Karyotype Reorganization in Wheat–Rye Hybrids Obtained via Unreduced Gametes: Is There a Limit to the Chromosome Number in Triticale?

To date, few data have been accumulated on the contribution of meiotic restitution to the formation of Triticum aestivum hybrid karyotypes. In this study, based on FISH and C-banding, karyotype reorganization was observed in three groups of F5 wheat–rye hybrids 1R(1A) × R. Aberrations, including aneuploidy, telocentrics, and Robertsonian translocations, were detected in all groups. Some of the Group 1 plants and all of the Group 2 plants only had a 4R4R pair (in addition to 1R1R), which was either added or substituted for its homeolog in ABD subgenomes. In about 82% of meiocytes, 4R4R formed bivalents, which indicates its competitiveness. The rest of the Group 1 plants had 2R and 7R chromosomes in addition to 1R1R. Group 3 retained all their rye chromosomes, with a small aneuploidy on the wheat chromosomes. A feature of the meiosis in the Group 3 plants was asynchronous cell division and omission of the second division. Diploid gametes did not form because of the significant disturbances during gametogenesis. As a result, the frequency of occurrence of the formed dyads was negatively correlated (r = −0.73) with the seed sets. Thus, meiotic restitution in the 8n triticale does not contribute to fertility or increased ploidy in subsequent generations.


Introduction
Polyploidy plays a central role in plant genome evolution and in the formation of new species [1,2]. In addition to the ancient process of genome-wide duplication in all seed plants, in most plant species, including cultivated ones, two or more divergent genomes can merge via hybridization in a single nucleus [3]. The high heterozygosity of such allopolyploid species ensures the high genetic diversity of their progeny [4,5].
The formation of polyploids is followed by their passing through a bottleneck of instability [6]. When two parental genomes join to form an allopolyploid genome, a "genomic shock" is experienced [7]. A multitude of evolutionary processes affects polyploid genomes, including rapid and substantial genome reorganization, transgressive gene expression alterations, gene fractionation, gene conversion, genome downsizing, and the sub-and neofunctionalization of duplicate genes [2,5,[8][9][10][11][12][13][14][15]. Thus, new polyploid species, most of which have experienced several cycles of polyploidization [16], end up suffering a massive loss of "redundant" DNA and the restructuring of their chromosomes, as well as a repeated reduction in genome size [17]. Other changes in genomes at the chromosomal level involve duplications, deletions, fissions, fusions, translocations, and inversions of whole chromosomes, chromosome arms, or smaller segments [18].
The Poaceae family includes many typical allopolyploids. It has been found that the bread wheat subgenomes A, B, and D were originally derived from three diploid (2x; haploid (n) chromosome numbers [3,[60][61][62]. Polyploids may arise in one step via the fusion of two unreduced gametes or through a so-called triploid bridge. The triploid bridge mechanism seems to occur more often than the single-step pathway because of the low probability of the fusion of two unreduced gametes in natural populations [63]. Functional gametes in wheat-rye F 1 hybrids are formed via meiotic restitution, in which there is no pairing of chromosomes, and univalents are divided into sister chromatids in the first meiosis after which the meiosis ends [64][65][66]. It was found that meiotic restitution is genetically controlled in wheat-rye hybrids [64][65][66][67][68] and inherited in durum wheat-rye hybrids [54,66,67].
