Introduction

The species representing the tribe Brassicae of the Brassicaceae family are of great economic importance, as they are domesticated for edible oil, vegetables, spices, forage crops and ornamental plants (Pratap and Gupta 2009). Interspecific and intergeneric hybridisation between Brassicae species makes the transferring of valuable resistance and quality traits of commercial interest, as well as broadening the gene pool for genetic improvement, challenging. Hybridisation has been extensively performed to develop more potential varieties with desired morphological characteristics, resistance to biotic and abiotic stresses, or male sterility suitable for breeding (Kaneko et al. 2009; Rahman 2001). The mechanism of chromosome elimination in interspecific Brassica species has not been completely described. However, the elimination of chromosomes that fall behind during mitosis (Sybenga 1992), abnormal spindle fibres that are formed during mitosis (Maiato et al. 2002), genetic controls that lead to chromosome elimination (Liu and Wendel 2000) and the responsibility of some genetic sequences for the non-random elimination of homologous sequences from different parental lines (Feldman et al. 1997) have been suggested. The hybridisation in Brassica proved to be very useful for the development of cytoplasmic male-sterile (CMS) crops by transferring the radish Ogura cytoplasm to Brassica oleracea and Brassica napus (Pelletier et al. 1988). The introgression of the alloplasmic CMS systems for the production of Brassica oilseeds allowed for the effective commercialisation, meeting plant breeding goals. With the development in 1994 of the improved CMS Ogu-INRA system in oilseed Brassica, seed companies have developed a number of hybrid varieties in Europe, Canada and Australia. This CMS sterility has also been widely used to produce commercial hybrids of B. oleracea vegetables (Prakash et al. 2009). Interspecific crosses attempted by plant breeders and geneticists between B. napus and B. rapa can occur readily owing to the high degree of relatedness, albeit with varying degrees of difficulty (Elling et al. 2010). Natural hybrids were found in fields where species are in physical proximity and have flowering synchrony (Anonymous 2002; Bing et al. 1996; Jorgensen and Andersen 1994).

Brassica rapa is a highly polymorphic species, represented by numerous important vegetables, and oilseed and fodder crops. Chinese cabbage (B. rapa ssp. pekinensis L.) is one of the most popular vegetables with a worldwide distribution, available for consumers year-round as a valuable source of carbohydrates, proteins, dietary fibre, vitamins A, C, K and B6, minerals, and glucosinolates (Dickson 2007). In Poland, the production area of Chinese cabbage ranges from 5000 to 7000 ha and depends entirely on F1 cultivars supplied by foreign seed companies. For this reason, a breeding programme to obtain new sources of genetic diversity and develop domestic Chinese cabbage F1 hybrids began at the Research Institute of Horticulture, Skierniewice, Poland in 2008. Production of commercial cabbage F1 hybrids requires an effective, inexpensive and reliable system of generative propagation (Faulkner et al. 1977; Hodgkin 1980a, b; Ockendon 1973, 1975) and is currently based on either self-incompatibility or CMS. Maternally inherited CMS traits have been ascertained to be more successful than self-incompatibility in the ability to stably express pollen sterility without any evident changes in vegetative growth or female fertility (Kaneko and Bang 2014). Among several male sterility types of Brassicae, the CMS discovered by Ogura (1968) is used worldwide for the breeding of B. napus, Brassica juncea, B. oleracea and Raphanus sativus F1 hybrids (Bartkowiak-Broda et al. 1979; Black et al. 2006). Ogura-CMS originated from a wild Raphanus that spontaneously hybridised with Japanese radish. An effective and reliable method for obtaining interspecific hybrids with the CMS trait is also important for the breeders of Chinese cabbage. Sterile and semi-fertile F1 and back-cross 1 (BC1) plants were obtained by interspecific sexual hybridisation between maternal B. napus var. oleifera and paternal B. rapa (Harsh et al. 2007). Interspecific F1 hybrids consecutively back-crossed from a cross between male-sterile B. napus (eru CMS) and an inbreeding line of tsai–tsai (B. rapa) were successfully obtained by Wan et al. (2013). Rapeseed with pol CMS has also been used to produce pol CMS Chinese cabbage lines (Ke et al. 1992; Yang et al. 2006). The cold-tolerant Ogura cytoplasm was successfully incorporated from B. napus rapeseed into B. rapa (Heath et al. 1994); however, there are no further reports about the use of CMS-Ogura for the development of Chinese cabbage F1 hybrids. The restorer fertility gene, which was introduced into Brassica from R. sativus by protoplast fusion (Sakai et al. 1996; Carlsson and Glimelius 2011), is also essential for the F1 hybrid production of oilseed crops; however, the restoration of the fertility is not necessary for F1 hybrids of B. oleracea vegetables with the CMS trait (Yamagishi and Bhat 2014).

