Morphological, Molecular, and Biochemical Characterization of a Unique Lentil (Lens culinaris Medik.) Genotype Showing Seed-Coat Color Anomalies Due to Altered Anthocyanin Pathway

This study reports the identification of a unique lentil (Lens culinaris Medik.) genotype L4717-NM, a natural mutant (NM) derived from a variety L4717, producing brown, black, and spotted seed-coat colored seeds in a single plant, generation after generation, in different frequencies. The genetic similarity of L4717 with that of L4717-NM expressing anomalous seed-coat color was established using 54 SSR markers. In addition, various biochemical parameters such as TPC (total phenolic content), TFC (total flavonoid content), DPPH (2,2-diphenyl-1-picrylhydrazyl), FRAP (ferric reducing antioxidant power), H2O2 (peroxide quantification), TCC (total carotenoids content), TAC (total anthocyanin content), and TAA (total ascorbic acid) were also studied in the seeds, sprouts, and seedlings of L4717, brown, black, and spotted seed-coat colored seeds. Stage-specific variations for the key biochemical parameters were recorded, and seedling stage was found the best for many parameters. Moreover, seeds with black seed coat showed better nutraceutical values for most of the studied traits. A highly significant (p ≤ 0.01) and positive correlation was observed between DPPH and TPC, TAA, TFC, etc., whereas, protein content showed a negative correlation with the other studied parameters. The seed coat is maternal tissue and we expect expression of seed-coat color as per the maternal genotype. However, such an anomalous seed-coat expression, which seems to probably be governed by some transposable element in the identified genotype, warrants more detailed studies involving exploitation of the anthocyanin pathway.


Introduction
Lentil is an important cool-season legume crop of India and is cultivated mainly as a rainfed crop [1]. Globally, it is cultivated in an area of 4.8 million ha and yields 5.73 Mt [2]. Canada (2.17 Mt) is the largest producer of lentils and is followed by India (1.23 Mt) and Australia (0.53 Mt). The market acceptability and value of lentils are very much dependent on seed traits such as testa color, testa pattern, and cotyledon color. Thus, to breed varieties having a specific testa color, a detailed understanding of the genetic mechanism controlling these traits is needed [3]. A very early report mentioning the monogenic nature of seed-coat color in lentil was presented by Tschermak Seysenegg [4], wherein he also observed a green seed coat (with purple spots) to be dominant over a light bland seed coat [5]. Erskine and Witcomb [6] identified five major classes of seed background color and three classes of seed-coat pattern in lentils. Monogenic expression of testa color with a dominance of brown color over tan and green was reported by Emami and Sharma [7]. Furthermore, cotyledon color was found to influence the testa color, which may be due to the diffusion of some soluble pigments from cotyledons to the testa. However, black testa color was found to be governed by two independent genes [8]. Similarly, the digenic inheritance of testa color was also reported by Vandenberg and Slinkard [9], while the testa pattern was found to be regulated by multiple alleles. In the case of cotyledon color, red was dominant over green and showed monogenic inheritance. Interestingly, green cotyledons showed the involvement of a recessive color inhibitor [10], and the Dg gene resulted in the formation of a dark-green cotyledon [11].
For the red lentil industry, efficient milling is one of the key economic traits; among various seed traits, seed-coat color is also known to influence milling efficiency [12]. Interestingly the seed-coat color and its composition are known to influence the seed-coat thickness. Genotypes with no tannins in their seeds (homozygous tan) show much lower polyphenol content [13]. Such seed coats are nearly 20% thinner and also nearly 20% lighter in weight than the genotypes having normal seed-coat color [14]. Two independently located genes (Ggc and Tgc) are known to determine the seed-coat color in lentils. Dominant and recessive combinations of these alleles at each locus determine the four basic seed-coat colors, namely, brown (Ggc Tgc), gray (Ggc tgc), tan (ggc Tgc), and green (ggc tgc). Differences in the contents (amount and type) of polyphenolic compounds in the seed coat ultimately determine the color of lentils. Higher contents of flavan-3-ols, proanthocyanidins, and certain flavones result in gray and green seed-coat color. Dark or red-colored genotypes of barley, sorghum, millet, and rye are colored due to the presence of anthocyanin compounds, while the reddish/brown color of red wheat, rice, and maize is due to oxidized proanthocyanidins [15][16][17][18][19]. In purple maize kernel, pigments are mainly concentrated in the external part of the seed (pericarp), while pigments in the aleurone layers of some species, such as blue maize kernel, have a colorless pericarp [20][21][22].
Generally, most of the colored seed-coat genotypes contain proanthocyanidins; however, all may not contain anthocyanins. On the other hand, black rice contains no or very small amounts of proanthocyanidins [18]. This is due to the activity of polyphenol oxidase, as well as the innate genetic differences present in the genotypes expressing anthocyanin reductase and/or leucoanthocyanin reductase. Anthocyanin reductase converts the anthocyanins into flavan-3-ols, which consequently can be polymerized into proanthocyanidins, whereas leucoanthocyanin reductase transforms leucoanthocyanins into flavan-3-ols, which subsequently polymerize into proanthocyanidins [23][24][25]. In both cases, proanthocyanidins after oxidation result in the formation of reddish/brown compounds. In a number of cereals, purple, blue, or black colors are due to the presence of anthocyanins [18,[26][27][28]. A similar process was also recorded in legumes such as broad bean [29] and soybean [30].
In lentils, polyphenol oxidase is known to oxidize the colorless proanthocyanidins into colored compounds, while it also results in structural changes in the seeds [31]. It is the varying contents of various polyphenolic compounds which ultimately affect the dehulling process and, in turn, cause differing levels of loosening of seed coats from the cotyledons during milling. Seed parameter uniformity in lentils is essential for the marketing and preparation of various food products. Thus, seed-coat color variations in lentils have become a major research concern as these are reportedly affected by genetic, pathological, and various environmental factors [32]. Seed-coat background colors in lentils are an important trait for both growers and consumers. Multicolored lentil grains fetch lower prices over single-and uniform-colored lentils. Across the world, the preference varies for various types of lentils, such as uniform white, green, gray, brown, red, or black. Consumers in the United States and Europe prefer light, uniformly colored lentils [33], while Indian consumers prefer lentils with a dark seed coat [34].
Today, consumers are also becoming more health cautious and want nutritionally rich or biofortified lentil varieties. Black seed-coat lentils are a candidate for such consumers if thoroughly investigated for their nutraceutical qualities. Proanthocyanidins are the major group of polyphenols concentrated in the seed coats of lentils. The presence of high phenolic content in the lentil seed coat is desired for both the food and the pharmaceutical industries primarily as a source of natural antioxidants [35]. Therefore, breeders must know the genes or QTLs regulating the seed-coat color formation in lentils. This will help in the development of customized seed-coat colored varieties in lentils [33]. With this background, this study was aimed at characterizing a unique lentil genotype having an anomalous expression of seed-coat color (brown, spotted, and black in a single plant), along with its parent (L4717), at the morphological, molecular, and biochemical levels at different growth stages (seeds, sprouts, and seedlings).

