A vertebrate N-end rule degron reveals that Orc6 is required in mitosis for daughter cell abscission

Origin recognition complex proteins form a hexameric complex that helps to load replication initiation factors at origins of replication (Bell and Stillman, 1992; Rowles and Blow, 1997; Quintana and Dutta, 1999). Orthologues occur in Saccharomyces cerevisiae (Li and Herskowitz, 1993), Drosophila melanogaster (Gossen et al., 1995), Xenopus laevis (Gillespie et al., 2001), or Homo sapiens (Dhar and Dutta, 2000) and probably exist in all eukaryotes. Orc1–Orc6 are essential for the initiation of DNA replication in all systems so far tested (Duncker et al., 2009) but may also have additional biological functions (Bell et al., 1995; Pak et al., 1997; Bell, 2002; Prasanth et al., 2004; Shimada and Gasser, 2007). The most striking example is Orc6 (Prasanth et al., 2002; Chesnokov, 2007), which has no homology with the other origin recognition complex proteins and has rapidly diverged from budding yeast to humans (Dhar and Dutta, 2000).

Mammalian Orc6 has been implicated in mitosis because its depletion using RNAi causes centrosome amplification and multinuclear cell division (Prasanth et al., 2002). However, Orc6 depletion from asynchronously dividing cells triggers delayed and diverse mitotic phenotypes whose prominence depends on the duration of RNAi treatment. Genetic experiments in budding yeast fail to identify a corresponding role (Semple et al., 2006). Moreover, aberrant replication can itself induce G2 checkpoint activation (Dodson et al., 2004), which when persistent, provokes both centrosome amplification and abnormal cell division (Meraldi et al., 2002). Thus, whether the mitotic abnormalities triggered by mammalian Orc6 depletion reflect direct functions of the protein or are simply the indirect consequences of defective replication remains unresolved. To address this issue, we created a temperature-sensitive N-end rule degron form of vertebrate Orc6 (degron-Orc6) in the avian DT40 cell line using a method we have recently established (Su et al., 2008).

We made DT40 cells in which endogenous Orc6 was replaced by degron-Orc6 using the approach described in Fig. S1 (Su et al., 2008) to fuse endogenous Orc6 in frame at its 5′ end to a FLAG tag and a thermosensitive N-end rule mutant of dihydrofolate reductase (Dohmen et al., 1994; Labib et al., 2000; Dohmen and Varshavsky, 2005). In cells exclusively expressing degron-Orc6 (Orc6^−/deg), the fusion protein is rapidly degraded to <10% of its initial levels within 90 min when cells are shifted from a permissive temperature of 35°C to a nonpermissive temperature of 42°C (Fig. 1 A). Several lines of evidence confirm that degron-Orc6 functionally replaces endogenous Orc6. The viability and proliferation of Orc6^−/deg cells is similar to that of controls (Fig. 1 B) at 35°C. However, when shifted to 42°C, Orc6^−/deg cells fail to survive. Complementation with an EGFP-Orc6 fusion protein (which is similar in size to the degron-Orc6 fusion protein but thermostable at 42°C; Fig. 1 D) was sufficient to revert this phenotype (Fig. 1 C). This confirms that the death of Orc6^−/deg cells at 35°C is a direct effect of Orc6 depletion. Interestingly, complementation with EGFP-Orc6Δ25, a mutant lacking an evolutionarily conserved domain spanning the C-terminal 25 residues of Orc6 (Balasov et al., 2009), did not rescue cell viability (Fig. 1 C).

