NudC, a nuclear movement protein that associates with dynein, was originally cloned as a mitogen-inducible early growth response gene. NudC forms a biochemical complex with components of the dynein/dynactin complex and is suggested to play a role in translocation of nuclei in proliferating neuronal progenitors as well as in migrating neurons in culture. Here, we show that NudC plays multiple roles in mitosis and cytokinesis in cultured mammalian cells. Altering NudC levels by either small interfering RNA-mediated gene silencing or adenovirus-mediated overexpression resulted in multinucleated cells and cells with persistent intercellular connections and disorganized midzone and midbody matrix. These phenotypes suggest a failure in cytokinesis in NudC altered cells. Further, a key mitotic enzyme, polo-like kinase, is mislocalized from the centrosomes and the midbody in NudC altered cells. Gene silencing of nud-1, the Caenorhabditis elegans ortholog of NudC, led to a loss of midzone microtubules and the rapid regression of the cleavage furrow, which resulted in one-celled embryos containing two nuclei. The loss of midzone microtubule organization owing to silencing of the NudC/nud-1 gene in two systems, coupled with the loss of Plk1 from mitotic structures in mammalian cells, provide clues to the cytokinesis defect and the multinucleation phenotype. Our findings suggest that NudC functions in mitosis and cytokinesis, in part by regulating microtubule organization at the midzone and midbody.
NudC was first identified as a gene that regulates nuclear movement in the asexual reproductive cycle of the filamentous fungus Aspergillus nidulans. In A. nidulans, nudC is required for nuclear distribution, cell wall deposition, colony growth and viability (Osmani et al., 1990; Chiu et al., 1997). Other nud loci encode either components or proposed regulators of the minus-end-directed microtubule (MT)-dependent motor dynein and its activator dynactin (Morris, 2000). The mammalian homologues of NUD proteins include Lis1 (NUDF) (Reiner et al., 1995; Smith et al., 2000; Faulkner et al., 2000; Xiang et al., 1995), NudE/NudeL (NUDE) (Niethammer et al., 2000; Feng et al., 2000; Efimov and Morris, 2000), cytoplasmic dynein heavy (NUDA) (Xiang et al., 1994) and light (NUDG) chains (Beckwith et al., 1998), and the actin-related protein Arp1 (NUDK) (Xiang et al., 2000), which is a component of dynactin. In mammalian cells, we showed that NudC forms a biochemical complex with Lis1, associates with components of the dynein/dynactin complex and colocalizes with dynein/dynactin at the microtubule-organizing center (MTOC) in neurons and fibroblasts (Aumais et al., 2001). The colocalization of NudC with Lis1 in the proliferative zone of the developing neocortex (Morris et al., 1998) and in migrating neurons in culture (Aumais et al., 2001) suggest that the NudC/Lis1 complex plays a role in cell division and cell migration. The high degree of sequence conservation in NudC from fungus to human suggests an important function that has been conserved over evolution (Moreau et al., 2001). The functional interactions between NudC, Lis1, NudeL/NudE and dynein/dynactin are likely to contribute to motor functions and microtubule dynamics during such diverse cellular processes as nuclear movement (Xiang et al., 1994; Morris et al., 1998; Aumais et al., 2001), neuronal migration (Niethammer et al., 2000; Sasaki et al., 2000; Smith et al., 2000; Leventer et al., 2001; Aumais et al., 2001; Faulkner et al., 2000; Vallee et al., 2000), axonal transport (Liu et al., 2000), intracellular transport (Karki et al., 1998; Karki and Holzbaur, 1999; Swan et al., 1999) and cell division (Faulkner et al., 2000; Tai et al., 2002; Coquelle et al., 2002; Zhang et al., 2002; Liu et al., 1999; Liu et al., 2000).
We originally cloned mammalian NudC as an early mitogen-responsive gene in T cells and showed that, in synchronized cultures, NudC mRNA levels increase during mid-G1 and NudC protein levels double during G2/M (Axtell et al., 1995). NudC expression is elevated in cells with a high growth rate (Gocke et al., 2000a) and in prostate tumors derived from transgenic mouse models (Aumais et al., 2000). Others have shown that NudC is highly expressed in clinical bone marrow isolates from patients with acute lymphoblastic or acute myelogenous leukemia (Miller et al., 1999; Gocke et al., 2000b). These studies show that NudC levels correlate with the proliferative status of various cell types, tissues and tumors. To address a role of NudC in mitosis, one study reduced NudC mRNA levels with an inducible-ribozyme strategy and showed an increase in the number of cells exhibiting multipolar spindles (Zhang et al., 2002). These results suggest that NudC plays a role in spindle formation during mitosis but the mechanism involved is unknown. Dawe et al. (Dawe et al., 2001) used Caenorhabditis elegans to silence nud-1, the nematode NudC ortholog, and showed that this led to embryonic lethality with defects in centrosome-pronuclear rotation at the time of the first embryonic cell division. In further support of a mitotic role for NudC, we recently demonstrated that NudC is phosphorylated at the onset of mitosis, and that NudC is both an in vitro and in vivo substrate for the mitotic polo-like kinase Plk1 (T. H. Zhou, J. P. Aumais, X. Liu, L.-y. Yu-Lee and R. L. Erikson, unpublished). Plk1 phosphorylates NudC in vitro on two serine residues, S274 and S326, in the highly conserved C terminus, and phosphorylation of NudC on these two sites in vivo appears to influence cytokinesis. How phosphorylated NudC mediates this effect is unclear.
