INVESTIGATION OF THE GENETIC STRUCTURE OF SOME ANATOLIAN Achillea L . ( ANTHEMIDEAE , ASTERACEAE ) POPULATIONS USING THE ISSR MARKERS

Due to high level of hybridization and polyploidy, the perennial and allogamic Achillea L. genus with a complex phyletic structure has about 142 species widely distributed over in the Northern Hemisphere. The genus is widely distributed in Turkey with 48 species half of which are endemic. The gene and diversification center of the Santolinoideae section including 18 endemic species is thought to be in Anatolia. There exist no comprehensive molecular phylogenetic study on Achillea whose morphological revision in Turkey was completed in 2006. In the study, phylogenetic analyzes were performed using 10 different oligonucleotides for amplification of ISSR bands based on 74 samples of 18 species from 3 sections of Achillea genus. All the oligonucleotides analyzed were found to be polymorphic. The total number of loci is 135 and 44 (33%) of them are parsimony informative. Serious topological differences showing that Achillea taxa include both monophyletic and polyphyletic lineages were revealed in phylogenetic trees obtained under UPGMA, MP and BI methods. Sections were not clearly separated in trees with clear species separations. The results of UPGMA and MP analyses showed that A. vermicularis Trin. was placed as the outgroup while A. sipirokorensis Hausskn. & Bornm. and A. sintenisii Hub.-Mor. formed the outgroup together in Bayesian Inference analysis (BI). The obtained clusters of PCA based on binary genetic distance values were consistent with the result of BI analysis. Molecular variation analysis showed that almost all of the molecular variation was completely resulted from variations within populations.


Introduction
Turkey has a rich flora because of its geological features, soil types and climate conditions in addition to the fact that it is located at the intersection point of Asian, European and African continents.It has a moving geological structure formed by the closure of the Tethyan Sea and it played an important role during the glacial periods The presence of charecteristics of the Mediterranean, Euro-Siberian and Irano-Turanian plant geographical regions in the country is the most important factor increasing species diversity.The flora of Turkey includes about 9222 vascular plant species of which 138 are cultured.Turkey is also an important gene centre 2 E. Bağda (%33.27 with 2991 species) with a high endemism rate (Arabacı 2006).However, despite the floral richness and high endemism rate in the country thenumber of molecular phylogenetic studies on floral members is limited.
The first work on Turkey's flora is "Flora Orientalis" written by Geneva's famous botanist Edmond Boissier in 1867-1888, and the most comprehensive work, the book "Flora of Turkey and the East Aegean Islands", was written by P. H. Davis.This book consists of 10 volumes together with additional volume.A second additional volume with an increase in Turkish flora studies was also added to this book (Güner et al. 2000).
Studies on Anatolian flora have gained a pronounced acceleration recently and new taxa have been identified and/or the available taxonomic groups have been redesigned in studies carried out with numerous samples collected during intensive floristic studies.Revision studies have been increasingly carried out, particularly at genus level, to solve the existing taxonomic problems.The identification of new taxa, the determination of species boundaries and the rewriting of species keys are important consequences of these studies.On the other hand, complex phyletic relationships that are frequently encountered, especially due to high hybridization and polyploidy rates weaken the solving power of morphological revision studies.Molecular systematic approaches are often preferred to overcome these problems with the special aim of revealing evolutionary relationships.
Members of the family Asteraceae (Compositae) are composed of 24080 species distributed in 1545 genera belonging to 21 tribes and three subfamilies.Most of the species in the family are members of the subfamily Asteroideae in which 12 tribes, 1176 genera and about 17025 species are gathered (Arabacı 2006).In the Flora of Turkey, Asteraceae is represented by 11 tribes, 136 genera and 1195 species and is the richest family of flora in terms of both species and genus levels.The family also includes most of the endemic species (endemism rate is %37.3 with 446 species) of the country (Arabacı 2006).
The genus Achillea L. is represented with 142 species from the Anthemideae tribe of Asteroideae subfamily and is one of the most recently evolved genus of the family (Arabacı 2006, Rahimmalek et al. 2009).Members of the genus can grow in almost all habitat types, from the sea level up to altitudes of 3000m a.s.l., mainly in the temperate zone.It is characterized by perennial and allogamic plants adapted to various ecological environments ranging from deserts to sea shores, steady snowy hills and rocky habitats.(Guo et al. 2004).The genus is widely distributed in Europe and West Asia, but it is represented with several species in North America, Australia, New Zealand and North Africa (Rechinger 1963).High biodiversity and natural hybrids in Achillea make it difficult to identify plant samples.Molecular markers are powerful alternative solutions especially for systematic problems which cannot be solved by morphological approaches.In this study, we aimed to contribute to the information about the genus systematic by investigating the biogenetic loci and genetic structures and phylogenetic relationships of 18 cultivars (10 endemic) of Achillea L (Asteraceae) genus spreading in and around Sivas using ISSR markers (Zietkiewicz 1994).