Seed-Coat Color Anomalies
Development of L4717 began in the year 2011-2012 wherein a nursery derived from the cross ILL7617 × 91516 was received from ICARDA, Lebanon for its further evaluation in India. It is interesting to mention that the progenies of this nursery always produced some black seeds (<1%) during seed production since the development of this variety in the year 2011-12 until now. Each year before sowing, we manually removed the black seeds and then performed sowing. However, to our surprise, we always found some black seeds (<1%) in the final seed lot.
In the year 2017-2018, we selected one plant (L4717-NM) which, when threshed, showed a mixed type of seed-coat expression having brown (98), black (29), and spotted seeds (10). Such an anomalous seed-coat color expression was unexpected and motivated a single-plant analysis. The seeds were again sown in the years 2018-2019, 2019-2020, and 2020-2021, and data were recorded for the seed-coat color (Table 1). To avoid any kind of chance cross-pollination, the flower buds were covered with small cotton balls of three plants in each category. Interestingly, brown, black, and spotted seeds produced all three types of progenies having brown, black, and/or spotted seeds. Some of the brown, black, and mixed genotypes produced a few seeds having spotted seed-coat patterns. The detailed plant data are presented in Table 2. The plants of L4717-NM look morphologically very similar to those of L4717. Moreover, the cotyledons are red irrespective of their seed-coat color ( Figure 1). The L4717-NM pods having black seeds possess anthocyanin pigmentation only at the time of seed development (before maturity), after which they become faded as pods fully mature. Figure 1. Phenotypic details of seeds, sprouts, seedlings, cotyledons, and seed coats of the genotypes L4717 and L4717-NM (brown, black, and spotted seed-coat types).  From the 3-year data, it is very clear that the plants producing mixed (brown, black, and spotted) types of seeds formed all three categories of plants (brown, black, and mixed in different frequencies). Moreover, when brown, black, and spotted seed-coat colored seeds were separately sown in different rows, they produced all three types of progenies (black, brown, and mixed). Overall, the number of plants producing only black seed-coated seeds was much lower compared to brown or mixed type. Interestingly, no trend was recorded between the seed-coat color of the parent and the progenies (Figures 2-4). This could be the reason why we are unable to completely purify the L4717 seed lot, as the brown-looking seeds produce black and spotted seed-coat colored seeds in the next generation (although in very low frequency).
To gain more clarity about the anomalous seed-coat color expression, we also analyzed the progenies derived from the brown, black, and spotted seed-coat seeds ( Table 2, Figures 2-4). Interestingly, plants derived from the brown, black, and spotted seeds (2018-2019) produced all three types of progenies (black, brown, and mixed) in 2019-2020. In addition, the ratio of brown, black, and spotted seeds varied in different plant progenies. Again, no specific trend could be established in the seed-coat color of the progenies when compared with that of parental seed-coat color.

