Location: Home >> Detail
This work is licensed under aCreative Commons Attribution 4.0 International License
Crop Breed Genet Genom. 2019;1:e190007. https://doi.org/10.20900/cbgg20190007
1 Department of Plant Breeding, Justus Liebig University, Heinrich-Buff-Ring 26-32, Giessen 35392, Germany
2 Department of Plant Production, Faculty of Agronomy, University of Concepción, Av. Vicente Méndez 595, Chillán, Chile
* Correspondence: Annaliese S. Mason.
This article belongs to the Virtual Special Issue "Genetic Gains in Plant Breeding"
Interspecific hybridization is widespread in nature, where it can lead to either the production of new species or to the introgression of useful adaptive traits between species. In agricultural systems, there is also great potential to take advantage of this process for targeted crop improvement. In the Brassica genus, several crop species share close relationships: rapeseed (Brassica napus) is an ancestral hybrid between turnip (B. rapa) and cabbage (B. oleracea), and mustard species B. juncea, B. carinata and B. nigra share genomes in common. This close relationship, plus the abundance of wild relatives and minor crop species in the wider Brassiceae tribe which readily hybridize with the Brassica crop species, makes this genus an interesting example of the use of interspecific hybridization for crop improvement. In this review we introduce the Brassica crop species and their wild relatives, barriers to interspecific and intergeneric hybridization and methods to overcome them, summarize previous successful and unsuccessful attempts at the use of interspecific hybridization for crop improvement in Brassica, and provide information about resources available to breeders wishing to take advantage of this method in the Brassica genus.
The Brassica genus belongs to the tribe Brassiceae (family Brassicaceae). This family comprises 338 genera (assigned to 25 tribes) and 3709 species [1,2]. The members of this family are mostly herbs with annual, biennial or perennial growth habits . Initially this family was known as “Cruciferae” due to its characteristic flower conformation of four petals arranged in a cross-shape . Most of the member species are distributed in temperate regions, with the first center of diversification located in the Irano-Turranian region (~150 genera and ~900 species), followed by a second center of diversification in the Mediterranean region (>110 genera and ~630 species).
Brassica is the most prominent genus in the Brassicaceae family and includes 39 species . Many of the species in this genus are cultivated for their edible roots, leaves, stems, buds, flowers, mustard and oilseeds . For 33 of the species the chromosome number has been determined, and ranges from n = 7 up to n = 20 . During the 1930s, the chromosome number and genetic relationships between the cultivated Brassica species was established [6,7]. The diploid species B. rapa (AA, n = 10), B. nigra (BB, n = 8) and B. oleracea (CC, n = 9) were determined to be the progenitors of the allopolyploid species B. juncea (AABB, n = 18), B. napus (AACC, n = 19), and B. carinata (BBCC, n = 17), in a relationship known as “U’s Triangle” . Based on chloroplast DNA data it was determined that B. nigra belongs to a different lineage (Nigra lineage) than B. rapa and B. oleracea (Rapa/Oleracea lineage), with the two lineages diverging approximately 7.9 Mya . The divergence between B. rapa and B. oleracea has been estimated to have occurred perhaps 3.75 Mya  to about 5 Mya . Later on, approximately 7500 years ago or less, diploid species B. rapa and B. oleracea hybridized to produce B. napus L. .
Genetic diversity within Brassica species has been broadly studied, with a special focus on the six crop species that form the U’s triangle. Of these species, three are highly diverse: B. oleracea, B. rapa and B. juncea [13,14]. These species are quite morphologically variable, presenting different leaf types, numbers of branches per stem, inflorescence types, and stem thicknesses; these variations also lead to different end-product usage (e.g., oil or vegetable type). Genetic diversity observed in the Brassica allopolyploids can be due to (i) multiple hybridization events with diverse parents (or possibly subsequent backcrossing of the newly formed allotetraploids to the parent species) and (ii) genome changes occurring after polyploidization . Four Brassica species are mainly used as oilseed crops: B. juncea, B. rapa, B. carinata and B. napus .
Brassica napus (rapeseed, oilseed rape, swede) is the most economically important of the Brassica crop species, occupying the third position worldwide in the oil vegetable market, after soybean and palm oil. In the year 2016, worldwide production of rapeseed was over 68 million tons (Mt) (www.fao.org/faostat/, November 2018): In Germany, a large proportion of the rapeseed oil produced is used to generate biodiesel (2017: 4 Mt of biodiesel produced, source: European Biodiesel Board). Rapeseed, as well as other members of the Brassicaceae, naturally contain 20–40% erucic acid  and high glucosinolates in the seed meal. However, rapeseed has been extensively bred for low erucic acid and low glucosinolates  to produce a type of rapeseed better known as canola. The main producers of rapeseed are Canada, China and India, which together represent almost 60% of the total production worldwide (www.fao.org/faostat/, November 2018). Winter-type rapeseed is mainly grown in Europe, and spring types are mostly grown in Canada, China and Australia . Brassica napus (AACC, 2n = 4x = 38) is thought to have originated in the last 7500 years via at least two different hybridization events between B. oleracea and B. rapa in agricultural systems . Unfortunately, most of the genetic variation in oilseed rape has been reduced due to intensive selection for low erucic acid and low glucosinolate content traits . Rapeseed is not found in nature as a wild type, and most of the diversity existing nowadays comes from breeding programs or cultivars from different countries .