Brassica species differ in their numbers of chromosomes (Nagaharu 1935; Johnston et al. 2005). This results in variations in nuclear DNA content (genome size) that were clearly confirmed using flow cytometry (FCM) by Sabharwal and Doležel (1993). The numbers of chromosomes and the nuclear DNA contents (2C) in B. napus are 2n = 38 and a 2C of approximately 2.3 pg, respectively, while in B. rapa they are 2n = 20 and a 2C of approximately 1.1 pg, respectively (Bennett and Smith 1976; Johnston et al. 2005). Sabharwal and Doležel (1993) showed that F1 hybrids between various pairs of Brassica species differing in genome sizes have intermediate nuclear DNA content values. The authors stated that the evaluation of the genome size by FCM can be a useful method to quickly identify Brassica interspecific hybrids. In this study, FCM was used as an additional tool to confirm the status of F1 hybrids between B. napus and B. rapa.

The aims of the study were to (1) introduce the CMS Ogu-INRA trait from rapeseed into Chinese cabbage inbred lines through sexual interspecific hybridisation, (2) evaluate the cytological and morphological characteristics, as well as the ability for generative propagation, of the interspecific F1 hybrids and BC1 generations, (3) obtain seeds of the consecutive BC1–BC3 generations of CMS B. rapa lines and (4) evaluate the agronomic traits of the new CMS Chinese cabbage lines of the BC3 generation in comparison with those of the recurrent parents and interspecific F1 hybrids in the field.

Materials and methods

Obtaining rapeseed × Chinese cabbage interspecific F1 hybrids and their consecutive BC1–BC3 generations

The male-sterile C1162 rapeseed line (B. napus) with Ogura cytoplasm and an isogenic P1162 fertile line were obtained from the Plant Breeding and Acclimatization National Research Institute, Poznan, Poland in 2010. Three F6 inbred lines of Chinese cabbage (B. rapa) (KW, HG and MR) were developed at the Research Institute of Horticulture, Department of Genetics, Breeding and Biotechnology of Vegetable Crops, Skierniewice, Poland. Selected Chinese cabbage lines were self-compatible, and they had the desired agronomic characteristics of a 71–77-days vegetation period from planting to harvest, good internal uniformity, good (HG) or moderate (KW and MR) vigour, and barrel (HG and KW) or cylindrical shaped heads (Fig. 1a–c). Seeds of the B. napus and B. rapa parental lines were sown at the beginning of September 2010 in a greenhouse. Three-week-old seedlings were transplanted into 10-cm plastic pots, fertilised and protected against pests and pathogens. Vernalisation started at the beginning of December when plants reached the 10–12-true-leaf stage, and it was conducted under a natural photoperiod at a temperature of 5–7 °C until the end of February 2011. At the beginning of March, three vernalised plants from each of the B. napus and B. rapa lines were selected, transplanted into 5-L plastic pots filled with substrate and placed on the ground in a greenhouse at a temperature of 16–20 °C. To avoid random and undesired cross-pollination, flower stacks were isolated using plastic and paper bags 2 days before anthesis. Pollen was collected individually from each of the Chinese cabbage inbred lines (KW, HG and MR) and transferred by hand to the stigmas of pistils of maternal rapeseed plants. The interspecific hybridisations KW × C1162, HG × C1162 and MR × C1162 were made on 7–10 freshly opened flowers of the CMS rapeseed C1162 line, while crosses KW × P1162, HG × P1162 and MR × P1162 were performed after emasculation at the green bud stage, 2 days before anthesis, of the fertile rapeseed P1162 line. From the second week of March until the end of April, three independent crosses were made for each of the combinations on consecutively developed flower stacks of maternal rapeseed plants. Mature siliques for all of the interspecific F1 hybrids and inbred lines were harvested by hand in the middle of June. After the siliques were dried, seeds were extracted and counted for each combination.

