﻿Intraspecific divergence of diploid grass Aegilopscomosa is associated with structural chromosome changes

﻿Abstract Aegilopscomosa Smith in Sibthorp et Smith, 1806 is diploid grass with MM genome constitution occurring mainly in Greece. Two morphologically distinct subspecies – Ae.c.comosa Chennaveeraiah, 1960 and Ae.c.heldreichii (Holzmann ex Boissier) Eig, 1929 are discriminated within Ae.comosa, however, genetic and karyotypic bases of their divergence are not fully understood. We used Fluorescence in situ hybridization (FISH) with repetitive DNA probes and electrophoretic analysis of gliadins to characterize the genome and karyotype of Ae.comosa to assess the level of their genetic diversity and uncover mechanisms leading to radiation of subspecies. We show that two subspecies differ in size and morphology of chromosomes 3M and 6M, which can be due to reciprocal translocation. Subspecies also differ in the amount and distribution of microsatellite and satellite DNA sequences, the number and position of minor NORs, especially on 3M and 6M, and gliadin spectra mainly in the a-zone. Frequent occurrence of hybrids can be caused by open pollination, which, along with genetic heterogeneity of accessions and, probably, the lack of geographic or genetic barrier between the subspecies, may contribute to extremely broad intraspecific variation of GAAn and gliadin patterns in Ae.comosa, which are usually not observed in endemic plant species.

In earlier classifications C-banded Ae. comosa chromosomes were arranged in a decreasing length and designated with capital letters A -G Georgiou et al. 1992). First genetic nomenclature of the M-genome chromosomes was developed by B. Friebe et al. (1996) based on similarities of their C-banding patterns with homoeologous chromosomes of other Aegilops species. This system was later proved by chromosome sorting and single-gene FISH (Molnár et al. 2016;Said et al. 2021). Some controversy remained in classifying chromosomes 2M and 5M, which are indistinguishable by flow sorting due to similar morphology and same DNA content. Even more discrepancies in chromosome designations exist with the nomenclature developed by C. Liu et al. (2019) on the basis of analysis of addition and substitution wheat-Ae. comosa lines using FISH and PLUG markers. Thus, chromosome classification of this species still needs verification.
Among a broad range of botanical, cytogenetic, biochemical and molecular markers employed for evaluating intraspecific and interspecific diversity of wild and cultivated plant species, seed storage proteins (gliadins) appear to be relatively cheap, but informative markers for polymorphism analysis. Gliadins (Gli) belong to protein fraction prolamines, which is characterized by high glutamine and proline amino acid content and by specific molecular structure (size, domen composition, biochemical properties) (Shewry and Halford 2002). Electrophoretic spectra of gliadins allow discrimination, with high effectiveness, of lines, cultivars and varieties of tetraploid and hexaploid wheat; gliadin profiles are used to assess samples' heterogeneity and to evaluate phylogenetic relationships between species and accessions (Metakovsky et al. 2018). So far, polymorphism analyses of Aegilops based on gliadin loci were mainly focused on Ae. tauschii Cosson, 1850, the D-genome donor of common wheat (Yan et al. 2003;Dudnikov 2018;Badaeva et al. 2019a), and only few publications dealt with other Aegilops species (Cole et al. 1981;Medouri et al. 2015; Garg et al. 2016). On the other hand, owing to similar structure of gliadin loci in wheat and Aegilops (Dong et al. 2016;Huo et al. 2018) and extremely high polymorphism, gliadins can serve as supplementary markers for the assessing genetic variability of Ae. comosa and clarifying phylogenetic relationships between the subspecies.
Aegilops comosa possesses a number of agronomically valuable traits such as pest and disease resistance (Riley et al. 1968;Gill et al. 1985;Boguslavsky and Golik 2003) and salt tolerance (Xu et al. 1996), which can potentially be used for wheat improvement. In late 60 th of XX R. Riley, V. Chapman, and R.O.Y.Johnson introduced resistance to yellow rust from Ae. comosa into wheat cultivar "Compare" by genetically induced homoeologous recombination (Riley et al. 1968). Several wheat-Ae. comosa amphiploid and introgression lines have been developed in China and UK (Weng 1995;Liu et al. 2019;Zuo et al. 2020); some of these lines showed good resistance to yellow rust and powdery mildew (Liu et al. 2019). However, modifications of the M-genome chromosomes over the course of evolution, in particular, species-specific translocations, prevent the direct utilization of Ae. comosa gene pool in wheat breeding (Nasuda et al. 1998). Manipulations with genetic material of Ae. comosa require a deeper understanding of the genome of this species, its chromosomal structure and the range of polymorphism.
The aim of our study was a comparative analysis of Ae. comosa subsp. comosa and subsp. heldreichii on a broad sample of accessions of different geographical origins using FISH with fifteen DNA probes and electrophoretic analysis of seed storage proteins (gliadins) in order to characterize polymorphism and reveal mechanisms leading to divergence of subspecies.

