Preslia 82: 127–148, 2010 127 Genome size correlates with growth form, habitat and phylogeny in the Andean genus Lasiocephalus (Asteraceae) Velikost genomu andského rodu Lasiocephalus (Asteraceae) souvisí s životní formou, ekologií a fylogenezí Eva D u š k o v á1, Filip K o l á ř1,2, Petr S k l e n á ř1, Jana R a u c h o v á2,1, Magdalena K u b e š o v á1,2, Tomáš F é r1, Jan S u d a1,2 & Karol M a r h o l d3,1 1 Department of Botany, Faculty of Science, Charles University in Prague, Benátská 2, CZ128 01 Prague, Czech Republic, e-mail: jezanek@centrum.cz, filip.kolar@gmail.com, petr@natur.cuni.cz, kubesovm@gmail.com, tomas.fer@centrum.cz, suda@natur.cuni.cz; 2 Institute of Botany, Academy of Sciences of the Czech Republic, Zámek 1, CZ-252 43 Průhonice, Czech Republic, e-mail: jrauchova@post.cz; 3Institute of Botany, Slovak Academy of Sciences, Dúbravská cesta 9, SK-845 23 Bratislava, Slovak Republic, e-mail: Karol.Marhold@savba.sk Dušková E., Kolář F., Sklenář P., Rauchová J., Kubešová M., Fér T., Suda J. & Marhold K. (2010): Genome size correlates with growth form, habitat and phylogeny in the Andean genus Lasiocephalus (Asteraceae). – Preslia 82: 127–148. Variation in genome size in a particular taxonomic group can reflect different evolutionary processes including polyploidy, hybridization and natural selection but also neutral evolution. Using flow cytometry, karyology, ITS sequencing and field surveys, the causes of variation in genome size in the ecologically and morphologically diverse high-Andean genus Lasiocephalus (Asteraceae, Senecioneae) were examined. There was a 1.64-fold variation in holoploid genome size (C-values) among 189 samples belonging to 20 taxa. The most distinct was a group of plants with large genomes corresponding to DNA triploids. Disregarding the DNA triploids, the remaining samples exhibited a pronounced (up to 1.32-fold) and rather continuous variation. Plants with the smallest genomes most likely represent intergeneric hybrids with the closely related and sympatric Culcitium nivale, which has a smaller genome than Lasiocephalus. The variation in genome size in samples of diploid Lasiocephalus was strongly correlated with several environmental and life history traits (altitude, habitat and growth form). However, all these factors, as well as genome size itself, were correlated with phylogeny (main split into the so-called ‘forest’ and ‘páramo’ clades), which most probably represents the true cause of the differentiation in intrageneric genome size. In contrast, relationships between genome size and phylogeny were not apparent at lower divergence levels. Instead, here we suggest that ecological conditions have played a role in driving shifts in genome size between closely related species inhabiting different environments. Collectively, this study demonstrates that various evolutionary forces and processes have shaped the variation in genome size and indicates that there is a need for multi-approach analyses when searching for the causes and consequences of changes in genome size. K e y w o r d s: Andes, Compositae, flow cytometry, genome size, growth form, habitat preference, hybridization, Lasiocephalus, nuclear DNA amount, nucleotypic effect, phylogeny, polyploidy Introduction One of the current challenging questions in plant evolutionary biology is the functional significance of the approximately 2000-fold variation in genome size among angiosperms (Leitch & Bennett 2007). This tremendous variation is largely caused by different proportions of non-coding repetitive DNA, such as transposable elements, satellite DNA, introns, 128 Preslia 82: 127–148, 2010 and pseudogenes (Bennett & Leitch 2005). Several mechanisms, both at the chromosomal and molecular levels, are thought to be capable of generating genome size changes (Bennetzen 2005). Polyploidy, chromosome gain, presence of supernumerary or sex chromosomes, amplification of retrotransposons and genomic duplications are included among the processes that cause genome expansion. Counterbalancing mechanisms involve the loss of entire chromosomes, illegitimate recombination, unequal intrastrand homologous recombination or a higher overall rate of nucleotide deletion over insertion (reviewed in Bennett & Leitch 2005). While chromosome-level events typically cause sudden and marked changes in genome size, molecular mechanisms are more gradual with each event producing only slight modifications in genome size. Variation in genome size has been interpreted in relation to various phenotypic and/or life-history traits. Classical observations documenting a positive relationship between genome size and cell size and duration of cell division (reviewed by Leitch & Bennett 2007) were followed by studies that revealed correlations between genome size and seed mass, leaf mass per unit area, growth rate and/or photosynthetic rate (Knight & Beaulieu 2008). In addition, genome size can also have significant consequences at the ecological level, influencing the range of environmental conditions a plant can tolerate (the so-called “large genome constraint hypothesis”; Knight et al. 2005). Over the years, relationships between genome size and temperature, water availability, latitude and altitude of the habitats have been reported (reviewed by Knight & Ackerly 2002). However, variation in genome size may not always be adaptive as documented for instance in a highly variable population of Festuca pallens (Šmarda et al. 2007). Recent studies have also emphasized the importance of phylogeny in variation in genome size. For example, genome size in Hieracium subg. Hieracium was strongly correlated with the two major phylogenetic groups (Chrtek et al. 2009) and congruency between genome size and evolutionary lineages is also found in the holoparasitic Orobanche (Weiss-Schneeweiss et al. 2005). If phylogeny is not considered, the role of other factors (e.g. environmental conditions) in shaping genome size may be overemphasized. Therefore, when searching for the causes of variation in genome size, both adaptive and non-adaptive components need to be taken into account. Knowledge of the patterns and dynamics of variation in genome size is largely based on the analysis of European and North American floras, and species from the world’s main centres of plant diversity are neglected. One of the geographic regions that hosts an extremely-rich plant biota is the Andes of South America, which rank among the most conspicuous physiographic features on Earth (Vuilleumier & Monasterio 1986, Luteyn 1999, Young et al. 2002). Andean plant diversity largely resulted from the adaptive radiation of numerous plant groups, such as Espeletia, Huperzia, Lupinus and Valeriana (Cuatrecasas 1986, Wilkström et al. 1999, Bell & Donoghue 2005, Hughes & Eastwood 2006). Most researchers relate the adaptive radiation to the final uplift of the mountains and emergence of high-altitude non-forest habitats (páramo) about 5–3 mya and the PlioPleistocene climatic oscillations. Although there is an increasing number of phylogenetic studies on Andean plants (e.g. Young et al. 2002, Struwe et al. 2009) the role of genomewide evolutionary processes (such as polyploidization and genome size change) in species radiation is largely unknown. Lasiocephalus Willd. ex Schltdl. (Asteraceae, Senecioneae) is a neotropical highAndean genus of ca 30 species distributed from Venezuela to Bolivia, with the highest species Dušková et al.: Correlates of genome size in Lasiocephalus 129 richness in Ecuador (Cuatrecasas 1978). Two main growth forms are recognized in the genus: (i) the broad-leaved suffrutescent climbers of montane forests and the tree-line ecoton, and (ii) erect and ascending narrow-leaved herbaceous plants to subshrubs of the high-altitude páramo. These growth forms plausibly reflect adaptive changes that occurred during the colonization of the páramo habitats (Cuatrecasas 1978). It is likely that Lasiocephalus is a monophyletic group nested within Senecio s. str. with the high-Andean genus Culcitium as a sister taxon (Pelser et al. 2007). The high number of species