|Home | About | Journals | Submit | Contact Us | Français|
Tight regulation of microtubule (MT) dynamics is essential for proper chromosome movement during mitosis. Here we show, using mammalian cells, that structure-specific recognition protein 1 (SSRP1) is a novel regulator of MT dynamics. SSRP1 colocalizes with the spindle and midbody MTs, and associates with MTs both in vitro and in vivo. Purified SSRP1 facilitates tubulin polymerization and MT bundling in vitro. Knockdown of SSRP1 inhibits the growth of MTs and leads to disorganized spindle structures, reduction of K-fibers and midbody fibers, disrupted chromosome movement, and attenuated cytokinesis in vivo. These results demonstrate that SSRP1 is crucial for MT growth and spindle assembly during mitosis.
Mitosis is the final and critical stage of cell division essential for cell proliferation, embryogenesis, and tumorigenesis. During mitosis, duplicated chromosomes are condensed, aligned, segregated, and equally packed into two daughter cells through cytokinesis (36). These chromosome movements are driven by bipolar spindles working in concert with stabilizing and destabilizing proteins (22). At the core of this mitotic machinery are microtubules (MTs) (22), which consist of polymerized α-tubulin/β-tubulin heterodimers (45). At prometaphase, MT nucleation mediated by a number of spindle assembly factors, such as the γ-tubulin ring complexes (47), is organized into bipolar arrays of MT bundles of the mitotic spindle that functions to capture, align, and segregate chromosomes in dividing cells (1). Thus, MT polymerization and bundling are necessary for the assembly of the mitotic machinery (22, 36, 45), although this machinery is dynamic with constant polymerization (rescue) and depolymerization (catastrophe) at the MT plus ends (34). These activities are finely tuned through the balancing action between plus- and minus-end regulators (22).
To understand how bipolar MTs are exactly regulated during mitosis, a number of proteins crucial for spindle assembly and midbody formation have been identified (13, 21, 30, 37, 38, 40). One group of proteins important for spindle assembly is the family of plus-end motor kinesin-5 proteins (2, 33, 36, 43). In vitro studies suggest that kinesin-5 proteins may utilize a “sliding” mechanism to recruit MTs into bundles (18). Also, two nuclear proteins, NuMA and TPX2, were shown to stabilize MT and to organize polar MTs during mitosis (12, 25, 49). In addition, the nuclear lamin B and RanGTP proteins were reported to facilitate spindle assembly (5, 16, 27, 38, 48). Later on, another nuclear protein, importin-β, was shown as a downstream target of RanGTP in the regulation of spindle assembly (11, 15, 26, 46). Although much insight has been gained into how these proteins may affect spindle assembly, illustration of the whole picture about and the detailed mechanisms underlying the sophisticated mitosis will not be completed until all of the mitotic regulators are uncovered.
Here we report the identification of nuclear protein SSRP1 as a novel MT-binding protein that facilitates MT growth and bundling and is essential for mitosis. SSRP1 is a member of the abundant, nonhistone, high-mobility group (HMG) family of proteins that are associated with chromatin in interphase cells (10). SSRP1 initially was identified as a protein that bound to DNA modified by the anticancer drug cisplatin (3) and later found in a heterodimic complex with SPT16, which regulates transcription elongation (28, 32) and possibly DNA replication (50). Also, this heterodimer binds to the protein kinase CK2, forming a specific kinase complex for the tumor suppressor protein p53 (19, 20). In addition, SSRP1 acts as a transcriptional coactivator (53), can physically modify chromatin (29), and is cleaved during apoptosis (23). However, the precise biological role of SSRP1 remains unclear, as murine ssrp1 knockouts are lethal at E3.5 (4). SSRP1 is expressed at high levels in proliferating tissues in the mouse (14) and human cancerous tissues (52), but at low levels in less-renewable and differentiated tissues (14) or cells (our unpublished data). These observations suggest that SSRP1 may be important for cell cycle progression. The findings of the present study support this hypothesis.
Lysis buffer, radioimmunoprecipitation assay buffer, and Buffer C 100 (BC-100) were as previously described (20). Buffer C (nuclear extract [NE] buffer) was composed of 20 mM Tris (pH 7.9) at 4°C, 25% glycerol, 1.5 mM MgCl2, 0.42 M NaCl, 0.2 mM EDTA, 0.5 mM dithiothreitol (DTT), and 0.5 mM phenylmethylsulfonyl fluoride (PMSF). Buffer A was composed of 10 mM Tris (pH 7.9) at 4°C, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, and 0.2 mM PMSF. Buffer B (10×) was composed of 0.3 M Tris (pH 7.9) at 4°C, 0.03 M MgCl2, and 1.4 M KCl. BC-100 contained 20 mM Tris (pH 7.9) at 4°C, 20% glycerol, 100 mM KCl, 0.2 mM EDTA, 10 mM mercaptoethanol, and 0.2 mM PMSF. PEMG buffer was 78 mM PIPES (pH 6.9), 2 mM EDTA, 1 mM MgSO4, 6% glycerol, and 0.4 mM GTP. PEM buffer consisted of 100 mM PIPES (pH 6.9), 2 mM EDTA, 1 mM MgSO4, and 0.2 mM GTP. Tubulin assembly buffer was composed of 80 mM PIPES (pH 6.9), 0.5 mM EGTA, 2 mM MgSO4, and 5% (vol/vol) glycerol. All of these buffers contained 1 mM DTT and the protease inhibitors 0.2 mM PMSF, 4 μM pepstatin A, 1 μg of leupeptin/ml, and 1 μg of aprotinin/ml. PTEMF buffer contains 20 mM PIPES (pH 6.8), 10 mM EGTA, 1 mM MgCl2, 0.2% Triton X-100, and 4% formaldehyde.
