Introduction

Lentil (Lens culinaris Medik.) is an annual food legume and one of the four most important legume crop in the world. It is a self-pollinated diploid plant with 2n = 2x = 14 chromosomes and a genome size of 4 Gbp (Arumuganathan and Earle 1991). This species has been cultivated for 10,000 years in several regions worldwide, especially in West Asia, Australia, North and South America, Mediterranean basin, Middle East, and in the Indian sub-continent (Cubero et al. 2009). Lentil seeds are a good source of human nutrient, containing proteins, carbohydrates, fibers, minerals and antioxidant compounds (Solanki et al. 1999; Boye et al. 2010; Alghamdi et al. 2014; Migliozzi et al. 2015). Moreover, lentil straw is used as high-quality animal feed (Erskine et al. 1990; Devendra 1997; Abd El Moneim and Ryan 2004; Lardy and Anderson 2009). In addition, lentil provides interesting possibilities for sustainable agriculture due to its nitrogen fixing capacity that enable low use of fertilizers in cereal-based cropping system (Prakesh et al. 1986; Shah et al. 2003).

In Morocco, lentil is essentially produced in constraining environments, using landraces maintained by farmers. Landraces are characterized by specific and evolutionary adaptation as well as high nutritional and organoleptic qualities (Benbrahim et al. 2017). However, their low yield implies the development of novel cultivars with high yield potential and resistance to biotic and abiotic stresses. Breeding could therefore lead to reduction or loss of lentil genetic diversity over time (Kumar et al. 2018). Thus, an assessment of plant genetic diversity for an efficient use in both breeding and conservation programs is worth carrying out.

Morphological and isozymes markers were first used for lentil genetic diversity assessment (Ladizinsky 1979; Pinkas et al. 1985; Erskine and Muehlbauer 1991; Ahmad and McNeil 1996). However, those markers are influenced by environmental conditions (Marić et al. 2004). For that reason, molecular markers are being largely used for this purpose. These markers are advantageous as they cover the entire genome, hence their high number, and are unaffected either by environmental factors or by plant development stages. Different molecular markers such as random amplified polymorphism DNA (RAPD), amplified fragment length polymorphism (AFLP), simple sequence repeats (SSR), inter simple sequence repeats (ISSR), sequence-related amplified polymorphism (SRAP), and single nucleotide polymorphism (SNP) has been used to assess genetic diversity of lentil (Abo-elwafa et al. 1995; Toklu et al. 2009; Tewari et al. 2012; Zaccardelli et al. 2012; Joshi et al. 2013; Rana et al. 2013; Alghamdi et al. 2014; Bermejo et al. 2014; Lombardi et al. 2014; Seyedimoradi and Talebi 2014; Khazaei et al. 2016). It is noteworthy that the genetic diversity of Moroccan lentil is not yet well studied. Recently, some Moroccan lentil genotypes, including landraces from five regions and two local released varieties, have been assessed by using SSR and AFLP markers (Idrissi et al. 2015, 2016). These studies, which focused merely on diversity among cultivars rather than within cultivars, revealed moderate to high genetic variations (Idrissi et al. 2015). Thus, the principal purpose of this study was to assess genetic variations among and within Moroccan lentil cultivars using SRAP and RAPD markers, as genetic variations within populations is a fundamental component to guide decision-making by breeders. Certain invaluable information on the basis of this investigation regarding breeding programs is expected.

Materials and methods

Plant materials

A set of nine landraces from different origins in Morocco and eight local registered varieties from National Institute of Agricultural Research (INRA) were assessed (Table 1).

Table 1 List of Moroccan lentil (landraces or varieties) and number of samples used for molecular analysis
Full size table

DNA extraction

Molecular analyses were performed using young leaves collected from lentil seedlings, placed in liquid nitrogen and stored at − 80 °C before lyophilization. Genomic DNA was extracted according to CTAB protocol, described by Lassner et al. (1989). Extracted DNAs from 89 individual plants were evaluated for quality by electrophoresis in agarose gel, and then quantified using Nano − 100 micro spectrophotometer (Xian Yima Opto-electrical Technology, Shaanxi, China).

