Within this study, different accessions (populations) belonging to the genus Hypericum and covering four species namely Hypericum perforatum (Hp), Hypericum maculatum (Hm), H. umbellatum (Hu) and H. hircinum (Hh) from different botanical gardens of Europe were analyzed. In most of the accessions, seeds were from plants cultivated in botanical gardens (Hp1-8, Hp11-13, Hp15, Hm23, Hh24) but some of them were collected from plants from natural populations (Hp9-10, Hp14, Hp16-17, Hu21, Hm22) according to Tab. 1. Seeds were germinated in soil and plants grown in the greenhouse (at least 30 seeds/accession).

Estimation of DNA content and ploidy by FCM was performed using fresh leaf material from young plants. Small amounts of leaf tissue of a sample and the internal standard were chopped together in 0.5 mL of Galbraith’s buffer (Galbraith et al. 1983), supplemented with 0.5 mL Partec CyStain propidium iodide solution containing DNase-free RNase following the manufacturer’s instructions (Partec) and filtered through a cell-strainer cap (BD FalconTM) with 35 µm pore size. Analyses were performed with a flow cytometer BD FACS Calibur (USA). For each sample, no fewer than 5000 particles were registered by 488 nm laser beam. For estimation of the nuclear DNA content, Raphanus sativus L. was used as an internal standard (2C = 1.11 pg; Doležel et al. 1992). Analyses were carried out for at least five randomly selected individuals per accession and three repetitions for each individual. For accessions with differing cytotypes up to 30 individual plants were tested (Tab. 2). Data evaluation was accomplished with BD software CellQuestTMPro. Nuclear DNA content was calculated using the linear relationship between the ratio of the 2C peak positions of the target species/internal standard on the area histogram of fluorescence intensities. The statistics were carried out using the Real Statistics Resource Pack Version 4.9 for Microsoft Excel.

Root tips were harvested from greenhouse plants and pretreated with 8-hydroxyquinoline for 2.5 hours at room temperature and fixed overnight in ethanol: glacial acetic acid (3:1 v/v) at 4 °C. The fixed root tips were transferred to 70% ethanol and stored at 4 °C. For further investigation fixed root tips were washed in distilled water for 10 min and then incubated at 37 °C for 45 min in an enzyme mix consisting of 4% celullase, 1% pectolyase and 45% acetic acid, pH 4.0. After being washed with distilled water, the tips were squashed on the slide in 45% acetic acid. If under a bright microscope in phase contrast metaphase chromosomes were observed, the slide was frozen for at least 15 min at -80 °C and the cover slip was blown off. After an air drying for 10 min or longer 15 µl DAPI VECTASHIELD® antifade mounting medium with DAPI (4’, 6-diamidino-2-phenylindol, Vector Laboratories) was added. The chromosomes were detected and photographed in fluorescent light (microscope Axioimager Z1 with CCD-camera Axiocam, Zeiss). The image analysis was carried out with the software program Isis (MetaSystems, Germany).

Marker analysis was conducted with 90 individuals (5 individuals/available accession), a representative sub-set of accessions and individuals screened by flow cytometry. Genomic DNA was isolated from young leaves of plants grown under standard greenhouse conditions, by using the CTAB method described by Doyle and Doyle (1987). DNA concentration was estimated using a UV-Vis spectral photometer Nanodrop 8000. ISSR (Inter Simple Sequence Repeats) amplification was performed with 11 primers (Rostami-Ahmadvandi et al. 2013). For SRAP (Sequence Related Amplified Polymorphism) analysis, sixteen primer combinations (Li and Quiros 2001) were used. Primer sequences are given in Tab. 3. PCR amplification was performed in a 0.2 mL tube containing 12.5 μL 2x DreamTaq Green PCR master mix (Thermo Fisher Scientific, USA), 10.25 μL nuclease-free water (Lonza, Switzerland), 25 pmol of each primer (Eurogentec, Belgium) and 5 ng of genomic DNA in a final volume of 25 μL. For ISSR analysis the GeneAmp PCR system 9700 (Applied Biosystems, Forster City, USA) was programmed as follows: 94 °C for 4 min, 35 cycles of 94 °C for 30 sec, 46 °C for 30 sec and 72 °C for 55 sec followed by a final elongation step at 72 °C for 5 min. For SRAP analysis, the following program was used: 94 °C for 4 min, 5 cycles of 94 °C for 30 sec, 35 °C for 30 sec, and 72 °C for 55 sec followed by 30 cycles of 94 °C for 30 sec, 50 °C for 30 sec, 72 °C for 55 sec, and a final elongation at 72 °C for 5 min. Amplification products were separated in 1.5% w/v agarose (Cleaver Scientific, United Kingdom) gel in 1×TBE buffer (Lonza, Switzerland) and stained with 0.5 μg/mL ethidium bromide (Thermo Fisher Scientific, USA).