The pH1 and pHTO2 small interfering RNA (siRNA) cloning vectors were as previously described (25a). siRNA derived from the SSRP1 gene sequence 5′-GAATGGCCATGTCTACAAGTT-3′ (nucleotides [nt] 201 to 220) was cloned into the pH 1 vector, and siRNA derived from the SSRP1 gene sequence 5′-GCTCAGGACTGCTCTACCC-3′ (nt 1043 to 1062) was cloned into the pHTO2 vector as described previously (24a). SSRP1 siRNA and scrambled siRNA (5′-AAGCGCGCTTTGTAGGATTC-3′) oligomers were synthesized (Dharmacon). pcDNA3-Flag-SSRP1 plasmid was previously described (53). An REF-H2B fusion protein expression vector was created by inserting an H2B expression insert into the pDSRED2-C1 vector with BamHI and BglII, and this vector was named pDSRED2-H2B. The psiRNA-h7SKGFPzeo vector (Invivogen, San Diego, CA) was used to generate a vector coexpressing either scramble (psiRNA-h7SKGFPzeo-scramble-siRNA) as shown above or ssrp1 siRNA (psiRNA-h7SKGFPzeo-ssrp1-siRNA, 5′-GGCCATGTCTACAAGTATGAT-3′), together with green fluorescent protein (GFP). Also polyclonal and monoclonal anti-SSRP1 (5B10) antibodies were as previously described (19, 23). Monoclonal anti-Flag and anti-α-tubulin were purchased from Sigma. Polyclonal anti-phosphorylated serine 10 histone 3 (H3) antibodies were from Upstate. Anti-EB1 (catalog no. 610534; BD Biosciences, NJ) and anti-γ-tubulin (catalog no. 620901 [Poly6209], polyclonal peptide antibody raised against amino acid KLH; BioLegend, San Diego, CA) were commercially purchased. Polyclonal and monoclonal (8E2) anti-survivin antibodies were from Santa Cruz Biotechnology, Novus, and NeoMarker, respectively. For immunostaining procedures, fluorescent secondary goat anti-rabbit Alexa-Fluor (AF) 488, goat anti-rabbit AF 546, and goat anti-mouse AF 488 (Molecular Probes, Eugene, OR) were used.
Human embryonic kidney (HEK) 293, human lung small cell carcinoma H1299, and human cervical carcinoma HeLa cells were cultured in Dulbecco modified Eagle medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS), 10 U of penicillin/ml, and 0.1 mg of streptomycin/ml at 37°C in a 5% CO2 humidified atmosphere. HeLa GFP-H2B stable cells were a gift from Geoffrey Wahl (17) and Susan Smith (8). All transfections were performed with Lipofectamine 2000 (Invitrogen).
H1299 pcDNA6-TR or HEK 293 pcDNA6-TR cells, which express Tet repressor, were transfected with 3 μg of pHTO2-SSRP1siRNA plasmid. At 24 h posttransfection, cells were treated with trypsin and transferred to 10-cm plates at low density. A total of 90 μg of hygromycin/ml was added to cells for a 2- to 3-week selection until colonies became visible. SSRP1 siRNA was induced by the addition of doxycycline (5 μg/ml), and cells were harvested for IF staining and Western blotting (WB) with the anti-SSRP1 antibody. The colonies with a marked reduction in SSRP1 were maintained for future use.
HEK 293 cells were transfected on 60-mm plates with SSRP1 siRNA or scrambled siRNA oligomers (5 μg per plate). At 5 h posttransfection, FBS was added to bring the medium's serum content to 10%, and 2 ml of fresh serum-containing DMEM was added. At 48 h posttransfection, cells were harvested. A portion of the cells was immunostained with anti-phosphorylated H3 and secondary anti-rabbit AF 488 antibodies and analyzed by flow cytometry. The mitotic index was measured by scoring H3-positive cells. Another portion of the cells was used for straight WB to analyze protein levels.
HEK 293 cells were transfected with 3 μg of pcDNA3-Flag-SSRP1 plasmid, and selected with 0.5 mg of G418/ml as previously described (6). The 293 cell line that stably expressed Flag-SSRP1 was used for affinity purification as described below. Also, pDSRED2-H2B was transfected into H1299 cells to generate an H1299 REF-H2B stable cell line by using the same strategy.
Flag-SSRP1 HEK 293 cells (108 cells) were cultured in suspension and harvested for nuclear (NE) and cytoplasmic (S100) extracts as previously described (7). Briefly, cells were harvested and washed in phosphate-buffered saline (PBS), and five cell pellet volumes of buffer A. The cells were suspended in two cell pellet volumes of buffer A, allowed to stand for 10 min in ice, and homogenized by 10 strokes of a Kontes all-glass Dounce homogenizer (B type pestle). The homogenates were centrifuged for 10 min at 2,000 rpm to pellet nuclei. The supernatants were carefully decanted, mixed with 0.11 volumes of buffer B (10×), and centrifuged for 60 min at 12,000 rpm. The high-speed supernatant from this step was dialyzed 5 to 8 h against 20 volumes of BC-100. After dialysis, the samples were centrifuged at 12,000 rpm for 20 min, and the resulting precipitates were discarded. The supernatant was designated as the S100 fraction. The crude nuclei obtained from the low-speed centrifugation of the homogenates were resuspended in 3 ml of buffer C with a Kontes all-glass Dounce homogenizer (10 strokes with a type B pestle). The resulting suspension was adjusted to the conductivity equal to 300 mM KCl by adding 3 M KCl solution, stirred gently with a magnetic stirring bar for 30 min, and then centrifuged for 30 min at 13,000 rpm. The resulting clear supernatant was dialyzed against 50 volumes of BC-100 for 5 h. The dialysates were centrifuged at 12,000 rpm for 20 min, and the resulting precipitate was discarded. The supernatants were designated the NE fraction.