To investigate further the role of NudC in mitosis and cytokinesis in mammalian cells, we altered NudC levels either by small interfering RNA (siRNA) silencing (Elbashir et al., 2001) to reduce endogenous levels of NudC, or adenovirus-mediated gene transfer to overexpress NudC ectopically. Here, we report that altering NudC levels inhibited cell proliferation and led to an increase in the proportion of multinucleated cells. Complementary studies using time-lapse video microscopy on C. elegans embryos extended our findings by demonstrating that cytokinetic furrows regressed when the nud-1 gene was silenced by RNA-mediated interference (RNAi). In both NudC altered mammalian cells and C. elegans embryos, a remarkably similar and consistent phenotype was the loss of midzone microtubule organization, an event that probably contributes to the observed failure in cytokinesis. Further, altering NudC levels resulted in the mislocalization of Plk1 from various mitotic structures, including the midbody during cytokinesis. These results show that a proper level of NudC is crucial for mitotic progression and completion of cytokinesis, and further suggest that NudC might function by regulating the stability of the cytokinetic furrow and microtubule organization during cytokinesis.
Materials and Methods
Small interfering RNAs
Deprotected and double-stranded 21-nucleotide RNAs were synthesized by Dharmacon Research (Lafayette, CO) (Elbashir et al., 2001). The targeted sequence of rat NudC cDNA (accession no. X82445) was AACACCTTCTTCAGCTTCCTT, corresponding to the protein-coding region 88-108 bp relative to the first nucleotide of the start codon (Axtell et al., 1995). This sequence is conserved in rat, mouse and human NudC (Moreau et al., 2001). A firefly (Photinus pyralis) luciferase siRNA (GL2, accession no. X65324) served as a negative control. siRNAs were resuspended in RNase-free water at a final concentration of 20 pmol μl–1 and stored at –20°C until use.
Cell culture and transfection
HeLa cells were grown in Dulbecco's modified Eagle Medium (DMEM, Gibco) supplemented with 5% fetal bovine serum (FBS, Atlanta Biologicals), 100 U ml–1 of penicillin/streptomycin (Gibco), 50 μg ml–1 Gentamycin (Sigma) and 1% (v/v) L-glutamine (Gibco). Cells were plated 24 hours prior to transfection in 12-well plates at a density of 25,000-30,000 cells per well. For immunocytochemical studies, cells were plated on poly-D-lysine-coated coverslips (Fisher). For each transfection, 200 pmol (10 μl) of the appropriate siRNA was diluted in 80μ l Opti-MEM I (Gibco). In a separate tube, 3 μl Oligofectamine (Invitrogen) was diluted in 7 μl Opti-MEM I and incubated at room temperature for 5 minutes. The diluted Oligofectamine was mixed gently with the diluted siRNA and incubated at room temperature for 20 minutes. RNA complexes were added to cells with 400 μl serum-free Opti-MEM I and incubated for 4 hours at 37°C, 5% CO2, after which time FBS was added to the transfections to a final concentration of 4% (v/v) and a final volume of 0.75 ml.
Cells were harvested by scraping in 200 μl TEN (20 mM Tris, pH 7.4, 100 mM NaCl, 5 mM EDTA and 0.5% Triton X-100) supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF) and protease-inhibitor cocktail (both from Sigma), and subjected to two cycles of freeze-thaw. Cell lysates were clarified by centrifugation. Protein concentrations were determined using Bradford reagent (Bio-Rad, Hercules, CA), and 5-10 μg total protein were resolved by SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad). Membranes were immunoblotted with affinity-purified rabbit anti-rat NudC C-terminus peptide (C peptide) antibodies or rabbit anti-rat MBP-NudC fusion protein antibodies (Morris and Yu-Lee, 1998) as indicated, anti-Plk1 monoclonal antibodies (Zymed), or mouse anti-β-actin monoclonal antibodies (AC-40, Sigma). Immunoblots were developed by enhanced chemiluminescence.
