The molecular taxonomy of three endemic Central Asian species of Ranunculus(Ranunculaceae)

Worldwide, the genus Ranunculus includes approximately 600 species and is highly genetically diverse. Recent taxonomic reports suggest that the genus has a monophyletic origin, divided into two subgenera, and consists of 17 sections. The Central Asian country of Kazakhstan has 62 species of the genus that have primarily been collected in the central part of the country. The latest collection trips in southern parts of the country have led to the description of a wider distribution area for Ranunculus and the identification of a new species Ranunculus talassicus Schegol. et A.L. Ebel from Western Tien Shan. Therefore, in this study, attempts were made to assess the molecular taxonomic positions of R. talassicus and two other species endemic to the Central Asian region R. karkaralensis Schegol. and R. pskemensis V.N. Pavlov in relation to other species of the genus, using internal transcribed spacer (ITS) molecular genetic markers. The ITS-aligned sequences of 22 local Central Asian accessions and 43 accession sequences available in the National Center for Biotechnology Information (NCBI) database allowed the construction of a maximum parsimony phylogenetic tree and a Neighbor-Net network. The results indicated that R. talassicus and R. pskemensis could be assigned to section Ranunculastrum. Additionally, an assessment of the network suggested that R. pskemensis was the rooting taxon for the group of species containing R. talassicus, and that R. illyricus L. and R. pedatus Waldst. & Kit. were founders of a prime rooting node for the Ranunculastrum section of the genus. The ITS-aligned sequences showed that R. karkaralensis was indifferent with respect to three other species in the Ranunculus section of the genus, i.e., R. acris L., R. grandifolius C.A. Mey., and R. subborealis Tzvelev. The study indicated that the assessments of ITS-based phylogenetic tree and Neighbor-Net network provided new insights into the taxonomic positions of three endemic species from Central Asia.


Introduction
The Ranunculaceae Juss., or buttercup family, has a worldwide distribution and comprises more than 2500 species belonging to 59 genera. The family members span a wide range of a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 different ecological conditions, especially in the Northern Hemisphere. The largest cosmopolitan genus of the family is Ranunculus L., which contains approximately 600 species distributed worldwide [1,2]. Due to its wide distribution, this genus is highly genetically diverse. Consequently, classification of the genus Ranunculus is very complex and has not been fully completed. Initial worldwide classifications of Ranunculus were based on descriptions of achenes, flowers, roots [3], and fruit anatomy [4]. Later, Tamura revised the classification of the genus based on a reassessment of the achene structure [1,2]. Tamura's classification subdivided the genus into seven subgenera: Pallasiantha, Coptidium, Ficaria, Batrachium, Crymodes, Gampsoceras, and Ranunculus. In this classification, the subgenus Ranunculus was further subdivided into 20 sections [2]. Soon thereafter, DNA markers were implemented to clarify and discuss the phylogenetic relationships within Ranunculaceae [5][6][7][8]. Several publications using molecular markers have been dedicated to the Ranunculus genus and suggest that the molecular taxonomy of the genus is not congruent with the classifications based on botanical studies [8][9][10]. For instance, phylogenetic analyses of c. 200 species of Ranunculus based on sequences of the muclear ITS (nrITS) region [9] showed a high level of similarity to previous chloroplast DNA restriction site analyses [8] and differed from the previous classification based on phenetic studies.
DNA barcoding studies have often been used by different scientists for evaluating the generic delimitation and infrageneric classification of the genus Ranunculus [9][10][11][12][13][14][15]. Most of these studies have relied on the use of nucleotide sequences of the internal transcribed spacer (ITS) of the nuclear ribosomal DNA (rDNA) region, as well as plastid regions. ITS are widely used for differentiating between the generic and infrageneric levels of plant phylogenetics [16] and are one of the most powerful barcode DNA markers available [17]. ITS markers have been confirmed to be useful for clarifying the phylogenetic relationships between closely related species and genera in Ranunculaceae [18][19][20][21]. However, a complete study of the taxonomy of the genus was based on using both DNA markers and morphological data, with the division of 226 taxa into two subgenera and 17 sections [14].