These fractions were loaded separately onto phosphocellulose chromatography columns and step eluted at 0.1 M KCl and 0.3 M KCl. The fractions were dialyzed against BC-100 buffer and incubated in coimmunoprecipitation reactions using anti-Flag M2-agarose affinity gel (Sigma). After a washing step, the immunoprecipitates were eluted from the beads under acidic conditions, run on SDS-PAGE, and stained with colloidal blue. Polypeptides were digested in gel and analyzed by using matrix-assisted laser desorption ionization time-of-flight mass spectrometry and electrospray ion trap mass spectrometry in combination with capillary high-pressure liquid chromatography. Analysis of p50 polypeptides from three fractions indicated that they were α-tubulin, with matching sequences from one fraction being DVNAAIATIK and from the other two fractions LISQIVSSITAS (see Fig. Fig.2A2A).
Equal amount (2 μg) of glutathione S-transferase (GST) fusion proteins immobilized on glutathione agarose beads were incubated with 0.5 mg of S100 extract from HEK 293 cells at 30°C for 1 h. Beads were washed three times with lysis buffer. Proteins bound to GST fusion protein were resolved onto an SDS-PAGE gel and analyzed by immunoblot, as previously described (19).
Tubulin was partially purified from HeLa cytoplasmic S100 extracts using alternate 37°C heating to polymerize the microtubules and 4°C chilling to depolymerize them as described previously (42, 44). A HeLa S100 fraction was centrifuged at 100,000 × g for 1 h at 4°C. The supernatant was removed, diluted 1:1 with PEMG buffer, and incubated at 37°C for 45 min. The centrifuge rotor was prewarmed to 37°C, and the tubulin solution was centrifuged for 45 min at 100,000 × g at 29°C. The supernatant from this spin was discarded, and the pellet was resuspended in PEM buffer to 20% of the original volume of the PEMG-diluted supernatant. The resuspension was incubated on ice 30 min and centrifuged at 100,000 × g for 45 min at 4°C. Partially purified tubulin (700 μg, material from the 4°C pellet above) containing endogenous SSRP1 was loaded onto an 11-ml, 12.5 to 25% glycerol gradient and centrifuged at 35,000 rpm in a Beckman SW41 rotor for 20 h. A total of 40 fractions were collected from the bottom.
For mitotic analysis, HeLa, H1299, or 293 cells were plated on glass coverslip four-well chamber slides, synchronized by double-thymidine block, and fixed with 2% paraformaldehyde when 70 to 80% of the cells were in G2/M phase as previously determined by flow cytometry. The cells were permeabilized in 0.1% Triton X-100 (in PBS) and blocked with 2% goat serum. The antibodies used for staining are described in the figure legend, and cells were visualized by deconvolution microscopy using a DeltaVision Applied Precision Nikon TE200 inverted fluorescence microscope at ×60 magnification. For all other staining, cells were visualized with a Zeiss Axiovert 200M microscope at magnification of 40× or 60×. Representative images for HeLa cells were shown in Fig. Fig.1C.1C. The same procedure was used for examining SSRP1 and tubulin after doxycycline-induced SSRP1 siRNA in 293 pHTO2-SSRP1siRNA Tet-inducible cells.
Live mitotic cells were monitored and recorded under a Zeiss Axiovert 200M microscope at magnification of 40× at different hours after introduction of anti-SSRP1 or scramble siRNA into GFP-H2B-expressing HeLa cells. Also, the same mitotic analysis of live cells was carried out at different hours after transient introduction of psiRNA-h7SKGFPzeo-scramble-siRNA for the expression of GFP and scramble siRNA or psiRNA-h7SKGFPzeo-ssrp1-siRNA for the expression of GFP and ssrp1 siRNA into H1299 REF-H2B cells. The same batch of the cells in each case was used for WB analysis of SSRP1 and tubulin.
His-SSRP1 was purified by Ni-NTA beads and Mono-Q column. Pure tubulin dimmers were purchased from Cytoskeleton, Inc. The purity of these proteins was shown in Fig. Fig.3A.3A. Polymerized microtubules were prepared by incubating tubulin with 1 mM GTP and 10 μM Taxol at 37°C for 40 min. Various concentrations of His-SSRP1 were then added to Taxol-stabilized microtubules (final concentration of tubulin dimers was 1 μM) at 25°C for 30 min. The mixtures were subjected to centrifugation over a 40% sucrose cushion at 100,000 × g for 40 min at 25°C in a Beckman (Fullerton, CA) Optima TLA 100.3 rotor. Supernatants and pellets were analyzed by SDS-10% PAGE and immunoblotted as indicated (see the legend of Fig. Fig.2F2F).
The in vitro kinetics of tubulin polymerization was measured by using the pure tubulin dimer purchased from Cytoskeleton, Inc., according to the manufacturer's instructions. In brief, 20 or 5 μM (final concentration) of pure tubulin dimers in tubulin assembly buffer and 1 mM GTP (final concentration) were mixed on ice with recombinant proteins (as indicated in Fig. Fig.3),3), which were dialyzed against tubulin assembly buffer, in 50 μl (final volume) of reaction. The mixtures were transferred to a 96-well plate. Polymerization was started by incubation at 37°C, followed by an optical density reading at the wavelength of 340 nm every minute for up to 30 min in a temperature-controlled 96-well microtiter plate spectrophotometer. Duplicate reaction mixtures also were processed for IF staining with antibodies against α-tubulin and SSRP1, as well as electron microscopic (EM) analysis, as described below.
