New chromosome counts and genome size estimates for 28 species of Taraxacum sect. Taraxacum

Abstract The species-rich and widespread genus Taraxacum F. H. Wiggers, 1780 (Asteraceae subfamily Cichorioideae) is one of the most taxonomically complex plant genera in the world, mainly due to its combination of different sexual and asexual reproduction strategies. Polyploidy is usually confined to apomictic microspecies, varying from 3x to 6x (rarely 10x). In this study, we focused on Taraxacum sect. Taraxacum (= T.sect.Ruderalia; T.officinale group), i.e., the largest group within the genus. We counted chromosome numbers and measured the DNA content for species sampled in Central Europe, mainly in Czechia. The chromosome number of the 28 species (T.aberrans Hagendijk, Soest & Zevenbergen, 1974, T.atroviride Štěpánek & Trávníček, 2008, T.atrox Kirschner & Štěpánek, 1997, T.baeckiiforme Sahlin, 1971, T.chrysophaenum Railonsala, 1957, T.coartatum G.E. Haglund, 1942, T.corynodes G.E. Haglund, 1943, T.crassum H. Øllgaard & Trávníček, 2003, T.deltoidifrons H. Øllgaard, 2003, T.diastematicum Marklund, 1940, T.gesticulans H. Øllgaard, 1978, T.glossodon Sonck & H. Øllgaard, 1999, T.guttigestans H. Øllgaard in Kirschner & Štěpánek, 1992, T.huelphersianum G.E. Haglund, 1935, T.ingens Palmgren, 1910, T.jugiferum H. Øllgaard, 2003, T.laticordatum Marklund, 1938, T.lojoense H. Lindberg, 1944 (= T.debrayi Hagendijk, Soest & Zevenbergen, 1972, T.lippertianum Sahlin, 1979), T.lucidifrons Trávníček, ineditus, T.obtusifrons Marklund, 1938, T.ochrochlorum G.E. Haglund, 1942, T.ohlsenii G.E. Haglund, 1936, T.perdubium Trávníček, ineditus, T.praestabile Railonsala, 1962, T.sepulcrilobum Trávníček, ineditus, T.sertatum Kirschner, H. Øllgaard & Štěpánek, 1997, T.subhuelphersianum M.P. Christiansen, 1971, T.valens Marklund, 1938) is 2n = 3x = 24. The DNA content ranged from 2C = 2.60 pg (T.atrox) to 2C = 2.86 pg (T.perdubium), with an average value of 2C = 2.72 pg. Chromosome numbers are reported for the first time for 26 species (all but T.diastematicum and T.obtusifrons), and genome size estimates for 26 species are now published for the first time.


Introduction
Taraxacum F. H. Wiggers, 1780 (Asteraceae subfamily Cichorioideae) is a species-rich genus of common and widespread perennial grassland herbs growing from the subtropics to subarctic (arctic/alpine) regions across the world. Rough estimates suggest the genus contains approximately 2,800 species in approximately 60 sections (Kirschner et al. 2015), with the higher diversity in the mountains of Eurasia (Ge et al. 2011); a total of 1,900 species in 35 sections are listed for Europe (Kirschner et al. 2007). The complexity of Taraxacum taxonomy is caused by its combination of different reproduction strategies, including sexual reproduction (mainly outcrossing, less frequently selfing) and apomixis (meiotic diplospory; Richards 1973, Asker and Jerling 1992. The vast majority of Taraxacum taxa are apomictic polyploid microspecies, only a few species are sexual diploids. The phenomenon of apomixis itself (i.e. clonal reproduction by seeds) attracts the attention of plant systematists as well as plant breeders for its possible application in crop breeding.
Taraxacum sect. Taraxacum (formerly known as T. sect. Ruderalia; generally known as Taraxacum officinale group; see Kirschner and Štěpánek 2011) has a strongly prevailing triploid ploidy level of 2n = 3x = 24, by which it differs from other closely related sections (Erythrosperma, Palustria, and Celtica) with known ploidies of 3x and 4x or even higher. In this study, we aimed to count the chromosome number of 28 species for which knowledge was lacking and to detect the ploidy level for these species. Fur-thermore, we searched for variability in genome size among these species to determine whether we can detect variation in DNA content among species similar to that found in a sample of unidentified taxa of T. officinale group.

