Maternal karyogene and cytoplasmic genotype affect the induction efficiency of doubled haploid inducer in Brassica napus

Background Artificial synthesis of octoploid rapeseed double haploid (DH) induction lines Y3380 and Y3560 was made possible by interspecific hybridization and genome doubling techniques. Production of pure lines by DH induction provides a new way to achieve homozygosity earlier in B.napus. Previously, the mechanism of induction, and whether the induction has obvious maternal genotypic differences or not, are not known so far. Results In this study, different karyogene and cytoplasmic genotype of B.napus were pollinated with the previously reported DH inducers e.g. Y3380 and Y3560. Our study presents a fine comparison of different cytoplasmic genotypes hybridization to unravel the mechanism of DH induction. Ploidy identification, fertility and SSR marker analysis of induced F1 generation, revealed that ploidy and phenotype of the induced F1 plants were consistent with that type of maternal, rather than paternal parent. The SNP chip analysis revealed that induction efficiency of DH inducers were affected by the karyogene when the maternal cytoplasmic genotypes were the same. However, DH induction efficiency was also affected by cytoplasmic genotype when the karyogenes were same, and the offspring of the ogura cytoplasm showed high frequency inducer gene hybridization or low-frequency infiltration. Conclusion The induction effect is influenced by the interaction between maternal karyogene and cytoplasmic genotype, and the results from the partial hybridization of progeny chromosomes indicate that the induction process may be attributed to the selective elimination of paternal chromosome. This study provides a basis for exploring the mechanism of DH inducer in B.napus, and provides new insights for utilization of inducers in molecular breeding. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-021-02981-z.


Conclusion
The induction effect is in uenced by the interaction between maternal karyogene and cytoplasmic genotype, and the results from the partial hybridization of progeny chromosomes indicate that the induction process may be attributed to the selective elimination of paternal chromosome. This study provides a basis for exploring the mechanism of DH inducer in Brassica napus, and provides new insights for utilization of inducers inbreeding.

Background
Brassica napus (AACC, 2n = 4×=38) is an allotetraploid plant derived from Brassica rape (AA, 2n = 20) and Brassica oleracea (CC, 2n = 18) through interspeci c cross and natural doubling of chromosomes that happened about 75 million years ago [1,2]. Breeding of Brassica napus varieties has appreciated the utilization of heterosis that mainly includes different technical methods e.g. polima cytoplasmic sterility (pol CMS) and ogura cytoplasmic sterility (ogu CMS). Before the advent of successful application of microspore culture in Brassica napus, the pure lines were used to obtained by means of multi-generational sel ng [3][4][5], breeding cycle was long. Isolated microspore culture is usually in uenced by maternal genotype, environment, temperature and many other abiotic factors [6][7][8]. Rapeseed scientists are keen to nd that is there a simpler and more e cient technique than isolated microspore culture that can quickly and e ciently obtain pure lines of Brassica napus? In recent years, in-vivo haploid induction line was derived from the maize Stock6 [9] and the haploid induction line mediated by the Arabidopsis CENH3 gene have been successfully used in maize [10], barley [11], and rice [12]. Maize haploid induction gene ZmPLA1 [13] and Arabidopsis gene CENH3 [14][15][16] have been applied to wheat, and haploid induction have been achieved. In addition, the use of barley bulb method and distant pollination of corn pollens in wheat have also achieved haploid induction, in which induction rate was about 20-45% [14]. It is easier to get pure lines induced by crossing with haploids than isolated microspore culture. A recent study reported, in which an allotetraploid Brassica napus was crossed with an allo-octaploid rape (AAAACCCC, 2n = 8x ≈ 76) [17] and two allo-octaploid rapes had induced the function of the maternal parent to produce double haploids (DH) and named as the DH induction lines in Brassica napus: Y3560 and Y3380 [18]. SSR molecular markers, plant ploidy, and morphological identi cation revealed that a higher proportion of plants in the F1 generation were similar to the maternal parent and induction e ciency ranges from 34.09%~98.66%. What accounts for such a huge difference in induction e ciency? And it was observed that there were different induction effects according to maternal cytoplasmic genotypes of Brassica napus. Whether the induction e ciency was related to the cytoplasmic genotype of the maternal parent? Therefore, in this study we used DH inducer lines as paternal parent to pollinate three types of pol, ogu and nap cytoplasmic maternal parents, and ploidy, phenotype and genotypes of the induced offspring were observed. SNP analysis was performed to evaluate the relationship of maternal parent cytoplasmic effect. This study lays the foundation for the application of the DH lines in Brassica napus and contribute to the understanding of maternal karyogenic and cytoplasmic genetypic effects.