Molecular Marker Analysis
To confirm the genetic similarity or dissimilarity of the identified genotype (L4717-NM) with that of the L4717, molecular analysis was performed using SSR primers. A total of 54 SSR primer pairs were used for the amplification of four genotypes (based on seed-coat color expression), which included (1) L4717-and L4717-NM-derived seeds, (2) brown seed-coat seeds, (3) black seed-coat seeds, and (4) spotted seed-coat seeds. As mentioned earlier, the seeds of L4717-NM were obtained from a single plant which was selected in the year 2017-2018. All the amplified primers showed monomorphic expression ( Figure S1) which confirmed the genetic similarity of L4717-NM with that of L4717 even at the DNA level.

Biochemical Studies
Biochemical studies including the estimation of TPC, TFC, DPPH scavenging, FRAP, peroxide quantification, TCC, crude protein, TAC, and TAA were performed for four distinct lentil genotypes (based on the seed-coat color) including L4717 at three stages, namely, seeds, sprouts (5 days), and seedlings or microgreens (11 days). Significant variations were recorded for most of the studied biochemical traits, especially for the growth stages. The TPC was recorded ranging from 1047.64 to 2434.27 (mg GAE/100g DW) with a maximum in the seedlings derived from black seed-coat seeds and a minimum in spotted seed-coat seeds. The range of other biochemical parameters such as TFC (1.91 to 3.28 mg/g DW), FRAP (7.35 to 46.04 µmol Fe 2+ /g DW), TCC (2.55 to 45.35 mg/100 g DW), TAA (27.21 to 44.12 mg/100 g), and DPPH scavenging activity (35.97 to 66.46%) also varied across different growth stages. Peroxide content was recorded to be maximum in brown seed-coat sprouts (2.95 nmol/g DW) and minimum in L4717 seeds (1.64 nmol/g DW). As expected, crude protein was found to be maximum in the seeds (26.43 g/100 g DW), followed by the sprouts, and minimum in the seedlings (9.52 g/100 g DW) of the studied genotypes. The TAC was recorded as maximum in black seed-coat seeds (0.042 mg/g DW as delphinidin equivalent), followed by spotted seed-coat seeds (0.033 mg/g DW), and minimum in L4717 seeds (0.016 mg/g DW). Interestingly, TAC was not recorded either in the sprouts or in the seedlings. Overall, TPC, TFC, DPPH, FRAP, H 2 O 2 , and TAA were found to be maximum in the seedlings and minimum in the seeds of the studied genotypes. TCC was recorded to be maximum in the sprouts and minimum in the lentil seeds. However, protein was recorded to be maximum in the seeds and minimum in the seedlings. The genotype-based details of all parameters are presented in Table 3.

Cluster and PCA Analysis of Functional Features of Lentil Genotypes
Hierarchical cluster analysis and heatmap visualization of physiochemical (DPPH, FRAP, and peroxide), nutritional (protein), and nutraceutical data (TAA, TCC, and TFC) of 12 lentil samples were performed. The results showed that the lentil genotypes could be grouped into two major clusters. Interestingly, the seeds, sprouts, and seedlings were clearly grouped into different clusters or subclusters ( Figure 6). The AoA changed significantly, from seeds to sprouts and then to microgreens; this was seemingly the reason for the clear grouping of the genotypes into very distinct stage-specific clusters. The clustering also showed that, within the studied genotype, there was not much significant difference at any particular growth stage. Furthermore, the functional characteristics such as AoA, protein, and polyphenols of lentil genotypes having different seed-coat colors were used to perform the PCA (Figure 7). The PCA showed that the first two principal components (PCs) explained very high variance (95%), wherein PC1 and PC2 explained 76% and 19% of the total variance, respectively.

Discussion
The improvement of seed morphology associated with economic traits such as seedcoat color in legumes needs an oriented exploitation of existing genetic diversity [36]. The development of a lentil genotype having desirable seed-coat color needs a proper understanding of the genetic basis of that trait, which will help in the acceptance of such varieties by different market segments, as well as in fetching a better price. This study aimed to characterize a lentil genotype having anomalous seed-coat color expression at morphological, molecular, and biochemical levels. It is essential to mention that we observed the presence of black and spotted seeds in the L4717 seed lot since the very beginning when we received the parental nursery of this genotype from ICARDA in the year 2011-2012. Each year, we manually removed the black and spotted seeds from the seed lot before sowing. However, up until 2021, we were unable to purify our L4717 seed lot, with <1% seeds still having either black or spotted seed-coat color.

