Fine mapping and expression characteristics analysis of male-sterile gene BrRNR1 in Chinese cabbage (Brassica rapa L. ssp. pekinensis)

Abstract

Background

Chinese cabbage is a cross-pollinated crop with remarkable heterosis, and male-sterile line is an important mean to produce its hybrids. In this study, a male-sterile mutant msm7 was isolated from a Chinese cabbage DH line ‘FT’ by using EMS-mutagenesis.

Results

Compared with the wild-type ‘FT’, the anthers of mutant msm7 were completely aborted, accompanied by the defects in leaf and petal development. Genetic analysis showed that a single recessive nuclear gene controlled the sterile phenotype of mutant msm7. Cytological observation indicated that the anther abortion of mutant msm7 was caused by the degenarated microspores and premature degradation of tapetum. MutMap and KASP analyses identified that BraA01g038270.3 C, encoding the large subunit of ribonucleotide reductase (RNR1) which involved in the biosynthesis of dNTPs, was the candidate gene, named BrRNR1. Compared with the wild-type ‘FT’, a G-A mutation occurred on the 4th exon of the BrRNR1 gene, leading to the premature termination of encoded amino acid in mutant msm7. Expression analysis indicated that the BrRNR1 gene was ubiquitously expressed in all organs and was significantly decreased in flower bud and anther of mutant msm7 compared with the wild-type ‘FT’. Subcellular localization revealed that BrRNR1 was an endoplasmic reticulum localization protein.

Conclusion

Our study is the first to characterize the function of BrRNR1 gene associated with male sterility and lays a foundation for exploring the molecular mechanism of anther abortion caused by the mutation in BrRNR1 gene of Chinese cabbage.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-06076-x.

Background

Heterosis is the phenomenon whereby hybrid offspring are superior to their parents in terms of quality, yield and resistance. The utilization of plant heterosis can significantly increase crop yields, improve crop quality and increase crop resistance [1, 2]. Male sterility is of great value in plant breeding and has become one of the important ways in the heterosis utilization [3, 4]. Based on the inheritance pattern, male sterility was divided into two categories: cytoplasmic male sterility (CMS) and genetic male sterility (GMS) [5]. CMS exhibited the maternal inheritance, however, GMS can be divided into dominant sterility and recessive sterility [6].

In flowering plants, the occurrence of male sterility is the result of different tissue differentiation and gene coordination expression. The pollen development is a complex process, involving in multiple processes and expression regulation of multiple genes. Once a certain process is abnormal, the anther development stops, which may cause male sterility [7]. In Arabidopsis thaliana, numerous genes related to anther development have been identified, mainly including SPOROCYTELESS/NOZZLE (SPL/NZZ), ABORTED MICROSPORE (AMS), EXCESS MICROSPOROCYTES1/EXTRA SPOROGENOUS CELLS (EMS1/EXS), DYSFUNCTIONAL TAPETUM 1 (DYT1), TAPETUM DETERMINANT 1 (TPD1), MYB33/MYB65, MYB80, DEFECTIVE IN TAPETAL DEVELOPMENT AND FUNCTION 1 (TDF1) and MALE STERILITY 1 (MS1). Among these genes, SPL is expressed during spore development and associated with the MADS-box protein. SPL gene hinders spore cell formation, and its product is a transcription regulator of spore cell development [8]. AMS gene encodes basic helix-loop-helix (bHLH) transcription factor, which plays a crucial role in the process of microspore formation [9]. EMS1/EXS and TPD1 are early genes during anther meiosis, which determine the differentiation of microspore cells and tapetum cells, and they both can control the tapetum cell recognition [9, 10]. DYT1 gene regulates the tapetum development. In addition, DYT1 acts on the downstream of EMS1/EXS and SPL/NZZ [11]. TDF1 encodes R2R3-MYB transcription factor, which can regulate callose dissolution. TDF1 gene plays a role in regulating tapetum development, acting on the upstream of AMS and AtMYB103 gene and downstream of DYT1 gene, respectively [12]. The expression of sporophytic pollen coat proteins (sPCPs) depends on MS188 and MS1. MS188 directly activates MS1, while MS1 directly regulates the expression of sPCPs, thus regulating the pollen exine formation [13]. Therefore, the formation of male gametophytes is very important for plant reproduction.

