Recreating Stable Brachypodium hybridum Allotetraploids by Uniting the Divergent Genomes of B. distachyon and B. stacei

Brachypodium hybridum (2n = 30) is a natural allopolyploid with highly divergent sub-genomes derived from two extant diploid species, B. distachyon (2n = 10) and B. stacei (2n = 20) that differ in chromosome evolution and number. We created synthetic B. hybridum allotetraploids by hybridizing various lines of B. distachyon and B. stacei. The initial amphihaploid F1 interspecific hybrids were obtained at low frequencies when B. distachyon was used as the maternal parent (0.15% or 0.245% depending on the line used) and were sterile. No hybrids were obtained from reciprocal crosses or when autotetraploids of the parental species were crossed. Colchicine treatment was used to double the genome of the F1 amphihaploid lines leading to allotetraploids. The genome-doubled F1 plants produced a few S1 (first selfed generation) seeds after self-pollination. S1 plants from one parental combination (Bd3-1×Bsta5) were fertile and gave rise to further generations whereas those of another parental combination (Bd21×ABR114) were sterile, illustrating the importance of the parental lineages crossed. The synthetic allotetraploids were stable and resembled the natural B. hybridum at the phenotypic, cytogenetic and genomic levels. The successful creation of synthetic B. hybridum offers the possibility to study changes in genome structure and regulation at the earliest stages of allopolyploid formation in comparison with the parental species and natural B. hybridum.


Introduction
of unreduced gametes, which are produced at low frequencies in diploid species, or the hybridization between two different autotetraploid species. By contrast, the two-step model requires the formation of amphihaploid interspecific hybrids from reduced gametes of two different species followed by chromosome doubling [38,39]. Synthesizing allotetraploids by applying the two-step model has been successfully reported for various species from different genera [15,[40][41][42][43]. Allotetraploids have also been synthesized by a variant of the one step method where the genomes of the parental species were artificially doubled by colchicine treatment to form autotetraploids. These autotetraploids, each producing 2n gametes, were then hybridized to obtain allotetraploids [40,[44][45][46].
The aim of this study is to further develop a Brachypodium polyploid model system by synthesizing Brachypodium allotetraploids, through hybridization between B. distachyon and B. stacei (Fig 1), and characterizing their stability at the genomic, phenotypic and cytogenetic levels in comparison to the parental species and the natural B. hybridum.
After removal of the lemmas and paleas, seeds were surface sterilized using a 10% bleach solution containing a drop of Tween-20 for three minutes. The seeds were then rinsed in sterile water three times. Germination was synchronized by incubating the seeds in Petri dishes at 4˚C for 3 days, and then at room temperature for five days. The seedlings were transferred into pots (10×7 cm) containing a mixture of equal volumes of peat moss and sand supplemented with a slow release fertilizer (2 g/L, Osmocote1 Standard 14-14-14, Scotts-Sierra Horticulture, Marysville, OH, USA). Greenhouse conditions were set at day temperature of 22˚C and night temperature of 18˚C, with a 16 h photoperiod.
Vegetative propagation F1 interspecific hybrids, the colchicine-treated F1 (S0), and the S1 allotetraploids were vegetatively propagated to create large numbers of plants from these sterile or nearly sterile lines (S1 Fig). Root development from secondary tillers was stimulated by covering the base of the plants with soil and adding solution 0.25% of the auxin indole-3-butyric Acid (IBA) [48] to the irrigation solution. Two to three weeks later, secondary tillers with roots were removed, cut and placed in new pots (S1C and S1D Fig). were used (Fig 1). While hybridizing diploid lines from the two species would create amphihaploid (n) F1 hybrids (Fig 1B), crossing the autotetraploid lines would lead directly to (2n) hybrids, i.e. alloteraploids (Fig 1C).
Reciprocal crosses between diploid B. distachyon (lines Bd21 and Bd3-1) and B. stacei (lines ABR114 and Bsta5) were performed over four years (2011)(2012)(2013)(2014) in the spring and fall seasons, the best seasons for flowering and pollination in our greenhouses. Flowering time was variable among lines to be crossed. Thus, in order to ensure simultaneous flowering, multiple sowings were done for each of the lines at 15 day intervals (from January to March).
Emasculation and pollination methods were adapted from Steinwand and Vogel (http://jgi. doe.gov/our-science/science-programs/plant-genomics/brachypodium/). Emasculations were accomplished by removal of the indehiscent anthers from the female parent plants on the two or three basal florets of the spikelet in the morning (10:00 am to midday, Fig 2A-2C). The emasculated flowers were bagged (NatureflexTM 70x130mm bags) to avoid contamination by non-selected pollen. Pollen from the selected paternal parent was collected from the male parent in the afternoon of the same day or one day later by placing nearly ripe anthers on a glass slide for 5-10 min. Most ripe anthers became more turgid and some of them dehisced on the slide ( Fig 2D). Pollen grains were transferred to the emasculated flowers ( Fig 2E) and the pollinated inflorescences were bagged ( Fig 2F) to avoid pollination by stray pollen. Seed formation was recorded 5 to 6 days after pollination The putative F1 amphihaploid interspecific hybrid seeds were collected at maturity (at least 4 weeks after pollination). The seeds were kept at 4˚C for three weeks and then at room temperature for two months. They were germinated to produce plants as described above. True interspecific hybrid plants were validated through cytological analysis and PCR markers.
All F1 amphihaploid interspecific hybrid plants were vegetatively propagated as described above (S1 Fig and S1 Table). The propagated plants were split into two groups. One group was grown in a greenhouse without colchicine treatment as control to see if spontaneous chromosome doubling would restore fertility and lead to the production of allotetraploid seeds. The second group was treated with colchicine to induce chromosome doubling.

