Spatiotemporal dynamics of Spc105 regulates the assembly of the Drosophila kinetochore

The formation of kinetochores shortly before each cell division is a prerequisite for proper chromosome segregation. The synchronous mitoses of Drosophila syncytial embryos have provided an ideal in vivo system to follow kinetochore assembly kinetics and so address the question of how kinetochore formation is regulated. We found that the nuclear exclusion of the Spc105/KNL1 protein during interphase prevents precocious assembly of the Mis12 complex. The nuclear import of Spc105 in early prophase and its immediate association with the Mis12 complex on centromeres are thus the first steps in kinetochore assembly. The cumulative kinetochore levels of Spc105 and Mis12 complex then determine the rate of Ndc80 complex recruitment commencing only after nuclear envelope breakdown. The carboxy-terminal part of Spc105 directs its nuclear import and is sufficient for the assembly of all core kinetochore components and CENP-C, when localized ectopically to centrosomes. Super-resolution microscopy shows that carboxy-terminus of Spc105 lies at the junction of the Mis12 and Ndc80 complexes on stretched kinetochores. Our study thus indicates that physical accessibility of kinetochore components plays a crucial role in the regulation of Drosophila kinetochore assembly and leads us to a model in which Spc105 is a licensing factor for its onset.

where XY is the level of the XY complex at kinetochore which is measured by fluorescently tagged Y protein. In this case we describe a two-step, sequential binding process; in the first step X is the Mis12 complex and Spc105 is Y and in the second step, X and Y are the Mis12 and Ndc80 complexes, respectively. The right hand-side of the equation can be rearranged based on the following considerations. First, the level of X not bound to Y at a kinetochore can be expressed as the difference between total X at the kinetochore (X T ) minus the complex: X=X t -XY. Secondly, assuming the free pool of Y (Y free ) is large and constant, it was grouped together with the association rate constant thus giving a first order association rate constant (k ass ' = k ass . Y free ). Because of sequential binding of molecules to the kinetochore, X T is not constant, but rather it is a monotonously increasing function of time. Therefore two extreme cases can be distinguished depending on the relative rate of change in X T increase and of Y binding to X. If Y binding to X is fast compared to the change in level of X T , then d XY/dt =0 and Equation 1 reduces to: 2 In this case, the level of Y at the kinetochore (XY) will be proportional to the total level of X (X T ). However, if Y binding is relatively slow compared to the change in total X at a kinetochore (X T ) then for small XY values: i.e. the rate of Y accumulation at kinetochores will be proportional to X T .
Based on these considerations, we analyzed the kinetics of kinetochore protein recruitment and Spc105 bind to each other very fast and that k diss << k ass '. In contrast, Ndc80 complex level (assessed as the mean value for its two equimolar components (Nuf2 and Mitch) at kinetochores does not show linear correlation with Mis12 complex (not shown). However, when we calculated the rate of accumulation of Ndc80 complex at kinetochores (d XY/dt), the slope of the blue curves from Fig.2a, and the level of Nsl1 (X T ) as a function of time, we found a strong correlation (Fig.2d). This suggests that Ndc80 binding to Mis12 is a slow and reversible process. The dissociation and the apparent first order association rate constants of their binding was estimated after rearranging Equation 3 by dividing both sides with XY: By plotting the specific rate of Ndc80 accumulation (1/XY . dXY/dt) as a function of Mis12 Ndc80 ratio gives straight with a slope and intercept corresponding to k ass ' and k diss values of 4.6 . 10 -3 and 1.8 . 10 -3 sec -1 (Fig.2e).

Supplementary Figures
5 Figure S1. Time frames of single nuclei in cleavage division cycle 12 from Mitch::EGFP and EGFP::Spc105 embryos. Maximum intensity projected z stacks were aligned in time to anaphase onset. The first time frame after anaphase onset was defined as 0 sec time point. In every 18 sec images were captured on two channels to detect EGFP fusion protein and rhodamine-conjugated tubulin. Circle at -342 sec shows the ROI (region of interest) used on green channel in every time point to quantify EGFP signal accumulation at the kinetochores.
ROIs with exactly the same coordinates were used on the red channel to define average rhodamine-tubulin signal intensity from interphase to anaphase onset (AO) in the nucleus until nuclear envelope break down (NEB) and around the chromosomes from NEB. In the same time interval average rhodamine-tubulin signal intensity in the cytoplasm was measured in the area between the two dotted line circles shown at -342 sec. After black bars between -180 sec and -162 sec the average rhodamine-tubulin signal intensity in the nucleus turns to be higher than in the cytoplasm. Scale bar represents 10 µm. Relative scale was defined for each nucleus separated as above in (e).
(h) The average of values from panel (c). Error bar represents standard deviation.
(i) Quantitative comparison of Spc105 kinetochore recruitment and accumulation into the nucleus/spindle matrix in mitosis 12. Maximal signal intensity in the nucleus and at the chromosomes after NEB represents the accumulation of GFP::Spc105 at the kinetochore (blue line). Maximum signal intensities in the nucleoplasm before NEB and in the spindle matrix after NEB represent the dynamics of Spc105 nuclear import (red line). Data indicate no delay in kinetochore accumulation after nuclear import. 9 Figure S3. Localization of Spc105 and Nnf1a/b throughout the cell cycle. Specificity of polyclonal antibodies against Spc105 and Nnf1a/b were determined by Western Blotting before (Przewloka et al., 2011). Fixed D-mel2 cells were stained for Spc105 or Nnf1a/b (in green), tubulin (in red), DNA (in blue) and CID (in greyscale, not merged). Spc105 localizes to the kinetochore in mitosis. Nnf1a/b is at the centromere throughout the entire cell cycle. Nnf1a/b signal intensity at the centromere is the weakest in metaphase.
Spc105-C over expression causes developmental arrest and lethality before cellularization of the embryo.
(a) Elongated anaphase spindle with non segregating chromosomes and with DNA fragments in the spindle poles are characteristic for early arresting embryos (vast majority of the embryos).
(b) Characteristic phenotypes in the embryos which completed the first mitotic divisions: centrosome detachment in mitosis and spindles collapse.
(c) Terminally arrested phenotype -enlarged cortical region (d) Terminally arrested phenotype in full view of an embryo with late stage developmental arrest.
Most embryos arrested before mitosis 10 when Spc105-C expression was driven by mat-tub-Gal4. When expression was weaker, under nanos-Gal4 control, embryos mainly arrested in mitosis 10-13 (Video 6 is also of such an embryo).
13 Figure S5. Quantification of intrakinetochore distances and the response to loss of tension.
(a) Maximum intensity projection of 38 z-sections, acquired by the OMX microscope in SI mode about a D-mel2 cell in metaphase. The equatorial plane of the mitotic spindle is perpendicular to the focal plane of the objective. CID in red, Spc105-N in green and DNA in blue. Scale bar represents 5 μm, z-step were 125 nm.
(b) Maximum intensity projection of 3 z-slides from the marked box area of (a). The middle slide has the highest signal intensities both for CID and Spc105-N. Scale bar represents 500 nm.