Interaction and Feedback Regulation between STK15/BTAK/Aurora-A Kinase and Protein Phosphatase 1 through Mitotic Cell Division Cycle*

STK15 is an Aurora/Ipl-1 related serine/threonine kinase that is associated with centrosomes and induces aneuploidy when overexpressed in mammalian cells. It is well known that phosphorylation and dephosphorylation of kinases are important for regulation of their activity. But mechanisms by which STK15 activity is regulated have not been elucidated. We report that STK15 contains two functional binding sites for protein phosphatase type 1 (PP1), and the binding of these proteins is cell cycle-regulated peaking at mitosis. Activated STK15 at mitosis phosphorylates PP1 and inhibits PP1 activity in vitro . In vivo , PP1 activity co-immuno-precipitated with STK15 is also reduced. These data indicate that STK15 inhibits PP1 activity during mitosis. Also, PP1 is shown to dephosphorylate active STK15 and abolish its activity in vitro . Furthermore, we show that non-binding mutants of STK15 for PP1 are superphos-phorylated, but their kinase activities are markedly reduced. Cells transfected with these non-binding mutants manifest aberrant chromosome alignment during mitosis. Our results suggest that a feedback regulation through phosphorylation/dephosphorylation events between STK15 kinase and PP1 phosphatase operates

Mitosis is morphologically the most dynamic phase in the cell cycle, and a large number of events including chromosome and microtubule dynamics are precisely coordinated at temporal and spatial levels. Reversible phosphorylation of proteins by protein kinases and protein phosphatases regulates these events, and alterations in the phosphorylated state of proteins have effects on mitotic progression (1,2). Whereas protein kinases have been studied well, relatively less is known about protein phosphatases that act antagonistically to kinases. In many cases except for the regulation of cdk1, it remains to be shown how phosphorylation regulates target proteins including regulatory proteins themselves.
STK15 (also known as BTAK, Aurora-A) is a human serine/ threonine kinase that belongs to the Drosophila aurora and Saccharomyces cerevisiae Ipl1 kinase family, and both are essential for chromosome segregation and centrosome functions (3)(4)(5)(6). The STK15 gene is amplified, and its transcript is also highly expressed in various cancers. Overexpression of STK15 induces increased numbers of centrosome, aneuploidy, and transformation of the cells (3,4). On the other hand, both loss-of-function by the addition of antibody against pEg2, the Xenopus orthologue of STK15, and overexpression of a dominant-negative kinase-inactive pEg2 cause defects on centrosome separation and mitotic spindle stabilization (7,8). These data suggest that STK15 plays critical roles in regulation of centrosome functions. Furthermore it has been reported that CDC20 and kinesin-related proteins associate with STK15, although their physiological functions have not yet been elucidated (9,10). Recently, it has been shown that in S. cerevisiae and Caenorhabditis elegans ipl1/aurora kinase phosphorylates histone H3 and a kinetochore protein Ndc10, suggesting that these interactions regulate chromosome condensation and kinetochore-microtubule interactions, respectively, acting in opposition to type 1 protein phosphatase Glc7/PP1 1 (11)(12)(13).
In mammals, there are four isoforms of protein phosphatase type 1 (PP1␣, PP1␥1, PP1␥2, and PP1␦) (14). The activity and subcellular compartmentalization of these PP1 isoforms are controlled by their association with a variety of regulatory subunits (15,16). It is known that these regulatory subunits have a common motif -(R/K)(V/I)XF-to associate with PP1, and mutation in this motif disrupts the association (17,18). Requirement of PP1 in mitosis has been shown from genetic studies in S. cerevisiae, Schizosaccharomyces pombe, Aspergillus nidulans, and Drosophila, in which mutants show defects in spindle organization, chromosome condensation, and segregation (19 -24). Inhibition of PP1 activity by drug treatment or injection of antisense oligonucleotides indicates that PP1 plays a regulatory role in microtubule dynamics during transitions into and out of mitosis and cytokinesis (25,26). Furthermore, PP1 is also prerequired for nuclear lamina assembly via the protein kinase A anchoring protein AKAP149 (27). On the other hand, PP1 activity is inactivated by Cdc2 phosphorylation in early to mid-mitosis (28 -30). The genetic and biochemical findings on PP1-interacting proteins described above can partly explain the complex mutant phenotypes but not all.
