Identification of a consensus motif for Plk (Polo-like kinase) phosphorylation reveals Myt1 as a Plk1 substrate.

Plk1 (Polo-like kinase 1), an evolutionarily conserved serine/threonine kinase, is crucially involved in multiple events during the M phase. Here we have identified a consensus phosphorylation sequence for Plk1, by testing the ability of systematically mutated peptides derived from human Cdc25C to serve as a substrate for Plk1. The obtained results show that a hydrophobic amino acid at position +1 carboxyl-terminal of phosphorylated Ser/Thr and an acidic amino acid at position -2 are important for optimal phosphorylation by Plk1. We have then found that Myt1, an inhibitory kinase for MPF, has a number of putative phosphorylation sites for Plk1 in its COOH-terminal portion. While wild-type Myt1 (Myt1-WT) served as a good substrate for Plk1 in vitro, a mutant Myt1 (Myt1-4A), in which the four putative phosphorylation sites are replaced by alanines, did not. In nocodazole-treated cells, Myt1-WT, but not Myt1-4A, displayed its mobility shift in gel electrophoresis, due to phosphorylation. These results suggest that Plk1 phosphorylates Myt1 during M phase. Thus, this study identifies a novel substrate for Plk1 by determining a consensus phosphorylation sequence by Plk1.

The activity of MPF is regulated by phosphorylation and dephosphorylation of Cdc2 and accumulation of cyclin B protein (24,25). Until the end of G 2 phase, Cdc2, in higher eukaryotes, remains inactive through inhibitory phosphorylation on Thr-14 and Tyr-15. At M phase entry, the Cdc25 phosphatase dephosphorylates Thr-14 and Tyr-15, thereby activating MPF (26 -28). Wee1 and Myt1 are responsible for such inhibitory phosphorylation of Cdc2. Wee1, a nuclear protein, is capable of phosphorylating Tyr-15 of Cdc2, but not Thr-14 (30 -33). Myt1 is a membrane-associated, dual-specific protein kinase that phosphorylates both Thr-14 and Tyr-15 of Cdc2 (34 -37). Myt1 is shown to be hyperphosphorylated during M phase, which is coincident with its inactivation (35,37). It is reported that p90 rsk and Akt can phosphorylate and down-regulate Myt1 during miosis in Xenopus and Asterina oocytes, respectively (38,39). However, a kinase(s) responsible for the regulation of Myt1 in the somatic cell cycle has been unknown.
In this study, we have identified a consensus sequence for Plk1 phosphorylation and found Myt1 a Plk1 substrate.
Cell Culture and Synchronization-HeLa cells were cultured in Dulbecco's modified Eagle's medium with 10% bovine calf serum. Cells were synchronized with a double-thymidine block. Exponentially growing cells were arrested in S phase by treatment with thymidine (2 mM) for 17 h and were released from the arrest by washing twice with fresh medium. Cells were grown in fresh medium for 9 h and then re-treated with thymidine (2 mM) for 15 h.
Kinase Assays-In the kinase assay for His-tagged Plk1 or immunoprecipitates, 0.5-1 g of His-tagged Plk1 or immunoprecipitates were mixed with substrate (0.1-3 g), 50 M ATP, and 15 mM MgCl 2 in a final volume of 15 l and incubated for 20 min at 30°C in the presence of 3 Ci of [␥-32 P]ATP. The reactions were stopped by addition of Laemmli's sample buffer and boiling. Histone H1 kinase assay was conducted as described previously (40).
Transfection and Immunoblotting-HeLa cells were transiently transfected by the use of FuGENE6 according to the manufacturer's instructions. To arrest cells, cells were treated with 250 ng/ml nocodazole or 2 mM thymidine for 18 h at 20 h after transfection. Cells were lysed in buffer B (50 mM Tris (pH 8.0), 100 mM NaCl, 5 mM EDTA, and 0.5% Nonidet P-40, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , and 2 g/ml aprotinin) (41) and centrifuged at 20,000 ϫ g for 15 min. The cell extracts were subjected to immunoblotting with anti-Myc (Santa Cruz) or anti-HA antibody (Santa Cruz). 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.

