Mouse SAS-6 is required for centriole formation in embryos and integrity in embryonic stem cells

SAS-6 (SASS6) is essential for centriole formation in human cells and other organisms but its functions in the mouse are unclear. Here, we report that Sass6-mutant mouse embryos lack centrioles, activate the mitotic surveillance cell death pathway, and arrest at mid-gestation. In contrast, SAS-6 is not required for centriole formation in mouse embryonic stem cells (mESCs), but is essential to maintain centriole architecture. Of note, centrioles appeared after just one day of culture of Sass6-mutant blastocysts, from which mESCs are derived. Conversely, the number of cells with centrosomes is drastically decreased upon the exit from a mESC pluripotent state. At the mechanistic level, the activity of the master kinase in centriole formation, PLK4, associated with increased centriolar and centrosomal protein levels, endow mESCs with the robustness in using a SAS-6-independent centriole-biogenesis pathway. Collectively, our data suggest a differential requirement for mouse SAS-6 in centriole formation or integrity depending on PLK4 activity and centrosome composition.


Introduction
Proliferating cells rely on stringent controls of cell division fidelity, primarily through the proper assembly, organization, and polarity of the microtubule-based mitotic spindle.In most animal cells, these functions are ensured by centrosomes, the major microtubule organizing centers (MTOCs).At the core of centrosomes are the microtubule-based centrioles that are highly conserved in evolution (Bornens, 2012;Nabais et al., 2020).During interphase or upon differentiation, centrioles provide the essential template to form cilia (Conduit et al., 2015).Cycling cells in G1 have two centrioles whose duplication is controlled by the centriole formation machinery, with the master kinase Polo-Like Kinase 4 (PLK4) regulating the early initiating steps (Bettencourt-Dias et al., 2005;Habedanck et al., 2005).Once per cycle, procentrioles assemble on the existing centrioles and form new daughter Table 1.Description of CRISPR/Cas9-mediated knockouts of Sass6 in the mouse in vivo.

Results
Mutation in mouse Sass6 leads to embryonic arrest around midgestation To determine the functions of mouse SAS-6 in vivo, we used CRISPR/Cas9 to generate Sass6 knockout mice by targeting exon 4 (Materials and methods, Table 1).The resulting Sass6 mutant allele (Sass6 em4/ em4 ) had a frameshifting deletion, which is predicted to lead to a premature stop codon (Table 1).Sass6 em4/em4 embryos arrested development ~E9.5, when they still formed a heart but did not show somites or undergo embryonic turning that are typical in wild-type (WT) embryos (Figure 1A).The phenotype of Sass6 em4/em4 embryos resembled our previously reported Cenpj −/− embryos without centrioles (Bazzi and Anderson, 2014a), suggesting a crucial role for SAS-6 in centriole formation and mouse development.
The loss of SAS-6 activates the 53BP1-USP28-p53 mitotic surveillance pathway In order to assess whether the loss of SAS-6 leads to p53 upregulation and cell death, as in Cenpj −/− mutants (Bazzi and Anderson, 2014a), we performed immunostaining for p53 and active cleavedcaspase 3 (Cl.CASP3) on sections from WT and Sass6 em4/em4 embryos at E8.5 (Figure 1B and C).
For higher resolution analyses of centriole formation in Sass6 mutant embryos, we utilized Ultrastructure-Expansion Microscopy (U-ExM), a technique that relies on isometrically expanding the sample ~4 times and has been recently widely implemented for centriole analyses (Gambarotto et al., 2019).We combined U-ExM with immunostaining for the centriolar wall marker, acetylated tubulin (Ac-TUB) (Figure 2G).We observed that in WT embryo sections (E9.0), each pole of a mitotic spindle contained a pair of centrioles, while in Sass6 em4/em4 mutants, only rare and single centrioles were detected (11%), suggesting centriole duplication failure (Figure 2G and H).Of note, no centrioles were detected in the mitotic poles of Sass6 em5/em5 embryos (Figure 2G and H).Overall, the data suggested that the Sass6 em4/em4 is a severe hypomorphic allele of Sass6, whereas the Sass6 em5/em5 is likely to be a null allele of Sass6, and that SAS-6 is essential for centriole formation in mouse embryos in vivo.

