Authors: Marine Brunet, Joëlle Thomas, Jean-André Lapart, Léo Krüttli, Marine H Laporte, Maria Giovanna Riparbelli, Giuliano Callaini, Bénédicte Durand, Véronique Morel
Categories: Article, Centriole Duplication, Ana2, Plk4, Alms1, Alström Syndrome, Cell Adhesion, Polarity & Cytoskeleton, Cell Cycle
Source: The EMBO Journal
Authors: Marine Brunet, Joëlle Thomas, Jean-André Lapart, Léo Krüttli, Marine H Laporte, Maria Giovanna Riparbelli, Giuliano Callaini, Bénédicte Durand, Véronique Morel
Centrioles play a central role in cell division by recruiting pericentriolar material (PCM) to form the centrosome. Alterations in centriole number or function lead to various diseases including cancer or microcephaly. Centriole duplication is a highly conserved mechanism in eukaryotes. Here, we show that the two Drosophila orthologs of the Alström syndrome protein 1 (Alms1a and Alms1b) are unexpected novel players of centriole duplication in fly. Using Ultrastructure Expansion Microscopy, we reveal that Alms1a is a PCM protein that is loaded proximally on centrioles at the onset of procentriole formation, whereas Alms1b caps the base of mature centrioles. We demonstrate that chronic loss of Alms1 proteins (with RNA null alleles) affects PCM maturation, whereas their acute loss (in RNAi KD) completely disrupts procentriole formation before Sas-6 cartwheel assembly. We establish that Alms1 proteins are required for the amplification of the Plk4-Ana2 pool at the duplication site and the subsequent Sas-6 recruitment. Thus, Alms1 proteins are novel critical but highly buffered regulators of PCM and cartwheel assembly in flies.
Centrioles are highly conserved, small, cylindrical organelles composed of nine triplets of microtubules. They are surrounded by a complex array of proteins, the pericentriolar material (PCM). Together they form the centrosomes that serve as microtubule-organising centres (MTOC), playing a central role in cell division. Alterations of centriole number are associated with numerous pathologies, including cancers (Levine et al, 2017). Centriole duplication must therefore be tightly regulated and coordinated with the cell cycle (Nigg and Holland, 2018).
Non-dividing cells normally contain a pair of connected centrioles, an older, the mother, and a younger, the daughter, arranged perpendicularly. As the cell enters G1-phase, centrioles disconnect in a process called disengagement and initiate their duplication. A procentriole forms on the side of each centriole, grows to become a daughter centriole and recruits PCM. The two duplicated centrosomes hence formed organise the mitotic spindle and, ultimately as cells divide, each daughter cell inherits a single centrosome made of a mother and a daughter centriole.
Centriole duplication relies on a few sets of proteins that are highly conserved across species and which functions are highly hierarchised. It begins with the recruitment of the Plk4 kinase (respectively Sak in Drosophila and PLK4 in mammals) by the PCM proteins Asl (CEP152) and Spd2 (CEP192) at the proximal side of the centriole (Dzhindzhev et al, 2010; Kim et al, 2013; Sonnen et al, 2013). Plk4 initially forms a cylinder around the centriole and next resolves to a single dot which corresponds to the future site of procentriole assembly (Dzhindzhev et al, 2017; Yamamoto and Kitagawa, 2019). This reorganisation relies on the stabilisation of Plk4 by Ana2 (STIL) (Moyer et al, 2015; Ohta et al, 2014, 2018) and a complex cascade of trans autophosphorylations (Sillibourne and Bornens, 2010; Lopes et al, 2015; Park et al, 2019) leading to Plk4 activation. This initiates the sequential phosphorylation on multiple sites of Ana2 which can then recruit Sas-6 (SAS-6) (Dzhindzhev et al, 2017; McLamarrah et al, 2018; Moyer and Holland, 2019) to form the cylindrical ninefold symmetry array of the cartwheel. The latter serves as a template for Sas-4 (CPAP)-controlled deposition of microtubule triplets, initiating procentriole elongation (Kitagawa et al, 2011; van Breugel et al, 2011).
ALMS1 is a centrosome and cilium-associated protein characterised by a C-terminal ALMS domain involved in the localisation of the protein at centrioles and basal bodies (Hearn et al, 2005; Knorz et al, 2010). Mutations in ALMS1 are responsible for a rare ciliopathy, the Alström syndrome (Collin et al, 2002; Hearn et al, 2002), but how ALMS1 controls ciliary functions is still enigmatic. Two alms1 genes have been identified in Drosophila, alms1a and alms1b, which share more than 80% sequence homology. Alms1a was shown to localise at the centrosome in the Drosophila male germline and to be involved in centriole duplication in asymmetrically dividing male germline stem cells (GSCs), at the exclusion of other cell or division types (Chen and Yamashita, 2020).
Using Ultrastructure Expansion Microscopy (U-ExM), we characterised Alms1a and Alms1b localisation with unprecedented resolution. We show that Alms1a and Alms1b are sequentially recruited at the proximal end of centrioles and that Alms1a is also a component of the PCM. We observe that sudden and severe, hereafter referred to as acute, depletion of both Alms1a and b by RNAi is associated with complete centriole duplication failure in several Drosophila tissues, while chronic loss of alms1a (using an RNA null alms1a mutant) leads to reduced PCM recruitment and chronic loss of both alms1a and 1b (RNA null alleles) leads to premature centriole disengagement. Finally, we place Alms1a,b in the molecular hierarchy of centriole duplication. We show that they are involved in the efficient recruitment/stabilisation of Plk4 and Ana2 at the centriole duplication site and that the loss of Alms1a,b leads to the complete failure of Sas-6 recruitment. Collectively, our work demonstrates that Alms1a and b are key but highly buffered players in the initiation of cartwheel formation and PCM assembly in Drosophila.
To characterise the centriolar localisation of Alms1a and b proteins and their dynamics of recruitment during centriole duplication in Drosophila (Fig. 1), we adapted the U-ExM protocol (Gambarotto et al, 2019) to several Drosophila tissues (testes, larval brains and early embryos) expressing either Tomato-tagged Alms1a or GFP-tagged Alms1b (Appendix Fig. S1A).Figure 1Alms1a and Alms1b show different spatial-temporal localisations.U-ExM images of (A, B) the centriolar localisation of (A) Alms1b, as shown with the expression of the fusion protein Alms1b-GFP under alms1b native promotor in alms1b deletion background (magenta; alms1b^del2^, Alms1b-GFP) and (B) Alms1a, as shown with the expression of the fusion protein Alms1a-Tom under alms1a native promotor in alms1a deletion background (red; alms1a^del1^, Alms1a-Tom) at early stages of spermatogenesis (spermatogonia SG; spermatocytes SC). (A) SC#1: early SC, bottom view exposing the proximal ring of Alms1b on mother centrioles (MC). SC#2: late SC, Alms1b is also detected at the proximal end of daughter centrioles (DC) (white arrow). (B) SG#1, bottom view with the focus plane on the proximal end of the MC, revealing the two concentric localisations of Alms1a. SG#2, Alms1a is also detected around the DC. Centriolar walls are revealed with acetylated α-tubulin (green, ac α-tub). MC are shown on the left, DC on the right. MC-masked insets enable better visualisation of faint localisations. (C) Alms1a-Tomato (red) localisation at centrioles in larval brain neuroblasts (NBs) and ganglion mother cells (GMCs). NB#1, shows a NB in interphase. Insets: close-up of the centriolar pair, MC at the top and DC at the bottom. Alms1a (red) forms a ring superposed to the ring of Asterless (Asl, grey) and is less abundant at the DC compared to the MC. NB#2, shows a dividing NB. The two centrioles have separated and migrated on both sides of the cell. DC in inset 1, MC in 2. (D) Alms1a-Tomato (red) localisation at centrioles in early embryos. Left prophase. In the inset, the two centrioles have initiated their duplication as evidenced by the faint Alms1a concentration on the side of the centriole (white arrows) preceding the recruitment of Asl. Right metaphase with advanced centriole duplication as shown in the insets by the accumulation of both Alms1a and Asl on the side of the MCs. In these two stages, the cells are dividing, forming both asters and mitotic spindle. This explains the wide and intense Asl staining with respect to Alms1a. This contrasts with NBs, GSCs and SCs where cells are not in the process of forming a spindle and where Alms1a localisation is slightly wider than that of Asl. Asl (grey) labels the centriole, α-tubulin (yellow) the cytoskeleton and Hoechst (cyan) the nuclei. In all images, mother centrioles are marked with an asterisk. In all panels, the fluorescence intensities along the white dotted line are plotted as a function of distance to further document the width of Alms1a or b localisation with respect to centriolar makers acetylated α-tubulin or Asl (A–D) Brackets show the centriolar wall width of Mother (MC) or Daughter (DC) Centrioles. Scale bars (corrected from expansion factor): (A, B): 250 nm, (C, D): 1 μm, 250 nm.
Drosophila spermatogenesis starts with the asymmetric division of a germline stem cells (GSC) which gives rise to a GSC and a goniablast that initiates differentiation into a spermatogonium (SG) (Appendix Fig. S2A). SG undergo four symmetric divisions to generate a cyst of 16 cells, each containing two extremely small (minute) centrioles. As SGs complete a pre-meiotic S phase and duplicate their centrioles, they differentiate into spermatocytes (SCs) (Appendix Fig. S2A). In SCs, the 4 centrioles grow unusually long, reaching about 1.3 μm in length, and all centrioles mature into basal bodies nucleating a primary-like-cilium (Jana et al, 2016). SCs finally undergo two meiotic divisions, forming 4 spermatids each (resulting in 64 spermatids for each initial goniablast) with one basal body from which the sperm flagellum emanates (Demarco et al, 2014).
We observed that Alms1b-GFP is not detected at centrosomes in spermatogonia (Fig. 1A, SG) and is first recruited to the mother centriole (MC) in early spermatocytes, forming a ring which caps the proximal end of centriolar microtubules, while no Alms1b-GFP is observed on the daughter centriole (DC, Fig. 1A, early SC: SC#1). This strong asymmetry within the centrosome is maintained until the end of the SC stage (Fig. 1A, late SC: SC#2). As SC undergo the second meiotic division, both centrioles exhibit similar amounts of proximal Alms1b-GFP (Fig. EV1A), which suggests that its recruitment at daughter centrioles occurs rapidly at the very end of the SC stage or during meiotic divisions (Fig. 1A, late SC: SC#2, white arrow). In contrast, Alms1a-Tomato localises at both mother and daughter centrioles from the GSC stage onwards and forms a ring at the base of the centriole wall and a sleeve surrounding it (Fig. 1B). It is loaded on the nascent procentriole during each centriole duplication event (Fig. 1B, SG#1 and #2). Alms1a-Tomato then accumulates on the daughter centriole and at mid-late SC stages both mother and daughter centrioles present similar amounts of Alms1a-Tomato at their proximal end (Fig. 1B, SC). Together, these results indicate that Alms1a is associated with each centriolar duplication event in spermatogenesis, from asymmetrically dividing stem cells (GSCs) (Chen and Yamashita, 2020) to symmetrically dividing spermatogonia, whereas Alms1b is only associated with post-duplication mature centrioles. Similar dynamics of localisation were observed for endogenous Alms1a and Alms1b (Fig. EV1B) using previously published antibodies (Chen and Yamashita, 2020).