Identification of a Consensus Motif for Plk1
Phosphorylation-We tested the ability of systematically mutated peptides derived from human Cdc25C to serve as a substrate for Plk1 in vitro. We have recently shown that human Cdc25C is phosphorylated on Ser-198 by Plk1 (9). Then, we used GST-fused Cdc25C peptides (residues 173-206) as test substrates. We constructed various GST-fused Cdc25C peptides (173-206), in which amino acids in residues 192-203 are mutated, to examine which amino acids surrounding Ser198 are important for the phosphorylation by Plk1.
First, we replaced each amino acid surrounding Ser-198 by Ala or Gly. The obtained results showed that the ability of the peptides to serve as a substrate for Plk1 was markedly reduced when Glu-196, Leu-199, or Asp-201 was replaced by Ala or Gly (Fig. 1A). When Lys-200 was mutated, the resultant peptides were phosphorylated more efficiently (Fig. 1A). These results suggest that amino acid residues at positions Ϫ2 to ϩ3 of the phosphorylated residue (Ser-198) are primarily important for phosphorylation by Plk1.
When we performed a single amino acid exchange at Glu-196, a peptide with Asp-196 was phosphorylated as efficiently as the original peptide, whereas other peptides with Leu-, Gln-, Lys-, Ala-, or Gly-196 were poorly phosphorylated (Fig. 1B). This suggests that an acidic amino acid at position Ϫ2 is important for optimal phosphorylation. Replacement of Phe-197 by Ala, Glu, Leu, Arg, or Lys did not affect significantly the efficiency of phosphorylation, while replacement by Gly reduced the efficiency of phosphorylation (Fig. 1C). Thus, Gly at position Ϫ1 is inhibitory for the phosphorylation. Replacement of Leu-199 by a hydrophobic amino acid such as Val, Ile, Phe, Trp, or Met did not decrease, or rather

Identification of a Consensus Motif for Plk1 Phosphorylation 25278
increase, the phosphorylation, whereas replacement by Pro, Arg, Glu, Gln, Ala, Gly, or Lys significantly decreased the phosphorylation (Fig. 1D). This indicates the significance of a hydrophobic amino acid at position ϩ1. Replacement of Lys-200 by Ala or Glu significantly increased the phosphorylation, while replacement by Arg or Gly did not significantly affect the phosphorylation efficiency (Fig. 1E), suggesting that a basic amino acid at position ϩ2 is slightly inhibitory. Replacement of Asp-201 by Glu did not affect the phosphorylation, while replacement by Ala or Gly decreased the phosphorylation (Fig. 1F). Thus, an acidic amino acid at position ϩ3 is important for optimal phosphorylation. When the target Ser-198 was mutated into Thr, the efficiency of phosphorylation did not change markedly (Fig. 1G), suggesting that Thr as well as Ser is able to be phosphorylated by Plk1.
To examine the importance of an acidic amino acid at position Ϫ2 in more detail, we made four new peptide sequences in which residues 195-197 were Ala-Ala-Ala, Glu-Ala-Ala, Ala-Glu-Ala, or Ala-Ala-Glu as shown in an upper panel of Fig. 2A.