Chinese cabbage (Brassica rapa L. ssp. pekinensis) is a cross-pollinated crop with remarkable heterosis. The discovery and utilization of genes related to male sterility can provide theoretical basis for improving sterile lines, solving fertility stability and producing hybrids in Chinese cabbage [14]. Zhang et al. [15] identified a series of pollen development-related genes by transcriptome sequencing and cDNA-AFLP analysis in the flower buds of Chinese cabbage “Bajh97-01A/B”. Of which, sequence alignment analysis showed that BcMF17 encoded polygalacturonase PG, which may involve in the pollen wall formation. Quantitative real-time PCR (qRT-PCR) analysis showed that BcMF17 was highly expressed at pollen maturation stage. Xu et al. [16] identified the genes related to male sterility using cDNA-AFLP and qRT-PCR technology in CMS7311 of Chinese cabbage. The results indicated that Glycine Rich Protein 17 (BcGRP17), BcMS2 and Pectin methylesterase 31 (BcPME31) were involved in pollen development. The transcriptome and proteome analysis of fertile and sterile flower buds were performed in Chinese cabbage GMS line ‘AB01’, and 4,795 differentially expressed genes (DEGs) and 1,494 differentially expressed proteins were obtained, of which, a number of genes and proteins associated with male sterility were further identified, including BrAMS, BrMS1, BrbHLH089, BrbHLH091, BrAtMYB103 and BrANAC025 [4]. Tan et al. [17] obtained a male-sterile mutant ftms using ⁶⁰Co-γ ray mutagenesis in Chinese cabbage, and Bra010198 was identified as a candidate gene by BSR-Seq and molecular marker technology. Bra010198 is homologous to KNS4/UPEX1 and encodes β-(1,3)-galactosyltransferase, which can regulate the pollen exine development. The RNA-Seq analysis of the stamens from sterile and fertile line (CCR20000 and CCR20001) in Chinese cabbage was conducted, and the transcription activity of many genes in the sterile line was weak, which may be the cause of abnormal stamen development. These DEGs mainly focused on plant hormones, carbon metabolism and amino acid biosynthesis pathways. In addition, transcription factors exhibited high activity in the structural gene regulation of pollen fertility, such as APETALA 2 (AP2), MYB, bHLH and WRKY [18]. Dong et al. [19] created a number of male-sterile mutants of Chinese cabbage by EMS mutagenesis, and selected three allelic male-sterile mutants msm2-1, msm2-2 and msm2-3 as research materials. BraA10g019050.3 C was ascertained as the candidate gene by MutMap and Kompetitive Allele Specific PCR (KASP) genotyping techniques, and the function of BraA10g019050.3 C was proved by allelic variation. BraA10g019050.3 C is homologous to MS1 in Arabidopsis thaliana and encodes a PHD-finger transcription factor, which can regulate the tapetum development and biosynthesis of pollen exine. These results laid a foundation for further exploring the mechanisms regulating stamen development in Chinese cabbage.

In our previous study, we employed a doubled haploid (DH) line ‘FT’ as the experimental material, and constructed a Chinese cabbage mutant library by EMS mutagenesis [20]. In this study, a male-sterile mutant msm7 was isolated from the mutant library. The objectives of this study was to clarify the cause of anther abortion of mutant msm7, identify the candidate gene and explore its expression characteristics. These results are conducive to explore the role of candidate gene in stamen development of Chinese cabbage.

Results

Phenotypic characteristics of mutant msm7

Compared with the wild-type ‘FT’, at the reproductive growth stage, the stamen development of mutant msm7 was abnormal, anthers were shrivelled, with no pollen. The petals crinkled in appearance with lighter color (Fig. 1a, b); at the vegetative growth stage, the leaves of mutant msm7 were variegated and slender, and the leaf margin had serrated character (Fig. 1c). Observation and comparison of floral organs showed except for anther abortion, all floral organs of mutant msm7 were smaller than those of wild-type ‘FT’ (Fig. 1d-k). The pollen viability detection showed that numerous mature and viable pollen grains were detected in the wild-type ‘FT’ anthers (Fig. 1a), however, there was no pollen grains in mutant msm7 (Fig. 1b).