After 24-96 hours, cells were fixed for 20 minutes at room temperature in PEM (80 mM K-PIPES, 5 mM EGTA, 2 mM MgCl2) with 4% (w/v) polyethylene glycol-6000 (PEG), 4% (v/v) ultrapure electron-microscopy-grade formaldehyde (Electron Microscopy Sciences, Ft Washington, PA) and 0.5% Triton X-100. To localize NudC to microtubule structures, cells were briefly treated with 0.5% Trion-X-100 in PEM/PEG for 1 minute at room temperature to remove soluble NudC and then fixed as described above. Cells were then blocked in PBS with 2% (w/v) bovine serum albumin (BSA) and 2% (v/v) normal goat serum (Sigma) and incubated with 0.5 μg ml–1 anti-NudC C peptide, anti-β-tubulin (tub2.1, Sigma) or anti-Plk1 antibodies followed by incubation with Texas-red goat-anti-rabbit IgG (Molecular Probes) or FITC-conjugated goat-anti-mouse IgG (Southern Biotechnologies, Birmingham, AL). Slides were mounted in Vectashield mounting medium (Vector Labs, Burlingame, CA) containing 0.2 mg ml–1 4′, 6-diamidino-2-phenylindole (DAPI). Images were acquired using a RT Color Spot digital camera (Diagnostic Instruments) mounted on a Leica DME microscope with red, green or blue emission filters. Figures were compiled using Spot Software (version 3.0) and Photoshop (version 5.5; Adobe, Mountainview, CA).
Untransfected HeLa cells or HeLa cells transfected with 200 pmol NudC or luciferase (Luc) siRNA were harvested at 24 hour intervals in 100 μl Trypsin/EDTA (Gibco) and resuspended in 400 μl growth medium. Cells were pipetted repeatedly to dislodge cell clumps and were counted in quadruplicates with a HY Lite Neubauer hemacytometer (Hausser Scientific) with Trypan blue exclusion. Values are presented as (cell number× 104) ml–1. HeLa cells transduced with recombinant adenovirus containing either the NudC or the Luc gene were also counted in a similar manner.
Recombinant adenovirus constructions and transduction
Full-length NudC was inserted into the adenovirus (Ad) shuttle vector pXCMV at the HindIII/NotI sites to generate pXCMV-NudC. The recombinant adenovirus Ad-NudC was generated by co-transfecting pXCMV-NudC and pJM17, which contains the adenoviral genome with the E1 gene deleted, into 293 cells as previously described (Estrera et al., 2001). Recombinant virus (Ad-NudC or Ad-Luc) was transduced into HeLa cells at an M.O.I. of 1. After 24 hours, the growth medium was changed and the cells were cultured for 96 hours.
C. elegans culturing methods
Wild-type C. elegans strain N2 (Bristol variety) was cultured under standard conditions (Brenner, 1974). To perform RNAi studies, the full-length cDNA of nud-1 (F53A2.4), the C. elegans ortholog of nudC, was cloned into the vector L4440 (Timmons et al., 2001) and transformed into HT115 (DE3). Transformed bacteria were grown for 14 hours at 37°C in 2 ml LB supplemented with 12.5μ g ml–1 tetracycline and 50 μg ml–1 ampicillin. These cultures were plated onto standard worm plates containing 0.8 mM IPTG and 50 μg ml–1 carbenicillin and allowed to grow overnight at room temperature. Dauer N2 larvae were then placed onto these plates and their offspring examined after 48 hours at 26°C (Kamath et al., 2000). Embryos analysed by time-lapse video microscopy were dissected in M9 medium from N2 worms treated with nud-1 dsRNA and placed on a 2% agarose pad with a coverslip mounted on top. Early development was typically analyzed in wild-type and RNAi-treated embryos from shortly after fertilization through two rounds of cell division (∼40 minutes).
Immunostaining of C. elegans embryos
Immunostaining of embryos was performed as previously described (Guo and Kemphues, 1995). Monoclonal anti-α-tubulin antibody (T9026, Sigma) was used at 1:400 and visualized using goat-anti-mouse AlexaFluor488 (Molecular Probes) at 1:800 dilution. DNA was visualized by staining with DAPI. Embryos were analysed using a Nikon Eclipse E800 epifluorescence microscope equipped with DIC optics and Endow GFP HYQ and UV-2E/C DAPI filter cubes (Chroma). Images were captured with a Spot RT CCD camera (Diagnostic Instruments). MetaMorph Software (Universal Imaging) was used for pseuodocoloration of images and image overlays. Images were processed with Adobe Photoshop 7.0.