Material
Thirty-six accessions of Ae. comosa including 20 accessions of comosa and 16 accessions of heldreichii collected from different regions of Greece and Turkey ( Fig. 1) and maintained in genetic collections of N. I. Vavilov All-Russian Institute of Plant Genetic Resources (VIR), St.-Petersburg, Russia, and Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Gatersleben, Germany were used in our study (Suppl. material 9). Accessions of comosa and heldreichii show clear differences in spike morphology (Fig. 2).
Thus, spikes of subsp. comosa plants are slender, narrowly cylindrical, tapering toward apex, with 3-4 fertile and 0-2 rudimentary spikelets. Glumes of lateral spikelets have one tooth and one short awn, the apical spikelet has three well-developed awns, the central one of 4-11 cm long and lateral -2.5-3.5 cm long. Spikes of subsp. heldreichii plants are shorter and stouter, not or hardly tapering toward the apex, with one rudimentary and 1-3 fertile spikelets. Lateral spikelets are urceolate, the apical one is obconical. Glumes are ovate, the lateral glumes with broadly triangular tooth on abaxial site and short awn on adaxial side. Apex of apical glume extends into three 3-3.5 cm-long cetulose awns. Lateral awns are shorter and more slender, often reduced to teeth or even absent (Van Slageren 1994). Accessions of both subspecies significantly vary in spike length and color (Fig. 2), and accession K-3809 (subsp. comosa) is characterized by longest spike with black color.

FISH analysis
The seeds are germinated on moist filter paper in Petri dishes at 24 °C. The seedlings with ~0.5 cm roots are transferred into 1.25 mM solution of hydroxyurea for 18 h, washed thoroughly with distilled water and grown in Petri dishes with distilled water for 5 h, as described in . The roots are cut and pretreated in ice water for 24 h and then fixed in the solution of ethanol : glacial acetic acid (3 : 1). Fixed roots are kept in fixative solution at -20 °C until use.
Metaphase cells are prepared by squashing, coverslips are removed after freezing in liquid nitrogen, and slides are kept in 96% ethanol at -20 °C. Fluorescence in situ hybridization is carried out according to previously published protocol . The slides are examined on a Zeiss Imager D1 epifluorescent microscope. Metaphase plates are captured with a 100 x objective using black and white digital camera Axiocam HRm using a software AxioVision, release 4.8. The images are processed using Adobe Photoshop, version 7.0.

Electrophoretic analysis of seed storage proteins (gliadins)
Electrophoresis (EP) in polyacrylamide gel (PAAG) according to the previously published protocol (Metakovsky and Novoselskaya 1991) was employed to obtain gliadin spectra of 26 accessions of Ae. comosa. Since no information on the composition and inheritance of the blocks of gliadin components in this species was available from literature, electrophoretic spectra of all accessions must be compared with an etalon sample with the known genetic control of components. In wheat and related species the gliadin spectrum of bread wheat cultivar Bezostaya-1 serves as an etalon (Novoselskaya-Dragovich et al. 2018).
The gliadin spectra of the Triticeae are traditionally divided into four zones, α, β, γ and ω-zones, depending on electrophoretic mobility of individual polypeptides (Woychik et al. 1961). Peptides from the ω-zone are coded by genes located on group 1 chromosomes, while those from the α-zone -by genes of group 6 chromosomes. Components from β and γ zones are controlled by chromosomes of both genetic groups (Dong et al. 2016;Huo et al. 2018). Based on this information we presume that electrophoretic components from the α-zone of the spectra of Ae. comosa accessions are encoded by 6M, whereas from the ω-zone, by 1M chromosome.

Results
Intraspecific diversity of Ae. comosa in karyotype structure Two subspecies (comosa and heldreichii) of Ae. comosa have similar karyotype structures, which include metacentric, submetacentric and subacrocentric chromosomes (Fig. 3 a-c). Despite overall karyotypic similarity, we observe variation in the number and morphology of satellite (SAT) chromosomes (Fig. 3d). Most accessions have two pairs of satellite chromosomes differing in morphology (Fig. 3a, b), but few genotypes carry only one SAT pair (Fig. 3c). The satellite on one pair is always small, and this chromosome is classified as 1M. The satellite on the second pair -6M, is much larger and appears on physically longer arm. Comparison of the SAT chromosomes allows to divide Ae. comosa accessions into four groups (Fig. 3d). Group I is characterized by long submetacentric chromosome 1M with a small SAT attached to the short arm. The SAT on 6M is very large and occupies nearly half of the arm length. Most accessions from this group belong to subsp. heldreichii, and this combination of satellite chromosomes is designated "heldreichiilike". Group II differs from Group I in shorter satellite length (approximately 1/3 to 1/4 of the arm) on the chromosome 6M. Accessions from this group belong predominantly to subsp. comosa and we designate this combination of satellite chromosomes as "comosa-like". Groups III and IV include representatives of both subspecies. Group III shows altered morphology of both SAT chromosome pairs, whereas Group IV contains just one pair of SAT chromosomes -1M (K-3308, K-3309, K-4873; Fig. 3d). Group IV includes one of the three analyzed genotypes of AE 1254, one of the two genotypes of each K-3824, K-3920 (all comosa), and all K-3806 genotypes (heldreichii). All these genotypes are characterized by heteromorphic pair of 6M chromosomes: one homolog carries large, while the second -much smaller satellite on the long arms ( Fig.  3, "heterozygote"), this group presumably represents hybrids between the subspecies.