The Central Asian country of Kazakhstan is the ninth largest country in the world and contains a rich array of flora, including a wide diversity of endemic and rare plant species. According to the Flora of Kazakhstan, the country has 62 species of Ranunculus [22], including the genus Batrachium (DC.) Gray. Previously, the botanical descriptions of Kazakh buttercups were related to the spatial distribution of the species within different floristic regions of Kazakhstan [23]. The botanical taxonomy of the genus Ranunculus has primarily been studied using samples collected in Central Kazakhstan [24,25]. For instance, Schegoleva (2014) [24] presented the taxonomic composition of Ranunculus in Central Kazakhstan as consisting of 19 species, which were part of the subgenera Ranunculus and Batrachium. As a result of the collection trips in this region, a new Ranunculus species R. karkaralensis Schegoleva was claimed to have been identified [26]. The author noted that R. karkaralensis is similar to R. grandifolius C.A. Meyer and differs from it in terms of the form and proportions of the sheet plate of basal leaves and the absence of a well-developed horizontal rhizome [26].
Until recently, very little effort has been made to compile a botanical description of Ranunculus species in the southern part of the country, where a new species R. talassicus Schegol. et A.L. Ebel from Western Tien Shan was recently described [27]. It was suggested that R. talassicus is morphologically close to R. alaiensis Ostenf. from Pamir-Alai and differs from it in terms of having triple-dissected basal leaves [27]. As these newly identified species have not been analyzed using DNA markers, the purpose of this study was to clarify the molecular taxonomy of R. karkaralensis and R. talassicus based on ITS markers. This work is part of a national program aiming to evaluate the genetic variation in the wild flora of Kazakhstan [28]. The project, along with DNA barcoding analyses of wild plant species [29][30][31][32], is also incorporating studies using simple sequence repeats (SSR) [33], amplified fragment length polymorphism (AFLP) [34], and single nucleotide polymorphism (SNP) [35] markers in a collaboration between local botanists and geneticists.

Plant materials
Plant materials for genetic analyses were collected during the period 2015 to 2019 from different parts of Kazakhstan (Fig 1). The geographic locations of the collection sites are given in S1 Table. The permission for field work at the Sairam-Ugam National Nature Park (Turkestan region, Kazakhstan) was given by the authorities of the Park (S1 Appendix).
Two endemic Kazakhstan species R. talassicus and R. karkaralensis (Fig 2), which were recently described as new species, were included in the phylogenetic analysis.
The study also included R. pskemensis V.N. Pavlov (species endemic to Uzbekistan), first sampled in Kazakhstan (Turkestan region) in 2019. The holotypes of the new species were deposited in TK (Herbarium at the Tomsk State University, Tomsk, Russian Federation), and isotypes were deposited in TK, AA (Herbarium at the Institute of Botany and Phytointroduction, Almaty, Kazakhstan), and TASH (Herbarium at the Institute of Botany, Tashkent, Uzbekistan) for R. talassicus and in TK for R. karkaralensis.

DNA isolation, amplification, and sequencing
Total genomic DNA was isolated from dried leaf material of the species studied using the CTAB (cetyl trimethylammonium bromide) protocol [36]. The quality and concentration of DNA were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). The internal transcribed spacers 1 and 2 (ITS) and the 5.8S rRNA gene were amplified using primers ITS1nF (5'-AGAAGTCGTAACAAGGTTTCCGTAGG-3') and ITS4nR (5'-T CCTCCGCTTATTGATATGC-3') with an annealing temperature of 58˚C [37]. Polymerase chain reaction (PCR) products were checked on 1.5% agarose gels. Single bands with the expected size of approximately 650 base pairs (bp) were cut out from the gel and subsequently purified using ULTRAPrep1 Agarose Gel Extraction Mini-Prep Kits (AHN Biotechnologie GmbH, Nordhausen, Germany) according to the protocol provided by the company. Purified PCR products were amplified with forward and reverse primers separately using the BigDye Terminator v.3.1Cycle Sequencing Kit (Applied Biosystems, USA). The ethanol/EDTA method [38] was used for precipitation of the products. Further purified PCR products were sequenced in forward and reverse directions using the ABI 3130 DNA sequencer (Applied Biosystems, USA).
