Fluorescence chromosome banding and FISH mapping in perennial ryegrass, Lolium perenne L.

The development and application of genomic methods to plant genetics and breeding involves the integration of information across the spectrum from DNA sequences to chromosomes, the units of transmission of genetic information [1]. A complete genomic analysis thus requires an ability to unambiguously identify the individual chromosomes of a species. Before 1970, the only available parameters for chromosome identification were the relative lengths and arm ratios [2]. Such parameters were not sufficient to identify chromosomes in species where chromosomes exhibited gross morphological similarities. The development of chromosome banding techniques in the 1970’s resolved this limitation to a large extent and promoted rapid advances in mammalian and human cytogenetics [3]. The patterns of longitudinal differentiation along chromosomes, produced through G-, Q- or R-banding protocols, are specific for each pair of chromosomes in a given species. This unambiguous chromosome identification has led to the development of standardized banded karyotypes incorporating a chromosome numbering system and banding nomenclature for several mammalian species including human [4], cattle [5], sheep [6] and dog [7]. As modern cytogenetics now incorporates several molecular-based technologies, fluorescence in situ hybridization (FISH), which allows direct visualisation of specific DNA sequences on chromosomes, plays an important role in bridging the gap between classical cytogenetics and molecular genetics. The application of a ‘toolbox’ of standardized banded karyotypes in combination with FISH has revolutionised structural, functional and comparative genomics of several mammalian species. On the rare occasions when the chromosome banding pattern cannot resolve chromosome pairs with similar morphologies, FISH mapping of molecular marker sequences can be applied [8].

Similar chromosome banding protocols have not yet been developed for plant species. Several variations of chromosome banding techniques have been tried with plant chromosomes [9–12] but consistent patterns necessary for sequential mapping have not been achieved [13]. It has been argued that the relatively enhanced compactness of the chromatin of plant metaphase chromosomes and/or the adherence of cytoplasmic debris over chromosome preparations obtained from traditional squash preparations obliterates the resolution of G-bands [9, 14]. A fluorescence DAPI (4’, 6-diamidino-2-phenylindole)-banding technique has recently been reported for plant chromosomes which is compatible with FISH mapping experiments using repetitive DNA sequences [15]. The banding patterns were resolved by digital deconvolution and reconstruction of 3D images of chromosomes.

Perennial ryegrass, Lolium perenne (2n = 2x = 14) (subfamily Pooideae; tribe Poeae) is native to the temperate parts of Asia, Europe and North Africa. It belongs to the same grass family as several important food crop plants including wheat (Triticum spp.), barley (Hordeum vulgare), rice (Oryza sativa), maize (Zea mays), sorghum (Sorghum bicolor), as well as the model grass, Brachypodium distachyon. It is widely cultivated for forage and turf because of its rapid establishment from seed, high yield, and good digestibility. However, perennial ryegrass has only limited variation for tolerance to biotic and abiotic stresses [16, 17]. Genomics-supported breeding strategies provide a promising approach to the improvement of the agronomic performances and stress tolerances of many crops, including L. perenne. The complexity of the perennial ryegrass genome, posed by its large size (2.6 Gbp/1Cx; [18]) and consisting of a significant proportion of repetitive DNA sequences, has hampered progress in sequencing of this genome. Nevertheless, ryegrass genomics is advancing rapidly, with the availability of a draft genome sequence [19] and several linkage maps [20–23]. Despite these advances, it has not so far been possible to unambiguously relate LGs with individual L. perenne chromosomes. A key reason is that all seven pairs of conventionally stained chromosomes of ryegrass, biarmed in morphology, display a symmetrical karyotype [24, 25] which makes their individual identification extremely difficult [1].

In this work, we have successfully obtained for the first time fluorescence Q-banding patterns using quinacrine mustard (QM) on ryegrass chromosomes prepared by an improved flame drying technique previously developed in our laboratory [26]. Importantly, the banding method is compatible with the FISH technique for regional chromosomal mapping of DNA sequences so that the banding and the FISH signals can be visualised in the same cell. These results have subsequently led us to produce a banded karyotype for the first time and develop a corresponding banded ideogram for perennial ryegrass. We have used this resource as a tool to map 18S rDNA and 5S rDNA sequences with FISH to precise chromosome bands. In addition, we have mapped a single copy gene to a L. perenne cytogenetic band, and so have demonstrated the anchorage of a linkage map marker to a chromosome region. Potential applications are discussed, including the combining of genetic and cytological information into an integrated cytogenetic map.

