Stk10, a New Member of the Polo-like Kinase Kinase Family Highly Expressed in Hematopoietic Tissue*

The Ste20 family of serine/threonine kinases plays an important role in numerous cellular functions such as growth, apoptosis, and morphogenesis. We have identified a previously cloned but uncharacterized family member termed Stk10, which is a human homolog of murine Lok, a serine/threonine kinase highly expressed in lymphocytes. Northern analysis demonstrated that the Stk10 transcript is present in many tissues, although highest expression levels are seen in hematopoietic cells. Due to close sequence homology to human Slk andXenopus laevis xPlkk1, two polo-like kinase kinases, we investigated whether Stk10 might also play a role as a Plk1 activator. Plk1 has been shown to be overexpressed in multiple tumor types, thus attracting high interest to its potential upstream regulators. We show here that Stk10 can associate with Plk1 in cells and furthermore can phosphorylate Plk1 in vitro. Engineered NIH-3T3 cell lines that overexpress a dominant negative version of Stk10 display an altered cell cycle phenotype characterized by increased DNA content, raising the possibility that expression of a dominant negative Stk10 may impinge upon Plk1 function in vivo; it has previously been shown that unregulated expression of Plk1 can result in a variety of nuclear defects. We suggest, therefore, that Stk10 is a novel polo-like kinase kinase that cooperates with hSlk to regulate Plk1 function in human cells.

The regulation of cell division is complex and tightly controlled. A number of protein kinases and phosphatases play important roles in the coordination of this process. Maturation promoting factor, a complex of the serine/threonine kinase Cdc2 and its partner cyclin B, has been shown to be essential for successful progression from G 2 to M phase (reviewed in Ref. 1). In turn, a number of other proteins have important functions in helping regulate the activity of Cdc2/cyclin B. In Xenopus laevis, the polo-like kinase Plx1 has emerged as potentially impacting this complex by phosphorylating and thus increasing the activity of Cdc25C, the phosphatase responsible for activating the Cdc2/B kinase (2). There is currently debate in the field whether this is an organism-specific function; experiments with small interfering RNA in mammalian cells do not point to polo-like kinase (Plk1) having a similar role (3). However, experiments with Plk1 immunoprecipitated from Jurkat T leukemia cells suggest that human Plk1 can indeed perform this function in vitro (4). Furthermore, recent evidence suggests that Plk1 phosphorylation of Cdc25C leads to nuclear localization of the mitotic phosphatase (5), thus contributing to its regulation in a much different manner. Despite this controversy over Cdc25C activation, it has been well established that Plk1 activity is critical for many other mitotic processes (for review, see Ref. 6).
Through a number of genetic studies in yeast and Drosophila, the role that Plk1 plays in regulating numerous aspects of mitosis has become more clear (7). Monopolar spindles, misaligned chromosomes, and a failure to undergo cytokinesis are all hallmarks of Plk1 mutants (cdc5 in budding yeast and polo in Drosophila) (8,9). How these steppingstones to successful cytokinesis are dependent upon Plk1 function is more ambiguous and only partially understood. Plk1 has been shown to phosphorylate components of the anaphase promoting complex (APC), 1 thus activating the complex and allowing ubiquitinmediated proteolysis to occur at the end of anaphase (10 -14). In addition, a direct link has also been demonstrated to exist between Plk1 and Cdc2/cyclin B independent of Cdc25 activation. Several groups have shown that Plk1 is capable of phosphorylating key residues on cyclin B that are responsible for its nuclear relocation during prophase (15,16). Thus, what is emerging is that this kinase may be a component of multiple key checkpoints that must be passed for normal cell division. Just as intriguing are the number of studies that have recently pointed to the large overexpression of Plk1 in tumors; correlations have been made between the relative level of Plk1 expression and long term prognosis/survival (17)(18)(19)(20)(21)(22).
Regulation of Plk1 activity appears to occur through two separate mechanisms. The cell controls the activity of Plk1 dramatically through transcriptional repression (23). At G 1 phase and in quiescent cells, the level and activity of Plk1 is barely detectable (24,25). As the cell progresses through the cell cycle and undergoes replication, the level of Plk1 rises markedly and reaches a peak between G 2 and M phases. Following successful cytokinesis, Plk1 is ubiquitinated and destroyed by the same complex it helped activate, the APC, thereby returning to the low basal level seen in G 1 phase (26). In addition to this tight transcriptional control, the kinase is also up-regulated by phosphorylation (24,27). Recent work in X. laevis has provided us with the most information in this regard. Maller and co-workers (28) purified an activity from Xenopus eggs they termed xPlkk1, for Xenopus Plk kinase, which they suggest is responsible for activating Plx during oocyte maturation. Following this discovery, a mammalian pro-tein, Slk, which shares significant homology to xPlkk1, was shown to phosphorylate and activate Plk1 in vitro (29).