Brassica juncea (AABB, 2n = 4x = 36) is also used as a vegetable, with leaf mustard or Indian mustard as the common name . A huge diversity of leaf morphotypes is present in this species that is thought to have been influenced by human selection , with two representative gene pools: East Europe and Indian . Mustard is mainly grown in India due to climate conditions, where the breeding objectives are mainly focused on improving seed yield . Although genetic resources available for B. juncea are not as comprehensive as those available for B. napus and its progenitor species, a reference B. juncea genome was published in the year 2016 .
Brassica rapa (AA, 2n = 2x = 20), initially named B. campestris and commonly known as turnip or Chinese cabbage, has its origins in the Mediterranean and Central Asia . The different subspecies of B. rapa can be used as a fodder (e.g., subsp. rapifera), vegetables (e.g., subsp. chinensis or pekinensis), or as an oilseed crop (e.g., subsp. oleifera). Brassica rapa, Chinese cabbage accession Chiifu-401-42, was the first Brassica species to get its genome sequenced . Of the estimated genome size of 485 Mb, 283.8 Mb was initially assembled . Later on, an improved assembly was released (v2.0) that increased the size of the genome assembly to 389.2 Mb . The B. rapa genome is rich in transposable elements, accounting for 32.3% (~54 Mb) of the assembled sequence , much more than the 10.0% observed in the related genome of Arabidopsis thaliana .
Brassica oleracea (CC, 2n = 2x = 18) is mainly used as an edible vegetable. This species is composed of several varieties and morphotypes are usually referred to as coles. These vegetables are rich in vitamin C, folate and calcium . Different varieties include Brussels sprouts (var. gemmifera), cabbage (var. capitata), cauliflower (var. botrytis), and Chinese kale (var. alboglabra). In the year 2016, the worldwide production of cauliflower and broccoli surpassed 25 million tons (www.fao.org/faostat/, November 2018). Some new vegetables have also been produced by crossing different varieties within this genus, such as broccolini . Two draft genome references for B. oleracea were published in 2014 [28,29].
Brassica carinata (BBCC, 2n = 4x = 34), also called Ethiopian mustard, possesses wide genetic variability and is also used as an oilseed crop . This crop has also been considered for use in biodiesel production  and for other purposes including as a condiment, medicine and vegetable .
Brassica nigra (BB, 2n = 2x = 16) was previously used as a condiment mustard but has now been mostly replaced by B. juncea . Compared to the major Brassica crops, B. nigra contains little variety in physical appearance , but it nevertheless possesses different agronomical traits of great value such as resistance to Phoma lingam . Although B. nigra is the least agriculturally significant of the six Brassica crop species, a scaffolded genome assembly (not yet assembled into pseudomolecules) was made available in 2016 alongside the B. juncea genome , and a new chromosome-level assembly was released in 2019 .
In the 1970s, Harberd defined the term “coenospecies” for those species or genera that have sufficient relatedness to the six Brassica crops to be able to exchange genetic material with them [34,35]. The coenospecies are composed of almost 100 wild species and genera that can potentially be used to increase diversity, and to introgress useful traits such as disease resistance or abiotic stress . Harberd also classified the Brassica coenospecies into biological units called “cytodemes” [34,35,37]. Each cytodeme can contain more than one genus or species, but all species within a cytodeme should have the same chromosome number, and readily cross with other species in the same cytodeme to produce fertile, vigorous hybrids. Based on these criteria, the Brassica coenospecies were initially classified into 38 cytodemes , covering nine genera from the subtribe Brassiceae (Brassica, Coincya, Diplotaxis, Eruca, Erucastrum, Hirschfeldia, Sinapis, Sinapidendron, and Trachystoma) and two genera from subtribe Raphaninae (Enarthrocarpus and Raphanus). This was later updated to 63 , after the addition of three genera (Moricandia, Pseuderucaria, and Rytidocarpus) from the related subtribe Moricandiinae . The crossability between cytodemes is low, but certain tools can be used to increase success rates (as discussed in later sections of this review). Crossability can also be influenced by the direction of the cross, i.e., which species is used as the maternal parent, which is referred to as “unilateral incompatibility” .
An extended list of potentially useful agronomic traits for crop improvement present in wild allies of the Brassica species can be found in . Examples include resistance to white rust (Albugo candida) in Brassica maurorum  and Eruca versicaria ssp. sativa , resistance to Alternaria blight in Brassica fruticulosa  and Trachystoma ballii , resistance to beet cyst nematode in Raphanus sativus  and Sinapis alba , and resistance to blackleg/Phoma disease (Leptosphaeria maculans) in Sinapis arvensis , Sinapis alba , Thlaspi arvense , and B. tournefortii . The Brassica crop species also contain unique, useful traits: examples include resistance to powdery mildew (Hyaloperonospora parasitica) in Brassica oleracea , resistance to clubroot disease (Plasmodiophora brassicae) in B. rapa, B. oleracea and B. napus , and pod shatter resistance and tolerance to heavy metals in B. juncea . More exotic traits of interest include a C3–C4 intermediate photosynthetic system in Moricandia  and Diplotaxis species [56,57], and high erucic acid levels in Crambe abyssinica . Cytoplasmic male sterility in Brassica could also be conferred by hybridization with Sinapis incana  and Diplotaxis siifolia , among other examples.