SAS-6 is required for centriole integrity, but not formation, in mESCs
To study the roles of Sass6 in an in vitro setting that mimics mouse embryonic development, we chose to knockout Sass6 in mESCs.To accomplish this without any reasonable doubt of residual SAS-6 protein, we used CRISPR/Cas9 with a pair of guide RNAs (gRNAs) flanking the open reading frame (ORF) of Sass6, and engineered a null allele lacking the entire Sass6 ORF (Sass6 −/− ) (Figure 3A, Materials and methods, Table 2).The deletion of the Sass6 ORF in Sass6 −/− mESCs was confirmed at the level of DNA using PCR (Figure 3A, bottom panel), the loss of Sass6 mRNA validated by RT-PCR (Figure 3B), and the lack of detectable SAS-6 protein corroborated by Western blot (Figure 3C) and immunostaining (Figure 3-figure supplement 1A).We next used immunostaining for TUBG and FOP, a centriole marker, to assess centrosome and centriole formation in Sass6 −/− mESCs.While centrosomes, as defined by the co-localization of TUBG and FOP, were evident in the vast majority of WT mESCs (94%), it was surprising that more than half of Sass6 −/− mESCs still possessed centrosomes (56%) (Figure 3D and E).This unexpected observation suggested that unlike in mouse embryos, SAS-6 may not be essential for centrosome formation in mESCs.
To probe whether the TUBG-marked centrosomes in Sass6 −/− mESCs contained centrioles at their core, we used higher-resolution U-ExM combined with immunostaining for bona fide centriolar wall markers (Ac-Tub and α/β-tubulin, TUB) (Figure 3F).Almost all of the centrosomes in WT cells contained two or more centrioles (99%) and only a rare fraction contained one centriole (1%) (Figure 3F and G).In contrast, in Sass6 −/− mESCs, about half of the centrosomes had two or more centrioles (46%) and a comparable fraction had one centriole (45%) (Figure 3F and G).Additionally, in a minor fraction of Sass6 −/− centrosomes (9%), aberrant centriolar threads were detected (Figure 3F and G).Structurally, we classified the centrioles in Sass6 −/− mESCs into the following three categories: normal-like centrioles (18%), abnormal centrioles (65%), and thread-like structures (17%) (Figure 3F and H).Quantitatively, we measured the length of the normal-like centrioles and found that they were significantly longer in in Sass6 −/− than in WT (Figure 3I).In summary, the data suggested that Sass6 −/− centrioles in mESCs had a compromised ability to duplicate and/or were unstable with mostly abnormal structures.
The online version of this article includes the following figure supplement(s) for figure 1:
The centrioles in Sass6 −/− mESCs have proximal and distal defects SAS-6 has been shown to cooperate with STIL, an essential component for centriole duplication, to initiate procentriole formation (Kratz et al., 2015).To address whether the fraction of centrioles that failed to duplicate in Sass6 −/− mESCs is associated with an impairment of STIL recruitment, we used U-ExM combined with Ac-TUB and STIL immunostaining (Figure 4A, Figure 4-figure supplement 1A).STIL localized to three-quarters of centrosomes in WT mESCs undergoing centriole duplication (74%) (Figure 4A and B;.In Sass6 −/− mESCs, STIL localized to less than one-third of the centrosomes (29%) (Figure 4A and B;, which is roughly half of the STIL-positive centrosomes in WT cells, and might account for the duplication failure in almost half of the Sass6 −/− centrosomes (single centrioles in Figure 3G).To assess whether the loss of centriole integrity in Sass6 −/− mESCs is associated with disruption of the proximal centriole end or internal structural scaffold, we immunostained centrioles in U-ExM for CEP135 (proximal end) and POC5 (scaffold) (Figure 4C and E;.We found that CEP135 localized to the proximal centriole in all WT cells, while the number of centrioles with CEP135 was decreased in Sass6 −/− cells (73%) (Figure 4C and D).Notably, only a minority of CEP135-positive centrioles in Sass6 −/− mESCs showed normal CEP135 localization (12%) (Figure 4-figure supplement  1B, C).On the other hand, POC5 was present in WT and Sass6 −/− intact centrioles, but also lining the abnormal centrioles and centriolar threads in Sass6 −/− cells, further confirming their centriolar nature (Figure 4E).
Next, we examined whether the mother centrioles in Sass6 −/− mESCs were decorated with distal appendages and used U-ExM combined with Ac-TUB and CEP164 immunostaining (Figure 4H; Figure 4-figure supplement 1F).The data showed, as expected, that CEP164 mostly localized to the mother centrioles in WT centrosomes (94%) (Figure 4H   To assess whether the abnormal centrioles in Sass6 −/− mESCs retained the ability to template cilia, we used immunostaining against the ciliary axoneme marker Ac-TUB and ciliary membrane protein ARL13B (Figure 4J).Although cilia were present in only a small fraction of WT mESCs (11%), no cilia were detected in Sass6 −/− mESCs (Figure 4J and K), suggesting that SAS-6 is not only required for centriole integrity, but also distal appendage recruitment and cilia formation in mESCs.

