Study of phylogenetic relationship of Turkish species of Klasea (Asteraceae) based on ISSR amplification

Abstract Klasea is a taxonomically complex genus in which there are many problems, mostly with Klasea kotschyi and Klasea hakkiarica. It is challenging to differentiate the genera based on morphological characters alone. Revision studies performed on the basis of molecular data obtained from studies conducted in recent years have made the phylogenetic relationships and systematic positions of the taxa more apparent and reliable. In this study, Klasea, Serratula, Jurinea and Centaurea species native to Turkey, were collected from different localities of Anatolia and DNA was isolated from the collected samples. The data were analyzed ordination analyses including UPGMA and PCA using NTSYSpc 2.1. The infrageneric and intergeneric phylogenetic relationships between Klasea and other related genera were also characterized. The Klasea species were grouped into three clusters. It was determined that taxa Klasea kotschyi and Klasea hakkiarica are separate but closely related. Moreover, it was observed that the Klasea lasiocephala a separate group within the genera. Clearly the genera Klasea, Serratula, Jurinea and Centaurea are phylogenetically differentiated on the dendogram.


Introduction
Th e tribe Cardueae (Asteraceae) is generally accepted to be classifi ed into fi ve subtribes named Echinopinae, Carlininae, Carduinae, Centaureinae and Cardopatiinae (Susanna et al. 2006). Cardueae include perennial, biennial, or monocarpic herbs and shrubs and, less often, annual herbs or small trees (Barres et al. 2013). However, delineation of these taxonomic entities is highly problematic. Beyond the limits of the tribes, the boundaries between these units are also very diffi cult to establish. Also, some large genera of the tribes have generic delimitation problems: Carduus L. (90 species), Cirsium Mill. (250 species), Centaurea L. (400 species), Cousinia Cass. (800 species), Serratula L. (70 species), and Saussurea DC. (more than 300 species) . Extensive work conducted recently by Garcia-Jacas et al. (2000 and Font et al. (2002) have clarifi ed the delineation of Centaurea. Limited studies also exist on Cirsium and Carduus (Haff ner and Hellwig 1999), but most of the taxonomic problems persist. Th e genus Klasea Cass. constitutes a taxonomically complex group of plants with generic boundaries are unclear, especially at the generic level surrounding genus Serratula (Martins and Hellwig 2005). Klasea Cass., traditionally treated as a section within Serratula L., is widely accepted at the generic level (Martins 2006). Klasea is naturally distributed in Central Asia, Iran, Turkey, China, Himalayas, south east Europe and south Russia. Klasea is located within the monophyletic tribe Cardueae, in the subtribe Centaureinae (Susanna et al. 2006). 16 species were reported for the genus Serratula in Turkey (Davis and Kupicha 1975;Davis et al. 1988). Th en all Turkish Serratula species were transferred to Klasea except Serratula tinctoria (Greuter 2003;Martins 2006). Th us, Klasea is represented by 15 species and Serratula is represented by one species within the Mediterranean and Irano-Turanian phytogeographic regions of Turkey (Dogan et al. 2012). Five of these species are endemic to Turkey, resulting in an endemism ratio of 33.3% (Dogan et al. 2012).
Currently, morphological revisions of various plant taxa are often supported by molecular data (APG 2003). As compared with morphological data, DNA sequences are not infl uenced by the environmental conditions in which the plants have grown; hence they serve as a powerful tool in resolving taxonomical and systematical problems. When compared with the phenotypic characters, by using diff erent molecular marker systems, more reliable results were also obtained by a number of researchers that used diff erent plant groups (Yang et al. 1996;Joel et al. 1998;Soranzo et al. 1999;Bremer et al. 2001;Mengitsu et al. 2002;Ash et al. 2003;Jump et al. 2003;Pharmawati et al. 2004;Dogan et al. 2007;Ali et al. 2013).
Th e RAPD (Randomly Amplifi ed Polymorphic DNA) fi ngerprinting method is widely used and has a wide range of applications (Williams et al. 1990). However, because RAPD is a highly sensitive method, it should be used with great care. Th e ISSR (Inter Simple Sequence Repeat) has much higher levels of reproducibility than RAPD, for which reason it is preferable (Zietkiewicz et al. 1994, Prevost andWilkinson 1999;Dogan et al. 2007;Hakki et al. 2010). Th e ISSR method is very widely used for the analysis of genetic diversity (Prevost and Wilkinson 1999).
Simple sequence repeats (SSRs), also known as microsatellites, are tandemly repeated di-, tri, tetra-or penta-nuclotide sequences (mainly within the range of 10-80 repeats of the core unit) that are abundant within eukaryotic genomes. A high level of genomic variation is generated by the more or less evenly distributed microsatellite sequences present within the plant and animal genomes. Th e high levels of genomic variation are widely used for genetic variation analysis of both wild plants (Wolfe et al. 1998;Dogan et al. 2010;Laosatit et al. 2013;Khalik et al. 2014) and crop plants (Vosman and Arens 1997;Hakki et al. 2001;Mohammadzadeh et al. 2011). Microsatellites can be used in inter-as well as intra-species analyses (Soranzo et al. 1999). However, the technique requires prior sequence information for the locus-specifi c primers, a feature that renders it diffi cult to be applied to plants for which no adequate genomic sequencing studies exist. Without considering their diffi culty or cost (Hakki and Akkaya 2000), numerous microsatellite loci have been identifi ed for economically important crops such as wheat, rice or maize. In Klasea, however, they have not been utilized.
In this study, Klasea species, which are diffi cult to delineate using morphological traits, were collected from their natural habitats in Turkey. DNA was isolated and fi ngerprinting was performed using a highly reliable and reproducible technique that mimics the application ease of RAPDs. Th e method employed to assess the genetic diversity and to resolve the genetic relationships among the species is a technique derived from SSR characterization based on PCR amplifi cation of ISSR regions primed by a single oligonucleotide corresponding to the targeted repeat motif. Th e SSR-containing primers are usually 16-25 base pair long oligonucleotides anchored at the 3'-or 5'-end by two to four arbitrary, and often degenerate, nucleotides (Fang et al. 1997). Th e primer can be based on any of the motifs found at SSR loci. In these conditions, only sequence regions fl anked by the two adjacent identical and inversely oriented microsatellites are amplifi ed. Overall, the technique does not require prior sequence information (an advantage against microsatellites) and its reliability is higher than RAPD's.
Th e aim of this study was to determine the genetic relationships among selected Anatolian-originated Klasea, Serratula, Jurinea and Centaurea species collected from diverse regions of Turkey and to use a DNA-based molecular marker system to resolve the unclear and controversial status of these species based on conventional morphological characters.