Here we show that another protein kinase may contribute to the regulation of Plk1. The murine gene Lok shows a great deal of sequence similarity to both xPlkk1 and Slk. We, along with another group, recently cloned the human homolog of Lok called Stk10 (30). In addition to sharing significant sequence similarity with both of these polo-like kinase kinases, we show here that Stk10 is also able to phosphorylate Plk1. Unlike Lok, Stk10 has a widespread tissue distribution, showing expression in a number of proliferative tissues as well as in multiple tumor lines. These data raise the intriguing possibility that Stk10 may play a role in regulating Plk1 in these specific biological settings.

EXPERIMENTAL PROCEDURES
Cloning/Subcloning-A human Stk10 clone, AA459448, was first identified using a Smith-Waterman search of the EST data base with the human Slk gene (GenBank TM accession number AB002804) as a query. Sequence analysis of the 1286-bp insert identified a 1227-bp open reading frame (409 amino acids) with the potential to encode the N terminus of a novel human STK related to the human Slk gene product. An additional Smith-Waterman search using the C terminus of the Slk gene as a query yielded three additional EST numbers, AA323687, AA380492, and AA168869, that encode the C-terminal region of human Stk10. PCR using single-stranded cDNA from human testis and the H23 tumor cell line enabled a complete sequence of Stk10 to be assembled. PCR primers were designed based upon the ESTs recovered as well as murine Lok (m_Lok).
To generate a kinase-dead mutant of Stk10, primers were designed to convert the conserved lysine (amino acid 66) in the kinase domain to an isoleucine using the Stratagene QuikChange kit (primers used were 5Ј-GCTTTGGCTGCGGCCATAGTCATTGAAACCAAG-3Ј and 5Ј-CTT-GGTTTCAATGACTATGGCCGCAGCCAAAGC-3Ј). Clones were sequenced to ensure that only the correct mutation was present. Wild type and the kinase-dead mutant Stk10, tagged on the C terminus with the HA epitope, were also subcloned into pBabepuro for stable cell line generation in NIH-3T3 cells.
Northern Blots-A Clontech Northern blot (#7780-1, 12-lane multiple tissue Northern blot) was incubated with a 659-bp fragment of Stk10 (1263-1922, EcoRI/BamHI) according to the manufacturer's protocol with slight modifications (200 Ci of [␣-32 P]ATP was used to label the probe, which was incubated with the blot overnight at 65°C). After multiple washes, the blot was exposed to a PhosphorImager screen and later read on a Amersham Biosciences PhosphorImager.
Immunoprecipitations, Kinase Assays, and Western Analysis-Cells were lysed in Nonidet P-40 lysis buffer (20 mM Tris, pH 8.0, 137 mM NaCl, 10% glycerol, 1% Nonidet P-40, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, aprotinin (0.15 unit/ml), 20 M leupeptin, 5 mM sodium vanadate) for 10 min prior to a clarifying centrifugation at 14,000 rpm. Protein concentrations were determined using a Pierce BCA kit (Rockford, IL). Equivalent amounts of protein lysate were pre-cleared by incubation with protein A-agarose for 1 h. The cleared lysates were then transferred to a fresh tube containing protein Aagarose and 2 g of affinity-purified Stk10 antibody and rocked for 5-16 h at 4°C. Immune complexes were washed three times in lysis buffer and three times in cold kinase buffer (3 mm MnCl 2 , 10 mM MgCl 2 , 20 mM Hepes, pH 7.6). 10 l of kinase buffer containing 10 Ci of [␥-32 P]ATP, 200 M ATP, and 5 g of histone H2A were then added to each tube and incubated for 20 min at 30°C. Kinase assays with Stk10 and Plk1 were washed an additional three times in a high salt buffer (500 mM LiCl, 100 mM Tris, pH 7.4) before the final washes with cold kinase buffer.