Neuroblasts (NBs) are somatic stem cells that generate the central nervous system of flies through asymmetric divisions (Januschke et al, 2011). In larval brain NBs, Alms1a-Tomato localises at both mother and daughter centrioles, labelled with the PCM protein Asterless (Asl, Fig. 1C), with a higher concentration of Alms1a-Tomato on the mother compared to the daughter centriole in early interphase (Fig. 1C, left panel NB#1, MC at the top, DC at the bottom). This asymmetry is still observed in pre-mitosis, after migration to the opposite pole of the NB of the centrioles, the mother centriole being inherited by the differentiating cell (ganglion mother cell, GMC) (Januschke et al, 2011) while the daughter centriole is maintained in the NB (Fig. 1C, right panel NB#2, DC: #1, MC: #2). In contrast to Alms1a, we were unable to detect Alms1b-GFP in the larval brain neither in NBs nor GMCs (Fig. EV1C), in agreement with RNA-seq data showing low expression of alms1b in neurons and glial cells (Li et al, 2022; Berger et al, 2012).
In syncytial embryos, which undergo 13 symmetric and synchronous mitosis (Foe and Alberts, 1983), Alms1a-Tomato is observed at both centrioles from the first mitosis onwards. As in the male germline, we detected a faint Alms1a-Tomato staining at the onset of procentriole assembly in prophase (Fig. 1D, embryo #1, white arrow), the signal becoming stronger on daughter centrioles at the beginning of the centriole-to-centrosome conversion in metaphase (Fig. 1D, embryo #2). In comparison, Alms1b-GFP was not detected in syncytial embryos and only observed at centrioles after cellularisation (Fig. EV1D).
Thus, despite their high degree of identity (>80%), Alms1a and Alms1b display very different spatiotemporal dynamics of centriolar recruitment in all tissues observed. Alms1a is observed on all centrioles from the onset of procentriole assembly, while Alms1b is only detected on mature centrioles.
As only one ALMS1 gene has been identified in humans, we wondered whether Alms1a or b was more similar to human ALMS1 by comparing their localisation profiles. Using U-ExM on cycling human retinal pigment epithelial cells (RPE-1), we observed that ALMS1 forms a ring capping the proximal end of both mother and daughter centrioles and slightly overlaps with the proximal segment of the centriolar wall (Fig. 2A,B). Careful examination of the dynamics of ALMS1 recruitment further reveals that ALMS1 is recruited after the onset of procentriole formation, as the procentriole reaches a length of 120 nm (Fig. 2C–E), which corresponds to the beginning of the procentriole elongation phase (Laporte et al, 2024). Thus, the localisation dynamics of human ALMS1 is more closely related to Alms1a profile, even though the fly Alms1a is recruited earlier during procentriole formation in the spermatic lineage than is human ALMS1 during procentriole elongation in RPE-1 cells.Figure 2Localisation of human ALMS1 during procentriole formation and at mature centriole.U-ExM images of hTERT-RPE-1 cells showing (A) side and (B) bottom views of a mature centriole (green) revealing ALMS1 (detected with an antibody directed against ALMS1, magenta) localisation as a ring capping the proximal extremity of the centriolar wall. (A, B) Graphs showing the position of ALMS1 with respect to the centriolar wall in longitudinal (A) and radial (B) dimensions. Data from N = 3 independent experiments, n = 72 centrioles (A) and n = 46 (B) ALMS1 107.6 nm, Tubulin 389.1 nm (A). ALMS1 240.1 nm, Tubulin 241.7 nm. Error bars correspond to s.d. (B) The fluorescence intensity along the white dotted line is plotted as a function of distance. Error bars correspond to s.d. (C) Dynamics of ALMS1 recruitment at forming procentrioles. (D, E) Quantifications of ALMS1 localisation with respect to the centriolar wall revealing the onset of ALMS1 recruitment as procentrioles reach 120 nm in length. (D) Size of procentrioles without (n = 39) or with (n = 54) ALMS1 capping the proximal side. (E) Evolution of the start and end position of ALMS1 signal relative to tubulin during procentriole growth. The light green and pink regions depict the centriole length (green) and region of the centriole which is covered by ALMS1 signal (pink). Scale bars (after expansion factor correction): 100 nm. Source data are available online for this figure.
Spatial analysis of Alms1a in U-ExM shows that Alms1a-Tomato colocalises with Asterless (Asl, CEP152) and Pericentrin-like protein (Plp, PCNT) (Fig. 3A,B), which are proposed to link centrioles and PCM and are hence called inner or bridge PCM proteins (Lattao et al, 2017; Varadarajan and Rusan, 2018). While Alms1a-Tomato fully surrounds the minute centrioles of SG (Fig. 1B), it remains restricted to the proximal third of the growing centriole of SC, a behaviour similar to Plp localisation (Figs. 1B, SC and 3B).Figure 3Alms1a is an inner PCM protein.(A, B) U-ExM images of Alms1a-Tomato (red, in alms1a^del1^;;Alms1a-Tom) relative localisation with (A) Asl (grey) and (B) Pericentrin-like protein (Plp, grey) in SC. The fluorescence intensity along the white dotted line is plotted as a function of distance. Bracket: centriolar wall width. Standard images of (C) Asl and (D) Plp in w^1118^ (control) or alms1a^del1^ in SG and SC and related fluorescence intensity quantification. In both conditions, Bld10-GFP (magenta) is expressed and used as an internal control for the quantification. Ten centrioles were quantified per stage and per testis. Here, we show two independent experiments that were quantified blind and pooled. For Asl: n = 70 centrioles per stage, 7 testes, alms1a^del1^: n = 80 centrioles per stage, 8 testes. For Plp: control n = 80 centrioles per stage, 8 testes, alms1a^del1^: n = 80 centrioles per stage, 8 testes. The box plots show the interquartile range (IQR), with the median (50th percentile) indicated by the horizontal line inside the box and the whiskers extending to the 10th and 90th percentiles. A two-sided unpaired Wilcoxon test was performed. (A, B) Scale bars (corrected from expansion factor): 250 nm; (C, D) scale 1 μm. Source data are available online for this figure.
To confirm a possible function of Alms1a in PCM assembly, we generated fly lines with an alms1a deletion by CRISPR/Cas9 (alms1a^del1^, Appendix Fig. S1) and quantified the fluorescence intensity of Asl and Plp at centrioles in control and alms1a^del1^ spermatogonia, early and late spermatocytes. We observed a significant decrease of both Asl and Plp fluorescence intensity in alms1a^del1^ cells at all stages (Figs. 3C,D and EV2A). Expression of Alms1a-Tom in alms1a^del1^ restored Asl intensity to normal, hence showing the specificity of the phenotype (Fig. EV2B). Altogether, these observations indicate that Alms1a is a novel inner PCM protein required to either efficiently recruit or stabilise the inner PCM proteins Asl and Plp at centrioles.
A previous study demonstrated that depletion of both alms1a and alms1b by RNAi in asymmetrically dividing GSCs results in a major loss of centrosomes and proposed, based on the absence of Alms1b in GSC, that Alms1a alone was involved (Chen and Yamashita, 2020). Here, despite the full deletion of alms1a locus in alms1a^del1^ mutant (an mRNA null allele, Appendix Fig. S1C), we failed to detect any centriole duplication loss (Fig. EV3A, middle). To exclude a possible compensation of alms1a loss by alms1b, we generated flies carrying a genomic deletion of both alms1a,b genes (alms1^del3^, mRNA null allele, see map and characterisation in Appendix Fig. S1). While centriole duplication in alms1^del3^ flies is largely normal (Fig. EV3B), we observed a premature disjunction of centrioles in 26% of centriole pairs of alms1^del3^ testes (Fig. EV3B middle and focus b, EV3C). This phenotype was fully rescued by expression of Alms1a-Tom (alms1^del3^, Alms1a-Tom), together validating the specificity of the alms1^del3^ CRISPR deletion and the functionality of our Alms1a-Tom transgene (Fig. EV3B, right, EV3C). In contrast, we confirmed that alms1a,b depletion by RNAi in the whole germline, using a nanos-Gal4 driver expressed in the GSCs and an RNAi targeting both alms1a and alms1b, results in the complete loss of centriole duplication (Fig. 4A, middle, only one or two single centrioles observed per testis) (Chen and Yamashita, 2020). In this condition as well, the introduction of extra copies of Alms1a with an Alms1a-Tomato transgene was sufficient to restore centriole duplication in 59% of observed testes (Fig. EV3D). The discrepancy between the two extreme phenotypes resulting from the acute alms1 depletion by RNAi or the chronic alms1 loss in alms1^del3^ flies could be due to either an off-target effect of the alms1^RNAi^ used or to compensation mechanisms triggered by the chronic loss of alms1a,b. To discriminate between these two possibilities, we performed alms1 depletion by RNAi in alms1^del3^ mutants (Fig. 4A). If off-target effect was to occur, alms1^RNAi^ should induce identical loss of centrioles in WT and alms1^del3^ backgrounds. However, expressing alms1^RNAi^ in alms1^del3^ flies leads to the same phenotype as alms1^del3^ flies, i.e. no centriolar loss (Fig. 4A, a compared to c), whereas expressing alms1^RNAi^ in a control background leads to complete centriolar loss (Fig. 4Ab). alms1^del3^ flies have thus developed compensatory mechanisms which alleviate the loss of Alms1 proteins, indicating that Alms1 proteins play a critical role in centriole biogenesis during male spermatogenesis which can be highly buffered during development.Figure 4The RNAi depletion of Alms1 proteins leads to centriole duplication failure in various cell lineages.(A) Unexpanded conventionally fixed testis expressing nos-Gal4>alms1^RNAi^ in alms1^del3^ background (c) characterised by numerous centrioles per testis, a phenotype identical to the one displayed by alms1^del3^ testis (a). In contrast, nos-Gal4>alms1^RNAi^ driven in w^1118^ (b) shows a sharp decrease in the centriole number, with only few single centrioles observed per testis. (B) Images of conventionally fixed bam-Gal4>lacZ^RNAi^ or alms1^RNAi^ SC (outlined with white dotted lines) expressing Ana1-GFP (magenta) as a centriolar marker. Nuclei (grey). The observed distribution of the number of centrioles per spermatocytes (4C, 2C, 1C or 0 C) in the indicated genotypes (lacZ^RNAi^: n = 36 cells, 3 testes; alms1^RNAi^: n = 89 cells, 5 testes). bam-Gal4>lacZ^RNAi^ and bam-Gal4>alms1^RNAi^ distributions are significantly different, two-sided unpaired Fisher exact P = 2 × 10^−16^. (C) U-ExM images of bam-Gal4>lacZ^RNAi^ or bam-Gal4>alms1^RNAi^ centrioles in SG and SC. Centriolar walls are revealed with acetylated α-tubulin (green, tub). Asl (magenta) as a marker of PCM. On each image, numbers indicate the occurrence of the phenotype (lacZ^RNAi^: n = 37 centrioles, 2 testes; alms1^RNAi^: n = 34 centrioles, 2 testes). alms1^RNAi^ KD results in a significant loss of centriole duplication compared to lacZ^RNAi^, two-sided unpaired Fisher exact P = 2 × 10^−14^. The fluorescence intensity along the white dotted line is plotted as a function of distance. (D) TEM images of bam-Gal4>lacZ^RNAi^ or bam-Gal4>alms1^RNAi^ mature primary spermatocytes confirming the absence of procentriole formation in alms1^RNAi^. Unduplicated centrioles show no structural defects in the centriole wall (bottom right image) and induce cilia formation (bottom left and middle images). (E) Unexpanded images of wor-Gal4>lacZ^RNAi^ or wor-Gal4>alms1^RNAi^ larval brain (lacZ^RNAi^: 45/45 observed NBs with centrosomes, 3 brains; alms1^RNAi^: 51/55 observed NBs lack centrosomes, 5 brains). alms1 RNAi KD results in a significant loss of centriole duplication compared to lacZ^RNAi^, two-sided unpaired Fisher exact P = 2 × 10^−16^. NBs (outlined with white dotted lines) are identified based on Miranda-GFP expression (yellow). Asl as a centriolar marker (magenta). (F) Unexpanded images of nos-Gal4>lacZ^RNAi^ or nos-Gal4>alms1^RNAi^ early embryos. β-tubulin for the mitotic spindle (magenta) and DNA (cyan). Scale (A, B, E, F) 5 μm, (C) (after expansion factor correction) 250 nm, (D) 5 μm on top and middle images and 200 nm on bottom images. Source data are available online for this figure.
