Similarities and differences in the nuclear genome organization within Pooideae species revealed by comparative genomic in situ hybridization (GISH)

In this paper, we highlight the affinity between the genomes of key representatives of the Pooideae subfamily, revealed at the chromosomal level by genomic in situ hybridization (GISH). The analyses were conducted using labeled probes from each species to hybridize with chromosomes of every species used in this study based on a “round robin” rule. As a result, the whole chromosomes or chromosome regions were distinguished or variable types of signals were visualized to prove the different levels of the relationships between genomes used in this study. We observed the unexpected lack of signals in secondary constrictions of rye (RR) chromosomes probed by triticale (AABBRR) genomic DNA. We have also identified unlabeled chromosome regions, which point to species-specific sequences connected with disparate pathways of chromosome differentiation. Our results revealed a conservative character of coding sequence of 35S rDNA among selected species of the genera Aegilops, Brachypodium, Festuca, Hordeum, Lolium, Secale, and Triticum. In summary, we showed strong relationships in genomic DNA sequences between species which have been previously reported to be phylogenetically distant.


Introduction
The Poaceae family (Barnh., Gramineae Juss.) is the fourth largest group of plants on Earth and can be found in nearly all regions and climate zones. More than 10,000 species and 600 genera constitute this family, but only three subfamilies, namely Ehrhartoideae, Pooideae, and Panicoideae, play the principal role of providing the human race with nutrition and a source of renewable energy. The Pooideae subfamily includes some economically important cereals, such as bread wheat (Triticum aestivum L.), barley (Hordeum vulgare L.), rye (Secale cereale L.), and triticale (×Triticosecale Wittm.), as well as many lawn and pasture grasses, such as perennial ryegrass (Lolium perenne L.). The genome composition of the species within Pooideae differs in respect to their size, ploidy level, basic chromosome number, and repetitive sequences content. For example, bread wheat (Triticum aestivum L.; 17 Gb; 2n = 6x = 42 chromosomes) is an allopolyploid, composed of six sets of chromosomes belonging to A, B, and D genomes with high content of non-coding, repetitive DNA. Contrastingly, Brachypodium distachyon has a small genome (272 Mb; 2n = 2x = 10 chromosomes) and low amount of repetitive sequences (Vogel and Hill 2008;International Brachypodium Initiative (IBI) 2010), making it a good model for structural genomic studies of grasses. It was previously reported that B. distachyon is a comparative and functional genetic model and its potential relevance for research on cereals, on the basis of its physiological and genetic advantages (Draper et al. 2001;Vogel and Bragg 2009). For example, the high degree of genetic synteny between wheat, B. sylvaticum, and B. distachyon enabled the identification of the genes present in the Ph1 (Pairing homeologous 1) locus of wheat (Griffiths et al. 2006;Huo et al. 2009), demonstrating the potential application of the genus Brachypodium in comparative genomics. The phylogenetic relationships between Communicated by: Andrzej Górny B. distachyon and other cereals and grasses have been derived on the basis of the analysis of the internal transcribed spacer (ITS) and rDNA coding sequences (Hsiao et al. 1994), genomic markers, such as random amplified polymorphic DNA (RAPD), restriction fragment length polymorphism (RFLP) (Catalán et al. 1995), chloroplast restriction sites, and ndhF gene sequence (Kellogg 2001;Mochida and Shinozaki 2013). Those approaches showed that B. distachyon can be considered as a kind of missing link between rice (Oryza sativa L.) and temperate grasses.
Exploring the genomes of a wide range of plant species revealed the existence of highly conserved sequences and synteny in gene order. Comparative genomics can distinguish synteny, which means the presence of two or more loci on the same chromosome, and collinearity, which refers to the similarities in the physical order of loci (McCouch 2001). However, recently, the synteny is defined as the conservation of blocks of order within the sets of chromosomes which are being compared to each other. The first study of chromosome evolution and comparative genomics was performed between potato (Solanum tuberosum) and tomato (Lycopersicon esculentum) (Bonierbale et al. 1988). Initially, low copy RFLP markers and low resolution genetic maps were used to determine the phylogenetic relationships between Poaceae species, and established that most of the grass genomes are composed from 30 rice-independent linkage blocks (Gale and Devos 1998;Keller and Feuillet 2000). These methods enabled the macro-collinearity of grass genomes to be determined; however, they also overestimated the level of synteny. Recently, the development of additional genomic resources such as EST marker databases and whole genome sequences, combined with improved software, has enabled comparative genomic studies to reach a higher level of resolution, uncovering the micro-collinearity of sequences, which form the basis to establish a model of grass genome evolution (Devos 2005;Wei et al. 2007;Salse et al. 2008;Murat et al. 2014).
One of the main factors leading to chromosome evolution and speciation are changes in the amount and distribution of repetitive DNA sequences (Flavell et al. 1979;Cuadrado and Jouve 2002). It is hypothesized that these sequences play an important role in chromosome organization, stabilization of chromosome structure, recognition and segregation of chromosomes in mitosis and meiosis, and regulation of gene activity (Vershinin et al. 1995). Repetitive sequences account for up to 90 % of plant genomes (Heslop-Harrison 2000), and show diverse compositions in different genomes. Changes in repetitive sequences could be particularly useful in studies of genomic evolution. Some families of those sequences, e.g., Afa family (Nagaki et al. 1995), are present in more than one genus, which could be an evidence of their genomic relationship. Genomic in situ hybridization (GISH) can distinguish particular genomes at the chromosomal level. GISH can result in chromosome painting and show similarities of repetitive DNA distribution between related species. Moreover, the physical locations of conserved sequences can be visualized on chromosomes as well.
In this work, we compare both the model species (Brachypodium distachyon) as well as other important grasses, using genomic DNA probes. Our intention was to verify and classify cytologically visible similarities and differences in repetitive non-coding DNA sequences locations in given Pooideae species. A comparative approach using GISH, instead of comparison of selected sequences or specific chromosome regions, was performed to widen the understanding of the relationships within the Pooideae subfamily at the chromosomal level.