SRAP analysis

A total of 30 SRAP combinations of five forward (ME) and six reverse (EM) primers (Table 2) (Li and Quiros 2001), purchased in commercially available kits from Eurogentec, were screened for polymorphism on three randomly chosen lentil genotypes. SRAP fragments were generated in a reaction volume of 20 µl containing 1X PCR buffer (BIOLINE, London, UK), 250 µM of each dNTP, 3 mM of MgCl2, 1 U of Taq™ DNA polymerase (BIOLINE, London, UK), 0.5 µM of each forward and reverse primers, and 50 ng of genomic DNA sample. Amplifications were carried out according to Li and Quiros (2001) PCR program, with an initial denaturation at 94 °C for 5 min, followed by 5 cycles of 94 °C, 35 °C, and 72 °C for 1 min each, as denaturation, annealing and extension respectively. Then the annealing temperature was set to 50 °C for other 35 cycles, followed by 5 min of final extension at 72 °C. PCR products were resolved using 6% polyacrylamide gel. After electrophoresis, the gel was carefully removed from the plates and stained in an ethidium bromide solution for 5 min. It was then rinsed with tap water before viewing the bands under ultraviolet light in Molecular Imager® Gel Doc™ XR System.

Table 2 Forward and reverse primers names and sequences used for SRAP analysis
Full size table

RAPD analysis

A set of 62 decamer primers (Online resource 1), purchased in commercially available kits from OPERON (OPERON Technologies, Alameda, CA, USA), were screened on three lentil samples. Primers revealing clear polymorphisms were used on the 89 lentil DNA samples. RAPD-PCR was performed in a final volume of 20 µl containing 25 ng of DNA, 1X buffer (Bioline, London, UK), 250 µM of each dNTP, 3 mM of MgCl2, 1 U of Taq™ DNA polymerase (Bioline, London, UK), and 0.5 µM of primer. PCR amplification was set in a thermocycler Applied Biosystems (Foster City, CA, USA) programmed as followed: 2 min of initial denaturation at 94 °C, followed by 45 cycles of 94 °C for 1 min, 36 °C for 1 min and 72 °C for 1 min with a final elongation at 72 °C for 5 min. Amplification products were fractioned by 1.5% agarose gel electrophoresis. Gels were stained with ethidium bromide before the visualization of amplified fragments using Molecular Imager® Gel Doc™ XR System.

Statistical analyses

RAPD and SRAP profiles were scored as present (1) or absent (0) for each primer or primer combination to generate a binary format matrix. Only clearest and reproducible bands were recorded. Polymorphic Information Content value (PICv) was calculated for each primer according to Roldán-Ruiz et al. (2000), in order to measure the efficiency of polymorphic loci. GenAlex ver. 6.5 software (Peakall and Smouse 2006) was used to perform genetic diversity analysis within landraces by calculating: percentage of polymorphic loci (PPL), expected heterozygoty (He), and Shannon’s information index (I). Analysis of molecular variance (AMOVA) was performed, using data from SRAP and RAPD, to estimate variations within and among landraces. Comparison of genetic distance from SRAP and RAPD markers of all single DNA samples (landraces and varieties) was computed using Mantel test. For cluster and relationships analysis among cultivars, SRAP and RAPD data of all samples were combined in a single matrix. GenAlEx ver. 6.5 software was used to calculate Nei genetic distance among cultivars. Their relationships were visualized through principal coordinates analysis (PCoA), and a dendrogram based on an unweighted pair-group method with arithmetic averages (UPGMA) was constructed on XLSTAT 5.14 software (XLSTAT 2017) using Jaccard’s similarity index. To infer population structure and assign individuals to populations, molecular data were analyzed using STRUCTURE software v.2.3.4 (Pritchard et al. 2000) based on a Bayesian clustering approach. The number of assumed clusters (K) was set from two to eight. Each simulation consisted in 100,000 Monte-Carlo Markov Chain (MCMC) iterations after a burn-in period length of 50,000. The output of STRUCTURE software was then uploaded in the STRUCTURE HARVESTER online tool (Evanno et al. 2005) implementing the Evanno method to identify the most probable structure with the best K value.

Results

A total of eight SRAP primers pairs (ME1-EM3, ME1-EM4, ME1-EM5, ME2-EM6, ME3-EM4, ME4-EM1, ME5-EM3, and ME5-EM5) and nine RAPD primers (OPD 13, OPL 12, OPL 16, OPL 19, OPX 04, OPX 08, OPX 12, OPX 13, and OPX 17) were selected as they generated robust amplification products and showed clear polymorphism.