As mentioned above, these accessions carry only one pair of SAT chromosomes; thus, inactivation of NORs on 6M is associated with/ or caused by elimination of the 45S rRNA gene sequences from the respective loci.
The pTa71 probe produces distinct signals in subtelomeric regions of short arms of 2M, 3M, and 5M chromosomes (minor NORs) in all accessions of both comosa and heldreichii, however, these loci are not visualized by o-18S probe. The latter ribosomal probe however reveals weak signals in pericentromeric region of 1ML, distal region of 2ML and 4MS, and two minor loci in the proximal half of 7M short arm; all these sites are common for both subspecies. By contrast, several sites discriminate comosa from heldreichii (Fig. 6, Suppl. material 1: fig. S1). These are: (1) the locus of variable intensity in a proximal part of chromosome 3M is observed in nearly all (with one exception) comosa genotypes, but not in heldreichii; (2) most comosa carry two minor pericentromeric o-18S sites in the opposite arms of chromosome 1M, while heldreichi possesses signal only in the long arm; (3) minor pericentromeric rDNA site on chromosome 4M is usually located in the long arm of comosa, but in the short arm of 4M h of heldreichii; (4) most comosa accessions possess two weak intercalary pTa71 sites in the long arm of 2M, and only one distal site is present in heldreichii. Two pairs of 5S rDNA loci with unequal size are revealed in all Ae. comosa accessions. The signal on 1M is much larger than that on 5M and arrears distally to NOR. The signal on 5M is very faint, especially in heldreichii, and occurs in the middle of short arm (Fig. 6).
Based on hybridization pattern of 5S and 45S rDNA probes we identify a reciprocal translocation between 1M and 6M chromosomes in K-3308 (Suppl. material 1:
Most obvious differences between the subspecies show chromosomes 3M and 3M h (Fig. 6), which differ even in size and morphology. 3M of comosa is larger and more asymmetric (submetacentric) than 3M h of heldreichii (metacentric), due to the loss of a distal part of the long arm. The short arm of 3M (comosa) carries two prominent, often fused GAA n clusters in a proximal part and several smaller sites in the distal and, rarer, in the proximal third of the long arm. A slightly deviant hybridization pattern of 3M in K-3857 (comosa) can be caused by large pericentric inversion (Fig. 4q). Chromosome 3M h of heldreichii shows highly polymorphic labeling patterns (Suppl. material 3: fig.  S3), however three intercalary GAA n -sites in the long arm are constantly present. Most proximal GAA n site varies in signal intensity from huge (K-669, K-1601, K-2272, K-2432) to medium or even small (K-3914, K-3919).
Among all comosa accessions, K-3857, AE 1376 and AE 1377 from Greece exhibit most deviant GAA n patterns (Suppl. material 2: fig. S2n, p, q). According to karyotype analysis, K-1601 is highly heterogeneous and consists of genotypes, belonging to comosa and heldreichii types, and some seedling prove to be hybrids, including hybrids between the subspecies (Suppl. material 3: fig. S3e). Hybrids between the subspecies and even with an unknown tetraploid wheat species have been identified within the accession K-3809 (Fig. 3m, n).
Distribution of GTT 10 probe discriminates subspecies comosa from heldreichii. Two small, but sharp intercalary signals appear in short and long arms of 4M in all comosa accessions (Fig. 4e). These GTT n sites are located at the borders of large pericentromeric GAA n complex (Suppl. material 2: fig. S2g). All studied heldreichii accessions possess distinct GTT n signals on three chromosome pairs, 4M (middle of short arm), 5M and 7M (pericentromeric regions) (Fig. 5b, e, g; Suppl. material 3: fig. S3a). In addition, very weak, fuzzy hybridization signals of GTT n probe may appear in regions overlapping with GAA n in both comosa and heldreichii accessions.
Poor hybridization with ACT 10 probe is observed in eight Ae. comosa accessions examined in our study; chromosomal regions corresponding to C-bands/ GAA n sites show a little brighter intensity (Figs 4c,h,5k). Sharp signals appear on chromosomes 6ML and 7ML of only one comosa accession -AE 1377 (Fig. 4c).

