Genome Size and Chromosome Number Evolution in Korean Iris L. Species (Iridaceae Juss.)

Chromosome numbers, karyotypes, and genome sizes of 14 Iris L. (Iridaceae Juss.) species in Korea and their closely related taxon, Sisyrinchium rosulatum, are presented and analyzed in a phylogenetic framework. To date, understanding the chromosomal evolution of Korean irises has been hampered by their high chromosome numbers. Here, we report analyses of chromosome numbers and karyotypes obtained via classic Feulgen staining and genome sizes measured using flow cytometry in Korean irises. More than a two-fold variation in chromosome numbers (2n = 22 to 2n = 50) and over a three-fold genome size variation (2.39 pg to 7.86 pg/1 C) suggest the putative polyploid and/or dysploid origin of some taxa. Our study demonstrates that the patterns of genome size variation and chromosome number changes in Korean irises do not correlate with the phylogenetic relationships and could have been affected by different evolutionary processes involving polyploidy or dysploidy. This study presents the first comprehensive chromosomal and genome size data for Korean Iris species. Further studies involving molecular cytogenetic and phylogenomic analyses are needed to interpret the mechanisms involved in the origin of chromosomal variation in the Iris.


Introduction
Chromosomal changes play a major role in plant evolution and are thus important in diversification and speciation in angiosperms [1][2][3]. Variation of chromosome numbers, ploidy levels, and genome sizes have been frequently analyzed for a better understanding of evolutionary patterns and species relationships in plants [4][5][6][7]. The genome size values, in combination with data on chromosome numbers, allow ploidy levels to be inferred and can also provide insights into evolutionary relationships between closely related taxa in wild plant groups [7][8][9][10][11][12]. While the monocots are known to have the widest range of genome sizes (0.2-152.2 pg/1 C) among angiosperms [13][14][15], the known genome sizes in the family Iridaceae range from 0.5 pg/1 C in Hesperantha and Sisyrinchium species to 31.4 pg/1 C in Moraea taxa [15,16].
Among monocots, Iris is an excellent system in which to study the evolutionary patterns of chromosome number and genome size evolution because this group has an exceptionally high diversity of chromosome numbers (n = 7, 8,9,10,11,12,13,15,16,17,18,19,20,21,22,24,25,27,36,54) [17] and genome sizes (1.05-28.2 pg/1 C) [15]. The genus Iris L. is perennial and comprises approximately 300 species worldwide, with the greatest number of endemic species occurring in the to the endangered or vulnerable species, we refrain from providing details of the locality information (GPS coordinates) in this study.

Chromosome Numbers and Karyotype Analysis
Fresh root tips were pretreated in 0.05% colchicine (Sigma-Aldrich Co., St. Louis, USA) solution at room temperature in the dark for 4.5 h. The sampled roots were fixed in ethanol:acetic acid (3:1) for two hours at room temperature and then stored at −20 • C. For karyotype and chromosome number analyses, meristematic root cells were used (Table 1). Prior to staining, the fixed roots were washed with tap water, and hydrolyzed in 5 N HCL at room temperature for 30 min. Subsequently, to perform Feulgen staining, the root tips were treated in Schiff s reagent (Merck KGaA, Darmstadt, Germany) in darkness for 1 h. Microscopic slides were prepared by squashing the meristematic tissue in 60% acetic acid. A minimum of three well spread chromosome plates were selected for the analyses [32]. The chromosomes were cut using Corel Photo-Paint 12.0 (Corel Company, Ottawa, Canada). For chromosomal measurements, MicroMeasure ver.3.3 program (https://micromeasure.software.informer.com/3.3/) [33] was used, and a minimum of three chromosomal spreads were measured with a comparable medium degree of chromosomal condensation (Table 2).

Genome Size Measurement by Flow Cytometry (FCM)
Genome sizes of 76 individuals of 14 Iris species and Sisyrinchium rosulatum were measured using flow cytometry with Solanum pseudocapsicum (1 C = 1.29 pg) [34] or Pisum sativum "Kleine Rheinländerin" (1 C = 4.42 pg) [35] as an internal standard. Approximately 20-30 mg of fresh leaves of each sample were chopped with an internal reference standard using Otto s buffer I [36]. Subsequently, each sample was filtered using a nylon mesh, and incubated with RNase A in a water-bath at 37 • C for 30 min. The nuclei staining was conducted using propidium iodide, which contains Otto´s buffer II, and the sample was stored at 4 • C. A total of three measurements of DNA content were estimated for each sample using a flow cytometer (Partec, Münster, Germany). 1 C-values were calculated using estimation of linear fluorescence intensity of nuclei of the examined taxon and internal standard. The coefficient of variation (CV) of all estimations were mostly below 3% on average, and never exceeded 7% [7].

