Chromosome Research (2007) 15:115–121 DOI: 10.1007/s10577-006-1101-2

# Springer 2007

Karyotype diversity and the origin of grapefruit Ana Paula de Moraes1, Walter dos Santos Soares Filho2 & Marcelo Guerra1* 1 Laborato´rio de Citogene´tica Vegetal, Departamento de Botaˆnica, Universidade Federal de Pernambuco, Recife, Brazil; Tel: +55-81-21268846; Fax: +55-81-2126834; E-mail: [email protected] 2 Embrapa Mandioca e Fruticultura Tropical, Cruz das Almas, Brazil * Correspondence

Key words: Citrus, CMA/DAPI banding, grapefruit, karyotype evolution, rDNA

Abstract Grapefruit is a group of citrus of recent origin, probably resulting from a cross between pummelo and sweetorange. Aiming to investigate this putative origin and the genetic variability among grapefruit cultivars, the karyotype of six grapefruits, two pummelos, and one tangelo cultivar (grapefruit  tangerine) were analyzed using sequential CMA/DAPI double staining and FISH with rDNA probes. The karyotypes of grapefruit FDuncan_ and FFoster_ differ from those of FFlame_, FHenderson_, FMarsh_ and FRio Red_. The former have two chromosomes with a single CMA+ band in both terminal regions (C type chromosome) and six chromosomes with only one CMA+ terminal band (D type), whereas the latter have three C and five D type chromosomes. All accessions investigated exhibited two chromosomes with 5S rDNA but a variable number of 45S rDNA. The two former grapefruits displayed four 45S rDNA sites, whereas the remaining grapefruit cultivars had five. The two pummelos showed identical karyotypes, homozygous for CMA+ bands and their four rDNA sites. From each pummelo chromosome pair one chromosome seems to be present in grapefruit karyotypes. The different grapefruit karyotypes might result from independent crosses between pummelos of different karyotypic constitution and sweet-oranges. The chromosome markers found in the tangelo FOrlando_ and the position of their two 45S rDNA confirm the grapefruit FDuncan_ and the tangerine FDancy_ as their parents.

Introduction Grapefruit (Citrus paradisi Macf.) is a commercial crop of great importance, mainly for the juice industry, but also as a source of essential oil and pectin (Hodgson 1967). The origin of grapefruit is attributed to hybridization between pummelo [C. grandis (L.) Osbeck] and sweet-orange [C. sinensis (L.) Osbeck], which might have occurred in the Caribbean õ200 years ago (Webber 1943, Barrett & Rhodes 1976, Scora et al. 1982). Thus, grapefruit might be more correctly classified as an interspecific hybrid than as a species (Gmitter 1995). The first of such hybrids were described in 1750 from Barbados Island (Kumamoto

et al. 1986) and referred to as Fforbidden fruits_ for reasons that are not clear at present. The relation between these hybrids and current grapefruit cultivars has been a source of controversy (reviewed by Bowman & Gmitter 1990). The scientific name C. paradisi was given by Macfadyen in 1830 to a citrus tree known as Fforbidden fruit_ in Jamaica. Although considered as a natural hybrid in the Caribbean, grapefruit has never been found to occur naturally in Asia, where its supposed ancestors are native (Bowman & Gmitter 1990) and from where other natural hybrids of pummelo are known (Hodgson 1967). Current grapefruits are apparently derived from a stock of wild plants introduced to Florida in 1823 by Count Odette Philippi (Bowman &

A. P. de Moraes et al.

116 Gmitter 1990). From this material FDuncan_, FMarsh_, and FWalters_ were selected, from which all other cultivars are derived (Gmitter 1993). The hybrid nature of several citrus species, as well as the genetic similarity among different cultivars and the origin of commercially important cultivars, have been investigated by different methods, mainly by molecular markers (reviewed by Moore 2001). Karyotype analysis revealed chromosome polymorphisms between (and among) species of the genus Citrus (Guerra 1993, Befu et al. 2000, Yamamoto & Tominaga 2003, Carvalho et al. 2005). Frost (1925) counted the correct haploid chromosome number (n = 9) for several Citrus species, which was subsequently confirmed by several authors (e.g. Guerra et al. 1997). In spite of the stable chromosome number and similarities in chromosome size and morphology, Citrus species and related genera show high variability as to the number and distribution of heterochromatic blocks (Guerra et al. 2000). Double staining with the fluorochromes chromomycin A3 (CMA) and 40 -6-diamidino-2-phenyl-indole (DAPI), with preferential affinity to C-G and A-T base pairs, respectively, revealed CMA+/DAPIj heterochromatic blocks in many chromosomes of all citrus species as well as interspecific polymorphism and a high level of chromosome heterozygosity in nearly all investigated cultivars (e.g. Miranda et al. 1997, Yamamoto & Tominaga 2003, Carvalho et al. 2005). Among grapefruit cultivars, only for FDuncan_ was a karyotype formula described (Befu et al. 2001). FDuncan_ possesses some chromosome markers also present in pummelo. Among the pummelo cultivars, karyotypes differ (Befu et al. 2000, 2001, Yang et al. 2002), whereas sweet-orange accessions revealed stable, but heteromorphic, karyotypes (Matsuyama et al. 1996, Miranda et al. 1997, Befu et al. 2000, Guerra et al. 2000, Pedrosa et al. 2000). Besides chromosome banding, the number and location of 5S and 45S rDNA sites provide additional chromosome markers. The location of these sites was reported for some sweet-orange cultivars (Pedrosa et al. 2000) and for a group of closely related species including lime, lemon, and citron (Carvalho et al. 2005). In both papers, heteromorphism as to position and number of 5S and 45S rDNA loci was observed in all cultivars, except for citron. In the present work the CMA+ band pattern and the distribution of 5S and 45S rDNA sites were analyzed for six grapefruit cultivars, in order to esti-

mate the karyotypic similarities within this group. Additionally, two pummelos, FIsrael_ and FPink_, and one tangelo, FOrlando_ (resulting from C. paradisi cv. Duncan  C. tangerina cv. Dancy), were investigated, aiming to understand the relationship between grapefruit and its possible parents, pummelo and sweetorange, on the one hand, and to prove the transmission of these chromosome markers to tangelo, an artificial citrus hybrid between karyotypically heterozygous parents, on the other. Materials and methods Materials Ovary walls and foliar meristems from adult individuals or shoot apical meristems of seedlings or germinated seeds of six cultivars of grapefruit (FDuncan_, FFlame_, FFoster_, FHenderson_, FMarsh_, FRio Red_), two of pummelo (FIsrael_ and FPink_) and one of tangelo (FOrlando_) were analyzed. All accessions were from the germplasm collection of the experimental station Embrapa Cassava & Tropical Fruits. Chromosome preparation The materials were pretreated with 8-hydroxyquinoline (0.002 M) for 5 h at 18-C, fixed in 3:1 ethanolYacetic acid (v/v) for 2Y24 h at room temperature and stored at j20-C. The materials were washed in distilled water and digested at 37-C for 2 h with a 2% cellulase (Onozuka)Y20% pectinase (Sigma) (w/v) solution. Then the meristems were squashed in a drop of 45% acetic acid and the coverslip was removed in liquid nitrogen. The slides were stained with DAPI (2 mg/ml): glycerol (1:1, v/v) solution to allow selection of the best preparations. Subsequently, they were destained in ethanolYacetic acid (3:1) for 30 min at room temperature and transferred to absolute ethanol overnight at 10-C. Slides were air-dried and aged for 3 days at room temperature. CMA/DAPI staining After aging, the slides were stained with CMA and DAPI as described previously (Carvalho et al. 2005). The best metaphases were captured with a Cohu video camera using Leica QFISH software. The slides were

Karyotype diversity and the origin of grapefruit destained again and stored at j20-C for posterior in-situ hybridization with 5S and 45S rDNA as probe. Fluorescence in situ hybridization (FISH) The 5S rDNA probe was obtained from genomic DNA of C. aurantifolia (Christm.) Swingle by PCR

117 using the primers 50 -GTG CGA TCA TAC CAG C(A/G)(C/T) TAA TGC ACC GG- 30 and 50 -GAG GTG CAA CAC GAG GAC TTC CCA GGA GG-30 (Gottlob-McHugh et al. 1990). The 45S rDNA sites were located using the R2 probe, which is a 6.5 kb fragment of an 18S-5.8S-25S rDNA repeat unit from

Figure 1. CMA+ band pattern and distribution of 5S and 45S rDNA sites in grapefruits (aYf), pummelo (gYi) and tangelo (jYl). aYc: FFoster_. dYf: FHenderson_. gYi: FPink_ pummelo. jYl: FOrlando_ tangelo. Photographs in the left column show CMA+ bands. The other two columns show DAPI stained chromosomes (pseudocolored in gray) with 5S rDNA sites in red (middle) and 45S rDNA sites in green (right). Capital letters indicate chromosome types. Arrowheads and arrows indicate the positions of 5S rDNA sites and small 45S rDNA sites, respectively. Bar in l corresponds to 5 mm.

A. P. de Moraes et al.

118

nologies). The FISH procedure was based on Moscone et al. (1996) with minor modifications. The hybridization mixture contained 60% v/v formamide, 5% w/v dextran sulfate, 2 SSC, salmon sperm DNA (0.3 ng/ml), and 2Y5 ng/ml of probe. The 5S rDNA was detected with mouse anti-biotin (Roche) and the signals were amplified with rabbit anti-mouse TRITC conjugate (Dako). The 45S rDNA was detected with sheep anti-digoxigenin FITC conjugate (Roche) and amplified with rabbit antisheep FITC conjugate (Dako). All preparations were counterstained with DAPI (2 mg/ml) and mounted in Vectashield (Vector). The cell images were acquired as above.

Results

Figure 2. Schematic representation of the chromosomal distribution of CMA+ bands and 5S and 45S rDNA site in the cultivars analyzed. The relative chromosome size and morphology in these idiograms are only an approximation. Yellow blocks = CMA+ bands, red blocks = 5S rDNA sites, and green blocks = 45S rDNA site.

Arabidopsis thaliana, kindly provided by Professor D. Schweizer of the University of Vienna. The 5S rDNA probe was labeled by PCR with biotin-16-dUTP (Roche), while the 45S rDNA probe was labeled by nick translation with digoxigenin-11-dUTP (Life Tech-

All analyzed accessions showed 18 meta- to submetacentric chromosomes. The heterochromatin stained brightly with CMA+ (CMA bands), while DAPI stained all chromosomes uniformly, except for some CMA+ blocks, which were less intensely stained by DAPI (DAPIj bands). In many cases the DAPIj bands were very useful to identify the small or pale CMA+ bands. The variation in number and position of CMA+ bands allowed identification of several chromosome types (Corne´lio et al. 2003): A type, one proximal and two terminal bands; B type, one proximal and one terminal band; C type, two terminal bands; D type, one terminal band on the long arm, and F type, without bands, to which the largest pair belongs FL (Figures 1a and 2). The chromosome types E (one interstitial band) and G (two nearly terminal bands), reported for other Citrus species (Carvalho et al. 2005), were not found. The chro-

Table 1. Karyotype formulae based on CMA+ bands and 5S and 45S rDNA sites of grapefruits, pummelos and tangelo FOrlando_ Distribution of rDNA sites Cultivars Grapefruit Duncan/Foster Flame/Henderson/ Marsh/Rio Red Pummelo Israel/Pink Tangelo FOrlando_

Karyotype formulae

5S

45S

5SY45S

2A+1B+2C+6D+7F 2A+1B+3C+5D+7F

1F 1F

2A+1B 2A+1B+1D

1D 1D

4A+2C+4D+8F

2F

4A

V

1B+3C+8D+4F+2FL

1F

1B

1D

Karyotype diversity and the origin of grapefruit mosome types A, B and C are more conspicuous and less frequent and therefore useful as marker chromosomes, while D and F types are widely distributed among Citrus species. The investigated grapefruit accessions showed similar karyotype formulae and could be divided into two groups: FDuncan_ and FFoster_, with 2A + 1B + 2C + 6D + 7F (Figure 1aYc) and FMarsh_, FFlame_, FHenderson_, and FRio Red_, with 2A + 1B + 3C + 5D + 7F (Figure 1dYf). The pummelos FIsrael_ and FPink_ displayed 4A + 2C + 4D + 8F, while the tangelo FOrlando_ exhibited a very different karyotype (1B + 3C + 8D + 6F), without A type chromosomes, characteristic for the other cultivars (Table 1). Figure 1a shows all chromosome types identified by letters, while in the remaining cells only the marker chromosomes A, B, and C were identified. Chromosomes D and F were also pointed out when they exhibited rDNA sites. Figure 2 displays the corresponding idiograms. The A type chromosomes showed a two-fold variation in band size, mainly on the long arm (Figure 1a). On B type chromosome the proximal CMA+ band was smaller than the terminal one (Figure 1a, d and j). The C type chromosomes also exhibited heteromorphism in chromosome length and band size, except for pummelos (compare Figure 1a and g). In the cultivars with three C chromosomes, two of them presented identical size and band intensity, while the third was larger (Figure 1d and j). All grapefruit and tangelo cultivars revealed one D chromosome with a significantly larger CMA+ band that sometimes apparently comprises a large part of the long arm. In all karyotypes one F L chromosome pair was observed (Figures 1 and 2). All 45S rDNA sites coincided with CMA+ bands located in proximal bands of the A and B chromosomes or in the terminal band of the largest D chromosome (Table 1 and Figures 1 and 2). Among grapefruits with 2C + 6D, four 45S rDNA sites were observed: three on A and B type chromosomes and the fourth one on the biggest D chromosome. The accessions with 3C + 5D displayed five 45S rDNA sites, four as in the previous group and an additional site located on another D chromosome. The pummelos FIsrael_ and FPink_ showed four 45S rDNA sites on A chromosomes, and the tangelo FOrlando_ possesses only two sites, one on a B and the other on a D chromosome. The 5S rDNA sites displayed no variation among the analyzed grapefruit and tangelo cultivars. They had two 5S rDNA sites, one in the subterminal region

119 of one F chromosome and the other on the biggest D, adjacent to a 45S rDNA site. This D chromosome was denominated D/5S-45S (Carvalho et al. 2005). Pummelos displayed 5S rDNA sites on two F chromosomes (Table 1).

Discussion CMA/DAPI double staining provided a clear karyotype differentiation of all accessions analyzed. Besides the A, B, and C chromosome types, D chromosomes may also serve as markers when combined with rDNA sites distribution. Three different D chromosome types are known: (a) D/5S type, bearing a 5S rDNA site in the euchromatic region; (b) D/45S, with a 45S rDNA site co-localized with the CMA+ band and, (c) D/5S-45S, with adjacent 5S and 45S rDNA sites co-localized with the CMA+ band (Pedrosa et al. 2000, Carvalho et al. 2005). Similarly, F chromosomes can be useful as a good marker when having a 5S rDNA site (F/5S), known previously only in C. sinensis (Pedrosa et al. 2000). In the cultivars analyzed, all these chromosomes types were found, except for D/5S, and contributed to differentiate their karyotypes. The band pattern analyzed showed that the six grapefruit cultivars investigated were not cytologically identical. The grapefruit FDuncan_ and FFoster_ showed the karyotype formula 2A + 1B + 2C + 6D + 7F, while FFlame_, FHenderson_, FMarsh_ and FRio Red_ exhibited 3C + 5D instead of 2C + 6D. Bearing in mind the different D chromosome types, the difference between these two groups becomes more important: FDuncan_ and FFoster_ are characterized by 5D + 1D/5S-45S, whereas the other cultivars displayed 3D + 1D/5S45S + 1D/45S. Since all 45S rDNA sites in Citrus are CMA+/DAPIj bands, five of the six D chromosomes of FDuncan_ and FFoster_ should carry a GC-rich satellite sequence not related to rDNA, while the other group of cultivars have only three D chromosomes with this sequence. Considering that the six grapefruit cultivars form two groups with different karyotypes, it seems unlikely that all grapefruit cultivars are derived apomictically from the same hybrid, as proposed by Barrett & Rhodes (1976). The only previous report of a grapefruit karyotype formula was published by Befu et al. (2001) with 1C less and 1F more than we observed for FDuncan_. Such a variation within a cultivar was unexpected. Either these authors overlooked weak CMA+ bands, due to the lack

A. P. de Moraes et al.

120 of DAPI counterstaining, or some heterochromatic bands were strongly reduced/ amplified in these apomictic accessions. The karyotype of the putative grapefruit ancestor, C. sinensis, was analyzed in at least 13 different cultivars (Guerra 1993, Matsuyama et al. 1996, Miranda et al. 1997, Befu et al. 2000, Pedrosa et al. 2000). It always showed the karyotype formula 2B + 2C + 7D + 7F, with different sized C chromosomes and rDNA sites located on two B/45S, one D/5S-45S, and one F/5S chromosome (Pedrosa et al. 2000). This species generally produces nucellar embryos, allowing the maintenance of karyotype heterozygosity. The pummelo C. grandis, on the other hand, is a monoembryonic species, producing exclusively, or predominantly, zygotic embryos (Barrett & Rhodes 1976). The karyotype formula and the distribution of rDNA sites for the two pummelo cultivars analyzed indicate karyotypic homozygosity, supporting the hypothesis that C. grandis could be one of the pure species of the subgenus Citrus, as proposed by Barrett & Rhodes (1976). However, some heteromorphic karyotypes have been previously described for pummelos, suggesting that this species is naturally polymorphic for some chromosome types. Guerra (1993) and Befu et al. (2002) reported the presence of 3A + 3C and 3A + 2C marker chromosomes, respectively, in different pummelos. These data suggest that A and C chromosomes can be found in more than one chromosome pair of this species (see also Yang et al. 2002). Therefore, homomorphic and heteromorphic karyotypes should occur naturally in C. grandis, differing from other accessions that are also considered as true species and display homomorphic karyotypes (Befu et al. 2001, Corne´lio et al. 2003, Carvalho et al. 2005). The most outstanding karyotypic feature of C. grandis is the presence of A chromosomes, which were not found in sweet-oranges, citrons, limes, lemons (Carvalho et al. 2005), and the majority of the tangerines tested so far (Corne´lio et al. 2003, Yamamoto & Tominaga 2003). The presence of two A chromosomes in all grapefruit cultivars confirms their relationship with pummelos. These two A chromosomes are likely nonhomologous, since C. sinensis has no A chromosome and the C. grandis cultivars analyzed here have two pairs. Therefore FDuncan_ and FFoster_ might have originated by fusion of a C. sinensis gamete (1B/45S + 1C + 3D + 1D/5S - 45S + 3F) and a C. grandis gamete (2A/45S + 1C + 2D + 3F + 1F/5S). The other grapefruit accessions with 3C + 5D could have been formed in a

similar way, involving, for example, an identical C. sinensis gamete and a C. grandis gamete with 2A/ 45S + 2C + 2D + 2F + 1F/5S. However, these grapefruit cultivars with 3C + 5D displayed a D/45S chromosome, which was not found in either pummelo or sweetorange. Since up to seven D chromosomes have been found in a single pummelo cultivar (Miranda et al. 1997), it is possible that D/45S chromosomes may exist in some non-analyzed pummelos. Alternatively, FMarsh_ could have originated from a cross between FDuncan_, or a similar cultivar, and one of several hybrids (C. grandis  C. sinensis) which have been reported from the Caribbean (Robinson 1933), while FFlame_, FHenderson_, and FRio Red_ are derived from FMarsh_ by sport mutation (Bowman & Gmitter 1990) conserving the same karyotype. According to Webber (1943), tangelos descend from a cross between grapefruit (C. paradisi cv. Duncan) and tangerine (C. tangerina cv. Dancy) made in 1827 at the experimental station of Florida, and the majority of resulting cultivars, with some economic importance, resemble the tangerine. Indeed, the FDuncan_ grapefruit karyotype formula (2A + 1B + 2C + 5D + 8F) and the FDancy_ tangerine (1C + 12D + 5F, see Corne´lio et al. 2003), are compatible with the FOrlando_ karyotype (1B + 3C + 8D + 6F). The lack of A chromosome in FOrlando_ should be attributed to the independent segregation of these two nonhomologous chromosomes in FDuncan_. Similarly, both FDuncan_ C chromosomes should belong to distinct heteromorphic pairs, since both of them were transmitted to tangelo. Additionally, the presence of an F/5S chromosome, seen in grapefruit and sweet-orange, but rarely in other citrus species, supports a relationship between tangelo and grapefruit. The concordance between ancestral and descendant karyotypes on one hand, and with the historical record on the other, demonstrates the value of cytogenetic analyses for reconstructing the origin of hybrids and for understanding phylogeny and evolution of Citrus.

Acknowledgements This work was supported by the Brazilian agencies Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico, CNPq, (n. 474589/03-0, 307916/03-0, and 140273/04-2) and Fundac¸a˜o de Amparo a` Cieˆncia e Tecnologia, FACEPE, (Edt. 0005-05-03/04).

Karyotype diversity and the origin of grapefruit References Barrett HC, Rhodes AM (1976) A numerical taxonomic study of affinity relationships in cultivated Citrus and its close relatives. Syst Bot 1: 105Y136. Befu M, Kitajima A, Ling YX, Hasegawa K (2000) Classification of FTosa-Butan_ pummelo (Citrus grandis [L.] Osb.), FWashington_ Navel orange (C. sinensis [L.] Osb.) and trifoliate orange (Poncirus trifoliata [L.] Raf.) chromosomes using young leaves. J Jpn Soc Hort Sci 69: 22Y28. Befu M, Kitajima A, Hasegawa K (2001) Chromosome composition of some citrus species and cultivars based on the chromomycin A3 (CMA) banding patterns. J Jpn Soc Hort Sci 70: 83Y88. Befu M, Kitajima A, Hasegawa K (2002) Classification of the Citrus chromosomes with same types of chromomycin A banding patterns. J Jpn Soc Hort Sci 71: 394Y400. Bowman KD, Gmitter FG, Jr (1990) Caribbean Forbidden Fruit: grapefruit_s missing link with the past and bridge to the future? Fruit Var J 44: 41Y44. Carvalho R, Soares Filho WS, Brasileiro-Vidal AC, Guerra M (2005) The relationship among lemons, limes and citron: a chromosomal comparison. Cytogenet Genome Res 109: 276Y282. Corne´lio MTMN, Figueiroˆa ARS, Santos KGB, Carvalho R, Soares Filho WS, Guerra M (2003) Chromosomal relationships among cultivars of Citrus reticulata Blanco, its hybrids and related species. Plant Syst Evol 240: 149Y161. Frost HB (1925) The chromosomes of Citrus. J Wash Acad Sci 15: 1Y2. Gmitter FG, Jr (1993) FMarsh_ grapefruit. Fruit Var J 47: 130Y133. Gmitter FG, Jr (1995) Origin, evolution, and breeding of the grapefruit. Plant Breed Rev 13: 345Y363. Gottlob-McHugh SG, Le´vesque M, MacKenzie K, Olson M, Yarosh O, Johnson DA (1990) Organization of the 5S rRNA genes in the soybean Glycine max (L.) Merrill and conservation of the 5S rDNA repeat structure in higher plants. Genome 33: 486Y494. Guerra M (1993) Cytogenetics of Rutaceae. V. High chromosomal variability in Citrus species revealed by CMA/DAPI staining. Heredity 71: 234Y241. Guerra M, Pedrosa A, Silva AEB, Corne´lio MTM, Santos KGB, Soares Filho WS (1997) Chromosome number and secondary constriction variation in 51 accessions of a Citrus germplasm bank. Braz J Genet 20: 489Y496.

121 Guerra M, Santos KGB, Silva AEB, Ehrendorfer F (2000) Heterochromatin banding patterns in Rutaceae-Aurantioideae Y a case of parallel chromosomal evolution. Am J Bot 87: 735Y747. Hodgson RW (1967) Horticultural varieties of Citrus. In Reuter W, Webber HJ, Batchelor LD, eds., The Citrus Industry. History, World Distribution, Botany and Varieties, vol. I. University of California Press, pp. 431Y591. Kumamoto J, Scora RW, Lawton HW, Clerx WA (1986) Mystery of the forbidden fruit: historical epilogue on the origin of the grapefruit, Citrus paradisi (Rutaceae). Econ Bot 41: 97Y107. Matsuyama T, Akihama T, Ito Y, Omura M, Fukui K (1996) Characterization of heterochromatic regions in FTrovita_ orange (Citrus sinensis Osbeck) chromosomes by the fluorescent staining and FISH methods. Genome 39: 941Y945. Miranda M, Ikeda F, Endo T, Moriguchi T, Omura M (1997) Comparative analysis on the distribution of heterochromatin in Citrus, Poncirus and Fortunella chromosomes. Chromosome Res 5: 86Y92. Moore GA (2001) Oranges and lemons: clues to the taxonomy of Citrus from molecular markers. Trends Genet 17: 536Y540. Moscone EA, Matzke MA, Matzke AJM (1996) The use of combined FISH/GISH in conjunction with DAPI counterstaining to identify chromosomes containing transgene inserts in amphidiploid tobacco. Chromosoma 105: 231Y236. Pedrosa A, Schweizer D, Guerra M (2000) Cytological heterozygosity and the hybrid origin of sweet-orange [Citrus sinensis (L.) Osbeck]. Theor Appl Genet 100: 361Y367. Robinson TR (1933) The origin of the Marsh seedless grapefruit. J Heredity 24: 437Y439. Scora RW, Kumamoto J, Soost RK, Nauer EM (1982) Contribution to the origin of the grapefruit Citrus paradisi (Rutaceae). Syst Bot 7: 170Y177. Webber HJ (1943) Cultivated varieties of citrus. In Webber HJ, ed., The Citrus Industry. History, World Distribution, Botany and Varieties, vol. I. University of California Press, pp. 475Y668. Yamamoto M, Tominaga S (2003) High chromosomal variability of mandarins (Citrus spp.) revealed by CMA banding. Euphytica 129: 267Y274. Yang X, Kitajima A, Hasegawa K (2002) Chromosome pairing set and the presence of unreduced gametes explain the possible origin of polyploid progenies from the diploids FTosa-Butan_  FSuisho-Butan_ pummelo. J Jpn Soc Hort Sci 71: 538Y543.