| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2 Universidade Federal de Pernambuco, CCB, Departamento de Botânica, Recife, PE, Brazil; and 3 Institute of Botany and Botanical Garden, University of Vienna, Rennweg 14, A-1030 Vienna, Austria
Received for publication December 3, 1998. Accepted for publication August 27, 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: chromosomal parallel evolution • Citrus • CMA/DAPI staining • heterochromatin banding patterns • Rutaceae-Aurantioideae • systematics
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
).
According to Swingle and Reece (1967)
, the subfamily comprises 33 genera and ~200 species. The first general review of the subfamily was presented by Engler (1931)
who recognized a single tribe Aurantieae with two subtribes, Hesperethusinae and Citrinae. Later, Swingle (updated by Swingle and Reece, 1967
, catalog: Carpenter and Reece, 1969
), on the basis of morphological and anatomical analyses, divided the subfamily into two tribes: Clauseneae (I), composed of subtribes Micromelinae (IA), Clauseninae (IB), and Merrilliinae (IC), and Citreae (II) with the subtribes Triphasiinae (IIA), Citrinae (IIB), and Balsamocitrinae (IIC). Since then, many authors have used other approaches to clarify the relationships within Aurantioideae, mainly by analyzing secondary metabolites (e.g., Grieve and Scora, 1980
; Waterman, 1983, 1990
; Da Silva, Gottlieb, and Ehrendorfer, 1988
), but also isozymes and other proteins (Esen and Scora, 1977
; Torres, Soost, and Mau-Lastovicka, 1982
; Iwamasa and Nito, 1988
; Fang, 1993a, b
; Fang, Zhang, and Xiao, 1993
; Zhong and Ye, 1993
). In spite of some new suggestions the concept of the subfamily by Swingle and Reece (1967) is still accepted and in current use. Nevertheless, recent analyses of cpDNA (Samuel et al., 1999a, b,
and unpublished data) will necessitate considerable systematic changes in the future.
Karyosystematic analysis of the Aurantioideae was first discouraged by observations that nearly all its taxa have small chromosomes of similar size and morphology (Krug, 1943
). Only subsequent karyotype analysis with C-banding revealed that most of the chromosomes had heterochromatic blocks that allow a better differentiation (Guerra, 1985
; Liang, 1988
; Wei, Cheng, and Duan, 1988
). The distribution of C-bands was similar to the pattern of heteropycnotic regions observed by conventional Giemsa or Feulgen staining in prometaphase chromosomes (Guerra, 1985
). More recently, it was demonstrated for some representatives of Citrus that their chromosomes, when stained simultaneously with the fluorochromes chromomycin A3 (CMA) and 4'-6-diamidino-2-phenylindol (DAPI), revealed a variable number of regions that appeared bright or positive with CMA, and faint or negative with DAPI. Chromomycin binds preferentially to sequences of repetitive DNA rich in cytosine-guanine, whereas DAPI has a reverse affinity, binding preferentially to adenine-thymine rich sequences (Schweizer, 1976
). The resulting fluorochrome banding allowed for an even better karyotype characterization than did previous methods (Guerra, 1993
).
Improved chromosomal analyses of Aurantioideae also have important implications for the citrus breeding program. Representatives of several wild relatives of Citrus, including such distantly related genera, as Murraya, Severinia, Atalantia, and Swinglea, have been hybridized with cultivated Citrus species in order to introduce desirable traits, mainly resistance to pests and pathogens (Barrett, 1977
; Motomura et al., 1995
). Since many species of this group have potentials for nucellar embryony and polyembryony (Cameron and Frost, 1968
), the genetic identification of intervarieties or intergeneric hybrids plays a fundamental role and is mainly done by isoenzyme profiles (Torres, Soost, and Mau-Lastovicka, 1982
). The chromosomal identification of different genomes may be an additional and simple method of identifying citrus hybrids and is thus of importance for future work on substitution lines.
In the present work we concentrate on the cytogenetical characterization of the most commonly used non-citrus species of Aurantioideae, mainly by CMA/DAPI staining, and investigate the heterochromatin patterns of 17 species from 15 genera, representing both tribes and five of the six subtribes of the subfamily. Discussing our results against the background of available morphological, phytochemical, and DNA-analytical data we hope to contribute to the chromosomal evolution of the Aurantioideae and to improve the phylogenetic classification of the subfamily.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and Carpenter and Reece (1969)
, numbered as outlined in the introduction, and listed with provenances, chromosome numbers, and cell sources. All tribes and subtribes of Aurantioideae are represented with the exception of the monogeneric Micromelinae (IA). Seeds from open-pollinated mother plants were germinated in petri dishes, and seedlings were cultivated in our experimental garden at the Federal University of Pernambuco, in Recife. Seeds of Poncirus and Citrus were highly polyembryonic, but since root tips of different embryos were analyzed and exhibited the same banding pattern we believe that in those cases we have worked with nucellar rather than zygotic embryos. Seed samples of other genera were nearly exclusively monoembryonic, and their embryos were probably zygotic. Adult plants were grown in pots. For cytological analyses, mostly actively growing root tips from seedlings and some adult individuals were collected. Voucher specimens are deposited in the herbarium of the respective institutions, and copies of most of them are maintained in the herbarium of the Federal University of Pernambuco (UFP).
|
Slide preparation for fluorochrome staining follows Schweizer (1976)
, described in detail by Deumling and Greilhuber (1982)
. For CMA/DAPI staining the meristems were washed twice in distilled water (10 min each), digested with a 2% cellulase-10% pectinase solution (1.5 h), and squashed in 45% acetic acid. After coverslip removal the slides were aged for 3 d, stained with CMA (1 h), counterstained with DAPI (30 min), and mounted in McIlvaine's (pH 7.0) buffer-glycerol v/v 1:1. For conventional staining the root tips were hydrolyzed in 5mol/L HCl (20 min) and squashed in a drop of 45% acetic acid. The coverslip was removed in liquid nitrogen, and the slide was stained in 2% Giemsa in Sorensen buffer (20 min) and mounted in Euparal (Guerra, 1983
).