Plant Material
We studied a total of 28 Taraxacum species (25 formally described and three still undescribed, referred to by their working names) belonging to Taraxacum sect. Taraxacum (Table 1). Plants and achenes of the investigated species were collected in natural habitats of several localities of Central Europe in the period 2014-2016. A detailed description of the localities, date, and collectors of samples is provided in Table 1. The studied plant material was documented by herbarium specimens and is deposited in the herbarium of the Department of Botany, Palacký University in Olomouc, Czech Republic (OL). All studied species are apomictic (agamospermous); thus, maternal plants and offspring plants (grown from seeds) are taxonomically (genetically) identical.
For karyological analyses, achenes were sown in Petri dishes containing 1% agar solution and germinated at room temperature. Fresh young leaves for flow cytometric analyses were collected from juvenile plants cultivated in a greenhouse at the Department of Botany, Faculty of Science, Palacký University in Olomouc.

Karyology
For chromosome counts, we used mitotically active root tip meristems of dandelion seedlings. Seedlings of the investigated species with 1-2 cm long roots were collected in the morning. To obtain the desired metaphase index, the roots were pre-treated in a 2 mM solution of 8-hydroxyquinoline for two hours at room temperature and an additional two hours at 4 °C in the dark. Then, the material was fixed in Carnoy's fixative (a mixture (3:1, v/v) of absolute ethanol and acetic acid) and stored in a refrigerator (4 °C) until further processing (Hasterok and Maluszynska 2000). For slide preparation, a combination of protocols in Hasterok and Maluszynska (2000) and van Baarlen et al. (2000) was used with the following changes for the investigated species of dandelions. Fixed root tips were washed in citrate buffer (0.01 M, pH 4.8) for 5 min and then enzymatically digested in a mixture of 0.1% cellulose Onozuka RS (Trichoderma Persoon, 1794; Sigma), 0.1% pectolyase (Aspergillus japonicus Saito, 1906; Sigma) and 0.1% cytohelicase (Helix pomatia Linnaeus, 1758; Sigma) in the citrate buffer for 90 min at 37-40 °C. To remove trace amounts of the enzymatic mixture, the root tips were then gently washed in citrate buffer for 5 min. Only the mitotically active meristematic tissue of a root tip was cut off under a stereoscopic microscope, transferred into a drop of 50% acetic acid on a slide and covered by a cov- Table 1. List of species used in this study, with sampling details. Country codes according to ISO 3166-1 alpha-2 (AT = Austria; CZ = Czechia, DE = Germany, HU = Hungary, IT = Italy, SK = Slovakia); Collectors: BT = Bohumil Trávníček; RJV = Radim Jan Vašut. erslip. After heating the preparation to 42 °C for 1-2 min, cells were spread between a glass slide and coverslip in a drop of 50% acetic acid. The coverslip was mechanically removed by a razor blade after deep freezing in liquid nitrogen, and the slide was air dried. To increase the contrast of metaphase chromosomes for counting, the preparations were stained with DAPI (4',6-Diamidine-2'-phenylindole dihydrochloride; Vectashield Mounting Medium with DAPI, Vector Laboratories). For each species, at least ten metaphases were analysed to determine the chromosome number. Wellspread metaphase images were captured using Olympus BX 60 and Axio Imager Z.2 Zeiss fluorescence microscopes, both equipped with a CCD camera and ISIS software (Metasystems, Altlussheim, Germany).

