Patterns of Genetic Diversity and Structure at Fine Scale of an Endangered Moroccan Endemic Tree (Argania spinosa L. Skeels) Based on ISSR Polymorphism

The preservation of the diversity of endangered populations of argan trees, in their natural habitat, is a crucial step toward their conservation. The aim of the present study was to evaluate the genetic diversity of the argan trees in the wild, and to establish a phylogenetic map using DNA fingerprints. The ultimate goal was to develop a core set that would represent the existing diversity in the whole germplasm. In regard to this, 200 samples of Argania spinosa individual trees were collected from 10 different provenances in the region of Essaouira (Morocco). The genetic variation between and within these argan trees was investigated using previously described Inter-Simple Sequence Repeat markers. These markers generated a total of 149 fragments, in which 148 (99.33%) were polymorphic. The samples collected in the ‘Ouled Lhaj’ provenance showed the lowest diversity (% of polymorphic locus P=48.32%; genetic diversity Nei h=0.153; allelic richness A=1.483), compared to those collected in the ‘Mramer’ provenance (%P=68.46%; h=0.233; A=1.685). Also, the results showed a high level of genetic differentiation among provenances (AMOVA=44%, Gst=0.40), and a limited gene flow (Nm=0.73) between the provenances. In addition, these data suggested a low correlation between the genetic diversity of the tree and their respective geographical location in relation to the proximity to the littoral. Finally, a core collection of 13 genotypes that represent the essential of the detected diversity was established. The distribution pattern of this genetic diversity provides an important baseline data for the conservation strategies of argan tree species in the wild.


Introduction
The argan tree (Argania spinosa L. Skeels) is an endemic forest species in Morocco (Northern West of Africa), covering approximately 828000 hectares (Mhirit et al., 1998).This species belongs to Sapotaceae family and it is a diploid species (2n=24) (Majourhat et al., 2007), monoecious and allogamous (Msanda et al., 2005) which is believed to be pollinated by insects (El Mousadik and Petit, 1996a).Argania spinosa is a multi-purpose tree (oil-agriculture-forestry-pastoral) with a great genetic diversity (El Mousadik and Petit, 1996a;1996b).Due to its diversity, the argan tree has a remarkable resilience to withstand climate change adverse, especially in geographical area with increasingly accentuated dryness.In Morocco, the argan tree plays an important role in preserving the local environment, and ensuring the socio-economical equilibrium in the area where it grows (Nouaïm et al., 1991).The area of Essaouira produces between 1000 to 2000 tonnes of oil per year, corresponding very roughly to a local population of 60000-120000 trees, producing 142 million to 286 million fruits per year (Moussouris and Pierce, 2000).
Knowledge of genetic diversity within and among populations is particularly important for conservation management (Jian et al., 2006).Indeed, species that lack an appropriate amount of genetic diversity are generally unable to survive in a changing environment (Qian et al., 2001).The levels date.The earliest investigations studied the isozymes markers and the DNA of chloroplast using PCR-RFLP (El Mousadik and Petit, 1996a;1996b), followed by studying the Random Amplified Polymorphic DNA (RAPD) (Bani-Aameur and Benlahbil, 2004;Majourhat et al., 2008), and more recently studies on the Single Sequence Repeats (SSR) or microsatellites (Majourhat et al., 2008).Among these various molecular tools, the Inter-Simple Sequence Repeats (ISSR) using the PCR, remains the most reproducible, and gives a larger number of polymorphic fragments per primer.Indeed, it allows the identification of a very high number of loci, using only a few primers (Zietkiewicz et al., 1994).This technique also does not require prior knowledge of DNA sequence in order to design primers, straightforward to perform and cost effective compared to SSR and SNP.ISSR markers have been often successfully used to evaluate the genetic diversity of many species, such as Morus spp.(Kar et al., 2008), Barby fig (Opuntia ficus indica) (Ganopoulos et al., 2015), Quercus infectoria (Rahmani et al., 2015).
The aim of the present study is to examine the genetic diversity of naturel populations of Argania spinosa in the wild.The ultimate goal of this work is to establish a core collection representing the diversity that still remains conserved in the argan germplasm located in the region of Essaouira.

