Introduction

Modern agricultural science has significantly improved seed production of crops, such as oil crops with an increase in yield of 125 % between 1985 and 2005 (Foley et al. 2011). However, the genetic variation in crops has been continuously reduced by modern plant breeding (Tanksley and McCouch 1997). For example, Brassica napus (rapeseed, AACC, 2n = 38), one of the main oil crops in the world, possesses a narrower genetic basis than its progenitor species, B. rapa (AA, 2n = 20) and B. oleracea (CC, 2n = 18) (Becker et al. 1995; Seyis et al. 2003). It has been suggested that an approach to broadening the genetic diversity of rapeseed could be the introgression of genetic components from its parental species (Jesske et al. 2013; Seyis et al. 2003).

In the practical breeding program of rapeseed, besides sexual crosses and protoplast fusion between two parental species (Eickermann et al. 2011; Ren et al. 2000; Wen et al. 2008), the cross of rapeseed with B. rapa has been widely adopted due to the high crossability between them and the high frequency of euploids produced from such a cross (Leflon et al. 2006, 2010; Liu 2000; Lu and Kato 2001; Mikkelsen et al. 1996; Olsson 1960; Qian et al. 2005; Shiga 1970; Zhou and Scarth 1995). However, there are few examples of interspecific hybridization between rapeseed and B. oleracea (Ayotte et al. 1988a, b; Chiang et al. 1978; Namai 1987; Quazi 1988; Ripley and Beversdorf 2003). Ultimately, this has left us with a poor understanding of the genetic and cytogenetic characterizations of interspecific hybridization between rapeseed and B. oleracea. Furthermore, the impacts of such activities on yield heterosis require more research, since it has been repeatedly observed that genetic diversity is lower in the C genome than the A genome of rapeseed (Bancroft et al. 2011; Bus et al. 2011; Delourme et al. 2013; Zhao et al. 2013), thus making the C genome as an obvious breeding target (Rahman et al. 2011). In the present study, the chromosomal behavior of an interspecific hybrid created between rapeseed and B. oleracea was characterized, and the B. napus-like progeny, referred as new-type rapeseed, were evaluated for genetic variation and heterotic potential. Our data suggest that it is easy to develop new-type rapeseed from the hybrids between B. napus and B. oleracea, and it possesses diverse genetic variation from current rapeseed and strong heterotic potential.

Materials and methods

Plant materials and field trials

An interspecific hybrid between B. napus and B. oleracea was developed as described in previous research (Li et al. 2013). Briefly, a line of B. oleracea var. acephala, SWU01 was employed as a female to cross with an elite cultivar of B. napus var. Zhongshuang 9, and the immature embryos were rescued 10 days after pollination and reproduced via asexual culture on MS agar medium. Clones of the F1 interspecific hybrid were transplanted in an isolated environment to develop F2 via open pollination, while the successive generations (F3 and F4) were developed via self-pollination.

Fifty-one new-type rapeseed lines (N) in the F4 generation demonstrating good fertility were employed to evaluate genetic diversity relative to a set of current accessions, which were selected randomly from three diverse gene pools of rapeseed (Becker et al. 1995; Bus et al. 2011; Diers and Osborn 1994; Hasan et al. 2006; Qian et al. 2006). This group comprised 15 spring (S), 16 winter (W) and 16 semi-winter (SW) accessions (Supplementary Materials S1).

In order to evaluate heterotic potential, 20 of the 51 new-type rapeseed lines were selected randomly to be crossed with the tester line Zhongshuang 9. The field trial was run in 2011 at two locations in the Yangtze River regions, Chongqing and Nanchang. A split-plot design comprising two blocks, hybrid block and parental line block, was employed with two replications. The plant density was designed according to farming practice, i.e. 10 plants were grown in a 2.5-m row with 0.3-m spacing, and each plot contained 30 plants in three rows. Agronomic traits, including plant height (PH), the height of first branch (HB), the inflorescence length (IL), number of branches (NB), seed yield (SY), seeds per pod (SP) and 1,000-seed weight (TSW), were evaluated at maturity.

Cytological analysis

The ovaries from young buds were collected and treated with 8-hydroxyquinoline for 3–4 h at room temperature, then fixed in Carnoy’s solution (ethanol : acetic acid = 3:1 v/v) and stored at 4 °C for chromosome number counting. The young flower buds were collected and fixed directly in Carnoy’s solution and stored at 4 °C, and the behavior of chromosomes at meiosis in pollen mother cells (PMCs) was studied according to the methods of Li et al. (1995). The fitness of chromosome segregation patterns in the F1 generation was calculated using the ratio of actual frequency to theoretical value according to the methods of Lu and Kato (2001).