A total of 20 μl of tubulin polymerization reaction mixture containing 100 μM (final concentration) of tubulin in tubulin assembly buffer and 1 mM GTP was incubated at 37°C for 30 min. Then, 180 μl of prewarmed tubulin assembly buffer containing 20 μM Taxol (final concentration) was added to the reaction, followed by incubation for an additional 10 min. This step diluted the final concentration of the tubulin to 10 μM in the reaction. Then, 10 μl of the Taxol-stabilized MTs (the final concentration of tubulin is 5 μM) was mixed with 10 μl of tubulin assembly buffer containing 0, 2, or 10 μM His-SSRP1 (0, 1, or 5 μM in the final concentration) at 37°C for 15 more min. The mixture was dropped onto poly-l-lysine-treated glass slides for 5 min and fixed in 4% paraformaldehyde for 30 min. The slides were preblocked with 8% bovine serum albumin (BSA) for 30 min and incubated with anti-SSRP1 and anti-α-tubulin antibodies overnight and secondary antibodies for 40 min. The slides were washed with PBS three times after incubation with each antibody. Polymerized MTs were analyzed under a fluorescence microscope.
Tubulin polymerization mixtures were incubated at 37°C for 30 min and immediately used for EM analysis. Briefly, 300 mesh copper grids were coated with a thin carbon film and irradiated for 15 min with UV light prior to sample application. Grids were floated on top of 9-μl sample drops for 3 min, after which excess sample was removed gently by wicking with Whatman 1M filter paper. The grids were stained by incubation on 2% uranyl acetate for 1 min and then wicked and dried. Samples were imaged on a Philips CM120 transmission electron microscope equipped with a Gatan Multiscan 625 charge-coupled device (CCD) camera.
The MT regrowth assay was conducted as previously described (31). Briefly, Tet-inducible SSRP1 or scramble siRNA H1299 cells were cultured on plates in the presence of 5 μg of doxycycline/ml for 3 days and then treated with 8 μM nocodazole for additional 4 h and 30 more min in ice before washing. Cultured cells on plates were washed with 50 ml of ice-cold PBS twice, incubated 37°C for 2 to 6 min to allow MT growth, fixed in −20°C methanol, washed three times in PBS, and stained with antibodies against EB1 and γ-tubulin, or EB1 and SSRP1, or α-tubulin and DAPI (4′,6′-diamidino-2-phenylindole), followed by AF 546 and AF 488 (highly cross-absorbed) secondary antibodies (Molecular Probes), respectively. IF images were examined and taken by acquiring 16 optical sections (200 ± 50 nm) using a deconvolution microscopy (a Zeiss Axiovert 200M microscope at ×60 magnification).
To determine a possible role for SSRP1 in the cell cycle, we analyzed the effect of SSRP1 knockdown on this cellular process. A statistically significant increase of mitotic cells was observed by staining phosphorylated serine 10 of histone H3 (see Fig. S1A in the supplemental material), a mitotic marker, when SSRP1 levels were reduced by small interfering RNA (siRNA) in human lung non-small-cell-carcinoma H1299 cells that are p53 deficient and express inducible scramble or SSRP1 siRNA (see Fig. S1B in the supplemental material). This result was reproduced in p53-null MEF cells (data not shown) and suggests a p53-independent role for SSRP1 in the progression through M phase.
These results prompted us to examine the cellular localization of SSRP1 during mitosis using IF staining. The specificity of polyclonal (Fig. (Fig.1A)1A) or monoclonal (Fig. (Fig.1B)1B) SSRP1 antibodies was demonstrated by Western blotting (WB) using HeLa whole-cell lysates or NEs. These two antibodies detected a specific band that migrated at ~80 kDa (Fig. 1A and B). Of note, a light band at ~50 kDa detected by the polyclonal antibody (Fig. (Fig.1A)1A) has been shown to be a cleaved SSRP1 fragment (23). In interphase cells, SSRP1 predominantly localized to the nucleus (Fig. (Fig.1C).1C). Surprisingly, immediately after prophase, most of SSRP1 colocalized with the astral and spindle MTs (Fig. (Fig.1C).1C). Conversely, SSRP1 was not detectable in the condensed chromosomes (Fig. (Fig.1C).1C). This persisted until the nucleus was reestablished during late telophase, when SSRP1 was observed in two locations: the midbody, which is the separation point between the two daughter cells and important for cytokinesis, and the nucleus (Fig. (Fig.1C).1C). The colocalization of SSRP1 with spindle and midbody MTs was also verified in H1299 cells (see Fig. S2A in the supplemental material) and with a monoclonal antibody (5B10) specifically against SSRP1 (Fig. 1B and D and see Fig. S2B in the supplemental material; of note, Fig. Fig.1D1D is also a control for Fig. Fig.4B).4B). This localization was specific to SSRP1, since its partner SPT16 did not colocalize with spindle and midbody MTs (see Fig. S2C in the supplemental material). These results, consistent with a proteomic study listing SSRP1, but not SPT16, as one of the components identified in the midbody complex (37), suggest that SSRP1 may associate with the spindle and midbody MTs.
