Nuclear Translocation of Plk1 Mediated by Its Bipartite Nuclear Localization Signal*

Polo-like kinase 1 (Plk1), a mammalian ortholog of Drosophila Polo, is a serine-threonine protein kinase implicated in the regulation of multiple aspects of mitosis. The protein level, activity, and localization of Plk1 change during the cell cycle, and its proper subcellular localization is thought to be crucial for its function. Although localization of Plk1 to the centrosome has been established, nuclear localization or nucleocytoplasmic translocation of Plk1 has not been fully addressed. Here we show that Plk1 accumulates in both the nucleus and the cytoplasm in addition to its localization to the centrosome during S and G2 phases. Our results identify a conserved region in the kinase domain of Plk1 (residues 134–146) as a functional bipartite nuclear localization signal (NLS) sequence that regulates nuclear translocation of Plk1. The identified NLS is necessary and sufficient for directing nuclear localization of Plk1. This bipartite NLS has an unusually short spacer sequence between two clusters of basic amino acids but is sensitive to RanQ69L, a dominant negative form of Ran, similar to ordinary bipartite NLS. Remarkably, the expression of an NLS-disrupted mutant of Plk1 during S phase was found to arrest the cells in G2 phase. These results suggest that the bipartite NLS-dependent nuclear localization of Plk1 before mitosis is important for ensuring normal cell cycle progression.

Polo-like kinase 1 (Plk1), 1 a mammalian ortholog of Drosophila Polo, is a serine-threonine kinase implicated in the regulation of multiple aspects of mitosis including centrosome maturation and regulation, spindle assembly, sister chromatin separation, cytokinesis, and exit from M phase (1)(2)(3)(4)(5). The Plk1 protein level is low in G 1 phase, increases during S phase, remains high through G 2 -M phase, and is rapidly decreased after mitosis. During the G 2 to M phase transition, Plk1 is phosphorylated and its kinase function is stimulated. In M phase, Plk1 is distributed at spindle poles, kinetochore during prophase and metaphase, and redistributed to the spindle equatorial region during anaphase and concentrated within post-mitotic bridge during telophase (6 -11). It is believed that this complex localization pattern may reflect multiple roles for Plk1 throughout mitosis. Additionally, previous work shows that Plk1 is able to activate Cdc2 by phosphorylating and activating Cdc25C (12)(13)(14)(15)(16)(17). Recently, we reported that Plk1 phosphorylates cyclin B1 and Cdc25C and targets them to the nucleus during prophase (18,19). These observations suggest that Plk1 plays a role in the regulation of entry into M phase as well. Thus, we have been interested in subcellular localization of Plk1 during S and G 2 phases that has not been examined in detail with the exception of the localization of Plk1 to the centrosome.
Proteins larger than 45 kDa often require a specific sequence called nuclear localization signal (NLS) to be targeted to the nucleus. NLSs are defined as the sequence sufficient and necessary for nuclear import of their respective proteins and are generally functional in targeting heterologous proteins to the nucleus. Classical NLSs are either monopartite or bipartite. Of the two basic types of NLSs, a monopartite NLS consists of a single short consecutive basic amino acid, the first example of which was the SV40 large T-antigen NLS (PKKKRKV), and a bipartite NLS comprises two clusters of basic amino acids separated by a 10 -12-amino acid spacer, which was first reported in the NLS of nucleoplasmin (KRPAATKK-AGQAKKKKLDK) (20 -22).
In this study, we have demonstrated localization of Plk1 to the nucleus as well as to the cytoplasm during S and G 2 phases and identified for the first time a bipartite nuclear localization signal sequence in Plk1 that regulates nuclear translocation of Plk1. In addition, we present several lines of evidence suggesting that nuclear localization of Plk1 before mitosis is important for ensuring normal cell cycle progression.

EXPERIMENTAL PROCEDURES
Cell Culture, Synchronization, and Transient Transfection-HeLa cells were cultured and synchronized with a double thymidine block as previously described (23). HeLa cells were transiently transfected by the use of FuGENE 6 according to the manufacturer's instructions with the use of 2 g of total DNA/35-mm dish. After 20 h, cells were fixed and stained.
