Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

6.1 Introduction

The global economic position of lentil (Lens culinaris Medik.) among grain legumes has increased in international trade and currently ranks sixth in terms of production after dry bean, pea, chickpea, faba bean, and cowpea. During the period 2003–2006, lentil constituted 6 % of the total world dry pulse production. Lentil production increased more than fourfold (413 %) in a period of 40 years, from 917,000 t in 1961–1963, with an average yield of 560 kg/ha, to a world harvest in 2004–2006 of 3,787,000 t with a mean yield of 950 kg/ha (Erskine 2009). According to FAOSTAT (http://faostat.fao.org), in (2007), lentils were being produced in 52 countries with approximately 3.8 million hectares under cultivation. The spread of lentil from its center of origin in the “Fertile Crescent” after its early domestication about 10,000 years ago (Cubero et al. 2009) has been accompanied by selection for traits that enhance adaptation to a wide range of agroecological environments. The crop is now being cultivated from cold-winter continental climates to subtropical areas and high altitudes in tropical regions. With an initial selection against pod dehiscence, the main objective of human selection has remained seed size together with flowering response to photoperiod and temperature, as well as resistance or tolerance to abiotic and biotic stresses. If lentils have been maintained by farmers through ages, it is most likely because they grow in poor soils, harsh climates, and conditions unfavorable for humans, animals, and crops (Cubero et al. 2009). In many cases, lentils may have been the only source of protein available to farmers. Lentil cultivation and adaptation to different environments and human preferences have resulted in a wealth of phenotypic seed variation together with the development of different culinary customs. Lentils are used worldwide as a starter, as a main dish, as a side dish, or in salads. For instance, according to Sarker and Kumar (2011), in West Asia and North Africa, Mujaddarah is a popular dish made out of whole lentil and immature wheat seed. Koshary is a commonly served dish in Egypt, made out of a mixture of rice and red lentil. In North Africa, lentil is prepared with vegetables and the recipe is known as Lentil Tagine. Of course, red lentil soup is popular all over West Asia, but most particularly in Turkey, Lebanon, Jordan, Palestine, and Syria. Wot is a traditional dish of Ethiopia. In South Asia, as indicated by Saha and Muehlbauer (2011), lentil is an important source of protein and is consumed almost daily as Dal, a typical stew seasoned with turmeric, ginger, onion, and other spices, by all strata of people irrespective of social and economic status. While in Western Europe, particularly in Spain, the brown lentil typified by variety pardina (greenish-brown seed coat with some dark speckling and mottling with yellow cotyledons) is preferred to be cooked as lentil stew. In addition, lentil may be deep fried and eaten as a snack or combined with cereal flour in the preparation of bread and cake. Large-seeded green lentils are commonly used in salads. The newly developed product Genki Energy Bar is an example of a potential wider use of the wholesome lentil (Vandenberg 2011). On a global scale, lentil consumption has augmented more than twice the rate of human population growth, with lentil consumption over the past 40 years having increased more than any other food crop. Among the cool season pulses, lentil is by far the fastest-growing crop, while many of the other grain legumes are actually in decline. This is most likely due to the convenient fast cooking coupled with a saving of fuel and time; dehulled lentils cook even faster than milled rice. Moreover, lentil requires less processing when compared to soya beans and cereals. As part of a diet, lentils are rich in protein (20–30 %), complex carbohydrates, and dietary fiber and are an excellent source of a large range of micronutrients (Thavarajah and Thavarajah 2011).

In order to introduce genetic variation, there remains no doubt as to the need to undertake manual crosses in a highly autogamous species such as lentil, allowing for segregation and selection, while introgression and genetic-based broadening go hand in hand. Germplasm evaluation together with a successful and targeted hybridization allows for an efficient breeding and selection of varieties adapted to specific environments. Therefore, knowledge on germplasm and specific practical skills regarding ways to optimize hybridization constitute vital factors to advance breeding programs. Our aim here is to provide insight into these issues; simultaneously we describe the evolutionary aspects of lentil and assess the gene pool and flow for crop improvement; consider levels of diversity and agronomic traits of importance, with emphasis placed on wild species and interspecific Lens hybridization; and finally conclude with molecular aspects and future prospects of lentil breeding.

6.2 Crop Gene Pool, Evolutionary Relationships, and Systematics

6.2.1 The Genus Lens Miller

The genus Lens Miller is taxonomically classified within the tribe Vicieae DC. (syn. Fabeae Rchb.) (syn. Fabaceae, Papilionoideae) together with three other genera: Lathyrus L., Pisum L., and Vicia L. The tribe Vicieae is included in the inverted repeat-lacking clade (IRLC) (Wojciechowski et al. 2000), accordingly denominated because the chloroplast genome is uniquely marked by the loss of one copy of the large inverted repeat (approx. 25 kb). The clade includes all the members of the tribes Cicereae, Hedysareae, Trifolieae, and Vicieae, as well as at least three other genera (Wojciechowski 2006; Jansen et al. 2008). Thus, the interesting crop legume genera such as Trifolium, Medicago, and Cicer all hold a close phylogenetic relation to lentil and other Vicieae species. Although all botanical texts since the sixteenth century have used the name Lens for the species to which lentil belongs, the first botanist to truly assign the status of the genus was Tournefort in 1700. Miller in 1740 verified the designation and became the authority on the genus, while he is also responsible for producing the oldest available botanical description. In 1787, the German botanist and physician Medikus assigned lentil the scientific name Lens culinaris and the nomenclature is currently accepted and used. Numerous taxonomic treatments during the nineteenth century were published based on similarities of Lens with other taxa such as Ervum, Ervilia, Vicia, Lathyrus, Orobus, and even Cicer (all of which belong to a relatively young group of plants that are still in active evolution). Nonetheless, by the end of the nineteenth century, the genus Lens was finally well established (for historical references and synonyms, see Cubero et al. 2009).

The taxonomy of the genus Lens is far from easy given the close relationships among species, and the phylogenetic relationships between Lens species continue to be a subject of scientific discussion (Cubero et al. 2009). The classical structure of the genus based on morphological analyses and/or biochemical markers included four species: Lens culinaris, L. orientalis, L. nigricans, and L. ervoides (Ladizinsky 1979). In 1984, accessions of nigricans were reclassified as odemensis, and a new taxonomy of the genus was proposed, Lens culinaris with cultigen subspecies culinaris and wild ssp. orientalis and odemensis, and Lens nigricans with two subspecies, nigricans and ervoides (Ladizinsky et al. 1984). However, Ladizinsky later recognized the following taxa: L. culinaris, with subspecies culinaris and orientalis, Lens odemensis, L. ervoides, and L. nigricans (Ladizinsky 1993). In 1997, two new species were added to the genus, Lens tomentosus (ex L. orientalis; Ladizinsky 1997) and Lens lamottei (ex L. nigricans; Van Oss et al. 1997). Ferguson et al. (2000), using morphological as well as molecular markers, considered both odemensis and tomentosus as subspecies of L. culinaris. The following classification of the genus Lens was proposed by these authors: L. culinaris, with four subspecies, namely, culinaris, orientalis, tomentosus, and odemensis; L. ervoides; L. nigricans; and L. lamottei. This nomenclature is currently being used by the National Center for Biotechnology Information (NCBI). A phylogenetic nomenclature based on SDS-PAGE analysis of seed proteins undertaken by Zimniak-Przybylska et al. (2001) clustered culinaris, orientalis, tomentosus, and odemensis in one group; L. ervoides and L. lamottei in a second group; and L. nigricans in a third group; this result conforms in grosso modo with the previous study of Ferguson et al. (2000).

To summarize, taxonomic analyses based on morphological and/or biochemical markers ranged from four species in 1979, namely, L. culinaris, L. orientalis, L. nigricans, and L. ervoides, to two species in 1984: L. culinaris (with subspecies culinaris (the cultigen), orientalis, and odemensis) and L. nigricans (with ssp. nigricans and ervoides); again to four species in 1993, i.e., L. culinaris (ssp. culinaris and orientalis), L. odemensis, L. nigricans, and L. ervoides; and to the previous four species plus the addition of L. tomentosus and L. lamottei to comprise a total of six species in 1997. The number of six species is now widely accepted: L. culinaris [ssp. culinaris, the cultigen and ssp. orientalis (Boiss.) Ponert, the wild ancestor of the cultivated forms], L. odemensis (Godr.) Ladiz., L. tomentosus Ladiz., L. nigricans (Bieb.) Godr., L. ervoides (Bring.) Grande., and L. lamottei Czfr. The geographical distribution of the Lens species which are all self-pollinated is described by Cubero et al. (2009).