Short-term culture of Sass6 em5/em5 blastocysts induces centriole formation
The finding that Sass6 em5/em5 Cetn2-eGFP mESCs derived from E3.5 blastocysts are also able to form centrioles, prompted us to investigate whether these centrioles formed de novo in the absence of SAS-6.To begin to address this question, we combined the Cetn-eGFP with TUBG immunostaining in control Cetn2-eGFP blastocysts, which showed that the majority of cells had foci positive for both markers (73%) (Figure 5A and B).In contrast, in mutant Sass6 em5/em5 Cetn2-eGFP blastocysts, only small TUBG accumulations were observed in a quarter of the cells (23%), but they did not contain CETN2-eGFP, suggesting that they were devoid of centrioles (Figure 5A and B).We next asked how early the centrioles form during the derivation of mESCs from the Sass6 em5/em5 Cetn2-eGFP blastocysts.After 24 hr (hr) of culture, almost all of the cells in the cultured WT blastocysts contained centrioles (98% with both CETN2-eGFP and TUBG), and remarkably, centrioles were already detectable in one-third gRNA (5′ F and 5′ R, band = 977 bp), Ex8 (Ex8 F and Ex8 R1, band = 281 bp), 3′ gRNA (3′ F and 3′ R, band = 992 bp), Sass6 ORF (5′ F and 3′ R, 825 bp in Sass6 −/− , 34,349 bp in wild-type (WT), product too long to be amplified).(B) RT-PCR analyses of Sass6 transcripts in WT and Sass6 −/− mESCs.The picture shows the PCR products from RT-PCR using the following primers: from Ex1 to Ex8 (Ex1 F and Ex8 R2, band = 734 bp), from Ex9 to Ex14 (Ex9 F and Ex14 R, band = 617 bp), Tbp Ctrl (Tbp F and Tbp R, band = 156 bp).(C) Western blot analysis using a SAS-6-specific antibody on WT and Sass6 −/− mESCs extracts.Asterisks mark non-specific bands.GAPDH is used as a loading control.     of the cells in Sass6 em5/em5 Cetn2-eGFP blastocysts (33%) (Figure 5C and D).The data suggested that the mESC culture conditions are conducive to de novo centriole formation in the absence of SAS-6.

The differentiation of Sass6 mutant mESCs leads to centriole loss
To test whether the ability to form centrioles, albeit mostly abnormal, via a SAS-6-independent pathway is a characteristic of the pluripotent mESCs, we analyzed centriole formation in mESCs differentiated and enriched for neural progenitor cells (NPCs).As expected, centrosomes immunostained for TUBG and FOP were detected in almost all WT NPCs characterized by the expression of the intermediate filament NESTIN (97%, Figure 5E and F; Figure 5-figure supplement 1A).Notably, in Sass6 −/− NPCs, the number of cells with centrosomes sharply decreased upon differentiation (from 56%, Figure 3D and E, down to 6%, Figure 5E and F).The data suggested that the SAS-6-independent centriole formation pathway is a property of pluripotent mESCs that is largely lost upon differentiation.

