APC/C^FZR-1 regulates centrosomal ZYG-1 to limit centrosome number

Centrosomes, as the primary microtubule-organizing centers, establish bipolar mitotic spindles that ensure accurate transmission of genomic content into two daughter cells. To sustain genomic integrity, centrosome number must be strictly controlled by duplicating only once per cell cycle. The abundance of centrosome factors directly influences centrosome number. Blocking degradation of centrosome factors causes extra centrosomes, while their depletion inhibits centrosome duplication (Nigg and Holland, 2018). The ubiquitin-proteasome system provides a key mechanism to control centrosome protein levels (Nakayama and Nakayama, 2006). Levels of centrosome factors are regulated by proteasomal degradation through E3 ubiquitin ligases, including the anaphase promoting complex/cyclosome (APC/C) and SKP1–CUL1–F-box-protein complex (SCF), and their regulatory mechanisms appear to be conserved (Arquint et al., 2012; Cunha-Ferreira et al., 2009; Holland et al., 2010; Medley et al., 2017a; Meghini et al., 2016; Peel et al., 2012; Puklowski et al., 2011; Rogers et al., 2009; Strnad et al., 2007; Tang et al., 2009). In C. elegans, the kinase ZYG-1 is a key centrosome regulator and ZYG-1 levels are critical for normal centrosome number and function (O'Connell et al., 2001; Song et al., 2008). The abundance of Plk4 (human homolog of ZYG-1) is regulated by the SCF-mediated proteolysis through the F-box protein Slimb and βTrCP (in Drosophila and human, respectively) (Cunha-Ferreira et al., 2009; Holland et al., 2010; Rogers et al., 2009). Such a mechanism is conserved in C. elegans where ZYG-1 levels are controlled by the SCF complex with Slimb/βTrCP (SCF^Slimb/βTrCP) (Peel et al., 2012).

Another E3 ubiquitin ligase, APC/C, also regulates levels of centrosome regulators. In C. elegans, FZR-1 (Cdh1 in humans), a coactivator of APC/C, negatively influences centrosome assembly, and together, denoted APC/C^FZR-1, regulate SAS-5 levels through a KEN box (Kemp et al., 2007; Medley et al., 2017a), which is the conserved motif in humans and flies where APC/C^Cdh1/Fzr targets centrosome factors (human STIL/SAS-5, SAS-6, CPAP/SAS-4 and Drosophila Spd2) via KEN-box (Arquint and Nigg, 2014; Meghini et al., 2016; Strnad et al., 2007; Tang et al., 2009). Here, we investigate ZYG-1 as a potential APC/C^FZR-1 substrate, and how the E3 ubiquitin ligases APC/C^FZR-1 and SCF^Slimb/βTrCP influence ZYG-1 levels during centrosome assembly.

Prior study has identified SAS-5 as an APC/C^FZR-1 substrate, and implicated APC/C^FZR-1 as targeting additional centrosome regulators in C. elegans (Medley et al., 2017a). FZR-1 was identified as a genetic suppressor of ZYG-1 since loss of FZR-1 restores embryonic viability and centrosome duplication to zyg-1(it25) mutants (Kemp et al., 2007; Medley et al., 2017a). Given the strong genetic interaction between zyg-1 and fzr-1, a compelling possibility is that ZYG-1 might be an APC/C^FZR-1 substrate. In C. elegans embryos, while cellular ZYG-1 protein remains low, centrosome-associated ZYG-1 is regulated during cell cycle, peaking at anaphase and declining toward mitotic exit (Song et al., 2008, 2011). ZYG-1 levels are known to be regulated by protein phosphorylation (Song et al., 2011) and proteasomal degradation (Peel et al., 2012).

