The gap-free genome of Forsythia suspensa illuminates the intricate landscape of centromeres

Abstract

Forsythia suspensa, commonly known as weeping forsythia, holds significance in traditional medicine and horticulture. Despite its ecological and cultural importance, the existing reference genome presents challenges with duplications and gaps, hindering in-depth genomic analyses. Here, we present a Telomere-to-Telomere (T2T) assembly of the F. suspensa genome, integrating Oxford Nanopore Technologies (ONT) ultra-long, Hi-C datasets, and high-fidelity (HiFi) sequencing data. The T2T reference genome (Fsus-CHAU) consists of 14 chromosomes, totaling 688.79 Mb, and encompasses 33 932 predicted protein-coding genes. Additionally, we characterize functional centromeres in the F. suspensa genome by developing a specific CENH3 antibody. We demonstrate that centromeric regions in F. suspensa exhibit a diverse array of satellites, showcasing distinctive types with unconventional lengths across various chromosomes. This discovery offers implications for the adaptability of CENH3 and the potential influence on centromere dynamics. Furthermore, after assessing the insertion time of full-length LTRs within centromeric regions, we found that they are older compared to those across the entire genome, contrasting with observations in other species where centromeric retrotransposons are typically young. We hypothesize that asexual reproduction may impact retrotransposon dynamics, influencing centromere evolution. In conclusion, our T2T assembly of the F. suspensa genome, accompanied by detailed genomic annotations and centromere analysis, significantly enhances F. suspensa potential as a subject of study in fields ranging from ecology and horticulture to traditional medicine.

Figure 1

Summary of the Fsus-CHAU genome and its associated phenotypic characteristics. A Circos plot. Starting from the outermost to the innermost ring: (a) chromosome, (b) gene number density, (c) LTR density, (d) GC content, and (e) syntenic block. B The phenotypes of different tissues, including folwer, fruit, stem, leaf, and the whole plant in Forsythia suspensa. C The karyotype analysis of 14 chromosome pairs in F. suspensa. D Chromatin interaction map generated by Hi-C for the Fsus-CHAU genome.

Figure 2

Collinearity between Fsus-CHAU and weeping forsythia. A Structural variations between the Fsus-CHAU and weeping forsythia genome. B Summary of genomic annotation: the ‘Other’ genomic composition encompasses non-protein-coding genes, unclassified repeat regions, and genomic segments that defy annotation as either genes or repeats. C Proportion of different repeat types in the Fsus-CHAU genome.

Figure 3

Genomic mapping of functional centromeres in Forsythia suspensa. A Multiple alignments of CENH3 homologs from different plant lineages including Zea mays, Oryza sativa, Arabidopsis thaliana, Glycine max, and F. suspensa. The peptide used to generate the anti-CENH3 antibody, with the N-terminal tail and the conserved CENP-A Targeting Domain (CATD) marked on the sequence. B Immunostaining of nucleus in the F. suspensa using the anti-CENH3 antibody. Arrows indicate CENH3 signals. Bar = 10 μm. C FISH assay using CENH3-ChIPed DNA as probes. The boxed insets show high-magnification image of the chromosome indicated by the arrow. Chromosomes are stained with DAPI. Bar = 10 μm. D This schematic overview illustrates the results of the CENH3 ChIP-seq assay in F. suspensa. The mapped reads density was plotted in 100-kb windows along the 14 chromosomes. The x-axis shows the positions on each chromosome. The y-axis represents ChIP/input ratio of the normalized density of uniquely mapped reads on each position, which was calculated based on the ratio of reads per million/100.

Figure 4

Characterization of the diverse centromeric satellites in Forsythia suspensa. A Characteristics of centromeres within F. suspensa genome assemblies. The enrichment of CENH3 [log₂(ChIP/Input)] depicted across centromere 9. B A circos plot depicting tandem repeats identified by TRASH in de novo mode within the Fsus-CHAU genome assembly. C Dot plots comparing the centromeres using a search window of 100 bp. Forward- and reverse-strand similarity are indicated by different points. D FISH signals of CentFs356, CentFs365, and CentFs733 in F. suspensa. Bar = 10 μm. E A histogram illustrating the lengths (bp) of tandem repeats identified in the Fsus-CHAU genome assembly by TRASH. F Proportion of the genome occupied by different tandem repeats.

Figure 5

The LTR-RTs in the centromeres of Forsythia suspensa. A The distribution of full-length LTR-RTs, fragmented LTRs, and centromeric satellites within the extended 20 Mb region surrounding the centromere of Chromosome 7. B Phylogenetic tree of full-length LTR-RTs from the Fsus-CHAU genome, with branches differentiated by their locations within and outside the centromeres. C Dot plot of centromeric full-length LTR-RTs from the Fsus-CHAU genome using a 100-bp search window. D CENH3 enrichment level [log2(ChIP/Input)] around CentFs satellites, centromeric full-length LTR-RTs (n = 295) and non-centromeric full-length LTR-RTs (n = 3176). E Comparison of the insertion timing between full-length LTR-RTs located within centromeres (Cen_LTR-RTs) and those outside of centromeres (Non-cen_LTR-RTs). F Non-centromeric full-length LTR-RTs (Non-cen_LTR-RTs): the histogram displays the distribution of LTR identity percentages for non-cen_LTR-RTs, with a sample size of 3176 elements. The median LTR identity is 0.982, as indicated by the vertical red line. The x-axis represents the percentage of LTR identity, while the y-axis shows the count of LTR-RTs within each identity range. G Centromeric full-length LTR-RTs (Cen_LTR-RTs): the histogram shows the distribution of LTR identity percentages for centromeric LTR-RTs, with a sample size of 295 elements. The median LTR identity is 0.979, marked by the vertical red line. The x-axis represents the percentage of LTR identity, and the y-axis indicates the count of LTR-RTs within each identity range.