NudC localization during M phase
To determine the subcellular localization of NudC in mitotic cells, randomly growing HeLa cells were stained with anti-NudC antibodies (red) (Morris and Yu-Lee, 1998), anti-β-tubulin antibodies (green) to visualize microtubule-containing subcellular structures and DAPI (blue) to visualize chromatin/DNA and to stage the cells in mitosis. NudC staining localized with β-tubulin staining on multiple mitotic structures that are important for M-phase progression (Fig. 1). NudC is localized at the microtubule organizing center (MTOC) during prophase, prometaphase and metaphase (Fig.1A-I) and on the mitotic spindles during prometaphase and metaphase (Fig. 1D-I). NudC is also found on the midzone microtubules at anaphase/telophase (Fig. 1J-L) and at the midbody during cytokinesis (Fig. 1M-O). These results suggest that NudC is associated with microtubule-based structures important for mitosis progression and cytokinesis.
NudC downregulation by siRNA inhibits cell proliferation
To determine the role of NudC during cellular replication in mammalian cells, we downregulated endogenous NudC by RNAi (Elbashir et al., 2001) and examined the effects of NudC gene silencing on NudC protein levels and cell growth. HeLa cells were transfected with either NudC siRNA or a firefly luciferase Luc siRNA, which served as a control against nonspecific gene silencing. Total cell lysates were prepared 24 hours, 48 hours, 72 hours and 96 hours post transfection and immunoblotted with anti-NudC antibodies. Cells transfected with NudC siRNA showed a reproducible and significant (80-90%) reduction in the amount of steady-state 42 kDa NudC protein 48 hours after transfection (Fig. 2A, lane 3). The reduction in NudC levels was sustained for up to 96 hours after transfection (Fig. 2A, lanes 4-5). By contrast, NudC levels in cells transfected with Luc siRNA were unaffected over 96 hours of incubation (Fig. 2A, lanes 6-9) and remained comparable to those found in the untransfected control cells (Fig. 2A, lane 1). Total lysates were also blotted with an anti-β-actin antibody to show equal loading of proteins (Fig. 2A, lower panel). Several doses of NudC siRNA were applied over a 3-day period to determine whether repeated applications of the siRNA duplex could enhance NudC gene silencing. Repeated dosing with NudC siRNA was no more efficient in depleting NudC levels than a single initial application (data not shown). Therefore, the regimen for subsequent experiments was a single dose of NudC siRNA transfection followed by a 72-hour incubation. NudC siRNA transfected cells showed reduced proliferation after 72-96 hours post transfection when compared with Luc-siRNA-treated or untransfected control cells (Fig. 2B). These results support the idea that NudC plays a role in mitosis and cell proliferation.
NudC depletion induces multinucleation and results in a failure in cytokinesis
To investigate the basis of the growth inhibition of NudC-depleted cells, HeLa cells were transfected with either Luc siRNA or NudC siRNA for 72 hours and examined by immunofluorescence microscopy. siRNA-mediated NudC gene silencing was observed in over 90% of cells as determined by a lack of NudC staining (red) (Fig. 3B), which agrees with reduced NudC protein expression by western-blot analysis (Fig. 2A). In control cells, NudC staining was generally observed in all the cells, with higher levels in mitotic cells (Fig. 3A, arrow), in agreement with previous studies showing that NudC protein levels double during mitosis (Morris and Yu-Lee, 1998; Aumais et al., 2000). Downregulation of NudC by siRNA led to several morphological phenotypes. MTOC structures were not distinct in NudC-siRNA-treated cells, in contrast to control cells, which showed MTOCs located characteristically on one side of the nucleus of interphase cells (Fig. 3A, asterisk). A significant (sixfold) increase in the number of large, flat cells containing multiple nuclei (Figs 3B, arrows) was observed in NudC-siRNA-treated cells (12%) compared with Luc-siRNA-treated controls (2%) (Fig. 3C). A 60% increase in the number of cells connected to each other by cytoplasmic bridges was found in NudC-siRNA-compared with Luc-siRNA-treated cells (Fig. 3B, arrowheads), suggesting a delay or inhibition in cleavage, cell separation or cytokinesis.