At present, little is known about the possibility of inheriting meiotic restitution, or about the contribution of meiotic restitution to the patterns of karyotype formation and the rate of meiotic stability restoration in common wheat-rye hybrids. Previously, we examined the chromosome sets, structures, and behavior in meiosis of the selected progenies (with good seed-setting ability) of F 2-3 wheat-rye hybrids, obtained using the bread wheat cv. Saratovskaya 29 and the wheat-rye substitution line 1Rv(1A), which determines meiotic restitution [64]. Karyotype analysis of the F 2 Triticum aestivum L. cv. Saratovskaya 29 × Secale cereale L. var. Onochoyskaya (S29 × R) hybrid revealed 56 chromosomes; among them were 42 wheat chromosomes and 14 rye chromosomes [69]. The karyotype of the F 2 1Rv(1A) × R hybrid contained 46 chromosomes, of which three pairs of rye chromosomes 1R1R4R4R2RL2RL, 1R1R replaced the chromosomes 1A1A, and 2RL2RL and 4R4R were added. In the F 3 generation of S29 × R hybrids, the octoploid number of chromosomes with aneuploidy of single rye and wheat chromosomes was preserved, while in the 1Rv(1A) × R hybrids, the number of chromosomes varied from 42 to 49, but in most plants, 2n = 46 was retained. The main meiotic disorders in hybrids F 3 1Rv(1A) × R and S29 × R was the presence of univalents in the first division and micronuclei in the second. Most disturbances are terminated by the fifth generation of allopolyploid hybrids [46,70]. Based on this, in the current work, we examined the chromosome sets, structures, and behaviors in meiosis of three groups of F 5 wheat-rye hybrids. Each group is the progeny of a single F 1 plant obtained by crossing the 1Rv(1A) common wheat disomic substitution line with the rye Secale cereale var. Onochoyskaya. The significant elimination of rye chromosomes was observed in the first two groups. The chromosome numbers in Group 3 varied from 52 to 56. All rye chromosomes were preserved there, but the wheat chromosomes showed insignificant aneuploidy. Our observations suggest that genome reorganization is not finished in any group of F 5 descendants. The meiosis in the hybrids was unstable. Specific features of meiosis in the plants of the third group included asynchronous cell division and the omission of the second division, followed by significant disturbances during mitosis in gametogenesis. In the other two groups, the second division of meiosis took place. Therefore, meiotic restitution in 8n triticale was inherited but did not contribute to increases in ploidy in subsequent generations.
Six seeds were set in the F 1 4-3 plant, and they produced only 2 fertile F 2 plants (6-1 and 6-2), which were taken for further study (Table S1, Figure 1). As the seeds of each plant were sown separately, starting from F 1 , their progeny was designated as lines. In the F 3 of plant 6-1 (subgroup 1a), only 1 plant  was fertile, while 4 high-yield plants were obtained in F 4 . One low-yield plant  and 2 plants with different numbers of grains (23-10 and 23-13) were chosen from the F 3 progeny of plant 6-2 (subgroup 1b). A total of 12 high-yield plants were chosen from F 4 in subgroup 1b ( Figure 1). As a result, the chromosome sets were analyzed in the F 5 plants that originated from 16 F 4 plants. Some of the grains of one plant were sown in the greenhouse; karyotypes were analyzed in vegetative plants using FISH. Other grains from the same plant were transferred for karyotype analysis using C-banding. Five seeds were set in plant F 1 4-7 (Group 2). They yielded 5 plants, and 3 high-yield plants were chosen for further crosses. To obtain generation F 3 , 36 seeds were taken from each plant, and 5 high-yield plants from these 36 were selected to obtain F 4 (Table S1, Figure 2). Finally, chromosome sets were analyzed in F 5 plants originating from 15 F 4 plants via C-banding. One seed was set in plant F 1 73-1 (Table S1, Group 3), and 35 seeds in F 2 (26-1). Generation F 3 plants were grown from these seeds, only 8 of which were fertile ( Figure 3). The 3 plants with the most seeds were selected. In total, 4 plants with the most seeds were chosen from F 4 . Some of the grains of each of the 4 plants were sown in the greenhouse; karyotypes were analyzed in the vegetative plants using FISH. Other grains from the same plants were transferred for karyotype analysis using C-banding.

Routine Meiosis Analysis
To analyze meiotic division, young spikes were fixed in ethyl alcohol-acetic acid 3:1 and stored at 4 • C. Pollen mother cells (PMCs) were stained with and squashed in 3% acetocarmine. All of the anthers with PMCs at metaphase I-anaphase I and anaphase II-telophase II, and with separate microspores, were examined (Table 1). Each anther was analyzed individually, assaying all PMCs in each anther.

Giemsa C-banding
C-banding was carried out as in [72]. The slides were examined under an Amplival microscope (Carl Zeiss Jena). Images were recorded with a LeicaDS300 camera (Leica Microsystems) and processed using the Adobe Photoshop CC2017 software.