Fig. 1.

Morphological characteristics of mutant msm7. a. Inflorescence and pollen viability of wild-type ‘FT’. b. Inflorescence and pollen viability of mutant msm7. c. Plants of wild-type ‘FT’ and mutant msm7. d. Flower of wild-type ‘FT’. e. Flower of mutant msm7. f. Flower bud. g. Petal. h. Sepal. i. Pistil. j). Long stamen. k. Short stamens. Wild-type ‘FT’ (left) and mutant msm7 (right) for f-k. Scale bar for comparison of floral organs (d-k) is 1000 μm; Scale bar for pollen viability analysis is 500 μm

Cytological observation on anther development in mutant msm7

In order to clarify the period and cause of anther abortion in mutant msm7, the cytological observation of anthers at different stages were performed. The anthers of wild-type ‘FT’ and mutant msm7 were normal at the tetrad stage (Fig. 2a, e). With the anther development, the pollen grains of wild-type ‘FT’ were gradually mature, and the tapetum was degraded normally. After anther dehiscence, mature pollen grains were released (Fig. 2b-d). Different from the wild-type ‘FT’, the tapetum of mutant msm7 began to degrade at the early uninucleate microspore stage, and the microspores were gradually degenerated (Fig. 2f). At the late uninucleate microspore stage, the tapetum of mutant msm7 was completely degraded, and the microspores were degenerated, while partial microspore degeneration was linearized (Fig. 2g). At the mature pollen stage, the pollen sac of mutant msm7 began to collapse inward, and the microspores in anther locules were deformed and died (Fig. 2h). In conclusion, the degenerated microspores and premature degradation of tapetum led to the anther abortion of mutant msm7.

Fig. 2.

Cytological observation of anther at different stages of the wild-type ‘FT’ (a-d) and mutant msm7 (e-h). T tapetum; Tds tetrad; Msp microspore; MP mature pollen; dMsp degenerated microspore; Se septum; eLo empty locule; dTa degraded tapetum. Scale bar 150 μm

Inheritance of mutant msm7

All plants in F₁ generation showed stamen fertility. In F₂ population, there were 54 fertile plants and 20 sterile plants, and the segregation ratio was 2.7:1 (χ² = 0.16 < χ²_0.05 = 3.84), which conformed with Mendelian inheritance. Therefore, the above results showed that a single recessive nuclear gene controlled the sterile phenotype of mutant msm7 (Table 1).

Table 1.

Genetic analysis of male-sterile mutant msm7

Generations	Total	Fertility plants	Sterility plants	Segregation ratio	Chi-square (χ2)
P1 (‘FT’)	50	50	0
P2 (msm7)	25	0	25
F1 (P2 × P1)	50	50	0
F2	74	54	20	2.7: 1	0.16

BraA08g022480.3C is the candidate gene for mutant msm7

High-throughput sequencing of the wild-type ‘FT’, mutant msm7 and F₂ mutant gene pools was conducted by using MutMap method, and the high-quality reads aligned to the reference genome were 90,640,910, 52,859,159 and 131,420,852, with the alignment rates of 98.05%, 99.14% and 98.99%, respectively. The polymorphic SNP sites were further filtered and screened for calculating the SNP-Index values.

The region exceeding the threshold line corresponding to the 95th percentile was selected as candidate region, which was located in a 3.37 Mb region on chromosome A08 (A08: 14,130,000–17,500,000) (Fig. 3), which contained three candidate SNPs (Additional file 1: Table S1). Candidate SNPs were further screened, and the criteria were: (1) SNP-Index > 0.95; (2) Non-synonymous mutation or premature termination occurred on exons; (3) C-T or G-A mutation. Finally, two SNPs were selected, SNPA08:15,336,393 and SNPA08:16,870,790, which were located in the genes BraA08g019940.3 C and BraA08g022480.3 C, respectively.