Fig. 1
figure 1

a–c Mature plants of the three Chinese cabbage inbred lines used as recurrent parents (from the left): lines MR, HG and KW, d morphological differences between KW (from the left) and rapeseed C1162 parental lines in the greenhouse, e interspecific F1 hybrid of Brassica rapa × Brassica napus, f–h segregation of the back-cross 1 (BC1) population of B. rapa × B. napus, i C1162 rapeseed line in the field, j interspecific C1162 × KW F1 hybrid in the field, and k CKW line of the BC3 generation (from the left) and KW B. rapa superseding line in the field

Full size image

In 2012, the consecutive BC1 generation ((B. napus/B. rapa) × B. rapa) was obtained through the use of propagation and vernalisation, as described for the seed development of the interspecific F1 hybrids. Three plants from each of the male-sterile interspecific F1 hybrids (KW × C1162, HG × C1162 and MR × C1162) having flower structures typical of the CMS Ogu-INRA-type were selected for the BC pollination and crossed with superseding Chinese cabbage lines at the open flower stage. Plants of the KW × P1162, HG × P1162 and MR × P1162 F1 hybrids were male fertile and, for this reason, were hybridised with superseding Chinese cabbage lines after emasculation at the green bud stage, 2 days before anthesis. In both experimental years, the self-pollination of the male-fertile parental lines (P1162, KW, HG and MR) and the sib-pollination of the sterile C1162 line with pollen of the complementary P1162 line were made at the open flower stage as controls. In contrast to the development of the F1 and BC1 interspecific crosses, the BC1/BC2 (year of propagation 2013) and BC2/BC3 (year of propagation 2014) generations were pollinized in the field by the red mason bee (Osmia rufa L.). Vernalised plants of the three superseding Chinese cabbage genotypes (KW, HG and MR) and CMS plants of the BC1 generation, [(C1162 × KW) × KW] = CKW BC1, [(C1162 × HG) × HG] = CHG BC1 and [(C1162 × MR) × MR] = CMR BC1, and BC2 generation, [((C1162 × KW) × KW) × KW] = CKW BC2, [((C1162 × HG) ≥ HG) × HG] = CHG BC2 and [((C1162 × MR) × MR) × MR] = CMR BC2, were cross-pollinated in three separate 9 m2 growth cages, each covered by a nylon fibre grid. In the second week of April, for each pair of CMS and male-fertile superseding lines, five plants with developed generative stacks were transplanted to the open field into two rows, with 100 cm between rows and 50 cm between plants. In the first week of May, when plants started to bloom, 90 red mason bees were placed into each cage to ensure appropriate cross-pollination between male-sterile and fertile components. From the second week of July until the beginning of August, mature siliques were harvested from CMS and male-fertile lines separately from each plant. After siliques dried, seeds were extracted, cleaned and weighed. For the seed productivity of interspecific crosses and their parental components, the standard deviation (SD) and coefficient of variation (V%) were calculated.

FCM analysis

Since B. rapa and B. napus differ in their nuclear DNA contents, their F1 hybrids should possess intermediate genomes of intermediate sizes compared with those of the two parental species. Therefore, FCM was used as an additional tool to confirm the status of B. napus × B. rapa F1 hybrids. Samples were taken in February and March from the fully developed young leaves of plants grown during the winter in a cold greenhouse at a vernalisation temperature of 5–7 °C. Leaf tissue (0.5–1 cm2) was chopped together with a piece (1 cm2) of an internal plant standard in a Petri dish in 0.5 mL nuclei isolation Partec buffer supplemented with propidium iodide (50 μg/mL) and RNasa (50 μg/mL) (Śliwińska 2008). As an internal standard for the FCM analysis, the young leaves of Zea mays CE-777 (2C = 5.43 pg DNA) (Institute of Experimental Botany, Olomouc, Czech Republic) were used (Lysák and Doležel 1998). After adding 1.5 mL of the isolation buffer, the samples were passed through a 30-μm filter and incubated for 80 min at room temperature. The fluorescence of the nuclei was measured using a CyFlow ploidy analyser with CyView software (CyFlow PA, Partec, Germany) and an Nd–YAG green laser at 532 nm. Data were analysed by means of CyView software (Partec). The 2C DNA content of a sample was calculated as the sample peak’s mean divided by the standard plant peak and multiplied by the amount of DNA of the standard plant. Five leaf samples from each plant having at least 5000 nuclei were measured, and each nucleic isolation extract was run twice. Data were subjected to a one-way analysis of variance (ANOVA) (STATISTICA package StatSoft v. 10). The means were compared by Tukey’s test at P = 0.05. SDs (n = 5) were calculated.

