Cohesin subunit SMC1 associates with mitotic microtubules at the spindle pole

Abstract

Accurate mitotic chromosome segregation depends on the formation of a microtubule-based bipolar spindle apparatus. We report that the cohesin subunit structural maintenance of chromosomes subunit 1 (SMC1) is recruited to microtubule-bound RNA export factor 1 (Rae1) at the mitotic spindle pole. We locate the Rae1-binding site to a 21-residue-long region, SMC1^947-967 and provide several lines of evidence that phosphorylation of Ser⁹⁵⁷ and Ser⁹⁶⁶ of SMC1 stimulates binding to Rae1. Imbalances in these assembly pathways caused formation of multipolar spindles. Our data suggest that cohesin's known bundling function for chromatids in mitotic and interphase cells extends to microtubules at the spindle pole.

During the cell cycle, Rae1 (also termed Gle2 or mrnp41) (1–3) dynamically partitions between nuclear pore complexes (NPCs) (4), mRNPs (1), and microtubules (2, 5). As a beta propeller protein, Rae1 is ideally suited for serving as a protein interaction platform. It binds to the Gle2-binding site (GLEBS) of the nucleoporin Nup98 and to the Ser/Thr protein kinase budding uninhibited by benzimidazole (Bub1) (3, 6). Rae1–Nup98 interaction prevents aneuploidy by inhibiting securin degradation (7). We previously reported that microtubule-bound Rae1 recruits nuclear mitotic apparatus protein (NuMA) (8) to the mitotic spindle. We located NuMA's binding site for Rae1 to an N-terminal segment of NuMA and showed that in vivo expression of this fragment and imbalances created in this assembly pathway, caused formation of multipolar spindles (2).

SMC1 has been reported to play a role in spindle pole formation (9). SMC1 is a large protein with a globular ATP-binding site at both ends and a long helical domain that is interrupted near its center by a nonhelical region. The two long helical regions loop back on each other and are predicted to form a 50-nm-long intramolecular coiled coil, yielding two vicinal ATP-binding sites at one end with the central nonhelical region forming a globular domain at the other end of the folded molecule (10–12). SMC1 pairs with a similarly built SMC3 molecule at their respective central globular regions called the hinge domain (see Fig. 3A). The heterodimer's ATP-binding site terminals are reversibly circularized by two proteins, Scc1/Rad21, and Scc3/SA (12). The resulting heteromultimeric complex is termed cohesin and functions in tethering mitotic and meiotic sister chromatids (13). In addition, cohesin has been reported to function in DNA repair (14, 15), in the morphogenesis of nondividing neurons (16, 17) and in the regulation of gene transcription (18–21) in postmitotic cells. After ionizing radiation, SMC1 is phosphorylated by the kinase Ataxia Telangiectasia Mutated (ATM) at Ser ⁹⁵⁷ and Ser ⁹⁶⁶ (15, 22, 23).

Fig. 3.

Mapping of the SMC1 domain interacting with Rae1. Schematic drawing of the hinge-like structure known to be formed by the two cohesin subunits, SMC1 and SMC3; purple stars indicates the location of the two Ser residues that have been reported to be phosphorylated by ATM (A Upper). Domain structure of SMC1 and the constructs 1–5 covering the entire length of SMC1; numbers refer to amino acid residues (A Lower). Each of the five untagged SMC1 fragments (see above) and Flag-tagged Rae1 were synthesized in ³⁵S methionine containing reticulocyte lysate followed by pulldown using anti-Flag coated beads. Pulldown samples were resolved in a 4–20% SDS-polyacrylamide gradient gel and visualized by autoradiography. Note that Rae1 pulled down SMC1 fragment 5, containing residues 947-1233 (A Lower; B, lane 5). Numbers on left indicate molecular mass markers in kilodaltons. Further mapping of SMC1–5; three constructs (5A-C) representing residues 947-1100, 947-1025, and 947–967 were made; note that all three contain Ser⁹⁵⁷ and Ser⁹⁶⁶, indicated in red (C). Fragments 5A-C were cosynthesized with Flag-Rae1 as in B and pulldowns were analyzed by SDS/PAGE by using an 18% acrylamide gel; arrows indicate pulled down fragments 5A, 5B, and 5C (D). In vivo expression of HA-tagged SMC1–5B and SMC1–5C and empty HA vector in HeLa cells, followed by anti-HA IP and immunoblotting with anti-Rae1 (E).