6.2.2 Genome Size, Ploidy, Chromosomes, and Karyotypes

The C value (unreplicated haploid) of the lentil genome was determined by flow cytometry to amount to 4.41 pg, which is equivalent to 4,063 Mbp (Arumuganathan and Earle 1991). This result agrees with a previous evaluation undertaken by Bennet and Smith (1976) using Feulgen microdensitometry of 4.6 pg, comparable also to the genome size of closely related Vicieae species such as pea (3,947–4,397 Mbp). Neither the mitochondrial nor the chloroplast genomes of lentil have been sequenced to date according to the data available in the NCBI Eukaryote Organelles and the Organelle Genome Megasequencing Program websites. However, Van Oss et al. (1997) carried out a comparative restriction fragment length polymorphism (RFLP) analysis of the Lens species chloroplasts.

All Lens species possess the same chromosome number (2n = 14) and share similar karyotypes, which consists of three pairs of metacentric and submetacentric chromosomes, three pairs of acrocentric chromosomes, and a pair of metacentric chromosomes with the nucleolar organizing region (NOR) secondary constriction site proximal to the centromere location (Ladizinsky 1993), although some karyotype variants have been described (Balyan et al. 2002). In addition to the standard ssp. culinaris karyogram, Ladizinsky (1993) also described other karyotypes in ssp. orientalis: a single accession (No 113) showed a karyotype in which about three-quarters of the satellite had been translocated to another chromosome and the metacentric-satellited chromosome had become acrocentric. Accessions from Iran differed by a chromosomal rearrangement involving two metacentric chromosomes, and accessions from Turkey differed by a paracentric inversion and inclusive of two or even three chromosome rearrangements. Accordingly, ssp. orientalis appeared to be the most chromosomally variable Lens taxon (Ladizinsky 1993). Balyan et al. (2002) using fluorescence in situ hybridization (FISH) identified the NOR region to be located in chromosome 3; chromosome 2 consisted of a submetacentric chromosome with a 5S locus distally located on the long arm (the major site of hybridization), whereas chromosome 6 comprised an acrocentric chromosome with the other 5S locus located proximal on the short arm (minor site). Further data based on Lens homologous probes, which differentially hybridized to the 5S non-transcribed spacer (NTS), allowed a better differentiation between the two 5S loci (Fernández et al. 2005). Locus 5S on chromosome 2 corresponded to the longer NTS family sequence, while the second 5S locus located on chromosome 6 corresponded to the shorter NTS family. Furthermore, FISH data indicated that in some wild orientalis accessions (e.g., ILWL-7), three-quarters of the satellite had been transferred to another chromosome, the metacentric-satellited chromosome had become acrocentric, and one of the submetacentric chromosomes had been lengthened. In this orientalis accession, only a single 5S locus was detected using the longer NTS probe on chromosome 6; thus, it is possible that some ssp. orientalis karyotypes lack one of the 5S loci (Fernández et al. 2005).

6.2.3 Gene Pools

The previous discussion regarding the genus Lens and their different karyotypes can be confusing; in addition, as will be described further on, no clear hybridization barriers have been defined between the accepted species. The levels of fertility and sterility so far found have depended on the taxa involved in a specific cross but also, to a greater or lesser extent, on the particular populations used within these taxa. This phenomenon is common as it happens in Lens, since the taxa are close relatives. Lens is a genus that belongs to a very active group from an evolutionary point of view. Even with the exceptions described above, of which more details are provided in the next sections, hybridization experiments have supported the idea of six differentiated species. In line with the above and particularly with hybridization barriers in mind, support for the idea of six differentiated Lens species has gained. Besides of the forms of the cultigen (L. culinaris ssp. culinaris), L. culinaris ssp. orientalis obviously belongs to the primary gene pool and L. odemensis to the second, although success of L. odemensis crosses with the cultigen may or may not require embryo rescue depending on the specific accessions used. The tertiary gene pool is composed of L. tomentosus, L. lamottei, L. nigricans, and L. ervoides, but these can become part of the secondary gene pool by means of embryo rescue. In any case, further hybridization studies are needed to truly establish whether L. tomentosus and L. lamottei belong definitely to the secondary or tertiary gene pool (Cubero et al. 2009).

6.2.4 Origin and Evolution

Seed size seems to be the only character which has allowed domestication events to be traced in archaeological remains. The oldest remains of wild lentils have been found in Mureybit (Syria) dated around 10,000 BP, while those of the cultigens have been dated around 9,000 BP and were discovered in aceramic Neolithic layers located in the Near East (see Cubero et al. 2009 for archaeological data). Given the coexistence of the wild and domesticated forms which is not found elsewhere, coupled with archaeological evidence, the Fertile Crescent is the most likely candidate to be the center of origin of the cultivated lentil. In addition, lentil diversity in the center of origin is still very high for both the cultivated primitive forms and the wild relatives. Using molecular markers, Ferguson et al. (1998a, b) located areas of high diversity for ssp. orientalis alongside the border between Turkey and Syria as well as in areas of southern Syria and Jordan; this data supports the archaeological evidence and indicates the Near East region as the most likely center of lentil domestication. Ladizinsky (1999) also suggested the Near East as the center of origin based on the polymorphism found in wild accessions of ssp. orientalis and the monomorphism of culinaris. Indicators of human activity suggest that some populations of orientalis were unconsciously subjected to selection (Harlan 1992) which resulted in the crop that we know today as lentil. According to Zohary (1999) and based on chromosome and DNA polymorphisms, the domestication event of lentil happened once or only a few times. Toklu et al. (2009) used ISSRs and AFLPs to undertake a variability analysis of lentil landraces from Southeast Turkey concluding that they exhibited a high genetic variation index (He = 0.180); nonetheless one particular landrace from the Karacadag/Diyarbakir region was found to be significantly different to the rest of the germplasm evaluated. The Karacadag mountain range located in Southeast Turkey is placed within the core area of plant domestication of the Fertile Crescent. Lev-Yadun et al. (2000) suggested a lentil domestication site close to or overlapping the area where einkorn and emmer wheats had been domesticated in the Fertile Crescent, concretely, the Karacadag/Diyarbakir region as the domestication center of lentil. Furthermore, two centers of diversity for L. culinaris ssp. orientalis have been described: (1) Southeast Turkey and Northwest Syria and (2) West and North Jordan and southern Syria (Ferguson et al. 1998a).

Compared to L. orientalis, cultivated lentils have a greater stem and rachis length, more leaflets per leaf, a greater leaf area, an increased number of flowers per peduncle, as well as an increased number of pods and seeds. In addition, peduncles of the cultivated forms are generally shorter or equal in length to the rachis when compared to the wild forms. Aside from all of the above, pod indehiscence of the cultivated forms in contrast to the dehiscent mode of the wild species, together with the erect growth habit of the cultivated lentil compared to the wild procumbent and prostrate plant structure, typifies the most noticeable differences between the domesticated and wild relatives, being traits of great economic value and with many implications on yield and ease of harvesting. All of the abovementioned characters are strongly associated with increased yield levels, in a similar way as observed for the other domesticated food legumes. Lentil cultigens show a higher frequency of white flowers compared to the dominant purple-colored inflorescences of the wild relatives, most probably a character associated to a better culinary quality and fixed by indirect selection of lighter-colored seed coats (Fratini et al. 2011). As is also the case of chickpea and faba bean, there exists a clear regional distribution pattern of the cultivated lentil forms (displayed in Cubero et al. 2009). The trend from eastern to western lentils comprises an increase in seed size, a higher number and size of leaflets in addition to the length of the calyx teeth relative to the corolla length. To explain this cline, it has been postulated that the introgression into western forms of lentil came from odemensis (more likely than from nigricans as odemensis was given its specific status because of its crossability with culinaris), while an introgression from orientalis played the leading role with regard to the eastern forms. For the short calyx of the prevalent lentil forms found in East Africa and mainly Ethiopia, the genetic influence of ervoides has also been postulated. A comparison between the geographical patterns of the wild species and the cultivated forms (see Cubero et al. 2009) seems to verify the introgression hypothesis as orientalis is the only wild form spreading eastward, ervoides to the Ethiopian region, and both nigricans and ervoides to the west. Nonetheless, the latter two species do not readily cross with culinaris as hybrids result in embryo abortion (Ladizinsky et al. 1984; Fratini and Ruiz 2006), but sporadic crosses through a long period of time cannot be readily dismissed. Besides, in the same way as odemensis was separated from nigricans because of the differential fertility level with cultigens, other strains of nigricans and ervoides could have behaved in a similar way.

Thus, although crosses between culinaris and odemensis are feasible and produced longer (sometimes branched) tendrils than those observed for culinaris (Ladizinsky 1979; Ladizinsky et al. 1984), more experimental work, including molecular biology analysis, is necessary to prove the historical introgression of traits derived from odemensis into the cultigens. The geographical pattern could simply be an indirect (correlated) response to different human selection approaches in different parts of the world accompanied with the usual sources of genetic variation (mutation, migration, and genetic drift) and crosses with companion weeds. In fact, molecular marker analyses have indicated that the genetic variability within cultivated lentil forms is relatively low (Durán and Pérez de la Vega 2004; Sonnante et al. 2003), suggesting that the two great groups of cultivated lentils, microsperma and macrosperma, could only be variants for quantitative traits resulting from disruptive selection.

