Festuca pratensis-like Subgenome Reassembly from a “Chromosomal Cocktail” in the Intergeneric Festulolium (Poaceae) Hybrid: A Rare Chromoanagenesis Event in Grasses

Festuca and Lolium grass species are used for Festulolium hybrid variety production where they display trait complementarities. However, at the genome level, they show antagonisms and a broad scale of rearrangements. A rare case of an unstable hybrid, a donor plant manifesting pronounced variability of its clonal parts, was discovered in the F2 group of 682 plants of Lolium multiflorum × Festuca arundinacea (2n = 6x = 42). Five phenotypically distinct clonal plants were determined to be diploids, having only 14 chromosomes out of the 42 in the donor. GISH defined the diploids as having the basic genome from F. pratensis (2n = 2x = 14), one of the progenitors of F. arundinacea (2n = 6x = 42), with minor components from L. multiflorum and another subgenome, F. glaucescens. The 45S rDNA position on two chromosomes also corresponded to the variant of F. pratensis in the F. arundinacea parent. In the highly unbalanced donor genome, F. pratensis was the least represented, but the most involved in numerous recombinant chromosomes. Specifically, FISH highlighted 45S rDNA-containing clusters involved in the formation of unusual chromosomal associations in the donor plant, suggesting their active role in karyotype realignment. The results of this study show that F. pratensis chromosomes have a particular fundamental drive for restructuring, which prompts the disassembly/reassembly processes. The finding of F. pratensis “escaping” and rebuilding itself from the chaotic “chromosomal cocktail” of the donor plant points to a rare chromoanagenesis event and extends the view of plant genome plasticity.


Introduction
In the plant kingdom, cross-hybridization between species is a rather common event, which in many instances results in the establishment of new allopolyploid taxa [1][2][3]. Polyploids are also frequent in domesticated and economically important crops [4][5][6]. It has been widely demonstrated that the newly formed allopolyploids are more than just the sum of their parents and that they have gone through complex processes of genome adjustment. Hybridization challenges the integrity of genomes and leads to a cascade of structural changes, and these responses are species-and genotype-dependent [7][8][9][10][11][12][13]. Studies in the lineages of nascent allopolyploids show numerous rearrangements described as biased fractionation [4], subgenome dominance [14,15], karyotype reshuffling [16,17], and genomic mosaicism [18].
In plants, new individuals can arise not only in a sexual way (through meiosis, gametes, fertilization), but somatic clones (through mitosis) are also readily available in vitro and in vivo. Grasses have a capacity for tillering; new tillers emerge at basal nodes, a small zone between the root and the stem. During their life, plants, and distant hybrids especially, accumulate somatic irregularities at the level of cell populations in their body, which further leads to the development of chimeric clones genetically differing from the donor plant [19,20]. Saccharum spp. and Phalaris spp. are classical examples of clonally propagated species where ploidy changes and chromosome structure alterations have been widely recorded [21,22].
In Poaceae, Lolium and Festuca genera are considered to have diverged from the common ancestor that had a basic chromosome number of x = 7 [23,24]. Several Lolium and Festuca species are very closely related. They hybridize easily, and the crosses between L. multiflorum and diploid F. pratensis, or hexaploid F. arundinacea, demonstrate particularly high compatibility, providing the germplasm for the development of commercial grass varieties [25][26][27][28]. F. arundinacea (2n = 6x = 42) is a natural allohexaploid species originating as a hybrid between F. pratensis and F. glaucescens, structurally defined as FpF-pFgFgFgFg [29]. In addition, L. multiflorum was found to be the third species involved in the development of F. arundinacea by its introgression within the F. pratensis subgenome [30]. Moreover, this study has shown that there is much closer homology between L. multiflorum and F. pratensis genomes than between L. multiflorum and F. glaucescens. In grass breeding, L. multiflorum and F. arundinacea show very high crossability and provide a particularly valuable combination of complementary traits [28,31,32].