To investigate the potential association of SSRP1 with MTs, we attempted to identify SSRP1-interacting proteins by coimmunoprecipitation, followed by mass spectrometry (Fig. (Fig.2A).2A). Indeed, α-tubulin was found in an SSRP1-associated protein complex from HeLa cytoplasmic extracts S100, but not NE (Fig. (Fig.2A),2A), indicating that SSRP1 interacts with cytoplasmic α-tubulin (see the Materials and Methods for experimental details and the sequences of α-tubulin peptides). This pulldown of tubulin with Flag-SSRP1 by the anti-Flag antibody was specific, since this antibody did not pull down tubulin in either HEK 293 parental cells or HEK 293 Flag-MDMX stable cells (see Fig. S3A in the supplemental material). Consistent with this result, SSRP1 copurified with MTs throughout MT purification when HeLa S100 fractions were cycled in an alternating pattern of polymerization by warming to 37°C and depolymerization by chilling at 4°C (Fig. (Fig.2B).2B). Of note, two SSRP1 bands were detected. This might be caused by phosphorylation, since the upper band could be shifted to the lower position when the fraction was treated with the calf intestine phosphatase (data not shown). Analysis of purified MTs from HeLa cells after repolymerization at 37°C using glycerol gradient sedimentation centrifugation revealed that SSRP1 also associated with polymerized MTs in fractions at 2,000 kDa or greater (Fig. (Fig.2C).2C). Although IF staining in Fig. Fig.11 mostly detected nuclear SSRP1 proteins, the results in Fig. 2A to C indicate that SSRP1 associates with cytoplasmic polymerized tubulin. An explanation for this might be that the SSRP1-tubulin complex in the fraction comes from two sources: (i) the spindle MTs after the breakdown of nuclear membranes of mitotic cells (as shown in Fig. Fig.1)1) and (ii) some tubulin-associated cytoplasmic SSRP1 proteins that were not detected by the anti-SSRP1 antibodies in IF staining of routinely fixed interphase cells but could be detected after nocodazole treatment and ice-cold buffer washing of nonsynchronized cells, as shown in Fig. Fig.6B6B (see Materials and Methods for details).
To map the domain of SSRP1 required for the interaction with MTs, cytoplasmic S100 extracts from HEK 293 cells were incubated in vitro with wild type (WT) or deletion mutant SSRP1s fused with GST (19) (see Fig. S3B in the supplemental material; see also Materials and Methods for detailed assays). GST-WT-SSRP1 (amino acids [aa] 1 to 709), GST-N-SSRP1 (aa 1 to 242), and GST-Mid-SSRP1 (aa 235 to 475), but not the GST-0 control and GST-C-SSRP1 (aa 471 to 709), bound to α-, β-, and γ-tubulins (Fig. (Fig.2D).2D). This mapping result was not due to the different levels of GST, since GST alone at a equivalent level to or greater level than GST fragments in GST-SSPR1, or GST-SSRP1 N- or mid-domain-containing tubes (see Fig. S3D in the supplemental material) did not pull down any tubulin, as repeated in Fig. S3C in the supplemental material. To investigate the binding affinity and stoichiometry of SSRP1 to MTs, purified SSRP1 (Fig. (Fig.3A)3A) was cosedimented with MTs. Band intensities were plotted against SSRP1 concentration to yield a hyperbolic curve that was fitted with the binding model as described previously (5a). SSRP1 as calculated bound to tubulin dimmers on MTs in a ratio of 1 to 1 with Kd of 0.2 ± 0.1 μM (Fig. (Fig.2F).2F). SSRP1-MT binding was further confirmed by IF staining of MTs that were formed in vitro in the presence of SSRP1 as shown in Fig. 3F to G and Fig. S4A and B in the supplemental material. The costaining signals of tubulin and SSRP1 in these figures was specific to these proteins, since these highly purified proteins as shown in Fig. Fig.3A3A were the only proteins used in the in vitro tubulin polymerization reactions, and the antibodies used are specific to these two proteins. Also, the SSRP1 IF signals as detected with SSRP1-specific antibodies aligning with the polymerized microtubule fibers were disrupted in the presence of nocodazole (Fig. (Fig.3F).3F). Altogether, these data demonstrate that SSRP1 directly binds to MTs through the highly conserved N-terminal two-thirds of SSRP1 (Fig. (Fig.2E2E).
To determine whether the SSRP1-tubulin binding affects MT formation in vitro, we performed in vitro tubulin polymerization assays using purified proteins (Fig. (Fig.3A).3A). After incubation of protein cocktails in the presence of GTP at 37°C for different times or 30 min (see Materials and Methods), MTs were detected by UV-spectrophotometry at 340 nm, fluorescence microscopy, and EM, respectively. Interestingly, SSRP1 stimulated MT formation in a dose-dependent fashion, since 2 μM (1:10 molar ratio of SSRP1 to tubulin in the reaction) of SSRP1 was more effective than 1 μM SSRP1 in promoting MT formation (Fig. (Fig.3B).3B). In the presence of Taxol, a chemical that stabilizes MTs (35), SSRP1 was still able to promote MT formation at a higher rate (compare the readouts at an optical density at 340 nm in Fig. Fig.3C3C to that seen in Fig. Fig.3B).3B). This stimulation required the full-length protein, since the SSRP1 C-terminal fragment and BSA were both ineffective (Fig. (Fig.3D).3D). Also, this stimulation by SSRP1 occurred even when only 5 μM tubulin was used (Fig. (Fig.3E).3E). This result of Fig. Fig.3E3E suggests that even though the concentration of tubulin is a rate-limiting factor, SSRP1 at a 1:1 molar ratio to tubulin in the reaction is still able to overcome this obstacle in facilitating tubulin polymerization. Control experiments using nocodazole, a chemical that destabilizes MTs, showed that SSRP1 stimulation of tubulin polymerization was sensitive to this drug (Fig. (Fig.3D,3D, ,3F,3F, S4A). IF (Fig. (Fig.3F3F and see Fig. S4A in the supplemental material) and EM (Fig. (Fig.3H3H and see Fig. S4C in the supplemental material) analyses further validated the SSRP1-induced MT formation.