Summing up the available information on crop evolution, lentils were domesticated in the foothills of the mountains of southern Turkey and northern Syria (Ferguson et al. 1998a, b, c; Ladizinsky 1999; Lev-Yadun et al. 2000) most likely by unconscious selection (Harlan 1992) within populations of ssp. orientalis. Nonetheless, the influence of other wild relatives cannot be excluded as has been shown to have occurred in the origin of most cultivated crops. In the concrete case of lentil, the similarity among wild species could have been a factor in favor of producing companion weeds and maintaining them in cultivated stocks. The ssp. orientalis and L. odemensis are the most likely candidates to have constituted the main origin of extra-specific variability for the cultigens; however, more experimental proof is further needed. The genetic variability found with molecular markers seems to be low (Durán and Pérez de la Vega 2004; Sonnante et al. 2003), suggesting a common origin for all the cultivated forms of lentil and a narrow range of achieved artificial selection. The observed differences detected among different geographical groups could be the plain result of a correlated response exerted by selection for higher yield, compared to a direct consequence of a more basic genetic difference. All in all, the role of introgression from wild forms, however, requires further study.

6.3 Assessment of Gene Flow for Crop Improvement

Ladizinsky (1979) made crosses between culinaris (three accessions representing extreme values for seed characteristics and other traits), orientalis (four), and nigricans (one). The F1 generation of the culinaris × orientalis crosses grew normally, and the F2 generation segregated for growth habit, flower color, pod dehiscence, and seed coat color. Meiosis was almost normal with the presence of a quadrivalent in most of the cells examined and, in very few cases, a trivalent and one univalent. This result supported previous proposals considering orientalis as a subspecies of culinaris. Other hybridization experiments have indicated that culinaris × orientalis crossing successes are equal or even higher than cross combinations within L. c. culinaris (Fratini et al. 2004), so far not having observed embryo abortion using different orientalis accessions, although clear signs of varying degrees of plant deterioration, such as chlorosis and other weakness signs, are clearly detectable in recombinant inbred populations (RILs) most likely due to the fixing of unfavorable allele associations (Fratini, unpublished). Altogether, the orientalis taxon remains the best source among the wild species for the introgression of traits to the cultigen. According to Ladizinsky (1979), culinaris and orientalis share the same karyotype, although karyotype variants in orientalis have been described (Ladizinsky 1993; Balyan et al. 2002; Fernández et al. 2005). Likewise, some extent of intraspecific karyotype variations in several Lens species has also been described (Galasso 2003; Fernández et al. 2005). Since fertility of the inter-subspecific hybrids between culinaris and orientalis depends on the chromosome arrangement of parents (Ladizinsky 1979; Ladizinsky et al. 1984), three crossability groups have been identified in the wild ssp. orientalis: a common, a unique, and an intermediate. Crosses between members of the common and unique groups yield aborted seeds which can be rescued by embryo culture, while members of the intermediate group are cross compatible with the other two groups (Ladizinsky 1993; van Oss et al. 1997). Subspecies orientalis is readily crossed with L. odemensis, the hybrids are vegetatively normal but partially sterile due to meiotic irregularities resulting from three chromosome rearrangements between the parental strains (Ladizinsky 1993). The first hybrids between culinaris and nigricans developed normally (Ladizinsky 1979). However, later on crosses with other accessions of nigricans were observed to be unviable (Ladizinsky et al. 1984). Therefore, in 1984, the former nigricans accession was reclassified as L. odemensis. Lens tomentosus is morphologically closer to L. c. ssp. orientalis than to any other Lens taxon; nevertheless, they are isolated by hybrid embryo breakdown, complete sterility, and five chromosomal rearrangements (van Oss et al. 1997), which supports the idea of a specific status for tomentosus. Likewise, L. tomentosus is reproductively isolated from L. lamottei and L. odemensis by hybrid embryo abortion (van Oss et al. 1997). Linkage studies have revealed chromosomal rearrangements between L. culinaris and L. odemensis (Ladizinsky 1993), which could possibly explain why certain accessions of odemensis are freely crossed with culinaris (Ladizinsky 1979) while others need embryo rescue (Fratini and Ruiz 2006). Lens odemensis is cross incompatible with nigricans and ervoides due to hybrid embryo abortion (Ladizinsky et al. 1984; Ladizinsky 1993).

The morphological differences between L. nigricans and L. lamottei are limited to stipule shape; however, the two species differ by four reciprocal translocations and one paracentric inversion, resulting in the complete sterility of their hybrids (Ladizinsky et al. 1984). Lens nigricans × L. ervoides interspecific hybrids are vegetatively normal but completely sterile (Ladizinsky 1993). Fratini and Ruiz (2006) carried out extensive crosses between L. c. ssp. culinaris and L. nigricans and L. ervoides and L. odemensis. Hybrids between the cultigens and the other species were viable only through embryo rescue; the rates of adult plants obtained were low: 9 % with odemensis and 3 % with nigricans and ervoides. Previously, it had been already shown that crosses between culinaris and nigricans or ervoides needed embryo rescue to recover interspecific hybrids (Ladizinsky et al. 1984; Ladizinsky 1993). In summary, hybridization barriers support the idea of six differentiated species. Together with Lens c. culinaris, the cultivated form, Lens c. orientalis obviously belongs to the primary gene pool, whereas L. odemensis to the second. Although the success in such crosses (needing embryo rescue or not) depends on the particular odemensis accessions used in the study, L. tomentosus, L. nigricans, L. ervoides, and L. lamottei belong to the tertiary gene pool, but can become part of the secondary gene pool by means of embryo rescue; however, further hybridization studies are needed to establish whether L. tomentosus and L. lamottei truly belong to the secondary or tertiary gene pool (Cubero et al. 2009).

6.3.1 Gene Flow Constraints

Mating system is a major determinant of the genetic structure of plant populations and of the breeding methods in crops. Lentil is a highly self-pollinating species in which outcrossing rates are usually low. Wilson and Law (1972) reported less than 0.8 % of natural cross-pollination as discerned by morphological markers. More recently, the outcrossing rate of lentils was established to range from 0.06 to 5.12 % among three varieties as measured in Central Europe using a dominant orange cotyledon allele as a genetic marker (Horneburg 2006). The outcrossing rate of individual plants ranged from 0 to 22.2 %. Results were strongly influenced by cultivar, year, and location; thus, for instance, the outcrossing rates of the cultivar Pisarecka Perla in the same locality were 0.89, 5.12, and 3.01 % in the years 1999, 2000, and 2001, respectively (Horneburg 2006).

The large variation detected in natural outcrossing rates of lentil, which for a given cultivar depend upon year and location (Horneburg 2006), indicates that in addition to the genotype, the environmental conditions can play a determinant role with regard to hybridization success. The effect of environmental conditions on intra- and inter-subspecific hybridization success was analyzed by Fratini et al. (2004). Fall crossing success under controlled greenhouse conditions among lentil cultivars or with subspecies orientalis was substantially higher with regard to the same field crosses carried out in the spring. In the first published literature on intraspecific lentil hybridization (Wilson 1972), it was observed that lentil crosses in the greenhouse were most successful during the winter months, when the temperatures were maintained in a range of 18–24 °C, and the relative humidity was above 50 %, while manual crossing in the field was unsuccessful. The study of Fratini et al. (2004) established that the fall crossing success in the greenhouse was favored by temperatures of 20–25 °C, high humidity, and sunny mornings; crosses performed in the morning and early afternoon (10–14 h) usually gave better results than crosses late in the afternoon (16–17 h); light conditions also influenced crossing success as crosses carried out on cloudy days (with lower temperatures around 15 °C) yielded a significantly lower pod and seed set; also a temperature range from 20 to 25 °C under high humidity conditions with an optimum temperature at 24 °C was found to be the most suitable for crossing lentil. Previously, Malhorta et al. (1978) had also determined that morning pollinations (9–11 a.m.) in India produced the highest percentage of pod set in intraspecific crosses. The field results of the spring crosses undertaken by Fratini et al. (2004) agreed with previous field data obtained for such crosses (Wilson 1972; Malhorta et al. 1978; Solh et al. 1980; Mera and Erskine 1982; Malaviya and Shukla 1990; Kumar and Singh 1998); all studies pointing out that conditions of high humidity and mild temperatures favored crossing success. In addition, for spring crosses in the field, it has been determined that the bagging of pollinated flowers or plants, together with a lower illumination, such as applying artificial shading after pollination, increased the intraspecific field hybridization success (Solh et al. 1980; Mera and Erskine 1982). For instance, Mera and Erskine (1982) obtained an average percentage of pod set of 27.1 %, ranging from 60.9 % with plastic bags left on shaded plants until harvest to 7.7 % for the unshaded control. During winter in India, Kumar and Singh (1998) found an overall pod set of 23 %, observing that evening crosses, with temperatures of 23–26 °C and a relative humidity of 85 %, were the most successful.