To investigate further the basis of multinucleation, we examined more closely the phenotypes of NudC-downregulated cells during M-phase progression. During prometaphase, spindles appeared to be less focused in NudC-siRNA-treated cells (Fig. 4A,E). Misaligned chromosomes were seen during metaphase in NudC-siRNA-treated cells (Fig. 4F, arrow and inset). Furthermore, striking abnormalities in the midzone/midbody structure at anaphase/telophase and cytokinesis were observed. Midzone/midbody microtubules were either missing or poorly organized (Fig. 4G-I, arrowheads) compared with control cells (Fig. 4C,D, arrowheads) as revealed by β-tubulin staining (green). Many cells had divided and re-entered interphase, as judged by the decondensed DNA, but they remained connected to each other by either thick (Fig. 4J) or very long and thin (Fig. 4K) intercellular cytoplasmic bridges. These cells failed to separate, resulting in a cell structure containing two or more nuclei of abnormal size and shape (Fig. 4J). These results suggest that multinucleation in NudC-downregulated cells might be due to incomplete cytokinesis.
NudC downregulation results in mislocalization of the mitotic kinase Plk1
To determine the consequence of a lack of microtubule organization at the midzone/midbody in NudC-depleted cells, we examined the presence of proteins that are known to localize in the midzone/midbody matrix and to play a role in cytokinesis (Field et al., 1999). One such factor is the mitotic kinase polo-like kinase (Plk1) (Glover et al., 1998), which we have recently shown to interact with and phosphorylate NudC during mitosis (T. H. Zhou, J. P. Aumais, X. Liu, L.-y. Yu-Lee and R. L. Erikson, unpublished). We determined the effect of NudC downregulation on the level and subcellular distribution of Plk1. HeLa cells were transfected with either Luc or NudC siRNA for 72 hours and the cell lysates were immunoblotted with anti-NudC, anti-Plk1 or anti-β-actin antibodies. In NudC-siRNA-treated cells, NudC levels were downregulated (Fig. 5A, top), as also observed in Fig. 2A, but Plk1 levels remained essentially unchanged (Fig. 5A, middle). β-Actin was used as a loading control (Fig. 5A, bottom).
We next examined Plk1 localization during M phase in control and NudC-downregulated cells. Plk1 staining was clearly localized to the centrosomes (Fig. 5Ba,c) and midbody (Fig. 5Be) in control cells, consistent with the multiple well-described roles for Plk1 during M-phase progression (Glover et al., 1998; Donaldson et al., 2001). By contrast, Plk1 was mislocalized in 40-45% of cells progressing through various stages of M phase in the NudC-downregulated cells. In NudC-siRNA-transfected cells, Plk1 staining was diffuse and not localized to the centrosomes (Fig. 5Bg,i) or midzone (Fig. 5Bk). Lagging chromosomes were observed in the midzone region in NudC-downregulated cells (Fig. 5Bl, arrow), suggesting a defect in DNA segregation. These results show for the first time that, in NudC-downregulated cells, the mitotic kinase Plk1 is not properly localized to the mitotic structures in dividing mammalian cells.
NudC overexpression by adenovirus-NudC transduction inhibited cell proliferation
We next determined the effect of overexpressing NudC on cell proliferation. Recombinant adenovirus containing the NudC gene (Ad-NudC) was used for high efficiency expression of NudC in mammalian cells. Recombinant adenovirus containing the firefly luciferase gene (Ad-Luc) was used as a control. At 24 hours, 48 hours, 72 hours and 96 hours post transduction, total cell lysates were immunoblotted with anti-NudC antibodies. Cells transduced with Ad-NudC showed a reproducible increase in NudC protein levels 24-96 hours post transduction (Fig. 6A, lanes 5-8, top). By contrast, NudC levels remained unchanged in cells transduced with the Ad-Luc control (Fig. 6A, lanes 1-4). Reblotting with anti-β-actin antibodies showed equal loading of proteins (Fig. 6A, bottom). Ad-NudC-transduced HeLa cells grew at the same rate as the Ad-Luc-transduced control cells for the first 48 hours (Fig. 6B). However, by 72-96 hours post transduction, overexpression of NudC led to an inhibition of cell proliferation in the Ad-NudC-transduced cells as compared with the control cells (Fig. 6B). These results show that overexpression of NudC also led to an inhibition of cell proliferation in mammalian cells and is not the result of viral transduction.
NudC overexpression induces multinucleation and a failure in cytokinesis
To determine the basis of Ad-NudC mediated growth inhibition, HeLa cells were transduced with Ad-NudC or Ad-Luc for 96 hours and examined by immunofluorescence microscopy. Overexpression of NudC led to an increase in multinucleated cells as early as 24 hours after Ad-NudC transduction (Fig. 7D, arrows). At 72 hours post Ad-NudC transduction, most of the cells showed significant increases in both cell size and nuclear size (Fig. 7F), and appeared spread out and flat. Some of the large cells contained micronuclei (Fig. 7F, arrow and inset). Indeed, by 96 hours post Ad-NudC transduction, greater than 90% of the cells have become very large and multinucleated. Typically, these large cells contained multiple nuclei (Fig. 7G) or a few enlarged nuclei (Fig. 7H), which suggests a failure in cytokinesis. Micronuclei were observed in these large multinucleated cells (Fig. 7G,H, arrow). Multinucleated cells with persistent cytoplasmic connections were also observed using transient transfection of NudC expression vectors (data not shown), thus suggesting that the observed multinucleation phenotype is unlikely to be a result of viral transduction. By contrast, in the Ad-Luc transduced control cells nuclear size remained constant (Fig. 7A-C) and the cell number increased between 24 hours and 72 hours, in agreement with the cell proliferation data (Fig. 6B). These results show that overexpression of NudC also caused inhibition in cell proliferation and led to an increase in multinucleation.