Centriole formation in Sass6 −/− mESCs relies on PLK4 activity
To understand the mechanism of how Sass6 −/− mESCs are able to form centrioles in the absence of SAS-6, we tested the hypothesis that a higher concentration of centriolar components and a robust activity of PLK4 allow for centriole formation in Sass6 −/− mESCs.We first examined whether the loss of centrioles upon mESCs differentiation correlated with a decrease in the recruitment of centriolar and centrosomal proteins that are important for the initial events of centriole assembly.Thus, we used immunostaining and quantified the centrosomal signal intensity of TUBG, CEP152, STIL, and SAS-4 in WT mESCs and NPCs (Figure 6A-D).Compared to mESCs, and in support of our hypothesis, NPCs showed highly reduced levels of all four investigated proteins (~ fourfold) (Figure 6A-D).
To functionally test whether SAS-6-independent centriole formation requires a potent activity of PLK4 in mESCs, we treated WT and Sass6 −/− mESCs with different doses of the specific PLK4 inhibitor, centrinone B (Wong et al., 2015), and performed immunostaining for TUBG and FOP (Figure 7A).In WT mESCs treated with 100 nM centrinone B, the fraction of cells with centrosomes, as defined by the co-localization of TUBG and FOP, was not significantly different than in DMSO-treated control cells (82% and 90%, respectively) (Figure 7A and B).In contrast, the number of cells with centrosomes in Sass6 −/− mESCs treated with 100 nM centrinone B was greatly reduced (6%) compared to DMSO-treated control cells (46%) (Figure 7A and B).In both WT and Sass6 −/− mESCs treated with 500 nM centrinone B, centrosomes were identified only in a minor fraction of cells (10% and 4%,    respectively) (Figure 7A and B).The data suggested that SAS-6-independent centriole formation in mESCs depends on high PLK4 activity.

Discussion
In this study, we report that mutations in mouse Sass6 cause embryonic arrest at mid-gestation with elevated levels of p53 and cell death, as well as the activation of the p53-, 53BP1-, and USP28dependent mitotic surveillance pathway (Figure 1).We have previously reported similar phenotypes of elevated p53 and cell death for mutations in other genes essential for centriole duplication, such as Cenpj and Cep152 (Bazzi and Anderson, 2014a).The current data demonstrated that mouse SAS-6 is required for centriole formation in developing mouse embryos (Figure 2), as expected from the established role of its orthologs in C. elegans and human cells (Gupta et al., 2020;Leidel et al., 2005;Wang et al., 2015).Together, the data provide further support that the activation of the mitotic surveillance pathway is not specific to the loss of specific centriolar proteins but rather the loss of the centriole/centrosome structure and function per se (Bazzi and Anderson, 2014b).

