Distinct evolutionary trajectories of subgenomic centromeres in polyploid wheat

Abstract

Background: Centromeres are crucial for precise chromosome segregation and maintaining genome stability during cell division. However, their evolutionary dynamics, particularly in polyploid organisms with complex genomic architectures, remain largely enigmatic. Allopolyploid wheat, with its well-defined hierarchical ploidy series and recent polyploidization history, serves as an excellent model to explore centromere evolution.

Results: In this study, we perform a systematic comparative analysis of centromeres in common wheat and its corresponding ancestral species, utilizing the latest comprehensive reference genome assembly available. Our findings reveal that wheat centromeres predominantly consist of five types of centromeric-specific retrotransposon elements (CRWs), with CRW1 and CRW2 being the most prevalent. We identify distinct evolutionary trajectories in the functional centromeres of each subgenome, characterized by variations in copy number, insertion age, and CRW composition. By utilizing CENH3-ChIP data across various ploidy levels, we uncover a series of CRW invasion events that have shaped the evolution of AA subgenome centromeres. Conversely, the evolutionary process of the DD subgenome centromeres involves their expansion from diploid to hexaploid wheat, facilitating adaptation to a larger genomic context. Integration of complete einkorn centromere assemblies and Aegilops tauschii pan-genomes further revealed subgenome-specific centromere evolutionary trajectories. By inclusion of synthetic hexaploid from S₂-S₃ generations, alongside 2x/6 × natural accessions, we demonstrate that DD subgenome centromere expansion represents a gradual evolutionary process rather than an immediate response to polyploidization.

Conclusions: Our study provides a comprehensive landscape of centromere adaptation, evolution, and maturation, along with insights into how retrotransposon invasions drive centromere evolution in polyploid wheat.

Keywords: Triticum aestivum; CENH3; CRW; Centromere evolution; FlLTR-RTs.

Fig. 1

Fine structure of centromere repeat arrays in hexaploid wheat T. aestivum. Characteristics of Cen5A (A), Cen5B (B), and Cen5D (C) are shown. Different layers display the CENH3 enrichment [log₂(ChIP/Input)], centromeric satellites centT550 and CentT566 distribution, centromeric flLTR-RTs annotations, and a heatmap of pairwise satellite sequence similarity. D Dot plots comparing the centromeres from AA, BB, and DD Subgenome assembly using a 500-bp search window. Red, green, and blue lines/boxes represent the centromeric regions from the AA, BB, and DD subgenomes, respectively. Red and blue points indicate forward- and reverse-strand similarity. E Distribution of LTR-RTs in the T. aestivum AA, BB, and DD Subgenomes across the 21 centromeres. F Gypsy Superfamily is evenly distributed across the 21 functional centromere regions. G Copia superfamily is enriched in the first homoeologous group, especially Cen1D

Fig. 2

Centromeric-specific retrotransposon element (CRWs) composition and insertion history in common wheat. A Phylogenetic tree based on flLTR-RTs from T. aestivum centromeres, with centromere-specific RTs sequences from corn, rice, and rye as outgroups. Among the five CRWs specifically presenting in the functional centromeres of T. aestivum, CRW1 is shown in red, CRW2 is in green, CRW3 is in blue, CRW4 is in yellow, and CRW5 is in black. B The insertion time of CRWs on homoeologous group Chr1 to Chr7, shown as histograms. Different colors represent CRWs from different wheat sub-centromeres. C Older CRWs were detected in the D subgenomic centromeres compared to the A and B subgenomic centromere. Statistical analyses were completed using the Mann–Whitney U test. Asterisks represent statistically significant differences (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001) between the indicated groups. D Copy number of incipient CRWs among three sub-centromeres. E The insertion time of CRWs, as sampled in three sub-centromeres of T. aestivum. CenA in red; CenB in green; and CenD in blue. The x-axis represents the time (Ma) of CRW insertion into the host genome; the y-axis represents the density of insert time. The bottom blue triangle Marks the time when CRW burst began to be detected in CenD, with yellow Marking CenA and CenB. The second polyploid event occurred 8000 years ago (ka)

Fig. 3

A prospective evolutionary model for common wheat CenA-CS. A The detailed layout of the CENH3 enrichment [log₂(ChIP/Input)] ratio along the pericentromeric region of bread wheat chromosome 4A. The first four tracks (1–4) are T. monococcum (A^mA^m, accessions TA299, TA10622, KT003-002, and KU104). The next three (5–7) are T. urartu (A^uA^u, accessions TMU06, G1812, and TMU38). Tracks eight and nine (8–9) are wild emmer wheat (WEW, accession Zavitan) and domesticated emmer wheat (DEW, accession Svevo), belonging to T. turgidum (BBAA). Tracks ten to twelve (10–12) are T. aestivum (accessions CS, TAA10, and Aikang58 (AK58)). The distribution of CRW1 (red), CRW2 (green), CRW3 (blue), CRW4 (yellow), CRW5 (black), and other flLTR-RTs (gray) are shown at the following tracks, respectively. B Dot plot alignments (300-bp window) of pericentromeric and centromeric regions for chromosomes 4 A, comparing T. monococcum (domesticated einkorn TA10622; wild einkorn TA299), T. urartu (G1812), and T. aestivum (CS). Genomic coordinates (top) and CENH3 ChIP-seq coverage profiles (middle) demarcate centromere boundaries. Distributions of CRWs are annotated as colored bars. Sequence similarities on forward and reverse strands are indicated in red and blue, respectively. C The insertion time of CRWs associated with corresponding subregions, analyzed by boxplot. Two successive polyploid events are indicated on the timescale in the left panel. The evolutionary trajectories of T. aestivum AA subgenome (TaA) and T. urartu (Tu) are shown as separate black lines. D A predictive evolutionary model of AA sub-centromeres in allohexaploid wheat. Episodic invasions by CRWs (black arrows) displace ancestral centromeric repeats to pericentromeric domains, driving dynamic redistribution of CENH3 nucleosomes between ancestral CRW clusters (blue shading) and emerging CRW arrays (green shading). Epigenetic competition culminates in dominant CENH3 occupancy at nascent retrotransposon arrays, while relocating ancient CRWs to unilateral or bilateral pericentromeric regions. Through continuous integration of novel CRW insertions and reorganization of ancestral repeats, this process establishes an evolutionarily dynamic CRW cluster architecture that ultimately shapes the functional centromeres of the AA subgenome