For reconstruction of the phylogenetic tree, 29 samples of 22 species collected in Kazakhstan, Uzbekistan, and Tajikistan (S1 Table) were sequenced, and 43 ITS sequences were downloaded from NCBI. The set of sequences consisted of 67 Ranunculus samples, including 22 local samples (one per species); 43 samples from NCBI; and 2 outgroup species, i.e., Trollius altaicus C.A. Mey. and T. ledebourii Rchb. New sequences were deposited in GenBank and are listed in S1 Table. In the phylogenetic analyses, the names of the sections and subgenera are given, according to Hörandl and Embadzade [14] and Schegoleva [43]. Genetic relationships between haplotypes were estimated using the Neighbor-Net algorithm in SplitsTree4 software [44]. The aligned sequences were converted into a Nexus file format in DNASP v5.10 [45] for analysis by SplitsTree software.

Results
The length of the ITS sequences varied from 573 to 577 base pairs (bp) for 65 Ranunculus samples, without including the outgroup species, and up to 591 bp including the two outgroups. In general, out of the 591 aligned ITS nucleotides, 206 bp were polymorphic, including 121 bp in ITS1, 14 in 5.8S, and 71 in ITS2. The total variability of the ITS sequences was 34.85%, with 26 nucleotide sites appearing as singletons, and 159 sites registered as parsimonious informative sites.
The alignment of ITS sequences for the 65 Ranunculus samples collected in Kazakhstan, neighbouring Central Asian countries, and Genbank was used for the construction of ML, MP, and NJ phylogenetic trees. An assessment of the resulting phylogenetic trees of the three methods suggested that the MP method provided clusters with the highest bootstrap values (Fig 3, S2 Appendix). Therefore, in the following steps of the study, the MP phylogenetic tree was selected for the analysis. The MP phylogenetic tree classified samples into two large Ranunculus and Auricomus subgenera clades, with further separation into seven clusters, where three clusters corresponded to the Auricomus subgenus and four clusters to the Ranunculus subgenus (Fig 3).
In the Auricomus subgenus, R. sceleratus was the closest taxon to the first cluster consisting of six species from the sections Hecatonia and Batrachium, where four species of Batrachium were confidently separated from species of Hecatonia. Notably, R. rionii Lagger, the species with a previously unknown position in genus taxonomy, was clustered together with R. natans from the section Hecatonia. The second cluster includes 16 taxa of the section Auricomus, including four local species (R. songaricus Schrenk, R. albertii Regel & Schmalh., R. polyrhizos Stephan ex Willd., and R. rubrocalyx Regel ex Kom.). The third cluster consists of two species of the section Flammula that was the closest group to representatives of the subgenus Ranunculus (Fig 3).
The subgenus Ranunculus clade comprises taxa of the sections Ranunculus, Echinella, Polyanthemos, and Ranunculastrum (Fig 3). The section Ranunculus was the first clustered group in the subgenus after the section Flammula and consists of eight taxa, including four species from Kazakhstan. The list of four species includes R. karkaralensis from Central Kazakhstan, which was previously described as a new species in the genus. The ITS sequence alignment of six taxa that were grouped in the section Ranunculus showed that R. acris, R. grandifolius, R. subborealis, and R. karkaralenis are invariant in all three segments of the ITS (ITS1, 5.8S, and ITS2). This group of species has two nucleotide differences from R. turneri and five nucleotide differences from R. japonicus Thunb. (Table 1). As three out of those five polymorphic nucleotides are in ITS2, it is interesting to note that ITS2 is a more informative segment for this subclade, although the total number of polymorphic nucleotides for the entire genus was higher in ITS1 in comparison to ITS2.