Morphological Characterization
In general, the seed-coat ground color and pattern are reportedly under the control of dominance and codominance effects, while a black seed coat is independent of any true ground color [9]. In contrast, this study reports the formation of black, brown, and spotted seed-coat colored seeds in a single lentil plant, generation after generation. Interestingly Wilson and Hudson [32] reported the expression of black, mottled, and bicolored seed coat in early generations (F 2 and F 3 ) from a cross derived from black and beige seedcoat colored parents. However, our study reports such seed-coat expression in a fixed genotype generation after generation without any specific pattern. Mottled, bicolor, or spotted patterns were observed on lentil seed backgrounds ranging from black to brown. Furthermore, a range of color patterns were recorded at a single-plant basis at random frequencies, as also observed by Wilson and Hudson [32]. Early generation (F 1 ) seedcoat mottling has also been reported in other pulses such as Phaseolus vulgaris [37] and in Vigna sinensis (L.) Savie × Hassk [38]. Fruwirth [39], while studying the lentil seed-coat color expression for eight generations in a genotype 'Krain' having partly solid brown and partly black spotted seeds, recorded the appearance of both kinds of seeds on the same plant. Interestingly, he also could not succeed in getting the black spots to full heredity, and even the selection of plants with no spotted seeds failed to produce a line having unspotted seeds. A similar observation was also recorded for other crops such as vetch (Vicia sativa), snap beans (Phaseolus vulgaris), fodder pea (Pisutn arvense), and white mustard (Sinapis alba), as well as in another lentil genotype 'Puy' after rounds of selfing and selections [39]. Such an anomalous seed-coat color in the cases of mottled beans and spotted peas was also observed and termed as a pseudo-mutation by Imai [40] (1935). Even for the chlorina and variegated mutants of okra (Abelmoschus esculentus), an unstable gene was reported [41]. Interestingly, in previous studies, researchers did not study the mottling pattern, especially in later generations using seed-coat color-based planting and a plant-by-plant study. Our study signifies the importance that detailed studies are needed for the expression of seed-coat color especially in the later generations (of certain crosses) using plant-by-plant and pod-by-pod approaches in lentils and other pulses. This will certainly help in reconfirming the presence of such a phenomenon in other pulses and will open a new frontier for unfolding the exact reason(s) behind such an anomalous expression of the seed-coat color in a few genotypes.
Punia et al. [42] obtained a natural mutant plant from an Indian variety DPL-62 in the year 2007-2008, which showed variation for earliness and other morphological traits including seed-coat color. They also reported that the progenies of this mutant plant were segregated until the sixth generation of its identification. Due to such an abundance of the variations in the studied traits, these identified genotypes are named hypervariable spontaneously generated plants (HVSGPs) and are considered the source of new alleles. In contrast, in our study, the anomalous expression of seed-coat color is not limited to a single plant and has persisted in the population since we received the genotype as a nursery from ICARDA, Lebanon in the year 2011-2012. In addition, we studied this anomalous seed-coat color expression in the identified genotype (L4717-NM) on a single-plant and podto-pod basis from 2017-2018 until 2020-2021. Furthermore, except for the seed-coat color, L4717-NM is the same as its parental genotype L4717. Thus, a few reports of anomalous expression of several traits have been reported in lentil [42], but such an anomalous seed coat-color expression has never been reported in lentil or in any other crop.

Molecular Characterization
Molecular analysis was performed for L4717 and L4717-NM (brown, black, and spotted seeds) using 54 SSR primers, which showed a monomorphic banding pattern for all SSR primer pairs in the studied genotypes. This again confirmed the genetic similarity of L4717-NM with that of L4717, from which this genotype was identified. In addition, L4717 is a very early maturing variety that matures under Delhi conditions in nearly 100-102 days. Interestingly, the maturity duration and other morphological features of the identified genotype are also the same, which again confirms that L4717-NM is derived from L4717. Otherwise, lentil is a naturally highly self-pollinated crop having cleistogamous flowers [43,44], and, with <0.8% natural cross-pollination [45], the possibility of intragenic recombination is negligible [42]. In soybean, the seed color (black, brown and brown striped) was found to be regulated via w4 and R loci [46,47], and methylation of the R locus was found to affect the transposition and splicing of a large CACTA transposon from an MYB transcription factor regulating anthocyanin synthase genes [48]. We may expect such epigenetic factors regulating seed-coat color expression in the identified lentil genotype, which can be tested in future studies using this genotype.