In addition, when examining electron micrographs of MTs formed in vitro, we found that the MTs formed in the presence of SSRP1 were much longer and more likely to occur in parallel bundles than those assembled in the presence of Taxol (Fig. (Fig.3H3H and see Fig. S4C in the supplemental material). This observation suggests that SSRP1 may possess an activity that promotes MT elongation and bundling. To test this idea, we conducted MT bundling assays. MTs were examined by fluorescence microscopy and EM. Remarkably, SSRP1 not only extended the short and needlelike MTs formed in the presence of Taxol but also organized netlike MTs into rootlike architectures in a dose-dependent manner (Fig. (Fig.3G3G and see Fig. S4B in the supplemental material). At an equal molar ratio of SSRP1 to tubulin, MTs were regrouped into large bundles (bottom panels of these figures). By counting the number of MT fibers per bundle in EM images, it was much apparent that SSRP1 plays a crucial role in facilitating MT bundling, since there were ~80% of 600 bundles with two or more MT fibers, ~50% of the bundles with three or more MT fibers, and ~40% of the bundles with four or more MT fibers in the presence of SSRP1 in comparison with ~60%, ~15%, and ~4% of 202 bundles in the same categories without SSRP1, respectively (Fig. (Fig.3I).3I). Kinetic analysis of MT bundle formation in vitro clearly showed that SSRP1 effectively facilitated MT bundling in a time-dependent fashion (Fig. (Fig.3J).3J). This effect was specific to SSRP1, since anti-SSRP1, but not anti-GFP, antibodies specifically inhibited MT bundling (see Fig. S4D in the supplemental material). Interestingly, this anti-SSRP1 antibody also disrupted MT architectures (see Fig. S4D in the supplemental material), suggesting that SSRP1 is essential for stabilizing MTs in vitro. These results strongly demonstrate that SSRP1 possesses a dual activity to promote MT elongation and bundling.
To determine whether SSRP1 is important for the formation of spindle and midbody structures and the progression of mitosis, we analyzed mitotic cells using H1299 cells that harbor Tet-inducible SSRP1 siRNA. As shown in Fig. Fig.4A,4A, both SSRP1 mRNA and protein levels were reduced dramatically after siRNA induction. This reduction was specific to SSRP1, since the levels of MDM2, aurora B, survivin, L23, Eg5, and α-tubulin were not significantly changed (Fig. (Fig.4A).4A). In the absence of doxycycline, mitotic cells with SSRP1 colocalizing to mitotic spindle and the midbody showed a normal mitotic phenotype (Fig. (Fig.1D1D and left column of Fig. Fig.4B).4B). In contrast, in representative SSRP1-knocked-down cells, MTs growing from the centrosomes were much fewer, shorter, and disorganized at prophase; the architectures of spindle MTs were also disordered and uneven at abnormal prometaphase, metaphase, and anaphase (compare Fig. Fig.4B4B to Fig. Fig.1D;1D; see also the recapitulated and enlarged images for side-by-side comparison between normal and defective mitotic spindle MTs in Fig. S5A in the supplemental material). Although residual SSRP1 proteins were detected by IF after doxycycline induction, they were clearly colocalized with tubulin in the disorganized spindle architectures and most concentrated at the centrosomes (Fig. (Fig.4B4B and see Fig. S5B in the supplemental material). Also, the level of either SSRP1 or tubulin and the size of the spindles after knockdown of SSRP1 were remarkably reduced (Fig. (Fig.4B4B and see Fig. S5A and S5B in the supplemental material). Perhaps due to the presence of residual SSRP1 molecules in the cells, ca. 50% of SSRP1-ablated H1299 cells were found to display these defects to different extents (47% among them displayed defective MTs at prometaphase/metaphase, 27% showed defective MTs at anaphase, and ca. 13% showed defective MTs at prophase and telophase), which are reflected in the differences in mitotic phenotypes of the cells, as described in the following paragraph. These results suggest that SSRP1 is more critical for mitotic MT growth and thus spindle formation at the early phases of mitosis, although the relatively low resolution of the images partially due to the reduction of spindle MT levels prevented us from analyzing the detailed defect in the spindle MTs after knockdown of SSRP1. In contrast, the structure of midbody in this SSRP1-deficient cell was similar to the normal controls (Fig. (Fig.1D1D and and4B4B and see Fig. S5A in the supplemental material), although some of midbodies were shorter and less dense than the normal controls, suggesting that SSRP1 may be less critical for the formation of midbodies or that residual SSRP1 proteins may be sufficient for maintaining the midbody structure.
Because SSRP1 plays a role in MT bundling in vitro (Fig. (Fig.3),3), we wanted to determine whether SSRP1 is also crucial for the formation of stable K-fibers and midbody fibers, since both of these mitotic MT structures are formed with bundled MT. To this end, synchronized H1299 cells were treated in ice water for 15 min as described in the figure legend of Fig. Fig.4C.4C. After this cold treatment, K-fibers and midbody fibers are relatively stable and thus can be visualized, whereas the rest of unbundled spindle MTs undergo active depolymerization (30a). Indeed, as shown in Fig. Fig.4C,4C, K-fibers and midbody fibers were clearly visible in cells without SSRP1 siRNA induction. In striking contrast, the numbers, shapes, and sizes of K-fibers in SSRP1-ablated cells were remarkably reduced, and to a lesser extent, the numbers and sizes of midbody fibers were also reduced (right panels). These phenomena can be seen in most of the SSRP1-ablated cells and demonstrate that SSRP1 is essential for MT bundling and thus for the formation of functional K-fibers and midbody fibers in cells.
