|Home | About | Journals | Submit | Contact Us | Français|
Tension sensing of bi-oriented chromosomes is essential for the fidelity of chromosome segregation. The spindle assembly checkpoint (SAC) conveys lack of tension or attachment to the anaphase promoting complex. Components of the SAC (Bub1) phosphorylate histone H2A (S121) and recruit the protector of cohesin, Shugoshin (Sgo1) to the inner centromere. How the chromatin structural modifications of the inner centromere are integrated into the tension sensing mechanisms and the checkpoint are not known.
We have identified a Bub1/Sgo1 dependent structural change in the geometry and dynamics of kinetochores and the pericentric chromatin upon reduction of microtubule dynamics. The cluster of inner kinetochores contract while the pericentric chromatin and cohesin that encircle spindle microtubules undergo a radial expansion. Despite its increased spatial distribution, the pericentric chromatin is less dynamic. The change in dynamics is due to histone H2A phosphorylation and Sgo1 recruitment to the pericentric chromatin, rather than microtubule dynamics.
Bub1 and Sgo1 act as a rheostat to regulate the chromatin spring and maintain force balance. Through Histone H2A S121 phosphorylation and recruitment of Sgo1, Bub1 kinase softens the chromatin spring in response to changes in microtubule dynamics. The geometric alteration of all 16 kinetochores and pericentric chromatin reflect global changes in the pericentromeric region and provide mechanisms for mechanically amplifying damage at a single kinetochore microtubule.
The fidelity of chromosome segregation relies upon the ability of the cell to monitor chromosome biorientation on the mitotic spindle. This monitoring system is known as the spindle assembly checkpoint (SAC). The checkpoint relays structural information from the centromere and/or kinetochore to the cell cycle regulatory machinery. The requirements for accurate segregation include attachment of spindle microtubules to the kinetochore and bi-orientation of sister kinetochores to opposite poles [1, 2]. The centromere designates the position of the kinetochore within the chromosome. The kinetochore is a multi-subunit protein/DNA complex containing more than 65 proteins organized into 8–9 distinct biochemical complexes [3, 4]. In mammals, the kinetochore is a multiple microtubule attachment site (20–25 microtubules per kinetochore) while in budding yeast the kinetochore binds a single microtubule . The kinetochore couples chromosome movement to microtubule dynamics and undergoes conformational changes as a function of tension between the sister kinetochores .
Cohesin, the protein complex that binds sister DNA strands following DNA replication is required for generating tension between sister chromatids [7, 8]. Cohesin, together with a related complex, condensin are enriched in pericentric chromatin in budding yeast [7–9]. Cohesin and condensin are cylindrically arrayed around the mitotic spindle with cohesin radially displaced and condensin proximal to the microtubule spindle axis [10, 11]. The complexes are required for chromatin compaction and spindle length regulation and therefore contribute to mitotic force balance mechanisms . Cohesin at the pericentromere is protected from premature degradation in meiosis I by Shugoshin (Sgo1). Shugoshin is recruited to the centromere via the Bub1 kinase [12–14] where it contributes to mechanisms responsible for orienting sister centromeres to opposite poles .
In metaphase, bi-oriented sister kinetochores separate into two foci that reside at the kinetochore microtubule plus-ends (Fig. 1A). Each focus represents the aggregate of 16 individual microtubule attachment sites. The separation between sister kinetochores is about 800nm, similar to the separation found between sister kinetochores in mammalian cells [16–19]. This finding, together with recent studies on the position of proteins within the kinetochore reveals remarkable conservation in both their number and spatial positions in yeast and mammals [20–22]. Thus the study of how the 16 clustered kinetochores behave in budding yeast is likely to inform our understanding of how a kinetochore with multiple attachments behaves. The focus containing 16 kinetochores is larger than a diffraction limited spot and the architecture (radius, shape) of the cluster is dictated by the position of individual sites within the multiple attachment site. Likewise the dimensions of the cylindrical array of cohesin provide positional information for the organization of pericentric chromatin. While individual kinetochore proteins and cohesin subunits are stable throughout mitosis [23–25] the aggregate kinetochore cluster as well as the cohesin/chromatin barrel, are dynamic and exhibit geometric changes in response to reduced microtubule dynamics. We demonstrate that the Bub1 kinase regulates these structures through histone H2A phosphorylation and recruitment of Sgo1 to the pericentric chromatin. This study reveals a new mechanism for how histone modification can amplify small changes in tension into large scale geometric changes in living cells.