Materials and Methods
18 Achillea species collected during the period from 2011 to 2012 were used as the study material (Fig. 1, Table 1).All collected specimens are kept in the herbarium of the Cumhuriyet University Faculty of Science (CÜFH).Air dried samples were used for DNA isolation.The total DNA isolations were done using 74 different specimens of 18 species distributed in 59 populations and ISSR-PCR was run for all samples.

Determination of Quality and Quantity of Genomic DNA
The quality and quantitation of DNA after isolation were determined by both agarose gel (1%) electrophoresis technique and spectrophotometric measurements at 260 and 280nm wavelengths.The quality and quantity of the DNA samples were estimated by electrophoresis band pattern and by comparing with DNA marker.

Amplification of ISSR Fragment by PCR
ISSR fragments were amplified by PCR using 10 different oligonucleotides (Table 2).

Analysis of ISSR Data
The phylogenies of Achillea populations were investigated using different algorithms.Analyzes were performed on three different approaches; DNA distance, maximum parsimony (MP) and Bayesian inference.Analyzes were evaluated according to the band profiles obtained from the ISSR markers.The presence of the bands is indicated by "1" and the absence of bands by "0".The presence of the band represents a dominant, non-existent recessive phenotype.Because ISSR markers are dominant, the genotype and allele frequencies can not be calculated since the alleles in the same locus can not be distinguished.Therefore, ISSR data is calculated based on the ratio of bands that are common to any locust to all bands.The ratios of existing bands (1) or non-existing bands (0) and common bands were used for the calculations.All these calculations are based on the assumption that the bands moving in the gel for the same distance (Rf), that is, of the same size, are similar.In fact, bands of similar length are directly proportional to the kinship grades of the compared individuals.So, we can say that individuals with more common bands are closer and those with less common bands are farther away.The obtained trees were drawn using FigTree v1.3.1 (Rambaut 2009).
Analysis of the obtained data was performed using a computer program called Popgene 32 (version 1.3.1)(Yeh et al. 1999).Assuming that the populations were in the Hardy-Weinberg equilibrium, the genetic distance (Fst) values between population pairs were calculated to determine population differentiation (Nei 1972).
The unweighted pair-group method with arithmetic mean (UPGMA) dendograms (10000 replicates) were generated by varying the Fst values of the Neighbor-Joining (Saitou & Nei 1987) procedure using the same analysis program.PCA (Principal Component Analysis) was used in the GenAlEx 6.3 package program (Peakall & Smouse 2006) to create a visual representation of the genetic relationship between populations.
Hierarchical analysis of the molecular variation (AMOVA) in the genetic construction of Achillea populations was performed using the Arlequin 3.11 (Excoffier et al. 2005) program using clusters obtained from biogeographic areas, phylogenetic and principal component analyzes.The significance ratings of fixation indices determined by AMOVA were determined by testing with 1000 proposed permutations given by Excoffier et al. (1992).
The MP analysis of the data sets was performed by applying the tree bisection-reconnection (TBR) branch swapping, random addition sequence replicates and 50% majority rule using PAUP * 4.0 beta 10 (Swofford 2002) with heuristic search algorithm.Character states were unordered and unweighted.The bootstrap values of the branches were investigated using heuristic search with 1000 bootstrap replicates.

E. Bağda
In MP analysis, A. vermicularis population was determined as the outgroup from UPGMA dendogram and consensus tree was formed.In this way, the most reliable tree was obtained by re-establishing the branches in the consensus tree.Genetic distance and parsinomic results were compared and common and non-common points were evaluated.
Phylogenetic analyses based on Bayesian and Markov Chain Monte Carlo (MCMC) were carried out using the program MrBayes 3.1.2(Huelsenbeck & Ronquist 2001, Ronquist & Huelsenbeck 2003).Analyzes were performed by presence or absence of markers in the populations (field analysis).In the ISSR analysis, "nst=1" and the rate is "rates=equal" was used.For the generation of 10 7 , two independent runs of four chains (3 heated and 1 cold chain) were carried out and the trees were sampled every 1000 cycles.Convergence on stationary distribution was verified by checking whether the mean standard deviation of the separation frequencies is less than 0.05 between two independent executions.Bayesian posterior probabilities were estimated by constructing a Majority-Rule Consensus Tree among the last 750 sampled trees (25% of the samples, ie, 250 samples were pre-tested or burn-in removed).

Results
Amplifications of ISSR markers were performed using 10 oligonucleotides from 74 samples of 18 species distributed in 3 sections of the Achillea genus (Table 3).All of the oligonucleotides analyzed were found to be polymorphic.The total number of loci was 135 (at most 24, at least 9 bands) and it was determined that 50 of them were fixed, 41 of them were not informative and 44 of them were parsinomically informative.The smallest (0.0676) genetic distance was found between A. schischkinii and A. coarctata populations and the largest genetic distance (0.8149) was found (Nei 1972) between A. sipikorensis and A. wilhelmsii populations (Table 4).
It was determined from the UPGMA dendograms based on the genetic distance values between the populations (Fig. 3) that the Anatolian Achillea populations were monophyletic and A. sipikorensis was the sister group and A. vermicularis was the outgroup.

