APC/CFZR-1 Controls SAS-5 Levels To Regulate Centrosome Duplication in Caenorhabditis elegans

As the primary microtubule-organizing center, centrosomes play a key role in establishing mitotic bipolar spindles that secure correct transmission of genomic content. For the fidelity of cell division, centrosome number must be strictly controlled by duplicating only once per cell cycle. Proper levels of centrosome proteins are shown to be critical for normal centrosome number and function. Overexpressing core centrosome factors leads to extra centrosomes, while depleting these factors results in centrosome duplication failure. In this regard, protein turnover by the ubiquitin-proteasome system provides a vital mechanism for the regulation of centrosome protein levels. Here, we report that FZR-1, the Caenorhabditis elegans homolog of Cdh1/Hct1/Fzr, a coactivator of the anaphase promoting complex/cyclosome (APC/C), an E3 ubiquitin ligase, functions as a negative regulator of centrosome duplication in the C. elegans embryo. During mitotic cell division in the early embryo, FZR-1 is associated with centrosomes and enriched at nuclei. Loss of fzr-1 function restores centrosome duplication and embryonic viability to the hypomorphic zyg-1(it25) mutant, in part, through elevated levels of SAS-5 at centrosomes. Our data suggest that the APC/CFZR-1 regulates SAS-5 levels by directly recognizing the conserved KEN-box motif, contributing to proper centrosome duplication. Together, our work shows that FZR-1 plays a conserved role in regulating centrosome duplication in C. elegans.

The centrosome is a small, nonmembranous organelle that serves as the primary microtubule-organizing center in animal cells. Each centrosome consists of a pair of barrel-shaped centrioles that are surrounded by a network of proteins called pericentriolar material (PCM). During mitosis, two centrosomes organize bipolar spindles that segregate genomic content equally into two daughter cells. Thus, tight control of centrosome number is vital for the maintenance of genomic integrity during cell division, by restricting centrosome duplication to once, and only once, per cell cycle. Erroneous centrosome duplication results in aberrant centrosome number that leads to chromosome missegregation and abnormal cell proliferation, and is associated with human disorders including cancers and microcephaly (Nigg and Stearns 2011;Gönczy 2015).
Maintaining the proper levels of centrosome proteins is critical for normal centrosome number and function (Kleylein-Sohn et al. 2007;Strnad et al. 2007;Rogers et al. 2009;Tang et al. 2009Tang et al. , 2011Holland et al. 2010;Brownlee et al. 2011;Puklowski et al. 2011;Meghini et al. 2016;Levine et al. 2017). In light of this, protein turnover by proteolysis provides a key mechanism for regulating the abundance of centrosome factors. A mechanism regulating protein levels is their degradation by the 26S proteasome that catalyzes the proteolysis of polyubiquitinated substrates (Livneh et al. 2016). The anaphase promoting complex/cyclosome (APC/C) is a multi-subunit E3 ubiquitin ligase that targets substrates for degradation (Acquaviva and Pines 2006;Peters 2006;Chang and Barford 2014). The substrate specificity of the APC/C is directed through the sequential, cell-cycle-dependent activity of two coactivators, Cdc20/Fzy/FZY-1 (Hartwell and Smith 1985;Dawson et al. 1995;Kitagawa et al. 2002) and Cdh1/Fzr/Hct1/ FZR-1 (Schwab et al. 1997;Sigrist and Lehner 1997;Visintin et al. 1997;Fay et al. 2002). During early mitosis, Cdc20 acts as coactivator of the APC/C, and Cdh1 functions as coactivator to modulate the APC/ C-dependent events at late mitosis and in G1 (Irniger and Nasmyth 1997;Visintin et al. 1997;Fang et al. 1998;Prinz et al. 1998;Shirayama et al. 1998). Upregulated targets in Cdh1-deficient cells are shown to be associated with the genomic instability signature of human cancers, and show a high correlation with poor prognosis (Carter et al. 2006;García-Higuera et al. 2008). Furthermore, a mutation in SIL/STIL (a human homolog of SAS-5) linked to primary microcephaly (MCPH; Kumar et al. 2009) results in deletion of the Cdh1-dependent destruction motif (KEN-box), leading to deregulated accumulation of STIL protein and centrosome amplification (Arquint and Nigg 2014). In Drosophila, the APC/C Fzr/Cdh1 directly interacts with Spd2 through KEN-box recognition and targets Spd2 for degradation (Meghini et al. 2016). Therefore, the APC/C Cdh1/Fzr/Hct1 plays a critical role in regulating the levels of key centrosome duplication factors in mammalian cells and flies.