Quinacrine mustard staining displayed distinct and reproducible Q-banding patterns on flame-dried preparations of the somatic chromosomes of both diploid and haploid L. perenne. The same preparations were then subjected to FISH procedures to achieve successful regional mapping of 5S and 18S rDNA sequences (Fig. 1). The availability of large numbers of suitable cells allowed us to produce a Q-banded karyotype for L. perenne (Fig. 2). Individual chromosome pairs in each cell were identified based on morphology, FISH mapping of rDNA markers and banding patterns. Thereafter, morphometric analysis was carried out based on the 10 selected cells (Table 1). The seven pairs of diploid L. perenne chromosomes displayed a gradation in size and have been numbered according to their decreasing size order (Fig. 3). The chromosome sizes ranged from chromosome 1, with 17.93% of the total haploid set to chromosome 7, with 10.83% of the total haploid set. Based on the estimated nuclear genome size of L. perenne (2.623 Gbp/1Cx; [18]), the molecular sizes of the chromosomes ranged from 470 Mbp for chromosome 1 to 284 Mbp for chromosome 7. FISH mapping provided marker status to three chromosome pairs in the diploid L. perenne. Chromosomes 2, 3 and 5 displayed proximal hybridization of 18S rDNA representing nucleolus organizer regions (NORs) or secondary constrictions. While 18S rDNA hybridized on the long arm in chromosomes 3 and 5, it was localised on the short arm in chromosome 2 (Fig. 2). Among these three NOR carrying chromosomes, chromosome 3, with a secondary constriction on the long arm, co-localized with 5S rDNA on the short arm. In our chromosome preparations, the GC-rich NOR chromatin was often decondensed and occasionally stretched, generally with condensed flanking ends (Fig. 1). Consequently, in DAPI staining the two condensed components of the same chromosome often appeared to be separated, without any visible connection. FISH showed that these parts were joined by cloudy decondensed 18S rDNA (Figs. 1d and 4b) and that there was no chromosome breakage. In some cells the FISH signals of the highly decondensed 18S rDNA were not picked up by the camera, but they were always visible under the microscope.

Chromosome No.	%LR ± SEa	Molecular size of chromosome (Mbp)b	IC ± SEc	Classification of chromosomed
1	17.93 ± 0.30	470.30	46.92 ± 0.45	m
2	16.39 ± 0.28	429.90	46.60 ± 0.53	m
3	15.50 ± 0.24	406.56	36.44 ± 0.96	sm
4	14.07 ± 0.20	369.03	47.68 ± 0.28	m
5	12.94 ± 0.26	339.41	38.21 ± 0.56	m
6	12.30 ± 0.14	322.62	36.60 ± 0.78	sm
7	10.83 ± 0.16	284.07	34.93 ± 0.81	sm

^a%L^R (%relative length) = Length of a chromosome/Total haploid chromosome length x 100

^bBased on reported genome size estimation [18]

^cI^C(centromeric index) = Length of short arm/Chromosome length x 100

^dAs previously reported [2]

The seven chromosomes of the haploid L. perenne plant displayed identical Q-banding patterns to those of the seven pairs of the diploid plant. However, instead of three marker chromosomes, four chromosomes displayed 18S rDNA hybridization (Fig. 1) with an additional NOR being detected proximally on the long arm of chromosome 1 (Fig. 2). Regional hybridization of the 18S and/or 5S rDNA sequences on chromosomes 2, 3 and 5 was identical to that observed in the diploid plant.

All seven chromosomes from each cell of the haploid plant were identified and karyotyped as for the diploid cells. An ideogram was drawn based on the data provided in Table 1. The Q-banding patterns for all seven chromosomes were analysed from cells selected from both the diploid and haploid plants. Landmark bands were identified, and the banding patterns and their nomenclature were diagrammatically represented in the banded ideogram with 156 bands (Fig. 3). This enabled us to assign the 5S and 18S rDNA loci in L. perenne to a specific chromosome, chromosome arm and band position. The 18S rDNA sequences have been assigned to chromosome positions 1q21, 2p21, 3q21 and 5q21 while 5S rDNA was on 3p21 (Fig. 2).