In this paper we present evidence that STK15 interacts with PP1 via functional binding domains within STK15 in vivo and in vitro, and this binding is higher at mitosis compared with interphase. These data indicate that interaction between these proteins is cell cycle-regulated. In addition, mitotic active STK15 phosphorylates PP1, and this phosphorylation reduces PP1 activity. On the other hand, PP1 dephosphorylates hyperphosphorylated active STK15 that results in loss of STK15 activity. Furthermore, we indicate that this interaction is necessary for activation of STK15 and that transient transfection of PP1 non-binding mutants of STK15 induces aberrant chromosome alignments at mitotic metaphase at higher rate compared with those of wild type STK15 and kinase-inactive STK15 mutant. Our results suggest that a feedback regulation through phosphorylation/dephosphorylation cycles exists between STK15 kinase and PP1 phosphatase and that overexpression of STK15 leads to deregulation of this phosphorylation/dephosphorylation balance causing aberration of this normal feedback regulation in cancer cells. The findings further suggest that aberrant STK15 expression affects an unknown STK15-PP1-mediated regulatory pathway in mitosis to induce mitotic anomalies in cancer cells.

EXPERIMENTAL PROCEDURES
Cell Culture, Synchronization, and Fluorescence-activated Cell Sorter Analysis-HeLa cells were grown in Dulbecco's modified Eagle's medium with 10% heat-inactivated fetal bovine serum. For synchronization at the G 1 /S boundary, HeLa cells were synchronized by a double thymidine block and release (31). Exponentially growing cells were blocked for 16 h with 2.5 mM thymidine, released from the block by three washes in phosphate-buffered saline, and incubated with fresh medium. After 8 h, cells were re-incubated for 16 h with 2.5 mM thymidine. For synchronization at the M phase, cells were treated with 25 ng/ml nocodazole for 16 h and then mitotic cells were collected by shake-off. At each time point cells were fixed with 70% ethanol at Ϫ20°C, treated with RNase, stained with propidium iodide, and then analyzed with EPICS XL-MCL (Coulter Corp.) in the MultiCycle program (Phoenix Flow Systems).
Expression and Purification of GST-PP1 and GST-STK15-GST-PP1 isoforms and GST-STK15 proteins were produced in BL21 pLys bacteria according to the manufacturer's protocol (Amersham Pharmacia Biotech). The protein bound to the glutathione-Sepharose beads was used for pull-down assay and activity assay. For in vitro kinase assay, the proteins were eluted with 20 mM glutathione solution in 50 mM Tris-HCl, pH 8.0.
Immunoprecipitation, Immunoblotting, and in Vitro Kinase Assay-250 -500 g of cell lysate was incubated for 3 h at 4°C with 6 l of anti-STK15 antibody-protein G-agarose or 10 g of anti-FLAG M2 monoclonal antibody-protein G-agarose (Sigma), and the immunocomplex was washed 4 times with lysis buffer and then subjected to SDS-PAGE. After transfer to nitrocellulose membrane, endogenous PP1 and exogenously expressed Myc-PP1 constructs were probed with anti-PP1 monoclonal antibody (Santa Cruz Biotechnology) and c-Myc monoclonal antibody (CLONTECH), respectively. For in vitro kinase assay, immunocomplex was washed 4 times with IP washing buffer (80 mM ␤-glycerophosphate, 15 mM MgCl 2 , 20 mM EGTA, 500 mM NaCl, 0.5% Nonidet P-40) and then washed 3 times with kinase buffer (50 mM Tris-HCl, pH 7.5, 15 mM MgCl 2 , 2 mM EGTA, 0.5 mM Na 3 VO 4 , 1 mM dithiothreitol). In vitro kinase assays to dephosphorylated ␤-casein, GST-PP1␣, -␥1, and -␦ were performed for 30 min at 30°C in kinase buffer containing 5 mM cold ATP and 5 Ci of [␥-32 P]ATP. The reaction was terminated by addition of sample buffer, and the samples were subjected to SDS-PAGE, followed by drying of the gel and autoradiography.