along with the variation in growth forms and their altitudinal distribution make Lasiocephalus a suitable candidate for studying the evolutionary outcomes associated with adaptive radiation. We use Lasiocephalus as a representative example of a group that has radiated in the Andes to address general questions concerning the extent and possible sources of the variation in DNA content. Flow cytometry, karyology and ITS sequencing were employed to address the following specific questions: (i) Does genome duplication play a role in the diversification of Lasiocephalus? Which cytotypes are found in this genus? Are there any mixed-ploidy populations? (ii) What is the variation in genome size (C- and Cx-values)? (iii) How does the observed pattern in DNA ploidy and variation in genome size correlate with growth forms and habitat preferences? (iv) How does the observed pattern in DNA ploidy and variation in genome size correlate with the phylogeny of the group? Materials and methods Plant material Plants were collected between 2007 and 2009 in Bolivia, Ecuador and Venezuela (see Electronic Appendix 1 for locality details). In total, 189 individuals from 101 populations corresponding to 20 species of Lasiocephalus sensu Cuatrecasas (1978) were sampled (Table 1). Species were identified according to Cuatrecasas (1978) and by comparison with herbarium vouchers deposited at AAU, COL, G, MERF, NY, P, PRC, QCA, S and VEN. Seven distinct morphotypes of Lasiocephalus, which could not be assigned to any hitherto described species, were treated as separate morphospecies and referred to as Lasiocephalus sp. 1 to 7 (Table 1). GPS coordinates and altitude (WGS 84 system), habitat type and growth form of the plants were recorded at each locality. Young intact leaves were collected and quickly desiccated for DNA analyses using silica gel. The following plant material was sampled for flow cytometry: (i) fresh tissue (stems and leaves) of adult plants stored in plastic bags at 58 °C until analyses (up to 14 days), (ii) silica-dried leaf tissue of adult plants, and (iii) achenes from which seedlings were raised. In addition, three samples of closely related and often sympatric Culcitium nivale were also collected and analyzed for comparative purposes. Herbarium vouchers are deposited in PRC, QCA, QCNE and VEN. Flow cytometry DNA ploidy levels (Suda et al. 2006) were determined using flow cytometry and a Partec PA II instrument (Partec GmbH) equipped with a HBO mercury arc lamp. Genome sizes (Cand Cx-values; Greilhuber et al. 2005) were determined using CyFlow SL instrument (Partec GmbH, Münster, Germany) equipped with a green solid-state laser (Cobolt Samba, 532 nm, 100 mW). Sample preparation generally followed the simplified two-step proce- 130 Preslia 82: 127–148, 2010 Table 1. – Nuclear DNA contents of 13 species and seven unidentified morphospecies of Lasiocephalus and the closely related Culcitium nivale. Samples analysed = no. of populations/no. of individials. Values are mean±SD. Vicia faba ‘Inovec’ (2C = 26.9 pg) served as an internal reference standard. In the Grouping columns, different letters indicate groups of taxa that are significantly different at α = 0.05 using Tukey’s HSD multiple comparison Species Lasiocephalus campanulatus (Sch. Bip. ex Klatt) Cuatrec. Lasiocephalus cuencanus (Hieron.) Cuatrec. Lasiocephalus decipiens (Benoist) Cuatrec. Lasiocephalus heterophyllus (Turcz.) Cuatrec. Lasiocephalus involucratus (Kunth) Cuatrec. Lasiocephalus lingulatus Schltdl. – 2x Lasiocephalus lingulatus Schltdl. – 3x Lasiocephalus longipenicillatus (Sch. Bip. ex Sandw.) Cuatrec. Lasiocephalus mojandensis (Hieron.) Cuatrec. Lasiocephalus otophorus (Wedd.) Cuatrec. Lasiocephalus ovatus Schltdl. – 2x Lasiocephalus ovatus Schltdl. – DNA 3x Lasiocephalus patens (Kunth) Cuatrec. Lasiocephalus pichinchensis (Cuatrec.) Cuatrec. Lasiocephalus sodiroi (Hieron.) Cuatrec. Lasiocephalus L. sp. 1 (liana) Lasiocephalus L. sp. 2 (shrub) Lasiocephalus L. sp. 3 (shrubby liana) Lasiocephalus L. sp. 4 (cf. L. ledifolius (Kunth) C. Jeffrey) Lasiocephalus L. sp. 5 (L. involucratus-like) Lasiocephalus L. sp. 6 (L. sodiroi-like) – 2x Lasiocephalus L. sp. 6 (L. sodiroi-like) – DNA 3x Lasiocephalus L. sp. 7 (L. sodiroi-hairy) Culcitium nivale Kunth Genome size (‘fresh’ dataset) Samples analysed 2C-value (pg of DNA) Grouping – – 2/2 – 5/5 3/3 1/1 – – 3/3 3/3 1/1 5/6 1/1 1/1 4/6 1/2 1/1 2/2 – – – – 3/3 – – 16.10±0.08 – 15.63±0.10 15.07±0.15 22.69 – – 16.10±0.47 14.99±0.04 22.45 17.27±0.15 15.02 14.58 15.59±0.08 17.27±0.42 15.70 13.82±0.10 – – – – 12.35±0.10 – – ef – def bc – – – f bc – g bcd ab de g cdef a – – – – – dure using Otto buffers (Doležel et al. 2007). Briefly, intact leaf tissue of the Lasiocephalus plant to be analyzed and an internal reference standard (Vicia faba ‘Inovec’, 2C = 26.9 pg; Doležel et al. 1998) were mixed together and chopped with a sharp razor blade in a Petri dish containing 0.5 ml of ice-cold Otto I buffer (0.1 M citric acid, 0.5% Tween 20; Otto 1990). The suspension was filtered through a 42-μm nylon mesh and incubated for approx. 30 min at room temperature. The staining solution consisted of 1 ml of Otto II buffer (0.4 M Na2HPO4·12 H2O), β-mercaptoethanol (final concentration of 2 μl/ml) and a fluorochrome. Propidium iodide (PI) and RNase IIA (both at final concentrations of 50 μg/ml) were used to determine the genome size in absolute values (pg of DNA) in fresh samples, while AT-selective DAPI (4',6-diamidino-2-phenylindole) at a final concentration of 4 μg/ml was used for estimating the DNA content in relative units in silica-dried samples. Samples were stained for 10 min at room temperature and run on the flow cytometer. Isolated nuclei were excited either with a laser (for PI staining) or a mercury arc lamp (for DAPI staining) and the fluorescence intensity of 5000 particles recorded. 131 Dušková et al.: Correlates of genome size in Lasiocephalus test. DNA-triploids were excluded from statistical comparison. ITS clade: CN = ‘C. nivale’ clade; F = ‘forest’ clade; P = ‘páramo’ clade. Growth form: BH = broad-leaved herb; L = liana; NH = narrow-leaved herb; S = shrub. Habitat type: F = Andean forest; G = grass páramo; SB = subpáramo; SP = superpáramo. Relative DNA content (‘silica’ dataset) Samples analysed Sample/ standard ratio Grouping 2/2 2/4 1/1 2/8 9/18 6/15 2/2 4/7 2/6 1/1 7/19 3/4 4/14 10/21 2/3 4/5 – 1/2 4/6 1/6 1/2 1/2 1/4 1/1 0.667±0.000 0.705±0.013 0.646 0.665±0.012 0.628±0.019 0.620±0.018 0.855±0.031 0.661±0.021 0.617±0.026 0.671 0.596±0.011 0.860±0.057 0.688±0.023 0.612±0.016 0.599±0.003 0.622±0.011 – 0.602±0.012 0.535±0.007 0.626±0.010 0.556±0.024 0.771±0.009 0.545±0.007 0.494 fghi i cefghi gh ef de – g ce efghi bc – hi ce bcd cef – bce a def ab – a – ITS clade Growth form Habitat type F F – F F P P F – F P P F F+P P – – – CN F – – – CN L L L L L NH NH NH BH L NH NH L NH NH L S L NH L NH NH NH NH F F F F SB G G SP G SB SP SP F G SP SB SB SB SP SB G G SP SP Two distinct datasets were produced and treated separately in subsequent statistical analyses. Firstly, relative DNA content was determined for 152 silica-dried Lasiocephalus and one Culcitium nivale sample (further referred to as the ‘silica’ dataset) using DAPI flow cytometry. The desiccated samples were not more than 18 months old when analyzed. DAPI was chosen because it is less sensitive to secondary metabolites and chromatin arrangement, which allowed a high-resolution analysis even of dehydrated plant tissues (Suda & Trávníček 2006). Only histograms with coefficients of variation (CVs) of the G0/G1 peak of the analyzed sample below 5% were considered. Secondly, absolute values of genome size of a subset of 37 fresh accessions of Lasiocephalus and three accessions of Culcitium nivale (‘fresh’ dataset) were determined using PI flow cytometry. More stringent criteria for the quality of the flow analysis were applied in this case: (i) peaks of both the sample and the internal standard of approximately the same height, (ii) CV of G0/G1 peak of the analyzed sample below 4%, (iii) three replicates of each sample on different days to