ISSR and SRAP patterns were assessed as dominant markers. Band patterns for both marker systems were recorded in 1/0 matrices to determine the level of genetic similarity between the different accessions on the basis of Jaccard's coefficient (Jaccard 1908). The resulting matrix of similarity was analyzed by the unweighted pairgroup method with arithmetic mean (UPGMA) and the dendrogram was obtained using MultiVariate Statistical Package 3.21 (Kovach 2007). Shannon’s information index (I) (Shannon and Weaver 1949) and expected heterozygosity (He) were calculated, using GenAlEx software version 6.5 (Peakall and Smouse 2012). The polymorphism information content (PIC) value of each individual locus was calculated according to Sehgal et al. (2009) as:

PIC j = 1-Ʃpi2

Where i is the i^th allele of the j^th marker, n is the number of alleles at the j^th marker and p is the allele frequency.

Resolving power (Rp) for the individual marker system was bases on the distribution of detected bands within the sampled clones and was calculated based on the formula described by Prevost and Wilkinson (1999):

Rp = ΣIb

Where Ib (informativeness) is 1- [2 x |0.5-p|] and p is the ratio of present bands among the analyzed accessions.

The flow cytometric measurements consistently showed a high reproducibility in repeated measurements of the same plant despite the presence of numerous secondary metabolites, which often influence the peak performance. The CV (coefficient of variation) of histogram peaks was below 5.0 in each case. Raphanus sativus served as an internal standard for genome size estimation. Both 2C peaks from internal standard and Hypericum sample were well separated from one another (Fig. 1). Fig. 1

Flow cytometric histograms (1: 2C Peak of internal standard Raphanus sativus, 1.11 pg). a – Hypericum perforatum Hp7-1: 2: 2C = 4Cx peak, 1.55 pg, b – Hp7-2: 2: 2C = 6Cx peak, 2.34 pg, c – H. maculatum Hm23: 2: 2C = 2Cx peak, 0.76 pg..

The genome size of 17 Hp accessions and the cultivar Topaz of the tetraploid genotypes amounted to between 1.56 pg and 1.62 pg (Tab. 2, Fig. 1a) with an average genome size of 2C = 4Cx = 1.58 pg. The Hypericum accessions were obtained from various European botanical gardens and from Romanian regions (Tab. 1). The accession Hp8 from the Salzburg botanical garden, Austria, revealed for five tested plants a hexaploid cytotype with a DNA content of 6Cx = 2.41 pg (Tab. 2). Besides tetraploid plants we found in Hp6, Hp7, Hp9 and Hp10 at least one plant which showed a genome size corresponding to the hexaploid chromosome level (Fig. 1a, b), taking it into account that the 1Cx DNA content for all Hp accessions amounted to 0.40 pg. However, one exception was noticed. In the Hp3 from the botanical garden in Regensburg, Germany, one plant with a genome size of 1.99 pg was observed. Only speculation could be offered about the nature of this cytotype.

