Painting of Parental Chromatin in Beta Hybrids by Multi‐colour Fluorescent in situ Hybridization

Abstract

Sugar beet (Beta vulgaris L.) is a relatively young crop and has a narrow gene pool. In order to introduce genetic variability into the crop, interspecific hybrids, selected from crosses with wild beets of the sections Corollinae and Procumbentes, have been generated. The introgressed B. procumbens chromatin carries resistance genes to beet cyst nematode Heterodera schachtii Schm. These lines are important for breeding of nematode‐resistant sugar beet, while Corollinae species are potential donors of tolerance to biotic and abiotic stresses such as drought or saline soils. We have used in situ hybridization of genomic DNA to discriminate the parental chromosomes in these interspecific hybrids. Suppression of cross‐hybridization by blocking DNA was not necessary indicating that the investigated Beta genomes contain sufficient species‐specific DNA enabling the unequivocal determination of the genomic composition of the hybrids. Interspecific hybrid lines with an additional chromosome (2n = 18 + 1), chromosome fragment (2n = 18 + fragment) or translocation of B. procumbens (2n = 18) were analysed by genomic in situ hybridization (GISH) at mitosis and meiosis. Species‐specific satellites and ribosomal genes used in combination with genomic DNA or in rehybridization experiments served as landmark probes for chromosome identification in hybrid genomes. The detection of a B. procumbens translocation of approx. 1 Mbp demonstrated the sensitivity and resolution of GISH and showed that this approach is a powerful method in genome analysis projects of the genus Beta.

INTRODUCTION

Cultivated beet (Beta vulgaris L.) belongs to the genus Beta, which is divided into the sections Beta, Corollinae, Nanae and Procumbentes. While different beet cultivars such as beet root, mangold and fodder beet have been used as vegetables or for animal feeding for a long time, sugar beet is a relatively young crop. Present day sugar beets can be traced back to the late 18th century and have been selected from a few crosses between mangold and fodder beet (Fischer, 1989). Hence, sugar beet has a narrow genetic basis, and wild Beta species, in particular species of the sections Procumbentes and Corollinae, are valuable genetic resources to broaden the gene pool and introduce genetic diversity (Van Geytet al., 1990).

For example, resistance to the beet cyst nematode (Heterodera schachtii Schm.) has been transferred to B. vulgaris from the section Procumbentes. Interspecific hybrids between B. vulgaris and B. procumbens include nematode‐resistant sugar beet lines containing an alien wild beet chromosome (Savitsky, 1975) or chromosome fragment (De Jonget al., 1986; Brandeset al., 1987) as a monosome. Among the offspring of these monosomic addition lines, diploid resistant sugar beets carrying wild beet translocations have been selected (Jung and Wricke, 1987; Heijbroeket al., 1988). The translocation A906001 carries a very small wild beet introgression and has been used for the positional cloning of the nematode resistance gene Hs1^pro–1 (Caiet al., 1997).

Species of the section Corollinae, in particular B. corolliflora, are of interest because of their tolerance to low temperatures, drought and saline soils as well as their resistance against viruses and fungi. Furthermore, apomictic reproduction has been found in most polyploid Corollinae species (Barocka, 1966), which, if transferred to cultivated beet, has the potential to improve the fixation of desired genotypes in breeding programmes.

Species of the section Corollinae are relatively closely related to B. vulgaris as revealed by analysis of chloroplast DNA (Fritzscheet al., 1987). Interspecific hybrids have been described by Van Geytet al. (1990). Recently, Liuet al. (1996) reported crosses between the diploid B. vulgaris (2n = 2x = 18) and the tetraploid B. corolliflora (2n = 4x = 36). Offspring lines are interspecific hybrids, including amphitriploids and monosomic additions containing either a haploid set or single chromosome of B. corolliflora added to the B. vulgaris complement. Some of these hybrid lines are tolerant to the pathogenic fungus Cercospora beticola, which causes beet leaf spot disease and considerable loss in sugar beet production.

The selection of B. vulgaris hybrids is time‐consuming and difficult, while resistance tests often lead to ambiguous results. Cytogenetic analyses in Beta species are hampered by the small size and similar morphology of the chromosomes. Neither the identification of alien chromosome(s) nor the detection of translocations in hybrids is possible by classical cytogenetics such as staining or banding techniques. Rapid screening approaches using species‐specific repetitive sequences like satellite DNAs in squash–dot or dot–blot hybridization have been described in the genus Beta (Schmidtet al., 1990). These analyses are based on molecular markers and are suitable for fast selection of a large number of plants. However, they give no information about the chromosomal composition of the hybrid genomes.

In contrast, fluorescent in situ hybridization (FISH) to metaphase chromosomes or interphase nuclei enables the unequivocal determination of the genomic constitution of hybrid plants. In particular, genomic in situ hybridization (GISH) is a sensitive technique to paint and identify genomes at the chromosomal level using genomic DNA as a labelled probe. The specificity is essentially based on the amount and divergence of repeated sequences, which account for the majority of the DNA in plant nuclear genomes.