Stk10 Antibody Generation/Affinity Purification-A glutathione Stransferase (GST) fusion protein was constructed by subcloning an internal, poorly conserved region of Stk10 (amino acids 352-477) into pGEX-4T utilizing BamHI and SalI restriction sites. BL21 cells were transformed with this construct, and protein production was induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. The protein was then purified on glutathione-Sepharose 4B (Amersham Biosciences), eluted with excess free glutathione and dialyzed overnight with phosphate-buffered saline. Naïve rabbits were injected every 3 weeks with 0.5 mg of GST fusion protein. The first injection was mixed 1:1 with Freund's complete adjuvant, and all others were mixed 1:1 with Freund's incomplete adjuvant. To affinity purify the crude serum, it was incubated with CnBr-activated Sepharose 4B covalently bound to the GST fusion protein according to manufacturer's instructions (Amersham Biosciences). After numerous washes, the antibody was eluted off of the resin with propionic acid, neutralized, and dialyzed overnight in PBS.
Xenopus Assays and Manipulations-Xenopus females were purchased from Nasco (Fort Wilkinson, WI). Oocyte removal and injection were done as previously described (31). Briefly, oocytes were removed and defolliculated by incubation in MMR (100 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 5 mM HEPES, pH 7.5) containing collagenase A (Roche Applied Science, Indianapolis, IN) (1.5 mg/ml) for 1.5 h. Oocytes were then washed several times in MMR and cultured overnight in 50% Leibovitz-15 medium (Invitrogen, San Diego, CA). Eighteen hours after isolation, oocytes were injected with ϳ30 ng of in vitro transcribed RNA encoding the Stk10 proteins (wild type or kinasedead). Capped RNA was transcribed using the Ambion Message Machine kit (Austin, TX). Following injection of RNA, progesterone (Sig- FIG. 1. Stk10 expression. A, a Clontech multiple tissue Northern blot was incubated with an Stk10-specific probe. Highest expression is seen in rapidly proliferating tissue (spleen, placenta, and peripheral blood leukocytes), although multiple transcripts are also seen in the other tissues profiled. B, expression in tumor lines was examined by Western analysis. Proteins (50 g of whole cell lysate) were resolved by SDS-PAGE, transferred to nitrocellulose, and probed with an Stk10specific antibody. ma) was added to the Leibovitz medium at 1 g/ml. GVBD was scored by the appearance of a white dot in the animal pole of the oocyte.
Oocytes were lysed in Nonidet P-40 buffer by gentle pipetting through a micropipette tip and spinning briefly at 14,000 to remove insoluble material. Lysates were then used for Western analysis.
Cell Culture, Transfection, and Synchronization-HeLa, 293T, and COS-7 cells were maintained in Dulbecco's minimal essential medium (Invitrogen) supplemented with 10% fetal bovine serum and penicillin/ streptomycin. HEK-293T and COS-7 cells were transiently transfected using SuperFect (Qiagen, Valencia, CA) according to the manufactur-er's specifications and incubated for 48 h before being lysed and analyzed for protein expression. HeLa cells were synchronized by a thymidine/aphidicolin block and then released into fresh media, and samples were collected periodically for biochemical and flow cytometry analysis. Cells were plated at a low density (5 ϫ 10 5 /10-cm plate), allowed to attach, and then incubated with 2 mM thymidine (Sigma) for 12 h. 8 h after being washed and released into fresh media, 1 g/ml aphidicolin (Sigma) was added to the media and the cells were incubated for another 12 h. The cells were then washed and released to progress through the cell cycle by incubation with fresh media. NIH-3T3 cells were synchronized by a sequential serum starvation (0.1% FBS) for 36 h, followed by incubation with fresh media (10% FBS) containing 1 g/ml aphidicolin for 16 h. The cells were then washed twice with PBS, and fresh media was added (10% FBS) to allow free progression through the cell cycle. Cells were periodically collected for flow cytometry analysis.
Cell Cycle Analysis by Flow Cytometry-Cells were detached with trypsin and washed once with PBS. The cell pellets were then resuspended in 0.5 ml of PBS, fixed by the slow addition of ice-cold 70% ethanol, and stored at Ϫ20°C for at least 2 h. The fixed cells were centrifuged at 300 ϫ g for 5 min and washed once in PBS. The cell pellets were then stained in 0.1% Triton X-100, 20 g/ml propidium iodide (Roche Applied Science), 200 g/ml RNase (Sigma) in PBS and incubated for 30 min in the dark.