Plant material
Ten species (Table 1) were selected for comparative GISH analysis: Aegilops speltoides Tausch., Aegilops tauschii Coss., Brachypodium distachyon L., Festuca pratensis L., Hordeum vulgare L., Lolium perenne L., Secale cereale L., Triticosecale Wittm., Triticum aestivum L., and Triticum urartu Tumanian ex Gandilyan. Brachypodium distachyon material was sourced from the University of Silesia in Katowice, Poland. Seeds of Ae. speltoides, S. cereale, T. aestivum, and T. urartu were kindly supplied for the study from the National Small Grains Germplasm Research Facility, National Small Grains Collection (Aberdeen, Idaho, USA). The remaining species were provided by a collection of the Institute of Plant Genetics, Polish Academy of Sciences in Poznan, Poland.

Chromosome preparation
Seeds of each species were germinated on filter paper in Petri dishes for 3-4 days in the dark. The root tips were immersed in ice-cooled water for 26 h, fixed in ethanol and acetic acid (3:1, v/v), and stored at −20°C until required. Mitotic chromosome preparations were made from root tips digested in a mixture of enzymes, diluted with 0.01 M sodium citric buffer, containing 20 % (v/v) pectinase (Sigma), 1 % (w/v) cellulose (Calbiochem), and 1 % (w/v) cellulase 'Onozuka R-10' (Serva). Meristems were dissected from root tips, squashed in drops of 45 % acetic acid, and the good quality preparation was frozen (Hasterok et al. 2006).

DNA isolation and probe labeling
Genomic DNA (gDNA) from young leaves of all selected species were isolated with the C-TAB method (Doyle and Doyle 1990). DNA from all plants were obtained at the same stage. After the extraction of DNA, samples were labeled by nick translation with tetramethylrhodamine-5-dUTP (Sigma-Aldrich). The ribosomal sequences 35S rDNA and 5S rDNA were labeled with digoxigenin-11-dUTP by nick translation and with tetramethyl-rhodamine-5-dUTP (Sigma-Aldrich) by polymerase chain reaction (PCR), respectively (Kwiatek et al. 2016b).