Morphological characteristics of the interspecific B. napus × B. rapa F1 hybrids and their BC1 generation

In 2012, the morphological characteristics of the two rapeseed lines (C1162 and P1162), three Chinese cabbage lines (KW, HG and MR), six interspecific F1 hybrids (P1162 × KW, P1162 × HG, P1162 × MR, C1162 × KW, C1162 × HG and C1162 × MR) and six BC1 ((B. napus/B. rapa) × B. rapa) populations [((P1162 × KW) × KW), ((P1162 × MR) × MR), ((P1162 × HG) × HG), ((C1162 × KW) × KW), ((C1162 × HG) × HG) and ((C1162 × MR) × MR))] were analysed. Each line was represented by 10 plants. Seeds were sown at the beginning of September 2011 in the greenhouse, and 3-week-old seedlings were then transplanted and cultivated in 0.33-L plastic pots filled with substrate. In the second week of January 2012, the internal uniformity and morphological characteristics, such as diameter, height and type of plant, position of leaves, heights of stacks, size, type, colour, innervation, waxiness and shapes of leaves, width, and blistering and edges of leaf blades, were analysed in the parental lines and interspecific crosses of the F1/BC1 generations. At the beginning of March, for the development of the generative phase, five vernalised plants from each genotype were selected, transplanted into 5-L plastic pots filled with substrate and placed on the ground in a greenhouse at 20 °C. Average plant heights (cm), times of blooming, shapes and lengths of green buds, sizes of flowers (mm), shapes, positions and colours of petals, and the presence of pollen for Chinese cabbage, rapeseed and their F1 and BC1 generations in the generative phase were analysed during the blooming season. The coefficients of variance for the average plant heights and average bud lengths were calculated, and other traits were classified according to the multigrade International Union for the Protection of New Varieties of Plants (UPOV) scale.

Agronomic characteristics of the CMS Chinese cabbage BC3 generation and interspecific (B. napus and B. rapa) F1 hybrids in the field

In 2014, the agronomic characteristics of three Chinese cabbage CMS lines of the BC3 generation (CKW, CHG and CMR) and three interspecific F1 hybrids (C1162 × KW, C1162 × HG and C1162 × MR) were compared with those of their parents, B. rapa (KW, HG and MR) and CMS B. napus (C1162). The experiment was performed in the field at the Research Institute of Horticulture, Skierniewice, Poland. The soil type was a pseudopodsolic over loamy sand (1.5 % organic matter, pH 6.5). The tested plants were developed from seeds in the greenhouse in mid-June 2014. Seeds were sown at the beginning of July in the greenhouse. Three-week-old seedlings were planted 50 cm apart in rows with 60 cm between rows. The experiment was a randomised complete block design with three replications. Each plot consisted of 10 plants in one row. Fertilisation, irrigation, and pest and disease control followed the current recommendations for cabbage plants. Cabbage heads were harvested gradually from the second week of September until the end of October, when they reached maturity. Masses, and lengths and widths of heads were measured, and the head shape coefficient (length/width) was calculated. The results of the head mass and head shape coefficients for the parental lines, F1 plants and BC3 generation were elaborated statistically using an ANOVA with the STATISTICA package StatSoft v. 10. The significances of differences among the means were evaluated by Tukey’s test at P = 0.05. Morphological characteristics of the Chinese cabbage population, such as lengths of vegetation periods from planting to harvest, intraline uniformity, plant vigour, head quality, and the colour and waxiness of external leaves, were classified separately for each plot according to the multigrade UPOV scale.

Results

Generative propagation

Rapeseed (B. napus) and Chinese cabbage (B. rapa) genotypes selected for interspecific hybridisation were characterised by a high self-compatibility and a good capability for generative propagation when pollinated at the open flower stage in 2011 (Online Resource, ESM_1.pdf). The total number of seeds (900) obtained for the C1162 male-sterile rapeseed line was higher than that for the P1162 fertile line (205). The average number of seeds/silique (18.0) for the sib pollination (C1162 × P1162) was fourfold higher than for self-pollination of the P1162 line (4.1 seeds/silique). Three B. rapa inbred lines developed 144 (MR) to 259 (KW) seeds after self-pollination in 2011 and 72 (HG) to 266 (KW) seeds in 2012. Line KW set also the highest average number of seeds/silique in both years of propagation (16.6 and 11.1 in 2011 and 2012, respectively). Cross-pollination between the male-sterile C1162 rapeseed line and three fertile Chinese cabbage lines resulted in the development of 373 (C1162 × KW) to 191 (C1162 × HG) seeds of interspecific F1 hybrids. The highest average number of seeds/silique (16.3) was obtained for the (C1162 × KW) F1 hybrid. Lower seed productivity was found in the interspecific crosses between fertile rapeseed line P1162 and Chinese cabbage superseding lines [12 seeds (P1162 × HG) to 95 seeds (P1162 × KW)].