In Vitro Binding Assays-For detecting the binding of STK15 to PP1, 5 g each of GST-PP1␣, GST-PP1␥1, and GST-PP1␦ bound to glutathione-Sepharose beads was mixed with 1 mg of cell lysate for 1 h at 4°C. The beads were washed 5 times with lysis buffer and then subjected to SDS-PAGE. The protein was transferred to nitrocellulose membrane and probed with anti-STK15 antibody. For detecting the binding of PP1 to GST-STK15 WT, GST-STK15 Lys 3 Arg, and NB, the same condition was used except that 2 mg of cell lysate was used and probed with anti-PP1 antibody.
PP1 and PP2 Treatment-Immunoprecipitated SKT15 from mitotic cell lysate was washed with IP washing buffer and then washed 4 times with phosphatase buffer (50 mM Tris-HCl, pH 7.5, 1 mM MnCl 2 , 0.1 mM EDTA, 5 mM dithiothreitol, 0.01% Brij 35 for PP1, and 50 mM Tris-HCl, pH 8.5, 20 mM MgCl 2 , 1 mM dithiothreitol for PP2A). The immunocomplex was incubated for 30 min at 30°C with 0.5 units of PP1 (New England Biolabs) or 0.5 units of PP2A (Promega) and then washed 5 times with kinase buffer, and subsequently an in vitro kinase assay to dephosphorylated ␤-casein was performed as described above. The samples were subjected to SDS-PAGE, transferred to nitrocellulose membrane, and probed with anti-STK15 antibody, followed by autoradiography.
Protein Phosphatase Assay-Phosphatase activity of PP1 associated with anti-STK15 immunocomplex and in vitro phosphorylated GST-PP1␣, -␥1, and -␦ by STK15 was measured by using 32 P-labeled myelin basic protein ( 32 P-MBP) as the substrate according to the manufacturer's protocol (New England Biolabs). After beads were washed 5 times with phosphatase buffer, beads were incubated for 10 -30 min at 30°C in phosphatase buffer containing 50 mM 32 P-MBP. The reactions were stopped by 20% trichloroacetic acid precipitation. After removing the precipitated protein and beads by centrifugation, the relative phosphatase activity was determined by the amount of soluble 32 P release.

STK15 Interacts with PP1 in Vivo and in Vitro-
To study the regulation of phosphorylation of STK15, we searched STK15 amino acid sequence and found two similar motifs to consensus motif for binding to PP1 (Fig. 1A). To investigate whether STK15 interacts with PP1 and if the two motifs identified are responsible for interaction with PP1, we constructed FLAGtagged mutants in which phenylalanine 165 and phenylalanine 346 in the putative consensus motifs were substituted to alanine, respectively. After transient transfection into HeLa cells, cells were cultured in the presence or absence of nocodazole. The expressed tagged proteins were immunoprecipitated with anti-FLAG antibody and then immunoblotted with anti-STK15 antibody and anti-PP1 antibody that recognizes all the PP1 isoforms. In nocodazole-treated cells arrested at mitosis, the interaction to PP1 of both F165A (NB1) and F346A (NB2) mutants was significantly decreased compared with wild type (WT) and kinase-inactive mutant (Lys 3 Arg) (Fig. 1B). In contrast, the differences in the degree of interactions between WT, Lys 3 Arg, and the non-binding mutants with PP1 were minimal in the exponentially growing cells, although WT indicated higher interaction to PP1 than mutants (Fig. 1B). To confirm further the role of the consensus binding motifs in the interaction, pull-down assays were performed using bacterially expressed GST fusion STK15 proteins. GST and the GST fusion proteins were incubated with mitotic extracts, and the PP1 bound to fusion proteins was immunoblotted with anti-PP1 antibody. Fig. 1C clearly shows that GST-NB1 and GST-NB2 failed to interact with PP1. These results show that STK15 interacts with PP1 through the functional binding domains, and STK15 kinase activity is not necessary for this interaction.
