Comparison of Genetic Diversity in Species and Cultivars of Pistachio (Pistacia sp. L.) Based on Amplified Fragment Length Polymorphism (AFLP) Markers

INTRODUCTION

Pistachio (Pistacia sp. L.) belongs to a genus of the Anacardiaceae family. It is one of the most prominent horticultural plants from an economic and commercial point of view, so that vast pieces of land in Iran are allocated to the growth of this plant (Anonymous, 2001).
 All pistachio species are dioecious and wind-pollinated. The genus Pistacia consists of eleven species which only has edible nuts and is commercially important (Zohary, 1952). In addition, there are two other wild species in Iran; Pistacia atlantica subsp. mutica and Pistacia  khinjuk. These species are used mainly as rootstock for Pistacia vera and rarely for oil extraction in some countries (Kafkas and Perl-Treves, 2002b).
 Molecular studies addressing the genus Pistacia are few. Most studies regarding the analysis of the genetic diversity of Iranian pistachios have been based on morphological characteristics (Kafkas et al., 2002a; Tajabadipur, 1997). 
 Isozyme markers have also been used to investigate the genetic diversity of pistachios (Aalami et al., 1996; Rovira et al., 1995). Mirzaei et al. (2005) used Random Amplified Polymorphic DNA (RAPD) markers for analyzing the genetic diversity of 22 pistachio species in Iran. RAPD markers have also been used for studying genetic relationships of two wild pistachios (Kafkas and Perl-Treves, 2002b). Hormaza et al. (1994) identified a RAPD marker linked to the sex determinant gene.
 Genetic relationship among native and introduced pistachio in Greece has been studied by RAPD and AFLP markers (Katsiotis et al. 2003). Golan-Goldhirsh et al. (2004) has assessed polymorphism among the Pistacia species of the Mediterranean basin and accessions within the same species using RAPD and AFLP markers.  Ahmad et al. (2003) used microsatellite markers to identify Pistachio varieties grow in kerman province. These isolated microsatellite loci along with sequence-related amplified polymorphism (SRAP) markers were used for assaying genetic relationships among pistachio species (Ahmad et al., 2005).
 Pistachio as an important agricultural plant has several cultivars, but there have been few studies regarding the genetic diversity of Pistacia using molecular markers in Iran. The objective of the present study is to investigate genetic diversity among Pistacia species grows in Iran. In this study, AFLP markers with 13 primers have been used to investigate the extent of diversity in 45 species of Pistacia.

MATERIALS AND METHODS

Plant materials and DNA extraction: Plant materials (leaves) were collected from Iran’s Pistachio Research Institute (IPRI) in Rafsanjan. Young leaf samples were kept in liquid nitrogen tanks for the purpose of  DNA extraction and AFLP analyses. In total, 45 pistachio species were used in this study which consisted of three wild species (Pistacia palaestina, Pistacia atlantica subsp. mutica and Sarakhs), and 42 cultivars (Table 1). Total genomic DNA was extracted from freez-dried leaf tissue using the Cetyl Trimethyl Ammonium Bromide (CTAB) mini-extraction protocol based on the method by Hormaza et al. (1998), but with minor modification (Mirzaei et al., 2005). Quality and quantity of DNA were determined by a spectrophotometer (Beckman DU 530) and the concentration of DNA was also confirmed by electrophoresis using 1% (v/v) agarose for 45 min at 80 V in 0.5X TAE buffer, followed by  visualization under UV light after staining with ethidium bromide.

AFLP Analysis: The AFLP reactions were carried out according to the method by Vos et al. (1995). Two hundred nanograms of genomic DNA from each sample was digested with EcoRI and MseI in a total volume of 20 ml at 37°C for 4 hour and then double stranded adaptors were ligated to the fragment ends as well. The digested and ligated DNA were then diluted by the addition of 125 ml ddH2O and pre-amplified using EcoRI and MseI primers with no additional selective nucleotide. Pre-amplification was performed in a total volume of 25 ml containing 50 ng of each primer, 0.2 mM of each dNTP (Roche), 1X PCR buffer (Roche), 1 U of Taq DNA Polymerase (Roche) and 3 ml of the diluted digested and ligated DNA.
 The temperature profile for pre-amplification was as follows; denaturing for 1 min at 94°C, 1 cycle; annealing for 1 min at 56°C, extension for 1 min at 72°C, 30 cycles; 72°C for 7 min, 1 cycle. For selective amplification, the product of pre-amplification was diluted by addition of 125 ml ddH2O. Selective amplification was carried out using thirteen primers with three selective nucleotides (Table 2). The selective amplification reactions were carried out in a total volume of 15 ml comprising 15 ng of each primer, 1X PCR buffer, 0.25 mM of each dNTP, 1 U of Taq DNA Polymerase and 4 µl of diluted pre-amplification product using a ‘Touchdown’ cycle programmed as follows; 30 s at 94ºC, 30 s at 65ºC and 60 s at 72ºC. The 65ºC annealing temperature was reduced by 0.7ºC per cycle for 12 consecutive cycles and then maintained at 56ºC for the remaining 24 cycles. The amplified products were separated on a denaturing 6% (v/v) polyacrylamide sequencing gel. After electrophoresis, the gel was stained with silver nitrate and then visualized. 