RESULTS
Expression Pattern of Stk10 -The closest characterized ortholog to Stk10 (sharing 88% identity) is murine lymphocyteoriented kinase (Lok). Western analysis has suggested that Lok expression is relatively restricted to lymphoid tissue (32). To ascertain whether Stk10 shares a common expression pattern with the mouse homolog we performed Northern blot analysis with a Clontech 12 tissue array blot and an Stk10-specific probe. As shown in Fig. 1A, high expression was seen in certain rapidly proliferating tissues (spleen, placenta, and peripheral blood leukocytes), but transcripts were also detectable in the other tissues profiled (brain, heart, skeletal muscle, colon, thymus, kidney, liver, small intestine, and lung) indicating that Stk10 has a more widespread pattern of expression than has been reported so far for its mouse homolog, Lok (32). In all tissues examined, multiple transcripts were observed, although we saw no evidence that different splice variants are translated. In addition, TaqMan analysis was performed on a similar set of tissues; results correlated well with that seen by Northern analysis (data not shown).
To detect Stk10 protein expression, a rabbit polyclonal antibody was generated against the internal domain of Stk10 as described under "Experimental Procedures." Given the elevated level of Stk10 mRNA in highly proliferative tissue, we examined a set of tumor cell lines for Stk10 protein expression. It is clear from Fig. 1B that Stk10 is expressed in many different tumor lines; interestingly, the highest level of expression does not occur in the leukemic lines as one might predict based upon the tissue distribution of the murine ortholog.
Kinase Activity of Stk10 in Vitro-To verify the kinase activity of recombinant Stk10, we transiently transfected HEK-293T cells with DNA encoding either wild type or a kinase-dead mutant of Stk10. The proteins were immunoprecipitated using an HA-epitope tag engineered onto the C terminus, and a kinase assay was performed as described under "Experimental Procedures." The results of this assay are shown in Fig. 2A. Clear autophosphorylation of wild type Stk10 as well as robust phosphorylation of histone H2A was observed. No autophos-phorylation or phosphorylation of substrate by the kinase-dead mutant was detected. The choice of substrate was based upon earlier published results with Lok (32). Unlike Lok, however, Stk10 does not appear to phosphorylate myelin basic protein (data not shown). Western analysis (Fig. 2B) shows the levels of wild type and kinase-dead Stk10 assayed for activity. Slightly less kinase-dead than wild type protein was assayed in this experiment; however, the difference in histone phosphorylation between the two samples far outweighs the minor difference in expression.
Stk10 as a Potential Polo-like Kinase Kinase-To understand the potential role for Stk10 in human cells, we examined two similar homologs with known functions found in humans and X. laevis (hSlk and xPlkk1, respectively). As Fig. 3 illustrates, both genes have a domain structure similar to Stk10, i.e. they contain an N-terminal kinase domain as well as a coiled-coil domain at the C terminus. Most importantly, hSlk and xPlkk1 have a common function in cells; they both act as polo-like kinase kinases, phosphorylating and activating polo-like kinase. These sequence similarities raised the possibility that Stk10 might act in a like fashion to regulate the activity of Plk1.
Co-association of Plk and Stk10 in Cells-We next examined whether Plk and Stk10 can associate in cells; such an interaction would suggest that Plk is an in vivo substrate for Stk10. We therefore performed immunoprecipitations from both exponentially growing and nocodazole-treated HeLa cells using the Stk10-specific antibody we had generated. The resulting immunoprecipitations were analyzed by SDS-PAGE, transferred to nitrocellulose, and probed with both Stk10 and Plk1 antibodies. The association between Plk1 and Stk10 was clearly seen in nocodazole-treated cells (Fig. 4A) but not visible in exponentially growing cultures. A straightforward explanation for this may be the difference in Plk1 protein levels between the two samples. Plk1 expression is low in most phases of the cell cycle, and it reaches maximal expression between G 2 and M phases. To examine this interaction in non-arrested cells, we synchronized HeLa cells at the G 1 /S interface and collected samples every 2 h after their release with fresh media. Stk10 was immunoprecipitated from lysates containing equal amounts of protein, and Western analysis was performed. Consistent with the data from nocodazole-treated cells, Plk1 co-association with Stk10 was most clearly visible between 10 and 12 h during G 2 and M phases, when Plk1 levels are highest (Fig. 4B, upper  panel). Interestingly, there is a detectable amount of Plk1 present in the lysates at 14 h when little association between Stk10 and Plk1 is seen. We cannot rule out the possibility, therefore, that phosphorylation of either protein during M phase may contribute to their association at this time.
