Chromosome studies in Orchidaceae from Argentina

The center of diversity of Argentinean orchids is in the northeast region of the country. Chromosome numbers and karyotype features of 43 species belonging to 28 genera are presented here. Five chromosome records are the first ones at the genus level; these taxa are Aspidogyne kuckzinskii (2n = 42), Eurystyles actinosophila (2n = 56), Skeptrostachys paraguayensis (2n = 46), Stigmatosema polyaden (2n = 40) and Zygostates alleniana (2n = 54). In addition, a chromosome number is presented for the first time for 15 species: Corymborkis flava (2n = 56), Cyclopogon callophyllus (2n = 28), C. oliganthus (2n = 64), Cyrtopodium hatschbachii (2n = 46), C. palmifrons (2n = 46), Galeandra beyrichii (2n = 54), Habenaria bractescens (2n = 44), Oncidium edwallii (2n = 42), O. fimbriatum (2n = 56), O. pubes (2n = 84), O. riograndense (2n = 56), Pelexia ekmanii (2n = 46), P. lindmanii (2n = 46) and Warrea warreana (2n = 48). For Oncidium longicornu (2n = 42), O. divaricatum (2n = 56) and Sarcoglottis fasciculata (2n = 46+1B?, 46+3B?), a new cytotype was found. Chromosome data support phylogenetic relationships proposed by previous cytological, morphologic and molecular analyses, and in all the cases cover some gaps in the South American literature on orchid chromosomes.


Introduction
Orchidaceae Juss. is one of the largest families in the plant kingdom, distributed throughout the tropical and subtropical areas of both hemispheres (Correa, 1955). The family comprises around 850 genera and 20000 species, and nearly 75% of them are epiphytes (Dressler, 1993;Johnson, 2001). In Argentina, this family comprises around 74 genera and c. 280 species (Johnson, 1992;Correa, 1996) distributed throughout the country. Numerous orchids are part of the northeast Argentinean flora, particularly that of the provinces of Misiones (154 species), Corrientes (76), and Chaco (34) (Correa, 1996;Zuloaga et al., 1999;Insaurralde and Radins, 2007), and most of them have ornamental value.
Cytogenetic studies on Orchidaceae are few, disperse and incomplete. It is estimated that so far no more than about 10% of all species were chromosomally analyzed, therefore the chromosome evolution in this family remains in debate. The chromosome numbers vary from 2n = 10 to 2n = 240 and high numbers are common, the most frequent being n = 19 or 20 (Jones, 1974). Regarding chromosome studies of South American orchids, the most important contributions were made by Blumenschein (Blumenschein A, PhD Thesis, University of São Paulo, Piracicaba, 1957;Blumenschein, 1960), Martínez (1981Martínez ( , 1985, Dematteis and Daviña (1999) and Felix and Guerra (1998, but the data are still considered insufficient. The cytology of orchids has been considered in the past to have a great potential in studies of taxonomic affinities and evolutionary bonds (Jones, 1974), and this is still true today, mainly if we take into account the phylogenetic gaps or misinterpretations of evolutionary trends. Moreover, in northeast Argentina, a great number of orchid populations are currently limited to remnants of the subtropical forest. Both epiphytes and terrestrial species grow in vegetation fragments, and their conservation is difficult due to changes in the habitat, patchy distribution and specific pollination strategies. For these reasons and as a first step towards characterization of the genetic variability of the species belonging to South American orchid genera, a cytological study was carried out.

Material and Methods
Forty-three species of Orchidaceae from Argentina were chromosomally studied (Table 1). Samples were cultivated in the greenhouse of the Programa de Estudios Floristicos y Genetica Vegetal (FCEQyN -UNaM), and some of them have not flowered yet. Voucher specimens were deposited at the herbarium of the Universidad Nacional de Misiones (MNES). The taxonomic nomenclature adopted for the species was that of Govaerts et al. (2008). The results obtained are presented and discussed according to the classifications of Dressler (1993) and/or Szlachetko (1995 at suprageneric levels. Mitotic studies were performed in root tips pretreated with saturated 1-bromonaphthalene for 2-3 hs at room temperature, fixed in absolute ethanol:glacial acetic acid (3:1) for 12 hs at 4°C, and stained according to the Feulgen technique. The meristems were macerated in a drop of 2% aceto-orcein and then squashed. Permanent slides were made using euparal as a mounting medium. Slides were analyzed with a Leica DMLS optical photomicroscope, photographs were taken with Imagelink HQ 25 ASA Kodak film, and negatives were digitalized with a Genius ColorPage-HR8 scanner. 812 Daviña et al.  Sinotô (1962Sinotô ( , 1969, Charanasri et al. (1973), Félix and Guerra (2000) Chromosome measurements were made using the MicroMeasure 3.3 computer program (Reeves, 2001).