If ZYG-1 is degraded by APC/C^FZR-1, downregulation of APC/C^FZR-1 would be expected to protect ZYG-1 from destruction, leading to protein accumulation. To address this, we stained embryos with anti-ZYG-1 and quantified the fluorescence intensity of centrosome-associated ZYG-1 signal (Fig. 1). Quantitative immunofluorescence (IF) reveals that, compared to wild-type (WT) controls (1.00±0.25 fold; mean±s.d.), fzr-1 mutant embryos exhibit increased levels of centrosomal ZYG-1 during the first anaphase (1.28±0.27 fold). zyg-1(it25) contains a temperature-sensitive point mutation (P442L) within the cryptic polo-box (CPB) domain that is critical for centrosomal localization of ZYG-1 (Fig. 2A) (Kemphues et al., 1988; O'Connell et al., 2001; Shimanovskaya et al., 2014). In zyg-1(it25) embryos, centrosomal ZYG-1 is reduced to ∼40% of the level of WT controls. By contrast, centrosomal ZYG-1 levels in zyg-1(it25) fzr-1(bs31) double-mutants are significantly increased (0.80±0.33 fold) compared to zyg-1(it25) mutants, indicating the fzr-1 mutation leads to partial restoration of centrosomal ZYG-1 to zyg-1(it25) embryos. Furthermore, inhibition of MAT-3 (also known as APC8 or CDC23), an essential subunit of the APC/C complex (Golden et al., 2000) results in elevated centrosomal ZYG-1, supporting that FZR-1 regulates centrosomal ZYG-1 through an APC/C-dependent mechanism (Fig. S1A).

These results support a hypothesis that ZYG-1 levels are regulated by APC/C^FZR-1-dependent proteolysis. Despite exhaustive efforts to address how loss of FZR-1 impacted cellular ZYG-1 levels, we were unable to detect endogenous ZYG-1 by immunoblot or immunoprecipitation (IP) due to extremely low abundance of ZYG-1. However, IP data suggest a possible physical interaction between ZYG-1 and FZR-1 (Fig. S1B), consistent with a hypothesis that ZYG-1 could be a potential APC/C^FZR-1 target.

The APC/C^FZR-1 E3 ubiquitin ligase targets substrates via interaction of the coactivator FZR-1 with conserved motifs, predominantly the destruction (D)-box and KEN box degrons, within the substrate (Glotzer et al., 1991; Pfleger and Kirschner, 2000). While ZYG-1 contains no KEN box, in silico analysis identified four putative D-box motifs (RxxL) within ZYG-1. We then asked whether APC/C^FZR-1 might target ZYG-1 for degradation through the D-box. In ZYG-1 (Fig. 2A), D-box1 resides within the kinase domain, D-box2 in the linker 1 (L1) domain, and D-box3 at the junction between the L1 and CPB domains, and D-box4 within the CPB domain (Lettman et al., 2013; O'Connell et al., 2001; Shimanovskaya et al., 2014). Sequence alignments show D-box1 is the only conserved motif in closely related nematodes (Fig. S2).

To determine whether any of putative D-boxes might be functional degron, we mutated each D-box (RxxL) by replacing the two critical residues with alanine (AxxA) at the endogenous locus in wild-type(N2) animals (Fig. 2A). Using CRISPR/Cas9 genome editing, we generated D-box2 and D-box3 mutants with two alanine substitution mutations (2A; AxxA), termed the ZYG-1^DB2(2A) and ZYG-1^DB3(2A) mutation. However, the 2A mutation of D-box1 or D-box4 produced sterile hermaphrodites. In human cells, L45 in the cyclin B1 D-box (42RxxL45) is critical for cyclin B1 destruction (Matsusaka et al., 2014). Thus, we generated D-box1 and D-box4 mutants with single alanine substitution (1A; RxxA), termed the ZYG-1^DB1(1A) and ZYG-1^DB4(1A) mutation. All four ZYG-1 D-box mutations in WT background produce viable progeny (Table 1).

Next, we utilized the hypomorphic zyg-1(it25) background to test how D-box mutations affected ZYG-1 function. In zyg-1(it25) mutants, centrosome duplication fails during the first cell cycle, leading to monopolar spindles in the second mitosis and 100% embryonic lethality at the restrictive temperature of 24°C (O'Connell et al., 2001). If APC/C^FZR-1 targets ZYG-1 through a D-box, mutating a functional D-box degron should inhibit APC/C^FZR-1 binding, thereby protecting ZYG-1 from destruction. Then, ZYG-1 accumulation should compensate for impaired ZYG-1 function in zyg-1(it25) mutants. By introducing the same D-box mutations as above in zyg-1(it25) mutants, we asked whether any D-box mutation rescued zyg-1(it25) phenotypes (Fig. 2B,C; Table 1). Intriguingly, the ZYG-1^DB3(2A) mutation significantly restores embryonic viability and bipolar spindles to zyg-1 mutants by >6-fold. Notably, the ZYG-1^DB3(2A) mutation restores embryonic viability more robustly at the semi-restrictive condition (22.5°C) compared to the restrictive temperature (24°C), indicating that the ZYG-1^DB3(2A) mutation requires ZYG-1 activity for the zyg-1 suppression (Table 1).