Figure 7 continued
Figure 8. Graphical model depicting the consequences of SAS-6 loss in mouse embryos, mouse embryonic stem cells (mESCs), and neural progenitor cells (NPCs).Compared to mouse embryos and in vitro differentiated NPCs, mESCs exhibit a higher concentration of centrosomal components and a robust Polo-Like Kinase 4 (PLK4) activity, as indicated by changes in pericentriolar material color and size.This difference permits the formation of abnormal centrioles in Sass6 −/− mESCs, while it results in the loss of centrioles in developing mouse embryos and NPCs.
C. Reinhardtii and D. melanogaster (Nakazawa et al., 2007;Rodrigues-Martins et al., 2007).Similar phenotypes of abnormal centrioles were also described in human RPE-1 cells lacking the SAS-6 oligomerization domain; however, the loss of the entire SAS-6 protein in these cells leads to the loss of centrioles (Wang et al., 2015).Interestingly, we demonstrate that centrioles appear during the derivation process of SAS-6-deficient mESCs but are lost again upon differentiation (Figure 5).The findings extend the observations on abnormal centrioles or centriole loss upon SAS-6 depletion and reveal the differential requirements for SAS-6 even within cells of the same species (Figure 8).
In D. melanogaster, C. elegans, and human cells, SAS-6 has been shown to directly interact with a downstream protein STIL, and provides a structural basis for the recruitment of other centriole duplication proteins, such as CEP135 and SAS-4 (Lin et al., 2013;Qiao et al., 2012;Tang et al., 2011).The loss of STIL or SAS-4 leads to centriole duplication failure in all organisms and cell types studied to date (Bazzi and Anderson, 2014a;Vulprecht et al., 2012).Notably, unlike SAS-6, SAS-4 is essential for centriole formation in mESCs (Xiao et al., 2021).Our data show that STIL localized to half of the centrosomes in Sass6-null mESCs, which might account for centriole duplication failure in almost half of the centrosomes that contain only single centrioles.Our data also suggest that pluripotent mESCs without SAS-6 have a bypass pathway of SAS-4 recruitment.Given that centrioles in mESCs show higher levels of centrosomal components compared to NPCs (Figure 6), and that SAS-6-independent centriole formation in mESCs depends on high PLK4 activity (Figure 7), we propose that these factors drive SAS-6-independent centriole formation by supporting the stability of centriole intermediate structures in mESCs, while embryos lacking such a mechanism experience rapid disassembly of the intermediate assemblies.
Although centrioles are still present in Sass6 −/− mESCs, they exhibit profound proximal and distal defects.Whether this phenotype arises as a consequence of an improper initiation of assembly or a later destabilization of the microtubule triplets, are still open questions.The cartwheel protein CEP135 has been shown to be important for centriole stability in T. thermophila and human cells (Bayless et al., 2012;Lin et al., 2013).Our data show that centrioles in Sass6 −/− mESCs exhibit abnormal localization of CEP135 (Figure 4-figure supplement 1B, C).We speculate that the lack of an initial stable cartwheel scaffold that provides a ninefold symmetry template and the ensuing mis-localization of CEP135 may account for the abnormal centriole architecture and its instability.
In conclusion, our work provides new fundamental insights into alternative and SAS-6-independent pathways of centriole formation in mammalian cells.It also highlights that mESCs are a special in vitro model for centriole biology that can tolerate centriolar aberrations, such as in Sass6 mutants, or even the loss of centrioles, as in Cenpj mutants, without undergoing apoptosis or cell cycle arrest (Lambrus and Holland, 2017;Xiao et al., 2021).The difference in centrosome biology between mouse embryos and mESCs adds to a growing body of evidence of variable centrosome composition and function among different cell types (O'Neill et al., 2022;Xie et al., 2021).

Generation of Sass6 mutant mESCs using CRISPR/Cas9
For the generation of the CRISPR/Cas9-mediated Sass6 knockout mESCs line, a pair of gRNAs targeting the 5′ and 3′ ends of the entire Sass6 ORF (Figure 2-figure supplement 1 and Table 2) were cloned as double-stranded oligo DNA into BbsI and SapI sites in pX330-U6-Chimeric_BB-CBh-hSpCas9 vector (Addgene; Watertown, MA, USA) modified with a Puro-T2K-GFP cassette containing puromycin-resistance by Dr. Leo Kurian's research group (Center for Molecular Medicine Cologne).mESCs in suspension were transfected with the modified pX330 vector containing the pair of gRNAs using Lipofectamine 3000 (Thermo Fisher Scientific).The cells were then subjected to selection using puromycin (2 μg/ml, Sigma-Aldrich; St. Louis, MO, USA) 24 hr post-transfection for 2 days.After recovery for 5 days, individual colonies were picked and screened for the Sass6 locus deletion by PCR (Figure 3A and Table 2; Supplementary file 1).The cells were used for experiments after four passages.

