Assessment of genetic variation and species relationships in a collection of Lens using RAPD and ISSR

The genetic variation of a collection of twenty-two Lens accessions was assessed using RAPD and ISSR markers. The collection included accessions of the cultivated lentil, Lens culinaris ssp. culinaris, and its wild ancestor L. c. ssp. orientalis, and the other wild species of the genus: L. odemensis, L. ervoides, L. nigricans, L. tomentosus, and L. lamottei. Both types of markers produced a relatively high number of polymorphic markers in the whole collection, although the degree of variation within the cultivated materials was lower. ISSR markers produced on average more bands and useful polymorphisms than RAPD markers. The Fitch-Margoliash dendrogram, based on Jaccard indices taking jointly RAPD and ISSR into account, showed that the cultivated materials were grouped according to their macroand microsperma type and their geographical origin, and was compatible with the hypothesis of the existence of six different species in the genus Lens, with L. tomentosus being the closest species to L. culinaris.


Introduction
Lentil is included in the cool season food legume group of pulses and is an important source of protein and fiber in human diet.Moreover, they are also valuable as feed and fodder for livestock, and play an important role in crop rotations because their nitrogen fixing capability.This crop was domesticated in the Fertile Crescent where it has been cultivated since at least the seventh century B.C. (Ladizinky, 1979), and its cultivation area expanded around the Mediterranean Basin, Middle East, Ethiopia and the Indian Subcontinent.The species is currently an important pulse grown in many areas of temperate climate such as the Mediterranean Basin, Central and Western Asia, and some areas in North and South America and Australia, where the crop is grown in semiarid regions, usually in rotation with cereals.
The species Lens culinaris Medik. is included in the genus Lens Miller, tribe Vicieae, subfamily Papilionacea, family Leguminosae.The tribe Vicieae includes another three additional genera of agronomic interest,
(synonymous, Lens esculenta Moench), is classified on the basis of seed size as macrospermas or microspermas (Barulina, 1930), categories which have been considered as subspecies, races or varieties by different researchers and breeders.
Lentil is a naturally self-pollinated species due to its cleistogamous flowers, with a level of natural crosspollination amounting to as low as 0.8% (Wilson and Law, 1972).Therefore, most of the individuals of this species are homozygous for most, if not all, of their genes, although wild populations and cultivated landraces can be formed by a combination of different genotypes (Alvarez et al., 1997).
One of the most widely used DNA markers are RAPDs (random amplified polymorphic DNA), generally obtained by the random amplification of DNA sequences using 10-mer primers (Pérez de la Vega, 1997).Intersimple sequence repeats (ISSR) are DNA markers which involve the direct use of microsatellite sequences as primers in the PCR (Gupta et al., 1994).According to Zietkiewicz et al. (1994) and Ratnaparkhe et al. (1998) this technique is more reliable than the RAPD technique and generates larger numbers of polymorphisms per primer.ISSRs have been used in many crop species including legumes (Ratnaparkhe et al., 1998;Bornet and Branchard, 2001;Iruela et al., 2002;Rajesh et al., 2003), and have proved their usefulness in genetic mapping and marker assisted selection (Zietkiewicz et al., 1994;Ratnaparkhe et al., 1998;Rubeena et al., 2003).ISSR markers have already been used in lentil to evaluate genetic variation in collections of cultivated lentils (Závodná et al., 2000;Sonante and Pignone, 2001).

Plant material and DNA extraction
The materials used in this work are listed in Table 1.They include modern cultivars, land races and samples of wild species, hereafter all of them referred to as accessions.Since we were mainly interested in the degree of differentiation between accessions, we pooled equal amounts of DNA from each of 4 seedlings per sample to amplify RAPDs and ISSRs.This was based on the previous knowledge that the level of variation within accessions was relatively low in cultivated lentils (Alvarez et al., 1997), and in that this method of pooling DNA is an appropriate strategy in assessing genetic variability in plant germplasm collections (Yu and Pauls, 1993;Gilbert et al., 1999).DNA was extracted from fresh leaf tissue of 15-dayold seedlings as described by Edwards et al. (1991) with minor modifications.