minimize potential random instrumental drift, and (iv) the between-day variation in fluorescence intensity below 3%; otherwise, the most extreme 132 Preslia 82: 127–148, 2010 125 200 B 75 80 50 40 25 Vicia faba 120 100 L. patens L. sp. 1 Number of nuclei 160 L. campanulatus A 0 0 0 200 400 600 800 Relative fluorescence 1000 0 200 400 600 800 1000 Relative fluorescence Fig. 1. – Fluorescence histograms for Lasiocephalus samples analyzed using flow cytometry. (A) Simultaneous analysis of propidium iodide stained nuclei isolated from fresh tissue of L. sp. 1 (2C = 15.59 pg, pop. no. P2) and L. patens (2C = 17.27 pg, pop. no. P12), which reveal interspecific differences in genome size. (B) Analysis of relative DNA content of silica gel preserved samples of L. campanulatus (pop. no. C1A). Isolated nuclei were stained with DAPI and fresh leaf tissue of Vicia faba was used as an internal reference standard. See Electronic Appendix 1 for population details. value was discarded and the sample re-analyzed. The reliability of FCM measurements (i.e. between-plant differences) was repeatedly confirmed by simultaneous analyses of Lasiocephalus samples with distinct genome sizes (Fig. 1). The presence of two distinct peaks in analyses of co-processed samples is considered the most convincing evidence for genuine differences in nuclear DNA content (Greilhuber 2005). To assess relationships between DAPI (AT-selective fluorochrome) and PI (intercalating fluorochrome without base bias) staining, a subset of thirteen fresh DNA diploid Lasiocephalus accessions covering the whole range of monoploid genome sizes was analyzed using both PI and DAPI flow cytometry concurrently. The analyses followed the above-defined stringent criteria with three independent replicates for each sample and each stain. Chromosome counts To confirm the FCM results, three individuals covering the whole genome size range at the diploid level (Table 1, Electronic Appendix 1) were subjected to conventional karyological counts using rapid squash methods. The apical root meristem of young seedlings was pre-treated with p-dichlorbenzene for 3 h, fixed in ice-cold 3 : 1 ethanol : acetic acid for 12–14 h, macerated for 1 min in 1 : 1 ethanol : hydrochloric acid, and stained with lacto-propionic-orcein (Dyer 1963). ITS cloning and sequencing Sequences of the nuclear ribosomal internal transcribed spacer (ITS) region were analyzed for 13 Lasiocephalus taxa (Table 1). Total genomic DNA was extracted from silicadried material using the Invisorb Spin Plant Mini Kit (Invitek) according to the manufacturer’s instructions. The ITS1-5.8S rDNA-ITS2 region of the nuclear ribosomal DNA was Dušková et al.: Correlates of genome size in Lasiocephalus 133 amplified using the primers ITS5 (White et al. 1990) and ITS_Lasio (5’AGTCGRAGCATCGTCATGAGA-3’). PCR amplifications were done in 22 μl reaction containing 0.18 mM of each dNTP (Fermentas), 0.23 mM of each primer (Sigma), 0.5 units of JumpStart REDTaq polymerase (Sigma), 1 × PCR buffer for JumpStart REDTaq (Sigma) and 5 ng of genomic DNA. An initial denaturation step at 94 °C for 1 min was followed by 35 cycles of denaturation (94 °C for 50 s), annealing (53 °C for 50 s) and extension (72 °C for 1 min) steps and a final extension at 72 °C for 10 min. Each sample was amplified in two separate reactions. Both PCR products were mixed and purified using JETQUICK PCR Product Purification Spin Kit (Genomed). The purified ITS templates were ligated to the pGEM-T Easy vector (Promega) following the manufacturer’s protocol but downscaled to half volume reactions and subsequently transformed into competent cells. After an overnight culture at 37 °C on the LB ampicillin/IPTG/X-gal selective plates, colonies carrying the ITS insert were identified by colour. 10–12 colonies from each reaction were selected for colony PCR using the same conditions as in the first PCR reaction. PCR products were purified using JETQUICK PCR Product Purification Spin Kit (Genomed) and subsequently sequenced (Macrogen, lll) using the ITS5 primer. Phylogenetic and statistical analyses Chromatograms of the cloned sequences were inspected by eye using Finch TV sequence viewer. In addition, direct GenBank sequences of nine species identified as putative closest relatives of Lasiocephalus by Pelser et al. (2007) were included as an outgroup reference. Initial sequence alignment was done with ClustalW algorithm incorporated in BioEdit v5.0.6 (Hall 1999) and manually improved. The same program was used for further manual editing. Autapomorphies that were found only in one clone within the whole dataset were considered to be polymerase errors and discarded (Popp et al. 2005). Two sequences of a highly ribotype-variable species L. pichinchensis, which combined apomorphies of the two main clades (see Results), were identified as chimeras and therefore omitted from analyses. Sequences were submitted to GenBank (GU289931 – GU290036). Bayesian analysis (MrBayes v3.1.2; Huelsenbeck & Ronquist 2001, Ronquist & Huelsenbeck 2003) was run on the complete dataset with GTR model of nucleotide substitutions, gamma model of rate heterogeneity and variable proportion of invariable sites. This model was found in hierarchical likelihood ratio tests as the model of molecular evolution best fitting the data as implemented in Modeltest v3.5 (Posada & Crandall 1998). Two parallel runs with four chains each were used, sampling every 1000th tree for 3 million generations until convergence. The first 750 trees per run were discarded as burn-in and the remaining 2251 trees per run were summarized. Correlation between DAPI and PI measurements of the same accessions was examined in Statistica 8. Differences in nuclear DNA contents were tested using an analysis of variance (ANOVA) and further assessed by Tukey’s HSD multiple comparisons incorporated in the R software 2.9.2 (R Development Core Team 2009). The differences between growth forms and preferred habitat types were assessed by ANOVA with species identity considered as a factor with random effect (nested in the particular factor) in order to avoid pseudoreplication. The relationship between DNA content and altitude was tested by linear mixed-effect model (species identity treated as a factor with random effect) incorporated in the nlme package in R (Lindstrom & Bates 1990). In all analyses the random effect 134 Preslia 82: 127–148, 2010 was modelled only for the intercept, as was suggested by the likelihood ratio test (only non-significant improvement when a more complex model taking into account also the effect of the random factor for the slope was considered). The relationships between genome size and phylogeny were examined by testing the association of nuclear DNA content with the two main clades detected (i.e. the ‘páramo’ and the ‘forest’ clade; see Results). Due to non-homogeneity of variances (indicated by highly significant Bartlett test: P < 0.002 in both datasets), differences in DNA content (mean values per species) between clades were tested by the non-parametric Mann-Whitney test implemented in Statistica 8 (StatSoft 2007). Species for which there was no sequence information were omitted from this kind of analysis. Within the ‘forest’ clade, differences in relative genome size (the ‘silica’ dataset) among the three main subclades (Table 1) were tested using the Kruskal-Wallis test implemented in Statistica 8. The relationship between DNA content and altitude was assessed separately for each of the two main clades (Table 1) using the same procedure as in the whole dataset (see above). In order to quantify the correlation between environmental factors (habitat and growth form) and phylogeny (split into two main clades), variation partitioning based on three-way ANOVA with sequential (Type I) sum of squares was done in Statistica 8. Mean DNA content values of each species were subjected to six three-way ANOVAs with different input sequence of factors and model sums of squares used as variance component measures. Results DAPI versus PI staining A highly linear DAPI/PI correlation was observed across the