RT-PCR
RNA was extracted from mESCs using the RNeasy Plus Mini Kit (Qiagen; Hilden, Germany).Reverse transcription was performed using SuperScript III reverse transcriptase (Invitrogen; Waltham, MA, USA) and oligo(dT) primer per the manufacturer's recommendation.The cDNA was used for PCR analysis using the primers shown in Supplementary file 1.
Pre-implantation E3.5 blastocysts were recovered by flushing the uterine horns with EmbryoMax M2 Medium (Sigma-Aldrich), and fixed in 4% PFA for 20 min at room temperature (RT) and 20 min at 4 °C.Blastocysts were then permeabilized for 3 min with 0.5% Triton X-100 (Sigma-Aldrich) in PBS and three times for 10 min with immunofluorescence (IF) buffer containing 0.2% Triton X-100 in PBS.After blocking with 10% heat-inactivated goat serum in IF buffer, the blastocysts were incubated overnight with the primary antibodies at 4 °C, followed by a 2 hr incubation with the secondary antibodies and DAPI at RT (1:1000, AppliChem).Blastocysts were imaged in single drops of PBS covered with mineral oil, followed by genotyping.
For IF staining of embryo sections from the brachial region (forelimb and heart level), the slides were post-fixed in methanol for 10 min at -20 °C, washed with IF buffer, and blocked with 5% heat-inactivated goat serum in IF buffer.The slides were incubated overnight with primary antibodies at 4 °C followed by 1 hr incubation with secondary antibodies and DAPI at RT (1:1000, AppliChem), then mounted with ProLong Gold Antifade reagent (Cell Signaling Technology; Danvers, MA, USA).
For IF staining of mESCs, the cells were cultured in Lab-Tek II chamber slides or ibiTreat µ slides (Figure 6), coated with 0.1% gelatin, fixed with 4% PFA for 10 min at RT, and post-fixed with methanol for 10 min at -20 °C.The cells were then permeabilized for 5 min using 0.5% Triton X-100 in PBS.After blocking with IF buffer with 5% heat-inactivated goat serum, the cells were incubated overnight with the primary antibodies at 4 °C, followed by 1 hr incubation with secondary antibodies and DAPI (1:1000, AppliChem).The slides were mounted with ProLong Gold (Cell Signaling Technology).Images were obtained using TCS SP8 (Leica Microsystems) confocal microscope with PL Apo 63 x/1.40Oil CS2 objective and Stellaris 5 (Leica Microsystems) confocal microscope with HC PL APO 63 x/1.30GLYC CORR CS2 objective.

Ultrastructure-expansion microscopy (U-ExM)
For U-ExM of mESCs, the cells were cultured on 12 mm glass cover slips (VWR) coated with 0.1% gelatin and fixed with methanol for 10 min at -20 °C.For U-ExM of embryos, cryo-sections were collected on 12 mm glass coverslips, O.C.T. was washed away in PBS.Sample expansion was performed as described in Gambarotto et al., 2019.Briefly, the cover slips were incubated in 1.4% formaldehyde (Sigma-Aldrich), 2% acrylamide (Sigma-Aldrich) in PBS for 5 hr at 37 °C.Gelation was carried out in 35 µl of monomer solution 23% (w/v) sodium acrylate (Sigma-Aldrich), 10% (w/v) acrylamide, 0.1% (w/v) N,N′-methylenebisacrylamid (Sigma-Aldrich) in 1x PBS supplemented with 0.5% APS (Bio-Rad; Feldkirchen, Germany) and 0.5% TEMED (Bio-Rad) for 5 min on ice and 1 h at 37 °C, followed by gel incubation in denaturation buffer 200 mM sodium dodecyl sulfate (SDS; AppliChem), 200 mM NaCl and 50 mM Tris in H 2 O (pH 9) for 1.5 hr at 95 °C, and initial overnight expansion in ddH 2 O at RT.The gels were incubated with primary antibodies for 3 hr at 37 °C on an orbital shaker, then with secondary antibodies for 2.5 hr at 37 °C.After the final overnight expansion in ddH 2 O at RT, the expanded gel size was accurately measured using a caliper, and then mounted on 12 mm glass cover slip coated with Poly-D-Lysine (Thermo Fisher Scientific) and imaged using a TCS SP8 confocal microscope with a Lightning suite (Leica Microsystems) to generate deconvolved images.

Blastocyst in vitro culture
Blastocysts at E3.5 were cultured for 24 hr at 37 °C and 6% CO 2 , in ibiTreat µ-Slides (Ibidi GmbH, Munich, Germany) on feeder cells that were previously growth-arrested with a 2 hr mitomycin C treatment (10 μg/ml, Sigma-Aldrich), in mESCs derivation media mESCs media except with the FBS replaced with Knockout Serum Replacement (15%, Thermo Fisher Scientific).The cultures were fixed and processed for IF staining as described above for the E3.5 blastocysts.Slides were mounted with VectaShield mounting medium (Linaris; Dossenheim, Germany).