Colchicine treatment of F1 interspecific hybrids
We treated plants with colchicine using a protocol adapted from the method described by Jahier [49], which was successfully used for wheat [43].
Vegetatively-propagated plants at 4-5 leaf stage were completely immersed for three hours in an aqueous solution of colchicine (Sigma-Aldrich Co., cat. no. C9754), at concentrations of 2.5 g/l, 5g/l or 7.5 g/l, and containing 2% DMSO (dimethyl sulfoxide, Sigma-Aldrich Co., cat. no. D8418). The colchicine-treated plants were then transplanted into fresh soil, without rinsing, and grown in a greenhouse. By seven to ten days after treatment surviving plants were producing new growth. Necrotic lesions observed on the treated leaves suggested that the treatment was effective.

Estimation of pollen abundance and viability
To estimate pollen viability, anthers were sampled the day of anthesis and pollen was stained with acetocarmine as described in [49]. Mature anthers were macerated in a drop of acetocarmine to release the pollen grains and pollen viability was estimated based on the amount of stain taken up under light microscope. Viable pollen grains appear dark purple because they take up acetocarmine whereas non-viable pollen grains do not take up acetocarmine and appear light.

DNA marker development and analysis
Genomic DNA was extracted from young leaves sampled as described previously [58].
Two types of polymorphic markers were used, simple sequence repeats (SSRs) and gene sequence-derived markers.
All SSR and gene-derived markers were tested to determine if they were polymorphic between the B. distachyon and B. stacei lines used in this study. In addition, they were also tested to make sure they amplify both the B. distachyon and the B. stacei sized bands on DNA from B. hybridum (lines ABR113 and Bhyb30) or a mixture of equal amount of B. distachyon and B. stacei DNA (Bd21 and ABR114; Bd3-1 and Bsta5). For the fertile allotetraploid allo3-1×5, the S1 plant and 118 plants from S2 generation were analyzed.
A marker was considered rearranged in a synthetic allopolyploid plant if its PCR amplification pattern was different from that observed in the mixture of parental DNA and/or sister allopolyploid plants from the same generation [43].

Phenotypic analysis
Fifteen morphological characters were measured and compared between synthetic allotetraploids, B. distachyon, B. stacei and natural B. hybridum (S2 Fig and S3 Table). Three inflorescence traits that could impact seed production were recorded: number of spikelets per inflorescence, number of florets per spikelet and number of florets per inflorescence. For synthetic polyploids with low fertility, we also recorded percent of fertile florets, seed number per inflorescence and 1,000 seed weight. Five additional inflorescence characters were also measured: inflorescence length (total length and length without awns), spikelet length (total length of spikelet excluding awns and averaging all spikelet lengths per each inflorescence), the distance between two spikelets on the inflorescence (the average of all distances in one inflorescence), upper glume length and upper glume width. Four floral characters were measured, floret length, lemma length, lemma width from the basal floret, and awn length (the longest within the spikelet) (S2 Fig). At least five plants per genotype were analyzed as replicates. Statistical analysis was done using non-parametric Kruskal-wallis test [63].