Characterization and efficiency of SRAP markers

SRAP primer combinations produced 115 fragments among which 113 (98.3%) were polymorphic. The number of polymorphic bands by primer combination ranged from 11 (ME5-EM5) to 18 (ME2-EM6) with an average of 14.1 bands. Six primer combinations (ME1-EM4, ME1-EM5, ME2-EM6, ME3-EM4, ME5-EM3, and ME5-EM5) presented 100% of polymorphic rates. The PICv varied from 0.31 (ME5-EM3) to 0.39 (ME5-EM5) with an average of 0.35 (Table 3).

Table 3 Bands number, polymorphism rate, and PIC values generated with eight SRAP primers pairs and nine RAPD primers on 17 lentil cultivars
Full size table

Characterization and efficiency of RAPD markers

RAPD primers generated 84 polymorphic bands out of 90 identified (93.3%). The number of fragments amplified by each primer varied from five (OPX 12) to 13 (OPL16). Polymorphism rates generated with RAPD primer ranged from 71.4% (OPX12) to 100% (OPD13, OPL12, OPX04, OPX08, OP13, and OPX17). The PICv ranged from 0.24 (OPL 12) to 0.40 (OPX 08), with an average of 0.33 (Table 3).

Genetic diversity of lentil landraces

SRAP and RAPD markers revealed various levels of genetic diversity within and among lentil landraces (Table 4). Considering SRAP markers, the highest values of number of different alleles (Na), number effective alleles (Ne), Shannon’s information index (I), and expected heterozygosity (He) were recorded in LR9, with values 1.26, 1.33, 0.25, and 0.18, respectively. While, the lowest values were observed in LR2 and LR3, with values 1.10, 1.26, 0.20, and 0.14, respectively. Percentage of polymorphic loci (PPL) of landraces with SRAP markers ranged from 31.30 (LR2 and LR3) to 40.0 (LR4). RAPD markers revealed high diversity indexes values (Na, Ne, I, He, and PPL) in LR8 with values 1.23, 1.28, 0.23, 0.16, and 41.1 (Table 4). Combination of data from both markers produced high genetic diversity indices in LR8 and low values in LR2.

Table 4 Genetic diversity indices for each lentil landraces revealed by SRAP and RAPD markers
Full size table

Analysis of molecular variance performed with SRAP and RAPD matrix revealed high variance among lentil landraces (61% and 70% respectively) than within landraces (39% and 30% respectively). Combining data from SRAP and RAPD showed 65% of variation among landraces and 35% of variations within landraces, with a significative PhiPT value (PhiPT = 0.652, p = 0.001) (Table 5).

Table 5 Analysis of molecular variance (AMOVA) of nine landraces based on data from SRAP, RAPD and combined SRAP and RAPD markers
Full size table

Genetic distance

Results of Mantel test showed high correlation among genetic distance revealed by SRAP and RAPD (r = 0.68, p = 0.01). Thus, the genetic distance of 17 tested lentils, based on combined SRAP and RAPD matrix revealed a wide range of values (Table 6). Genetic distance (GD) varied from 0.149 and 0.630. LR1 and LR9 presented the lowest GD, while LR2 and V05 exhibited the highest GD.

Table 6 Pairwise matrix of Nei genetic distance among 17 lentil cultivars, calculated using SRAP and RAPD data
Full size table

Cluster analysis

The Jaccard’s index calculated with data from combined matrix of SRAP and RAPD markers, and used to build an UPGMA tree ranged from 0.503 to 0.969. At 60% of similarity, the dendrogram clustered all cultivars into four main groups, which fit their origin (Online resource 2). Cluster I included four landraces (LR4, LR5, LR9, and LR1) and five varieties (V07, V02, V03, V05, and V04). Cluster II grouped three landraces (LR3, LR6, and LR7) and one variety (V06). Cluster III consisted of individuals from one landrace (LR2) and two varieties (V01 and V08). The last cluster contains only representatives of LR8.

Principal coordinate analysis (PCoA) and population structure

The principal component analysis (PCoA) of 89 lentil samples was performed to establish relationships among individuals. Total variation explained by PCoA was 37.08%, with 14.83% for the first principal component (PCo1), 12.55% for the second (PCo2), and 9.70% for the third (PCo3). Biplot generated by PCoA showed a clear repartition of landraces according to their origins (Fig. 1). Varieties were widely scattered in the plot.