Overexpression of NudC resulted in cytokinetic abnormalities at 96 hours post Ad-NudC transduction. The most striking phenotype was that many cells showed abnormal midbody and cytoskeletal structures (Fig. 8A,B, compare with Fig. 4C,D). Lagging chromosomes were observed in cells exhibiting unusual midbody structures (Fig. 8A, inset), suggesting a defect in DNA segregation. Many cells were connected by an intercellular bridge that contained disorganized microtubules (Fig. 8C, arrow). Large multinucleated cells were connected by an abnormally dense midbody-like structure (Fig. 8D, DNA and micronuclei in inset). These abnormal cellular connections and cytoskeletal morphologies, which persisted after the cells have re-entered interphase, probably resulted from abortive cytokinesis. There was also an increase in cells displaying numerous MTOCs (Fig. 8E) or multipolar spindles with unattached metaphase chromosomes (Fig. 8F, inset). Unattached, apolar, fragmented and sometimes curving microtubules at the cell cortex were found in NudC-overexpressing cells (Fig. 8C,G, arrows). Thus, overexpression of NudC disrupted cytoskeletal structures and microtubule organization in the cytoplasm and at the midbody.
NudC overexpression results in the mislocalization of the mitotic kinase Plk1
Because reduction of NudC levels by RNAi resulted in Plk1 mislocalization during M phase (Fig. 5B), we next determined whether NudC overexpression has a similar effect on the localization of Plk1. HeLa cells were transduced with Ad-NudC or Ad-Luc for 96 hours and examined by immunofluorescence microscopy. In Ad-Luc-transduced control cells, normal Plk1 staining was observed at the centrosomes during metaphase (Fig. 9A) and the midbody during cytokinesis (Fig. 9C). By contrast, Plk1 was mislocalized in 30% of Ad-NudC-transduced cells. In NudC-overexpressing cells, Plk1 exhibited punctate and irregular staining, as shown in cells connected by an intercellular bridge (Fig. 9E) and in large multinucleated cells (Fig. 9G). These results show that Plk1 exhibits aberrant localization in NudC-overexpressing cells.
NUD-1 depletion in C. elegans causes cleavage-furrow regression
We further used a different model genetic system, the nematode C. elegans, to investigate the functional consequence of depletion of NUD-1, the C. elegans ortholog of mammalian NudC (Dawe et al., 2001), on mitosis. Using injection of nud-1 dsRNA, Dawe et al. (Dawe et al., 2001) previously showed that the putative zygotic function of NUD-1 involves the development of the germ line and nervous system. Embryonic cell division defects were observed using this method. However, oogenesis fails rapidly after injection of nud-1 dsRNA, precluding a detailed analysis of embryonic phenotypes by this method. We used the new technique of RNAi feeding to expand upon these prior analyses by examining cell division defects associated with NUD-1 knockdown during development (Kamath et al., 2000; Timmons et al., 2001). In addition to a defect in pronuclear rotation in the one-celled embryos, as previously described (Dawe et al., 2001), nud-1 RNAi feeding yields a more severe, highly reproducible defect in late cytokinesis during the first cell cycle following fertilization (Fig. 10) (see also movies online: http://jcs.biologists.org/supplemental). In both wild-type and NUD-1-depleted embryos, a spindle developed and elongated along the longitudinal axis (Fig. 10A,B,E,F). The first cleavage furrow appeared at the appropriate time and position in both types of embryos (Fig. 10C,G). The wild-type embryo underwent cleavage-furrow ingression, which resulted in the generation of two differently sized cells (Fig. 10D). By contrast, in the nud-1 RNAi embryo, the cleavage furrow stalled (Fig. 9H, arrow indicates the furthest point of ingression of the furrow) and then quickly regressed (Fig. 10I-K), resulting in a multinucleated one-celled embryo (Fig. 10L).