Cell Fractionation-Fractionation of HeLa cell extracts in each time after release from double thymidine block was performed as described previously (24). The nuclear and cytoplasmic fractions were analyzed by SDS-PAGE and immunoblotting. The nuclear and cytoplasmic fractions were normalized to contain equal quantity of total proteins. DNA Constructs-The mutagenesis of Plk1 3A was performed using a mutagenic primer 5Ј-GTTTTGGAGCTCTGTGCCGCGGCGTCCCTC-CTGGAGCTG-3Ј by the use of QuikChange site-directed mutagenesis kit (Stratagene). To yield Plk1 4A, a mutagenic primer 5Ј-CTGTCGC-GCGGCGGCCGCCCTGGAGCTGC-3Ј was used. PCR products of Plk1 WT, 3A, 4A, and 7A were subcloned into pSR␣-HA vector. The mutations were confirmed by DNA sequencing. A sequence corresponding to residues 124 -148 of human Plk1 was amplified by PCR with a 5Ј primer 5Ј-CGGGATCCGACTTTGTATTTGTAG-3Ј and a 3Ј primer 5Ј-ACGCGTCGACTCACAGTGCCTTCCTC-3Ј, generating BamHI and SalI sites at the 5Ј and 3Ј ends, respectively, and the fragment was * This work was supported by grants from the Ministry of Education, Science and Culture of Japan (to E. N.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Microinjection-GST-Plk1 proteins were injected into the cytoplasm of HeLa cells as described previously (25). At 1 h after injection, cells were fixed and stained with anti-GST antibody (Santa Cruz Biotechnology). SR␣-HA-Plk1 WT or 7A was injected into the nucleus of HeLa cells at 4 h after the release from double thymidine block.

RESULTS AND DISCUSSION
Plk1 Enters the Nucleus during Interphase-To examine subcellular localization of endogenous Plk1 during interphase, we used the indirect immunofluorescent staining and the biochemical cell fractionation method in synchronized HeLa cells. Consistent with previous observations (6 -11), the protein level of Plk1 is very low in G 1 /S phase, but its localization to the centrosome is detectable (Fig. 1, 0h). In S and G 2 phases, Plk1 increases dramatically and localizes to both the cytoplasm and the nucleus in addition to the centrosome (Fig. 1, 4h and 6h). At prophase, the protein level of Plk1 is maximal and it shows both nuclear and cytoplasmic localization (Fig. 1, 8h). Thus, these observations indicate that Plk1 has the ability to enter the nucleus constitutively during interphase.
Identification of a Bipartite Nuclear Localization Signal in Plk1-It is generally believed that those proteins whose molec-ular mass is larger than 50 kDa cannot pass through the nuclear pore by passive diffusion. Because Plk1 is ϳ67 kDa, Plk1 should have some mechanism to enter the nucleus. We found a region in human Plk1 where two clusters of basic amino acids, residues 134 -136 (RRR) and residues 143-146 (KRRK), are aligned in tandem ( Fig. 2A). This region lies in the kinase catalytic domain in the primary sequence, and its se- quence is highly conserved in Xenopus Plx1 and Drosophila Polo (Fig. 2A). The yeast orthologs, Cdc5 and Plo1, lack the first cluster of basic amino acids but contain the second one ( Fig.  2A). We hypothesized that this region in mammalian Plk1 might function as a bipartite NLS sequence.