In the next segment of the phylogenetic tree, a solo species R. arvensis L. in the section Echinella, and four species of the section Polyanthemos were placed after the section Ranunculus (Fig 3). Finally, the last subclade contained 27 specimens, including 20 taxa that were previously assigned to the section Ranunculastrum. Therefore, it was hypothesized that seven unassigned species in this subcluster also belong to the section Ranunculastrum. Three out of seven unassigned taxa (R. talassicus, R. alaiensis, and R. pskemensis) are species endemic to Central Asia, with two of them (R. talassicus and R. pskemensis) growing in southern Kazakhstan (S1 Table). A separate group consisting of these three Central Asian species was analyzed against taxa from eight sections of the subgenera Auricomus and Ranunculus using Nei's unbiased genetic distance ( Table 2). The results suggest that this group of species is genetically closer to the section Ranunculastrum (0.005), followed by sections Polyanthemos (0.035) and Echinella (0.051), which is an additional indication that all previously unassigned species in this subclade belong to the section Ranunculastrum.
In a separate study, the alignment of ITS sequences from R. talassicus, R. alaiensis, and R. pskemensis with R. illyricus, which is a key species in the section Ranunculastrum, suggested that R. talassicusis is the most distant species from R. illyricus (25 base pairs), followed by R. alaiensis (24 bp) and R. pskemensis (20 bp) ( Table 3). A direct comparison of ITS sequences in R. talassicus and R. alaiensis revealed eight nucleotide differences.

PLOS ONE
The molecular taxonomy of endemic Central Asian species of Ranunculus In general, Nei's genetic distance ( Table 2) suggested that the nearest sections were Hecatonia and Batrachium (0.013), and the most distant sections were Auricomus and Ranunculus (0.109). The analysis of molecular variance (AMOVA) for the ITS sequences of specimens from eight sections suggested that the total genetic variation was partitioned into 73% of the variation within the sections and 27% between sections.
The second analysis was related to haplotype network assessment, where nucleotide sequences were converted into haplotypes. In total, the network reconstruction allowed the identification of 61 haplotypes, including outgroups (Fig 4, S2 Table).
The haplotype diversity for this set of data, Hd, was 0.996; the nucleotide diversity, π, was 0.064; and the average number of nucleotide differences, k, was 33.033. The haplotype groups generated corresponded to the MP phylogenetic tree subclades (Fig 3), and the profile of the network confirmed the monophyletic origin of the genus. The network, with outgroups, confirmed the separation of the sections in the subgenera Auricomus and Ranunculus. The Auricomus profile hinted that species from the sections Hecatonia and Batrachium are closest to the most common predecessors of the genus, as their profiles were next to the nodes, which linked directly with the outgroups (Fig 4). Notably, R. rionii (Hap 43, Gene bank accession FM242855), which had previously not been studied in the genus taxonomy, was positioned in the middle between Hecatonia and Batrachium species, unlike its position in the phylogenetic tree (Fig 3). The subgenera Ranunculus profile of the network is nearly congruent with the phylogenetic tree, except for the section Echinella, which, similar to Hecatonia and Batrachium, is located closer to the hypothetical most common evolutionary ancestor (MCEA) of the genus from the opposite side of the subgenera Auricomus (Fig 4). Further in the network, in the section Ranunculus, the key haplotype is Hap 1, which not only consists of the four invariable species R. grandifolius, R. acris, R. subborealis, and R. karkaralensis, but is also close to Hap 47 (R. turneri) and Hap 49 (R. japonicus) (Fig 4). In the section Ranunculastrum, the network profile indicates that the first haplotypes linked with the median vector on the main axis were Hap 6 (R. pedatus) and Hap 54 (R. illyricus). The network confirms that haplotype Hap 15 of the newly discovered species R. talassicus is connected to Hap 11 (R. alaiensis). The branch with these two species has a root with R. paucidentatus (Hap 18). A larger group of branches containing ten species, including R. talassicus, has a root with a more ancient node connected with R. pskemensis (Hap 4).