To determine whether this function in centriole duplication is a general feature of Alms1 proteins, we extended our study to later stages of centriole duplication in the male germline and to the two other tissues studied above (Fig. 4B–F). We depleted both alms1a and b in these tissues by expressing the alms1^RNAi^ with the appropriate tissue-specific Gal4 drivers. We used bam-Gal4 (bag of marbles) (Chen and McKearin, 2003) to induce alms1^RNAi^ expression in symmetrically dividing spermatogonial cysts, between 4 and 16-cells stages (Demarco et al, 2014) (Fig. 4B–D; Appendix Fig. S2A). While control SC (bam-Gal4>lacZ^RNAi^) contain two centrosomes with one centriole pair each, for a total of 4 centrioles per cell (Fig. 4B: 4C, 100% of cells), bam-Gal4>alms1^RNAi^ cells contain only one centriole (Fig. 4B: 1C, 92.1% of cells) or no centriole (Fig. 4B: 0C, 6.7% of cells). This distribution is consistent with a centriole duplication failure event (i) all cells within 8-cells spermatogonial cysts (each giving 2 cells with one centriole after the 4th round of mitosis); and (ii) in some cells at the 4-cells stage (giving two cells without centrioles and two with one centriole after the last round of mitosis, Appendix Fig. S2B). More, using U-ExM, we confirmed that most of the centrioles still present in bam-Gal4>alms1^RNAi^ are mother centrioles that fail to assemble procentrioles as visualised with acetylated α-tubulin (94.1% of MC) (Fig. 4C). The remaining centriole pairs (5.9% of centrioles) likely arise from a duplication event at the SG stage, in cells where the alms1^RNAi^ was not activated yet. Together, these findings demonstrate that Alms1 proteins are crucial for centriole duplication in Drosophila male germline and that this role is not restricted to GSCs as previously proposed (Chen and Yamashita, 2020) but applies to all dividing germline cells. Strikingly, we observe that all the unduplicated centrioles present a normal distribution of Alms1a (Fig. EV4A,B), indicating that centriolar duplication events do not depend on the pool of Alms1a surrounding the mother centriole but more likely rely on a new pool of Alms1a recruited at procentrioles (Fig. 1B, SG). In agreement with these observations, unduplicated centrioles also show a normal ultrastructure by EM (Fig. 4D), while no procentriolar structures could be detected. Together, these observations indicate that Alms1 proteins are involved in the initiation of all centriole duplication events in the male germline.
We then used the worniu-Gal4 driver (wor-Gal4) (Lai et al, 2012) to express the alms1^RNAi^ in the larval brain. Control larval brains (wor-Gal4>lacZ^RNAi^) display centrosomes in GMCs and in NBs while wor-Gal4>alms1^RNAi^ brains show a strong reduction of GMCs with centrosomes and a total loss of centrosomes in NBs (Fig. 4E), consistent with a role of Alms1 proteins in centriole duplication during NBs asymmetric division. We also generated embryos depleted in maternally provided alms1a,b transcripts (from females expressing alms1^RNAi^ under the control of nanos-Gal4) (Van Doren et al, 1998) to investigate the contribution of Alms1 proteins in symmetric divisions in somatic tissues. While control syncytial embryos (from nanos-Gal4>lacZ^RNAi^ females) undergo symmetrical and synchronous nuclear division cycles, alms1^RNAi^-derived embryos exhibit strong mitosis defects including multipolar spindles characteristic of centriole duplication failure (Fig. 4F).
Collectively, our results thus demonstrate that Alms1 proteins are required for centriole duplication in both symmetrically and asymmetrically dividing cells of the soma or germline.
The formation of the procentriole is the result of a complex cascade of phosphorylations initiated by the recruitment, concentration and transactivation of Plk4 as a single dot on the side of the mother centriole (Lopes et al, 2015; Sonnen et al, 2012). Activated Plk4 phosphorylates Ana2 in two first on serine 38 (localised in the ANST motif at Ana2 N-terminus), allowing the efficient loading of Ana2 and the reinforcement of Plk4 localisation and activation at the duplication site (Dzhindzhev et al, 2017; McLamarrah et al, 2018; Moyer and Holland, 2019; Moyer et al, 2015; Ohta et al, 2014, 2018), then on the STAN motif, triggering the recruitment of Sas-6 by Ana2 (Dzhindzhev et al, 2014; McLamarrah et al, 2018; Moyer et al, 2015) and the formation of the cartwheel onto which the procentriole assembles (Fig. 5A).Figure 5Alms1 proteins are essential for Sas-6 recruitment and required to stabilise Ana2 and Plk4.(A) Scheme representing part of the known protein hierarchy responsible for centriole duplication, leading to procentriole formation. (B) U-ExM images of Sas-6-GFP (in magenta) in bam-Gal4>lacZ^RNAi^ in SG (a) before Sas-6 recruitment at the duplication site, (b) when Sas-6 is at the duplication site, (c) when procentriole has formed and in SC. (C) U-ExM images of Sas-6-GFP (in magenta) in bam-Gal4>alms1^RNAi^ in SG and SC. (B, C) Centriolar walls are revealed with acetylated α-tubulin (green, ac α-tub), (lacZ^RNAi^: 4 testes; alms1^RNAi^: 6 testes). (D) U-ExM images of Ana2-mNeonGreen knock-in (Ana2-eNG, in magenta) in bam-Gal4>lacZ^RNAi^ or bam-Gal4>alms1^RNAi^. Centrioles at the SG stage are divided into two SG short, when the length of the centrioles does not exceed their width and SG long, when their length exceeds their width (lacZ^RNAi^: 3 testes; alms1^RNAi^: 5 testes). The fluorescence intensity along the dotted line is plotted as a function of distance. Mother centriole (MC, on the left of all images) and daughter centriole (DC, on the right) are delineated by brackets, duplication site (ds) by an arrow. (E) U-ExM images of Plk4-meGFP knock-in (in magenta) in bam-Gal4>lacZ^RNAi^ or bam-Gal4>alms1^RNAi^ in SG (lacZ^RNAi^: 3 testes; alms1^RNAi^: 6 testes). Quantification of Plk4-meGFP fluorescence intensity performed on unexpended testis (lacZ^RNAi^: n = 4 testes; alms1^RNAi^: n = 5 testes) is shown on the right. The box plots show the interquartile range (IQR), with the median (50th percentile) indicated by the horizontal line inside the box and the whiskers extending to the 10th and 90th percentiles. Two-sided unpaired Wilcoxon test. In all images, the MC is on the left and marked with an asterisk. Scale bars (after expansion factor correction): 250 nm. Source data are available online for this figure.
In control testes, a dot of Sas-6-GFP is observed on the side of most mother centrioles initiating centriole duplication and subsequently in the proximal lumen of all growing procentrioles (Fig. 5B, U-ExM in bam-Gal4> lacZ^RNAi^). Sas-6-GFP is detected at the proximal end of the mother centriole in both alms1^RNAi^ and control conditions (U-ExM, Fig. 5C), in agreement with the conservation of the cartwheel after centriole elongation in Drosophila (Nigg and Holland, 2018). In contrast, we never observed Sas-6-GFP affixed to mother centrioles in late spermatogonia nor in spermatocytes in bam-Gal4>alms1^RNAi^ (Fig. 5C). Hence, in the absence of Alms1 proteins, the cartwheel of Sas-6 fails to assemble at the duplication site.
The lack of cartwheel in alms1^RNAi^ might reflect a role of Alms1 proteins either in the recruitment of Ana2 at the site of centriole duplication or in the recruitment of Sas-6 by Ana2.
We thus characterised Ana2 localisation in control and alms1^RNAi^ conditions by U-ExM using a Ana2-mNeonGreen knock-in line (Ana2-eNG) (Steinacker et al, 2022). In control SG, Ana2-eNG localises at the proximal inner end of the mother centriole and forms a dot at the base of the forming procentriole (Fig. 5Da,b) (Dzhindzhev et al, 2017; McLamarrah et al, 2020). As daughter centrioles elongate, we observed a shift in Ana2 abundance from the mother to the daughter centriole. Whereas in early duplicating centrioles Ana2 is more abundant at the mother centriole than at the nascent procentriole (Fig. 5Da, 55% of SG short centrioles), it becomes more abundant at daughter centrioles as they grow (all centrioles in SG long, Fig. 5Dc), to finally fully disappear from some mother centrioles (n = 8/32 SC, Fig. EV5A) (McLamarrah et al, 2018, 2020).