Ploidy analysis
The offspring of a hybrid between high-ploidy and low-ploidy that are prone to intermediate ploidy [19]. In order to understand whether there were differences in ploidy level before and after induction, we selected different Brassica napus induced F1, and tetraploid Brassica napus ZS11, which was used as control. The detection results are as follows (Fig 1, Additional le 1, Additional le 2). The uorescence intensity of F1 generation obtained by pollination of Y3560 and four maternal parents was about 409.5~510.5D thousand lines. The uorescence intensity of F1 generation of Y3380 pollinated with ve maternal parents was about 398.9~521.1D thousand lines, and the detection results were almost the same as peak value of 423.14 87.7D thousand lines in control hybrid offsprings, so they were all tetraploid plants, too. At the same time, we tested the ploidy of pollinated maternal parents and DH inducer. The uorescence intensity of the maternal parent plants ranged from 406.9~ 502.1D thousand lines, all of which were tetraploid plants. The uorescence intensity of inducer was about 753.2~852.5D thousand lines, which was two times ZS11 as control, so all of them were octoploid plants. However, in this study we selected octoploid rapeseed DH lines as the paternal parent of the tetraploid Brassica napus. The offspring ploidy were consistent their maternal parent ploidy (tetraploid), the preliminary explanation of octoploid and tetraploid rape crossing does not make sense but the tetraploid Bassica napus were pollinated by octoploid rape, which has played an induced or partial chromosomal hybrid effect.