Only the sequence Ala-Glu-Ala was phosphorylated efficiently ( Fig. 2A, lower). Therefore, an acidic amino acid should locate at position Ϫ2 for optimal phosphorylation. To examine the importance of a hydrophobic amino acid at position ϩ1, we made four new peptide sequences in which residues 199 -201 were Ala-Ala-Ala, Leu-Ala-Ala, Ala-Leu-Ala, or Ala-Ala-Leu as shown in an upper panel of Fig. 2B. Only the sequence Leu-Ala-Ala was phosphorylated efficiently (Fig. 2B, lower), showing that a hydrophobic amino acid should locate at position ϩ1 for optimal phosphorylation. Finally, to examine the importance of an acidic amino acid at position ϩ3, we made four new peptide sequences in which residues 199 -202 were Leu-Ala-Ala-Ala, Leu-Asp-Ala-Ala, Leu-Ala-Asp-Ala, or Leu-Ala-Ala-Asp as shown in an upper panel of Fig. 2C. The sequence Leu-Ala-Asp-Ala was phosphorylated most efficiently among these peptides (Fig. 2C, lower). Thus, when an acidic amino acid locates at position ϩ3, the sequence is optimal for phosphorylation by Plk1. All these results taken together suggest a sequence D/E-X-S/T-⌽-X-D/E (X, any amino acid; ⌽, a hydrophobic amino acid) as an optimal phosphorylation sequence by Plk1 (Fig. 2D).
Phosphorylation of Myt1 by Plk1 in Vitro-We noted that Myt1, a negative regulator of MPF (cdc2/cyclin B), has multiple putative phosphorylation sites for Plk1 on its C-terminal region (Fig. 3A). Because Myt1 was reported to be highly phosphorylated during M phase (35,37), we hypothesized that Myt1 could be a substrate of Plk1. We fused an NH 2 -terminal region (residues 1-377) or a COOH-terminal region (residues 401-499) of Myt1 to GST and used these bacterially produced fusion proteins. As shown in Fig. 3B, recombinant His-Plk1 efficiently phosphorylated the COOH-terminal region of Myt1, but not the NH 2 -terminal region. Phosphorylation seen in lanes 1 and 2 results from autophosphorylation of the GST-Myt1(N), as a kinase-dead form of GST-Myt1(N), GST-Myt1(N,KD), did not undergo phosphorylation upon incubation with His-Plk1 ( lanes  3 and 4). There are four possible phosphorylation sites (Ser-426, Ser-435, Ser-469, and Thr-495) for Plk1 in the COOHterminal region of Myt1 (see Fig. 3A). When a mutant form of Myt1 (GST-Myt1(C) 4A), in which the four putative phosphorylation sites were replaced by Ala, was tested for phosphorylation by Plk1, it was not phosphorylated by Plk1 (Fig. 3C). HA-tagged Plk1, which was expressed in HeLa cells and purified by immunoprecipitation with anti-HA antibody, was also able to phosphorylate GST-Myt1(C) WT, but not GST-Myt1(C) 4A (Fig. 3D). A kinase-dead form of Plk1 (HA-Plk1 KD) did not phosphorylate

Identification of a Consensus Motif for Plk1 Phosphorylation 25279
GST-Myt1(C) WT at all (Fig. 3D). Thus, Plk1 phosphorylates Myt1 in vitro on several or all of these four residues. Phosphorylation of Myt1 during G 2 -M Phase in HeLa Cells-Endogenous Plk1 was immunoprecipitated from synchronized HeLa cells with anti-Plk1 antibody and tested for the ability to phosphorylate GST-Myt1(C). Phosphorylation of Myt1(C) WT by immunoprecipitated Plk1 was increased during G 2 -M phase, i.e. 9 -11 h after release from a double thymidine block (Fig.  4A). GST-Myt1(C) 4A was not phosphorylated at all (Fig. 4A). Thus, Plk1 in G 2 -M phase is able to phosphorylate Myt1.
To examine whether Plk1 is able to phosphorylate Myt1 in intact cells, we co-expressed Myc-tagged full-length Myt1 with HA-tagged Plk1. We performed immnunoblotting analysis, as Myt1 is shown to display a mobility shift upon phosphorylation (35,37). A mobility-shifted band of Myt1 was detected when wild-type Plk1 was co-expressed (Fig. 4B, Myc-Myt1 WT/HA-Plk1 WT). When kinase-dead Plk1 was co-expressed, the mobility shift of Myt1 did not occur (Fig. 4B, HA-Plk1 KD). Importantly, the mobility shift of Myt1 was not observed when a mutant form of Myt1, Myt1 4A, was expressed (Fig. 4B, Myc-Myt1 4A).
