Chromosome engineering allows the efficient isolation of vertebrate neocentromeres

Abstract

Centromeres are specified by sequence-independent epigenetic mechanisms in most organisms. Rarely, centromere repositioning results in neocentromere formation at ectopic sites. However, the mechanisms governing how and where neocentromeres form are unknown. Here, we established a chromosome-engineering system in chicken DT40 cells that allowed us to efficiently isolate neocentromere-containing chromosomes. Neocentromeres appear to be structurally and functionally equivalent to native centromeres. Chromatin immunoprecipitation sequencing (ChIP-seq) analysis with 18 neocentromeres revealed that the centromere-specific histone H3 variant CENP-A occupies an ∼40 kb region at each neocentromere, which has no preference for specific DNA sequence motifs. Furthermore, we found that neocentromeres were not associated with histone modifications H3K9me3, H3K4me2, and H3K36me3 or with early replication timing. Importantly, low but significant levels of CENP-A are detected around endogenous centromeres, which are capable of seeding neocentromere assembly if the centromere core is removed. In summary, our experimental system provides valuable insights for understanding how neocentromeres form.

Graphical abstract

Figure 1

Formation of Neocentromeres on Chromosome Z (A) A strategy to isolate surviving cells after removal of endogenous centromere of chromosome Z with positive (Zeocin resistance) and negative (FIAU resistance) selections. Surviving cells are expected to have chromosome Z with a neocentromere or fusion chromosome Z with another chromosome. (B) Southern hybridization analysis to confirm that endogenous centromere sequence is removed. Probe information (probe #1–5) for Southern analysis to confirm removal of endogenous centromere is also shown. (C) Frequency of isolation of surviving cells without endogenous centromere sequence. We independently performed four trials. (D) Classification of surviving clones based on cytological experiments using anti-CENP-T (CT) antibodies as a centromere marker. Among 136 clones, ten clones have fusion chromosome Z with another chromosome. We classified the centromere position defined by CENP-T staining as five types (type I–V). (E) Immuno-FISH images of various types of chromosomes. Red shows signals from the satellite sequence on q-arm of chromosome Z. Centromeres (green) are visualized by anti-CENP-T antibodies. See also Figure S1.

Figure 2

Cells with Neocentromere Show Normal Chromosome Segregation (A) Growth curve of cells with neocentromere. Z#3 has endogenous centromere on chromosome Z (control). #BM23, #0303, #0727, #BL02, #0514, and #1024 cells, each of which has a neocentromere on chromosome Z, grew well as Z#3 cells. (B) Time to complete mitotic progression, population of cells with misaligned chromosome, percentage of aneuploid cells, and numbers of cells in which sister chromatids of chromosome Z are prematurely separated in each neocentromere-containing cell line. (C) Typical images of cells with either aligned or misaligned chromosome Z. Chromosome Z was detected by FISH the satellite sequence on q-arm of chromosome Z as a probe (red). (D) A typical image of cell in which sister chromatids of chromosome Z are prematurely separated during prometaphase (right). Left shows a cohered chromosome Z (control). (E) Immunofluorescence analysis on neocentromeres with antibodies against CENP-A, -C, -O, -E, -T, Ndc80, Aurora B, and KNL2. All tested centromere proteins were detected on all neocentromeres. See also Figure S2.

Figure 3

Neocentromeres Are ∼40 kb Long (A) ChIP-seq analysis with anti-FLAG or anti-CENP-A antibodies in cells containing neocentromere. Size of the CENP-A domain in each cell line was shown. We used anti-FLAG antibodies for cells expressing CENP-A-FLAG (FLAG-IP). For cells not expressing CENP-A-FLAG, we used native CENP-A antibodies (CENP-A-IP). IP DNAs were deeply sequenced and sequence data were mapped on chicken genome database. We first identified a major peak as a neocentromere (each position is indicated) from entire chromosome Z in each cell line and examined detail distribution around the peak with a higher resolution. GC% contents and distribution of transposons, repeat sequence, and genes in CENP-A-associated DNAs are shown. Arrow in data of #0514 cells indicates a gap of CENP-A distribution, which corresponds to exon 1 of MAMDC2 gene (also see Figure 7). (B) Counts of sequence reads (gray bar) for CENP-A-IP DNAs around centromere (Z#3) and noncentromere (#BM23) region. Ranking for these counts are also shown (black line). Pink dots are shown as neocentromere loci. CENP-A-associated DNAs were enriched in 2 Mb around the major CENP-A peak (Z#3). However, numbers of sequence tags associated with CENP-A were reduced in this region, when an endogenous centromere was removed (#BM23). (C) Counts of sequence reads (gray bar) for CENP-A-IP and their ranking (black line) around centromere (#BM23) and noncentromere (Z#3) region. CENP-A cluster was observed around centromere (#BM23). (D) Genome-wide ranking of numbers of sequence tags associated with CENP-A for Z#3 cells were subtracted from that for #BM23. A CENP-A cluster was observed around neocentromere region of #BM23. See also Figures S3 and S4.