However, the other D-box mutations increase embryonic lethality and monopolar spindles to zyg-1(it25) mutants. It appears that the zyg-1(it25) mutation provides a sensitized genetic environment for ZYG-1^DB1(1A), ZYG-1^DB2(2A) and ZYG-1^DB4(1A) mutations as the same ZYG-1 D-box mutations in WT background have little effect on embryonic survival. ZYG-1 function might be impaired by alanine substitutions at these sites, implicating that these residues are important for ZYG-1 function, presumably as part of the functional domain (Fig. 2A, Fig. S2; Lettman et al., 2013; O'Connell et al., 2001; Shimanovskaya et al., 2014). Although we cannot exclude the possibility of the other D-boxes functioning as degrons, our data support a model where ZYG-1 D-box3, a functional degron, mediates the interaction of APC/C coactivator FZR-1 with ZYG-1 for proteasomal degradation.

The ZYG-1^DB3(2A) mutation should inhibit APC/C^FZR-1-mediated proteolysis of ZYG-1, leading to ZYG-1 hyperstabilization. To test this, we examined centrosome-associated ZYG-1 by quantitative IF (Fig. 2D). Consistent with the zyg-1 suppression, the ZYG-1^DB3(2A) mutation leads to elevated centrosomal ZYG-1 levels (1.48±0.4 fold mean±s.d.) compared to zyg-1(it25) controls (1.00±0.32 fold) while the ZYG-1^DB2(2A) mutation results in modestly decreased centrosomal ZYG-1 (0.87±0.28 fold), suggesting that the ZYG-1^DB3(2A) mutation renders ZYG-1 resistant to degradation. It remains unclear whether APC/C^FZR-1 regulates cellular ZYG-1 levels or affects centrosomal ZYG-1 locally. However, we favor a model where elevated centrosomal ZYG-1 is a direct consequence of increased cellular ZYG-1 levels (Decker et al., 2011; Song et al., 2008). Such correlation between cellular and centrosomal levels of Plk4 has been reported previously (Guderian et al., 2010).

Since the D-box (1A) mutations may not completely disrupt APC/C^FZR-1 binding, it remains possible that D-box1 and D-box4 motifs are functional degrons. To address this, we mutated L352 of D-box3 (349RxxL352) to alanine (ZYG-1^DB3(1A); Fig. 2A), and examined its consequences (Fig. S1C,D). The ZYG-1^DB3(1A) mutation produces neither change on centrosomal ZYG-1 nor rescue of zyg-1(it25) lethality, indicating that both the arginine (R) and leucine (L) residues of D-box3 are required for ZYG-1 destruction, unlike what is seen in human cyclin B1 (Matsusaka et al., 2014). Furthermore, the ZYG-1^DB1(1A) or ZYG-1^DB4(1A) mutation leads to decreased centrosomal ZYG-1, suggesting alanine substitutions at these sites negatively impacted ZYG-1 localization (Fig. S1C,D). While we cannot exclude the role of D-box1 and D-box4 as functional degrons, our results support a model where APC/C^FZR-1 targets ZYG-1 for proteasomal destruction, at least partially, through D-box3.

Increased centrosomal ZYG-1 by the ZYG-1^DB3(2A) mutation closely correlates with the zyg-1 suppression. The zyg-1(it25) embryo exhibits drastically reduced levels of centrosomal ZYG-1, which negatively affects recruitment of downstream centrosome factors. ZYG-1 is required for SAS-5 and SAS-6 loading to centrosomes, and SAS-5 and SAS-6 together recruit SAS-4 (Delattre et al., 2006; Pelletier et al., 2006). SAS-7 was recently identified as a centrosome factor critical for centrosomal targeting of SPD-2, and both SAS-7 and SPD-2 act upstream of ZYG-1 (Sugioka et al., 2017).

Elevated centrosomal ZYG-1 by the ZYG-1^DB3(2A) mutation should restore recruitment of downstream factors, rendering centrosomes competent to assemble new centrioles. To address this, we quantified centrosome-associated protein levels in D-box mutants. To facilitate quantification, we generated epitope-tagged SAS-5::V5 and Ollas::SAS-6 strains at the endogenous locus using CRISPR/Cas9 genome editing (Fig. S1F,G). By co-staining embryos with anti-V5 and anti-SAS-4, we first quantified the fluorescence intensity of centrosomal SAS-5::V5 and SAS-4 signals (Fig. 2E). At anaphase, the ZYG-1^DB3(2A) mutation leads to elevated centrosomal SAS-5 (1.26±0.39 fold; mean±s.d.) compared to WT controls (1.00±0.27), which is nearly equivalent to the increase in SAS-5^KEN(3A) mutants (1.26±0.37 fold) where APC/C^FZR-1-mediated degradation of SAS-5 is blocked (Medley et al., 2017a). However, centrosomal SAS-5 levels are decreased in ZYG-1^DB2(2A) mutants (0.70±0.19 fold). Similarly, centrosomal SAS-4 levels are increased in ZYG-1^DB3(2A) (1.18±0.22 fold) and SAS-5^KEN(3A) (1.14±0.17 fold) mutants, but reduced in the ZYG-1^DB2(2A) mutant (0.79±0.22 fold). Furthermore, D-box mutations produce similar effects on SAS-5 and SAS-4 levels in zyg-1(it25) mutants (Fig. S3A,C). However, SAS-5 and SAS-4 levels are unaffected at metaphase since APC/C^FZR-1 is active in late mitosis (Fig. S3B,D).

Next, we examined SAS-6 levels in ZYG-1 D-box(2A) mutants (Fig. 2F). At anaphase, the ZYG-1^DB3(2A) mutation leads to increased centrosomal SAS-6 (1.26±0.23 fold) compared to WT controls (1.00±0.19 fold), which is comparable to the change in SAS-5^KEN(3A) mutants (1.22±0.23 fold), whereas the ZYG-1^DB2(2A) mutation has little effect on SAS-6 (0.94±0.13 fold). Finally, we monitored dynamic changes of centrosomal GFP::SAS-7 by live imaging (Fig. S3E). As expected for an upstream factor, centrosomal SAS-7 levels are unaffected in ZYG-1 D-box mutants. Together, our data show the ZYG-1^DB3(2A) mutation leads to elevated centrosomal ZYG-1, which promotes centrosomal loading of downstream factors, thereby restoring centrosome duplication and viability to zyg-1 mutants.

While the ZYG-1^DB3(2A) mutation leads to elevated centrosomal ZYG-1, it is evident that the fzr-1 mutation produces a greater impact on ZYG-1 levels (1.86±0.44 fold) than the ZYG-1^DB3(2A) mutation (Fig. 2D). Similar trends between fzr-1(bs31) and ZYG-1^DB3(2A) mutations were observed for the zyg-1 suppression (Fig. 2B,C, Table 1).

Since SAS-5 is an APC/C^FZR-1 substrate and mutating the SAS-5 KEN-box leads to the zyg-1 suppression (Medley et al., 2017a), we asked whether mutating the ZYG-1 D-box3 and SAS-5 KEN-box degrons simultaneously produces effects comparable to those resulting from the fzr-1 mutation. Co-mutating ZYG-1 and SAS-5 degrons led to zyg-1 suppression at a level nearly equivalent to the ZYG-1^DB3(2A) mutation alone (Fig. 2B,C, Table 1). Consistently, the effects on centrosomal ZYG-1 levels are comparable between the ZYG-1^DB3(2A); SAS-5^KEN(3A) double mutation (1.55±0.42 fold; mean±s.d.) and ZYG-1^DB3(2A) single mutation (1.48±0.42 fold; Fig. 2D). Since SAS-5 acts downstream of ZYG-1, mutating the SAS-5 KEN-box did not affect centrosomal ZYG-1 (1.02±0.3 fold) although it increased SAS-5 levels (Fig. 2D,E). Thus, it seems unlikely that SAS-5 stabilization has an indirect effect on increased centrosomal ZYG-1.

Our data illustrate the epistatic relationship between zyg-1 and sas-5 in centrosome assembly. Co-mutating ZYG-1 and SAS-5 degrons produces impacts that are comparable to the ZYG-1^DB3(2A) single mutation and less potent than the fzr-1 mutation. These results suggest that APC/C^FZR-1 targets additional centrosome factors, likely acting upstream of ZYG-1, that support ZYG-1 activity. An intriguing candidate is SPD-2, which promotes centrosomal targeting of ZYG-1 (Shimanovskaya et al., 2014). In support of this, C. elegans SPD-2 contains putative D-boxes and Drosophila Spd2 has been shown to be an APC/C^Fzr/FZR-1 substrate (Meghini et al., 2016).