Plant morphology and fertility investigation
Fertility and plant morphology investigations (Fig 2, Additional le 3) of F1 showed that most of plant were steriles, while only a few were fertile. Which sterile lines were pollinated by the DH inducer lines (used as paternal parent), while those of the fertile plants have shown no obvious difference from the maternal parents except for the ower morphology. Fertility identi cation results of paternal parent Y3560, Y3380 and induced F1 were shown as follows Additional le 3. The homozygous and stable pol CMS of 0068A, 0933A, and D717A that were used as the maternal parent crossed with Y3560 and six, two, and one fertile plant were appeared in their offspring, respectively. Only one fertile plant was appeared in F1, when Y3380 was used as the paternal parent and homozygous pol CMS 0933A was used as maternal parent; fertile plants were not found in F1, when Y3380 was crossed with L0068A and L0933A as maternal parent. At same time, the hybrid progeny ZS11 was also used as paternal parent for crossing with different maternal parents, and all resulting progeny produced were semi-sterile or sterile plants. Results of fertility identi cation revealed that among the induced offspring from paternal parents Y3560 and Y3380, the fertile plants appeared in the offspring of pol CMS, when it was used as maternal parent with a probability of 2.22% to 30.00%. Among them, the probability of offspring (0068A×Y3560) was highest (30.00%) while lowest (2.22%) was found in the offspring (D717A×Y3560). The induction of fertile plants in pol CMS maternal parent may be attributed to the hybridization compatibility of Y3380 and Y3560 with pol cytoplasmic restoration genes. According to the probability of fertile plants, the induction rate of Y3380 and Y3560 was estimated to vary from 70% to 100%.
Screening results of SSR molecular markers and SNP chips for induction line F1 600 pairs of SSR primers were used to differentiate the polymorphism between parental and maternal parents and their progeny (Fig 3, Additional le 4, Additional le 5). When Y3560 was used the parental parent, three pairs of speci c primers were ampli ed with 0068A; four pairs of speci c primers were ampli ed with 0933A and L0933A. Those speci c primers were also ampli ed in their offsprings. Ampli cation of SSR analysis revealed that induced F1 generation plants were consistent with maternal band type, without paternal or heterozygous band type. At the same time, combined with the eld fertility survey, the M1-1, M1-8 and M1-9, the progeny of 0068A×Y3560 were tentatively judged as the induced plants. While that of Y3380, three pairs of speci c primers were ampli ed with L0068A, ve pairs of speci c primers were ampli ed with 0933A and L0933A, and ampli cation results showed that in addition to L0068A× Y3380 offspring M4-5, M4-9 revealed heterozygous band type, while remaining offspring and maternal band type. ZS11 was used a parental parent, and 2-3 pairs of speci c primers were ampli ed in each of maternal parent. The ampli cation of primers revealed that the progeny contained the band type of both parents, hence all were hybrid plants. Since the polymorphic primers were not enough, so 6K SNP chip analysis covering more marker sites in whole of genome should be used to verify whether the progeny of each combination is homozygous or heterozygous induced plant? In view of the fact that ploidy, morphology and SSR molecular marker identi cation would not produce accurate results, 62 plants were selected for 6K SNP chip analysis (Additional les 6). The results of SNP homozygosity and genetic similarity are given in Additional le 7 and the 7-materials (Additional les 8) used this study showed a homozygosity rate that ranges from 97.93% ~ 99.29%. Such higher homozygosity rate was close to ZS11 as control (99.16%), suggesting that these seven maternal parents were homozygous. The homozygosity rate of F1 hybrid was 63.73%~ 68.74%, and genotyping (Fig 4d, Additional le 9c, Additional le 10e) also indicated that these offspring were F1 hybrids produced by ZS11 as the paternal parent. At the same time, the homozygosity rate of some F1 plants after induction ranged from 56.39% 7 8.87%, and genotyping (Fig 4a-c, Additional le 9a-b, Additional le 10a-d) was used to identify the F1 as hybrid or partial chromosome hybrid offspring. The rate of homozygous SNP sites in the remaining single plants ranged from 98.48% ~ 99.33%. Genotyping (Fig 4a-c, Additional le 9a-b, Additional le 10a-d) con rmed that these F1 were homozygous plants. Therefore, the analysis of the homozygous rate between different individual plants showed that used the DH lines as the paternal parent, the offspring produced by the homozygous rate of about 60% were hybrid offspring, and the homozygous rate greater than 95% was the induced offspring. Subsequently, we analyzed the genetic distance between the maternal parent and 62 materials, and calculated the genetic similarity rate between the maternal parents and their F1 offspring. The genetic similarity rate between hybrid offspring and the maternal parent was 64.64% ~ 68.74%, while that of genetic similarity rate between the induced homozygous individual plant and the maternal parent was as high as 99.33%.
Genotyping revealed paternal chromosome hybridization and in ltration during induction process In order to better understand the paternal chromosome hybridization and in ltration in induction process, homozygous SNP sites with parental differences were screened and genotyping revealed SNP sites in the offspring (Fig 4, Additional le 9, Additional le 10, Additional le 11). The number of homozygous SNP sites in parental differences ranged from 873-1227 (Table 1, Fig 5a, Additional le 11). Taking the maternal parent 0933 as an example, compared with the hybrid combination 0933A× ZS11 (Fig 4d), 97.91% heterozygosity, when 0933B (Fig 4a) and 0933A (Fig 4b) were the maternal parent, genotype of the induced offspring was consistent with those of the maternal parent. Hybridization of 33.84% ~ 44.26% paternal genes and paternal in ltration rate of 0.09~0.19% were observed in the hybrid offspring. When L0933A (Fig 4c) was used as maternal parent, the hybrid offspring revealed 37.87% ~ 53.03% (Table 1) paternal genes heterozygosity and 0.09% ~ 0.18% (Table 1) paternal introgression. These results indicated that the induction e ciency was in uenced by cytoplasm genotype, and the hybridization of ogu cytoplasmic genotype was more prone to occur in the maternal parent when the karyogene was same. When the cytoplasmic genotype was pol (0068A, 0933A, D717A), the offspring of 0068A (33.82% ~ 47.10%) ( Table 1) were more likely to cross with the paternal parent; and the cytoplasmic type was nap (0933B, D717B), the offspring of D717B (24.84% 4 2.86%) ( Table 1) were most likely to in ltrate the paternal gene; while the cytoplasm genotype was ogu (L0068A, L0933A), the L0933A was easier to cross or exchange with the paternal parent, and was more likely to be on the C-genome (Fig 5b), indicating that the cytoplasm genotype was the same, the induction effect was affected by the maternal parent nuclear genotype. In conclusion, the induction e ciency is in uenced by both the maternal karyogene and the cytoplasm genotype, and karyogene > cytoplasmic genotype.