In order to estimate the variation in chromosome complement size and CMA+ heterochromatin proportion, a sample of six species (Table 2) were chosen for chromosome measurements. A hexaploid cytotype of Glycosmis pentaphylla agg. and one diploid representative of each subtribe studied were included. Swinglea glutinosa and Murraya paniculata were also chosen because they presented, respectively, the smallest and the largest chromosomes among all the species analyzed. CMA+ blocks and chromosome size were measured from chromosome drawings of amplified projections of film negatives of the best five metaphases of each species. In a few species, with different chromosome number and heterochromatin amount (Glycosmis pentaphylla agg., Clausena excavata, Severinia buxifolia, Eremocitrus glauca, and Swinglea glutinosa), the prophase condensation pattern was analyzed for comparison with the CMA banding pattern.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
After conventional Giemsa staining, prophase chromosomes exhibited a very similar condensation pattern in all species investigated. Proximal heteropycnosis was observed in most chromosomes and terminal heteropycnotic blocks were present in those species that possessed terminal blocks of CMA+ heterochromatin (Figs. 1–3). The nuclear interphase structure in the whole subfamily was always of the areticulate type, with a variable number of chromocentres. Two types of chromocentres were observed: a smaller number of deeply stained, large, well-defined chromocentres and a larger number of smaller, less densely stained ones (Figs. 1–2); the latter were more clearly observed in meristematic cells. The number of larger chromocentres in each species was proportional to the number of heterochromatic blocks observed after CMA staining.
|
Species differ widely in terms of chromosome size, number, and distribution of heterochromatin blocks positive to CMA. The hexaploid G. pentaphylla exhibits a mean chromosome size proportional to the chromosome size of the diploid species. Size variation between the largest and the smallest chromosomes of each karyotype was very small in absolute values (one or two micrometres), although proportionally this represents a variation of 100–200%. Variation in chromosome complement length also was relatively small and was far exceeded by variation in heterochromatin amount (Table 2).
The chromosome banding with CMA appeared more pronounced in two or three of the largest pairs, where generally a higher amount of heterochromatin was concentrated. Medium and small-sized pairs showed an apparent greater similarity in banding pattern due to their smaller number of bands. A noteworthy feature was the presence of a large chromosome pair common to practically all species, independent of the heterochromatin amount within the karyotype. This largest or second largest pair, metacentric, without bands or only with very small telomeric ones, is here referred to as FL (large chromosome of the F type described for Citrus by Guerra, 1993
).
In the schematic idiograms (Table 3) no differences in ploidy, chromosome size, and centromeric position within and between the genomes of the various taxa are shown in order to facilitate comparison. Only the location and relative size of CMA+ blocks are demonstrated. When two or more different banding patterns were observed for a single chromosome pair, only that with the highest amount of heterochromatin was shown. The following detailed specific presentation of our karyological findings follow the taxonomic classification of Swingle and Reece (1967)
and the (sub)tribal numbers outlined in the introduction.
|
|
Another species of Clausena, C. excavata, had 2n = 36 and differed by a larger number of small CMA+ bands, most of them very small. One chromosome pair had a terminal band in each telomere region, six pairs were with a single terminal CMA+ block, and 11 pairs were without bands (Fig. 7), including the two largest ones (FL type).
Two species of Murraya (M. koenigii and M. paniculata) were analyzed, both with 2n = 18, but with very distinct karyotypes. Murraya koenigii had less contrasted bands, terminally located on pairs II and either on IV or V (Figs. 8–9), which were associated with the nucleolus in prophase and interphase cells. In some cells, fine proximal bands in pair II and terminal ones in pair I (FL type) were observed.
On the other hand, Murraya paniculata was the species with the largest chromosomes (Table 2) and the highest number of bands in the tribe Clauseneae (Figs. 10–11). CMA+ blocks were found in all chromosomes except in one of the smallest pairs. The FL chromosome type corresponded to one of the three largest pairs, showing a small but always visible telomeric CMA+ block. The other two large pairs were far richer in heterochromatin. One of them presented a very brilliant band in each telomeric region, whereas the other showed a large telomeric and a small proximal heterochromatic block. The latter sometimes appeared decondensed like a secondary constriction. Each of the remaining five pairs exhibited a single large terminal band. The analysis of four provenances from far distant locations did not show any karyotype variation.
Merrillia caloxylon, the only representative of subtribe Merrilliinae (IC), showed a heterochromatin-rich karyotype. Bands were found in six of its nine chromosome pairs. Pair I (FL) with a small terminal band, pair II with large CMA+ blocks in both telomeres, and pair III with a large terminal block as well as a proximal band could be distinguished. The latter band exhibited heterozygotic variation in size and strongly quenched fluorescence with DAPI; in some metaphases it appeared finer and distended as a secondary constriction. In some cells, analyzed with DAPI or Giemsa, pair III seemed to be split in two (Figs. 12–13). The other three larger chromosome pairs showed single major terminal bands, whereas the three smaller were wholly euchromatic.
|
Seven of the 13 genera of subtribe Citrinae (IIB) were analyzed. The least amounts of heterochromatin were found in Severinia and Atalantia, genera with rather similar banding patterns, both with pair I of the FL type. In S. buxifolia, CMA+ blocks were restricted to the three pairs II, III, and IV, always located at terminal positions (Figs. 16–17). When the nucleolus was visible in prometaphases, it was always associated with two or three of these blocks. In some cells, pair I showed a very small terminal CMA+ region. Two samples of S. buxifolia (BGRJ and CNPMF) did not exhibit karyotype differences. A third, received as "Atalantia buxifolia" (BGV), differed by bands in both telomeric regions of pair II (Figs. 18–19). Atalantia monophylla had four pairs showing CMA+ blocks on the long arms of pairs II, III, IV, and V (Figs. 20–21).