GISH has been applied for the identification of parental genomes in natural polyploids such as Triticum aestivum, Milium montianum and other plant species (Bennettet al., 1992; Mukaiet al., 1993; Fukuiet al., 1997) and in ornamental plants (Ørgaardet al., 1995; Kuiperset al., 1997). The detection of alien chromatin in interspecific crop hybrids has been described and, at higher resolution, chromosomal rearrangements such as translocations of chromosome arms or segments are detectable (Heslop‐Harrisonet al., 1990; Schwarzacheret al., 1992; Kosina and Heslop‐Harrison, 1996). Most of these species have relatively large chromosomes. Nevertheless, GISH analyses of Nicotiana and sugarcane cultivars have shown that the small chromosomes of these species originate from different ancestral genomes, which can clearly be discriminated (Lim et al., 2000; Piperidiset al., 2000).

B. vulgaris has a relatively small genome of 758 Mbp (Arumuganathan and Earle, 1991) and has been used as model species to study the large‐scale organization of the major classes of repetitive DNA of plant genomes. Genomes of Beta species are characterized by many satellite DNA families and contain different Ty1‐copia‐ and Ty3‐gypsy‐like retrotransposons and non‐LTR‐retrotransposons such as LINEs in high copy numbers (Schmidtet al., 1995; Kubiset al., 1998). These repeat families have been characterized extensively at the molecular level and by high‐resolution FISH, resulting in a structural model of a plant chromosome (Schmidt and Heslop‐Harrison, 1998).

In this paper, we report the successful application of GISH to discriminate parental chromosomes in interspecific hybrids between B. vulgaris and either B. procumbens or B. corolliflora at mitosis and meiosis. We demonstrate that tandemly repeated DNA sequences such as species‐specific satellites and ribosomal genes can serve as landmark probes for chromosome identification in hybrid genomes. Using genomic DNA, we were able to detect unequivocally a B. procumbens translocation of approx. 1 Mbp in B. vulgaris, thus demonstrating the resolution potential of GISH in the genus Beta.

MATERIALS AND METHODS

Plant material

Plants were grown under greenhouse conditions. Seeds of the wild beets B. corolliflora and B. procumbens from collections in natural habitats were obtained from L. Frese (Institut für Pflanzenbau der Bundesforschungsanstalt für Landwirtschaft Braunschweig‐Völkenrode; Germany).

Interspecific hybrids between diploid B. vulgaris (2n = 2x = 18) and tetraploid B. corolliflora (2n = 4x = 36) were established at the Department of Biotechnology, Heilongjiang University Harbin (China). Among their offspring, an amphitriploid hybrid (2n = 3x = 18 + 9 = 27; accession 970009/4) was selected because of the apomictic phenotype (Liuet al., 1996). Crossing of this amphitriploid hybrid with B. vulgaris resulted in aneuploid progenies. In particular, monosomic addition lines (2n = 2x = 18 + 1; accession 970006/61; 970026/33) were chosen and kindly provided by Dongjie Gao (Christian‐Albrechts‐University of Kiel, Germany).

A series of nematode‐resistant B. vulgaris (Junget al., 1992) including a monosomic addition line (2n = 18 + 1; accession 950039), the chromosome fragment addition PRO1 (2n = 18 + fragment; accession 930145) and a translocation (2n = 18; accession 940045) was investigated in this study. These resistant lines contain alien chromatin from B. procumbens.

Preparation of mitotic and meiotic chromosomes

The meristem of young leaves (5–10 mm in length) of vigorously growing plants was used for the preparation of mitotic chromosomes. Before fixation in methanol/acetic acid (3 : 1) leaves were incubated for 3–5 h in 2 mm 8‐hydroxyquinoline. Preparation of mitotic and meiotic chromosomes were carried out as described by Deselet al. (2001). Briefly, leaves were washed in enzyme buffer (4 mm citric acid/6 mm sodium citrate, pH 4·8) and digested in a mixture of cell wall‐degrading enzymes consisting of 1·8 % (w/v) cellulase from Aspergillus niger, 0·2 % (w/v) cellulase Onozuka‐R10 and 20 % (v/v) pectinase from A. niger in enzyme buffer. The resulting protoplast suspension was washed three times with enzyme buffer and twice with fresh fixative. Protoplasts were collected by centrifugation (3000 g, 4 min) and the pellet was resuspended in 100 µl of the methanol/acetic acid mixture. An appropriate volume (10–15 µl) of the cell suspension was applied to acid‐cleaned slides and spread.

Young flower buds were used for the preparation of meiotic chromosomes. The material was equilibrated in enzyme buffer for 1 h and incubated in an enzyme mixture consisting of 0·3 % (w/v) cytohelicase, 1·8 % (w/v) cellulase from A. niger, 0·2 % (w/v) cellulase Onozuka‐R10 and 20 % (v/v) pectinase from A. niger in enzyme buffer at 4 °C overnight. After several washes in water, a single anther was dissected carefully and placed on an acid‐cleaned slide. The tissue was teased apart in 60 % acetic acid and the pollen mother cells were incubated for 4 min until the cytoplasm was transparent. Chromosomes were spread by addition of cold fixative. The slides were immersed in absolute ethanol and air‐dried.

DNA probes and labelling

Genomic DNA from B. vulgaris, B. procumbens and B. corolliflora was prepared as described by Junghans and Metzlaff (1988). Genomic DNA and the 18S‐5·8S‐25S rRNA gene probe pTa71 (Gerlach and Bedbrook, 1979) were labelled with digoxigenin‐11‐dUTP (Roche Diagnostics, Germany) or biotin‐dUTP (Roche Diagnostics) by nick translation as recommended by the supplier.

The genome‐specific satellite pBV1, localized exclusively in the pericentromeric regions of all B. vulgaris chromosomes, and pXV1 containing the B. vulgaris 5S rRNA gene with spacer (Schmidtet al., 1991, 1994) were included. Both probes were labelled by PCR in a 50 µl reaction volume in the presence of biotin‐11‐dUTP or digoxigenin‐dUTP using the M13 universal primers. PCR products were precipitated in ethanol and dissolved in 20 µl sterile water.

FISH

FISH was performed as described by Schmidtet al. (1994) and Deselet al. (2001) with some modifications.

After pretreatment with RNAseA, pepsin and paraformaldehyde, chromosome preparations were covered with 30 µl predenatured hybridization solution containing 100–150 ng of labelled genomic DNA, 50 % (v/v) formamide; 2 × SSC (300 mm sodium chloride, 30 mm trisodium citrate); 20 % (w/v) dextran sulfate; 0·15 % (w/v) sodium dodecyl sulfate (SDS) and 250 ng µl^–1 salmon‐sperm DNA. The slides were placed in an in situ thermocycler (Hybaid) and denaturated at 70 °C for 8 min. The temperature was gradually reduced to 37 °C, and the chromosome spreads were incubated for 14–16 h in a humid chamber. After hybridization, the slides were washed at a stringency of 76 % in 20 % (v/v) formamide/0·2 × SSC at 42 °C. Probes labelled with digoxigenin or biotin were detected with 2·6 µg ml^–1 anti‐digoxigenin‐FITC (fluorescein isothiocyanate) or 5·6 µg ml^–1 streptavidin conjugated to Cy3, respectively. Antibody reaction was performed at 37 °C in 3 % bovine serum albumin (BSA)/4 × SSC/0·2 % Tween20 for 1 h, and chromosomes were counterstained with DAPI (4′,6‐diamidine‐2‐phenylindole) after removal of excess antibody.

Rehybridization of chromosomes was performed as described by Heslop‐Harrisonet al. (1992). Chromosome preparations were washed several times for 60 min in 0·1 % Tween20/4 × SSC and once in 2 × SSC. In order to stabilize chromosomes, slides were immersed in 4 % paraformaldehyde for 10 min and washed in 2 × SSC before re‐probing.

Examination of slides was performed with a Zeiss Axioplan2 Imaging fluorescence microscope equipped with a Filter 09 (FITC), Filter 15 (Cy3), Filter 01 (DAPI) and Filter 25 (triple‐band‐pass). Photographs were taken on Fujicolor SUPERIA 400 colour print film. Film negatives were digitized on a Nikon LS‐1000 scanner, contrast‐optimized using only functions affecting the whole image equally, and printed using Adobe Photoshop software.

RESULTS

Discrimination of Beta vulgaris and Beta corolliflora genomes in interspecific hybrids

Amphitriploid plants (2n = 3x = 18 + 9 = 27) containing a full complement of B. vulgaris and a haploid genome of B. corolliflora are an interesting material to study the evolutionary divergence of repetitive sequences within related plant species. In particular, satellite DNA is rapidly evolving both in copy number and sequence and, if species‐specific, can be used to monitor the chromosomal composition of interspecific hybrid genomes.

In interspecific Beta hybrids, parental genomes were discriminated by hybridization of genomic DNA and the sugar beet‐specific satellite DNA pBV1. This satellite repeat has been isolated from B. vulgaris and belongs to a BamHI sequence family that extends over large regions up to 3000 kb at the centromere of all sugar beet chromosomes (Schmidtet al., 1991; Kubiset al., 1998). The satellite pBV1 labelled all sugar beet chromosomes in the amphiploid hybrid (Fig. 1A, right), while the nine wild beet chromosomes could be identified by hybridization of genomic DNA of B. corolliflora (Fig. 1B, left). In interphases, the majority of B. vulgaris centromeres detectable by strong pBV1 signals tend to cluster in a loose Rabl configuration (Fig. 1C, right, red signals).

Fig. 1. Discrimination of chromosomes originating from B. vulgaris or B. corolliflora in the genome of (A–F), an amphitriploid hybrid line (2n = 3 x = 18 + 9 = 27) and (H–J), a monosomic addition line (2n = 18 + 1 = 19) by GISH and FISH. Probes were labelled with biotin‐dUTP or digoxigenin‐dUTP and detected by streptavidin‐cy3 (red fluorescence) or FITC‐labelled antibodies (green fluorescence), respectively. Blue fluorescence in all panels shows chromosomes counterstained with DAPI. Arrowheads mark hybridization signals resulting from cross‐hybridization to conserved 18S‐5·8S‐25S rRNA genes. A–C, Eighteen sugar beet chromosomes were identified by in situ hybridization of the pericentromeric, B. vulgaris‐specific BamHI satellite pBV1. B, Nine wild beet chromosomes were labelled green after simultaneous hybridization of genomic DNA of B. corolliflora. A digital overlay (B, right) confirms the results of the double‐target hybridization. At interphase C, B. vulgaris centromeres tend to cluster in one domain of the nucleus. D–F, Multi‐colour FISH with differently labelled genomic DNA of both B. corolliflora and B. vulgaris. Chromosomes of B. corolliflora were labelled red while chromosomes of B. vulgaris were detected by green fluorescence. After rehybridization (E, right) of rDNA probes, the B. vulgaris chromosome I with the terminal loci for 18S‐5·8S‐25S rRNA genes (arrowheads) and chromosome IV carrying the 5S rRNA genes were identified. Homoeologous chromosomes (designated I^C and IV^C) from B. corolliflora were identified by comparison with the GISH results. The hybridization at interphase (F) shows that parental genomes are located in discrete domains. G–H, FISH of pBV1 enabled the discrimination and selection of (G) trisomic plants with 19 sugar beet chromosomes and (H, centre) a monosomic addition line (18 + 1). H (right), Genomic DNA of B. corolliflora hybridized strongly with the alien wild beet chromosome. Chromosomes of B. vulgaris showed weak dispersed cross‐hybridization with signal depletion in centromeric regions (arrow). I, Hybridization of genomic DNA of B. vulgaris to less‐condensed chromosomes produced strong signals around the centromeres. Arrow shows the alien chromosome without hybridization signal. J (left), Double‐target in situ hybridization of differently labelled genomic DNA of the parental genomes. J (right), Identification of the alien chromosome as chromosome IV^C of B. corolliflora after rehybridization of the 5S rDNA probe (green signals) and 18S‐5·8S‐25S rDNA repeats (red signals) used as probes.

Identification of sugar beet and wild beet chromosomes was also possible by simultaneous hybridization of differently labelled genomic DNA of both Beta species. Hybridization with DNA of B. vulgaris resulted in green fluorescence signals on 18 sugar beet chromosomes (Fig. 1D, right). The signals were mainly restricted to the centromeric regions where large satellite arrays of pBV1 reside. In contrast, uniform painting of nine wild beet chromosomes was observed after hybridization with genomic DNA of B. corolliflora (Fig. 1E, left, red signals) indicating no regions of sequence depletion or amplification.

At interphase, GISH with genomic DNA of B. vulgaris and B. corolliflora demonstrated that chromosomes of the parental genomes are located in discrete domains of the cell nucleus (Fig. 1F, right). Two yellow signals, resulting from cross‐hybridizing 18S‐5·8S.25S rRNA genes are visible (Fig. 1F, arrowheads).

The amphitriploid B. vulgaris × B. corolliflora hybrids serve as starting material to establish monosomic addition lines, which are useful for the introgression of genes carrying valuable agronomical traits such as pathogene resistance or stress tolerance into cultivated beet. To stimulate recombination of parental chromosomes during chromosome segregation, aneuploid, in particular trisomic B. vulgaris lines (2n = 19), have been backcrossed with the amphitriploid hybrids.

Among the progenies, plants with different genomic constitutions including monosomic addition lines (2n = 18 + 1), as well as trisomic B. vulgaris plants (2n = 19), were expected. Because of its species‐specificity and exclusive occurence on all B. vulgaris chromosomes, the satellite monomer pBV1 was used to differentiate aneuploid B. vulgaris lines from interspecific hybrids. Trisomic plants can be clearly selected by FISH with pBV1: all chromosomes of the complement are labelled at the centromere showing their B. vulgaris origin (Fig. 1G). These plants were excluded from further analysis.

Signals on the 18 B. vulgaris chromosomes are detectable in monosomic addition lines, while the single alien B. corolliflora chromosome showed no signal with pBV1 (Fig. 1H, red, centre) but strong hybridization with differently labelled genomic DNA of B. corolliflora, thus verifying the hybrid character of this plant (Fig. 1H, green, right).

Only weak and randomly dispersed hybridization of B. corolliflora DNA to B. vulgaris chromosomes was observed; therefore the use of unlabelled blocking DNA was not necessary. However, it is noteworthy that depletion of the weak signal was observed in the centromeric regions of most sugar beet chromosomes, which consist largely of pBV1 arrays (examples arrowed in Fig. 1H). Consistently, GISH with labelled B. vulgaris DNA to less‐condensed metaphase chromosomes revealed strong signals around the centromeres of the sugar beet chromosomes (Fig. 1I). Hybridization signals correspond largely to DAPI positive regions, while the distal euchromatin is only weakly labelled. No hybridization of the wild beet chromosome was observed (Fig. 1I, arrow).

In general, GISH with genomic DNA showed that strongly cross‐hybridizing sites are confined to the chromosomal regions consisting of conserved 18S‐5·8S‐25S ribosomal genes. Two additional signals in the distal region of one pair of B. vulgaris chromosomes of the monosomic addition line were detected after hybridization with B. corolliflora DNA (Fig. 1J, left, red signals), while three ribosomal sites were observed in the amphitriploid hybrid (arrowheads in Fig. 1E, left).

Rehybridization of both metaphases with pTa71—containing the heterologous rRNA genes from Triticum aestivum—revealed terminal signals on a pair of sugar beet chromosomes (Fig. 1J, right, red signals) and three rDNA sites in the amphidiploid hybrid (Fig. 1E, right, red signals). Rehybridization also showed that the tandem arrays of the 5S rRNA genes are detectable by pXV1 in pericentromeric regions of a pair of sugar beet chromosomes and two intercalary loci of the alien B. corolliflora chromosome (Fig. 1E and J, right, yellow‐green signal).

Beta procumbens chromatin in nematode‐resistant sugar beet

A source of resistance to the beet cyst nematode H. schachtii, which causes damage in sugar beet cultivation, is the wild beet B. procumbens. This wild beet species has been used to establish nematode‐resistant sugar beet hybrids including monosomic addition (2n = 18 + 1), fragment addition (2n = 18 + fragment) and translocation lines (2n = 18). These different lines were investigated by FISH using genome‐specific satellites and genomic DNA as probes aiming to detect the alien B. procumbens chromatin.

In monosomic additions, all 18 B. vulgaris chromosomes were labelled at pericentromeric regions by the sugar beet‐specific satellite pBV1 (Fig. 2A, green signals). No hybridization signal was visible on the alien B. procumbens chromosome. Similarly, strong signals corresponding to the sugar beet chromosomes were observed after probing monosomic addition lines with genomic DNA of B. vulgaris (Fig. 2B, right, red signals). The fluorescence signals were mainly confined to the centromeric heterochromatic regions of the sugar beet chromosomes, which consist largely of pBV1 satellite arrays; only little hybridization was observed in the terminal euchromatic regions, which are only weakly stainable with DAPI. The alien wild beet chromosome was unequivocally detectable by simultaneous hybridization with differently labelled genomic DNA of B. procumbens, (arrow in Fig. 2A red, 2B green).

Fig. 2. Detection of B. procumbens chromatin in nematode‐resistant sugar beet hybrids. Chromosome preparations of (A–E) monosomic addition line of B. vulgaris with a wild beet chromosome of B. procumbens, (F–G) a fragment addition line of B. vulgaris with a fragment of B. procumbens (PRO1) and (H–L) a homozygous translocation line of B. vulgaris with introgressed chromatin of B. procumbens (A906001) were analysed. Arrows in all panels indicate specific hybridization of genomic DNA of B. procumbens, while arrowheads indicated cross‐hybridization signals to 18S‐5·8S‐25S rRNA genes. Preparations were counterstained with DAPI (blue fluorescence) to visualize chromosome morphology. A, Discrimination of the chromosomes from B. vulgaris or B. procumbens by simultaneous hybridization of the pericentromeric, B. vulgaris‐specific BamHI satellite pBV1 (green) and genomic DNA of B. procumbens (red). B–C, Hybridization with genomic DNA from B. vulgaris (red) and B. procumbens (green) enabled the alien chromosome to be detected at metaphase (B) and interphase (C). Cross‐hybridization to rRNA genes resulted in orange signals. D–E, During meiosis, the alien chromosome is detectable by GISH. The B. procumbens chromosome is labelled over its entire length at pachytene (D) and remains as a univalent at diakinesis (E). F, The chromosome fragment of B. procumbens could be identified by GISH in the genome of the interspecific hybrid line PRO1. F–I (right), Digital overlay of the green and blue image show the position of the wild beet fragment or translocation (G). The acrocentric fragment was detected in the genome of PRO1 by hybridization of the centromeric, wild beet‐specific Sau 3A satellite pTS5 at pachytene. (H–L), The wild beet translocation of 1 Mb was localized at the physical end of one pair of chromosomes at mitotic metaphase (H), pachytene (I), zygotene (J–K) and early leptotene (L). Simultaneous hybridization of genomic DNA of B. vulgaris (J–L, red) supports the detection of the weak wild beet‐specific signals. K, Close‐up view revealed an orange signal adjacent to the wild beet translocation, which is presumably caused by cross‐hybridization of terminal conserved sequences.

At interphase, the region occupied by the B. procumbens chromosome is visible in most nuclei (example shown Fig. 2C, arrow). The green wild beet‐specific signal is readily discernible from the orange fluorescence of cross‐hybridization of the conserved 18S‐5·8S‐25S rRNA genes, which were also observed at metaphase in terminal position on one pair of chromosomes (arrowhead in Fig. 2A and B).

The alien B. procumbens chromosome was also studied at meiotic stages. During pachytene, the chromosomes are less condensed and much longer than in mitotic metaphases. They show specific patterns of DAPI‐positive chromomeres, which are mostly clustered around the centromeres. However, the alien wild beet chromosome cannot be differentiated from the completely synapsed homologues of the sugar beet complement. Hybridization with genomic DNA of B. procumbens clearly revealed the position of the alien wild beet chromosome, which was labelled over its entire length (Fig. 2D, right). At diakinesis, the sugar beet chromosomes form ring bivalents and rod bivalents. The B. procumbens chromosome remains as a univalent and is labelled by genomic DNA (Fig. 2E, arrow). Like the FISH investigation of mitotic chromosomes, B. procumbens DNA as probe was sufficient to discriminate the alien chromosome in pachytene and diakinesis, and unlabelled genomic DNA of B. vulgaris to suppress unspecific hybridization was not necessary.

One of the resistant sugar beet hybrids investigated is the fragment addition line PRO1 (accession 930145), which contains a B. procumbens chromosome fragment of approx. 6–9 Mbp. The fragment forms a mitotically stable minichromosome in the B. vulgaris complement. However, in some cells, the fragment appeared to be attached to a B. vulgaris chromosome; careful inspection of DAPI‐stained metaphases can therefore be useful for the selection of fragment addition lines (unpubl. res.).

The wild beet fragment is detectable after hybridization with B. procumbens DNA (Fig. 2F). Although cross‐hybridization of sugar beet 18S‐5·8S‐25S rRNA genes was observed after probing with labelled wild beet DNA, the morphology of the alien fragment enables unequivocal assignment of the fluorescence signals during metaphase. Differentiation is not possible at interphase of PRO1, where three hybridization sites were observed, although the juxtaposition of two strong signals indicated the putative rDNA sites (Fig. 2F, top). A positive identification of the PRO1 chromosome fragment is achieved by using the B. procumbens satellite repeat pTS5 as a probe. This satellite family is specific for the genomes of Procumbentes wild beets and showed only hybridization with the fragment (Fig. 2G).

In order to explore the sensitivity and resolution of GISH, mitotic and meiotic chromosomes of a B. vulgaris line carrying a small wild beet translocation were investigated. This B. procumbens translocation has a size of approx. 1 Mbp and confers full nematode resistance to sugar beet.

The wild beet chromatin was detected by GISH at the physical end of two chromosomes (Fig. 2H, arrows). The terminal position of the introgressed B. procumbens chromatin was evident from GISH to pachytene chromosomes, showing that the translocation occupies the distal region. Since homologous chromosomes are completely paired at pachytene, only a single translocation‐specific signal is visible (Fig. 2I, arrow). Strong hybridization of the paired rDNA sites was observed (Fig. 2I, arrowhead).

GISH with differently labelled DNA from both Beta species resulted in strong hybridization to the B. vulgaris chromosomes (Fig. 2J, red), while the B. procumbens translocation was visualized as green decondensed chromatin (Fig. 2J, green signal). The close‐up view revealed an orange signal adjacent to the wild beet translocation, which resulted from hybridization of the differently labelled genomic DNAs and indicates genomic sequences conserved between B. vulgaris and B. procumbens (Fig. 2K).

In meiotic interphases, the nucleolus is clearly detectable and contains 18S‐5·8S‐25S rRNA genes of both Beta species as observed by green (B. procumbens) and red (B. vulgaris) signals (Fig. 2L). Minor green signals outside the nucleolus indicate the position of the B. procumbens translocation (Fig. 2L, arrow).

DISCUSSION

Painting of Beta chromosomes

Interspecific hybridization including genome rearrangements or polyploidization is an important process in the evolution of plant species, but also a source for the creation of genetic variability for crop improvement.

Crosses of cultivated beet with wild beets of the sections Corollinae and Procumbentes have been performed to introduce pest resistance or tolerance into the crop (Van Geytet al., 1990). Furthermore, during the last 10 years, cultivated beet has increasingly become a suitable model species for molecular–cytogenetic analyses of the structure, organization and evolution of plant genomes (Kubiset al., 1998; Schmidt and Heslop‐Harrison, 1998). Large insert libraries have been constructed and, in particular, a recently established BAC (bacterial artificial chromosome) library with deep genome coverage is valuable for the isolation of genes and repeats, and for the structural analysis of chromosome domains such as the centromere (Gindulliset al., 2001a). Despite this, no comprehensive GISH study of the sugar beet genome and its wild relatives has yet been performed.

GISH using genomic DNA of B. vulgaris and either B. corolliflora or B. procumbens as probes enabled the clear discrimination of the three genomes in various hybrids of B. vulgaris. Blocking DNA was unnecessary for the painting of alien chromosomes or genomes in interspecific Beta hybrids. Usually, blocking DNA is used in excess ratios, ranging in final concentrations from 25‐fold up to 100‐fold the amount of that of the probe. For example, closely related ornamental species of the genus Alstroemeria contain large chromosomes and have some of the largest genomes in the plant kingdom; here, parental genomes were discriminated in interspecific hybrids at probe‐block ratios of 1 : 100 to 1 : 200 (Kuiperset al., 1997). In contrast, parental genomes of interspecific or intergeneric cereal hybrids and widely grown Crocus cultivars were differentiated by lower amounts of blocking DNA, although these species have also relatively large chromosomes (Schwarzacheret al., 1989; Ørgaardet al., 1995; Kosina and Heslop‐Harrison, 1996).

GISH relies largely on the hybridization of genome‐specific repetitive DNA sequences. While genes are often similar over large taxonomic distances, repetitive DNA motifs vary in both sequence and abundance even between closely related species. It has been suggested that these evolutionary changes in repetitive DNA, its genome organization and chromosomal structure are possibly correlated with speciation (Dean and Schmidt, 1995; Heslop‐Harrison, 2000).

The chromosomes of Beta species are relatively small in size (2–3 µm). Analyses of many families of repetitive sequences have demonstrated that some repeat families are highly amplified in single sections of the genus Beta, while others are conserved across taxonomic boundaries (Kubiset al., 1998; Schmidt and Heslop‐Harrison, 1998). Most of these repetitive sequence families consist of satellite DNA at prominent chromosomal sites. However, the genomes of the Beta species investigated contain sufficient amounts of dispersed species‐specific repetitive sequences to enable clear distinction of genomes (Fig. 1D, E), chromosomes (Figs 1H, 2A), chromosome fragments (Fig. 2F) and translocations (Fig. 2H–L). The differential painting of chromosomes and genomes indicates the phylogenetic distances between the Beta species. The analysis of chloroplast DNA (Fritzscheet al., 1987) suggested that species of the section Corollinae such as B. corolliflora have a closer phylogenetic relationship to sugar beet than B. procumbens of the section Procumbentes.

Nevertheless, the genomes of B. vulgaris and B. corolliflora in interspecific hybrids can be distinguished by GISH. Moreover, inspection of interphase nuclei of the amphitriploid hybrid indicates that the two Beta genomes are not randomly intermixed, and that the chromosomes of each species are clustered and occupy discrete domains (Fig. 1F, green and orange). The separation of parental genomes is similar to that observed in grasses (Schwarzacheret al., 1989; Leitchet al., 1991; Kosina and Heslop‐Harrison, 1996).

Double‐target rehybridization of the chromosome preparations of the amphitriploid hybrid enabled the comparative analysis of the rDNA loci. In diploid B. vulgaris, two major 18S‐5·8S‐25S rDNA sites have been mapped on chromosome I, while a single minor site was detectable only in a small number of cells (Schmidtet al., 1994). Three chromosomes, corresponding to a diploid B. vulgaris complement and a haploid set of B. corolliflora chromosomes, showed terminal signals indicating a conservation of the chromosomal position of the 18S‐5·8S‐25S rDNA in both Beta genomes. No minor rDNA sites were observed.

In contrast, the chromosomal position of the 5S rRNA genes differs between both species, although the basic number of 5S rDNA chromosomes is conserved between B. vulgaris and B. corolliflora. In B. vulgaris, the 5S rRNA genes have been physically and genetically mapped close to the centromeric region of the B. vulgaris chromosome IV (Schmidtet al., 1994; Schondelmaieret al., 1997). The occurrence at two intercalary loci per chromosome is a typical feature of the 5S rRNA genes in Corollinae species (Steffensen, 1997). Therefore, we concluded that the monosomic addition line shown in Fig. 1J contains an alien B. corolliflora chromosome, which is homoeologous to the B. vulgaris chromosome IV.

The alien wild beet chromosome can be unequivocally recognized in monosomic addition lines when B. corolliflora DNA is used as a probe. It is noteworthy that most of the remaining B. vulgaris chromosomes showed weak hybridization with B. corolliflora DNA at pericentromeric DNA sequences and depletion of the fluorescence signal in centromeric regions. These DAPI‐positive regions consist of long arrays of the BamHI satellite repeat pBV1, which has been amplified relatively late in the phylogeny of the section Beta (Schmidtet al., 1991). The weak hybridization on sugar beet chromosomes originates most probably from repetitive sequences that are conserved between B. vulgaris and B. corolliflora. In particular, Ty1‐copia‐like and Ty3‐gypsy‐like retrotransposons are ubiquitous sequence elements of plant genomes (Flavellet al., 1992; Hirochika and Hirochika, 1993; Brandeset al., 1997), and it has recently been shown that Ty3‐gypsy‐retrotransposons are clustered in the centromeres of B. vulgaris (Gindulliset al., 2001b).