Although our data suggest APC/C^FZR-1 controls ZYG-1 stability via D-box3, mutating D-box3 alone produces a modest effect on ZYG-1 levels (Fig. 2D), raising the possibility that ZYG-1 levels are regulated via additional degrons (D-boxes or non-canonical degrons) recognized by APC/C, or by other proteolytic pathways. Indeed, ZYG-1 levels are known to be regulated by SCF^Slimb/βTrCP-mediated proteolysis (Peel et al., 2012), the mechanism of which is conserved (Cunha-Ferreira et al., 2009; Holland et al., 2010; Rogers et al., 2009). In Drosophila and human cells, the SCF^Slimb/βTrCP E3 ubiquitin ligase controls Plk4 abundance through the conserved Slimb/βTrCP-binding motif (DSGxxS/T) within Plk4, and autophosphorylation of this motif is critical for SCF^Slimb/βTrCP binding and Plk4 degradation (Cunha-Ferreira et al., 2013; Guderian et al., 2010; Klebba et al., 2013).

In C. elegans, although inhibiting LIN-23 (the Slimb/βTrCP homolog) stabilizes ZYG-1, a direct involvement of the ZYG-1 Slimb/βTrCP-binding (SB) motif (334DSGxxT339) in ZYG-1 stability remains unclear (Peel et al., 2012). We thus asked whether the ZYG-1 SB motif influences ZYG-1 levels. Toward this, we generated zyg-1(it25) mutant strains carrying alanine substitution mutations for two critical residues within the ZYG-1 SB motif by CRISPR/Cas9 genome-editing (Fig. 3A), equivalent to the known phosphosites (DSGxxT) of the SB motif in Plk4 (Cunha-Ferreira et al., 2013; Guderian et al., 2010; Holland est al., 2010; Klebba et al., 2013). Single alanine substitution (S335A or T339A) led to an ∼3-fold increase in embryonic viability in zyg-1(it25) mutants compared to controls at 22.5°C (Fig. 3B, Table 1). Intriguingly, mutating both residues (334DAGxxA339), termed the ZYG-1^SB(2A) mutation, restored embryonic viability to zyg-1(it25) mutants to >10-fold higher than controls (Fig. 3B, Table 1). Therefore, both S335 and T339 in the ZYG-1 SB motif are critical for ZYG-1 degradation, supporting a model that SCF^Slimb/βTrCP targets ZYG-1 for proteasomal degradation through the ZYG-1 SB motif. Our data substantiate the conserved mechanism by which Plk4/ZYG-1 levels are regulated by SCF^Slimb/βTrCP-mediated proteolysis through the SB motif.

Given that both APC/C^FZR-1 and SCF^Slimb/βTrCP regulate ZYG-1 levels, we asked whether co-mutating ZYG-1 SB and D-box3 motifs could enhance ZYG-1 protection from degradation, and produced a strain carrying a double mutation of ZYG-1 SB and D-box3 motifs at the endogenous locus (ZYG-1^SB+DB3(4A); Fig. 3A). The ZYG-1 SB (334DSGxxT339) and D-box3 (349RxxL352) motifs reside in close proximity within the L1 domain (Fig. 3A). Whereas neither motif is conserved in closely related nematodes (Fig. S2D), the flanking region of these motifs appears to be somewhat conserved (Klebba et al., 2013).

In zyg-1(it25) background, co-mutating ZYG-1 SB and D-box3 motifs further increases embryonic viability and bipolar spindles, compared to ZYG-1^SB(2A) or ZYG-1^DB3(2A) single mutation (Fig. 3B,C, Table 1). Consistently, mutating both degrons leads to further elevated centrosomal ZYG-1 levels (1.52±0.37 fold; mean±s.d.) than ZYG-1^SB(2A) (1.20±0.31 fold) or ZYG-1^DB3(2A) (1.29±0.34 fold) mutation (Fig. 3D). Thus, co-inhibiting APC/C^FZR-1 and SCF^Slimb/βTrCP enhances ZYG-1 stability in an additive manner, suggesting that APC/C^FZR-1 and SCF^Slimb/βTrCP-mediated proteolysis cooperatively controls ZYG-1 stability via conserved motifs.

