The long myosin light chain kinase is differentially phosphorylated during interphase and mitosis

Experimental Cell Research 299 (2004) 303 – 314 www.elsevier.com/locate/yexcr

The long myosin light chain kinase is differentially phosphorylated during interphase and mitosis Natalya G. Dulyaninova, Anne R. Bresnick* Department of Biochemistry, Albert Einstein College of Medicine, Bronx, NY 10461, United States Received 20 April 2004, revised version received 14 June 2004 Available online 22 July 2004

Abstract We have shown previously that the activity of the long myosin light chain kinase (MLCK) is cell cycle regulated with a decrease in specific activity during mitosis that can be restored following treatment with alkaline phosphatase. To better understand the role and significance of phosphorylation in regulating MLCK function during mitosis, we examined the phosphorylation state of in vivo derived MLCK. Phosphoamino acid analysis and phosphopeptide mapping demonstrate that the long MLCK is differentially phosphorylated on serine residues during interphase and mitosis with the majority of the phosphorylation sites located within the N-terminal IgG domain. Biochemical assays show that Aurora B binds and phosphorylates the IgG domain of the long MLCK. In addition, phosphopeptide maps of the endogenous full-length MLCK from mitotic cells and in vitro phosphorylated IgG domain demonstrate that Aurora B phosphorylates the same sites as those observed in vivo. Altogether, these studies suggest that the long MLCK may be a cellular target for Aurora B during mitosis. D 2004 Elsevier Inc. All rights reserved. Keywords: Myosin-II; Phosphorylation; Myosin light chain kinase; Cell division; Mitosis; Aurora B

Introduction Myosin light chain kinase (MLCK) is a calcium and calmodulin-dependent enzyme that exclusively phosphorylates the regulatory light chain of myosin-II on Thr-18 and Ser-19. In nonmuscle cells, MLCK is a key regulator in processes involving myosin-II activation such as cell spreading and motility, stress fiber formation, and cytokinesis [1]. In vertebrates, the smooth/nonmuscle MLCK gene expresses two MLCK transcripts; a high molecular weight MLCK (long MLCK, approximately 210 kDa) and a low molecular weight MLCK (short MLCK, 108–125 kDa). Both isoforms are expressed in smooth and nonmuscle tissues; however, the

Abbreviations: MLCK, myosin light chain kinase; Ig, immunoglobulin; GFP, green fluorescent protein. * Corresponding author. Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. Fax: +1 718 430 8565. E-mail address: [email protected] (A.R. Bresnick). 0014-4827/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2004.06.015

long MLCK is the predominant isoform expressed in cultured cells [2]. The long MLCK encompasses the short MLCK and also contains an N-terminal extension of approximately 900 amino acids [3,4]. The N-terminal extension of the long MLCK is comprised of six immunoglobulin (Ig) motifs, which we refer to collectively as the IgG domain, and two additional actin-binding motifs in tandem (DXRs). Although the functions of the six Ig motifs in the N-terminal extension have yet to be elucidated, our previous studies have shown that the IgG domain contributes to targeting the long MLCK to stress fibers and the cleavage furrow during interphase and mitosis, respectively [5,6]. In addition, given that the long and short MLCKs have similar K m and V max values [7], the IgG domain is not essential for the catalytic function of the long isoform. Our previous studies demonstrated that the activity of the long MLCK is cell cycle regulated with a two-fold decrease in specific activity during mitosis that can be restored following treatment with alkaline phosphatase [5]. In addition, expres-

304

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

sion of the IgG domain + 5 DXRs of the long MLCK [6] or inhibition of Aurora B [8–10] induces extension of the astral microtubules, bundling of the kinetochore microtubules, and chromosome miscongregation. Aurora B kinase is an evolutionarily conserved serine/threonine kinase that functions in both early and late mitotic events [11–13], and is a component of the chromosomal passenger complex that contains INCENP and survivin. This complex localizes to the centromeres during prometaphase and then relocalizes to the central spindle in anaphase [14]. Recently, Aurora B function has been implicated in the regulation of numerous mitotic events including chromosome congression, kinetochore assembly and the promotion of amphitelic kinetochoremicrotubule attachments, the establishment and maintenance of the spindle-assembly checkpoint, and the final stages of cytokinesis [9,10,13]. To better understand the role and significance of phosphorylation in regulating MLCK function during mitosis, we examined the phosphorylation state of in vivo derived MLCK and the potential role of Aurora B in mediating MLCK phosphorylation during mitosis. Our observations indicate that the long MLCK is highly phosphorylated on serine residues with the majority of the phosphorylation sites located within the IgG domain. In addition, in vitro assays demonstrate that Aurora B binds to the long MLCK and phosphorylates serine residues within the IgG domain. These observations suggest that Aurora B may regulate MLCK during mitosis.

Materials and methods Cell culture, synchronization, and

32

P-labeling

HeLa cells were grown as monolayer cultures at 378C in a humidified atmosphere of 5% CO2 in DMEM containing 10% FBS. Cells were synchronized by the addition of 250 ng/ ml nocodazole for 16 h [6]. For 32P-labeling interphase cells, HeLa cells were plated on 150-mm culture dishes at 30–50% confluence and left to attach overnight. The next morning, the medium was replaced with phosphate-free DMEM. After 30 min, the phosphate-free medium was removed and replaced with 8 ml of phosphate-free medium containing 1% dialyzed FBS and 250–500 ACi/ml [32P]-orthophosphate (NEN Life Sciences). After 14–24 h of labeling, mitotic cells were removed by selective detachment and the attached cells were washed three times with 10 ml phosphate-buffered saline (1.5 mM KH2PO4, 8 mM Na2HPO4d 7H2O, 2.6 mM KCl, 137 mM NaCl), scraped into 10 ml ice-cold PBS, and harvested by centrifugation. To radiolabel mitotic cells, synchronized HeLa cells were collected by selective detachment, centrifuged, washed two times with phosphate-free medium, and incubated in phosphate-free medium for 30 min. The medium was then replaced with phosphate-free medium containing 1% dialyzed FBS, 250 ng/ml nocodazole, and 250–400 ACi/ ml [32P]-orthophosphate. After 2–4 h, the cells were collected