In situ hybridization
The GISH procedures were performed according to the protocol of Kosmala et al. (2006) and Kwiatek et al. (2016a, b), with minor modifications. The GISH protocol was standardized by several repeats to ensure that the obtained results were comparable and reproducible. The hybridization mixture consisted of 50 % deionized formamide, 10 % dextran sulfate, 2 × SSC, 0.5 % SDS, as well as 100-120 ng/slide gDNA probe. The blocking DNA was not used. Several initial GISH experiments were carried out to reveal optimal specifications, such as probe concentrations and wash temperatures. After establishing optimal GISH conditions, we used only those conditions for all experiments described in this paper. Chromosome preparations and hybridization mixture were denatured together at 80°C for 2 min and then hybridized overnight at 37°C. The post-hybridization washes were performed, according to Heslop-Harrison (2000), at 42°C in 2 × SSC buffer. Probes labeled with tetramethylrhodamine-5-dUTP were directly visualized. After the acquisition of images, selected slides were washed off and reprobed with a new set of probes (35S rDNA and 5S rDNA). The reprobing procedures were performed according to the protocol of Heslop-Harrison et al. (1992), with minor modifications. Immunodetection of the digoxigenated probe (35S rDNA) was performed using fluorescein isothiocyanate-conjugated anti-digoxigenin antibody (Sigma-Aldrich). Mitotic cells were examined with an Olympus XM10 CCD camera attached to an Olympus BX61 automatic epifluorescence microscope. Image processing was carried out using Olympus Cell-F imaging software and Micrografx Picture Publisher software.

Results
GISH analyses were performed in order to determine the relationships between total genomic DNA of key crops and forage species at the chromosomal level. The experiments were carried out using labeled probes from each species to hybridize with chromosomes of every species used in this study based on a Bround robin^rule. Overall, 100 GISH experiments were used (10 probes × 10 species, Table 2). The experiments showed various types of signal locations, including: telomeric and centromeric regions, rDNA loci, subcentromeric and subtelomeric regions, as well as chromosome labeling or signals dispersed along whole chromosomes. Additionally, rDNA-FISH (ribosomal DNA fluorescent in situ hybridization) was carried out when GISH revealed rDNAlike signals.

Secale cereale chromosomes
All rye (S. cereale; 2n = 2x = 14) chromosomes were labeled by H. vulgare, S. cereale, and ×Triticosecale probes. The triticale probe did not hybridize with two 35S rDNA loci (Fig. 3c). The H. vulgare probe also identified an additional two signals located on the pericentromeric regions of the chromosomes lacking rDNA loci. Hybridizations with the rest of probes, except for B. distachyon probe, resulted in labeling of all the rye chromosome except for the telomeric regions. Similarly to triticale, the wheat probe also did not label two

Triticum aestivum chromosomes
All wheat chromosomes (T. aestivum; 2n = 6x = 42) were labeled by T. aestivum, T. urartu, Ae. tauschii (Fig. 1c), and S. cereale gDNA probes, although the T. urartu probe resulted in 14 chromosomes of A genome being marked stronger. The rye gDNA probe showed strong and dispersed signals in most of the wheat chromosomes. Triticale and barley gDNA probes also labeled all chromosomes; however, the centromeric and telomeric regions as well as two 35S rDNA loci (only in the case of triticale probe) remained unlabeled. Contrastingly, the Ae. speltoides gDNA probe gave only 14 signals in 5S rDNA loci and 14 centromeric regions in B genome chromosomes. The B. distachyon probe resulted in labeled telomeric sites and four signals in 35S rDNA loci, while L. perenne yielded four 35S rDNA loci and telomeric regions. Finally, weak and dispersed signals were observed using F. pratensis gDNA probe.