To assess precisely if this interaction is cell cycle-dependent, HeLa cells synchronized by a double thymidine block and release protocol were analyzed. Whereas protein expression, phosphorylation, and activity of STK15 peak at mitosis and decline during exit from mitosis through G 1 ( Fig. 2A), protein expression of PP1 in the cell cycle progression of HeLa cells was constant (Fig. 2B). Similar expression pattern of PP1 protein has been earlier reported in other cell types (32). Endogenous STK15 was immunoprecipitated with anti-STK15 antibody, and the immunocomplex was immunoblotted with anti-PP1 and anti-STK15 antibodies. Consistent with the data shown in Fig. 1B, PP1 binding to STK15 increased with cell cycle progression to mitosis ( Fig. 2A). Interestingly, this interaction seems to be correlated with the appearance of hyperphosphorylated STK15 that has kinase activity ( Fig. 2A). To investigate the role of STK15 phosphorylation in PP1 interaction, mitotic extracts were prepared in the presence or absence of phosphatase inhibitor (okadaic acid, NaF, and EDTA), and the amount of immunoprecipitated PP1 was compared. As shown in Fig.  2C, when STK15 was dephosphorylated in the absence of phosphatase inhibitor (indicated by arrow), the interaction was decreased compared with the phosphorylated form of STK15 detected in the presence of phosphatase inhibitor (indicated by arrowhead). This indicates that phosphorylated active STK15 interacts with PP1 more effectively than inactive STK15. Paradoxically, however, we detected similar quantitative binding between STK15 and PP1 at 8, 9, and 12 h post-thymidine release, despite the fact that STK15 was predominantly seen in active hyperphosphorylated form at 8 and 9 h post-release unlike the hypophosphorylated form detected at 12 h postrelease ( Fig. 2A). At 12 h post-release only 4% of cells are in G 2 /M phase of mitosis compared with 74 and 37% of cells in G 2 /M phase at 8 and 9 h post-release. The reason why this interaction is still high in vivo at 12 h post-thymidine release when almost 88% of the cells have progressed to G 1 phase after mitosis remains to be elucidated.
There are four isoforms of PP1 (PP1␣, PP1␥1, PP1␥2, and PP1␦) in mammals. To examine which isoform interacts to STK15, pull-down assay was performed using bacterially expressed GST fusion PP1␣, PP1␥1, and PP1␦. GST and each GST fusion protein were incubated with mitotic extracts and then STK15 bound to fusion proteins was immunoblotted with anti-STK15 antibody. Fig. 3 shows STK15 is able to interact to all PP1 isoforms. Similar result was also obtained using Myctagged constructs of all the PP1 isoform-transfected cells (data not shown). These results taken together indicate that STK15 interacts with PP1 both in vivo and in vitro, and this interaction is regulated in cell cycle-dependent manner.
Dephosphorylation and Inactivation of STK15 by PP1 Not PP2A-To investigate the biochemical significance of interaction between STK15 and PP1, we first examined whether PP1 dephosphorylates STK15 and, if so, the effect on kinase activity. Immunoprecipitated STK15 from mitotic extracts was incubated with purified PP1 and PP2A, respectively, and in vitro kinase assay was subsequently performed. Interphase extract and mitotic extract were used as positive control to confirm phosphorylation status of STK15. Whereas incubation with PP2A did not affect phosphorylation status and kinase activity of STK15 similar to control reaction, PP1 caused significant dephosphorylation and inactivation of STK15 (Fig. 4).
