A 15.8‐Mb Alien Radish Chromosomal Fragment Inversion Drives Fertility Restoration and Telomere Loss of <scp>C09</scp> in <i>Brassica oleracea</i>

Cabbage (Brassica oleracea L. var. capitata) is one of the most important vegetables grown globally (Li et al. 2024). Currently, more than 90% of cabbage cultivars are male sterile. Male sterility-based hybridisation facilitates heterosis utilisation and yield enhancement (Han et al. 2023). Ogura cytoplasmic male sterility (CMS) is more commonly used due to its maternal inheritance and easy transferability (Ren et al. 2020, 2022). However, CMS varieties cannot be self-pollinated, thereby limiting germplasm innovation and restricting genetic diversity, which can be solved by fertility restorer lines (Li et al. 2021; Ren et al. 2022). The Rfo gene exists only in radish but not in any B. oleracea crops. This gene has been transferred from radish into rapeseed, a related species of B. oleracea, providing a promising resource for creating cabbage Ogura CMS fertility restorer lines (FRLs) (Yu et al. 2020). In our previous studies, Rfo was successfully transferred from rapeseed into Chinese kale and cabbage through distant hybridisation (Yu et al. 2020; Ren et al. 2020). However, critical knowledge gaps remain regarding the genome composition and genetic effects of the cabbage Ogura CMS restorer lines. Initially, the absence of Rfo-homozygous lines and high heterozygosity hindered de novo genome assembly of Ogura CMS FRLs (Table S1). In this study, BC7 high-generation Ogura CMS FRLs were developed via successive backcrossing (Figure 1A). To overcome this challenge, we generated a chromosome-level genome assembly for the BC7 FRL individual RFO7GH by combining Oxford Nanopore ultra-long reads, Illumina short reads for error correction and Hi-C scaffolding for chromosomal phasing (Figure S1A; Figure S2; Table S2). The de novo assembly yielded 40 contigs (N50 = 32.3 Mb) spanning 601.8 Mb from 59 Gb raw data. The final assembly contains more than 98.3% of the genome sequences (591.8 Mb) anchored to nine pseudochromosomes. Crucially, a 15.8-Mb alien radish genomic fragment (R09: 0–15 789 821 bp) containing the Rfo gene was inverted and integrated into the long-arm terminal region of C09 in RFO7GH, replacing the homologous 28.2-Mb cabbage genomic fragment (C09: 0–28 239 208 bp) (Figure 1B). Genome analysis revealed the loss of the telomere on the C09 long arm (Figure 1C). To validate this structural rearrangement, a specific primer set (GR-4F/4R) targeting the recombination breakpoint was designed, which successfully amplified a 4.6-kb junction sequence (Table S3). Sanger sequencing of the amplified products showed 100% concordance with Nanopore data, confirming the assembly accuracy. Our Ogura CMS FRLs not only restored the fertility of Ogura CMS varieties but also exhibited three novel benefits arising from the telomere dysfunction and chromosomal inversion, enabling the rapid loss of alien genomic fragments after cytoplasmic replacement without requiring additional marker-assisted selection: (i) Transmission Barrier Ensures the Rapid Loss of Alien Fragment: Multigenerational analysis of 285 populations (n = 51 787) revealed severe transmission barriers of the Rfo gene. The mean transmission rate (TR) was 8.05% in self-pollinated populations and 1.18% in backcross populations (Table S1). Fluorescence in situ hybridisation (FISH) of meiotic prophase I cells, using a 9-kb probe specific to the alien radish fragment, revealed that only a very small proportion of PMCs (29/412, 7%) carried the alien genomic fragment (Figure S3; Table S3). The large chromosomal inversion and, crucially, the telomere loss appear to severely disrupt genomic stability, making it difficult for cells carrying the 15.8-Mb alien fragment to undergo meiosis and produce gametes, thereby leading to rapid loss. (ii) The Intact Alien Fragment Enables One-Step Purging: C09 heterozygosity progressively declined from BC1 (63%) to BC4 (38%), stabilising after BC4 with complete recombination suppression (Figure S1B). Twenty-six pairs of primers (BoRa1p-BoRa26p) were designed, with target sequences evenly distributed across the 15.8-Mb alien radish genomic fragment (Table S3). These primers, together with the breakpoint-specific primers GR-4F/4R, were used to genotype all Rfo-positive individuals. Genotyping confirmed that the 15.8-Mb alien radish fragment was the shortest alien fragment that completely co-segregated with fertility restoration in all Ogura CMS FRLs (Figure S4) and is inherited as a single, intact unit in all Rfo-positive individuals. This intactness—combined with the transmission barrier (Advantage i)—enables the rapid loss of the entire alien fragment. Screening of over 5000 Rfo-negative individuals confirmed that none contained any alien radish fragments, further demonstrating the rapid loss of alien fragment (Figure S4). (iii) Growth Penalty Provides Fail-Safe Elimination of Alien Fragment-Carriers: Individuals without the alien fragment show stronger growth potential. Specifically, Rfo-positive individuals showed significantly decreased outer leaf length (LL) and economic yield (EY), and an extremely significant