To test if DAPI-banding on the flame-dried chromosome preparations could also be produced while maintaining compatibility with the FISH procedure, a modified protocol was developed to map a circadian clock-associated gene of L. perenne, GIGANTEA (LpGI), using genomic as well as cDNA sequences as probes. Although a defined DAPI differential staining was not observed on mid-metaphase chromosomes, low levels of Q-like DAPI-banding were generated in pro-metaphase and early metaphase chromosomes. Specific and identical hybridization of both the probes was observed on a chromosome arm with 18S rDNA. Using the combined banding and marker patterns, LpGI was unambiguously assigned to 2p31 (Fig. 4).

A simple fluorescence based Q-banding technique using quinacrine mustard for plant chromosomes, which is compatible with sequential FISH mapping of cloned DNA sequences, is reported here for the first time. This technique enabled the visualization of chromosome banding patterns and FISH signals in the same cell. The longitudinal differentiation of Q-bands produced along the length of each chromosome facilitated unambiguous identification of perennial ryegrass chromosomes. Based on these results, a banded karyotype and ideogram of perennial ryegrass was developed and used as a cytogenomic tool. Based on only two individuals from different cultivars, this represents a first step, and analyses of more populations is needed before a standardized karyotype can be established.

Success was also achieved in producing DAPI-banding patterns while maintaining compatibility with FISH mapping of a single copy sequence on the same cytological preparation. While the DAPI-bands could be visualised only in pro-metaphase and early metaphase cells (Fig. 4), the banding patterns were similar to Q-banding as would be expected because both QM and DAPI are AT-specific fluorochromes [27]. Although less defined than Q-bands, the DAPI bands were sufficiently clear to identify commonality. DAPI-bands were not observed by Rocha et al. [25] who restricted their observations to metaphase cells of L. perenne. Cytochemical pre-treatments for the production of bands on fixed chromosomes may cause minor destruction or loss of chromosomal DNA [15, 28]. Such preparations may then be incompatible with FISH mapping, especially with low-copy targets. Because the Q- or DAPI-banding techniques presented here did not involve any specific pre-treatment before staining with QM or DAPI, they were compatible with FISH mapping.

Technical limitations in chromosome preparation have previously hampered progress in developing plant chromosome banding techniques. The traditional squash method causes adherence of cytoplasmic debris from clumped meristematic tissue and may have been the main barrier to generating resolvable bands on plant chromosomes [9, 29]. The flame-drying method reported previously [26], and adopted in this work, did not involve the traditional squashing of tissue. Instead, the enzymatically macerated meristematic tissue was teased into a droplet of cell suspension, given a temperature shock with chilled fixative and then flame-dried. This left the chromosomes well-spread and completely clean on the glass slide without any cytoplasmic debris. An additional advantage of the well-spread and debris-free chromosome preparation has been the enhanced sensitivity of FISH signals [30, 31]. Here, we have successfully mapped a small target sequence, LpGI, onto a specific chromosome band of L. perenne using both genomic DNA and cDNA probes.

Variability in the number of NOR loci in L. perenne has been described in earlier reports [25, 32–34]. Six NOR loci have been common and an additional NOR FISH signal has been recorded in a few reports [25, 33, 34], but eight NOR loci have not been encountered. The chromosomes carrying these sequences were either not numbered or were inconsistently numbered. In the present work, the three pairs of chromosomes in the diploid L. perenne plant consistently carrying NOR loci were designated according to the proposed karyotype. The NOR (18S rDNA) loci and the single 5S rDNA locus were then regionally mapped and assigned to cytogenetic bands. Identical assignments were made in the haploid L. perenne but with the addition of a fourth 18S rDNA signal on the largest chromosome. This was consistent with the previous observation [25] that a hemizygous 18S rDNA locus occurred on the largest chromosome.

Seven pairs of biarmed chromosomes of L. perenne, were numbered according to decreasing size order in the karyotype presented here. A combination of gross chromosome morphology, Q-banding patterns and molecular cytogenetic markers (two rDNA sequences) has helped in individualizing all seven pairs for the first time (Fig. 2). An estimation of the molecular size of each chromosome has also been presented, based on the nuclear genome size determined earlier by flow cytometry [18]. The chromosome designations given here (Table 1) are similar, but not identical, to those assigned by Rocha et al. [25]. Chromosomes 1–5 are probably the same in both classifications, but the potential confusion of chromosomes 6 and 7, which are of similar morphology and size, can now be resolved by Q-bands.