The other genera analyzed from this subtribe (the group of "true Citrus fruit trees" according to Swingle and Reece, 1967
; also called "Citrus clade" by Samuel et al., 1999a, b,
and unpublished data) had karyotypes with a much higher heterochromatin content. Eremocitrus glauca showed bands on all chromosomes except the smallest pair. One of the largest pairs appeared with CMA+ blocks in both telomeres, whereas the others had only single terminal blocks (Figs. 22–23). This species was the only one in the tribe where the FL chromosome was not identifiable.
|
Citrus reticulata showed only three chromosome pairs without bands, including one of the FL type. Five pairs exhibited a single terminal CMA+ block, conspicuously large in three of them, and one pair with bands in both telomeres (Figs. 26–27). In some cells, one of the single banded chromosomes showed an additional heterochromatic block in the opposite telomere.
Fortunella crassifolia exhibited the largest number of banded chromosomes per karyotype observed in the whole subfamily, with CMA+ telomeric blocks in each of its chromosomes. In two of them, the bands were found in both telomeric regions, and in some cells one of these pairs exhibited an additional proximal band, stained slightly more with CMA than the euchromatin. When observed with DAPI, this band was strongly negative, giving the impression that each homologue chromosome was divided in two (Figs. 28–29). The other chromosome pairs displayed terminal bands of very variable size. The largest chromosome pair was of the FL type, showing a single, very fine but constant terminal band.
In Microcitrus australasica we were not able to obtain well-condensed metaphases from the meristematic tissue of young leaves, although it was possible to identify the banding pattern in many prometaphases (Figs. 30–31). One chromosome pair had terminal bands in both telomeres, six pairs had single terminal bands of variable size, whereas two pairs were wholly euchromatic, including the largest one. In some cells, weak proximal or terminal bands were additionally found in some of the single banded chromosomes.
Of the seven genera that make up subtribe Balsamocitrinae (IIC), only two, Swinglea and Feroniella, are represented in this study. Swinglea glutinosa showed a very distinct, heterochromatin-poor karyotype with CMA+ blocks restricted to one of the telomere regions of pair III (Fig. 32). At interphase, these heterochromatic blocks were found associated with the nucleolus. This species had the smallest chromosomes known in the whole subfamily Aurantioideae (Table 2).
In Feroniella sp., on the other hand, there was a large amount of heterochromatin. Although only meiotic analysis was performed on this species, the banding pattern observed in metaphases I and II (Figs. 33–34) suggests that CMA+ blocks were located in both telomeres and in the centromeric region of five chromosome pairs, in a single telomere of two pairs, and absent in two other pairs. The meiotic behavior was regular in spite of the occurrence of a few anaphase bridges.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Stace, 1995
). From ~60 species of Aurantioideae studied so far, the vast majority has 2n = 18 (2x) and infraspecific autopolyploids (3x, 4x. ... ) occur only occasionally. Proof for stabilized polyploids is very limited: Glycosmis pentaphylla agg. (2x, 5x, 6x), G. parviflora agg. (= G. citrifolia agg.) (4x), Clausena excavata (4x), C. dentata (= C. willdenowii) (4x), and Paramignya monophylla (4x). There is only one, obviously erroneous indication for a deviation from x = 9 (Krug, 1943
; Mehra, 1976
; Guerra, 1984
; Stace, Armstrong, and James, 1993
; Stace, 1995
). Karyotype stability in Aurantioideae is apparently linked with a high capacity for interspecific hybridization.
The areticulate nuclear structure of Aurantioideae is clearly correlated with the chromosome organization observed in prophase. During this stage, heteropycnotic blocks are visible after conventional Giemsa or Feulgen staining in the proximal region of many chromosomes, but in some also in the terminal region (Guerra, 1985, 1987
). Such heteropycnotic regions may change depending on the condensation stage and the conditions of fixation. However, CMA/DAPI staining reveals the true heterochromatic nature of all of these terminal, but only of very few of the proximal prophase blocks, and helps to detect a large variation in banding patterns.
The maintenance of the areticulate nuclear structure throughout the whole subfamily is not surprising since this kind of nuclear organization is commonly found in species with a low DNA amount and small chromosome size (Barlow, 1977
; Guerra, 1987
). The chromosome complement size shows only limited variation (Table 2): Swinglea glutinosa seems to have the smallest chromosome complement size (11.11 µm) of the subfamily (see also Sharma and Bal, 1957
; Banerji and Pal, 1957
; Ghosh, 1966
; Wei, Cheng, and Duan, 1988
; Miranda et al., 1997
), only about half of that observed in Murraya paniculata (21.24 µm). This latter species has one of the largest chromosome complement sizes among diploid species of the subfamily. Like other Aurantioideae with DNA amounts known (M. koenigii, Citrus sinensis, and Atalantia monophylla—Guerra, 1984
; Ohri and Kumar, 1986
), M. paniculata, nevertheless, is among the angiosperms with low DNA content (Bennett and Leitch, 1995
).
Heterochromatin quantity and banding
Relative heterochromatin amounts per karyotype as measured by relative length of bands vary largely among the species analyzed (Table 2). The highest heterochromatin content (32.35%) was observed in Murraya paniculata, an amount tenfold higher than in Swinglea glutinosa (3.12%). However, the highest number of bands per karyotype was observed in Fortunella crassifolia. Miranda et al. (1997)
observed an identical CMA-banded karyotype for this species and estimated a heterochromatin proportion of 34.25%, whereas for Citrus species they found 20.58–22.74%. Liang (1988)
also found higher heterochromatin proportions in Fortunella species (~20%) than in Citrus (~16%). This variability is proportional to that observed in the satellite DNA percentage by Ingle, Pearson, and Sinclair (1973)
in Fortunella sp. (24%) and Citrus spp. (19–23%), which were among the highest values observed in a large sample of angiosperms.