Biological materials
Fresh leaves were collected from 20 individual trees per each provenance, and then stored at -20 °C.The trees were selected randomly and their locations were individually registered using the Global Positioning System (GPS).

DNA extraction
The genomic DNA was extracted from 40 mg of grounded leaves, using a modified CTAB (Cethyltrimethyl Ammonium Bromide) procedure, initially described by Doyle and Doyle (1987).Briefly, 1 ml of CTAB extraction buffer (1 M Tris-HCl pH 8.0, 1.36 M NaCl; 0.5 M EDTA pH 8.0, 0.5 M Poly-Vinyl Pyrrolidone (PVP) and CTAB 2%) and 3 µl of βmercaptoéthanol were added to the leaves and the mixture was incubated at 65 °C for 30 min.Genomic DNA was extracted using chloroform: iso-amyl alcohol (24:1) and centrifugation at 10000 rpm for 10 min, at 4 °C was performed.The DNA was precipitated, from the supernatant, using isopropanol (2/3 volume of supernatant) and incubate overnight at -20 °C.The precipitated DNA was collected by centrifugation at 14000 rpm for 15 min, and the pellet was washed successively in 76% ethanol containing 10 mM ammonium acetate and in 75% ethanol.Finally, the dried pellets of DNA were dissolved in 100 of genetic variation between different populations and within the same population are the result of a dynamic interplay between several factors that include the gene flow, the selection effect, the inbreeding, the genetic drift and mutations (Hartl and Clark, 1994).Recently, molecular methods have provided accurate and reproducible tools to analyze the genetic variation of species, which is proven to be of great importance in planning efficient strategies for their conservation (Hamrick and Godt, 1996).
In Morocco, the argan ecosystem is subjected to a continuous degradation due to the anthropozoïc pressure, the dry climate, and the lack of natural regeneration.By consequence, the density of the argan tree forest suffered a regression from an average of 200 trees/ha to less than 30 trees/ha (Dupin, 1949;Nouaïm et al., 1991).As a result, the conservation of the remaining genetic resources of argan forest became a priority, especially after the UNESCO declared the argan ecosystem as a Biosphere Reserve in 1998.
Most of the studies regarding the genetic variability of the argan tree used agromorphological approaches (Ferradous, 1995;Aameur and Ferradous, 2001;Benlahbil, 2003;Bani-Aameur, 2004;Ait Aabd et al., 2011).These studies strongly highlighted the existence of a high phenotypic diversity.However, morphological characteristics are usually not sufficient to give a clear picture of the existing genetic diversity.By contrast, the assessment of genetic diversity using DNA markers was proven to be more accurate and reliable.
The level of the genetic diversity of the argan trees in the wild remains poorly investigated.In fact, only few studies that focused on the genetic diversity of Argania spiniosa were published to µl of sterile ultrapure water.The quantity and the quality of DNA obtained were controlled by spectrophotometer and by electrophoresis in agarose gel (0.8%) containing ethidium bromide.Aliquots of DNA samples were prepared at a concentration of 50 ng/µl, and stored at -20 °C before use.

ISSR PCR amplification
In this study, ten ISSR primers that produced a clear, reproducible and relatively high polymorphism were used to estimate the genetic diversity of 200 samples (Table 2).These primers, previously described by Souto Alves et al. (2009), were successfully used to study a wide range of forest species (e.g., Apterosperma oblata (Su et al., 2008), Murraya koenigii (Verma and Rana, 2011).
The DNA amplification by PCR was performed in a 10 µl reaction mixture containing 15 ng of genomic DNA, Taq polymerase buffer (1 ), 0.2 mM of dNTP, 0.2 mM of primers (Eurofins) and 1U of Taq polymerase (Promega).The reaction mixture was supplemented with 1.5 to 2 mM of MgCl2.The concentration of MgCl2 varied according to the nature of the primers (Table 2).The PCR was performed in a Thermal Cycler (Applied Biosystems 2720), using a program that start with an initial denaturation cycle at 94 °C for 3 min; followed by 30 cycles of 45 s denaturation at 94 °C, annealing at 50-60 °C for 30 s (depending on the type of primers) and 1 min extension at 72 °C; and ending with final extension cycle at 72 °C, for 7 min.The PCR products were visualized by separation for 2 h at a constant voltage (90 V), on agarose gel (2%) (containing ethidium bromide).