Biochemical Characterization
Various antioxidant assays including anthocyanin estimation and total crude protein were performed in four lentil samples differing in terms of seed-coat color to identify the significant differences (if any) on the basis of genotype and growth stages. Different AoA-associated mechanisms can be exploited by determining the varied modes of action such as reducing ability (FRAP), free-radical-scavenging ability (DPPH), preventive effect against lipid peroxidation (H 2 O 2 ), and ability to chelate transition metals ions [49]. Normally, flavonoids possess more AoA than phenolics [50] and play a key role in the celldivision-related signal transduction pathways, thereby imparting anticancer properties [51]. Interestingly, a significant stage-wise increase was recorded for TPC (1047.64 to 2434.27 mg GAE/100 g DW) and TFC (01.91 to 3.28 mg/g DW) when studied in the lentil samples (differing in terms of seed-coat color) at different stages (seeds, sprouts, and seedlings). Similarly, in the 6-day old lentil sprouts, nearly similar TPC (19.4 mg/g DW) was reported by Swieca et al. [52]. Furthermore, in the lentil seeds, Fouad et al. [53] and Altuner [54] reported TPC to the tune of 1510.10 and 385.44 mg/100 g DW, respectively. A similar range of TPC was also reported for mustard (1889.76 mg/100 g FW) and radish (211.19 mg/100 g FW) seedlings [55]. Additionally, in lentil microgreens (seedlings), a similar range of TPC (212.60 to 321.35 mg GAE/100 g FW) was reported [56].
H 2 O 2 in leaf tissue is known to respond to various environmental stimuli [60]. Hydrogen peroxide acts as a signaling molecule in the plant system and imparts a second line of defense; an increased level of H 2 O 2 is associated with cellular oxidative damage [61]. Interestingly, the H 2 O 2 was recorded to be the minimum (1.64 nmol/g DW) in the seeds, followed by seedlings, and the maximum (3.04 nmol/g DW) in the sprouts. A similar range was also recorded for microgreens (1.63-5.92 nmol/g FW) by Priti et al. [56]. In seven lentil genotypes, the peroxide activity was reported in the range of 1.22-2.22 µmol/g [62]. The conjugated double bonds in the carotenoid result in AoA via free-radical scavenging [63]. In our study, TCC (2.55 to 45.35 mg/100 g DW) was recorded to be minimum in the seeds, followed by the seedlings, and maximum in the sprouts. Seedlings of the black genotype showed the highest value among the 12 samples studied. A similar range of carotenoids was reported in the seeds of red lentil (5.31-28.1 µg/g DW) and green lentil cultivars (6.20-16.4 µg/g DW) [64]. Additionally, in peas (5.41-28.19 µg/g DW) and chickpea seeds (8.86-30 µg/g DW), nearly similar rage were recorded for TCC [65]. Interestingly, in lentil microgreens, a similar range of TCC was recorded. i.e., 28.37 mg/100 g FW [66] and 14.40-22.22 mg/100 g FW [56]. However, in lentil sprouts, Wojdylo et al. [59] also reported a similar TCC (6.4 mg/100 g FW).
Our study recorded the maximum protein in seeds (26.43 g/100 g DW), followed by sprouts, and the minimum in microgreens (9.52 g/100 DW). Similarly, seed protein content was recorded as 25.63 g/100 g DW [53], while in 4-and 6-day sprouts, the protein level was 22.7 g/100 g DW and 22.1 g/100 g DW, respectively [67]. In microgreens, the protein level was 1.84-2.67 mg/100 g FW [56]. Fouad et al. [53] reported more free amino acids in the sprouts (13.80 mg/g DW) than in the lentil seeds (1.86 mg/g DW). Thus, the reduction in the total protein content from seeds to seedlings is due to the breaking down of proteins into free amino acids as required for the growth and development of lentil seedlings.
As recorded in this study, nearly similar TAA (289.2 µg/g DW) was also reported in 4-day old lentil sprouts [67] and in microgreens (16.15-26.86 mg/100 g FW) of lentils [56]. Since ascorbic acid is present mostly in chloroplasts, the increased value in seedlings is due to the formation of chloroplasts [68,69]. In addition, vitamin C content is usually high in photosynthetic tissues but low in non-photosynthetic tissues. Some orthodox categories of seeds, which reach maturity in a state of strong dehydration, contain little vitamin C [69]. A similar trend of TAA was recorded in this study, wherein the maximum TAA was recorded in seedlings, followed by sprouts and seeds.