decrease in outer leaf width (LW) (Figure S1C,D). Although alien fragments were transmitted to offspring in the rare instance (~1%), individuals with alien radish genomic fragments were easily eliminated during selective breeding due to poor agronomic performance. The efficacy of this Ogura CMS FRL system was validated by restoring fertility in seven genetically diverse cultivars carrying resistance genes to clubroot (CRa), black rot (Xcc3), and Fusarium wilt (Foc1) diseases. Using CRa-containing germplasm as an example: after fertility restoration and cytoplasm replacement in BC1, 1044 CRa-positive individuals were obtained from 11 combinations (n = 2880) (Table S4; Figure S1E). Among these, 59 individuals exhibiting substantial biomass reductions (23.4%–38.1%, p < 0.001) were phenotypically eliminated during selection (Table S4). Comprehensive screening revealed that only 28/1044 individuals retained the 15.8-Mb fragment; all of these 28 individuals were part of the 59 previously eliminated based solely on phenotypic screening. This confirms the system's efficiency in purging alien fragments through phenotypic screening alone. In summary, our study generated high-generation Ogura CMS FRLs via distant hybridisation and successive backcrossing. The high-quality genome of the FRL line RFO7GH was de novo assembled. An Ogura CMS fertility restoration system for cabbage breeding without alien fragment interference was constructed. The Ogura CMS FRLs can make the male-sterile disease-resistant varieties into male-fertile germplasms and have been proven to be valuable in cabbage breeding. Y.Z. and W.R. conceived the project and supervised the overall experiments. W.R. performed data analysis and experimentation and wrote the manuscript. All authors participated in the study and approved the final manuscript. The genome data set is available in the NCBI Sequence Read Archive (SRA) (https://www.ncbi.nlm.nih.gov/sra) under accession number PRJNA991960. Table S1: Comparison of the transmission rates (TRs) for the Rfo gene among generations in backcross and self-pollinated populations during distant hybridisation. Table S2: Statistical results of the sequencing data and genome assembly features of RFO7GH. Table S3: Primers used in this manuscript. Table S4: Creation of CRa-positive cabbage varieties among different generations. Figure S1: The characteristics of Ogura CMS FRLs. (A) Circos plot of important features of the assembled RFO7GH genome. The five layers show the chromosome names and sizes (a), the GC content (b), the gene density (c), the density of LTR retrotransposons (d) and the links between syntenic genes (e). (B) Extent of C09 chromosome heterozygosity in different generations of Ogura CMS FRLs. The black symbol indicates the band pattern specific to the male rapeseed parent Y403 under distant hybridization, the white symbol indicates the band pattern specific to the recurrent maternal parent CMS2254 under distant hybridization, and the grey symbol indicates the heterozygous genotype. From the BC4 generation onward, the heterozygous segment no longer decreased. (C) Individuals with alien radish genomic fragment exhibit typical growth penalty (a) compared with normal individuals (b). Scale bar = 10 cm. (D) Analysis of three agronomic traits in 236 individuals from 18 combinations. The horizontal axis represents different kinds of individuals: Rfo-positive individuals and Rfo-negative individuals in Ogura CMS lines. The vertical axis shows trait values. Differences were considered significant (*) at p < 0.05 and extremely significant (**) at p < 0.01. (E) Clubroot disease resistance of the susceptible control (a) and obtained cabbage germplasms with the CRa gene (b) using the proposed system. Scale bar = 1 cm. Figure S2: Heatmap showing the Hi–C interactions. A global view of chromosome conformation (A) and enlarged view of chromosome C09 (B). Creen arrow indicates the chromosome C09. Red arrow indicates the breakpoint. Black arrow indicates the 15.8-Mb alien radish genomic fragment. Figure S3: Representative results of FISH analysis in prophase of the first meiotic division. Representative pollen mother cell (PMC) with FISH signal (green) at the end of the chromosome (A) and Representative PMC without FISH signal (B) in Rfo-positive individuals from high generation Ogura CMS FRLs. Scale bars = 10 μm. Figure S4: Identification of the size of alien radish fragment in all obtained Rfo-positive individuals and over 5000 Rfo-negative individuals randomly selected from backcross populations, self-pollinated populations, and populations obtained after fertility restoration and cytoplasmic replacement. No Rfo-positive individuals contained a smaller alien radish fragment. No Rfo-negative individuals contained any alien radish fragments. Results are shown for six representative individuals. Complete linkage was observed across the entire 15.8-Mb introgressed alien radish fragment. M: marker. 1: Cabbage variety ‘CMS2254’ (negative control). 2: radish variety ‘RFRs’ (positive control). 3–5: Three representative Rfo-negative individuals. 6–8: Three representative Rfo-positive individuals. 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.