Polymorphism in the distribution of satellite DNAs: pTa-713 family
Most prominent pTa-713 clusters appear in the short arm of 2M and in very proximal parts of both arms of 5M. Two smaller sites are observed in subtelomeric and pericentromeric regions of the long arm of 7M of all studied accessions of Ae. comosa (Figs 4j,k,5c,6). Signals in a terminal part of 1MS, a distal part of 1ML, and in subtelomeric region of 5ML are detected in several accessions belonging to both subspecies, however, some polymorphic pTa-713 sites are specific for either comosa or heldreichii (Suppl. material 5: fig. S5, Suppl. material 6: fig. S6).
Thus, sixteen of 20 accessions of comosa carry small signal in a proximal third of 3ML, while the signal in the middle of 5ML is detected in more than a half comosa accessions (Suppl. material 5: fig. S5). AE 1376 and AE 1377 (comosa) show the unique pTa-713 patterns on 3M chromosome consisting of clear signal in the terminus of short arm and fuzzy signal in a distal part of the long arm (Suppl. material 5: fig. S5g, h, shown with pink arrows), which are not observed in comosa and heldreichii accessions, except K-3811 (heldreichii) (Suppl. material 6: fig. S6e). Most comosa also possess two faint signals in the short arm of 4M. Two comosa's lack these marker sites, but carry a small signal in the distal part of 4ML arm (Suppl. material 5: fig. S5o, t). Many small sites in unusual positions appear in Turkish accession K-3787 (comosa).
All heldreichii accessions show characteristic labeling pattern of chromosome 6M h , which carries clear pTa713 signal in the short arm, either adjacent to secondary constriction (8 of 13 accessions) or in the middle of satellite (2 accessions), and two signals in the proximal half of the long arm (Suppl. material 6: fig. S6). We consider such pattern as heldreichii-specific, although it is not observed in two "comosa-like" genotypes of K-1601. These genotypes carry "short" satellite on 6M h , similar to chromosome 6M of comosa. A unique prominent pTa-713 site is detected in a distal part of 5M h L of K-3914 (heldreichii) (Suppl. material 6: fig. S6a).

Polymorphism in the distribution of satellite DNAs: pTa-k566 and oligo-45 families
The pTa-k566 sequence hybridizes to all chromosomes of Ae. comosa and labeling patterns varied between the accessions. Polymorphism observed on 1M, 2M, 3M, 4M, 6M, and 7M chromosomes is found to be subspecies-specific (Fig. 6, Suppl. material 4: fig. S4). Chromosomes 7M of comosa and heldreichii with similar probe distribution differ only in signal intensities, whereas other chromosomes carry pTa-k566 sites specific for only one of the subspecies. Chromosome 1M of comosa contains pTa-k566 sites at the both sides of the centromere, but only one site occurs on 1M h L of heldreichii. Median and distal pTa-k566 signals are present in the long arm of 2M of comosa, while only distal of them appears in heldreichii. Chromosome 3M of comosa carries prominent pTa-k566 site in the proximal region of the long arm, while the respective site on 3M h of heldreichii is much weaker and the chromosome possess the second site distally in the short arm. Chromosome 4M of comosa lacks small proximal pTa-k566 site in the short arm, which occurs on 4M h of all heldreichii accessions. comosa differs from heldreichii in the presence of clear pTa-k566 site in the proximal regions of 6M short arm, although it is also detected in two of 11 studied comosa accessions, namely K-3787 and AE 1377 (Suppl. material 4: fig. S4). Noteworthy that all pTa-k566 sites overlap with o-18S loci.

Polymorphism in the distribution of satellite DNAs: oligo-42 and oligo-44 families
No signals of oligo-42 probe are detected on chromosomes of five heldreichii accessions, and very weak, inconsistent signals are found on chromosomes of very few comosa accessions. Most frequently, signals occur in the long arm of 3M (Fig. 4l), but in AE 1377 oligo-42 signals are also observed on other chromosomes (Fig. 6). Positions of these sites coincide with the signals of o-18S probe. Owing to weakness and inconsistency of oligo-42 signals, this probe is considered non-informative for FISH-analysis of Ae. comosa.
All Ae. comosa accessions contain clear oligo-44 site approximately in the middle of 5M short arm (Figs 4l, 5j; Suppl. material 4: fig. S4). Additional, weaker signals emerge in the proximal part of 4M short arm and in a proximal third of 7M short arm. Besides them, many, but not all heldreichii accessions carry minor oligo-44 signals proximally in 3M short arm and in pericentromeric region of 6M short arm. Both heldreichii-specific sites are recorded in AE 1377 (comosa), which however exhibits also several comosa-specific karyotype features.

Polymorphisms of electrophoretic patterns of gliadins
Genetic variability of Ae. comosa was also assessed using electrophoretic analysis of gliadins in 13 comosa and 13 heldreichii accessions. Comparison of gliadin profiles of all 26 accessions reveals an extremely broad intraspecific variation of Ae. comosa: each accession shows the unique profile. Spectra of comosa accessions are usually more "enriched in components" compared to heldreichii accessions (Fig. 7). This trend is manifested in a larger number of protein bands and their higher intensities, which is especially clear in the a-zone of electrophoretic spectra controlled by chromosomes of homoeologous group 6. From the other side, this trend is not mandatory because gliadin profiles of some comosa accessions (e.g., AE 1257, AE 1376, AE 1377) are very poor, whereas several heldreichii accessions exhibit rich spectra (K-3919; K-4498, some genotypes of K-1601).