In C. elegans, FZR-1 has been shown to be required for fertility, cell cycle progression and cell proliferation during embryonic and postembryonic development via synthetic interaction with lin-35/Rb (Fay et al. 2002;The et al. 2015). However, the role of FZR-1 in centrosome assembly has not been described. In this study, we molecularly identified fzr-1 as a genetic suppressor of zyg-1. Our results suggest that APC/ C FZR-1 negatively regulates centrosome duplication, in part, through proteasomal degradation of SAS-5 in a KEN-box dependent fashion. Therefore, FZR-1, the C. elegans homolog of Cdh1/Hct1/Fzr, plays a conserved role in centrosome duplication.

MATERIALS AND METHODS
C. elegans strains and genetics A full list of C. elegans strains used in this study is listed in Supplemental Material, Table S1 in File S1. All strains were derived from the wild-type Bristol N2 strain using standard genetic methods (Brenner 1974;Church et al. 1995).
Strains were maintained on MYOB plates seeded with Escherichia coli OP50 and grown at 19°unless otherwise indicated. The fzr-1:: gfp::3xflag construct containing 21.6 Kbp of the fzr-1 59UTR and 6 Kbp of the fzr-1 39UTR was acquired from TransgenOme (construct number: 7127141463160758 F11, Sarov et al. 2012), which was used to generate the transgenic line, MTU10, expressing C-terminal GFPtagged FZR-1. For the generation of N-terminal GFP-tagged FZR-1 (OC190), we used Gateway cloning (Invitrogen, Carlsbad, CA) to generate the construct. Coding sequence of fzr-1 was PCR amplified from the cDNA clone yk1338f2, and cloned into pDONR221 (Invitrogen) and then the resulting entry clone was recombined into pID3.01 (pMS9.3), which is driven by the pie-1 promotor. The transgenes were introduced into worms by standard particle bombardment (Praitis et al. 2001). For embryonic viability and brood size assays, individual L4 animals were transferred to clean plates, and allowed to self-fertilize for 24 hr at the temperatures indicated. For brood size assays, this was repeated until animals no longer produced embryos. Progeny were allowed at least 24 hr to complete embryogenesis before counting the number of progeny. The fzr-1(RNAi) experiments were performed by RNAi soaking . To produce dsRNA for RNAi soaking, we amplified a DNA template from the cDNA clone yk1338f2 using the primers 59-ATGGATGAGCAACCGCC-39 and 59-GCACTGTACGTAAAAAGTGATC-39 that contained a T7 promoter sequence at their 59 ends. In vitro transcription was performed using the T7-MEGAscript kit (Thermo-Fisher, Hanover park, IL). L4 animals were soaked overnight in M9 buffer containing either 0.1-0.4 mg dsRNA/ml or no dsRNA (control).
