CENP-A Is Dispensable for Mitotic Centromere Function after Initial Centromere/Kinetochore Assembly

Abstract

Human centromeres are defined by chromatin containing the histone H3 variant CENP-A assembled onto repetitive alphoid DNA sequences. By inducing rapid, complete degradation of endogenous CENP-A, we now demonstrate that once the first steps of centromere assembly have been completed in G1/S, continued CENP-A binding is not required for maintaining kinetochore attachment to centromeres or for centromere function in the next mitosis. Degradation of CENP-A prior to kinetochore assembly is found to block deposition of CENP-C and CENP-N, but not CENP-T, thereby producing defective kinetochores and failure of chromosome segregation. Without the continuing presence of CENP-A, CENP-B binding to alphoid DNA sequences becomes essential to preserve anchoring of CENP-C and the kinetochore to each centromere. Thus, there is a reciprocal interdependency of CENP-A chromatin and the underlying repetitive centromere DNA sequences bound by CENP-B in the maintenance of human chromosome segregation.

Keywords: CENP-A; CENP-B; CENP-C; auxin; centromere; chromosome segregation; epigenetic; kinetochore; mitosis; protein degradation.

Figure 1. Complete and rapid removal of the centromere epigenetic mark CENP-A in human cells

(A) Schematic of the TALEN-mediated genome editing to endogenously tag CENP-A with AID (A) and EYFP (E) in the indicated cell lines. Position of exons, introns and start/stop codons are indicated. (B) Immuno-blot of RPE-1 cells following treatment with Auxin (IAA) at the indicated time points (hours). α-tubulin was used as a loading control. A schematic of the soluble CENP-A-associated factors is also shown. (C) Representative immunofluorescence images on DLD-1 cells showing CENP-A depletion after 24 hours treatment with IAA. CENP-B was used to mark centromere position. (D) Quantification of the experiment in C by using an antibody against CENP-A or by monitoring EYFP intensity. Unpaired t test: *** p < 0.0001. (E) Degradation kinetics of CENP-A in RPE-1 cells with or without (+/-) IAA treatment measured by EYFP intensity during live cell imaging. IAA was added at the microscope stage. Σn = 10 cells (F) Representative images show immuno-fluorescence on RPE-1 cells expressing the cell cycle indicator FUCCI. Red (G1), yellow (S) or green (S/G2) circles indicate the cell cycle position +/− IAA treatment. A schematic of the experimental designed is also shown. (G) Images of representative crystal violet–stained colonies from the colony formation assay +/− IAA treatment in RPE-1 cells. (H) Cell counting experiment on DLD-1 CENP-A^−/EA cells +/− IAA treatment. IAA was added at day 0 and kept for a maximum of 7 days. Error bars represent the SEM of five independent experiments. Unpaired t test: ** p = 0.0026. See also Supplementary Fig. S1. Scale bars = 5 um.

Figure 2. CENP-A is dispensable for centromere/kinetochore maintenance and chromosome segregation once centromere is assembled

(A) Schematic of the experiments shown in B–C. (B) Bar graph shows the percentage of chromosome mis-segregation events by live cell imaging in non-treated conditions or following IAA treatment for 2 or 24 hours, respectively. Error bars represent the SEM of three independent experiments. Individual Σn = ~35 cells. Unpaired t test: ** p = 0.0099 and 0.0065. (C) Each individual point represents a single cell. Time in mitosis was defined as the period from NEBD to chromosome decondensation. Error bars represent the SEM of three independent experiments. Unpaired t test: **** p < 0.0001. (D) Bar graph shows the frequency of chromosome mis-segregation (micronuclei and mis-aligned chromosomes, 2 hour IAA treatment condition) versus the cell cycle phase in which CENP-A was depleted (determined by measuring the time required for the cells to enter into mitosis in asynchronous population, using cell co-expressing PCNA^GFP as an indicator of the cell cycle or by synchronizing cell in G1 by adding IAA either at t= 0 hr or at t = 9 hr) in DLD-1 cells. Unpaired t test: * p = 0.0142, ** p = 0.0024. (E) Schematic of the experiments shown in F–G. (F) Bar graphs showing centromere intensities of CENP-A, CENP-B, CENP-C, CENP-I and CENP-T for the indicated cell line following IAA treatment. Values represent the mean of six independent experiments combining analysis performed in RPE-1 and DLD-1 cells. Individual Σn = ~30 cells, Σn = 25 centromeres per cell. Error bars represent the SEM (standard error of the mean). Unpaired t test: *** p < 0.0001, ** p < 0.07, * p < 0.05. (G) Box & whisker plots of Dsn1 or Hec1 intensities at the centromere measured on metaphase spreads. Unpaired t test: * p = 0.017, *** p = 0.0005, **** p < 0.0001. See also Supplementary Fig S2