The highly variable heterochromatin amounts and banding patterns observed in different genera of Aurantioideae make them very useful tools for identification of intergeneric hybrids. For instance, the sexually obtained hybrids of Citrus x Severinia (Medina-Filho, Bordignon, and Ballvé, 1998
) can be easily recognized by chromosomal CMA staining, since the two genera have widely different banding patterns. Similarly, zygotic and nucellar embryos of such hybrids would differ by at least three chromosome pairs. The heterozygosity observed in some Aurantioideae species, mainly of Citrus and Poncirus, is probably due to the natural and artificial hybridization that has occurred during their intensive breeding and cultivation. However, in most cases a few chromosome markers can distinguish the karyotypes of closely related species (Guerra, 1993
; Miranda et al., 1997
).
Different heterochromatin fractions
At least three types of heterochromatin can be recognized in species of Aurantioideae: (1) C-banding+/CMA+ heterochromatin associated with the NORs (nuleolus organizing regions); (2) C-banding+/CMA+ heterochromatin not associated with the NORs, mostly terminal; and (3) C-banding+/CMA- heterochromatin, observed in proximal regions of some species (Guerra, 1985, 1993
). That the heterochromatin type (1) is always associated with the nucleolus is most easily seen in species of Aurantioideae with only one pair of CMA+ blocks per monoploid complement. It sometimes appears partially decondensed and as a secondary constriction. In several species this heterochromatin was observed as a strongly DAPI-negative block and as a CMA band brighter or fainter than the remaining ones. In general, the chromatin associated to the nucleolus is CMA+ and GC rich (Schweizer, 1976
; Deumling and Greilhuber, 1982
) and represents a special kind of heterochromatin.
On the other hand, in Aurantioideae species with a large number of CMA+ blocks the majority of them are not nucleolus-associated bands of type 2. Matsuyama et al. (1996)
have shown by in situ hybridization that only three of the 16–18 CMA+ blocks of Citrus sinensis correspond to rDNA sites. These data suggest that the majority of the CMA+ blocks are constituted by another DNA sequence involved in the heterochromatin amplification of Aurantioideae, which, although rich in GC (CMA+), is independent of the rDNA sequence. In some other angiosperms that have many CMA+ blocks, it has also been demonstrated that only a few are composed of rDNA repeats (Deumling and Greilhuber, 1982
; Parokonny et al., 1992
).
Quantity and distribution of the C-banding+/CMA- heterochromatin (3) within Aurantioideae still need further analysis. In most species the proximal region of prophase chromosomes is heteropycnotic after conventional staining and seems to correspond to small, DAPI-brilliant chromocenters observed in all genera and species analyzed here (see also Guerra, 1993
). This proximal chromatin was detected in only a few chromosomes with the C-banding method (Guerra, 1985
) but, at least in the ‘Trovita’ orange, it has been clearly demonstrated to occur in every chromosome by HKG banding (Ito, Omura, and Nesume, 1993
).
Constraints in banding patterns
A comparative analysis of the CMA+ heterochromatin within the karyotypes of Aurantioideae reveals remarkable constraints. Despite the high diversity found in representatives of both tribes, CMA+ blocks evidently are not distributed at random. Chromosome banding patterns of all species analyzed and of those previously described (Matsuyama et al., 1996
; Miranda et al., 1997
), correspond to one of the six chromosome types reported for Citrus (Guerra, 1993
) and suggest quite limited variation. Terminal bands dominate, whereas proximal ones are rare. Such proximal CMA+ bands have only been found in chromosomes with at least one telomeric band. Interstitial bands were never observed or have been misinterpreted as being proximal or terminal, due to the small size of chromosome arms.
In proximal positions, generally, CMA+ bands are less common in plant chromosomes than DAPI+ bands (see, e.g., Schweizer, 1976
; Deumling and Greilhuber, 1982
; Moscone, Lambrou, and Ehrendorfer, 1996
), although they are known in some species (Röser, 1994
). In general, when there is proximal and terminal heterochromatin in a banding pattern, single proximal bands also occur. In the tribe Anthemidae (Asteraceae), for example, Schweizer and Ehrendorfer (1983)
have demonstrated a banding pattern very similar to that of Aurantioideae, but there are also chromosomes with single proximal heterochromatic blocks.
The Aurantioideae banding patterns reveal a remarkable constraint of heterochromatin development in one of the three largest chromosomes of each set, designated as FL. This can be easily identified in nearly all species, independent of genome size and ploidy level. CMA+ heterochromatin is usually absent in this chromosome, although in some species a small CMA+ band may be observed in one or both homologues. This was previously reported also for some Citrus species (see idiograms in Guerra, 1993
, and Miranda et al., 1997
). Obviously, the structure of the FL chromosome has been strongly conserved during Aurantioideae evolution. Linkage maps of Citrus species obtained through recombination of molecular markers also suggest high conservation of linkage groups (Durham et al., 1992
; Jarrel et al., 1992
), and detailed molecular maps in other plant groups suggest linear conservation of genome structures even in spite of variable basic chromosome numbers (Gill, 1995
; Moore et al., 1995
).
Another aspect of chromosome evolution in Aurantioideae is that CMA+ heterochromatin is concentrated in three of the largest chromosomes of each haploid complement. Such chromosomes display banding patterns of types A to D (according to the classification proposed for Citrus chromosomes by Guerra, 1993
). Furthermore, the most commonly banded chromosome types had a single terminal band (types D and E), preferentially located in the long arm (see also, Guerra, 1993
; Matsuyama et al., 1996
; Miranda et al., 1997
).
The above data partly disagree with the widely accepted model of heterochromatin dispersion, proposed by Schweizer and Loidl (1987)
. This model predicts that "telomeric bands tend to be preferentially associated with short chromosomes or chromosome arms." However, in Aurantioideae the bands are preferentially distributed in large chromosomes and large arms. Furthermore, the model assumes a coevolution of the C-band pattern in nonhomologous chromosome arms of similar size. In Aurantioideae this fits the preferential distribution of CMA+ blocks in the larger arms, but it does not apply to the FL chromosomes with similar arm size, which remain heterochromatin poor.