In most of the analized taxa, the karyotype was found to be of the bimodal type. Bimodality is present to varying degrees, as evidenced by the short/large chromosome ratio (S/L). Some Epidendroideae species showed small chromosomes, ranging from 0.5 to 2.5 mm (S/L = 0.20). The chromosomes of Habenaria (Orchidoideae) were small, ranging from 1 to 3 mm (S/L = 0.33), whereas those of 816 Daviña et al.

Discussion
A wide diversity in chromosome numbers and karyotype features was found among the species of orchids that inhabit the northeast region of Argentina. In order to clarify any possible relationships among taxa, the results obtained are discussed, first circumscribed to each subfamily as starting point and then as a whole.
Subfamily Spiranthoideae encompasses about 95 genera and 1140 species, predominantly terrestrial (Salazar et al., 2003), with 2n = 24, 26, 28, 30, 32, 36, 44, and 46 as the most common chromosome numbers (Szlachetko, 1995). However, Cameron et al. (1999) did not recognize this clade as a separate subfamily and included it within Orchidoideae. We analyzed 17 species of 10 genera and found a wide diversity of chromosome numbers, i.e., 2n = 26, 28, 32, 40, 42, 46, 56, 64. This is the first time chromosome counts of Aspidogyne, Eurystyles and Stigmatosema species were made, but more studies are needed on these neglected genera. In natural populations of Aspidogyne kuczynskii, we observed polymorphism for the color of the leaves among different plants, but all phenotypes presented the same chromosome number.
The type of karyotype bimodality observed in Sacoila, Skeptrostachys, Mesadenella and Eltroplectris, with a chromosome pair twice as large as the mean chromosome size and carrying a macrosatellite on the short arm, along with previous cytogenetical data (Cocucci, 1956;Martínez, 1985;Dematteis and Daviña, 1999;Felix and Guerra, 2005), support their inclusion in a separate clade (subtribe Stenorrynchidinae sensu Szlachetko, 1995; Stenorrhynchos clade sensu Salazar et al., 2003). On the other hand, Pelexia bonariensis, P. ekmanii, P. lindmanii, Sarcoglottis fasciculata, S. grandiflora and S. ventricosa also share the chromosome number (2n = 46) and the general karyotype features, which, together with previous cytogenetic data for those genera (Martínez, 1985;Dematteis and Daviña, 1999;Felix and Guerra, 2005), reinforces the proposal incorporating Pelexia and Sarcoglottis into the same clade (Dressler, 1993;Szlachetko, 1995;Salazar et al., 2003). Moreover, in S. fasciculata we found 2n = 46, 47 and 49 in the same plant. This variation in the somatic chromosome number may be due to the presence of B chromosomes with mitotic instability, as reported in several angiosperms (Jones and Rees, 1982), or aneusomaty.
Tropidoideae is a small subfamily that comprises three genera including Corymborkis (tribe Tropideae) (Szlachetko, 1995). Cameron et al. (1999), however, did not support this clade as a separate subfamily, but as a tribe of Epidendroideae. The chromosome count for Corymborkis flava (2n = 56) reported here is the first one for this species. Only other two chromosome counts were made in this genus (2n = 40, 58), for C. veratrifolia, by Pancho (1965) and Ono and Masuda (1981), respectively.
Among other genera, the tribe Cymbidieae encompasses Cyrtopodium, Galeandra and Oeceoclades (Dressler, 1993;Szlachetko, 1995). For Cyrtopodium, only 10 species had their chromosomes previously counted, showing n = 22 and 2n = 46, 92 (Aoyama, 1989;Felix and Guerra, 2000). Cyrtopodium hatschbachii and C. palmifrons presented 2n = 46, which, combined with the 23II observed at diakinesis in the former species (unpublished data), allows us to propose a basic chromosome number of x = 23 for this genus, although it might be a derived number. For Galeandra, there are only two previous chromosome reports concerning G. baueri and G. devoniana, two epiphytic species with 2n = 56 (Aoyama, 1989). Our count (2n = 54) in G. beyrichii, a terrestrial species, suggests that there is probably more than one basic chromosome number in this genus. As in Galeandra, Felix and Guerra (2000) report different chromosome numbers for Brazilian Cyrtopodium and Oncidium orchids with different habitats. Genus Oeceoclades comprises about 39 species distributed over Africa, and O. maculata is the only one that also lives in America (Govaerts et al., 2008). In Brazilian O. maculata populations, 2n = 48, c.52, 54, 58 cytotypes were found (Guerra, 1986;Felix and Guerra, 2000). However, for Argentinean populations, we consistently observed a 2n = 56 cytotype, in agreement with our previous work (Dematteis and Daviña, 1999). Our results regarding O. maculata agree with the proposal of the primary basic number x 1 = 7 for Cymbidieae, presented by Felix and Guerra (2000), so this species could be considered an octoploid.