The geometric size and shape of the clustered 16 kinetochores in metaphase can be determined using the mitotic spindle as reference. The metaphase spindle is on average 1.3 microns in length and 0.25 microns in diameter. The full width of a 2-D Gaussian function fitted to each kinetochore cluster is measured perpendicular (y-axis) and parallel (x-axis) to the spindle axis (denoted as height and width, respectively)(Fig. 1, S1). Quantitative analysis reveals structural features not apparent in singly attached kinetochores that are smaller than the diffraction limit. An example of paired kinetochore clusters in metaphase is shown in Figure 1A (top)(Cse4-GFP, Nuf2-GFP green, spindle pole body Spc29-RFP red). The height of kinetochore complex (Nuf2-GFP) is ~580nm in metaphase (Fig. 1A,B) and 575nm in anaphase (Fig. S1). The aspect ratio in anaphase (height/width) is 1.07 (Fig. S1). The MIND complex (Dsn1, Nsl1) is 580nm in height, with an aspect ratio in anaphase of 1.06 – 1.12 respectively (Fig. 1B, S1). Spc105 and Mif2 are 567nm and 581nm in height, aspect ratio of 1.04 (Fig. 1B, S1). The outer kinetochore components exhibit aspect ratios ranging from 1.04 to 1.13 in anaphase (Fig. S1).
The geometry of the inner kinetochore complexes including COMA (Ctf19, Okp1, Ame1), DNA binding complex CBF3 (Ndc10) and the centromere specific histone variant (Cse4) are significantly different (Fig. 1A, S1). Foci containing these components extend 619– 666 nm in height (Fig. 1A, B, S1) versus 504–545nm in width (Fig. S1). The aspect ratio of inner components ranges from 1.22 to 1.27 in anaphase (Fig. S1). The aggregate of 16 kinetochores is not a uniform cylinder surrounding the spindle axis. Rather, there is an anisotropic flare (aspect ratio 1.22–1.27, Fig. S1) in the inner kinetochore cluster as it approaches the chromosomal DNA interface.
The kinetochore undergoes conformational changes in the presence or absence of tension exerted by kinetochore microtubules [22, 23, 26, 27]. To test whether the anisotropy of the inner kinetochore is tension-dependent, we used low concentrations of benomyl to reduce microtubule dynamics and consequently the distance between sister centromeres . At low concentrations of benomyl (10–20µg/ml), cell cycle progression is delayed due to activation of the SAC but cell growth and viability are robust. The qualitative distribution of spindle microtubules and sister kinetochores (spindle length, Cse4 and LacO spot separation) are similar in the presence or absence of benomyl (Table S1 Rows 1,2), even though spindle length is reduced (1.3 to 1.0µm). The size and shape of the outer kinetochore components (NDC80) are unchanged in these concentrations of benomyl (Fig. 1C). In contrast the inner kinetochore components (Cse4, Ame1) contract in height upon exposure to benomyl, from 655–666nm to 565–568nm (Fig. 1C). The height of the inner kinetochore is dynamic and can be attenuated by reducing spindle tension. The anisotropy observed in anaphase reflects mechanical linkages and therefore tension, that persist from metaphase to anaphase [29, 30].