Finally, we asked whether ZYG-1^DB3(2A) mutation influences ZYG-1 function through the 26S proteasome (Fig. 3E). We depleted the proteasome component RPT-4 using rpt-4(RNAi) to partially inhibit proteasome function, allowing completion of meiosis and mitotic cycles (Song et al., 2011). Given that rpt-4(RNAi) leads to increased centrosomal ZYG-1 (Peel et al., 2012), we examined how proteasomal inhibition by rpt-4(RNAi) affected ZYG-1 function in centrosome duplication. As expected, rpt-4(RNAi) in zyg-1(it25) mutants significantly restores bipolar spindles compared to control RNAi (50±28% versus 15±4%; 3.33 fold; mean±s.d.). By contrast, rpt-4(RNAi) combined with ZYG-1^DB3(2A) mutation in zyg-1(it25) mutants led to a small increase in bipolarity relative to control RNAi (70±8% vs 55±21%; 1.16 fold). Similarly, rpt-4(RNAi) with ZYG-1^SB(2A) (70±25% versus 56±5%; 1.25 fold) or ZYG-1^SB+DB3(4A) (70±5% vs 61±4%; 1.15 fold) mutation exhibits slightly higher bipolarity than controls. These results further support that ZYG-1^DB3(2A) mutation renders ZYG-1 partially resistant to proteasomal degradation, consistent with ZYG-1^SB(2A) and ZYG-1^DB3(2A) mutations partially blocking proteolysis, leading to ZYG-1 stabilization and elevated centrosomal ZYG-1.

In summary, our work reports, for the first time, that centrosomal ZYG-1/Plk4 levels are influenced by APC/C^FZR-1. Our data also suggest that two E3 ubiquitin ligases, APC/C^FZR-1 and SCF^Slimb/βTrCP, cooperate for timely degradation of ZYG-1 during early cell cycle in C. elegans embryos, preventing aberrant centrosome numbers (Fig. S1H,I). Along with proteolytic regulation of ZYG-1, ZYG-1 activity is known to be modulated by kinase and phosphatases (Kitagawa et al., 2011; Medley et al., 2017b; Peel et al., 2017; Song et al., 2011). Thus, multiple mechanisms appear to govern precise control of ZYG-1 activity, which might account for rare observations of centrosome amplification in our ZYG-1 degron mutants (Fig. S1I). Notably, centrosome amplification phenotypes were observed in neither fzr-1 nor SAS-5^KEN(3A) mutants (Medley et al., 2017a). Finally, our results suggest that APC/C^FZR-1 targets additional centrosome factors acting upstream or parallel of ZYG-1. Additional work is required to further understand the regulatory mechanisms of APC/C^FZR-1 in centrosome assembly. It will be also interesting to explore the role of APC/C^Cdh1/FZR-1 in regulating Plk4 stability and its association with human cancers.

The C. elegans strains used in this study are listed in Table S1. All strains were derived from the wild-type Bristol N2 strain (Brenner, 1974; Church et al., 1995) and maintained on MYOB plates seeded with Escherichia coli OP50 at 16 or 19°C. Some strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). For genetic analysis, individual L4 hermaphrodites were transferred to new plates and allowed to produce progeny for 24–48 h at the temperatures indicated. Progeny were allowed to develop for 18–24 h before counting the number of larvae and dead eggs. RNAi feeding was performed as described previously (Kamath and Ahringer, 2003) and the L4440 empty feeding vector was used as a negative control. Animals were fed RNAi for 12–18 h prior to imaging.