Phylogeny, systematics, karyotypes, and banding patterns
Current studies on cpDNA sequences and resulting trees (Samuel et al., 1999a, b,
and unpublished data) clearly show the Aurantioideae to be monophyletic, but the resulting phylogeny of the subfamily is at variance in several respects with the classical systematic treatment by Swingle and Reece (1967),
which was based on morphological and anatomical characters and may need considerable modifications. Earlier phytochemical evidence had already been used to improve the Aurantioideae classification and to calculate advancement parameters for several biogenetic classes of compounds (e.g., Waterman, 1983, 1990
; But et al., 1998;
Da Silva et al., 1988
). This is supplemented by suggestions for character progressions (plesiomorphic 
apomorphic), both from phytochemistry (e.g., the stepwise replacement of anthranilic acid derived alkaloids by coumarins and/or limonoids) and from morphology (e.g., development of axillary spines; pinnate simple leaves; petioles and rachis roundish winged; stamens 5 + 5 more numerous; fruits many, small few, large; exocarp thin, soft thick, leathery, or woody; fruit locules without with ± specialized pulp-vesicles; etc.) (e.g., Tanaka, 1936
; Swingle and Reece, 1967
; Da Silva et al., 1988
). Against this background we propose to discuss the karyological findings presented in this study.
According to the cpDNA data the subdivision of Aurantioideae into Clauseneae and Citreae (Swingle and Reece, 1967
) is justified, if the genera Murraya (with exception of the species segregated as Bergera, e.g., Murraya koenigii = Bergera koenigii) and Merrillia (Merrilliinae) are transferred to the Citreae. This narrower concept includes only genera within Clauseneae s.s., which have carbazoles and only a limited quantity of heterochromatin per haploid chromosome set (<5% in one to four banded chromosomes). In contrast, the genera of Citreae s.l. lack carbazoles and have either little or more often considerable quantities of heterochromatin (>10% in up to eight banded chromosomes) (Tables 2–3).
One of the core genera of Clauseneae s.s. is Glycosmis, a large and widespread Australasian genus, morphologically and phytochemically quite plesiomorphic, well circumscribed, and relatively isolated. In no other genus of Aurantioideae have so many polyploids accumulated: G. pentaphylla agg. and G. parviflora agg. represent polyploid complexes (with 2x, 4x, 5x, and 6x populations), still very insufficiently understood (H. Greger, Inst. Bot. Univ. Vienna, personal communication). The banding pattern of Glycosmis is simple, with only a single CMA+ block per monoploid complement on chromosome II. This results in the low heterochromatin value of 4.36% (Table 2).
Clausena forms a somewhat more advanced group of Clauseneae s.s., but the diploid C. lansium still exhibits a karyotype very similar to Glycosmis with only one CMA+ block on chromosome pair II and a heterochromatin value of 3.41%. The distinct tetraploid C. excavata has small bands in seven of its 18 chromosome pairs, but only a slightly higher amount of heterochromatin.
Morphological and phytochemical evidence (But et al., 1988
; Waterman, 1990
) as well as recent cpDNA data (Samuel et al., 1999a, b,
and unpublished data) demonstrate that the genus Murraya s.l. in its present and wide circumscription (Swingle and Reece, 1967
) is an "artifact" and should be split at least into Murraya s.s. and Bergera. This suspicion is strongly supported by our karyological data: whereas the type species of Murraya, M. paniculata, exhibits an elaborate banding pattern with CMA+ blocks on eight of its nine chromosome pairs and the highest heterochromatin index (>32%) in the Aurantioideae, Bergera koenigii has bands on only two chromosome pairs (II, IV) and a low heterochromatin value (Table 3).
In conclusion, the Clauseneae s.s. are constituted by two sister clades, one formed by Glycosmis and the more distantly related Micromelum (not studied here), and the other by Clausena and Bergera. These Clauseneae s.s. represent an assembly of genera with a dominance of plesiomorphic characters. In contrast, Murraya s.s., together with Merrillia, should be removed to the Citreae s.l. As described by Swingle and Reece (1967)
, the three subtribes of Clauseneae s.l., Micromelinae (IA), Clauseninae (IB), and Merilliinae (IC), appear heterogeneous and obsolete.
Within their Citreae s.s. (II) Swingle and Reece (1967)
have differentiated three subtribes, i.e., Triphasiinae (IIA), Citrinae (IIB), and Balsamocitrinae (IIC). Available cpDNA data (Samuel, 1999a, b
, and unpublished data) demonstrate that these subtribes also are "artificial." The genera analyzed in the present study apparently belong to three major clades of Citreae s.l., the first two have no or only plesiomorphic, broadly based pulp-vesicles and often lack limonoids, and the third, the Citrus clade, develops specialized pulp-vesicles with a slender stalk and mostly exhibits a variety of limonoids (Da Silva, Gottlieb, and Ehrendorfer 1988
). Both of the first two clades include taxa from the Clauseneae (i.e., IB and IC) and from all three of the Citreae subtribes (IIA, IIB, and IIC) as recognized by Swingle and Reece (1967)
.
In the first and most basal of the cpDNA-supported clades we propose to include Swinglea (IIC), Severinia (IIB), Murraya s.s. (IB), and Merrillia (IC) as already mentioned, and possibly also Atalantia (IIB). Both Tanaka (1936)
and Swingle and Reece (1967)
consider Swinglea as morphologically plesiomorphic and basal within Balsamocitrineae. The only species, S. glutinosa, has a single CMA+ block in its haploid chromosome complement and the lowest heterochromatin content yet recorded for the Aurantioideae (3.12%). Severinia within Citrinae (IIB) also exhibits relatively plesiomorphic features. Its banding pattern consists of three to four CMA+ bands in three chromosome pairs, and the heterochomatin amount is intermediate. Murraya s.s. (IB) with M. paniculata with its extremely high heterochromatin value of 32.35% has been discussed already. There are close relationships with the monotypic Merrillia (IC). Both genera are former members of Clauseneae s.l., exhibit relatively apomorphic features, have quite similar karyotypes with chromosome bands in six to eight of their chromosome pairs, and should be transferred to the Citreae s.l.