To test this possibility, we first constructed a fusion protein between GST and a sequence corresponding to residues 124 -148 of Plk1 GST-Plk1(25 a.a.)) and co-injected this fusion protein and TRITC-BSA into the cytoplasm of HeLa cells. Whereas TRITC-BSA remained in the cytoplasm, the fusion protein translocated to and accumulated in the nucleus (Fig. 3A). Under the same conditions, GST alone did not translocate to the nucleus (Fig. 3A). The nuclear import of this fusion protein was almost completely suppressed by RanQ69L, a dominant negative form of Ran (Fig. 3A). Thus, a putative NLS region of Plk1 is able to function as a typical NLS that is dependent on the Ran/importin system. We then produced a fusion protein between GST and full-length Plk1 (Fig. 3B, GST-Plk1 WT) and injected it into the cytoplasm. Within 1 h after injection, GST-Plk1 WT translocated to and accumulated in the nucleus, whereas co-injected TRITC-BSA remained in the cytoplasm (Fig. 3B). This nuclear import of GST-Plk1 WT was also almost completely suppressed by RanQ69L. We next produced a GST fusion protein of a mutant Plk1 in which all of the seven arginine and lysine residues present in the identified NLS sequence were mutated to alanines (GST-Plk1 7A; see also Fig. 2). When injected into the cytoplasm, GST-Plk1 7A did not translocate to the nucleus and remained in the cytoplasm (Fig.  3B). These results indicate that the identified NLS sequence is required for Plk1 to enter the nucleus.
To assess the importance of the first and second clusters of basic amino acids in the identified NLS sequence of Plk1, we made several mutant forms of Plk1 in which the first, second, or both basic amino acid cluster(s) was mutated to alanines, respectively (see Fig. 2B). We then expressed HA-tagged forms of these mutants (3A, 4A, or 7A) and wild-type Plk1 in HeLa cells and examined their subcellular distribution by staining with anti-HA antibody. In more than half of the transfected cells, Plk1 WT accumulated in the nucleus, and in the remainder, Plk1 WT showed pan-cellular distribution (Fig. 4, WT). The mutation at the first basic cluster slightly affected the subcellular distribution. In ϳ40% of the transfected cells, Plk1 3A showed nuclear accumulation (Fig. 4, 3A). In contrast, nuclear accumulation was markedly impaired in the mutation at the second basic cluster. Plk1 4A showed pan-cellular distribution in ϳ80% of transfected cells, and in ϳ20% of the cells, it showed exclusively cytoplasmic localization (Fig. 4, 4A). The double mutation abolished completely nuclear localization. Plk1 7A showed exclusively cytoplasmic localization (Fig. 4,  7A). These results indicate that both the first and the second basic clusters contribute to nuclear translocation of Plk1, and therefore the identified NLS is in fact a bipartite NLS. In addition, the obtained data suggest that the second basic cluster alone can work as NLS although not so strong. It should be noted that the second basic cluster is conserved even in yeast orthologs of Plk1 (see Fig. 2A).
Expression of Plk1 7A Arrests the Cells in G 2 Phase-To know possible physiological significance of nuclear localization of Plk1 during interphase, we examined the affect of expression of the NLS-disrupted mutant of Plk1 (Plk1 7A) on the cell cycle progression. We injected a plasmid harboring Plk1 7A or Plk1 WT into the nucleus of HeLa cells at 4 h after release from double thymidine block, and we followed the cell cycle progression by examining the state of chromosomes. A pGFP empty vector was injected as a control. Although uninjected HeLa cells entered the M phase at ϳ10 h after the release from the double thymidine block, the injection procedure under our conditions appeared to induce a slight delay (from 1 to ϳ 2 h) in the cell cycle progression. Approximately half of the empty vectorinjected cells (ϳ56%) was in M phase at 12 h (Fig. 5B, Vector). The expression of Plk1 WT did not affect the timing of M phase entry significantly. 48% of the cells expressing Plk1 WT were in M phase at 12 h ( Fig. 5A and B, WT). In contrast, the expression of Plk1 7A seemed to block M phase entry. Only 5 to ϳ10% of the cells expressing Plk1 7A were in M phase at 12 h ( Fig. 5A and B, 7A), and the majority of the cells had a large G 2 -like nucleus and grew to cover a large surface area (Fig. 5A, 7A). At 14 h, there was no increase in M phase cells in the cells expressing Plk1 7A, whereas most of the cells expressing Plk1 WT or the pGFP vector-injected cells went through mitosis and entered the G 1 phase. Rather surprisingly, the cells expressing Plk1 7A did not enter the M phase and appeared to stay in G 2 phase even at 24 h after release (data not shown).