Results
We used two approaches to synthesize allotetraploids from B. distachyon and B. stacei. The first approach was to cross diploid B. distachyon and B. stacei to produce an amphihaploid F1, followed by colchicine treatment to double the chromosomes. The second approach was to first produce B. distachyon and B. stacei autotetraploid plants and then cross them. Since autotetraploids should have 2n gametes, the expected F1 progeny would be allotetraploid without need of further chromosome doubling.
A total of 9,388 crosses between the two diploid species were performed over a four year period and 122 mature seeds were obtained (Table 1). Among these, 68 were obtained from 4,587 crosses where B. distachyon was the maternal parent and 54 from 4,801 crosses where B. stacei was the female parent (Table 1). Only 38 of the 122 mature seeds (31%) germinated and produced viable plants. In comparison, the germination rates of B. distachyon, B. stacei and natural B. hybridum were usually around 96%.
To determine which of the 38 putative F1 plants were true hybrids, we first used codominant SSR markers that differentiate B. distachyon and B. stacei (see below). This analysis identified six bona-fide F1 interspecific hybrids, four arising from the 2,664 crosses between B. distachyon Bd21 and B. stacei ABR114 (designated hereafter as F1_21×114) and two from the 817 crosses between B. distachyon Bd3-1 and B. stacei Bsta5 (designated hereafter as F1_3-1×5). The final success rate for these crosses was 0.15% and 0.245%, respectively. Interestingly, we failed to obtain any true F1 interspecific hybrids from the 4,801 crosses where B. stacei was the female partner (all four genotype combinations) as well as from crosses between the two other genotype combinations where B. distachyon was the female partner (Table 1).
One F1_3-1×5 hybrid plant died before flowering. The five remaining F1 hybrids were vegetatively-propagated (S1 Fig) and separated into two batches (S1 Table). The first batch of 99 plants were grown without colchicine treatment to test if spontaneous chromosome doubling would occur and lead to fertile sectors as has been observed in other systems [43,64]. The second batch of 226 plants was treated with colchicine to induce chromosome doubling and fertility.
Overall, the F1_21×114 and F1_3-1×5 amphihaploid F1 plants were phenotypically similar to natural B. hybridum and intermediate between B. distachyon and B. stacei (e.g. inflorescence architecture and flag leaf morphology) (Fig 3A and 3B). However, for some phenotypic traits the F1s more closely resembled one of the parents. Floret hairiness and floret shape of the F1 amphihaploid hybrids were more similar to B. distachyon than to B. stacei, whereas the number of stamens and stigma structure were more similar to B. stacei (Fig 3C-3E).
Amphihaploid F1 interspecific hybrids were sterile. F1 interspecific amphihaploid hybrids are normally sterile, presumably because of defective chromosome pairing at meiosis [65][66][67]. However, in several cases amphihaploid interspecific hybrids have been reported to produce seeds, most likely by spontaneous genome doubling prior to flowering [43,64]. We tested this possibility with 99 vegetatively propagated plants from the five different F1 interspecific hybrids. Over a period of 2 years no seeds were produced (S1 Table). Each individual plant produced about 20-30 tillers, with two to three inflorescences per tiller and an average of 33 florets per inflorescence. Thus, about 1,320 to 2,970 florets were checked for each individual plant and a total of approximately 128,040 to 288,090 florets for all F1 plants combined.
