Independence of centromeric and pericentromeric chromatin stability on CCAN components

Abstract

The chromatin of the centromere provides the assembly site for the mitotic kinetochore that couples microtubule attachment and force production to chromosome movement in mitosis. The chromatin of the centromere is specified by nucleosomes containing the histone H3 variant, CENP-A. The constitutive centromeric-associated network (CCAN) and kinetochore are assembled on CENP-A chromatin to enable chromosome separation. CENP-A chromatin is surrounded by pericentromeric heterochromatin, which itself is bound by the sequence specific binding protein, CENP-B. We performed mechanical experiments on mitotic chromosomes while tracking CENP-A and CENP-B to observe the centromere's stiffness and the role of the CCAN. We degraded CENP-C and CENP-N containing auxin-inducible degrons, which we verified compromises the CCAN via observation of CENP-T loss. Chromosome stretching revealed that the CENP-A domain does not visibly stretch, even in the absence of CENP-C and/or CENP-N. Pericentromeric chromatin deforms upon force application, stretching ∼3-fold less than the entire chromosome. CENP-C and/or CENP-N loss has no impact on pericentromere stretching. Chromosome-disconnecting nuclease treatments showed no structural effects on CENP-A. Our experiments show that the core-centromeric chromatin is more resilient and likely mechanically disconnected from the underlying pericentromeric chromatin, while the pericentric chromatin is deformable yet stiffer than the chromosome arms.

FIGURE 1:

Model of the relative positioning of the centromere and associated complexes (One chromatid). Components of the CCAN (inner kinetochore, blue) connect the underlying pericentromeric chromatin (green) to the outer kinetochore (purple). CENP-A nucleosomes (red) are associated with CENP-C, with adjacent DNA (gray) contacting CENP-N/L and CENP-TWSX.

FIGURE 2:

Fluorescence imaging and degradation of the used AID-CENP cell lines. (A) Example images of mitotic cells with centromere fluorescence. Top row shows phase-contrast imaging of target mitotic cells. Second row shows corresponding DNA signal via Hoechst staining in cyan (used here to show relative positions of centromere and DNA; Hoescht staining was not used in experiments that measured mechanical properties of chromosomes or the centromere). Third and fourth rows show endogenous, CENP protein in green (CENP-A or CENP-C) and magenta (CENP-C or CENP-N); labels are shown above each pair of images. Bottom row shows merged centromere images. Overlap signal shown in white. The first two columns show the AID-CENP-C–containing line, the middle two columns show the AID-CENP-N–containing lines, and the last two columns show the AID-CENP-C– and AID-CENP-N–containing line. Odd columns show untreated cells and even columns show auxin-treated cells. Auxin treatment removed the AID-tagged fluorophores’ signal but did not affect the fluorescence of non-AID-tagged CENP-A. Bar indicates 10 µm. (B) Fluorescence quantification method and example image. Representative cell showing endogenous CENP-A fluorescence with areas of interest for centromere and background intensities. In short, centromere fluorescent values are measured as the fluorescence average of the boxed centromere minus the fluorescence average of the background (inside the cell) at 500 ms exposure. Bar indicates 10 µm. (C) Quantification of fluorescence in channel 1. Numbers shown are averages of the centromere signal minus the cell background signal at 500 ms exposure. In the CENP-C line, the CENP-A fluorescence was 624 ± 48 counts (N = 43) in untreated cells and 856 ± 70 counts above background (N = 20) (statistically significantly higher) in auxin-treated cells, demonstrating auxin does not have a diminishing effect on the CENP-A signal. In the CENP-N (409) line, the CENP-A fluorescence was 261 ± 48 counts above background (N = 13) and 211 ± 21 counts (N = 12) (not statistically different) in auxin-treated cells. In the CENP-C/N line, CENP-C fluorescence was 396 ± 37 counts above background (N = 26) and −7 ± 2 counts above background (N = 21) (statistically significantly lower) in 4-h auxin-treated cells. Single data points corresponding to these averages are shown in Supplemental Figure S1A. (D) Quantification of auxin degradation (CENP-C, CENP-N (409), and CENP-N, respectively). In the CENP-C line, the CENP-C fluorescence was 338 ± 16 counts above background (N = 42) in untreated cells and 9 ± 6 counts above background (N = 20) (statistically significantly lower) in auxin-treated cells. In the CENP-N line, the CENP-N fluorescence was 208 ± 33 counts above background (N = 26) in untreated cells and 21 ± 22 counts above background (N = 24) (statistically significantly lower) in auxin-treated cells. In the CENP-C/N line, the CENP-N fluorescence was 348 ± 28 counts above background (N = 26) in untreated cells and 3 ± 5 counts above background (N = 21) (statistically significantly lower) in 4-h auxin-treated cells. Single data points are shown in Supplemental Figure S1B.