To test whether the conformation of the inner kinetochore depends on the SAC we treated cells lacking either Bub1 or Mad2 with benomyl and examined the dimension of Ame1 (COMA complex). In the absence of Bub1 kinase the inner kinetochore cluster (Ame1) is 699nm in height (Fig. 1C). The geometry of Ame1 persists following exposure to low concentration of benomyl in bub1Δ (689nm, Fig. 1C). Thus Bub1 is required for attenuation of the height of the kinetochore in reduced tension. In the absence of Mad2, the geometry of Ame1 is reduced upon exposure to low benomyl, as observed in wild-type (566 vs. 568nm Fig. 1C). Bub1 kinase and not the spindle checkpoint are responsible for the conformational change of the inner kinetochore upon reduced tension.
Bub1 kinase is known to directly phosphorylate Histone H2A (S121) [13, 14], that serves as a mark for Sgo1 localization in the centromere in both yeast and mammals . To test whether histone phosphorylation is required for conformational change in response to reduced tension, we examined inner kinetochore structure in cells expressing the unphosphorylated form of histone H2A-S121A. In cells expressing H2A-S121A, the geometry of Ame1 persists following exposure to low benomyl (689/697nm, −/+ benomyl, respectively). Likewise, in sgo1Δ mutants the geometry of Ame1 is unchanged upon benomyl treatment (Ame1, sgo1Δ; Fig. 1C). The kinetochore localization of Sgo1 through Bub1 phosphorylation of histone H2A is required to mediate tension dependent shape changes of the inner kinetochore (Fig. 1C).
Cohesin is enriched in the pericentric chromatin [31, 32] and encircles the mitotic spindle in the form of a cylindrical barrel [24, 25]. The cohesin barrel spans the distance between the cluster of sister kinetochores and has a diameter approximately 417nm (sagittal view). To determine whether the structural change of the inner kinetochore is propagated to adjacent chromatin we examined the structure of pericentric cohesin (Smc3-GFP). In the presence of benomyl, the diameter of the barrel expands to 483nm (Fig. 2). To estimate the increase in volume we have compared cohesin barrel width from spindles of comparable length (~1.0µm) in benomyl treated and untreated cells. The volume of the cohesin cylinder increases ~ 34% upon exposure to benomyl (Benomyl treated (π ((483/2)2)605) ÷ wt (π ((417/2)2)605) = 1.34). The concentration of cohesin measured by quantitative fluorescence microscopy increases approx. ~15% (Fig. S2). In the absence of Bub1, the barrel has a diameter of 442nm (vs. 417nm for wt). Upon treatment with benomyl, the barrel does not expand and instead exhibits wild-type dimensions (Fig. 2 bub1Δ, ben 426nm) and cohesin concentration (Fig. S2). The barrel does not expand when exposed to low benomyl in sgoΔ and the H2A-S121A mutant (Fig. 2). In contrast, the cohesion barrel expands in benomyl treated mad2Δ mutants (Fig. 2 mad2Δ, 441/484nm −/+ben). The geometry and abundance of cohesin are not just a consequence of spindle shortening (Table S1, Fig. S2). In the presence of benomyl, spindles shorten in wild-type as well as the mutants, but cylinder expansion depends on Bub1, Sgo1 and not Mad2 (columns 1, 2, Table S1). Thus, like Ame1 in the inner kinetochore cluster, the structure of cohesin within the pericentric chromatin responds to reduced tension and is dependent on the Bub1 kinase, Sgo1, H2A S121 phosphorylation, but not the spindle assembly checkpoint (Fig. 2).
As an alternative method to reduce microtubule dynamics, we utilized a temperature-sensitive mutation stu2-145. Stu2 is the homolog of XMAP215 and functions as a regulator of microtubule dynamics . Repression of Stu2 leads to a severe reduction of microtubule dynamics . To examine the response of the cohesin barrel to loss of Stu2, we introduced Smc3-GFP into cells containing stu2-145. Stu2-145 contains a single point mutant in the HEAT repeat (V145E) that renders the strain temperature sensitive. Like the repression of Stu2, microtubule dynamics in stu2-145 are suppressed (Pearson, C.G. and Bloom, K., data not shown). Upon shift of these cells to restrictive temperature for stu2-145, the cohesin barrel expands from 393nm to 478nm (Fig. 2). Barrel expansion in stu2-145 is dependent upon Bub1 (stu2-145, bub1Δ, 25°C vs 37°C Fig. 2).