By immunofluorescence microscopy, SMC3 was colocalized with NuMA at spindle poles, although its association with these structures has not been further defined (9). Moreover, immunodepletion of SMC1 from mitotic extracts resulted in failure of spindle aster assembly (9, 24). It was therefore suggested that cohesin may be involved in spindle assembly during mitosis (9, 24, 25). However, the molecular mechanism by which cohesin functions in spindle assembly was not addressed in these studies.

Results and Discussion

SMC1 Interacts with Rae1 During Mitosis.

To investigate whether microtubule-bound Rae1 might be involved in the recruitment of SMC1 to mitotic spindle poles, we carried out immunoprecipitation of mitotic HeLa cell extracts. Coimmunoprecipitated proteins were separated by SDS/PAGE and proteins of interest identified by immunoblotting. We found that Rae1 antibodies coimmunoprecipitated both SMC1 and SMC3, along with Nup98 and tubulin (Fig. 1A); conversely, SMC1 antibodies coimmunoprecipitated Rae1 and SMC3 but not Nup98 (Fig. 1B). To test the strength of the associations with Rae1, we treated the Rae1 immunoprecipitate with 2 M guanidine·HCl and found that SMC3 was solubilized, whereas SMC1 remained associated with Rae1 (Fig. 1C). These data suggested that SMC1 interacts with Rae1.

Fig. 1.

Human cohesin subunit SMC1 interacts with Rae1 during mitosis. IP from mitotic HeLa cell extracts with anti-Rae1, anti-SMC1, or nonspecific rabbit antibodies (IgG) were analyzed by SDS/PAGE, followed by immunoblotting with Rae1, SMC1, SMC3, Nup98 or tubulin (DM1A) antibodies. In lanes marked “2% input,” 5 μl of 250 μl of extract that was used per IP was analyzed directly (A and B). Anti-Rae1 IP was washed with 2.0 M guanidine·HCl before being analyzed by SDS/PAGE and immunoblotting with anti-SMC1 or -SMC3 antibodies (C).

SMC1 and Rae1 Colocalize at Spindle Pole.

To visualize and localize these interactions, we carried out two assays, immunofluorescence of permeabilized HeLa cells (Fig. 2 A and B) and in vitro assembly of asters from mitotic extracts of HeLa cells [Fig. 2 C and D and supporting information (SI) Fig. S1]. In digitonin-permeabilized HeLa cells, SMC1 localized to the spindle (Fig. 2A), and colocalized with Rae1 at the spindle pole region (Fig. 2B). In the in vitro aster assembly assays, Rae1 localized along the aster microtubules, showing a noticeable concentration in the center of the aster (Fig. 2C). Strikingly, SMC1 located almost exclusively to the center of the aster (Fig. 2D). In mitotic extracts that were immunodepleted of Rae1, SMC1, or SMC3, microtubules failed to organize into recognizable mitotic asters (Figs. S2–S4). Extracts that were mock-depleted by rabbit IgG had no effect on aster formation (data not shown). Together with the immunoprecipitation data, these results suggested that SMC1 and SMC3 interact with microtubules and Rae1 at the spindle pole region.

Fig. 2.

Colocalization of SMC1 and Rae1 at the spindle pole. Confocal microscopy of mitotic HeLa cells that were fixed and permeabilized with cold methanol and 0.3% digitonin and costained with anti-SMC1 (red) and -tubulin (green) (A) or anti-SMC1 (green) and -Rae1 (red) (B). DNA was visualized by using DAPI (blue). Confocal microscopy of asters that were formed in vitro from mitotic HeLa cell extracts and that were visualized by anti-tubulin (green) and anti-Rae1 (red) (C) or anti-tubulin (green) and -SMC1 (red) (D). (Scale bar, 5 μm.)

Biochemical Mapping of Rae1–SMC1 Interaction Domain.