Phosphorylation and Inhibition of PP1 by STK15-We did in vitro kinase assay to test if active hyperphosphorylated STK15 from mitotic cells can phosphorylate PP1 using GST-PP1 proteins as substrates. To detect phosphorylation, assay was performed in the presence of sodium vanadate or okadaic acid to prevent GST-PP1 from dephosphorylating hyperphosphorylated STK15. Active STK15 from mitotic cells phosphorylates all the isoforms of GST-PP1 in vitro in the presence of phosphatase inhibitors, but we could never detect phosphorylation of PP1 in the absence of these inhibitors (Fig. 5A). These data indicate that STK15 may phosphorylate common serine and/or threonine residues in all the isoforms. Next, to examine the effect of STK15-mediated phosphorylation on PP1 activity, PP1 assay was performed using 32 P-labeled MBP as substrate. For this objective, we used interphase extract as negative control because interphase extracts have less STK15 kinase activity than mitotic extracts. STK15 immunoprecipitated from both extracts were reacted with GST-PP1␣ in the presence of okadaic acid, immediately followed by extensive washing to remove okadaic acid, and then GST-PP1␣ activity was assayed. As shown in Fig. 5B, activity of GST-PP1␣ reacted with mitotic STK15 was decreased about 25% in comparison with interphase STK15. Furthermore, to examine if this decrement was directly because of phosphorylation by STK15, we transfected WT and Lys 3 Arg constructs to HeLa cells, respectively. PP1 activity present in immunocomplex that was immunoprecipitated with anti-FLAG antibody was measured in the presence of 2 nM okadaic acid to inhibit PP2A activity. PP1 activity bound to WT was decreased about 25% compared with that of Lys 3 Arg mutant (Fig. 5C). We next measured PP1 activity bound to endogenous STK15 derived from interphase and mitotic extracts. PP1 activity bound to mitotic STK15 was decreased compared with that of interphase STK15 (Fig. 5D). Because it is known that Cdk1-phosphorylated PP1 at threonine 320 results in inhibition of PP1 activity, we examined whether STK15 phosphorylates the same site using mutant GST-PP1␣ substituted at position 320 from threonine to alanine. We detected similar phosphorylation signals from both normal and mutant fusion proteins in the presence of active mitotic STK15 (data not shown). These results suggest that STK15 phosphorylates PP1 at different site(s) from Cdk1, and its phosphorylation leads to inhibition of PP1 activity at mitosis.
Interaction with PP1 Is Essential for STK15 Kinase Activation-Based on the results from Fig. 4, we speculated that both NB mutants might be more phosphorylated compared with WT at mitosis and display higher kinase activity. To address the effect of reduced interaction for PP1 on the kinase activity and phosphorylation status of STK15, we performed immunoblotting and in vitro kinase assay of immunoprecipitated NB mutants from mitotic extract. As expected, both NB mutants were more phosphorylated compared with WT and Lys 3 Arg (Fig.  6, lanes 2 and 3). Surprisingly, both hyperphosphorylated (indicated by arrow) and super-phosphorylated (the slowest migrating top band indicated by arrowhead) NB mutants together had much less kinase activity than that detected for the WT form, which displayed strong kinase activity, as expected (Fig.  6). This result provides evidence that inhibitory phosphorylation sites exist in STK15 and that dephosphorylation of these sites by PP1 is required for activation of STK15.
