2012, vol. 68, 89–100
Imen Ben El Hadj Ali, Arbi Guetat, Mohamed Boussaid
Genetic diversity of wild Thymus capitatus (Lamiaceae) in Tunisia using molecular markers
Received: 02 April 2012; Accepted: 04 June 2012
Abstract: In Tunisia, Thymus capitatus L. populations are severely destroyed due to deforestation and over-col- lecting. The species occurs in small scatteredpopulations decreasingprogressively in size. Yet, no conserva- tion or improvement programs are attemptedto preserve andpromote the potential value of this resource. In this work, we assessedthe genetic diversityof nine Tunisian populations of Thymus capitatus L. from different bioclimates, using 103 polymorphic randomly amplified polymorphic DNA (RAPD) loci. The analysis of the genetic variation within andamong populations is primordialto elaborate conservation andimprovement programs. The species showed a low diversity within populations (0.276 causedby the habitat fragmentation. A high genetic differentiation(G ST=0.359 and FST=0.284) anda low gene flow (0.435 basedon FST distance matrix showed that most populations clustered independently to bioclimate or geo- graphical distance indicating that genetic differentiation mainly occurs at local space scale due to genetic drift. Given the high proportion of the within-population diversity and the high amount of among populations vari- ation, the ex-situ conservation should be made by the collection of seeds/cuttings rather within than among populations. The in-situ conservation should be made appropriately according to populations in their bioclimate. Additional key words: Habitat fragmentation, conservation strategies, genetic structure. Address: I. Ben El Hadj Ali, A. Guetat, M. Boussaid, National Institute of Applied Sciences and Technology. Laboratory of Plant Biotechnology B.P.676, 1080 Tunis Cedex, Tunisia, e-mail: [email protected] Introduction comes from natural populations that are decreasing progressively in number andsize. Studieson thyme Thymus L. genus comprises more than 350 species conservation strategies andpopulation’s genetic growing in wildthroughout the world(Mabberley structure in relationship with the ploidy level, geo- 1997). Species can be diploids (2n=2x=24, 26, 28, graphical origin andfloral biology, remain rather 30, 32), tetraploids (2n=4x=52, 54, 56, 58, 60) or scarce in contrast to those assessing their chemical hexaploids (2n=6x=42, 48, 84, 90) (Mártonfi and composition andbiological properties (Ehlers and Mártonfiová 1996; Morales 1996). A growing interest Thompson 2004; Lopez-Pujol et al. 2004). in the use of these species has been recently reported In Tunisia, the genus Thymus L. is representedby for industrial, pharmaceutical and chemical fields four species: Thymus capitatus (L.) Hoffm. et Link., (Hazzit et al. 2009). However, the usedmaterial often Thymus numidicus Poir., Thymus hirtus Willd. ssp. 90 Imen Ben El Hadj Ali, Arbi Guetat, Mohamed Boussaid algeriensis Boiss. et Reut. var. cinerescens Murb. and among Tunisian populations of Thymus capitatus from Thymus vulgaris L. (Pottier-Alapetite 1981). Species different bioclimatic zones using RAPDs (Random can be sympatric in a wide part of the country (Ben El AmplifiedPolymorphic DNA) (Williams et al. 1993), Hadj Ali et al. 2010; Ben El Hadj Ali et al. 2012). Thy- for information on the genetic variation of the species mus capitatus (L.) Hoffm. et Link. [Synonym Thymbra in order to work out the conservation strategy and im- capitata (L.) Cav., Corydothymus capitatus Rechenb. f.] provement programs. (Figueiredo et al. 2008), is a diploid (2n=2x=30) RAPDs are extensively appliedto assess the ge - (Morales 1996) andoutcrossing shrub predominantly netic diversity and to address conservation and pro- bee-pollinated. It’s a typical plant with the Mediterra- tection purposes in woody plant populations nean region andit’s usedas medicinalherbs. It is con - (Monteleone et al. 2006; Solouki et al. 2008; Zheng et sists of perennial andornamental plant (Le Floc’h al. 2008). They involve the analysis of a large number 1983). The species exhibits both vegetative propaga- of loci, thus providing a more complete evaluation of tion and sexual reproduction (Petanidou and Vokou the genome as comparedwith other biochemical and 1993). However, natural regeneration of genets from molecular markers (Monteleone et al. 2006). RAPDs seeds is extremely rare because of the difficulty of are selectively neutral andable to detecthigh varia - seedling survival, mainly consequent to the low habi- tion both in coding and non-coding DNA regions tat quality (Pérez-Garcia et al. 2003). without prior knowledge of the genome (Allnutt