Chromosome numbers of Carex (Cyperaceae) and their taxonomic implications

Counting chromosomes is the first step towards a better understanding of the karyotype evolution and the role of chromosome evolution in species diversification within Carex; however, the chromosome count is not known yet for numerous sedges. In this paper chromosome counts were performed for 23 Carex taxa from Armenia, Austria, the Czech Republic, and Poland. Chromosome numbers were determined for the first time in three species (Carex cilicica, 2n = 54; C. phyllostachys, 2n = 56; C. randalpina, 2n = 78), two subspecies (C. muricata subsp. ashokae, 2n = 58; C. nigra subsp. transcaucasica, 2n = 84) and two hybrids (C. ×decolorans, 2n = 74; C. ×walasii, 2n = 108). Among the taxa whose number of chromosomes had been known before, the largest difference was found in C. hartmaniorum (here 2n = 52) and C. aterrima subsp. medwedewii (here 2n = 52). A difference in the chromosome count was demonstrated for C. cilicica (2n = 54) versus the species of the section Aulocystis (2n = 30 to 40) and for C. tomentosa (2n = 48) versus the species of the section Acrocystis (2n = 18 to 38). The results of this study indicate that the position of C. cilicica in Aulocystis section may raise doubts. Attention was paid to the relationship between C. phyllostachys and taxa of the subgenus Carex section Gynobasidae.


Introduction
With about 2000 species described worldwide Carex L. (Cyperaceae) represents one of the most species-rich angiosperm genera [1]. The taxonomic richness is accompanied by an extreme variability in the number of chromosomes [2]. Sedges have holocentric chromosomes, which-in theory-guarantee a rapid karyotype evolution [3,4]. If a monocentric chromosome is fragmented, fragments lacking the centromere cannot be normally segregated during meiosis, which results in a loss of genetic material, the gametes produced being potentially non-viable [3]. The situation is different with holocentric chromosomes, because chromosome fragments are not lost, and a change in the chromosome count can be offset by, e.g., self-pollination or back-crossing. In addition, associations between non-homologous chromosomes during meiosis frequently do not disturb segregation, thus reducing selective pressure against chromosome rearrangements [5,6]. Although not all organisms with holocentric chromosomes show highly variable chromosome counts [7,8], the genus Carex is an ideal object to study the chromosome number variability [3,9]. The chromosome number variability in different species within a genus with monokinetic chromosomes is usually a result of polyploidy or aneuploidy [6,10,11]. In those species with holokinetic chromosomes, the frequent aneuploidy is complemented by two additional mechanisms which may lead to differences in the chromosome count: agmatoploidy (fission of chromosomes) and symploidy (fusion of chromosomes) [10,[12][13][14][15]. It seems that evolution of karyotypes in sedges, important for species diversification, is driven by fusion and fission of chromosomes [16]. On the other hand, polyploidy is most likely rare in Carex [11,17].
The number of chromosomes in Carex varies from n = 6 to n = 62 and spans actually a continuous series from n = 6 to n = 47, more than 100 species showing different cytotypes [2,17]. In addition, the chromosome count is not known yet for numerous sedges [2,18]. Counting chromosomes is the first step towards a better understanding of the karyotype evolution and the role of chromosome evolution in species diversification within Carex. Therefore, the present work was aimed at: (i) analysing the chromosome counts in 23 Carex taxa, including 7 with hitherto unknown chromosome numbers, and (ii) exploring relationships between the number of chromosomes and taxonomy of the genus Carex.  [29] 62-64 [30] c. 60 [31] c. 62 [32] (Continued ) Caucasus (except Ciscaucasia) and Turkey [19] (Continued )  Table 1), while C. cilicica Boiss. was collected from the only site known in Armenia, which has been well-documented in the herbarium of ERE. Voucher specimens for each taxon were deposited in the Herbarium Stetinensis at the University of Szczecin (SZUB). The nomenclature used follows Egorova [19] and Koopman [20], except for C. curvata [21], C. hartmaniorum A.Cajander [22], and C. nigra subsp. transcaucasica (T.V.Egorova) Jim.Mejías, G.E.Rodr.-Pal., Amini Rad & Martín-Bravo [23]. The names of sections used follow Egorova [19], Reznicek [24], and Ball & Reznicek [1].