All slides were examined under an Axio Imager M1 (Carl Zeiss) microscope. Images were recorded with a ProgRes MF camera (Meta Systems, Jenoptic) and processed using the Adobe Photoshop CS2 software.

Statistical Analysis
Associations between 2 traits (the number of grains and the percentage of micronuclei at telophase II, the number of grains, and the percentage of dyads at telophase II) were determined using the Pearson correlation coefficient (Microsoft Excel program). The significance of the correlation was determined using the Student's t-test and the Chaddock scale (Table 2).  The FISH and C-banding data indicate that the chromosome sets of the F 5 descendants of two sister lines (subgroups 1a and 1b) differ. Plants with 2n = 44 and 2n = 43 (47.05 and 36.76%, respectively) were predominant in subgroup 1a ( Figure 4). The numbers of rye chromosomes varied from three to six, and most sets (79.4%) displayed four rye chromosomes, namely, disomic 1R and 4R ( Figure 5). Chromosome 1R was trisomic or tetrasomic in sets with six rye chromosomes. Wheat chromosomes were mostly present in the disomic state, and chromosomes 3B, 4A, 4D, 5D, and 7D were monosomic in sets one, two, one, one, and one, respectively. The rye chromosome 1R always replaced wheat 1A. Chromosome 4R replaced chromosome 4D in one plant and 4A in two. In other plants, this chromosome was added to the wheat chromosomes.  Subgroup 1b included chromosome sets of F 5 plants obtained from three F 3 plants (Figures 1 and 6). This subgroup was distinguished by the presence of the three rye chromosomes 1R, 2R, and 7R, while 4R was absent. Chromosome 1R was found in the disomic state in all plants, and it replaced wheat chromosome 1A. The chromosomes 2R and 7R were monosomic. This group was also marked by wheat-rye Robertsonian translocations and wheat and rye telocentric chromosomes ( Figure 6). A different pattern was observed in the progeny of plant 23-10. Their chromosome sets were more diverse: a total of 20 different sets were found ( Figure 6b). The most frequent chromosome sets were 39W+1R1R+2R+7R (18.75%), 38W+1R1R+2R2R+T2RL.W (12.5%), and 40W+1R1R+2R+7R (10.41%). The numbers of rye chromosomes varied from two to five. The 40W+1R1R set was found in only one plant. Sets with wheat and rye telocentrics and with Robertsonian translocations were found in 33.33% of plants. The translocated chromosomes had hybrid centromeres because the centromeric repeats pAet6-09 and pAWRc did not overlap ( Figure 8).

Group 2
The chromosome sets of the F 5 plants in Group 2 were relatively uniform. As shown by C-banding, only two rye chromosomes were present in the disomic state: 1R1R and 4R4R (Figures 9 and 10).  The intergenomic substitution 1R(1A) was preserved in the first homeologous group in all sets. A 4R4R pair was added to the whole set of common wheat chromosomes in 58.59% of sets. Alternatively, it replaced one of three wheat chromosomes of the fourth homeologous group in 13.15% of sets. Monosomic was observed in 11 chromosome sets (Figure 10), and disomic substitution in 6: in 1 plant with 4R(4A) substitution and in 5 with 4R(4B) ( Table 3). Chromosome 4B was the most commonly eliminated or rearranged (19 plants, Table 3). Aneuploidy for chromosomes 6A, 5B, and 7D was detected in 11 plants-7, 1, and 3, respectively.  Figure 11). The presence of 16 rye chromosomes owing to chromosome 1R tetrasomy (68.4% of plants) was a specific feature of the chromosome sets ( Figure 11). The 40W+16R chromosome combination was found in 28.07% of the sets. Four sets lacked one pair of 1R chromosomes, and one set lacked one 6R chromosome. Intergenomic substitution 1R(1A) was preserved in homeologous group 1. It was found in all chromosome sets but four. Disomic and monosomic intergenomic substitutions of wheat chromosomes 3R(3A), 6R(6A), 2R(2B), 3R(3D), and 4R(4D) were detected in nine plants, of which six showed substitutions of the chromosomes of homeologous group 3 ( Figure 12). Chromosomes with altered structures were identified by GISH in only six plants-five with a rye telocentric and one with a wheat telocentric.