Analysis of the interaction effect between the inducer line and the maternal karyogene
Since maternal karyogene and cytoplasm genotype jointly affects the induction e ciency, it is not clear whether the inducer hybridization and in ltration of the sites is random. Based on the comparison of SNP in genotyping (Table 1, Fig 5a), it was found that the SNP sites of normal hybrid offspring with ZS11 as the paternal parents were heterozygous with the exception of individual sites in ltrated by the paternal parents, and the heterozygous rate were 96.48%~ 99.19% (Table 1), and the paternal in ltration rate was 0.54% 3 .34% (Table 1). While using the inducer as the paternal parent, although hybrid progeny will be produced, the heterozygous rate of these hybrid progeny is only 29.78% ~ 53.03% (Table 1), and most of the sites are the same as the maternal parent. Subsequent analysis of the chromosomes in which these hybridization sites are located found that when ZS11 is the paternal parent, hybridization mainly occurs on the C03 and C04 chromosomes of the C genome, and the heterozygosity rate is as high as 99.2% ~ 100% (Table 1); When the inducer was used as the paternal parent, although the same as the normal hybridization, it mainly occurred on the C03 and C04 chromosomes of the C genome, but the heterozygosity rate on C03 and C04 in the hybrid progeny was 28.1% ~ 68.14% and 30.71% ~ 86.2% (Table 1), which were signi cantly lower than the normal hybridization level. Therefore, the cross produced by using the inducer as the paternal parent is not an ordinary cross.
At the same time, the chromosome where the paternal in ltration site is located is analyzed (Table 1, Fig 5b-c).
0068 and 0933 have similar genetic background (from part of one same parent), when used as maternal parent, the inducer was the paternal parent, the in ltration of the paternal parent gene was more likely to be on the C04 and C06 chromosomes of the C genome; while ZS11 was the paternal parent, the sites of paternal parent in ltration were mainly on the C05 and C06 chromosomes of the C-genome, and the paternal parent of the inducer is more likely to cause the in ltration (Fig 5b). However, since the site on chromosome C06 mostly the same site, thus eliminating C06 penetration site on the chromosome, and presumed DH lines is more likely to cause parent gene penetration on the C04 chromosome of the C genome, and the paternal in ltration caused by normal hybridization is more likely to occur on the C05 chromosome. When the maternal parent was D717, the paternal parent in ltration sites of the induced was more likely to appear on the C07 chromosome of the C genome, while the parent of ZS11 was mainly on the A06 chromosome of the A genome and the C06 chromosome of the C genome. In summary, the analysis of the paternal cross and in ltration sites, it is indicated that the DH inducer has an interaction effect with the maternal karyogene, and has the biases on different genomes.