Fig. 1
figure 1

Principal coordinates analysis (PCoA) of 89 lentil samples from 17 lentil cultivars based on combined data from SRAP and RAPD markers

Full size image

Characterization of genetic structure of 89 individuals using STRUCTURE software, following the Bayesian approach, allowed identification of the best K value. In fact, STRUCTURE output extracted with Structure Harvester, detected the highest delta K value with K equals four, suggesting that the most probable structure of tested samples is their distribution into four main populations (Fig. 2). This population structure is associated with a high gene flow within populations and a low gene flow among them.

Fig. 2
figure 2

Genetic assignment of individuals to populations according to the Bayesian method implemented in the program STRUCTURE at K = 4 based on SRAP and RAPD markers

Full size image

Individuals are represented by vertical colored columns. Same color in different individuals indicates that they belong to the same genetic group. Different colors in the same individual indicate the posterior probability to belong to different genetic clusters.

Discussion

Various molecular markers were used in lentil genetic studies, including RFLP, RAPD, SSR, ISSR, AFLP, and SRAP. The choice of molecular marker is a critical step for geneticists and breeders. Our study reports genetic diversity and population structure of Moroccan lentil, including landraces and released varieties, using two PCR-based marker systems. SRAP and RAPD markers cover the entire genome and are easy techniques that can detect DNA polymorphism without knowledge of genome sequence. Whereas RAPD uses a single primer that amplifies DNA arbitrarily (Williams et al. 1990), SRAP detects polymorphism with a primer pair that targets open reading frames (ORFs) (Li and Quiros 2001).

Both molecular markers generated high polymorphism levels, although the average number of fragments per primer, polymorphism rates, and PICv were slightly higher with SRAP primer combinations (14.4, 98.3, and 0.35 respectively) compared to those of RAPD primers (10, 93.3, and 0.33 respectively). This finding is in accordance with several studies reporting the high efficiency of SRAP and RAPD markers (Keify and Beiki 2012; Yin et al. 2014). Polymorphism rates obtained with SRAP primers combinations during this study were consistent with those observed with 35 lentil breeding lines by Alghamdi et al. (2014). However, the polymorphism levels were higher than those obtained by Bermejo et al. (2014) using SRAP on 25 lentil recombinant inbred lines. The average of polymorphism rate (93.3%) generated with RAPD markers was similar to those reported by Tewari et al. (2012) in 83 lentil genotypes and Abo-elwafa et al. (1995) in 36 genotypes of wild and cultivated lentils, reaching 91.3% and 90% respectively. In contrast, low polymorphism rates (54%) were detected in two different studies using RAPD in lentil (Sonnante and Pignone 2001; Yüzbaşıoğlu et al. 2006). In another study including Italian lentil landraces, 57% of bands were polymorphic using AFLP markers (Torricelli et al. 2012). Furthermore, Moroccan lentil expressed moderate polymorphism (54.78%) and PICv (0.3391) with AFLP markers on 51 landraces (Idrissi et al. 2015). Discrepancies of polymorphism indices among studies could be attributed to the informativeness level of selected primers.

Landraces expressed different genetic diversity levels, which varied. The genetic diversity indexes recorded within landraces with SRAP and RAPD markers were relatively close (0.15 and 0.13 respectively). This is in accordance with Idrissi et al. (2015) study in Moroccan landraces where they detected a genetic diversity of 0.13 with AFLP markers. Similarly, Toklu et al. (2009) found a mean of 0.18 after analyzing Turkish lentil landraces using AFLP and ISSR markers. In contrast, higher genetic diversity values were obtained by Tewari et al. (2012) with RAPD and SSR in Indian lentil (0.29, 0.24 respectively) and by Zaccardelli et al. (2012) using SSR markers in some Italian landraces (0.29). Among our landraces, LR2 expressed the lowest genetic diversity with the combined data from SRAP and RAPD markers. This could be the consequence of its geographic origin. LR2 is a landrace collected in Beni Mellal, a region that is geographically distant in comparison with the original regions of other landraces used in this study. Furthermore, this area is characterized by low rainfall which affects negatively the yield. Reduction of the plant yield could probably contribute to the loss of available diversity. This is in accordance with Sonnante and Pignone (2007) who indicated that low genetic diversity of a landrace is a result of a bottleneck effect when grown in an isolated area.