Cell Cycle Expression/Activity Profile of Stk10 -To examine the cell cycle profile of Stk10, we again synchronized HeLa cells at the G 1 /S interface using a thymidine/aphidicolin block. Upon release from this arrest, samples were collected every 2 h to assay for kinase activity and Western analysis. In contrast with the pattern seen with Plk1, the abundance of Stk10 did not vary significantly during the time course (Figs. 4B and 5B) suggesting that the expression profile of Stk10 is constant throughout the cell cycle. Similarly, the ability of Stk10 to both autophosphorylate and phosphorylate histone H2A was relatively stable (Fig. 5A). This pattern of activity and expression is very similar to that seen with another human polo-like kinase kinase, hSlk.
Stk10 Can Phosphorylate Plk1 in Vitro-To confirm the potential role of Stk10 as a polo-like kinase kinase, we performed in vitro kinase assays with Plk1 as the substrate. Briefly, a kinase-dead mutant of Plk1, tagged on the C terminus with an HA epitope, was transfected into COS-7 cells. Kinase-dead Plk1 was used to avoid signal generated by autophosphorylation. Separate dishes of cells were transfected with either a wild type or kinase-dead form of Stk10 tagged on the N terminus with a FLAG epitope. 48 h after transfection the cells were lysed, clarified lysates were mixed, and the proteins were immunoprecipitated with their respective tags. After a stringent set of washes, an in vitro kinase assay was performed. Fig. 6A (upper panel) clearly shows the high incorporation of 32 P into the kinase-dead Plk1 only occurs in the presence of wild type Stk10. A Plk immunoblot indicates that equal amounts of kinase-dead Plk were present in each sample assayed (Fig. 6A, lower panel).
In addition, to demonstrate the efficiency of the reaction, the amount of substrate (kinase-dead Plk1) was titrated and found to directly modulate the reaction rate (data not shown).
To extend these findings, we then assayed endogenous Stk10 precipitated from aphidicolin/thymidine-arrested HeLa cells for the ability to phosphorylate kinase-dead Plk1. Synchronized, G 1 phase cells were used to reduce the background from endogenous, active, wild type Plk1, which could co-precipitate with Stk10 during other phases of the cell cycle. As shown in Fig. 6B, endogenous Stk10 is able to phosphorylate Plk1 in vitro, lending support to the model that Stk10 is an upstream regulator of Plk1 in vivo.
Overexpression of WT Stk10 in Xenopus Oocytes Accelerates GVBD-Microinjection of RNA encoding xPlkk1 into immature, Stage VI Xenopus oocytes can impinge on the normal maturation process that occurs after progesterone exposure. Overexpression of wild type xPlkk1 results in an acceleration of germinal vesicle breakdown (GVBD) compared with uninjected controls or expression of a kinase-dead mutant (28). Based upon the high degree of similarity between Stk10 and xPlkk1, we investigated whether injection of Stk10 RNA could similarly increase the rate of progesterone-induced maturation. RNA corresponding to wild type and kinase-dead Stk10 was microinjected into Stage VI immature oocytes, the oocytes recovered overnight and were then treated with progesterone. As displayed in Fig. 7A, overexpression of wild type Stk10 resulted in an acceleration of GVBD. Maturation was scored by the appearance of a white dot at the animal pole, which is formed when the nucleus breaks down and pigment granules are displaced. Oocytes lysed for Western analysis are displayed in Fig.  7B. Clear expression of both wild type and kinase-dead Stk10 is visible. The correlation of a biological phenotype between xPlkk1 and Stk10 strongly suggests the two proteins share a common function in cells.

Overexpression of Kinase-dead Stk10 Results in Abnormal DNA Content-Misregulation of Plk expression in cells results
in a number of nuclear defects. We determined whether overexpressing wild type or kinase-dead Stk10 could interfere with normal Plk function and thus lead to nuclear aberrations. NIH-3T3 cells were engineered to stably overexpress either a wild type or a kinase-dead mutant of Stk10, tagged on the C terminus with an HA epitope. We noted that it was much more difficult to isolate single clones expressing the kinase-dead mutant than to find wild type-expressing clones. Growth curves confirmed that cells expressing kinase-dead Stk10 grew more slowly than those expressing the wild type protein (data not shown).