Chromosome Numbers and Karyotypes of Iris L.
Chromosome numbers for 74 individuals of 14 currently recognized the Korean Iris species and two individuals of Sisyrinchium rosulatum are provided in Table 1 and Figures 1 and 2. The karyotype analyses of each species with detailed chromosome size measurements are presented in Table 2 and Figure 2. The previously reported chromosome number for the genus Iris varied from 2n = 14 (I. vorobievii) to 2n = 108 (I. versicolor) representing approximately a 7.7-fold difference (Table S1). To date, the chromosome numbers have been determined for approximately 168 taxa of ca. 300 currently recognized species in the genus Iris worldwide (c. 56%; Table S1) [17].   Figure 2F). HKL of this species was 86.4 ± 3.8 µm. All seven plants of I. pseudacorus had 2n = 34 ( Figure 1G), AsI was 62.0% and the karyotype formula was 26 m + 8 sm ( Figure 2G) with HKL of 107.1 ± 4.2 µm. All nine plants of I. rossii had 2n = 32 ( Figure 1H). This species had AsI of 60.9%, karyotype formula of 2n = 24 m + 8 sm ( Figure 2H) and HKL of 100.5 ± 3.0 µm. All seven plants of I. sanguinea possessed 2n = 28 ( Figure 1I), AsI of 62.5%, and the karyotype formula of 18 m + 10 sm ( Figure 2I) with HKL of 87.2 ± 2.8 µm. All five plants of I. setosa had 2n = 38 ( Figure 1J), in agreement with previous reports [17] except for the one study reporting 2n = 40 for this species [31]. This deviating count might have potentially referred to another taxon (either a species or a variety), but it is difficult to verify due to unknown origin of the examined sample (e.g., cultivated material) [31]. Most previous studies used cultivated material. Thus, any discrepancies that might exist between different reports might have been affected by domestication and associated potential hybridization [31], as also suggested for Iberian irises [37]. Iris setosa had AsI of 61.6% and karyotype formula of 2n = 20 m + 18 sm ( Figure 2J) with HKL of 122.9 ± 6.9 µm.

Iris L. Subgenus Pardanthopis (Hance) Baker
All four examined individuals of I. dichotoma had 2n = 32 ( Figure 1M). This number was consistent with previous records for this species from Russian and Chinese populations [17], except for the study of Kim et al. [31], in which analyzed cultivated material was reported to have 2n = 34. This species had AsI of 61.9% and karyotype formula of 2n = 18 m + 14 sm ( Figure 2M) with HKL of 97.0 ± 4.4 µm. Iris domestica had 2n = 32 ( Figure 1N), which was in agreement with previous reports (Table S1) [17]. Its AsI was 59.7% and karyotype formula was 2n = 22 m + 10 sm ( Figure 2N) with HKL of 92.8 ± 4.2 µm.

Sisyrinchium rosulatum E.P.Bicknell.
Both of the S. rosulatum individuals analyzed had the chromosome number of 2n = 32 ( Figure 1O), and the results were in agreement with previous reports [17]. The AsI of this species was 59.9% and karyotype formula was 2n = 24 m + 8 sm ( Figure 2O) with HKL of 43.4 ± 1.9 µm.