whole range of genome sizes of the 13 selected fresh diploid Lasiocephalus samples (the relationship was: DAPI fluorescence = 1.126 × PI fluorescence, R2 = 0.99, P < 0.001). This indicates a constant AT/GC content within the genus and enables a reliable conversion and interpretation of the estimated values. Chromosome counts Karyological analyses of three putatively diploid plants with markedly different genome sizes [Lasiocephalus sp. 4 (cf. L. ledifolius)], pop. G18, 2C = 13.82 pg; Lasiocephalus sp. 3, pop. no. 11567, 2C = 15.70 pg; and shrubby Lasiocephalus sp. 2, pop. no. 11584, 2C = 17.27 pg) confirmed the diploid number of chromosomes (2n = 2x = 40) in all samples (Fig. 2). DNA ploidy levels The analysis of 37 fresh and 152 silica-dried Lasiocephalus samples revealed a large variation in nuclear DNA content within this genus, spanning 1.64-fold range (2C = 13.82–22.69 pg) and 1.60-fold range (sample/Vicia ratio = 0.535–0.860), respectively (Table 1). In each dataset, two well-defined genome size groups (separated by a distinct gap) were detected. The fluorescence values of the two groups differed by ~1.5-fold, indicating that DNA diploids and DNA triploids were involved (Table 1). DNA triploids were represented by ten samples belonging to three species (L. lingulatus, L. ovatus and L. sp. 6) Dušková et al.: Correlates of genome size in Lasiocephalus 135 Fig. 2. – Mitotic chromosomes of Lasiocephalus sp. 3 (shrubby liana, pop. no. 11567), 2n = 40. Scale bar = 10 µm (photo V. Jarolímová). and occurred in eight populations (~8% of all sampled populations). In most cases, DNA triploids grew in sympatry with more abundant diploid plants. Genome size variation Disregarding the DNA triploids, the remaining Lasiocephalus accessions still exhibit a pronounced variation in genome size. Mean 2C-values of fresh accessions ranged from 13.82 pg in Lasiocephalus sp. 4 (cf. L. ledifolius) to 17.27 pg in shrubby Lasiocephalus sp. 2 and L. patens, (1.25-fold variation). The variation in silica-dried samples was 1.32-fold, with L. sp. 4 (cf. L. ledifolius) possessing the smallest and L. cuencanus the largest genomes (Table 1, Fig. 3). The closely related Culcitium nivale possessed a smaller genome than any species of Lasiocephalus (mean 2C = 12.35 pg; Table 1). In both datasets, samples with the smallest genome sizes formed a discrete group (Fig. 3). This group consisted of two taxa of putative hybrid origin – Lasiocephalus sp. 4 (cf. L. ledifolius) (a putative intergeneric hybrid between L. ovatus and Culcitium nivale) and Lasiocephalus sp. 7 (a putative hybrid between L. sodiroi and C. nivale). DNA content values of suspected hybrids were intermediate between values of their putative parents. A continuous variation in nuclear DNA content was observed in the remaining samples. Nevertheless, this variation was sorted according to the growth form and preferred habitat (Fig. 3). Analysis of variance supported significant differences in genome size among the different habitats (F3,23 = 7.16, P = 0.01 and F3,125 = 9.38, P = 0.001 for ‘fresh’ and ‘silica’ datasets, respectively) and growth forms (F2,23 = 6.23, P = 0.02 and F2,125 = 5.10, P = 0.02 for ‘fresh’ and ‘silica’ datasets, respectively). Generally, the genome size of the Lasiocephalus species analyzed increased with decreasing altitude as documented by 136 Preslia 82: 127–148, 2010 Fig. 3. – Variation in holoploid genome size of putatively diploid Lasiocephalus. (A) Absolute genome sizes (2C values) in pg DNA of 35 accessions of 13 Lasiocephalus species (‘fresh’ dataset, staining with propidium iodide). (B) Relative fluorescence intensity (expressed as a ratio of the sample and the internal reference standard, Vicia faba) in 144 accessions of 19 Lasiocephalus species (‘silica’ dataset, stained with DAPI). Different symbols refer to different habitat types (growth forms correlate with the habitat types, see Table 1). Circles depict putative hybrids between Lasiocephalus spp. and Culcitium nivale. Note the generally increasing DNA content in species inhabiting different habitats in the sequence: superpáramo < grass páramo < subpáramo < Andean forest. Dušková et al.: Correlates of genome size in Lasiocephalus 137 the following habitat sequence: 1) superpáramo plants (mean 2C = 14.53 pg, mean sample/standard ratio = 0.591), 2) grass páramo plants (mean 2C = 15.06 pg DNA, mean sample/standard ratio = 0.616), 3) subpáramo plants (mean 2C = 15.71 pg, mean sample/standard ratio = 0.626) and 4) forest plants (mean 2C = 16.97 pg DNA, mean sample/standard ratio = 0.681). Similarly, a progressive increase in DNA content was observed among different growth forms. The narrow-leaved herbaceous plants had the lowest mean DNA values (2C = 14.74 pg, sample/standard ratio = 0.603), followed by broad-leaved herbaceous plants (sample/standard ratio = 0.617), lianas (2C = 16.15 pg, sample/standard ratio = 0.652) and a shrub (2C = 17.27 pg). The only exception was L. longipenicillatus, a narrow-leaved superpáramo herbaceous plant, with a distinctly higher genome size than other species of the same growth form and habitat (see Fig. 3). It should, however, be noted that the negative correlation between genome size and altitude was statistically supported only by the ‘silica’ dataset (F1,124 = 7.611, P = 0.007) but not by the less representative dataset of fresh samples (F1,22 = 0.001, P = 0.996). ITS phylogeny and genome size Bayesian phylogenetic analysis of ITS sequences run on a subset of 13 Lasiocephalus species revealed two well-supported lineages (Fig. 4), referred to as ‘forest’ and ‘páramo’ clades because of the markedly different habitat preferences of their members. The ‘forest’ clade consists almost entirely of lianas inhabiting Andean forest and subpáramo habitats (see Table 1). The only exception was L. longipenicillatus, a superpáramo herbaceous plant from Venezuela. The ‘forest’ clade appears to be further subdivided into three sub-clades: (i) forest + superpáramo taxa (sub-clade ‘a’ at Fig. 4), (ii) entirely forest species (sub-clade ‘b’), and (iii) forest + subpáramo taxa (sub-clade ‘c’). The ‘páramo’ clade is composed of four exclusively grass páramo and superpáramo narrow-leaved herbaceous plants with unresolved relationships. The closely related Culcitium nivale and one putatively hybrid taxon Lasiocephalus sp. 4 (cf. L. ledifolius) with similar growth form and habitat preferences, are a sister group to the ‘páramo’ clade. One species (L. pichinchensis) possessed ITS ribotypes characteristic of both major clades (Fig. 4); sequences were similar to L. otophorus and L. ovatus from the ‘forest’ and ‘páramo’ clades, respectively, and were isolated from a single cloned individual. For this reason, L. pichinchensis was omitted from subsequent tests of correlation between genome size and phylogeny. Our data do not support the monophyly of the genus Lasiocephalus sensu Cuatrecasas (1978) because of the polytomy at the base of the whole group, which also includes several species of the closely related genus Culcitium (Fig. 4). The two main clades differ markedly in DNA ploidy and DNA content variation. DNA triploids were entirely confined to the ‘páramo’ clade (polyploids accounted for 17% of all samples in this clade) whereas the ‘forest’ clade was ploidy-uniform. In addition, diploid species from the ‘páramo’ clade have significantly smaller genome sizes than their ‘forest’ counterparts based on both the ‘fresh’ (Mann-Whitney Z = 2.121, P = 0.034; mean 2C = 14.97 and 16.43 pg, respectively) and ‘silica’ datasets (Z = 2.717, P = 0.007; mean sample/standard ratio = 0.606 and 0.658, respectively). Relationships between DNA content and growth form, habitat preferences and altitudinal range within each of the two main clades were further analyzed in the more representative ‘silica’ dataset. Species occupying different habitats also differed significantly 138 Preslia 82: 127–148, 2010 Dušková et al.: Correlates of genome size in Lasiocephalus 139 Fig. 4. – Phylogenetic analysis of ITS sequences of thirteen Lasiocephalus species and their closest relatives in the Senecio s.l. group. A Bayesian 50% probability consensus tree of 4502 trees is shown with posterior probabilities indicated above the branches. (A) Basal part of the tree and ‘páramo’ clade. (B) ‘forest’ clade further divided into three subclades (designated by a, b and c letters). Mean 2C-values for ‘forest’ and ‘páramo’ clades were 16.43 pg and 14.97 pg DNA, respectively. Different patterns of the branches indicate different habitat preferences of the respective accession/lineage. 