Biologia 66/6: 1003—1010, 2011 Section Botany DOI: 10.2478/s11756-011-0106-2 Population genetic structure of Tunisian Hypericum humifusum assessed by RAPD markers Afef Béjaoui, Abdennacer Boulila,ChokriMessaoud & Mohamed Boussaid* Laboratory of Plant Biotechnology, National Institute of Applied Science and Technology. Centre Urbain Nord, B.P. 676, 1080 Tunis Cedex, Tunisia; e-mail: [email protected] Abstract: The genetic variation within and among seven Tunisian natural Hypericum humifusum L. populations belonging to three bioclimatic zones (sub-humid, upper semi-arid, and lower semi-arid) was assessed using random amplified poly- morphic DNA markers. Eight selected primers produced a total of 166 bands, of which 153 were polymorphic. The genetic diversity within a population, based on Shannon’s index and percentage of polymorphic loci, was relatively high. The level of variation among populations did not differ significantly. However, the variation among populations grouped according to their bioclimates was significant. A high differentiation and a low gene flow were observed at all spatial scales among all populations. The upper semi- arid populations exhibited the highest differentiation. The relationship between genetic and geographic distances was not significant indicating that structuring occurred due to founding events. The UPGMA analysis based on Nei & Li’s coefficients showed that individuals from each population clustered together. The cluster analysis based on genetic distances among populations did not show clear groupings relevant to geographical distances or bioclimates. The high differentiation among populations even through a small geographic range implies the collection of seeds from all populations to preserve, ex-situ, extant variation in the species. Populations from the upper semi-arid zone showing the highest genetic diversity should be first prospected. Key words: Hypericum humifusum; genetic structure; RAPD markers; Tunisia Introduction humid to the upper arid, usually below 300 m a.s.l. It is a diploid species (2n =2x = 16) and belongs to The genus Hypericum (Hypericaceae) is represented the Oligostema section (Carine & Christenhusz 2010; by more than 458 species (trees, shrubs and herbs) Robson 2010; Crockett & Robson 2011). Stems (20–70 subdivided into 36 sections (Robson 2003; Robson cm in height) are glabrous. Leaves are sessile, opposite 2010; Crockett & Robson 2011). Species are distributed with dark glands at the edges. Flowers produced from mainly in temperate regions of the Northern hemi- April to September are bright yellow, and arranged sphere and tropical high elevation habitats (Robson in branched cymes. They are regular with five short 2003; C¨uneyt et al. 2007). Most previous studies on sepals, five petals, and numerous stamens. The gynoe- Hypericum species have focused on ploidy level, breed- cium has 3–4 locules. Fruits are small capsules, dark ing system, genetic diversity (Dlugosch & Parker 2007; purple, producing ∼20 dark brown seeds (∼2–5 mm in Percifield et al. 2007; Poutaraud & Giradin 2008), phy- size). The species reproduces via seeds and shows some logenetic relationships (Robson 2003; Crockett & Rob- capacity to propagate vegetatively. It contains amounts son 2011), chemical composition and biological activ- of hypericin and hyperforin (Tanaka & Takaishi 2006; ity (Meral & Karabay 2002; Conforti et al. 2005). The Smelcerovic et al. 2006), and is used to treat several most studied species is H. perforatum L. known mainly diseases such as wounds and burns (Smelcerovic et al. for its antidepressive, anticancer and antiviral activities 2006). In Tunisia, the species tends to occur in scat- (Gartner et al. 2005; Kubin et al. 2005). tered metapopulations, often with a low size and be- H. humifusum L. is found mainly in wet habi- comes abundant, forming extensive populations, within tats in the North and the Center of Europe, Argentina and at the edges of the northern crop fields. Populations and several Mediterranean regions (Robson 2010). In are heavily harvested and affected by agricultural prac- Tunisia, the species grows widely in the north and in tices (i.e. ploughing, weeding, crop rotation schemes). the center of the country on sandy soils under a rain- The dynamic of populations was influenced by the dis- fall ranging from 350 to 992mm/year (Pottier-Alapetite persion of dry fragments or entire individuals and seeds 1979). It grows in bioclimates extending from the sub- by wind. The fragmentation of populations and site dis- * Corresponding author c 2011 Institute of Botany, Slovak Academy of Sciences 1004 A. Béjaoui et al. Table 1. Location and main ecological traits for the seven Tunisian H. humifusum populations analysed. b Bioclimatic Population Simple Q2 Geographic Altitude Rainfall zonea Site code size (site) region (m) Longitude Latitude (mm/year) Sh Edkhila 1 10 144.31 Kroumiri and Mogods 500 8◦38E36◦50N 992 Touiref 2 10 49.77 High Tell 550 8◦32E36◦33N 466 Triaate 3 10 49.77 High Tell 400 8◦22E36◦28N 466 Usa ◦ ◦ Testour 4 10 65.39 High