Direct wide hybridization has been attempted many times between Brassica and various wild relative species, with different levels of success (reviewed in ). Originally such hybrids were produced to resolve chromosome homoeology (phylogenetic relationships) or simply out of curiosity . However, crossing with distant relatives is today attracting increasing recognition as a method with which to improve agronomic traits in high-end varieties. There are many examples of the successful introgression of new traits into Brassica crops. Initial attempts to create hybrids between Brassica species started in the early 1800s. At this time, some crosses were made between B. napus × B. rapa and B. oleracea × B. rapa. Different success rates were reported and the results were published in 1925 by . Later on, a compilation of crossability between species in the Brassica, Raphanus and Sinapis genera was published, showing that interspecific hybrids can be made between the Brassica crops and many closely-related wild species .
The occurrence of natural hybridization between distant relatives in natural conditions is low. For instance,  found that hybridization between Brassica napus, B. rapa and B. juncea and their two weedy relatives B. nigra and Sinapis arvensis does not occur under open pollination conditions in the field, although B. rapa, B. juncea and B. napus all readily produce hybrid progeny with each other under the same conditions. The cross between B. napus (2n = 38) and Raphanus raphanistrum (2n = 18) has also been assessed under field conditions. In this case, just two allopolyploid hybrids (2n = 56) were obtained from more than 52 million B. napus seedlings when this species was used as a female, showing a hybridization frequency of 4 × 10−8 in field conditions. These results indicate that the likelihood of this cross in the wild is low, which shows the importance of conducting such hybridizations under controlled conditions.
The introgression of genes for disease resistance between species has been widely studied in Brassica. One example is the utilization of the B genome as a source of resistance against Leptosphaeria maculans (blackleg) from diploid and tetraploid species. For instance, chromosome B4 from B. nigra was introgressed into rapeseed variety “Darmor”, which showed high resistance with the addition of this chromosome . Similarly, high resistance from B. juncea was obtained in selected recombinant lines of B. napus carrying a resistance gene located on chromosome B8 . A similar study  successfully introgressed a B-genome chromosome from B. carinata to B. napus, with plants carrying this chromosome showing variation in traits such as blackleg resistance, days to flowering, days of maturity, and fatty acid composition. Another example is the improvement of resistance against Erysiphe polygoni (which can cause powdery mildew disease). Resistance in 100% of BC1 progeny was successfully demonstrated in hybrids obtained by hand crossing and embryo rescue between B. carinata (donor) × B. oleracea . Other cases of resistance transfer include transfer of blackrot resistance from B. carinata to B. oleracea , resistance to Brassica leaf blight caused by Alternaria brassicae from B. hirta to B. juncea  and transfer of powdery mildew resistance from B. carinata to B. oleracea through embryo rescue followed by backcrossing to B. oleracea .Yellow Seededness
Yellow seededness is a desirable trait in Brassica, as yellow seeds have less fiber, higher protein, and higher oil content than black seeds. Although B. juncea and B. rapa contain yellow-seeded traits, this trait is not found in rapeseed (B. napus). Using monosomic alien addition lines from the cross B. rapa × B. oleracea, Heneen et al.  found that seven of the nine C chromosomes carry genes that affect seed color, showing the complexity of this phenotype. Interspecific crosses between B. alboglabra, B. rapa var. “yellow sarson”, yellow seeded B. carinata and black seeded B. napus have been carried out previously to attempt to produce yellow-seeded B. napus, with interspecific hybrid progeny showing different degrees of seed colour . However, this study demonstrated that the combination of the C genome of yellow-seeded B. carinata with the A genome of “yellow sarson” does not result in a yellow-seeded B. napus. The expression of this trait also appears to be heavily affected by the environment. Rashid et al.  crossed [(B. napus × B. juncea) × B. napus] × [(B. napus × B. carinata) × B. napus] and successfully obtained yellow seeds. However, when these plants were tested in the field the color was found to be highly affected by temperature .Male Sterility
A common use of wild relatives for Brassica crop improvement is in the production of male sterile lines to facilitate hybrid production. Male sterility is often conferred when cytoplasm from an alien species is present in the genetic background of another species: this is referred to as cytoplasmic male sterility, or CMS. The most successful example of this approach in Brassica is the Ogura CMS system, where alien cytoplasm was obtained from crossing Brassica napus to Japanese radish (Raphanus sativus). This system was subsequently widely used in B. napus, B. juncea and B. oleracea . Several other CMS systems have also been successfully developed from interspecific hybridization events, including a novel CMS system in B. juncea incorporating the cytoplasm of B. fruticulosa , and the Nsa CMS system in B. napus utilizing Sinapis arvensis cytoplasm . On the other hand, several attempts to produce additional CMS lines through interspecific hybridization have also been unsuccessful. Seventeen crosses between Diplotaxis species and B. napus were done in order to introgress CMS, but out of hundreds of crossings using conventional techniques only crosses with D. muralis and D. erucoides were successful, and no CMS system was consequently established . Protoplast fusion has been used to transfer Ogu cytoplasmic male sterility factor from Brassica napus to Brassica juncea and for the improvement of male sterile lines in hybrid breeding systems . Somatic hybridization between B. juncea and B. oleracea has also been used to transfer cytoplasmic male sterility and resistance to Turnip mosaic virus from B. oleracea to B. juncea [82,83]. Prakash et al.  successfully obtained both stable CMS B. juncea and an introgression line carrying the restorer gene via somatic hybridization between M. arvensis and B. juncea followed by backcrossing with B. juncea.Oil Quality Traits
Interestingly, oil quality traits have also been successfully transferred between species for crop improvement in Brassica. In the case of rapeseed, low erucic acid and low glucosinolate content originate from two B. napus cultivars: “Liho” with low erucic acid and “Bronowski” with low glucosinolate content . Another possible source of these oil quality traits is Capsella bursa-pastoris, which can show less than 1% erucic acid and less than 16 µmol/g of glucosinolates in the seeds, as well as high resistance to Sclerotinia sclerotiorum . Previously, several chromosomes and chromosomal fragments from C. bursa-pastoris were successfully introgressed into B. napus and B. rapa . Another wild relative with favorable fatty acid content is Orychophragmus violaceus, which has been successfully crossed with B. napus [86,87]. From this cross, advanced progenies with 2n = 38 chromosomes, ≥70% oleic acid, 28% linoleic acid and low glucosinolate content in the seeds (<30 µmol/g oil free meal) were produced .Other Traits of Agronomic Interest
Moricandia arvensis is a plant that expresses an intermediate C3–C4 photosynthetic mechanism . This trait was introgressed into B. napus by somatic hybridization by , who obtained three hybrid plants that expressed C3–C4 intermediate photosynthesis characteristics. Dwarfism is a useful agronomic characteristic which helps avoid lodging, and which was introgressed by  from a mutant B. rapa into natural B. napus via production of a resynthesized B. napus from the mutant B. rapa with a normal B. oleracea, followed by four generations of backcrossing with natural B. napus. Pod shatter resistance has also been introgressed into B. napus from B. juncea via direct hybridization . Finally, drought tolerance has been introgressed from Sinapis alba into B. napus by somatic hybridization, and was identified at the vegetative stage in the BC3F1 vegetation, although the original target was yellow-seededness .
Interspecific hybridization has two major outcomes: introgression and speciation. While introgression transfers just a limited number of alleles, hybrid speciation produces a new hybrid species. Resynthesis is the process of reproducing an already existing species from its progenitor species. This is most often done to increase the genetic diversity of the existing allotetraploid species by incorporating some of the greater genetic diversity of the progenitor species. Resynthesis as a tool of crop improvement has many benefits. Polyploidy induced during the process of resynthesis can overcome crossing barriers due to endosperm failure in interploidy crosses . The genetic diversity of some Brassica allotetraploid crops is limited due to the few hybridization events that gave rise to these species . In the case of B. napus, geographic isolation, extensive breeding and selection for low erucic acid and glucosinolate content has further eroded the genetic diversity of this species [37,94]. Resynthesizing the Brassica allotetraploids from their diploid parents is a means of increasing the genetic diversity of these species. Studies of this method abound: Seyis et al.  resynthesized 165 Brassica napus lines by crossing B. rapa and B. oleracea progenitor species; analysis of these resynthesized lines using RFLP markers showed they were highly genetically divergent from established oilseed rape cultivars, and also showed a high degree of morphological diversity. Abel et al.  also developed resynthesized Brassica napus to study fixed heterosis by crossing 21 B. rapa and 16 B. oleracea species, and showed that the direction of the cross affects hybridization outcome, although the diversity of this population and its effect on fixed heterosis was not reported in this study. Several other studies have also reported on resynthesis of B. napus in order to expand the available gene pool [97–100], and to test for new traits such as resistance to cabbage stem weevil Ceutorhnchus pallidactylus . Brassica juncea has also been resynthesized by crossing its progenitor species B. rapa and B. nigra to broaden the genetic base of this species [102–104]. Bansal et al.  resynthesized new B. juncea genotypes, and found the resynthesized B. juncea to be morphologically diverse compared to natural B. juncea. Brassica carinata has also been resynthesized from its progenitor species, with hybrids showing morphological variation potentially useful for crop improvement [106,107].
These new synthetic polyploids are not usually being bred to become a new crop nor in competition with the elite varieties, but rather as a source of diverse new agronomic traits, where they are used to cross with and introgress these traits into high-performance cultivars . One successful example is the synthetic clubroot-resistant allotetraploid B. napus RS 15/04, which was created by crossing a resistant kale (B. oleracea ECD-15) and turnip rape (B. rapa ECD-04). This synthetic B. napus was subsequently crossed with WOSR cv. “Falcon”, and a DH line created from the F1. This line was then backcrossed with cv. “Falcon” until the BC2F1 where three dominant genes specific to a particular race of the clubroot pathogen were present. Further breeding was done, and in 2001 the clubroot-resistant winter oilseed rape cv. “Mendel” was released . Newly synthesized Brassica polyploids can also present extensive genome change at very early stages and also throughout further generations (F1–F5). This variation can also be phenotypically observed in traits like flowering time  and hybrid vigor in synthetic B. juncea , and may comprise a means of generating entirely new traits.