FIGURE 3:

Methods for chromosome isolation, mechanics, and quantification. (A) Example of chromosome isolation by bundle removal method. Upper left- Target mitotic cell. Upper middle- Target cell after lysis by detergent. Upper right- Chromosome bundle after aspiration into stabilizing pipette. Lower left- Target chromosome aspirated into chromosome-aspirating pipettes. Lower right- Isolated target chromosome. Bar indicates 10 µm. (B) Example experiment stretching a whole chromosome in phase-contrast imaging. Chromosomes were stretched by moving the pull pipette (bottom) away from the deflecting pipette (top). An example of a live stretch-deflection trace is shown in C, while the corresponding fluorescent images are shown in D. The chromosome versus centromere stretch is plotted in F, where the chromosome stretch is manually measured as the distance between the middle of the pipettes. Bar indicates10 µm. (C) Example trace of a stretch-deflection experiment for a whole chromosome. A computer tracked the two pipettes’ positions simultaneously through the experiment and directed a micromanipulator to move the pull pipette away from the deflecting pipette at a specified speed and distance (typically 1.0 µm/s and 20 µm in these experiments). The linear regression slope of chromosome stretching versus pipette deflection was multiplied by the spring constant of the deflection pipette to obtain the spring constant of the whole chromosome. Supplemental Figure S2A shows an additional force-extension plot, including past its linear-stretching regime and Supplemental Figure S2B shows the individual data points of the mechanics of all chromosomes. (D) CENP-A fluorescence during chromosome stretching. CENP-A images were taken while stretching the chromosome (B). The yellow box was placed around the centromere and averaged over each column of the box's length and plotted in E to obtain its length. Bar indicates 10 µm. (E) FWHM/fluorescent trace of CENP-A. Black dots represent the raw data of the box plot seen in D. Red dashed line shows a Gaussian fit, used to calculate FWHM, giving the centromere length in pixels. The FWHM was converted from pixels to microns when reporting CENP-A lengths and corrected for its point-spread function. This was repeated for each image of the centromere and was plotted against chromosome stretch in F. (F) Example plot of centromere stretch versus whole-chromosome stretch. Chromosome length (B) was plotted against centromere FWHM (D, E), both normalized to initial unstretched chromosome length and centromere FWHM. A linear fit gave a slope of 0.057 ± 0.046 for the CENP-A example stretch (black line and black squares). Another example chromosome stretched with CENP-B staining is also shown in the graph plotted against chromosome stretch. A linear fit yielded slope of 0.407 ± 0.038 (orange line and orange circles). Slopes were determined for each centromere and CENP-B stretching experiment (Figures 4, B and D and 5B; one data point for each average of a series of individual chromosome stretches is shown in Supplemental Figure S3).