Tell 450 9 25 E3635 N 585 El Kssour 5 10 49.77 High Tell 400 8◦48E36◦06N 466 Bir Mchergua 6 8 50.09 Sahel 150 9◦55E36◦25N 486 Lsa ◦ ◦ Ain Errahma 7 10 35.24 Sahel 50 10 22 E3613 N 250 a b Bioclimatic zone; Sh – sub-humid; Usa – upper semi-arid and Lsa – lower semi-arid; Q2 – Emberger’s pluviothermic coefficient 2 2 ◦ (1966). Q2 = 2000P/M − m where P is the mean of annual rainfall (mm). M (K ) is the mean of maximal temperatures for the warmest month (July) and m (K◦) is the mean of minimal temperatures for the coldest month (February). P , M and m values were calculated as the average for the period from 1953 to 2003. Bizerte on the amount of genetic diversity maintained in the species to face environmental changes. Tunis Molecular markers such as AFLP, RFLP and ◙ 1 °Le Kef ALGERIA RAPD have been widely used to assess the genetic di- ♦ 2 ♦ 4 Hammamet MEDITERRANEAN SEA versity within and among populations in Hypericum ♦ 3 ▲ 6 species (Arnholdt-Schmidt 2000; Mayo 2003; Percifield ♦ 5 ▲ 7 et al. 2007; Dlugosch & Parker 2007). RAPD mark- ers are well established in population genetics and have proven to be useful tools in the assessment of genetic variation and population structure of plant species. They provide high number of loci and reveal a large Sfax amount of variation with good coverage of the entire genome (Mirali & Nabulsi 2003; Zhang et al. 2005; Gabès Sreekumar et al. 2006). They have been found to be use- ful in differentiating medicinal plant populations, and in establishing phylogenetic relationships among species (Taylor-Grant & Soliman 1999; Rout 2006). However, RAPD markers are dominant, and thus are less suitable LIBYA for linkage analysis than codominant markers. This dis- advantage could be overcome by using several statisti- cal methods such as the analysis of molecular variance (Excoffier et al. 1992). The aim of this study is to determine the level of genetic diversity within and among Tunisian H. humi- fusum populations sampled from different bioclimatic zones throughout the geographic range of the species. For this purpose RAPD markers were used. This anal- ysis, a complementary to that previously published on the genetic diversity of the same populations using al- lozymes (Béjaoui et al. 2010), should be important for Fig. 1. Map of Tunisia. Geographical distribution of the seven informing decisions about conservation and improve- Tunisian Hypericum humifusum populations analysed. Numbers (1. 2...) – population code. Symbols before numbers indicate the ment strategy. bioclimatic zone: – Sub-humid, –Uppersemi-arid, –Lower semi-arid. Material and methods Surveyed populations and sampling turbance are the main factors causing random genetic Tunisian H. humifusum populations were collected, from drift which enhances the risk of extinction of the species seven sites belonging to the sub-humid, upper semi-arid and lower semi-arid bioclimatic zones according to Emberger’s (Young et al. 1996; van Rossum & Prentice 2004). Thus, Q2 (1966) pluviothermic coefficient (Table 1). Altitudes of conservation strategies taking into account the patterns sites varied from 50 m (population 7 from Ain Errahma) of genetic variation within and among populations are to 550 m (population 2 from Touiref). The average annual crucial to evaluate the actual status of populations. rainfall ranged from 250 mm (population 7 from Ain Er- The capacity of these populations to evolve will depend rahma) to 992 mm (population 1 of Edkhila). The popu- Genetic diversity of Hypericum humifusum 1005 lation size ranged from 100 (population of Ain Errahma) also was used to estimate average diversity within all pop- to > 1000 individuals (populations of Twiref and Triaate). ulations HPOP [HPOP =(1/nΣH )wheren is the number Almost all populations (2–7) growing in suboptimal habi- of populations], and the total diversity among all individu- tats for the species suffered some degree of human impact, als within the species HSP [HSP = −Σps log2 ps where ps is (e.g., ploughing, weeding, harvesting) and collected within the frequency of RAPD fragments in the entire sample]. The Triticum durum and T. aestivum fields after harvesting, and proportion of diversity within populations is HPOP/HSP and in roadsides near these fields. The main ecological traits of the proportion of diversity among populations is GST [GST sites are reported in Table 1 and Fig. 1. Eight to ten plants =(HSP−HPOP)/HSP]. The components of variability at the were sampled at distances exceeding 50 m to avoid the sam- ecological group level were estimated by HEG (the diversity pling of closely related individuals. within a group), HEG/HSPG (The proportion of diversity within ecological groups), and GSTG (The proportion of di- RAPD procedure versity among ecological groups). DNA extraction The different indices were calculated by the POP- For each plant, DNA was isolated from young leaves (0.3g) GENE computer package (Yeh et al. 1999). The compari- using 2 mL CTAB extraction buffer and 50 mg PVP son, among Shannon’s diversity indices for populations, and 40000.