Abstract

Agromorphological traits have immense importance in breeding lentils for higher yield and stability. We studied the genetics and identified map positions of some important agro-morphological traits including days to 50% flowering, plant height, seed diameter, 100 seed weight, cotyledon color, and growth habit in Lens culinaris. Earlier developed RILs for stemphylium blight resistance (ILL-5888 × ILL-6002), contrasted for those agro-morphological traits, were used in our study. Three QTLs for days to 50% flowering were detected with additive and epistatic effects. One QTL for days to 50% flowering, QLG483 (QTL at linkage group 4 at 83 cM position), accounted for an estimated 20.2% of the variation, while QLG124 × QLG1352 and QLG484 × QLG138 accounted for 15.6% and 24.2% of the variation, respectively. Epistatic effects accounted for most of the variation in plant height, but the main effect of one QTL, QLG84, accounted for 15.3%. For seed diameter, three QTLs were detected, and one QTL, QLG482, accounted for 32.6% of the variation. For 100 seed weight, five QTLs were identified with significant additive effects and four with significant interaction effects. The main effect of one QTL, QLG482, also accounted for 17.5% of the variation in seed diameter. QLG which appears to affect days to 50% flowering, seed diameter, and 100 seed weight is flanked by RAPD markers, UBC 34 and UBC1. Growth habit and cotyledon color are controlled by single genes with prostrate dominant to erect and red cotyledon dominant to yellow. The QTL information presented here will assist in the selection of breeding lines for early maturity, upright growth habit, and improved seed quality.

1. Introduction

Agromorphological traits have immense importance in crop breeding. Crop adaptation, field performance, market value, and demands for specific uses are major factors that drive breeding goals. Linkage maps and QTL analysis are valuable tools for plant breeders to improve breeding efficiency by tagging genes with markers and analyzing the association between markers and traits. The inheritance of quantitative traits and tagging genes such as days to 50% flowering, plant height, seed diameter, and seed weight and qualitative trait genes such as growth habit and cotyledon color in lentil (Lens culinaris Medikus subsp. culinaris) will help breeders in the selection process and understanding interrelationships among traits.

Lentil is quantitative long-day plant flowering in progressively longer days [1]. Sometimes it is hard to determine days to maturity at or near the end of a crop season due to weather conditions. It has been reported that soybean (Glycine max) breeders select lines for optimum maturity based on days to flowering [2]. Sarker et al. [3] reported that flowering time is sensitive to photoperiod and temperature, and that a more complete understanding of genetic control of flowering time in lentil is needed. Roberts et al. [1] proposed four developmental phases of flowering as: preemergence, preinductive, inductive, and postinductive. When lentil plants were transferred from short days (either 8 or 10 h) to long days (16 h), or vice versa, the first two phases and the last are insensitive to photoperiod but are probably sensitive to temperature. So, it is evident that a complex phenomenon controls days to flowering in lentil.

Plant height of lentil ranges from 25 to 30 cm for the majority of genotypes but may vary from extremes of 15 to 75 cm depending on genotype and environment [4]. Most lentils grown in South Asia, Middle East, and Africa are land races. These are generally short in plant stature, prostrate in growth habit, lack uniformity in pod maturation, have a high incidence of pod shattering, and are low yielding. Tall upright lentil cultivars with high basal pod positions are always preferred by farmers for mechanical harvesting [5]. Sakar [6] reported that three genes were responsible for variation in plant height from the cross of the two lentil cultivars, Laird and Precoz. Tullu et al. [7] reported that plant height is a polygenic trait and QTL detected for plant height has environmental effect.

Growth habit has received a great deal of attention from breeders attempting to develop cultivars with more upright stature that are also lodging resistant and adaptable to mechanized harvesting. Ladizinsky [8] indicated that a single gene with incomplete dominance controls growth habit in lentil.