Besides environmental factors determining hybridization success, genotype differences related to crossing efficiency have also been detected. For instance, Fratini et al. (2004) found that the cultivar Lupa was on average the worst male parent (5.9 % of pod and 7.1 % of seed set per pollinated flower); however, this cultivar was on average to be the best female parent (29.1 and 39.4 %, respectively). Under controlled greenhouse conditions, it was concluded that the higher significance of the variable Male relative to the Female pointed out the importance of pollen quality (Fratini et al. 2004). As to specific genotype influences, it has been noted that certain intraspecific cross combinations are easier to obtain than others, inclusive in the case of reciprocal crosses, especially with regard to crosses between microsperma and macrosperma varieties of lentil (Fratini, unpublished). Previously three crossability groups had been identified in the wild ssp. orientalis: a common, a unique, and an intermediate; only members of the intermediate group were cross compatible with the other two groups (Ladizinsky 1993; van Oss et al. 1997). Therefore, it seems that lentil genotypes influence crossing success, although further studies are needed to clarify the relative influence of the two main factors, genotype and environment. Besides from all of the above considerations regarding the mating system, environmental conditions, and parental genotype cross combinations, all of which constrain gene flow in lentil breeding; the operator efficiency undertaking manual hybridizations is also a determinant factor. A skilled operator needs to possess the experience to discern highly fertile female lentil plants and appropriate male pollen donors. Identifying flowers at the correct physiological stage of ¾ sepals to petals to emasculate and pollinate or to gather pollen is of vital importance. For crosses undertaken in the greenhouse under controlled conditions, out of 247 emasculations and intra- and inter-subspecific pollinations, a total of 60 hybrids were obtained as discerned by morphological and molecular markers, equal to an overall hybridization efficiency of 24 % (Fratini et al. 2004).

As a summary, gene flow in lentil breeding is constrained by (1) the mating system with the inherent complexity to undertake crosses by hand emasculation and pollination (which means, for instance, that gene introgression using the backcross method of breeding is difficult and tedious); (2) the need to dispose of controlled environmental greenhouse conditions to assure hybridization success; (3) certain parental cross combinations result in higher crossing efficiencies than other genotype combinations; (4) need of skilled manual hybridization operators; and (5) in the case of certain inter-subspecific crosses and in most interspecific crosses, embryo abortion must be overcome by in vitro embryo rescue, although this may not preclude a future hybrid breakdown during the juvenile stage or sterility during the adult phase. More data on this aspect will be provided in the following sections. On the other hand, traditional lentil breeding is still limited by: (1) the apparent low genetic variability and the relatively low number of described Mendelian genes; (2) the lack of information on genetic, physiological, and other aspects of the crop; and (3) the lack of screening methods under controlled conditions. Traditional breeding has often been directed to address major production constraints, which in turn has limited to take into consideration other possible breeding goals.

6.4 Level of Diversity in Crop Germplasm

6.4.1 Morphological Diversity

The first detailed and complete study of the cultivated lentil was made by Barulina (1930), a disciple and subsequently wife of N. I. Vavilov. She considered two subspecies, microsperma and macrosperma, according to seed size, the main objective of human selection. Barulina also considered the geographical distribution of a cluster of characters, defining regional groups or greges. The main characters used to define the subspecies consisted of pod and seed traits together with distinct differences in flower length. Important characters used to define the groups within the subspecies included dehiscence, length of calyx teeth, and number of flowers per peduncle (Cubero et al. 2009). More than half a century later, a study of the quantitative agro-morphological variation found in the germplasm of 13 major lentil-producing countries identified three major regional groups (Erskine et al. 1989). The close resemblance detected between germplasm from adjacent climatically similar countries indicated that environmental conditions were the major evolutionary driving force in the case of the cultivated lentil. Furthermore, it appeared that the phenological adaptation to the environment was central to crop evolution. Landrace accessions from India and Ethiopia are grouped together with striking similar phenotypes for the nine quantitative traits surveyed. The group was characterized by early flowering and maturity, low biological yield, short stature, low pod height, and small seeds. This similarity for quantitative traits had already been noted by Barulina (1930).

South Asia is the world’s largest lentil-growing region; nevertheless, indigenous lentils (with specific ecotype pilosae) show a noticeable lack of variability. According to Erskine et al. (1998), this lack of variability is the result of a genetic bottleneck (most likely a founder effect) derived from lentil introduction into the Indian subcontinent from Afghanistan around 2,000 BC. Further evidence of the variation paucity within Indian germplasm resides in a sensitivity study of a world lentil collection to temperature and photoperiod (Erskine et al. 1990). In the agro-morphological study (Erskine et al. 1989), germplasm from Afghanistan was found markedly different to the Indo-Ethiopian group and had a greater similarity to lentil germplasm from Iran and Turkey. Of all the accessions tested, the Afghan accessions were among the latest to mature. Germplasm from Pakistan was also of the pilosae group for qualitative characters, yet intermediate between Afghan and Indian material for quantitative traits. Remarkably, countries with a much smaller area dedicated to lentil production (e.g., Lebanon and Chile) compared to South Asian countries (e.g., India and Pakistan) showed a higher mean coefficient of variation for the nine morpho-agronomic characters scored (Erskine et al. 1998).

6.4.2 Biochemical and Molecular Diversity

The introduction of biochemical and molecular markers in Lens diversity studies represented a leap forward, overcoming the previous limitations imposed by the relative low number of morphological characteristics recorded. Isozymes and storage proteins were first used in the 1980s to study genetic control and differences between Lens species and among lentil germplasm collections and to lay out geographical distribution (Skibinski and Warren 1984; Zamir and Ladizinsky 1984; Pinkas et al. 1985; Hoffman et al. 1986; Muehlbauer et al. 1989; Rosa and Jouve 1992). The use of isoenzymatic and protein SDS-electrophoresis procedures rapidly declined when DNA markers were introduced. Restriction fragment length polymorphism (RFLP) markers were the first molecular markers used to undertake lentil genetic and linkage analyses (Havey and Muehlbauer 1989). Subsequently, random amplified polymorphic DNA (RAPD) was used, among other purposes, to study the diversity, phylogeny, and taxonomy of Lens (Ford et al. 1997; Ferguson et al. 1998a, b, c). The most extensive analysis of the genetic variation in cultivated lentil was carried out by Ferguson et al. (1998c) who analyzed isozyme and RAPD variation in accessions derived from 16 different countries. The germplasm of Afghanistan clustered with that of South Asian countries. Accessions from these two areas were strikingly different to all other countries studied and showed low levels of variability. Another remarkable result was that the classification into macrosperma or microsperma types did not reflect overall country relationships, thus suggesting that seed size variation has little relation to between-country genetic similarities. However, different results in relation to seed size were observed by Duran and Pérez de la Vega (2004). Using RAPD and ISSR markers, they found that continental Spanish macrosperma and microsperma were more similar between them than with microsperma from different origins. Most of the studies on genetic variability after 1998 have been limited to local or regional germplasm samples (Sonnante and Pignone 2001, 2007; Durán and Pérez de la Vega 2004; Babayeva et al. 2009; Fiocchetti et al. 2009; Toklu et al. 2009). The general overview is that genetic variability in lentil is higher in or near the area where the domestication of lentil occurred in the Fertile Crescent.

In relation to the wild Lens species, the analysis of molecular variation with isozyme and RAPD data has revealed that between 78 and 99 % of the variation is attributable to between-population differences (Ferguson et al. 1998b). Areas of high diversity and unique diversity were located for four wild species, thus indicating regions where further germplasm collection is likely to yield novel genetic variation (Ferguson et al. 1998a). A systematic comparative study on the extent of genetic variation in Lens is still lacking, but in general it is considered that wild species hold more genetic variation compared to cultivated materials. In addition, when compared to other diploid sexually reproduced and cultivated crop species, lentils have less variation but show similar levels to that of the other cool season grain legumes.

Arbitrarily produced amplified fragment length polymorphism (AFLP) markers have also been used to study genetic diversity (Sharma et al. 1996) and to differentiate cultivars (Závodná et al. 2000; Fiocchetti et al. 2009). More recently, simple sequence repeat (SSR or microsatellite) and inter-simple sequence repeat (ISSR) markers have been included in the list of markers managed in lentils. Sets of microsatellite markers have been developed by Závodná et al. (2000) and Hamwieh et al. (2005, 2009) at the ICARDA, while other sets (also functional markers) are currently being produced at the Universidad de León (de la Puente 2012). In addition, cross-genera SSRs have been used in a diversity analysis of Lens species (Reddy et al. 2010). Muehlbauer et al. (2006) reviewed the molecular marker use in lentil. During the first decade of the twenty-first century, numerous studies regarding lentil diversity using biochemical and/or molecular markers have analyzed local or regional germplasm collections (Sonnante and Pignone 2001, 2007; Durán and Pérez de la Vega 2004; Scippa et al. 2008; Yuzbasioglu et al. 2008; Inder et al. 2008; Fiocchetti et al. 2009; Babayeva et al. 2009; Hamwieh et al. 2009; Toklu et al. 2009). The extensive use of single nucleotide polymorphisms (SNPs) will further improve the usefulness of molecular markers applied to lentil for basic genetic studies and breeding.