In this study, we follow up on the story of rediploidization in the next F2C1 progeny, concentrating on a particular "super-recombinant" L. multiflorum × F. arundinacea (2n = 6x = 42) hexaploid genotype characterized as a "chromosomal cocktail". We reveal the changes and trace them from plant morphology going deeper into the karyotype/genotype level. Once again, as in F1, we record diploid plants in F2 as a result of the split allohexaploid genome in clonal plants and show these "escapees" to be close to the F. pratensis subgenome of F. arundinacea.
Summarizing the results, we discuss a possible scenario of the re-establishment of diploid segregants occurring throughout the hybrid genome response that could reach the status of unpredictable chromosome shattering and reassembly known under the term chromoanagenesis. Chromoanagenesis was first defined by Holland and Cleveland [37] from the "cellular cataclysm" studies observed in human cancer research. Chromoanagenesis (from chromo for chromosomes and anagenesis for "rebirth") is a phenomenon in which large numbers of complex rearrangements occur involving one or a few chromosomes in a single catastrophic event [37][38][39]. So far, there are only a few instances of chromoanagenesis described in plants, in grapevine [40] and Populus [41]. The results of this study add to the conceptual ideas on how dramatically genomes can rearrange themselves when fused together in intergeneric hybrids.

Plant Material
The pedigree and experimental chronology are shown in Figure 1. To produce F1 hybrids, the anthers were removed from diploid L. multiflorum (2n = 2x = 14) plants, var. 'Prima Roskilde' and var. 'S.22'. The inflorescences were bagged and pollinated with the pollen of hexaploid F. arundinacea var. 'Barundi' genotype no. 870. The hybrids required embryo rescue; they were isolated after 16-18 days and grown on LS medium. The germinated embryos were treated with 0.3% colchicine solution, and F1C0 hybrids were obtained [36]. One hundred sixty-seven F1C0 fertile plants were planted in an isolation plot to allow for cross-pollination and seed setting. From the bulk of seeds obtained, a group of 682 F2C1 plants was grown in the experimental field of The Lithuanian Institute of Agriculture. Plants were grown at a distance of 50 × 50 cm and evaluated visually for their

Phosphoglucoisomerase-2 Isozyme Assessment
The preparation of the buffers, starch gel, PGI-2 detection, and phenotype assessment were performed according to Humphreys [31]. In brief, from each plant, about 0.2 g of young fresh leaves were crushed in a cooled mortar with 0.1 M TRIS, pH 7.2 extraction buffer. The filter paper wicks were soaked in the extract, the samples were placed on the starch gel (Sigma) for electrophoresis, and the gels were treated with the staining solution for PGI-2 detection and analysis.

Chromosome Preparation
The tillers collected from the plants were shortened, old roots were removed and set up in tap water in glass tubes for about 5-7 days, then fresh roots were collected and placed in ice-cold water for 24 h, followed by fixation in 1:3 acetic acid-ethanol. The roots were softened in a mixture of 0.1% pectolyase Y-23 and 0.1% cellulase R-10 and squashed in 45% acetic acid. Mitotic chromosome spreads were fixed on glass slides by freezing at −80 • C, coverslips removed and used fresh or stored at −20 • C. For the chromosome counts, ≥3 metaphases with a complete chromosome spread were analyzed in each individual. For GISH karyotyping, only the cells (3)(4)(5) with clear signals throughout the chromosome spreads were analyzed, and the most representative ones were chosen to be included in the Results and Supplementary data.