Genome size estimation
The absolute genome size (2C-value; Doležel et al. 2007) of the fresh plant samples was quantified using a BD Accuri C6 flow cytometer (BD Biosciences, San Jose) equipped with a blue laser (488 nm, 20 mW, BD Accuri; BD Biosciences, San Jose). Sample preparation followed the standard protocol using LB01 isolation buffer supplemented with PVP (polyvinylpyrrolidone, 10 g/500 ml of buffer) to suppress interference of phenolic compounds with DNA staining Bartoš 2005, Doležel et al. 2007). Approximately 0.2 cm 2 of the plant tissue between secondary veins was chopped in 500 μl of LB01 buffer together with a similar amount of tissue of an internal standard. Due to peak overlap in some accessions, Solanum lycopersicum Linnaeus, 1753 'Stupické polní rané' (2C = 1.96 pg; Doležel et al. 2007) served as the primary reference standard, and Glycine max (Linnaeus, 1753) Merrill, 1917 'Polanka' (2C = 2.33 pg, re-calculated against a primary standard) served as the secondary standard. The suspension was filtered through a 42 μm nylon mesh, supplemented with 20 μl of RNase A type II-A (with a final concentration of 50 μg/ml) and incubated at room temperature for approximately 10 min. The sample was then stained with 20 μl of propidium iodide (PI; final concentration of 50 μg/ml) and incubated with occasional shaking for approximately 5 min at room temperature. A flow-through fraction was then run on the flow cytometer, and the relative fluorescence intensity of at least 5,000 particles was recorded. Each sample was analysed at least three times. If the range of variation in the three measurements exceeded the 2% threshold, then the outlying value was discarded, and the sample was re-analysed. Only G0/G1 peaks with coefficients of variation < 4% were accepted. The 2C-value was calculated by multiplying the 2C-value of the standard with the sample/standard fluorescence ratio. Monoploid genome size (1Cx-value) was calculated by dividing the 2C-value by the inferred chromosome number.

Results
The chromosome number of all 28 studied species of Taraxacum sect. Taraxacum  (Figs 1, 2). With respect to the position of the centromere, the chromosomes of all studied species were predominantly submetacentric or metacentric. The chromosome sizes were relatively small (Figs 1, 2). The smallest chromosome size in this study was 1.02 μm (T. ochrochlorum), and the largest one was 4.94 μm (T. baeckiiforme).
The DNA content of the twenty-six studied Taraxacum species (two species, i.e., T. chrysophaenum and T. subhuelphersianum, were not analysed due to low-quality fresh material) ranged 1.08-fold from 2C = 2.60 pg in T. atrox to 2C = 2.86 pg in T. perdubium ( Table 2). The average and median 2C-values for Taraxacum sect. Taraxacum (based on these 26 species) are 2.72 pg and 2.71 pg, respectively.