To determine whether the change in cohesin structure reflects the conformation of pericentric DNA, we examined the spatial position of centromere-linked LacO arrays (6.8kb from the centroid of the LacO to CEN15) visualized by LacI-GFP. The LacO arrays appear as two fluorescent spots that lie between the spindle poles in metaphase (Fig. 3, Inset, top left panel) [33, 34]. To determine their spatial distribution, we compiled a two-dimensional density map for LacO foci in a population of metaphase cells [11, 35]. In cells where the kinetochores had bioriented, the peak intensity (pixel) of each diffraction limited focus was determined, and the x,y coordinates were plotted relative to the spindle pole body (marked with Spc29-RFP). The frequency distribution is indicated in the color-coded map (Fig. 3, yellow and white, highest density; red and black, least density). Because the rotation of the spindle is random in individual cells, the coordinates obtained are one quadrant of the cylindrical arrangement of pericentric chromatin around the spindle. To account for this geometry, we mirrored the heat map about the spindle axis to project the area as it would be viewed from a single plane cut through the middle of the spindle. The pericentric LacO DNA is cylindrically arrayed and radially displaced from the central spindle axis, with a diameter of 194nm. The least frequently occupied positions are along the spindle axis (note low density in black, along the spindle axis, spindle pole marked by red spot). Relative to the spindle (x-axis), the LacO array is 398 nm from the spindle pole (Fig. 3, top left). In the presence of low concentrations of benomyl the LacO arrays move closer to the spindle pole (332nm, Fig. 3, bottom left) reflecting the shortening of kinetochore microtubules in the presence of benomyl . Most notable is the increased radial displacement (324nm, Fig. 3, bottom left). The separated LacO arrays are no longer confined to a region near the microtubule plus-end rather they occupy an area comparable to the dimensions of the expanded cohesin barrel. Their radial position is expanded from 194nm to 324nm. The volume occupied by pericentric chromatin increases ~100% upon exposure to benomyl (π *(324/2)2*332) ÷ (π*(194/2)2*398 = 2.3). The radial expansion does not reflect a change in the volume due to spindle shortening since the spindle shortens and lacO spots contract along the x-axis, but do not expand radially in bub1Δ mutants treated with benomyl (Fig. 3, middle panels and Table S1). Thus pericentric chromatin as well as cohesin occupies a greater spatial area following reduction of microtubule dynamics. Like the cohesin barrel, radial expansion of pericentric chromatin upon benomyl treatment is dependent upon the Bub1 kinase (230 vs. 235nm bub1Δ, −/+ low benomyl Fig. 3 middle panels), and Sgo1 (240 vs. 259nm sgo1Δ, −/+ low benomyl Fig. 3 right panels). Pericentric chromatin lies closer to the spindle axis in sgo1Δ mutants (Fig. 3 top right), indicating that Sgo1 is likely to have additional functions in the organization of pericentric chromatin.
The increase in spatial occupancy of the pericentric LacO DNA suggests that chromatin structure may be modulated in response to the change in microtubule dynamics. To deduce whether the dynamics of the pericentric chromatin have been altered, we examined the fluctuation in pericentric LacO DNA motion on individual chromosomes over time . In wild-type cells the LacO DNA spots move relative to the spindle poles as well as to each other. They are often separated and transiently come together [33, 34]. In metaphase, the fluctuation (variance) is 0.132 microns (measured in single cells at 20sec intervals over 12min) (Fig. 4, Table S1). Upon treatment with low concentrations of benomyl, LacO spot movement is dampened and the variance is reduced to 0.07µm. This reduction correlates with reduced microtubule dynamics (Fig. S3) , and has led to the paradigm that centromere-linked LacO movements serve as a proxy for microtubule dynamics .