In alms1^RNAi^ condition, Ana2-eNG is also nested in the proximal end of all remaining mother centrioles in SG and SC. However, we failed to detect any Ana2-eNG on the side of 58% of the unduplicated centrioles (Fig. 5Dd,f for SG, SC see Fig. EV5B), whereas for the remaining 42% we observed a very faint Ana2-eNG staining at the centriole duplication site, almost always less abundant than the concentration at the base of the unduplicated centriole (Fig. 5De,g for SG, SC see Fig. EV5B). Altogether our observations thus suggest that Alms1 proteins are involved in the efficient recruitment or stabilisation of Ana2 at the duplication site and that this reduced Ana2 pool is responsible for the lack of Sas-6 recruitment.
Ana2 recruitment at the procentriole is under the direct control of Plk4 (Dzhindzhev et al, 2017). In wild-type conditions, Plk4 activity is tightly buffered to avoid the formation of supernumerary duplication sites and hence centriole overduplication. This control can be overcome by Plk4 overexpression (Dzhindzhev et al, 2017; Habedanck et al, 2005; Park et al, 2019). Hence, in Drosophila spermatogonia, increased expression of Plk4 leads to extra-centrioles organised in rosettes and de novo centriole formation (Lopes et al, 2015). In agreement with this latter result, Plk4 overexpression (Plk4^OE^) in our control background (bam-Gal4 > {lacZ^RNAi^, Plk4}) induces rosettes containing more than three centrioles (Fig. 6A, >3C). In contrast, when Plk4 is overexpressed in the absence of Alms1 proteins (Fig. 6A, bam-Gal4 > {alms1^RNAi^, Plk4}), the centrioles remain unduplicated except for an extremely low percentage of centriole pairs (less than 1%,). Thus, Alms1 proteins are required for Plk4-driven centriole overduplication. More, while overexpression of Alms1a alone doesn’t induce centriole overduplication, co-expression of Alms1a together with Plk4 enhances centriole overduplication compared to overexpression of Plk4 alone (respectively 76% and 57% of centrosomes with overduplicated centrioles, Fig. 6B,C). Altogether these results suggest a role of Alms1 proteins in either Plk4 activity or recruitment at duplication sites.Figure 6Alms1 proteins are required for Plk4 function in centriole duplication.(A) Images of unexpanded bam-Gal4>lacZ^RNAi^ or bam-Gal4>alms1^RNAi^ SC overexpressing Plk4 (Plk4^OE^, not tagged). Asl (grey) as a centriolar marker. The different phenotypes observed are shown by 1, 2, 3 centrioles or more (> 3) per centrosome (MTOC) (lacZ^RNAi^: n = 7 testes; alms1^RNAi^: n = 9 testes). (B) Quantification of centrosomes with or without centriole duplication in SC overexpressing Plk4 (Plk4^OE^-GFP), Alms1a (Alms1a^OE^-V5) or both (Alms1a^OE^-V5 +Plk4^OE^-GFP) (two independent experiments are pooled, Plk4^OE^-GFP: n = 734 centrioles, 5 testes; Alms1a^OE^-V5: n = 505 centrioles, 5 testes; Alms1a^OE^-V5 +Plk4^OE^-GFP: n = 1054 centrioles, 5 testes), Fischer’s exact test, ****P = 2 × 10^−16^. (C) Images of unexpanded SC overexpressing Plk4 (Plk4^OE^-GFP), Alms1a (Alms1a^OE^-V5) or both (Alms1a^OE^-V5 +Plk4^OE^-GFP). Percentage of centrosomes with 2, 3 or more (> 3) centrioles (revealed by Asl) is indicated on the images. The overexpressed proteins are revealed with their respective tag (V5 and GFP). (D) U-ExM images of bam-Gal4 > {lacZ^RNAi^, ND-Plk4-GFP} or bam-Gal4 > {alms1^RNAi^, ND-Plk4-GFP}. A close-up of the region with centrosomes (white arrow on the right panel) is presented in the inset and the phenotypes observed are indicated. Acetylated α-tubulin in green as a centriolar marker. (E) Proportion of centrosomes observed with 1 to >3 centrioles per indicated genotypes (lacZ^RNAi^: n = 133 centrosomes, 3 testes; alms1^RNAi^: n = 55 centrosomes, 4 testes). Fischer’s exact test, P = 2 × 10^−16^. (F) U-ExM image focusing on a centrosome with the predominant composition per genotype (> 3 centrioles for lacZ^RNAi^ and 1 centriole for alms1^RNAi^). On alms1^RNAi^, the white arrow points to ND-Plk4-GFP faint concentration on the side of the remaining centriole. ND-Plk4-GFP in magenta, acetylated α-tubulin in green. (G) Images of unexpanded control or alms1^del3^ SC overexpressing ND-Plk4-GFP (bam-Gal4>ND-Plk4-GFP). Overexpression of ND-Plk4-GFP in control background induces a strong overduplication of centrioles with centrosomes often containing more than 3 centrioles while alms1^del3^ SC overexpressing ND-Plk4-GFP present fewer overduplicated centrioles containing less often more than 3 centrioles. Centrioles labelled with anti Asl (magenta), centrosomes with 3 or more centrioles marked with an asterisk. Scale (A, C, G) 1 μm; after expansion factor correction (D) 1 μm, (D, inset, F): 250 nm. Source data are available online for this figure.
In spermatogonia depleted for Alms1 proteins, the endogenous spot of Plk4 (Plk4-eGFP) (Nabais et al, 2021) is still detected on the side of 91.7% of the unduplicated centrioles but with reduced concentration when compared to control condition (Fig. 5E; Plk4-eGFP staining intensity at ds/daughter centriole, SG: alms1^RNAi^ = 0.54 A.U., lacZ^RNAi^ = 1 A.U.; SC: alms1^RNAi^ = 0.67 A.U., lacZ^RNAi^ = 1.92 A.U.). Together, these results indicate that Alms1 proteins are not strictly required for the recruitment of Plk4 at the duplication site, but are necessary to reach a concentration of Plk4 sufficient to elicit duplication, either by enhancing its recruitment or by stabilising it once it is recruited.
Plk4 is a suicidal above a local threshold concentration, Plk4 trans-autophosphorylates two residues of its degron motif (Ser293 and Thr297), thus eliciting its own degradation by the SCF-Slimb/βTrCP-E3 ubiquitin ligase complex and the proteasome machinery (Cunha-Ferreira et al, 2013, 2009). We reasoned that if Alms1 proteins are involved in stabilising Plk4 protein, expression of a non-degradable form of Plk4 (with the point mutations S293A and T297A) (Cunha-Ferreira et al, 2009) could rescue the centriole duplication defects associated with alms1^RNAi^. In the control condition, overexpression of a non-degradable Plk4 (bam-Gal4 > {lacZ^RNAi^, Plk4^ND-GFP^}) induces numerous ectopic centrioles in SG, either organised in clusters or isolated, suggesting that we can achieve overduplication of centrioles (Fig. 6D,E) (Cunha-Ferreira et al, 2009). In contrast, when Plk4^ND-GFP^ is overexpressed in conditions where Alms1 is depleted (bam-Gal4 > {alms1^RNAi^, Plk4^ND-GFP^}), centrioles remain unduplicated (Fig. 6D,E). Interestingly, overexpression of ND-Plk4 in alms1a,b deleted flies (alms1^del3^) leads to less centrosomes with overduplication (three centrioles) than overexpression in control conditions in which numerous rosettes are observed (alms1^del3^, nos-Gal4 > ND-Plk4-GFP compared to nos-Gal4 > ND-Plk4-GFP, Fig. 6G). This indicates that despite the compensation of Alms1 loss operating in the chronic (RNA null) mutants, Alms1 remains required for over-amplification of centrioles in these overexpression assays.
In addition, in the control condition, Plk4^ND-GFP^ localises at the base of most centrioles contained within a rosette (Fig. 6F) and forms large and very bright aggregates in SG undergoing the 3^rd^ and 4^th^ mitosis and early SC as previously described (Cunha-Ferreira et al, 2013) (Fig. EV6A). In contrast, we only observed a faint Plk4^ND-GFP^ concentration on the side of the unduplicated centrioles in alms1^RNAi^ (Fig. 6F, white arrow), in agreement with the localisation observed for endogenous Plk4-eGFP (Fig. 5E), and failed to detect Plk4^ND-GFP^ cytoplasmic clustering (Fig. EV6A).
Altogether, these results suggest that Alms1 proteins are required to stabilise Plk4 clusters and that this function is independent of the degron-mediated regulation of Plk4 stability.
These experiments thus demonstrate that, in Drosophila, Alms1 proteins are required for Plk4 function in centriole duplication, in the regulation pathway that contributes to Plk4 recruitment or stabilisation at mother centrioles.