by centrifugation and washed three times with PBS to remove the nocodazole. MLCK-GFP fusion constructs and transfections The HeLa and rabbit MLCK-GFP fusions were prepared as described previously [6]. The rabbit short MLCK-GFP construct was generously provided by Dr. Jim Stull (University of Texas Southwestern Medical Center). The rabbit short MLCK corresponds to residues 923–1914 of the HeLa long MLCK and displays 81% identity with the HeLa short MLCK. The rabbit 5 DXRs correspond to residues 862–994 of the HeLa long MLCK and display 93% identity with this region of the HeLa long MLCK. HeLa cells were transfected with the MLCK-GFP constructs using lipofectin (Invitrogen) or polyfect (Qiagen). Immunoprecipitation For immunoprecipitation of the endogenous HeLa long MLCK, 32P-labeled interphase and mitotic cells were washed twice with PBS and resuspended in a lysis buffer composed of 50 mM Tris–HCl pH 8.2, 400 mM NaCl, 0.5% Triton X100, 0.5 mM EDTA, 0.5 mM EGTA, 1 mM DTT, 1 mM PMSF, 50 Ag/ml each of chymostatin, leupeptin, pepstatin A, and 1:100 dilution of phosphatase inhibitor cocktail I (microcystine LR, cantharidin, and ( )-p-bromotetramisole), and phosphatase inhibitor cocktail II (sodium vanadate, sodium molybdate, sodium tartrate, and imidazole) (Sigma). After 20 min on ice, the lysate was clarified by centrifugation at 48C for 15 min at 16,000  g. The supernatant was diluted 1:3 in 20 mM Tris–HCl pH 7.5, 0.5% Triton X-100, 0.5 mM EDTA, 1 mM DTT, 1 mM PMSF with protease and phosphatase inhibitors as described above. Before immunoprecipitation, MLCK polyclonal antibodies were bound to protein A-Sepharose (Sigma) in PBS (10 mM sodium phosphate, pH 7.5, 150 mM NaCl, and 1 mM EGTA) containing 1 mg/ml bovine serum albumin (BSA). The diluted supernatant was incubated with the antibody-protein A-Sepharose for 2 h at 48C. Immune complexes were collected by centrifugation and washed five times with 20 mM Tris–HCl pH 7.5, 150 mM NaCl, 10 mM magnesium acetate, 0.5 mM DTT. Laemmli sample buffer (2) [15] was added to the immune complexes, which were boiled, and the proteins were separated by 6% or 10% SDS-PAGE, 32Plabeled MLCK was visualized by autoradiography of the dried gel. For immunoprecipitation of GFP, the GFP-IgG domain + 5 DXRs fusion, or the rabbit short MLCK-GFP fusion, HeLa interphase extracts were prepared as described above from transiently transfected cells. GFP, the GFP-IgG domain + 5 DXRs fusion, and the rabbit short MLCK-GFP fusion were immunoprecipitated with a GFP polyclonal antibody (Clontech), affinity-purified IgG domain polyclonal antibodies [6], or affinity-purified short MLCK polyclonal antibodies [5], respectively, coupled to protein-A-Sepharose.

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

Phosphoamino acid analysis and phosphopeptide mapping Two-dimensional phosphopeptide mapping and phosphoamino acid analyses were performed using standard methods. Briefly, immunoprecipitates of 32P-labeled endogenous HeLa long MLCK, HeLa IgG domain + 5 DXRs, rabbit short MLCK, or in vitro phosphorylated HeLa IgG domain were subjected to SDS-PAGE and transferred to Hybond nitrocellulose membrane (Amersham Pharmacia Biotech). The membrane was stained with Ponceau S, autoradiographed, and 32P-labeled protein bands were excised and blocked for 30 min at 378C in 0.5% PVP-40 in 100 mM acetic acid. In addition, a portion of the membrane was probed with the appropriate antibodies on immunoblots to confirm that the correct protein bands were analyzed. Tryptic digests were performed in 50 mM NH4HCO3 pH 8.0 at 378C overnight by the addition of 50 Ag TPCK-trypsin (Worthington Biochemical Corp). The following morning, a second aliquot of TPCK-trypsin was added and the digest continued for an additional 3 h. The solution, which typically contained 90% of the starting cpm, was washed, dried, and the pellet was dissolved in 20 Al of Electrophoresis Buffer (formic acid/ acetic acid/dH2O, 1:3:36, by volume) pH 1.9. Approximately 3000 cpm of the digest was applied to a 100 Am  20 cm  20 cm silica gel plate (Merck, UK) and subjected to electrophoresis at 650 V for 47 min to resolve peptides in the first dimension, followed by chromatography in Regular Buffer (n-butanol/pyridine/acetic acid/dH2O, 6.5:5:1:4, by volume) in the second dimension. Phosphopeptides were visualized by autoradiography. For phosphoamino acid analysis, dried tryptic digests were resuspended in 100 Al of constant boiling HCL (5.7 M) and incubated at 1108C for 60 min. The hydrolysate was dried and resuspended in 10 Al of Electrophoresis Buffer, containing 1/15 part of cold phosphoamino acid standards (1.0 mg/ml each of phosphoserine, phosphothreonine, and phosphotyrosine), and spotted (approximately 600 cpm) onto a silica gel plate as described above. Electrophoresis was carried out for 21 min at 1500 V in the first dimension using the pH 1.9 Electrophoresis buffer as the running buffer, and for 21 min at 1500 V in the second dimension using a pH 3.5 Running buffer composed of pyridine/acetic acid/dH2O (1:10:189, by volume). Phosphoamino acid standards were visualized with 0.25% ninhydrin in acetone and the 32P-labeled phosphoamino acids were visualized by autoradiography. GST fusions and protein purification Rabbit short MLCK was generously provided by Dr. Jim Stull (University of Texas Southwestern Medical Center). The GST-Aurora B construct was kindly provided by Dr. William Earnshaw (University of Edinburgh) [16]. The expression and purification of GST-Aurora B were performed as described in Goto et al. [17] and the purified GST-Aurora B was stored at 208C before use.