Statistical analysis
The amplified fragments with the same mobility, on the agarose gel, according to their molecular weight (bp) were scored using a binary code [present (1) or absent (0)].Each of these fragments represents a locus.The parameters related to the genetic diversity within and between different provenances [i.e. the percentage of polymorphic bands (PPB), Nei's gene diversity (h), Nei's genetic differentiation index among populations (G st ), allelic richness (A) and gene flow (Nm)] were calculated using POPGENE 1.32 (Yeh et al., 2000).An estimate of Nm among populations was calculated using the formula of Nm=0.5 (1-G st )/G st (McDermott and McDonald, 1993).The genetic distance matrix was employed to construct UPGMA tree (unweighted pair group method using arithmetic mean) using MEGA version 5 (Tamura et al., 2011).The molecular variance (AMOVA), using GenALEX version 6.5b3 (Peakall and Smouse, 2012) was used to estimate the distribution of the genetic variation between different provenances and within the same provenance.The correlation between the genetic and geographic distances was estimated using the Mantel test.A Bayesian analysis of ISSR population structure on the entire data set, using STRUCTURE 3.2 (Pritchard et al., 2000), was conducted to test for genetic admixture across provenances.STRUCTURE algorithm was run using the basic model with admixture and correlated allele frequencies, with the assumed number of genetic K clusters varying from 1 to 10 (the total number of provenances), ten replicate runs per K value(number of populations), and the burn-in period and Markov Chain Monte Carlo (MCMC) were set to 500,000 and 5000,000 iterations, respectively.To identify the number of K clusters explaining the observed genetic structure, statistical parameters defined by Evanno et al. (2005) based on the rate of change in the log probability of data between successive K values were used.Finally, it assembled a core collection that should represent the entire genetic diversity explored in the study using Powercore software (Kim et al., 2007).

Genetic diversity
The results showed (i.e.Fig. 2) a total of 149 bands detected by using 10 screened primers across 200 individuals, from the 10 provenances studied.Among them, 148 (99.33%) were polymorphic.
The 149 locus generated by 10 ISSR primers were sufficient to genetically characterize the studied provenances.To define the ISSR variation within the provenances, the POPGENE software was used to calculate five parameters that represent estimation for each locus, and the average represent the diversity in the loci.The analysis of the diversity within provenances (Table 3) showed that all the parameters followed the same direction (A; h; NPB; %P).The

Genetic differentiation and gene flow
The AMOVA analysis showed that 56% of the total genetic variability resided between provenances (Table 4).This result was confirmed by the estimate of gene differentiation coefficient (Gst=0.4053),which indicates that 40.53% of the total genetic variability exist between different groups of trees and 59.47% was within the groups.The level of gene flow (Nm) was estimated to 0.7336, indicating a limited rate of gene exchange among provenances.
To measure the genetic divergence between these populations, it has been used the G st pairwise index (Palsbøll et al., 2007).The index values showed that all the populations studied were clearly differentiated.The highest genetic differentiation (0.3435) was particularly observed between 'Retmana' (RT) and 'Tamanar'(TM) (Table 5).
In the ISSR admixture analysis using STRUCTURE, the real K value with the highest value of ∆K for the 200 individuals was K=6 (Fig. 4a).The proportions of each provenance assigned into six clusters (cluster I and cluster II) (Fig. 4b) which result is in agreement with UPGMA dendrogram based on Nei's genetic identities.The Mantel's test (Mantel, 1967) results showed a positive but non-significant correlation between genetic structure and geographical distance (r=0.181,P=0.120, 999 permutations).