Stabilisation of Plk4 at the duplication site relies on a positive feedback loop between Plk4 and Ana2 which mutually contribute to their concentration and activation (Dzhindzhev et al, 2017; McLamarrah et al, 2018; Moyer and Holland, 2019; Moyer et al, 2015; Ohta et al, 2014, 2018). We thus investigated the possibility that Alms1 could stabilise Ana2 and hence indirectly Plk4. Ana2 overexpression (Ana2^OE^) in WT background (bam-Gal4>Ana2) results in mild overduplication of centrioles with 15% of centrosomes containing 3 centrioles. In contrast to what is observed for Plk4, overexpressing Alms1a with Ana2 does not potentialize the overduplication observed, but rather results in a small but significant reduction of centriole overduplication (Fig. 7A), suggesting that other factors are limiting. Considering the mutual reinforcement between Ana2 and Plk4, we reasoned that increasing both Plk4 and Ana2 could be sufficient to by-pass Alms1 requirement. Indeed, while overexpression of ND-Plk4 is not sufficient to rescue the centriole duplication defect associated with alms1 depletion (Fig. 7B,C, bam-Gal4 > {alms1^RNAi^, ND-Plk4}), and overexpression of Ana2 only rescues centriole duplication in few centrosomes (10%, Fig. 7B,C, bam-Gal4 > {alms1^RNAi^, Ana2}), overexpression of both Ana2 and ND-Plk4 largely rescues the centriole duplication defect associated with alms1 ^RNAi^, with 64% of the centrosomes composed of 2 centrioles (Fig. 7B,C, bam-Gal4 > {alms1^RNAi^, ND-Plk4, Ana2}).Figure 7Alms1 proteins are required for Plk4-Ana2 interaction.(A) Quantification of centriole number per centrosome in SC from testis overexpressing Ana2-GFP, Alms1a-V5 or both. Ana2^OE^ results in 13.6% of centrosomes with centriole duplication while 6.6% of centrosomes present centriole duplication upon co-overexpression of Alms1a^OE^-V5 and Ana2^OE^-GFP (Ana2^OE^-GFP: n = 509 centrosomes, 7 testes; Alms1a^OE^-V5: n = 277 centrosomes, 3 testes; Alms1a^OE^-V5 +Ana2^OE^-GFP: n = 652 centrosomes, 3 testes). Two-sided unpaired Fisher exact test. (B) Quantification of centriole number per centrosome in SC from testis bam-Gal4>alms1^RNAi^ or bam-Gal4>alms1^RNAi^ overexpressing ND-Plk4-GFP, Ana2-GFP or both (bam-Gal4>alms1^RNAi^: n = 303 centrosomes; +ND-Plk4-GFP: n = 552 centrosomes; +Ana2-GFP: n = 753 centrosomes; +{ND-Plk4-GFP, Ana2-GFP}: n = 1348 centrosomes). Two-sided unpaired Fisher exact test. (C) Images of unexpanded SC bam-Gal4>alms1^RNAi^ or bam-Gal4>alms1^RNAi^ overexpressing ND-Plk4-GFP, Ana2-GFP or both. While only unduplicated centrioles are observed in bam-Gal4>alms1^RNAi^ overexpressing ND-Plk4-GFP, 10% of centrosomes contain two centrioles upon Ana2-GFP overexpression while 64% of centrosomes are composed of doublet in bam-Gal4>alms1^RNAi^ overexpressing both Ana2-GFP and ND-Plk4-GFP. Related controls are shown in Fig. EV6B. Asl (magenta) as a centriolar marker. Scale 1 μm. (D) Model of Alms1 role in centriole duplication. We propose that Alms1 facilitates/potentializes the interaction between Plk4 and Ana2 at the onset of centriole duplication. By promoting the positive feedback loop between Ana2 and Plk4, Alms1 would hence allow for an efficient recruitment of Plk4 at the duplication site (ds), the concentration and activation of Ana2 and the subsequent formation of the cartwheel of Sas-6. In absence of Alms1, the amplification between Plk4 and Ana2 is not sufficient and Ana2 fails to recruit Sas-6. Source data are available online for this figure.
We therefore propose that, in Drosophila, Alms1 proteins contribute to centriole duplication by potentializing Plk4-Ana2 interaction required for the positive feedback loop (Fig. 7D), thus explaining how Alms1 can be both upstream of Plk4 (required for its localisation/stabilisation at the duplication site) and downstream of Plk4 (required for its activity in centriole duplication). Hence, Alms1 contributes to Plk4 recruitment or stabilisation at mother centrioles, increased Ana2 activation and subsequent cartwheel formation.
Here we show that, in Drosophila, Alms1 proteins are regulators of centriole duplication as the acute depletion of both alms1a and alms1b by RNAi results in complete failure of centriole duplication in all asymmetrically and symmetrically dividing somatic and germline cells analysed. Even though we cannot discriminate between Alms1a and b by RNAi due to the extreme conservation of their transcripts sequence, we anticipate that Alms1a is the one involved in the duplication process since we show during spermatogenesis that Alms1a is recruited at the onset of procentriole formation whereas Alms1b is recruited after the initiation of procentriole formation (Fig. 1), and centriole duplication is lost upon alms1 RNAi in early embryos or neuroblasts in which Alms1b is not expressed (Fig. 4E,F).
We propose that Alms1 proteins promote centriole duplication by enhancing Plk4-Ana2 functional interactions, thus contributing to the stabilisation of Plk4 at the centriole duplication site. In the absence of Alms1 proteins the concentration of both Plk4 and Ana2 at the duplication site is reduced, leading to dramatic loss of Sas-6 cartwheel assembly and of nascent procentriole formation.
It has been shown that centriole duplication at a single position relies on the functional interaction of Ana2 with the Plk4 ring initially formed around the centriole. The binding at a single position of Ana2/STIL to Plk4 (Dzhindzhev et al, 2017; Ohta et al, 2018) protects it from degradation by the ubiquitin–proteasome pathway. Ana2/STIL binding indeed prevents Plk4 from auto-phosphorylating its degron motif, which is required for its ubiquitination (Rogers et al, 2009), but also prevents Plk4 ubiquitination independently of the degron motif (Ohta et al, 2014). As a consequence, Plk4 is degraded in the whole ring with the exception of the single Plk4/Ana2 interaction site. This initial interaction of Ana2 to Plk4 also releases a positive reinforcement loop whereby Plk4 phosphorylation of Ana2 stabilises it and, in turn, phosphorylated Ana2 stabilises activated Plk4 (Dzhindzhev et al, 2017; McLamarrah et al, 2020; Moyer and Holland, 2019; Ohta et al, 2014). This loop results in the phosphorylation of Ana2 on its STAN domain, making it competent to recruit Sas-6 and trigger the formation of the cartwheel.
In contrast to isolated mammalian or Drosophila cells, we never observed the initial ring of Plk4 around the centriole in Drosophila tissues, but directly observed a single dot of Plk4 at the base of the procentriole from the very early steps of centriole duplication (Fig. 5E). After acute depletion of alms1 by RNAi, we observed that Plk4 is present at a reduced level on centrioles compared to control (Fig. 5E) and Ana2 is either absent in 58% of the centrioles or reduced in the remaining 42% (Fig. 5D), suggesting that the Ana2-Plk4 amplification loop is not set efficiently in absence of Alms1. Two hypotheses can support this in the absence of Alms1 Plk4 accumulation does not reach a concentration threshold required to switch on the amplification loop, or Alms1 is required to promote Plk4 and Ana2 interactions.
We observed, in agreement with the literature, that overexpression of wild-type or non-degradable (Cunha-Ferreira et al, 2009) Plk4 (with a non-phosphorylable degron motif, ND-Plk4) induces strong centriole overduplication, associated with centriole and cytoplasmic Plk4 accumulation in control flies. It however failed to restore Plk4 concentration and centriole duplication in flies depleted for Alms1 proteins (Fig. 6D–F). More, in absence of Alms1 proteins, we did not observe Plk4 accumulation in the cytoplasm as observed in control conditions (Fig. EV6A). These observations indicate that Alms1 proteins play a key role in stabilising PLK4 clusters independently of the phosphorylation of its degron motif. They thus suggest that Alms1 proteins are crucial for the formation of a dot of Plk4 sufficiently concentrated and /or stable to trigger the rest of the duplication pathway.
Stabilisation of Plk4-Ana2 has been proposed to rely on Sas-4/CPAP, which acts as a platform to bring the two proteins in proximity (Dzhindzhev et al, 2017; McLamarrah et al, 2020; Moyer and Holland, 2019). Interestingly, ALMS1 has been identified in BioID screens as a potential interactor of CPAP (Firat-Karalar et al, 2014; Gupta et al, 2015) and is a direct interactor of Plk4 (Chen and Yamashita, 2020). As well, Cep131/AZI1 was recently shown to be a substrate of Plk4 facilitating the Plk4-STIL interaction when phosphorylated (Kim et al, 2019). We speculated that, as proposed for Sas-4/CPAP or for CEP131/AZI1, Alms1 proteins could be involved in stabilising Plk4-Ana2 at the duplication site. In agreement with this hypothesis, while overexpression of Plk4 does not rescue the centriole duplication defect associated with alms1 depletion and overexpression of Ana2 in alms1^RNAi^ rescues centriole duplication for a few centrosomes, we observed a strong rescue of centriole duplication upon co-overexpression of Plk4 and Ana2 in alms1^RNAi^ background. We therefore propose that Alms1 contributes to the amplification loop by promoting Plk4 and Ana2 interactions. Future experiments will be required to position Alms1 proteins relative to the Sas-4-PLK4-Ana2 or the CEP131-PLK4-Ana2 modules.
Our work demonstrates that Alms1a,b are either essential (following acute RNAi depletion) or dispensable (upon chronic loss) intermediate players between Plk4 and Ana2. Such apparent contradictory properties illustrate striking buffering capacities of cells or tissues to maintain centriole numbers during development. This could also explain why Alms1 proteins have remained unidentified so far despite the extensive screens in various models, and in particular in C. elegans and Drosophila, that led to the identification of the set of conserved proteins at the core of the centriole duplication process (Plk4 (ZYG-1 in C.e.), Ana2 (SAS-5), Sas-6 (SAS-6) and Sas-4 (SAS-4)) (Bettencourt-Dias et al, 2005; Delattre et al, 2006; Dzhindzhev et al, 2014; Rodrigues-Martins et al, 2007; Shimanovskaya et al, 2014; Stevens et al, 2010). Several observations in the literature support the hypothesis that centriole duplication is very sensitive to variations in the expression of core molecular players. In particular, it has been shown that overexpression of centriole duplication factors can induce centrosome amplification and centriole overduplication. Thus, we anticipate that during chronic loss of Alms1a and b, the expression of some core players and/or other components still to be identified are modulated to fine-tune centriolar duplication. Investigations of the compensatory mechanisms involved in alms1 loss-of-function will likely provide fascinating future knowledge on centriole homoeostasis in cells. Interestingly, such compensatory mechanisms have already been described for other cellular processes in whole organisms (Hall et al, 2013) and thus represent a challenging complexity for their understanding.
The key role of Alms1 in centriole duplication in Drosophila that we uncovered in this work raises the question of the conservation of such function in other species and most specifically in humans, as mutations in ALMS1 are known to lead to Alström syndrome, a severe disorder classified as a ciliopathy (Collin et al, 2002; Hearn et al, 2002). The centriole disjunction observed after chronic loss of alms1a,b is also observed in human cells after RNAi depletion of ALMS1 (Knorz et al, 2010), thus indicating that at least this function of Alms1 proteins is likely conserved across evolution. In contrast, detailed examination of ALMS1 dynamics by U-ExM in RPE-1 cells revealed that it is not recruited at the initiation of procentriole assembly but rather appears at the beginning of the procentriole elongation phase, where it forms a ring capping the base of the microtubule wall (Fig. 2B,C). This observation, together with the fact that acute loss (RNAi KD) of ALMS1 is not sufficient to reveal centriolar duplication defects in mammalian cells (Li et al, 2007), suggests that ALMS1 role in centriole duplication in mammals might diverge from the one identified in Drosophila or else be shared with other centriolar proteins along which the two other proteins containing an ALMS domain, FATS and CEP295 (Knorz et al, 2010). No centriolar functions have been described for FATS to date. CEP295, on the other hand, is recruited at the onset of procentriole formation and progressively decorates the outer surface of the proximal third of the new centriole (Laporte et al, 2024). This localisation dynamics, together with the role of CEP295 in centriole stability and elongation (Izquierdo et al, 2014; Chang et al, 2016), led to propose that CEP295 is involved in the maintenance of the centriole integrity (Laporte et al, 2024; Meehl et al, 2016). Interestingly, CEP295 dynamics and localisation present similarities with the one we observed for Alms1a in Drosophila (Fig. 1B), questioning a role for Alms1a in daughter centriole stability and maturation in addition to its role in centriole duplication per se.