To distinguish whether the histone chromatin modification or decrease in microtubule dynamics is responsible for reduced chromatin dynamics we examined fluctuations of pericentric LacO arrays in bub1Δ mutants treated with low concentrations of benomyl. In this situation the inner kinetochore does not contract (Fig. 1), the cohesin barrel does not expand (Fig. 2) and microtubule dynamics remain suppressed (Fig. S3). Pericentric LacO arrays in bub1Δ mutants treated with benomyl exhibit variance comparable to wildtype dynamics (wt 0.132 µm; bub1Δ, ben 0.121µm, Fig. 4, Table S1). Thus decreased variance in LacO motion in benomyl reflects a Bub1-dependent chromatin state rather than the change in microtubule dynamics. The decreased variance of pericentric LacO in benomyl is also dependent upon Sgo1, though less so than Bub1 (sgo1Δ, ben, 0.103µm vs. wt, ben 0.07µm Fig. 4, Table S1). The increase in spatial distribution of the LacO array observed in low benomyl (Fig. 3, middle right) does not reflect increased motion. Rather there is an increased area available for occupancy in response to treatment with benomyl. Once a position is achieved, the LacO array is relatively stationary, hence the reduced variance. This behavior is non-ergodic since the behavior of the ensemble (increased spatial occupancy of LacO arrays, Fig. 3) does not reflect the behavior of an individual LacO DNA array exhibiting reduced variance (Fig. 4). Following histone modification the pericentric DNA occupies a larger area, but movement of a chromosomal locus is reduced.
Bub1- and Sgo1-GFP fusion proteins are found at or between the clusters of 16 kinetochores in metaphase [12, 13](Fig. 5, Bub1, 67%; Sgo1, 90% 1 or 2 spot). To deduce the position at nanometer resolution of proteins that do not cluster into diffraction-limited spots we determined the statistical distribution from a large population. We applied this method to kinetochore proteins of known position to validate and assess the accuracy (Fig. S4, Table S3). The statistical maps recapitulate the linear distances determined by pairwise centroid mapping  with an error of 5nm (Table S3, column 2). In cells with two foci (48% of total), Bub1-GFP is ~5nm from Cse4, toward the inner centromere (Fig. 5B). In cells with two Sgo1 foci (23%), Sgo1 is ~61nm from Cse4, toward the inner centromere. Sgo1 is closer to the sister chromatid axis (center of spindle) relative to Bub1, thereby decreasing the opportunity to resolve two spots (Fig. 5A). Upon exposure to benomyl, both Sgo1 and Bub1 change their distribution along the spindle. There is an increase in concentration of Sgo1-GFP as well as a shift of Bub1-GFP to primarily a single spot along the sister chromatid axis (Fig. 5A). To distinguish whether the increase in Sgo1 reflects chromatin re-distribution or Sgo1 recruitment, we examined the position of pericentric lacO in the spindle. The distribution of LacO spots in spindles of the same absolute length (0.7 – 1.0 microns) in wild-type and benomyl treated cells is identical (Fig. S5). Likewise Cse4 to Cse4 distance in short (1µM) spindles does not change significantly after benomyl treatment (612 nm and 606 nm, respectively, p = 0.878 as determined by Student’s T-Test).
If there was an increase in chromatin re-distribution upon spindle damage, there would be an increase in the fraction of cells in which two lacO spots are observed. Thus Sgo1 increases linearly in abundance with increasing concentrations of benomyl (approximately 2-fold increase with 4-fold increase in benomyl, Fig. 5A). Bub1-GFP shifts from a predominance of two spots (48%) to one spot (41%) in benomyl. The recruitment of Sgo1 and cohesin (Fig. 5, Fig. S2) reflects a mechanism to modify the physical properties of the pericentric chromatin of all 16 chromosomes that lie between the clusters of bi-oriented kinetochores.