Chromosome Behavior in Meiosis
Chromosome behavior was studied in plants from Groups 1 and 3. The predominant meiotic aberrations in Group 1 were (i) the formation of univalents and their improper disjunction, leading to the formation of micronuclei, and (ii) cell cycle asynchronization ( Figure 13). Meiocytes with univalents constituted 68.3 ± 2.64% to 100% of subgroup 1a and 60 ± 3.05 to 92 ± 2.16% of subgroup 1b (progeny of plant 23-10), and made up 32.4 ± 1.05% of the progeny of plant 23-8. Univalents were lagged on the metaphase plate at anaphases I and II. The arrest of chromosomes on the metaphase plate at AII caused the formation of micronuclei at the tetrad stage. The counting of micronuclei in tetrads revealed differences within Group 1 and within each of its subgroups. In subgroup 1a, meiocytes with micronuclei constituted 47 ± 3.75 to 70.26 ± 6.2% (Table 4). The asynchronous cell cycle manifested itself as the presence of meiocytes from metaphase I to telophase II inside the same anther ( Figure 13). Such anthers were noted in all plants, but their frequencies varied broadly, from 13.2 to 64%. The frequency of anthers with asynchronous meiocyte division did not correlate with poor seed sets (r = 0.19).
In subgroup 1b, the lowest number of meiocytes with micronuclei was found in the progeny of plant 23-8: 13.93 ± 1.89%. In the progeny of 23-10, the percentage of meiocytes with micronuclei varied from 46.5 ± 2.45 to 65.21 ± 9.4%. Seed sets varied among plants within subgroups, and showed negligible correlations with micronucleus numbers, according to the Chaddock scale (Table 1, Figure 14). The rye chromosomes 1R1R and 4R4R were identified in the chromosome sets of subgroup 1a. To understand the cause of the preservation of chromosomes 4R4R up to generation F 5 , we analyzed the behavior of rye chromosomes at metaphase I and found that chromosomes 1R1R and 4R4R together formed 1.82 ± 0.05 rod and ring bivalents per meiocyte ( Figure 15). Univalent rye chromosomes were detected in 17.8% of cells. Chromatin migration was also detected in meiosis. Cytomixis was identified in MI meiocytes and in pollen grains ( Figure 16). Chromosome disjunction was not disturbed in donor cells (with fewer chromosomes). In contrast, the division machinery was inoperative in the recipient cells, and chromosomes occurred in indistinct clusters. Univalents were found in 77.07 ± 2.01% of the meiocytes in metaphase I in Group 3 plants. They were formed from both wheat and rye chromosomes (Figure 17a,b). Micronuclei were detected in 68.9 ± 3.86% of the microsporocytes in telophase II (Figure 17d). Some meiocytes showed no chromosome pairing at all (Figure 17c). Some cells displayed chromatin breakage and cytomixis ( Figure 18). The migration of chromatin between a tapetum cell and a meiocyte was detected at prophase I, at which point its compaction changed. Asynchronous cell cycles were characteristic of meiosis in hybrids of this group. At the leptotene-zygotene stage, all anther meiocytes corresponded to this state (Figure 19a), whereas meiocytes corresponding to leptotene-zygotene was present in all subsequent meiotic phases, from pachytene to telophase II (Figure 19b-f). The omission of meiotic division II was observed in some plants, and dyads were identified among tetrads (Figure 20a). The percentage of dyads varied from 21.9 to 100 (Figure 20c). Significant aberrations in the mitotic division and chromatin structure were observed during pollen grain formation (Figure 20b). To summarize, the data on meiosis in particular plants and their fertility show that the high frequency of dyad formation is negatively correlated (r = −0.73) with the seed set ( Figure 20c). According to the Chaddock scale, a correlation coefficient r = −0.73 indicates high correlations with marked associations. Dyad formation and seed set were significantly associated (Student's t-test, p ≤ 0.001, df = 15).

