A telomere-to-telomere assembly of the white mulberry (Morus alba) genome provides insights into karyotype evolution and the white-fruit phenotype in Morus

Dear Editor,

Mulberries (Morus spp.) are economically important fruit trees valued for their nutritional, medicinal, and sericultural applications [1]. The genus exhibits remarkable chromosomal diversity, with haploid chromosome numbers of either n = 7 or n = 14 [2]. To date, telomere-to-telomere (T2T) genomes have been reported for the wild species Morus notabilis (n = 7) [3] and Morus mongolica (n = 14) [4]. However, white mulberry (M. alba), the most widely cultivated species within the n = 14 group, lacks a complete T2T reference genome. Such a resource would address two major limitations. First, unlike most mulberry species that produce pigmented fruits, the M. alba cultivar ‘Zhenzhubai’ produces white fruits, providing a valuable system for investigating the molecular basis of anthocyanin deficiency. Second, previous chromosome-level assemblies of M. alba contain unresolved gaps and lack some centromeric and telomeric regions [5], limiting analyses of genome architecture and structural variation. A complete T2T assembly would enable comprehensive characterization of chromosome structure and facilitate downstream functional and evolutionary analyses of the Morus genus.

In this study, we constructed a gap-free, T2T genome assembly for the M. alba cultivar ’Zhenzhubai’ (n = 14), hereafter referred to as MZB (Fig. 1A). We performed a genome survey using Illumina short-read data to assess K-mer frequency distribution, yielding an estimated genome size of 318.9 Mb, a heterozygosity rate of 1.4%, and a repeat content of 45.8% (Table S1, Fig. S1). To improve assembly contiguity and completeness, we then integrated 34.7 Gb of PacBio HiFi reads, 21.3 Gb of Oxford Nanopore ultra-long reads, and 44.4 Gb of Hi-C data (Table S2, Fig. S2 and S3, Supplementary Methods), successfully bridging the seven gaps initially present among contigs (Table S3 and S4, Fig. S4). The final assembled genome spans 317.6 Mb with a contig N50 value of 21.7 Mb, with all sequences successfully anchored onto 14 chromosomes (Fig. 1B–Table S5). The Hi-C-based chromosome interaction heatmap revealed no conspicuous aberrant interaction loci (Fig. S5A). The assembly exhibits high completeness and accuracy, with a Benchmarking Universal Single-Copy Ortholog (BUSCO) completeness score of 98.0% (Table S6) and a quality value (QV) of 52.9 (Table S7 and S8). The mean long terminal repeat (LTR) Assembly Index (LAI) of the genome assembly is 23.8, above the gold standard level (LAI ≥ 20), confirming the high contiguity of this genome assembly (Table S9, Fig. S5B).

Fig. 1.

Overview of the telomere-to-telomere genome assembly, structural features, comparative genomics, evolutionary history, and anthocyanin-related analyses of MZB. A Field photograph of MZB. B Genome landscape displayed by a Circos plot. From outer to inner rings: chromosomes, gene density, repeat sequence density, and GC content. The lines in the center of the circle indicate syntenic blocks. C Chromosomal landscape of the 14 pseudo-chromosomes. Gene-density distribution is shown along each chromosome, with blue indicating gene-sparse regions and orange indicating gene-rich regions. Gene-sparse blocks correspond to predicted centromeric regions enriched in tandem repeats. D Whole-genome alignment and structural variant identification between MZB and the published M. alba genome using MUMmer and SyRI. E Distribution profiles of transposable elements (TE), tandem repeats (TR) and genes across the chromosome. F Time-calibrated divergence time estimation inferred from single-copy orthologs. G Karyotype Analysis of Morus. H Variations of genes enriched in the flavonoid biosynthesis pathway identified via SV analysis. I Expression profiles of key anthocyanin biosynthetic genes in purple-fruited (‘Da-Shi’) and white-fruited (MZB) accessions across three developmental stages, namely S1, S2, and S3. Different colors represent gene expression levels, ranging from high (orange) to low (green).