6.5 Germplasm Collection

Managed germplasm collections are available for many legume species, including lentil; characterization of the genetic diversity within these collections is a necessary prelude to their efficient use (Varshney et al. 2009). At least 43 lentil germplasm collections are maintained across the globe for breeding and research purposes. Muehlbauer et al. (1995) reviewed the most important institutions holding lentil collections. With a conservation capacity of more than one thousand accessions, the International Center for Agricultural Research in the Dry Areas (ICARDA; http://www.icarda.org), the Indian Agricultural Research Institute (http://www.iari.res.in), the N I Vavilov Research Institute of Plant Industry (http://www.vir.nw.ru), the National Bureau of Plant Genetic Resources, India, and finally the USDA collection at the Regional Plant Introduction Station, Pullman, Washington (http://www.ars.usda.gov) stand out. Another national germplasm bank holding an important collection of lentils is the Spanish Plant Genetic Resource Center (http://www.inia.es) with more than 600 accessions. A complete picture of the genetic resources available can be obtained at the germplasm collection directory web page of Bioversity International (http://www.bioversityinternational.org/), which integrates information regarding most of the world germplasm collections. For instance, the query on Lens culinaris landraces and traditional cultivars generated in the year 2012 a list of 43 gene banks and institutions holding a total of 12,097 accessions, whereas the query on wild or weedy forms of Lens yielded a list of six institutions maintaining only a total of 296 accessions (Pérez de la Vega et al. 2012). The GapAnalysis tool (http://gisweb.ciat.cgiar.org/GapAnalysis/) under a wild lentil search describes gaps in the collected Lens gene pool and provides recommendations for new recollection priorities, particularly L. tomentosus, which is given a high priority and of which only 4 gene bank accessions exist worldwide. Aside from the relatively low number of less than 300 wild accessions indexed by the germplasm collection directory web page of Bioversity International and maintained at six institutions, it is also remarkable that no genetic stocks, introgressed forms, or mutants were listed in 2012 among the available materials (Pérez de la Vega et al. 2012). Nevertheless, obtaining new useful mutants by artificial mutagenesis has been attended to in lentils. Chemical mutagens and ionizing radiation have been used to produce mutant lentil lines released for commercial production (review by Toker et al. 2007), and gamma ray-induced mutations have been used to obtain and release new varieties of outstanding performance (“NIA- MASOOR-05”) in Pakistan (Ali and Shaikh 2007). Dixit and Dubey (1986) described the effect of alkylating agents to generate chromosome mutations, mainly translocations. Moreover, the world lentil core collections have been mainly developed from the ICARDA and USDA-ARS collections (Varshney et al. 2009), implying a great extent of overlapping; in addition, the possibility might exist misclassification for certain wild accessions given their great morphological similarity.

Ladizinsky et al. (1990) described tertiary trisomic lines of lentils with L. orientalis and L. ervoides. Trisomic plants were weaker than euploid plants, and many did not reach maturity. There are no further records of their use in genetic analyses or breeding. Colchicine-induced lentil polyploids have been reported (Singh et al. 1992). Polyploidy results generally in gigantism for most determinate plant parts yet causes reduced fertility and yield; nonetheless, fertility and yield increased dramatically after selection from C1 to C4 generations. However, no commercial tetraploid lentil varieties have so far been released.

Wild Lens species can constitute an important source of donor genes. For instance, resistance to Ascochyta blight was found in an ICARDA wild germplasm collection: 24 out of 86 accessions of L. culinaris ssp. orientalis were resistant, as were 12 out of 35 accessions of L. odemensis, 3 out of 35 accessions of L. nigricans, and 36 out of 89 accessions of L. ervoides (Bayaa et al. 1994). Ye et al. (2002) listed a series of cultivated and wild lentil materials as sources of partial resistance to Ascochyta blight, and Muehlbauer and Chen (2007) reported that the most prominent among the partially resistant germplasm accessions consisted of lentil accessions PI 339283, PI 374118, ILL5588, ILL5684, PR86-360, and ILL7537. Wild materials have also been used as a source of genes for winter hardiness. Assessment of 245 wild lentil accessions and 10 cultivated lentil lines in Turkey showed that, on average, L. orientalis exhibited the highest level of winter hardiness, whereas accessions of L. ervoides were the most susceptible (Hamdi et al. 1996). Cultivated materials with large seeds were found in West Asia to be less susceptible to winter cold than those with small seeds, with average temperatures below 10 °C from January to April; the yield of the large-seeded group was higher than the group of the small-seeded lentils, while the converse was true at higher temperatures (Erskine 1996).

L. nigricans provided the maximum number of resistant accessions for biotic (wilt, powdery mildew, and rust) and tolerance to abiotic stresses (Gupta and Sharma 2006). Tullu et al. (2006) tested all six wild Lens species for resistance to Colletotrichum truncatum. Accessions of L. ervoides revealed the highest level of resistance (3–5) to races Ct1 and Ct0 followed by L. lamottei in both field and greenhouse assays. L. orientalis, L. odemensis, L. nigricans, and L. tomentosus were highly susceptible (8–9) to race Ct0 in the greenhouse. The highest levels of resistance were found among L. ervoides accessions originating from Syria and Turkey.

Resistance against the parasitic plant Orobanche crenata was found in wild Lens materials (Fernández-Aparicio et al. 2007, 2009). The screening of 23 wild accessions for resistance to O. crenata under field conditions showed a wide range of responses from complete resistance to high susceptibility. The highest levels of resistance were observed in accessions of L. ervoides, L. odemensis, and L. orientalis. Likewise, resistance against Sitona crinitus Herbst, a major insect pest limiting lentil productivity mainly in the countries of West Asia and the North African region, was found in wild species (L. ervoides, L. odemensis, L. nigricans, and L. c. ssp. orientalis) (El-Bouhssini et al. 2008). A lentil germplasm collection of 571 accessions from 27 countries including wild species was screened for susceptibility to seed bruchids under natural field conditions in Central Spain (Laserna-Ruiz et al. 2012); a total of 32 accessions including cultivated landraces, L. c. ssp. orientalis, L. nigricans, and L. lamottei showed lower infestation rates than the susceptible check and were selected as potential sources of resistance to seed weevil (Bruchus spp.).

Resistance to fusarium wilt has been described in germplasm from India (Kamboj et al. 1990); and Bayaa et al. (1997) screened a core collection of 577 germplasm accessions from 33 countries and rescreened a subset of the 88 mostly resistant accessions. A total of 1,771 lentil accessions from the US lentil collection (Pullman, WA) and the Institut für Pflanzengenetik und Kulturpflanzenforschung (Gatersleben, Germany) were screened for resistance to C. truncatum; about 95 % of the accessions were susceptible when inoculated with a single isolate in the field, and 16 accessions displayed resistance (Buchwaldt et al. 2004). Tullu et al. (2006) searched for sources of resistance to anthracnose in cultivated and wild materials.

6.6 Production-Related Problems

Muehlbauer et al. (2006) pointed out that several abiotic stresses (drought, heat, salinity, iron deficiency, among others) are important constrains to lentil production worldwide. The severity of these stresses is unpredictable in field experiments; hence, field trials are increasingly being supplemented with controlled environment testing and physiological screening. Lentils are mainly grown as a rainfed crop and often are seeded in marginal areas where limited rainfall and deficient soil moisture conditions are encountered. Thus, lentil breeding for improved drought resistance is a major goal for most regions where lentils are grown and particularly in areas of limited and erratic rainfall such as the Middle East and areas of the Mediterranean climate. Drought-tolerant cultivars are required to stabilize lentil production and also to extend cultivation to further dry areas. Muehlbauer et al. (1995) indicated that screening for drought tolerance was difficult due to the lack of knowledge of what to screen for and the absence of efficient screening techniques.

Screening for drought tolerance has improved with controlled environment testing and physiological screening. In a recent review on screening techniques and sources of resistance to abiotic stresses in cool season food legumes (Stoddard et al. 2006), it was stated that over the last decade, there has been considerable progress to increase the adaptation of lentil to autumn sowing, and a large body of literature exists regarding drought responses, particularly in chickpea and faba bean. Yet the available literature on salinity tolerance and waterlogging is minimal. The review of Stoddard et al. (2006) also summarizes information on identified lentil sources of resistance or tolerance to drought and heat stress [e.g., MI-30, a mutant line showing high yields under limited water supply identified by Salam and Islam (1994)], to chilling or freezing stress (Hamdi et al. 1996; Ali et al. 1999; Kahraman et al. 2004a), to salinity stress (Mamo et al. 1996; Ashraf and Zafar 1997; Rai and Singh 1999), to waterlogging (Ashraf and Chishti 1993), and to high boron concentrations (Hobson et al. 2006). Another very important component of drought resistance is the plant root system. Sarker et al. (2005) analyzed the variation in shoot and root characteristics and their association with drought tolerance in several lentil landraces. Stem length, taproot length, and lateral root number were highly correlated, both among themselves and with yield. High heritability estimates provided reliability for screening based on these traits. Among a total of 40 genotypes, only one line (ILL 6002) was strikingly different to all other tested genotypes. Another approach to drought resistance screening is to select for drought avoidance through early maturity, particularly when the major moisture stress occurs at the end of the growing season. Early planting can contribute to this avoidance since lentil plants usually develop and initiate flowering before serious drought conditions prevail. This is the case for fall versus spring lentil sowing in cold areas. On the other hand, among cool season legumes, lentils seem to be the least tolerant to transient waterlogging in greenhouse experiments (Solaiman et al. 2007; Singh et al. 2013a).