Chromosome Instability in F 5 1Rv(1A) × R Hybrids
Meiotic restitution (spontaneous chromosome duplication in gametes, yielding 2n gametes) in F 1 of interspecies and intergeneric hybrids is the means by which new polyploids (allopolyploids) arise in angiosperms [78][79][80]. The earliest wheat-rye hybrids, octoploid triticales, were also obtained by means of spontaneous chromosome duplication, including anthers as well as ovules [47]. Although triticales have a complete chromosome set, almost all newly formed triticales produce some chromosomally variable progeny [47,49,65,81,82]. In this study, the F 1 hybrids from which F 5 were obtained via self-pollination were produced using unreduced gametes. The unreduction occurs as follows: univalents congregate on the metaphase plate and separate into sister chromatids at AI. Then, two daughter nuclei are formed, and meiosis ends after meiosis I [64]. It was expected that the fusion of 2n gametes would give rise to octoploid triticales. However, the analysis of the chromosome sets of the three groups studied revealed their different means of reorganization. A common feature of the chromosome sets of the three groups was the predominant transmission of chromosome 1R, which replaces wheat chromosome 1A. This was predictable, as line 1Rv(1A) was used for hybridization to rye. One-third of Group 3 plants were octoploids (2n = 56) bearing the tetrasome rye chromosome 1R. Other chromosome sets showed aneuploidy for two rye chromosomes and five intergenomic substitutions for wheat chromosomes. In contrast, the chromosome sets in Groups 1 and 2 reverted to the ancestral substitution line 1Rv(1A). They retained 37 to 40 wheat chromosomes and eliminated four to five rye chromosome pairs. The F 2 hybrid 7-4, the ancestor of the F 5 of Group 2, comprised a set of 46 chromosomes, with 40 wheat and 6 rye chromosomes (1R1R4R4R2RL2RL), as shown by C-banding [69]. The long arms 2RL2RL were preserved in the F 3 generation [69] and were eliminated by F 5 . One reason for the absence of rye chromosomes from F 2 might be their elimination in hybrid embryogenesis [69]. Alien chromosome elimination has been reported in crosses of wheat with Secale cereale, species of genus Hordeum, and species more distantly related to wheat, such as maize (Zea mays), pearl millet (Pennisetum glaucum), sorghum (Sorghum bicolor), and Imperata cylindrica [83][84][85][86][87][88][89][90][91][92][93][94][95][96][97]. The preferred elimination of wheat D-genome chromosomes in the first generations after synthetic wheat (BBAADD) × rye (RR) hybridization was also observed, and an F 2 seedling carrying 48 chromosomes was observed [56]. Hexaploid triticales with 28 intact A/B and 14 intact R chromosomes, and with other chromosome constitutions, including monosomic, substitution, and translocation lines, were found in F 5 of these hybrids [54,56]. Another reason for the absence of rye chromosomes from F 2 may be meiotic irregularities in F 1 , which produce laggard chromosomes and aneuploid gametes [98]. Alterations in chromosome disjunction in wheat-rye hybrids may produce gametes with chromosome numbers other than 21 or 28 [65,81]. The analysis of F 2 T. aestivum L. × S. cereale L. indicated chromosome number variability. Plants in one group had the euploid chromosome number 2n = 56, and others had aneuploid numbers 49 to 54 [81]. The same study showed that most of megaspores in F 1 hybrids were aneuploid for one to four chromosomes, wheat or rye. Aneuploid plants were detected among hybrids T. turgidum L. × S. cereale L. using the meiotic restitution pathway: one was monosomic (41 chromosomes) for rye chromosome 3R, two had 42 chromosomes each, one was a nullisomic-1R-tetrasomic-1B heterozygous for deficiency in approximately half of the long arm of chromosome 2B, and one was nullisomic-1R-trisomic-2A-trisomic-1B [65].