305

For mammalian expression and purification of a GST fusion of the HeLa long MLCK IgG domain, residues 1– 867 were subcloned into the XhoI–KpnI sites of pXJ-GST [18], kindly provided by Dr. Louis Lim (The National University of Singapore). The final construct contained GST on the N-terminus followed by eight additional amino acids (Gly-Ser-Lys-Leu-Leu-Glu-Ala-Thr) and was confirmed by DNA sequencing. For purification of the GST-IgG domain, approximately 30 h post-transfection, HeLa cells were washed twice with chilled PBS and processed as described above for the MLCK immunoprecipitations. The diluted supernatant was incubated with glutathione-Sepharose for 1 h at 48C. The resin was collected by centrifugation and washed five times with 20 mM Tris–HCl pH 7.5, 500 mM NaCl. The glutathione-Sepharose containing bound GSTIgG domain was then used for the GST pull-down assay. SDS-PAGE analysis and Coomassie staining confirmed that the glutathione-Sepharose contained only the bound GSTIgG domain and no additional proteins (Fig. 4B, left panel). Co-immunoprecipitation assay HeLa cells were synchronized with nocodazole for 16–18 h and mitotic cell lysates prepared as described above. Aurora B was immunoprecipitated using an Aurora B monoclonal antibody (Transduction Laboratories) and the endogenous MLCK was immunoprecipitated using a MLCK monoclonal antibody (Sigma, K-36). The immune complexes were washed five times with 10 mM Tris–HCl pH 7.5, 150 mM NaCl, 10 mM magnesium acetate, 0.5 mM DTT, and then boiled in 2 Laemmli sample buffer [15]. Proteins were separated by 10% SDS-PAGE followed by immunoblotting with the Aurora B and MLCK monoclonal antibodies, respectively. GST pull-down assay Glutathione-Sepharose containing bound GST-Aurora B or GST alone was incubated with a cell lysate prepared from mitotic HeLa cells transiently transfected with the IgG domain + 5 DXRs, or glutathione-Sepharose containing bound GST-IgG domain or GST alone was incubated with a mitotic lysate from untransfected HeLa cells for 2.5 h at 48C. The cell lysates were prepared essentially as described by Crosio et al. [19] by lysing the HeLa cells in EBC buffer (50 mM Tris–HCl, pH 8.0, 400 mM NaCl, 0,5% NP-40, 1 mM PMSF) supplemented with protease and phosphatase inhibitor cocktails as described above. After 20 min on ice, the lysate was clarified by centrifugation at 48C for 15 min at 16,000  g. Before incubation with glutathione-Sepharose, the supernatant was diluted with salt-free EBC buffer to adjust the final salt concentration to 170 mM. The glutathione-Sepharose was washed with EBC buffer containing 170 mM NaCl and the bound proteins were subjected to 10% SDS-PAGE followed by immunoblotting with Aurora B and MLCK monoclonal antibodies, respectively.

306

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

In vitro Aurora B kinase assay The bacterially expressed and purified HeLa IgG domain (residues 2–867) [6] was used as a substrate in an in vitro Aurora B kinase assay. The phosphorylation reaction was performed at 258C in 25 mM Tris–HCl, pH 7.5, 2 mM MgCl2, 3.6 mM EDTA, 100 AM [g-32P]ATP, phosphatase inhibitor cocktails I and II (Sigma), and 17 Ag/ml GST-Aurora B in the presence of 22 Ag/ml calf thymus histone H3 (Roche) or 135 Ag/ml recombinant HeLa IgG domain. Aliquots of the reaction mix were boiled in Laemmli sample buffer [15] and subjected to 8% or 12% SDS-PAGE. The gels were Coomassie-stained, dried, and the bands corresponding to the IgG domain or histone H3 were excised, dissolved in Solvable (Packard BioScience Company), and measured for 32 P incorporation using a Beckman LS 5000 TD scintillation counter.

with alkaline phosphatase [5]. However, these experiments did not distinguish between direct phosphorylation of the long MLCK and phosphorylation of a co-precipitating accessory molecule that modulates kinase activity. To establish that MLCK itself is phosphorylated, interphase and mitotic HeLa cells were labeled with 32P-orthophosphate and the long MLCK was examined for phosphate incorporation. Phosphoamino acid analysis of the long MLCK immunoprecipitates demonstrated that the interphase and mitotic MLCKs are phosphorylated on serine (Fig. 1A). Twodimensional tryptic phosphopeptide maps showed differences in the phosphorylation state of the interphase and mitotic long MLCKs (Fig. 1B). In the interphase MLCK at least four spots (a–d) were detected; whereas in the mitotic MLCK spots a, b, and d were not observed and two additional

Immunofluorescence and microscopy HeLa cells were transfected with the GFP-IgG Domain + 5 DXRs construct as described above. For localization of Aurora B and tubulin, cells were fixed 28 h post-transfection in freshly prepared 4% formaldehyde in PBS for 15 min at room temperature, permeabilized for 3 min in 0.5% Triton X100 in PBS, rinsed with PBS containing 0.02% azide, and blocked in PBS containing 1% BSA. Cells were then incubated with primary antibodies against h-tubulin (clone tub 2.1, Sigma) or an Aurora B mouse antibody (BD Transduction Laboratories) at a dilution of 1:300 in PBS containing 1% BSA. After washing with 1% BSA in PBS, the cells were incubated with the appropriate secondary antibodies in PBS containing 1% BSA and rinsed again. Monoclonal antibodies were visualized using a Cy3-conjugated donkey anti-mouse antibody (Jackson ImmunoResearch Lab) at a dilution of 1:500. Coverslips were briefly incubated with DAPI (Sigma) at 0.5 Ag/ml in 1% BSA in PBS before mounting in Pro-Long Anti-Fade (Molecular Probes). Conventional fluorescence images were acquired using IPLab Spectrum software (Scanalytics, Inc.) with a CoolSNAP HQ interline 12-bit, cooled CCD camera (Roper Scientific) mounted on an Olympus IX70 microscope with a PlanApo 60X, 1.4 N.A. objective (Olympus) and HiQ bandpass filters (Chroma Technology Corp). Images were processed using Photoshop (Adobe Systems).

Results The long MLCK is differentially phosphorylated during interphase and mitosis Our previous studies demonstrated that the activity of the long MLCK is cell cycle regulated; specifically, MLCK exhibits a two-fold decrease in specific activity during mitosis and full activity can be restored following treatment

Fig. 1. The long MLCK is differentially phosphorylated during interphase and mitosis. (A) Phosphoamino acid analysis. The interphase and mitotic MLCKs are phosphorylated on serine residues. Circles indicate the positions of phosphoamino acid standards. (B) Two-dimensional phosphopeptide maps of the mitotic MLCK (m), MLCK from HeLa cells 3.5 h after release from nocodazole arrest (mYi), and the interphase MLCK (i) were resolved by electrophoresis at pH 1.9 followed by ascending chromatography in regular chromatography buffer.