A factor for increasing production in the Mediterranean region and other areas with cool winters consists of earlier sowing in the fall rather than at the end of the winter. Available winter-hardy lentil germplasm has prompted interest in the development and use of cultivars that can be sown in the fall in cold highland areas. This change in production of lentils from normal spring sowing to fall sowing increases yield potential (Kahraman et al. 2004a; Barrios 2012). Early sowing has proved to increase lentil production in several countries, for instance, by shifting sowing from spring to early spring or to autumn sowing in areas such as the highlands of Turkey, Iran, Afghanistan, the Baluchistan area in Pakistan, and also Spain. Early fall-seeded lentil increased yield from 480 to 590 kg ha−1 when compared to early spring sowing of non-winter-hardy varieties in the US Pacific Northwest and northern Great Plains of Canada. Fall sowing in Australia allowed for a longer period of vegetative and reproductive growth, rapid canopy development, greater absorption of photosynthetically active radiation, as well as a better water use efficiency, increased dry matter, and improved seed production when compared to a later sowing. Early sown lentils began flowering and forming seeds ahead in the growing season, at a time when vapor pressure deficits and air temperatures were lower, using more water in the post-flowering period compared to the delayed sowing treatments (Siddique et al. 1998; Chen et al. 2006). Delaying spring seeding in the northern Great Plains of Canada by 4 weeks negatively affected crop development parameters due to the exacerbation of summer drought stress. Seed yields decreased at the majority of site years and the production reduction averaged 38 %. Seeding delays of 2 and 4 weeks decreased average seed size by 10 % (Miller et al. 2006). However, early sowing increased the incidence of some fungal and bacterial diseases (Davidson and Kimber 2007). According to Muehlbauer et al. (1995), diseases of lentil are in general less damaging than those of most other food legume crops. However, there are some important and potentially devastating diseases (Taylor et al. 2007). Several reviews on screening techniques and sources of resistance to root and foliar diseases in cool season grain legumes have been published (Infantino et al. 2006; Sillero et al. 2006; Tivoli et al. 2006).

The root rot/wilt complex is probably the most important lentil disease problem worldwide (Muehlbauer et al. 1995). This set of diseases is caused by several fungi of the genera Fusarium, Pythium, Rhizoctonia, and Sclerotinia, in particular F. oxysporum f. sp. lentis, R. solani, and S. sclerotiorum. Another fungus which can be included in this complex is Aphanomyces euteiches, a soil-borne pathogen causing Aphanomyces root rot of several legumes, including alfalfa, bean, lentil, and pea. Relatively little is known about the population biology of this legume pathogen (Grunwald and Hoheisel 2006). The inheritance of resistance to Fusarium vascular wilt was investigated by Eujayl et al. (1998a, b) who found several random amplified polymorphic DNA (RAPD) markers linked to the Fw locus. Wang et al. (2006) found small but consistent differences among 14 lentil cultivars with regard to their reaction to Rhizoctonia seedling blight and root rot caused by Rhizoctonia solani in greenhouse trials and in field experiments conducted in Canada. Fusarium redolens has also been described as a causal agent of Fusarium wilt in lentil (Riccioni et al. 2008).

Rust caused by Uromyces viciae-fabae Pers. is a serious problem in areas where mild temperatures and humid conditions favor the development of this fungus (Muehlbauer et al. 1995). Incomplete (partial) resistance to Uromyces has been described in lentil cultivars (Negussie et al. 2005). These authors analyzed the response to four components of resistance, namely, latent period, infection efficiency, pustule size, and spore production and concluded that, since there was an interdependence of the components, selection based on more than one component should help to obtain lines with higher levels of partial resistance. Uromyces-resistant cultivars, such as Laird or Calpun-INIA derived from Laird, have been described (Penazola et al. 2007). Ascochyta blight, a seed-borne disease described in at least 16 countries across five continents, is probably one of the most widely distributed and devastating diseases. The fungus, Ascochyta lentis (syn. A. fabae f.sp. lentis; teleomorph: Didymella lentis), attacks leaves, stems, and pods causing necrotic spots. In some cases it kills plants before harvest or infects seeds to such a degree that lentils are unmarketable. Ascospores of Didymella lentis can be found on lentil stubble under natural conditions, and the presence of the teleomorph has implications in the long-distance dispersal of A. lentis (Galloway et al. 2004). Species of the genus Ascochyta cause most of the major diseases of many cool season grain legumes, and a considerable amount of information about the pathogens, plant hosts, and disease resistance responses is available. Recently, Phytopathologia Mediterranea (Millan 2013) has brought out a special issue on Ascochyta.

The mating type (MAT) locus of Ascochyta lentis was cloned and characterized, and a multiplex PCR assay for mating type was developed based on the MAT idiomorph and flanking sequences (Cherif et al. 2006). Molecular diagnostic techniques have been developed to differentiate the Ascochyta pathogens that infect cool season food and feed legumes, as well as to improve the sensitivity of detecting latent infection in plant tissues (Taylor and Ford 2007). Recent multilocus phylogenetic analyses of a worldwide sample of Ascochyta fungi causing Ascochyta blights of chickpea (Cicer arietinum), faba bean (Vicia faba), lentil (Lens culinaris), and pea (Pisum sativum) have revealed that the fungi causing disease of each host formed a monophyletic group. Host inoculations of these fungi demonstrated that they were host-specific, causing disease only on the host species from which they were isolated (Hernández-Bello et al. 2006; Peever 2007). Tivoli et al. (2006) provided a thorough review on the sources of resistance and screening techniques for Ascochyta blight in lentil, and Ye et al. (2002) gave an account of suitable breeding techniques for the selection of lentils with resistance to Ascochyta blight; partial resistance to the disease is available in germplasm (Muehlbauer and Chen 2007).

Anthracnose is another important disease in lentil caused by the fungus Colletotrichum truncatum (Schwein) Andrus and Moore [teleomorph Glomerella truncate, described by Armstrong-Cho and Banniza (2006)]. The presence of two distinct pathogen races (Ct0 and Ct1) was demonstrated (Buchwaldt et al. 2004), and Tullu et al. (2006) searched for sources of resistance to anthracnose in cultivated and wild materials, localizing two resistant cultivars (Indianhead and PI 320937) to race Ct1 but none to race Ct0. Some accessions of wild Lens species are potential sources of genes for resistance to both races. Transfer of resistance to both races, Ct1 and Ct0, from L. ervoides accession L-01-827A to the cultivar Eston by single-seed descent from F2 individuals to F7:8 RILs was described by Fiala et al. (2009). Following rooting and hybrid clonal propagation procedures developed for lentil (Fratini and Ruiz 2003, 2008), Vail and Vandenberg (2010) produced a clonal propagation protocol to evaluate disease data of lentil infections. These authors tested the feasibility of cloning individual lentil plants to generate replicated ratings in order to evaluate the response to anthracnose, proposing that this method is more dependable than using single-plant evaluations. Powdery mildew is a minor disease in lentil, predominantly present under controlled greenhouse conditions, even though it can severely affect certain lentil cultivars in some parts of the world. Erysiphe pisi is the pathogen which has been detected in several countries, although E. trifolii has recently been added as a causal agent of this disease in the Pacific Northwest of the United States (Attanayake et al. 2009).

Genotypes with resistance to various biotic stresses, particularly resistance to vascular wilt, rust, and Ascochyta blight, have been identified and directly exploited or used in breeding programs (Sarker and Erskine 2006). Inclusive multiple resistant materials exist among both breeding lines and landraces. For instance, Laird is a macrosperma multiresistant cultivar, and Assano is a recently released cultivar highly resistant to rust caused by Uromyces viciae-fabae with also moderate resistance to Ascochyta blight and to the wilt-root rot complex (Fusarium sp. and Rhizoctonia sp.) (Fikru et al. 2007). A Pakistani accession, 66013-6, was found to have a high level of resistance to blight, rust, and pea seed-borne mosaic virus (Hussain et al. 2008).