To more closely examine binding of Rae1 to SMC1 that was suggested by the immunoprecipitation data (Fig. 1C) and to biochemically map the region of SMC1 that interacts with Rae1, we constructed a series of cDNA fragments covering the entire length of SMC1 (Fig. 3A Lower) and coexpressed these constructs together with Flag-tagged Rae1 cDNA in a cell-free reticulocyte translation system. We found that Flag-tagged Rae1 pulled down only the C-terminal fragment 5, containing residues 947-1233 of SMC1 (see lanes 1–5 of Fig. 3B). This fragment, SMC1–5, contains the entire C-terminal ATP binding domain and part of the upstream alpha-helical region (Fig. 3A Lower). To further fine-map the Rae1 association site on SMC1–5, we shortened its C-terminal end to obtain fragments 5A (residues 947-1100), 5B (residues 947-1025), and 5C (residues 947–967) (Fig. 3C). All three fragments of SMC1 associated with Flag-tagged Rae1 that was cosynthesized in the reticulocyte translation system (Fig. 3D). To test for interaction with Rae1 in vivo, we expressed the shorter two fragments, HA-SMC1^947-1025, or HA-SMC1^947-967, or an empty HA-vector in HeLa cells. As shown in Fig. 3E, Rae1 was specifically pulled down by either SMC1 fragments, but not by an empty HA vector alone. We conclude that the shortest fragment containing only 21 residues, SMC1^947-967, was sufficient to associate with Rae1, both in vitro and in vivo. Close inspection of the 21-residues-long SMC1–5C fragment revealed that it contains a consensus WD-40 repeat-binding site (SXXXS) (26), strongly supporting our biochemical mapping data that this region of SMC1 binds the WD-40 beta propeller Rae1.

Phosphorylation of SMC1^947-967.

It was previously reported that the protein kinase ATM phosphorylates SMC1 at two Ser residues, S⁹⁵⁷ and S⁹⁶⁶, and that this phosphorylation is required in DNA repair (14, 22). To investigate whether phosphorylation of SMC1 is also involved in subsequent interaction of phosphorylated SMC1 with Rae1, we used two kinase inhibitors, LY294002 and caffeine, that inhibit ATM and related kinases. We observed that the interaction between Rae1 and SMC1 in the reticulocyte translation system was abolished at the higher concentration of these inhibitors (Fig. 4A and Fig. S5). We conclude that phosphorylation of SMC1, whether by ATM or another kinase that is present in the reticulocyte lysate, greatly stimulates interaction with Rae1 and that this interaction is abolished in the presence of two kinase inhibitors.

Fig. 4.

Phosphorylation of Ser⁹⁵⁷ and Ser⁹⁶⁶ of SMC1 stimulate binding to Rae1. The kinase inhibitor LY294002 was added at the indicated concentrations to the reticulocyte system that was programmed with Rae1-Flag and full length SMC1 cDNA; Rae1-Flag pulldowns were analyzed by SDS/PAGE (18% acryalmide gel) and autoradiography (A). (B) As in A, except that the incubations contained the indicated SMC1 wild type or mutant constructs. HeLa cells were cotransfected with Rae1-HA and Myc-SMC1-WT, Myc-SMC1-S957A, Myc-SMC1-S966A, or Myc-SMC1- DBA (double S957A and S966A) and treated with nocodazole (2); anti-Myc IPs of lysates were analyzed by SDS/PAGE and immunoblotted with HA antibodies (C Upper). In a control, cotransfected cell lysates were blotted with anti-HA and -Myc antibodies (C Lower); aster formed in vitro and stained with tubulin antibodies (green) and ATM antibodies (red) (D). (E) As in D but using SMC1 antibodies (green) and ATM antibodies (red).

To investigate whether it is specifically phosphorylation of S⁹⁵⁷ or S⁹⁶⁶ alone, or of both, enhances binding to Rae1, we tested SMC1-S957A, SMC1-S966A, and the double SMC1 mutant, SMC1-S957A-S966A (14). Translation together with Flag-tagged Rae1 in the reticulocyte system and subsequent pulldown with Flag-tagged Rae1 (Fig. 4B), showed that conversion of Ser into Ala at these sites greatly reduced, but did not completely abolish interaction with Rae1: interaction of the SMC1 mutants with Rae1 was still detected after longer exposure of the gel (compare lane 1 with lanes 2–4). To examine whether the SMC1 mutants could interact with Rae1 in vivo, we cotransfected HeLa cells with myc-tagged wild-type or the mutant SMC1 constructs together with HA-Rae1. Myc antibody-conjugated beads were used for pulling down associated proteins that were in turn detected by immunoblotting (Fig. 4C). SMC1 mutants were clearly less efficient in pulling out Rae1, compared with wild-type SMC1. However, as observed for the in vitro experiments (Fig. 4B), the interaction of SMC1 mutants with Rae1 were not abolished. Together, these experiments suggested that in vitro and in vivo phosphorylation of SMC1 at Ser⁹⁵⁷ and S⁹⁶⁶ stimulates interaction of Rae1 with SMC1. It will be interesting to determine whether the reported ATM-mediated specific phosphorylation of SMC1 in a DNA repair pathway also leads to the recruitment of Rae1.

