The KLP-7 Residue S546 Is a Putative Aurora Kinase Site Required for Microtubule Regulation at the Centrosome in C. elegans

Abstract

Regulation of microtubule dynamics is essential for many cellular processes, including proper assembly and function of the mitotic spindle. The kinesin-13 microtubule-depolymerizing enzymes provide one mechanism to regulate microtubule behaviour temporally and spatially. Vertebrate MCAK locates to chromatin, kinetochores, spindle poles, microtubule tips, and the cytoplasm, implying that the regulation of kinesin-13 activity and subcellular targeting is complex. Phosphorylation of kinesin-13 by Aurora kinase inhibits microtubule depolymerization activity and some Aurora phosphorylation sites on kinesin-13 are required for subcellular localization. Herein, we determine that a C. elegans deletion mutant klp-7(tm2143) causes meiotic and mitotic defects that are consistent with an increase in the amount of microtubules in the cytoplasmic and spindle regions of meiotic embryos, and an increase in microtubules emanating from centrosomes. We show that KLP-7 is phosphorylated by Aurora A and Aurora B kinases in vitro, and that the phosphorylation by Aurora A is stimulated by TPXL-1. Using a structure-function approach, we establish that one putative Aurora kinase site, S546, within the C-terminal part of the core domain is required for the function, but not subcellular localization, of KLP-7 in vivo. Furthermore, FRAP analysis reveals microtubule-dependent differences in the turnover of KLP-7(S546A) and KLP-7(S546E) mutant proteins at the centrosome, suggesting a possible mechanism for the regulation of KLP-7 by Aurora kinase.

Fig 1. KLP-7 locates to female meiotic spindles and limits MT growth near the spindle and cortex.

(A) A western blot probed for tubulin and KLP-7 demonstrates that KLP-7 is not detected in klp-7(tm2143) worm lysate. (B) Wild-type (upper panels, also S1 Movie) and klp-7(tm2143) (lower panels, also S2 Movie) meiotic spindles during meiosis I and II are displayed. GFP::tubulin is green and mCherry::histone is magenta. In klp-7(tm2143) embryos, many long MTs were visible near the meiotic spindle (arrowhead), and chromosome scattering was often observed after anaphase. Scale bar is 5 μm. (C) MTs were more prevalent near the cortex in meiotic klp-7(tm2143) embryos. The average ratio of cortical GFP::tubulin to a central region (a:b) in metaphase embryos is shown for WT (n = 16) and klp-7(tm2143) (n = 24). Error bars are SEM. (D) Wild-type embryos were fixed and stained to visualize MTs (green), KLP-7 (red) and DNA (blue). KLP-7 associates with chromosomes and spindle poles. Higher magnification insets show a higher concentration of KLP-7 at the region between meiotic chromosomes in metaphase I and II (arrows), and to the spindle midzone between the separating chromosomes in anaphase II (arrowhead). Scale bars are 5 μm.

Fig 2. klp-7(tm2143) embryos exhibit increased levels of centrosomal MTs and a spindle-break phenotype during mitotic anaphase.

(A) Representative images of wild-type and klp-7(tm2143) embryos expressing GFP::tubulin (green), mCherry::histone (magenta), in mitotic metaphase. Bar is 10 μm. (B) Quantification of centrosomal MTs in wild-type (n = 8) and klp-7(tm2143) (n = 12) embryos. The integrated intensity of GFP::tubulin per centrosome was measured. Levels are displayed relative to the average intensity of wild-type centrosomes (Error bars are SEM). (C) Still images of mitotic embryos from a time-lapse movie are shown. Top panel: wild-type embryo expressing GFP::γ-tubulin, GFP::histone (S3 Movie). Middle panel: klp-7(tm2143) embryo expressing GFP::γ-tubulin, mCherry::histone (S4 Movie). Lower panel: klp-7(tm2143) embryo expressing GFP::KLP-7(WT) (S5 Movie).

Fig 3. gpr-1/2 RNAi increases spindle midzone MT levels in klp-7(tm2143) embryos.

(A) Shown are representative images of wild-type, gpr-1/2(RNAi), klp-7(tm2143) and klp-7(tm2143); gpr-1/2(RNAi) embryos at anaphase of mitosis. GFP::tubulin is white and mCherry::histone is magenta. (B) Quantification of GFP::tubulin fluorescence (arbitrary units, relative to WT) within a rectangular region at the midzone (inset) are shown for wild-type (n = 8), gpr-1/2(RNAi) (n = 9), klp-7(tm2143) (n = 12) and klp-7(tm2143); gpr-1/2(RNAi) (n = 12) embryos. Bar is 10 μm.

Fig 4. KLP-7 is phosphorylated in vivo.