Chromosome doubling of amphihaploid F1 plants and generation of allopolyploids. One-hundred-fifty-three vegetatively propagated plants from the four original F1_21×114 interspecific hybrids and one of the F1_3-1×5 interspecific hybrids were treated with colchicine to induce chromosome doubling. The majority of plants treated with 2.5 g/l and 5 g/l colchicine solution survived (74% and 87% survival, respectively). By contrast, only 23% of the plants treated with 7.5 g/l colchicine solution survived (S1 Table). We compared FCM profiles from leaves of colchicine-treated F1 interspecific hybrid plants with those of the non-treated plants and of the parental lines. The results revealed the expected average c-value of~0.6 pg for the F1 interspecific hybrids, which is similar to the c-values of B. distachyon and B. stacei. The positions of G1 and G2 peaks in the F1 hybrids were also similar to their counterparts in the diploid parental species (Fig 4A-4C). The 24 colchicine-treated F1 interspecific hybrid plants (23 plants vegetatively-multiplied from the F1_21×114 initial plant and one from the initial F1_3-1×5 plant) showed G1 and G2 peaks at similar positions to those of the natural allotetraploid B. hybridum, indicating that the genomes of these colchicine-treated hybrids have been partially or completely doubled (Fig 4D and 4E; Table 2).
Metaphase chromosome counting of root-tip cells revealed that six plants derived from F1_21×114 and one plant derived from F1_3.1×5 had a chromosome number of 30 in all cells examined which is consistent with whole genome duplication. Four plants derived from F1_21×114 had variable chromosome number in different cells indicating that these plants were a mosaic of cells with doubled and non-doubled genomes ( Table 2). The seven colchicine-treated F1 plants with 30 chromosomes in all cells observed were considered to be zero-selfed (S0) generation of the allotetraploid allo21×114 derived from F1_21×114 (six plants), and of the allotetraploid allo3-1×5, derived from F1_3-1×5 (one plant). These S0 plants were maintained by vegetative propagation. Only two S1 (selfed generation subsequent to S0) seeds were obtained from more than 200,000 flowers from 153 S0 allo21×114 plants, whereas one S1 seed was obtained from the single S0 plant of allo3-1×5. This indicated an overall low fertility in the first generation of the synthetic allotetraploids. Only one of the two S1 seeds of allo21×114 and the single S1 seed of allo3-1×5 germinated and  produced mature S1 plants. The S1 plants were also vegetatively propagated to produce 161 allo21×114 S1 and 48 allo3-1×5 S1 plants.
All S1 plants were taller and more vigorous than the F1 amphihaploid hybrids and the parental species. The two synthetic allotetraploids exhibited similar morphology during the early stages of leaf development and tillering (as defined by [68] (Fig 5A). However, they showed differences in stem elongation. Allo21×114 stems tended to have longer and more internodes, leading to taller plants, than allo3-1×5 (Fig 5B-5D). Inflorescence architecture was similar for both lines with long inflorescences and three to five spikelets (Fig 5E). Florets of both lines were also similar. They had long hairy lemmas, three stamens and feathery stigmas (Fig 5F-5G). Floral characteristics of the F1 hybrids, S1 allotetraploids, the parental lines and natural B. hybridum, are summarized in S4 Table. Genome size of the S1 allotetraploid plants was assessed by FCM and the positions of their G1 and G2 peaks (Fig 4D) were similar to those observed in natural B. hybridum (Fig 4C). S1 plants of allo3-1×5 were fertile with 23% of florets producing seed. While this is much greater than the S0 plants, it is lower than wild B. hybridum lines ABR113 (91%) and Bhyb30 (68%) ( Table S3). More than 100 S2 seeds of allo3-1×5 were sown and almost all of them germinated and produced plants for further cytogenetic and genomic characterization. Surprisingly, all 135 vegetatively propagated S1 allo21×114 plants were sterile.
To examine the female fertility of S1 allo21×114 plants we pollinated emasculated S1 allo21×114 flowers (as well as F1_21×114) with pollen from the two diploid parents and two natural B. hybridum lines but no seeds were obtained (S5 Table). By comparison, 10 seeds were obtained from 35 crosses between the two natural B. hybridum lines.
Our results suggest that S1 allo21×114 plants may be both male and female sterile.