Genome Size Variation in Iris L.
DNA content analyses using an internal standard revealed clear and well-defined peaks for all samples with the coefficients of variation (CV) for the internal standard and sample peaks ranging from 1.03% to 5.40% (mean ± S.D. = 2.24 ± 0.80) and from 1.23% to 6.50% (mean ± S.D. = 2.87 ± 1.06), respectively.
Genome sizes for all Korean irises, except for Iris pseudacorus, are reported here for the first time (Table 1; Figure 3). The analysis revealed a 3.29-fold variation in nuclear DNA content among the analyzed taxa, with genome sizes ranging from 2.39 pg/1 C in Iris ruthenica to 7.86 pg/1 C in I. ensata. To date, the 1 C-values of 44 Iris species have been reported ranging from 1.05 pg/1 C in I. versicolor to 28.20 pg/1 C in I. histrio [15]. Genome size of Iris pseudacorus reported earlier (1 C = 5.67 pg) [38] is very similar with the current data (5.77-6.02 pg/1 C; Table 1). The outgroup species Sisyrinchium rosulatum had 0.90 pg/1 C (Table 1; Figure 3). The 1 C-values in Korean irises vary significantly among species, but are constant within populations and species, as previously observed in closely related species within a genus [10,39]. This suggests that ploidy levels in the populations of Korean irises examined in this study are at least partially stable within species. The new data presented here will allow for better understanding of the dynamics and patterns of chromosomal evolutionary changes in the genus Iris [15].

Evolution of Chromosome Numbers and Genome Sizes
The newly obtained chromosome numbers and genome sizes were mapped on nrITS phylogeny of Korean irises (modified from [27]). The data revealed that there is no correlation between chromosome numbers and genome sizes among the phylogenetic lineages of the analyzed Korean Iris species. As an example, Iris ensata with low chromosome number of 2n = 24 has the highest 1 C-value (7.76 pg/1 C) whereas its close relative I. setosa has 2n = 38, but 1.4-fold lower genome size ( Figure 4). It suggests that dysploidy and polyploidy are not solely responsible for the observed variation of genome sizes and chromosome numbers, but also other processes, like independent accumulation or reduction of repetitive DNAs amount (e.g., satellite DNAs and/or transposable elements) play an important role, as previously reported in other plant groups [7,[40][41][42][43].
Other lineages within the genus Iris might be more stable evolutionarily. Both Iris ruthenica and I. uniflora (subgenus Limniris section Ioniris) have the same chromosome numbers of 2n = 42 and similar genome sizes (ca. 2.4 pg/1 C). Thus, the current cytogenetic data support the monophyly of this section, as proposed based on phylogenetic analyses of plastid markers [44][45][46], but also suggest low levels of genome dynamics accompanying their evolution. Iris dichotoma and I. domestica of the subgenus Pardanthopsis have also been recovered in monophyletic clade in ITS-and plastid-based phylogenetic analyses [46,47]. However, although they share the same chromosome numbers of 2n = 32, they differ nearly a 1.5-fold in genome sizes. Such variation of genome sizes suggests independent accumulation/loss of repetitive DNAs in the absence of numerical chromosome changes as reported also in other monocot genera [7,40].
In contrast, allopolyploidy can be hypothesized for the origin of Iris koreana (2n = 50) from closely related I. minutoaurea (2n = 22) and I. odaesanensis (2n = 28) [30]. It is also supported from the additive genome size of this taxon. Allopolyploidy has been implied to be an important process in the early diversification of the whole family of Iridaceae [48,49]. Thus, further data employing GISH (genomic in situ hybridization) will allow testing the allopolyploid origin of this and other species in the section Limniris. The subgenus Limniris section Limniris has previously been shown to be polyphyletic [27,30,[44][45][46]. This study shows that there is also quite wide variation of chromosome numbers and genome sizes within the newly identified polyphyletic clades of this subgenus (Figures 3 and 4). It suggests that the genome size and chromosome number evolved rather dynamically in those evolutionary lineages, as indeed also observed in other monocot genera [11,50,51].

Conclusions
The present study provides the first comprehensive analyses of chromosome numbers, karyotype features and genome size variation of Korean irises from natural populations. Despite the large variation of chromosome numbers (n = 11, 12,14,16,17,19,20,21,25), some variation of genome sizes (2.42-7.76 pg/1 C) has been demonstrated. This variation is only partially phylogenetically structured. Numerical chromosomal changes, both polyploidy and dysploidy, seem to be accompanied to a different degree by the likely independent changes in the amounts of repetitive DNAs in different phylogenetic lineages. Various combinations of these processes acting on the genomes of closely related species create complex patterns of genome size and chromosome number variation observed in Korean irises. Further analyses are needed to elucidate how these processes affect the evolution of the genomes of individual species, not only of the Iris species in Korea but ideally in a broader phylogenetic context of whole subgenera/sections.