No report of variations within Moroccan landraces is available; previous study focused on variations among landraces rather than within landraces (Idrissi et al. 2015). Results of AMOVA test showed larger variations among than within landraces (65% and 35% respectively). The same pattern was reported by Tewari et al. (2012), who observed that variation among lentil accessions (63%) were higher than within them (37%). This suggests that landraces have a very low overlap, which may lead to a low gene flow among them. Furthermore, since lentil is a self-pollinating plant and farmers generally maintain their own seeds, reduction in variations within population are observed over time, resulting in an increase of inter-population variation. This is generally the case in Moroccan lentil, which is essentially produced by farmers using their own landraces. In this context, LR1 landrace was labelled with a Geographical Indication (GI) in 2015 because of its seeds nutritional quality that is associated to specific environment and agricultural practices of its production (Benbrahim et al. 2017).

Variable genetic distances among lentil cultivars were recorded in this study. Results revealed that landraces with the lowest distance (LR1 and LR9: 0.149; LR4 and LR5: 0.168; LR3 and LR8: 0.193) are cultivated in adjacent geographic regions. This suggests a possible existence of seed movement among these areas, unlike populations of distant geographic origins (LR2 and LR8; LR2 and LR3), which exhibited higher genetic distances (0.546 and 0.508 respectively).

PCoA and cluster analysis displayed relationships among individuals and grouped landraces into four clusters. STRUCTURE analysis consolidated this result, and confirmed that population genetic structure of tested landraces is essentially based on their geographic origin. It also highlighted high gene flow within populations, especially between LR1 and LR9 (from Zaer), LR4 and LR5 (from Gharb), LR3, LR6, LR7 and LR8 (from Rif and Saïs), and a high inter-population differentiation among them. This result is in accordance with the positive and significant correlation between the genetic and the geographic distances, calculated with Mantel test (r = 0.673, p = 0.01). Separation of landraces could be the result of a long evolution process caused by various climatic conditions of the regions where they are cultivated, associated with agricultural practices. Geographically isolated populations accumulate genetic differences as they acclimatize themselves to different environments (Tewari et al. 2012). This is consistent with the studies that established, immigration rates in plants are inversely proportional to the distance between populations and suggested that the most common gene flow pattern is isolation by distance (IBD) (Sexton et al. 2014). The genetic variation among the plant material studied reflects varied responses observed by Mbasani-Mansi et al. (2019) while assessing its resistance to Orobanche crenata, a parasitic plant that is threatening to lentil production. These authors detected some sources of resistance to include in a breeding program.

The limited number of populations revealed by the structure analysis in this study can help to select lentils to include in a core collection for conservation in Moroccan lentil gene bank. A core collection is important, as it allows working with fewer accessions, representative of a crop species with minimum repetitiveness (Arber et al. 1984; Khazaei et al. 2013).

Improved varieties were not affected to a unique cluster; however, most of them were grouped in cluster I. Results of our investigation also confirmed the origin of V07 (L56) that is a released cultivar from LR9 (Precoce), as supposed by Benbrahim et al. (2017). Similar result was reported by Idrissi et al. (2015) who used combined data sets from the SSR and AFLP analyses and clustered this variety into the same group with a landrace from Zaer.

The genetic diversity detected in the plant material studied is consistent with varying responses it expressed to Orobanche crenata infection.

Conclusion

This study reports population structure and genetic diversity among and within a Moroccan lentil collection using SRAP and RAPD markers. These makers revealed high polymorphism rates and PIC values. Combination of the two markers targeting different regions of the genome provided adequate information on the genetic diversity among genotypes. Within landraces variation was lower than among landraces. PCoA and cluster analysis displayed groups based on their geographic origins. The genetic diversity and the differentiation among populations associated with characterization of this plant material will offer breeders opportunities to develop new varieties with beneficial agronomic traits to cope with climate changes and biotic stresses. In parallel, ex situ conservation of local landraces should be a priority. In this study, structure analysis indicated a low gene flow among populations. Regrouping the studied genotypes into reduced number of populations consolidates the usefulness of structure analysis while managing conservation strategies and developing core collections.