The pollen grains from three flowers of each plant were stained with 1 % acetocarmine, and more than 300 pollen grains of each flower were observed under the microscope. The percentage of round and stainable pollen grains was calculated to measure pollen fertility.

Genetic diversity evaluation

The genomic DNA was isolated from young leaves, and the fingerprints of genotypes were developed with 155 primers of simple sequence repeats (SSRs) (Supplementary Material S2). The SSR bands were described by absence (0) or presence (1). The genetic distance (GD) between accessions X and Y was calculated using the formula GDxy = 1 − Nxy/(Nx + Ny), where Nxy is the number of common bands shared by accessions X and Y, and Nx and Ny are the total number of bands in accessions X and Y, respectively (Nei and Li 1979). The data from the GD matrix of 98 genotypes (51 new-type lines and 47 current rapeseed accessions) were subjected to principal component analysis (PCA) using NTSYS-pc software version 2.1 (Rohlf 1997).

The genomic components of B. oleracea in new-type rapeseed were described by the index of subgenomic components (ISG) according to the method of Qian et al. (2005): ISG = n/N × 100, where n represents the number of bands present in both new-type rapeseed and the B. oleracea parent, but absent in the B. napus parent, and N represents the total number of polymorphic SSR bands between the B. napus and B. oleracea parents.

Statistical analysis

Analysis of variance (ANOVA) was performed between hybrids across all environments using the GLM procedure of the Statistical Analysis System (SAS) (SAS Institute 1999). The comparisons between hybrid and control were performed using the F test. Pearson’s simple correlation coefficients were calculated between variables of interest. Mid-parent heterosis (MPH) was calculated as follows: MPH = 100 × (F1 − MP)/MP, where F1 = hybrid performance and MP = mean performance of both parents.

Results

Development of new-type B. napus

One embryo derived from the cross between B. napus and B. oleracea was rescued and reproduced clones in MS medium. A total of 184 clones were transplanted to the field. The clones exhibited intermediate phenotypes between the two parents (Fig. 1a–c). Seven clones with relatively larger flowers had been identified as hexaploid (AACCCC, 2n = 56) in our previous study (Li et al. 2013), and the remaining 177 clones were identified as triploid (ACC, 2n = 28) in the somatic cells (Fig. 1m). The hexaploid clones were isolated from triploid clones during the flowering stage.

Fig. 1
figure 1

Morphological and cytological characterizations of progenies between B. napus and B. oleracea. Seedling of parental B. napus, var. Zhongshuang 9 (a), parental B. oleracea var. acephala, SWU01 (b), hybrid F1 (c) and F4 (d); pollen fertility of F1 (e), F2 (f), F3 (g) and F4 (h); pods of F1 (i), F2 (j), F3 (k) and F4 (l); m one ovary cell of F1 with 28 chromosomes; n one pollen mother cell (PMC) of F1 at diakinesis with chromosome confugation 9II + 10I (the star symbols indicate bivalents); o one PMC of F1 at anaphase I with chromosome segregation of 9:19; p one ovary cell of a new-type rapeseed with 38 chromosomes. Bar 10 µm

Full size image

The young buds of triploid clones were analyzed for chromosome behavior at meiosis. Of 234 PMCs observed at diakinesis, 141 (60.26 %) exhibited 9II +10I (Fig. 1n) and 52 exhibited 10II + 8I (22.22 %), accounting for more than 80 % of chromosome pairings. The average chromosomal configuration at diakinesis was 9.06I + 9.12II + 0.10III + 0.08IV + 0.01 V.

In order to calculate the fitness of each chromosome segregation pattern, 366 PMCs were observed at anaphase I. Despite the higher frequency of the 13:15 (45.90 %) and 14:14 (13.39 %) patterns, the fitness of the 9:19 pattern was the highest among the six patterns observed, 16.4-fold more than theoretical expectation (Figs. 1o, 2). This pattern is indicative of a relatively high possibility of producing euploid progeny (C/AC, n = 9/19).