Correlation, Cluster and PCA Analysis
The AoA in the studied lentil samples was found to be linearly correlated according to the FRAP and DPPH methods. On a similar note, Priti et al. [56] also reported a positive correlation between DPPH and TFC (p ≤ 0.01) in lentil microgreens and between FRAP and DPPH activity (p ≤ 0.01) in mungbean microgreens. In sweet potatoes, a significant positive correlation between DPPH and FRAP (r = 0.966) was also recorded [70]. This means that the phenolic and flavonoid compounds mainly contribute to the overall AoA in the studied lentil genotypes. Such observations were also reported for Capparis leaves [71,72]. Similarly, FRAP was found to be significantly correlated with TAA (p ≤ 0.05) in lentil microgreens [56] indicating their role in the AoA of lentil seeds, sprouts, and microgreens. Thus, correlation studies have provided a clear understanding of the parameter(s) contributing to the total AoA [73][74][75]. Proteins are the main energy source for lentil seeds, and the negative correlation between total protein content and all other studied parameters indicates that the protein is being used to produce various types of AoA in the sprouts and seedlings.
Cluster analysis distinctly grouped the lentil genotypes according to their growth stages (seeds, sprouts, and seedlings). A similar kind of clustering pattern was observed for the AoA in lentil and mungbean genotypes when studied at the seedling stage [56]. Furthermore, for different caper (leaves) genotypes when collected from different parts of Ladakh, a similar kind of genotypic clustering was recorded for the total AoA [71]. In plants, phenolics are well reported to have a key role in their total bioactive activities [76]; thus, a positive association between phenols and the total AoA is a common trend in many crop species [77].
PCA is useful for understanding the variations at the species level across different attributes in response to diverse factors such as genetic makeup, maturity duration, and growing conditions [78][79][80]. PCA showed that 95% of the variation was explained by the first two PCs in the studied lentil samples. Similarly, for the lentil and mungbean seedlings, when grown under Delhi conditions, the first two PCs explained 43.68% and 52.7% of the total variance, respectively [56]. Moreover, for lettuce seedlings, the first two PCs explained 99.1% of the total variance [81].

Probable Effect on the Anthocyanin Biosynthesis Pathway
Seed-coat pigmentation in pulses happens through the deposition of various flavonoids which are mainly derived from the anthocyanin biosynthesis branch of the phenylpropanoid pathway [82]. Anthocyanins are water-soluble colored pigments having health benefits due to their antioxidative and antimicrobial activities [83]. The presence of a number of mutants affecting the expression of several anthocyanin biosynthetic pathway genes altering seed pigmentation in a number of crops such as maize, petunia, snapdragon, and gerbera has facilitated the cloning of various regulators [84]. Black seed-coat lentils are known to have delphinidin-3-glucoside as the major anthocyanin component [85]. In our study, TAC was (estimated as delphinidin equivalent) recorded in the range of 0.016 (L4717) to 0.042 mg/g DW (black seed-coat seeds). A nearly similar range of anthocyanins was reported in green lentil (0.018 mg/g DW) and red lentil (0.02 mg/g DW) seeds [86]. Interestingly, anthocyanin could not be detected in sprouts or microgreens/seedlings. Similarly, the nutritional profiling of 14 different microgreens species including lentils reported the presence of anthocyanins only in purple radish and red cabbage [66]. As expected, significantly more TAC was recorded in black and spotted seed-coat seeds than in L4717 or brown seed-coat seeds. This means that the anthocyanin pathway is more active in the black and spotted seed-coat seeds compared to brown seed-coat seeds (even in the same plant). It is also very clear that the flavonoid pathway is affected very easily and frequently, even in the same plant of L4717-NM, by certain unknown factors, thereby resulting in the anomalous seed-coat color expression.
In the flavonoid pathway, there are two major groups of genes: early biosynthetic genes (CHS: chalcone synthase, CHI: chalcone isomerase, F3'H: flavonoid 3 -hydroxylase, and F3 5 H: flavonoid 3 ,5 -hydroxylase) and late biosynthetic genes (DFR: dihydroflavanol reductase and AS: anthocyanidin synthase). EBGs are associated with the formation of common precursors (dihydroflavonols) [87]. The corresponding genes are associated with the formation of seed-coat pigment, and mutation in these genes may result in transparent testa as reported in Arabidopsis [88]. Furthermore, in Chinese cabbage, whole-genome sequencing identified TRANSPARENT TESTA GLABRA 1 (TTG1) as the likely candidate gene controlling seed-coat color [89]. Similarly, in soybean, the R and T loci are known to control various seed-coat colors including black (i, R, T), imperfect black (i, R, t), brown (i, r, T), and buff (i, r, t) through the controlled production of different types of anthocyanin and proanthocyanidin pigments. Similar to lentils [85], the main anthocyanin pigments present in black seed-coat soybean is delphinidin-3-monoglucoside [82]. Unfortunately, in lentils, such detailed studies on the pathways/genes governing seed-coat color are lacking. Thus, the identified genotype may prove to be an asset for the identification of the key switches which smoothly regulate the enzymes resulting in such a drastic expression of seed-coat color in the same plant or even in the same branch.