Catasetum (tribe Catasetinae) is an American genus with 163 species (Govaerts et al., 2008), and 2n = 54 is the most common chromosome number. In C. fimbriatum, we observed 2n = 108. The available chromosome data for the genus (Jones and Darker, 1968;Dematteis and Daviña, 1999;Felix and Guerra, 2000) suggest a polyploid series, probably based on x = 6. Cameron et al. (1999) and Cameron (2004) support the idea that genus Galeandra (Cymbidieae) is closer to Catasetum than to Cyrtopodium (Cymbidieae), and our chromosome counts reinforce their proposal.
Warrea warreana and Zygopetalum maxillare presented 2n = 48 median-size chromosomes. Both genera are included in the same clade, tribe Maxillarieae, according to Dressler (1993), or Zygopetaleae sensu Szlachetko (1995). Our cytogenetic data, along with previous chromosome number reports for both genera (Blumenschein, 1960;Tanaka and Kamemoto, 1984;Aoyama, 1989;Aoyama et al., 1994), support the idea of their inclusion in the same clade and reinforce the hypothesis of x = 24 or 26 in this group (Felix and Guerra, 2000).
Campylocentrum neglectum (tribe Vandeae) showed a 2n = 38, and this is the first count for natural populations from Argentina. The chromosome number agrees with that from Paraguay reported by Dematteis and Daviña (1999). No other Campylocentrum species were chromosomally analyzed so far, although 64 tropical and subtropical American species of this genus were described.
Oncidium is a large neotropical genus with showy epiphytic, litophytic and terrestrial species, and some of them can be misidentified when in vegetative state. This genus has been extensively studied, showing 2n = 56, 42, 28 as the most common chromosome numbers (Felix and Guerra, 2000). All the Oncidium species analyzed here showed small chromosome size (0.5 to 3 mm) and presented 2n = 42, 56 as the most common numbers, with the exception of O. bifolium (2n = 108) and O. pubes (2n = 84). Cytotype 2n = 84 is present in at least six species (Tanaka and Kamemoto, 1984;Felix and Guerra, 2000) that could be considered 12-ploid, whereas cytotype 2n = 108 could be considered a 16-ploid derived from 2n = 112, both based on x = 7.
At any stage of analysis, Orchidaceae is a complex taxon. Its large number of species, along with a wide variation at the morphological, ecological and cytological levels, have made it difficult to establish an accurate phylogenetic scenario using classical approaches. Therefore, different criteria for the classification of these orchids as a whole have been applied, such as those of Dressler (1993) and Szlachetko (1995) as the most recent analyses, resulting in different evolutionary pictures. Molecular analysis of the orchids as a whole (i.e., Cameron et al., 1999;Cameron, 2004;Chase et al.,2005) or of any of their larger clades (i.e., Cameron and Chase, 2000;Chase et al., 2005;Van den Berg et al., 2005) proved to aid the classical approaches, but differing at some point. Regarding the number of subfamilies comprised by Orchidaceae, several contradictions have arisen between classical and molecular analysis. Dressler (1993) and Szlachetko (1995) do not assign the genera to the same taxonomic intra-family categories, whereas Cameron et al. (1999) support five major monophyletic clades, delimiting some taxa sustained by the former authors (such as inclusion of Vandoideae and Tropidoideae into Epidendroideae and inclusion of Spiranthoideae into Orchidoideae). In addition, whatever the classification criterion, the wide diversity in chromosome number within each subfamily of Orchidaceae added to the few cytologic data available, make it difficult to discuss the taxo-nomic and molecular contradictions at this level based on cytogenetic findings, although the latter proved to be a powerful tool for solving doubts regarding lower taxonomic ranks (i.e., Martínez, 1985;Aoyama, 1989;Aoyama et al., 1994;Felix and Guerra, 1998Luo, 2004; this work).
Regarding chromosome size and interchromosomal asymmetry, our results point out that in Epidendroideae (sensu Cameron et al., 1999) small chromosomes and definitely bimodal karyotypes predominate, with the exception of Corymborkis, of uncertain phylogenetic position, which has mostly median-size chromosomes. Furthermore, our results on Orchidoideae (sensu Cameron et al., 1999) reveal the presence of small, median and large chromosomes, and, with the exception of Cyclopogon and Stigmatosema, distinctly bimodal karyotypes, which is more evident in Eltroplectris, Mesadenella, Sacoila and Skeptrostachys. Particularly, the chromosome size and karyotype features of the species of the Orchidoideae tribe Cranichideae [Spiranthoideae sensu Dressler (1993) and Szlachetko (1995)] are good markers for supporting the current lower-rank clades proposed by the latter author.
The genetic variability of the orchids is also expressed by the diversity of their chromosome number and karyotype features. Progress in the cytogenetic research of orchids is essential to solve the still persisting contradictions between the results of morphologic and molecular analyses and to help developing an improved taxonomic approach.