Atalantia (IIB) was formerly considered close to Severinia, and the two were even united into one genus (Engler, 1931
). However, Swingle and Reece (1967
, p. 283) used the old generic name Atalantia to denominate those species with more apomorphic characters (as larger flowers and well-formed conical pulp-vesicles, thus approaching the Citrus clade), whereas the species with more plesiomorphic features were left in Severinia. As Atalantia monophylla has bands in four of its nine chromosome pairs, the banding patterns and heterochromatin values of the two genera are similar. Nevertheless, the cpDNA sequences suggest a separate position from the Citrus clade, in spite of the occurrence of several limonoids in Atalantia.
The second cpDNA-supported clade of Citreae s.l. includes among others the genera Triphasia (IIA) and Feroniella (IIC). Nevertheless, the two genera are by no means close. Whereas the first is relatively plesiomorphic in its characters, the second exhibits quite apomorphic features. Triphasia has a relatively high amount of heterochromatin with single telomeric bands on seven of its nine chromosome pairs. The genus was placed into Triphasiinae (IIA), a heterogeneous subtribe whose cytogenetics still are poorly understood. The only other IIA-taxon karyologically investigated is the tetraploid Paramignya monophylla (Mehra, 1976
; Stace, 1995
). The karyotype of Ferionella is also rich in heterochromatin, with bands in seven of the nine chromosome pairs, but the pattern differs considerably from Triphasia, because five of these pairs have two telomeric and one centromeric band (Table 3).
The core of the Citrinae (IIB) are the "true Citrus fruit trees" of Swingle and Reece (1967
; corresponding to the Aurantieae of Tanaka, 1936
). As the third clade of the Citreae s.l., this Citrus clade is well supported by the available cpDNA sequences (Samuel et al., 1999a, b
, and unpublished data), and also quite coherent morphologically, phytochemically, and in its crossing potentials. At the same time, this clade obviously is the most advanced in the tribe with the greatest number of apomorphic features. This corresponds well with its strongly banded karyotypes and the increased heterochromatin amount (28.71% in Citrus). There are one or more bands on six to all nine chromosome pairs of Citrus, Poncirus, Microcitrus, Eremocitrus, and Fortunella (Table 3). In spite of some clear-cut differences between species of Citrus and Fortunella (the present data correspond to those in Guerra, 1993
, and Miranda et al., 1997
), the characteristic overall banding pattern of the Citrus clade is not obscured.
In view of the possible links of the Citrus clade to more basal members of the Aurantioideae with more plesiomorphic characters (e.g., Clausena, Severinia, Atalantia), it would be important to analyze the banding patterns of other possibly related genera like Pleiospermium, Burkillanthus, Limnocitrus, Hesperethusa, and Citropsis for a better resolution of their phylogeny.
In retrospect, the Citreae s.l. are the clearly more advanced and diverse of the two Aurantioidae tribes. The relationships within and between its two more basal clades with relatively plesiomorphic features need further study. In contrast, the third clade with Citrus and five other closely related genera is clearly circumscribed, obviously monophyletic, and strongly apomorphic in most of its characters. Our results suggest an "artificial" nature of the subtribes Triphasiinae, Citrinae, and Balsamocitrinae.
Chromosomal evolution
On the basis of the karyological and other data available for Aurantioideae, one may speculate about the ancestral karyotype and the evolution of banding patterns of the subfamily. The ancestral karyotype was certainly composed of nine chromosome pairs, ranging in size from 1.0 to 3.5 µm. Concerning the heterochromatin pattern, there is an apparent correlation between heterochromatin-rich karyotypes and relatively advanced genera, more or less apomorphic in respect to morphology, anatomy, and phytochemistry. What is noteworthy is the fact that all heterochromatin-rich karyotypes from different and unrelated clades within Aurantioideae exhibit remarkably similar banding patterns, as shown in the foregoing discussion about "constraints." There are at least four instances of such heterochromatin-rich groups with similar banding patterns that are not closely related, i.e., the Murraya s.s./Merrillia clade (IB/C), the genera Triphasia (IIA) and Feroniella (IIC), and the Citrus clade (IIB).
The correlation of plesiomorphic/heterochromatin-poor and apomorphic/hetero-chromatin-rich karyotypes seems to be rather general in plant groups at supra- and infrageneric levels and is commonly interpreted as heterochromatin accumulation during evolution (Ikeda, 1988
; Morawetz and Samuel, 1989
; Röser, 1994
). In some groups this trend does not result in a parallel evolution of banding patterns. For example, Cyphomandra and Capsicum (tribe Solaneae of the Solanaceae) exhibit increased heterochromatin content but have completely different banding patterns (Pringle and Murray, 1993
; Moscone et al., 1993
; Moscone, Lambrou, and Ehrendorfer, 1996
). On the other hand, there are cases like Aurantioideae with apparently independent lines showing heterochromatin increase following very similar parallel patterns, as in Scilla (Hyacinthaceae), where heterochromatin amounts have increased independently in the species groups of S. vindobonensis and of S. luciliae. Here, remarkable similarities in banding patterns, except in the NOR-bearing chromosomes, have developed despite large differences in chromosome size and morphology (Greilhuber, 1979
). Such parallelisms can only be a consequence of special structures in the ancestral karyotype, subsequent constraints, and karyotypic orthoselection (White, 1973
). This author admitted three distinct ways in which selection might canalize structural changes: similar environmental pressure, similar cellular adaptation, or similar distribution of chiasmata. This kind of orthoselection has frequently been correlated with heterochromatin distribution (John and Miklos, 1979
) and could be a reasonable explanation for the parallel evolution of banding patterns. Furthermore, if optimal regions for the occurrence of heterochromatin and chiasmata exist, as predicted by the chromosome field hypothesis (reviewed by Lima-de-Faria, 1983
), heterochromatin would tend to be selectively accumulated through amplification of pre-existing repeats (library hypothesis; Fry and Salser, 1977
) in the same optimal regions of homeologous chromosomes of different clades.
|
FOOTNOTES |
|---|
4 Author for correspondence (FAX ++43-1-4277-9541; e-mail: friedrich.ehrendorfer{at}univie.ac.at
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Barlow, P. W. 1977 Determinants of nuclear chromatin structure in angiosperms. Annales des Sciences Naturelles, Botanique et Biologie Végétale (Paris). 18: 193–206.