The construction of the core collection
Core collection is a subset of accessions from the entire collection that capture, with minimum redundancy, most of the available genetic diversity of a crop, a wild species or a group of   long-term survival of a species.In this study, the high level of genetic diversity has reported in 'Mramer ' provenance A=1.6846; h=0.2332; %P=68.46).By contrast, the argan tree from the provenance of 'Ouled Lhaj' (OH), located in the North and closest to the ocean (less than 3 km), were the less genetically diverse (A=1.4832;h=0.1526; %P=48.32).However, if considering all the provenances that were studied as a single source, a high genetic diversity is noted (A=1.9933,h=0.353) (table 3).In this work, the ISSR analysis using ten primers generated a higher polymorphism compared to other molecular markers such as isozymes (El Mousadik and Petit, 1996) and RAPD (Bani-Aameur and Benlahbil, 2004;Majourhat et al., 2008).In addition, the Argania spinosa showed a relatively higher mean of gene diversity using these ISSR markers compared to other species of trees (Zhao et al., 2006;JunMin et al., 2007;Sun et al., 2013;Vaishali et al., 2014, Rahmani et al., 2015).However, there was no significant difference between genetic diversity and altitude; high genetic diversity can be found in mountains and plains.Low and high genetic diversity between provenances could be attributed to many factors including, the geographical distribution combined with the level of isolation of the populations, the heavy deforestation and the extensive inhabitation.

Genetic differentiation and gene flow
The analysis of multi locus profiles has shown a very high level of differentiation inside the 10 provenances studied (G st =0.4053; AMOVA=0, 44; Nm=0.7336), which is consistent with our previous reports that focused on isoenzymes and chloroplast DNA of the argan tree (Gstn=0.25;G stc=0.60) (El Mousadik and Petit, 1996a;1996b).Compared to other forest species, for instance the sessile oak that covers most of Europe species (Brown, 1989a, b;Van Hintum et al., 2000, Belaj et al., 2012).Generally, information on geographic data, genetic data, or phenotypic data were used for grouping.To achieve this objective, an approach of maximization was adopted, using a heuristic algorithm implemented in the Power Core 1.0 software (Kim et al., 2007).This program allowed the selection of the most diverse genotypes covering all alleles existing in the original collection.The content of the core collection obtained using this software is shown in the Table 6.The results showed a Core collection representing existing genetic diversity that includes 13 genotypes from different regions selected by this program.This collection represents 6.5% of the total collection.The constructed core collection capture about 95% of the total diversity (A=1.9396;h=0.3630; %P=93.96).Indeed, the comparison of the genetic diversity indices between the genotypes included in the collection (13 genotypes) and the entire collection (200 genotypes) did not show any significant differences (Table 7).

Genetic diversity
Genetic diversity is critical for the adaptation and the resistance to environmental changes and consequently for the present in the entire collection (Zhang et al., 2011;Khaing et al., 2013).In this sense, core collections exist for some fruit tree species (Volk et al., 2005;Escribano et al., 2008;Santesteban et al., 2009;El Bakkali et al., 2013;Miyamoto et al., 2014).In a comparative study, Belaj et al. (2012) used Mstrat and PowerCore to construct a core collection of Olea europaea L. their results showed the high efficiency of PowerCore to capture all alleles present in the source collection.This is consistent with the results obtained in multiple previous studies that have also suggested the efficiency of PowerCore (Agrama et al., 2009;Zhao et al., 2011;Kaga et al., 2012;Zhang et al., 2012;Khaing et al., 2013).Considering the present results, this collection might be an appropriate choice for applications involving the conservation of genetic resources of the argan trees.Field investigations must be taken as complementary criteria to ensure optimal management of the argan forest, in particular, criteria of phenotypic and agronomic aspect, but also of socioeconomic and environmental nature as this will help to utilize germplasm in breeding programmes more effectively.