In conclusion, our work shows that Alms1 proteins are general regulators of centriole duplication in flies, involved in the complex step of Plk4-Ana2 stabilisation and amplification required to initiate cartwheel formation. It also reveals striking, but still to be understood, buffering capacities of cells and tissues during development to compensate for alms1 loss-of-function, highlighting their critical role in centriole duplication. Understanding if Alms1 function in centriole duplication is conserved in humans is a future challenge as ALMS1 is the only gene associated with the extremely rare Alström syndrome in humans (Collin et al, 2002; Hearn et al, 2002).
Flies were raised at 18 °C, 25 °C or 29 °C on a standard nutrient medium (cornmeal, yeast, agar, nipagin, ethanol).
All experiments involving flies conform to the relevant regulatory standards (DUO6403). Fly lines generated and stocks used are listed in Tables 1 and 2.Table 1Drosophila melanogaster stocks.Alm1a-Tomato in rescue conditionalms1a^del1^/FM7h; ; Alms1a-Tomato/TM3, SerThis studyalms1a^del1^alms1a^del1^/FM7hThis studyAlms1a-Tomato; ; Alms1a-Tomato/TM3, SerThis studyAlms1a^R^-Tomato; ; Alms1a^R^-Tomato/TM3, SerThis studyAlms1b-GFP in rescue conditionalms1b^del2^/FM7; Alms1b-GFP/CyOThis studyalms1b^del2^alms1b^del2^/FM7This studyAlms1b-GFP; Alms1b-GFP/CyOThis studyalms1^del3^alms1a^del1^, alms1b^del2^/FM7hThis studylacZ^RNAi^w^1118^; P{GD936}v51446VDRC 51446alms1^RNAi^; UAS-P{TRiP.HMJ30289}/CyOBL 63721bam-Gal4; If/CyO; bam-Gal4, UAS-Dcr2/TM6Bbam-Gal4bam-Gal4, UAS-Dcr2/FM7hnos-Gal4w^1118^ ; ; P{w[+mC]=GAL4::VP16-nanos.UTR}CG6325[MVD1]BL 4937wor-Gal4, Miranda-GFPw^^; P{w[+mC]=wor.GAL4.A}2, P{w[+mC]=UAS-mira.GFP}1.2/CyO ; Dr/TM6BModified from BL 56555Ana1-GFPw^^; Bl/CyO ; P{ana1-GFP.B}/TM6BBld10-GFP; UASp-endo-Bld10-GFP [142.1]/CyOFrom T. Megraw (Mottier-Pavie and Megraw, 2009)Plk4-eGFPw^^; If/CyO; meGFP-Plk4/TM6BFrom M. Bettencourt-Dias (Nabais et al, 2021)Ana2-eNGw^^; eAna2-mNG 3.72/SM5From J. Raff (Steinacker et al, 2022)Sas-6-GFPw^^; pUbq-GFP-Sas-6/CyOFrom R. BastoPlk4^OE^; UASp-Plk4/CyOFrom M. Bettencourt-DiasPlk4^ND-GFP^; UAS-ND-Plk4-GFP/CyOFrom M. Bettencourt-DiasAna2^OE^; ; UASp-Ana2-GFP/TM6BFrom J. RaffAna2^OE^; UASp-Ana2-GFP/CyOFrom J. RaffAlms1a^OE^; ; UASp-v5-Alms1a/TM3, SbThis studyTable 2Genotypes of Drosophila melanogaster flies analysed.FigureFlies as in textGenotypeFigure 1B–D Figure 3A,BAlms1a-Tomalms1a*^del1^/Y; ; Alms1a-TomFigure 1AAlms1b-GFPalms1b^del2^/Y; Alms1b-GFPFigure 3C,DControl expressing Bld10-GFPw^1118^; UASp-endo-Bld10-GFP [142.1]/+Figure 3C,Dalms1a ^del1^ expressing Bld10-GFPalms1a^del1^/Y; UASp-endo-Bld10-GFP [142.1]/+Figure 4Aalms1^del3^; nos-Gal4>alms1^RNAi^alms1^del3^ /Y ; UAS-P{TRiP.HMJ30289}/+P{w[+mC]=GAL4::VP16-nanos.UTR}CG6325[MVD1]/+Figure 4Anos-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/+P{w[+mC]=GAL4::VP16-nanos.UTR}CG6325[MVD1]/+Figure 4Aalms1^del3^; nos-Gal4alms1^del3^ /Y ; ; P{w[+mC]=GAL4::VP16-nanos.UTR}CG6325[MVD1]/+Figure 4Bmales bam-Gal4>lacZ^RNAi^expressing Ana1-GFPw/Y ; P{GD936}v51446/+P{ana1-GFP.B}/bam-Gal4, UAS-Dcr2Figure 4Bmales bam-Gal4>alms1^RNAi^expressing Ana1-GFPw/Y; UAS-P{TRiP.HMJ30289}/+P{ana1-GFP.B}/bam-Gal4, UAS-Dcr2Figure 4C,Dbam-Gal4>lacZ^RNAi^w/Y ; P{GD936}v51446/+ ;bam-Gal4, UAS-Dcr2/+Figure 4C,Dbam-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/+; bam-Gal4, UAS-Dcr2/+Figure 4Ewor-Gal4>lacZ^RNAi^w^^/Y; P{w[+mC]=wor.GAL4.A}2,P{w[+mC]=UAS-mira.GFP}1.2/P{GD936}v51446Figure 4Ewor*-Gal4>alms1^RNAi^w^^/Y; P{w[+mC]=wor.GAL4.A}2, P{w[+mC]=UAS-mira.GFP}1.2/UAS-P{TRiP.HMJ30289}Figure 4Fnos-Gal4>lacZ^RNAi^w/Y; P{GD936}v51446/+P{w[+mC]=GAL4::VP16-nanos.UTR}CG6325[MVD1]/+Figure 4Fnos-Gal4>alms1^RNAi^w/Y; UAS-P{TRiP.HMJ30289}/+P{w[+mC]=GAL4::VP16-nanos.UTR}CG6325[MVD1]/+Figure 5BSas-6-GFP in bam-Gal4>lacZ^RNAi^w/Y ; P{GD936}v51446/ pUbq-GFP-Sas-6;bam-Gal4, UAS-Dcr2/+Figure 5CSas-6-GFP in bam-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/ pUbq-GFP-Sas-6; bam-Gal4, UAS-Dcr2/+Figure 5DAna2-mNeonGreen knock-in bam-Gal4>lacZ^RNAi^w/Y ; P{GD936}v51446/ eAna2-mNG 3.72; bam-Gal4, UAS-Dcr2/+Figure 5DAna2-mNeonGreen knock-in bam-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/ eAna2-mNG 3.72; bam-Gal4, UAS-Dcr2/+Figure 5EPlk4-meGFP knock-in in bam-Gal4>lacZ^RNAi^w/Y ; P{GD936}v51446/ +; meGFP-Plk4/ bam-Gal4, UAS-Dcr2Figure 5EPlk4-meGFP knock-in bam-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/ +; meGFP-Plk4/ bam-Gal4, UAS-Dcr2Figure 6Abam*-Gal4 > {lacZ^RNAi^; Plk4 ^OE^}w/Y ; P{GD936}v51446/ UASp-Plk4; bam-Gal4, UAS-Dcr2/+Figure 6Abam-Gal4 > {alms1^RNAi^; Plk4 ^OE^}w/Y ; UAS-P{TRiP.HMJ30289}/ UASp-Plk4; bam-Gal4, UAS-Dcr2/+Figure 6B,Cbam-Gal4 > {Plk4 ^OE^-GFP}w/Y ; UASp-GFP-Plk4/(CyO or If); bam-Gal4, UAS-Dcr2/+Figure 6B,Cbam-Gal4 > {Alms1a ^OE^-V5}w/Y ; +/(CyO or If); bam-Gal4, UAS-Dcr2/UASp-v5-Alms1aFigure 6B,Cbam-Gal4 > {Plk4 ^OE^-GFP; Alms1a ^OE^-V5}}w/Y ; UASp-GFP-Plk4/(CyO or If); bam-Gal4, UAS-Dcr2/ UASp-v5-Alms1aFigure 6D,Fbam-Gal4 > {lacZ^RNAi^; ND-Plk4-GFP}w/Y ; P{GD936}v51446/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/+Figure 6D,Fbam-Gal4 > {alms1^RNAi^, ND-Plk4-GFP}w/Y ; UAS-P{TRiP.HMJ30289}/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/+Figure 6Galms1^del3;^ nos-Gal4>ND-Plk4-GFPalms1^del3^ /Y ; UAS-ND-Plk4-GFP/+; nos-Gal4/+Figure 6Gnos-Gal4> ND-Plk4-GFPw/Y ; UAS-ND-Plk4-GFP/+; nos-Gal4,/+Figure 7Abam-Gal4>Ana2-GFPw/Y ; UASp-Ana2-GFP/(CyO or If ;bam-Gal4, UAS-Dcr2/ +Figure 7Abam-Gal4>Alms1a-V5w/Y ; +/(CyO or If); bam-Gal4, UAS-Dcr2/UASp-v5-Alms1aFigure 7Abam-Gal4 > {Ana2-GFP; Alms1a ^OE^-V5}}w/Y ; UASp-Ana2-GFP /(CyO or If); bam-Gal4, UAS-Dcr2/ UASp-v5-Alms1aFigure 7B,Cbam-Gal4 > {alms1^RNAi^}w/Y ; UAS-P{TRiP.HMJ30289}/ (CyO or If); bam-Gal4, UAS-Dcr2/+Figure 7B,Cbam-Gal4 > {alms1^RNAi^, ND-Plk4-GFP}w/Y ; UAS-P{TRiP.HMJ30289}/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/+Figure 7B,Cbam-Gal4 > {alms1^RNAi^, Ana2-GFP}w/Y ; UAS-P{TRiP.HMJ30289}/(CyO or If); bam-Gal4, UAS-Dcr2/ UASp-Ana2-GFPFigure 7B,Cbam-Gal4 > {alms1^RNAi^, ND-Plk4-GFP, Ana2-GFP}w/Y ; UAS-P{TRiP.HMJ30289}/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/ UASp-Ana2-GFPFigure EV1AAlms1b-GFPalms1b^del2^/Y; Alms1b-GFPFigure EV1Bcontrolw^1118^/YFigure EV1CAlms1b-GFPalms1b^del2^/Y; Alms1b-GFP / P{w[+mC]=wor.GAL4.A}2,P{w[+mC]=UAS-mira.cherry}2Figure EV1DAlms1b-GFPalms1b^del2^/Y; Alms1b-GFPFigure EV2A,BControlw^1118^/YFigure EV2A,Balms1a ^del1^alms1a^del1^/YFigure EV2Balms1a ^del1^ ;;Alms1a-Tomalms1a^del1^/Y; ; Alms1a-TomFigure EV3Acontrolalms1b^del2^/Y; Alms1b-GFPFigure EV3Aalms1a ^del1^alms1a^del1^/YFigure EV3Aalms1b^del2^alms1b^del2^/YFigure EV3B,Ccontrolw^1118^/YFigure EV3B,Calms1^del3^alms1^del3^ /YFigure EV3B,Calms1^del3^;;,Alms1a-Tom (rescue)alms1^del3^ /Y ; ; Alms1a-TomFigure EV3Dnos>alms1^RNAi^, Alms1a^R^-Tomw/Y ; UAS-P{TRiP.HMJ30289}/(CyO or If); nos-Gal4 /Alms1a^R^-TomFigure EV4A,Bbam-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/ (CyO or If); bam-Gal4, UAS-Dcr2/+Figure EV4A,Bbam-Gal4>lacZ^RNAi^w/Y ; P{GD936}v51446/ (CyO or If); bam-Gal4, UAS-Dcr2/+Figure EV5AAna2-mNeonGreen knock-in bam-Gal4>lacZ^RNAi^w/Y ; P{GD936}v51446/ eAna2-mNG 3.72; bam-Gal4, UAS-Dcr2/+Figure EV5BAna2-mNeonGreen knock-in bam-Gal4>alms1^RNAi^w/Y ; UAS-P{TRiP.HMJ30289}/ eAna2-mNG 3.72; bam-Gal4, UAS-Dcr2/+Figure EV6Abam-Gal4 > {lacZ^RNAi^; ND-Plk4-GFP}w/Y ; P{GD936}v51446/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/+Figure EV6Abam-Gal4 > {alms1^RNAi^, ND-Plk4-GFP}w/Y ; UAS-P{TRiP.HMJ30289}/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/+Figure EV6Bbam-Gal4 > {lacZ^RNAi^}w/Y ; P{GD936}v51446/ (CyO or If); bam-Gal4, UAS-Dcr2/+Figure EV6Bbam-Gal4 > {lacZ^RNAi^, ND-Plk4-GFP}w/Y ; P{GD936}v51446/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/+Figure EV6Bbam-Gal4 > {lacZ^RNAi^, Ana2-GFP}w/Y ; P{GD936}v51446/(CyO or If); bam-Gal4, UAS-Dcr2/ UASp-Ana2-GFPFigure EV6Bbam-Gal4 > {lacZ^RNAi^, ND-Plk4-GFP, Ana2-GFP}w/Y ; P{GD936}v51446/ UAS-ND-Plk4-GFP; bam-Gal4, UAS-Dcr2/ UASp-Ana2-GFPThese flies result from crosses with stocks listed in Table 1.
