Binding of Drosophila Polo kinase to its regulator Matrimony is noncanonical and involves two separate functional domains

Drosophila melanogaster Polo kinase physically interacts with, and is repressed by, the Matrimony (Mtrm) protein during oogenesis. Females heterozygous for a deletion of the mtrm gene display defects in chromosome segregation at meiosis I. However, a complete absence of Mtrm results in both meiotic catastrophe and female sterility. We show that three phosphorylated residues in an N-terminal region in Mtrm are required for Mtrm::Polo binding. However, this binding is noncanonical; it does not require either a complete S-pS/pT-P motif in Mtrm or key residues in the Polo-box domain of Polo that allow Polo to bind phosphorylated substrates. By using fluorescence cross-correlation spectroscopy to characterize the Mtrm::Polo interaction in vivo, we show that a sterile α-motif (SAM) domain located at the C terminus of Mtrm increases the stability of Mtrm::Polo binding. Although Mtrm’s C-terminal SAM domain is not required to rescue the chromosome segregation defects observed in mtrm/+ females, it is essential to prevent both meiotic catastrophe and the female sterility observed in mtrm/mtrm females. We propose that Polo’s interaction with the cluster of phosphorylated residues alone is sufficient to rescue the meiosis I defect. However, the strengthening of Mtrm::Polo binding mediated by the SAM domain is necessary to prevent meiotic catastrophe and ensure female fertility. Characterization of the Mtrm::Polo interaction, as well as that of other Polo regulators, may assist in the design of a new class of Polo inhibitors to be used as targeted anticancer therapeutic agents.Successful cell division requires the careful coordination of multiple processes, such as DNA replication, nuclear envelope breakdown (NEB), alignment and segregation of chromosomes on the spindle, and finally, cytokinesis. Temporal and spatial control of these events is partially regulated by Polo-like kinases, hereafter referred to as Polo, which are conserved from budding yeast (Cdc5) to humans (Plk1) (1–4). Polo is comprised of two functional domains—a canonical N-terminal serine/threonine kinase domain and a unique C-terminal Polo-box domain (PBD)—that are separated by a flexible linker region. The noncatalytic PBD consists of two Polo boxes, PB1 and PB2, which function as a single protein-binding unit.Studies in several organisms have identified a number of mutations within the PBD that abrogate its ability to selectively bind phosphorylated proteins containing the core consensus motif S-pS/pT-P/X (5–8). The classic “pincer mutant,” which fails to bind the core consensus motif S-pS/pT-P/X of Polo targets, was first identified as an H538A, K540M double-mutant residing within PB2 of human Plk1 (5, 6). (The equivalent mutations within the highly conserved PB2 of Drosophila Polo are H518A and K520M.) One example of this type of binding is the interaction between Plk1 and Bub1, a spindle checkpoint protein, in mitotic HeLa cells. This interaction requires the phosphorylation of Bub1 at a threonine within a conserved S-pT-P PBD binding motif and is critical for the proper localization of endogenous Plk1 to kinetochores during mitosis. The Polo::Bub1 interaction is also dependent on the PBD, as the Plk1 H538A, K540M pincer mutant is unable to coimmunoprecipitate (co-IP) with the Bub1 protein (8).However, it is becoming increasingly clear that Polo is also capable of interacting with regulatory proteins via noncanonical mechanisms (7). For example, in Drosophila S2 cells, Polo has been shown to robustly interact with the microtubule-associated protein Map205 during interphase of the cell cycle (9). Although a functional PBD and some structural elements of the Polo kinase domain, as well as a 162-aa region in Map205, are required for this interaction, it occurs by a phospho-independent mechanism (9). Although other examples of noncanonical Polo binding will be discussed below (Discussion), we focus here on the interaction between Drosophila Polo and its regulator Matrimony (Mtrm).Although the Mtrm protein was first shown to be a Polo binding protein by a global yeast two-hybrid (Y2H) interaction screen (10), the mtrm gene was identified in a genetic screen for genes that were haplo-insufficient with respect to the proper segregation of achiasmate homologs at the first meiotic division (11). Loss-of-function alleles of mtrm cause three distinct phenotypes: (i) high levels of achiasmate chromosome missegregation in mtrm/+ females, (ii) precocious NEB in mtrm/+ and mtrm/mtrm females, and (iii) sterility in females with a mtrm-null mutant background (12). The interaction between Mtrm and Polo was demonstrated genetically by the finding that reducing the dose of the polo⁺ gene to one copy fully suppressed the chromosome segregation defect seen in mtrm/+ females, while increasing the dose of the polo⁺ gene to three copies in females with two functional copies of mtrm⁺ induced defects in chromosome segregation (12). Xiang et al. (12) confirmed the physical interaction of Mtrm and Polo observed in Y2H studies with Drosophila oocytes by co-IP and proteomic analysis (12). These observations support a model in which Mtrm binding serves to inhibit the activity of Polo, and lend evidence to the idea that unbound Polo is deleterious to proper chromosome segregation during