Genetic Similarities Among Iranian Populations of Festuca, Lolium, Bromus and Agropyron Using Amplified Fragments Length Polymorphism (AFLP) Markers

INTRODUCTION

Forage and turf grasses have an important role in sustainable agriculture and contribute extensively to the world economy. They play a major role in meat and dairy production and are important in soil conservation, environmental protection, and outdoor recreation (Wang et al., 2001).
The genus Festuca contains approximately 100 species, some of which are commonly used as forage and turf grasses. They belong to the grass family Poaceae, subfamily Pooideae, and tribe Poeae (Soreng and Davis, 1998). Based on leaf texture these are divided in two subgeneric types, including the coarse fescues (e.g. Festuca arundinacea and Festuca pratensis) and fine fescues (e.g. Festuca rubra and Festuca ovina) (Turgen, 1985). Tall fescue (Festuca arundinacea) is the most important perennial forage and turf grass species of this genus and is widely grown throughout the temperate regions of the world (Saha et al., 2005; Sleper, 1985). Tall fescue is a hexaploid (2n=6x=42) consisting of three genomes (PPG1G1G2G2) with the P genome derived from diploid Festuca pratensis (2n=2x=14) and the G1G2 genome from the tetraploid  Festuca arundinacea var glaucescens (2n=4x=28) (Sleper, 1985; Seal, 1983). The genus Lolium, which is closely related to the Festuca, contains several diploid species including the outcrossing perennial Lolium perenne, the annual Lolium multiflorum, and the self-pollinating annual Lolium temulentum, all of which are cross-compatible (Saha et al., 2004). Rigid ryegrass (Lolium rigidum) is another annual diploid in this group.
Russian bromegrass (Bromus tomentellus) is a perennial grass with a wide geographical distribution in the most arid and semiarid regions of Iran, as well as neighboring countries (Rechinger, 1973). Ecotypes of B. tomentellus are well adapted to both cool-moist and cool-dry environments and are grown at different densities on approximately 15 million hectares of native rangelands in Iran, mixed with other grasses and legumes (Mirlohi et al., 2006). Crested wheatgrass (Agropyron cristatum) is a cross-pollinating diploid, containing the basic P-genome (Dewey, 1984; Love, 1984), and is widely distributed in the natural habitats of Iran.
Among molecular techniques for genetic assessment, amplified fragment length polymorphisms (AFLP) is a DNA marker system based on a combination of polymerase chain reaction (PCR) and restriction enzyme analyses and reveals high levels of polymorphism. AFLP is highly reproducible, less sensitive to reaction conditions and does not require DNA sequence information (Krauss and Peakall, 1998; Vos et al., 1995). This technique have been effectively used to study genetic variation, characterize accessions and asses genetic similarities between and within genera and species of grasses (Vergara and Bughrara, 2004; Wu et al., 2004; Vergara and Bughrara, 2003; Guthridge et al., 2001). Mian et al. (2002) used AFLP markers to determine genetic diversity and to distinguish 18 populations of tall fescue from USA, using the DNA bulk strategy. Fjellheim and Rognli (2005) assessed 12 Nordic cultivars and one Icelandic natural population of meadow fescue (Festuca pratensis Huds.) by the AFLP marker technology and found high levels of genetic diversity. Guthridge et al. (2001) assessed genetic diversity within and between perennial ryegrass (Lolium perenne L.) populations using AFLP markers and interpreted their findings in terms of breeding history of genotypes.
Reliable characterization of native plant germplasm is an essential step towards better use of wild genetic resources in plant improvement programs. The aim of the present study was to analyze genetic diversity and inter-population variability in different species of Festuca, Lolium, Bromus, and Agropyron native to Iran using AFLP markers.

MATERIALS AND METHODS

Plant materials: Forty two populations belonging to eight species of the genera Festuca, Lolium, Bromus, and Agropyron were used in this work (Table 1). Iranian accessions were collected from different geographical regions nation wide. Out of the 42 accessions, 13 originated from Hungary, Poland and USA, kindly provided by the gene bank of the Hungarian Institute of Agrobotany (HIFA), Tapioszele, Hungary. All accessions were germinated and grown in a greenhouse before use in DNA extraction.