Crossing autotetraploid B. distachyon and B. stacei
We performed 4,384 reciprocal crosses between autotetraploids from two lines of B. stacei (ABR114 and Bsta5) and two lines of B. distachyon (Bd21 and Bd3-1) in four genotype combinations and obtained only 48 seeds (Table 3). However, only 11 germinated and none of those were true interspecific hybrids (Table 3).

Phenotypic characterization of synthetic allotetraploids
Fifteen morphological characteristics of inflorescences, flowers and seeds were measured and compared between S1 generation allotetraploids, B. distachyon, B. stacei and B. hybridum ( Fig  6; S3 Table). In general, the synthetic allotetraploids were more similar to natural B. hybridum and usually exceeded the parental species. More comparisons for each of the individual traits are detailed in S1 Text.

Karyotype characterization
Metaphase chromosomal analysis was done in F1 interspecific amphihaploid hybrids and in S1 and S2 generation plants of the synthetic allotetraploid (Fig 7; Table 2). The comparative analysis showed the expected 10 large chromosomes in B. distachyon (Fig 7A), 20 small chromosomes in B. stacei ( Fig 7B) and 30 (large and small) chromosomes in the natural allotetraploid B. hybridum (Fig 7C). Amphihaploid F1 interspecific hybrids contained 15 chromosomes, five derived from B. distachyon and 10 from B. stacei, (Fig 7D). As expected, chromosomes were duplicated in the two derived S1, S2 plants of the synthetic allopolyploid allo3-1×5 (Fig 7E and 7F) that had similar karyotypes to those of natural B. hybridum. FISH with the 45S rDNA probe labelled the expected number of spots in all lines: two spots in B. distachyon, B. stacei and their F1 amphihaploid hybrids (Fig 7A, 7B and 7D) and four spots in the S1 and S2 allotetraploid plants (Fig 7E and 7F) and natural B. hybridum (Fig 7C). Genome-specific chromosome discrimination with the BAC ABR1-63-E6 probe demonstrated the presence of five chromosomes from B. distachyon in the amphihaploid F1 interspecific hybrids ( Fig 7D) and a doubled number (10) in their derived S1 and S2 synthetic allotetraploids (Fig 7E and 7F) and in the natural B. hybridum (Fig 7C). We analyzed karyotype of 53 cells from 10 different S2 allo3-1×5 plants.All karyotypes had the expected number of 30 chromosomes indicating that chromosomes are stably inherited ( Table 2).

Genetic characterization of synthetic allopolyploids
SSR-and gene-derived PCR markers were used to characterize the genetic stability of synthetic allotetraploids. The single allo3-1×5 S1 plant was fertile, allowing us to examine 118 individual S2 progeny, whereas only a single sterile allo21×114 S1 plant was analyzed. The genetic markers were classified based on the polymorphism observed between parental species, their pooled  DNAs, and the natural allotetraploid B. hybridum using the momenclature recommended in [43]. The polymorphism pattern observed in the synthetic allotetraploids was identical to natural B. hybridum and was the same for all S2 individuals examined, as illustrated in Fig 8 and described in more details in S2 Text A total of 151 markers (129 gene-based and 22 SSR markers) and 140 markers (123 genebased and 17 SSR markers) were analyzed for allo21×114 and allo3-1×5, respectively. The amplification patterns observed for F1 interspecific amphihaploid hybrids, F1_21×114 and F1_3-1×5, were the same as those observed for their derived S0 and S1 synthetic allopolyploids (Table 4).
No evidence of DNA rearrangements in the F1 interspecific amphihaploid hybrids or the derived synthetic S0 and S1 plants was found (Fig 8; Table 4). Similarly, none of the 118 analyzed S2 allo3-1×5 plants showed rearrangements of parental alleles (Fig 8; Table 4), indicating genomic stability.
The genetic patterns in the synthetic allotetraploids were almost identical to the genetic profiles observed in natural B. hybridum ABR113 and Bhyb30 lines (S2 Table). However, consistent with the parental genotypes, we observed slightly more differences with line Bhyb30 than with line ABR113.