Given our proliferation data and the potential role that Stk10 plays in cell cycle regulation, we next examined the ability of these various clones to cycle in a synchronous manner. The cells were blocked at the G 1 /S interface with aphidicolin, then fixed, stained, and analyzed by flow cytometry. The results from a typical experiment are shown in Fig. 8. A dramatic effect was observed in the clones expressing kinase-dead Stk10. In a dose-dependent fashion, expression of the mutant Stk10 resulted in an abnormal DNA content. At the time sampled, when the majority of each population should be arrested in G 1 /S phase with 2N DNA, it is clear that most of the cells expressing the kinase-dead allele of Stk10 have 4N DNA content (Fig. 8B, top panels). This effect continued as the cells progressed through the cell cycle; when control cells had duplicated their DNA to 4N, the kinase-dead expressing cells also increased their DNA content to ϳ8N (Fig. 8B, bottom panels). In addition, these cells were not multinucleate; instead, a process related to endoreduplication appeared to have occurred. In support of this theory we found that the nuclear size of the kinase-dead expressing cells was, on average, larger than those of the vector-transfected cells or wild type-expressing cells (data not shown). Expression of the kinase-dead Stk10 clearly interfered with normal cell cycle progression and cytokinesis. DISCUSSION Over the past few years it has become clear that Plk1 plays a key role in controlling the cell cycle, specifically in regulating many aspects of mitosis. Entrance to mitosis may rely on this kinase; polo-like kinase has been implicated in indirectly activating the mitotic kinase complex Cdc2/cyclin B by phosphorylating the activating phosphatase, Cdc25C (2,4,33). Exit from mitosis has also been associated with Plk function; activation of the APC, the proteasome responsible for destroying a number of mitotic proteins, has been shown to occur through Plk1 (10 -14). In this study we examined the function of Stk10, a previously uncharacterized member of the Ste20 family of serine/threonine kinases. Based upon sequence homology with Slk and xPlkk1, known polo-like kinase kinases in humans and X. laevis, we posited that Stk10, too, might exert control over the activity of Plk1. Toward that end, we show here that Stk10 is expressed, like Plk1, in highly proliferative normal tissue as well as in a number of tumor cell lines. In addition, Stk10 co-associates with Plk1 in cells and is capable of phosphorylating a catalytically inactive mutant version of Plk1 in vitro. More compelling is the observed phenotype of NIH-3T3 cells engineered to overexpress a dominant negative form of Stk10. These cells display an abnormal cell cycle profile and increased DNA content in an Stk10-dose dependent manner. Such a phenotype is consistent with a partial loss of Plk1 function in these cells. We presume that expression of kinase-dead Stk10 does not completely inhibit Plk activity; we would expect that full inhibition of the kinase would result in a cell cycle block.
Experiments in a number of genetic organisms such as Drosophila and Saccharomyces cerevisiae have shown convincingly that expression of mutant forms of Plk1 can lead to monopolar spindles, abnormal chromosome segregation, and other gross errors in cytokinesis. Furthermore, recent studies utilizing small interfering RNA in HeLa and U2OS cells suggest that depletion of Plk1 transcript can also interfere with sister chromatid separation and normal cytokinesis in mammalian systems. In fact, additional studies imply that the absolute amount of Plk1 in the cell is of critical importance; fluctuations in either direction can lead to abnormalities. Interestingly, Mundt et al. (34) has shown that overexpression of wild type Plk1 in HeLa cells can lead to multiple and fragmented nuclei. All of these studies support the idea that interfering with the normal activity level of Plk1, whether through the introduction of a dominant negative mutant or the supplementation of normal activity with exogenous wild type Plk1, can have deleterious effects on the cell. Perturbation of this balance results in a host of cell duplication mistakes. Clearly, the tight regulation of Plk1 that exists on multiple levels has evolved for a distinct reason.