Conclusions
In the current study, the ISSR analysis was a key tool to demonstrate a high level of genetic diversity between the argan trees from twelve different areas, which provides a potential resource of germplasm that might be used in the future conservation programs.Indeed, the determination of the genetic relationships between argan tree will further improve the efficiency of sampling and using of germplasm resources.and that is characterized by a low differentiation index (G st =0.026) (Kremer et al., 2002), it has been found that the argan tree is among the most genetically differentiated forest resources.This is consistent with reports in relation to other threatened or endemic species, like: Pinus sylvestris (G st =0.3965) (Hui-yu et al., 2005); Cupressus gigantea (G st =0.36) (Xia et al., 2008); Machilus thunbergii (G st =0.4118) (Liu et al., 2013) and Rheum palmatum and Rheum tanguticum (G st palmatum =0.537; G st tanguticum =497) (Wang et al., 2012).The high genetic differentiation of the argan trees could be explained by several factors, including their geographical distribution, their breeding system (entomogame) and the genetic isolation of their populations.Indeed, the scattered distribution of species added to the topographic barriers can lead to difficulties in the dispersal of pollen and seeds, and consequently limit gene flow among populations (Lu et al., 2008).Also, it was suggested that endemism and limited distributions of populations within a species favor a high genetic differentiation (Hamrick and Godt, 1990).

Genetic relationships
The knowledge of the genetic relationships among provenances of Argania spinosa is important information for the efficient use and ex situ conservation.According to the present data analysis, using UPGMA dendrogram, the argan trees included in this study could be divided into 6 distinct groups or clusters.Four of them were formed by trees from two different provenances, (i.e.'Retmana' and 'Ouled Lhaj', 'Jbel Kourati' and 'Ait Issi', 'Rbaï' and 'Meknafa', 'Tamsrourt' and 'Tamanar').By contrast, the remaining two groups were formed by trees coming from a single provenance (i.e.'Mramer' and 'Timzgida Ouftass').Bayesian analysis showed that the germplasm collection of argan tree of Essaouira could be represented by six large groups, which is consistent with UPGMA cluster analysis.Further analysis like G st pairwise confirmed this genetic structure.However, there was no correlation between the genetic structures of these populations and their respective geographic distance (r=0.181,P=0.120, 999 permutations), indicating that geographic isolation was not the main factors inducing genetic difference.
The genetic structure of plant populations reflects the interactions of various evolutionary processes including the longterm evolutionary history of the species (e.g., shifts in distribution, habitat fragmentation, and population isolation), genetic drift, mutation, breeding system, and selection (Slatkin, 1987;Hamrick and Godt, 1996;Schaal et al., 1998).Therefore, the development of genetic structure within and between the populations of the argan tree in the region of Essaouira is probably strongly associated with the gene flow caused by the movement of pollen and seeds.

Construction of nested core collections maximizing diversity
The purpose of core collections is to facilitate the use of germplasm by providing a set of accessions displaying the genetic diversity available in the larger collection (Brown, 1989a), and this core collection facilitate experimental trials to assess germplasm under contrasting environmental conditions (El Bakkali et al., 2013).The current data analysis provides 13 individuals (6.5% of the origin collection) defined by 10 ISSR markers (Table 6), which represent 100% of the genetic diversity of the entire collection (Table 7).The core collection established with heuristic analysis method could maintain specific alleles

Fig. 3 .
Fig. 3. UPGMA dendrogram based on Nei's genetic distances among the populations of the argan trees

Table 1 .
Geographic coordinates of selected argan trees Fig.1.The geographical distribution of the provenances of argan tree samples (see Table1for abbreviation)

Table 2 .
List of ISSR primers used in this study Temperature of annealing; TNB: Total number of band.

Table 3 .
Genetic diversity within the argan provenances

Table 4 .
Analysis of the molecular variance (AMOVA) among the argan trees provenances DF: degree of freedom; SS: sum of squares; MS: Mean Square; Est.var: estimation variance; %:Percentage variance.

Table 5 .
The determination of the genetic differentiation between the different argan trees provenances using Gst pairwise index (see Table1for abbreviation

Table 6 .
List of 13 trees out of 200 trees of entire collection included in core set formed by PowerCore

Table 7 .
Comparison of genetic diversity indices among the entire set and the core established