Fig. 3.

MutMap analysis. The orange line represents the threshold line corresponding to the 95th percentile

Based on the analysis results of MutMap, the candidate genes were further determined by KASP genotyping technology. The results showed that SNPA08:16,870,790 was co-segregated with the sterile phenotype, and only T: T genotype was identified in plants with sterile phenotype from F₂ population. The genotyping results of SNPA08:15,336,393 indicated that not only T: T genotype but also C: T genotype were detected (Additional file 2–4: Figure S1, Table S2, S3). Therefore, BraA08g022480.3 C gene containing SNPA08:16,870,790 was predicted as the most possible candidate gene. BraA08g022480.3 C is homologous to AT2G21790 in Arabidopsis, encoding RNR1 which is involved in the biosynthesis of dNTPs [21].

Cloning of candidate gene

The full-length of the candidate gene BraA08g022480.3 C is 3,989 bp, including 17 exons and 16 introns. Sequence alignment showed that compared with wild-type ‘FT’, mutant msm7 harbored a G-A mutation on the 4th exon of BraA08g022480.3 C gene (Fig. 4a), which resulted in premature termination of encoded amino acid (TGG-TAG). A truncated protein containing 197 amino acids was produced, leading to the lack of the C-terminal ribonucleotide reductase domain (Fig. 4b, c). In addition, the promoter sequence of BraA08g022480.3 C gene was further cloned, and there were several differences between wild-type ‘FT’ and mutant msm7, comprising the base substitutions of G-A and C-T and the insertions of base A (Additional file 5: Figure S2).

Fig. 4.

Gene structure and phylogenetic tree of BrRNR1. a. Gene structure of BrRNR1. b. Prediction protein domain of BrRNR1. c. Partial sequence alignment. d. Phylogenetic tree of BrRNR1.

To clarify the evolutionary relationship of BraA08g022480.3 C gene in different species, phylogenetic tree indicated that BraA08g022480.3 C was highly similar to Arabidopsis AtRNR1, and the homology reached 91.18% (Fig. 4d). Therefore, we named BraA08g022480.3 C as BrRNR1.

Expression analysis of candidate gene BrRNR1

As shown in Fig. 5, the expression of BrRNR1 gene was detected in roots, stems, leaves, flower buds and flowers, and compared with wild-type ‘FT’, the expression of BrRNR1 gene was significantly reduced in flower buds and flowers of mutant msm7, while the expression in stems and leaves was significantly increased. Furthermore, the expression of BrRNR1 gene in anthers, filaments, stigmas and petals of mutant msm7 was significantly reduced compared with wild-type ‘FT’, and no significant difference was detected in sepals (Fig. 6).

Fig. 5.

Expression patterns of BrRNR1 gene in different organs. Note: * indicates a significant difference determined by the t-test (at the 0.05 level)

Fig. 6.

Expression patterns of BrRNR1 gene in different floral organs. Note: * indicates a significant difference determined by the t-test (at the 0.05 level)

In Chinese cabbage, RNR1 gene is encoded by four genes, including BraA08g022480.3 C (BrRNR1), BraA01g000410.3 C, BraA06g042290.3 C and BraA10g029930.3 C. Except for BrRNR1, the expression levels of other three genes had no significant difference in the anthers of wild-type ‘FT’ and mutant msm7. In addition, the expression patterns of RNR2 genes were further analyzed, and there was no significant difference between wild-type ‘FT’ and mutant msm7 (Additional file 6: Figure S3).

Subcellular localization of BrRNR1 protein

The constructed vector pCAMBIA1301S-BrRNR1-EGFP was transformed into the leaves of Nicotiana benthamiana, and HDEL-mCherry acted as an endoplasmic reticulum marker protein. As shown in Fig. 7, BrRNR1 and HDELm-Cherry were co-localized. Therefore, the BrRNR1 protein is an endoplasmic reticulum localization protein.

Fig. 7.