Immunofluorescence and confocal microscopy were performed as described previously (Song et al., 2008). Primary and secondary antibodies against the following proteins were used at 1:3000 dilutions: DM1a (Sigma, #T9026), ZYG-1 (Stubenvoll et al., 2016), SAS-4 (Song et al., 2008), Myc (Thermo Fisher Scientific, #PA1-981; Fig. S1E), and V5 (MBL, #M167-3; Fig. S1F), Ollas (Thermo Fisher Scientific, #MA5-16125; Fig. S1G), Alexa Fluor 488 and 568-conjugated secondary antibodies (Thermo Fisher Scientific, #A11001, A11004, A11006, A11034, A11036). Time-lapse recordings for GFP::SAS-7 embryos were performed as described previously (Medley et al., 2017b). Images of GFP::SAS-7 were acquired at 30 s intervals starting from prometaphase until the completion of first cell division. Confocal microscopy was performed using a Nikon Eclipse Ti-U microscope equipped with a Plan Apo 60×1.4 NA lens, a spinning disk confocal (CSU X1) and a Photometrics Evolve 512 camera. MetaMorph software (Molecular Devices, Sunnyvale, CA, USA) was used for image acquisition and quantification of the fluorescence intensity, and Adobe Photoshop/Illustrator 2020 for image processing. To quantify centrosomal signals, the average intensity within an 8- or 9-pixel (1 pixel=0.151 µm) diameter region was recorded for the highest intensity of the focal plane within an area centered on the centrosome. The average intensity within a 25-pixel diameter region drawn outside of the embryo was used for background subtraction.

For genome editing, we used the co-CRISPR technique described previously (Arribere et al., 2014; Paix et al., 2015). crRNA was designed using the CRISPOR web server (http://crispor.tefor.net/; Concordet and Haeussler, 2018). Animals were microinjected with a mixture of commercially available SpCas9 (IDT, Coralville, IA) and custom-designed oligonucleotides (Tables S2 and S3) including crRNAs at 0.4–0.8 µg/µl, tracrRNA at 12 µg/µl, and single-stranded DNA oligonucleotides at 25–100 ng/µl. The amount of crRNA was tripled for low-efficiency crRNAs (ZYG-1 D-box1 and D-box4, Ollas::FZR-1). After injection, we screened for dpy-10(cn64) II/+ rollers in F1 progeny and genotyping F2. The genome editing was verified by Sanger Sequencing (GeneWiz, South Plainfield, NJ).

IP experiments were performed as described in Stubenvoll et al. (2016). Embryos were collected from young gravid worms using hypochlorite treatment (1:2:1 ratio of M9 buffer, 5.25% sodium hypochlorite and 5 M NaOH), frozen in liquid nitrogen and stored at −80°C until use. Embryos were suspended in lysis buffer [50 mM HEPES, pH 7.4, 1 mM EDTA, 1 mM MgCl₂, 200 mM KCl, and 10% glycerol (v/v)] with complete protease inhibitor cocktail (Roche) and MG132 (Tocris, Avonmouth, Bristol, UK), milled for 5 min (repeated three times) at 30 Hz using a Retsch MM 400 mixer-mill (Verder Scientific, Newtown, PA), then sonicated for 3 min in ultrasonic bath (Thermo Fisher Scientific). Lysates were spun at 45,000 rpm for 45 min using a Sorvall RC M120EX ultracentrifuge (Thermo Fisher Scientific), then the supernatant was recovered to clean tubes. The equivalent amount of total protein lysates was used for IP. Embryonic protein lysates mixed with anti-GFP or anti-Myc magnetic beads (MBL, # D153-11; M047-11) were incubated by rotation for 1 h at 4°C, and washed (three times for 5 min each time) with PBST (PBS plus 0.1% Triton-X 100). IP with beads and input samples were resuspended in 2× Laemmli sample buffer (Sigma), and boiled for 5 min before fractionating on a 4–12% NuPAGE Bis-Tris gel (Invitrogen). Proteins on a gel were transferred to a nitrocellulose membrane and analyzed by using the antibodies against the indicated proteins at 1:3000–10,000 dilutions: DM1a (Sigma, #T9026), GFP (Roche, #11814460001), Myc (Genscript, #A00704), SAS-5 (Medley et al., 2017b), IRDye secondary antibodies (LI-COR Biosciences). Blots were imaged using an Odyssey infrared scanner (LI-COR Biosciences), and analyzed using Image Studio software (LI-COR Biosciences).

Statistics were produced using R statistical software and presented as mean±standard deviation (s.d.). Dot plots were generated using the R ‘beeswarm’ package. In the dotplots, box ranges from the first through third quartile of the data. Thick bar indicates the median. The solid gray line extends 1.5 times from the inter-quartile range or to the minimum and maximum data point. All P-values were calculated using two-tailed unpaired t-tests: ns, not significant (P>0.05), *P<0.05, **P<0.01, ***P<0.001.

We thank members of Song laboratory for technical support and helpful discussions.