Fig. 4

The evolutionary trajectory of Cen1D. A-B The detailed layout of the CENH3 enrichment [log₂(ChIP/Input)] ratio in centromeric regions of T. aestivum CS (A) and Ae. tauschii DD (B). Ae. tauschii (accession Y2282) is shown in cyan; CS in orange. The distribution of CRW1 (red), CRW2 (green), CRW3 (blue), CRW4 (yellow), CRW5 (black), and other flLTR-RTs (gray) in pericentromeres is shown in the respective tracks. C The insertion time of CRWs within chromosome 1D Subregions 1 and 2 in CS. The median insertion time in Region 1 is 1.20 Ma, and 0.23 Ma in Region 2. D,E Box plot (D) and kernel density estimate plot (E) of CRW insertion time in Ae. tauschii cen1D (cyan) and T. aestivum cen1D (orange). Two polyploid events are marked by dashed lines

Fig. 5

The evolutionary trajectory of Cen4D. A The detailed layout of the CENH3 enrichment [log₂(ChIP/Input)] ratio in the centromeric regions of CS (above) and Ae. tauschii (below). Ae. tauschii (accession Y2282) is shown in cyan; CS in orange. The distribution of CRW1 (red), CRW2 (green), CRW3 (blue), CRW4 (yellow), CRW5 (black), and other flLTR-RTs (gray) in pericentromeres is shown at their respective tracks. B The insertion time of CRWs associated with Ae. tauschii CENH3-ChIP samples and CS CENH3-ChIP samples in corresponding chromosome 4D subregions, analyzed by boxplot. The enriched region observed from mapping CENH3-ChIP samples of Ae. tauschii to the CS reference genome (DD-CS-Cen4D) is highlighted in violet, while the centromere 4D in Ae. tauschii (DD-Cen4D) is marked in brown. C RT phylogenetic relationship of the CRWs in DD-CS-Cen4D and CS-Cen4D. CRWs from DD-CS-Cen4D are shown in red and green, and CRWs from CS-Cen4D are shown in black. D Kernel density estimate plot of the time when flLTR-RTs inserted in the 4D pericentromere of T. aestivum. CRW1 is shown in red, CRW2 in green, CRW3 in blue, CRW4 in yellow, and all other types of flLTR-RTs in gray

Fig. 6

Evolutionary model of DD sub-centromeres in common wheat. A The detailed layout of the CENH3 enrichment [log₂(ChIP/Input)] ratio in the centromeric regions of CS (above) and Ae. tauschii (below). Ae. tauschii (accession Y2282) is shown in cyan; CS in orange. The distribution of CRW1 (red), CRW2 (green), CRW3 (blue), CRW4 (yellow), CRW5 (black), and other flLTR-RTs (gray) in pericentromeres is shown at the following tracks, respectively. B The insertion time of CRWs associated with Ae. tauschii CENH3-ChIP samples and CS CENH3-ChIP samples in corresponding chromosome 6D subregions, analyzed by boxplot. C Relative plot showing the difference in size between CENH3-binding region of four Ae. tauschii (cyan) and T. aestivum (orange) among Chr1-7D. The axis represents centromere size (megabases). The light color represents Ae. tauschii from the L2 lineage, the dark color represents the L1 lineage, and different accessions are marked after the lineage name. D Dynamics of DD subgenome centromeres in allohexaploid wheat. Expansion arrows (purple) indicate epigenetic expansion of CENH3 domains into pericentromeres (purple hatched areas: ancestral CRW regions). Partial CENH3 occupancy (red circles) in ancestral pericentromeric regions reflects functional plasticity, facilitating adaptation to genomic complexity (see bimodal CENH3 distribution in Fig. 2E). E The box plot of CRW insertion time in Ae. tauschii (cyan) and T. aestivum (orange) among homoeologous groups Chr1 to Chr7. F Kernel density estimate plot of flLTR-RTs along DD sub-pericentromeres of T. aestivum in homoeologous groups Chr1 to Chr7. The pericentromere 1D (Peri1D) is red, Peri2D is green, Peri3D is blue, Peri4D is black, Peri5D is navy blue, Peri6D is cyan, and Peri7D is purple

Fig. 7

CENH3 nucleosomes showed limited deposition in pericentromeres of resynthesized hexaploid wheat. A The detailed layout of the CENH3 profile in Ae. tauschii (Chr1-7D). From top to bottom are shown as [log₂(ChIP/Input)] ratio data of paternal line Y2282, S₂, and S₃. B, C Correlation analysis. The centromeric localizations of CENH3 nucleosomes on the paternal Ae. tauschii Chr1-7 chromosomes showed a significant positive correlation between the paternal line Y2282 and S₂ (B), and the paternal line Y2282 and S₃ (C), respectively