et al. Thymus capitatus L., known in Tunisia under the 2003; Nybom andBartish 2000). Nevertheless, vernacular name of “Zaater”, is usedas herbal teas, RAPDs have same disadvantages, the most signifi- condiment and as a food additive. It’s an aromatic cant being their dominant allelic expression and their plant very important for the horticultural andmedici - low level of reproducibility. Thus, allelic frequencies nal industryandit’s the most important andexpen - estimatedfor loci are less accurate than those ob - sive of all commercial origanum oils (Bauer et al. tainedwith codominantmarkers allowing to bias in 1997; Hedhili et al. 2005). In Mediterranean cooking, the evaluation of the genetic diversity (Nybom 2004). it’s extensively usedfor culinary purposes, for its use However, these disadvantages could be overcome by in flavouring sauces, stews andsoups. In Tunisian the use of appropriate statistical methods such as the folk medicine, the species is used as decoction in case analysis of the molecular variance (Nybom and of colic, ulcer andarterial hypertension (Bouraoui Bartish 2000). Besides, the detection of a high num- 2000). It’s mainly known for its antibacterial, antioxi- ber of polymorphic loci, the analysis of bands with dant, antimycotic and spasmolytic activities frequency higher than 3/N (N was the number of the (Bounatirou et al. 2007). Infusion anddecoctionof analysedplants) andthe reproducibilityof experi- leaves andflowers were usedagainst diarrhoea,diges- ments couldbuffer significantly the bias of estimation tive and respiratory system disorders (Stahl-Biskup of the genetic variation (Hovmalm et al. 2004; Nybom and Sáez 2002). and Bartish 2000). In Tunisia, the species grows mainly in dry and sunny garrigues issued from the degradation of Pinus halepensis L. and Juniperus phoenicea L. primary forests. Material and Methods Populations are locatedin differentbioclimatic zones (extending from the sub-humid to the upper arid) on Surveyed populations and sampling sandy and often on rocky soils, under a rainfall vary- Nine Thymus capitatus L. populations in the whole ing from 300 to 1000 mm/year andat altitudesrang - distribution area of the species in Tunisia were cho- ing from 150 to 500 m (Nabli 1995). The species oc- sen on the basis of previous work using isozyme and curs as mosaic of local populations in little patches chemical markers (Bel Hadj Ali et al. 2006; Ben El along the landscape. The populations have been se- Hadj Ali 2010; Ben El Hadj Ali et al. 2012). They have verely depleted and fragmented due to been collectedfrom differentbioclimatic zones overexploitation andhabitat destructioncausedby (sub-humid, upper semi-arid, mean semi-arid, lower grazing, clearing, low soil quality andirregularity of semi-arid and upper arid) according to Emberger’s Q2 2 2 rainfall. The habitat fragmentation andthe spatial coffecient (1966) [Q2= 2000 P/(M -m ), where P is isolation of populations increase genetic drift and dif- the mean of annual rainfall (mm), M (K) is the mean ferentiation between populations, andreducetheir of maximal temperatures for the warmest month future adaptation to environmental changes (June) andm (K) is the mean of minimum tempera - (Ellstrand and Elam 1993). Thus, understanding the tures for the coldest month (February)]. The alti- patterns of the genetic variation within andamong tudes of sites varied from 100 m to 600 m. The aver- populations is crucial to evaluate the present status of age of the annual rainfall rangedfrom 125 mm to 550 these populations anddevelopconservation pro - mm (Table 1). The soil is calcareous or marno-calcar- grams. The aim of this study is to assess the fragmen- eous. Ten plants, from each population, were ran- tation andthe extent of genetic diversitywithin and domly sampled at distances exceeding 20 m to avoid Genetic diversity of wild Thymus capitatus (Lamiaceae) in Tunisia using molecular markers 91 Table 1. Main ecological traits for the nine Thymus capitatus populations analysed Population Bioclimatic Rainfall Population Latitude Longitude Q coefficient Altitude (m) code zone* 2 (mm/year) 1 El Hairech Jeb. Mt. * Sh 36°10’N 8°4’E 64.52 440 500–600 2 Abderrahmen Jeb. Mt. Sh 36°40’N 10°41’E 65.64 420 500–600 3 Bou Argoub Usa 36°31’N 10°3’E 59.72 100 400–500 4 Mansour Jeb. Mt. Usa 36°17’N 9°36’E 45.72 600 400–500 5 Essers Msa 36°76’N 9°40’E 44.24 604 400–500 6 Goubellat Jeb. Mt. Msa 36°60’N 9°52’E 39.75 300 400–500 7 Khenis Lsa 35°45’N 10°49’E 34.05 300 300–400 8 Aïn Errahma Lsa 36°13’N 10°22’E 33.24 100 300–400 9 Toujene Ua 33°27’N 9°58’E 29.19 600 100–150 Sh: Sub-humid, Usa : Upper semi-arid, Msa: Mean semi-arid, Lsa: Lower semi-arid, Ua: Upper arid. *Jeb. Mt.: Jebel Mountain. = 2000P/M2-m2, where P is the average of an *Bioclimatic zones were defined according to Emberger’s (1966) pluviothermic quotient Q2 - nual rainfall (mm). M is average of maximal temperature (K: Kelvin degree) for the hottest month (June) and m the mean of minimal tem- was calculatedfor each site using P, M andm values for the periodfrom 1953 to 2008 perature (K) for the coldest month (February). Q2 (Tunisian National Institute of Meteorology). the sampling of closely related individuals. Samples (5’TCGTTCCGCA3’), OPJ-08 (5’CATACCGTGG3’), were storedin plastic bags at -80°C before DNA ex - OPJ-10 (5’AAGCCCGAGG3’), OPJ-12 (5’GTCCCG- traction. TGGT3’), OPJ-13 (5’CCACACTACC3’), OPJ-14 (5’CACCCGGATG3’) andOPJ-16 (5’CTGCTTAG- DNA extraction GG3’). Amplification products were separated by DNA was preparedby grinding500 mg of young electrophoresis in 1.5% agarose gels in TAE buffer leaves from each plant in liquidnitrogen, 2 ml of pre- (pH 8), stained with ethidium bromide, and visual- heated CTAB extraction buffer added with 1% PVP izedunderUV light using a DOC PRINT Photo Docu- (40000). Samples were incubatedat 65°C for 1 hour mentation System. Molecular weights were estimated in a water bath. Subsequently, the mixture was using a 200 bp DNA Promega ladder. treatedtwice with 600 µL chloroform-isoamyl alcohol (24:1) andcentrifugedfor 10 min at 12000 rpm. DNA Data analysis precipitation was performedfollowing the method described by Lodhi et al. (1994). The quality of the RAPD markers were scoredas presence (1) or ab - DNA was estimatedon an agarose gel (0.8%) stained sence (0) of bands, and then transformed into a bi- with ethidium bromide. nary matrix. Each marker bandwas assumedas being a single locus. For each primer the percentage of poly- RAPD assay and PCR conditions morphic bands was estimated for each population. PCR was performedin 25 µL reaction volume con - The genetic diversity within a population was esti- taining 50 ng DNA, 5 µL of 5 X reaction buffer, 40 matedusing the percentage of polymorphic bandsP% pmoles of primer, 200 µM of each dNTP, 2.5 mM [(number of polymorphic bands/total number of MgCl2 and1.5 U Taq polymerase (Promega). The bands) x 100] and Shannon’s index (H’) for each mixture was overlaidwith one dropof mineral oil. RAPD locus [H’=- Spi log2pi, where pi is the frequency Reactions were performedin Stuart Thermal Cycler of RAPD bandin i th population]. Shannon’s index was (Maxi-Gene) programmedfor an initial denaturation also usedto estimate the average diversityH pop over step of 94°C for 2 min, followedby 45 cycles of 30 s at all populations [Hpop=-1/nSH’; where n is the number 94°C, 1 min at 36°C (annealing step), and2 min at of populations]. The species diversity was estimated 72°C (elongation step). An additional 10 min period as Hsp [Hsp=-S pslog2 ps; where ps is the frequency of for elongation at 72°C followedthis cycle. To test presence or absence of the RAPD in the whole sam- reproducibility between and within runs, DNA from ple]. The proportion of diversity within populations same two plants was included in every PCR run. A was estimatedas H pop/Hsp, that among populations negative control without DNA was also usedin every was evaluatedby G ST [GST=(Hsp–Hpop)/ Hsp]. Shan- PCR run. Eighteen RAPD 10-mer oligonucleotide non’s diversity indices and GST were also usedto par - primers (kit OPJ, Promega) were screened. After opti- tition the variation within andamong ecological mizing the PCR conditions, seven RAPD primers groups (each ecological group includes populations were selected according to consistency, reproducibi- from the same bioclimate). The different indices were lity andpolymorphism level of electrophoretic bands calculatedby the POPGENE computer package soft - to amplify DNAs. Primers selectedare: OPJ-06 ware (Yeh et al. 1999). The comparison among Shan- 92 Imen Ben El Hadj Ali, Arbi Guetat, Mohamed Boussaid non’s diversity indices at the population and ecologi- age of polymorphic loci per primer variedfrom cal group levels was performedusing a variance analy - 64.71% (OPJ-14) to 100% (OPJ-08). Specific bands sis (ANOVA procedure, SAS 1990) and Duncan’s test were revealedaccordingto populations and (Dagnelie 1975). bioclimates. For example, bands 340 bp (OPJ-06), The partitioning of the genetic variation within 460 bp (OPJ-08) and400 bp (OPJ-12) were re - andamong populations or within andamong ecologi - strictedto populations from the sub-humidandthe cal groups, besides Hpop/Hsp andG ST estimates, was mean semi-aridzones (populations 1, 2, 5 and6). evaluatedby AMOVA using WINAMOVA