Fluorescent Probes
L. multiflorum, F. pratensis and F. glaucescens genomic DNA was sonicated for 5 min in an ultrasonic bath, checked on 1% agarose gel, and the DNA samples, sheared to 500-2500 bp fraction, were chosen for labelling. DNA was labelled with rhodamine-11-dUTP (Roche) or fluorescein-12-dUTP (Roche) by nick translation, using Nick Translation Mix kit (Roche). L. multiflorum rhodamine labelled probe was used specifically to highlight simultaneously both L. multiflorum (red) and F. pratensis (mauve), as described by Pašakinskienė et al. [30]. The pTA71 plasmid containing wheat 45S ribosomal DNA repeats was cleaved with the EcoRI restriction enzyme to release the insert and labelled with fluorescein-12-dUTP by nick translation.

In Situ Hybridization
The slides were soaked in 45% acetic acid for 5 min at room temperature, and at 48-50 • C for 3 min. Denaturation of genomic DNA was performed at 70 • C in 70% deionized formamide in 2xSSC for 2 min, followed by dehydration with cold ethanol series (70%, 90% and 100%), 2 min each, and air-drying. A 25 µL volume of hybridization mix made of about 100 ng of DNA probe, 60% formamide, 25% dextran sulfate, 10% 20xSSC, and 5% SDS was denatured at 70 • C for 10 min, then placed on slides and set up for incubation at 37 • C for 16 h in a moist chamber. After hybridization, the slides were washed in 20% formamide in 0.1xSSC twice at 42 • C for 5 min, and three times in 2xSSC at 42 • C for 3 min. The slides were mounted with Vectashield antifade and DAPI (4,6-diamidino-2-phenylindole) for DNA counterstaining. For reprobing with F. glaucescens genomic DNA and 45S rDNA probe, the L. multiflorum probe was washed off by soaking the slides in 4xSSC with Tween 20, the coverslips were removed, and the slides were washed three times in 2xSSC.

Somatic Tiller Chimerism in F2 3-18 Allohexaploid and Re-Establishment of Diploid Clonal Plants
A peculiar "super-recombinant" genotype F2 3-18 was detected among 682 individuals from the cross L. multiflorum × F. arundinacea in F2C1 plant group screened morphologically. This genotype was multiplied clonally and showed a phenomenal effect of chimerism and emerging variable tiller phenotype (growth habit, leaf shape) different from that of the donor (Figures 1 and 2a). Plants N1, N2, N4, and N5 had a typical inflorescence (panicle) characteristic of F. pratensis, while plant N3 never produced inflorescences. This particular instance of visually observed, distinctly different clonal segregants (Figure 2a) is the first case described in the Festuca × Lolium hybrid literature.

Somatic Tiller Chimerism in F2 3-18 Allohexaploid and Re-Establishment of Diploid Clonal Plants
A peculiar "super-recombinant" genotype F2 3-18 was detected among 682 individuals from the cross L. multiflorum × F. arundinacea in F2C1 plant group screened morphologically. This genotype was multiplied clonally and showed a phenomenal effect of chimerism and emerging variable tiller phenotype (growth habit, leaf shape) different from that of the donor (Figures 1 and 2a). Plants N1, N2, N4, and N5 had a typical inflorescence (panicle) characteristic of F. pratensis, while plant N3 never produced inflorescences. This particular instance of visually observed, distinctly different clonal segregants ( Figure 2a) is the first case described in the Festuca × Lolium hybrid literature.
The somatic chromosome number of the F2 3-18 donor was determined as hexaploid 2n = 6x = 42 in the root meristematic cells cytologically examined. All five unusual plants, N1-N5, phenotypically contrasting with the donor plant, were determined as diploids (2n = 2x = 14), whereas the rest of the 22 clonal plants, phenotypically identical to the donor, remained hexaploid (2n = 6x = 42). Hyperploid nuclei, the split spindles' anaphases and cells of different sizes were observed in the F2 3-18 donor (Figure 2c,d), which could presumably lead to heterogeneous cell populations in meristematic tissues.  The somatic chromosome number of the F2 3-18 donor was determined as hexaploid 2n = 6x = 42 in the root meristematic cells cytologically examined. All five unusual plants, N1-N5, phenotypically contrasting with the donor plant, were determined as diploids (2n = 2x = 14), whereas the rest of the 22 clonal plants, phenotypically identical to the donor, remained hexaploid (2n = 6x = 42). Hyperploid nuclei, the split spindles' anaphases and cells of different sizes were observed in the F2 3-18 donor (Figure 2c,d), which could presumably lead to heterogeneous cell populations in meristematic tissues.