Sowing and Generation-Based Seed-Coat Color Analysis
A natural mutant plant was identified from an Indian lentil variety named L4717 (Pusa Ageti Masur) in the year 2017-2018 which showed three different types of seeds (namely, brown, black, and spotted) in the same plant (Figures 8 and 9). This unique genotype looks like a natural mutant and was, thus, named L4717-NM (natural mutant). The sowing of lentil genotypes was performed at IARI, New Delhi (28.6331 • N and 77.1525 • E) during the last week of November each year using standard practice. The row-to row-distance was maintained as 30 cm, while the plant-to-plant spacing was maintained as 5 cm, and the row length was kept as 2 m. Both L4717 and L4717-NM generally mature in 100-103 days under Delhi conditions. Three types of seeds based on the seed-coat color (derived from a single plant of L4717-NM) were again sown during the next year (2018-2019), and data were recorded for the seed-coat color pattern in the individual plants. A similar type of sowing was repeated in the rabi season of the years 2019-2020 and 2020-2021. Data recording was conducted for individual plants. Morphologically, L4717-NM was most similar to the variety L4717.

SSR Marker Analysis
Seedlings of the studied lentil genotypes were grown in a paper towel and subjected to genomic DNA isolation using the CTAB method [90]. The DNA was quantified using NanoDrop 2000c (Thermo Fisher Scientific, Waltham, MA, USA), while a quality check was performed on 0.8% agarose gel, and samples were diluted to 20 ng/µL and then used for PCR based amplification. A total of 54 EST-SSRs [91] which were synthesized from Eurofins Genomics, India (Table S1) were used for the polymorphism survey [92]. The amplified products were separated by electrophoresis (3 h at 100 V in 1 × TBE buffer) on 3% Metaphor agarose gel (Lonza, Rockland, ME, USA). The gel was stained by ethidium bromide and then photographed using a gel documentation system (Alpha Imager) [93].

Biochemical Studies
For antioxidant activity (AoA) studies, we took four different types of seeds, i.e., L4717 along with the brown, black, and spotted seed-coat seeds produced by the identified genotype L4717-NM. The analysis was performed at three different growth stages (seeds, sprouts, and seedlings or microgreens) (Figure 1). Freshly harvested seeds from the year 2019-2020 were used for this analysis. For sprouting, the seeds were wrapped in germination paper, soaked in water, and kept in the dark for 4 days. However, for seedlings, the seeds were again wrapped in germination and butter paper, soaked in water, and kept for 11 days at the National Phytotron Facility (New Delhi) with 21/18 • C as the day/night temperature and natural day and night cycles. The parameters studied were TPC, TFC, DPPH assay, FRAP assay, peroxide quantification, TCC, total crude protein, TAC, and TAA.

Total Phenolic Content (TPC)
A modified Folin-Ciocâlteu colorimetric method [94] was used to determine the TPC. Briefly, 0.5 mL of extract was mixed with 2.5 mL of distilled water, 0.25 mL of 1 N Folin-Ciocâlteu reagent, and 0.75 mL 20% sodium carbonate (Na 2 CO 3 ), with the final volume made up to 5 mL. This was then vortexed and incubated at room temperature for 2 h (in dark), and absorbance was measured at 750 nm using a UV/Vis spectrophotometer. Then, a calibration curve of gallic acid (conc. from 10 µg/mL to 60 µg/mL) was generated using a regression equation. The equation of the curve was y = 0.0157x − 0.0012, and the linearity coefficient was R 2 = 0.9981. Final values were expressed as mg GAE/100 g DW (dry weight).

Total Flavonoids Content (TFC)
The ethanolic extract was made by crushing 0.05 g of the samples in 80% ethanol (2 mL), and, to this extract (0.5 mL), 2% AlCl 3 (0.5 mL) was added before incubating for 1 h at room temperature. The absorbance was then measured at 420 nm, and TFC was measured using the calibration curve as quercetin equivalent [95]. The equation of the curve was y = 0.0021x + 0.004, and the linearity coefficient was R 2 = 0.9983.

DPPH (1,1-Diphenyl-2-picrylhydrazyl) Scavenging Activity
For the preparation of methanolic extract, 0.05 g of sample was crushed in 2 mL of methanol and centrifuged (12,000 rpm; 10 min). An aliquot (200 µL) was added to 3.8 mL of DPPH and incubated at room temperature for 1 h in dark; absorbance was measured at 517 nm [96]. The DDPH scavenging activity was estimated using the following equation: (DPPH scavenging activity (%) = Abs Control -Abs Sample /(Abs Control ) × 100 , (1) where, Abs Control is the absorbance of DPPH radical + methanol, and Abs Sample is the absorbance of DPPH radical + sample extract.

Peroxide Quantification
The sample (0.025 g) was homogenized in 1% (w/v) trichloroacetic acid (1 mL) and centrifuged (12,000 rpm, 15 min, 4 • C). The supernatant (0.75 mL) was then mixed with 10 mM potassium phosphate buffer (0.75 mL; pH 7) and freshly prepared 1 M potassium iodide (1.5 mL) solution. The peroxide quantification was performed by comparing the absorbance at 390 nm with that of the standard calibration curve (10 to 200 µmol/mL of H 2 O 2 ), and the concentration was expressed as nmol/g [98].