program, Bands 700 bp and 1700 bp (OPJ-16) were only ob- version 1.55 (Excoffier et al. 1992). F-statistics: FST servedin populations from the sub-humidandthe (differentiation among populations), FCT (differenti- upper aridbioclimates (populations 1, 2 and9). The ation among ecological groups) and FSC (differentia- band650 bp (OPJ-12), which was foundto be widely tion among populations within groups) were calcu- represented,is not foundin the populations 1 and2 lated. The significance of variance components and from the sub-humidarea. Bands1000 bp (OPJ-08) that of F-statistics were estimatedusing permutation and1400 bp (OPJ-16) were not observedin the lower procedures (NTSYS-pc, version 2.0, Rohlf 1998). The semi-arid populations 7 and 8. gene flow (Nm) between populations was estimated The number anddistributionof polymorphic prod - as Nm = [(1/FST)–1]/4 (Wright, 1951). The correla- ucts detected with each primer in each population or tions between the matrices of genetic distances (FST) ecological group were given in Table 2. At the popula- andgeographic distances(Km), FST andaltitudesor tion level, the percentage of polymorphic bands for FST andEmberger’s Q 2 pluviothermic coefficients each primer variedfrom 30.77% to 86.36%. Primers (Emberger 1966) was estimatedby a Mantel test differed in their ability to distinguish individuals (Mantel 1967) using ZT program (Bonnet andVan de within andamong populations. OPJ-08 followedby Peer 2002). The significance of the correlation was OPJ-10 andOPJ-14 revealedthe highest percentage of tested after 1000 permutations. polymorphism for all populations (63.64%, 58.89% The genetic similarity between individuals was es- and55.53% respectively), while OPJ-06 andOPJ-16 timatedusing the Nei andLi’s (1979) similarity coef- identified only 42.74 and 48.15% of polymorphism. ficient Sxy [Sxy =2mxy/(mx+my), where mxy is the The highest number of polymorphic fragments within number of bands shared by samples x and y, mx and populations was scoredfor OPJ-08 (populations 1, 4, my are the number of bands in samples x and y, re- 7, 8 and9), OPJ-10 (population 6), OPJ-13 (popula- spectively]. The genetic distance (Dxy) between indi- tions 3 and5) andOPJ-16 (population 2). The aver- viduals was estimated using the complementary value age percentage of the polymorphic loci detected with Sxy [Dxy =1-Sxy]. A Neighbour-joining tree (Saitou and all primer rangedfrom 61.16% (population 1) to Nei 1987), basedon Nei andLi’s distancematrix be - 50.41% (population 3) with an average of 69.26%. tween individuals, was constructed to ordinate rela- The percentage of polymorphic loci scoredin each tionships among individuals and construction of ecological group over all primers variedfrom 53.72% phylogenetic trees using the Win95/98/NT program (Upper aridzone) to 76.86% (Sub-humidzone). FreeTree (Hampl et al. 2001). Support values of the OPJ-16, OPJ-12, OPJ-13, OPJ-10 andOPJ-08 are char - internal branches of NJ were evaluatedthrough boot - acterizedby being more efficient in the detectionof strap method(1000 replicates) (Hampl et al. 2001). A polymorphism according to bioclimates. PCO basedon Nei andLi’s genetic similarity was per - Shannon’s diversity index (H’) estimating the formed to ordinate the relationship among all individ- within-population variability rangedfrom 0.276 uals using the MVSP program (version 3.1). UPGMA (population 3 from the upper semi-aridzone) to dendrogram analysis displaying the relationship 0.327 (population 8 from the lower semi-aridzone). among populations was also produced using pairwise However, H’ values did not show significant differ- FST matrix among populations. ences among populations (ANOVA test, p>0.05). The average of H’ value for all populations was 0.303 indicating a low level of the within-population ge- Results netic diversity (Table 3). At the species level, the Shannon index was moderate (H =0.473). The The genetic diversity sp within-ecological group diversity (H’gp) variedfrom The seven primers usedfor all populations gener - 0.296 (Upper aridzone) to 0.417 (Mean semi-arid ated121 discernibleandreproducibleDNA frag - zone) with an average of 0.384. The average of ments, out of which 103 (85.12 %) were polymor- within-group diversity did not differ statistically phic. The bands ranged in size from 200 to 2000 bp. among groups (ANOVA; p>0.05). The proportions The number of bands produced by primer varied of within-populations (Hpop/Hsp=0.641) and from 13 (OPJ-06 andOPJ-13) to 22 (OPJ-08), with within-ecological groups (Hgp/Hsp=0.791) genetic di- an average of 17.28 at the species level. The percent- versity were high. Genetic diversity