Considering that the PGI-2 isozyme isoform profile is a unique and constant feature for a given plant, the F2 3-18 hybrid and clonal plants were analyzed for the PGI-2 phenotype. The PGI-2 profile in the F2 3-18 hybrid was found to be very complex, including all possible allelic isoforms, a + , a, b, c and d, whereas only b and c isoforms were clearly detected in the diploid segregants (Figure 2b). This pattern in the diploid segregants confirmed them as descendants of F2 3-18 and highlighted a tendency for a regular genome split. The distribution of the a + isozyme fraction was not retained in the clonal diploids as it was in the original donor plant. Presumably, a ++ instead of a + was present in diploids, which explains the bands below b as the hybrid bands a ++ /c and a ++ /b. More importantly, the complexity of the pattern in four (N1-N4) out of five diploid segregants clearly shows them not being "plain" diploids since a maximum of two PGI-2 alleles are possible for diploids, e.g., b and c in a heterozygote. In other words, it is not possible to have three alleles in a diploid unless there is a third PGI-2 locus involved from the multiple gene set present in the donor hybrid. Taking into account the PGI-2 allele specificity throughout the parental species [32,42], these results point out that two components representing F. glaucescens had a different fate in the processes of rediploidization, namely a + appeared by its modified a ++ version, whereas the genome fraction carrying allele d was the one that was always left out.
Meiosis in the F3 3-18 plant was extremely irregular, with the multivalents of variable sizes and shapes prevalent in MI (Supplement Figure S1). Micronuclei, cytomixis and chromatin "bridges" were common in tetrads (Supplement Figure S1). Nevertheless, even though the meiosis view was not directly related to the discovery of somatic diploid descendants, it added data to nuclear content disruptions characteristic of the F2 3-18 donor plant.

GISH/FISH of Diploid Clonal Plants
F. arundinacea is an allohexaploid comprising two Festuca subgenomes, diploid F. pratensis and tetraploid F. glaucescens, defined as FpFpFgFgFgFg and having L. multiflorum introgressions within F. pratensis chromosomes [29,30]. GISH images of initial F1 and colchicine doubled hybrids are shown in the pedigree scheme ( Figure 1). GISH analysis using an L. multiflorum genomic probe that simultaneously highlights F. pratensis and L. multiflorum chromosomes and chromosomal segments has shown diploid clonal plants to be new genomic variants. The karyotype of these newly split diploids was preferentially based on the F. pratensis subgenome, with L. multiflorum found in small variable-sized blocks (Figure 3a-c, Table 1). F. glaucescens segments were found as rare tiny insertions (image data not included). All of these F. pratensis-like diploids had two 45S rDNA sites at the interstitial position near the centromere on two chromosomes (Figure 3a-c, Table 1), which is a distinct feature of F. pratensis subgenome in F. arundinacea (Figure 3d).
Thus, GISH revealed the genome composition of the diploids and presented them as F. pratensis-like subgenome segregants established throughout somatic tiller chimerism. From these results, we can conclude that on the basis of genome irregularities that occurred in the particular somatically unstable allohexaploid L. multiflorum × F. arundinacea (2n = 6x = 42) F2 3-18 plant (one out of 682 in the F2 plant group), the diploid F. pratensis (2n = 2x = 14) subgenome of F. arundinacea was able to recompose and stabilize itself with minor footprints of other genomic partners, L. multiflorum and F. glaucescens.     Figure 4 for two representative mitotic chromosome spreads probed in two ways: (i) the total genomic DNA of L. multiflorum, which highlights both L. multiflorum (red) and F. pratensis (mauve) simultaneously (Figure 4a,c); (ii) the combined probe of total F. glaucescens DNA and pTa71 probe for 45S rDNA cluster (Figure 4b,d). In the F2 3-18 karyotype, L. multiflorum chromosomes and chromosomal parts were found to be the most prevalent (Figure 4a,c), followed by F. glaucescens (Figure 4b,d) and F. pratensis (Figure 4a,c), summarized in Table 2. About 1 / 4 of the chromosomes were species-recombinant.