NuMA and SMC1 Bind to Different Sites of Rae1.

A previously identified fragment of NuMA, NuMA^325-829, that also binds Rae1 (2), did not compete, indicating that SMC1 and NuMA bind to different sites of Rae1 (Fig. S6).

Colocalization of ATM and SMC1 at the Spindle Pole.

Although ATM has been shown to be localized at the spindle pole by immunofluorescence confocal microscopy (27), we investigated whether SMC1 colocalizes with ATM using the in vitro aster assembly assay. Strikingly, both ATM and SMC1 are colocalized at the center of the aster (Fig. 4 D and E). Moreover, after immunodepletion of ATM from the mitotic extract, we observed only a few irregular, and unbundled microtubules (Figs. S2 and S7), pointing to a critical function of ATM, although ATM immunodepletion may have codepleted other kinases. Nevertheless, colocalization of ATM and SMC1 suggest that a spindle pole-localized ATM is a plausible candidate for phosphorylating SMC1's S⁹⁵⁷ and S⁹⁶⁶, thereby promoting binding of phosphorylated SMC1 to microtubule-bound Rae1.

Abnormal Spindle Formation.

Because imbalances in the concentration of reaction partners in the mitotic spindle formation pathways interfere with the assembly of normal bipolar spindles, we carried out overexpression experiments (Fig. S8) and checked for the formation of supernumerary spindles (Table 1). Transfection of HeLa cells with SMC1 or SMC3 caused an increase of multipolar spindles from a base line of 5% to ≈25% (Table 1). In contrast, overexpression of the corresponding S to A mutants led only to a slight increase in supernumerary spindles (Table 1). Strikingly, transfection with SMC1 fragments 5B or 5C yielded an increase in supernumerary spindles comparable to that of transfection with full length SMC1 or SMC3 (Table 1 and Fig. S8 A and B). These data suggest that overexpression of the 21-residues-long region of SMC1 might compete for phosphorylation with full-length SMC1 and/or that a phosphorylated 21-mer in turn might compete with binding of full length SMC1 to microtubule bound Rae1.

Table 1.

Spindle phenotypes in HeLa cells overexpressing SMC3 or various SMC1 constructs (n = 3 independent experiments)

Vectors	Mitotic cells	Percent bipolar	Percent monopolar	Percent multipolar
Control	100	95 ± 2	0	5 ± 1
Myc-SMC1	100	67 ± 5	5 ± 3	28 ± 2
Myc-SMC1-DBA	100	90 ± 1	0	10 ± 1
Myc-SMC1-S957A	100	91 ± 3	0	9 ± 2
Myc-SMC1-S966A	100	89 ± 2	0	11 ± 3
HA-SMC3	100	71 ± 5	2 ± 2	27 ± 4
HA-SMC1–5B	100	73 ± 5	2 ± 2	25 ± 4 (22 ± 2)
HA-SMC1–5C	100	76 ± 2	1 ± 2	23 ± 3 (4 ± 1)

Numbers in brackets in the last two lines indicate percentage of cells with irregular chromosome masses.

When examining chromosome condensation figures as revealed by the shape of DAPI-stained metaphase chromosome masses, we observed an interesting difference in the in vivo expression of the two short SMC1 fragments. Expression of the longer fragment, SMC1^947-1025, in addition to causing formation of multipolar spindles, also yielded defects in the normal arrangement of chromosome masses; instead of the characteristic metaphase plate, the chromosome masses were arranged in circular or fragmented figures (Fig. S8A). In contrast, the shorter fragment SMC1^947-967 yielded chromosome masses arranged in a more normal metaphase plate (Fig. S8B). One possible interpretation of these observations is that the shorter fragment interferes only with phosphorylation and the resulting bundling of microtubules at the spindle pole, whereas the longer fragment, perhaps by invading the coiled-coil domain of SMC1, also interferes with cohesin's function in chromosome condensation pathways and in the formation of a normal metaphase plate arrangement of the chromosome masses.