To define the consequences of Orc6 deficiency during DNA replication, we conditionally depleted degron-Orc6 from cells synchronously released into the S phase of the cell cycle. Centrosome number increased sharply from 16 h after degron-Orc6 depletion (22.8 ± 2.9% vs. control 6.6 ± 2.2%; Fig. 2 A), just as after depletion of the DNA recombinase RAD51 (Fig. S2). Staining with β-tubulin, pericentrin, or γ-tubulin (Fig. 2 B) detected multipolar spindles and supernumerary centrosomes 24 h after degron-Orc6 depletion (79.8 ± 4.7%, mean ± SEM), which were absent from Orc6^+/− parental cells at 42°C or Orc6^−/deg control cells at 35°C (9.8 ± 1.9% and 5.6 ± 1%, respectively). The 16–24-h interval after S-phase release when these anomalies occur corresponds to ∼1.5–2 times the typical 10-h doubling time of DT40 cells (Li and Dodgson, 1995), which is indicative of cell cycle prolongation. Indeed, the fraction of degron-Orc6–depleted cells with 4N DNA content increased from 25 to 66% over 24 h (Fig. 2 C), but there was no increase in M-phase cells stained with anti-MPM2 (1.3% at 24 h). Thus, these observations suggest that degron-Orc6 depletion during DNA replication causes arrest in the G2 phase, allowing the accumulation of supernumerary centrosomes and multipolar spindles.

Supporting this notion, phosphorylated Chk1, a G2 checkpoint effector that is induced by defective replication (Heffernan et al., 2002; Xiao et al., 2003), accumulated in degron-Orc6–depleted cells (Fig. 2 D). Treatment of degron-Orc6–depleted cells with caffeine, an inhibitor of G2 checkpoint kinases known to bypass G2 arrest, decreased the percentage of 4N cells from 66 to 42% after 24 h; it correspondingly reduced supernumerary centrosomes and multipolar spindle formation between 16 and 20 h after degron-Orc6 depletion (Fig. 2 E). Finally, cells synchronized with the microtubule poison nocodazole (Nz; Fig. 2 F) before release into mitosis without degron-Orc6 did not exhibit supernumerary centrosome formation (Fig. 2 G). Collectively, these findings argue against a direct role of Orc6 in centrosome duplication or mitotic spindle formation and suggest instead that the occurrence of centrosome and spindle anomalies in degron-Orc6–depleted cells is the result of persistent G2 arrest after defective DNA replication.

Does Orc6 have a direct role in cytokinesis independent of its role in DNA replication? To answer this question, we depleted degron-Orc6 from Nz-arrested cells by incubation at 42°C before release into mitosis (as in Fig. 2 F). Representative frames from serial time-lapse images show that Orc6^−/deg cells lacking degron-Orc6 displayed asymmetric cell cleavage and failure in cytokinesis in contrast to the symmetric division of controls (Fig. 3, A–E). Moreover, during cytokinetic furrow ingression, degron-Orc6–depleted cells exhibited cortical blebbing, furrow regression, and failure to divide, forming multinucleated cells (Fig. 3, B and C). Defective cell division was strikingly more frequent in cells lacking full-length degron-Orc6 than in controls (Fig. 3 F, compare Orc6^−/deg and Orc6^{−/deg/EGFP-Orc6Δ25} with Orc6^−/deg and Orc6^{−/deg/EGFP-Orc6}). Full-length EGFP-Orc6 could complement these defects in Orc6^−/deg cells, whereas EGFP alone or the EGFP-Orc6Δ25C mutant could not (Fig. 3, E and F). Thus collectively, our findings show that the C-terminal region of Orc6 is essential during mitosis for the late stages of cell division and, in particular, that it promotes correct abscission and cell separation. However, EGFP-Orc6Δ25C (like EGFP-Orc6) decorates chromosomes from anaphase to telophase independent of cell division fate and also accumulates in interphase nuclei (Fig. 3, G and H), suggesting that mislocalization does not explain its defective function.