Induction effect of rape DH inducer lines
When the Maize haploid inducer line induces the maternal plant to produce the haploid, it has obvious maternal genotype in uence, the induction e ciency ranges from 2% to 15%, there is also about 2% of the paternal gene in ltration phenomenon, and whether the Bassica napus DH inducer lines has a similar situation has not been reported. In this study, two DH inducer lines of Brassica napus, Y3380 and Y3560, were used for pollination of maternal parents with different cytoplasmic genotype, and multiple methods were used for identi cation (Additional le 12). Identi cation methods were consistent to re ect the effects of induction and hybridization, fertile offspring, heterozygous offspring marker with SSR, and heterozygous offspring detected with SNP, indicating that the detection results were completely reliable. Combining these test results, it is found that the induction system has different induction effects on different genotype of Brassica napus. When the karyogene were the same, the induction difference was mainly affected by cytoplasmic genotype. In the same cytoplasmic genotype, the induced differences are mainly affected by the karyogene. Therefore, the induction process of DH inducer line has interaction effects with maternal karyogene and cytoplasmic genotype. At same time, it was found that there was a very signi cant positive correlation between the genetic similarity rate and homozygous rate of SNP sites of induced F1 generation and the maternal parent (Additional le 7), Such as F1 of 0068A× Y3560 and the maternal parent genetic similarity of 99.79%, but its SNP sites homozygous rate and parent SNP sites homozygous rate of 98.79%, 97.93%, respectively (Additional le 7); F1 of 0933A× Y3560 has a genetic similarity of 98.86% with the maternal parent, but its SNP sites homozygous rate was 99.23%, while the maternal parent SNP sites homozygous rate was 98.27% (Additional le 7). It shows that the DH inducer lines can make homozygous parents in nitely closer to homozygous. In addition, there was some heterozygosity in the paternal parent genes of the F1 generation, but the heterozygosity varied greatly, ranging from 0.09 to 64.52% (Table 1). For the induced offspring, some individual plants will also have a small amount of paternal gene in ltration (0.09-0.18%). In the induced offspring, all the pol CMS and nap cytoplasm did not have paternal gene in ltration, while in the induced offspring of ogu CMS as the maternal parent, there was low frequency in ltration of the DH inducer gene, and high frequency hybridization. (Fig 5a, Table 1). These results indicated that the DH inducer line gene was more likely to penetrate into the progeny when the mother containing ogu cytoplasm was induced.

Mechanism analysis of induction of DH inducer line
The study of inducing vivo induced plants to produce haploid or double haploid has been explored in crops such as maize [20][21][22], Brassica napus [17,18]. However, the mechanism of the induction function of the DH inducer lines is not yet clear, and it is generally believed that uniparental chromosome elimination [14,23,24] found in maize that inducer genes were in ltrated into haploids Xu et al [25] con rmed that double fertilization occurred during the induction of maize haploids, and showed that chromosome elimination was the basis of maize haploid induction; Zhao et al [22] Found that within 7 days after pollination, most of the chromosomes of the inducible lines were excreted from the cell, and about 44 Mb of paternal parent chromosomal fragments were also found in the haploid offspring, further con rming the phenomenon of chromosome in ltration. Therefore, based on the investigation results of the mechanism of maize haploid induction line, this is similar to our ndings in Brassica napus of DH inducer.
In this study, allo-octoploid Y3380 and Y3560 were used for pollination of tetraploid Brassica napus, and the progenies were tetraploid. Except for fertility, the progenies were almost the same as the maternal parents, but were signi cantly different from the paternal parents. From the results of genotyping (Table 1), the SNP hybridization rate of ZS11 as a paternal parent hybrid is over 96.48%, and the maternal parent in ltration rate is 0.19% ~ 0.27%. And the use of inducer as paternal parent, SNP hybrid highest rate of 64.52%, the lowest was 0%, and the maternal parent in ltration rate is 35.39% ~ 100%, indicating that even if there is hybridization between the inducer and maternal parent, it is not general cross, but partial chromosome or genes cross between the inducer and maternal parent. Therefore, we speculate that the reason for the induction of double haploid in Brassica napus may also be the selective elimination of inducer chromosomes. A little part of Chromosome or large fragment cross with the maternal parent may be caused by incomplete or partial chromosome loss of inducer. The in ltration of a small number of inducer genes in the progeny may be caused by gene exchange. Further studies are needed to determine whether the induction mechanism of Y3560 and Y3380 is related to some speci c genes.