In conclusion, we have detected a distinct pathway that participates in spindle pole formation and that involves recruitment of SMC1 by microtubule-bound Rae1 at the spindle pole (Figs. 1 and 2). The region of SMC1 that interacts with Rae1 has been mapped to a 21-residue-long segment of SMC1, SMC1^947-967 (Fig. 3). The spindle-pole-anchored ATM or a related kinase presumably phosphorylates two Ser residues, Ser⁹⁵⁷ and Ser⁹⁶⁷ of SMC1^947–967 (Fig. 4), which in turn stimulates binding of SMC1 to microtubule bound Rae1 (Fig. 4 and Figs. S2, S5, and S7). We propose that only those SMC1/SMC3 heterodimers that have access to a spindle pole-anchored ATM will be phosphorylated in situ. We envision that such assemblies occlude further access by additional SMC1/SMC3 heterodimers to spindle pole-anchored ATM thereby preventing the spread of SMC1/SMC3 interactions along microtubule-bound Rae1 away from the spindle pole. It is likely that the spindle pole bound SMC1/SMC3 heterodimer recruits the other members of the cohesin complex, Scc1 and Scc3, thereby completing closure into a circular structure. Hence, cohesin would function to embrace microtubules at the spindle pole, extending its known function in encircling chromatids.

A detailed biochemical characterization of the interactions between the numerous proteins that have so far been reported to be associated with normal bipolar spindles by microscopical method, will considerably advance our understanding of the mechanisms underlying the formation and the dynamic structure of bipolar spindles.

Methods

Plasmids.

The mammalian expression constructs encoding full-length human SMC1 and its variants at the serine phosphorylation sites [SMC1 S957A, SMC1 S966A, and double-mutant SMC1 S957A:S966A (SMC1-DBA) were a generous gift of Michael B. Kastan (St. Jude Children's Research Hospital, Memphis, TN]. Rae1 with Flag tag was subcloned into pET 28a vector. The SMC1 domains were subcloned by PCR from pJET vector. All constructs were confirmed by DNA sequencing.

Preparation of Mitotic Extracts for Mitotic Aster Formation.

Mitotic extracts from HeLa cells were prepared according to refs. 28 and 29 (for overview, see Fig. S1). Briefly, HeLa cells were synchronized by double block with 2 mM thymidine. After a release from thymidine block, the cells were allowed to grow for 7 h, and then nocodazole was added to a final concentration of 40 ng/ml. The mitotic cells that accumulated over the next 4 h were collected by mitotic shake-off and incubated for 30 min at 37°C with 20 μg/ml cytochalasin B. The mitotic index of the population of cells isolated in this fashion was >90% and we typically obtained 10⁷ cells from eight 15-cm² tissue culture dishes. The cells were then collected by centrifugation at 2,000 × g and washed twice with cold PBS containing 20 μg/ml cytochalasin B. Cells were washed one last time in cold KHM buffer (78 mM KC1/50 mM Hepes, pH 7.0/4 mM MgC1₂/2 mM EGTA/1 mM DTT) containing 20 μg/ml cytochalasin B and finally sonicated at a concentration of ≈3 × 10⁶ cells per ml in KHM buffer containing 20 μg/ml cytochalasin B, 20 μg/ml phenylmethylsulfonyl fluoride and 1/4 pill of Protease inhibitor (Roche). The crude cell extract was then subjected to sedimentation at 100,000 × g for 30 min at 4°C (Beckman airfuge, 25 psi). The supernatant was recovered and supplemented with 2.5 mM ATP-Mg and 10 mM taxol. The total protein concentration of the extract prepared in this fashion was routinely ≈0.8–1.0 mg/ml. The mitotic extract was incubated at 33°C for 30–60 min to form asters. After incubation, 5 μl of the extract was diluted into 25 μl of KHM buffer, spotted onto a polyL-lysine-coated glass coverslip, fixed by immersion in −20°C methanol, and then processed for confocal microscopy as described (2).

Immunodepletion of Mitotic Extracts.

Antibodies to Rae1, SMC1, SMC3, ATM, or for mock depletion, nonspecific rabbit IgG, were each coupled to protein-A Dynabeads (Dynal, Invitrogen). Using a 1:10 ratio of beads to extract, 500-μl aliquots were incubated with agitation for 1 h and the beads removed by sedimentation at 15,000 × g for 10 s. After three repeat cycles, leached antibodies were removed by one cycle of incubation with 10 μl of washed protein-A Dynabeads. The immunodepleted extracts were then incubated for aster formation as described above. Depletion efficiency was assayed by SDS/PAGE and immunoblot.

Preparation of Flag-Rae1.