DNA extraction and amplified fragments length polymorphism (AFLP) profiling: For DNA extraction, young leaf tissue was equally sampled from 30 plants of each accession and bulked together. Genomic DNA was isolated according to the procedure described by Dellaporta et al., (1983). DNA was quantified by spectrophotometer (Beckman-DU 530, Germany) and its quality was checked by agarose gel electrophoresis (Biorad, Germany).
Isolated genomic DNA (approximately 300ng) was digested with EcoRI and MseI restriction enzymes at 37ºC for 3 h. The restricted DNA fragments were ligated to EcoRI and MseI adaptors overnight at 37ºC and the product was then diluted (1:5). Pre-amplification reactions were performed with EcoRI+C and MseI+C AFLP primers. The amplification products were diluted (1:5) and stored at -20ºC until use for selective amplification. Selective amplification was carried out with 12 combinations of EcoRI+3 and MseI+3 primers (Table 2) in a final volume of 20 ml containing 4 ml of the diluted pre-amplification product, 15 ng of the EcoRI and MseI primers, 1X PCR buffer, 20 mM MgCl2, 1.0 U of Taq DNA polymerase (Roche company, Germany) and 0.2 mM dNTPs (deoxynucleotide triphosphates).
The selective amplification product was mixed with 10 ml of the loading buffer, and the mixture was denatured at 95ºC for 4 min and immediately placed on ice. Five ml of each of the denatured samples was loaded onto a 6% (w/v) polyacrylamide gel containing 7 M urea and electrophoresis was conducted with constant power (100 W) at a constant temperature of 50ºC for 2.5 h in a Biometra S2 sequencing gel. After electrophoresis, gels were fixed for 30 min in 10% (V/V) acetic acid solution and immediately afterwards, stained with silver nitrate (Pillay and Myers, 1999). A typical sample of stained gel is shown in Figure 1.

Data analysis: For data analysis, AFLP bands throughout the gel profile were scored as present (1), absent (0) or ambiguous (9), at least twice. The NTSYSpc v.2.02 software was used to generate genetic similarity matrices, create a dendrogram and corresponding cophenetic matrices, and calculate the cophenetic correlation (Rohlf, 1997). Cophenetic matrix correlation values were calculated to measure goodness of fit of the tree matrices and were interpreted according to Rohlf (1997) as follows: less than 0.7, very poor fit; 0.7-0.8, poor fit; 0.8-0.9, good fit; and 0.9-1.0, very good fit. Genetic similarities were calculated based on the Jaccard coefficients (Rohlf, 1997). Dendrograms were generated with the unweighted pair group method using arithmetic average (UPGMA) clustering method. Principal coordinate analysis (PCOA) was also conducted to identify the number of groups based on the Eigen vectors.

RESULTS

A total of 497 fragments were scored from 12-primer combinations, ranging in size from 50 to 500-bp (Table 2). Out of the 497 scored bands, 457 (92%) were polymorphic. The number of polymorphic bands for each primer combination varied from 10 to 74. The E-ATG/M-CCT primer combination produced the greatest number of polymorphic fragments (74 bands), while the E-AGG/M-CCC primer combination produced the lowest number (10 bands) (Table 2).
Genetic similarity coefficients (SC) based on AFLP markers ranged from 0.15 to 0.88 in these accessions. The highest SC (0.88) for pair wise comparisons among the genotypes was obtained between two tall fescue accessions (FAM6 and FAO10) from Hungary. The lowest SC value (0.15) was for the pair wise comparisons of A. cristatum and L. rigidum.
Cluster analysis provided a better illustration of genetic similarities among accessions. The UPGMA cluster tree generated by similarity coefficient matrix is shown in Figure 2. To test the dendrogram goodness of fit, the cophenetic correlation was calculated and interpreted according to Rohlf (1997). The cophenetic correlation was 0.96 indicating the high goodness of fit of the similarity indices. According to this interpretation the evident patterns, were considered significant because the correlation between the SC matrix and the cophenetic matrix was as r = 0.96 (data not shown).
The AFLP analyses provide measure of genetic variation within and among studied grass genera. All accessions within a species were clearly clustered in a single group. At the similarity coefficient of 0.26, cluster analysis grouped Agropyron and Bromus populations in separate clusters and distinguished them from other genera. Agropyron and Bromus belong to tribes Triticeae and Bromeae respectively while other genera fall in the Poeae tribe. At the similarity coefficient of 0.30, fine leaf fescues (F. ovina and F. rubra) were separated from coarse leaf fescues (F. arundinacea and F. pratensis) and Lolium species. The AFLP profile showed that coarse fescues have more genomic similarity with Lolium species than with fine fescues.
At the similarity coefficient of 0.42, the accessions were grouped into seven major clusters, each corresponding to a separate species except for the two Lolium species in that were placed in one cluster (Fig. 2). Festuca ovina was the only accession in this study that was separated from the other accessions, showing greater inter-specific than intra-spesific variation at the genomic level.
In the cluster containing F. pratensis accessions, the two accessions, FPM4 and FPN11, had the highest similarities. These two accessions were also related in terms of geographical locations (Table 1). The last and largest cluster in this grouping included all 25 accessions of the tall fescue. The accessions of this cluster were subdivided into six subclusters, most of which fell into subclusters congruent with their geographical origins. Two accessions, FAM9 and FAG9 (both from Iran), did not group with any other entry and consisted of separate clusters.