meiosis I.Although it remains to be shown biochemically, multiple lines of genetic evidence strongly suggest that Mtrm acts by inhibiting the activity of Polo kinase, and that Polo is likely to be the only target of Mtrm. First, as noted above, mutants in polo dominantly suppress the meiotic defects seen in mtrm/+ heterozygotes (12). Second, others have demonstrated a genetic interaction of Mtrm with other regulators of Polo, most notably, Scant (a dominant allele of greatwall) and endos (13, 14). Third, Von Stetina et al. (14) demonstrated that reducing polo⁺ dosage suppressed the decreased fertility observed in mtrm endos double heterozygotes. The authors also showed that increasing the dosage of polo⁺ to three copies resulted in almost complete sterility of mtrm heterozygous females (14). Fourth, in the proteomic studies performed by Xiang et al. (12) in which Mtrm was used as the bait protein, Polo was the only consistent high-affinity interactor.In Xiang et al. (12), immunofluorescence studies suggested that the amount of Mtrm was greatly decreased at the end of stage 12, coincident with NEB (12, 15). However, Von Stetina et al. (14) clearly showed by Western blot analysis that Mtrm levels substantially increase in stages 13 and 14. Indeed, we show below that the concentration of both Mtrm and Polo increases by approximately fivefold during the transition from stage 13 to stage 14. In this article we characterize both the role of Mtrm as a regulator of Polo at stages 13 and 14 and the nature of the physical Mtrm::Polo interaction.Although Xiang et al. (12) initially suggested that Mtrm and Polo interacted via a canonical interaction of the PBD of Polo with the S-pT-P motif of Mtrm, we use three techniques [co-IP from Drosophila ovaries, Y2H analysis, and fluorescence cross-correlation spectroscopy (FCCS) studies in living stage 13 and 14 oocytes] to demonstrate that the interaction of Mtrm and Polo is actually mediated by at least two highly conserved regions in Mtrm, an 18-aa N-terminal region that includes the S-pT40-P motif and a 61-aa C-terminal sterile α-motif (SAM) domain (Fig. 1A). The Mtrm::Polo interaction does not require the serine at the −1 position of the S-pT40-P motif, demonstrating that this interaction does not explicitly follow the canonical mechanism previously described (6).Open in a separate windowFig. 1.The binding of Mtrm to Polo is noncanonical. Residues in two distinct regions of Mtrm are required for Mtrm to fully bind Polo (A–C). The pincer residues in the PBD of Polo, which are generally essential for Polo to bind its substrates, are not required for full Mtrm::Polo binding. (D–F). (A) A schematic representation of the Mtrm protein depicting the two regions of conservation analyzed: the N-terminal region shown in yellow and the SAM domain shown in blue. (B) Coimmunoprecipitation was performed on FLAG-tagged overexpression transgenic lines expressing either full-length or a mutated form of the Mtrm protein. Both truncations, Mtrm^{Nterm-deletion} (lane 3) and Mtrm^SAM-deletion (lane 4) were unable to bind Polo. The Mtrm::Polo interaction was also disrupted for Mtrm^T40A (lane 6), Mtrm^S48A (lane 7), and Mtrm^S52A (lane 8); Mtrm^S39A (lane 5) and Mtrm^S137A (lane 9) showed Mtrm::Polo binding similar to Mtrm^FL (lane 2). Although Mtrm^S52A levels were reduced in this co-IP it showed higher expression in replicate experiments, and Polo binding was still decreased. Full genotypes of the transgenic lines were y w;transgene/+;nanosGAL4:VP16/+ and no-construct flies were y w;spa^pol. (C) FCCS was performed on mCherry-tagged Mtrm proteins with GFP-tagged Polo protein. Tagged constructs were expressed with their genomic promoters and flies were y w/w;mCherry-mtrm/GFP-polo. (D) A schematic representation of the Polo kinase protein depicting the kinase domain in green and the two PBDs in yellow and blue. (E) Co-IP was performed on GFP-tagged transgenic lines expressing either full-length or a mutant form of the Polo protein. The classic pincer mutant, polo^H518AK520M was able to bind Mtrm at levels comparable to polo^wt. Full genotypes of the transgenic lines were w;GFP-polo, and no-construct flies were y w;spa^pol. (F) FCCS was performed on mCherry-tagged Mtrm protein with GFP-tagged Polo proteins. Flies were y w/w;mCherry-mtrm/GFP-polo. Negative cross-correlation may be observed when diffusion of two species is not random, but some factor actively causes them to be segregated in space or time. However, in this case, cross-correlation values in select cases appear negative because of uncertainties in the measurement, as none of the negative cross-correlation values are statistically indistinguishable from the null hypothesis.Finally, we demonstrate that Polo and Mtrm use a noncanonical mechanism of interaction with regard to the PBD. Mutational ablation of the classic pincer residues within PB2 does not prevent Polo from binding Mtrm in co-IP experiments. Moreover, FCCS studies show that the Polo pincer mutant retains a strong ability to bind Mtrm in vivo, providing further evidence for noncanonical binding. We propose a model in which Polo’s interaction with the N-terminal cluster of three phosphorylatable Mtrm residues alone is sufficient to rescue the meiosis I defect, but strengthening of Mtrm::Polo binding, which is mediated by the SAM domain, is necessary to prevent meiotic catastrophe and ensure female fertility.