Discussion
The taxonomic order in the study was determined using 226 species of the genus and by the application of DNA markers and morphological data. In the current study, the molecular  phylogeny of three Central Asian endemic species, R. karkaralensis, R. talassicus, and R. pskemensis was assessed using 65 Ranunculus species, including 43 taxa from NCBI, two outgroup species, and ITS nuclear genome marker data. The analysis of polymorphism levels in the ITS indicated that for 591 aligned ITS nucleotides, 206 bases were polymorphic, including 121 bases in ITS1, 14 in 5.8S, and 71 in ITS2. Recently, two reports have discussed the level of variability in different DNA barcoding markers for the differentiation of medicinal plants [46,47]. The authors suggested that ITS2 is one of the most appropriate markers for species identification and is more informative than other types of markers, including ITS1. However, the results in this study indicate that the number of polymorphic sites in ITS1 is higher than that in ITS2. Hence, the outcome suggests that it is useful and possibly necessary to use both ITS1 and ITS2 spacers for phylogenetic studies.

PLOS ONE
The molecular taxonomy of endemic Central Asian species of Ranunculus Despite the smaller number of samples and lower bootstrap values in the current study in comparison with the report of Hörandl and Emadzade (2012) [14], the profiles generated in the phylogenetic trees were similar. The profiles of both the MP phylogenetic tree and the Neighbor-Net network within the present study were also nearly congruent (Figs 3 and 4). Therefore, these similarities suggest taxonomic assessment of these three endemic species is reliable. However, there were minor differences in positions between the phylogenetic tree and the Neighbor-Net network. These differences include the closer distance of the section Echinella from the edge of subgenus Ranunculus to the MCEA, unlike the result from the phylogenetic tree, where Ranunculus is looked as early-diverging section in the subgenus (Fig 3).
The intragenera differences of taxa identified by the Neighbor-Net network for this group of sections were most probably predetermined by species evolution, genetic distinction, geographic distribution, and ecological conditions of growth [14]. For instance, the section Ranunculus consists of approximately 30, mostly perennial, species, both diploid and polyploid, widely distributed in Eurasia, North America, and Africa, with the basal number of chromosomes x = 7 (rarely 8) [14]. The section Echinella is rather monotypic and includes R. arvensis, which is an annual species with enlarged nuts that have a size of more than 5 mm. R. arvensis is a very cosmopolitan species with a wide distribution in Eurasia and is characterized by the basal number of chromosomes x = 8 [14]. The section Polyanthemos is the largest group (80-90 species); is very morphologically diverse; has both annual and perennial species; has diploids and polyploids with the basal number of chromosomes x = 8 (rarely 7); and is spread over all continents, except Antarctica [9,48]. The phylogenetic studies suggest that Polyanthemos includes 46 species from the Southern Hemisphere, with regional clades formed in Australia, Malaysia, and South America [9,14,48]. These findings may explain the phylogenetic position of the section Polyanthemos, which contrasts with the positions of Ranunculus and Echinella (Figs 3 and 4), since the origin of these two sections is associated with the Eurasian continent.