Abscission defects late in cytokinesis are difficult to detect by conventional light microscopy. To determine whether degron-Orc6 depletion during early mitosis induces delayed abscission, we used a fluorescence-bleaching approach to test whether there was a cytoplasmic connection between daughter cells transfected with free monomomeric EGFP (Fig. 4 A). We first depleted degron-Orc6 during mitosis (as in Fig. 2 F) and, during division, photobleached EGFP from one daughter cell and measured over time the mean fluorescence intensity (MFI) of EGFP in its partner (Fig. 4 B). As long as a cytoplasmic connection exists between the two cells, the MFI of the unbleached cell will equilibrate with that of its partner (see example in Fig. 4 C). When abscission is complete and no cytoplasmic connection exists, the MFI of the bleached and unbleached cells will not equilibrate (Fig. 4 C). 2 h after release into mitosis, 92% of Orc^−/+/EGFP cells (n = 25) have completed abscission, whereas only 24% of Orc6^−/deg/EGFP cells (n = 25) lacking degron-Orc6 did so in the same time frame (Fig. 4 D). To exclude that this difference was not simply the result of delays in mitosis before the abscission step, we measured the precise timing of abscission and the completion of cell separation by modifying a recently developed fluorescence-bleaching assay (Steigemann and Gerlich, 2009; Steigemann et al., 2009). Repeated bleaching of EGFP in one daughter cell decreases the MFI in the unbleached daughter cell for as long as they maintain their cytoplasmic connection. The decrease in MFI terminates after abscission, when the daughter cells separate (Fig. 4 E). When comparing Orc6^−/+/EGFP and Orc6^−/deg/EGFP cells, we find a significant delay in the timing of abscission after degron-Orc6 depletion (78 ± 3.8 min, mean ± SEM; n = 15) compared with control cells (44 ± 1.1 min, mean ± SEM; n = 15; Fig. 4, F–H).

This delay could not be explained by the persistence of DNA bridges between the two daughter cells that might impede abscission (Fig. S3, A and B). Moreover, the Aurora B kinase, which delays abscission when chromosome bridges are present (Steigemann et al., 2009), localizes normally to the midbody in degron-Orc6–depleted cells and did not colocalize with DAPI-stained DNA (Fig. S3 B).

How vertebrate Orc6 may regulate cytokinesis is unclear, although we show that a C-terminal region of 25 residues is necessary for abscission. Silencing of the D. melanogaster Orc6 orthologue also causes multinuclear cell division; its conserved C-terminal region binds to the septin “peanut” (Chesnokov et al., 2003; Huijbregts et al., 2009). However, we could not detect an interaction between vertebrate Orc6 and the peanut homologue septin 7 (Fig. S3). Thus, vertebrate Orc6 might exert its functions during abscission through a different mechanism.

In summary, we have exploited an N-end rule degron to show that vertebrate Orc6 is necessary during mitosis for symmetric cell division and cytokinetic abscission; this is distinct from its established function in DNA replication. In contrast, our results suggest that centrosome amplification induced by Orc6 depletion reflects an indirect consequence of aberrant replication, rather than the direct role suggested by previous studies (Prasanth et al., 2002; Chesnokov et al., 2003). Thus, our work illustrates the utility of vertebrate N-end rule degrons to unravel complex phenotypes associated with different functions of a single protein during different phases in the cell cycle.

Our work highlights an evolutionary difference between vertebrate and yeast Orc6 (Fig. 5). Conditional genetic experiments in yeast indicate that yeast Orc6 exclusively functions during DNA replication. In contrast, our findings indicate that Orc6 has acquired an additional and distinct role in the completion of cell division during the transition from unicellular to multicellular organisms, providing fresh insight into the evolution of the mechanisms that coordinate DNA replication and mitosis in eukaryotes.

The Gallus gallus Orc6l (ENSGALT00000006929, 1261 bp; ENSGALP00000006918, 267 aa) was cloned from DT40 first-strand cDNA by PCR using PF1 and PR1 primers. The PCR product was digested with NotI and KpnI and ligated into pEGFP-N1 digested with the same enzymes to create pN1-GgOrc6. To generate pEGFP-Orc6 and pEGFP-Orc6Δ25, the vector pEGFP-C1 was digested with EcoRI and KpnI and ligated to the PCR products generated with the primers EGF and EGR1 or EGR2, respectively, cut with the same enzymes.