Loss of Interaction to PP1 Induces Aberrant Chromosome Alignment during Mitosis-Next, to study whether overexpression of NB mutants affects mitotic progression, indirect immunofluorescence microscopy of HeLa cells transiently transfected with pEGFP-NB mutants was performed. Interestingly, NB mutant-transfected cells frequently demonstrated misalignment of chromosomes at metaphase plate compared with control empty vector and WT-transfected cells (Fig. 7 A, panels  B, F, J, and N). GFP-NB mutants showed similar centrosome localization as in WT-transfected cells (Fig. 7A, panels E, I, and M) with no apparent influence on phosphorylation of histone H3 (Fig. 7A, panel C, G, K, and O). The proportions of cells showing aberrantly aligned chromosomes in the NB mutanttransfected cells were, however, significantly increased compared with those of WT and Lys 3 Arg-transfected cells (Fig.  7B). These results indicate that interaction between STK15 and PP1 is involved in regulation of chromosome alignment/ segregation at metaphase. DISCUSSION STK15 and its related kinases have been implicated in regulating equal segregation of chromosomes during mitotic cell division cycle. The mechanisms underlying activation of STK15 through phosphorylation and regulation of centrosome as well as mitotic microtubule functions by activated STK15, however, remain to be elucidated. In this paper, we present the first evidence in human cells that (i) STK15 interacts with PP1 via consensus binding domains, and this interaction is necessary for STK15 activation; (ii) activated STK15 at mitosis phosphorylates PP1 which leads to reduction in PP1 phosphatase activity; (iii) dephosphorylation of STK15 by PP1 causes inhibition of its kinase activity; and (iv) interaction between STK15 and PP1 is involved in regulating chromosome alignment/segregation at metaphase.
We demonstrated that non-binding mutant STK15 is highly phosphorylated at mitosis, resulting in significant inhibition of its kinase activity. This result implies two points. First, single or multiple inhibitory phosphorylation sites as well as activation site(s) exist on STK15. Second, these sites phosphorylated by unknown inhibitory kinase(s) are dephosphorylated by PP1 at mitosis, leading to STK15 activation. This idea is supported by a report by Walter et al. (33) which shows that threonine 288 residue on STK15, only phosphorylated in interphase but not in mitosis, can be dephosphorylated by PP1 in vitro (33). Although Thr-288 appears to be at least one candidate responsible for preventing STK15 activation in interphase, we cannot exclude the possibility of the existence of other inhibitory phosphorylation sites as well. Further studies are needed to identify both activation and inhibitory phosphorylation sites and to elucidate their functional significance during cell cycle progression. It is known that PP1 is phosphorylated at threonine 320 by Cdk1 and is inactivated through G 2 /M transition during mitosis (28,30). It has also been reported that this inactivation is necessary for maintaining the microtubule dynamics through inactivation of unidentified proteins. In this study we showed that phosphorylation of PP1 by STK15 at still unidentified site(s) can inhibit PP1 activity. This implies that STK15 kinase is involved in negative regulation of PP1 in addition to Cdk1 kinase. Taken together, our observations suggest that a feedback regulation through phosphorylation/dephosphorylation cycles exists between STK15 kinase and PP1 phosphatase.
Genetic studies in yeast have shown that Ipl1 kinase acts in opposition to Glc-7/type 1 phosphatase. Recently, Ndc10 and histone H3 have been identified as the common target proteins for Ipl1 kinase and Glc-7 phosphatase (11)(12)(13). Ndc10 is a component of CBF3 kinetochore protein complex, and reversible phosphorylation of Ndc10 by Ipl1 and Glc-7 plays an important role in the attachment of kinetochores to mitotic spindles (11,12). We observed abnormal chromosome alignment at metaphase in non-binding mutant-transfected cells. This indi-cates the possibility that STK15 and PP1 regulatory feedback mechanism is involved in regulating kinetochore/microtubule attachment in human cells. Because we did not investigate the phosphorylation status of kinetochore proteins in this study, further studies are required to elucidate this possibility. On the other hand, reversible phosphorylation of histone H3 at serine 10 by Ipl1 and Glc-7 is considered to be crucial for chromosome condensation/decondensation (13). In C. elegans it was reported that AIR-2, an orthologue of human AIM-1/Aurora-B kinase that regulates cytokinesis (34), but not AIR-1, an orthologue of human STK15/Auorora-A kinase, is involved in the phosphorylation of histone H3 at serine 10 (13). Recently, it has shown (35) that Xenopus Aurora-B kinase regulates phosphorylation of histone H3 in concert with chromatin-associated PP1. These findings indicate that chromosome condensation through phosphorylation of histone H3 by Aurora-B kinase but not Aurora-A kinase is evolutionarily conserved. Immunofluorescence detection of histone H3 phosphorylation on misaligned chromosomes of NB mutant-transfected cells, as reported in this study, reinforces the idea that STK15/Aurora A is not the primary kinase involved in phosphorylating histone H3 during G 2 /M phase.