Several accessions analyzed in this study are genetically homogeneous and consist of genotypes with identical spectra (e.g., K-3914, K-3309, K-3920, AE 1376). Other accessions are found to be heterogeneous and show two or even more gliadin profiles (K-3857, K-3809 (comosa), K-1601 and K-2432 (heldreichii)). The broadest variation of gliadin patterns is detected in K-1601. We identified six variants of electrophoretic spectra among eight individual grains taken from four spikes; they differ in position of polypeptide bands in γand ω-zones (Fig. 8). The spectra 1a and 1b are more typical for comosa, whereas 2a and 3b (different spikes) -for heldreichii. The spectrum 3a contains bands present on 2a and 3b and probably represents a hybrid between these two genotypes. Similarly, the spectrum 1a may correspond to hybrid between 1b and 3b (Fig. 8). Another heterogeneous accession -K-3809 (comosa) is characterized by spectra highly enriched with gliadin components, which differ significantly from other Ae. comosa in the number, position and size of protein bands. This genotype can represent a hybrid of Ae. comosa with unknown 4× wheat species.
Comparison of gliadin patterns of different accessions of Ae. comosa shows that several combinations of polypeptide bands on electrophoretic spectra always appear together and are inherited as blocks of components (Fig. 7). Such coincidence means that (1) these bands are controlled by a common locus, which, in turn, indicates (2) that accessions under comparison are genetically related. Based on the presence of such common blocks we discriminate six gliadin Families among 26 Ae. comosa accessions (Fig. 7): (1) Families include accessions mainly from one subspecies. Thus, Families 1 and 2, which exhibited the richest spectra, consist of predominantly comosa accessions. Families 4 and 6 and partially Family 5 having relatively poor spectra are mainly composed by heldrechii accessions. Although both accessions from Family 6 have been assigned to subspecies comosa, their gliadin spectra share more common features with the spectra of heldrechii accessions AE 783, K-669, K-3804, K-3811, and K-3897 than with those of comosa. Each of the five abovementioned heldrechii accessions show unique gliadin profile, which cannot be assigned to either one of the families due to dissimilarities in position and intensity of polypeptide bands on electrophoretic spectra. High ratio of heldrechii accessions with the unique gliadin spectra is an indicative of higher variability of this subspecies compared to subspecies comosa.

Discussion
Our current results and the data available from literature ) demonstrate that the levels of intraspecific diversity in Ae. comosa significantly vary depending on markers used for their assessment. Thus, FISH with GAA n probe and gliadin electrophoresis uncover an extremely broad polymorphism of this species. High effectiveness of these markers for the analysis of intra-and interspecific diversity, evaluation of population structure, for characterization of individual genotypes has been proved in many publications (Metakovsky et al. 1989;Badaeva et al. 1990Badaeva et al. , 2015bBadaeva et al. , 2022Masci et al. 1992;Pomortsev et al. 2011;Keskin et al. 2015;Jiang et al. 2017;Majka et al. 2017;Song et al. 2020). We find that, from one side, distribution of GAA n sites is species-and chromosomes specific allowing identification of all individual chromosomes. From the other side, each accession carries a unique combination of polymorphic GAA n sites and unique gliadin spectrum comprising their individuality.
Comparison of the GAA n patterns of Ae. comosa chromosomes obtained in a current study and those reported previously (Molnár et al. 2016;Song et al. 2020) with the C-banding patterns Friebe et al. 1996;Badaeva et al. 1999) show that positions of GAA n overlap with location of C-bands. Two other microsatellite sequences -GTT n and ACT n , which are known to be the components of constitutive heterochromatin in chromosomes of wheat and other grass species (Pedersen and Langridge 1997;Cuadrado et al. 2000;Cuadrado et al. 2008a, b;Luo et al. 2017Luo et al. , 2018Zhang et al. 2022) are detected in minor quantities. The (AC) n sites co-localize with positions of minor NORs visualized with pTa71 probe on 1MS, 2MS, 3MS/3ML, and 5MS arms. These facts indicate that heterochromatin regions detected using Giemsa C-banding on Ae. comosa chromosomes are composed by predominantly GAA n repeat, which is typical for the Triticeae, except diploid Thinopyrum Á.Löve, 1980 einkorn wheat and Ae. tauschii. Chromosomes of these species contain only little amounts of microsatellite sequences or do not possess them at all (Badaeva et al. 2015a(Badaeva et al. , 2019aLinc et al. 2017;Li et al. 2018;Zhao et al. 2018;Ebrahimzadegan et al. 2021).