Flag-Rae1 was expressed in 600 μl of rabbit reticulocyte lysate by using the Promega TNT coupled transcription/translation system according to the manufacturer's protocol; after posttranslational incubation with 20 μl of anti-Flag beads (Sigma-Aldrich), the beads were washed three times with Binding buffer (20 mM Hepes, pH 7.5/100 mM KCl/5 mM MgCl₂/0.1% Tween 20/20% glycerol/0.01% BSA/1 mM DTT/1 mM PMSF/1× complete protease inhibitor mixture) and Flag-Rae1 was eluted by 100 μg/ml Flag peptide (Sigma-Aldrich) or with 0.1 M glycine·HCl, pH 3.5, according to the manufacturer's protocol. Ten micrograms per milliliter of Flag-Rae1 was routinely collected by the above methods. Flag-Rae1 was further concentrated to 100 μg/ml by using Microcon-30 concentrators (Amicon), and where indicated, was added to Rae1 immunodepleted extracts at 1:10 in volume.

Immunoprecipitations (IPs).

HeLa cells (10⁷) were seeded and synchronized as described in ref. 2. Mitotic cells were collected, washed with PBS, spun at 400 × g for 10 min, and lysed in 1 ml of cold lysis buffer [50 mM Tris·HCl, pH 7.2/300 mM NaCl/0.1% Nonidet P-40/5 mM MgCl₂/2 mM EDTA/10% glycerol) containing 1x protease inhibitor mixture (Roche)] and 1 mM PMSF. Lysates were centrifuged for 30 min at 4°C at 14,000 × g. The supernatant was cleared with 50 μl of Protein A/G bead slurry (Santa Cruz Biotechnology), then mixed with 10 μl of various antibodies as specified, and incubated for 1 h at 4°C with rocking. The beads were then washed five times with 500 μl of lysis buffer. After the last wash, proteins were extracted by 50 μl of 1× SDS/PAGE Blue Loading buffer (New England Biolabs) and analyzed by SDS/PAGE.

Immunofluorescence Confocal Microscopy.

Synchronized HeLa cells were washed in PBS and fixed for 10 min in methanol at −20°C. Cells were then permeabilized with 0.1% Triton X-100 or 0.3% Digitonin in PBS for 10 min at room temperature. Samples were examined on a Zeiss LSM510 MEGA confocal microscope, and all images were acquired by using a plan-Apochromat 100 × 1.4-n.a. objective.

In Vitro Binding Assays.

Proteins were expressed by using the Promega TNT-coupled transcription/translation system according to the manufacturer's protocol. Five microliters of Flag beads (ANTI-FLAG M1 Agarose Affinity Gel from Sigma–Aldrich) was washed three times with binding buffer (see above), preblocked for 10 min with 10 μl of nonspecific rabbit IgG, washed with binding buffer, and resuspended in 60 μl of binding buffer. Then, 10 μl of various indicated combinations of in vitro transcribed and translated [³⁵S] methionine-labeled SMC1, SMC3, SMC1S957A, SMC1S966A, or double S/A mutants and Flag-Rae1 were incubated with beads at 4°C for 1 h. Beads were washed six times with binding buffer and boiled in 15 μl of SDS/PAGE sample buffer. Samples were analyzed by SDS/PAGE (4–20% Tris-glycine gels; Invitrogen), followed by autoradiography. All other chemicals were from Sigma-Aldrich, including kinase inhibitors (LY294002 and Caffeine).

Mammalian Cell Culture and Transfection.

Mammalian cell culture, preparation of mitotic HeLa cells, and plasmid transfection of HeLa cells, were as described (2).

Antibodies.

Anti-Rae1 and -Nup98 polyclonal antibodies were prepared as described (2). Anti-NuMA polyclonal antibodies were from D. Compton (Dartmouth Medical School, Hanover, NH); anti-NuMA monoclonal antibodies (clone 22) from BD Biosciences; anti-Myc monoclonal antibodies, DM1A monoclonal α-tubulin antibodies were from Sigma–Aldrich; anti-hSMC1, anti-hSMC3 polyclonal antibodies were from Chemicon; anti-SMC1(C2M) monoclonal antibody was from Santa Cruz Biotechnology; anti-Histone H3, anti-Phospho-Histone H3 antibodies were from Upstate Biotechnology; anti-ATM, anti-HA, anti-GFP antibodies were from Abcam. Secondary antibodies were from Molecular Probes.