Subcellular localization of BrRNR1 protein. Scale bar for GFP positive control is 50 μm; Scale bar for BrRNR1-GFP is 75 μm

Promoter activity analysis of BrRNR1 gene

The constructed vector carrying GUS reporter gene was transformed into the Columbia-0, and a total of 20 transgenic plants were screened (Additional file 7: Figure S4) and further verified by PCR amplification (Additional file 8: Figure S5). GUS staining showed that blue signals were detected in the flower buds, leaves, stems and roots of transgenic plants (T₁ generation) compared with the wild-type Columbia-0, suggesting that BrRNR1 gene was ubiquitously expressed in all organs (Fig. 8).

Fig. 8.

GUS staining of transgenic plants. a-d. Flower bud, leaf, stem and root of Columbia-0 plants. e-h. Flower bud, leaf, stem and root of transgenic plants. Scale bar 2500 μm

Discussion

Male sterility is an important trait in hybrid breeding and variety improvement [22]. Male sterility is one of the important means to utilize the heterosis in Chinese cabbage, due to its small floral organs and the difficulty of artificial pollination [23]. In this study, we isolated a male-sterile mutant msm7 from a Chinese cabbage DH line ‘FT’ by using EMS-mutagenesis. MutMap combined with KASP technology were applied to locate the mutant gene, and it was predicted that BraA08g022480.3 C (BrRNR1) was the candidate gene, encoding RNR1 and participating in the biosynthesis of dNTPs during the DNA replication and repair. Compared with the wild-type ‘FT’, the mutant msm7 contained a G-A base substitution on the 4th exon of BrRNR1 gene, which led to the premature termination of encoded amino acid. The BrRNR1 gene was ubiquitously expressed in all organs and the expression of BrRNR1 gene was significantly reduced in flower buds and anthers of mutant msm7 compared with wild-type ‘FT’. It is the first time that BrRNR1 gene was associated with stamen fertility of Chinese cabbage, which provides a framework for further revealing the regulatory mechanism of anther abortion caused by the mutation of BrRNR1.

During anther development, the tapetum is an important nutrient source for microspore development and can release uninucleate microspores through timely programmed death [24, 25]. Abnormal degradation of tapetum cells can lead to insufficient carbohydrate and lipid metabolism, impaired pollen development and incomplete pollen wall structure, and eventually induce male sterility [26, 27]. Zhou et al. [4] performed the cytological observation on the anthers of a Chinese cabbage GMS line ‘AB01’, and the results showed that the abnormal development of pollen mother cells and abnormal expansion of tapetum cells resulted in male sterility. Huang et al. [5] observed the anthers of a male-sterile mutant msm by paraffin section in Chinese cabbage. The anther abortion occurred at the tetrad stage, with abnormal expansion and high vacuolation of tapetum cells, which eventually resulted in microspore abortion. Guo et al. [28] employed the GMS near-isogenic line 10L03 of Chinese cabbage as the experimental material. Cytological observation showed that anther abortion started at the tetrad stage, and the abnormal programmed cell death of tapetum caused microspore abortion. Zou et al. [29] conducted cytological observation on the anthers of three allelic male-sterile mutants msm1-1/1–2/1–3 in Chinese cabbage, and found that anther abortion occurred at the early uninucleate microspores stage, and microspore development was abnormal due to the premature degradation of tapetum. Similar to previous studies, in this study, due to the premature degradation of tapetum, the microspore development of mutant msm7 was affected, which eventually led to anther abortion.