Fig. 2
figure 2

The fitness of chromosome segregation patterns in pollen mother cells of F1 triploid crosses between B. napus and B. oleracea at anaphase I. The red line indicates where the frequency of chromosome segregation fits well with the theoretical value

Full size image

The F1 plants had low pollen fertility (10 %), and produced less than one seed per pod (Fig. 1e, i). Likewise, very few F2 seeds were produced via open pollination. However, there was a dramatic increase in the F2 fertility, where an average of 55.5 % pollen fertility (from 10.8 to 95.6 %) was observed along with an average of 13.4 seeds per pod (from 2.1 to 24.5) among the 485 individuals tested (Fig. 1f, j). The F2 individuals with good seed-set (>15 seeds per pod) were selected for development of the next generations. The seed-set of F3 and F4 plants exhibited normal levels without significant differences from parental B. napus (Zhongshuang 9) and were significantly higher than those of F1 and F2 (Figs. 1e–l, 3). In order to identify chromosome numbers of new-type lines, 14 F3 lines were randomly selected and it was discovered that all of them had the same chromosome numbers as Zhongshuang 9 (2n = 38) (Fig. 1p). It appeared from these results that it was feasible to select new-type B. napus from the progeny of the cross between B. napus and B. oleracea.

Fig. 3
figure 3

Seed-set of progenies derived from B. napus and B. oleracea. The same letter above a column means no significant difference (P = 0.05), and the number of plants tested is given in parentheses

Full size image

Genetic diversity of new-type B. napus

To detect genetic diversity of new-type B. napus, 495 polymorphism bands revealed with 155 SSR primers were employed to calculate genetic distances between 98 accessions, comprising 51 new-type (N), 16 semi-winter (SW), 16 winter (W) and 15 spring (S) ecotype lines. Obvious genetic diversity was detected by PCA (Fig. 4), where the total variation explained by the first and second principal components was 24.57 and 12.10 %, respectively. Using these two principal components, spring types, winter types, semi-winter types and new-type rapeseed lines were assigned to three major clusters (Fig. 4). The average genetic distance between different types of rapeseed was greater than that within them (Table 1). Since the B. napus parent of the new-type lines was a semi-winter ecotype, a comparison of the genetic differentiation between new-type and semi-winter B. napus was performed relative to lines in the winter and spring ecotype groups. The average genetic distance between new-type lines and lines in the winter (N/W, 0.404) and spring (N/S, 0.445) ecotype groups was significantly larger than the same comparison made between lines in the semi-winter group (SW/W, 0.383; SW/S, 0.397) (P < 0.01) (Table 1). This indicated that the introgression of B. oleracea presented a viable option for widening the genetic diversity of current rapeseed.

Fig. 4
figure 4

Association between 98 genotypes, comprising 51 new-type (light green circles) and 47 natural B. napus from winter (blue triangles), spring (orange triangles) and semi-winter ecotypes (purple triangles). The proportion of variance explained by the principal coordinates is given in parentheses

Full size image
Table 1 Genetic distance (cM; mean ± SD) between and within 51 new-type (N) and 47 natural B. napus lines, comprising 16 semi-winter (SW), 16 winter (W) and 15 spring (S) accessions
Full size table

A significant and positive correlation was detected between the index of subgenomic components (ISG) of B. oleracea in new-type lines and the genetic distance between new-type lines and semi-winter rapeseed (r = 0.79, P < 0.01), spring rapeseed (r = 0.65, P < 0.01) and winter rapeseed (r = 0.46, P < 0.01), although partial genetic components of B. oleracea were transferred into those new-type lines, with an average ISG of 29.9 %, varying from 18.41 to 49.73 %. These findings indicated that new-type rapeseed carrying more genomic components from B. oleracea was more genetically distant from current rapeseed.

Field performance of new-type rapeseed

To evaluate the potential of new-type rapeseed for seed production, 20 F4 new-type lines were randomly selected and evaluated for seed yield and yield-related traits, together with the hybrids between them and Zhongshuang 9 at two locations, Chongqing and Nanchang. Heavy rain in Nanchang during flowering significantly reduced the seed yield at Nanchang (av. 1.41 T/ha) relative to Chongqing (av. 1.96 T/ha) (P < 0.01). Nevertheless, a positive correlation was detected for the yield of lines grown in the two environments (r = 0.35, P = 0.035). There was no significant difference between the seed yield of the new-type line and Zhongshuang 9. However, the new-type line was higher than Zhongshuang 9 by 17.09 % for number of branches (P = 0.018), 4.62 % for seeds per pod, 3.99 % for plant height (P = 0.015) and 5.16 % for main inflorescence length. Furthermore, the hybrids made between new-type lines and Zhongshuang 9 were significantly better than Zhongshuang 9 for seed yield (31.63 %, P = 0.05), seeds per pod (11.61 %, P = 0.002), number of branches (16.11 %, P = 0.01), plant height (10.61 %, P < 0.01), height of first branch (6.17 %, P = 0.01) and main inflorescence length (12.90 %, P < 0.01) (Fig. 5). A significant positive correlation was noted between ISG and mid-parental heterosis between the new-type line and Zhongshuang 9 for seed yield (r = 0.54, P = 0.02).