In this study, univalents and micronuclei were also formed in the meiosis of F 5 hybrids, regardless of chromosome number and set completeness. Thus, genome reorganization was not completed in any of the three groups of wheat-rye F 5 generation. Little is known about how long an allopolyploid chromosome set can remain unstable for, and how it can affect the allopolyploid evolution. Artificially resynthesized allopolyploids obtained by meiotic restitution mimic the allopolyploidization process. Studies in this field show different advances in the increase in stable allopolyploids in different taxa. An attempt to resynthesize the ancient polyploid Arabidopsis suecica by crossing A. thaliana and A. arenosa produced a viable hybrid, which showed homologous pairing and no important structural reorganization of the homeologous genomes in F 5 [99]. In contrast, chromosomal variation is ubiquitous in newly developed synthetic hexaploid wheat (SHW) created by crossing T. turgidum with A. tauschii [100][101][102]. The common occurrence of univalency during meiotic metaphase I was associated with chromosome instability [102]. Young allopolyploids, termed neopolyploids, are appropriate evolutionary model systems for understanding early allopolyploid formation. Chromosome set instability is exemplified by the natural neoallotetraploids Tragopogon mirus and T. miscellus (about 40 generations). Aneuploids constitute 38 and 69% of these plants, respectively [103].
Cytological instability and aneuploidy in wheat-rye octoploid and hexaploid allopolyploids have presented problems since their creation [47,49,[104][105][106]. The cytological study of triticale demonstrates that the interaction of wheat and rye genomes in the cells of one plant leads to profound derangements in cell physiology, which are maintained for decades at least. Thus, the same irregularities in meiosis and mitosis are noted in the triticale produced by Rimpau in 1889 as are found in triticale derived in later studies, including this one. In spite of the complete chromosome set, univalents are abundant in the meiosis of triticales of different ploidies [47,65,[104][105][106]. In a comprehensive study of this phenomenon, only bivalents were found at diakinesis, but at MI, a pattern was established that can be interpreted as either chromosome lagging or the presence of a univalent. Aneuploid cells may arise in triticale as a result of the asynchronized functioning of rye and wheat chromosomes, and from chromosome lagging at the anaphase and telophase [105]. Chromosome disjunction depends on the proper functioning of the kinetochore [107]. As such, in stable hybrids, the CENH3 produced by one parent must be able to support the functionality of the other parent's centromeres, despite differences in each parent's centromere sequences [83]. Thus, the conservation of chromosome sets of the parental subgenomes in octoploid triticale over generations may be associated with the increased expression of rye centromeric histone CENH3 variants in the new genomic environment [108].

Rye Chromosome 4R Is Preserved until F 5 in 1Rv(1A) × R Hybrids
Unexpectedly, we detected the preservation of rye 4R chromosomes in a monosomic or disomic state up until generation F 5 . While in F 2 , the chromosome pair 4R4R only supplemented the wheat chromosome set [69], in F 5 , 4R4R was added to wheat chromosomes in 58.59% of plants and replaced chromosomes of the fourth homeologous group in 13.15% of plants, which implies its compensational and competitive activity in a new genomic environment. The short arm of chromosome 4R is known to be homeologous to wheat chromosome arms 4BS and 4DS, and partially homeologous to chromosome arm 4AL [109], which is itself involved in evolutionary translocations between chromosome arms 4AL, 5AL, and 7BS [37,39]. On the other hand, it has been shown that rye chromosomes are incorporated into the wheat genome at different frequencies depending on cross direction and genotype [110,111]. This is true for chromosome 4R as well, which can be eliminated at high frequencies from triticale [111,112] or from disomic addition lines [113], but at the same time it can be successfully transmitted in crosses of wheat and octoploid triticale, which results in the 4R addition line [114] and in offspring from the substitution line [115]. The transmission rate of the 4R chromosome pair was consistent at 98% in subsequent generations [114]. In a study summarizing the genetic stability of several wheat-rye disomic addition lines, the frequency of progeny plants being disomic for 4R ranged from 74% to 93% [116]. In our study, the preservation of the 4R chromosome copy in F 5 1Rv(1A) × R likely results from the great similarity to corresponding homeologs in the genomes of wheat, similarly to the preservation of rye genes in allohexaploid triticale with a high similarity to their homeologs in Triticum genomes [117].