FIGURE 4:

CENP-A does not stretch but may become smaller when CENP-C or -N is degraded. (A) Example chromosome stretch showing centromere behavior (endogenous CENP-A). The whole chromosome in phase-contrast imaging is shown above each image of the endogenous centromere in its corresponding fluorescent channel. The example image shows a CENP-A stretch ratio of 0.058 (centromere stretch: chromosome stretch) with an initial length of 0.59 µm after point-spread correction. Bar indicates 10 µm. (B) Quantification of endogenous CENP-A mechanics dependence on CENP degradation. Left panel shows the average stretch of the centromere to the whole chromosome. In CENP-C cells, the endogenous CENP-A: chromosome stretch ratio was 0.035 ± 0.032 (N = 28) in untreated cells, and −0.018 ± 0.058 (N = 14) (statistically insignificantly different) in auxin-treated cells. In CENP-N cells, the endogenous CENP-A: chromosome stretch ratio was −0.033 ± 0.058 (N = 14) in untreated cells, and −0.022 ± 0.039 (N = 15) (statistically insignificantly different) in auxin-treated cells. The individual data corresponding to these averages are shown in the first four columns of Supplemental Figure S3A. The right panel shows the average of the initial length of the centromere (endogenous fluorescent CENP-A signal) (the centromere's length when the chromosome is at its initial length and under no tension). Average lengths are corrected for the microscope fluorescence point-spread function (Supplemental Figure 7, A and B). The initial length of the endogenous CENP-A in the CENP-C line was 0.71 ± 0.04 µm (N = 41) in untreated cells, and 0.63 ± 0.07 µm (N = 14) (smaller, but not statistically different) in auxin-treated cells. The initial length of the endogenous CENP-A in the CENP-N line was 0.63 ± 0.06 µm (N = 18) in untreated cells, and 0.48 ± 0.04 µm (N = 14) (marginally statistically significantly lower [P = 0.054]) in auxin-treated cells. The single data points can be seen in the first four columns of Supplemental Figure S3B. (C) Example stretching of a chromosome with the centromere (antibody-labeled CENP-A). The whole chromosome in phase-contrast imaging is shown above each image of the antibody-stained centromere in its corresponding fluorescent channel. The example image shows a CENP-A stretch ratio of −0.045 (centromere stretch: chromosome stretch) with an initial length of 0.36 µm after point-spread correction. While there is nonspecific binding, as seen by the frequency of fluorescent signal along the chromosome, an increase in fluorescence can be seen in a punctate area, corresponding to the centromere, used in the quantification shown in Figure 4D. Bar indicates 10 µm. (D) Quantification of antibody CENP-A mechanics dependence on CENP degradation. The left panel shows the average stretch ratio of the centromere to the whole chromosome. In CENP-C/N cells, the antibody-stained CENP-A: centromere ratio was 0.011 ± 0.028 (N = 16) in untreated cells, 0.022 ± 0.024 (N = 17) (statistically insignificantly different from all treatments) in 4-h auxin-treated cells, and −0.007 ± 0.020 (N = 7) (statistically insignificantly different from all treatments) in 24-h auxin-treated cells. Single data points are shown in the last three columns of Supplemental Figure S3A. The right panel shows the average of the initial length of the centromere (CENP-A signal after fluorescent antibody spraying) (the centromere's length when the chromosome is at its initial length and under no tension). The reported lengths are corrected for the microscope fluorescence point-spread function (Supplemental Figure S7A, B). The initial length of the antibody-stained CENP-A in the CENP-C/N line was 0.91 ± 0.13 µm (N = 16) in untreated cells, 0.58 ± 0.08 µm (N = 17) (statistically significantly smaller compared with untreated) in 4-h auxin-treated cells, and 0.62 ± 0.18 µm (N = 8) (statistically insignificantly different from other treatments) in 24-h auxin-treated cells. The single data points can be seen in the last three columns of Supplemental Figure S3B.

FIGURE 5:

CENP-B stretches less than the chromosome, but is not affected by CENP degradation. (A) Example stretching of a chromosome with the pericentromere, antibody-labeled CENP-B. The whole chromosome in phase-contrast imaging is shown above each image of the CENP-B antibody-stained pericentromere in its corresponding fluorescent channel. The example image shows CENP-B stretching via a smear of fluorescence. The example image shows a CENP-B stretch ratio of 0.262 (centromere stretch: chromosome stretch) with an initial length of 1.20 µm after point-spread correction. Bar indicates 10 µm. (B) Quantification of antibody-labeled CENP-B mechanical dependence on CENP degradation. The top panel shows the average stretch of the pericentromere to the whole chromosome. In CENP-C cells, the antibody-stained CENP-B: chromosome stretch ratio was 0.322 ± 0.083 (N = 8) in untreated cells, and 0.244 ± 0.036 (N = 10) (statistically insignificantly different) in auxin-treated cells. In CENP-N cells, the antibody-stained CENP-B: chromosome stretch ratio was 0.233 ± 0.072 (N = 9) in untreated cells, and 0.376 ± 0.115 (N = 9) (statistically insignificantly different) in auxin-treated cells. In CENP-C/N cells, the antibody-stained CENP-B: chromosome stretch ratio was 0.332 ± 0.083 (N = 12) in untreated cells, 0.450 ± 0.072 (N = 8) (statistically insignificantly different from all treatments) in 4-h auxin-treated cells, and 0.413 ± 0.131 (N = 10) (statistically insignificantly different from all treatments) in 24-h auxin-treated cells. Single data points are shown in Supplemental Figure S3C. The bottom panel shows the average initial length of the pericentromere (length of CENP-B signal after fluorescent antibody spraying when chromosome is at its initial length and under no tension). Reported lengths are corrected for the microscope fluorescence point-spread function (Supplemental Figure S7A, B). Initial length of the antibody-stained CENP-B in the CENP-C line was 1.34 ± 0.14 µm (N = 9) in untreated cells, and 1.11 ± 0.14 µm (N = 11) (statistically insignificantly different) in auxin-treated cells. The initial length of the antibody-stained CENP-B in the CENP-N line was 1.22 ± 0.12 µm (N = 10) in untreated cells, and 1.16 ± 0.07 µm (N = 8) (statistically insignificantly different) in auxin-treated cells. The initial length of the antibody-stained CENP-B in the CENP-C/N line was 1.34 ± 0.13 µm (N = 12) in untreated cells, 1.09 ± 0.08 µm (N = 7) (statistically insignificantly different from all treatments) in 4-h auxin-treated cells, and 1.48 ± 0.21 µm (N = 10) (statistically insignificantly different from all treatments) in 24-h auxin-treated cells. The single data points are shown in Supplemental Figure S3D. (C) Example image of CENP-A localization with CENP-B signal. After spraying with CENP-B antibody in the CENP-C line, we observed colocalization of CENP-A with CENP-B. The CENP-A signal was periodically seen separating into distinct entities as each of the chromatid centromeres moved away from each other. These individual centromeres were still seen inside the CENP-B signal. This image also illustrates that even when the CENP-B signal was observed to stretch, the CENP-A signal was not seen deforming as the entire chromosome was stretched. More examples are shown in Supplemental Figure S5. Main scale bar is 10 µm, middle lower right. 5x Merge (right) is five times the size of the other images with CENP-A in green and CENP-B in magenta; scale bar for these images is 1 µm (lower right).

FIGURE 6:

CENP-C and -N degradation affects CENP-T protein levels while DNA digestion does not affect centromere length. (A) Fluorescence of bundles after CENP-T fluorescent antibody spraying in all conditions. The whole chromosome bundle in phase-contrast imaging is shown above each image of the anti-CENP-T fluorescent antibody in its corresponding fluorescent channel across CENP-C, N, and C/N lines across untreated, 4-h auxin-treated cells, and 24-h auxin-treated cells in the CENP-C/N line. Bars represent 10 µm. (B) CENP-C, N, and C/N degradation over 4 and 24 h decreases CENP-T fluorescence. Chromosome bundles were imaged in the anti-CENP-T fluorescence channel, where five random and in focus centromeres had their intensities measured at 500 ms exposure time. In untreated cells, the bundles were bleached of fluorescence in the CENP-T fluorescent channel so there was no interference of the endogenous fluorescence with the CENP-T fluorescence. In the CENP-C line bundles, the CENP-T fluorescence was 4300 ± 600 counts above background (N = 5) in untreated cells 1600 ± 180 counts above background (N = 3) (statistically significantly lower) in auxin-treated cells. In the CENP-N line bundles, the CENP-T fluorescence was 4410 ± 730 counts above background (N = 6) in untreated cells and 1930 ± 460 counts above background (N = 6) (statistically significantly lower) in auxin-treated cells. In the CENP-C/N line bundles, the CENP-T fluorescence was 5640 ± 780 counts above background (N = 4) in untreated cells, 1620 ± 330 counts above background (N = 3) (statistically significantly lower than untreated cells) in 4-h auxin-treated cells, and 2360 ± 780 counts above background (N = 4) (statistically significantly lower than untreated cells) in 24-h auxin-treated cells. Single data points are shown in Supplemental Figure S6A. (C) Example images pre- and post-AluI and MNase digestion on CENP-A length. AluI can digest isolated chromosomes and snap chromosome bundles (breaks AG^CT DNA sequences, which is also contained in human centromere repeat sequences). The isolated bundles were shown in both phase-contrast imaging and the CENP-A fluorescent channel, then sprayed with AluI mix and imaged again in phase-contrast imaging and the CENP-A fluorescent channel. An example of both pre and post digestion are shown in this image. MNase treatment is shown in the bottom panels and causes a near complete digestion of the chromosome except for the CENP-A region in phase-contrast. Bar represents 10 µm. (D) The centromere width does not change with AluI digestion. Bundles of chromosomes were extracted from the CENP-C line for use in digestion experiments, due to its higher intensity and endogenous CENP-A fluorescence. Before digestion, the bundles were imaged in phase contrast and the CENP-A fluorescent channel, sprayed with AluI or MNase, then imaged again in phase contrast and the CENP-A fluorescent channel. Five random, in-focus centromeres were scanned with a boxplot (Figure 3D and E) in both the pre and post digestion images, then averaged for the experimental data point. In these experiments, the CENP-A width was 0.41 ± 0.02 µm (N = 3) in undigested bundles, and 0.39 ± 0.01 µm (N = 3) (statistically insignificantly different) following AluI digestion. The CENP-A width was 0.42 ± 0.02 µm (N = 3) in undigested bundles and 0.49 ± 0.03 µm (N = 3) (statistically insignificantly different) in post-MNase-treated bundles (there was more interference with the background in the post-MNase digested bundles due to the thorough digestion by MNase, complicating this result). Single data points are shown in Supplemental Figure S6B.

FIGURE 7:

Model of chromosome stretching with centromere tracking and alternate possibilities. (A) Stretching schematic of the chromosome from untreated cells. The top panel shows a cartoon schematic of the unstretched chromosome arms colored in black. The chromosome arms are connected to the larger green CENP-B binding/α-satellite DNA. The red CENP-A nucleosomes are smaller and lateral to the CENP-B region. The CENP-A nucleosomes are connected to the blue CCAN structures that also invade the CENP-A domain, representing CENP-T binding to centromeric chromatin. The left side of the chromosome shows the model as if the CENP-A region is still along the chromosome axis and experiences stretching forces when the whole chromosome is stretched, while the right-hand side shows the model if the CENP-A region is mechanically disconnected from the central axis of the chromosome and thus does not experience forces when the whole chromosome is stretched. The bottom panel shows the example chromosomes under stress via stretching of the whole chromosome. The chromosome arms stretch the most, while the CENP-B region stretches less. This model shows the sliding of the two chromatid centromeres from each other in the model due to them not being locked into a specific connected region in the centromere relative to each other. Both panels show no change in the distance of the CENP-A region from the unstretched picture in the top panel. The left-hand panel shows how an increase in stiffness in the CENP-A region compared with the CENP-B region allows CENP-A to maintain its integrity despite being under stress, in that it resists the pulling tension more than the CENP-B region. The right-hand panel shows how if not in line with the chromosome-axis stretching, CENP-A would be under no/minimal stress, and thus not stretch. (B) Stretching schematic of chromosome from CCAN-degraded (auxin-treated) cells. The model shows the same effects of stretching as in A, but without the CCAN. Because our experiments show that there is no change in stretching, nor in initial length, we show that the stretching effects are near-identical in the model but missing the structures in between the CENP-A nucleosomes.