Discussion
Chromosome number variation differs among sections of the genus Taraxacum and more frequently occurs in sections such as Palustria or Celtica, whereas in section Taraxacum (and also section Hamata), it is nearly unknown. In our study, we aimed to either find variation in ploidy or confirm the prevailing triploid level. Our findings confirmed previously published records of 2n = 3x = 24 for T. diastematicum and T.
A tetraploid chromosome number (2n = 4x = 32) was counted for only a few species of the 165 species of T. sect. Taraxacum with known chromosome numbers in the Chromosome Counts Database (CDDB, version 1.45; Rice et al. 2015). None of the records can be considered fully reliable due to frequent misidentifications of the Taraxacum microspecies (lack of identification by specialists). Den Nijs and Sterk (1984) published two chromosome counts, i.e., triploid (2n = 3x = 24) and tetraploid (2n = 4x = 32), for species named as T. lacistrum Sahlin, 1982, and collected in France; however, the tetraploid number is listed as a question mark, and this chromosome number must therefore be considered dubious (due to the apomictic behaviour of microspecies, it is implausible to have 2 different ploidy levels for the same species). The chromosome number for a species from the High Atlas, T. atlantis-majoris H. Lindberg, 1932 was counted as tetraploid, but the species identification is mentioned as "T. cf. atlantis-majoris", and misidentification as other species (even from other sections, such as Piesis) cannot be excluded (Oberprieler and Vogt 1993). The tetraploid record for T. albertshoferi Sahlin, 1984(Sahlin 1984) cannot be accepted without doubt either, because in the same paper, T. franconicum Sahlin, 1984 (which is now considered a synonym of T. plumbeum Dahlstedt, 1911) is also described with a tetraploid chromosome number, which was confirmed to be erroneous; the correct one is triploid (e.g., Vašut 2003). The tetraploid record for T. mediterraneum Soest, 1954 (Cardona andContandriopoulos 1983;identified as T. balearicum Soest, 1961) does not refer how the taxon was determined. Rousi et al. (1985) published a tetraploid record for T. penicilliforme H. Lindberg, 1907 as a member of T. sect. Vulgaria (= T. sect. Taraxacum), but this species belongs to T. sect. Borea. Thus, the only somewhat reliable record of a tetraploid in Taraxacum sect. Taraxacum is for the alpine species T. venticola A. J. Richards, 1972(Richards 1972. Our list of species of T. sect. Taraxacum mainly includes typical members of the section, which differ slightly in their eco-geographic preferences. Some species have (in Central Europe) a preference for wet and sub-oceanic regions (such as T. corynodes, T. chrysophaenum, T. lucidifrons and T. ochrochlorum); on the other hand, some occupy more xerothermic regions (e.g., T. atrox, T. baeckiiforme, and T. lojoense). Some species resemble members of T. sect. Celtica (T. lucidifrons) or T. sect. Palustria (T. perdubium and T. sepulcrilobum). However, although the species in our study differ somewhat in ecology and geography, there is no variation in their ploidy levels. This is in agreement with previous studies in which only a triploid level was undoubtedly recorded for Nordic ("Atlantic") and Pannonian or Mediterranean species.
Genome size estimates in Taraxacum sect. Taraxacum are very limited. Only a few papers dealt with its genome size (Bennett et al. 1982, Záveský et al. 2005, Bainard et al. 2011, Iaffaldano et al. 2017, but none of these papers studied known apomictic microspecies; only unknown species of the T. officinale group were measured. Gener-  (Bennett et al. 1982, Záveský et al. 2005, Vidic et al. 2009, Temsch et al. 2010, Bainard et al. 2011, Iaffaldano et al. 2017summarized in , Chrtek et al. 2009). Genome size estimates vary in all taxa. Multiple factors can affect the measurement of genome size, e.g., differences in instrument settings among the instruments used (Doležel et al. 1998), using inadequate dye (DAPI vs. PI; Doležel et al. 1992), interactions between the dye and other molecules that lead to cytosolic effects (Noirot et al. 2000), and discrepancies in standardization (Doležel and Greilhuber 2010). Applying different laboratory procedures to the same species can lead to up to <10% variation; in the T. officinale group, different treatments led to a difference of up to 8.7% (Bainard et al. 2011). Therefore, at least part of the difference among published records can be attributed to a bias due to differences in laboratory procedures. We used a standardized procedure (buffers, tissue treatments, etc.) in our lab; therefore, the observed variation among the species used in this study likely reflects the real variation in DNA content.
Our study provided new data for 26 species of T. sect. Taraxacum, which confirmed no variation in chromosome number and ploidy level (2n = 3x = 24) and revealed only minor variation in DNA content that roughly equalled a possible methodological bias. The species sampled cover variation within the section: a sample of typical T. sect. Taraxacum species (most of the studied species) but also species that by morphology or ecology are intermediates of other sections, i.e., T. perdubium and T. sepulcrilobum, which are morphological and ecological intermediates between the studied section and T. sect. Palustria; or T. lucidifrons, which is morphologically similar to T. sect. Celtica or species resembling members of T. sect. Borea (T. ohlsenii, T. lojoense and T. atrox). Two species in our list are apolliniferous (T. atrox and T. subhuelphersianum). Such unusual homogeneity among species in T. sect. Taraxacum rather than great morphological (and ecological) variability might reflect a young evolutionary origin, which is likely in contrast to sections Palustria, Erythrosperma and others that may partly consist of evolutionarily older species (Wittzell 1999, Majeský et al. 2012, Kirschner et al. 2015. Although there is no evidence for the potential evolutionary scenario in European Taraxacum sections, we can speculate that the origin of apomictic species of T. sect. Taraxacum (T. officinale group) may be a result of "recent" hybridization between triploid apomicts and diploid sexuals in the sexual-asexual cycle in a mixed dandelion population, a phenomenon experimentally described in this group (Tas and van Dijk 1999, van Dijk 2003, van Dijk and Vijverberg 2005. In a mixed population (2x and 3x cytotypes; sexual and apomictic types), triploids are results of hybridization between triploid apomicts (diploid pollen) and diploid sexuals (haploid egg cell); however, a rare occurrence of tetraploidy (probably of temporary occurrence) can accelerate the formation of novel triploids (Verduijn et al. 2004b). These tetraploids probably occur in nature as a (rare) product of hybridization in mixed populations (probably discovered in the papers of Sato et al. 2014or Iaffaldano et al. 2017Ľ. Majeský, unpublished results) and function as a bridge in the formation of novel stable apomictic microspecies, but probably no such temporary tetraploid hybrids evolved in stable microspecies.