Chromosome counts
Plant cuttings were transferred from soil to hydroponic cultures. When the new roots emerged, they were excised and immersed in ice-cold water for 16 h. The roots were subsequently fixed in Carnoy's solution (absolute ethanol: glacial acetic acid 3:1 v/v) for 24 hours at 4˚C. They were carefully washed in distilled water, and the root tips were dissected. Each root tip was macerated directly on a microscope slide in a mixture of 4% (w/v) pectinase (Fluka, Buchs, Switzerland), 6% (w/v) hemicellulase (Sigma-Aldrich, St. Louis, USA) and 4% (w/v) cellulase (Sigma-Aldrich, St. Louis, USA) in 0.01 M citric acid-sodium citrate buffer (pH 4.8), for 5 hours at 37˚C in a humidity chamber. Root tips were washed with 0.01 M citric acid-sodium citrate buffer (pH 4.8) and then with 45% acetic acid. Root tips were squashed under a cover glass. The cover slip was removed after freezing over dry ice, and the slides were air-dried overnight. The slides were dehydrated in a graded ethanol series (70%, 96%, and 99.8%) at room temperature, air-dried and stained with DAPI (1 μg/mL) (Sigma-Aldrich, St. Louis, USA) for 15 min. The slides were rinsed 3× with distilled water, air-dried and mounted in Vectashield1 Hard Set mounting medium for fluorescence (Vector Laboratories, Burlingame, USA) and analysed with the Axio Imager Z2 epifluorescence microscope (Carl Zeiss, Oberkochen, Germany). The resultant images were captured and analysed using the GenASIs software (Applied Spectral Imaging). About 60 slides per taxon were prepared and analysed (2 preparations × 30 plants per taxon). The accurate counting was carried out in at least 60 metaphase spreads per each taxon.
Chromosomes of all Carex species are very small making it impossible either karyotyping or determining the presence of structural aberration. Their identification based upon morphological features and size is unreliable. We have measured the chromosome lengths in 25 randomly chosen metaphase spreads of different species. The mean length of a Carex chromosome, based on 1600 measurements, was 1.01 μM (σ = 0.27) with minimum 0.48 μM and maximum 1.92 μM. It is because of their size that the analysis of the number of chromosomes was carried out in as many as 60 metaphase plates from each taxon. Only in this way the error can be avoided and the results are authenticated.

Chromosome numbers
The records, 2n = 32 for Carex aterrima subsp. medwedewii, cited by Gvinianidze & Avazneli [25], and 2n = 68 for C. hartmaniorum, reported by Lipnerová et al. [11] are doubtful because a similar chromosome number has not been recorded within the section Racemosae G. Don to which these taxa belong. Generally, within this section, two groups of cytotypes are given: the first group with 2n between 50-60 and the second with 2n between 100-106 [2]. Lipnerová et al. [11] addresses the section Racemosae as the product of polyploidy. Identification of polyploids in Carex is extremely difficult. In the case of autopolyploidy, a tetraploid species is expected to have twice as many chromosomes (in this work: C. buxbaumii, 2n = 102 and C. hartmaniorum, 2n = 52) and twice as big a genome than the initial diploid species. However, should the polyploidy event be relatively ancient evolutionarily, this direct relationship is most often blurred by a DNA sequence loss/acquisition, aneupolyploidy etc. occurring during evolution [57]. That is why different evolutionary scenarios in case of Carex aterrima subsp. medwedewii and C. hartmaniorum are possible. It can be hypothesised, that among Carex aterrima subsp. medwedewii there exist a diploid form (2n = 32) and a polyploidy one, which during its evolution has undergone different aneuploidy events, reaching the chromosome number of 52. It is confirmed by many studies that in neopoliploids a "genomic shock" occurs, which leads to many dysploidy and aneuploidy [58]. These changes are often inevitable to make the polyploid genome stable, properly functioning. Moreover, because Carex chromosomes are holocentric it can be expected that aneuploidy may occur on a larger scale. Therefore, it cannot be excluded, that large discrepancies in the number of chromosomes exist in one species, like for example in C. hartmaniorum (2n = 52 in this work, 2n = 68 in [11]).