140 Preslia 82: 127–148, 2010 in their DNA content values both within the ‘forest’ clade (F2,52 = 12.681, P = 0.011) and within the ‘páramo’ clade (F1,34 = 170.16, P = 0.048). On the other hand, C-values in the ‘forest’ clade were not associated with different growth forms (F1,52 = 0.013, P = 0.915); this test was not performed for the ‘páramo’ clade because all species were narrow-leaved herbaceous plants. The DNA content of ‘forest’ clade taxa increased with decrease in altitude (F1,51 = 4.342, P = 0.042) while no such association was found within the ‘páramo’ clade (F1,33 = 0.048, P = 0.828). In addition, a clear phylogenetic structure within the ‘forest’ clade was subjected to further analysis. However, only non-significant differences in fluorescence values among the three subclades (Fig. 4) were detected (Kruskal-Wallis H = 0.125, P = 0.939). Relations among the explanatory variables Variation partitioning provided a better insight into the relations among habitat, growth form and phylogeny that were selected as possible explanatory variables for the intrageneric variation in genome size in Lasiocephalus. Strong correlations among the variables were obtained (Fig. 5). A large proportion of the variation was explained by a combination of all three factors (71% and 48% for ‘fresh’ and ‘silica’ datasets, respectively). Habitat seemed to be the single most important explanatory factor, although this result may be biased due to an unequal number of degrees of freedom for different variables (3 for habitat type versus 1 for ITS clade and 2 for growth form). Discussion This study aimed at addressing the role of different evolutionary forces in shaping genome size in a diverse plant group. We estimated relative or absolute genome size values in a representative sample set of the high-Andean genus Lasiocephalus (20 taxa, 189 samples) and interpreted the results in the light of the phylogenetic relationships, growth form and ecological preferences. DNA ploidy level variation Polyploidy is widely acknowledged as the key force in angiosperm evolution with recent or ancient genome duplication being detected in virtually all angiosperms (Soltis et al. 2009). Although no polyploids have previously been reported in Lasiocephalus, our FCM data provide clear evidence for intrageneric cytotype polymorphism. The most parsimonious explanation for the existence of two discrete groups of genome sizes (differing by ~1.5-fold) is the incidence of diploid and triploid plants. Because exact chromosome counts for putatively triploid individuals are not available we refer to these large-genome accessions as DNA triploids (Suda et al. 2006). DNA triploids are most probably of autopolyploid origin as suggested by their: (i) close morphological similarity with diploid individuals of the corresponding species, and (ii) regular sympatric growth of 2x and 3x cytotypes in the same populations. Indeed, no population for which several samples were analyzed was composed entirely of DNA triploids (but triploids were detected in two populations of L. lingulatus for which only one individual was analyzed). Most likely, triploids originated recurrently in each population as Dušková et al.: Correlates of genome size in Lasiocephalus 141 Fig. 5. – Partitioning of variance in genome size of Lasiocephalus samples explained by habitat, growth form and phylogeny. (A) ‘fresh’ dataset, (B) ‘silica’ dataset. The numbers represent percentages of explained variance (which was 96% and 83% of the total variance in A and B, respectively). The relative importance of a particular factor or factor combination is distinguished by different shades of grey. Note the strong correlation of all three factors. High proportion of variance explained solely by the habitat may be related to the higher number of degrees of freedom for this variable. Differences in the percentages of explained variance between both datasets reflect unequal sample sizes. a result of fusion of reduced and unreduced gametes of diploids, such as in Picris (Slovák et al. 2009) and Quercus (Dzialuk et al. 2007). Interestingly, all ten DNA triploids belonged to only three narrow-leaved Lasiocephalus taxa, namely L. lingulatus, L. ovatus and L. sp. 6 (resembling L. sodiroi). All of them grow in the high-altitude grass páramo or superpáramo and are a part of the ‘páramo’ clade (Lasiocephalus sp. 6 was not sequenced but is believed to belong to this clade because of close phenotypic similarities to another clade member, L. sodiroi). Genome duplication thus seems to be restricted to a small group of closely related taxa inhabiting the highest and most stressful environments. This pattern can be explained by the higher viability of polyploids in harsh conditions and/or by the higher frequency of unreduced gamete formation in a stressful and fluctuating climate (regular frost, high solar radiation, etc.; Sarmiento 1986). Relatively higher incidence of triploid plants at high altitudes is also recorded for instance in the Cardamine pratensis group (C. rivularis; Marhold 1994). Genome size variation at diploid level The three chromosome counts (2n = 40) determined in this study confirm the previous numbers recorded for Venezuelan Lasiocephalus longipenicillatus and Ecuadorian L. involucratus, L. patens and Culcitium nivale (Powell & Cuatrecasas 1970, Robinson et al. 1997). Despite quite a high number of somatic chromosomes, all these counts are regarded as diploid based on x = 20. This number seems to be the base chromosome number also in 142 Preslia 82: 127–148, 2010 some other representatives of the widely conceived genus Senecio (see also Suda et al. 2007), although some authors suggest x = 10 (e.g. Hodálová et al. 2007 and references therein). Because the karyologically-confirmed diploid plants are distributed across the entire range of genome sizes that remain after the exclusion of DNA triploids (Table 1), the 1.32-fold difference was interpreted as a genuine variation in monoploid genome size. It is likely that amplification or deletion of non-genic repetitive DNA (transposable elements in particular), among other mechanisms, drive this variation (Bennetzen et al. 2005, Cavalier-Smith 2005, Piegu et al. 2006). Genome size and homoploid hybridization Genome size is mostly stable at the species level, while it often shows considerable differences even among closely related species (Greilhuber et al. 2005, Leong-Škorničková et al. 2007, Ekrt et al. 2009, Loureiro et al. 2010). As such, genome size can be used as a supportive species-specific characteristic and is also a helpful marker for investigating interspecific hybridization. The value of genome size data lies in the fact that both heteroploid and homoploid hybrids can be identified based on intermediate DNA amounts (e.g. Šiško et al. 2003, Mahelka et al. 2005, Yahata et al. 2006, Kron et al. 2007, but see Bureš et al. 2004). In this study, two different intergeneric hybrids between Lasiocephalus spp. and closely related Culcitium nivale were assumed on the basis of field observations and plant phenotypes. In both cases, FCM data confirmed our expectation. The average genome size of Lasiocephalus sp. 4 (cf. L. ledifolius) (2C = 13.82 pg), a putative hybrid between L. ovatus (2C = 14.99 pg) and C. nivale (2C = 12.35 pg), closely matched the theoretical DNA amount of the F1 hybrid (13.67 pg). Putative crosses showed intermediate morphological characteristics with respect to the size and shape of leaves, type of indumentum, capitula diameter, etc. Further support for their hybrid origin comes from the fact that only a small number of often non-flowering