Efforts on Brassica improvement through polyploid synthesis have not only been limited to the naturally occurring allotetraploids. Several attempts have been made to synthesize a new, fertile and meiotically stable allohexaploid Brassica (2n = AABBCC), with varying success rates that appear dependent on both genotype and method used (reviewed by ). Synthetic allohexaploids produced from crosses between B. carinata and B. rapa followed by chromosome doubling showed bigger flowers, high silique setting and high fertility, the latter increasing from the F2 to F4: this trend is expected to continue across generations, leading to a potentially stable species which could be of benefit to agriculture . Other studies on allohexaploid Brassica have focused on using these hybrids as a bridge between species (reviewed by ), such as in the creation of novel Brassica napus genotypes exhibiting useful traits like yellow seededness via hybridization between B. rapa and B. carinata to produce 2n = AABBCC types followed by backcrossing to B. napus and elimination of the B genome [112–114].
The Raphanus genome has also been used to develop synthetic allotetraploids, as radicole (CCRR, 2n = 36) or Raparadish (AARR, 2n = 38). Both of these hybrids feature a fodder-like crop with the advantage of resistance to the beet cyst nematode. Although B. napus has some resistance to this nematode, transfer of this high-resistance trait from Raparadish to B. napus was attempted in 1993 . Surprisingly, there was no significant difference in the number of hybrids produced based on the B. napus cultivar or accession used in the crosses. In the F1 population (AACR, 2n = 38), nematode resistance was found to be intermediate between the two parental species. The meiosis observed in the F1 plants was also very variable, producing a high frequency of unbalanced and unreduced gametes.
Near and far relatives of major crop species provide us with an enormous untapped reservoir of agriculturally important traits. Transferring this genetic variation to crops through introgression breeding has helped produce improved, high yielding crops resilient to prevailing climatic conditions . The Brassica A, B and C genome species and other wild relatives contain valuable genetic variation for crop improvement, including genes or alleles for defense against pests and diseases [67,69] and drought tolerance [119,120]. Extensive interspecific and intergeneric hybridization has been performed between cultivated species, and between cultivated species and wild relatives, to develop more potentially useful cultivars with improved biotic and abiotic stress tolerances .
However, despite the potential of using hybridization to transfer useful traits from related crop species or wild relatives, there are barriers that limit the usefulness of this process. Interspecific and intergeneric hybridization barriers can be divided into two categories: pre-fertilization and post-fertilization barriers. Pre-fertilization barriers can arise due to failure of pollen germination, pollen tube growth or pollen tube penetration of the embryo [121,122]. Degradation or death of the hybrid embryo and male and female sterility in hybrid plants are some of the causes of post-hybridization barriers and hybrid sterility . Fertilization in interspecific crosses can still occur, but later on can produce embryo abortion related to problems with endosperm development . This often happens in one direction (i.e., when one species is used as the maternal parent, but not when it is used as the paternal parent) and it can be overcome when the reciprocal cross direction is tested . This has been recorded, and some examples show more success when B. napus is used as a female in interspecific hybridization events . Similarly, in some attempted crosses between B. carinata and B. rapa, F1 hybrids were only obtained when B. carinata was used as the female . The challenge of creating interspecific hybrids increases as the phylogenetic distance between the combining species increases . Opportunities for and success of interspecific crosses are also dependent on a number of other factors: physical distance between the species/parent plants, synchrony of flowering, the specific parental genotypes used, the method of pollen dissemination, the direction of the cross (which parent is female), environmental factors, and whether one parent is male-sterile [127,128].
In Brassica it is difficult to make a simple statement about reproductive compatibility and incompatibility, as reproductive compatibility relationships are complicated, with partial reproductive barriers between many species [61,127]. Despite years of research on hybridization in Brassica, the degree of reproductive compatibility between many species combinations remains untested. Detailed summaries of the extent of interspecific hybridization in Brassica have been reported by various sources [13,41,61]. Given that several factors need to be considered in creating successful interspecific hybrids, different methods have been developed to transfer useful traits between different Brassica species and to increase the genetic diversity of Brassica crops.
Failure of foreign pollen to germinate on the stigma, to grow pollen tubes or to subsequently fertilize ovules, and for fertilized ovules to develop into seeds, are all commonly observed in interspecific hybridization attempts. However, a number of strategies exist to overcome these pre- and post-fertilization barriers (reviewed by ). Early pollination of stigmas (before buds open and before full maturity) or stump pollination can help in overcoming reproductive incompatibilities between some genotypes of Brassica species , while in other cases in vitro pollination of the stigma or pistils and/or opened ovules and ovaries may facilitate the interspecific fertilization event . Seed abortion post-fertilization is also often observed in crosses between plants of different species or ploidy levels . In cases where seeds cannot be obtained from crossing, a technique where the embryo is “rescued” from the putatively hostile maternal environment, usually into tissue culture or a sterile medium, can sometimes allow the production of hybrid plants. The technique of in vitro culture to rescue interspecific hybrid embryos was first used in crosses between Lolium perenne and L. austriacum . Wide crosses between many crop plants and their wild relatives have now become possible through the use of embryo rescue techniques, as embryo rescue and subsequent culture in vitro helps to overcome post-fertilization barriers . In the production of Brassica interspecific hybrids, embryo rescue is commonly used to overcome natural reproductive barriers [94,134]. Embryo rescue was first used in Brassica by . Following this study, extensive investigations have been carried out to improve this method [136,137]. The successful application of this technique depends on the developmental stage of the embryo being rescued .