Bub1 and Sgo1 modify the cohesin barrel in response to altered microtubule dynamics during mitosis. Phosphorylation of histone H2A by Bub1 and the recruitment of Sgo1 within the kinetochore/pericentric region result in a conformational change in the inner kinetochore and pericentric cohesin. The paradox in these results is that the pericentric chromatin expands radially from the spindle axis, while the inner kinetochore contracts toward the spindle axis. We propose that this change is due to the physical expansion of pericentric chromatin and cohesin surrounding the kinetochore and spindle microtubules (Table 1, Fig.6). Increasing the area occupied by pericentric chromatin could exert an isotropic force that expands the outer dimension (pericentric chromatin and cohesin) and contracts the inner dimension (inner kinetochore) of the cylindrical array of DNA and cohesin (Fig. 6, bottom). In metaphase, pericentric cohesin together with histones and condensin restrain the pericentric chromatin to an entropically disfavored area proximal to the spindle . Chromatin modification provides a mechanism to tune the force exerted from the constrained DNA polymer and maintain force balance with spindle microtubules.
Upon exposure to agents that damage the spindle or chromosome attachments to the spindle, it may be important for the cell to modulate the chromatin spring in an effort to maintain force balance. The chromatin spring behaves like a worm-like chain in vivo [11, 37]. The parameters that dictate the spring constant are the chain length (Lc) and persistence length (Lp)(Table 1). Persistence length is the length scale over which the ends of links in the chain are correlated. The force required to extend a polymer chain is inversely proportional to the persistence length (F ~ KBT/Lp, KBT, Boltzmann constant). This makes intuitive sense if one considers a polymer chain. The shorter the persistence length the greater the number of available states for the random coil (# of entropic states, Table 1); the longer the persistence length the fewer number of states for the random coil (Table 1). Less force is required to extend a chain that has fewer available states, and therefore the spring constant decreases. The persistence length also dictates the physical size of the random coil (radius of gyration Rg ~ √(Lc Lp)). Increase of persistence length would change the dimensions of the pericentric chromatin (Table 1, Fig. 6) and soften the spring. We propose that the cell is able to tune the chromatin spring constant in response to mitotic spindle damage. In addition, the reduction in spring constant would also lead to a reduction in tension–based rescue mechanisms that regulate microtubule growth . It may be important for the cell to shift the equilibrium to shorter kinetochore microtubules to maintain critical concentrations of tubulin polymer for the integrity of interpolar microtubules. The recruitment of additional Sgo1 to the kinetochore and pericentromere when the spindle is perturbed is suggestive of the continuous modulation of pericentric chromatin in response to damage. A quantitative increase in chromatin modification could reflect a biological rheostat for regulating the spring as a function of the extent of spindle damage and/or chromosome loss.
The mechanism by which Bub1 kinase and Sgo1 modulate chromatin structure and promote chromosome bi-orientation is poorly understood. Bub1 resides on the chromatin side of the kinetochore, several nanometers from the Histone H3-variant, Cse4 (Fig. 5). Sgo1 resides another 60nm towards the sister chromatid axis (Fig. 5). It has been previously proposed that Sgo1 is responsible for the intrinsic property of the kinetochore to bi- orient although a mechanism for this geometric bias has not been well understood . Sgo1 resides at a critical junction defined by the packing ratio of pericentric chromatin. The packing ratio of pericentric chromatin from 1.7kb to 8.8kb is 106bp/nm . From 1.7kb to the position of the centromere (CEN, Cse4) the packing ratio is 25bp/nm, equivalent to nucleosomal chromatin. The position of Sgo1 coincides with this transition (Fig. 5B, S4), and indicates that Sgo1 may stabilize the C-loop cruciform structure for geometric bias of sister centromeres [15, 25].