individuals scattered among abundant and spatially intermingled putative parents was observed at all the localities sampled. An analogous situation exists in a mixed population of C. nivale and another páramo species, L. sodiroi. The population of plants provisionally called Lasiocephalus sp. 7 (no. 11088 in Electronic Appendix 1) is comprised of morphologically transient individuals with fluorescence values (mean sample/standard ratio of 0.545) intermediate between the values of putative parents (0.494 and 0.599, respectively). In contrast to the FCM results, the ITS sequences did not provide direct evidence supporting the hybrid origin of Lasiocephalus sp. 4 (cf. L. ledifolius). All clones isolated from two L. sp. 4 individuals originating from two different populations were clearly assigned to the Culcitium nivale lineage (most of them fell directly among the C. nivale accessions). The absence of L. ovatus-type ribotypes can be explained by at least two non-exclusive hypotheses: (i) the L. ovatus-belonging paralogues were not sampled due to either the limited number of clones (10 clones in total) or because of biased PCR amplification (Wagner et al. 1994, Kanagawa 2003), (ii) concerted evolution could have homogenized the paralogues in putative hybrid accessions towards a C. nivale copy. Rapid homogenization of rDNA in several generations is reported for instance in allopolyploid Nicotiana (Kovarik et al. 2004) and Tragopogon (Kovarik et al. 2005). Ongoing molecular work aims to shed more light on the evolutionary history of Lasiocephalus sp. 4 (cf. L. ledifolius). Dušková et al.: Correlates of genome size in Lasiocephalus 143 Genome size and phylogeny The recent burst of phylogeny-based studies on genome size variation revealed a strong phylogenetic component in this trait at various taxonomic levels, from seed plants as a whole (Leitch et al. 2005), through family (Leitch et al. 2007) and (sub)genus (Weiss-Schneeweiss et al. 2005, Garcia et al. 2008, Chrtek et al. 2009) up to the intraspecific level (Popp et al. 2008). Neutral evolution (i.e. random and passive changes in genome size) may often be as important as selective processes in shaping genome size (Lysak et al. 2009). Two main clades were identified in the Lasiocephalus accessions analyzed (Fig. 4) – one containing lianas from the Andean forest and subpáramo habitats (plus one superpáramo species, L. longipenicillatus) (‘forest’ clade) and the other exclusively grass páramo and superpáramo narrow-leaved herbaceous plants (‘páramo’ clade). Interestingly, both clades differed significantly in genome size, with smaller genomes recorded in the ‘páramo’ clade. On the other hand, genome size in Lasiocephalus did not appear to be affected so much by phylogeny at the lower divergence levels as documented by the non-significant differences in genome size between the three subclades within the ‘forest’ clade. Any further considerations of the ancestral state and possible direction of genome size evolution in the two main clades are, however, largely obscured by a polytomy at the base of the whole group that involves several Culcitium species. Moreover, C. nivale appears to be more closely related to the Lasiocephalus ‘páramo’ clade than to other members of its own genus (Fig. 4). Thus the data indicate that in order to re-establish the group as monophyletic C. nivale, at least, has to be included in Lasiocephalus. Nevertheless, at this stage of the investigation it is not possible to exclude that the closeness of the relationship between Lasiocephalus and C. nivale is overestimated due to the incidence of hybridization and introgression. In addition, only direct non-cloned sequences of other Culcitium species were obtained from the GenBank (Pelser et al. 2007). As noted by Soltis et al. (2008), the use of direct ITS sequences could mislead the phylogenetic inference especially in paralogue-rich groups. We also admit that the explanatory value of the current phylogenetic tree may be limited by the incomplete taxon sampling (only 13 out of 20 FCM-analyzed Lasiocephalus taxa were included). On the other hand, the sequenced dataset is sufficiently representative. It covers the whole range of estimated genome sizes and all major types of growth forms and habitats (perhaps with a single exception of the broad-leaved herbaceous L. mojandensis). A multi-marker molecular analysis applied to a larger dataset is currently in progress in order to draw more robust conclusions regarding the phylogenetic relationships between Culcitium and Lasiocephalus. Genome size and environmental correlates The influence of environmental factors on genome size evolution has been debated for years and is still a matter of controversy. Significant correlations of genome size with various ecological factors (see Leitch & Bennett 2007 for a review) differ when different model systems are compared. For example, of the correlations of genome size with altitude for 24 genera, approximately one third are positive, one third negative and one third not significant (Knight et al. 2005). Similarly, while Suda et al. (2003) report correlations between genome size and altitude, average annual temperature, rainfall and/or humidity in a number of Macaronesian genera, these relationships disappear when the sampling is more representative (Suda et al. 2005). In addition, rigorous studies on variation in 144 Preslia 82: 127–148, 2010 genome size at the lowest divergence levels (within species or even within population) often fail to detect any link between genome size and environmental conditions (Šmarda et al. 2007). Significant correlations between genome size and growth form, habitat and altitude were also found in Lasiocephalus. On average, species from high-altitude harsh environments (superpáramo) have the smallest genomes while their counterparts from lower altitudes and more favourable conditions (Andean forest) have the largest genomes, which accords with the ’large genome constraint’ theory (Knight et al. 2005). Different growth forms are associated with different altitudes: narrow-leaved herbaceous plants on the superpáramo and lianas in forests. However, it should be noted that all these variables were strongly correlated with the major phylogenetic split in Lasiocephalus (i.e. the division into two lineages referred to as ‘páramo’ and ‘forest’ clades; Fig. 4, see also Fig. 5). Although we do not apriori reject that environmental factors may have somehow shaped the genome size, we are convinced that phylogenetic relationships were the key determinant of the observed divergence in genome size at the highest divergence levels. This assumption is strongly supported by the fact that the only superpáramo herbaceous plant, L. longipenicillatus with a large genome, was included in the ‘forest’ clade. These findings emphasize that for the correct interpretation of the environmental correlates of genome size it is crucially important to incorporate phylogenetic information into analyses (Grotkopp et al. 2004, Jakob et al. 2004, Chase et al. 2005, Beaulieu et al. 2007, Chrtek et al. 2009). Environmental factors seem to be more important in shaping the genome size of Lasiocephalus at lower divergence levels. While the three main sub-clades within the ‘forest’ clade do not significantly differ from each other in terms of genome size, the variation is structured according to the habitat preferences of the respective species (as also reflected by the negative correlation between genome size and altitude). Even within the sub-clades, high-altitude species have smaller DNA C-values than their low-altitude counterparts. Similar patterns are observed, for example, in the genus Hordeum where the environmental variable (length of vegetation period) is correlated with genome size only at the lowest taxonomic levels (Jakob et al. 2004). Conclusions This study is the first attempt to assess genome size variation in a radiating plant group in the tropics and interpret the results in the light of ecological conditions, life history traits and phylogeny. It indicates that the genome size