Data analysis: Banding patterns were scored as presence (1) and absence (0) using the cross Checker software ver 2.91. The genetic similarity metrices were constructed using Simple Matching (Sokal and Sneath 1963), Jaccard (Jaccard 1908) and Dice coefficients (Nei and Li 1997). Dendrograms were constructed by the unweighted pair-group method using arithmetic average (UPGMA) and complete linkage algorithms. In addition to cluster analysis, principal component analysis was used to confirm the results of cluster analysis. The efficiency of clustering algorithms and their goodness of fit were determined based on the cophenetic correlation coefficient. Data analyses were performed by the NTSYS software ver 2.02.
RESULTS

AFLP profiling of 45 pistachio genotypes with 13 primer combinations revealed a total of 527 scorable bands ranging in size from 100-1000 nucleotides. A total of 506 AFLP fragments were polymorphic across all the genotypes for the 13 primer pairs. The 506 polymorphic fragments accounted for 95.2% of the total amplified fragments.
 The degree of polymorphism among primers was different (Table 2). The E(AAC)-M(CAA) primers showed the highest polymorphism whereas the minimum number of bands and the lowest degree of polymorphism was observed with the E(AAG)-M(CCA) primers. However, the E(AAG)-M(CAC) primer combination showed the maximum number of bands.
 The cophenetic correlation, a measure of the correlation between the similarity represented on the dendrograms and the actual degree of similarity, was calculated for each dendrogram. Among the different methods, the highest value (r= 0.8855) was observed for UPGMA based on Jaccard’s coefficient (Table 3). Therefore, the dendrogram constructed based on this method was used for depicting genetic diversity of genotypes (Figure 1).
 Cutting the dendrogram at the point with maximum distance among groups resulted into four main groups (Figure 1). All pistachio cultivars were placed in group I and II. Group III included two wild pistachios, Baneh (P. atlantica subsp. mutica) and Baneh Baghi. A wild pistachio species, P. palaestinia was placed in a separated cluster with very low similarity to other groups.
 In group I the highest similarity value was observed between Rezaii Zodras and Hasan Zadeh cultivars. Group I was further divided into two subgroups in which Sarakhs as a wild variety of P. vera was placed in subgroup II. In group II, the highest similarity value was found between two cultivated pistachios; Javad Aghaii and Mosa Abadi cultivars (0.88). Comparing with group I, the pairwise similarities in group II were higher. For example, the highest observed similarity value in group I was 0.78 which was found between Rezaii Zodras and Hasan Zadeh cultivars. Cultivars in group II were clustered together at a higher similarity value. The lowest similarity values were between        P. palaestinia and all other species.
 Considering morphological characteristics, cultivars in group I had fruits with orbicular or elongated shapes except for the Harati cultivar which had an ovate-shaped fruit, however, most of the cultivars with ovate-shaped fruits were mainly clustered in group II. According to fruit ripening time, cultivars were almost separated by cluster analysis. Early ripening cultivars were placed in group I except for the Ghafouri which as a late ripening cultivar was also placed in this group. Late ripening cultivars were placed in group II which also included the early ripening cultivars of Ghazvini Zodras and Italiaii Zodras.
 Principal coordinate analysis (PCA) based on genetic similarity matrices were used to visualize the genetic relationships among species. The first two eigenvectors accounted for 44.03% of the total molecular variation. Therefore, PCA results confirmed the results of cluster analysis. Based on PCA, cultivars and wild species were almost separated from each other (Figure 2).

DISCUSSION

The utility of DNA-based markers such as AFLP as a reliable technique for assaying genetic variation among plant species has widely been reported (Blears et al., 1998). This technique is more informative and reproducible compared to previously used biochemical and molecular methods such as isozyme, and RAPD markers in detecting genetic relationships of pistachio genotypes (Golan-Goldhirsh et al., 2004; Katsiotis et al., 2003).
 Cluster analysis of AFLP data could separate cultivars and wild species of pistachio. In this grouping, wild species P. atlantica subsp. mutica was closer to the cultivated pistachio than P. palaestinia, supporting the earlier report of Kafkas and Perl-Treves, (2002b).
 All the cultivated pistachios were clustered in groups I and II. These cultivars are mostly separated according to two morphological traits; fruit shape and time of fruit ripening. Results showed that, cultivars with orbicular and elongated nut shapes are close to each other but separated from cultivars having ovate shapes. Hence, orbicular and elongated nuts are different form nuts with ovate shape, from molecular point of view. Mirzaei et al. (2005) who analyzed the genetic relationships of Iranian cultivated pistachio and wild species also could not separate orbicular nuts from elongated ones. It seems that primers used in the RAPD and also in the present study did not cover the genomic region(s) consisting of gene(s) controlling the orbicular and elongated nut shapes.
 Based on AFLP data, the cultivars with early and late ripening fruits were distinguished from each other. Therefore, the primer used in this study could be employed in further studies to separate cultivars having these two morphological traits. It was hypothesized that regions amplified with these primers probably contain gene(s) which control(s) fruit ripening and specify ovate fruit shape. Nonetheless confirming the location of these genes and relating them to AFLP markers needs further research.
 P. vera subsp. Sarakhs is a wild species of Pistacia which grows as self-grown forests in North-east of Iran. It showed the highest genetic similarity with cultivated species. The results of this study as in previous studies (Mirzaei et al., 2005; Aalami et al., 1996) suggest that the cultivated pistachio in Iran may have originated from P. vera subsp. Sarakhs. Baneh (P. atlantica subsp. mutica) and Baneh Baghi have been placed in a single cluster. These two species are similar based on morphological characteristics such as tree height, crown shape and number of leaflets. As a result, it can be hypothesized that Baneh Baghi is a hybrid of Baneh and cultivated pistachio that has acquired certain traits from cultivars through time.

Acknowledgments

The Financial support of Isfahan University of Technology, Isfahan, Iran and provision of plant materials from the Pistachio Research Institute of Iran, Rafsanjan, Iran are gratefully acknowledged.