of wild Thymus capitatus (Lamiaceae) in Tunisia using molecular markers 93 Table 2. Percentage of polymorphic loci (Pt%) andnumber of polymorphic bands(in parenthesis) per primer in each popu - lation and bioclimatic zone Primer Population Pt% OPJ06 OPJ08 OPJ10 OPJ12 OPJ13 OPJ14 OPJ16 1 53.85 (7) 86.36 (19) 50 (10) 53.33 (8) 61.54 (8) 47.06 (8) 66.67 (14) 61.16 2 53.85 (7) 59.09 (13) 45 (9) 53.33 (8) 38.46 (5) 58.82 (10) 76.19 (16) 56.20 3 38.46 (3) 40.91 (9) 60 (12) 53.33 (8) 69.23 (9) 47.06 (8) 47.62 (10) 50.41 4 46.15 (6) 72.73 (16) 35 (7) 53.33 (8) 61.54 (8) 58.82 (10) 38.10 (8) 52.07 5 53.85 (7) 63.64 (14) 50 (10) 46.67 (7) 69.23 (9) 58.82 (10) 42.86 (9) 54.55 6 38.46 (5) 54.55 (12) 85 (17) 60.00 (9) 38.46 (5) 58.82 (10) 28.57 (6) 52.89 7 23.08 (3) 68.18 (15) 60 (12) 46.67 (7) 38.46 (5) 64.71 (11) 42.86 (9) 51.24 8 30.77 (4) 59.09 (13) 80 (16) 53.33 (8) 46.15 (6) 52.94 (9) 42.86 (9) 53.72 9 46.15 (6) 68.18 (15) 65 (13) 53.33 (8) 30.77 (4) 52.94 (9) 47.62 (10) 53.72 Average 42.74 63.64 58.89 52.59 50.43 55.55 48.15 69.26 Bioclimatic zone Sh 84.62 (11) 86.36 (19) 60.00 (12) 66.67 (10) 84.62 (11) 64.71 (11) 90.48 (19) 76.86 Usa 53.85 (7) 81.82 (18) 75.00 (15) 86.67 (13) 76.92 (11) 70.59 (12) 76.19 (16) 75.21 Msa 69.23 (9) 68.18 (15) 90.00 (18) 80.00 (12) 92.31 (12) 64.71 (11) 57.14 (12) 73.55 Lsa 46.15 (6) 81.82 (18) 85.00 (17) 66.67 (10) 46.15 (6) 64.71 (11) 61.90 (13) 66.94 Ua 46.15 (6) 68.18 (15) 65.00 (13) 53.33 (8) 30.77 (4) 52.94 (9) 47.62 (10) 53.72 Average 42.74 63.64 58.89 52.59 50.43 55.55 48.15 69.26 Number of polymorphic bands are given in parentheses; Pt%: Percentage of polymorphic loci Table 3. Shannon’s index and ratio of genetic diversity in each population and each ecological group Population 123456789 H’ 0.321 a 0.309 a 0.276 a 0.302 a 0.314 a 0.294 a 0.291 a 0.327 a 0.296 a Hpop 0.303 ns Hsp 0.473 Hpop/Hsp 0.641 GST 0.359 Ecological group Sh Usa Msa Lsa Ua H’gp 0.408 a 0.412 a 0.417 a 0.389 a 0.296 a Hgp 0.384 ns Hsp 0.486 Hgp/Hsp 0.791 GSTg 0.209 Population codes as identified in Table 1. Sh. Sub-humid, Usa. Upper semi-arid, Msa. Mean semi-arid, Lsa. Lower semi-arid, Ua. Upper arid, La. Lower arid. Numbers with the same letter are not significantly different (Duncun’s test, p> 0.05). H’, H ,Hsp,H /H andG are the average per primer values of genetic diversity for each primer within each population (H’), over all popu- pop pop sp ST lations (H ), whole sample (Hsp) andtheir partition within- (H /H ) andbetween-populations (G ) components respectively, calculated pop pop sp ST for all primers. H’ ,H ,H ,H /H andG are the average per primer values of genetic diversity for each primer within each ecological gp gp sp gp sp STg group (H’ ), over all groups (H ), andtheir partition within- (H /H ) andbetween-groups (G ) respectively, calculatedfor all primers. gp grp gp sp STg F value among all populations was 0.284, andall The genetic structure and divergence ST pairwise F values were significantly different from among populations ST zero (Table 5). The highest FST values were found, re- The GST value estimating the genetic structure spectively, between populations 3 and4 ( FST=0.365), among populations was 0.359 andamong groups which of 90 Km distant and between the populations (GSTg) was 0.209 (Table 3). The within-population 3 and8 ( FST=0.353), which of 40 Km apart. The low- component of variance estimatedthrough AMOVA est FST value (0.205) was foundamong the popula - accountedfor 71.61% of overall variation, while that tions 7 and8, separatedby 50 km. The effective num - among populations was 28.39% (Table 4). The mean ber of migrants among populations is low 94 Imen Ben El Hadj Ali, Arbi Guetat, Mohamed Boussaid Table 4. Analysis of molecular variance within andamong Thymus capitatus populations andwithin andamong ecological groups Source of variation d.f. Msd Variance Total variance (%) F-statistics Population Among populations 8 65.10 5.25 28.39 FST = 0.284** Within populations 80 13.23 13.23 71.61 Total 88 Ecological group ns Among groups 4 68.26 0.35 1.89 FCT = 0.019 Among populations/group 4 61.93 4.93 26.65 FSC = 0.272** Within populations 80 13.23 13.23 71.46 Total 88 d.f.: degree of freedom; Msd: Mean squared; ns: not significant; **: significant at p<0.001 after 1000 permutations F : differentiation among populations, F : differentiation among ecological groups, F : differentiation among populations within groups. ST CT SC (0.435 (FCT=0.019). The dendrogram generated from pairwise FST ma- Neighbour-joining dendrogram generated from trix among populations showedthree distinctgroups Nei and Li’s genetic distances between all individuals without evident relationship