Uniformity of seed size, shape, and color is important for marketing of lentil. A wide range of lentil cultivars are used throughout the world. Small diameter red cotyledon types account for most of the lentil production followed by the large-seeded and small-seeded yellow cotyledon types. Lentil seeds are lens shaped and generally weigh between 20 and 80 mg. Their diameter generally ranges from 2 to 9 mm. Seed size differs according to genotype, and researchers frequently follow the classification of Barulina [9] who grouped lentils as macrosperma with large seeds that range from 6 to 9 mm in diameter and microsperma with smaller seeds that range from 2 to 6 mm in diameter. It has been reported that dry seed weight of lentil is controlled by two genes [6], whereas polygenic control of seed weight was reported by Abbo et al. [10]. Cotyledon color of lentil can be red/orange, yellow, or green. Large green lentils with yellow cotyledons are marketed to countries of southern Europe, particularly Spain, Italy, and Greece, and small red cotyledon types are exported to South Asia and the Middle East [11]. The first report of genetics of lentil cotyledon color was studied by Tschermak-Seysenegg et al. [12, 13] and by Wilson et al. [14] and confirmed that cotyledon color is controlled by a single gene and red/yellow cotyledon is dominant over yellow. Singh [15] and Slinkard [16] reported that red cotyledon color is completely dominant over green and yellow. Sharma and Emami [17] detected monogenic and digenic control of cotyledon coloration in lentil.

The objectives of the present study were to identify regions of the lentil genome associated with agromorphological traits including days to 50% flowering, plant height, growth habit, seed diameter, seed weight, and cotyledon color that could be utilized in marker-assisted breeding and to improve our understanding of the genetics of these traits.

2. Materials and Methods

2.1. Development of Inbred Lines

In order to identify and map the agromorphological trait genes, a lentil mapping population (F6 derived F7 recombinant inbred lines (RILs)) that was previously developed to determine the genomic locations of the genes for stemphylium blight (caused by Stemphylium botryosum Wallr.) resistance was used [18]. The mapping population is comprised of 206 RILs from the cross ILL-5888 (“Uthfola” the popular name and described as a glex pilosae microsperma type; short stature with prostrate growth habit) by ILL-6002 (developed as a pure line selection from Argentinian variety, Precoz, macrosperma type; tall and erect in growth habit). These parents were contrasting for the agromorphological traits (quantitative and qualitative traits) under study.

2.2. Phenotyping

The parents and the mapping population of 206 RILs were grown in silty loam soil at the Washington State University Spillman Agronomy Farm near Pullman, Washington (46°73′N latitude and 117°73′W longitude), USA, in the 2007 and 2008 cropping year. The experimental design was a randomized complete block with three replications. Individual plots were single rows 1 meter long and spaced 60 cm apart and within row plant spacing of approximately 3 cm. Flowering was recorded as number of days from planting to 50% of the plants in the plot with at least one open flower. Plant height was taken at the 50% flowering stage and measured from the soil surface to the tip of the central axis. Seed diameter was measured by using sliding calipers, and 100 seed weight was taken using a Mettler digital balance. Growth habit data were taken based on plant canopy spreading, and cotyledon color was determined visually. All the qualitative and quantitative trait data taken were used for statistical analysis. The combined analysis of variance (ANOVA) of the quantitative traits was done using SAS 9.1.

2.3. Genotyping, Linkage Analysis, and QTL Mapping

DNA extraction and genotyping using simple sequence repeats (SSR), randomly amplified polymorphic markers (RAPD), and sequence-related amplified polymorphic (SRAP) markers were performed following protocols described in Saha et al. [18]. Mapmaker Macintosh V2.0 was used for linkage analysis, and QTL analysis for quantitative traits were conducted following inclusive composite interval mapping (ICIM) method with the following software: QTL IciMapping v2.2 [19] and Q gene 4.2.3 [20] as described by Saha et al. [18]. The intraspecific linkage map was developed with an LOD score of 4.0 and at maximum recombination fraction 0.25. The ICIM methods were used to detect QTLs for the quantitative traits due to the greater efficiency of detection. This method uses stepwise regression models for identifying the significant flanking markers and one-way scanning to identify additive effect and two-way scanning to identify digenic epistatic effect [19].

3. Results and Discussion

The agromorphological data of two parents (ILL-5888 and ILL-6002) and 206 RILs were collected at Spillman Agronomy Farm near Pullman in 2007 and 2008 cropping years. The combined ANOVA showed highly significant differences between the parents and among RILs for the four quantitative traits analyzed (Table 1). Highly significant ( ) correlations exist between seed diameter and 100 seed weight, and days to 50% flowering and seed diameter and 100 seed weight (Table 2). Plant height was highly significantly correlated with 100 seed weight ( ) and significantly correlated with days to 50% flowering and seed diameter ( ) (Table 2). The photoperiod sensitivity between two parental lines was also observed due to high genetic differences (data not shown).

To map the genes, an intra-specific lentil map of 14 linkage groups comprised of 23 SSRs (simple sequence repeats), 30 RAPDs and 108 SRAPs (sequence-related amplified polymorphisms) and two morphophysiological markers (cotyledon color and growth habit) was used [18].