The chromosome numbers in the remaining taxa examined in this work proved identical with or similar to literature data. Although the somatic chromosome number in C. bohemica was reported to be about 60 [30][31][32], other authors [18,[26][27][28] provided data indicating the chromosome count to be identical with that found in this work (2n = 80). However, as stated above, parallel existence of different cytotypes, even with a very diverse number of chromosomes is possible.
The difference in the chromosome counts, between this study and data reported in the literature, for C. buxbaumii, C. diluta, C. hordeistichos, C. otomana, C. pairae and C. pallidula could have resulted from a number of reasons. The first involves the technical difficulty of counting the very small chromosomes, whereby some authors report their counts as approximate, using "±" or "ca.". The Carex chromosomes are indeed small (ca. 1μm), which greatly hinders accurate counting; the relatively high number of chromosomes is an additional difficulty. This is, however, not the reason with which to plausibly explain such large discrepancies in the chromosome numbers in C. hartmaniorum and C. aterrima subsp. medwedewii. Another possible explanation of the discrepancy is a potential species misidentification. The third reason, probably the most important one, is the between-populations [59,60] or even between-individuals [61,62] variability. In addition, some species show a correlation between distribution at certain latitudes and the chromosome count variation [17,59,63]. However, the latitude-chromosome number correlation is not direct, and there is no pattern indicating an increase or a reduction in the chromosome number with latitude [62,64]. As we were comparing the chromosome numbers between sedges collected in Armenia and Poland (C. hartmaniorum) as well as in Armenia and the Czech Republic (C. secalina), we found no betweenpopulations differences.

Relationship between chromosome numbers and taxonomy of the genus Carex
The genus Carex seldom shows discontinuities in the chromosome count series at the intraspecific level or in species aggregates; discontinuities, however, do usually occur between sections or subsections [2,68]. This is in line with the scenario whereby sedge species gradually accumulate chromosome rearrangements, which is reflected in the selection dynamics at the cellular level or in non-random cytotype extinction, and generates discontinuities usually observed at the level of section or subsection [68]. However, the subgenus Vignea frequently shows similar (or even identical) chromosome counts at the section level, e.g. sections Ammoglochin and Phaestoglochin both have the dominant cytotype 2n = 58 [2,18]. In this case, the numbers of chromosomes are hardly suitable for species identification, e.g. C. brizoides L., C. curvata and C. praecox Schreb. (section Ammoglochin Dumort.) as well as C. muricata subsp. ashokae, C. pairae and C. divulsa Stokes (section Phaestoglochin). Within the section Ammoglochin, a clearly different chromosome number occurs in C. repens (2n = 70), most probably of hybrid origin [69]. The subgenus Vignea is regarded as monophyletic, whereas the remaining subgenera established earlier (Carex, Indocarex and Psyllophora) are considered polyphyletic [70]. Results of recent phylogenetic studies showed the genus Carex to encompass five groups: the Siderostictae clade, the Schoenoxiphium clade, the core Unispicate, Vignea and the core Carex [70; see also Fig 3].