Consistent with these results, chromosomes were also much less condensed at abnormal prophase and metaphase and hardly aligned at abnormal metaphase (compare Fig. Fig.4B4B to Fig. Fig.1D).1D). Although a set of chromosomes was detected in telophase with different orientations (Fig. (Fig.4B),4B), their segregation might have occurred prior to the reduction of SSRP1 levels by siRNA. These abnormal spindle structures were observed in 47% of mitotic cells with the reduction of SSRP1, whereas only 4% of mitotic cells with normal levels of SSRP1 showed abnormal mitotic structures among 200 mitotic cells counted (Fig. (Fig.4D).4D). Correspondingly, ablation of SSRP1 by siRNA prevented cells from proceeding to telophase and cytokinesis, resulting in the accumulation of abnormal prometaphase/metaphase and anaphase cells, relative to normal mitotic cells (Fig. (Fig.4E).4E). Consistent with this result, as well as the results showing defective spindles (Fig. (Fig.4B),4B), 46.8 and 26.6% of abnormal mitotic cells were at prometaphase/metaphase and anaphase, respectively, whereas only 12.7 and 13.9% were at prophase and telophase, respectively, (Fig. (Fig.4F).4F). The effect of SSRP1 siRNA on spindle structures and mitotic function was reproduced in 293 and HeLa cells (Fig. (Fig.5A5A and data not shown) and also p53 independent, since H1299 cells are p53 deficient. Hence, these results demonstrate that SSRP1 is required for the formation and maintenance of mitotic spindle architectures and essential for the progression of mitosis.
Next, we sought to determine whether SSRP1 siRNA would affect chromosome movement during mitosis, since the above assays did not address this issue. To do so, we analyzed live mitotic cells using a HeLa cell line that expressed the GFP-H2B fusion protein (17). GFP-H2B HeLa cells were transiently transfected with either scrambled siRNA or SSRP1 siRNA, after which cells were monitored by fluorescence microscopy at different time points (Fig. 5A and B). WB analysis for SSRP1 and tubulin was undertaken immediately after image collection (Fig. (Fig.5B).5B). As shown in Fig. Fig.5A,5A, HeLa cells expressing the scrambled siRNA displayed a normal mitotic process during which chromosome condensation, alignment, and segregation took approximately 30, 60, and 100 min, respectively, in a well-ordered fashion. However, chromosome movements were severely impaired in SSRP1-knocked down cells (Fig. (Fig.5A).5A). Four major defects were observed. (i) There was an ~2-h delay of chromosome alignment. (ii) Chromosomes misaligned with an excess small cluster of condensed chromosomes dissociated from a large cluster of condensed chromosomes. (iii) At 2.5 h, misaligned chromosomes appeared to change orientation but failed to segregate, and appeared to return to prometaphase at 4.5 h and then was abortively realigned at 5 h. (iv) Finally, there was an absence of chromosome segregation. To ensure the cells that were defective in chromosome movement were indeed SSRP1-ablated cells, we created another cell line that stably expressed the RFP-H2B fusion protein by using H1299 cells and then transfected this cell line with a mammalian expression vector that encodes both GFP and siRNA against scramble or ssrp1 mRNA sequences. Because GFP and the target siRNA are expressed from the same vector, siRNA should be expressed when GFP was detected in a cell (data not shown). By monitoring the GFP-positive cells under a microscope, we observed five live GFP-ssrp1-siRNA expression cells and ten live GFP-scramble siRNA expression cells. At the end of the experiment, the cells were immediately harvested for WB analysis to confirm the knockdown of SSRP1 (Fig. (Fig.5C).5C). As shown in the representative result in Fig. Fig.5D,5D, all of the 10 green cells that expressed scramble siRNA and GFP underwent normal mitosis. However, once again, defects in chromosome movement, including delayed alignment and lack of chromosome segregation, which were similar to those in SSRP1-knocked-down HeLa cells (Fig. (Fig.5A),5A), were observed in all of the five green cells that expressed ssrp1 siRNA and GFP (Fig. 5E and C). These mitosis-defective cells eventually underwent apoptosis, although they died at different time points (one died within 2 h and four died within 5 h, respectively).
Altogether, these results demonstrate that SSRP1 is essential for chromosome alignment and segregation during mitosis largely due to its critical role in MT growth and bundling. The explanation for why some of the SSRP1-ablated cells still displayed spindle structures, although mostly deformed, and underwent cytokinesis may be that either residual SSRP1 proteins could be sufficient for maintaining MT growth and that bundling or knockdown of SSRP1 by siRNA might occur after the initiation of spindle formation or during the process of spindle formation in some of the individual and fixed cells.
Centrosomal MTs nucleated from the γ-tubulin-containing centrosomes undergo active plus end growth in early phases of mitosis (39, 47). A marker of the plus end growth is EB1 (41). To determine whether the activity of SSRP1 to facilitate MT growth and bundling in vitro is correlated with the growth of MTs in cells, we performed MT regrowth assays (31) by using the Tet-inducible SSRP1 siRNA H1299 cell line. Knockdown of SSRP1 markedly eliminated the growth of centrosomal MTs, as revealed by IF staining with anti-EB1 and anti-α-tubulin antibodies (Fig. 6A to F), whereas γ-tubulin signals in the centrosomes were not affected (Fig. (Fig.6A).6A). MTs detected by EB1 (Fig. 6A and B) and α-tubulin (Fig. 6E and F) staining in SSRP1-ablated cells were much shorter than those from scramble siRNA-expressing cells (Fig. 6A and B and Fig. 6E and F). SSRP1 coresided with EB1 at the centrosomes, as well as the elongating MTs with starlike architectures in prophase cells, whereas the knockdown of SSRP1 abolished these structures (Fig. 6B, E, and F), indicating that SSRP1 plays a role in MT growth from the centrosomes at prophase. MTs regrew in a time-dependent manner, since EB1-stained (Fig. (Fig.6B)6B) and α-tubulin-stained MTs (Fig. 6E to F) grew much longer at 6 min than that at 2 min after nocodazole was withdrawn. Again, ablation of SSRP1 by siRNA (Fig. (Fig.6D)6D) markedly reduced the regrowth of MTs by ~6-fold (200 cells counted), as shown by staining α-tubulin under the same condition in two independent H1299 Tet-inducible ssrp1 siRNA cell clones (Fig. 6E to F). These results demonstrate that SSRP1 is required for the active growth of MTs, particularly at prophase when MTs are actively growing from two polar centrosomes for the formation of a mitotic spindle in each mitotic cell.