DNA extraction
Nuclear DNA of silica gel dried leaf samples were extracted according to the instructions of the Nucleon phytopure plant DNA extraction kit (RPN 8510, Amersham Life Science, England). For each sample, DNA was extracted from 100 mg of leaf. After concentrations were determined using an Eppendorf BioPhotometer, DNA samples were diluted to the working concentration of 25 ng/μL. To better quantify the DNA and to assess the quality of the DNA, samples were run on an agarose gel (0.9%), stained with ethidium bromide, against a DNA standard with known concentrations. Stock DNA was kept at -86 °C.

ISSR Amplifications
Of the 20 primers investigated during our initial screening, the primers that gave the most informative patterns (in terms of repeatability, scorability, and the ability to distinguish between varieties) were selected for fi ngerprinting. Th e characteristics of the primers used are given in Table 2. Each reaction contained 2.5 mM MgCl 2 , 10 mM Tris-HCl (pH 8.8), 50 mM KCl; 0.8% Nonidet P40, 200 mM of each of dNTP, 0.5 mM primer, 25 ng DNA template and 0.4 units of Taq DNA Polymerase (Bioron, Germany) in a final reaction volume of 25 μl. After a pre-denaturation step of 3 minutes at 94 °C, amplification reactions were cycled 40 times at 94 °C for 1 minute, at annealing temperature (Table 1) for 50 seconds and 72 °C for one minute followed by a final 10 minutes 72 °C extension in an Eppendorf Mastercycler gradient thermocycler. Upon completion of the reaction, aliquots of PCR products (15 μL) were mixed with 3 μL of loading buffer (50% glycerol, 0.25% bromophenol blue and 0.25% xylene cyanol), loaded onto a 2.0% agarose/1x Tris-Borate EDTA gel and electrophoresed at 4 V/cm. Amplifi cations were repeated at least twice at diff erent time periods for each primer using the same reagents and procedures.