Figure 4

Neocentromere Formation on Chromosome 5 (A) A strategy to isolate surviving cells after removal of endogenous centromere of chromosome 5 with positive (Zeocin resistance) and negative (FIAU resistance) selections. In this case, we used shorter deletion (67 kb) than for Z centromere deletion (127 kb). (B) Location of neocentromeres on chromosome 5. These locations were determined by ChIP-seq analysis with anti-CENP-A. Ninety-seven percent of neocentromeres are located in the 3 Mb region from the original centromere. (C) Counts of sequence reads (gray bar) for CENP-A-IP and their ranking (black line) around centromere 5. A CENP-A cluster was observed around centromere and neocentromeres (pink) are formed in the CENP-A cluster region.

Figure 5

Chromatin Features in Various Neocentromeres (A) ChIP-seq analysis with anti-H3K9me3 or anti-CENP-A in Z#3 cells. In chromosomes 1 and 2 that have repetitive centromeres, H3K9me3 was enriched around the CENP-A domain. H3K9me3 was not enriched around nonrepetitive centromeres of chromosomes Z, 5, and 27. (B) ChIP-seq analysis with anti-H3K9me3 or anti-CENP-A in neocentromere-containing #BM23 and #0514 cells. H3K9me3 was not enriched around neocentromere region in both cell lines. (C) ChIP-PCR analysis with antibodies against H3K4me2 and H3K36me2 around centromere regions in Z#3 (chromosome Z), #BM23 (Z neocentromere), and #0514 (Z neocentromere). In #BM23 and #0514 cells, these profiles are shown before and after neocentromere formation. Primer positions for this analysis and ChIP-seq data with anti-CENP-A antibodies in each cell line are also shown. Error bars show SEM. See also Figure S5.

Figure 6

DNA Reapplication Timing of Neocentromeres (A) DNA replication profiles around endogenous centromere regions on chromosomes Z, 5, and 27 in DT40 cells. Endogenous centromeres are on middle to late replication domain. (B) Comparison of data of entire DNA replication timing for Z#3 (before neocentromere formation) with those for BM23, #0514, and #1024 cells (after neocentromere formation). R² values are shown. (C) Changes of DNA replication timing at three neocentromere loci before and after neocentromere formation. By SAM, later shifts in replication timing of 100 kb segments at loci of neocentromere formation in both #0514 and #1024 cells were found to be statistically significant (#0514: p = 1.86 x 10⁻⁵, q = 0.0153; #1024: p = 7.79 x 10⁻³, q = 0.0903; q value is a FDR-based measure of significance; Storey and Tibshirani, 2003). Replication timing was not changed in #BM23 cells upon neocentromere formation.

Figure 7

Neocentromeres Are Efficiently Formed on Nontranscribed Region in Chromosome Z (A) Examination of expression for genes, which are located on each neocentromere, in early embryos or wild-type DT40 cells. Expression was analyzed by RT-PCR. Most genes except for MAMDC2 are not expressed in DT40 cells, but expressed in early embryonic cells. (B) Quantitative-RT-PCR analysis of MAMDC2 in wild-type DT40 or #0514 cells in which a neocentromere is formed on MAMDC2 gene locus (see Figure 3A). #0514-CAf cells are expressing CENP-A-FLAG. Expression level of MAMDC2 extremely was reduced in #0514 cells. Error bars show SD. (C) Position of each neocentromere and native centromere on chromosome 5. Location of genes around neocentromeres is shown. Many neocentromeres are formed on transcribed genes in chromosome 5.