We created pDegron-GgOrc6 by PCR amplifying the degron cassette from pKL187 using the primers gallus forward and gallus reverse before digestion with NheI and KpnI and ligation into pN1-GgOrc6. A targeting vector, pLoxNeo/KO-Orc6, for knocking out the entire Orc6l gene, including sequences 2 kb 5′ from the ATG (∼6.5 kb in all), was constructed as follows. A left arm from position −4 kb to −2 kb from the ATG (homology A) was amplified by PCR using the primers PF2 and PR2, digested with KpnI and SalI, and ligated into pLoxNeo digested with the same enzymes. This construct was then digested with SpeI and NotI and ligated to a right arm from the 3′ end of the gene to 2 kb downstream (homology B) prepared by PCR using primers PF3 and PR3.

A targeting vector, pLoxBsr/KI-deg-Orc6, for knocking in the degron sequence into the endogenous DT40 Orc6 gene was constructed as follows. A right arm from intron 1 to intron 4 (homology D) was amplified by PCR using the primers PF4 and PR4, digested with SalI, and ligated into pLoxBsr digested with the same enzymes (pLoxBsr-Right-arm). A left arm pLoxBsr/KI-deg-Orc6 was assembled using ZeroBlunt as a backbone in two steps. In the first step, a region from the 3′ end of exon 1 to –2 kb from the ATG (homology C) was amplified using the primers PF5 and PR5 and cloned into Zeroblunt (Invitrogen) to create Zeroblunt-Left-arm. In the second step, the degron in frame with the first exon of Orc6 was amplified using the primer pair PF6 and PR6 and cloned into Zeroblunt-Left-arm using NotI–NdeI. The cassette was digested with NdeI and SpeI and ligated into the pLoxBsr-Right-arm construct as described in this paragraph. All constructs were verified by nucleotide sequencing. Site-directed mutagenesis was used to restore a Kozak sequence 5′ of the ATG.

The following primers were used in this paper (sequences run 5′ to 3′): PF1, ­ATTGACGGTAC­CTATGGAGCGC­GGCGCGGTCCG­GGAG; PR1, ­TTAATCAGCGG­CCGCTTAACCA­CTGTTTGTCTC­TTGTGTCTTAG; PF2, ­TATGCATGGTA­CCCAAGGCATG­GGCCCTCCAAA­CAGCACTGAAG­C; PR2, ­TATTCGTCGAC­AAACGAGGGCG­TGTGAGCAAAC­AGGAGG; PF3, ­ATTCGAGACTA­GTGTCCTTGTG­CCTCTAGCTGG­CAGTAGGCAGT­G; PR3, ­ATTCATGCGGC­CGCTGCCTTTG­GCTCAGCTGGG­AGGTCGAAAC; PF4, ­GTCGACGGTGC­GGGACTCGGGA­GCAAGGAAGGG­GTG; PR4, ­GTCGACGTCCA­GCTCCAATCTA­CACTGACAAGA­GAAG; PF5, ­GCGGCCGCACC­ACAGCAGAATT­GAAGGTAAGCC­TGAGGAAC; PR5, ­CATATGCCCCG­CTCTCGGTTCC­CGCTCGCTCTG­C; PF6, ­CATATGCAGAT­TTTCGTCAAGA­CTTTGACCGGT­AAAACCA; PR6, ­ACTAGTCGGTA­CCTGACGACGC­GCGGCTCGCCG­AGCCCCA; gallus forward, GGCGCTAGCATGCAGATTTTCGTCAAGA; gallus reverse, ­GGCCGGTACCG­AGTCTTTCTTC­TCGTAGACTTC­AAAC; EGF, ­ATTCGTGAATT­CTATGGAGCGC­GGCGCGGTCCG­GG; EGR1, ­TCTAGCGGTAC­CTTAACCACTG­TTTGTCTCTTG­TGTCTTAGC; and EGR2, ­TATGGTACCTT­ATTCAGTCTTC­GGCTGTTTACA­TGG.