For both Romanian H. maculatum accessions as well as for H. umbellatum a small genome size with 0.76 pg was estimated (Fig. 1c). The genome size of H. hircinum was 1.00 pg (Tab. 2). Thus, the 1Cx value of H. hircinum differed considerably from the others and was only 0.25 pg.

For correct assignment of genome size to ploidy, the numbers of metaphase chromosomes were counted in stained root tips. The chromosomes of Hypericum are very small and morphologically similar, making chromosome counting difficult. In addition, the plant tissue digestion posed problems with regard to achieving well-spread chromosome plates. The unsatisfactory quality in connection with low mitotic index resulted in a fluctuating number of chromosomes being counted. In the accessions Hp3, Hp4, Hp7, and Hp9 we noticed 30 – 34 chromosomes, mainly 32 (Fig. 2a, Tab. 4). In H. umbellatum and H. maculatum 16 chromosomes were detected (Fig. 2b, Tab. 4). In squash preparations of H. hircinum 30 – 40 chromosomes were found, in the best preparations 32 chromosomes (Fig. 2c, Tab. 4). Especially in H. hircinum, chromosomes often so stuck together that even at different focal levels an unambiguous assessment of the chromosomes was problematic.

In all, 699 amplified DNA bands were scored using 11 ISSR primers (298 markers) and 16 SRAP primers (401 markers) for the 17 H. perforatum accessions and one H. maculatum accession (Hm22).

The overall number of detected individual bands per primer ranged from 9 (SRP17) to 53 (SRP30), with an average of 27.1/25.1 (ISSR/SRAP) (Tab. 5). The combined analysis of ISSR and SRAP markers revealed a total of 661 (94.6%) polymorphic bands. The marker performance was estimated by two parameters: PIC value and resolving power (RP). ISSR primers and SRAP primer combinations showed the same mean PIC value (0.38). The highest PIC value was determined for primers UBC808, UBC809, UBC857, UBC873 and SRP26 (0.49). The resolving power (RP) of primers tested varied between 0.44 (UBC818/ SRP29) and 1.69 (UBC112). The mean values for Shannon’s information index (I) and expected heterozygosity (He) were 0.39/0.25 for ISSR primers, and 0.42/0.28 for SRAP markers (Tab. 5).

Combined ISSR-SRAP UPGMA revealed two major clusters (Fig. 3): Cluster I comprising all H. perforatum accessions and cluster II represented by H. maculatum (Hm22) as a clearly separated outgroup. Within cluster I we observed 6 subclusters: Hp1 and Hp2 from Germany; Hp7 (Switzerland) and Hp9 (Austria); Hp13 (Poland) and Hp14 (Norway); Hp5 and Hp6 from Germany. French Hp10 and Hp11 and Italian Hp16 and Hp15 grouped with Hp17 from Estonia. Four H. perforatum accessions, Hp3 and Hp4 from Germany as well as Hp8 from Austria and Hp12 from France were separated in individual branches from the other genotypes. H. maculatum is clearly distinguished from all H. perforatum accessions.

The key to a successful breeding program is a better understanding of the extent and nature of genetic diversity present in wild, conserved and/or actively utilized germplasm of various species. There is a broad interest in gaining a better understanding of diversity within Hypericum species, especially in H. perforatum, due to its pharmaceutical importance but also for its remarkable evolutionary and adaptive capacities. Therefore, we employed FCM for analysis of ploidy and genome size as well as two types of molecular markers (ISSR and SRAP) to reveal the genetic diversity and relationships among several Hypericum accessions.

FCM is a powerful tool for genome size estimation and ploidy determination. With all the advantages and possibilities of FCM, however, it must be taken into account that the results are always expressed in relation to a known standard. For genome size estimation (pg value) a plant with an already defined DNA content as internal standard should be measured together with the sample plant in the staining solution. For ploidy estimation, plants with a cytologically verified chromosome number within the same species serve as standard, providing that different ploidy degrees are in linear dependency.