GISH/FISH of F2 3-18 Super-Recombinant Donor
GISH/FISH was carried out to reveal genomic features of the donor L. multiflorum × F. arundinacea hybrid F2 3-18. The results are shown in Figure 4 for two representative mitotic chromosome spreads probed in two ways: (i) the total genomic DNA of L. multiflorum, which highlights both L. multiflorum (red) and F. pratensis (mauve) simultaneously (Figure 4a,c); (ii) the combined probe of total F. glaucescens DNA and pTa71 probe for 45S rDNA cluster (Figure 4b,d). In the F2 3-18 karyotype, L. multiflorum chromosomes and chromosomal parts were found to be the most prevalent (Figure 4a,c), followed by F. glaucescens (Figure 4b,d) and F. pratensis (Figure 4a,c), summarized in Table 2. About ¼ of the chromosomes were species-recombinant.   Table 2). GISH revealed a specific representation of F. pratensis in this hybrid karyotype, which appears clearly different from the representations of the other two species, L. multiflorum and F. glaucescens. Namely, F. pratensis chromosomal components were the least represented, but the most involved in chromosomal rearrangements (Figure 4a-d, Table 3). Table 2. Genomic complexity of super-recombinant F2 3-18 L. multiflorum × F. arundinacea revealed by GISH for its true parental and recombinant chromosomal components (structural analysis from Figure 4).

A1 A2
Lm chromosomes and chromosomal parts 20 24 Ratio of broken   Figure 4a-d, and in Supplement Figure S2.
The 45S rDNA clusters were abundant in the karyotype of the F2 3-18 plant. There were two different patterns of metaphases observed in the donor plant: (i) in general, chromosomes are intact, and there are few or no breakages (Figures 4a,b and S2); (ii) chromosome breakages are abundant, predominantly at 45S rDNA positions of L. multiflorum chromosomes, and detached pieces of 45s DNA are present in and along the metaphase plate (Figures 4c,d, S2 and S3). In contrast, 45S rDNA sites at the terminal position of two F. glaucescens chromosomes appeared to be stable (Figure 4b,d). Thus, a specific pronounced fragility at 45S rDNA positions in the F2 3-18 donor plant was observed with numerous freely floating detached 45S rDNA pieces in some metaphase plates (Figures 4b,d, S2 and S3). This was a distinctive feature found in F23-18 metaphases, whereas no such patterns of scattered 45S rDNA components were observed in diploid metaphases (Figure 2a-c). When irregular-shaped enlarged recombinant chromosomes and sticking chromosome structures were analyzed (Figure 4o-r), it was found that some of them had a 45S rDNA block positioned at the junction point (Figure 4p). In addition, rDNA-enriched sites were seen to be involved in some sticking branched chromosome structures (Figure 4r).
Regarding visible F. pratensis chromatin disruptions, several F. pratensis chromatin protrusions at the sites of fusing chromosomal configurations (Figures 4q and S2) and specific breakages at the position of F. pratensis insertion within L. multiflorum chromosomes were detected (Figure 4a,c).
The results of FISH/GISH analysis in this study show that a particularly irregular L. multiflorum × F. arundinacea F2 3-18 genotype is prone to complex chromosomal aberrations in the somatic tissues: abundance of recombinant chromosomes, chromosome fragility and stickiness, 45S rDNAs containing detached structures, and their involvement in chromosome fusing configurations (Figures 4, S2 and S3). This suggests that recombinant chromosomes, with F. pratensis predominately involved, could be the source of vulnerable sites and could lead to extensive irregularities within the chromosome set of proliferating tissues in the plant body, which can finally turn out in split clonal plants of different ploidy, namely diploids (Figure 3).