to bioclimate and/or showed that the majority of individuals from the geographic distance (except for populations 7 and 8) same population groupedtogether andpopulations (Fig. 3). The first group (G1) includes populations 3 from the same bioclimate did not always clustered and5 from the upper semi-aridandmean semi-arid closely (Fig. 1). The PCO basedon Nei andLi’s simi - bioclimates, respectively. The population 4 from the larity matrix for all individuals showed that the first upper semi-aridzone constitutedthe secondgroup three principal axes accountedfor 19.9% of the total (G2). The thirdgroup (G3), which couldbe subdi - variation. The plot according to the first two PCO vided into two subclusters includes the lower axes (14.26%), revealedthree major population semi-aridpopulations 7, 8 andthe upper aridpopula - groups (Fig. 2). The first one, projectedat the nega - tion 9 (subcluster 1) andthe populations 1, 2, 6 from tive side of axes 1 and 2, is represented by individuals the sub-humidandmean semi-aridbioclimates, re - belonging to populations Essers (population 5, Mean spectively (subcluster 2). semi-aridzone) andBou Argoub (population 3, Up - Table 5. Pairwise gene flow (Nm) (above diagonal) and FST values (below diagonal) among populations of Thymus capitatus Population 123456789 1 0.796 0.605 0.534 0.744 0.674 0.737 0.717 0.720 2 0.239** 0.777 0.671 0.676 0.836 0.591 0.565 0.592 3 0.292** 0.243** 0.435 0.850 0.505 0.527 0.450 0.516 4 0.319** 0.271** 0.365** 0.547 0.593 0.555 0.562 0.579 5 0.252** 0.270** 0.227** 0.314** 0.704 0.479 0.552 0.549 6 0.271** 0.230** 0.331** 0.297** 0.262** 0.681 0.802 0.816 7 0.253** 0.297** 0.322** 0.311** 0.343** 0.269** 0.970 0.710 8 0.259** 0.307** 0.353** 0.308** 0.312** 0.238** 0.205** 0.828 9 0.258** 0.297** 0.326** 0.302** 0.313** 0.235** 0.260** 0.232** 1, 2, …: population code; **Highly significant at p < 0.001 (after 1000 permutations). Genetic diversity of wild Thymus capitatus (Lamiaceae) in Tunisia using molecular markers 95 Fig. 1. Neighbour-joining dendrogram generated for 90 individuals of Thymus capitatus analysedusing Nei andLi’s genetic distances Population code: Numbers (1, 2, …, 9). 96 Imen Ben El Hadj Ali, Arbi Guetat, Mohamed Boussaid Fig. 2. Principal Coordinates Analysis (PCO) generated from Nei and Li’s similarity coefficient for the 90 individuals ana- lysed of Thymus capitatus. Plot according to axes 1–2. Population code: Numbers (1, 2,…, 9). Fig. 3. UPGMA dendrogram of pair-wise FST values derived from AMOVA of RAPD profiles Population code: Numbers (1, 2,…, 9). Discussion couldbe bufferedby the detectionof a high number of loci (Aagaardet al. 1998) andthe use of appropriate Our results basedon RAPD markers revealeda low genetic population parameters such as FST which re- level of variation within populations andhigh differ - duces the bias in estimating the genetic variation. In entiation among them. The analysis of population ge- our study, the seven selected primers revealed 121 netic variation with RAPDs couldbe hamperedby a loci, of which 103 were polymorphic. This large num- loss of a part of genetic information (Kremer et al. ber of loci couldbe consideredsufficientto compen - 2005; Nybom 2004). However, this disadvantage sate loss of genetic information content at loci. Genetic diversity of wild Thymus capitatus (Lamiaceae) in Tunisia using molecular markers 97 Thymus capitatus populations maintain a moderate tween populations 3 and8 (40 Km), belonging to the genetic diversity (Hsp=0.473). It couldbe explained upper semi-aridandthe lower semi-aridbioclimates, by the outcrossing system associatedprobably with respectively. Besides, the highest Nm value and the the number of initial founders that differed geneti- lowest FST value were notedbetween populations 7 cally in a population (Dittbrenner et al. 2005; Thomp- and8 (50 Km), belonging to the lower semi-arid son 1999, 2002). Additionally, populations of Thymus bioclimate. The absence of correlation between popu- capitatus analysedwere fragmenteddueintensive for - lation genetic structure andgeographic distancessug - est clearing, grazing andoccur in small scattered gests the existence of important historical gene flow patches. The level of the genetic diversity may result between populations. Therefore, the current scatte- from loss of genetic variation through inbreeding reddistributionarea of Tunisian T. capitatus andthe and/or genetic drift caused by the restriction of the high differentiation among populations is probably species to small and degraded populations. However, recent due mainly to anthropic pressures. the outbreeding mating system and the