Several studies have demonstrated the importance and success of this technique in transferring useful traits between Brassica species. Using embryo rescue, triazine resistance has been transferred from B. napus to B. oleracea . Yao et al.  produced allohexaploids by crossing B. maurorum with all three Brassica allotetraploids. Herbicide resistance was transferred from Sinapis arvensis to B. juncea and B. rapa using embryo rescue . Cytoplasmic male sterility has been transferred from B. juncea and B. napus to B. oleracea . Zhang et al.  resynthesized B. napus from interspecific hybridization between B. rapa and B. oleracea, and new type B. napus types showing resistance to Verticillium longisporum were synthesized from a diverse set of B. rapa and B. oleracea through embryo rescue .Somatic Fusion
Somatic fusion is an important means of transferring useful traits from one species to another. Somatic fusion has the advantage that it can bypass these incompatibility barriers and transfer genes between sexually incompatible species . Besides the transfer of agronomically important traits, protoplast fusion can be used to modify organellar traits, as chloroplasts and mitochondria from both parental species are combined with somatic fusion, rather than only the maternal cytoplasm being inherited by the interspecific hybrid as is the case for sexual crosses. Brassica species were among the first crops used for protoplast isolation, as most parts of the plant are suitable for releasing totipotent protoplasts [142,143]. Regeneration of plants from isolated protoplasts has been reported in all Brassica species following the first report of successful plant regeneration from B. napus mesophyll tissue . Somatic hybridization has successfully been used to transfer traits such as disease resistance, oil quality, cold and drought tolerance and herbicide resistance between species [142,143]. In one example, somatic hybrids between B. rapa and B. oleracea were used to create improved B. rapa cultivars resistant to soft rot by backcrossing somatic hybrids to B. rapa . Asymmetric somatic hybridization has also been used to transfer resistance to blackleg disease from B. juncea, B. rapa and B. carinata into B. napus .Genetic Transformation
Genetic transformation can play an important role in variety improvement and functional analysis of Brassica crops. It has paved the way for the development of new Brassica varieties producing biodegradable plastics, pharmaceuticals and nutritive compounds by introducing new genes from unrelated sources . Conventional breeding of Brassica is time consuming, labor and resource intensive. On the other hand, genetic transformation provides a direct means of introducing specific genes or traits without negatively affecting the desirable genetic background . In addition, certain important traits may not be available in the existing germplasm . Under such circumstances, genetic transformation has shown to be a powerful means of effectively transferring genes across reproductive barriers .
Genetic transformation systems have been developed in almost all the economically important Brassica species, including B. napus , B. oleracea , B. juncea , B. nigra , B. carinata, and B. rapa . Different plant transformation methods exist. The direct method, where naked DNA is introduced into the protoplasts of intact cells, can be mediated by methods such as polyethylene glycol (PEG) treatment, microinjection and electroporation. Alternatively, indirect methods requiring an intermediate biological vector can be used; usually Agrobacterium tumefaciens transformation is suitable for this purpose in Brassica .
Genetic transformation has led to the introduction of new traits in to Brassica crops far beyond the species boundary: genes not present in the Brassica species. Traits improved through genetic transformation include resistance to herbicides such as glyphosate, glufosinate, sulfonylurea, bromoxynil, and bromoxynil resistance [155–157]. Oil quality improvement has also been a target of transformation. Brassica juncea and B. napus with high oleic acid have been produced by silencing the endogenous oleate desaturase . Also, transformation of the d12-desaturase genes from the fungus Mortierella alpina has led to the production of canola with high gamma-linolenic acid .
Insect and disease resistance have also been important target traits for improvement of Brassica crops. Brassica napus producing an endogenous endotoxin of Bacillus thuringiensis poisonous to the diamondback moth have been produced through transformation with the Bt cry1 gene [160,161]. Novel insect resistance in B. napus has also been developed by transformation of chitinase and scorpion genes . Transformation has been used to convert Brassica crops to biofactories producing pharmaceutical and industrial products such as biodegradable polymers ; the anticoagulant protein hirudin has been produced in B. carinata .