To examine further the late cytokinesis failure in the nud-1 RNAi embryos, we determined whether midzone microtubules were present by staining for α-tubulin (green) when NUD-1 was depleted. Midzone microtubules are typically robust in anaphase wild-type embryos (Fig. 11A-C). However, in nud-1 RNAi embryos, midzone microtubules were absent in 26% (10/39) of one-cell staged embryos (Fig. 11D-F). The remaining 74% of one-cell staged embryos (29/39) contained midzone microtubules that were less well defined (Fig. 11G-I). Frequently, nud-1 RNAi embryos (15/29) containing weak midzone microtubules exhibited chromatin bridges, which are indicative of DNA mis-segregation (Fig. 11H,I). After the first cell cycle, cytokinetic furrows in NUD-1-depleted embryos continued to form and regress dynamically without complete stabilization of the furrows, as observed in live embryos (data not shown). To confirm this, older embryos were analysed by α-tubulin staining (Fig. 11L). Multipolar spindles and additional DNA were detected in these embryos, suggesting that embryos depleted for NUD-1 underwent multiple rounds of the cell cycle without completing cytokinesis. Together, the combined mammalian and C. elegans studies show that NudC/NUD-1 is crucial for cleavage-furrow ingression, midzone microtubule organization and the completion of cytokinesis.
NudC encodes a protein that is highly conserved from fungus to human, suggesting that it has an evolutionarily conserved function. In this study, we show for the first time that NudC is involved in cleavage-furrow ingression and microtubule organization at the midbody, and thus plays an important role in cytokinesis. We first show that depleting endogenous NudC (Fig. 2) or ectopically overexpressing NudC (Fig. 6) inhibited proliferation in mammalian cells. The proliferation defect could be a result of increased incidences of both mitotic and cytokinetic abnormalities. In NudC-altered cells, we observed misaligned chromosomes during metaphase (Fig. 4) and multipolar spindles (Fig. 8), which was also observed by others in cells in which NudC mRNA was downregulated (Zhang et al., 2002). The most striking phenotype when NudC levels were altered was the increase in multinucleated cells (Figs 3, 7). Multinucleation can arise from centrosome aberrations (Brinkley, 2001; Nigg, 2002) or defects in cytokinesis (Meraldi et al., 2002; Nigg, 2002). Our data are consistent with the interpretation that the multinucleated cells resulted from a failure to complete cytokinesis. Our studies show that NudC is localized on midzone microtubules and at the midbody in epithelial cells (Fig. 1) and fibroblasts (data not shown). Altering NudC levels led to a disorganization of midzone/midbody microtubule matrices during anaphase/telophase and cytokinesis (Figs 4, 8), supporting a role for NudC in midzone microtubule dynamics. Midzone microtubules have an important role in promoting the ingression of the cleavage furrow and a continued presence of midzone microtubules is required for the completion of cytokinesis (Wheatley and Wang, 1996; Field et al., 1999; Straight and Field, 2000; Murata-Hori and Wang, 2002). Furthermore, in time-lapse recordings of developing one-cell C. elegans embryos in which nud-1 has been silenced by RNAi, cleavage furrows formed but quickly regressed (Fig. 10), which probably resulted from absent or weak midzone microtubules found in these telophase-staged one-cell embryos (Fig. 11). Previous studies have also localized NudC to both the animal and vegetal poles of the cleavage furrow in two-celled embryos of the amphibian Pleurodeles waltl (Moreau et al., 2001). Taken together, these observations support potential NudC functions in cleavage-furrow ingression, midbody microtubule organization and the completion of cytokinesis, thus suggesting an evolutionarily conserved role for NudC in midzone/midbody microtubule stability and matrix organization.
Consistent with midzone microtubule disruption, the mitotic kinase Plk1, which normally localizes to centrosomes and midbody microtubule matrix of mitotic cells, is mislocalized in cells in which NudC levels were altered (Figs 5, 9). Plk1 is known to play several roles during M phase progression, including mitotic entry, mitotic exit and cytokinesis (Glover et al., 1998; Nigg, 2001; Donaldson et al., 2001) and is suggested to affect spindle organization and microtubule dynamics (Tong et al., 2002). Plk1 depletion by RNAi resulted in an increase in binucleate cells and cells connected by a long cytoplasmic bridge (Liu and Erikson, 2002), whereas overexpression of Plk1 resulted in large cells containing multiple, fragmented nuclei (Mundt et al., 1997). These phenotypes are reminiscent of both depletion (Fig. 3, Fig. 4J,K) and overexpression (Figs 7, 8, 9) of NudC, respectively. Time-lapse microscopy studies revealed that Plk1 is involved in cleavage furrow ingression, which supports a role of Plk1 in cytokinesis (Seong et al., 2002; Chase et al., 2000). Notably, we recently demonstrated that NudC is phosphorylated by Plk1 in vitro and that NudC phosphorylation increases during M phase and correlates with increased Plk1 kinase activity in vivo (T. H. Zhou, J. P. Aumais, X. Liu, L.-y. Yu-Lee and R. L. Erikson, unpublished). Furthermore, phosphorylation of NudC by Plk1 on the highly conserved C-terminal S274 and S326 residues seems to be important during cytokinesis. These results suggest that, although NudC activity might be regulated by Plk1 phosphorylation, NudC might, in turn, be involved in the temporal and spatial localization of Plk1 during M phase. Although it is possible that Plk1 mislocalization from the midbody in NudC altered cells is a consequence of the effects of NudC on microtubule stability and/or organization, we suggest that Plk1 mislocalization contributes to cytokinesis failure and the increased incidence of multinucleation and aneuploidy in cells with altered NudC expression.