in high-Andean Lasiocephalus has been shaped by different evolutionary processes, including polyploidy, hybridization and gradual changes at the molecular level. The variation in genome size at the higher divergence level primarily reflects phylogenetic effects, while ecological factors are more important in shaping genome size at the lower divergence levels. This highlights the need to always evaluate ecological and phenotypic correlates of genome size within a phylogenetic framework to avoid misinterpretation (e.g. overestimating the significance of environmental factors). We believe that this study is likely to stimulate further research into the promising but still largely unexplored field of tropical polyploid evolutionary biology and cytogeography. See http://www.preslia.cz for Electronic Appendix 1. Dušková et al.: Correlates of genome size in Lasiocephalus 145 Acknowledgements This research was supported by the Czech Science Foundation (project no. 206/07/0273), Grant Agency of the Academy of Sciences of the Czech Republic (project no. KJB601110710) and partly also by projects MSM 0021620828 (Ministry of Education, Youth and Sports of the Czech Republic), GAUK 261211 (Charles University Prague) and AV0Z60050516 (Academy of Sciences of the Czech Republic). We thank V. Jarolímová and E. Záveská for their assistance with the karyological and molecular analyses, respectively, P. Ubiergo, K. Romoleroux, D. Vázquez and P. Macek for their help with the field work, and J. Lepš and Z. Münzbergová for advice with statistical analyses. Souhrn Mezidruhové rozdíly ve velikosti genomu jsou určovány souhrou různých mikroevolučních procesů, mezi něž patří např. polyploidie, hybridizace, selekční tlak, ale také neutrální evoluce. Článek se zabývá variabilitou v obsahu jaderné DNA u andského rodu Lasiocephalus (Asteraceae, Senecioneae) a snaží se objasnit příčiny této variability s využitím spektra metodických přístupů (průtoková cytometrie, klasická karyologie, sekvenace jaderné DNA, ekologická pozorování). V souboru 189 jedinců představujících 20 druhů byl zjištěn 1,64-násobný rozdíl mezi průměrnými C-hodnotami jednotlivých druhů. Rostliny s největšími genomy s veškerou pravděpodobností představují triploidy (první údaje v rámci celého rodu). Zbývající vzorky vykazovaly 1,32-násobnou a prakticky kontinuální variabilitu v obsahu jaderné DNA. Nejmenší genomy patrně patří mezirodovým hybridům mezi zástupci rodu Lasiocephalus a blízce příbuzným a sympatrickým druhem Culcitium nivale. Velikost genomu diploidních rostlin souvisela s různými ekologickými charakteristikami a biologickými vlastnostmi druhů (nadmořská výška, charakter stanoviště, životní forma). Nicméně všechny tyto charakteristiky (stejně jako samotná velikost genomu) výrazně korelovaly s fylogenezí skupiny. V rámci celého rodu proto považujeme fylogenetickou příslušnost za hlavní příčinu rozdílů ve velikosti genomu. Naproti tomu na nižších fylogenetických úrovních (v rámci jednotlivých evolučních linií) hrají významnou roli ekologické podmínky stanovišť, které mohou působit jako selekční faktor měnící velikost genomu. Zjištěné výsledky tedy dokládají vzájemné spolupůsobení mnoha různých evolučních činitelů při utvářejí a změnách obsahu jaderné DNA a poukazují na nutnost interdisciplinárního přístupu při studiu variability a dynamiky ve velikosti genomu. References Beaulieu J. M., Leitch I. J. & Knight C. A. (2007): Genome size evolution in relation to leaf strategy and metabolic rates revisited. – Ann. Bot. 99: 495–505. Bell C. D. & Donoghue M. J. (2005): Phylogeny and biogeography of Valerianaceae (Dipsacales) with special reference to the South American valerians. – Org. Divers. Evol. 5: 147–159. Bennett M. D. & Leitch I. J. (2005): Genome size evolution in plants. – In: Gregory T. R. (ed.), The evolution of the genome, p. 89–162, Elsevier, San Diego. Bennetzen J. L. (2005): Transposable elements, gene creation and genome rearrangement in flowering plants. – Curr. Opin. Genet. Dev. 15: 621–627. Bennetzen J. L., Ma J. & Devos K. M. (2005): Mechanisms of recent genome size variation in flowering plants. – Ann. Bot. 95: 127–132. Bureš P., Wang Y-F., Horová L. & Suda J. (2004): Genome size variation in Central European species of Cirsium (Compositae) and their natural hybrids. – Ann. Bot. 94: 353–363. Cavalier-Smith T. (2005): Economy, speed and size matter: evolutionary forces driving nuclear genome miniaturization and expansion. – Ann. Bot. 95: 147–175. Chase M. W., Hanson L., Albert V. A., Whitten W. M. & Williams N. H. (2005): Life history evolution and genome size in subtribe Oncidiinae (Orchidaceae). – Ann. Bot. 95: 191–199. Chrtek J. jun., Zahradníček J., Krak K. & Fehrer J. (2009): Genome size in Hieracium subgenus Hieracium (Asteraceae) is strongly correlated with major phylogenetic groups. – Ann. Bot. 104: 161–178. Cuatrecasas J. (1978): Studies in Neotropical Senecioneae, Compositae I. Reinstatement of genus Lasiocephalus. – Phytologia 40: 307–312. Cuatrecasas J. (1986): Speciation and radiation of the Espeletiinae in the Andes. – In: Vuilleumier F. & Monasterio M. (eds), High altitude tropical biogeography, p. 267–303, Oxford University Press, New York. Doležel J., Greilhuber J., Lucretti S., Lysák M. A., Nardi L. & Obermayer R. (1998): Plant genome size estimation by flow cytometry: inter-laboratory comparison. – Ann. Bot. 82 (Suppl A): 17-26. 146 Preslia 82: 127–148, 2010 Doležel J., Greilhuber J. & Suda J. (2007): Estimation of nuclear DNA content in plants using flow cytometry. – Nature Protoc. 2: 2233–2244. Dyer A. F. (1963): The use of lacto-propionic orcein in rapid squash methods for chromosome preparations. – Stain Technol. 38: 85–90. Dzialuk A., Chybicki I., Welc M., Śliwińska E. & Burczyk J. (2007): Presence of triploids among oak species. – Ann. Bot. 99: 959–964. Ekrt L., Trávníček P., Jarolímová V., Vít P. & Urfus T. (2009): Genome size and morphology of the Dryopteris affinis group in Central Europe. – Preslia 81: 261–280. Garcia S., Canela M. A., Garnatje T., McArthur E. D., Pellicer J., Sanderson S. C. & Vallès J. (2008): Evolutionary and ecological implications of genome size in the North American endemic sagebrushes and allies (Artemisia, Asteraceae). – Biol. J. Linn. Soc. 94: 631–649. Greilhuber J. (2005): Intraspecific variation in genome size in angiosperms: identifying its existence. – Ann. Bot. 95: 91–98. Greilhuber J., Doležel J., Lysák M. A. & Bennett M. D. (2005): The origin, evolution and proposed stabilization of the terms ‘genome size’ and ‘C-value’ to describe nuclear DNA contents. – Ann. Bot. 95: 255–260. Grotkopp E., Rejmánek M., Sanderson M. J. & Rost T. L. (2004): Evolution of genome size in pines (Pinus) and its life-history correlates: a supertree analyses. – Evolution 58: 1705–1729. Hall T. A. (1999): BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. – Nucl. Acids Symp. 41: 95–98. Hodálová I., Grulich V., Horová L., Valachovič M. & Marhold K. (2007): Occurence of tetraploid and octoploid cytotypes in Senecio jacobaea ssp. jacobaea (Asteraceae) in Pannonia and the Carpathians. – Bot. J. Linn. Soc. 153: 231–242. Huelsenbeck J. P. & Ronquist F. (2001): MRBAYES: Bayesian inference of phylogeny. – Bioinformatics 17: 754–755. Hughes C. & Eastwood F. (2006): Island radiation on a continental scale: exceptional rates of plant diversification after uplift of the Andes. – Proc. Natl. Acad. Sci. USA 103: 10334–10339. Jakob A., Meister A. & Blattner F. R. (2004): The considerable genome size variation of Hordeum species (Poaceae) is linked to phylogeny life form, ecology, and speciation rates. – Mol. Biol. Evol. 21: 860–869. Kanagawa T. (2003): Bias and artifacts in multitemplate polymerase chain reactions (PCR). – J. Biosc. Bioeng. 