Aegilops tauschii chromosomes
Aegilops tauschii (2n = 2x = 14) chromosomes were marked in GISH experiments using T. aestivum, Ae. tauschii, S. cereale, and Ae. speltoides gDNA probes. Additionally, Ae. tauschii probe hybridization resulted in strong telomeric signals and clear dispersed signals along all chromosomes. Six telomeric sites were detected using rye probe. The Ae. speltoides gDNA probe gave also two terminal signals in Ae. tauschii chromosomes. The ×Triticosecale and T. urartu gDNA probes also marked all chromosomes; however, the telomeric regions and two 35S rDNA loci (only in the case of triticale probe) remained unlabeled. Labeled segments of Ae. tauschii were observed when H. vulgare gDNA probe was used in the GISH experiments. In contrast, L. perenne (Fig. 1d) and F. pratensis gDNA probes resulted in weak dispersed sites observed in all chromosomes. Additionally, the L. perenne probe gave signals in two 35S and two 5S rDNA loci, and also labeled all telomeric regions (Fig. 1d). Two 35S rDNA loci and telomeric sites were also labeled when GISH was carried out using B. distachyon gDNA probe.

Aegilops speltoides chromosomes
The experiments with labeled DNA of Ae. tauschii, T. urartu (Fig. 3a), S. cereale, T. aestivum, and F. pratensis (Fig. 1e) revealed labeling of all Ae. speltoides (2n = 2x = 14) chromosomes with the exception of the telomeric regions, whereas gDNA probe of S. cereale revealed strong, distinct signals on several chromosomes. GISH with Ae. speltoides gDNA probe resulted in chromosome labeling and brighter signals in pericentromeric regions. Similar results for all chromosomes o f A e . s p e l t o i d e s w e r e o b t a i n e d w h e n g D N A of ×Triticosecale, T. aestivum, and H. vulgare were used for GISH. What is more, the subtelomeric region of a pair of chromosomes possessing the 5S loci remained unlabeled when hybridization with F. pratensis gDNA (Fig. 1e) was performed. Four 35S and two 5S rDNA loci were obtained for the DNA of L. perenne and B. distachyon (Fig. 2a) and centromeric regions were highlighted using B. distachyon only.

Triticum urartu chromosomes
Labeled gDNA of five species, S. cereale (Fig. 2d), Ae tauschii, T. urartu, F. pratensis, and H. vulgare lead to chromosome labeling on their entire length. Additionally, distinct signals in the pericentromeric regions in two chromosomes and unlabeled centromeric regions in all chromosomes were observed in experiments using H. vulgare DNA (Fig. 3b). Chromosome painting with lack of signals in the terminal part of one pair of chromosomes was obtained for the DNA of Ae. speltoides (Fig. 1f), T. aestivum, and ×Triticosecale. Additionally, Ae. speltoides probe resulted in some strong, clear signals in pericentromeric regions (Fig. 1f).
Brachypodium distachyon probe resulted in strong subtelomeric signals and gave four signals of 35S rDNA loci. The genomic DNA of L. perenne revealed only the 35S rDNA sequences.

×Triticosecale chromosomes
Hybridization with gDNA of B. distachyon and H. vulgare to chromosomes of triticale (×Triticosecale; 2n = 6x = 42) provided contrasting results to the remaining experiments, where chromosome labeling was observed. The gDNA of B. distachyon was mapped in telomeric sites and gave six signals of 35S rDNA loci and ten signals of 5S rDNA loci. The dispersed, strong signals located near the centromeric and telomeric regions in genome B were visible for the experiment with gDNA of H. vulgare. GISH with S. cereale probe enabled to distinguish the R genome chromosomes, while the DNA probe of Ae. speltoides visualized the B genome, and DNA of T. urartu the A genome chromosomes. What is more, the hybridization with DNA of T. aestivum enabled to identify A and B genome as well as R genome chromosomes in terms of negative discrimination (very weak labeling). Probing of homologous gDNA to triticale chromosomes resulted in entire chromosomes being labeled. Finally, probing with Ae. tauschii gDNA resulted in chromosome labeling, but with more intensity of the B genome of ×Triticosecale.