Our experiments in this current study further reinforce the idea that the amount of Plk1 activity is of critical importance to successful cell division. NIH-3T3 cells that express a kinasedead mutant of Stk10 have a greatly altered cell cycle phenotype. At each phase, the cells contain a larger amount of DNA than normal, suggesting that a process similar to endoredupli-cation has occurred. Not surprisingly, the cells with up to twice as much DNA as controls also have larger nuclei and divide more slowly than vector-transduced cells (data not shown). There are a number of possible mechanisms capable of producing this observed phenotype. The most straightforward is that the expressed Stk10 is exerting a dominant negative effect on the endogenous Plk1, preventing its normal activation. This would result in a condition similar to that of overexpressing a dominant negative form of Plk1. Alternatively, the exogenous protein may be binding to Plk1, allowing its normal activation to occur but slowing down or preventing the necessary proteolytic destruction that must take place at the end of mitosis. In this way, the result would mimic that seen when wild type Plk1 is overexpressed in cells, i.e. fragmented or duplicated nuclei (34). Additional experiments are needed to fully explain which mechanism is behind our observed phenotype. It is interesting to note that wild type Plk1 is overexpressed in a number of tumor cell lines (35). It has been shown that constitutive ex- pression of Plk1 can transform cells (36); certainly it is true that the type of genetic instability described above is a clear hallmark of the cancerous cell.
Given the stringent control that the cell must place on the activity of Plk1, it is not surprising that there now appear to be multiple kinases capable of phosphorylating the cell cycle regulator. Although the upstream activator of Plk1 has not been definitively identified, our current study demonstrating the ability of Stk10 to phosphorylate Plk1 in vitro suggests it is a candidate, along with Slk (29), for the physiological activator. Northern analyses of Stk10 and Slk imply that the expression pattern of the two genes is somewhat complementary. Highest levels of Slk transcript are observed in heart and skeletal muscle, whereas Stk10 RNA is less common in these tissues, showing higher levels instead in spleen and peripheral blood leukocytes. The two protein kinases may therefore provide the same function to Plk1 but in different cell types. Further evidence that Stk10 plays a role in activating Plk1 in vivo is provided by Erikson and co-workers. In a recent publication they state that unpublished experiments have demonstrated the ability of Lok, the mouse ortholog of Stk10, to activate Plk1 in vitro (37).
The expression pattern we observe for Stk10 is somewhat different than that reported for Lok. Kuramochi et al. (32) showed that the expression of Lok is quite restricted, detecting transcript and protein only in hematopoietic tissue. Although we do observe high levels of Stk10 transcript in lymphocytic organs, a number of other tissues showed measurable levels of Stk10 RNA as well. It is possible that the function of Stk10 is more complex in humans than in mice and so requires expression in additional tissues. A potential area for divergent complexity between the species lies within the Plk family itself. Although quite a bit of work has been done to elucidate the function of Plk1, less is known about Plk2 and Plk3 (also known as Snk and Fnk/Prk). Both of these genes are immediate-early gene products and so appear to play a role in regulating earlier phases of the cell cycle, G 1 and S phases, although recent work suggests that Fnk may also impact on M phase (38 -40). Given that the Plks share a relatively conserved kinase domain, it is possible that Stk10 and/or Slk may also act as regulators of these kinases in human cells, perhaps thus requiring the broader expression pattern observed.
We have also shown through Western analysis that there is a detectable amount of Stk10 present in a panel of tumor cell lines. A number of investigators have reported elevated levels of Plk1 protein and transcript in tumors and tumor cell lines. Because there have thus far been no reports of activating mutations in Plk1 in cancerous cells, there is clearly a need in these cells for an upstream activator. Experiments with Plx in X. laevis have demonstrated that phosphorylation plays an obligate role in the activation of polo-like kinase. Stk10 may provide this function in human tumors; it is possible that Slk also contributes to this activity, at this point expression profiles in tumors for Slk have not been reported. Interestingly, we found no evidence that the activity of Stk10 is regulated in a cell cycle-dependent fashion. Both protein level and activity remained relatively constant in all phases of the cell cycle. It is still possible that there are more subtle variations than we were able to measure; we note that the differences in activity reported for Slk were extremely modest, although investigators did present evidence that the protein is phosphorylated during M phase.
Our study provides strong evidence that Stk10 is one of the physiological activators of Plk1. It will be interesting in future studies to examine the specific sites in Plk1 phosphorylated by Stk10. As noted earlier, Jang et al. (37) stated that Lok, the mouse ortholog of Stk10, phosphorylated Plk1 in vitro on threonine 210, which they demonstrated was an in vivo site of phosphorylation during mitosis. It seems logical to hypothesize that Stk10 will also activate Plk1 in the same fashion; further experiments will help resolve this question.