In 2012, seeds of the BC1 generation were obtained for each combination of interspecific F1 hybrids and B. rapa superseding lines. The numbers of seeds of the BC1 generation were diverse and ranged from 47 [((P1162 × HG) × HG)] to 281 [((C1162 × MR) × MR)]. The average number of seeds/silique for the BC1 generation (1.9) was lower than that of the F1 hybrids (6.4) and those of their parental rapeseed and Chinese cabbage lines (11.1 and 9.8, respectively) (Online Resource, ESM_1.pdf). The V% for the propagated population was diverse, ranging from 6.5 % (P1162) to 64.4 % (KW).

FCM analysis

The FCM analysis of young, fully expanded leaves revealed that all of the Brassica genotypes were characterised by a high degree endoreplication, which was illustrated on the histograms by at least three distinct peaks corresponding to 2C, 4C and 8C DNA nuclear contents. The FCM analysis carried out with Z. mays (2C = 5.43) as an internal standard revealed that the relative 2C nuclear DNA contents of B. napus lines P1162 and C1162 were similar, at approximately 2.5 pg. Also, B. rapa genotypes did not differ significantly in 2C genome sizes that were approximately 1.1 pg. The genome sizes of all of the F1 hybrids (1.71–1.83 pg) demonstrated values approximately intermediate to those observed in the parental species (Online Resource, ESM_2.pdf).

Morphological characteristics for F1 and BC1 interspecific crosses of B. napus and B. rapa

Rapeseed and Chinese cabbage genotypes had morphological characteristics typical of their species and good intraline uniformity (Table 1). B. napus inbred lines (P1162 and C1162) were tall, with greater plant diameters in comparison with B. rapa lines (KW, MR and HG). Leaves of B. napus plants were large, lyrate-type, with a grey-green colour. Their leaf blades were wide with a strong waxiness, weak blistering and dentated or serrated leaf blade edges (Fig. 1d). All of the B. rapa lines had smaller plant diameters and were shorter in comparison with B. napus lines. They were also characterised by a semi-compact plant type with low (HG) or very low (KW and MR) stack heights. Leaves were typical for Chinese cabbage: small in size, whole with petioles, having green coloured blades and very weak, or absent, waxiness. The shape of the leaf blades for B. rapa lines was broad obovate (KW) or elliptical (MR and HG), with medium width, intermediate blistering and crenated leaf blade edges.

Table 1 Morphological characters of parental B. napus and B. rapa lines, their interspecific crosses of F1 and back-cross 1 (BC1) generations in the vegetative phase
Full size table

The interspecific F1 hybrids of B. napus × B. rapa were internally uniform with intermediate morphological traits typical of their parental components (Fig. 1e). All of the F1 generation plants were large in diameter, of medium height and had extensive leaf growth. Their stack heights, size, colour and waxiness of leaves were less than those of the rapeseed lines. The elliptical shaped leaves, with intermediate blistering, of the interspecific F1 hybrids were more typical of the Chinese cabbage than the rapeseed parental line. However, the erect position and lyrate type of leaf with dentated edges were more typical for rapeseed than for B. rapa (Table 1).

The BC1 generation ((B. napus/B. rapa) × B. rapa) was more diverse in the majority of the morphological characteristics. All of the BC1 plants had very small stack heights, semi-compact plant types and horizontal or half-erect leaf positions, typical of Chinese cabbage recurrent parents. Leaves with absent or very light leaf waxiness with elliptical, medium-sized leaf blades, intermediate blistering and crenated leaf edges, typical of B. rapa plants, were also found. Plant diameters of the interspecific BC1 population ranged from small to large, and plant heights were small or medium. The most evident diversity of the BC1 population occurred in the leaf morphological characteristics, such as size, innervation, waxiness, shape, width, edge and blistering of leaf blades (Fig. 1f–h). Three types of leaves—whole without petiole, semi-lyrate and lyrate—were present. Most of the BC1 plants had an intermediate level of leaf innervation; however, plants with weak or strong innervation, and leaf colours from light to dark green, were also observed. Parental lines evaluated for morphological characteristics in the greenhouse developed flower stacks typical for rapeseed and Chinese cabbage. The average size of blooming rapeseed plants was the highest of all of the tested genotypes and ranged from 180 to 190 cm (Table 2). Rapeseed lines started to bloom in the third week of March 2012, 2 weeks later than the three Chinese cabbage genotypes. Flowers of the P1162 rapeseed line were male fertile, with normally developed anthers and abundant pollen grains (Fig. 2a).The C1162 rapeseed line had flowers typical of Ogura-type CMS, having smaller anthers without pollen. The P1162 and C1162 lines developed bulgy-type buds 7.2–7.8 mm long and flowers in sizes of 19 and 17 mm, respectively. The petals of the rapeseed lines were close to one another, with round shapes and a yellow colour. The three Chinese cabbage parental lines were smaller (90–100 cm) compared with the rapeseed genotypes. They developed short (5 mm), cylindrical buds and male-fertile flowers that were small in size (12.1–12.9 mm), with ovate petals and a dark yellow colour (Fig. 2b).