All targeting constructs were transfected into DT40 as described previously (Su et al., 2008). In brief, 10⁶ cells were washed with PBS and resuspended in a final volume of 400 µl PBS. 20 µg of targeting construct was linearized with ApaLI and dissolved in 400 µl PBS. Cells and DNA were mixed in an electroporation cuvette (Bio-Rad Laboratories) and kept on ice for 10 min. Electroporation was performed with an electroporation system (Gene Pulser II; Bio-Rad Laboratories) at 950 µF and 250 V followed by a 10-min incubation on ice. Cells were seeded into five 96-well plates, and the relevant antibiotics were added 24 h later at final concentrations of 1.5 mg/ml G418 (Invitrogen) and 30 µg/ml blasticidin (InvivoGen). Orc6^+/− and Orc6^+/deg cells were generated by sequential transfection of pLoxNeo/KO-Orc6 and pLoxBsr/KI-deg-Orc6 targeting constructs into wild-type DT40 cells. Genomic DNA was extracted from the candidate clones by DNAzol (Invitrogen) according to the manufacturer’s instructions, and the identification of successful targeting events was performed by PCR product analysis generated by primers 5′ knockout forward (KOF) and 5′ knockout reverse (KOR; 5′ in the knockout) and 3′KOF and 3′KOR (3′ in the knockout) for the heterozygous allele. Degron-positive clones were reselected 2 wk later by their ability to grow at 35°C and lack of viability at 42°C under neomycin and blasticidin selection. These knockin clones were identified using the primers 5′ knockin forward (KIF) and 5′ knockin reverse (KIR; 5′ in the knockin) and 3′KIF and 3′KIR (3′ in the knockin).

The resistance cassettes were then floxed out by transient expression of Cre recombinase. In brief, cells were transfected with the pPGK-Cre plasmid (gift from K.J. Patel, Laboratory of Molecular Biology, Cambridge, England, UK) using solution from Nucleofector Kit T (Lonza) according to the manufacturer’s guidelines. Clones with restored sensitivity to Bsr and Neo were selected. Direct sequencing of the RT-PCR product of the observed clone was used to confirm fusion degron-Orc6 expression. Cells stably expressing EGFP-Orc6 or EGFP were transfected as described in this paper and selected with 1.5 mg/ml G418.

The following primers were used in this paper (sequences run 5′ to 3′): 5′KOF, ­TGCTTGGAGTT­CCCATAAGCAT­CTTTTAAGG; 5′KOR, GCCTTCTATCGCCTTCTTGACGAGTTCT; 3′KOF, GAGAGCAAGTTAGCTCTGCTAATCA; 3′KOR, AACGCACGGGTGTTGGGTCGTTTGTTC; 5′KIF, TAGCAATGCTCTGCAGGTGGCTAA; 5′KIR, TGCCTACCACCTCTTAGCCTTAGC; 3′KIF, TATACGAACGGTAGGGGATCCAAG; and 3′KIR, AGGTGCTCATAGCAGCAGCTATTC.

DT40 cells were maintained in RPMI 1640 medium (Invitrogen) supplemented with 10⁻⁵ M β-mercaptoethanol, 10% fetal calf serum, and 5% chicken serum (Invitrogen) between 35 and 42°C. DT40 cells were synchronized at early mitosis by the addition of 500 ng/ml⁻¹ Nz (Sigma-Aldrich) for 5 h or at the G1/S boundary by exposure to 400 µM mimosine (Sigma-Aldrich) for 5 h (Su et al., 2008). Cells were released by washing three times with warm media, and samples were taken after release at the time points indicated in the figures. Caffeine (Sigma-Aldrich) was used at 2 mM for 2 h to suppress the G2 checkpoint. The doubling time of DT40 cells was determined by counting the viable cell number enumerated after trypan blue staining with a cell viability analyzer (Vi-CELL; Beckman Coulter) according to the manufacturer’s guidelines.