Ribonucleotide reductase (RNR) catalyzes the reduction of ribonucleoside diphosphates (NDPs) to deoxyribonucleoside diphosphates (dNDPs) and then to deoxynucleoside triphosphates (dNTPs), thus providing the cell with precursors needed for DNA synthesis and repair [21, 30–32]. In eukaryotes, RNR enzyme is a ɑ₂β₂ heterodimer consisting of two large subunits RNR1 and two small subunits RNR2 [33]. In Arabidopsis thaliana, RNR1 is encoded by a single copy gene, while RNR2 is encoded by three genes, namely RNR2A, RNR2B and TSO2, which have a certain degree of functional redundancy [34, 35]. RNR1 gene was highly expressed in young leaves, shoot tips and inflorescences (especially carpels), while it was relatively low in cotyledons, stems, adult leaves and senescent leaves. In addition, RNR1 gene was also expressed during pollen development [35, 36]. Transient expression of RNR1-GFP indicated that RNR1 was localized in cytosol [37]. In this study, we proved that BrRNR1 is the candidate gene of msm7, which is homologous to AT2G21790 (RNR1) gene in Arabidopsis thaliana. The expression of BrRNR1 gene was detected in roots, stems, leaves, flower buds and flowers. Compared to wild-type ‘FT’, BrRNR1 gene was significantly reduced in flower buds and flowers of mutant msm7. Further analysis of floral organ expression showed that BrRNR1 gene was significantly reduced in the anther of mutant msm7 compared to wild-type ‘FT’. These results indicated that the expression of BrRNR1 gene was hindered in mutant msm7, revealing that the gene function of BrRNR1 may be impaired, and thus affected the anther development. In addition, unlike studies in Arabidopsis thaliana, subcellular localization results indicated that BrRNR1 was localized in the endoplasmic reticulum, and RNR1 gene was encoded by four genes, not a single copy in Chinese cabbage. However, the expression of other three genes had no significant difference between the wild-type ‘FT’ and mutant msm7. Therefore, we speculated that although RNR1 is encoded by four genes in Chinese cabbage, however, there is no functional redundancy among these genes.

Studies have shown that normal RNR function and dNTP pool are required for plant growth and development in Arabidopsis thaliana. Moreover, the RNR complex is also essential for the pollen normal development by maintaining the balance of dNTP pool [35, 37, 38]. Garton et al. [35] isolated a crinkled leaves8 (cls8) mutant from an EMS mutant library in Arabidopsis thaliana. The mutant cls8 exhibited abnormal leaf and floral organ morphology, curly and bleached leaves, asymmetrical and wrinkled petals, however, there was no difference in seed viability and seed set. Map-based cloning and complementary verification showed that RNR1 was the mutant gene, a G-A mutation occurred on the 15th exon of CLS8 (RNR1) gene, resulting in the amino acid substitution changing from glycine to glutamic acid (position 718). Tang et al. [37] obtained a mutant defective pollen organelle DNA degradation 2 (dpd2) in the EMS-mutagenized M₁ and M₂ populations in Arabidopsis thaliana. The mutant dpd2 showed a variety of defects at vegetative growth and reproductive growth stages, such as curly and bleached leaves. Although the stamen of mutant dpd2 was fertile, the number of stamens and petals was abnormal, with a significant decrease in normal pollen grains, different sizes and atrophy shapes of pollen grains, and reduced pollen viability. Since the phenotype of dpd2 was similar to cls8, and the map-based cloning results showed that the location regions were overlapped, further allelic detection proved that dpd2 and cls8 are allelic. The cloning results indicated that a G-A mutation existed on the 5th exon of RNR1 gene in mutant dpd2, resulting in the amino acid changing from glycine to aspartic acid (position 264). In addition to Arabidopsis, mutations of RNR1 gene have also been identified in rice (Oryza sativa) and maize (Zea mays). A temperature-conditional chlorophyll-deficient mutant virescent3 (v3) was reported in rice, and map-based cloning and complementation verification demonstrated that a single nucleotide change (G871A) in the 5th exon of the RNRL1 gene caused a missense mutation, Gly to Ser (G291S), underlies the chlorotic leaves of v3 mutant, indicating that RNRL1 is involved in chloroplast biogenesis [39]. The thermosensitive vanishing tassel1-R (tvt1-R) mutant, which is a temperature-sensitive mutant mainly affected in tassel formation, was identified in maize. Map-based cloning, allelism test and complementation verification demonstrated that a G-A mutation resulted in an amino acid change from Arg to His (R277H) in ZmRNRL1, confers the mutant phenotype of tvt1-R [40]. In this study, the mutant msm7 exhibited the variegated leaves, with shrivelled and light color petals, which was similar to the phenotypes of Arabidopsis mutants cls8 and dpd2. However, the stamens of mutant msm7 were completely sterile, with no pollen, which was different from those of Arabidopsis, rice and maize mutants. Sequence analysis showed that mutant msm7 harbored a G-A mutation on the 4th exon of BrRNR1 gene, leading to the premature termination of encoded amino acid, while only non-synonymous mutations occurred in Arabidopsis, rice and maize mutants. In view of this, we speculated that the mutation of BrRNR1 gene produced a truncated protein and affected the function of RNR, which led to the impaired ability to synthesize dNTPs, and ultimately impeded the pollen development of mutant msm7, resulting in male sterility and associated defects in leaf and petal development.

