Introduction

Harnessing alien genetic diversity is an important crop improvement activity. Wild Brassiceae, having evolved in diverse ecogeographical habitats (Tsunoda 1980), comprises a rich repository of variability especially for the defensive traits. Establishment of the phylogenetic relationships between cultivated Brassicas and contemporary crucifer genomes, through the interpretation of the chromosome pairing patterns or through genetic linkage maps anchored to the Arabidopsis genome, has underlined immense possibilities of directed gene exchange across taxonomic domains. Though somatic hybridization is the technique of choice in overcoming the fertilization barriers, various modifications of the embryo rescue techniques continue to be employed due to their simplicity and operational ease. Examples for successful gene introgressions abound, and include resistance to diseases (Hagimori et al. 1992), novel fatty acid composition (Fahlesson et al. 1994) and fertility restorer genes for a number of cytoplasmic male sterility (CMS) systems (Banga 2003). Commercialization of at least two alloplasmic CMS systems (ogu, mori) for hybrid seed production in Brassica oilseeds is a classic example of the efficacy of the introgression route for achieving plant breeding goals.

In this communication, we report on the development of hybrids of B. rapa with Diplotaxis erucoides and Brassica maurorum. Genetic homoeology between B. rapa and D. erucoides/B. maurorum genomes has been reported in the past (Vyas et al. 1995; Chrungu et al. 1999). Both these reports, however, relied solely on the meiotic chromosome pairing data of the F 1 hybrids to draw inferences on the intergenomic affinities. Such data have a limited value for genome analysis (Jauhar and Crane 1989) as they fail to differentiate between the allo- and auto-syndetic pairing; hence they tend to inflate the affinity relationships. In the absence of competition for pairing partners in interspecific F 1 hybrids, even distantly related homoeologues may frequently engage in pairing. Only higher ploidy hybrids, such as triploids and higher, or amphiploids provide a realistic test of genome affinity. Here we use tetraploid amphiploids to provide a more objective assessment of the genetic relationships between the A genome of B. rapa and the De and Bm genomes of D. erucoides and B. maurorum, respectively.

Materials and methods

Field grown plants of the cultivated B. rapa L. ssp. oleifera cv TH 68 (2n = 20, AA), and two wild accessions, B. maurorum (2n = 16, BmBm) and D. erucoides (2n = 14, DeDe) were used to produce interspecific hybrids between cultivated and wild species. Direct as well as reciprocal crosses were attempted in both the cross combinations. Flower buds of the three parental species were emasculated and pollinated with freshly collected pollen from the desired pollinator species. Some of the pollinated buds were left on the plant while others were excised and used for the in vitro experiments. For the ovary culture, pistils were excised 2–3 days after pollination (DAP), surface sterilized with mercuric chloride (0.01%) for 8 min and cultured on the Murashige and Skoog’s (MS) medium containing 5% sucrose, 0.8% agar and 500 mg/l of casein hydrolysate. Cultures were maintained at 25 ± 2°C under a 16-h light (2,000 lux)/8-h dark cycle as described earlier (Bhaskar et al. 2002). For the sequential culture, pollinated ovaries 8–9 days after the initial culture were dissected and enlarged ovules were re-cultured on a fresh MS medium containing gibberellic acid. Shoot tips from hybrid seedlings were multiplied in vitro through the shoot tip culture and the culture of nodal segments on the MS medium supplemented with benzyl amino purine (BAP) at 0.5 mg/l. The axillary shoots were then rooted on the half-strength MS or MS + IBA (0.5 mg/l) media and transferred to the field after 7–10 days of hardening under controlled environmental conditions. Chromosome doubling was induced by placing cotton swabs saturated with 0.1% colchicine on the meristematic sectors for 72 h. The hybrid nature of the recovered plants was verified by the Randomly Amplified Polymorphic DNA (RAPD) analysis. For this purpose, the total DNA was extracted from the young leaves of the parents and hybrids according to Doyle and Doyle (1990). Amplifications were performed in a MJ Research PTC 200 thermal cycler (MJ Research, Waltham, USA) using the method described by Bhaskar et al. (2002). The amplified products were separated on a 1.5% agarose gel in 1% TBE buffer, stained with 1% ethidium bromide and visualized under the UV light in a Gene Genius Gel documentation system. For cytological studies, young buds were fixed in the Carnoy’s solution II (ethanol:chloroform:acetic acid in the ratio of 6:3:1). Anthers from appropriate buds were squashed in 2% acetocarmine, and pollen mother cells (PMC’s) were viewed under Olympus microscope (XL-70) for meiotic configurations at diakinesis/metaphase. Mean bivalent, trivalent and quadrivalent frequencies were calculated as sum total of respective configurations divided by total number of PMC’s observed. Parental species, hybrids and amphiploids were tested for reaction to Alternaria brassicae and Albugo candida infestation under aided epiphytotic conditions of high humidity and repeated inoculum applications.