DISCUSSION

Most of the primer combinations tested in this study revealed workable patterns and high DNA polymorphism among grass populations. For this set of primers none of the accessions shared identical DNA marker profiles indicating that the collection did not contain duplications. The high level of polymorphism has facilitated analysis of the genetic diversity among accessions and in a few situations specific AFLP markers were also found for some species. Affirmation of these markers in other collections may assist in developing specific probes to effectively discriminate fescue species.
    The coarse fescues had more genomic similarity with Lolium species than fine fescues. Xu and Sleper (1994) have indicated that the genetic distance between F. Pratensis and L. perenne was the lowest when compared to these of other species of Festuca based on RFLP markers. Darbyshire and Warwick (1992) have evaluated the phylogeny of Festuca and related genera using chloroplast DNA restriction site variation and have reported high similarity between F. arundinacea and L. perenne. The closely related Festuca and Lolium genera have also been reported by Stammers et al. (1995).
     In this study, greater inter-specific than intra-spesific variation at the genomic level was observed. For example the only F. ovina accession clearly separated from the other accessions. Although the four F. rubra accessions were grouped in one cluster, AFLP could separate the Iranian F. rubra (P3) accession from the other three Hungarian populations. This reflects the possible role of geographical regions in intra-specific genetic variability.
Two Iranian accessions of F. arundinacea (M9 and G9) did not group with the entries of similar geographic origin and each was placed in a separate cluster. In many molecular systems, the lack of genetic differentiation between accessions of definite identity and distinct geographic origin is usually attributed to the random nature of genomic DNA amplification, which is the case in AFLP (Roldan-Ruiz et al., 2000). This may also be due to the uncultivated nature (lack of artificial selection) and high intr-aspecific genetic variation in the Iranian tall fescue populations.
     Principle coordinate analysis (Fig. 3), in which PC1 accounted for 50.5% of the total variation and PC2 accounted for 28.1%, was generally consistent with results from the cluster analysis in groupings of the species and accessions. The PC2 values for F. rubra and F. ovina accessions were high. These values were medium for accessions of F. pratensis and low for accessions of F. arundinacea.  The values of PC1 for accessions of F. ovina were low, medium for accessions of F. rubra and F. pratensis and high for accessions of F. arundinacea. The FAG9 and to a lesser extent FAM9 were located far apart from the other accessions (Fig. 3). This was very much in accordance with grouping of these two accessions in the clustering (Fig. 2), indicating their greater genetic divergence from other accessions. These accessions may be good candidates for breeding programs in constructing mapping populations or as parents of synthetic varieties.
In conclusion, genetic similarity coefficients between populations showed high levels of inter and intra-specific genetic diversity in the studied germplasm. Cluster analysis reflected the phylogenetic relationships among species and clearly demonstrated differences in the degree of similarity between genotypes. The close molecular homology between some grasses specially Lolium and coarse fescues, may reflect the possibility of producing new valuable germplasm through introgression breeding. In order to improve forage quality and digestibility, tall fescues and meadow fescues are sometimes intercrossed with other related species such as Lolium perenne. (Yamada et al., 2005). The untapped Iranian genetic resources of these grass species may be utilized for similar objectives. There was considerable congruity between AFLP groupings and geographical distribution of accessions. The high level of polymorphism indicated the capacity of AFLP markers in assessment of phylogenetic relationships and genetic variation studies. Genome specific amplified products will be useful molecular markers to discriminate species, verify hybrids and monitor gene introgression for inter-specific and intergeneric hybridization. Iranian tall fescue accessions contain a high degree of genetic variability, are very much diverged from accessions of other geographical regions, and thus can be exploited in breeding programs.