Although labeling patterns of GAA n probe prove to be highly informative for Ae. comosa chromosome identification and authentication of gene bank accessions, they are too polymorphic and complicated for broad-scale phylogenetic analyses. A similar complexity and ambiguity is found for gliadin profiles. The appropriate markers should be relatively simple and easy to score and should generate specific and reproducible patterns. Eight out of 15 FISH probes used in our study fit these criteria: the 5S and 45S rDNAs, pAs1, pSc119.2, pTa-713, pTa-k566, oligo-44, and oligo-45 probes (oligo-42 and ACT n are found to be low informative for the analysis of Ae. comosa chromosomes due to weak and inconsistent labeling patterns). We used these eight probes for verification of the M-genome chromosome classification and for the assessment of intraspecific diversity of Ae. comosa.
Our study reveals an interesting feature of two oligo-probes designed for the detection of 45S rDNA loci. The oligo-pTa71-2 probe developed by Tang et al. (2014) is homologous to wheat rDNA 25S-18S intergenic region EcoRI-BamHI fragment (X-7841.1). This oligo-probe proves to be effective for the analysis of wheat and Aegilops species, however it fails to detect NORs in most other plant taxa, including, for example, barley, oat, or Erantus (Ranunculaceae) (Mitrenina et al. 2021). Owing to this, we tried to design a new oligo-probe for rDNA loci (based on genome sequence of Aegilops tauschii), which will be applicable for many plant species. Nucleotide sequence of a newly designed oligo-18S is homologous to highly conservative region of 18S rDNA gene of Aegilops, Triticum, Hordeum, Musa, and Iris (Suppl. material 8: fig. S8d) and it was able to detect major NORs on barley chromosomes, although signals were very weak. The o-18S hybridizes to chromosomes of many Aegilops species, however it was more effective for the detection of minor NORs (Suppl. material 8: fig. S8c). In contrast to pTa71 probe obtained from plasmid DNA or oligo-pTa71, o-18S does not detect major NORs and fails to reveal marker terminal 45S rDNA loci on 2M, 3M and 5M chromosomes (Suppl. material 8: fig. S8a, b). Currently we have no explanation of this phenomenon.
Correct chromosome classification is essential for phylogenetic analyses. Nomenclatures suggested for Ae. comosa chromosome classification have been built on different principles. In early studies, the authors followed the rules of cytological nomenclature: chromosomes are arranged according to decreasing length and arm ratio (Chennaveeraiah 1960;Georgiou et al. 1992). Currently most chromosome classifications of the Triticeae species are based on homoeology with common wheat chromosomes -genetic nomenclature. Although genetic nomenclature of Ae. comosa chromosomes have been suggested in several publications (Badaeva et al. 1996a;Friebe et al. 1996;Molnár et al. 2011Molnár et al. , 2016Liu et al. 2019;Song et al. 2020;Said et al. 2021), classification of some chromosomes is still controversial. Two chromosomes -2M and 5M, prove to be most difficult for discrimination owing to same DNA content, similar morphology, pAs1 labeling patterns, heterochromatin content and distribution.
Chromosome 2M was first identified and assigned to genetic group 2 by S. Nasuda et al. (1998) using RFLP, GISH and C-banding techniques in Ae. comosa, wheat cultivar Compair, and wheat-Ae. comosa 2A/ 2M and 2D/ 2M translocation lines. Classification of 2M and 5M was further validated by M. Said et al. (2021) based on FISH mapping of tandem repeats and wheat single-gene probes. In a current study we confirmed classification of 2M and 5M by using DNA probes pTa794 (5S rDNA) and oligo-44. This is because the 5S rDNA loci in Triticum Linneaus, 1753 and Aegilops occur only on group 1 and 5 chromosomes (Dvořák et al. 1989), and in diploid Aegilops species the respective signals appear in a distal part of group 1 chromosomes and in the middle of short arm of group 5 chromosomes (Badaeva et al. 1996b). The 5S rDNA signal on 5M localizes in the middle of short arm, thus this chromosome belongs to genetic group 5. No signals of 5S rDNA probe are detected on another chromosome, which is designated 2M (Fig. 9).
Another validation of genetic group for chromosome 5M come from hybridization pattern of oligo-44 probe. This probe was developed from chromosome-specific tandem repeats of wheat (Tang et al. 2018b) and mapped on all group 5 chromosomes, 3A and 7A of Chinese Spring. A little later, T. Lang at al. (2019) using bioinformatics tools identified the homologous minisatellite repeat, Ta-3A1 from the wheat genome assembly and localized it on group 5 chromosomes of wheat (2x, 4x and 6x), rye, Aegilops, Dasypyrum (Cosson et Durieu) T.Durand, 1888 and Thinopyrum species. The CL241 probe, another homolog of oligo-44 isolated from Ae. crassa genome, was also mapped to all group 5 chromosomes of Ae. tauschii (2n=2x=14, DD), tetraploid and hexaploid Ae. crassa (2n=4x=28, D 1 D 1 X cr X cr and 2n=6x=42, D 1 D 1 X cr X cr D 2 D 2 (Kroupin et al. 2022).