Our understanding of kinetochore geometry and pericentric spring function derive largely from the discovery of Shugoshin (Sgo1) and the mechanism by which cohesin, condensin and centromere loops generate a molecular spring. The molecular spring, comprised of the DNA worm-like chain and cohesin and condensin protein springs is important for generating a counterforce to spindle microtubules such that tension can be measured between sister kinetochores . The pericentric region is also responsible for predisposing replicated centromeres to lie on the surface of the chromosome. This study provides mechanistic insight into the role of H2A phosphorylation and recruitment of Sgo1 in tuning the chromatin spring in vivo.
Proteins were tagged at the C terminus with GFP, CFP or RFP through use of PCR cassettes unless otherwise noted. The strains are listed in the Supplemental Material.
Cells were imaged at room temperature on a Nikon TE-2000E (Nikon Instruments, Melville, NY) inverted microscope equipped with a 1.4 NA, 100X Plan-Apo objective as previously described . For each cell, 13 images stepped through 200nm z-axis were obtained. Imaging and acquisition software was by Metamorph (Molecular Devices, Sunnyvale, CA)
To determine the spatial geometry of kinetochore and pericentric chromatin we compiled a two-dimensional density map of the fluorescence distribution of kinetochore proteins and LacO foci in the pericentric chromatin from a population of metaphase cells. The density map is a statistical distribution of a given protein’s residence. These maps were generated by taking the peak intensity of each diffraction limited focus, and plotting the coordinates in two dimensions relative to the spindle pole body (marked with Spc29p-RFP). The distribution of x,y coordinates represents the frequency of a given protein or pericentric LacO array relative to the spindle pole (Figure 3, ,55 and S4). Image analysis was conducted in a custom written GUI with point-and-click selector for the selecting fluorescently labeled SPBs. The program searches for the brightest pixel in a pre-defined vicinity of the user-selected pixel and uses this pixel as the center of the SPB image. The program rotates the spindle axis defined by the two SPB centers using the 'imrotate' routine in the Image processing toolbox in Matlab to align it with the image X axis. The width of the spot is determined by fitting its intensity distribution with a 2-D Gaussian function (Center X, Center Y, SigmaX, SigmaY, Peak Intensity, and background value are the free parameters). To generate a statistical positional density map, spindle pole bodies were tagged with RFP, and these foci were used to determine the spindle axis. The image was then rotated to bring the spindle axis to a horizontal level. This was done using bicubic interpolation in a custom graphical user interface (GUI) in Matlab 2008a. The rotational data was then saved and exported into images of kinetochores or LacO arrays tagged with GFP, taken concurrently with the RFP images. This allows for an identical rotational shift to be applied to the GFP images. Once all images were aligned and rotated, the coordinates of the brightest pixel in each foci were logged. To create a statistical positional density map from this dataset, the total number of pixels localized to any given coordinate was determined. These values were then imported into another custom GUI in Matlab, which generated a 2x interpolated color coded positional density map. High density areas, indicated by large numbers of pixels falling on these coordinates, are represented by the black body radiation spectrum (black represents zero probably, red and orange represents low probability, yellow and white represent high probability).
For benomyl treatment, mid-logarithmic cells in YC complete media were washed with sterile water and incubated for ~45 min on slabs containing 10–20µg/ml benomyl (DuPont, Wilmington, DE) in dimethylsulfoxide (DMSO). At these concentrations of benomyl the cell division cycle is not arrested, however tubulin dynamics are slowed from a t½= 50sec (untreated) to a t½= 260sec .
We thank our Mitosis group, including Leandra Vicci, Drs. Russell M. Taylor II (Dept of Computer Science, UNC-CH) and Mike Falvo (Dept of Physics, UNC-CH) for discussions and Dr. Ajit Joglekar (U Michigan) for expert assistance with image analysis and Matlab programming. We thank Drs. A. Strunnikov (NIH) for the Smc3-GFP plasmid, Y. Watanabe (University of Tokyo) for the Sgo1-GFP and Histone H2A S121A mutant strains and Drs. Lynne Cassimeris and Margaret Kenna (Lehigh University) for the stu2-145 (V145E) temperature sensitive allele. This work was supported by the National Institutes of Health R37 GM32238 to KB.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.