Data collection and cluster analysis
Amplifi ed fragments were visualized under a UV transiluminator and photographed using a gel documentation system (Vilbert Lourmat, Infi nity model). All of the amplifi ed fragments were treated as dominant genetic markers. Each DNA band generated was visually scored as an independent character or locus (1 for presence and 0 for absence). Qualitative diff erences in band intensities were not considered. Every gel was scored in triplicate (independent scorings) and only the fragments consistently scored were considered for analysis. A rectangular binary data matrix was prepared and all the data analysis was performed using the Numerical Taxonomy System, NTSYS-pc version 2.1 (Applied Biostatistic, Exeter Software, Setauket, New York, USA). In cluster analysis of the samples, the unweighted pair-group method with the arithmetic mean (UPGMA) procedure was followed (Rohlf 1992). Th e genetic distances were calculated with the SM coeffi cient. In order to determine the ability of ISSR data to display the inter-relationships among the samples, principle co-ordinate analysis (PCA) of pair-wise genetic distances (Nei 1972) was also conducted using the NTSYS-pc package.

Results and discussion
Silica gel dried plants collected from 19 diff erent natural habitats were taken to the laboratory. Th e total number of species collected and used in the phylogenetic analysis was 19. DNA extractions were fi rst attempted using a standard 2X CTAB method. Due to the poor DNA quality produced by the CTAB procedure, a commercial kit (Nucleon phytopure) was used in all isolations and repeated extractions were conducted whenever necessary.
From an initial screening of 20 ISSR primers, seven primers revealed high levels of polymorphisms. Th ese primers generated 85 highly polymorphic fragments that were consistently amplifi ed in repeated experiments conducted on separate dates. Th e GC percentages of the selected primers were within the range of 38.8-66.7%. Th e characteristics as well as the sequences of the primers revealing a polymorphism are shown in Table 2. Th e primer ISSR F3 amplifi ed the highest number of polymorphic fragments (14 bands) and primer ISSR F5 yielded the lowest number of fragments (10 bands). In total, the average number of polymorphic fragments per primer used was roughly 12. A representative fi gure containing ISSR F3 and ISSR F5 banding patterns is given in Figure 1.
A total of 15 Klasea, 1 Serratula, 1 Jurinea and 2 Centaurea taxa were used in the scoring analysis. Th e Jurinea and Centaurea taxa, which were used as the out-group, formed a cluster that was distinct from the Klasea and Serratula cluster in the constructed dendogram. Furthermore the Klasea and the Serratula taxa form clearly separate clusters among themselves (Figure 2).
K. lasiocephala is distinguished within the genus by its very short stems or the absent stems.. K. lasiocephala diff ers morphologically from other Klasea taxa in having absent or reduced stems and that it is also somewhat genetically distinct from other Klasea taxa, as the sole taxon in the cluster in which it is placed. Th e K. erucifolia and K. grandifolia taxa have similar leaf characteristics and were also located in the same sub-cluster owing to their molecular characteristics. K. bornmuelleri taxon does not have a morphologically close relative in the genus. Its position on the dendogram confi rmed this classifi cation. K. cerinthifolia is distinguished by its yellow fl owers and semiamplexicaul leaf structure and was also molecularly identifi ed to be distinct. All these fi ndings were consistent with the morphological classifi cations made in the Flora of Turkey (Davis and Kupicha 1975;Dogan et al. 2012). Martins and Hellwig (2005) showed that Klasea and Serratula taxa to belong to separate clusters in a molecular study conducted using the ITS and ETS sequences. Th e same study reported shorter distances on the dendogram constructed based on molecular similarities for the taxa, which showed morphological similarities.
Th e inspection of the dendogram indicated that molecularly similar taxa were also morphologically similar. Th is separation was also shown in the PCA plot (Figure 3).
Our study has demonstrated that ISSR is a powerful tool in resolving the genetic relationships within problematic taxonomical entities. In conclusion, the morphologically close taxa were, in the molecular aspect, also located in the same clade. Th e genera used as out-groups (Serratula, Jurinea, and Centaurea) were clearly separate from the genus Klasea. According to our knowledge, this is the fi rst report on the use of ISSR in Klasea.