The chromosome numbers in the subgenus Carex (core Carex) are usually different at section level and may be useful for establishing the status of a taxon in the sedge classification system. Chromosome numbers in the section Aulocystis usually range within 2n = 30-40, but the section contains also species with the cytotype 2n = 54 (C. cilicica; this study) or 2n = 56 and 58 (C. frigida All.; see [2]). This section, divided into numerous subsections [19], proves to be polyphyletic [70]. Some taxa, e.g. Carex frigida mentioned above, are-on the phylogenetic tree-far removed from the remaining members of the section Aulocystis [70,71]. Similarly, the taxonomic status of C. cilicica is not clear. Owing to differences in morphology [19] and the chromosome number between C. cilicica and taxa of the section Aulocystis, it seems hardly likely that the sedge is closely related to them. Kükenthal [72] assigned this species to the subsection Fuliginosae Tuckerm. within the section Frigidae Fries. (= Aulocystis), whereas Nilsson [73] put it in the section Fulvellae Fries ex Christ. The latter has been recently divided into two  [19,74]. The chromosome numbers in the section Spirostachyae are relatively well known and a substantial cytogenetic variability, 2n = 60-84, has been found [38,59,68,75,76]. The chromosome numbers in the section Ceratocystis range within 2n = 56-72 [61,77,78]. Most probably, the inclusion of C. cilicica in the section Ceratocystis or Spirostachyae rather than in the section Aulocystis would be more appropriate; therefore further studies-molecular analyses in particular-are necessary for unequivocal resolution of the taxonomic position of this species within the subgenus Carex.
The chromosome numbers in the section Acrocystis Dumort. usually range from 2n = 18 to 2n = 38 [2]. In this study, C. tomentosa was confirmed to belong to a cytotype of 2n = 48 which seems to be stable in this species [18]. According to Roalson, et al. [79], the section Acrocystis appears to be polyphyletic and some species, e.g. C. grioletii Roem. and C. tomentosa, should be excluded from it. This seems justified also because of differences in the chromosome numbers (2n = 48 in C. grioletii [80]). Kükenthal [72] included these species in the section Pachystylae Kükenth., whereas Egorova [19] assigned them to different subsections (the Elongatibracteatae Egor. and the Tomentosae Egor.) within the section Acrocystis. Phylogenetic studies carried out by the Gobal Carex Group [70] showed the species to be located at different sites on the phylogenetic tree: C. grioletii was within the section Thuringiaca G. Don., while C. tomentosa was placed in the vicinity of the section Paniceae G. Don [71]. In our opinion, the position of these species in the sedge classification system is not clear and requires further study.
The chromosome number in the Clandestinae G. Don is usually 2n = 35-56, except for C. callitrichos V.I.Krecz., C. lanceolata Boott and C. rhizina Blytt ex Lindblom which are all polyploid with 2n = 70 cytotype [11,18]. The section Clandestinae is a large and inhomogeneous group which is divided into numerous subsections [19]. Some taxa resemble one another morphologically and have similar distribution, e.g., C. digitata L. and C. pallidula, which renders their identification difficult [51]. Perhaps the chromosome numbers will prove useful in the identification of those species. The cytotype of Carex pallidula is 2n = 56 (as reported here) or 2n = 54 [49], whereas 2n = 52 appears to be the most frequent chromosome number in C. digitata throughout the whole natural range of the species [2]. Although Roalson [2] reported a cytotype variation (2n = 48, 2n = 50, 2n = 52, 2n = 54, and 2n = 56) in the latter taxon, the variation could have been caused by the fact that C. digitata s.l. has been split up recently in C. digitata s.s. and C. pallidula [81,82]. However, more detailed studies covering other areas of their occurrence are necessary to confirm that the number of chromosomes is appropriate for distinguishing these species.
Recent phylogenetic studies have demonstrated a close relationship between C. phyllostachys and the sedges of the subgenus Carex section Gynobasidae Trabut.: C. illegitima Ces. and C. oedipostyla Duval-Jouve within Schoenoxiphium clade [70; see also Fig 3], but this taxa substantially differ in morphology [74]. Most likely, C. phyllostachys is not closely related to the section Phyllostachyae Tuckerman ex Kükenthal species [70], the section grouping species occurring in North America [83]. The Phyllostachyae species' chromosome numbers range from 2n = 62 to 2n = 98 [84], the chromosome count in C. phyllostachys being 2n = 56 (determined in this study). The chromosome numbers in C. illegitima and C. oedipostyla are not known yet. Information on the chromosome counts in those species will most likely help to gain insight into the relationship between them and C. phyllostachys, because, as observed by Heilborn [35], closely related carices frequently show similar numbers of chromosomes.