For 17 H. perforatum accessions from different botanical gardens across Europe and the cultivar Topaz, the genome size was on average 1.58 pg ranging from 1.56 to 1.62 pg in the different accessions. Chromosome counting in root tips of different Hypericum plants with the corresponding genome size of 1.58 pg revealed 2n = 4x = 32 chromosomes. These results are in agreement with Temsch et al. (2010) (2C = 1.59 pg), and Alan et al. (2015) (2C = 1.50 pg). The slight differences result from employment of different standards, extraction / staining buffers and flow cytometers (Doležel and Bartoš 2005) but the use of various accessions and the physiological state of the plant material also have an influence on the measurement.

Matzk et al. (2003) developed the flow cytometric seed screen (FCSS), boosted the investigation of apomixes of Hypericum and broaden the knowledge about apomictic pathways in general. However, reports about the germination capacity of the differently developed seeds and their contribution to plant populations are very limited. Thought, such information would have an impact on germplasm management and preservation strategies of rare natural populations. Mixed cytotypes in populations could provide novel traits for crop development and therefore the seed samples in gene banks should be adapted to that situation. Hypericum perforatum is predominately tetraploid (Matzk et al. 2003, Galla et al. 2011). Barcaccia et al. (2007) describe wild populations of H. perforatum in more detail as populations with diploid and polyploid (mainly tetraploid) plants. Savaş Tuna et al. (2017) analyzed three seedlings from 39 Hp accessions each from different regions of Turkey and revealed a nuclear DNA content between 0.8 – 2.57 pg. Of the 39 accessions, one was diploid, 5 hexaploid and 33 tetraploid but no ploidy variation was noticed inside the accessions. Perhaps the test of only three seedlings per accession is not sufficient to provide reliable findings. Here we present flow cytometric DNA size determination of plants of 17 belonging to Hp accessions and the cultivar Topaz. Tetraploidy was revealed with only one exception: all plants from Hp8 from Salzburg have shown a DNA size of 2.41 pg, corresponding to the hexaploid level. This uniformity could be explained by the origin of seeds in plants cultivated in botanical garden, most probably the collection being not very diverse. In four accessions (Hp6-7 and Hp9-10), there were mixed cytotypes (4x + 6x). This is an interesting situation, as Hp 6-7 seeds were collected from plants cultivated in botanical gardens and Hp 9-10 belong to plants from natural populations. Similar results were found in three H. perforatum accessions (Qu et al. 2010). Despite the diverse embryo and endosperm ratios observed in seeds, Qu et al. (2010) found only tetraploids and hexaploids in seedling populations with tetraploids constituting 87 - 97% but a complete hexaploid population was not detected. Among the accessions presented here the accession Hp8 shows only hexaploid plants indicating that such accessions at least with a high proportion of hexaploid plants could exist and depend on the seed source. Moreover, in accession Hp3 a single plant was found with 1.99 pg. Since it does not fit the 1C content of 0.40 pg, one can assume that it is an aneuploid plant reflecting additionally the high plasticity of the H. perforatum genome, even if Hp3 comes from plants that are cultivated in a botanical garden.

To our knowledge we report for the first time the genome size of H. maculatum (two accessions, one from a natural habitat and one cultivated in a botanical garden), H. umbellatum (from natural habitat) (0.76 pg each) and H. hircinum (from botanical garden) (1.00 pg). The first three accessions are native to Romania. Hence, the DNA size of H. hircinum of 1.0 pg is between the estimated DNA size for the diploid species H. maculatum und H. umbellatum (0.76 pg) and the tetraploid H. perforatum (1.58 pg). The genome size for H. maculatum and H. umbellatum corresponds to the chromosome number 2n = 2x = 16. For H. hircinum Loon and Jong (1978) published a chromosome number 2n = 40. They explained the high chromosome number of 40 as a possible pentaploid cytotype. Castro and Rosselló (2006) found 2n = 32 in H. hircinum subsp. cambessedesii, an endemic plant from the Balearic Islands. The H. hircinium accession investigated in the present paper was provided by the Alexandru Borza Botanical Garden from Cluj-Napoca, Romania. In this accession, we observed 30 - 40 chromosomes; in the best metaphase plates 32, which is in agreement with the findings of Castro and Rosselló (2006). The result was surprising because H. hircinum has a much lower DNA size than H. perforatum (1.58 pg) although it also has 32 chromosomes. This fact underlines the high variability of the genus Hypericum.