The development of male sterile lines and restoration system has also been a significant advancement in Brassica transformation. Male sterile plants were obtained in B. juncea by introducing the barnase gene with tapetum-specific promoters, following which the fertility of the male sterile line was restored by crossing it with a barstar containing transgenic line .Genome Editing
Recently, the clustered regularly interspaced short palindromic repeat (CRISPR)-CRISPR associated protein 9 (CRISPR/Cas9) system has emerged as a versatile molecular tool for genome editing in different organisms . It has been shown that the CRISPR/Cas9 system is able to achieve efficient gene editing in plants through either transient experiments or in the production of transgenic plants . In this system, the endonuclease Cas9 is directed to a specific DNA target by a synthetic guide RNA . It is an innovative genetic tool that can modify the genome of any species with high precision and accuracy . Although this technology is still in its early stages, its application has been demonstrated not only in model species such as Arabidopsis thaliana , but also in crops such as tomato (Solanum lycopersicum) and wheat (Triticum aestivum). In Brassica napus, proof of concept was recently demonstrated by , who targeted the two homologues and four alleles of the BnALC gene, which is responsible for fruit dehiscence in Brassica. Similar reports have since followed, such as  who determined the mutation efficiency of CRISPR-Cas9 in 12 gene families. CRISPR/Cas 9 has also been used to modify the fatty acid desaturase 2 (FAD2) gene which catalyzes the desaturation of oleic acid in B. napus leading to the production of B. napus with high oleic acid . The application of this system has been demonstrated also in B. oleracea  and B. carinata . CRISPR/Cas9 therefore promises to be an important tool in Brassica improvement. In future, linking genetic and genomic information to germplasm bank resources could extend the reach of this genome editing technique to many genetic variants of agricultural significance present within the wild relatives of the Brassica crop species, allowing direct editing of crops to mimic wild relative variants.
Wild Brassiceae species can be found around the world in temperate climates , and hence may constitute a valuable source of locally-adapted germplasm for use in crop improvement. Although all of the cultivated Brassica species are thought to originate from roughly around the Mediterranean region, with wider distributions from Europe to North Africa to the Middle East and West Asia , Brassiceae germplasm has also been identified in North America in archeological and ethnobotanical studies , with wild mustard relative Sinapis arvensis widespread 2000 years ago in North Eastern American states ( as cited in ). Other Brassiceae weeds and crop species have been identified in weedy habitats in Canada , the United States and Mexico [180,181], as well as in Australia , and of course Europe and Asia . Germplasm resources and collections of Brassica crops and related species, which are either cultivated (domesticated lines) or growing in the natural environment, are mostly (90%) conserved as seeds in cold storage in gene banks . These collections generally comprise elite and domesticated breeding lines, plus a few wild relatives which are being conserved for breeding as well as for research purposes. Overall, conservation methods can basically be categorized into either in situ or ex situ conservation.In Situ Conservation of Germplasm Resources
In situ conservation is the primary form of conservation for crop wild relatives, and either takes place in farmers’ fields or in natural environments. In situ conservation is promoted because landraces can be an essential component of indigenous cultures and show highly specialized local adaptations [183,184]. Growth of plants in the natural environment also allows selection and adaptation to changing environmental conditions and is highly cost effective . Growing interest in the use of wild species in breeding [185,186] has underlined the need to also create national in situ inventories to encourage conservation. In situ conservation also includes conservation in natural or wilderness areas, national parks and special management areas. Understanding the genetic potential of Brassica crops and wild relatives is critical for the establishment of long term breeding programs. Useful agronomic traits which can potentially be introgressed from wild relatives into elite crops include resistance traits , salt tolerance [187,188] and cold tolerance . However, to date ex situ conservation remains the most common form of germplasm conservation.Ex Situ Conservation of Germplasm Resources: Genebanks
Ex situ conservation of plant genetic resources started in the mid-twentieth century, as an initiative to prevent the rapid loss of plant biodiversity resulting from the introduction of improved varieties to replace landraces [182,190,191]. Therefore, germplasm (or “gene”) banks were established with the intention to preserve genetic material which might be useful in future for cultivation or as material in breeding programs . The major world germplasm collections of Brassica today include the Centre for Genetic Resources (CGN, The Netherlands),
the Institute for Horticultural Plant Breeding (IVT, The Netherlands),
the Horticultural Research Institute (HRI, UK) and the Gene Bank of the Crop Research Institute (UK). Other genebanks include the United States Department of Agriculture (USDA) (https://www.ars-grin.gov/npgs/) in the United States, the Australian Grains Genebank (https://grdc.com.au/resources-and-publications/groundcover/gc110/australian-genebank), and the Nordic Genetic Resource Centre (NordGen) (https://www.nordgen.org/en/) in Norway. In Spain, the Brassica genebank MBG-CSIC (http://www.mbg.csic.es/es/) started its activities in 1985. This gene bank holds a collection of Galician Brassica crops belonging to the species B. oleracea L., B. rapa L. and B. napus L., and houses a total of 644 accessions. B. oleracea varieties include kales (B. oleracea var. acephala), cabbages (B. oleracea var. capitata), and Tronchuda cabbage (B. oleracea var. costata). Brassica rapa includes the turnips, turnip greens, and turnip tops; and B. napus appears only in the form known as “nabicol” or leaf rape . The United Kingdom Vegetable Genebank (UKVGB) managed by the University of Warwick conserves approximately 14,000 accessions of crops including Brassica types . Brassica genetic resources hosted at the UKVGB have been incorporated into several germplasm panels, including (amongst others) the European clubroot differential series (ECD) to help identify races of the clubroot-causing pathogen Plasmodiophora brassicae , Brassica S allele (self-incompatibility) collections  which comprise Brassica lines with characterized S-allele haplotypes, and other collections of B. oleracea and B. napus fixed diversity sets (homozygous doubled-haploid (DH) or inbred lines). In total, about 74,000 Brassica accessions from various sources have been identified: mostly conserved in Europe (41%) and Asia (41%) as well as a few in the Americas (12%). Brassica oleracea and B. rapa species, which comprise the most important Brassica vegetables, are represented worldwide by about 20,000 (27%) and 18,000 (25%) accessions, respectively . The European Brassica database (Bras-EDB;
www.cgn.wageningen-ur.nl/pgr/collections/brasedb/) contains detailed accession data on 32 collections from 22 European countries.