Application of DH inducer in Brassica napus breeding
Based on the above research results, due to the hybridization or gene in ltration of DH inducer to maternal chromosome (but the genotype is highly consistent with maternal karyogene, the maternal in ltration rate is 90.44% to 99.91%), these conditions may slightly change the maternal inheritance characteristics, and the overall consistency, especially for the same karyogene and different cytoplasmic maternal parent, hybridization and gene in ltration are quite different. We boldly predict that the DH inducer can provide a new model for the innovation of Brassica napus germplasm resources. The innovation the germplasm resources were from DH inducer speci c gene in ltration, and we observed in the eld that the purple leaf mutation was found during the induction of ogu cytoplasmic Brassica napus (the parent does not have purple traits) (Fig  2c). It is speculated that the in ltration of the inducer gene fragment or transposon during the induction process leads to the acquisition of certain functions of the maternal parent. Therefore, the phenomenon that the in ltration of the inducer gene enables the plant to obtain certain functions can be applied to development of rape germplasm resources, to nd in-depth special sites, and to modify the maternal genes at speci c sites.
The application of rape DH inducer can accelerate and change the rape breeding model, and create new ideas for development of germplasm resources, which have huge application potential and practical value.

Conclusions
This study explored the induction characteristics of two double-haploid inducible lines of Brassica napus. It was found that when the induced lines were used as the paternal parent to pollinate the maternal parent, the   [27]. The plant morphology of ogu CMS was not signi cantly different from fertile rapeseed plants, but the stamens were differentiated with thin anthers and did not produce normal pollen. We identi ed and counted the fertility and plant types of each F1 strain and photographed selected representative plants.

DNA extraction and PCR ampli cation
The total genomic DNA of leaves was extracted using the CTAB method. We grinded fresh leaves quickly in liquid nitrogen, added in 900 µl CTAB extract, and then heated in a water bath for 45 min, followed by placing on a shaker (80 rpm) for 15 min for mixing well. Then centrifuged the samples at 12000 rpm for 10 minutes, aspirated the supernatant, added an equal volume of pre-cooled isopropanol and placed on ice for 30 min. All samples were then centrifuged at 12000 rpm for 5 minutes, washed twice with 70% and absolute ethanol, dissolved in 100 uL ddH2O, and nally, DNA quality was detected by 0.8% agarose gel electrophoresis. The PCR reaction program is 94°C for 4min, 94°C for the 30s, TM30s, 72°C for 45s, 72°C for 8min, 35 Cycle. And 2% to 2.5% agarose gel electrophoresis for detection.

SNP chip identi cation and analysis
We used the 6K In nium SNP chip, which is evenly distributed across the rapeseed genome (Additional le 13, for distribution of the chip on the rape chromosome), to analyze the parental gene loci in the progeny. The reference genome was Brassica napus_v4.1 (http://brassicadb.org/brad/datasets/pub/Genomes/Brassica_napus/). The quality control of SNP sites was performed on samples with DQC ≥ 0.82 and CR ≥ 95. Finally, 3816 SNP sites with useful polymorphism were screened out, accounting for 75.44% of the total labeled sites. These loci were analyzed by using related software, and the homozygosity rate of individual gene sites and the genetic similarity rate of maternal parents were compared. By using Power Maker V3.25 software, we calculated the genetic distance between the maternal parents. We obtained the genetic similarity rate for which the formula is S = Nxy / Nx + Ny (Nxy is the same genotype in the two samples, Nx and Ny are the number of genotypes in X and Y samples, respectively). SNP gene analysis was based on the SNP locus of each pollination combination that was different and homozygous for the parents. The SNP of the same locus in the offspring using R studio for genotyping (Additional le 11). If the SNP banding pattern in offspring is the same as the paternal parent, it is judged to be in ltrated and consistent with the maternal parent. It is judged to be induced, and the parental heterozygous band type is judged to be hybrid. The 6Kchip detection was completed by China Golden Marker Biology Co., Ltd (Additional le 6).
Declarations Figure 2 Comparison of phenotype in orescence ower and leaf of 0933. a: Phenotype of maternal parent 0933B×Y3380 and 0933A×Y3380 (F: fertility, S: sterility). b: Phenotype of in orescence and ower of 0933A×Y3380 (F: fertility, S: sterility). c: Phenotype of leaf of induction line L0933A progeny with leaf contrasts in purple color.

Figure 4
Genotyping diagram of induced line before and after induction. a: 0933B × Y3380 . b: 0933A×Y3380. c: L0933A×Y3380. d: 0933A×ZS11. Schematic diagram of the genotyping of parents and progeny, the numbered band M is the progeny plant.