Previous reports have shown that Myt1 is phosphorylated during G 2 -M and displays mobility shifts (35,37,39). To test whether or not phosphorylation of Myt1 on the putative Plk1 phosphorylation sites occurs during G 2 -M phase, we expressed Myc-tagged Myt1 WT or Myt1 4A in HeLa cells and treated the cells with thymidine or nocodazole to arrest cells in S phase or M phase, respectively. Immunoblotting showed that in M phase-arrested cells, but not in S phase-arrested cells, Myt1 WT displayed mobility-shifted bands, whereas Myt1 4A did not (Fig. 4C). When Plk1 was specifically depleted in cells by siR-NAs, the mobility-shifted bands of Myc-Myt1 WT seen in M phase-arrested cells were markedly reduced (Fig. 4D). These results suggest that part of the phosphorylation of Myt1 during G 2 -M phase is mediated by Plk1.
Plk1 Phosphorylation Sites of Myt1-To examine on which sites Myt1 is phosphorylated by Plk1, we constructed several forms of Myt1, in which one, two, or three of the four putative Plk1 sites was replaced by alanines, and co-expressed them with Plk1. Immunoblotting showed that Myt1 4A, Myt1 426A-469A-495A, or Myt1 426A-495A did not display a markedly shifted band, whereas Myt1 495A displayed the shifted band significantly and Myt1 426A slightly when co-expressed with wild type Plk1 (Fig. 5A). These results suggest that Ser-426 is a major phosphorylation site by Plk1, and Thr-495 is a second major site. When HeLa cells expressing these constructs were arrested in M phase by nocodazole treatment, Myt1 4A, Myt1 426A-469A-495A, Myt1 426A-495A, or Myt1 426A did not display markedly shifted band, whereas Myt1 495A displayed the shifted band to almost the same extent as Myt1 WT did (Fig.  5B), suggesting that phosphorylation of Myt1 by Plk1 on Ser-426 occurs during M phase. We then tested the ability of these mutant forms of Myt1 (GST fusion proteins), to serve as a substrate for Plk1 in vitro. Quantification of the result (Fig. 5C) showed that Ser-426 is the major phosphorylation site by Plk1 in vitro and Thr-495 the second major site. This conclusion is identical to that obtained from the co-expression experiment (Fig. 5A).
In this study, we have identified a consensus motif for Plk1 phosphorylation. Our results show that a hydrophobic amino acid at position ϩ1 and an acidic amino acid at position Ϫ2 are important for optimal phosphorylation. The reported autophosphorylation sites on Plx1 (a Xenopus homolog of Plk1) (42) and the identified phosphorylation site of Scc1 by Cdc5 (a yeast homolog of Plk1) (12) match this optimal motif. This consensus sequence can be used for identification of novel targets for Plk1.
A previous report showed that inhibition of Plx1 by the injection of anti-Plx1 antibody impaired the mobility shift of Myt1 in the system of Xenopus cycling extracts (6), suggesting that the activity of Plx1 is required for phosphorylation of Myt1 in this system. Here, we have shown that Plk1 is responsible for part of the phosphorylation of Myt1 during M phase. The kinase activity of human Myt1 is reported to be decreased during M phase, and the decreased activity correlates with hyperphosphorylated forms of Myt1 (35,37). Previously, Myt1 was shown to be phosphorylated by Cdc2, but this phosphorylation did not decrease the kinase activity of Myt1 (37). Most recently, p90 rsk and Akt are reported to phosphorylate and down-regulate Myt1 at the onset of meiosis in Xenopus and Asterina oocytes, respectively (38,39). However, a kinase(s) responsible for the regulation of Myt1 during M phase in somatic cell cycles has not been fully identified. Functional consequences resulting from Plk1-mediated phosphorylation of Myt1 should be elucidated in the future studies.