(A) A western blot analysis of wild-type and phosphatase-treated lysates separated by Phos-tag SDS PAGE, probed with anti-KLP-7 antibodies. (B) Phos-tag western blots of wild-type and kinase knockdown lysates, probed with anti-KLP-7. Twenty RNAi-treated young adult hermaphrodites were loaded into each lane. (C) 2D gel electrophoresis of KLP-7 in wild-type (top), air-1(RNAi) (middle) and air-2(or207ts) (bottom) lysates, followed by immunoblotting with anti-KLP-7 are shown. The pH gradient is indicated. Putative phospho-isoforms that display differences in wild-type and kinase knockdown lysates are shown with white brackets.

Fig 5. KLP-7 is phosphorylated by AIR-1 and AIR-2 kinases in vitro.

(A) Shown is a schematic of the KLP-7 protein, with Aurora kinase phosphorylation sites (gray) predicted by GPS [39]. Fragments used for in vitro kinase reactions are indicated below. (B-E) AIR-1 or AIR-2 kinases were incubated with different substrates in vitro and γ[32P]-ATP incorporation was determined via SDS-PAGE and subsequent autoradiography. (B) AIR-2 kinase with coactivator ICP-1 phosphorylates KLP-7 within the N-terminus. (C-D) TPXL-1 N-terminal fragment (aa 1–63; [48]) stimulates AIR-1 phosphorylation of KLP-7 within the N-terminus and (D) the C-terminus. (E) AIR-2 kinase with coactivator ICP-1 phosphorylates KLP-7 within the C-terminus. KD = kinase dead; Ponceau-S stained membranes for B-E are provided in S3 Fig.

Fig 6. KLP-7 residue S546 is required for protein function but not subcellular localization.

(A) KLP-7 proteins mutated at different predicted Aurora kinase phosphorylation sites localize at centrosomes and kinetochores. Images of metaphase embryos expressing GFP::KLP-7(WT) or GFP::KLP-7(phospho-mutant) transgenes are shown. (B) Different versions of KLP-7-GFP were expressed in klp-7(tm2143) worms. The maximum velocity of the posterior centrosome during the first 40 seconds of mitotic anaphase was determined; also see S7 Fig. The relevant alteration to the klp-7 transgene is listed. Average maximum velocity was determined from GFP-KLP-7(WT), klp-7(tm2143), GFP-KLP-7(S546A), GFP-KLP-7(S546E), GFP-KLP-7(T182E), and GFP-KLP-7(T182S539ES546E). klp-7(tm2143) control worms expressed GFP::γ-tubulin and GFP::histone to enable centrosome tracking; all other strains were tracked with the GFP::KLP-7 signal. GFP::KLP-7(S546E) and GFP::KLP-7(S546A) transgenes failed to rescue klp-7(tm2143). P-values for WT vs experimental data were: klp-7(tm2143) P = 0.002; GFP-KLP-7(S546A) P = 0.017; GFP-KLP-7(S546E) P = 0.030; GFP-KLP-7(T182E) P = 0.498; GFP-KLP-7(T182S539ES546E) P = 0.135. ** (0.0005<P<0.005); * (0.005<p<0.05), Two-tailed Student’s t-test. (C) Similar to klp-7(tm2143), GFP::KLP-7(S546E) and GFP::KLP-7(S546A) embryos exhibited an increase in centrosomal MTs, relative to WT. At least 14 centrosomes were analyzed for each strain. **(0.0005<P<0.005); *(0.005<p<0.05), Two-tailed Student’s t-test. (D) S546 is within a conserved Aurora kinase motif (underlined).

Fig 7. Fluorescence recovery after photobleaching of centrosomal GFP::KLP-7.

(A) An example of fluorescence recovery of GFP::KLP-7 at the centrosome is shown. One centrosome (left) was photobleached and the fluorescence intensity was measured from images acquired at one second intervals. Bar: 10 μm. (B) Centrosomal recovery rates (in arbitrary units) are displayed for the first 50 sec for WT, S546A and S546E versions of GFP::KLP-7. Error bars are SEM (n ≥ 5). (C) Pre-treatment of the embryos with nocodazole revealed a slower recovery time for WT and S546A, compared to untreated embryos. S546E recovery is similar in nocodazole-treated and untreated samples. Normalized fitted recovery curves for B and C are shown below; also see S8 Fig. Insets display the calculated t(1/2) and mobile fractions (M.F.). (D) A model for two phases of KLP-7 behaviour related to the phosphorylation state of S546. Phosphorylated KLP-7 (green) exchanges rapidly between centrosomes and cytoplasm, by a MT-independent mechanism. Dephosphorylated KLP-7 (magenta) locates to the centrosome but must undergo phosphorylation (1) and subsequent dephosphorylation (2) before depolymerizing MTs and exiting the centrosome. The phosphorylation of KLP-7 at the centrosome could occur via TPXL-1-activated Aurora kinase A (AIR-1), based on in vitro data and data from Ozlü et al., [48].