Barrett, H. C. 1977 Intergeneric hybridization of Citrus and other genera in citrus cultivar improvement. Proceedings of the International Society of Citriculture 2: 586–589.
Bennett, M. D., and I. Leitch. 1995 Nuclear DNA amounts in angiosperms. Annals of Botany 76: 113–176.
But, P.P.-H., Y.-C. Kong, Q. Li, H.-T. Chang, K.-L. Chang,, K. M. Wong, A. I. Gray, and P. G. Waterman. 1988 Chemotaxonomic relationship between Murraya and Merrillia (Rutaceae). Acta Phytotaxonomica Sinica 26: 205–210.
Cameron, J. W., and H. B. Frost. 1968 Genetics, breeding and nucellar embryony in Citrus. In W. Reuther, L. D. Batchelor, and H. J. Webber [eds.], The citrus industry, 325–370. University of California Press, Berkeley, California, USA.
Carpenter, J. B., and P. C. Reece. 1969 Catalog of genera, species, and subordinate taxa in the orange subfamily Aurantioideae (Rutaceae). Crops Research, Agricultural Research Service, U.S. Department of Agriculture Beltsville, Maryland USA.
Da Silva, M. F. G. F., O. R. Gottlieb, and F. Ehrendorfer. 1988 Chemosystematics of the Rutaceae: suggestions for a more natural taxonomy and evolutionary interpretation of the family. Plant Systematics and Evolution 161: 97–134.[CrossRef][ISI]
Deumling, B., and J. Greilhuber. 1982 Chracterization of heterochromatin in different species of the Scilla siberica group (Liliaceae) by in situ hybridization of satellite DNAs and flurochrome banding. Chromosoma (Berlin) 84: 535–555.[CrossRef][ISI]
Durham, R. E., P. C. Liou, F. G. Gmitter, Jr., and G. A. Moore. 1992 Linkage of restriction fragment length polymorphisms and isozymes in Citrus. Theoretical and Applied Genetics 84: 39–48.
Engler, A. 1931 Rutaceae. In A. Engler and K. Prantl [eds.], Die natürlichen Pflanzenfamilien, 2. Auflage, 19a: 187–359. Engelmann, Leipzig, Germany.
Esen, A., and R. Scora. 1977 Amylase polymorphism in Citrus and some related genera. American Journal of Botany 64: 305–309.[CrossRef][ISI]
Fang, D. 1993a Intra- and intergeneric relationships of Poncirus polyandra: investigation by leaf isozymes. Journal of Wuhan Botanical Research (Wuhan, China) 11: 34–40.
———. 1993b Citrus taxonomy: past, present and future. Journal of Wuhan Botanical Research (Wuhan, China). 11: 375–382.
———, W. C. Zhang, and S. Y. Xiao. 1993 Studies on taxonomy and evolution of Citrus and related genera by isozyme analysis. Acta Phytotaxonomica Sinica 31: 329–352.
Fry, K., and W. Salser. 1977 Nucleotide sequences of HS satellite DNA from kangaroo rat Dipodomys ordii and characterization in other rodents. Cell 12: 1069–1084.[CrossRef][ISI][Medline]
Ghosh, R. B. 1966 A contribution to the karyotypic study in Glycosmis pentaphylla. Beiträge zur Biologie der Pflanzen 42: 139–143.
Gill, B. S. 1995 The molecular cytogenetic analysis of economically important traits in plant. Kew Chromosome Conference IV: 47–53.
Greilhuber, J. 1979 Evolutionary changes of DNA and heterochromatin amounts in Scilla bifolia group (Liliaceae). Plant Systematics and Evolution, Supplement 2: 263–280.
Grieve, C. M., and R. W. Scora. 1980 Flavonoid distribution in the Aurantioideae (Rutaceae). Systematic Botany 5: 39–53.
Guerra, M. 1983 O uso do corante Giemsa na citogenética vegetal: comparação entre a coloração simples e o bandeamento C. Ciência e Cultura 35: 190–193.
———. 1984 Cytogenetics of Rutaceae. II. Nuclear DNA content. Caryologia 37: 219–226.[ISI]
———. 1985 Cytogenetics of Rutaceae. III. Heterochromatin patterns. Caryologia 38: 335–346.[ISI]
———. 1987 Cytogenetics of Rutaceae. IV. Structure and systematic significance of the interphase nuclei. Cytologia 52: 213–222.[ISI]
———. 1993 Cytogenetics of Rutaceae. V. High chromosomal variability in Citrus species revealed by CMA/DAPI staining. Heredity 71: 234–241.[ISI]
Ikeda, H. 1988 Karyomorphological studies on the genus Crepis with special reference to C-banding pattern. Journal of Science, Hiroshima University, Series B, Division 2, 22: 65–117.
Ingle, J., G. G. Pearson, and J. Sinclair. 1973 Species distribution and properties of nuclear satellite DNA in higher plants. Nature New Biology 242: 193–197.[CrossRef][ISI][Medline]
Ito, Y., M. Omura, and H. Nesume. 1993 Improvement of chromosome observation methods for Citrus. In T. Hayashi, M. Omura, and N. S. Scott [eds.], Techniques on gene diagnosis and breeding, 31–38. FTRS, Tsukuba, Ibaraki, Japan.