There are four PP1 isoforms in mammalian cells, and their subcellular localizations are different at mitosis. PP1␣ is localized to centrosome; PP1␥ is associated with mitotic spindles, and PP1␦ associates with chromosomes (16). Such different subcellular localization of various isoforms of PP1 might explain why mutants in PP1 and inhibition of PP1 by chemical drugs show complex phenotypes with condensed chromosomes, formation of abnormal spindles, chromosome separation malfunction, and defect of cytokinesis (19 -24). STK15 binds with all the isoforms examined in vitro. Our previous study has shown that STK15 is localized to the centrosome (3). This suggests that STK15 could indeed associate with PP1␣ in vivo. In addition, it is possible that STK15 may also interact with PP1␦ as STK15 might be involved in kinetochore-microtubule in the presence of okadaic acid. Phosphorylated proteins were detected by autoradiography. B, STK15 were immunoprecipitated from interphase extracts (black bar) and mitotic extracts (gray bar) with anti-STK15 antibody, and then in vitro kinase assay was performed using GST-PP1␣ as substrate in the presence of okadaic acid. After extensively washing, phosphatase assay of GST-PP1␣ was carried out using 32 P-labeled MBP as substrate. Protein phosphatase activity was plotted as a percentage of relative activity. The error bars indicate the S.D. (n ϭ 3). C, transiently transfected HeLa cells with FLAG-tagged WT (black bar) and Lys 3 Arg mutant (K/R) (gray bar) were cultured in the presence of nocodazole for 16 h. Immunocomplexes of FLAG-tagged protein and PP1 were immunoprecipitated with anti-FLAG antibody, and phosphatase assay was performed. Protein phosphatase activity was plotted as a percentage of relative activity. The error bars indicate the S.D. (n ϭ 3). D, immunocomplexes of STK15 and PP1 from interphase extracts (black bar) and mitotic extracts (gray bar) were immunoprecipitated with anti-STK15 antibody, and then phosphatase assay was performed. Protein phosphatase activity was plotted as a percentage of relative activity. The error bars indicate the S.D. (n ϭ 2). 3 Arg mutant (lane 5) from mitotic extracts with anti-FLAG antibody were assayed for kinase activity using ␤-casein as substrate. Immunoblotting (IB) was performed with anti-STK15 antibody (top panel), and autophosphorylated STK15 (middle panel) and phosphorylated ␤-casein (bottom panel) were detected by autoradiography. Empty vector served as a control (lane 1). Super-phosphorylated STK15 and hyperphosphorylated STK15 are indicated by arrowhead and arrow, respectively. attachment mentioned above. Our results have revealed that STK15 shows strong binding to PP1 at metaphase and also during early stages of interphase, although its kinase activity peaks only during G 2 /M phase of the cell cycle. This apparent paradox can be explained by the fact that PP1 bound to activated STK15 at G 2 /M phase has reduced phosphatase activity that helps remove the inhibitory phosphates on STK15 to keep it activated, whereas stoichiometric over-representation of activated bound PP1 helps keep STK15 inactive during early interphase. It is also likely that still unidentified additional proteins play a key role in the feedback regulation of STK15 kinase and PP1 phosphatase activities.
In this paper we present evidence that the interaction between STK15 and PP1 is necessary for STK15 kinase activation, and the two enzymes inhibit each other both in vivo and in vitro. STK15 is overexpressed in various cancers, and such overexpression induces oncogenic transformation in mammalian cells. On the other hand, PP1 isoforms are also found to be overexpressed in various cancers, which is taken as evidence for a positive role of PP1 in tumorigenesis (14,36,37). These findings suggest that in cancer cells phosphorylation/dephosphorylation balance between STK15 and PP1 is deregulated. The findings reported in this paper offer an interesting lead toward elucidating the pathway responsible for causing mitotic anomalies in STK15 overexpressing cancer cells.