CRISPR/CAS-9 mediated genome editing For genome editing, we used the co-CRISPR technique as previously described in C. elegans (Arribere et al. 2014;Paix et al. 2015). In brief, we microinjected N2 and zyg-1(it25) animals using a mixture containing recombinant SpCas9 (Paix et al. 2015), crRNAs targeting sas-5 and dpy-10 at 0.4-0.8 mg/ml, tracrRNA at 12 mg/ml, and single-stranded DNA oligonucleotides to repair sas-5 and dpy-10 at 25-100 ng/ml. Microinjection was performed using the XenoWorks microinjector (Sutter Instruments, Novato, CA) with a continuous pulse setting at 400-800 hPa. All RNA and DNA oligonucleotides used in this study were synthesized by Integrated DNA Technologies (IDT, Coralville, IA) and are listed in Table S2 in File S1. As we were unable to engineer a silent mutation into the PAM sequence used by the sas-5 crRNA, we introduced six silent mutations to sas-5 (aa 201-206; Figure 5A) by mutating 8 out of 20 the nucleotides that comprise the sas-5 crRNA, in order to disrupt Cas9 recognition after homology-directed repair. After injection, animals were allowed to produce F1 progeny that were monitored for the presence of dpy-10(cn64)/+ rollers. To identify the sas-5 KEN-to-3A mutation, we extracted genomic DNA from broods containing the highest frequency of F1 rollers. Using the primers, forward: 59-TGCCCAAAATACGACAACG-39 and reverse: 59-TACACTACTCACGTCTGCT-39, we amplified the region of sas-5 containing the KEN-box sequence. As the repair template for the sas-5 KEN-to-3A mutation introduces an Hpy8I restriction enzyme (NEB, Ipswich, MA) cutting site, we used an Hpy8I enzyme digestion to test for the introduction of our targeted mutation. After isolating homozygotes based on the Hpy8I cutting, we confirmed the SAS-5 KEN-to-3A mutation by genomic DNA sequencing. Sequencing revealed that several lines were homozygous for the SAS-5 KEN-to-3A mutation ( Figure 5A and Table S1 in File S1). However, the strain MTU14, contained all of the silent mutations that we designed to disrupt Cas9 recognition without affecting the KEN-box ( Figure 5A and Table S1 in File S1). Thus, we used MTU14 as a control for our assays.

Cytological analysis
To perform immunostaining, the following antibodies were used at 1:2000-3000 dilutions: a-Tubulin  (Stubenvoll et al. 2016) 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. Images were acquired using MetaMorph software (Molecular Devices, Sunnyvale, CA). MetaMorph was used to draw and quantify regions of fluorescence intensity and Adobe Photoshop CS6 was used for image processing. To quantify centrosomal SAS-5 signals, the average intensity within an 8-pixel (1 pixel = 0.151 mm) diameter region was measured within an area centered on the centrosome and the focal plane with the highest average intensity was recorded. Centrosomal TBG-1 (g-tubulin) levels were quantified in the same manner, except that a 25-pixel diameter region was used. For both SAS-5 and TBG-1 quantification, the average fluorescence intensity within a 25-pixel diameter region drawn outside of the embryo was used for background subtraction.

Statistical analysis
All P-values were calculated using two-tailed t-tests assuming equal variance among sample groups. Statistics are presented as Average 6 SD unless otherwise specified. Data were independently replicated at least three times for all experiments and subsequently analyzed for statistical significance.

Data availability
All strains used in this study are available upon request. File S1 contains the following: Figure S1, Centrosome-associated TBG-1 levels are unaffected in fzr-1(bs31) and sas-5 KEN-to-3A mutant embryos; Figure S2, Brood size in sas-5 KEN-to-3A and fzr-1(bs31) mutants; Figure S3, SAS-5 levels are increased in sas-5 KEN-to-3A mutants; Table S1, List of strains used in this study; Table S2, List of oligonucleotides used for CRISPR/ Cas9 genome editing.
The fzr-1(bs38) mutation produces a missense mutation (R65C) at the conserved C-box of FZR-1. The C-box is known to be crucial for the physical interaction between FZR-1 and other APC/C subunits (Schwab et al. 2001;Thornton et al. 2006;Chang et al. 2015;Zhang et al. 2016). Thus, both fzr-1(bs31) and fzr-1(bs38) mutations appear to affect conserved domains that are critical for the function of the APC/C complex, suggesting that FZR-1 might regulate centrosome duplication through the APC/C complex.

FZR-1 localizes to nuclei and centrosomes during early cell division
To determine where FZR-1 might function during the early cell cycle, we produced two independent transgenic strains that express FZR-1 tagged with GFP at the N-or C-terminus (see Materials and Methods).