Table 2 Morphological characters of parental Brassica napus and Brassica rapa lines and their interspecific crosses of F1 and back-cross 1 (BC1) generations in the generative phase
Full size table
Fig. 2
figure 2

Flower stacks of the a P1162 rapeseed inbred line, b Chinese cabbage KW inbred line, c Ogu-cytoplasmic male-sterile (CMS) interspecific F1 hybrid Brassica napus × Brassica rapa, d flowers of the Ogu-CMS back-cross 3 (BC3) Chinese cabbage line CKW, and e flowers of the male-fertile recurrent parental KW line

Full size image

The interspecific F1 hybrids of B. napus × B. rapa had intermediate plant heights (140–170 cm) in the generative stage. They started to bloom approximately 5 days earlier than the rapeseed and 10 days later than the Chinese cabbage lines (Table 2). The buds of F1 hybrids were cylindrical, similar to Chinese cabbage but equal in size to the buds of rapeseed (7.3–8.0 mm). The flowers of F1 hybrids (18.9–20.1 mm) were larger than those of both of their parental components. The shapes and positions of petals were similar to those of Chinese cabbage, but the yellow colour of the F1 hybrids was more typical of rapeseed lines. The interspecific F1 hybrids obtained from the crosses of emasculated fertile P1162 rapeseed × Chinese cabbage were also male fertile. All interspecific F1 hybrids of male-sterile C1162 B. napus × B. rapa were male sterile (Fig. 2c).

The average height of plants for the BC1 generation (125–140 cm) was less than that of the F1 hybrids. The BC1 genotypes started to bloom at the same time as their Chinese cabbage recurrent parents, in the first week of March. The sizes and shapes of buds of the BC1 generation, as well as the characteristics of flowers, were more typical of Chinese cabbage than of the interspecific F1 hybrids (Table 2). Three BC1 lines, ((C1162 × KW) × KW), ((C1162 × HG) × HG) and ((C1162 × MR) × MR), were male sterile. Three other BC1 lines, ((P1162 × KW) × KW), ((P1162 × HG) × HG) and ((P1162 × MR) × MR), were male fertile with lower pollen development capabilities than those of their parental components.

Seed effectiveness of the BC2 and BC3 generations of CMS B. napus × B. rapa propagated in the field

The BC2 and BC3 generations of the three Chinese cabbage lines (CKW, CHG and CMR) with Ogura cytoplasm were 100 % male sterile. Besides having pistils without pollen, male-sterile lines developed flowers (Fig. 2d) that did not differ from their fertile recurrent components (KW, HG and MR) (Fig. 2e). The use of the red mason bee as a pollinator insect enables the effective cross-pollination between male-sterile and male-fertile superseding lines and sib-pollination of the fertile components. The capability for seed development in male-sterile lines was similar in comparison with their fertile components (Table 3). In 2013, the average mass of seeds/plant for B. rapa fertile lines ranged from 13.6 g (MR) to 16.6 g (HG); the mass of seeds for the male-sterile BC2 generation was slightly lower, from 9.2 g (CKW) to 14.0 g (CHG). In 2014, the average seed productivity for Chinese cabbage fertile lines (38.4 g) and male-sterile lines of the BC3 generation (56.0 g) was higher than that in 2013. The two male-sterile lines (CKW and CHG) of the BC3 generation developed a higher mass of seeds/plant (65.2 and 69.2 g) than did their recurrent fertile pollinators (41.6 and 39.1 g, respectively).

