended analysis of SPD-5(Δ610-640). (A) Viability of offspring in worms expressing wild-type (WT) or mutant (Δ610–640) RFP::SPD-5. Mean ± 9 5% C.I. (N = 10 worms). (B) Western blots depicting expression levels of indicated proteins. Alpha tubulin was detected as a loading control. The blot for WT protein is the same used for Fig. S5 B. (C) Time-lapse confocal images of embryos during spindle assembly, spindle rocking, and PCM disassembly. Images were taken every 20 s. Source data are available for this figure: SourceData FS4.

Figure S5.

Control experiments for SPD-5(Δ734–918) mutant analysis. (A) Viability of embryos expressing either RFP::SPD-5(WT) or RFP::SPD-5(Δ734–918). Mean ± 96% C.I.; n = 10 worms per condition, 23–44 embryos each. (B) Western blot showing expression of RFP::SPD-5(WT) or RFP::SPD-5(Δ734–918). The blot for WT protein is the same as in Fig. S4 B. Alpha tubulin was used as a loading control. Source data are available for this figure: SourceData FS5.

Figure 5.

A long central coiled-coil motif in SPD-5 is essential for PCM assembly and strength. (A) Endogenous RFP-tagged SPD-5 was modified by CRISPR to delete residues 734–918. Embryos expressing the unmodified (WT) and mutant RFP::SPD-5 were visualized with fluorescence confocal microscopy. 20 μM nocodazole or 1% DMSO were introduced via laser puncture of the eggshell. (B) Quantification of mean intensity of RFP signal localized at PCM during metaphase. Mean ± 95% C.I.; N = 20 centrosomes; P value from a Kruskal–Wallis test. (C) Quantification of PCM circularity during metaphase in indicated strains. Mean ± 95% C.I.; N = 20 centrosomes; P value from a Kruskal–Wallis test. (D) Quantification of fluorescence integrated density of RFP signal localized at PCM during metaphase. Mean ± 95% C.I.; N = 20 centrosomes; P value from a Kruskal–Wallis test. (E) Embryos were treated with RNAi against csnk-1, which increases microtubule-mediated pulling forces. Insets show zoomed in images of centrosomes. (F) PCM fragmentation in one-cell embryos measured from nuclear envelope breakdown onward. Mean ± 95% C.I.; N = 10 embryos. Arrows indicate embryos were treated with csnk-1 RNAi.

Figure 6.

The coiled-coil-containing N-terminus of SPD-5 is required for PCM strength and assembly. (A) Diagram of SPD-5 domains and truncation mutants. Predicted coiled-coil domains (MARCOIL, 50%) are indicated in orange. g-TuC, gamma-tubulin complex. (B) Fluorescence confocal images of embryos expressing full-length (a.a. 1–1198; FL) or truncated (a.a. 566–1198) spd-5 transgenes after knockdown of endogenous SPD-5 by RNAi. (C) Quantification of fluorescence integrated density of GFP signal localized at PCM during metaphase. Mean ± 95% C.I.; N = 10 (FL) and 17 (566–1198) centrosomes; P value from a Mann-Whitney test. (D) Images of csnk-1(RNAi) embryos expressing endogenous spd-5 and transgenic gfp::spd-5. (E) Quantification of PCM circularity during metaphase in B and D. Arrows indicate embryos were treated with csnk-1 RNAi. Mean ± 95% C.I.; N = 16–22 centrosomes; P value from a Kruskal–Wallis test. (F) PCM fragmentation from nuclear envelope breakdown onward in one-cell embryos. Mean ± 95% C.I.; N = 6–10 embryos. Arrows indicate embryos treated with csnk-1 RNAi. (G) Images of one-cell and multicell embryos expressing full-length (FL) or truncated SPD-5 (a.a. 1–566). The number of multicell embryos displaying GFP-positive foci is indicated below. (H) Embryos expressing GFP::SPD-5(1–566) in the absence of endogenous SPD-5. Fluorescence image on the left, DIC image on the right. Cells are labeled based on developmental stage. (I) Immunofluorescence of transgenic embryos stained for α-tubulin and DNA. Embryos were lightly fixed to preserve the native GFP fluorescence. (J) Immunofluorescence of transgenic embryos stained for a centriolar marker (SAS-4) and α-tubulin. (K) Left: Overlap scores between SAS-4 and GFP::SPD-5 signal. Right: Overlap scores between α-tubulin and GFP::SPD-5 signal. Mean ± 95% C.I.; n = 12–36 GFP-positive foci.

Figure S6.