To label microtubules, we mated GFP-tagged FZR-1 transgenic animals with the mCherry::b-tubulin expressing line, and performed 4D timelapse movies to observe subcellular localization of GFP::FZR-1 throughout the first cell cycle (Figure 2A). Confocal imaging illustrates that during interphase and early mitosis, GFP::FZR-1 is highly enriched at the nuclei. After the nuclear envelope breaks down (NEBD), GFP:: FZR-1 diffuses to the cytoplasm and reappears to the nuclei at late mitosis when the nuclear envelop reforms. After NEBD, GFP::FZR-1 becomes apparent at spindle microtubules, and centrosomes that colocalize with SPD-2, a centrosome protein ( Figure 2B). Both GFP-tagged FZR-1 transgenic embryos exhibit similar subcellular distributions, except a slight difference in fluorescent intensity (data not shown). While we do not exclude the possibility that FZR-1 functions in the cytoplasm to regulate cellular levels of centrosome factors, our observations suggest that C. elegans FZR-1 might direct APC/C activity at centrosomes during late mitosis in early embryos, which is consistent with the role of FZR-1 as the coactivator of the APC/C at late mitosis in other organisms (Raff et al. 2002;Zhou et al. 2003;Meghini et al. 2016).
Loss of FZR-1 results in elevated SAS-5 levels Next, we wanted to understand how FZR-1 contributes to centrosome duplication. Since FZR-1 appears to function through the APC/C complex in centrosome assembly, we hypothesized that the APC/C FZR-1 specifically targets one or more centrosome regulators for ubiquitinmediated degradation. If that is the case, depleting FZR-1 should protect substrates from degradation leading to accumulation of target proteins. To identify a direct substrate of APC/C FZR-1 that regulates centrosome assembly, we utilized the conserved FZR-1 coactivator specific recognition motif, KEN-box, to screen for a potential substrate (Glotzer et al. 1991;Pfleger and Kirschner 2000;Song and Rape 2011).
Elevated protein levels might influence centrosome-associated SAS-5 levels in fzr-1(bs31) mutants. To determine how inhibition of the APC/C FZR-1 affected overall protein levels, we performed quantitative western blot analysis using embryonic protein lysates and antibodies against centrosome proteins ( Figure 4C). Our data indicate that fzr-1(bs31) embryos possess increased SAS-5 levels ($1.5-fold), relative to wild-type embryos, while the levels of SAS-6 and TBG-1 are not significantly affected in fzr-1(bs31) mutants ( Figure 4C). Our observation on the SAS-6 levels in fzr-1(bs31) mutants is consistent with previous work by Miller et al. (2016), showing no increase in SAS-6 levels by the mat-3(bs29)/APC8 mutation that inhibits the APC/C function. These results suggest that C. elegans utilizes a different mechanism to control SAS-6 levels, unlike Human SAS-6, which is regulated by the APC/ C-mediated proteolysis (Strnad et al. 2007). Furthermore, our immunoprecipitation suggests a physical interaction between SAS-5 and FZR-1 in C. elegans embryos ( Figure 4D), supporting that SAS-5 might be a direct substrate of the APC/C FZR-1 . Consistent with our results in this study, prior study has shown that inhibiting the 26S proteasome leads to increased levels of SAS-5 . Thus, SAS-5 levels are likely to be controlled through the ubiquitin-proteasome system.
Collectively, our data show that the fzr-1 mutation leads to a significant increase in both cellular and centrosomal levels of SAS-5, suggesting that the APC/C FZR-1 might control SAS-5 levels via ubiquitin-mediated proteasomal degradation to regulate centrosome assembly in the C. elegans embryo.

Mutation of the KEN-box stabilizes SAS-5
If the APC/C FZR-1 directly targets substrates for destruction via the conserved KEN-box, mutating this motif should cause substantial resistance to ubiquitination-mediated degradation. To determine whether the APC/C FZR-1 targets SAS-5 through the KEN-box motif, we mutated the KEN-box at the endogenous sas-5 locus. By using CRISPR/CAS-9 mediated genome editing (Paix et al. 2015), we generated mutant lines (sas-5 KEN-to-3A ) carrying alanine substitutions of the SAS-5 KEN-box ( Figure 5A). The sas-5 KEN-to-3A mutant embryo exhibits no obvious cell cycle defects or embryonic lethality (Table 1), consistent with fzr-1 mutants (Kemp et al. 2007). sas-5 KEN-to-3A animals exhibit a slightly reduced ($80%) and irregular distribution of brood size within the population ( Figure S2 in File S1). Reduced brood size and slow growth phenotypes were previously reported in fzr-1 mutant alleles (Fay et al. 2002;Kemp et al. 2007).