Alterations of Centromeric Regions
Deletions and translocations of individual chromosomal regions and chromosome arms are also among the most common chromosomal alterations [57,106]. The chromosome sets of our F 5 hybrids contain rye and wheat telocentrics and Robertsonian translocations. The formation of inter-and intrachromosomal translocations in wheat-rye hybrids cause reductions, eliminations, or expansions in the centromeric retrotransposon sequences, and the formation of multiple centromeres [53,59]. In our experiments, a centromere carrying two nonoverlapping loci, rye-specific pAWRc and pAet06, was identified in a 1RL.1BL Robertsonian translocation. Multicentric chromosomes are frequently formed in hybrids of wheat and related species, such as Th. elongatum, Th. poticum, Th. intermedium, Agropyron cristatum, Hordeum vulgare, and S. cereale [59]. Wheat and Th. elongatum chromosomes with two regions containing centromeric sequences were observed in the F 1 hybrids of null-tetra lines N3AT3B, N5BT5A, N5DT5B, and N6AT6B, and in the hexaploid amphiploid 8802 (AABBEE), which originated from hybrids between T. durum and Th. elongatum [59].

Meiotic Restitution Does Not Increase Ploidy in Progenitors of Octoploid Triticale
The heritability of meiotic restitution has been described in wheat hybrids [54,67,101]. The genes for meiotic restitution in those studies originated from various accessions of durum wheat T. turgidum [65][66][67][68]. Lines of synthetic hexaploid wheat (SHW) were produced by spontaneous chromosome doubling via unreduced gametes resulting from meiotic restitution in T. turgidum × A. tauschii hybrids [66]. These hexaploids also inherited the gene(s) for meiotic restitution, because meiotic restitution also occurs in SHW-rye F 1 hybrids and gives rise to amphiploids or partial amphiploids [66]. Another example is line Do1, which was selected for its capacity to produce self-fertile F 1 hybrids with rye [65,67]. Spontaneous chromosome duplication in androgenic haploids was observed when crossing the F 1 hybrid to hexaploid triticale [67].
Three groups of F 5 1Rv(1A) × R hybrids were obtained via meiotic restitution, whereby chromosome pairing occurs sporadically if at all, univalents segregate into sister chromatids in meiosis I, and the second division is absent [64,118]. We found that the meiosis in some F 5 plants of Group 3 also ended after the first division; therefore, dyads formed after division completion instead of tetrads. Another feature of chromosome behavior was the asynchronization of the meiotic cycle within an anther. Asynchronous cell division was also noted in F 5 plants of Groups 1 and 2, but the second division did occur there. An Arabidopsis thaliana mutant named tardy asynchronous meiosis (tam), with a phenotype of delayed and asynchronous cell divisions during male meiosis, has been described [119]. The genes TAM (also known as CYCA1;2) and OSD1 (omission of second division) are essential for the meiosis I/meiosis II transition. A mutation in CYCA1;2/TAM or OSD1 leads to the premature ending of meiosis after meiosis I, and, as a consequence, to the production of diploid spores and gametes [120,121]. Hence, mutations in such genes as OSD1 and TAM may favor polyploidization, as demonstrated in common wheat. A QTL responsible for the unreduction in T. turgidum × Ae. tauschii hybrids, named QTug.sau-3B, was identified on wheat chromosome 3B [122]. Comparative genomic analysis indicates that QTug.sau-3B is a collinear homolog of cyca1;2/tam, which is known to be responsible for unreduced gamete formation in Arabidopsis thaliana [121].