The reference data on the pronounced capability of F. pratensis for chromosome restructuring observed in Festulolium hybrids are summarized in Table 4. A hypothetical model of a diploid F. pratensis-like subgenome reassembly de novo is presented in Figure 5.  Reassembly of the F. pratensis subgenome was discovered from a "chromosomal cocktail" represented in a hexaploid donor plant; five "novel diploids" were found out of 27 clonal descendants. Reassembly of the F. pratensis subgenome was discovered from a "chromosomal cocktail" represented in a hexaploid donor plant; five "novel diploids" were found out of 27 clonal descendants.

Discussion
Structural genomic changes in allopolyploids depending on the progenitor genomes were shown across different plant taxa by using GISH/FISH cytology and by sequence data assessment [3,11,16,46]. The breakage-fusion-bridge cycles may recompose the genome content in the somatic cells of the hybrids and account for the establishment of novel tiller phenotypes [47,48].
In the plethora of experimental plant hybrids, Lolium and Festuca species have been a very useful source to demonstrate how parental genomes adjust and rearrange themselves in intergeneric hybrids. In nearly three decades of FISH cytology studies, a wide spectrum of intergenomic adjustments has been determined in Festulolium hybrids. Lolium chromosomes dominate over Festuca; a high portion of chromosomes are species-recombinant; recombination events are significantly higher for F. pratensis-like chromosomes than for those of Lolium; 45S rRNA sites show high fragility; atypical interstitial positions of telomeric repeats appear in the hybrids [13,17,[25][26][27]43,45,49]. In addition, a rare and more unusual case of instant rediploidization of the F. pratensis subgenome was described in F1C0 of the L. multiflorum × F. arundinacea (2n = 8x = 56) octoploid plant group [36].
Here, we follow up on the story of somatic chimerism and present a particularly complex case of genomic turbulence in the hexaploid genome of F2C1 L. multiflorum × F. arundinacea hybrid F2 3-18 (2n = 6x = 42), comprising L. multiflorum genome and two subgenomes of F. arundinacea, F. pratensis and F. glaucescens. In the first instance, this plant attracted attention for the distinct chimerism of its clonal parts, differing from the donor by their tiller phenotype (growth habit, leaf shape), and, cytologically, diploids. Furthermore, the screen for of PGI-2 isozyme phenotype clearly confirmed diploids to be descendants of the donor plant, with a regular tendency of a certain allele set retained in these somatic segregants. Cytological GISH/FISH analysis of the F2 3-18 donor genome presented us with a broad colorful canvas of a "chromosomal cocktail". The recombinant chromosomes made up about 25% of the karyotype, and promiscuous exchange patterns between three genomes, L. multiflorum, F. pratensis, and F. glaucescens, were observed. Hence, it was rather unexpected to discover somatically segregating plants featuring a diploid (2n = 2x = 14) karyotype of a fairly defined genomic structure from the body of a donor hexaploid (2n = 6x = 42) plant representing such a chaotic "chromosomal cocktail". Our findings show that all of these diploid clonal descendants have a few things in common: (i) the basic genome comes from F. pratensis progenitor subgenome with minor L. multiflorum, and, to a lesser extent, F. glaucescens components; (ii) they have two 45S rDNA sites positioned interstitially near the centromere, which corresponds to the variant of F. pratensis subgenome in F. arundinacea.
