Gap‐free telomere‐to‐telomere assembly of the <i>Mangifera persiciforma</i> genome and its evolutionary insights on resistance

Mangifera persiciforma, a member of the Mangifera genus, is the same family as mangos (Mangifera indica) often referred to as the ‘king of fruits’. It is primarily distributed across the Guangxi Zhuang Autonomous Region, as well as Yunnan and Guizhou provinces in China (Wu et al., 2021), in which they thrive in barren regions with karst landforms, indicating they possess exceptional environmental adaptability. Compared with mangos, M. persiciforma produces smaller and unpalatable fruit (Figure 1a). Due to close ties of consanguinity, M. persiciforma can serve as a rootstock for mangos, helping to improve resistance and expand planting areas. Understanding the genetic background of M. persiciforma can provide new insights into the evolution of Mangifera species and help improve agricultural traits of mango. Here, we achieved a gapless telomere-to-telomere (T2T) assembly of the M. persiciforma genome, marking the first T2T genome in the Mangifera genus. We also conducted a comparative genomic analysis to investigate evolutionary trends, which provides important data for the cultivation and breeding of Mangifera genus. We assembled the M. persiciforma genome using a standard pipeline (Figure S1) and found that the estimated genome size is 373 534 488 bp, with a heterozygosity rate of 2.35% (Figure S2; Table S1). Preliminary assembly using PacBio HiFi reads and ONT ultra-long reads resulted in primary contigs with a size of 387 617 296 bp and an N50 value of 20 124 366 bp (Table S3). Subsequently, Hi-C technology was employed to anchor these contigs onto 20 pseudo-chromosomes (2 N = 2X = 40; Figure S4; Tables S4 and S5). Telomeres and centromeres of the scaffolds were identified, and gap regions were filled (Tables S6 and S7). The final gapless/gap-free assembled T2T M. persiciforma genome size is 382 840 341 bp (Figures S6 and S7; Tables S8 and S9). Evaluation using Benchmarking Universal Single-Copy Orthologs (BUSCO) indicated that the completeness of the assembled genome is 98.8% (Table S11). Finally, the size of repetitive sequences in the M. persiciforma genome is 158 564 556 bp, accounting for 41.41% of the total genome (Table S12). Gene density increases towards the ends of the chromosomes compared with non-coding elements. Additionally, the M. persiciforma chloroplast genome was also assembled, resulting in a chloroplast genome size of 158 844 bp with 152 predicted genes (Figure S8). In summary, this assembly represents a high-quality, gapless T2T M. persiciforma genome, providing a solid foundation for subsequent comparative genomic studies with mangos. The evolutionary relationship between M. persiciforma and mangos was analysed using single-copy orthologous genes (Figure 1b; Figure S9), revealing that five species from the Anacardiaceae family clustered into one group. Meanwhile, three species belonging to the Mangifera genus, including M. persiciforma, M. indca ‘Alphonso’ (Alphonso; Wang et al., 2020), and M. indca ‘Hongxiangya’ (Hongxiangya; Li et al., 2020), formed a clade (Figure 1b). It indicated that the divergence between M. persiciforma and mangos occurred ~ 14 million years ago. Further analysis showed that M. persiciforma experienced two whole-genome duplication events (Figures S10 and S11) same as mangos (Li et al., 2020). The results of the phylogenetic tree analysis suggest that M. persiciforma is a close relative of mangos and is more closely related to their ancestral species. Previous studies have indicated that the expansion and contraction of gene families are closely related to the strength or weakening of traits in species (Li et al., 2020). We found that the expansion of gene families in M. persiciforma is primarily associated with the pathways of environmental adaptability, including terpene synthesis, phenylpropanoid biosynthesis, and ABC transporter (Figure 1c, Figure S12; Tables S15 and S16). Terpenes are end products in the terpene synthesis pathway catalysed by terpene synthase (TPS) proteins that enhance biotic and abiotic resistance in plants (Jia et al., 2019). Notably, M. persiciforma contains significantly more TPS gene family members than mangos, with 74 TPS genes in M. persiciforma, while Hongxiangya and Alphonso have 48 and 53 TPS genes, respectively (Figure 1d; Figures S14 and S15). Additionally, the phylogenetic tree and sequence analysis revealed six TPS genes specific to M. persiciforma that exhibit certain exon sequence deletions (Figure S16). We compared the existing expression data of multi-tissue M. persiciforma and Alphonso, removing the non-expressed Mper22894. The results showed that these TPS genes in M. persiciforma were indeed specifically expressed in roots, which was completely different from the TPS gene pattern in mango. This indirectly indicates that the deletion of this sequence may indeed affect the expression level of genes (Figure S17). Moreover, understanding the various types of gene duplication can provide insights into crucial environmental adaptations (Qiao et al., 2019). The gene pairs resulting from proximal duplications (2207) and tandem duplications (3019) in M. persiciforma were subject to stronger positive selection (Figure 1e). Additionally, we find that TPS genes are unusually active during the evolution of M. persiciforma. Compared with whole-genome duplication, the expansion of TPS genes in the evolutionary development of M. persiciforma tends to occur through proximal duplications and tandem duplications (Figure S19). It implies that the specific expansion of the TPS gene family, along with gene sequence variants in M. persiciforma, may be an important factor contributing to its greater environmental adaptability compared with mangos. Unique genes in the genome often bring more possibilities (Liu et al., 2022). Through the comparison analysis of homologues and collinear genes between M. persiciforma and mango varieties, we identified 1009 unique genes in M. persiciforma (Figure 1f). Interestingly, we found these unique genes are mainly enriched in environmental-adaption pathways related to plant-pathogen interactions (Figure S20; Tables S29 and S30). Additionally, we screened 19 highly expressed genes through tissue-specific expression in M. persiciforma (Figure S21), of which 17 belonged to the NLR (nucleotide-binding leucine-rich repeat) gene family, and the plant NLR gene family functions as intracellular immune receptors that recognize pathogen effector proteins and activate immune responses to resist pathogen infection (Wang et al., 2023). These genes were expressed in roots, fruit, stems, and leaves, with particularly significant expression in stems. We further analysed the hypergeometric distribution and found that these NLR gene family members are more likely to have originated from proximal duplications (Figure S22), and here is a typical example that there are two most prominent NLRs located in the non-collinear segments of chromosome 2 (Figure 1g,h; Figure S23). These results suggest that the unique resistance genes (NLRs) of M. persiciforma are likely to have arisen from proximal duplications. This work was supported by the Guangxi Natural Science Foundation (2024GXNSFBA010399), Science and Technology Talent Special Project of Guangxi (Gui Ke AD23026320), Funding from State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources (SKLCUSA- b202302) and The CARSGIT—Guangxi Mango Industry Project (nycytxgxcxtd-2021-06-02). The authors declare no conflict of interest. B-J. L. and L.X. conceived and designed this project. J. W. and R-X. B. analyzed the data and wrote the manuscript. Other authors carried out experiments and revised the article. The raw sequencing data generated in this study has been uploaded to the GSA database with the accession number CRA021696. The genomes and annotations of M. persiciforma have been uploaded to Figshare under DOI: 10.6084/m9.figshare.28123553. Figure S1–S23 Supplementary Figures. Table S1–S30 Supplementary Tables. Data S1 Supplementary Data. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.