A total of 412 accessions of wild relatives have also been identified in gene banks (mostly European) including 179 species at the University of Madrid in Spain, and 97 species at the Leibniz-Institut für Pflanzengenetik und Kultur Pflanzenforschung (IPK) in Gatersleben, Germany . However, wild species are still under-represented in most ex situ collections .Information Databases
Brassica databases are another important resource for crop improvement. These comprise freely available online databases which provide genomic and genetic data for important Brassica crops, including genome sequence information, predicted genes and associated annotations, and genetic marker information. In addition, several databases provide cytogenetic and taxonomy data, such as Brassibase (https://brassibase.cos.uni-heidelberg.de/), or species distribution and observation data (usually for specific countries or regions) for Brassica crops and wild relatives growing in the natural environment. In Canada for example, an electronic database provides taxonomy and synonymy information for 338 Brassicaceae genera and 3709 species (14,000 taxonomic names) found distributed across Canada: http://www.cbif.gc.ca/eng/species-bank/?id=1370403266204 . The Brassica database (BRAD (https://brassicadb.org)) has a specific focus on genome annotations and deep mining of the assembled Brassica crop genomes to provide information for breeding and research . Another database, brassica.info, contains links to browsers and downloads for annotated reference genomes of B. napus, B. rapa and B. oleracea as well as Brassica linkage maps and molecular marker collections (www.brassica.info/genome/linkage_maps.html). The Brassica genome databases (http://www.plantgdb.org/BrGDB) mainly focus on genome data dissemination via CropStore and the Brassica Genome Database (BrassicaDB). The Brassica CropStore was initially developed to collate and disseminate information from crop research communities [201,202] as well as provide data information for Brassica phenotypic and genetic maps from different projects [203,204]. CropStore is an integral part of InterStoreDb which provides a platform for the utilization of a set of interlinked databases to assist linking phenotype to QTL regions for a particular trait. Data contained within CropStore can be accessed via a web interface [201,204].
In the era of fast growing technologies such as genome editing, sequencing and biotechnology tools, there is scope to improve the efficient utilization of information and resources provided by gene banks. Future gene banks should also aim to conserve DNA as well as products of genome editing and transgenic approaches, alongside genomic sequence information for plant accessions . If possible, current gene banks should aim to provide genotypic as well as phenotypic information on Brassica species and wild relative collections in the form of an online portal or databases. A number of online Brassica species databases have been in existence since the era of reduced cost genome sequencing: the incorporation of these online databases with traditional germplasm banks would provide breeders and scientists with considerable resources for efficient crop improvement.
In this review we describe the progress that has been made to date in the use of interspecific hybridization for Brassica crop improvement. But what may be possible in future? Recent technological advances in genome sequencing and editing have the potential to revolutionize the use of genetic diversity present in the wild relatives for Brassica crop improvement. Putatively, Brassica wild relatives with useful phenotypic diversity can be identified through screening of diverse populations under different environmental conditions, phenotype data then coupled with genome and resequencing data to link phenotypes to genotypes, followed by gene editing to directly install these genetic variants into the major Brassica crop species. Although this process may still be more speculative than realistic, the technological basis for this approach already exists today. High-throughput phenotyping platforms are available and under constant improvement for glasshouse and field environments [205–207]. In natural environments, traits have also been successfully linked to genetic loci through sequencing of contrasting species populations in different habitats . Whole genome sequencing and resequencing is becoming increasingly cheap and available, with major strides being made in both improving genomic resources available for the Brassica crop genomes [209–212] and in the availability of additional genomic resources for Brassica wild relatives [213,214]. As previously mentioned, genetic transformation and genome editing protocols have already been established for many of the Brassica crop species [172,174,175]. In future, we expect the true value of interspecific hybridization and the use of wild relatives for crop improvement in the agriculturally significant Brassica genus to be realized, with implementation of new technologies supported by gene banks and information resources for breeding and research outcomes.
All authors contributed to conceptualization and writing of the review.
The authors declare that they have no conflicts of interest.
ASM and EK are financially supported by Deutsche Forschungsgemeinschaft (DFG) Emmy Noether grant MA6473/1-1, EIK is supported by DFG grant MA6473/2-1, DQM is supported by DFG Sino-German Centre grant MA6473/3-1, and PVT is supported by a bilateral fellowship between the DAAD and CONICYT.
Katche E, Quezada-Martinez D, Ihien Katche E, Vasquez-Teuber P, Mason AS. Interspecific hybridization for Brassica crop improvement. Crop Breed Genet Genom. 2019;1:e190007. https://doi.org/10.20900/cbgg20190007