Fig. 5
figure 5

Performance of new-type lines and hybrids relative to check cultivar. TSW 1,000-seed weight (g), HB height of first branch (cm), PH plant height (cm), SP seeds per pod, IL inflorescence length (cm), NB number of branches, SY seed yield (g/plot)

Full size image

Discussion

Rapeseed is an important but young oil crop, which was domesticated no more than 400 years ago (Gomez-Campo and Prakash 1999; Toxopeus 1979). The short history of domestication and the reproductive isolation from the progenitor species might be critical causes for the narrow genetic diversity in rapeseed, particularly in the C subgenome of rapeseed (Bus et al. 2011; Jesske et al. 2013; Li et al. 2013). In comparison, B. oleracea possesses wide genetic variation, and differs substantially from the C subgenome of rapeseed (Mei et al. 2010, 2011a, b). In order to take advantage of this variation, the genomic components of B. oleracea were introgressed during the creation of new-type rapeseed developed in this study. Our data indicated that this new-type rapeseed exhibited diverse genetic variation from current rapeseed. This is likely due to the introgression of B. oleracea. In the F1 triploid derived from crossing B. oleracea with B. napus, variable chromosome pairings and segregation were observed in this and previous studies (Namai 1987), which indicated a strong possibility of homologous and homoeologous recombination and rearrangement. Subsequently, this could have induced genomic changes in the progeny, such as deletion, duplication and translocation as reported in previous studies (Cifuentes et al. 2010; Gaeta and Pires 2010).

The transfer of genetic variation from the progenitor species into rapeseed has traditionally occurred by resynthesizing B. napus from crosses between B. rapa and B. oleracea. However, those synthetic lines have limited use in breeding programs due to wild or disadvantageous characteristics from both parental species (Fujii and Ohmido 2011; Gaeta et al. 2007; Girke et al. 2012a, b; Jesske et al. 2013; Szadkowski et al. 2010, 2011; Xiong et al. 2011). In practical breeding programs, the synthetic lines are normally backcrossed with current rapeseed for several generations to improve selection against the negative traits originating from the parental species (Becker et al. 1995), resulting in a dilution in the genetic variance gained from parental species. In comparison, interspecific hybridization between rapeseed and the progenitor species can often minimize the negative impacts in lines derived from crosses between the diploid parental species. For example, interspecific hybridization between B. napus and B. rapa has been widely employed and resulted in the release of some cultivars in China and Japan (Liu 2000). However, few examples of interspecific hybridization between B. napus and B. oleracea exist in practical breeding programs, possibly due to the low crossability between B. napus and B. oleracea. Our data suggested that F1 lines could be developed between B. napus and B. oleracea via embryo rescue and new-type rapeseed could easily be developed from the progenies. It was interesting that those new-type lines possessed the strong potential to widen the genetic diversity and improve the heterosis of current rapeseed. Similar observations have been reported in other introgression lines of rapeseed carrying genetic components of B. rapa and B. carinata (BBCC) (Fu et al. 2012; Li et al. 2006; Qian et al. 2005; Xiao et al. 2010; Zou et al. 2010). Thus, it seems that harvesting genetic diversity from B. oleracea via interspecific hybridization between B. oleracea and B. napus is an important breeding strategy in rapeseed.

For many years, authors have suggested that B. rapa and B. oleracea are the parents of B. napus. UN (1935) proposed a process of origination via a cross between 2× B. rapa and 2× B. oleracea, followed by genome duplication. This scenario remains possible since it is available for direct synthesis of rapeseed via crossing between B. rapa and B. oleracea (Girke et al. 2012a, b; Jesske et al. 2013; Ren et al. 2000; Seyis et al. 2003; Wen et al. 2008). The relatively high frequency of euploid progency in ACC triploid hybrids observed in this study, as well as in AAC triploid hybrids detected in a previous study (Lu and Kato 2001), suggest that an alternative pathway, a cross between 2× B. rapa and 4× B. oleracea or between 4× B. rapa and 2× B. oleracea, followed by successive self-pollination, may be a more likely possibility for the origination of B. napus, although we did not perform those crosses.