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

spots (e and f) were detected. Phosphopeptide maps of MLCK immunoprecipitates from HeLa cells 3.5 h after release from nocodazole arrest showed an intermediate phosphopeptide pattern when compared to the phosphopeptide maps derived from interphase and mitotic cells. We observed the appearance of spots a, b, and d, and the loss of spots e and f (Fig. 1B, mYi). These maps indicate that mitotic and interphase MLCKs are differentially phosphorylated and also provide some information about the physical properties of the phosphorylated peptides. Spots a–d from interphase and mitotic cells consistently displayed low mobility in the second dimension, independent of the chromatography buffer used, which is characteristic of hydrophilic peptides. The HeLa long MLCK contains 167 serines, making identification of the phosphorylated residues a significant challenge. As a first step toward mapping the phosphorylated serines, GFP fusions of the HeLa IgG domain + 5 DXRs, the rabbit short MLCK (81% identity with the HeLa short MLCK), or rabbit 5 DXRs (93% identity with the HeLa long MLCK 5 DXRs) were immunoprecipitated from interphase HeLa cells labeled with 32P-orthophosphate (Fig. 2A). A comparison of the HeLa and rabbit short MLCKs indicates that they have Ser/Thr/Tyr contents of 81/57/20 and 77/77/ 20, respectively. Importantly, 85% of the serine, 89% of the threonine, and all of the tyrosine residues in the HeLa short MLCK are conserved in the rabbit short MLCK, suggesting that any phosphorylated serines in the C-terminal half of the HeLa long MLCK are likely to be conserved in the rabbit kinase. With respect to the 5 DXRs, the Ser/Thr/Tyr contents are 6/8/0 and 7/7/0 for the HeLa and rabbit 5 DXRs, respectively. All six serine residues and seven of the threonine residues in the HeLa 5 DXRs are conserved in the rabbit 5 DXRs, thus the rabbit 5 DXRs are a suitable substitute for the HeLa 5 DXRs. No phosphate incorporation was detected in immunoprecipitates of GFP alone or in a GFP fusion of the 5 DXRs (data not shown). Phosphoamino acid analyses showed that the IgG domain + 5 DXRs and short MLCK were both phosphorylated on serine (Fig. 2B). The short MLCK also contained a small amount of phosphothreonine. Two-dimensional phosphopeptide maps demonstrated that the IgG domain + 5 DXRs contain spots b, c, and d, whereas spots a and bV appear to be derived from the C-terminal half of the long MLCK (i.e., the short MLCK) (Fig. 2C). These observations suggest that the phosphopeptide pattern observed for the full-length interphase kinase is comprised of multiple phosphopeptides derived from the C-terminal (a and bV) and N-terminal (b, c and d) halves of the kinase. Phosphopeptide maps of the IgG domain + 5 DXRs and short MLCK could not be obtained from mitotic cells because (i) expression, synchronization, and isotope labeling required approximately 44 h and these constructs induced extensive cell death when expressed for periods greater than 34 h [6] (data not shown), and (ii) the mitotic cells were sensitive to the 32P-label and the transfectants even more so, with 40– 50% of the cells dying after 2 h in the presence of the

307

Fig. 2. The long MLCK contains phosphorylation sites in the N-terminal and C-terminal halves of the kinase. (A) Schematic of the MLCK truncations. The dark grey boxes represent the Ig motifs, the black bars represent the actin binding DXR motifs, the light grey box represents GFP, and the hatched box represents GST. The rabbit short MLCK corresponds to residues 923–1914 of the HeLa long MLCK and displays 81% identity with the HeLa short MLCK. The rabbit 5 DXRs correspond to residues 862–994 of the HeLa long MLCK and display 93% identity with this region of the HeLa long MLCK. (B) Phosphoamino acid analysis. The IgG domain + 5 DXRs and the short MLCK are both phosphorylated on serine residues in interphase. Circles indicate the positions of phosphoamino acid standards. (C) Two-dimensional phosphopeptide maps of the HeLa IgG domain + 5 DXRs, the rabbit short MLCK, and the HeLa long MLCK immunoprecipitated from interphase HeLa cells. Phosphopeptides were resolved by electrophoresis at pH 1.9 followed by ascending chromatography in regular chromatography buffer.

radioisotope. The sensitivity of mitotic cells to the 32P-label likely results from a p53-dependent cell cycle arrest as several studies have demonstrated that the doses of radioisotopes typically used for metabolic labeling induce p53 expression and a growth arrest [20,21]. Nonetheless, a comparison of the phosphopeptide maps of the full-length long MLCK and the IgG domain + 5 DXRs suggests that the majority of the interphase and mitotic phosphorylation sites are located in the IgG domain. Aurora B and the long MLCK colocalize in early mitosis Expression of the IgG domain + 5 DXRs of the long MLCK induces extension of the astral microtubules,

308

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

bundling of the kinetochore microtubules, and spindle pole fragmentation (Fig. 3A). Moreover, these microtubule defects are associated with abnormalities in

metaphase chromosome alignment and a subsequent metaphase arrest [6]. Interestingly, ablation of Aurora B function by RNAi [8], inhibitory antibodies [9], or

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

expression of a dominant-negative, kinase-dead Aurora B [10] produces similar defects in the microtubule spindle and in chromosome alignment as observed in cells expressing the IgG domain or the IgG domain + 5 DXRs. Given that the long MLCK is phosphorylated during metaphase with the majority of the phosphorylation sites located within the IgG domain, inhibition of Aurora B by overexpression of the IgG domain could be responsible for the observed phenotypic effects in IgG domain-expressing cells. Therefore, we examined the potential role of Aurora B in regulating MLCK function during mitosis. Fig. 3B shows the subcellular localization of Aurora B in mitotic HeLa cells. Consistent with previous reports, our immunofluorescence studies indicate that Aurora B associates primarily with mitotic structures [19,22–24]. In prophase, the kinase associated with regions of condensed chromatin (Fig. 3B, a). From prometaphase through metaphase Aurora B localized to the centromeres (Fig. 3B, b–c), then redistributed to the spindle midzone in anaphase (Fig. 3B, d), and concentrated in the cleavage furrow of telophase cells and at the postmitotic bridge between spreading daughter cells (Fig. 3B, e and f). We also detected cortical localization of Aurora B at the cleavage furrow during anaphase and telophase (Fig. 3B, e, arrows), confirming the localization of the kinase at the equatorial cortex that has been reported in other cell types [25]. In addition, we observed that during prophase and prometaphase, Aurora B had a punctuate appearance in the cell cortex (Fig. 3B, a–c), which persisted until anaphase when Aurora B redistributed to the cell equator (Fig. 3B, d). As reported by us previously, the long MLCK localized primarily to stress fibers in interphase cells (Fig. 3C) [5,6]. In mitotic cells, the long MLCK was enriched in the cell cortex just beneath the plasma membrane during prometaphase and metaphase, and concentrated in the cleavage furrow throughout anaphase and telophase. Similarly, a GFP fusion of the IgG domain + 5 DXRs, which displays a localization that is identical to the full-length MLCK (Fig. 3C) [5,6], was