All primer sequences are listed in Appendix Table S1. All transgenic constructs were injected by BestGene Inc.
Alms1a-Tomato was obtained by cloning the PCR product (F-Alms1a-Tom/R-Alms1a-Tom), containing 1.73 kb upstream regulatory sequences and the entire coding sequence (4.23 kb), in frame with Tomato in BglII-NotI sites of pJT108 (Vieillard et al, 2016).
For Alms1b-GFP construct, the PCR product (F-Alms1b-GFP/R-Alms1b-GFP) including 1.5 kb upstream regulatory sequences and the entire coding sequence (3.3 kb) was cloned in frame with GFP in the BglII site of pJT61 (Vieillard et al, 2016) using Gibson Assembly Master Mix (New England Biolabs Inc.).
Alms1a-Tomato was integrated in the 89E11 VK00027 landing site on the third chromosome and Alms1b-GFP on the 53B2 VK00018 landing site on the second chromosome by PhiC31 integrase (BestGene Inc.).
alms1a^del1^ allele was generated by CRISPR/Cas9-induced deletion (NHJE). Three couples of gRNAs (gRNA1Alms1a+gRNA2Alms1a, gRNA3Alms1a+gRNA4Alms1a and gRNA5Alms1a+gRNA6Alsm1a) were respectively cloned in pBFv-U6.2B vector (Kondo and Ueda, 2013) and injected together in CG12179[NP]/(FM7h);;vasa-Cas9 embryos. Flies were crossed to Hira^[-]^/FM7h females or FM7h/Y males and offspring were selected for white-eyed flies. Deletion in the alms1a locus was further characterised by PCR with primers F-Alms1aKO/R-Alms1aKO and confirmed by sequencing.
alms1b^del2^ allele was generated by CRISPR/Cas9-induced homologous direct repair (Gratz et al, 2015). The 1.5 kb 5’ homology arm and 1.4 kb 3’ homology arm were amplified by PCR (F-5’armAlms1b/R-5’armAlms1b and F-3’armAlms1b/R-3’armAlms1b) on genomic DNA from vasa-Cas9 flies and cloned, respectively, into the NheI-KpnI sites and BglII-AvrII sites of pJT38 plasmid (pRK2 plasmid (Huang et al, 2008) containing an attB cassette). Two gRNAs (gRNA1Alms1b and gRNA2Alms1b) were cloned in the pBFv-U6.2B vector (Kondo and Ueda, 2013). The two constructs were injected in vasa-Cas9 embryos. Flies were crossed to Hira^[-]/^FM7h females or FM7h/Y males and the offspring were screened for red-eyed flies. Homologous recombination was checked by PCR (F-5’Alms1bKO/R-5’Alms1bKO and F1-3’Alms1bKO/R-3’Alsm1bKO).
alms1^del3^ allele was generated from alms1a^del1^ flies by CRISPR/Cas9-induced homologous direct repair on the alms1b locus. Template repair vector was constructed as above, with the exception of the 5’ arm, which was obtained by PCR with primers F-5’armAlms1a^del1^/R-5’armAlms1b on genomic DNA from alms1a^del1^ flies. This new template repair was injected together with the previous pBFv-U6.EB vector expressing gRNA1Alms1b and gRNA2Alms1b in alms1a^del1^ embryo. Flies were crossed to Hira^[-]^/FM7h females or FM7h/Y males and the offspring were screened for red-eyed flies. Homologous recombination was checked by PCR (F2-5’Alms1bKO/R-5’Alms1bKO and F2-3’Alms1bKO/R-3’Alms1bKO).
alms1 deletions were validated by PCR on genomic DNA extracted from single hemizygous male (Appendix Fig. S1B) and on cDNA obtained by RT-PCR on RNA extracted from testes of hemizygous males (Appendix Fig. S1C). Primer couples specific of the control gene rbp49 (F-RBP49 and R-RBP49, 438 bp), alms1a (F2-Alms1a and R2-Alms1a, 345 bp) or alms1b (F3-Alms1b and R-3-Alms1b, 354 bp) were used.
Single males emerged from the day were crossed to 3 3-day-old w^1118^ females. Flies were left together at 25 °C for 5 days before being discarded. Total progeny was scored (Appendix Fig. S1D).
Testes from young adults were dissected in PBS and fixed 20 min in PFA 4%. For each fly, a single testis was kept to avoid analysing the same individual twice. After 20 min of permeabilisation in PBS-Triton X-100 0.1% (PBST), testes were blocked 1 h in PBST/BSA 2% (PBST-BSA) and then incubated with primary antibodies (in PBST-BSA, see Table 3) overnight at 4 °C. After four washes, testes were incubated 2 h with appropriate secondary antibodies and Hoechst. After four washes, testes were mounted in Vectashield Antifade Mounting Medium. If immunofluorescence was not required, testes were incubated 2 h in Hoechst in PBST.Table 3Antibodies used.Classical preparationU-ExMPrimary antibodiesAnti-acetylated α-TubulinMouse, IgG2bclone 6-11B-1, T6793, Sigma-Aldrich, lot 017M4806V–1:500Anti-RFPMouse, IgG2c6G6, ChromoTek, 6G6, lot 107272-07-02–1:100Anti-GFPRabbitTP401, Chemokine, lot 081211–1:100Anti-GFPRabbit632459, Living Colors, lot11050011:500–Anti-mNeonGreenRabbitCell Signaling Technology–1:100Anti-AsterlessRat/guinea pigfrom G. Rogers/ this study1:75,0001:5000Anti-PlpRabbitfrom G. Rogers1:20001:1000Anti-β-tubulinMouse, IgG1E7, Developmental Studies Hybridoma Bank1:1001:500Anti-α-tubulinMouse, IgG2aAA345, ABCD antibodies–1:250 (cells)Anti-β-tubulinMouse, IgG2aAA344, ABCD antibodies–1:250 (cells)Anti-ALMS1Rabbit27231-1-AP, Proteintech–1:500 (cells)Anti-Alms1aGuinea pigFrom Y. Yamashita (Chen and Yamashita, 2020)Anti-Alms1bGuinea pigFrom Y. Yamashita (Chen and Yamashita, 2020)Secondary antibodies****Anti-mouseAlexa fluor 488 goat anti-mouse IgG (H + L)A11001, lot 1810918, Invitrogen1:10001:1000Alexa fluor 488 F(ab’)2 donkey anti-mouse IgG (H + L)715-546-151, lot 152325, Jackson Lab1:10001:1000Alexa fluor 594 goat anti-mouse IgG (H + L)A11005, lot 1796406, Invitrogen1:10001:10001:625 (cells)Alexa fluor 594 goat anti-mouse IgG2bA21135, lot 898249, Invitrogen–1:1000Anti_rabbitAlexa fluor 488 goat anti-rabbit IgG (H + L)A11008, lot 1797971, Invitrogen–1:10001:625 (cells)Alexa fluor 488 F(ab’)2 goat anti-rabbit IgG (H + L)111-546-144, Lot 153510, Jackson Lab–1:1000Alexa fluor 594 F(ab’)2 goat anti-rabbit IgG (H + L)111-586-144, Lot 148526, Jackson Lab1:10001:1000Alexa fluor 647 goat anti-rabbitA21244, lot 1654324, Invitrogen1:10001:10001:625 (cell)Anti-ratAlexa fluor 488 goat anti-ratA11006, lot 940882, Invitrogen1:10001:1000Alexa fluor 647 F(ab’)2 donkey anti-rat IgG (H + L)712-606-153, Lot 150868, Jackson Lab1:10001:1000Unless specified, antibodies were used on Drosophila tissues.