Our studies demonstrate that SSRP1, in addition to its role in regulating transcription (28, 32), and replication (50) in interphase cells, also plays a direct role in mitosis. SSRP1 not only directly associates with polymerized tubulin in vitro and in cells but also promotes MT growth in vitro and in cells. To our surprise, SSRP1 exhibits an activity that tethers MTs together and organizes them into bundlelike architectures. These novel activities of SSRP1 are important for mitosis, since knockdown of SSRP1 impairs MT growth, the formation of mitotic machinery, and chromosome movement during mitosis. Therefore, this mitotic role of SSRP1 is crucial for cell division and possibly for embryogenesis (4).
The mitotic and nuclear functions of SSRP1 are not in direct contradiction to each other, because SSRP1 partially separates from the condensed chromosomes during mitosis when the nuclear membrane no longer exists (Fig. (Fig.1).1). Other nuclear proteins, such as RanGTP (5, 16, 27, 38, 48), importin (11, 15, 26, 46), the DNA replication factor Orc6 (30), tankyrase1 [a telomeric poly(ADP-ribose) polymerase] (8), NuMA (12, 25), TPX2 (49), and nuclear lamin B (38), have also been shown to play a role in mitosis. Thus, mammalian cells effectively utilize their limited resource of proteins for different cellular functions in order to maintain normal cell growth.
For some of its functions, such as transcription elongation and DNA replication, SSRP1 works with SPT16 as a heterodimer (28, 32, 50). SSRP1 may have SPT16-independent functions as well (24, 53). In mitosis, we found that SPT16 did not colocalized with SSRP1 to the spindle and midbody (see Fig. S2C in the supplemental material) and also SSRP1 associated with tubulins independently of SPT16 (Fig. (Fig.2A).2A). Consistent with these results, SSRP1, but not SPT16, was found to reside in the mitotic midbody complex (37). Although these preliminary results suggest that SSRP1 could function independently of Spt16 in mitosis, this speculation requires further and substantial analyses.
The mitotic role of SSRP1 is linked to its interaction with MTs during mitosis, since SSRP1 colocalized with the spindle and midbody MTs (Fig. (Fig.1)1) and associated with polymerized MTs in vitro and in cells (Fig. (Fig.22 and and33 and see Fig. S3 and S4 in the supplemental material). Through a direct association with MTs, SSRP1 enhanced MT polymerization (Fig. (Fig.33 and see Fig. S4 in the supplemental material) and was required for the growth of MTs in cells (Fig. (Fig.6).6). The growth of MTs is essential for maintaining the dynamics of mitotic MTs (41, 51). Thus, SSRP1 appears to play a role in MT dynamics. Because SSRP1 associates with MTs most apparently in mitotic cells (Fig. (Fig.1),1), its role in maintaining MT dynamics is more important for mitotic MTs. Consequently, depletion of SSRP1 by siRNA resulted in disorganized spindle structures and marked reduction of stable K-fibers and midbody fibers (Fig. (Fig.44 and see Fig. S5B in the supplemental material). However, it remains to be seen if SSRP1 may also have a role in regulating cytoplasmic MT dynamics, since SSRP1 was detected in cytoplasmic extracts (Fig. (Fig.2A)2A) and required for MT growth (Fig. (Fig.66).
It is still baffling how exactly SSRP1 organizes MTs into bundles. Perhaps SSRP1 may tether MTs together by cross-linking them, as SSRP1 associates with MTs throughout MT architecture (Fig. 3F to G). At substoichiometric levels (in the concentration of protein monomers), SSRP1 appeared to associate with a portion of MTs and to promote partial MT bundling (Fig. (Fig.3G),3G), whereas at a stoichiometry of 1 to 1, the effect of SSRP1 on MT bundling was more compelling (Fig. (Fig.3G).3G). This effect was time dependent, since SSRP1 started to convert netlike MTs into bundles at 5 min and the bundles became larger and fewer 30 min after incubation (Fig. (Fig.3J).3J). Therefore, our study reveals a previously unknown and direct role of SSRP1 in mitosis. This role is transcription independent (Fig. (Fig.11 to to6).6). Further supporting this conclusion is that depletion of Xenopus SSRP1 with its antibody from mitotic Xenopus egg extracts resulted in a complete loss of mitotic spindles in the extracts, indicating that xSSRP1 is tightly associated mitotic MTs and important for spindle MT formation independently of transcription (data not shown). Although SSRP1 might also regulate mitosis in part by controlling the expression of yet-unidentified target genes critical for mitosis, our recent studies using a cDNA microarray did not reveal any apparent SSRP1 target genes that encode known mitotic protein regulators (24a). Future studies are needed to elucidate how SSRP1 regulates the dynamics of mitotic MTs by working in concert with other regulatory proteins, such as importin, RanGTP, NuMA, TPX2, lamin B, Eg5, or survivin, and how this mitotic function of SSRP1 is regulated during mitosis. Finally, it still remains to firmly determine whether the mitotic function of SSRP1 is Spt16 independent or not.
We thank Mushui Dai for help with cloning the SSRP1 siRNA construct; Mathew Thayer, Geoffrey M. Wahl, and Susan Smith for reagents; and Stephen J. Doxsey, Yinghui Mao, Fengzhi Li, Dahong Zhang, and Yixian Zheng for helpful discussions.
This study was supported by NIH/NCI grants to H.L. (CA93614, CA095441, CA079721, and CA129828).
Published ahead of print on 7 December 2009.
†Supplemental material for this article may be found at http://mcb.asm.org/.