To examine whether the cells expressing Plk1 7A were arrested in G 2 , we first checked subcellular localization of cyclin B1, which normally translocates from the cytoplasm to the nucleus before nuclear envelope breakdown during prophase (23, 26 -29). At 12 h after the release, all of the Plk1 7A-injected cells (n ϭ 65) showed exclusively cytoplasmic localization of cyclin B1 (Fig. 6A), suggesting that the cells are arrested in G 2 . We then performed cell staining with anti-MPM2 antibody, which is believed to recognize mitotic phosphoproteins. Although ϳ50% pGFP-or Plk1 WT-injected cells showed strong staining with this antibody at 12 h, few cells expressing Plk1 7A (Ϲ4%) were MPM2-positive (Fig. 6B). This result also suggests that the cells expressing Plk1 7A did not enter the M phase. Finally, we assessed the activation state of M-phase promoting factor by examining the phosphorylation state of Cdc2 on Tyr-15. The cell staining with anti-phospho-Tyr-15 Cdc2 antibody demonstrated that although Ͼ60% pGFP-or Plk1 WT-injected cells were negative for this staining, only ϳ20% Plk1-7A-injected cells were negative (Fig. 6C). This suggests that Cdc2 was not activated in most of the cells expressing Plk1 7A, consistent with the idea that the cells are arrested in G 2 .
We have here demonstrated that Plk1 localizes to both the cytoplasm and the nucleus in addition to the centrosome in S and G 2 phases. Moreover, our results have identified a bipartite NLS that is responsible for nuclear localization of Plk1. Remarkably, the expression of an NLS-disrupted mutant of Plk1, Plk1 7A, during S phase is found to arrest the cells in G 2 phase. These results suggest that nuclear localization of Plk1 directed by its NLS during interphase is required for cells to enter mitosis. In a recent study in which the injection of anti-Plk1 antibodies into cultured cells arrested the cells in G 2 , Lane and Nigg (30) suggested the existence of a centrosomematuration checkpoint that is sensitive to the impairment of Plk1 function in cells (31). Similarly, we also hypothesize that there may be some checkpoint mechanism that senses nuclear localization of Plk1 or its interacting protein(s). Because the expression of Plk1 7A does not affect subcellular localization of endogenous Plk1, 2 it is possible that expression of Plk1 7A, which is exclusively cytoplasmic, may retain some interacting protein(s) in the cytoplasm by competing with endogenous Plk1 and thereby could trigger activation of an assumed checkpoint pathway, resulting in G 2 arrest. To examine whether localizing Plk1 7A in the nucleus rescues the G 2 arrest, we added an NLS sequence of SV40 large T-antigen (PKKKRKVEDP) to the C terminus of the Plk1 7A (Plk1 7A-SV40NLS) and injected the plasmid harboring Plk1 7A-SV40NLS into the nuclei of HeLa cells. Unfortunately, however, this mutant Plk1 did not accumulate strongly in the nucleus but showed pan-cellular distribution. In addition, the expression of this mutant also induced the G 2 arrest. 2 Therefore, this experiment could not determine propriety of our simple hypothesis. On the other hand, more complicated scenarios are also possible. We cannot completely exclude the possibility that Plk1 7A might not be activated properly during the G 2 to M phase transition, although Plk1 7A has essentially the same basal kinase activity as Plk1 WT. 2 The NLS of Plk1 described here is located within the subdomain V of the kinase catalytic domain. Only the sequence in Plk1 but not the sequence in the subdomain V of other protein kinases, matches the consensus sequence of NLS sequences. According to the predicted three-dimensional structure of the catalytic domain of Plk1 by the Swiss model, the NLS identified here forms a helix at the surface of the activation loop. Thus, the NLS may be well positioned for interactions with the nuclear import transporters.