Discussion
The origin and evolutionary relationships of the natural B. hybridum allotetraploid in relation to its progenitor species B. distachyon and B. stacei is now clearly elucidated [26,[32][33][34]69]. The creation of synthetic allotetraploid plants similar to natural B.hybridum provides empirical evidence and establishes the tractable Brachypodium allotetraploid model. This represents a unique allopolyploid model where one parental genome (B. distachyon) has similar genome size to the other one (B. stacei), but half the sporophytic (2n) chromosome number (2n = 10 and 2n = 20, respectively) whereas its individual chromosome size is approximately two times larger.
Moreover, the B. hybridum-type allotetraploids synthesized here appear highly stable from the earliest generations (S1 and S2) as characterized at the phenotypic, cytogenetic and genetic levels. The prominent differences in chromosome number and chromosome size of the two parental genomes likely serve as a barrier to homoeologous pairing which may contribute to the chromosomal stability of both natural and synthetic B. hybridum. We will investigate this further by studying meiosis and chromosome pairing in these model allotetraploids.
Evidence suggests that B. hybridum allotetraploids formed naturally more than once with both B. distachyon and B. stacei as the maternal parent [34]. In our experiments, all surviving  Table for further details. There were no seeds from S1 plants of allo21×114, therefore seed number per inflorescence and percent of fertile florets were not scored. doi:10.1371/journal.pone.0167171.g006 Successful Synthesis of Stable Brachypodium hybridum synthetic allotetraploids had B. distachyon as the maternal parent and no allopolyploids were obtained from reciprocal crosses. Moreover, the success rate of interspecific hybridization in the present study was very low. This is similar to other species where parental genotypes have been shown to have a large effect on the success of interspecific hybridization [72][73][74].
It has been suggested that a combination of factors, including differences in flowering time, pollinator behavior and floral structure, caused by both biological and genetic factors, limit hybridization between distantly-related species [75]. Even when pollination occurs, post-pollination barriers, such as differences in style structure and the arrest of the pollen tube growth, can inhibit the formation of zygotes between different species. These can be overcome by refining crossing methods [76]. As an example, in lilies (Lilium candidum L.) the pollen tubes arrest halfway down the style after interspecific pollination, a barrier that can be overcome by in vitro methods [77]. Post-pollination barriers have also been reported in species, such as Rhinanthus and Nicotiana [78,79], as the pollen tube developement at different rate in hetero-specific style or because of differences in pistil length between the crossed species.
In the present study, by performing a high number of interspecific crosses, we obtained viable F1 interspecific hybrids in only two out of four genotype combinations. In comparison, intraspecific hybridizations between divergent lines of each of the two parental species (including the ones used in this study) could be realized relatively easily with a high success rate (http://jgi.doe.gov/our-science/science-programs/plant-genomics/brachypodium/). Previously, other groups have failed to obtain F1 interspecific amphihaploid hybrids between other Brachypodium species [70]. Significantly, the successful combination of lines Bsta5 of B. stacei and Bd3-1 of B. distachyon was not tried in previous attemps (G. Linc and R. Hasterok, unpublished). This illustrates the importance of the parental genotypes and the need to conduct a very large number of crosses. Further characterization of the germination of the pollen on the stigma papilla as well as the progression of the pollen tubes in the style may elucidate the barriers limiting zygote and interspecific hybrid formation between B. distachyon and B. stacei.
Interspecific F1 amphihaploid hybrids are normally sterile because the parental chromosomes do not pair normally during meiosis leading to unbalanced non-viable gametes [80,81]. Doubling the genome of F1 amphihaploid plants often restores fertility and occasionally this occurs spontaneously as has been observed in a variety of plant species such as in wheat [43], Arabidopsis [64,[82][83][84], and rice [85]. In our study, no seeds were obtained from thousands of amphihaploid F1 interspecific hybrid flowers, indicating that restoration by spontaneous genome doubling does not occur or is exceedingly rare for the crosses we made. This would  stacei (ABR114 and Bsta5). In this case, while the parents have different sized bands, the B. distachyon allele is not amplified in MPVs (red suggest that B. hybridum may have been formed naturally by hybridization between unreduced gametes that are rarely produced by the diploid parents. Nevertheless, crosses between the autotetraploid lines of B. distachyon and B. stacei, that produce 2n gametes, were not also successful. A possible reason for the failure of the autotetraploid crosses is the low fertility of the B. stacei and B. distachyon autotetraploid lines (46% and 82%, respectively) and reduced pollen viability (data not shown). We were able to artificially double the genome of our two amphihaploid F1 hybrid plants leading to low fertility. Interestingly, for one cross, the next selfed generation (S1) was even more fertile. It will be very interesting to explore the changes responsible for such increasing fertility. Conversely, fertility did not increased in the S1 generation of the other allopolyploid and this contrast may provide mechanistic insight. Whilst reasons of arrows), natural B. hybridum or the F1, S1 and S2 generations. Absence of the B. distachyon allele is most likely due to competition for PCR amplification between progenitor alleles and not from DNA rearrangements. L: 50  sterility of allo21×114 in comparison to the fertile allo3-1×5 allotetraploid need to be investigated, these findings suggests the existence of pre-established genetic or structural fertility barriers that influence hybridization success and stability of allopolyploid genomes, as observed for hexaploid wheat [43].
In conclusion, the successful synthesis of allotetraploids similar to the natural B. hybridum provides a powerful new tool to an emerging polyploid model system. When combined with the experimental resources and experimental tractability of B. distachyon, B. stacei and B. hybridum, the ability to create allotetraploids opens up exciting possibilities to study various aspects of polyploidy in grasses at genomic, cytomolecular, epigenetic and physiological levels from the very earliest stages of their formation.   characters of B. distachyon, B. stacei, B. hybridum, interspecific F1 hybrids and plants of S1 generation of the synthetic allopolyploids allo21×114 and allo3-1×5.