A karyotypic study of three southern Brazilian Asteraceae species using fluorescence in situ hybridization with a 45 S rDNA probe and CCMA 3 banding

The Asteraceae, one of the largest families of flowering plants, contains about 1,100 genera and 20,000 species, and is well known for its extensive karyotypic variation. In this study, conventional Feulgen staining, C-CMA3 banding, and fluorescence in situ hybridization with a 45S rDNA probe were used to determine the chromosome number and the number and physical position of GC-rich heterochromatin and 45S rDNA sites in three Asteraceae weed species (Crepis japonica, Galinsoga parviflora and Chaptalia nutans). The three species exhibited karyotype differences in the chromosome number and shape, as a commom feature of Asteraceae. However, the 45S rDNA sites always occurred on the short chromosomal arms, associated with GC-heterochromatin. Althought of these differences, it suggests that commom features of plant karyotype are maintained.


Introduction
The Asteraceae is a large family of flowering plants that contains about 1,100 genera and 20,000 herbs, shrubs and, to a lesser extent, trees.The plants in this family are characterized by having reduced flowers that are organized into an involucrate pseudanthium in the form of a head or capitulum.This family is well known for its extensive variation in karyotype.Inter-and intraspecifc chromosomal variations have been reported in some Brazilian genera, such as in Mikania (Ruas and Ruas, 1987;Ruas and Aguiar-Perecin, 1997;Maffei et al., 1999).Other examples of karyotypic variation in Asteraceae include the occurrence of cytodemes in Brachycome dichromosomatica, with 2n = 4, 8, 10, 12 and 16 (Watanabe et al., 1999a), as well as variations in the number and distribution of 45S rDNA sites and the occurrence of B-chromosomes (Houben et al., 1999).
The search for chromosomal markers in wild and cultivated species of Asteraceae has contributed to our understanding of the karyotypic organization and has provided useful information for taxonomic applications.Of the DNA segments used for physical chromosome mapping, rDNA (45S and 5S) has been the most widely employed because it provides excellent chromosomal landmarks for investigating karyotypic evolution in diploid and polyploid species.According to Jiang and Gill (1994), the copy number of rDNA sites can vary among populations of common wheat, although the location of most of the loci is conserved in different diploid and polyploid wheat species.Our understanding of karyotypic features, including abnormalities and variations, can be improved by cytogenetic methods such as CMA 3 /DAPI banding and fluorescence in situ hybridization (FISH).
The cytogenetic information provided by a combination of chromosome banding and FISH can be useful for comparing species within a genus, as well as species of different genera.In this study, three annual Asteraceae weeds (Crepis japonica (L.) Benth., Galinsoga parviflora Cav. and Chaptalia nutans (L.) Pol.), which are very common, small plants found in grasslands and disturbed areas in southern Brazil where they colonize degraded or newly open habitats, were studied.The aim of this study was to determine the chromosome number and to establish the number and physical location of 45S rDNA sites and C-CMA 3 bands in these groups.The findings are discussed with regard to the number and location of rDNA sites and their relationship with C-CMA 3 bands, in both diploid and polyploid specimens.

Plant material and conventional analysis
Five samples of each species were collected (Crepis japonica and Galinsoga parviflora from Londrina in the State of Paraná -PR) and Chaptalia nutans from Londrina-PR and Laguna, in the State of Santa Catarina).The specimens were cultivated in a greenhouse and voucher material was deposited in the FUEL herbarium.Chromosomal preparations were obtained from root tips pretreated with 2 mM 8-hydroxyquinoline for 24 h followed by fixation in ethanol:acetic acid (3:1, v/v) for 12 h and storage at -20 °C.The Feulgen method was used for conventional analysis.Chromosomal recounts were done in at least five complete metaphases.

C-CMA 3 banding
The chromosomal preparations were processed for C-banding according to Schwarzacher et al. (1980), with modifications (see Vanzela et al., 2002).The samples were softened in a solution containing 4% cellulase and 40% pectinase at 37 °C for 1 h and then squashed in a drop of 45% acetic acid.The coverslips were subsequently removed in liquid nitrogen.After three days, the samples were hydrolyzed in 45% acetic acid at 60 °C for 10 min followed by incubation in 5% Ba(OH) 2 at 25 °C for 10 min and in 2xSSC, pH 7.0, at 60 °C for 80 min.After a further three days, the samples were stained for 1.5 h with a drop of solution containing 0.5 mg of CMA 3 /mL diluted in McIlvaine buffer (pH 7.0) and distilled water (1:1, v/v) containing 2.5 mM MgCl 2 .The slides were mounted in a solution of glycerol:McIlvaine buffer (1:1, v/v) containing 2.5 mM MgCl 2 .
To assess the relationship between C-CMA 3 blocks and 45S rDNA sites, C. nutans with 2n = 50 were processed sequentially for FISH and C-CMA 3 banding.Photographs were taken using Kodak Imagelink HQ ISO 25 film for conventional staining, T-Max ISO 100 film for C-CMA 3 banding and Proimage ISO 100 film for FISH.

Results and Discussion
Variations in the chromosomal number have been reported for many groups of Astereaceae, including diploidy as in Mikania (Maffei et al., 1999) and polyploidy as in Helianthus (Vanzela et al., 2002); other groups, such as Stevia, show numerical stability (Frederico et al., 1996).
As shown here, Crepis japonica (L.) Benth.had 2n = 16 chromosomes which were metacentric and submetacentric and of similar size (~1.9-2.2 µm); satellites were seen in only one submetacentric pair (Figure 1A).FISH revealed only two signals for 45S rDNA in interphase nuclei, and these were observed on almost the entire short arm of one submetacentric pair (Figure 1C).CMA 3 + NORassociated heterochromatin has been observed on the short arm of a submetacentric pair in three Crepis foetida subspecies (2n = 10), in addition to smaller pericentromeric and intercalary dots revealed by Giemsa-C banding (Dimitrova et al., 1999).Guerra (1982) and Noguchi and Ohno (1989) described large C-bands on the short arm of submetacentric pairs in C. vesicaria and C. capillaris, respectively, in which this large heterochromatic block was associated with NOR, at least in C. capillaris.The observations agree with the FISH results obtained here, and suggest a constancy in the localization of the NOR on the submetacentric short arm in these species.
Galinsoga parviflora Pav. had 2n = 16 with small meta-and submetacentric chromosomes, the smallest pair being of the acrocentric type.Chromosomal lengths varied from 1.4 µm to 1.9 µm.According to Strother and Panero (1994), x = 8 is the probable basic number for Galinsoga.The largest submetacentric pair had a pericentromeric CMA 3 + block that occupied most of the short arm.Additional C-CMA 3 + bands appeared as interstitial dots on the other two small chromosomes (Figure 1F).FISH with a 45S rDNA probe revealed two signals in interphase nuclei and a large hybridization site in the major submetacentric pair (Figure 1H), probably at the same position as the CMA 3 + block.
Chaptalia nutans (L.) Pol.showed polyploidy in two cytotypes, 2n = 50 collected in Londrina-PR and 2n = 100 from Laguna-SC (the conventional staining is not shown here).These chromosomal numbers agree with previous counts of n = 25 and 2n = 50 for this species (Teppner and Tropper 1984).The chromosomal number for C. graminifolia and C. pilloselloides has been reported to be 2n = ca.48 (Wulff et al., 1996) and 2n = ca.50 for C. arechavaletai (Waisman et al., 1984).The Laguna cytotypes studied here had karyotypes with duplicated chromosomal numbers, i.e., both were asymmetrical and consisted of large meta-submetacentric types (~2.5 µm in size) and several small acrocentric pairs ~1.3 µm long (Figure 1B).There were no major differences in chromosomal size and type.Neither cytotype showed differences in plant morphology.The C-CMA 3 -banding of the 2n = 50 and 100 cytotypes revealed four and eight terminal C-CMA 3 + blocks, respectively (Figure 1D and G).In the 2n = 50 cytotype, FISH with rDNA 45S revealed four hybridization signals of the same size (Figure 1E), whereas in the 2n = 100 cytotype one pair of chromosomes gave a much stronger hybridization signal than the other (Figure 1I).This size difference was also observed in these segments after C-CMA 3 banding (Figure 1G).
The results described here support the idea that the C-CMA 3 banding method is useful for examining the correlation between terminal GC-rich segments and 45S rDNA sites, as demonstrated by sequential FISH and C-CMA 3 banding in the C. nutans cytotype with 2n = 50 (Figures 1D, E).In addition, C-CMA 3 banding was effective in revealing fine bands (see arrows in Figure 1F) that were not visualized when stained directly with CMA 3 .Vanzela et al. (2002) used this same procedure in some Helianthus species and found terminal GC-rich segments associated with 45S rDNA, as well as fine GC-rich bands that were not visualized after CMA 3 and Giemsa C banding.GC-heterochromatin associated with 45S rDNA sites is a common karyotypic feature that has been reported for many plant species (see Guerra, 2000a).The occurrence of interstitial C-CMA 3 blocks in a smaller chromosomal pair of G. parviflora (Figure 1F) suggested that, at least in this species, not all GC-rich heterochromatin was associated with rDNA sites.
The three species showed a conserved location for the 45S rDNA sites on the short chromosomal arms, i.e., a terminal position in C. nutans and proximal-intercalary positions in C. japonica and G. parviflora.Clearer differences were observed in the size of the hybridization sites among the three species and within C. nutans with 2n = 100.The chromosomes bearing 45S rDNA were either submetacentric or acrocentric.Although FISH is a qualitative rather than a quantitative technique, an association between the size of the hybridization signal and the number of rDNA cistrons can be demonstrated (see Zurita et al., 1997).Thus, the most brightly stained 45S rDNA sites likely represented cistrons with more repetitions than those with less intense signals.
Polyploidy is very common in plants and occurs in 70% of angiosperms (Wendel, 2000).According to Greilhuber and Ehrendorfer (1988), when polyploids are established as a population, they can become reproductively isolated and morphologically distinct from their parent species.Some Vernonia species collected in South America have a high level of ploidy and numerous interspecific karyotypic differences, especially in karyotype form, total chromosomal length and degree of asymmetry (Dematteis and Fernández, 2000).The two cytotypes of C. nutans studied here (2n = 50 and 100) were morphologically indistinct and certainly polyploids since the basic chromosome numbers proposed for angiosperms are x = 4, 5 and 6 (Guerra, 2000b).However, there was no evidence indicating that they were auto-or allopolyploids, despite the larger number of C-CMA 3 blocks and 45S rDNA sites.
Cytotypes originating from polyploidy have also been reported for other Brazilian Asteraceae, including Mikania micrantha (Ruas and Ruas, 1987) and Vernonia polyphylla (Ruas et al., 1991).In some cases, the karyotypes of Asteraceae, such as those belonging to the Gnaphalieae group (Watanabe et al., 1999b), can vary considerably in number, from 2n = 6 in Podolepis to 2n = 120 in Craspedia, and in symmetry, from unimodal in Podotheca to bimodal in Schoenia.Thus, the karyotype differences observed among Crepis japonica, Galinsoga parviflora and Chaptalia nutans appear to be common in the Asteraceae family.In contrast, the constancy of the rDNA position on the short arm of submetacentric pairs, as well as the association of this chromosomal segment with CMA 3 + heterochromatin, are common features of most plant karyotypes.