Figure 3. CENP-A is required to regulate early steps of centromere assembly, such as for CENP-C and CENP-N, but not CENP-T

(A) Schematic of the experiments shown in B and C. (B) Representative immunofluorescence images show de novo CENP-C^AID-mRFP loading at centromeres. (C) Bar graphs showing centromere intensities of CENP-C^AID-mRFP in the indicated cell lines. Error bars represent the SEM of three independent experiments. Individual Σn = ~30 cells, Σn = 25 centromeres for cell. Unpaired t test: *** p < 0.0001. (D) Schematics of the experiments shown in E–G. (E) Representative immunofluorescence images show de novo CENP-N^3HA-SNAP loading at centromeres. HA antibody was used to identify unlabeled CENP-N. (F–G) Bar graphs showing centromere intensities of CENP-N^3HA-SNAP or CENP-T^3HA-SNAP respectively in the indicated cell lines. Values represent the mean of three independent experiments. Error bars represent the SEM. Individual Σn = ~30 cells, Σn = 25 centromeres for cell. Unpaired t test: * p = 0.01, ** p = 0.0028, *** p = 0.0003. See also Supplementary Fig S2. Scale bars = 5 um.

Figure 4. CENP-B is sufficient and essential to maintain kinetochore assembly and consequently faithful chromosome segregation in the absence of CENP-A

(A) (left) Representative images show a micronucleus containing the Y chromosome by dual FISH analysis. (Right) Graphs show the frequency of micronuclei formation (X axis) versus the frequency of a micronucleus containing the chromosome Y or chromosome 4 +/− IAA treatment for 24 hours. n = ~400 cells. Unpaired t test: * p = 0.01, ** p = 0.0094 (B) Representative images of an immuno-fluorescence coupled with FISH showed CENP-C binding to centromere on the Y or X chromosome +/− IAA treatment for 24 hours. Scale bar = 5 μm. (C) Schematic of the experimental design shown in D-G. (D) Immuno-blot shows depletion of endogenous CENP-B using the CRISPR technology. α-tubulin was used as a loading control. (E) Cell counting experiment on RPE-1 +/− IAA treatment and/or CENP-B gene. IAA was added at day 0 and kept for a maximum of 7 days. Error bars represent the SEM of four independent experiments. Unpaired t test: ** p = 0.0043; 0.0016. (F) Representative immuno-fluorescence FISH to measure CENP-C levels following CENP-A depletion (by IAA) in CENP-B depleted cells. A FISH probe against CENP-B boxes was used to mark centromere position. (G) Quantification of the experiment show in D. Each dot represents an average of 25 centromeres in a single cell. Unpaired t test: *** p < 0.0001 (H) Schematic of the experiments shown in I–K. (I) Bar graph shows the percentage of chromosome mis-segregation events observed by live cell imaging following siRNA depletion of GAPDH or CENP-B and IAA treatment for 2 hours, respectively. Error bars represent the SEM of three independent experiments. Individual Σn = ~60 cells. Unpaired t test: * p = 0.02, ** p = 0.0068 (J) Bar graph shows the number (1 or >2) of mis-aligned chromosomes in percentage from analysis in E. Error bars represent the SEM of three independent experiments. Unpaired t test: ** p = 0.0093. (K) Scatter plot graph shows the time in mitosis (from NEBD to chromosome decondensation). Each individual point represents a single cell. Error bars represent the SEM of three independent experiments. Unpaired t test: *** p < 0.0001. (L) Box & whisker plots of Dsn1 and Hec1 intensities at the centromere measured on metaphase spreads. Unpaired t test: ** p = 0.002, *** p = 0.0005. See also Supplementary Fig S3 and S4. Scale bars = 5 μm.

Figure 5. Model of centromere function mediated by centromeric chromatin and DNA sequences

At exit of mitosis, centromeric chromatin replication and identity is mediated by CENP-A (in red) deposition via interaction with HJURP. CENP-A then mediates the assembly of CENP-C (in green) in mid-G1 followed by CENP-N/L (in orange) during S-phase. These steps might be interconnected. At this point, CENP-A becomes dispensable for mitotic centromere function as long as CENP-B (in light blue) is stably bound to centromeric sequences to support CENP-C binding. Assembly of the other subunits of the CCAN, such as CENP-T/W (in yellow) and HIKM (in brown), allows the full recruitment of the kinetochore complex (in grey) required to mediate centromere function. In summary, we propose that the kinetochore is tethered to the centromere through a dual linkage of CENP-A chromatin and CENP-B-bound DNA sequences, as the two major links from the DNA to the kinetochore to mediate successful chromosome segregation.