Results and discussion

In the present study, two diploid wild crucifers were selected for hybridization with diploid B. rapa. Only in the crosses involving wild crucifers as female, 2–3% of the pollinated pistils formed pods, with the bulk of the pollinated pistils drying 5–7 days after pollination. Seed set in the surviving pods was low (<1%) and the seed was shriveled and failed to germinate. The failure to produce hybrid seeds from field pollinations appears to be largely due to post fertilization barriers, which cause embryo abortion (Shivanna 1996). The sequential ovary–ovule culture overcame post fertilization barriers in crosses involving B. rapa as the pollen parent. In the combination D. erucoides × B. rapa, 230 cultured ovaries yielded 34 ovules which upon subsequent culturing developed into 34 seeds. However, most of these seeds were shriveled and malformed and only three germinated, producing F 1 hybrids. In the B. maurorum × B. rapa cross, 144 cultured ovaries yielded 22 ovules, three of which germinated. Of the three, two were F 1 hybrids and remaining one maternal. All produced F 1 hybrids were multiplied through in vitro culture of shoot tips and nodal segments.

Attempts at genetic enrichment through interspecific hybridization in cultivated Brassica species date back to early nineteenth century when Sageret (1826) synthesized hybrids of B. oleracea with B. rapa. However, the intergeneric barriers were breached only in the early XX century with the synthesis of a hybrid between Raphanus sativus and Brassica oleracea (Karpechanko 1924). The development of the embryo rescue techniques, by overcoming the post-fertilization barriers, greatly expanded the scope of wide hybridization attempts. They provided the Brassica breeders an access to a range of potentially beneficial nuclear or cytoplasmic encoded traits present in the wild relatives. It also allowed the establishment of intergenomic relationships (Chandra et al. 2004). Because of the application ease, in vivo fertilization followed by in vitro culturing of fertilized ovaries has been especially rewarding. Alternatively, the fertilized ovules can be dissected out from the ovary after few days of culturing and recultured in a defined medium as discussed earlier.

Morphologically, both the F 1 hybrids were intermediate to the parents. Male fertility, as indicated by pollen stainability in 2% acetocarmine was very low (<2%). Induced amphiploids had normal-sized anthers and near normal pollen stainability (≈75%). While the cultivated B. rapa parent was susceptible to the Alternaria brassicae and Albugo candida, the two wild species, the F 1 hybrids with B. rapa and the induced amphiploids were resistant. In the RAPD analysis, primers OPA 16, OPF 01, OPW 19, and OPW 13 produced polymorphic DNA fragments between D. erucoides and B. rapa (Fig. 1A) while primers OPW 19, OPW 13, OPA 16, OPN 10, and PPF 01 generated polymorphic fragments between B. maurorum and B. rapa (Fig. 1B). Hybrid plants possessed many of the fragments specific to each of the parents, thus confirming their hybrid nature.

Fig. 1
figure 1

(A) Molecular characterization of D. erucoides × B. rapa hybrid using primers OPW 19 and OPW 13. Lanes 1: female parent, 2: F 1 hybrid, 3: amphiploid, 4: male parent. (B) Molecular characterization of B. maurorum × B. rapa hybrid using primers OPW 19, OPW 13 and OPA 16. Lanes 1: female parent, 2: F 1 hybrid, 3: male parent

Full size image

Somatic chromosome numbers (2n) of D. erucoides, B. maurorum and B. rapa were confirmed at 14, 16 and 20, respectively, with normal bivalent formation in the metaphase I of meiosis. The D. erucoides × B. rapa F 1 hybrid had the expected 2n = 17 chromosomes. In the pollen mother cells (PMC’s) of the F 1 hybrids (Fig. 2a–f), a range of chromosome configurations was observed in diakinesis/metaphase I with 5II + 7I as the predominant configuration (Fig. 2d) present in about 24% of the PMC’s (Table 1). A maximum of 8II were observed in 5% of PMC’s whereas 16% of PMC’s showed more than five bivalents. The mean bivalent frequency was 3.75. A single trivalent was also observed in about 4% of the PMC’s. The B. maurorum × B. rapa F 1 hybrid also had the expected somatic chromosome number (2n = 18). 18 I was the predominant meiotic configuration in the F 1 hybrid and it occurred in about 24% of PMC’s (Table 2, Fig. 3a). One to seven bivalents were observed per PMC, with mean bivalent frequency of 3.41 % (Table 2, Fig. 3b–f). Higher than expected frequency of bivalent formation in the two wide hybrids may not be construed as a reflection of high affinity of the De/Bm and A genomes as chromosome pairing in the absence of preferential pairing is not necessarily a function of homology. The archetype of the sub tribe Brassicinae is believed to have had the basic chromosome number of x = 5 or x = 6 (Quiros 1999). The increase in the chromosome number possibly occurred by whole genome duplications and subsequent divergence. Past cytogenetic investigations with B. rapa haploids (Armstrong and Keller 1981; Truco et al. 1996), have shown occurrence of two bivalents and one trivalent resulting from autosyndetic pairing. Similarly, D. erucoides has also been reported to form up to three bivalents in haploid state (Delourme et al. 1989). Thus from purely theoretical considerations, 5II + 1 III configuration in the D. erucoides × B. rapa hybrid would be expected due to autosyndetic pairing within the parental genomes alone.