Lentils have been reported to be susceptible to more than 25 viruses according to the list supplied by the Plant Virus Online (http://www.agls.uidaho.edu/ebi/vdie/). In a field experiment, 12 out of 16 lentil genotypes were ranked as highly susceptible and four as susceptible to the pea seed-borne mosaic virus (PSbMV). Once infected, plant sensitivities (symptom severities) were low in most lentil genotypes. Seed lots harvested from PSbMV-infected plants of lentil were usually less affected than other cool season grain legumes. Seed transmission of PSbMV (6 %) was detected in seed from infected lentil plants (Coutts et al. 2008). In a study (Bashir et al. 2005) to assess the amount of variation among lentil germplasm with regard to infection reaction against several viruses and also to identify sources of resistance, most of the lentil genotypes were susceptible to highly susceptible, and none was found to be highly resistant although some genotypes were found to be resistant and moderately resistant. Some differences in the sequence of the eIF4E lentil gene have been found in lentil (Saenz de Miera, pers. comm.); this gene is related to virus resistance in pea and other plants species (Robaglia and Caranta 2006).

Another major threat with regard to lentil cultivation in the Mediterranean area and Western Asia consists of crenata broomrape (Orobanche crenata). This weedy root parasite genus constitutes a serious threat to lentils. Broomrape control is difficult due to the large number of wind-blown seeds which are produced and dispersed each year and which may thereafter remain dormant in soils for several years. The control strategy has consisted in applying different agronomic practices and herbicides. Resistance breeding so far is hampered by a scarcity of proper resistance sources in addition to a reliable and practical screening procedure. A Spanish germplasm collection of 234 accessions of lentils was screened for resistance to crenate broomrape under field conditions (Fernández-Aparicio et al. 2007), followed by a germplasm screening of 23 wild accessions (Fernández-Aparicio et al. 2009). A wide range of responses was observed; however, complete resistance was not detected. Other Orobanche species can also parasitize lentil plants (Muehlbauer et al. 1995).

Lentil is damaged by many types of insects and other pests. Among insects, the major pests in the field are aphids (Aphis craccivora, Acyrthosiphon pisum), leaf weevil (Sitona spp.), lygus bugs (Lygus spp.), and the cutworm (Agrotis ipsilon). Another major pest problem causing great seed losses are the seed insect species: Bruchus ervi and B. lentis and also Callosobruchus chinensis and C. maculatus (Stevenson et al. 2007). Little progress in breeding for insect resistance was described by Muehlbauer et al. (1995), and compared to breeding for diseases, the situation continues to be similar nowadays. No sources of resistance to Sitona were detected, and ICARDA materials showed only to be effective against the adult insect while not to larvae (Stevenson et al. 2007). However, resistance to S. crinitus Herbst has been described in wild materials (El-Bouhssini et al. 2008). An extensive search on bruchid species infecting lentil was carried out in 10 years by the National Bureau of Plant Genetic Resources, New Delhi (Bhalla et al. 2004). A total of 2,517 imported lentil samples from 40 countries were X-ray screened for hidden insects in seeds. About 30 % of the samples were infected, and the species found consisted of Bruchus ervi Froelich, B. lentis Froelich, B. tristiculurs Fahraeus, Callosobruchus analis (Fabr.), C. chinensis (L.), and C. maculatus (Fabr.). After screening 571 accessions including wild species of the Spanish germplasm collection for susceptibility to seed bruchids under natural field conditions in Central Spain, a total of 32 accessions including cultivated landraces, L. c. ssp. orientalis, L. nigricans, and L. lamottei were selected as potential sources of resistance to Seed weevil (Laserna-Ruiz et al. 2012).

6.7 Traits of Importance for Base Broadening

6.7.1 Pod and Seed Characteristics

Reduced shattering is a trait of great economic value as pod dehiscence can cause significant losses before or during harvest; the trait is considered to have played a major role in lentil domestication. Pod dehiscence was found to be completely dominant over indehiscence and was assigned the gene symbol Pi (Ladizinsky 1979; Vaillancourt and Slinkard 1992); nonetheless, response to selection for pod indehiscence indicates the presence of quantitative variation (Erskine 1985), which was confirmed by quantitative trait loci (QTL) analysis that determined one major recessive QTL and two minor dominant QTLs which together accounted for 81 % of the observed variation (Fratini et al. 2007). Pod dehiscence in segregating populations shows several degrees between completely dehiscent and non-dehiscent plants, therefore classified since 1985 by the IBPGR/ICARDA lentil descriptor into four classes on a scale from 0 to 9 when scored 1 week after maturity.

Seed size is an important economic trait with special attributes in lentil consumption and trade, considering that wholesale prices of small-seeded (microsperma) and large-seeded (macrosperma) varieties differ by a large margin depending on consumer preferences and farmers’ choice. Cultivated lentils are divided on the basis of seed size differences into microsperma (<4.5 mm) and macrosperma (>4.5 mm and sometimes over 7 mm). Seed size has a continuous distri bution in F2 progenies following crossing of large- and small-seeded types (Ladizinsky 1979), and at least two major additive QTLs and one minor dominant QTL have been described to determine seed size variation (Fratini et al. 2007). Recombinant inbred populations (RILs) derived from the interspecific cross with L. ervoides (Fiala et al. 2009) revealed transgressive segregants with an 8 % increase in seed size (Tullu et al. 2011). With regard to seed mass, QTL analysis has shown a polygenic control with partial dominance for low seed weight alleles (Abbo et al. 1991), which has also been confirmed in a second study which describes two recessive and one additive QTL associated to low seed weight (Fratini et al. 2007).

Imbibition is one of the first steps in breaking seed dormancy to initiate seed germination; dormancy and viability can be maintained over long periods of time in hard-seeded legumes because seed coats are impermeable to water, some legume seeds being capable of remaining viable for more than 100 years (Rolston 1978). This is important ecologically in wild populations and economically in cultivated legumes, since seed coat hardness conditions cooking and palatability. Typically, wild Lens accessions are hard seeded, whereas cultivated varieties are not. Seed hardness is an undesired agronomic character, when lentil was domesticated from wild populations seed hardness most probably was selected against – especially considering that among pulse crops, lentil stands out for their fast imbibition and cooking time. Reports on the inheritance of seed hardness in lentil are highly conflicting, particularly considering that the quantitative nature of the trait has been thoroughly demonstrated since 1990 in the closely related grain legume Glycine max. (Keim et al. 1990); however, in lentil Ladizinsky (1985) intially described a single recessive gene in crosses with ssp. orientalis compared to a single dominant gene in crosses with L. ervoides, nonetheless thereafter it has persistently been stated to be a monogenic trait and seed hardness was designated the gene symbol Hsc which in addition is assured to be linked to the Pi locus for pod dehiscence (Sharma 2011).

Lentil is a rich source of dietary proteins among crops that are consumed without industrial processing. Furthermore, lentil is possibly the richest source of proteins among edible pulses that are cooked and directly consumed without prior processing for quality alterations. The usual protein content in dry lentil is around 26 %, although a germplasm range from 20.4 to 29.8 % has been reported (Sharma 2011). Studies to analyze the inheritance of seed protein concentration in lentil (Hamdi et al. 1991; Chauhan and Singh 1995; Tyagi and Sharma 1995) have revealed the quantitative nature of this trait and a nonsignificant correlation with grain yield and seed size (Hamdi et al. 1991; Tyagi and Sharma 1995). According to Sharma (2011), in most crops, especially cereals, protein content is invariably negatively correlated to seed size; however, lentil is possibly an exception and protein content has been claimed to be positively (although mildly) correlated to seed size.

6.7.2 Flowering and Plant Architecture

Flowering time is particularly important for adaptation and yield; it determines the length of the vegetative phase and conditions crop exposure during reproductive growth to climatic settings. Selection for response to specific regional balances between photoperiod and temperature for the onset of flowering has played a vital role in the adaptation of lentil to different regions and conditions around the globe (Erskine et al. 1990). Earliness is a desirable trait ensuring completion of the crop cycle in a relatively short period of time, thereby making more efficient use of resources as well as avoiding losses due to high temperatures during crop maturation. The inheritance of flowering time was described as monogenic, with earliness being recessive, and the gene symbol Sn was assigned (Sarker et al. 1999); in spite of the authors also observing transgressive segregation for early flowering, they considered the transgressive phenotypes to be a consequence of the interaction between the recessive Sn locus and a polygenic system of minor earliness genes.

Continuous polygenic variation had already been described in another analysis regarding flowering response in crosses between early maturing microsperma varieties from India and an early maturing macrosperma variety from Latin America (Tyagi and Sharma 1989). Transgressive segregation was observed for these crosses hinting strongly to the quantitative nature of flowering time. As crosses between microsperma varieties did not produce transgressive flowering segregants, they concluded that microsperma and macrosperma lentils possess different sets of genes controlling flowering time, while the Indian microsperma genotypes all share the same gene pool with regard to flowering response. Finally, three QTLs that accounted for more than 90 % of the observed variation in flowering response were detected, one was a major recessive gene while the remaining two were minor and dominant (Fratini et al. 2007). As the Sn locus was shown to be linked to Scp (seed coat pattern) and the latter major recessive QTL was located in the linkage group (LG) containing Gs (green stem), it seems likely that flowering response is under a complex genetic control with several qualitative and quantitative genes segregating differentially in diverse crosses. As a matter of fact, Tullu et al. (2008) reported two QTLs affecting earliness which accounted for 37–46 % of the variation observed in number of days to flower. Recently, a functional SSR marker in the TFL1 gene related to earliness has been described (de la Puente et al. 2013).

