PerspectiveSignal Transduction

Capturing Polo Kinase

See allHide authors and affiliations

Science  21 Feb 2003:
Vol. 299, Issue 5610, pp. 1190-1191
DOI: 10.1126/science.1082384

The reversible addition of phosphate groups (phosphorylation) to proteins is one of the principal ways in which cells regulate protein activity. Phosphorylation not only directs the allosteric regulation of enzyme activity, but also is important for controlling protein-protein interactions, particularly those that assemble the protein complexes of signaling pathways. Phosphorylation creates docking sites in the phosphorylated protein to which other proteins then bind through specific phosphopeptide-binding domains. Prominent examples of phosphopeptide-binding domains include SH2 and PTB domains, which interact with phosphotyrosine motifs, and Forkhead-associated and WW (“double tryptophan”) domains, which interact with phosphoserine (pS) and phosphothreonine (pT) motifs (1, 2). On page 1228 of this issue, Elia and co-workers (3) describe a clever proteomics screen that they use to identify a new phosphopeptide-binding domain specific for pS and pT motifs. This new domain, the Polo box domain (PBD), was identified in Polo-like kinase 1 (Plk1), a kinase involved in the onset of mitosis. The PBD module appears to be specific to members of the PLK family and may help to regulate both the activity and subcellular localization of these kinases.

The novelty of Elia et al.'s approach is their use of an immobilized library of degenerate phosphopeptides to screen protein pools (prepared by in vitro translation of a cDNA library) for interacting proteins. Any protein identified can then be used in turn to establish the phosphopeptide consensus that provides optimal binding. First, Elia and co-workers searched for proteins that would bind, in a phosphorylation-dependent manner, to substrates of proline-directed kinases, such as cyclin-dependent kinases (CDKs), which are key players in progression through the cell cycle, and mitogen-activated protein (MAP) kinases, which are components of many signaling pathways. The investigators discovered two proteins that bound to a phosphothreonine-proline (pT-P) peptide library but not to a corresponding library of dephosphorylated peptides. The first protein was Pin1, a WW domain protein known to interact with pT-P motifs; the second protein was the carboxyl-terminal domain of human Plk1. Acting in concert with Cdk1, the serine/threonine protein kinase Plk1 is an important regulator of mitosis (4, 5). It comprises an amino-terminal catalytic kinase domain and a carboxyl-terminal regulatory region containing two conserved motifs (the Polo boxes). The PBD in Plk1's noncatalytic domain (residues 326 to 603) contains both Polo boxes and a poorly conserved region amino-terminal to these motifs. The PBD of Plk1 appears to be considerably larger than other phosphorylation site-binding domains, which typically comprise about 75 to 150 residues, although further work might reveal a smaller minimal binding motif. This binding domain is also unusual in that it appears to be restricted to the PLK family, whereas other phosphopeptide-binding modules are found in a wide variety of proteins.

PLK recruitment to phosphorylated proteins.

The carboxyl-terminal region of Plk1 comprises a phosphopeptide-binding domain containing two Polo boxes (PBD). This domain enables Plk1 to be recruited to different subcellular structures or substrates (docking proteins) after their phosphorylation by one or more priming kinases. For example, the Plk1 substrate Cdc25 is first phosphorylated by the priming kinase Cdk1 on its ST/SP/X motif; this provides a binding site for Plk1, which then also phosphorylates Cdc25 at distinct sites. The PBD domain may also autoregulate Plk1 activity. CAT, catalytic domain.

To identify phosphopeptides with the highest affinity for Plk1, the investigators screened multiple phosphopeptide libraries using PBD. They defined the consensus motif S-pT/pS-P/X in potential interacting partners as an optimal binding site, with pT conferring a sevenfold increase in affinity over pS. The observed strong selection of a serine in position −1 was unexpected. Even more surprising was the finding that phosphopeptide binding to the PBD did not strictly require a proline in the +1 position, although proline conferred a slight advantage. This suggested that the generation of a PBD-binding site does not necessarily require a proline-directed kinase. Nonetheless, two experiments demonstrate the ability of PBD to bind to pS/pT-P-containing phosphoproteins. First, the PBD of Plk1 interacted with several proteins that are also recognized by the MPM-2 monoclonal antibody, which binds to the pS/pT-P motif. Second, and more relevant from a physiological perspective, Plk1 binds to an optimal PBD binding site in the human Cdc25 phosphatase, a key enzyme responsible for the dephosphorylation (and concomitant activation) of Cdk1 at the onset of mitosis. Both Cdk1 and Plk1 have previously been implicated in phosphorylating and activating Cdc25 and most likely cooperate in a positive feedback loop (6, 7). In light of the new data, it is attractive to envision that Cdk1 creates a PBD docking site on Cdc25, leading to the recruitment of Plk1 and further activation of Cdc25, culminating in the onset of mitosis.

It has been suggested that Polo boxes are involved in the targeting of PLKs to subcellular structures (8, 9). By showing that injection of phosphopeptide competitors prevents the association of a PBD fusion protein with the centrosome, Elia et al. provide direct support for such a targeting activity (3). In addition, it has been proposed that the carboxyl-terminal domain of Plk1 is required for autoinhibition of the catalytic amino-terminal domain (4, 9). This is reminiscent of the inhibition of Src kinase family members by their own SH2 domains (2). In Src kinases, tyrosine phosphorylation regulates the interactions of SH2 domains with phosphotyrosine motifs, but the mechanisms controlling purported PBD interactions with the catalytic domain of Plk1 remain to be clarified. As yet, there is no evidence for inhibitory phosphorylation of Plk1 that could provide a docking site for the PBD. However, as the presence of amino-terminal Plk1 sequences reduces the affinity of the PBD for phosphopeptides, the catalytic domain and phosphopeptides may compete for binding to the PBD. Furthermore, Plk1 is activated through phosphorylation of threonine residue 210, and this prevents binding of the carboxyl-terminal domain (9). Thus, this phosphorylation could be a prerequisite for a conformational change that liberates the PBD for phosphopeptide binding.

The study by Elia and co-workers raises several interesting issues. First, clearly it will be interesting to apply this screening approach to peptide libraries modeled after consensus sites for other kinases (or other protein-modifying enzymes). Second, it will be important to identify the proteins that provide physiological docking sites for Plk1 in different cellular locations (that is, at the centrosome, the kinetochore, and the spindle midzone) (10). These proteins may themselves constitute substrates for Plk1 or facilitate the phosphorylation of neighboring proteins (see the figure). Of equal importance will be the identification of priming kinases that generate PBD binding sites. Moreover, if PBDs are phosphopeptide-binding domains in all PLK family members including the four mammalian PLKs—Plk1, Snk (Plk2), Fnk (Plk3), and Sak (Plk4)—then which structural features confer distinct targeting properties on different PBDs? This question is particularly intriguing in the case of Sak (Plk4), as its recently solved crystal structure reveals only a single Polo box (11). Finally, detailed structural information on the interactions between PBDs and phosphopeptides should not only afford insights into PLK activities but might also set the stage for designing small molecules that target these important cell cycle regulatory kinases.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article