L. perenne exhibits extensive conserved synteny with the Triticeae and other members of subfamily Pooideae [21, 35]. Various gene maps of L. perenne have been developed through comparative mapping, mainly with Triticeae [20–22]. The numbering system of linkage groups (LGs) in these maps have directly been adapted from Triticeae [21, 22, 36, 37]. Due to the lack of a standardized karyotype of L. perenne, no correlation between LGs and individual chromosomes has been established so far through FISH mapping. Neither the analysis of alien substitution lines [38] nor alien introgression lines [39] has provided unambiguous identification of all L. perenne chromosomes. Indeed, as noted by Kopecky and Studer [1], only one of the seven chromosomes had so far been discriminated. GIGANTEA (LpGI) has been mapped to LG3 of L. perenne [40]. Here, this single copy gene has been precisely assigned through FISH mapping to a cytogenetic band on the short arm of chromosome 2 (Fig. 4). LG3 has been intensively analyzed by introgression mapping and the presence of an NOR as noted previously [41], was also found in the present work. Thus Q-banding combined with FISH mapping has enabled LG3 to be identified as chromosome 2 and described by its size, banding pattern and morphology. The accumulation of many such regional assignments should lead to the development of an integrative cytogenetic map. Such an integrative cytogenetic map will provide an alternative approach for analyses of recombination (cM) along the chromosome [39] and reveal the extent of unmapped genomic regions. It could also help to resolve contradictory chromosome assignments of markers [21] and LG map discrepancies [22]. Chromosomal mapping will also allow the placement of markers which cannot be genetically mapped [42].

To facilitate DNA sequencing of large grass genomes, individual chromosomes can be isolated by flow cytometry [31, 43, 44]. DNA preparations from flow-sorted chromosomes can then be used for the generation of chromosome-specific sequences and assemblies [45, 46]. Application of the chromosome banding technique presented here should provide unambiguous identification of flow sorted chromosomes and help to reveal contamination. Flow-sorted chromosomes can also be identified by FISH if appropriate markers are available [18]. However, the chromosome banding method is faster, cost effective and will also reveal the intact status of the sorted chromosomes.

Introgression of alien chromosome or chromosome segments have produced cytogenetic stocks of addition or substitution lines in a number of Pooideae, including Festulolium hybrids [38, 47, 48]. Application of the Q-banding technique in combination with sequential FISH and/or genomic in situ hybridization (GISH) should provide identification of specific chromosomes and, by extension, the nature and extent of introgression.

Transgenesis is often used by agricultural biotechnologists to introduce traits of agronomic interest into crops [49–54] but instability in the expression of the introduced transgene is common [55–57]. The transgenes may unpredictably integrate at different locations in the host genome and be subject to “position effect” variation in expression patterns [51, 58]. The chromosome banding technique developed here with the concomitant visualization of the integrated transgene through FISH should reveal the number of insertion events and the hemi- or homozygosity of each insertion. Further, the precise cytogenetic determination of transgene integration site(s) may throw light on the relationship between transgene expression pattern and the location of integration. Additionally, it should detect tissue culture-mediated chromosomal abnormalities which frequently occur in transformed plants [59, 60].

The Q-banding method described here has enabled the unambiguous identification of all seven chromosomes (and all 14 chromosome arms) of L. perenne. Further, the technique has been demonstrated to be compatible with FISH mapping, enabling the anchoring of a locus mapped by linkage analysis to a specific chromosome position on an identifiable chromosome arm. Similar work should quickly enable the anchoring of many more loci currently located by linkage analysis to specific chromosome positions, thus integrating cytogenetics and genetics and enabling LGs to be associated with identifiable chromosomes. Extension of the method to the close relatives of L. perenne is likely to greatly facilitate the development of addition/substitution lines and, by extension, the integration of introgression bins with known chromosome positions. The work thus provides an integrative cytogenomic resource that will contribute to a better understanding of genomic structures and functions in grasses.