The similarities of phylogeny between sections Batrachium and Hecatonia, shown in Figs 3 and 4, can presumably be linked to commonalities associated with aquatic and wetland conditions of growth. Previously, species of the section Batrachium were often classified as a separate genus or subgenus. The section is well-spread over all continents, except Antarctica, and is composed of annual and perennial species with white-flowered buttercups and sometimes with a high ability for cross-hybridization [49]. A poor clusterization of R. sceleratus and R. natans in the section Hecatonia (Fig 3) supported the observation that this section is more morphologically diverse than Batrachium [14]. The species in Hecatonia are mostly distributed in the Northern Hemisphere and also in South America and Antarctica. Hecatonia is represented by annual and perennial species and diploids and polyploids, with yellow flowers and well-developed stem leaves [14]. Hence, the clustering of species in these two sections may reflect visible morphological differences and geo-ecological conditions of growth.
The evaluation of the ITS nucleotide sequence in R. karkaralensis suggested no difference from R. acris, R. grandifolius, and R. subborealis, and all of these taxa were placed in the section Ranunculus. Morphologically, R. karkaralensis is the most close to R. grandifolius, as they only differ in terms of the absence of a well-developed horizontal rhizome and the shape and proportions of the leaf blade of the basal leaves. The lack of differences in the ITS region for these species could be explained by several reasons, such as (i) the low-resolution power of the selected marker for these species, (ii) recent speciation events in the newly identified species, and (iii) high hybridization rates among subspecies within a species. Therefore, despite obvious morphological differences between R. grandifolius and R. karkaralensis [26], additional evidence is likely to be required to accept the latter as a separate species. A very different outcome was revealed in the genetic analyses of R. talassicus and R. pskemensis in comparison with other taxa of the genus. The assessment of the genetic distances between species (Table 2) and the phylogenetic tree (Fig 3) clearly placed R. talassicus and R. pskemensis in the section Ranunculastrum within the subgenus Ranunculus (Fig 4). The phylogenetic tree, similar to the network, clustered R. talassicus together with R. alaiensis (Figs 3 and 4). The result supports the botanical assessment, which suggests strong similarities between these vicar species [27]. Morphologically, R. talassicus is very similar to R. alaiensis; however, these species are easily distinguished by the shape of the basal leaves. In addition to geographical fragmentation, they grow at different altitudes, although the ecological niches are similar. R. talassicus is found in the Western Tien-Shan at an altitude of 2400-3200 m above sea level. R. alaiensis is found in the high mountains of the Pamir-Alai, and its altitudinal confinement begins at 2700 m on the Darvaz Range and 3500 m above sea level on the Gissar Range. R. talassicus is also morphologically similar to R. pskemensis, and the latter species is also found in the western part of the Tien-Shan. However, unlike R. talassicus, R. pskemensis occupies other habitats [27]. The ITS sequence alignment revealed that R. talassicus was 20 nucleotide bases away from R. pskemensis and 25 bases from R. illyricus (Table 3). The difference was reflected in the network, with R. pskemensis forming the root for the larger branch with R. talassicus, and R. illyricus along with R. pedatus forming a prime rooting node for the section (Fig 4). Therefore, the taxonomic positions for both Central Asian endemic species R. talassicus and R. pskemensis are well defined.

Conclusions
Ranunculus is a very large and complex genus spread across moderately cold climatic regions worldwide. Studies of its molecular taxonomy have shown that this genus is well-classified and suggest a separation of 226 species into two subgenera and 17 sections. However, despite a thorough botanical classification of taxa in the genus, the molecular taxonomy of endemic species from Central Asia is poorly understood. Therefore, three previously unstudied Kazakh species, along with 19 other local species and 43 taxa from the NCBI, were analyzed using the ITS marker region of the nuclear genome. The analysis of ITS sequences confirmed the monophyletic origin of the genus and clustered the previously unstudied Central Asian species into their sections. It was determined that R. karkaralensis has invariant ITS sequences with three other species in the section Ranunculus, i.e., R. acris, R. grandifolius, and R. subborealis. The two other endemic Central Asian species R. talassicus, and R. pskemensis were assigned to the section Ranunculastrum. It was revealed that R. talassicus is most closely related to R. alaiensis. Therefore, the application of ITS-aligned sequences has provided new insights into the taxonomic positions of three endemic species from Central Asia.