Interestingly, that in Ae. comosa we reveal minor oligo-44 signals on chromosomes 3M and 7M, the homoeologs of 3A and 7A, which also possess the TA-3A1 sites in karyotypes of diploid, tetraploid and hexaploid wheat species, but not in Aegilops .
Another interesting outcome from our study is a possibility of discrimination of the two Ae. comosa subspecies using chromosomal markers. From one side, all thirtysix accessions of Ae. comosa included in the analyses show similar karyotype structures, distribution of rDNA loci, labeling patterns of repetitive DNA sequences indicating that they all belong to one biological species. Despite these similarities, we found clear Figure 9. Characterization of 2M and 5M chromosomes of Ae. comosa using sequential FISH with (round 1) pTa794 (5S rDNA), pTa71 (45S rDNA), followed by (round 2) GAA 10 and pTa-713 probes. Probe names are shown on the bottom, probe color corresponds to signal color. Arrows point to 5S rDNA loci on chromosome 5M. and reproducible differences between the subspecies in morphology, C-banding and FISH patterns of two chromosome pairs -3M and 6M. We hypothesize that these two chromosomes were involved in the subspecies-specific translocation identified earlier in Ae. comosa by mean of meiotic analysis (Kihara 1940). This translocation probably led to the shortage of the long arm of 3M and the increase in the satellite length on 6M of heldreichii compared to comosa.
Interestingly, all heldreichii accessions carrying large SAT on chromosome 6M showed relatively poor gliadin profiles (Fig.8). Considering these facts, we suggest that the 3M-6M translocation can change the functioning of gliadin loci on 6M h chromosome resulting in depletion of gliadin spectra in the α-zone. Intraspecific differentiation of Ae. comosa (Groups I and II) in the morphology of SAT chromosomes observed in our study and reported previously (Karataglis 1975) is caused by this subspecies-specific translocation, whereas the Group III probably evolved due to reciprocal 1M-6M translocation identified here in comosa K-3308, K-3309, K-3787 (all from Turkey) and K-4873 (heldreichii) accessions using FISH with rDNA, pTa-713, and GAA n probes.
Peculiarities of C-banding Friebe et al. 1996;Badaeva et al. 1999)/ GAA n patterns also allow discrimination of Ae. comosa subspecies indicating that their divergence was accompanied by amplification/ elimination and re-distribution of microsatellite repeats. These results however contradict the observations of Z. Song et al. (2020), who have not revealed any differences between the subspecies in GAA n labeling patterns. These authors, however, did not provide complete karyotypes of accessions used in their analyses, which does not permit the direct comparison of our data.
Differences between the subspecies are most clearly detected using a combination of pTa-k566 and oligo-45 probes; the pTa-713 also shows subspecies-specific patterns. These three probes prove to be best choice for the precise discrimination of comosa from heldreichii accessions using FISH markers. Most Ae. comosa chromosomes show rather conservative patterns, while diagnostic sites appear mainly on 3M or 6M. Different repeats are often accumulated in a single cluster. Such complex cluster, composed of 45S rDNA, pTa-713, pTa-k566, oligo-42, and, rarely, GAA n appears in a proximal third of the long arm of 3M (comosa) (Fig. 6). These repeats are mainly absent or found in minor quantities in the long arm of 3M h (heldreichii), which contains small cluster consisting of oligo-42 and oligo-45 in the short arm. Clear signals of pTa-713, oligo-44 and oligo-45 probes are detected on chromosome 6M h of heldreichii, but they are lacking in 6M of Ae. comosa. Although pAs1 probe cannot reliably discriminate comosa from heldreichii, pSc119.2-labeling pattern of 7M shows differences between the subspecies, although with few exceptions. Thus, two comosa accessions -AE 1377 and K-3809, carry two pSc119.2 sites on 7ML, which is typical for heldreichii. These accessions however are deviant and share karyotypic features of both subspecies. Cytogenetic analysis reveals that some grains of K-3809 represent hybrids of Ae. comosa with unknown tetraploid wheat, which can explain a deviant gliadin profile of this accession, and one genotype possessing heteromorphic chromosome 5M is probably a derivative of comosa × heldreichii cross.
The significant role of hybridization in evolution and diversification of Ae. comosa is supported by other studies (Georgiou et al. 1992;Song et al. 2020). Thus, these papers described many heterozygotes in Ae. comosa, and in a current study we found genotypes segregating in labeling patterns of one to all seven chromosome pairs (see Suppl. material 1: fig. S1, h10, Suppl. material 3: fig. S3e, Suppl. material 4: fig. S4j), which point to their hybrid origin. Accession K-1601 (heldreichii) shows the highest heterogeneity: each of the seven genotypes examined by FISH and 5/8 genotypes analyzed by gliadin electrophoresis show unique patterns. Karyotypic features of some K-1601 genotypes correspond to heldreichii subspecies, while other share similarities with comosa. A similar trend is uncovered by gliadin analysis. Both FISH and gliadin electrophoresis identify many heterozygotes in K-1601, which may represent recent hybrids, including hybrids between the subspecies. It should be mentioned, however, that no variation in spike morphology (all heldreichii-like, see Fig. 2) has been identified between individual plants of this accession.