Festuca pratensis chromosomes
GISH with gDNA probes of selected species, with the exception of F. pratensis and L. perenne (Fig. 2c), revealed two signals of 35S rDNA in the karyotype of F. pratensis (2n = 2x = 14). Furthermore, in nearly all cases where 35S rDNA signals were detected, complete chromosome labeling was also observed; the exceptions were B. distachyon and ×Triticosecale. However, B. distachyon probe produced weak signals in telomeric regions. Chromosome labeling was clearly visible when F. pratensis gDNA was used to probe chromosomes of F. pratensis. Hybridization of L. perenne DNA to F. pratensis revealed strong, distinct signals in some chromosomes (Fig. 2c).

Lolium perenne chromosomes
The hybridization of genomic DNA of B. distachyon, H. vulgare, T. aestivum, Ae. tauschii, Ae. speltoides, and T. urartu to chromosomes of L. perenne (2n = 2x = 14) identified the 35S rDNA loci. What is more, the B. distachyon gDNA probe labeled the telomeric sites of L. perenne. Strong, dense signals were observed for S. cereale probe. Lolium perenne chromosomes probed with gDNA from H. vulgare, Ae. tauschii, T. urartu, and Ae. speltoides resulted in both 35S rDNA signals and chromosome labeling. Probing gDNA of ×Triticosecale (Fig. 1g), F. pratensis, and L. perenne resulted in the labeling of all chromosomes. GISH with gDNA of L. perenne to L. perenne chromosomes showed an unlabeled terminal part of a pair of chromosomes that lacked rDNA loci.