Iwamasa, M., and N. Nito. 1988 Cytogenetics and the evolution of modern cultivated Citrus. In R. Goren and K. Mendel [eds.], Proceedings of the Sixth International Citrus Congress, 265–275. Margraf Scientific Books, Weikersheim, Germany.
Jarrell, D. C., M. L. Roose, S. N. Traugh, and R. S. Kupper. 1992 A genetic map of citrus based on the segragation of isozymes and RFLPs in an intergeneric cross. Theoretical and Applied Genetics 84: 49–56.[ISI]
John, B., and G. Miklos. 1979 Functional aspects of satellite DNA and heterochromatin. International Review of Cytology 58: 1–144.[Medline]
Krug, C. A. 1943 Chromosome numbers in the subfamily Aurantioideae with special reference to the genus Citrus. Botanical Gazette 104: 602–611.
Liang, G. 1988 Studies on the Giemsa C-banding patterns of some Citrus and its related genera. Acta Genetica Sinica 15: 409–415.
Lima-de-Faria, A. 1983 Molecular evolution and organization of the chromosome. Elsevier, Amsterdam, The Netherlands.
Matsuyama, T., T. Akihama, Y. Ito, M. Omura, and K. Fukui. 1996 Characterization of heterochromatic regions in ‘Trovita'orange (Citrus sinensis Osbeck) chromosomes by the fluorescent staining and FISH method. Genome 39: 941–945.[Medline]
Medina-Filho, H. P., R. Bordignon, and R. M. L. Ballvé. 1998 Sunkifolias and Buxisunkis: Sexually obtained reciprocal hybrids of Citrus sunki x Severinia buxifolia. Genetics and Molecular Biology 21: 129–133.
Mehra, P. N. 1976 Cytology of Himalayan hardwoods. Sree Saraswaty Press Ltd., Calcutta, India.
Miranda, M., F. Ikeda, T. Endo, T. Moriguchi, and M. Omura. 1997 Comparative analysis on the distribution of heterochromatin in Citrus, Poncirus and Fortunella chromosomes. Chromosome Research 5: 86–92.[CrossRef][ISI][Medline]
Moore, G., K. Devos, R. Dunford, T. Foote, and M. Gale. 1995 Comparative organization of the cereal genomes. Kew Chromosome Conference IV: 109–117.
Morawetz, W., and M. R. A. Samuel. 1989 Karyological patterns in the Hamamelidae. In P. R. Crane and S. Blackmore [eds.], Evolution, systematics and fossil history of the Hamamelidae, vol. 1, Introduction and "Lower Hamamelidae," 129–154. Clarendon Press, Oxford, UK.
Moscone, E. A., M. Lambrou, and F. Ehrendorfer. 1996 Fluorescent chromosome banding in the cultivated species of Capsicum (Solanaceae). Plant Systematics and Evolution 202: 37–63.[CrossRef][ISI]
———, ———, A. T. Hunziker, and F. Ehrendorfer. 1993 Giemsa C-banded karyotypes in Capsicum (Solanaceae). Plant Systematics and Evolution 186: 213–229.[CrossRef][ISI]
Motomura, T., T. Hidaka, T. Moriguchi, T. Akihama, and M. Omura. 1995 Intergeneric somatic hybrids between Citrus and Atalantia or Severinia by electrofusion, and recombination of mitochondrial genomes. Breeding Science 45: 309–314.[ISI]
Ohri, D., and A. Kumar. 1986 Nuclear DNA amounts in some tropical hardwoods. Caryologia 39: 303–307.[ISI]
Parokonny, A. S., A. Y. Kenton, L. Meredith, S. J. Owens, and M. Bennett. 1992 Genomic divergence of allopatric sibling species studied by molecular cytogenetics of their F1 hybrids. Plant Journal 2: 695–704.[ISI]
Pringle, G. J., and B. G. Murray. 1993 Karyotypes and C-banding patterns in species of Cyphomandra Mart. ex Sendtner (Solanaceae). Botanical Journal of the Linnean Society 111: 331–342.[CrossRef]
Röser, M. 1994 Pathways of karyological differentiation in palms (Arecaceae). Plant Systematics and Evolution 189: 83–122.[CrossRef][ISI]
Samuel, R. H., F. Ehrendorfer, M. Chase, and H. Greger. 1999b Plastid DNA sequences and secondary metabolites help to reconstruct the phylogeny of Aurantioideae (Rutaceae). Abstracts, Seventh Congress of the European Society for Evolutionary Biology, Barcelona, Spain, 23–28 August 1999.
———, H. Greger, M. Chase, and F. Ehrendorfer. 1999a Phylogeny of Rutaceae-Aurantioideae: multidisciplinary studies, in particular cpDNA sequences from the atpB/rbcL spacer. Abstracts, XVI International Botanical Congress, St. Louis, USA, 1–7 August, 1999, p. 438.
Schweizer, D. 1976 Reverse fluorescent chromosome banding with chromomycin and DAPI. Chromosoma 58: 307–324.[CrossRef][ISI][Medline]
———, and F. Ehrendorfer. 1983 Evolution of C-band patterns in Asteraceae (Anthemideae). Biologische Zentralblatt 102: 637–655.
———, and J. Loidl. 1987 A model for heterochromatin dispersion and the evolution of C-band patterns. Chromosomes Today 9: 61–73.
Sharma, A. K., and A. B. Bal. 1957 Chromosome studies in Citrus. I. Agronomia Lusitana 19: 101–126.
Stace, H. M. 1995 Primitive and advanced character states for chromosome number in Gondwanan angiosperm families of Australia. Kew Chromosome Conference IV: 223–232.
———, J. A. Armstrong, and S. H. James. 1993 Cytoevolutionary patterns in Rutaceae. Plant Systematics and Evolution 187: 1–28.[CrossRef][ISI]
Swingle, W. T., and P. C. Reece. 1967 The botany of Citrus and its wild relat