Total Carotenoid Content (TCC)
The sample (0.05 g) was homogenized in 6 mL of ethyl alcohol/BHT (1 mg of BHT/mL of ethanol), and then incubated (85 • C; 6 min) with continuous vortexing. To this, KOH (120 µL) was added, before incubating (5 min; 85 • C) and cooling on ice. Afterward, distilled water (4 mL) and 3 mL of PE:DE (2:1, v/v) were added and centrifuged (10 min; 12,000 rpm; room temperature). The upper phase was collected, and distilled water (4 mL) and PE:DE solution (3 mL) were added, before centrifuging. Then, the process was repeated twice, and samples were diluted to 10 mL (PE:DE solution). The increase in absorbance was measured at 470 nm [99]. The TCC was calculated using the following formula: where A total = absorbance at 470 nm, Vol (mL) = total volume of extract, and A 1% = absorbance coefficient for carotenoid by column mixture.

Total Ascorbic Acid (TAA)
The titration method was used for the ascorbic acid estimation using a 2,6-dichloroindo phenol (DCIP) dye solution and 4% oxalic acid as a stabilizing medium [100]. Preparation of the dye solution was performed using NaHCO 3 hot solution, and dye standardization was performed by titrating the standard ascorbic acid (1 mg/mL) until it showed a pale pink color. This color persisted for 15 s, and final estimation was performed using the following formula: where X mg = mg of standard ascorbic acid, V 1 = titer value of standard ascorbic acid against dye, V 2 = titer value of sample against dye, Y Ml = amount of aliquot taken (mL) for estimation, and Z mL = total amount (mL) of extracted sample.

Total (Crude) Soluble Protein
Total soluble protein was estimated using the Bradford assay [101]. Briefly, 0.05 g of sample was homogenized in 2 mL of phosphate buffer (pH = 7) and centrifuged (13,000 rpm for 20 min); then, to 20 µL of supernatant, 980 µL of distilled water and 2 mL of Bradford reagent were added, before incubating in the dark (10 min) at room temperature. The change in color of Coomassie dye occurs under acidic conditions when protein molecules bind to the dye and the color changes from brown to blue. This color change was then measured at 595 nm by generating a calibration graph using 25, 50, 75, 100, 125, 150, and 200 µg/mL BSA. The equation of the curve was y = 0.0029x + 0.1418, and the linearity coefficient was R 2 = 0.9895.

Statistical Analysis
The experiments were conducted in four replications, and the results were presented as the mean ± SD. SPSS11.5 was used to perform the one-way ANOVA (analysis of variance) to compare the groups, and Pearson's correlation test was used to assess the correlation between means, while mean comparisons were performed using Tukey's test. The correlation matrix of metabolites, hierarchical cluster analysis, and heatmap visualization were performed using R software (ver. 4.1.3). PCA (principal component analysis) was performed using the details of various biochemical parameters (TPC, TFC, DPPH, FRAP, H 2 O 2 , TCC, and TAA) in the lentil genotypes differing in terms of seed-coat color. The PCA plot was generated on the basis of the first and second principal components (PC1 and PC2) using STAR ver. 2.0.1.

Conclusions
The seed coat in the legume crop is derived from the maternal tissue, which is ultimately determined by the genotype of the plant producing the seeds. Therefore, the seed coat in any lentil plant should display only one seed-coat color. On the basis of the results, we strongly presume the possible presence of some hyperactive transposable element regulating seed-coat color in the studied genotype. In addition, the presence of some epigenetic factors regulating the expression of certain key enzymes of the anthocyanin pathway could also be a possibility. Furthermore, there could be some unknown factors or some hypervariable regions which seemingly regulate the anthocyanin pathway in the identified genotype. The presence of environmental factors such as temperature or photoperiod regulating the seed-coat color can be ruled out as even two adjacent pods showed both brown and black seed-coat colored seeds ( Figure 10). Interestingly, the seed-coat color and the cotyledon color are unrelated, and the embryo and cotyledon genome may interact and affect the surrounding tissue during morphogenesis in some yet unknown way. Thus, in the future, this genotype may serve as a focal point to understand this unique expression of seed-coat color in lentils. Thus, there is a need to identify a stable black seed-coat genotype which can then be crossed with the fixed brown seed-coat genotype for the formation of a mapping population. Furthermore, a transcriptomic approach can be exploited using black and brown seed-coat genotypes (acting as an isogenic line) at the time of seed formation for the identification of DEGs impacting the trait. Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/plants11141815/s1: Figure S1. A representative picture showing the amplification details in four lentil samples (differing in terms of seed-coat color) using SSR markers. PLC2, 12, 10, and 11 are the primers; 1: L4717, 2: brown seed coat, 3: black seed coat, 4: spotted seed coat (2-4 were obtained from the L4717-NM expressing a mixed phenotype); Table S1: List of primers used in the study (Jain et al. 2013).