Control experiments for SPD-5 truncation analysis. (A) Viability of offspring after mothers were fed for 24 h on spd-5(RNAi) plates. No transgene (N2) compared with MosSCI worms expressing transgenic gfp::spd-5. Mean ± 95% C.I.; n = 10 worms per condition, 25–40 embryos each. (B) Western blot showing expression of gfp::spd-5 transgenes. Alpha tubulin was used as a loading control. (C) Fluorescence confocal image of GFP::SPD-5(566–1198) in an embryo expressing endogenous SPD-5. (D) Quantification of cytoplasmic fluorescence of transgenic GFP::SPD-5 proteins in one-cell embryos with and without endogenous SPD-5. Mean ± 95% C.I.; n = 7–13 embryos. (E) Quantification of cytoplasmic fluorescence of transgenic GFP::SPD-5 proteins in one-cell stage versus multicell stage embryos (8-cell or greater). Mean ±95% C.I.; n = 5–12 embryos). (F) Immunofluorescence of transgenic embryos grown on control or spd-5(RNAi) feeding plates. GFP signal was preserved by light fixation. Endogenous SPD-5 was detected by an antibody that targets its C-terminus. Source data are available for this figure: SourceData FS6.

Figure 7.

Multiple SPD-5 regions are sufficient to form micron-scale assemblies in vitro. (A) Domain maps of nine different SPD-5 variants. (B) Circular dichroism spectroscopy of purified SPD-5 fragments. (C) SPD-5 self-assembly assay. 500 nM of purified GFP-labeled SPD-5 proteins were incubated for 10 min in 7.5% (wt/vol) PEG-3350 then imaged. Mean ± 95% C.I.; n = 11–22 images per condition. (D) Comparison of in vitro condensate mass with coiled-coil content. Mean ± 95% C.I.; n = 11–22 images per condition. The data are fit with a quadratic model (R² = 0.78). (E) SPD-5 recruitment assay. 1 μM SPD-5::RFP was incubated in 7.5% PEG-3350 for 2 min to form condensates. 10 nM SPD-5::GFP proteins were added, incubated for 8 min, then imaged. Mean ± 95% C.I.; n = 875–4,702 condensates per condition. (F) Comparison of in vitro recruitment with coiled-coil content. Mean ± 95% C.I.; n = 875–4,702 condensates per condition. The data are fit with an exponential model (R² = 0.82).

Figure S7.

Analysis of SPD-5 assembly and recruitment in vitro. (A) Coomassie-stained SDS-PAGE gels of SPD-5 proteins used in this study. (B) Contrast adjusted sections from images in Fig. 7 B highlighting detected assemblies. (C) In vitro SPD-5 assembly using equivalent molar concentrations of coiled-coil mass (500 nM (FL), 1,140 nM (F20), 1,140 nM (F21), 3,315 nM (F22), 890 nM (F24), 1,675 nM (F25), 1,120 nM (F26), 3,390 nM (F27), 710 nM (F28)). Mean ± 95% C.I.; n = 11–22 images. (D) In vitro SPD-5 recruitment using equivalent molar concentrations of coiled-coil mass (1,000 nM SPD-5::RFP; for GFP proteins: 10 nM (FL), 23 nM (F20), 23 nM (F21), 66 nM (F22), 18 nM (F24), 36 nM (F25), 22 nM (F26), 68 nM (F27), 14 nM (F28)). Mean ± 95% C.I. n = 532–4,702 condensates. Source data are available for this figure: SourceData FS7.

Figure 8.

Model for SPD-5 scaffold assembly via multivalent coiled-coil interactions. (A) Domain map of SPD-5. CM1, γ-tubulin complex binding domain; Phos. Hotspot, a region containing essential PLK-1 sites; PReM, phospho-regulated multimerization domain; CL, centriole-localization domain; CM2-like, essential region for SPD-5 assembly. The N-terminal assembly region, Helical Hairpin, and long coiled-coil domain (CC-Long) are new motifs described in this study. (B) SPD-5 is an elongated protein with coiled-coil domains separated by disordered linkers. Multiple coiled-coil domains interact to form intermolecular connections that drive multimerization of SPD-5 into micron-scale assemblies. The linkers allow chain flexibility so that SPD-5 can sample multiple configurations. Phosphorylation of SPD-5 prevents autoinhibitory intramolecular folding and promiscuous intermolecular interactions between linker regions. Strong interactions connect alpha-helical hairpins to form the structural core of SPD-5. The cumulative action of these domains enables full-scale PCM assembly and strength in C. elegans embryos.