Discussion
In this study, we detected and described the hybridization patterns of various genomic DNA probes generated from key representatives of cereals, forage grasses, and B. distachyon on mitotic chromosomes of those species. The aim of this approach was to score the affinity of the given species by using gDNA as probes, without blocking DNA from the species chosen for chromosome spreads preparations. As a result, the whole chromosomes or chromosome regions were distinguished or variable types of signals were visualized to prove the different levels of the relationships between genomes used in this study. A part of the presented results are either as they were expected or confirm the results reported in other papers. However, our study also reveals some novel insights, which require more specific analysis in the future. The most characteristic DNA motifs in the Pooideae subfamily are telomeric and rDNA sequences. Our results confirmed the previous study of Hasterok et al. (2004), who reported labeled pericentrom eric regions and 35S rDNA loci on B. distachyon chromosomes using B. distachyon genomic probe. This phenomenon could be explained by the small content of repetitive sequences in this genome. Besides, Mayer et al. (2011) compared the gene order of B. distachyon with H. vulgare and reported that the number of conserved syntenic loci was high (14,422), reflecting a closer phylogenetic affinity. What is more, Ma et al. (2010) carried out GISH reciprocal experiments between B. distachyon and H. vulgare and observed mainly 35S ribosomal DNA loci. In the present study, the same type of affinity was observed in GISH runs with all of the gDNA probes, indicating that this high copy DNA sequence is conserved between analyzed grasses. Ribosomal DNA loci, which are essential for all organisms, can be detected at the cytogenetic level; however, homogenization operates very strongly in these regions in all Poaceae genomes (Sallares and Brown 1999). Signals observed in the centromeric regions of B. distachyon led to the assumption that this species is more closely related to Ae. speltoides and Ae. tauschii than with T. urartu. Moreover, we showed a certain degree of genomic relationship between B. distachyon and F. pratensis. This point of view could also be confirmed by the phylogenetic analysis of grass species based on the ndfH chloroplast gene sequence made by Mochida and Shinozaki (2013).
Considering their importance in agriculture, the most analyzed temperate cereal species in comparative genomics are barley and wheat, along with their relatives. Based on comparison to the previously constructed high-density physical marker map of wheat (Qi et al. 2004), Mayer et al. (2011) reported that barley contains an archetypal Triticeae genome. Most of the chromosome arms showed well-conserved synteny with previously reported chromosomal translocations (Qi et al. 2006). In the present work, we observed labeled pericentromeric regions on H. vulgare chromosomes using Ae. speltoides, T. aestivum, and ×Triticosecale gDNA probes. These observations are correlated with the results of Icsó et al. (2015), where gDNA of H. vulgare was used for the chromosome probing of T. aestivum. They reported that gDNA of barley produced hybridization bands mainly in the pericentromeric and intercalary chromosome regions of B genome chromosomes and it corresponded to the GAA FISH pattern. Surprisingly, we observed differences in the labeling of wheat chromosomes using barley genomic probe in comparison to chromosomes of wheat labeled with gDNA probes of its ancestral species. The three genomes of wheat share similar repetitive sequence types, with the D genome specific repetitive sequences being the most frequent (Nagaki et al. 1995;Jia et al. 2013). In addition, the existence of tandemly repeated sequence dpTa1 presented in 58 species of the Triticeae tribe have been reported (Vershinin et al. 1994). The occurrence of signals detected in phylogenetically distant species may indicate strong relationships between repetitive sequences in selected grasses. For example, the pSc119.2 sequence from rye is widely used in the identification of wheat chromosomes and its relatives (Cuadrado and Jouve 1994;Schneider et al. 2003Schneider et al. , 2005Wiśniewska et al. 2013).
On the other hand, we have observed an unexpected lack of signals in secondary constrictions of rye chromosomes probed by triticale genomic DNA. This may be connected with the differences in rDNA sequences between 1R chromosomes of rye and 1R chromosomes of triticale. In general, unlabeled chromosome regions can be related with species-specific sequences or disparate pathways of chromosome differentiation, which was exposed by the lack of labeled segments of Ae. speltoides chromosomes probed with F. pratensis gDNA. The most likely explanation for the unlabeled telomeric regions could be the fact that the Ae. speltoides genome carries subtelomeric sequences Spelt1 and Spelt52 (Salina et al. 2006(Salina et al. , 2009. On the other hand, the lack of signals from the 5S chromosome segment of Ae. speltoides remains unknown but indicates that this region could have gone under structural changes during S genome evolution. What is more, we observed single signals in pericentromeric locations in some chromosomes of T. urartu probed with Ae. speltoides gDNA. Similarly, Belyayev et al. (2000) reported pericentromeric signals in the chromosomes of T. urartu and suggested that those regions are conserved in A and B genomes. Moreover, comparative GISH resulted in a lack of signals in the subtelomeric regions in the chromosomes of T. aestivum, Ae. speltoides, Ae. tauschii, and H. vulgare. This could be determined by the different constitution of subtelomeric regions. For example, the composition of the termini of S. cereale chromosomes is complex and, besides the basic telomeric sequence (T/A) 1-4 G 1-8 , it also contains tandemly organized DNA families such as pSc119.2, pSc200, and pSc250 (Salina et al. 2009).
Our comparative analysis of genomic relationships within crucial Pooideae species revealed some intriguing results, such as unexpected chromosome labeling between distant species (e.g., ×Triticosecale vs. L. perenne), dispersed signals (e.g., L. perenne vs. F. pratensis), or labeling of selective chromosome regions (e.g., S. cereale vs. Ae. speltoides). The strong, distinct signals observed for L. perenne and F. pratensis can be useful for the karyotyping of F. pratensis chromosomes. Till now, only two pairs of chromosomes, 2F with 35S rDNA and 3F with 5S rDNA loci, can be easily identifiable (Jones et al. 2002;Thomas 1981).
In conclusion, our results are in parallel with the comparative studies made on model species and key representatives of the Pooideae subfamily. However, some results indicated unknown intergenomic dependencies that need further analysis. Moreover, reciprocal GISH analysis on Aegilops, Hordeum, and Triticum species could shed light on the chromosome structure and evolution within the Triticeae tribe.