persistence of Shannon’s index indicates that the genetic diver- multiple individuals through generations issued from sity in T. capitatus was distributed within rather than large populations before fragmentation might be the among populations (Hpop/Hsp=0.641). AMOVA con- main contributing factors of this high level of ducted at the population and ecological group levels intrapopulational variation (Hpop/Hsp=0.641). The showedalso that most of the total variation was level of average within-population gene diversity was found among individuals within populations or relatively lower comparedto values derivedfrom within populations in their corresponding group RAPDs for other Lamiaceae plants with an outcross- (71.61%; 71.46%). This suggests that mating occurs ing mating system (Boulila et al. 2009; Verma et al. mainly among individuals within a sub-population 2007). thus favouring the divergence between populations. In our work several RAPD loci appear to be specific The PCO analysis showedthat populations plotted to populations according to the biocliamte (i.e. as a broadscatter, andgatheredoften without evident OPJ06-5, OPJ08-8, OPJ12-3, OPJ16-12 andOPJ16-21 relationship to the bioclimate with the exception of in the Sub-humidpopulations 1 and2, OPJ08-8 and populations 7 and8 (Lower semi-aridzone). Results OPJ12-3 in populations 5 and6 from the Mean from the UPGMA dendrogram also revealed a differ- semi-arid,andOPJ16-12 andOPJ16-21 in the Upper entiation among populations independently to aridpopulation 9). However, the detectionof these bioclimate exceptedfor populations 7 and8. Several loci might not reflect adaptability to ecological factors populations from continuous bioclimates (popula- since they were not detected in all populations from tions 3 and5) gatheredtogether, anda high distinc- the same bioclimate. Besides, the variation of RAPDs, tion between the two mean semi-aridpopulations 5 may not be a reliable measure of adaptive potential and 6 was revealed indicating that genetic differentia- (Volis et al. 2005). These markers were believedto re - tion mainly occurs at local space scale due to limited veal variation both in coding and non-coding genes gene flow and genetic drift. (Castro-Félix et al. 2008; Nybom andBartish 2000). Thymus capitatus occurs in small scatteredpopula - Thymus capitatus populations were highly differenti- tions due to habitat fragmentation. The within-popu- ated(G ST=0.359; FST=0.284). The observedamount lations genetic diversity is low and the among-popu- of differentiation was higher than the average for pe- lations differentiation is high (GST=0.359 and rennial outcrossing species (GST=0.22, FST=0.27) FST=0.284). These processes may increase genetic (Hamrick andGodt1996; Nybom 2004; Nybom and drift. Thus, efforts to conserve populations are ur- Bartish 2000). This high differentiation could be ex- gently needed. Similar patterns of genetic diversity as plained by genetic drift due to limited gene flow estimated by the Shannon’s index and no differences (0.435 et Zoologie, University of Sciences Montpellier 7: Mártonfi P., Mártonfiová L. 1996. Thymus chromo- 1–43. some numbers from Carpathians andPannonia. Excoffier L., Smouse P.E., Quattro J.M. 1992. Analy- Thaiszia Journal of Botany 6:25–38. sis of molecular variance inferredfrom metric dis - Monteleone I., Ferrazzini D., Belletti P. 2006. Effec- tances among DNA haplotypes: application to tiveness of Neutral RAPD Markers to Detect Ge- human mitochondrial DNA restriction data. Ge- netic Divergence between the Subspecies uncinata netics 131: 479–491. and mugo of Pinus mugo Turra. Silva Fennica 40: Figueiredo A.C., Barroso J.G., Pedro L.G., Salgueiro 391–406. L., Miguel M.G., Faleiro M.L. 2008. Portuguese Morales R. 1996. Studies on the genus Thymus L. Thymbra and Thymus Species Volatiles: Chemical Lamiales Newsletters 4: 6–8. Composition andBiological Activities. Curium Nabli M.A. 1995. Essai de synthèse sur la végétation Pharmaceutical Design 14: 3120–3140. et la phyto-écologie tunisiennes II et III: le milieu Hampl V., Pavlcek A., Flegr J. 2001. Construction and physique et la végétation. Ecologie végétale bootstrap analysis of DNA ?ngerprinting-based appliquée (I.O.R.) Tunisie. phylogenetic trees with the freeware program Nei M., Li W.H. 1979. Mathematical model for study- FreeTree: application to trichomonadparasites. ing genetic variation in terms of restriction International Journal of Systematic, Evolution endonucleases. Proceedings of the National and Microbiology 51: 731–735. Academy of Sciences of the United States of Hamrick J.L., Godt M.J.W. 1996. Effects of life history America 76:5269–5273. traits