We resolved all 14 centromeres (108.1 kb–3.2 Mb, 3.9% of the total chromosomal length) (Table S10) and 28 telomeres (4.5–20.0 kb, 0.2% of the total chromosomal length) (Table S11). Repetitive sequences accounted for 56.5% of the genome, with LTR retrotransposons being the most abundant (26.6%), mainly Gypsy and Copia elements (Table S12, Fig. S6). Using an integrated approach combining transcriptome-based, homology-based, and ab initio predictions, we predicted 25,069 protein-coding genes and subsequently functionally annotated them using public databases (Table S13), achieving a BUSCO completeness score of 98.7% (Table S14). An analysis of tandem ribosomal DNA (rDNA) repeats revealed distinct chromosomal distribution patterns: we identified a continuous 787.4-kb 45S rDNA array harboring 204 intact transcriptional units on the long arm of chromosome 1, whereas other chromosomes harbor only scattered and truncated 18S and 5.8S fragments without large-scale tandem assemblies. By contrast, 5S rDNA loci exhibited a heterogeneous distribution: a prominent 740.4-kb cluster comprising 2189 repeat units on the short arm of chromosome 3, in addition to multiple discrete 5S rDNA loci on chromosomes 1, 4, 5, 7, 9, 10, and 13 (Table S15, Fig. 1C). The two rDNA families thus show distinct chromosomal localizations, indicating divergent amplification and translocation patterns.

When compared with the previously published M. alba genome (the 2020 assembly) [5], our T2T genome assembly resolved many genomic regions across all 14 chromosomes, closing 250 assembly gaps and enabling the annotation of 2320 additional genes (Fig. 1D–Table S16). Although the new and the previous assemblies were broadly syntenic (Fig. S7), we observed substantial large-scale structural divergence on chromosome 1. We therefore performed a detailed annotation of this chromosome, which harbors a large centromeric region enriched with tandem repeats, as well as Gypsy- and Copia-type LTR retrotransposons (Fig. 1E, Fig. S8). The earlier assembly was generated from short reads, likely limiting the resolution of highly homologous repetitive sequences and leading to erroneous contig arrangements [6]. Such assembly artifacts were a major contributor to the detected inconsistencies between the two assemblies along chromosome 1 but were corrected in the new long-read–based assembly [7].

We identified 23,716 structural variants (SVs). Presence/absence variations (PAVs), including deletions (42.4%) and insertions (36.1%), were the most abundant class (Fig. S9A, Table S17). The corresponding PAV-associated genes (i.e., genes present in the T2T assembly but absent or misassembled in the prior gap-containing assembly) were significantly enriched in gene ontology (GO) terms related to ‘ADP binding’, ‘actin filament organization’, and ‘defense response’ (Fig. S9B). Collectively, the MZB assembly reported represents a substantial improvement in genome contiguity, completeness, and annotation accuracy over the previously published genome assembly, providing a robust resource for investigating structural variation and gene evolution in mulberry.

We investigated the evolutionary history of MZB through comparative genomics with 15 closely related species. We identified a total of 61,664 orthologous gene families comprising 471,360 genes, including 123 single-copy families (Table S18, Fig. S10A). Within the MZB genome assembly, we identified 7104 conserved core families and 1769 species-specific families (Table S18). Phylogenetic reconstruction revealed a close relationship between MZB and other M. alba genotypes, with an estimated divergence time of 10.1 (6.0–15.5) million years ago (MYA). The divergence time between MZB and M. notabilis was around 18.7 (12.0–26.0) MYA (Fig. 1F).Whole-genome duplication (WGD) analysis (Fig. S10B) indicated that MZB experienced an ancient duplication event approximately 124 MYA (estimated from the Ks distribution [8]), much earlier than the divergence between M. notabilis and MZB. We determined that chromosome 1 of M. notabilis corresponds to multiple chromosomes in the MZB assembly (Fig. S10C and D). To better understand the underlying karyotypic reorganization following the divergence of Moraceae, we traced the evolutionary trajectory from the ancestral eudicot karyotype (AK) to the ancestral mulberry karyotype (AMK). The transition from a 21-chromosome AK to a 14-chromosome AMK involved five end-to-end joining events, two nested chromosome fusions, and ten reciprocal translocations, collectively resulting in the loss of seven chromosomes (Fig. S11). Two additional Morus species, M. yunnanensis and M. notabilis, each possess six chromosomes. Dot-plot comparisons with the AMK showed that the six chromosomes of these two species are largely syntenic, aside from a few large-scale chromosomal inversions. The transition from the 14-chromosome AMK to six chromosomes in these two species entailed eight end-to-end joining events, accompanied by an eight-chromosome reduction (Fig. 1G, Fig. S12).