Admittedly, precise ways of the formation of F. pratensis-like diploids cannot be followed in time and space, and GISH/FISH analysis results are rather controversial. The split diploids are based on an F. pratensis-like genome, a minor component of the hybrid, whereas L. multiflorum and F. glaucescens chromosomes make up most of the donor plant genome. Therefore, some intriguing questions arise following a broader context of the genome conflict and adjustments in Festulolium hybrids. Why is it specifically the F. pratensis genome that splits out, "escapes" and re-establishes as diploid throughout vegetative tiller chimerism? An important point may be the involvement of 45S rDNA arrays and rDNAbearing chromosomes [44,45]. Interestingly, the earlier studies by Sokolov et al. [50] on Crepis capillaris (Asteraceae) experimentally demonstrated that the presence of at least one NOR (45S rDNA) site-bearing chromosome in the karyotype is sufficient for the cell to be functional for replication. It is likely that due to genomic disturbance in the highly irregular karyotype of the L. multiflorum × F. arundinacea hybrid, the re-establishment and stabilization of the F. pratensis subgenome carrying two 45S rDNA sites [30] seems an easier task to be achieved than a Lolium-like variant with the abundant, 5-7, site number [51,52]. Likewise, it is quite possible that Lolium spp. chromosomes have an advantage and become dominant in Lolium x Festuca hybrids [25][26][27]43] exactly because of the abundance of their 45S rDNA clusters.
The results of this study together with observations of other research groups on Festulolium support the involvement of 45S rDNA loci in the karyotype rearrangements [17,26,45,53], (Table 4). More supportive data come from studies on other plant species [54][55][56][57]. Notably, similarly to our findings showing the involvement of rDNA-enriched clusters in the formation of sticking chromosome structures, Gernand et al. [58] recorded patterns in end-to-end and branching chromosomal associations in Allium fistulosum (Allioideae, Amaryllidaceae) triggered by tissue culture. Furthermore, the occurrence of detached rDNAs similar to ours was detected in Tragopogon species and their intergenic hybrids, suggesting that rDNAs display an active role in genome restructuring [59]. Whereas specifically, the intergenic spacer (IGS) within the 45S rDNA arrays was FISH labelled and traced as dynamic "carrier" of genomic blocks throughout the evolutionary changes of the Senna tora (Fabaceae) karyotype [60]. On the other hand, from the massive sequencing data, Long et al. [61] concluded that the 45S rDNA sequence amount is greatly variable, and in some Arabidopsis thaliana lines (out of 180 tested), this has resulted in a substantial increase in genome size. Thus, plant cytogenetics and DNA data studies show that rDNA arrays appear as nomadic genomic blocks capable of proliferation and may form some kind of intermediate self-organized structural elements in the process of genome rearrangements.
Even though GISH provides us with a highly informative picture of L. multiflorum × F. arundinacea donor plant genome composition and reveals the complexity of the chromosomal structures, the enigma of F. pratensis-like subgenome "escaping" from this "chromosomal cocktail" and rebuilding itself remains mysterious. Theoretically, there must have been short-lived moments of massive chromosomal breakages and reassembly that have little chance of being detected visually in the cells of a growing plant.
Recently, in the story of the plant genome restructuring through gross rearrangements of its structural building blocks, a new and striking phenomenon of chromosomal shattering and reassembly, for a long time only known in human cancer genetics [37,62,63], comes into light. In plants, the first cases have come from observations in woody plants where the major chromosome disruption and reassembly occurred spontaneously in clonal descendants or as a result of mutagenesis.
In Tempranillo Tinto grapevine, Carbonell-Bejerano et al. [40] applied the whole genome scale approach and discovered catastrophically unbalanced genome rearrangements caused by chromosomal shattering in a spontaneously arisen somatic clone showing the loss of berry color, named Tempranillo Blanco. This random reassembly involving the structural reshuffling of three chromosomes was defined as a case of chromoanagenesis (chromosome rebirth). Further on, short-read sequencing in Populus after gamma radiation mutagenesis demonstrated an instance of shattered chromosomes in two out of 592 individuals [41]. From these Populus lines, exhibiting extreme chromosomal restructuring, stable clones were produced and maintained vegetatively for several years. This shows that massive chromosome shattering and reassembly events can be tolerated with no severe effect on life. It is likely that such changes, although rare, may occur in the somatic tissues of long-lived plants, trees or perennial grasses, and may result in novel phenotypes bringing some advantages with the new set of traits.