309

enriched in the cell cortex beginning in prometaphase where it partially colocalized with Aurora B (Fig. 3D). Aurora B phosphorylates the long MLCK in vitro In a co-immunoprecipitation assay, Aurora B was detected in long MLCK immunoprecipitates (Fig. 4A, lane 2); however, we did not detect MLCK in Aurora B immunoprecipitates (data not shown). This was anticipated as the Aurora B antibody recognizes the N-terminal 124 residues of Aurora B, which is the region of the kinase involved in binding substrates [26]. Thus, it is likely that the Aurora B antibody prevents the interaction of Aurora B with MLCK due to overlapping binding sites. Neither kinase was detected in immunoprecipitates using nonspecific IgG (Fig. 4A, lane 1). Consistent with the cell cycle-regulated expression of Aurora B, the kinase was not detected in lysates of interphase HeLa cells (Fig. 4A, lane 4) [13,19]. To further examine the association of Aurora B with the Nterminal half of the long MLCK, we performed a GST pulldown assay. Expression and purification of a GST fusion of the IgG domain from HeLa cells are shown in Fig. 4B (left panel, lane 2). Immunoblot analysis with antibodies to GST and the IgG domain demonstrated that the full-length GST fusion of the HeLa long MLCK IgG domain is expressed in HeLa cells (Fig. 4B, middle and right panels). GlutathioneSepharose containing the GST-IgG domain brought down endogenous Aurora B in a mitotic HeLa extract (Fig. 4C, lane 1), which did not interact with GST alone (Fig. 4C, lane 3). These observations suggest that Aurora-B binds to the Nterminal region of the long MLCK. In the converse experiment, glutathione-Sepharose containing bound GST-Aurora B or GST alone was incubated with a mitotic cell extract prepared from HeLa cells transiently expressing the IgG domain + 5 DXRs. We used this construct as our previous studies demonstrate that the IgG domain + 5 DXRs exhibits a localization that is identical to the full-length MLCK [6]. Glutathione-Sepharose containing GST-Aurora B specifically interacted with the IgG domain + 5 DXRs (Fig. 4C, lane 5), but not GST alone (Fig. 4C, lane 4).

Fig. 3. Microtubule defects observed in cells expressing the IgG domain + 5 DXRs and distribution of endogenous Aurora B and GFP fusions of the full-length long MLCK and IgG domain + 5 DXRs in mitotic HeLa cells. (A) Expression of the IgG domain + 5 DXRs of the long MLCK disrupts normal spindle morphology during mitosis; inducing extension of the astral microtubules, bundling of the kinetochore microtubules, and spindle pole fragmentation (lower panel). Moreover, these microtubule defects are associated with abnormalities in metaphase chromosome alignment and a subsequent metaphase arrest. The upper panel shows representative images of tubulin and DAPI staining in transfected cells. Metaphase cells were fixed and stained for microtubules with antibodies against h-tubulin (red) and for DNA (blue). In control cells expressing GFP, alignment of the chromosomes and well-organized bipolar spindles were observed (far right column). Scale bars, 10 Am. (B) Aurora B associates primarily with mitotic structures. Untransfected HeLa cells at different stages of mitosis were fixed and stained for endogenous Aurora B with AIM-1 antibodies (green) and for DNA with DAPI (red). From prophase through metaphase (a– c), the kinase associated with regions of condensed chromatin (a) and localized to centromeres (b and c). In addition, Aurora B displayed a punctate appearance in the cell cortex, which persisted until anaphase onset. In anaphase, Aurora B redistributed to the spindle midzone (d) and concentrated in the furrow of dividing cells and in the postmitotic bridge during late telophase (e and f). Aurora B also displayed a cortical localization in dividing cells (e, arrow). (C) Distribution of full-length long MLCK in interphase and mitotic cells. Fluorescence micrographs of HeLa cells transfected with the HeLa full-length long MLCK fused to GFP. In interphase, the long MLCK primarily localizes to stress fibers. From prophase to metaphase, the long MLCK is enriched in the cell cortex and concentrates in the cleavage furrow throughout anaphase and telophase. (D) Distribution of endogenous Aurora B (red) and a GFP fusion of the IgG domain + 5 DXRs (green) in HeLa cells during prometaphase. Aurora B partially colocalizes with the IgG domain + 5 DXRs at the cell cortex in early mitosis.

310

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

Aurora B efficiently phosphorylated the recombinant IgG domain and histone H3 with stoichiometries of approximately 1.1 and 1.75 mol phosphate per mol of substrate, respectively. More importantly, this assay demonstrates that Aurora B can bind and phosphorylate MLCK in the absence of other mammalian proteins. In addition, Aurora B did not phosphorylate the short MLCK (data not shown). Phosphoamino acid analysis revealed that Aurora B phosphorylated only serine residues in the IgG domain (Fig. 5B). This result is consistent with our in vivo data showing that the endogenous long MLCK and the IgG domain + 5 DXRs are phosphorylated only on serine (Figs. 1A and 2A). To further validate Aurora B as a kinase that phosphorylates the N-terminus of the long MLCK, we examined phosphopeptide maps of the endogenous fulllength MLCK from mitotic cells and the recombinant IgG domain phosphorylated in vitro by Aurora B (Fig. 5C). This analysis demonstrated that the two phosphopeptide maps are nearly identical, suggesting that Aurora B might be one of the kinases responsible for the phosphorylation of MLCK during mitosis. In addition, based on its closer migration to the positive electrode, we found that spot c, derived from the endogenous full-length MLCK, may be more highly phosphorylated than phosphopeptide cV derived from the in vitro phosphorylated IgG domain. This pattern could result from incomplete phosphorylation of the IgG domain by Aurora B or may indicate that additional kinases are involved in mitotic phosphorylation of the long MLCK in vivo. Fig. 4. Aurora B binds to the IgG domain of the HeLa long MLCK. (A) Long MLCK immunoprecipitates from a mitotic HeLa cell lysate were immunoblotted with an Aurora B monoclonal antibody (left panel) or a MLCK monoclonal antibody (right panel). Lane 1: control immunoprecipitate with nonspecific rabbit IgG; lane 2: MLCK immunoprecipitate; lane 3: whole cell lysate of mitotic HeLa cells; lane 4: whole cell lysate of interphase HeLa cells. (B) Mammalian expression and purification of a GST fusion of the IgG domain. Lane 1: whole cell lysate of HeLa cells expressing the GST-IgG domain; lane 2: the purified GST-IgG domain (molecular weight approximately 120 kDa). Left panel: Coomassie stain; middle panel: immunoblot with an IgG domain antibody and right panel: immunoblot with a GST antibody. (C) The IgG domain and Aurora B interact in a GST pull-down assay. Lanes 1–3: immunoblot with an Aurora B antibody. Lane 1: purified GST-IgG domain; lane 2: whole cell lysate of mitotic HeLa cells; lane 3: purified GST. Lanes 4–6: immunoblot with a MLCK antibody. Lane 4: purified GST; lane 5: purified GST-Aurora B; lane 6: whole cell lysate of mitotic HeLa cells transiently transfected with a GFP fusion of the IgG domain + 5 DXRs (molecular weight approximately 138 kDa). The topmost band in lane 6 is the endogenous long MLCK (210 kDa). The lower three bands correspond to proteolytic products of the IgG domain + 5 DXRs. The absence of degradation products in lane 5 suggests that the entire IgG domain may be required for interaction with Aurora B.