Three accessions, AE 1376, AE 1377 and K-3857 assigned to subspecies comosa based on botanical characters, combine chromosomal features of both Ae. comosa subspecies assuming that they might have hybrid origin. Gliadin analysis supports closer relations of AE 1376 and AE 1377 with heldreichii than with comosa indicating that taxonomical position of these accessions should be verified. Probably these forms emerged via hybridization of comosa and heldreichii followed by karyotype stabilization toward heldreichii (AE 1376, AE 1377) or comosa (K-3857) parent. In contrast to K-1601 or K-3809, these three accessions are cytogenetically stable and genetically uniform. Most likely, they emerged via comosa × heldreichii hybridization long time ago, and hybrid forms become stabilized over generations. Based on these facts we suggest that hybridization, including hybridization between subspecies, plays an important role in broadening genome diversity of this grass. It can be facilitated by following factors: • comosa and heldreichii often grow together in mix stands (Zhukovsky 1928;Van Slageren 1994); • heading and flowering time of comosa and heldreichii overlap (Boguslavsky and Golik 2003); • Ae. comosa is considered as autogamous species, however, open pollination could be more common event than it is usually believed (Georgiou et al. 1992;Boguslavsky and Golik 2003); • hybrids between the subspecies are partially fertile (Kihara 1940).
Summarizing results of a current study, we recommend the following set of markers for the precise identification of individual chromosomes and for discrimination of Ae. comosa subspecies using FISH markers (Table 2).

Conclusions
FISH and gliadin electrophoresis reveal broad intraspecific polymorphism of GAA n patterns and gliadin profiles of Ae. comosa allowing not only genetic authentication of gene bank accessions, but also discrimination between the subspecies. Application of these markers however will be too complicated for the broad-scale phylogenetic analyses. Table 2. Probe combinations for the M-genome chromosome identification and discrimination of Ae. comosa subspecies (according to Fig. 6).
two faint (comosa) vs one clear (heldreichii) GTT n sites in short arm 5M Minor NOR in short arm overlapping with (AC) n ; pericentromeric GTT n cluster (heldreichii); 5S rDNA site in the middle of short arm; oligo-45 site in short arm (comosa). Two prominent pTa-713 clusters; pTa-k566 site in long arm; oligo-44 site in short arm; pSc119.2 signals are mainly absent; pAs1 sites distally in the long arm and terminally in the short arm. 6M Satellite in physically longer arm carrying major NOR; Medium (comosa) vs large (heldreichii) satellite; terminal pSc119.2 and distal pAs1 sites in the arm, opposite to NOR. small pTa-713 sites in both arms (heldreichii); oligo-44, oligo-45, and pTa-k566 sites in the SAT arm (heldreichii). 7M proximal pTa71/ o-18S/ pTa-k566 sites in short arm; two (heldreichii) vs one (comosa) pSc119.2 sites in long arm; oligo-45 site in short arm; pTa-713 sites in subtelomeric and proximal regions of long arm. GTT n cluster in proximal part of short arm (heldreichii); o-18S, pTa-k566, oligo-45 signal intensities By using group-specific FISH markers, we justify classification of 2M and 5M chromosomes of Ae. comosa and suggest a set of DNA probes for the precise identification of each of the seven M-genome chromosomes.
Two subspecies of Ae. comosa -comosa and heldreichii, are karyotypically distinct and diverge from each other as a result of subspecies-specific translocation 3M-6M, which probably affects functioning of gliadin locus. Divergence of subspecies was accompanied with amplification/ elimination and re-distribution of the repeated DNA sequences.
Three FISH probes, pTa-k566, pTa-713, and oligo-45 generate clear and reproducible patterns specific for comosa or heldreichii accessions; they can serve as reliable markers for discrimination of Ae. comosa subspecies.
An extremely broad genetic variability of GAA n -FISH patterns and gliadin profiles revealed in Ae. comosa -an endemic autogamous plant species (Van Slageren 1994), can be due to frequent occurrence of hybridization, including hybridization of comosa with heldreichii or with other neighboring wheat or Aegilops species.

Author contributions
EB (Badaeva E.D.) planned and performed the experiments, analyzed data, wrote the first draft of the manuscript; KV: make chromosomal preparations and participates in FISH experiments; FA: carried gliadin electrophoresis, analyzed data; CN and BM: provide and characterized materials for this study; ZP: designed oligo probe; SS: synthesized oligo-probes; DA: analyzed gliadin spectra and wrote the manuscript. All authors read and approved the submitted version of the manuscript and agree to be personally accountable for their own contributions and for ensuring that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and documented in the literature.