Linking together the scientific outcomes of grapevine and Populus [40,41] and the patterns of genome restructuring in Lolium × Festuca hybrids [13,17,26,27,44,45,49], conclusions can be drawn about the evident differences between the species and a particular genotype in their ability for abrupt or/and massive chromosomal rearrangements. Meanwhile, it seems that F. pratensis chromosomes have a particular drive for chromosome restructuring [26,36,44,45]. There is something fundamental about the F. pratensis chromosome structure that prompts the disassembly/assembly process. Theoretically, besides rDNA arrays, the telomeric/centromeric repeats, and transposons involved, there could be some 3D properties of the chromatin that are still to be discovered.
In this study, a peculiar instance of F. pratensis subgenome reassembly and establishment was recorded as an outcome of the somatic tiller chimerism of an L. multiflorum × F. arundinacea F2C1 hybrid bearing a highly irregular karyotype. Though this phenomenon seems very rare, a similarly striking effect was noticed by Liv Østrem (pers. comm.) in another hybrid combination where a complete diploid F. pratensis (2n = 2x = 14) surprisingly re-emerged in the cross L. perenne × (L. perenne × Festulolium [var. Prior]) (2n = 4x = 28). Remarkably, there was no full set of F. pratensis involved in the first place, as it is known to be represented by only 3.5 (out of 14) chromosomes in the plants of var. Prior [43]. This observation and our results showing F. pratensis rebuilding itself from the minor component of the hybrid suggest the involvement of structural resources from the bulk of two/threespecies recombinant chromosomes. At the moment, it is difficult to imagine how exactly such a scenario could be governed and accomplished in the cell. Our results show that 45S rDNA-containing structures are likely to be dynamic intermediates involved in de novo F. pratensis-like chromosome rebuilding. Notably, rDNA resources were highly abundant both in the donor F2 3-18 karyotype and in the cross-combination L. perenne × (L. perenne × Festulolium [var. Prior]). Therefore, the abundance of rDNA seems to be a key condition for chromoanagenesis of F. pratensis-like karyotype from the complex "chromosomal cocktail" in the hybrid.
The extreme plasticity of the genomes, occurring through chromosome shattering and chromoanagenesis, and its significance in macroevolution has been well reviewed by Pellestor and co-authors [38,39], pointing towards the genome-centric (as opposed and complementary to gene-centric) approach in building up the full picture of speciation processes in eukaryotes. In plants, some ambiguous aberrant chromosomal structures detected by FISH in A. fistulosum tissue culture-derived plants [58] and extrachromosomal circular DNAs maintained in the herbicide-resistant weed Amaranthus palmeri (Amaranthaceae) [64] extend our knowledge of surprising ways of genome self-reorganizing responses and point to avenues of new discoveries. Very recently, Majka and co-authors [65] provided new insights into the mechanism of perpetual pushing around and competition within the Lolium × Festuca hybrid genome with an in-depth study of parental chromosome behavior at the meiosis level.
Since the work of Barbara McClintock [66], there are things still to be discovered about jumping genes and recombining genome components in a non-Mendelian way. This study demonstrates an instance of a mysterious reassembly ("rebirth") of the F. pratensis-like genome, which adds to the growing data proving the existence of sudden genomic alterations beyond genes. This comes in line with the intriguing chromoanagenesis phenomena and highlights the role of gross abrupt genome changes in the evolutionary processes of plant taxa [67]. In practical terms, such spontaneous genome-scale somatic variability may provide a new and unexpected set of traits in herbaceous and woody plants.
Funding: This research received no external funding. Data Availability Statement: All of the data are included in the main text.