DT40 cells (2 × 10⁶) were resuspended in 0.5 ml PBS before addition of 4.5 ml of ice-cold 70% (vol/vol) EtOH and incubation for ≥2 h. Before analysis, cells were washed once with PBS and stained with anti-MPM2 antibody (1:1,200; Upstate Cell Signaling) for 1 h at 37°C before washing and staining with Alexa Fluor 488–conjugated secondary antibody (1:500) for 1 h at room temperature. Cells were incubated in 1 ml of 20 mg/ml⁻¹ propidium iodide (Sigma-Aldrich), 50 mg/ml⁻¹ RNase A, and 0.1% (vol/vol) Triton X-100 in PBS for 15 min at 37°C. Cell cycle profiles were acquired on a flow cytometer (FACSCalibur; BD) using FACSDiva software 6.0 (BD).

Immunofluorescence analyses were performed as described previously (Su et al., 2008). In brief, ∼10⁵ cells were spun out onto glass slides in a Cytospin-2 cytocentrifuge (Shandon) for 5 min at 800 rpm. Slides were dried at 37°C for 5 min and fixed in 4% paraformaldehyde for 15 min at room temperature or directly in 100% MeOH at −20°C. Subsequently, slides were then rinsed in PBS and kept at −20°C in 70% EtOH until further analysis. Cells were blocked with 3% BSA in PBS, 0.01% Tween 20, and 0.01% Triton X-100, stained with primary antibodies (mouse anti–β-tubulin [Sigma-Aldrich], rabbit anti–Aurora B [ab2254], rabbit antipericentrin [ab4448; Abcam], or rabbit anti–γ-tubulin [Sigma-Aldrich]) followed by Alexa Fluor 488 or Alexa Fluor 568 secondary antibodies (Invitrogen), and mounted in Vectashield medium with DAPI (Vector Laboratories). Imaging was performed on a confocal microscope (LSM 510 Meta; Carl Zeiss, Inc.) using a 63× Apochromat 1.4 NA objective. For Western blotting, total cell extracts were prepared in ice-cold radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% [vol/vol] deoxycholate, 0.1% [vol/vol] SDS, and 1% [vol/vol] Nonidet P-40) supplemented with 1 mM DTT, 1 mM PMSF, and protease inhibitors (GE Healthcare). For fractionation, cells were lysed with buffer A (10 mM Hepes, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 0.034 M sucrose, 10% glycerol, 1 mM DTT, 100 µM PMSF, protease inhibitor cocktail [Complete; Roche], 5 mM Na₃VO₄, 10 mM NaF, and 1 µM okadaic acid) with 0.1% Triton X-100 for 5 min on ice. Cells were centrifuged at 1,300 g for 4 min, and the supernatant was collected as a soluble cytoplasmic protein. This fraction was further clarified by high speed centrifugation. The nuclei were washed in buffer A and then lysed by buffer B (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT, PMSF, Complete protease inhibitor cocktail, 5 mM Na₃VO₄, 10 mM NaF, and 1 µM okadaic acid) for 1 min on ice. Nuclei were centrifuged at 1,700 g for 4 min. The soluble nuclear fraction was collected and mixed with the cytoplasmic soluble protein to constitute the soluble fraction. The insoluble chromatin pellet was then resuspended in buffer A with 1 mM CaCl₂ and 0.2 U micrococcal nuclease incubated at 37°C for 10 min. The chromatin fraction was collected by high speed centrifugation. Protein concentration was quantified using the bicinchoninic acid assay (Sigma-Aldrich). Cell extracts were resolved by 4–12% SDS-PAGE (Invitrogen) and transferred to nylon membranes (Millipore) before blotting with the appropriate antibodies (anti-EGFP [JL-18; BD], anti-Mek2 [BD], anti–histone H3 [Cell Signaling Technology], anti-FLAG [M2; Sigma-Aldrich], anti-CHK1 [G-4], septin 7 [H-120; Santa Cruz Biotechnology, Inc.], anti-Mcm2 [ab4461; Abcam], anti–phospho-CHK1 [S345; Cell Signaling Technology], or anti–β-actin [Sigma-Aldrich]). HRP-conjugated secondary antibodies against rabbit and mouse were used at 1:20,000 or 1:40,000 dilutions, respectively. Signal intensities were quantified by integrated density measurements on ImageJ (National Institutes of Health).