E. Bağda
of the total genetic variation.The total variation ratio of the primary components to the first and third axes is 48.68%, and the total variation ratio of the axes 2 and 3 is 40.81%.While A. sipikorensis and A. sintenisii were located differently in both distributions, no significant distribution was observed in other populations.This distribution model is found compatible with the results of Bayesian analysis.

Discussion and Conclusion
Determination of the phylogenetic relationships of the genus Achillea, which has a complex phyletic structure due to hybrid and polyploidy frequency is problematic (Guo et al. 2004, Guo et al. 2012).Morphological revision of Turkey's Achillea taxa was conducted by Arabacı (2006) but molecular phylogenetic studies on the genus are limited and they were mostly based on inadequate sample size.Therefore, the present study is the most comprehensive molecular phylogenetic study so far on Anatolian Achillea species and aimed to contribute to the systematic information using ISSR markers of 18 species from three sections.
Turkey is an important evolutionary unit for the genus Achillea (ESU) considering the fact that 46 species of Achillea which constitute 1/3 of the total number of the species within the genus are found in Turkey and half of these species are endemic.Due to this evolutionary importance, species diversity needs to be well investigated.The Santolinoideae section has 38 species.Of these, 16 are endemic and 24 are located in Turkey.This indicates that the gene and the center of change of the Santolinoideae section is Anatolia Guo et al. (2004) suggested based on the results of ITS and trnL-F sequence analysis that the ancestral section may be Santolinoideae, providing an interesting detail for the origin of the genus.A. teretifolia and A. wilhelmsii, both belonging to the Santolinoideae section, were located in the ancestral clades in the study of Guo et al. (2004).The ISSR data obtained in the present study supports the Santolinoideae section as an outgroup, but A. teretifolia and A. wilhelmsii species were not found in the outgroup.UPGMA and MP support A. vermicularis (Santolinoideae) as an ancestral taxon, while BI results support the external group as A. sintenisii (Santolinoideae) and A. sipikorensis (Arthrolepis).It should be mentioned that, all sections are not included in the present study.
The phylogenetic tree patterns from the phylogenetic analysis of Achillea populations containing a large number of hybrids and polyploidy species/taxa are one of the most likely scenarios to be expected and this pattern has also been obtained from MP and BI cladograms generated from ISSR data.The data given by Guo et al. (2004) also supports a similar polyphyletic story based on analysis of ITS and trnL-F sequence data.Significant major differences were determined between the results of the alternative analysis, but Achillea populations were considered to be monophyletic by the analysis (without the choice of outgroup).Some of the results obtained from the analysis were found to overlap with some information given in the Anatolian revision study of Arabacı (2006).A. armenorum, an endemic species unique to the Berit Mountain, has distributed in rocky areas above 2400m.There is no Achillea species that are similar or closely related to A. armenorum (Arabacı 2006).BI cladogram and PCA, the pattern in which the monophyletic A. sintenisii and A. sipikorensis were evaluated together as an outgroup support the finding that two revised polyploidies are given in the revision.
A. sintenisii and A. sipikorensis which grow between 1200-2000m on gypsiferous peaks are endemic to Irano-Turanian region.They are distributed intensively in steppe and calcareous slopes in Sivas and its vicinity.The species commonly found in gypseous areas are close relatives and the spreading areas are the same or close together.There are significant differences between the populations of A. sipikorensis on the gypsifer bed rock and others (Arabacı 2006).
The ISSR data obtained from the study did not fully support the hypothesis that Anatolian Diagonal may have played an important role in the evolution of Anatolian Achillea populations.In addition to the sympatric speciation expected to be effective in the evolution of the Achillea taxa in Anatolia, the vicariance speciation model needs to be tested again using more extensive samplings.Thereby, looking at the nucleotide sequence variations, AFPL marker or chloroplast will increase the reliability of the result.
The present study is the first study on molecular phylogeny of non-endemic (A.schischkinii, A. vermicularis) and endemic (A.armenorum, A. cappadocica, A. cucullata, A. kotschyi, A. lycaonica, A. magnifica, A. phyrgia, A. sintenisii, A. sipikorensis, A. sivasica) Achillea species in Turkey.In addition, to the best of our knowledge, ISSR markers of 17 species (except A. millefolium) included in the study were obtained for the first time.
In conclusion, ISSR markers provided a comprehensive and significant contribution to the understanding of the genetic diversity and phylogenetic relationship of Achillea taxa in Turkey.The results obtained will help to understand the evolutionary dynamics of Achillea genus.

Fig. 1 .
Fig. 1.Distribution map of Achillea populations used in the study.

Fig. 4 .
Fig. 4. Compatibility tree with 50% majority rule based on MP method.All branches except two branches (95% and 89%) were shown to support 100%.Colored groups stands for sections represented by populations.

Fig. 5 .
Fig. 5. Phylogenetic relationship between populations based on ISSR haplotypes.The cladogram is a majority-based compliance tree based on Bayesian analysis.

Table 1 .
Herbarium numbers (Hrb.No.) and localities of Achillea specimens used in the study.

Table 2 .
Base sequences and annealing temperatures of the ISSR oligonucleotides used for amplification.