Conclusions

We obtained a stable genetic male-sterile mutant msm7 by EMS mutagenesis, and proved that BrRNR1 was the candidate gene controlling the male-sterile phenotype of mutant msm7. Further study on the function of BrRNR1 is helpful to elucidate its regulatory mechanism in stamen development of Chinese cabbage.

Materials and methods

Plant materials and genetic analysis

The mutant msm7 was crossed with the wild-type ‘FT’ to obtain the F₁ generation, and the F₂ segregation population was obtained by self-crossing F₁ generation. The segregation ratio was analyzed by chi-square (χ²) test.

Identification of pollen viability

At the flowering stage, the open flowers of wild-type ‘FT’ and mutant msm7 were collected, and the pollen grains from fresh anthers were immersed in the Alexander staining solution (malachite green, acid fuchsin and phenol) [41]. The samples were cultured at 37℃ for 30 min, and the pollen viability was compared using an optical microscope (Nikon ECLIPSE 80i, Japan).

Cytological observation of anther development

The flower buds at different developmental stages were collected and classified from the wild-type ‘FT’ and mutant msm7. The flower buds were fixed in FAA solution (50% ethanol, 5% acetic acid, 37% formaldehyde) [42], and then dehydrated with ethanol (different concentrations of 50–100%) and infiltrated with xylene. The flower buds were embedded in paraffin and cut into 10 μm sections. The sections were dewaxed with xylene, and then stained and observed under a light microscope (Nikon ECLIPSE 80i, Japan).

Identification of candidate genes by MutMap combined with KASP genotyping

The mutant msm7 and wild-type ‘FT’ were used to construct F₂ segregation population, and 50 male-sterile plants were screened and used to construct mutant gene pool. The wild-type ‘FT’, mutant msm7 and F₂ mutant gene pools were sequenced based on Illumina NovaSeq sequencing platform. The high-quality data were aligned to the reference genomes (http://brassicadb.cn; Brassica rapa Version 3.0), and then the SNP sites were detected and annotated. The SNP-index values of high-quality SNPs were further calculated, and the candidate regions were detected by sliding-window analysis.

According to the results of MutMap analysis, the candidate SNPs were identified by KASP genotyping. 184 male-sterile plants were selected from F₂ population for KASP genotyping, and the wild-type ‘FT’, mutant msm7 and F₁ plants were used as genotyping controls. Primers were shown in Additional file 9: Table S4.

Clone of candidate genes

Genomic DNA was extracted from wild-type ‘FT’ and mutant msm7 leaves using DNA plant kit (Tiangen, China) and applied to clone the full-length and promoter of candidate genes [42]. Total RNA was extracted from wild-type ‘FT’ and mutant msm7 flower buds using RNA extraction kit (Takara, China) and reverse transcribed to cDNA, which was used to clone the coding sequences of the candidate genes [42]. After amplification of the corresponding sequences, the PCR products were purified, and transformed into pGEM-T Easy vector (Promega, USA). The PCR reaction system and procedure were referred to the methods of Huang et al. [43]. The purified products were cultured on LB solid medium (anti-Ampicillin), and the white spots were selected into LB liquid medium and shaken for 24 h. The bacterial solution containing the target fragment was sequenced at Sangong Bioengineering Co., Ltd. (Shanghai, China). Finally, the DNAMAN V6 software (Lynnon BioSoft, Canada) was applied to align the sequences, and the primers of cloning candidate genes were in Additional file 10: Table S5.