96: 317–323. Knight C. A. & Ackerly D. D. (2002): Genome size variation across environmental gradients in the California flora. – Ecol. Lett. 5: 66–76. Knight C. A. & Beaulieu J. (2008): Genome size scaling in phenotype space. – Ann. Bot. 101: 759–766. Knight C. A., Molinari N. & Petrov D. (2005). The large genome constraint hypothesis: evolution, ecology, and phenotype. – Ann. Bot. 5: 177–190. Kovarik A., Matyasek R., Lim K. Y., Skalická K., Koukalová B., Knapp S., Chase M. & Leitch A. R. (2004): Concerted evolution of 18-5.8-26S rDNA repeats in Nicotiana allotetraploids. – Biol. J. Linn. Soc. 82: 615–625. Kovarik A., Pires J. C., Leitch A. R., Lim K. Y., Sherwood A. M., Matyasek R., Rocca J., Soltis D. E. & Soltis P. S. (2005): Rapid concerted evolution of nuclear ribosomal DNA in two Tragopogon allopolyploids of recent and recurrent origin. – Genetics 169: 931–944. Kron P., Suda J. & Husband B. C. (2007): Applications of flow cytometry to population biology. – Annu. Rev. Ecol. Evol. Syst. 38: 847–876. Leitch I. J., Beaulieu J. M., Cheung K., Hanson L., Lysak M. A. & Fay M. F. (2007): Punctuated genome size evolution in Liliaceae. – J. Evol. Biol. 20: 2296–2308. Leitch I. J. & Bennett M. D. (2007) Genome size and its uses: the impact of flow cytometry. – In: Doležel J., Greilhuber J. & Suda J. (eds), Flow cytometry with plant cells, p. 153–176, Wiley-VCH, Weinheim. Leitch I. J., Soltis D. E., Soltis P. S & Bennett M. D. (2005): Evolution of DNA amounts across land plants (Embryophyta). – Ann. Bot. 95: 207–217. Leong-Škorničková J., Šída O., Jarolímová V., Sabu M., Fér T., Trávníček P. & Suda J. (2007): Chromosome numbers and genome size variation in Indian species of Curcuma L. (Zingiberaceae). – Ann. Bot. 100: 500–526. Lindstrom M. J. & Bates D. M. (1990): Nonlinear mixed effects models for repeated measures data. – Biometrics 46: 673–687. Loureiro J., Trávníček P., Rauchová J., Urfus T., Vít P., Štech M., Castro S. & Suda J. (2010): The use of flow cytometry in the biosystematics, ecology and population biology of homoploid plants. – Preslia 82: 3–21. Luteyn J. L. (1999): Páramos: a checklist of plant diversity, geographical distribution, and botanical literature. – Mem. New York Bot. Gard. 84: 1–278. Dušková et al.: Correlates of genome size in Lasiocephalus 147 Lysak M. A., Koch M., Beaulieu J. M., Meister A. & Leitch I. J. (2009): The dynamic ups and downs of genome size evolution in Brassicaceae. – Mol. Biol. Evol. 26: 85–98. Mahelka V., Suda J., Jarolímová V., Trávníček P. & Krahulec F. (2005): Genome size discriminates between closely related taxa Elytrigia repens and E. intermedia (Poaceae: Triticeae) and their hybrid. – Folia Geobot. 40: 367–384. Marhold K. (1994): Chromosome numbers of the genus Cardamine L. (Cruciferae) in the Carpathians and in Pannonia. – Phyton (Horn) 34: 19–34. Otto F. (1990): DAPI staining of fixed cells for high-resolution flow cytometry of nuclear DNA. – In: Crissman H. A. & Darzynkiewicz Z. (eds), Methods in cell biology 33, p. 105–110, Academic Press, New York. Pelser P. B., Nordensatm B., Kadereit J. W. & Watson L. E. (2007): An ITS phylogeny of tribe Senecioneae (Asteraceae) and a new delimitation of Senecio L. – Taxon 56: 1077–1104. Piegu B., Guyot R., Picault N., Roulin A., Saniyal A., Kim H., Collura K., Brar D. S., Jackson S., Wing R. A. & Panaud O. (2006): Doubling genome size without polyploidization: dynamics of retrotransposition-driven genomic expansions in Oryza australiensis, a wild relative of rice. – Genome Res. 16: 1262–1269. Popp M., Erixon P., Eggens F. & Oxelman B. (2005): Origin and evolution of a circumpolar polyploid species complex in Silene (Caryophyllaceae) inferred from low copy nuclear RNA polymerase introns, rDNA, and chloroplast DNA. – Syst. Bot. 30: 302–313. Popp M., Gizaw A., Nemomissa S., Suda J. & Brochmann C. (2008): Colonization and diversification in the African “sky islands” by Eurasian Lychnis L. (Caryophyllaceae). – J. Biogeogr. 35: 1016–1029. Posada D. & Crandall K. A. (1998): MODELTEST: testing the model of DNA substitution. – Bioinformatics 14: 817–818. Powel A. M. & Cuatrecasas J. (1970): Chromosome numbers in Compositae: Colombian and Venezuelan species. – Ann. Mo. Bot. Garden 57: 374–379. R Development Core Team (2009): R: a language and environment for statistical computing. – R Foundation for Statistical Computing, Vienna, URL: [http://www.R-project.org]. Robinson H., Carr G. D., King R. M. & Powell A. M. (1997): Chromosome numbers in Compositae, XVII. Senecioneae III. – Ann. Mo. Bot. Gard. 84: 893–906. Ronquist F. &. Huelsenbeck J. P. (2003): MRBAYES 3: Bayesian phylogenetic inference under mixed models. – Bioinformatics 19: 1572–1574. Sarmiento G. (1986): Ecological features of climate in high tropical mountains. – In: Vuilleumier F. & Monasterio M. (eds), High altitude tropical biogeography, p. 11–45, Oxford University Press, New York. Šiško M., Ivančič A. & Bohanec B. (2003): Genome size analysis in the genus Cucurbita and its use for determination of interspecific hybrids obtained using the embryo-rescue technique. – Plant Sci. 165: 663–669. Slovák M., Vít P., Urfus T. & Suda J. (2009): Complex pattern of genome size variation in a polymorphic member of the Asteraceae. – J. Biogeogr. 36: 372–384. Šmarda P., Bureš P. & Horová L. (2007): Random distribution pattern and non-adaptivity of genome size in a highly variable population of Festuca pallens. – Ann. Bot. 100: 141–150. Soltis D. E., Albert V. A., Leebens-Mack J., Bell C. D., Paterson A. H., Zheng C., Sankoff D., dePamphilis C. W., Wall P. K. & Soltis P. S. (2009): Polyploidy and angiosperm diversification. – Am. J. Bot. 96: 336–348. Soltis D. E., Mavrodiev E. V., Doyle J. J., Rauscher J. & Soltis P. S. (2008): ITS and ETS sequence data and phylogeny reconstruction in allopolyploids and hybrids. – Syst. Bot. 33: 7–20. StatSoft, Inc. (2007): STATISTICA (data analysis software system), version 8.0. – URL: [http://www.statsoft.com]. Struwe L., Haag S., Heiberg E. & Grant J. R. (2009): Andean speciation and vicariance in neotropical Macrocarpaea (Gentianaceae-Helieae). – Ann. Mo. Bot. Gard. 96: 450–469. Suda J., Krahulcová A., Trávníček P. & Krahulec F. (2006): Ploidy level versus DNA ploidy level: an appeal for consistent terminology. – Taxon 55: 447–450. Suda J., Kyncl T. & Jarolímová V. (2005): Nuclear DNA amounts in Macaronesian angiosperms: forty percent of Canarian endemic flora completed. – Pl. Syst. Evol. 252: 215–238. Suda J., Kyncl T. & Wildová R. (2003): Nuclear DNA amount in Macaronesian angiosperms. – Ann. Bot. 92: 153–164. Suda J. & Trávníček P. (2006): Reliable DNA ploidy determination in dehydrated tissues of vascular plants by DAPI flow cytometry: new prospects for plant research. – Cytometry Part A 69A: 273–280. Suda J., Weiss-Schneeweiss H., Tribsch A., Schneeweiss G. M., Trávníček P. & Schönswetter P. (2007): Complex distribution patterns of di-, tetra- and hexaploid cytotypes in the European high mountain plant Senecio carniolicus (Asteraceae). – Am. J. Bot. 94: 1391–1401. 148 Preslia 82: 127–148, 2010 Vuilleumier F. & Monasterio M. (eds) (1986): High altitude tropical biogeography. – Oxford University Press, New York. Wagner A., Blackstone N., Cartwright P., Dick M., Misof B., Snow P., Wagner G. P., Bartels J., Murtha M. & Pendleton J. (1994): Surveys of gene families using polymerase chain reaction: PCR selection and PCR drift. – Syst. Biol. 43: 250–261. Weiss-Schneeweiss H., Greilhuber J. & Schneeweiss G. M. (2005): Genome size evolution in holoparasitic Orobanche (Orobanchaceae) and related genera. – Am. J. Bot. 93: 148–156. White T. J., Bruns T., Lee S. & Taylor J. (1990): Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. – In: Innis M., Gelfand D., Sninsky J. & White T. (eds), PCR protocols: a guide to methods and applications, p. 315–322, Academic Press, San Diego. Wilkström N., Kenrick P. & Chase M. (1999): Epiphytism and terrestrialization in tropical Huperzia (Lycopodiaceae). – Pl. Syst. Evol. 218: 221–243. Yahata M., Kunitake H., Yasuda K., Yamashita K., Komatsu H. & Matsumoto R (2006): Production of sexual hybrid progenies for clarifying the phylogenic relationship between Citrus and Citropsis species. – Am. Soc. Hort. Sci. 131: 764–769. Young K. R., Ulloa C. U., Luteyn L. J. & Knapp S. (2002): Plant evolution and endemism in Andean South America: an introduction. – Bot. Rev. 68: 4–21. Received 6 January 2010 Revision received 27 January 2010 Accepted 2 February 2010