3.1. Qualitative Trait Loci
3.1.1. Cotyledon Color

ILL-5888 and ILL-6002 have red and yellow cotyledon colors, respectively. The RILs segregated into a 1 : 1 ratio of red to yellow (Table 3). The clear 1 : 1 segregation ratio of the RILs was consistent with reports of the inheritance of cotyledon color by Tschermak-Seysenegg et al. [12, 13] and Wilson et al. [14]; that is, cotyledon color is controlled by a single gene. All F1 seeds were red and the F2 segregated in a 3 : 1 ratio of red to yellow (99 : 32), indicating that red cotyledon is completely dominant over yellow cotyledon. Singh [15] and Slinkard [16] also reported that red cotyledon color is completely dominant over green and yellow. The cotyledon color gene (Yc) was positioned on LG8 and flanked by RAPD marker UBC40b and SSR marker GLLC511a at 16.4 cM and 13 cM from Yc, respectively (Figure 1).

3.1.2. Growth Habit

ILL-5888 is a prostrate variety while ILL-6002 has an erect growth habit. The RILs segregated in a 1 : 1 ratio (Table 3). Based on segregation of the RILs, it can be inferred that a single gene controls plant growth habit in this population. The F1s (ILL-5888 × ILL-6002) had a prostrate growth habit indicating that prostrate is dominant over erect. The gene for growth habit (Gh) was located at the distal end of LG9 at 77.8 cM position and 35.1 cM apart from the F18XR9b SRAP marker, which is considered as loosely linked with the gene (Figure 1).

Ladizinsky [8] made crosses within and between lines of L. culinaris and L. orientalis that differed for growth habit: erect tall with few branches, erect bushy with many branches, and prostrate. He reported that prostrate growth habit was incompletely dominant over erect growth habit. Emami and Sharma [21] also showed that prostrate is dominant over erect growth habit. On the contrary, Kumar [22] and Mishra [23] reported erect growth habit as completely dominant over prostrate type. These suggest that multiple alleles may confer growth habit, so different inheritance patterns may be observed in different crosses.

3.2. Quantitative Traits and QTLs

The QTLs and epistatic interactions were identified in different linkage groups, but one of the QTLs, QLG (QTL at linkage group 4 at 82-83 cM), accounted for 20.2%, 32.6%, and 17.5% phenotypic variation for days to 50% flowering, seed diameter, and 100 seed weight, respectively.

3.2.1. Days to 50% Flowering

ILL-5888 flowered in 53 days compared to 60 days for ILL-6002. The range for days to 50% flowering among RILs was 47 to 58 days with a mean of days (Table 1). The frequency distribution of the RILs and the normality test (Shapiro-Wilk, ) of days to 50% flowering showed two discrete classes and the absence of a normal distribution pattern (Table 1, Figure 2(a)).

The ICIM analysis for days to 50% flowering (Table 4) showed the presence of additive and epistatic interaction effects that accounted for a significant amount of phenotypic variation. LODA determines the significance of total variation at the interaction data point, whereas the LODAA value indicates epistatic effects between QTLs.

Three significant QTLs were detected with two on linkage group 4 and one on linkage group 13. The QTL detected at the 83cM position of linkage group 4 (QLG483) showed significant additive effects (1.506) that accounted for 20.2% of the phenotypic variation (Table 4). The flanking markers, UBC34 and UBC1, were 1.4 and 1.6 cM from QLG483 (Figure 1). The other two QTLs, QLG47 and QLG1350, accounted for 14.4% and 10.4% of the variation, respectively (Table 4). QTL, QLG47, was 7 cM from the GLLC 556 marker on linkage group 4, and QLG1350 was flanked by two SRAP markers, ME5XR7b and F8XEM58b, on linkage group 13 separated by 2.9 cM and 7.2 cM, respectively (Figure 1).

The and interactions accounted for 24.2% and 15.6% of the phenotypic variation, respectively, and were due to combined additive and epistatic effects. Five other QTL interaction pairs had significant epistatic effects that accounted for 4.6 to 8.0% phenotypic variation each, but the additive effects of the individual QTLs were nonsignificant (Table 4). It should be noted that the interaction probably involves the main effect of QTL QLG483, and the interaction probably involves the QTL QLG1350.