Phylogenetic tree analysis and protein domain prediction

The protein sequence of candidate genes was used as searching points, and the homologous sequences in other species were searched by BLAST (http://www.ncbi.nlm.nih.gov/BLAST). Clustal W in MEGA6.0 software was applied to align the sequences, and the phylogenetic tree analysis was performed according to neighbor-joining method (1000 bootstrap replications). The protein domain prediction was conducted by SMART software (http://smart.embl.de/).

qRT-PCR analysis

At the flowering stage, different organs and different tissues of floral organ of wild-type ‘FT’ and mutant msm7 were respectively collected. The expression patterns of candidate genes and related genes were analyzed by qRT-PCR. Total RNA of the samples was extracted, and the first-strand cDNA was obtained by reverse transcription kit (Novizan, China). qRT-PCR was conducted by QuantStudio™ 6 Flex Real-time PCR System (ABI, USA), and the qRT-PCR reaction system and procedure were referred to the methods of Huang et al. [43]. The Actin gene was applied as internal control. Three biological replicates and three technical replicates were performed for each reaction, and the relative expression level was calculated by 2^−ΔΔCt method [44]. The data were analyzed by OriginPro9.0 software, and primers of qRT-PCR analysis were in Additional file 11: Table S6.

Subcellular localization

The coding sequence of candidate gene BrRNR1 was amplified in the wild-type ‘FT’, and the PCR product was purified to achieve the target fragment. The PCR reaction system and procedure were referred to the methods of Huang et al. [43]. The vector was digested with KpnI/SalI. Then, the ligation of the vector and target fragment was conducted by T4 ligase, and the constructed vector pCAMBIA1301S-BrRNR1-EGFP was sequenced and verified. The fusion expression vector and empty vector were respectively injected into the leaves of Nicotiana benthamiana by Agrobacterium injection method. Finally, the fluorescence signals of BrRNR1-GFP fusion expression vector and endoplasmic reticulum marker protein HDEL-mCherry were observed by a laser scanning confocal microscope (Leica, Germany). The primers were in Additional file 12: Table S7.

GUS staining

The promoter sequence (~ 2,000 bp) of the candidate gene BrRNR1 was amplified in the wild-type ‘FT’, and the purified PCR product was ligated into the vector pCambia1301S (del-CMV-gus) with HindIII/BglII double digestion. Finally, the constructed vector harboring GUS reporter gene was sequenced for verification, and introduced into Agrobacterium tumefaciens strain GV3101. The primers were in Additional file 13: Table S8.

Genetic transformation of Arabidopsis thaliana Columbia-0 was performed by using floral dip method [45]. The transgenic plants were detected by hygromycin and PCR amplification, and the seeds of transgenic plants were harvested, and then the plants of T₁ generation were stained by GUS staining solution. The staining organs were observed under a dissecting microscope (Nikon SMZ800, Japan).

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CMS

Cytoplasmic male sterility

GMS

Genetic male sterility

DEG

Differentially expressed gene

EMS

Ethyl methanesulfonate

DH

Double haploid

SNP

Single nucleotide polymorphism

Author contributions

S. Huang and H. Feng conceived and designed the experiments; M. Xue, R. Liao and J. Xu performed the experiments, J. Li and M. Zhou analyzed the data; M. Xue and S. Huang wrote and revised the paper; and R. Yao and Z. Liu coordinated and designed the study. All authors have read and approved the final manuscript.

Funding

This research was supported by grants from the National Natural Science Foundation of China (31801854) and Key Research and Development Project of Liaoning Province (2024JH2/102500048). The funding played roles in the design of the study and collection, analysis, and interpretation of data.

Data availability

The data charts supporting the results and conclusions are included in the article and additional files. All data generated and analyzed in this study are available upon request. The sequence data of MutMap have been deposited in the NCBI Sequence Read Archive (SRA) repository under accession numbers SRR15803269 in PRJNA761522, SRR31398300 and SRR31402234 in PRJNA1188086.

Declarations

Ethics approval and consent to participate

All the plant materials are from the Shenyang Agricultural University (SYAU, Shenyang, China). This study is not a clinical trial, and the utilization of these plant materials in this study complies with the guidelines and legislation of China.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.