To determine if the IgG domain of the long MLCK can serve as a substrate for Aurora B, we performed an in vitro kinase assay (Fig. 5A). Aurora B is well documented to phosphorylate histone H3 on two sites, Ser10 and Ser28 [17]. Our in vitro phosphorylation studies demonstrate that

Discussion Functional significance of MLCK phosphorylation Our in vivo studies demonstrate that the long MLCK is highly phosphorylated on serine residues during metaphase. On two-dimensional phosphopeptide maps, the majority of the interphase and mitotic MLCK phosphopeptides consistently demonstrated low mobility in the second dimension, suggesting that the phosphopeptides may be enriched in acidic residues. A comparison of the phosphopeptide maps of the endogenous full-length MLCK, the IgG domain + 5 DXRs, and short MLCK suggests that the phosphopeptide pattern observed for the endogenous interphase kinase consists of spots derived from the C-terminal half of the kinase (a and bV) and the N-terminal half of the kinase (b, c, and d). Although the C-terminal three DXR motifs present in the IgG domain + 5 DXRs construct are also found in the short MLCK construct, we believe it is highly unlikely that spots b and bV are derived from this overlapping region because (i) sequence analysis indicates that tryptic peptides derived from this region are not enriched in serine residues; (ii) radioactive phosphate incorporation was not observed in the 5 DXRs alone, and (iii) phosphate incorporation was detected in the IgG domain alone (unpublished observations).

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

311

Fig. 5. Aurora B phosphorylates the IgG domain in vitro. (A) Time course of phosphorylation of histone H3 (circles) and the IgG domain (diamonds) by Aurora B. Phosphorylated histone H3 and IgG domain were subjected to SDS-PAGE and visualized by autoradiography (insets). (B) Phosphoamino acid analysis demonstrates that Aurora B phosphorylates the IgG domain on serine residues. Circles indicate the positions of the phosphoamino acid standards. (C) Twodimensional phosphopeptide maps of the IgG domain phosphorylated in vitro by Aurora B and the endogenous mitotic MLCK were resolved by electrophoresis at pH 1.9 followed by ascending chromatography in regular chromatography buffer.

This is the first in vivo study directly demonstrating phosphorylation of the N-terminal extension that is unique to the long MLCK. Previous studies have suggested that Tyr464 and Tyr471 may be sites of p60src-mediated phosphorylation [7]; however, we have found no evidence for tyrosine phosphorylation in our in vivo labeling studies with HeLa cells. The identification of the phosphoserines in the long MLCK will be a considerable challenge as the full-length MLCK and IgG domain contain 167 and 84 serine residues, respectively. Sequence analysis of the long MLCK indicates that potential PKA, PKC, CaM kinase II, cdc2 kinase, and casein kinase II phosphorylation sites are located within the IgG domain. It is unlikely that cdc2 kinase mediates phosphorylation of the long MLCK during metaphase because the MPM-2 antibody, which recognizes the phosphoepitope p(S/T)P that overlaps with the cdc2 kinase consensus phosphorylation site [27], does not react with MLCK immunoprecipitates from nocodazole-synchronized HeLa cells (unpublished observations). Our peptide maps indicate that during interphase, the long MLCK is phosphorylated in the IgG domain as well as in the C-terminal half of the kinase (i.e., the segment corresponding to the short MLCK). The observation that multiple domains in the long MLCK are phosphorylated suggests that these phosphorylation events could have multiple effects on MLCK function. We predict that the

phosphorylation sites within the IgG domain observed during interphase will not directly affect the catalytic activity of the kinase, but rather regulate kinase targeting by mediating interactions between MLCK and any proteins that bind to the IgG domain. This proposal is based on our findings that the IgG domain is required for correct localization of the kinase as well as biochemical studies demonstrating that the interphase MLCK has a two-fold higher specific activity than the mitotic MLCK [5,6]. In contrast, we anticipate that phosphorylation of serine residues in the C-terminal half of the kinase will likely downregulate the catalytic activity of MLCK. This is supported by the observation that multiple kinases are capable of phosphorylating Ser1760 [28–30], which is adjacent to the calmodulin binding domain. Phosphorylation on this site increases the K calmodulin by 10-fold, thus desensitizing MLCK to activation by Ca2+/calmodulin. Interestingly, our peptide maps of the endogenous mitotic long MLCK and in vitro phosphorylated IgG domain indicate that during metaphase the long MLCK is phosphorylated primarily within the IgG domain, but at unique sites in addition to those observed during interphase. Moreover, our previous biochemical studies showed a two-fold decrease in the specific activity of the long MLCK during metaphase, which can be restored following treatment with alkaline phosphatase [5]. These observations

312

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

suggest that mitotic phosphorylation of the IgG domain downregulates the catalytic activity of the kinase. This proposal is not without precedent as tyrosine phosphorylation of the IgG domain has been shown to increase the catalytic activity of the long MLCK approximately two-fold [7]. Altogether, our phosphopeptide maps suggest that there are multiple phosphorylation sites within the IgG domain; mitotic sites that effect catalytic activity and interphase sites whose effect on MLCK function are unknown. Phosphorylation of MLCK by Aurora B kinase Because cells expressing the IgG domain of the long MLCK display similar defects in the microtubule spindle, chromosome alignment, and cell division [6] as Aurora Bdepleted cells [8], and in cells in which Aurora B function has been ablated [9,10], we hypothesize that overexpression of the IgG domain, which contains the majority of the mitotic phosphorylation sites present in MLCK, inhibits Aurora B. This proposal is supported by our localization studies, which show that MLCK and Aurora B partially colocalize in the cell cortex in early mitosis, and by biochemical data demonstrating that Aurora B may participate in the mitotic phosphorylation of the IgG domain. Coimmunoprecipitation and GST pull-down assays suggest that Aurora B can physically interact with the N-terminus of the long MLCK. In addition, our in vitro phosphorylation studies demonstrate that Aurora B efficiently phosphorylates the IgG domain. Most importantly, phosphopeptide maps of the endogenous full-length MLCK from mitotic cells and the recombinant IgG domain phosphorylated in vitro by Aurora B demonstrate that Aurora B phosphorylates the same sites as those observed in vivo. Altogether, these studies indicate that the long MLCK may be a cellular target for Aurora B. The consensus site for phosphorylation by the S. cerevisiae Aurora B, Ipl1p, has been identified as [R/K] [X] [T/S] [I/L/V], where X is any residue [31]. More recent studies have suggested that the human Aurora B is an arginine-directed kinase. Substitution of arginine with lysine

in the sequence ARKS, which encompasses the Ser10 and Ser28 histone H3 phosphorylation sites [17], significantly decreases the phosphorylation efficiency of Aurora B in vitro [32], suggesting a strong preference for arginine and not just any basic residue in the 2 position. In addition, several mammalian Aurora B substrates, including CENPA, vimentin, desmin, GFAP, and survivin, all have phosphorylation sites comprised of the sequence [R] [X] [T/S] [33–36], in which an arginine occupies the 2 position. Sequence analysis of the full-length long MLCK indicates 12 potential consensus sites ([R] [X] [S]) for phosphorylation by Aurora B; eight within the IgG domain (Ser145, 154, 181, 208, 320, 411, 482, and 609) and four in the C-terminal half of the kinase (Ser1346, 1759, 1772, and 1852). However, in our in vitro Aurora B kinase assay, no phosphate incorporation was detected when the short MLCK was used as a substrate. In addition to the eight potential Aurora B phosphorylation sites, the IgG domain contains another 76 serines, which are potential phosphorylation sites for other kinases. Although phosphopeptide maps of the endogenous full-length MLCK from mitotic cells and the Aurora B phosphorylated IgG domain are nearly identical, one spot derived from the endogenous fulllength MLCK (c) migrates closer to the positive electrode than the corresponding spot derived from the in vitro phosphorylated IgG domain (cV). This disparity may result from incomplete phosphorylation of the IgG domain by Aurora B or phosphorylation of the in vivo derived MLCK by additional kinases during mitosis. Model for MLCK regulation by phosphorylation Based on our observations that the long MLCK is differentially phosphorylated during interphase and mitosis, we propose the following model for the regulation of MLCK function by phosphorylation (Fig. 6). During interphase, phosphorylation of the IgG domain modulates kinase targeting while phosphorylation near the calmodulinbinding domain downregulates MLCK’s kinase activity.

Fig. 6. Model for the regulation of MLCK function by phosphorylation. A schematic of MLCK showing the phosphorylation sites in the IgG domain and in the C-terminal half of the kinase. The placement of the phosphate groups is meant to indicate their presence in the IgG and C-terminal domains and not the actual sites of phosphorylation. See Discussion for more details.

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314

These phosphorylation events likely occur on a subset of MLCK molecules, thus allowing for local regulation of myosin-II function in interphase cells. When cells enter mitosis, MLCK is dephosphorylated on sites adjacent to the calmodulin binding domain and is phosphorylated on additional sites within the IgG domain, allowing for global inhibition of MLCK function and a net decrease in phosphorylation on Ser-19 of the myosin-II regulatory light chain. Before the onset of cytokinesis, the inhibitory phosphorylation site on the IgG domain is dephosphorylated and MLCK is activated upon binding of Ca2+/calmodulin, thus allowing for increased levels of Ser-19 phosphorylation and activation of myosin-II’s motor activity for cytokinesis. Future studies aimed at identifying the phosphorylated serines on the long MLCK, the functional consequences of these phosphorylations, and the additional kinase or kinase(s) that direct these phosphorylation events will be required to establish the molecular basis for the regulation of MLCK function by phosphorylation.

Acknowledgments We thank Dr. Jim Stull for the rabbit short MLCK-GFP construct and purified rabbit short MLCK, Dr. William Earnshaw for the GST-Aurora B construct, and Dr. Louis Lim for the pXJ-GST vector. We are grateful to Dr. Jon Backer for the use of his HTLE 7000 thin-layer electrophoresis unit and Dr. George Orr for helpful discussions. This work was supported by the National Institutes of Health.

References [1] K.E. Kamm, J.T. Stull, Dedicated myosin light chain kinases with diverse cellular functions, J. Biol. Chem. 276 (2001) 4527 – 4530. [2] E.K. Blue, Z.M. Goeckeler, Y. Jin, L. Hou, S.A. Dixon, B.P. Herring, R.B. Wysolmerski, P.J. Gallagher, 220- and 130-kDa MLCKs have distinct tissue distributions and intracellular localization patterns, Am. J. Physiol.: Cell Physiol. 282 (2002) C451 – C460. [3] D.M. Watterson, M. Collinge, T.J. Lukas, L.J. Van Eldik, K.G. Birukov, O.V. Stepanova, V.P. Shirinsky, Multiple gene products are produced from a novel protein kinase transcription region, FEBS Lett. 373 (1995) 217 – 220. [4] J.G. Garcia, V. Lazar, L.I. Gilbert-McClain, P.J. Gallagher, A.D. Verin, Myosin light chain kinase in endothelium: molecular cloning and regulation, Am. J. Respir. Cell Mol. Biol. 16 (1997) 489 – 494. [5] A. Poperechnaya, O. Varlamova, P.J. Lin, J.T. Stull, A.R. Bresnick, Localization and activity of myosin light chain kinase isoforms during the cell cycle, J. Cell Biol. 151 (2000) 697 – 708. [6] N.G. Dulyaninova, Y.V. Patskovsky, A.R. Bresnick, The N-terminus of the long MLCK induces a disruption in normal spindle morphology and a metaphase arrest, J. Cell Sci. 117 (2004) 1481 – 1493. [7] K.G. Birukov, C. Csortos, L. Marzilli, S. Dudek, S.F. Ma, A.R. Bresnick, A.D. Verin, R.J. Cotter, J.G. Garcia, Differential regulation of alternatively spliced endothelial cell myosin light chain kinase isoforms by p60Src, J. Biol. Chem. 276 (2001) 8567 – 8573. [8] B. Delaval, A. Ferrand, N. Conte, C. Larroque, D. Hernandez-Verdun, C. Prigent, D. Birnbaum, Aurora B-TACC1 protein complex in cytokinesis, Oncogene 23 (2004) 4516 – 4522.

313

[9] M.J. Kallio, M.L. McCleland, P.T. Stukenberg, G.J. Gorbsky, Inhibition of aurora B kinase blocks chromosome segregation, overrides the spindle checkpoint, and perturbs microtubule dynamics in mitosis, Curr. Biol. 12 (2002) 900 – 905. [10] M. Murata-Hori, Y.L. Wang, The kinase activity of aurora B is required for kinetochore-microtubule interactions during mitosis, Curr. Biol. 12 (2002) 894 – 899. [11] J.R. Bischoff, L. Anderson, Y. Zhu, K. Mossie, L. Ng, B. Souza, B. Schryver, P. Flanagan, F. Clairvoyant, C. Ginther, C.S. Chan, M. Novotny, D.J. Slamon, G.D. Plowman, A homologue of Drosophila aurora kinase is oncogenic and amplified in human colorectal cancers, EMBO J. 17 (1998) 3052 – 3065. [12] J.M. Schumacher, A. Golden, P.J. Donovan, AIR-2: an Aurora/ Ipl1-related protein kinase associated with chromosomes and midbody microtubules is required for polar body extrusion and cytokinesis in Caenorhabditis elegans embryos, J. Cell Biol. 143 (1998) 1635 – 1646. [13] Y. Terada, M. Tatsuka, F. Suzuki, Y. Yasuda, S. Fujita, M. Otsu, AIM1: a mammalian midbody-associated protein required for cytokinesis, EMBO J. 17 (1998) 667 – 676. [14] M. Murata-Hori, M. Tatsuka, Y.L. Wang, Probing the dynamics and functions of aurora B kinase in living cells during mitosis and cytokinesis, Mol. Biol. Cell 13 (2002) 1099 – 1108. [15] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680 – 685. [16] C. Morrison, A.J. Henzing, O.N. Jensen, N. Osheroff, H. Dodson, S.E. Kandels-Lewis, R.R. Adams, W.C. Earnshaw, Proteomic analysis of human metaphase chromosomes reveals topoisomerase II alpha as an Aurora B substrate, Nucleic Acids Res. 30 (2002) 5318 – 5327. [17] H. Goto, Y. Yasui, E.A. Nigg, M. Inagaki, Aurora-B phosphorylates Histone H3 at serine28 with regard to the mitotic chromosome condensation, Genes Cells 7 (2002) 11 – 17. [18] E. Manser, H.Y. Huang, T.H. Loo, X.Q. Chen, J.M. Dong, T. Leung, L. Lim, Expression of constitutively active alpha-PAK reveals effects of the kinase on actin and focal complexes, Mol. Cell. Biol. 17 (1997) 1129 – 1143. [19] C. Crosio, G.M. Fimia, R. Loury, M. Kimura, Y. Okano, H. Zhou, S. Sen, C.D. Allis, P. Sassone-Corsi, Mitotic phosphorylation of histone H3: spatio-temporal regulation by mammalian Aurora kinases, Mol. Cell. Biol. 22 (2002) 874 – 885. [20] R. Dover, Y. Jayaram, K. Patel, R. Chinery, p53 expression in cultured cells following radioisotope labelling, J. Cell Sci. 107 (1994) 1181 – 1184. [21] J.A. Bond, K. Webley, F.S. Wyllie, C.J. Jones, A. Craig, T. Hupp, D. Wynford-Thomas, p53-Dependent growth arrest and altered p53-immunoreactivity following metabolic labelling with 32P ortho-phosphate in human fibroblasts, Oncogene 18 (1999) 3788 – 3792. [22] R. Giet, D.M. Glover, Drosophila aurora B kinase is required for histone H3 phosphorylation and condensin recruitment during chromosome condensation and to organize the central spindle during cytokinesis, J. Cell Biol. 152 (2001) 669 – 682. [23] Y. Minoshima, T. Kawashima, K. Hirose, Y. Tonozuka, A. Kawajiri, Y.C. Bao, X. Deng, M. Tatsuka, S. Narumiya, W.S.J. May, T. Nosaka, K. Semba, T. Inoue, T. Satoh, M. Nagaki, T. Kitamura, Phosphorylation by aurora B converts MgcRacGAP to a RhoGAP during cytokinesis, Dev. Cell 4 (2003) 549 – 560. [24] Y. Yasui, T. Urano, A. Kawajir, K.I. Nagata, M. Tatsuka, H. Saya, K. Furukawa, T. Takahashi, I. Izawa, M. Inagaki, Autophosphorylation of a newly identified site of Aurora-B is indispensable for cytokinesis, J. Biol. Chem. 279 (2004) 12997 – 13003. [25] M. Murata-Hori, Y.L. Wang, Both midzone and astral microtubules are involved in the delivery of cytokinesis signals: insights from the mobility of aurora B, J. Cell Biol. 159 (2002) 45 – 53. [26] M.A. Bolton, W. Lan, S.E. Powers, M.L. McCleland, J. Kuang, P.T. Stukenberg, Aurora B kinase exists in a complex with survivin and

314

[27]

[28]

[29]

[30]

[31]

N.G. Dulyaninova, A.R. Bresnick / Experimental Cell Research 299 (2004) 303–314 INCENP and its kinase activity is stimulated by survivin binding and phosphorylation, Mol. Biol. Cell 13 (2002) 3064 – 3077. J.M. Westendorf, P.N. Rao, L. Gerace, Cloning of cDNAs for Mphase phosphoproteins recognized by the MPM2 monoclonal antibody and determination of the phosphorylated epitope, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 714 – 718. M.A. Conti, R.S. Adelstein, The relationship between calmodulin binding and phosphorylation of smooth muscle myosin kinase by the catalytic subunit of 3V:5V cAMP-dependent protein kinase, J. Biol. Chem. 256 (1981) 3178 – 3181. M. Ikebe, S. Reardon, Phosphorylation of smooth myosin light chain kinase by smooth muscle Ca2+/calmodulin-dependent multifunctional protein kinase, J. Biol. Chem. 265 (1990) 8975 – 8978. Z.M. Goeckeler, R.A. Masaracchia, Q. Zeng, T.L. Chew, P. Gallagher, R.B. Wysolmersk, Phosphorylation of myosin light chain kinase by p21-activated kinase PAK2, J. Biol. Chem. 275 (2000) 18366 – 18374. I.M. Cheeseman, S. Anderson, M. Jwa, E.M. Green, J. Kang, J.R.r. Yates, C.S. Chan, D.G. Drubin, G. Barnes, Phospho-regulation of kinetochore-microtubule attachments by the Aurora kinase Ipl1p, Cell 111 (2002) 163 – 172.

[32] K. Sugiyama, K. Sugiura, T. Hara, K. Sugimoto, H. Shima, K. Honda, K. Furukawa, S. Yamashita, T. Urano, Aurora-B associated protein phosphatases as negative regulators of kinase activation, Oncogene 21 (2002) 3103 – 3111. [33] S.G. Zeitlin, R.D. Shelby, K.F. Sullivan, CENP-A is phosphorylated by Aurora B kinase and plays an unexpected role in completion of cytokinesis, J. Cell Biol. 155 (2001) 1147 – 1157. [34] H. Goto, Y. Yasui, A. Kawajiri, E.A. Nigg, Y. Terada, M. Tatsuka, K. Nagata, M. Inagaki, Aurora-B regulates the cleavage furrow-specific vimentin phosphorylation in the cytokinetic process, J. Biol. Chem. 278 (2003) 8526 – 8530. [35] A. Kawajiri, Y. Yasui, H. Goto, M. Tatsuka, M. Takahashi, K. Nagata, M. Inagaki, Functional significance of the specific sites phosphorylated in desmin at cleavage furrow: Aurora-B may phosphorylate and regulate type III intermediate filaments during cytokinesis coordinatedly with Rho-kinase, Mol. Biol. Cell. 14 (2003) 1489 – 1500. [36] S.P. Wheatley, A.J. Henzing, H. Dodson, W. Khaled, W.C. Earnshaw, Aurora-B phosphorylation in vitro identifies a residue of Survivin that is essential for its localization and binding to INCENP in vivo, J. Biol. Chem. 279 (2004) 5655 – 5660.