Random amplified polymorphic DNA (RAPD) analysis
A total of 160 decamer primers (Operon Technologies, Alameda, Calif., USA; Kits A, B, C, D, G, H, P and W) were assayed, and 36 of them which produced easily observable and repeatable fragments were used in further assays.PCR amplifications were carried out following the procedure described by Abo-Elwafa et al. (1995) for lentil.RAPD products were separated on 1.8% agarose gels and bands were visualized with ethidium bromide.

Inter-simple sequence repeats (ISSR) analysis
One hundred ISSR primers (University of British Columbia Biotechnology Laboratory, primer set #9) were used to screen for polymorphisms.Finally, 14 primers which produced amplification patterns with the highest number of intense bands were chosen.The ISSR technique was performed as described by Ratnaparkhe et al. (1998) with minor modifications as follows: the reaction mixture (25 µl) contained 40 ng of template DNA, 10 pmol of oligonucleotide primer, Taq-DNA polymerase buffer (20 mM Tris-HCl, pH 8.4; 50 mM KCl), 2 mM of MgCl 2 , 0.25 mM each of the four dNTPs and 1 U of Taq DNA Polymerase (GIBCO BRL).The DNA was first denatured for 1 min at 94°C, followed by 30 cycles of 30 s denaturation at 94°C, 30 s annealing at 50°C and 2 min elongation at 72°C, with a final elongation of 10 min at the same temperature.The reaction products were separated in 2.0% agarose gels and bands were visualized with ethidium bromide.

Statistical methods
The relationship between accessions was estimated using the Jaccard's similarity index (SI) (Sneath and Sokal, 1973), and transformed in distance as 1-SI.
Dendrograms were obtained with the Fitch and Margoliash (1967) method.Bootstrap resampling (Efron and Gong, 1983) was used to estimate branchlength standard errors and the support values of nodes.