Larva collection was performed as 20% sucrose solution was added in tubes containing larvae in food. The supernatant containing larvae was collected, rinsed in PBS and larvae were placed on a Petri dish with agar. Third-instar larvae were selected for dissection. Brains were dissected in PBS during sessions not exceeding 20 min and then fixed 25 min in PFA. After two baths of 10 min each in PBST 0.3%, brains were blocked 2 h in PBST-BSA 2% and then incubated with anti α-tubulin overnight at 4 °C. After four washes, brains were incubated with the appropriate secondary antibody and Hoechst overnight at 4 °C. Brains were last washed (4×) and mounted in Vectashield Antifade Mounting Medium.
In all, 30 min–2 h 30 synchronised embryos were collected on a petri dish with agar. Embryos were dechorionated for 5 min in bleach, devitellinised by vigorous shaking in heptane-methanol solution (1:1 ratio) and fixed in methanol. Embryos were blocked in two 30 min baths of PBST 0.1%-BSA 3% and incubated in anti β-tubulin overnight at 4 °C. After four washes, embryos were incubated with the appropriate secondary antibody and Hoechst for two hours at RT. Embryos were last washed 4x and mounted in Vectashield Antifade Mounting Medium.
w^1118^, alms1a^del1^ and alms1a^del1^;;Alms1a-Tom were prepared accordingly to the above protocols and incubated, for classical immunofluorescence with anti-Asterless or anti-Plp antibodies and for U-ExM with anti-acetylated α-tubulin and anti-Asterless antibodies. Samples of the different genotypes were processed identically and simultaneously. Adjustments of acquisition parameters were performed on one testis isolated from each genotype. Acquisitions were then made with the same exposure and laser parameters for all testes. Fluorescence quantification was performed after sample anonymisation by a third party. For each testis, Asl or Plp and Bld10-GFP (IF) or Asl and acetylated α-tubulin (U-ExM) fluorescence was quantified for 10 centrioles per stage (spermatogonia, early spermatocytes and late spermatocytes) using a threshold to delimit the area to be quantified and from which background noise was subtracted (macros available upon request).
bam-Gal4>lacZ^RNAi^ and bam-Gal4>alms1^RNAi^ testes expressing Plk4-meGFP were incubated with anti-Asterless antibody. Samples of the different genotypes were processed identically and simultaneously. Adjustments of acquisition parameters were performed on one testis isolated from each genotype. Acquisitions were then made with the same exposure and laser parameters for all testes. For each testis, Plk4 fluorescence was quantified for 10 centrioles per stage (spermatogonia and early spermatocytes) using a square surrounding the centriole to delimit the area to be quantified and from which background noise was subtracted (macros available upon request). (lacZ^RNAi^: n = 4 testes; alms1^RNAi^: n = 5 testes).
Testes from young adults were dissected in PBS and incubated PBS/1.4% Formaldehyde—0.3% Acrylamide (PBS-FA-AA) overnight at 18 °C. Prior to the gelation, testes were incubated 2 h in the monomer solution (PBS/19% Sodium Acrylate—10% Acrylamide—0.1% Bis-Acrylamide—0.5% TEMED – Hoechst 1/500) to allow reagents to penetrate the inner layers of the tissue. Testes were then deposited in packs of 3–4 on 12 mm coated coverslips (0.2 mg/mL Poly-d-Lysine during 2 h at 37 °C) and all the monomer solution was absorbed with filter paper. Coverslips were deposited on a silicone cushion made in 50-mL Falcon tube and spun down using a swing rotor for 5 min at 5000 rpm at 10 °C. In total, 38 μL of cold monomer solution with 0.5% APS was deposited on coverslips which were immediately put upside down on a parafilm tensed on a microscope slide laid on an ice-cold block. Gelation was started on cold for 5 min before transfer for 1 h at 37 °C. Gels were cut around groups of testes on the coverslip using a biopsy punch (4 mm diameter). Punches were incubated in denaturation buffer (SDS 200 mM—NaCl 200 mM—Tris pH 9 50 mM) for 1.5 h at 95 °C. After three washes in water, gels were preserved in PBS at 4 °C.
Gels were blocked 2 h in PBS/Tween 0.05%—BSA 1%—Sodium azide 0.02% (PBSTw-BSA) and then incubated with primary antibodies (in PBSTw-BSA) in a humid chamber overnight at 4 °C. After four washes, gels were incubated overnight with appropriate secondary antibodies and Hoechst and washed again. Gels were expanded for 2 h in two baths of water. Gels were mounted on 15 mm coated coverslips (0.2 mg/mL Poly-d-Lysine during 2 h at 37 °C).
Larvae were collected and selected as described above. Brains were dissected in PBS and immediately incubated in PBS-FA-AA. PBS-FA-AA solution was renewed before incubation overnight at 18 °C. Brains were then dissected again on 12-mm coated coverslips to remove the two optic lobes. The rest of the protocol is identical as described above.
Embryos were collected, dechorionated and devitellinised as described above. The rest of the protocol is identical to the testes one.
hTERT-RPE-1 cells (gift from A-M. Tassin; ATCC-CRL-4000) were grown on non-coated 12 mm diameter coverslips in DMEM-F12-Glutamax 10% SVF and penicillin/streptomycin 1% at 37 °C in 5% CO2. RPE-1 were seeded at 35,000 cell/cm^2^ for 24 h. Coverslips were rinsed in PBS 1× and incubated in 1 mL of a AA2% -FA1.4% in PBS, in a 12-well plate for 3 h at 37 °C. The coverslip was placed, cells facing down, on a drop of 35 μL monomer solution with APS deposited on a parafilm in a humid chamber placed on ice for 5 min before incubation at 37 °C for 30–60 min.
Gels were cut in 4–5 circles of using a biopsy punch (4 mm diameter) and incubated in denaturation buffer for 15 min at RT under agitation for the gel to detach from the coverslips. Gel pieces were transferred in 1.5-mL tube filled with denaturation buffer at 95 °C for 1.30 h. Gels were then washed two times 15 min in H2O and transferred in PBS 1×. Primary antibodies diluted in PBS-BSA 2% were incubated from 5 h to O/N at RT with agitation, washed 3 × 10 min at RT in PBS-Tween (0.1%) under agitation, incubated with secondary antibodies in PBS-BSA 3 to 5 h at RT and finally washed 3 × 10 min at RT in PBS-Tween (0.1%) under agitation. Expansion was performed as above before imaging. For centriole labelling on U-ExM cells, mouse IgG2a anti α-tubulin and anti β-tubulin were mixed.
Mother and daughter centrioles are identified based on their relative during centriole duplication, the daughter centriole is templated from the wall of the mother centriole. In U-ExM acquisition, when both centrioles retain their close, orthogonal organisation, the centriole which lumen faces the wall of the other centriole is identified as the daughter.
Expansion factor is calculated in each experiment by measuring the size of the gel disk obtained by standardised biopsy punch (4 mm diameter) after expansion. Centriole expansion isotropy and factor was validated by measuring centriole diameter (electron microscopy 200 nm ± 12 nm). Values presented in graphs and scale bars always correspond to “real” values after the application of the expansion factor.
All images (classical and U-ExM) were obtained using an IX 83 inverted microscope from Olympus, equipped with a Yokagawa CSU-X1 Spinning Disk Unit, Borealis technology for homogeneous illumination and Ixon3 888 EM-CCD camera from Andor. The oil immersion Plan Apochromat 60x/1.42 NA objective from Olympus was used for all acquisitions. All images were processed and analysed with FiJi (Schindelin et al, 2012).
For each image of human expanded centriole, a small z-projection (stack of 2–4 images maximum) was applied. Images were duplicated and rotated to orientate each centriole (pro or mature) in the proximal to distal orientation. A crop of 50 × 75 pixels was made around each centriole and images were sorted by tubulin length For bottom view-oriented centrioles, only single plane images are presented as a crop of 50 × 50 pixels.
Longitudinal and radial measurements from side viewed centrioles: Measurements of the length, diameter and relative position of ALMS1 were done on dual staining images where tubulin is always used as a proxy for the evaluation of the centriole length. From resized images where the pixel size was artificially decreased by 3 to improve the measurement precision (plugin “CropAndResize”, https://github.com/CentrioleLab), the fluorescent signal distribution of tubulin and ALMS1 were measured using the Fiji line scan (with a width covering the centriole for longitudinal measurements) and the plot profile tool. Using the plugin “PickCentrioleDim” (https://github.com/CentrioleLab) to facilitate the picking of the start and the end of the fluorescent signal (defined as 50% of the peak value at both extremities of the centriole), the user automatically generates an entry table with the coordinates of fluorescent signal extremities in each measured channel. For raw length and diameter measurements, the distance between the fluorescent signal extremities was calculated after application of the gel expansion factor and plotted using GraphPad Prism7. For the relative protein position, the tubulin was defined as the reference protein and its starting coordinate was shifted and set to 0. The same shift was applied to ALMS1. The entry table was read by a second plugin (“CentrioleGraph”, https://github.com/CentrioleLab) which generates a plot illustrating the relative average position of ALMS1 to the tubulin.
Appearance of ALMS1 during procentriole growth: The appearance ALMS1 was evaluated by measuring the length of the tubulin fluorescent signal in growing procentrioles. Measurements were sorted depending on whether the fluorescent signal of ALMS1 was present or not and plotted as two separate sets of data. The average between the highest centriole length without fluorescent signal and the lowest centriole length with fluorescent signal was considered as the appearance point of ALMS1.
Testes from pupae and 3–4 days old adults were dissected in phosphate-buffered saline (PBS), and fixed in 2.5% glutaraldehyde in PBS overnight at 4 °C. After rinsing for 30 min in PBS, the samples were post-fixed in 1% osmium tetroxide in PBS for 1 h. The samples were dehydrated in a graded series of ethanol and then infiltrated with a mixture of Epon–Araldite resin and polymerised at 60 °C for 48 h. Ultrathin sections (50–70 nm thick) were cut with a Reichert ultramicrotome equipped with a diamond knife. The sections were collected with formvar-coated copper slot grids and stained with 2% aqueous uranyl acetate for 20 min in the dark and then with lead citrate for 2 min. TEM preparations were observed with a Tecnai G2 Spirit EM (FEI Eindhoven, The Netherlands) equipped with a Morada CCD camera (Olympus, Tokyo, Japan).
Either non-parametric, two-sided, unpaired Wilcoxson or Fisher exact tests were performed using R or GraphPad Software. In all figures, the box plot shows the interquartile range (IQR), with the median (50th percentile) indicated by the horizontal line inside the box and the whiskers extending to the 10th and 90th percentiles.
Appendix Peer Review File Source data Fig. 2 Source data Fig. 3 Source data Fig. 4 Source data Fig. 5 Source data Fig. 6 Source data Fig. 7 Source data Fig. 6C EV and Appendix Figure Source Data Expanded View Figures