Cells were plated in glass-bottomed Mat-Tek dishes. Before imaging, growth medium was replaced with phenol red–free L15 imaging medium. Confocal live imaging was performed on a customized microscope (LSM 510 Axiovert; Carl Zeiss, Inc.) using a 100× 1.4 NA oil Plan Apochromat objective (Carl Zeiss, Inc.) or a microscope (IX81; Olympus) using a 40× 1.4 NA objective. Both microscopes were equipped with piezo focus drives and incubation chambers, providing a humidified atmosphere at 35 or 42°C. Long-term videos for Fig. 3 were acquired on an IX81 microscope (CellR software). Sample illumination was kept to a minimum and had no adverse effect on cell division and proliferation. Image analysis was preformed by LSM 510 and ImageJ software. No contrast adjustments were applied.

For experiments shown in Fig. 4, EGFP was bleached by radiating a defined region with a 488-nm laser at 100% transmission. Bleaching of EGFP on a microscope (LSM 510) was preformed by epifluorescence illumination. FRAP experiments used 50 iterations of photobleaching at 100% transmission of a 488-nm Ar ion laser at a 15-mW output at regions similar to the one indicated in Fig. 4 A. The recovery kinetics of MFI was measured in a region of constant size. Fluorescence loss in photobleaching experiments used 20 iterations of photobleaching at 100% transmission of a 488-nm laser at regions similar to the one indicated in Fig. 4 D. Bleaching was repeated every five cycles of 120 s. Mean fluorescence was measured in regions of constant size as indicated in Fig. 4 D.

Cells cytospun onto glass slides were stained with rabbit polyclonal anti–γ-tubulin and detected using an Alexa Fluor 488–conjugated anti–rabbit antibody. Slides were mounted with Vectashield medium. 20 representative fields were acquired using a 40× 1.4 NA objective on a confocal microscope (LSM 510 Meta) for each sample using constant zoom and imaging parameters (laser intensities and detector settings). Typically, 50–200 cells were analyzed in each sample to determine parameters, including mean number of foci per cell and number of cells with more than a defined foci count. Cells containing more than two γ-tubulin foci were counted as positive for supernumerary centrosomes. The data were plotted in Prism v5.0 (GraphPad Software).

All statistical analyses were performed using Prism v5.0 for Windows (Microsoft).

Immunoprecipitations were performed from 1 mg lysate using 4 µg of control mouse IgG (Sigma-Aldrich) or mouse anti-EGFP (Abcam). Immune complexes were captured on protein A–Sepharose (GE Healthcare). Whole-cell extracts and immunoprecipitates were resolved by 4–12% gradient SDS-PAGE, transferred to nylon membranes (Immobilon-P; Millipore), and blotted with polyclonal antisera against EGFP (Takara Bio Inc.) or septin 7.

Fig. S1 shows the generation and characterization of a cell line expressing functional degron-Orc6. Fig. S2 shows that degron-RAD51 depletion during the S phase triggers supernumerary centrosome formation. Fig. S3. shows that abscission defects after Orc6 depletion correlate neither with abnormal DNA bridges during division nor an interaction with septin 7.

We thank the members of the Venkitaraman laboratory for helpful discussion and technical advice and Dr. K.J. Patel for the gift of reagents.

This work was supported by grants to A.R. Venkitaraman from the Wellcome Trust and the Medical Research Council. J.A. Bernal was also supported by the Spanish Ramón y Cajal program (RYC-2009-04341) from the Ministry of Science and Innovation.