Study of the inheritance of flowering in Precoz under both Indian and Syrian environments determined that a dominant gene Sn played a major role in early flowering [3]. Tahir and Muehlbauer [24] identified four QTLs on different linkage groups, and Sarker et al. [3] reported single and polygenic systems control of days to 50% flowering based on F2 segregation in different lentil crosses. Tullu et al. [7] identified two QTLs (LG4 and LG12) from one location data and five QTLs (LG1, LG4, LG5, LG9, and LG12) from another location data for days to 50% flowering.

We could not compare our QTLs for days to flowering with the QTLs identified by Tullu et al. [7] because the UBC and SSR markers they were using were not polymorphic in our population. But all the QTL mapping studies on lentil indicate the presence of more than two QTLs for days to 50% flowering. In our study, three significant QTLs were detected, QLG483, QLG47, and QLG1350, that have additive effects and accounted for 45% of the phenotypic variation, and one QTL, QLG483 alone accounted for 20% of the phenotypic variation. The and interactions accounted for 24.2 and 15.6%, respectively, of the variation through combined additive and epistatic effects. We found that the 83-84 cM position on LG 4 and the 50–52 cM position on LG 13 contained major QTLs and accounted for significant phenotypic variation for days to 50% flowering.

3.2.2. Plant Height

The RILs varied in height from 15.7 cm to 34.3 cm with a mean of (Table 1). Frequency distribution and the normality test for plant height showed the presence of a normal distribution (Shapiro-Wilk, ) and indicated polygenic control (Table 1, Figure 2(b)).

A significant QTL (LOD = 4.9) was detected on LG 8 at the 4 cM position (QLG84), and the closest marker, SSR204b, was 4 cM away at the proximal end. QTL QLG84 accounted for an estimated 15.3% of the phenotypic variation for plant height (Table 5, Figure 1).

The epistatic effects of and accounted for 17.6 and 17.3% phenotypic variation, respectively, but additive effects of the interacting QTLs were insignificant (Table 5). The interaction of accounted for an estimated 14.6% of phenotypic variation by combined additive and epistatic gene action (Table 5). Six other pairs of QTLs displayed epistatic interactions each affecting 6.8 to 10.7% of the phenotypic variation (Table 5).

Epistatic effects of QTLs accounted for a major portion of the phenotypic variation for plant height in our population. Genes close to F13XR8 and Yc markers played a major role for the interactions, whereas, for the interactions, genes close to the same SRAP marker F13XR8 and GLLC 556 accounted for a significant portion of the variation (Table 5, Figure 1).

Tahir and Muehlbauer [24] reported that a gene linked to the Aat-p locus was responsible for increased plant height. Tullu et al. [7] reported that PI 320937 is taller than “Eston” but both contributed to reduced height of lentil plants, and they identified different QTLs of plant height at two locations, though there is a nonsignificant genotype × environment interaction. They concluded that the chance of simultaneous detection of QTLs at both environments was small due to lack of powerful statistical methods to detect minor QTLs.

3.2.3. Seed Diameter

The parents, ILL-5888 and ILL-6002, had seed diameters of 3.7 mm and 6.5 mm, respectively, and all the RILs were intermediate to the parents. The highest and the lowest seed diameters of the inbred lines were 4.0 and 6.3 mm, respectively, with a mean of  mm (Table 1). Frequency distribution of seed diameter of the RILs is bimodal but continuous (Figure 2(c)) indicating that a single major gene or QTL along with minor QTLs was involved in determining seed diameter. Three different QTLs, QLG1112, QLG482, and QLG598, were detected for seed diameter on LG1, LG4, and LG5, respectively. QLG482, at the 82 cM position of linkage group 4, accounted for 32.6% of the phenotypic variation through large additive effect of 0.293 with an LOD score of 22.2 (Table 6). The QLG482 QTL has the most significant effect by far and may be responsible for the bimodal frequency distribution of the mapping population observed for this trait (Table 6, Figure 2(c)). Two flanking markers, UBC34 and UBC1, were 0.4 cM and 2.6 cM from the QTL, respectively (Figure 1). The other two QTLs, QLG1112, and QLG598 accounted for 4.5 and 3.6% of the phenotypic variation for seed diameter, respectively (Table 6). Significant epistatic interactions of , , and with relatively minor effects were observed (Table 6).

Cubero [25] reported that a polygenic system governs seed shape and size in faba bean (Vicia faba L.). In chickpea (Cicer arietinum L.), small seed size was found to be dominant in desi × kabuli crosses [26], but Niknejad et al. [27] found polygenic control of seed size and partial dominance for large seededness. Davies et al. [28] reported that three genes control the seed size in peas (Pisum sativum L.). In our study, one major QTL with two minor QTLs and three minor epistatic interactions indicate that genetic control of seed diameter in lentil is controlled by a combination of genetic effects.