Characterization of Hypericum species by different molecular marker types has been performed over the years (Corral et al. 2011, He and Wang 2013). ISSR and SRAP markers were chosen due their advantages: cost efficiency, informativeness, versatility and reproducibility. We employed a set of 11 ISSR primers and 16 SRAP primer combinations to assess the genetic diversity among one Hm22 and 17 Hp accessions. ISSR and SRAP markers have proved to be highly polymorphic. The average number of polymorphic bands per primer was the same (24) for both types of markers. Several studies report that when several marker types are used, they can be complementary tools for genetic diversity analysis, because they can be used to amplify different regions of the genome (Chen et al. 2013). Our study revealed a significantly higher rate of polymorphism in the analyzed Hypericum germplasm for both SRAP (97.8%) and ISSR (90.2%) markers than previously reported (He and Wang 2013). This might be explained by a larger sampling area with very different Hp accessions from botanical gardens in Europe and native wild accessions from Romania. The two marker systems used in our study revealed close degrees of resolution (Tab. 5).

Species cluster analyses based on combined ISSR-SRAP data (Fig. 3) indicate that H. maculatum is closely related to some H. perforatum accessions. These taxa, belonging to section Hypericum “core Hypericum” (Nürk 2011), were proven to be in close contact and apparently hybridize frequently, which might explain the sympatric occurrence of morphologically similar taxa (Robson 2003, Koch et al. 2013). Thus, according to Brutovská et al. (2000), H. perforatum probably originates from autopolyploidization of an ancestor closely related to diploid H. maculatum, while Robson (2003) regards H. perforatum as an allopolyploid, derived from a cross between H. maculatum subsp. immaculatum and H. attenuatum. This close relationship between the two species was also supported by later studies based on different marker types, such as RFLP and cpDNA markers, as well as phylogenetic studies using nuclear internal transcribed spacer (ITS) sequences (Pilepić et al. 2011). However, it was recently implied that H. perforatum is not of hybrid origin (Koch et al. 2013). The authors suggest that H. perforatum has a single evolutionary origin arising from independent and recurrent polyploidization of two different ancestral gene pools along with occurrence of substantial gene flow within and between H. perforatum and H. maculatum.

Regarding the cluster analysis of H. perforatum accessions, we have noticed a partially regional-based relationship. Some accessions from the same regions shared similar genotypes and ploidy levels (Hp1 and Hp2 from Germany; HP15 and HP16 from Italy), and the same was noticed for accessions from different geographical locations (HP9 from Austria and HP7 from Switzerland; HP13 from Poland and HP14 from Norway) (Tab. 1, 2; Fig. 3). Moreover, mixed ploidy accessions (4x and 6x) (Hp5 and Hp6; Hp9 and Hp7; Hp10 and Hp11) shared similar genotypes, while the complete hexaploid accession HP8 from Austria is clearly of different genetic origin from that of the tetraploid Hp9 accession from the same country (Tab. 1, 2; Fig. 3). This endorses the high plasticity in ploidy and reproductive system of H. perforatum, regardless of geographic origin (Koch et al. 2013). Differences among H. perforatum genotypes were reported in different H. perforatum wild populations and landraces as confirmed by molecular, morphometric and cytogentic analyses (He and Wang 2013, Morshedloo et al. 2014). The high genetic diversity exhibited by our analysis might be explained by the diverse mating systems of H. perforatum from sexual cross to apomixis.