Comparative and evolutionary analyses provide a comprehensive framework for understanding the MZB genome. Building on the above results, we focused on gene families related to anthocyanin biosynthesis to explore the molecular basis of the white-fruit phenotype. Specifically, we searched for candidate genes in the flavonoid biosynthesis pathway and characterized SVs affecting these genes in the MZB genome, using the 2020 assembly as the reference. We identified severely damaging SVs in Chalcone synthase (CHS), including stop-gain variants, frameshift deletions, and frameshift insertions (Fig. 1H–Table S19). We also observed a marked expansion of the Flavonol synthase (FLS) gene family and a remarkable contraction of the CHS gene family in MZB (Table S20 and S21). We also performed reverse-transcription quantitative PCR (RT-qPCR) on the core anthocyanin biosynthetic genes in MZB and the purple-fruited cultivar ‘Da-Shi’, which revealed that these genes were consistently expressed at low levels throughout fruit development in MZB, reaching only 0.1% to 10.1% of the levels observed in ‘Da-Shi’. By contrast, FLS was more highly expressed in MZB than in ‘Da-Shi’ (Fig. 1I). Measurements of anthocyanin content confirmed that MZB fruits accumulate very low total anthocyanins during fruit ripening, reaching only 2.4% of the level in ‘Da-Shi’ at the S3 stage (Fig. S13).

Based on these findings, we propose that although most anthocyanin structural genes are present in the MZB genome, their transcription is severely impaired throughout fruit development. Moreover, CHS, the first committed enzyme of the flavonoid pathway, exhibited the highest mutational burden among the genes examined, with stop-gain and frameshift mutations likely impairing its function and limiting the conversion of p-coumaroyl-CoA into downstream flavonoid intermediates [9,10]. Furthermore, the elevated expression of FLS could divert substrates from dihydroflavonol reductase (DFR) [11,12], redirecting metabolic flux toward flavonol production and thereby contributing to the white-fruit phenotype.

This study presents the first T2T genome assembly of white mulberry, providing a high-quality reference for genomic studies of this economically important species. Although this study achieved a T2T-level assembly with high overall contiguity and single-nucleotide accuracy, highly repetitive genomic regions, including tandem rDNA arrays, centromeric repeats, and transposon-enriched regions, continue to pose technical obstacles for genome assembly. Resolving these complex regions remains a common challenge encountered in current T2T genome studies of diverse eukaryotes. Moreover, although we characterized the variation patterns of anthocyanin biosynthesis–related genes at the transcriptional and genomic levels in white mulberry, the underlying regulatory mechanisms—such as the involvement of specific transcription factors, promoter methylation, or chromatin accessibility—remain to be elucidated. Functional validation through transgenic approaches or genome editing will be necessary to establish causal relationships between the identified genomic features and the white-fruit phenotype.

CRediT authorship contribution statement

Fangfang Peng: Writing – review & editing, Writing – original draft, Conceptualization. Xunlan Li: Investigation, Data curation. Zhaoxin Wei: Visualization, Data curation. Guohui Han: Conceptualization. Xiaoyu Wang: Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA029803) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa.