Fax +41 61 306 12 34E-Mail karger@karger.chwww.karger.com

Original Article

Cytogenet Genome Res 119:275–281 (2007) DOI: 10.1159/000112074

Chromosomal markers distinguish hybrids and

non-hybrid accessions of mandarin

A.P. Moraes a R.R. Lemos a A.C. Brasileiro-Vidal b W. dos Santos Soares Filho c M. Guerra a Departments of a Botany and b Genetics, Federal University of Pernambuco, Recife c Citrus Active Germplasm Bank of Embrapa Cassava & Fruits, Cruz das Almas (Brazil)

most accessions are hybrids instead of pure species, where-as Tanaka (1969) separated the mandarins into 36 species.

In the last 20 years, several biochemical and molecular studies have tried to put order in this confusing group, with quite modest results. In general, these works revealed a large molecular heterogeneity, suggesting that most mandarin accessions belong to a single species, C. reticulata Blanco, composed of many genotypes and hybrids (Luro et al., 1995; Fang and Roose, 1997; Coletta Filho et al., 1998; Federiciet al., 1998; Nicolosi et al., 2000). For these authors thereare only three Citrus species: C. grandis , C. medica L., and C. reticulata (reviewed by Moore, 2001). On the other hand, two extensive chromosomal studies in mandarin accessions (Cornélio et al., 2003; Yamamoto and Tominaga, 2003) in-dependently revealed much larger variation than that ob-served in other Citrus species, suggesting a polyphyletic rather than a monophyletic status for the mandarin group.

In spite of the chromosome number stability observed in the genus (2n = 18), Citrus species are cytogenetically well characterized by distinct patterns of chromosomal bands rich in GC base pairs revealed by fluorochrome chromomy-cin A 3 , CMA (Guerra, 1993). These brilliant CMA + bands can also be identified as negative bands when stained with

Abstract. Mandarin is the common name of a heteroge-neous group of Citrus species with a large range of variation in morphological and molecular characters as well as in number of species. Aiming to identify chromosome mark-ers and to clarify the relationship within this group, the karyotype of 13 mandarin accessions were analyzed using CMA/DAPI staining and in situ hybridization with 5S and 45S rDNA probes. The CMA band pattern together withthe position of rDNA sites revealed that mandarins can be separated karyologically into three groups: a) C. sunki and

Request reprints from M. GuerraDepartamento de Botânica, Centro de Ciências BiológicasUniversidade Federal de Pernambuco, 50.670-420, Recife, PE (Brazil)telephone: +55 81 2126 8846; fax: +55 81 2126 8348e-mail: mguerra@ufpe.br.

© 2008 S. Karger AG, Basel1424–8581/07/1194–0275$23.50/0

Accessible online at:www.karger.com/cgr

C. reshni ; b) the Mediterranean mandarin, C. deliciosa , and the closely related C. tangerina cv. Dancy and C. reticulata cv. Cravo; c) the remaining cultivars, which are cytologi-cally heterozygous and most probably interspecific hybrids. The former two groups are assumed to be pure species to-gether with C. medica and C. grandis . A chromosome mark-er for mandarin species was identified and the relationship among the pure species and some hybrids is discussed.

Copyright © 2008 S. Karger AG, Basel

Mandarin is the common name of a group of Citrus L. species whose fruits are easily recognized by their usually thin, loose peel, and a characteristic flavor and aroma. In spite of this simple characterization, the range of variation in morphological characters observed among mandarin ac-cessions is much greater than among sweet orange ( C. sinen-sis (L.) Osbeck), pummelo ( C. grandis (L.) Osbeck), or grape-fruit ( C. paradisi Macfad.) cultivars (Hodgson, 1967). Swingle and Reece (1967) noted that mandarins include both monoembryonic and polyembryonic cultivars, as well as self-fertile and self-incompatible types. These authors recognized only three mandarin species and assumed that

This work was supported by the Brazilian agencies Conselho Nacional de De-senvolvimento Científico e Tecnológico – CNPq (n. 474589/03-0, 307916/03-0, and 140273/04-2) and Fundação de Amparo à Ciência e Tecnolo-gia – FACEPE (Edt. 0005-05-03/04).

Accepted in revised form for publication by B. Friebe, 4 September 2007.

Cytogenet Genome Res 119:275–281 (2007)276

4�,6-diamidino-2-phenylindole (DAPI), which binds pref-erentially to AT-rich DNA. The chromosomal distribution of CMA + bands in Citrus species allowed the characteriza-tion of seven different chromosome types (Carvalho et al., 2005), five of which were found in most cultivars. Chromo-some types A (two terminal and one proximal band), B (one proximal and one terminal), C (two terminal), E (one sub-terminal/interstitial band) and G (one terminal and one subterminal band) are relatively rare, whereas D (one termi-nal) and F (without bands) types are the most common.

The karyotype analysis of almost 100 accessions revealed that hybrids, such as grapefruits and tangelos, characteristi-cally present strong heteromorphism for one to three chro-mosome pairs, whereas those of non-hybrid origin, as citron (C. medica) and wild species from several Citrus related gen-era, are homozygous or only exhibit small heteromorphism for CMA + bands (Guerra et al., 2000; Carvalho et al., 2005; Brasileiro-Vidal et al., 2007; Moraes et al., 2007). Therefore, it is assumed that heteromorphism for chromosome types in Citrus species is a strong indication of hybridism. Befu et al. (2001, 2002), Cornélio et al. (2003), Yamamoto and Tom-inaga (2003), and Yamamoto et al. (2005) analyzed the CMA banded karyotype of 23 different mandarin cultivars and found that most of them were heterozygous. Cornélio et al. (2003) identified two mandarin groups which were homo-zygous and could represent true species. One group was rep-resented by two wild species, C. sunki (Hayata) Hort. ex Tan. and C. reshni Hort. ex Tan., both with the simplest karyo-type formula 14D + 4F. The second group was formed by C. deliciosa Ten. cv. Rio, Montenegrina, and Comum and C. reticulata cv. Cravo, all of them with 2C + 10D + 6F.

A further refinement of the karyotype analysis of Citrus species was obtained when CMA banded metaphases were sequentially hybridized with 5S and 45S rDNA probes by FISH (Fluorescent In Situ Hybridization). Although Citrus species have only a few rDNA sites, they are very useful for distinguishing chromosome subtypes (Pedrosa et al., 2000; Carvalho et al., 2005; Moraes et al., 2007). In this work, the distribution of 5S and 45S rDNA sites was investigated in ten mandarin species and one synthetic hybrid, according to Tanaka’s system, aiming to identify cytologically homo-zygous groups and to get further insight on the relationship among mandarins and between mandarins and other citrus species.

Material and methods

Seedlings or adult plants of ten mandarin species, represented by 12 cultivars ( C. sunki, C. reshni, C. reticulata cv. Cravo and Ponkan, C. deliciosa cv. Comum and Montenegrina, C. tangerina Hort. ex Tan. cv. Dancy, C. clementina Hort. ex Tan., C. unshiu Marcow., C. ambly-carpa (Hassk.) Ochse, C. tachibana (Makino) Tanaka, C. depressa Haya-ta), and the tangor cv. Murcott (C. sinensis ! C. reticulata) were ana-lyzed. In order to avoid working with zygotic embryos, five root tips from seedlings of each accession were analyzed individually and only those with identical karyotype formulae in at least three root tips were accepted for analysis as nucellar embryos (see also Cornélio et al., 2003; Carvalho et al., 2005). All accessions were from the Citrus Active Germ-plasm Bank of the Embrapa Cassava & Fruits, Cruz das Almas, Bahia.

Root tips were pre-treated with 8-hydroxyquinoline (0.002 M) for 5 h at 18 ° C, fixed in absolute alcohol/glacial acetic acid (3: 1, v/v) for 2–24 h at room temperature and stored at –20 ° C. Afterwards, they were washed three times in distilled water and digested in a 2% cellu-lase (Onozuka) and 20% pectinase (Sigma) solution for 2 h at 37 ° C. Meristems were squashed in a drop of 45% acetic acid and frozen in liquid nitrogen. The slides were stained with DAPI (2 � g/ml):glycerol (1: 1, v/v) solution to allow selection of the best cells. Subsequently, they were destained in ethanol:glacial acetic acid (3: 1, v/v) for 30 min at room temperature and transferred to absolute ethanol overnight at 10 ° C. Slides were air dried and aged for three days at room tempera-ture.

The aged preparations were stained with CMA (0.5 mg/ml) for 1 h and counterstained with DAPI (1 � g/ml) for 30 min. The best meta-phases were photographed with a Cohu camera using the QFISH soft-ware (Leica). The slides were destained again and stored at –20 ° C.

D2 probe from Lotus japonicus and R2 probe from Arabidopsis thaliana were used to localize 5S and 45S rDNA sites, respectively. The D2 probe was labeled with biotin-16-dUTP (Roche) or Cy3 (Invitro-gen) and the R2 probe was labeled with digoxigenin-11-dUTP (Life Technologies), both by nick translation. The FISH procedure was based on Jiang et al. (1995) with minor modifications. The hybridization mixture contained 60% formamide (v/v), 5% dextran sulfate (w/v), 2 ! SSC, and 2–5 ng/ � l of probe. After overnight hybridization, slides were washed in 2 ! SSC at room temperature for 5 min, 20% formamide in 0.1 ! SSC at 37 ° C for 15 min, and two washes in 2 ! SSC at 37 ° C for 5 and 10 min, respectively. The biotin labeled probe was detected with mouse anti-biotin (Dako) and the signal amplified with rabbit anti-mouse TRITC conjugate (Roche). The digoxigenin probe was detected with sheep anti-digoxigenin FITC conjugate (Dako) and amplified with rabbit anti-sheep FITC conjugate (Roche). All slides were coun-terstained with DAPI (1 � g/ml) and mounted in Vectashield (Vector). Metaphase images were acquired as indicated before.

Results

The accessions analyzed showed 2n = 18 chromosomes which could be differentiated by the size and distribution of CMA + /DAPI – bands. Chromosome types A to G were found ( Figs. 1–3 ). The largest F chromosome pair, also referred to as F L , was identified in all accessions and displayed fine ter-minal dot-like bands (F L + ) only in C. sunki and C. reshni ( Fig. 1 A).

The CMA + band pattern allowed the separation of the 13 accessions into the following four groups, according to Cornélio et al. (2003): I) only D and F chromosome types ( Figs. 1 A and 3A); II) only C, D, and F chromosome types ( Figs. 1 C, E, and 3B, C); III) with four or five different chro-mosome types, except A ( Figs. 1 F, 2A, C, D, F, and 3D–H); and IV) with one A chromosome type ( Figs. 2 H and 3I).

In some accessions, the marker chromosomes showed small heteromorphism in CMA + band size and position. The proximal CMA band of B chromosomes was often less brilliant or neutral (CMA 0 , insert in Fig. 2 C). In C. unshiu , one of the C chromosomes had a terminal band larger than the other ( Fig. 2 A). The position of the single CMA + band of the E chromosomes seemed to be more subterminal in C. amblycarpa ( Fig. 2 C), whereas in C. tachibana ( Fig. 2 D) and C. depressa ( Fig. 2 F) both E chromosomes were ob-served.

The 5S rDNA sites varied in size within the karyotype of several cultivars (see Figs. 1G, 2C, I), especially in group III.

Cytogenet Genome Res 119:275–281 (2007) 277

In all accessions, the 5S rDNA sites were localized in the euchromatic region, whereas 45S rDNA sites were colocal-ized with CMA + /DAPI – or CMA 0 /DAPI – bands ( Figs. 1 and 2). All A and B chromosome types possessed a 45S rDNA site co-localized with the proximal CMA band ( Figs. 1 F, G, 2, and 3) and were identified as A/45S and B/45S, respec-tively. On the other hand, only one or two D chromosomes per karyotype had a 45S rDNA site co-localized with the CMA band. In these cases, there was always a 5S rDNA ad-jacent to a 45S rDNA site and the chromosomes were identi-fied as D/5S-45S. Similarly, the F chromosome bearing a 5S rDNA site, found only in tangor ‘Murcott’, was denominat-ed F/5S ( Fig. 2 I). The 45S rDNA site of the D/5S-45S chro-mosome was frequently distended and, in this case, its CMA band was usually neutral ( Figs. 1 C–D and F–G). The 45S rDNA site on B chromosomes was also distended, some-times exhibiting a CMA 0 band ( Fig. 2 C).

The distribution of 5S and 45S rDNA sites was in agree-ment with the previous division in groups ( Table 1 and Fig. 3 ). The accessions of group I and II displayed one D/5S-45S chromosome pair ( Fig. 1 B, D, E), while those of group III showed one pair of D/5S-45S and a single B/45S

chromosome ( Figs. 1 G and 2B, C, E, G). The only acces-sion of group IV, tangor ‘Murcott’, was more complex, pos-sessing one A/45S, one B/45S, one D/5S-45S, and one F/5S ( Fig. 2 H, I).

Discussion

The karyotype formulae Most samples investigated here were heterozygous for

CMA bands, as previously reported by Cornélio et al. (2003) and Yamamoto and Tominaga (2003). The karyotype for-mula of C. amblycarpa was described here for the first time. All other accessions analyzed had CMA banding patterns very similar to those described by Befu et al. (2001, 2002), Cornélio et al. (2003), Yamamoto and Tominaga (2003), and Yamamoto et al. (2005).

Nevertheless, in a few accessions there were some appar-ent contradictions between the present results and the pre-vious reports. The karyotype formulae described here and by Cornélio et al. (2003) for C. sunki and C. tangerina were 14D + 4F and 1C + 11D + 6F, respectively, whereas Yama-

A B

C D

E F G

Fig. 1. CMA + band pattern and distribu-tion of 5S and 45S rDNA sites in mandarins of groups I, II, and C. clementina . ( A , B ) Citrus reshni . ( C , D ) C. reticulata cv. Cravo. ( E ) C. tangerina cv. Dancy. ( F , G ) C. clementina . A , C , and F show CMA stained metaphases, while B , D and G show the same cells counterstained with DAPI and hybridized with 5S (pink) and 45S (green) rDNA probes. E shows CMA + im-age merged with DAPI, Cy3 (5S rDNA) and FITC (45S rDNA) images. Capital letters in C , E , and F indicate B and C chromosome types. Arrows point out the chromosomes bearing 5S and 45S rDNA sites. Asterisk in A indicates the F L + . Inserts in E show D/5S-45S chromo-somes without FISH signals. Insert in G shows a D/5S-45S chromosome without 45S rDNA signal (yellow arrow). Bar = 5 � m.

Cytogenet Genome Res 119:275–281 (2007)278

A B C

D E

F G

H I

Fig. 2. CMA + band pattern and distribu-tion of 5S and 45S rDNA sites in mandarins. ( A , B ) C. unshiu . ( C ) C. amblycarpa. ( D , E ) C. tachibana . ( F , G ) C. depressa . ( H , I ) Tangor ‘Murcott’. A , D , F , and H show CMA + banded cells, while B , E , G , and I show the same cells counterstained with DAPI and hybridized with 5S (pink) and 45S (green) rDNA probes. In C , the CMA + image was merged with DAPI, Cy3 (5S rDNA) and FITC (45S rDNA) images. Capital letters in A , C , D , F , and H indicate rare chromosome types. Arrows point out the chromosomes bearing 5S and 45S rDNA sites. Inserts in C and G show the B/45S chromo-somes stained with CMA (yellow)/DAPI (blue) and with FITC, respectively. Asterisk in D indicates a partial overlapping between an E and a D chromosome. Bar = 5 � m.

Table 1. List of mandarin accessions investigated, grouped according to Cornélio et al. (2003), with respective karyotype formula and number and position of rDNA sites. Species name is according to Tanaka (1969).

Group Species Karyotype formula Chromosome types bearing rDNA sites

I Citrus reshni Hort. ex Tanaka 14D + 4F 2D/5S-45SC. sunki (Hayata) Hort. ex Tanaka 14D + 4F 2D/5S-45S

II C. deliciosa Ten. cv. Comum/Montenegrina 2C + 10D + 6F 2D/5S-45SC. reticulata Blanco cv. Cravo 2C + 10D + 6F 2D/5S-45SC. tangerina Hort. ex Tanaka cv. Dancy 1C + 11D + 6F 2D/5S-45S

III C. amblycarpa (Hassk.) Ochse 1B + 11D + 1E + 4F + 1G 1B/45S + 2D/5S-45SC. clementina Hort. ex Tanaka 1B + 1C + 11D + 5F 1B/45S + 2D/5S-45S C. reticulata Blanco cv. Ponkan 1B + 1C + 11D + 5F 1B/45S + 2D/5S-45SC. depressa Hayata 1B + 1C + 10D + 2E + 4F 1B/45S + 2D/5S-45SC. tachibana (Makino) Tanaka 1B + 1C + 10D + 3E + 3F 1B/45S + 2D/5S-45SC. unshiu Marcow 1B + 2C + 10D + 5F 1B/45S + 2D/5S-45S

IV Tangor (C. sinensis (L.) Osb. × C. reticulata Blanco) ‘Murcott’ 1A + 1B + 10D + 6F 1A/45S + 1B/45S + 1D/5S-45S + 1F/5S

Cytogenet Genome Res 119:275–281 (2007) 279

moto and Tominaga (2003) reported 12D + 6F for C. sunki and 1C + 10D + 7F for C. tangerina (following the nomen-clature of Carvalho et al., 2005). In both cases it seems that one or two D chromosomes were interpreted as F types by Yamamoto and Tominaga (2003). These differences may be attributed to the staining procedure used since the latter authors did not use DAPI counterstaining, although the possibility of band heteromorphism cannot be discarded. Citrus species did not display DAPI + bands, but the use of DAPI is very important as counterstaining since it intensi-fies the contrast of CMA + bands and allows the identifica-tion of CMA 0 bands as DAPI – bands (Cornélio et al., 2003; Carvalho et al., 2005). Furthermore, the CMA bands co-lo-calized with 45S rDNA may exhibit different brightness in-tensity, from positive to neutral, even in different cells on the same slide, but they always were DAPI – . Therefore, if DAPI counterstaining is not used, it is possible that one or two D/5S-45S chromosomes look like F chromosomes.

Similarly, the karyotype of C. deliciosa described here and by Cornélio et al . (2003) as 2C + 10D + 6F seemed to be the same reported by Yamamoto and Tominaga (2003) as 1C + 10D + 7F, where one C and one D chromosome were iden-tified as D and F, respectively. Indeed, Cornélio et al. (2003) observed that the C chromosomes of C. deliciosa had two

CMA + bands of different sizes, but one of them was small and pale, making this chromosome similar to a D chromo-some. Therefore, the karyotype of C. deliciosa seems to be stable in at least four cultivars: Rio, Montenegrina, Comum (Cornélio et al., 2003; and present report) and Tardivo di Ciaculli (Yamamoto and Tominaga, 2003).

The karyotype of C. reticulata cv. Ponkan was described by Cornélio et al. (2003) as having a small and weak termi-nal CMA + band on a C chromosome, which was not found either by Yamamoto and Tominaga (2003), nor here. Actu-ally, the karyotype of ‘Ponkan’ was identical to that of C. clementina . The similarity between them was also point-ed out by Federici et al. (1998) based on RFLP and RAPD markers.

The tangor ‘Murcott’ showed a heterozygous karyotype, as expected for a synthetic hybrid. However, the karyotype formula indicates that ‘Murcott’ is not a primary hybrid be-tween C. sinensis ! C. reticulata , as commonly admitted (Hodgson, 1967), since none of these species display the A chromosome type observed in ‘Murcott’. Probably this A chromosome was inherited from a citrus accession related to C. grandis , which characteristically bear A chromosomes (Moraes et al., 2007).

Fig. 3. Idiogram representing the distribution of CMA + bands and 5S and 45S rDNA sites in the cultivars analyzed. The karyotype formula is indicated under the idiogram. Black blocks = CMA + bands. Arrowheads = 5S rDNA sites. Hatched circles = 45S rDNA sites.

Cytogenet Genome Res 119:275–281 (2007)280

References

Befu M, Kitajima A, Hasegawa K: Chromosome composition of some Citrus species and culti-vars based on the chromomycin A 3 (CMA) banding patterns. J Japan Soc Hort Sci 70: 83–88 (2001).

Befu M, Kitajima A, Hasegawa K: Classification of the Citrus chromosomes with same types of chromomycin A banding patterns. J Japan Soc Hort Sci 71: 394–400 (2002).

Brasileiro-Vidal AC, dos Santos-Serejo JA, Soares Filho WS, Guerra M: A simple chromosomal marker can reliably distinguish Poncirus from Citrus species. Genetica 129: 273–279 (2007).

Carvalho R, Soares Filho WS, Brasileiro-Vidal AC, Guerra M: The relationship among lemons, limes and citron: a chromosomal comparison. Cytogenet Genome Res 109: 276–282 (2005).

Coletta Filho HD, Machado MA, Targon MLPN, Moreira MCPQDG, Pompeu Jr J: Analysis of the genetic diversity among mandarins ( Citrus spp .) using RAPD markers. Euphytica 102: 133–139 (1998).

Cornélio MTMN, Figueirôa ARS, Santos KGB, Carvalho R, Soares Filho WS, Guerra M: Chro-mosomal relationships among cultivars of Cit-rus reticulata Blanco, its hybrids and related species. Plant Syst Evol 240: 149–161 (2003).

The distribution of rDNA sites Among the accessions investigated, only C. sunki, C. resh-

ni , both cultivars of C. deliciosa, C. reticulata cv. Cravo, and C. tangerina cv. Dancy, displayed a single, homozygous pair of D/5S-45S chromosomes. These accessions were separat-ed by Cornélio et al. (2003) into group I, constituted only by C. sunki and C. reshni , with the basic karyotype formula 14D + 4F, and group II, represented here by C. deliciosa and C. reticulata cv. Cravo, with 2C + 10D + 6F. In C. tangerina cv. Dancy, one C chromosome was substituted by a D type, probably because one terminal band of the C chromosome was too small. This cultivar was also reported to be closely related to C. deliciosa by Nicolosi et al. (2000), using RAPDs and SCARs. All accessions of group III displayed a pair of D/5S-45S and a single B/45S chromosome while the acces-sion of group IV exhibited a complex composition of 5S and 45S rDNA sites. In a previous paper (Brasileiro-Vidal et al., 2007) we commented that two mandarin cultivars had a single B/5S-45S. However, these data were not confirmed here. The false signal was most probably caused by a con-taminated probe. Until now, the B/5S-45S has only been found in Poncirus trifoliata (L.) Raf. (Brasileiro-Vidal et al., 2007) and C. hystrix DC. (unpublished data).

Among Citrus species, the D/5S-45S chromosome was found to be homozygous in all mandarin accessions ana-lyzed, except in the tangor ‘Murcott’. Noteworthy, it was not found to be homozygous in any of the other citrus acces-sions analyzed previously. Therefore, the D/5S-45S chro-mosome type seems to be a good chromosome marker for this group. Two other true species of Citrus recognized by most authors (reviewed by Moore, 2001), C. grandis and C. medica , were also cytologically characterized by exclu-sive chromosome markers to be homozygous: A/45S and F/5S chromosomes in the former (Moraes et al., 2007) and D/5S in the latter (Carvalho et al., 2005).

The origin of mandarins Although Swingle and Reece (1967) considered only

three mandarin species ( C. reticulata, C. tachibana and C. indica Tanaka), from the cytological point of view C. reticulata sensu Swingle is comprised of at least three true species (C. sunki, C. reshni, and C. deliciosa) and many hy-brids. Citrus sunki , named by Swingle and Reece (1967) as C. reticulata var. austera, and C. reshni can be differentiated by morphological characters and geographical distribution, but cytologically and molecularly they seem to be very sim-

ilar (Cornélio et al., 2003). Out of the two other mandarin species (C. tachibana and C. indica) accepted by Swingle and Reece (1967), only the former is cytologically known and it was found here to be highly heteromorphic (see also Yama-moto and Tominaga, 2003) and probably of hybrid origin.

Comparing the assumed true species of mandarin, one is tempted to speculate that the acid and small fruited C. sunki and C. reshni were less probably involved in the origin of the sweet, edible accessions. Moreover, most man-darins analyzed had C type chromosomes, except C. sunki and C. reshni . On the other hand, C. deliciosa has a very at-tractive fruit and is the only mandarin species that displays a homozygous C chromosome. Therefore, it is a more likely candidate to be involved in the origin of many mandarin hybrids. Citrus tangerina cv. Dancy and C. reticulata cv. Cravo, considered by Tanaka (1969) as distinct species, seem to be variations of the Mediterranean mandarins. Accord-ing to Hodgson (1967), C. tangerina cv. Dancy could have been introduced from Tangiers (Morocco) into Florida (United States) in 1867, whereas the Brazilian C. reticulata cv. Cravo seems to be related to the Portuguese ‘Carvalhais’, a putative natural hybrid of the Mediterranean mandarin ‘Setubalense’. Therefore, both cultivars may have originated from other varieties cultivated in the Mediterranean region. The origin of the Mediterranean mandarin itself is unclear, but most likely it originated from a Chinese mandarin (Hodgson, 1967).

On the other hand, C. sunki and C. reshni may be in-volved in the origin of other acid hybrids. For example, C. jambhiri Lush., the rough lemon, is supposed to be a hy-brid between mandarin and C. medica (Scora, 1975; Potvin et al., 1983; Federici et al., 1998). Indeed, Carvalho et al. (2005) observed that one half of the rough lemon karyotype is identical to the haploid chromosome complement of C. medica , inheriting B/45S and D/5S chromosomes of this species. The other half chromosome set perfectly matches the haploid complement of C. reshni , including a D/5S-45S chromosome type. The same applies to Rangpur lime ( C. li-monia Osbeck), also a putative citron ! mandarin hybrid (Nicolosi et al., 2000), with half of the chromosome comple-ment similar to C. medica and the other half similar to C. reshni . Therefore, the D/5S-45S chromosome may be a good chromosome marker to test the supposed involvement of mandarin germplasm in the origin of several Citrus hy-brids. In addition, the B chromosome seems to be a further cytological marker in most mandarin hybrids.

Cytogenet Genome Res 119:275–281 (2007) 281

Fang DQ, Roose ML: Identification of closely re-lated Citrus cultivars with inter-simple se-quence repeat markers. Theor Appl Genet 95: 408–417 (1997).

Federici CT, Fang DQ, Scora RW, Roose ML: Phy-logenetic relationships within the genus Citrus (Rutaceae) and related genera as revealed by RFLP and RAPD analysis. Theor Appl Genet 96: 812–822 (1998).

Guerra M: Cytogenetics of Rutaceae. V. High chro-mosomal variability in Citrus species revealed by CMA/DAPI staining. Heredity 71: 234–241 (1993).

Guerra M, Santos KGB, Silva AEB, Ehrendorfer F: Heterochromatin banding patterns in Rutace-ae-Aurantioideae – A case of parallel chromo-somal evolution. Am J Bot 87: 735–747 (2000).

Hodgson RW: Horticultural varieties of Citrus , in Reuther W, Weber HJ, Batchelor LD (eds): The Citrus Industry. History, World Distribution, Botany and Varieties, Vol I, pp 431–591 (Uni-versity of California Press, Berkeley 1967).

Jiang J, Gill BS, Wang GL, Ronald PC, Ward DC: Metaphase and interphase f luorescence in situ hybridization mapping of the rice genome with bacterial artificial chromosomes. Proc Natl Acad Sci USA 92: 4487–4491 (1995).

Luro F, Laigret F, Bové JM, Ollitrault P: DNA ampli-fied fingerprinting, a useful tool for determina-tion of genetic origin and diversity analysis in Citrus . Hort Sci 30: 1063–1067 (1995).

Moore GA: Oranges and lemons: clues to the tax-onomy of Citrus from molecular markers. Trends Genet 17: 536–540 (2001).

Moraes AP, Soares Filho WS, Guerra M: Karyotype diversity and the origin of grapefruit. Chromo-some Res 15: 115–121 (2007).

Nicolosi E, Deng ZN, Gentile A, La Malfa S, Conti-nella G, Tribulato E: Citrus phylogeny and ge-netic origin of important species as investigated by molecular markers. Theor Appl Genet 100: 1155–1166 (2000).

Pedrosa A, Schweizer D, Guerra M: Cytological het-erozygosity and the hybrid origin of sweet or-ange ( Citrus sinensis (L.) Osbeck) Theor Appl Genet 100: 361–367 (2000).

Potvin C, Bergeron Y, Simon JP: A numerical taxo-nomic study of selected Citrus species (Ruta-ceae) based on biochemical characters. Syst Bot 8: 127–133 (1983).

Scora RW: IX. On the history and origin of Citrus . Bull Torrey Bot Club 102: 369–375 (1975).

Swingle WT, Reece PC: The botany of Citrus and its wild relatives, in Reuther W, Weber HJ, Batch-elor LD (eds): The Citrus Industry. History, World Distribution, Botany and Varieties, vol I, pp 190–430 (University of California Press, Berkeley 1967).

Tanaka T: Misunderstanding with regard to Citrus classification and nomenclature. Bull Uni Osa-ka Pref Ser B 21: 139–145 (1969).

Yamamoto M, Tominaga S: High chromosomal variability of mandarins ( Citrus spp.) revealed by CMA banding. Euphytica 129: 267–274 (2003).

Yamamoto M, Kubo T, Tominaga S: CMA banding patterns of chromosome of mid- and late-ma-turing citrus and acid citrus grown in Japan. J Japan Soc Hort Sci 74: 476–478 (2005).