GISH with genomic DNA of B. procumbens resulted in the detection of wild beet chromatin in mitotic and meiotic stages. Cross‐hybridization is confined to the conserved rRNA genes and no labelling of the B. vulgaris chromosomes was observed. We assume that this observation is a reflection of the large phylogenetic distance between B. vulgaris and B. procumbens. B. procumbens is considered as the most distantly related species of sugar beet (Fritzscheet al., 1987; Junget al., 1993; Sendaet al., 1995; Kubiset al., 1998). This is in agreement with our observation at diakinesis, where the alien wild beet chromosome was visible as a univalent indicating that there is only little homology between the chromosomes of B. vulgaris and B. procumbens.

One of the smallest B. procumbens segments was detected in a nematode‐resistant translocation line (Fig. 2H–L). These diploid sugar beets carry a translocation of approx. 1 Mbp from a B. procumbens chromosome. They are valuable plants for breeding because they have only a few wild beet traits and higher transmission of the introduced resistance gene than addition lines. The translocation line has been used for the genetic mapping of the nematode resistance locus (Helleret al., 1996) and positional cloning of the nematode resistance gene Hs1^pro‐1 (Caiet al., 1997).

The physical mapping of a YAC containing the Hs1^pro‐1 gene by FISH has shown that the wild translocation is terminal (Deselet al., 2001), which is consistent with the results presented here.

Satellite DNA

The B. vulgaris‐specific BamHI satellite pBV1 was used as probe for FISH to identify sugar beet chromosomes in interspecific hybrids. Southern hybridization has demonstrated that pBV1 occurs exclusively in the species of the section Beta (Schmidtet al., 1990). By FISH, the BamHI satellite labelled exclusively all sugar beet chromosomes at the pericentromeric region and did not hybridize with wild beet chromosomes in the various Beta hybrids investigated. The chromosomes of B. corolliflora and B. vulgaris are similar in size and morphology, and the determination of chromosome numbers alone is not sufficient to draw conclusions about the genomic constitution of the investigated hybrid plant. Therefore, this satellite is particularly valuable for the identification of sugar beet chromosomes in aneuploid plants (2n > 18). B. corolliflora‐specific satellite DNA sequences have been isolated and it will be interesting to use these probes in combination with pBV1 in double‐target FISH experiments (Schmidt and Heslop‐Harrison, 1993; Gaoet al., 2000).

The genome‐specific satellite pTS5 of the wild beet section Procumbentes has been used for the investigation of monosomic addition lines of sugar beet containing an alien B. procumbens chromosome (Schmidtet al., 1997). The satellite pTS5 belongs to a prominent family of tandem repeats and has been located in the centromeric region of ten to 12 of the 18 B. procumbens chromosomes (Schmidt and Heslop‐Harrison, 1996). This confined chromosomal distribution does not permit the detection of the nine possible monosomic addition lines of B. procumbens. Nevertheless, pTS5 can detect the small B. procumbens chromosome fragment in the nematode‐resistant sugar beet line PRO1. This indicates that the fragment most probably originates from one of the chromosomes containing the large pTS5 satellite arrays. Furthermore, the distal localization of the signal on the wild beet fragment suggests that the chromosomal breakpoint is located within one of the long pTS5 satellite arrays.

Satellite DNAs with species‐specificity are valuable probes for the analysis of interspecific hybrid genomes (Kammet al., 1995; Harrison and Heslop‐Harrison, 1996; Gaoet al., 2000). Genome‐specific satellite DNA has been successfully used for the selection of monosomic addition lines by rapid screening methods such as squash–dot or dot–blot hybridization (Schmidtet al., 1990). However, satellite repeats have to be cloned and their specificity for the genome under investigation has to be tested prior application in FISH. Moreover, satellite DNA is usually restricted to centromeric, subtelomeric or intercalary regions, or is present only on a subset of chromosomes of the genome. Therefore, the application of satellite probes has some limitations for the detection of alien chromatin such as intercalary introgressions or terminal translocations.

Genomic in situ hybridization is an alternative, sensitive and reliable approach to localize small alien introgressions that are not detectable by satellite DNA probes. The detection of small wild beet introgressions is of particular interest because they contain only a small number of alien genes, which are transferred into the crop.

Recombinant sugar beet chromosomes carrying a small translocation were observed only in hybrids between B. vulgaris and B. procumbens, but can be also expected in the interspecific hybrids with B. corolliflora currently used extensively to widen the gene pool of cultivated beet (Gaoet al., 2000). The Beta hybrids investigated here provide breeders with an extended range of genetic variability and it will be interesting to monitor alien chromosome segments from wild beets once sugar beet lines with desired agronomic traits have been selected.

ACKNOWLEDGEMENTS

This work was supported by the BMBF (BioFuture grant 0(11860) and DFG grants Schm1048/2–2 and Schm1048/2–3 to T. Schmidt. We thank D. Dechyeva for discussions.

Received: 24 August 2001; Returned for revision: 13 September 2001; Accepted: 20 October 2001.