Neural differentiation of mESCs
For neural differentiation, embryoid bodies were generated using the hanging drop method (Wang and Yang, 2008).Around 1000 mESCs were suspended in 20 µl of differentiation medium mESCs media without LIF, PD0325901, and CHIR99021, supplemented with 1 µM retinoic acid (Sigma-Aldrich).After 3 days, the resulting embryoid bodies were plated on ibiTreat µ-slides coated with fibronectin (30 μg/ml, Sigma-Aldrich) and cultured for 4 days.The cells were fixed in methanol for 10 min at -20 °C for centrosomal staining or 4% PFA for 10 min at RT for other stainings, and processed for IF similar to mESCs.Slides were mounted with VectaShield mounting medium (Linaris).

Western blotting
Western blots were performed according to standard procedures (Mahmood and Yang, 2012).Briefly, dissected embryos at E9.5 were lysed in Laemmli buffer, and the cells were lysed in RIPA buffer 150 mM NaCl, 50 mM Tris pH 7.6, 1% Triton X-100, 0.25% sodium deoxycholate, and 0.1% SDS (AppliChem) with an ethylenediaminetetraacetic acid (EDTA)-free protease inhibitor cocktail (Merck), phosphatase inhibitor cocktail sets II and IV (Merck), and phenylmethylsulfonyl fluoride (PMSF; Sigma-Aldrich).80 µg of protein per sample was used for SDS-PAGE.Samples were then blotted onto polyvinylidene difluoride membrane (Merck).After blocking in 5% non-fat milk (Carl Roth), the membrane was incubated overnight at 4 °C with a SAS-6 or GAPDH antibody.Secondary antibodies coupled to horseradish peroxidase were used for enhanced chemiluminescence signal detection with ECL Prime Western Blotting System (GE Healthcare).

Image analysis
For signal quantification using ImageJ (NIH), the signal intensity from the nuclear area, as determined by DAPI staining, was normalized to DAPI intensity.The signal intensity from the centrosomal area was determined by TUBG staining.The fold change of p53 or Cl.CASP3 was defined as a ratio between normalized signal intensity to the mean signal intensity of the WT from all replicates, and the fold change of centrosomal proteins was defined as a ratio between mean signal intensity to the mean signal intensity of the mESCs from all replicates.The percentage of cells with centrosomes in embryos and mESCs was defined as a ratio between centrosome number manually quantified using ImageJ and the number of nuclei quantified using the IMARIS software (Bitplane; Belfast, United Kingdom).Centriole length was measured using ImageJ from nearly parallel-oriented centriole walls stained for TUB.The length was corrected for the expansion factor obtained from dividing the gel size after expansion by the size of a cover slip used for gelation (12 mm).

Statistical analysis
To identify statistical differences between two or more groups, two-tailed Student's t-test or oneway ANOVA with Tukey's multiple comparisons was performed.p<0.05 was used as the cutoff for significance.The statistical analyses were performed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) or Prism (GraphPad; San Diego, CA, USA) and the graphs were generated using Prism.
downstream of the translation start site, with 78 amino acids not native to the protein 129 amino acids downstream of the translation start site, with six amino acids not native to the protein

Figure supplement 1 .
Figure supplement 1. Mutations in Sass6 lead to an increase in the mitotic index in mouse embryos.

Figure 4 .
Figure 4. Centrioles in Sass6 −/− mouse embryonic stem cells (mESCs) exhibit proximal and distal defects.(A) Immunostaining for Ac-TUB and STIL of U-ExM of centrioles from wild-type (WT) and Sass6 −/− mESCs.Examples of centrioles with or without STIL are shown.Scale bar = 200 nm.(B) Quantification of the percentage of centrosomes with (w.) STIL in (A) from four independent experiments.Error bars represent mean ± SD WT: 74 ± 6% Figure 4 continued on next page

Table 3 .
List of primary antibodies used in this study.