3.2.4. 100 Seed Weight

Seed weight is a major yield component. The 100 seed weight of ILL-5888 was 2.16 g and ILL-6002 was 5.59 g. The range of 100 seed weight of the RILs was 2.0 to 4.9 g with a mean of  g (Table 1). As with seed diameter, the 100 seed weight of the RILs was intermediate to the two parents. The frequency distribution showed a skewed distribution toward light seed (Figure 2(d)). Five QTLs for 100 seed weight were identified on four linkage groups (LG1, LG4, LG5, and LG8) that additively accounted for 5.6 to 17.5% of the phenotypic variation. Two QTLs, QLG482, and QLG1113 accounted for 17.5% and 12.8% of the phenotypic variation with an LOD score of 15.3 and 11.5, respectively (Table 7). QLG482 was flanked by UBC34 and UBC1 at 0.4 and 2.6 cM, respectively, and QLG1113 was flanked by UBC38b and UBC 24a at 1 cM and 7 cM, respectively (Figure 1). The epistatic interactions of , , , and accounted for a significant portion of the phenotypic variation, each accounting for about 9% of the variation (Table 7).

Genetics of seed weight of common bean (Phaseolus vulgaris) has been under investigation since the early studies of Johannsen [29]. According to Motto et al. [30], based on a classical quantitative genetics study, seed weight of common bean is quantitatively inherited and affected by at least ten genes with additive effects. In mung bean (Vigna radiata), seed weight is controlled by genes with additive and nonadditive actions and low seed weight is dominant [31]. It has been reported that dry seed weight of lentil is controlled by two genes [6]. Abbo et al. [10] found that seed weight of lentil is under polygenic control with additive and dominant gene action and partial dominance of low seed weight alleles. In our study, two QTLs accounted for relatively high levels of variation supporting the results of Sakar [6], and the frequency distribution was skewed toward low seed weight and indicated the polygenic nature of the control of seed weight and partial dominance for low seed weight, the results supporting Abbo et al.’s [10] findings.

4. Conclusion

Agromorphological traits have immense importance in breeding lentils for high yields, yield stability, and market acceptability. The ICIM (inclusive composite interval mapping) method opens the door for understanding quantitative inheritance with epistatic interactions. Now, it is possible to identify digenic interactions while developing polygenic interaction models to improve the efficiency and accuracy of QTL detection. Taking these interactions into account, it will be possible to formulate breeding and selection procedures for important agronomic and market value traits.

We report here the detection of three significant QTLs (QLG47, QLG483, and QLG1350) for days to 50% flowering, one significant QTL (QLG84) for plant height, three significant QTLs (QLG1112, QLG482, and QLG598) for seed diameter, and five significant QTLs (QLG1107, QLG1113, QLG482, QLG555, and QLG834) for 100 seed weight.

QLG accounted for 20.2%, 32.6%, and 17.5% of the phenotypic variation for days to 50% flowering, seed diameter, and 100 seed weight, respectively. QLG was flanked by two RAPD markers UBC 34 and UBC 1 at 0.4–1.4 and 1.6–2.6 cM, respectively. The three traits involved are all positively correlated with each other indicating the possibility for linkage or pleiotropic effect in this LG 4 QTL region.

Some interacting QTLs accounted for significant phenotypic variation for quantitative traits through additive or epistatic interactions or both. and accounted for 15.6% and 24.2% through additive and epistatic effects for days to 50% flowering. For plant height, , , , and interactions accounted for 17.6%, 17.3%, 10.6%, and 10.7% of the phenotypic variation, respectively. both additively as well as epistatically accounted for 14.6% of the phenotypic variation for plant height. Growth habit and cotyledon color were each controlled by single genes with prostrate growth dominant over erect plant type and red cotyledon dominant over yellow cotyledon.

Understanding genetics of the quantitative traits will help to develop the breeding strategy for selection. The significant correlations among days to 50% flowering, seed diameter, and 100 seed weight and the association between the gene-rich QTL region (QLG ) will help the breeders in selecting plants for early maturity and improved seed quality. The putative QTLs will be useful to locate the genes in the genome that are important for the traits and provide the guidance for marker-assisted selection and cloning of the genes.

Acknowledgments

The authors are indebted to Sheri McGrew, Tony Chen, Jarrod Pfaff, Lorna Burns, and Tracie Anderson for the field and lab support.