Fig. 2
figure 2

Meiotic studies in the hybrid D. erucoides × B. rapa (a–g) and the induced amphiploid (h–l), (a) 17I, (b) 1II + 15I, (c) 4II + 9I, (d) 5II + 7I, (e) 8II + 1I, (f) 1III + 3II + 8I, (g) 2–15 distribution at anaphase I, (h) 17II, (i) 2III + 14II, (j) 3III + 12II + 1I, (k) 1IV + 15II, (l) 19–15 distribution at anaphase I

Full size image
Table 1 Meiotic analysis in the intergeneric cross D. erucoides × B. rapa
Full size table
Table 2 Meiotic analysis in intergeneric hybrid B. maurorum × B. rapa
Full size table
Fig. 3
figure 3

Meiotic studies in the hybrid B. maurum × B. rapa (a–f) and induced amphiploid (g–l), (a) 18I, (b) 2II + 14I, (c) 3II + 12I, (d) 4II + 10I, (e) 6II + 6I, (f) 7II + 4I, (g) 18II, (h) 1III + 14II + 1I, (i) 1IV + 16II, (j) 1IV + 14II + 4I, (k) 16–2–18 distribution at anaphase I, (l) 19–17 distribution at anaphase I

Full size image

The amphiploid (AADeDe; 2n = 34) of D. erucoides × B. rapa had 17II as the predominant meiotic configuration, although PMC’s with varying numbers of univalents, trivalents and, occasional quadrivalent were also observed (Table 3, Fig. 2h–k). Presuming normal pairing control mechanism in the newly developed amphiploid, the occurrence of up to three trivalents (Fig. 2j) or one quadrivalent (Fig. 2k), under the conditions of preferential pairing available in tetraploid amphiploid, can be considered as an indicative of allosyndetic pairing between the A and De genomes. This is likely since both B. rapa and D. erucoides belong to Brassica lineage (Song et al. 1990; Warwick and Black 1991; Lysak et al. 2005) and the genus Diplotaxis is considered to be the closest wild relative of crop Brassica species. Unequal distribution was observed during anaphase in the amphiploid (Fig. 2l). The B. maurorum × B. rapa amphiploid (AABmBm; 2n = 36) had 18II as predominant meiotic configuration (Table 4, Fig. 3g) occurring in 74% of the PMC’s investigated, whereas the remaining 26% had either one trivalent (Fig. 3h) or one quadrivalent (Fig. 3i–j). Occurring under rigid regime of tetraploid test, these configurations were supportive of limited homology between A and Bm genomes as suggested by Takahata and Hinata (1983) in the interspecific hybrid they developed. Although the majority of the anaphase-I cells in the amphiploid showed normal 18–18 distribution of chromosomes, there was some evidence of unequal separation at anaphase I (Fig. 3k, l). Both the amphiploids showed no self-seed setting, however few open pollinated and cross seeds could be harvested.

Table 3 Meiotic analysis in the amphiploid D. erucoides × B. rapa
Full size table
Table 4 Meiotic analysis in amphiploid B. maurorum × B. rapa
Full size table

The successful synthesis of F 1 hybrids/amphiploids of B. rapa with D. erucoides and B. maurorum can be viewed as a significant step towards the development of alternaria resistant B. juncea/B. napus as both the amphiploids showed resistance to Alternaria brassicae and Albugo candida. Availability of amphiploids with genomic constitutions of AADeDe and AABmBm is expected to allow exchange of genetic information between wild and crop Brassica genomes via pairing of B/C genome of B. juncea/B. napus with Bm or De genomes of the two amphiploids. Backcross introgression of gene(s) for alternaria resistance from D. erucoides and B. maurorum into B. juncea and B. napus has already been initiated using the synthesized amphiploids as bridging species.