on genetic diversity in plant species. Philo Nybom H., Bartish I.V. 2000. Effects of life history Trans Roy Soc London B 35:1291–1298. traits andsampling strategies on genetic diversity Hazzit M., Baaliouamer A., Veríssimo A.R., Faleiro estimates obtainedwith RAPD markers in plants. M.L., Miguel M.G. 2009. Chemical composition Perspectives in Plant Ecology, Evolution andSys- andbiological activities of Algerian Thymus oils. tematics 3: 93–114. Food Chemistry 116: 714–721. Nybom H. 2004. Comparison of different nuclear Hedhili L., Romdhane M., Planche H., Abderrabba M. DNA markers for estimating intraspecific genetic 2005. Towards gas chromatography–mass spec- diversity in plant. Molecular Ecology 13: trometry coupling protocols for both identifying 1143–1155. andquantification essential oils of Thymus Petanidou T., Vokou D. 1993. Pollination ecology of capitatus Hoff et Link. Journal of Chromatography Labiatae in a phryganic (East Mediterranean) eco- A 1064:129–134. system. American Journal of Botany 80: 892–899. Hovmalm H.A.P., Jeppsson N., Bartish I.V., Nybom Pérez-García F.M., Hornero J., González-Benito E. H. 2004. RAPD analysis of diploid and tetraploid 2003. Interpopulation variation in seedgermina - populations of Aronia points to different repro- tion of five Mediterranean Labiatae shrubby spe- ductive strategies within the genus. Hereditas cies. Israel Journal of Plant Science 51: 117–124. 141: 301–312. Pottier-Alapetite G. 1981. Flore de la Tunisie Angio- Kremer A., Caron H., Cavers S., Colpaert N., Gheysen spermes Dicotylédones Gamopétales, pp. G., Gribel R., Lemes M., Lowe A.J., Margis R., 809–811. Navarro C., Salgueiro F. 2005. Monitoring ge- Rohlf F.J. 1998. NTSYSpc Version 2.0. User Guide. netic diversity in tropical trees with multilocus Applied Biostatistics Inc. dominant markers. Heredity 95: 274–280. Saitou N., Nei M. 1987. The neighbour-joining me- Le Floc’h E. 1983. Contribution à une étude thod: a new method for reconstructing phylogen- ethnobotanique de la flore tunisienne. Imp.Off. etic trees. Molecular Biology andEvolution 4: Rep. Tunisia. 406–425. Lodhi M.A., Ye G.N., Weeden N.F. 1994. A simple Sas (statistical analysis system) 1990. SAS user’s andefficient methodfor DNA extraction from guide: SAS STAT, SAS BASIC. Version 6 fourth grapevine cultivars and Vitis species. Plant Molec- edition. SAS incl, Box 8000. Cary, NC ular Biology Reporter 12: 6–13. 27512-8000, Cary: NC: SAS Institut Inc. López-Pujol J., Bosch M., Simon J., Blanché C. 2004. Solouki M., Mehdikhani H., Zeinali H., Emamjomeh Allozyme Diversity in the TetraploidEndemic A.A. 2008. Study of genetic diversity in Chamo- Thymus loscosii (Lamiaceae). Annals of Botany 93: mile (Matricaria chamomilla) basedon morpholog - 323–332. ical traits andmolecular markers. Scientia Mabberley D.J. 1997. The plant-book. Cambridge: Horticulturae 117: 281–287. Cambridge University Press Stahl-Biskup E., Sáez F. 2002. Thyme – The Genus Mantel N. 1967. The detection of disease clustering Thymus. Medicinal and Aromatic Plants-Indus- anda generalizedregression approach. Cancer trial profiles. Taylor andFrancis, London,New Research 27:209–220. York, pp. 75–124. 100 Imen Ben El Hadj Ali, Arbi Guetat, Mohamed Boussaid Thompson J.D. 1999. Population differentiation in daptive genetic differentiation: comparison of Mediterranean plants: insights into colonization QST and FST at two spatial scales. Heredity 95: history andthe evolution andconservation of en - 466–475. demic species. Heredity 82: 229–236. Williams J.G.K., Hanafey M.K., Rafalski J.A., Tinjey S.V. Thompson J.D. 2002. Population structure andthe 1993. Genetic analysis using random amplified spatial dynamics of genetic polymorphism in polymorphic DNA markers. Methods Enzymo- thyme. in Stahl-Biskup, E., Saez, F., (Eds.), logy 218: 704–740. Thyme: the Genus Thymus, London: Taylor and Wright S. 1951. The genetical structure of popula- Francis. pp. 44–74. tions. Annals Eugenics 15: 323–354. Verma S., Karihaloo J.L., Tiwari S.K., Magotra R., Koul Yeh F., Yang R., Boyle T. 1999. Popgene, version 1.31. A.K. 2007. Genetic diversity in Eremostachys Microsoft window-based freeware for population superba Royle ex Benth. (Lamiaceae), an endan- genetic analysis. University of Alberta, Edmon- geredHimalayan species, as assessedby RAPD. ton. Genetic Resources andCrop Evolution 54: Zheng W., Wang L., Meng L., Liu J. 2008. Genetic varia- 221–229. tion in the endangered Anisodus tanguticus (Sola- Volis S., Yakubov B., Shulgina I., Ward D., Mendlinger naceae), an alpine perennial endemic to the S. 2005. Distinguishing adaptive from nona- Qinghai-Tibetan Plateau. Genetica 132: 123–129.