Studies in the fungus A. nidulans have shown that NudC is genetically and functionally linked to other nud loci that encode components of the dynein/dynactin complex (Morris, 2000). Biochemical evidence in mammalian cells show that NudC is associated with the microtubule minus-end-directed dynein/dynactin motor complex (Aumais et al., 2001). NudC also forms a complex with Lis1 (A. nidulans NUDF) and the NudC/Lis1 complex is localized to the MTOC and the microtubule network in neurons and fibroblasts (Morris and Yu-Lee, 1998; Aumais et al., 2001). We previously suggested that the NudC/Lis1 complex plays a role in mediating nuclear movement and transport in neuronal progenitors as well as in migrating neurons (Aumais et al., 2001). In dividing cells, Lis1 potentiates dynein function and is suggested to regulate dynein/dynactin interactions with the cell cortex (Faulkner et al., 2000; Vallee et al., 2000) and with the kinetochore (Tai et al., 2002; Coquelle et al., 2002), through association with microtubule plus ends and CLIP-170 (Tai et al., 2002; Coquelle et al., 2002). Lis1 also regulates microtubule dynamics (Sapir et al., 1997; Smith et al., 2000). Given the observations that NudC/Lis1 interacts with the dynein/dynactin complex (Morris et al., 1998; Aumais et al., 2001), that dynein/dynactin subunits are localized to the midbody (Karki et al., 1998; Campbell et al., 1998; Karki and Holzbaur, 1999) and that NudC is localized to the cleavage furrow (Moreau et al., 2001) and midbody (this study), we speculate that the NudC/Lis1/dynein/dynactin complex functions in regulating cytokinesis either by regulating midbody matrix organization or by targeting regulatory and effector molecules to the midzone and midbody during cytokinesis. Our studies also raise the interesting possibility that Plk1 might affect the formation and/or function the NudC/Lis1/dynein/dynactin complex. In support of this notion, we recently observed that Plk1 associates with the dynein/dynactin complex upon entry into mitosis (T. H. Zhou, J. P. Aumais, X. Liu, L.-y. Yu-Lee and R. L. Erikson, unpublished), the significance of which awaits further experimentation.
Many proteins are localized to the midzone and play a role during cytokinesis. These include chromosomal passenger proteins, mitotic kinases such as Plk1 and aurora kinases, GTPases and kinesin-like motor proteins CHO1/MKLP1, which are needed for microtubule stability and microtubule-dependent protein and vesicular transport to support furrow ingression and cytokinesis (Straight and Field, 2000). Inappropriate expression of some of these proteins also results in cleavage furrow regression and cytokinesis failure (Matuliene and Kuriyama, 2002; Raich et al., 1998; Liu and Erikson, 2002; Seong et al., 2002; Chase et al., 2000). It would be interesting to determine whether these diverse groups of proteins, which are needed for the successful completion of cytokinesis, are also missing or mislocalized in NudC altered cells. Our findings that NudC is involved in Plk1 localization during M phase, cleavage-furrow ingression and midzone/midbody microtubule organization suggest novel functions for this highly conserved protein in mitosis and in cytokinesis.
We thank Farrah Kheradmand for use of the fluorescence microscope, Ming-Jer Tsai for support, Bill Brinkley and David Stenoien for critical comments, and Jenny Whitworth for assistance with C. elegans embryo preparations. This work was supported by a Basil O'Connor Scholar Award from The March of Dimes Birth Defects Foundation (5-FY00-592) to G.A.C., The University of Alabama Research Advisory Committee to G.A.C., an Undergraduate Research Program Grant from the Howard Hughes Medical Institute to The University of Alabama, a McNair Fellowship to M.N. and grants from the National Institutes of Health (RO1 CA86342) to S.-H.L. and (RO1 DK53176) to L.-y.Y.-L.
Movies available online
- Accepted February 5, 2003.
- © The Company of Biologists Limited 2003