Whilst dyads in F 1 hybrids 1Rv(1A) × R passed through mitotic divisions and formed functional pollen grains after meiosis, the mitotic division in octoploid F 5 hybrids was greatly disturbed. The disturbances affected chromatin's structure significantly. As a result, functional pollen grains were not formed, and plants either offered few seeds or were totally sterile. Arabidopsis thaliana osd1 mutants showed no somatic developmental defects, male or female gametophyte lethality, or reduced fertility. Only tetraploids and triploids were found in selfed progeny [123]. In plants homozygous for null alleles of CYCA1;2/TAM, the unreduced gametes were functional, giving rise to polyploid progeny [121,124]. The chromosome number in each generation of MiMe plants' selfing (mitosis instead of meiosis, triple osd1/Atrec8/Atspo11-1 mutants) doubled; as such, tetraploids (4n, 20 chromosomes) and octoploids (8n, 40 chromosomes) were obtained [123]. However, the increase in ploidy was accompanied by a seed set decrease. Fertility dropped from 25 ± 6 seeds/fruit in 2n plants and 19 ± 4 in 4n plants to < 0.1 in 8n plants. The causes of this phenomenon remain obscure. In the case of octoploid plants 1Rv(1A) × R, the formation of microspores with 56 chromosomes may induce stress, entailing a collapse in cell cycle regulation and, as a consequence, apoptosis [125]. The Poaceae family includes perennial plants with over 100 chromosomes [126]. They are characterized by low fertility and sterility, probably associated with meiotic anomalies induced by polyploidy. Studies of the genus Arundo L. (Poaceae) have demonstrated that the sterility of A. micrantha (2n = 12x) and A. donax (2n = 18x) is due to the early failure of gametogenesis [127]. In theory, unreduced gametes form during meiosis in these species; however, these gametes have not been proven to cause sterility.
Is there a limit to ploidy in flowering plants? Although the haploid chromosome numbers in 66120 angiosperm species with known chromosome sets vary from n = 2 to n = 320 [126], the chromosome numbers of 80% of angiosperms range from n = 5 to 20, and in 95% the haploid chromosome numbers are less than n = 34. A similar distribution in chromosome numbers is seen in the tribe Triticeae. All the species of this tribe have the basic haploid chromosome number x = 7. About 31% of the species are diploids (or rather paleopolyploids); 1% are triploids; 45%, tetraploids; 17%, hexaploids; 5%, octoploids; 0.2%, decaploids; and 0.2%, dodecaploids. Elymus displays the larger series and highest level of polyploidy, from 2x to 12x [20]. Owing to the cyclic mode of polyploidy, most angiosperm species have less than 14 chromosome pairs, which show no signs of exponential growth [128]. Genome synteny comparisons show that many ancient polyploidization events were followed by striking reductions in chromosome number [16], which in some cases are estimated to have occurred relatively soon after polyploidization [129]. For instance, an n = 7 monocot ancestor underwent four tetraploidy events in the lineage leading to Zea mays; had not it been for fusions, maize would have n = 112, but today it has n = 10 [130]. On the other hand, chromosome sizes cannot rise infinitely after fusion. Chromosome lengths are limited by the sizes of the dividing cells; chromosome arms longer than half the cell length are truncated by the new cell wall, causing damage and gene loss at the ends [131], as proven experimentally with artificial chromosomes in barley [132]. Ploidy increase may also be limited by other factors as well [133]. These include biochemical and energetical expenses, cell size limits, time limitations caused by longer mitosis and meiosis with larger genomes, and difficulties in support of gene expression diversity in giant genomes in response to environmental changes.

Conclusions
In this work, we studied the karyotypes and meiotic behavior of chromosomes in three groups of fifth-generation hybrids (1Rv(1A) × R) obtained via meiotic restitution. Our observations suggest that genome reorganization is not finished in any of the groups of F 5 hybrids. It was found that in two groups of karyotypes, one to three rye chromosomes were preserved in a disomic or monosomic state. The chromosome 4R in 13.15% of plants substituted the chromosomes of the fourth group of wheat genomes, ABD. The karyotypes of the plants of these groups were also characterized by the presence of Robertsonian translocations. The chromosome sets of Group 3 were near octoploid, varying from 52 to 56. The presence of 16 rye chromosomes owing to chromosome 1R tetrasomy (68.4% of plants) was a distinctive feature.
Meiosis in the hybrids was unstable. Univalents in the first division were found, characterized by a violation of segregation, which led to the formation of micronuclei in microspores. However, according to the correlation analysis, no connection was found between the presence of micronuclei and seed sets. The analysis of meiosis in Group 3 revealed asynchronous cell division and omission of the second division. Diploid gametes did not form because of significant disturbances during mitosis in gametogenesis. As a result, the frequency of the formed dyads was negatively correlated (r = −0.73) with the seed sets. Thus, the trait "meiotic restitution" is inherited in octoploid triticale; however, gametogenesis does not take place in dyads, and functional gametes are not formed.