Table 3 Seed effectiveness of cytoplasmic male-sterile (CMS) Brassica napus × B. rapa of back-cross 2 (BC2)/BC3 generations in comparison to Brassica rapa recurrent parental lines propagated in the field under isolated growth cages with the use of Osmia rufa as a pollinator
Full size table

Agronomic characteristics of the CMS Chinese cabbage lines of the BC3 generation

Rapeseed P1162 and C1162 inbred lines had the longest periods of vegetation (120 days from planting to the end of the growing season) when plants started their vernalisation (Table 4). Both B. napus lines were uniform with medium vigour, dark green coloured leaves and the medium waxiness typical of this species (Fig. 1i). Three Chinese cabbage parental lines had the shortest periods of vegetation, from 71 (HG) to 77 days (KW), good intraline uniformity, green coloured leaves and very weak, or absent, waxiness (Fig. 1k). Two lines (KW and MR) developed high-quality commercial heads with masses greater than 1 kg, while the HG line had smaller heads (0.71 kg). Interspecific F1 hybrids (CMS B. napus × B. rapa) developed plants with intermediate morphological characteristics of rapeseed and Chinese cabbage—without heads and with low vigour, dark green coloured leaves and strong waxiness (Fig. 1j). All of the BC3 populations were more similar to their recurrent parental components with respect to their agronomic traits (Table 4). The vegetative length for the BC3 generation was approximately 1 week longer (86 days from planting to harvest) than those of the fertile inbred lines of Chinese cabbage (71–77 days). The CMS lines of the BC3 generation were characterised by the partial uniformity of morphological traits, and high or medium level of plant vigour. The CKW BC1 line developed heads with significantly higher masses than those of the Chinese cabbage parental components, while the CHG and CMR lines were comparable with the fertile B. rapa inbred lines. The BC3 generation produced longer, more elliptical heads than its recurrent parental components. CMR BC3 lines developed green (CKW and CHG) or dark green (CMR) leaves. Line CMR had leaves without wax, but the CHG and CKW genotypes had strong or medium waxiness, which is atypical of parental Chinese cabbage lines.

Table 4 Agronomic characters of cytoplasmic male-sterile (CMS) Chinese cabbage lines of the back-cross 3 (BC3) generation in comparison to F1 CMS Brassica napus × Brassica rapa and their parental components
Full size table

Discussion

A method of developing novel and valuable breeding materials for Chinese cabbage with Ogura-CMS cytoplasm from B. napus (rapeseed) was presented in this study. We showed that interspecific F1 hybrids and consecutive BC1–BC3 generations had 100 % male fertile flowers only when the male sterile CPN1162 rapeseed line was the source of the Ogura-CMS trait. Interspecific F1 hybrids (B. napus × B. rapa) and BC1 generations ((B. napus × B. rapa) × B. rapa) obtained using the male-fertile maternal P1162 line (with B. napus cytoplasm) were male fertile. The C1162 and P1162 rapeseed lines were isogenic and differed only in respect to the cytoplasm type. For that reason the male fertility/sterility character of interspecific hybrids and their consecutive BCs was determined by the maternal component. The CMS-Ogura retained all of its characteristics in the consecutive F1–BC3 generations, irrespective of different environmental conditions and genetic backgrounds, with normal floral development in contrast with the pol CMS described by Fan et al. (1986). For plant breeders such stable expression of the maternally inherited male sterility through F1–BC3 generations should be sufficient proof for the presence of Ogu-INRA cytoplasm. The other factors, such as additional chromosomes from B. napus, probably had no impact on the expression of the CMS trait observed in consecutive BCs.

Our results have confirmed that evaluating genome sizes using FCM is a fast and reliable way to identify interspecific hybrids when parental genotypes differ considerably in genome size. The nuclear DNA contents of B. rapa and B. napus (approximately 1.1 and 2.5 pg, respectively) are comparable to the values reported by other authors (Arumuganathan and Earle 1991; Johnston et al. 2005), whereas the values of the F1 hybrids revealed that they had intermediate DNA contents compared with those of the parental species. In addition, the values are similar to the theoretically expected (i.e., approximately 1.8 pg). Previously, FCM was used to detect interspecific hybrids obtained from reciprocal crosses between B. napus and B. rapa, as well as B. napus and B. juncea (Sabharwal and Doležel 1993). FCM was also useful for the very early detection of interspecific hybrids within the genera of Lilium (van Tuyl and Boon 1996), Centaurium (Banjanac et al. 2014) and Hieracium (Morgan-Richards et al. 2004). In our study, the FCM of young, fully expanded leaves revealed that all of the Brassica genotypes were characterised by high degrees of endoreplication. This phenomenon, also called endoreduplication, endoploidisation or polysomaty, refers to a well-known process in which multiple rounds of DNA replication occur without subsequent chromosome separation and cytokinesis (De Veylder et al. 2011). Such a high degree of endoreplication was also found in root samples of B. oleracea var. capitata 3–4-day-old seedlings, but this phenomenon was not observed in leaf samples (Šesek et al. 2005). Since the degree of endoreplication may be dependent on physiological or developmental phases, and may vary in different plant organs (Šesek et al. 2005; Rewers et al. 2009; De Veylder et al. 2011), we suppose that such endoreplication could be related to the physiological phase of Brassica plants undergoing vernalisation from which the samples were collected.