Interestingly, although either inhibiting FZR-1 or mutating KENbox influences SAS-5 stability at a comparable level, we observe a notable difference in the suppression level by these two mutations. Weaker suppression by the sas-5 KEN-to-3A mutation suggests that the APC/C FZR-1 might target additional substrates that cooperatively support the zyg-1 suppression. In this scenario, APC/C FZR-1 might target other centrosome proteins outside core duplication factors through the conserved degron motifs, such as destruction (D)-box and KEN-box (Glotzer et al. 1991;Pfleger and Kirschner 2000). Alternatively, APC/ C FZR-1 might target additional core centrosome factors through other recognition motifs other than KEN-box, such as D-box (Glotzer et al. 1991) or unknown motif in the C. elegans system. In humans and flies, APC/C Cdh1/Fzr has been shown to regulate the levels of STIL/SAS-5, Spd2, HsSAS-6, and CPAP/SAS-4 (Strnad et al. 2007;Tang et al. 2009;Arquint and Nigg 2014;Meghini et al. 2016). While C. elegans homologs of these factors, except SAS-5, lack a KEN-box, all five centrosome proteins contain at least one putative D-box. An intriguing possibility, given the strong genetic interaction observed between fzr-1 and zyg-1, is that ZYG-1 could be a novel substrate of APC/C FZR-1 . Additional work will be required to understand the complete mechanism of APC/C FZR-1 -dependent regulation of centrosome duplication in C. elegans. In summary, our study shows the APC/C FZR-1 -dependent proteolysis of SAS-5 partially contributes to the suppression of the zyg-1 mutants, and we report that FZR-1 functions as a negative regulator of centrosome duplication in C. elegans.  Mutation of the SAS-5 KEN-box leads to increased SAS-5 levels at centrosomes and restores centrosome duplication to zyg-1(it25) mutants. (A) SAS-5 contains a KEN-box (aa 213-216) motif. Mutations (red) are introduced at multiple sites to make alanine substitutions (AAA; 3A) for the KEN-box and additional silent mutations for the CRISPR genome editing (see Materials and Methods). The KEN-box is highlighted in yellow. Note that the sas-5 KEN-to-KEN mutation contains the wild-type SAS-5 protein. (B) Quantification of bipolar spindle formation during the second cell cycle in zyg-1(it25); sas-5 KEN-to-KEN and zyg-1(it25); sas-5 KEN-to-3A embryos at 22.5°. zyg-1(it25); sas-5 KEN-to-3A double mutant embryos produce bipolar spindles at a higher rate (67.5 6 16.3%, n = 124, P = 0.02) than zyg-1(it25); sas-5 KEN-to-KEN controls (35.1 6 10.7%, n = 164). n is the number of blastomeres. Average values are presented and error bars are SD. (C) Centrosomes stained for SAS-5 (green) during the first anaphase. Bar, 5 mm. (D) Quantification of centrosomal SAS-5 levels during the first anaphase. We used two independently generated sas-5 KEN-to-3A mutant lines to quantify SAS-5 levels (MTU11 and 12, Table S1 in File S1). SAS-5 levels at centrosomes are normalized to the average fluorescence intensity in wild-type centrosmes. Mutating the SAS-5 KEN-box leads to increased levels of centrosomal SAS-5 in both MTU11 (1.54 6 0.63-fold, n = 16; P = 0.04) and MTU12 (1.48 6 0.50 fold, n = 8; P = 0.03), compared to wild type (1.00 6 0.29fold; n = 24). Consistently, there are a significant increase in centrosomal SAS-5 levels in both zyg-1(it25); sas-5 KEN-to-3A double mutant lines (MTU13: 0.85 6 0.24-fold, n = 16; P = 0.01 and MTU15: 1.09 6 0.59-fold; n = 12; P = 0.03), compared to zyg-1(it25); sas-5 KEN-to-KEN control that contains reduced levels of centrosomal SAS-5 (0.67 6 0.20-fold; n = 36). n is the number of centrosomes. Each dot represents a centrosome. Box ranges from the first through third quartile of the data. Thick bar indicates the median. Solid gray line extends 1.5 times the interquartile range, or to the minimum and maximum data point. Ã P , 0.05, ÃÃ P , 0.01 (two-tailed t-test).