Results and Discussion
The estimation of the degree of differentiation between materials included in a crossing program is useful since it can help in selecting the different parents.However, for this purpose, the use of a high number of markers represents a considerable amount of work.We decided to use two types of dominant markers (RAPD and ISSR) to estimate the variation and the differences between a set of lentil material included in a crossing program and some samples of wild materials, some of which are also included in the crossing program.Both techniques yielded a high number of markers ( between 300 and 2,000 bp for RAPDs, and between 400 and 4,000 bp for ISSRs.The average of RAPD polymorphic markers per primer was 8.1, which is almost twice higher than the values described by Abo-Elwafa et al. (1995) and Sharma et al. (1995).This difference is most probably due to a more adequate selection of primers.On average ISSR markers produced more markers and more polymorphisms per primer than RAPD markers (Table 2).Thus, this result supports the previous reports of the suitability of ISSR in the assessment of plant genetic variation (Zietkiewicz et al., 1994;Ratnaparkhe et al., 1998).ISSR markers have been used to evaluate genetic variation within collections of cultivated lentils (Závodná et al., 2000;Sonante and Pignone, 2001), but they have not been used in the assessment of the variability between wild Lens materials.The polymorphisms generated either by RAPDs or ISSRs were enough to differentiate each accession.The level of polymorphism is much lower in the cultivated materials than in the wild ones, as indicated by RAPD markers in previous works (Abo-Elwafa et al., 1995;Sharma et al., 1995;Alvarez et al., 1997) and in this work.For instance, only a 40% of the RAPDS were polymorphic in Spanish lentil accessions (Alvarez et al., 1997).
Both kinds of markers were used to estimate the similarity between accessions by the Jaccard similarity index.For RAPDs the similarity values ranged from 0.909 between Mosa and Castellana, two Spanish macrosperma samples, to 0.208 between Lens nigricans and L. ervoides; for ISSRs the values ranged from 0.929 between Pardina and Verdina, two microsperma samples from León (Spain), to 0.157 between L. nigricans and the sample ILWL 39 of L. odemensis.The correlation between the similarity value matrices calculated for RAPDs and for ISSRs was 0.969 (p < 0.001).This high correlation value points out that both kinds of markers are probably estimating the same type of genetic variability within the lentil genomes.The average RAPD and ISSR similarities between cultivated lentil accessions were similar, 0.768 and 0.758, respectively (Tables 3 and 4).The similarities between cultivated lentil for RAPDs ranged from 0.909 (between Mosa and  Castellana) to 0.586 (between Pardina and ILL 5588), and for ISSRs from 0.929 (between Pardina and Verdina) to 0.615 (also between Pardina and ILL 5588, as for RAPDs).The lowest similarities observed between samples of cultivated lentils were always higher than any similarity between the cultivated lentils and accessions of the wild ancestor L. culinaris ssp.orientalis.Distances were used to obtain dendrograms.First, RAPD and ISSR distances were used separately obtaining dendrograms with similar clustering (not shown).In both cases there was a perfect discrimination at the sub-specific and specific levels.Considering the cultivated materials, Lens culinaris ssp.culinaris, RAPD marker data generated two clear clusters, one including the microsperma accessions and the other the macrosperma accessions, while ISSR marker data differentiated Spanish lentil accessions (both, from Continental Spain and from the Canary Islands) from non-Spanish accessions.The ability of RAPDs to discriminate between microsperma and macrosperma accessions has already been pointed out by Alvarez et al. (1997).Figure 1 shows the dendrogram obtained considering RAPD and ISSR marker data together.It can be observed that species and subspecies are differentiated in the dendrogram, that cultivated materials are clearly differentiated from the wild accession of L. c. ssp.orientalis, and that both culinaris subspecies are grouped in the same cluster together with L. tomentosus, the latter being a species morphologically similar to L. c. ssp.orientalis (Ferguson et al., 2000).The cultivated materials were grouped according to their macro-and microsperma type and their geographical origin (Fig. 1), although this result must be considered as provisional because some branch lengths were not significantly different from 0 and the values of some nodes were low.Ferguson (2000) pointed out that conflicting specif ic relations have emerged which appear to depend on the accessions and the techniques used in measuring genetic variation.The clustering of L. tomentosus with L. culinaris agreed with previous reports pointing out the taxonomic proximity between both species (Van Oss et al., 1997;Ferguson et al., 2000).Morphologically L. lamottei is closely related to L. odemensis, it is practically equally associated with L. odemensis and L. culinaris on the basis of isozyme data, but it appears to be the most distantly related taxon to all other Lens species according to RAPDs (Ferguson, 2000).However, our data taking into account RAPD, ISSR or both together, clearly differentiated L. lamottei from L. culinaris, even if it always appeared as a sister species of L. odemensis.Although the dendrogram is unrooted (Fig. 1), our data suggest that L. nigricans is the most differentiated taxon in the genus, according to previous results based on chloroplast DNA and other polymorphisms (Van Oss et al., 1997;Ferguson, 2000).
To sum up, both RAPD and ISSR markers contribute with a significant number of polymorphic markers which could be useful in identifying lentil cultivars, contributing to saturate genetic maps, in markerassisted selection, but which also could contribute useful data in phylogenetic analyses in the genus Lens.

Figure 1 .
Figure 1.Dendrogram obtained using the Fitch-Margoliash method on the basis of Jaccard distances calculated considering jointly RAPDs and ISSRs.Figures on the nodes indicate the bootstrap values.Branches marked with and asterisk did not have a significantly different length from 0 with p < 0.05.All the samples inside box A are microspermas from mainland Spain, those in B are macrospermas from mainland Spain, those in C are microspermas from the Canary Islands, and those in D are microspermas from Lithuania.

Table 2 )
with a high degree of polymorphism in the whole collection: 91.3% of the RAPD bands and the 98.8% of the ISSR bands were polymorphic.Band sizes ranged 540 Y. Durán and M. Pérez de la Vega / Span J Agric Res (2004) 2(4), 538-544

Table 1 .
Lens materials used M: macrosperma.m: microsperma.Samples: 1 from the Spanish Germplasm Bank, 2 from the Germplasm Bank of the International Centre for Agricultural Research in Dry Areas (ICARDA), 3 kindly provided by Prof. G. Ladizinsky.

Table 2 .
Polymorphism of the two types of molecular markers

Table 3 .
Average similarity between accessions within and between species based on RAPDs Standard deviation in parentheses.NA: not applicable as there was a single accession.

Table 4 .
Average similarity between accessions within and between species based on ISSRs Standard deviation in parentheses.NA: not applicable as there was a single accession.