The results obtained in this study showed that interspecific F1 hybrids of B. rapa and B. napus had intermediate traits of both parents; however, Wan et al. (2013) reported that F1 hybrids of tsai–tsai (B. rapa) × B. napus were more similar to the female B. napus parent. The high level of diversity and evident segregation of morphological characteristics among the interspecific BC1 population of B. napus/B. rapa were also reported for the genotypes obtained by Wan et al. (2013). The relatively good quality of the CMS genotypes of the BC3 generation in comparison with their superseding Chinese cabbage lines reveals that most of the C chromosomes may have been eliminated in the very early BC stages of rapeseed × Chinese cabbage interspecific crosses. However, the lack of internal uniformity among all of the populations of the BC3 generation showed that back-crossing should be continued for at least two more generations to obtain homozygous inbred lines suitable for the breeding of F1 hybrids. Wan et al. (2013) also suggested that the recombination during back-crossing between A and C chromosomes may affect the higher-than-expected level of atypical traits among CMS B. rapa lines. This hypothesis was supported by the analysis of non-parental amplified fragments length polymorphism fragments, which were produced at much higher numbers than expected for the advanced BC populations of B. rapa and CMS B. napus.

The good seed productivity of BC1–BC3 generations of CMS Chinese cabbage lines obtained in this study showed that sexual hybridisation with B. napus can be effectively used for breeding (Kaneko and Bang 2014). The development of high volumes of seeds was possible because of the lack of the physiological barriers between B. napus and B. rapa, physical proximity and the flowering synchrony of both components (Anonymous 2002; Bing et al. 1996; Elling et al. 2010). The utilization of three highly self-compatible Chinese cabbage inbred lines (KW, HG and MR) for the interspecific hybridisation with rapeseed allowed for the efficient cross-pollination at the open flower stage in consecutive BCs. The use of the red mason bee to pollinate the BC2 and BC3 generations of CMS B. rapa lines in the field proved to be very effective and less time-consuming than pollinating by hand in the greenhouse, which is in agreement with the observation made by Steffan-Dewenter (2003). Greater masses of seeds for the CMS BC3 generation in comparison with the self-pollinated inbred lines of fertile Chinese cabbage were probably caused by their higher vigour and the development of stronger blooming plants with abundant flowers, which is typical of heterosis. Relatively lower seed setting rates in 2013 compared with 2014 were probably affected by the low temperatures during the blooming season in April and May, as well as the extremely high volume of rainfall during seed development in June 2013. Weather conditions during hand-pollination, especially temperature and humidity, may be the reason for the high V% for seed development in the greenhouse. The lower seed effectiveness for the interspecific F1 hybrids of fertile P1162 rapeseed crossed with B. rapa in comparison with interspecific F1 hybrids of the CMS C1162 rapeseed line crossed with B. rapa was probably affected by the different pollination techniques (pollination at the green bud stage after emasculation and pollination at the open flower stage, respectively).

The method presented in this study can play an important role in the development of new and valuable sources of genetic diversity with stable pollen sterility and without any evident changes in vegetative growth or female fertility for Chinese cabbage. Our results showed, for the first time in the available literature, that the CMS-Ogura system could also be very effective, inexpensive and reliable for the breeding of Chinese cabbage male-sterile F1 hybrids not only in Poland, but also in other regions where Chinese cabbage plays an important role as a vegetable crop.

Conclusions

Our paper presents the development of a novel and valuable Chinese cabbage breeding material with Ogura cytoplasm through the use of interspecific hybridization. The successful introgression of Ogura-CMS from B. napus (rapeseed) into B. rapa (Chinese cabbage), effectiveness of seed development and characteristics of interspecific F1 hybrids and BC1–BC3 populations were described in this study for the first time. The interspecific F1 hybrids had intermediate traits of both parents; however, the morphological characteristics segregated in the BC1 generation, and, in general, the majority of plants were more similar to B. rapa. A high seed effectiveness for the interspecific F1 hybrids and consecutive BCs was obtained both in the greenhouse and in the field. Plants of the B. rapa BC3 generation had a high level of vigour and the good morphological characteristics typical of Chinese cabbage recurrent lines, but they need further BCs to improve the uniformity and stability. As a result of the study, a valuable collection of Chinese cabbage genotypes with the CMS trait, which could be very useful for producing domestic F1 hybrids as well as for studies on their chromosomal and molecular characteristics, was obtained for the first time in Poland. The material presented in this paper will be supplied for research purposes upon request.