Plant structure holds many implications regarding yield and ease of harvesting. Incomplete dominance of a bushy/erect growth habit was reported and the gene symbol Gh proposed (Ladizinsky 1979), later, the recessive Ert gene was discovered by Emami and Sharma (1999) and mapped in the LG containing Gs together with Rdp (red pod) and Bl (brown leaf). The erect phenotype Ert is recessive to the most prevalent growth habit in cultivated lentil in which plants remain procumbent for about a month, and thereafter branches grow in a semierect or semi-spreading fashion, while the wild Lens species retain prostrate stems much longer. Although according to Sharma (2011) there is a gradation between spreading and completely erect growth habit among the cultivated lentil varieties, genotypes with spreading growth habit can be grouped into several categories from highly prostrate to more upright; the spreading genotypes are generally endowed with profuse branching (basal as well as secondary). With regard to other plant structure attributes, QTL analysis has revealed that the number of branches at the first node is mainly controlled by a dominant quantitative locus together with two minor loci of opposing effects which together explain 92 % of the observed variation. Two additive QTLs explained one-third of the variance for the height of the first node. Finally, one recessive and one dominant QTL jointly explained half of the variation encountered for the total number of branches (Fratini et al. 2007).

Plant height is strongly associated to yield index. Plants of erect and a very tall habit tend to lodge as they approach maturity; thus, erect varieties of medium height should be expected to be more lodging resistant, which results in higher yields and also to a more amenable mechanical harvesting. Tallness in plants has almost universally been shown to be dominant over dwarfness since Mendel’s times. In lentil, the gene Ph for plant height was first reported (Tahir et al. 1994) and found to be dominant over dwarfness. Ph is located in an LG comprising eight morphological (Kumar et al. 2005) and more than a dozen isozyme markers (Tahir et al. 1994). In addition, plant height was also found to be quantitatively inherited (Haddad et al. 1982), while another study described one additive QTL and two recessive QTLs, but the three QTLs explained only 38 % of the observed variation in plant height (Fratini et al. 2007); likewise, five QTL for plant height accounted for 31–40 % of the observed variation (Tullu et al. 2008).

Flower number per peduncle in lentil may retain significance in relation to productivity potential. This agronomic trait is known as prolificacy and is defined as the ability of a genotype to produce many flowers and ultimately pods on each peduncle (Sharma 2011). Genetic analysis is complicated because of an inconsistent expression of the trait which is highly influenced by environmental conditions and declines as plants get older. Nonetheless, monogenic inheritance with the two-flower phenotype dominant over the three flowered was initially described (Gill and Malhotra 1980), followed by contrasting results from different studies that concluded that a higher flower number was dominant over a low flower number per peduncle (Sharma 2009).

It can be concluded from this section that contrasting results have been obtained for several morphological markers; some studies considering monogenic inheritance while other studies have determined quantitative variation. These contrasting results are not unexpected since many characters are controlled by both qualitative (e.g., dwarf plants insensible to phytohormones) and quantitative genes (generating variation within dwarf and “normal” plants). Anyway, an extensive revision of the status of several morphological traits is necessary. A very comprehensive list of the gene symbols used for all of the lentil traits of simple monogenic inheritance described so far, some of which are oligogenic instead of being strictly monogenic, together with the different linkage studies in which they have been used can be found in Pérez de la Vega et al. (2012); the list also provides references of linked markers to biotic disease resistance loci and to the abiotic radiation-frost tolerance gene.

6.7.3 Abiotic Stresses

Tolerance to frost injury is an essential requirement when lentils are sown during the winter and cultivated in cooler climates. Monogenic inheritance of radiation-frost tolerance was reported and assigned gene symbol Frt; the gene was also tagged with a random amplified polymorphic DNA (RAPD) marker at a distance of 9.1 cm (Eujayl et al. 1999). On the other hand, it has been concluded that winter hardiness in lentil is a polygenic trait, and several QTLs together accounted for 42 % of the variation observed in recombinant inbred lines (RILs) (Kahraman et al. 2004a, b). At least four QTLs were detected under controlled frost conditions and field conditions in two RIL populations; furthermore, two QTLs related to frost response were also related to yield under winter-sown conditions (Barrios et al. 2007); QTLs with a major effect for winter hardiness and yield seem to be closely located within a single linkage group, and some molecular markers are potentially useful for their tracking (Barrios 2012). SuperSAGE expression profiling in response to some biotic and abiotic stresses has been carried out within the LEGRESIST and other projects, for instance, lentil SuperSAGE genomic analysis was used to analyze the allele-specific differential expression of transcripts potentially involved in frost tolerance by bulk segregant analysis among 90 F9 RILs derived from the Precoz × WA8649041 lentil cross (Barrios et al. 2010).

Lentil has been traditionally grown in semiarid regions under rainfed conditions; thus it combines a high degree of drought resistance and a low water requirement; in fact, excessive water supply is damaging to the crop. The genetics of drought tolerance or waterlogging are still to be explored.

6.7.4 Biotic Stresses

Most genetic studies regarding rust (Uromyces fabae) resistance in lentil have reported that resistance is under a monogenic control, resistance being dominant over susceptibility. However, reports of incomplete resistance (Negussie et al. 2005), as well as duplicate dominant genes controlling resistance, have frequently emerged; only one unconfirmed report has suggested rust resistance to be a recessive trait (see Sharma 2009 for a review). Furthermore, it has been observed that the dominant gene for resistance of a macrosperma variety from Latin America differs from that of an Indian microsperma variety (Sharma 2009, 2011). Therefore, it seems that at least two separate genes controlling rust resistance have evolved in spatially and temporally isolated lentil groups. Gene symbols urf1, urf2, and the unconfirmed urf3 have been proposed (Sharma 2009). Uromyces-resistant cultivars, such as Laird or Calpun-INIA derived from Laird, have been described (Penazola et al. 2007); in addition, Assano is a cultivar highly resistant to Uromyces with moderate resistance to Ascochyta blight and to the wilt-root rot complex (Fikru et al. 2007).

Resistance to Fusarium wilt was described in a germplasm collection from India (Kamboj et al. 1990). Allelism tests concluded that Fusarium resistance in lentil is conferred by five dominant genes (Kamboj et al. 1990). However, subsequently only one dominant gene (Fw) for wilt resistance was reported. It was tagged with a RAPD marker at 10.8 cm (Eujayl et al. 1998b); a microsatellite marker and an AFLP marker were further linked to the Fw locus at distances of 8.0 and 3.5 cm, respectively (Hamwieh et al. 2005). Moreover, a 3-year screening of lentil germplasm at the ICARDA yielded a total of 34 strains confirmed for resistance to Fusarium wilt. Evaluations from the F5 to F8 successfully identified 753 resistant lines. Among the small-seeded lines, 72 % were wilt resistant compared with 41 % of the large-seeded lines, suggesting that genes for small seed size might be loosely associated with genes for wilt resistance (Sarker et al. 2004).

Resistance to Ascochyta blight, screening techniques, and breeding methods have been reported (Bayaa et al. 1994; Ye et al. 2002; Muehlbauer and Chen 2007; Tivoli et al. 2006). The genetic control of Ascochyta blight resistance was described to be monogenic recessive in spite of having observed transgressive segregations (Tay and Slinkard 1989). However, two complementary dominant genes were further described in a cross between L. ervoides and L. odemensis (Ahmad et al. 1997), while only one dominant gene was found in crosses between L. culinaris accessions (Ahmad et al. 1997). The existence of two complementary dominant genes within cultivated lentils was thereafter established (Nguen et al. 2001). Two flanking RAPD markers at distances of 8.0 and 3.5 cm from the designated resistance locus Ral1 (Abr1) have been mapped (Ford et al. 1999); likewise, two additional RAPD markers have been located in flanking positions of the recessive gene for resistance ral1 at distances of 6.4 and 10.5 cm (Chowdhury et al. 2001). The first report on the use of the microarray technology to study gene expression in lentil was recently published (Mustafa et al. 2009). Highly resistant (ILL7537) and highly susceptible (ILL6002) lentil germplasm accessions were inoculated with Ascochyta lentis. Ninety genes were differentially expressed in ILL7537, and 95 genes were differentially expressed in ILL6002. The expression profiles of the two accessions showed substantial difference in the type of genes and the time of expression in response to the pathogen. The resistant variety showed early upregulation of PR4 and 10 proteins and other defense-related genes. The susceptible genotype showed early downregulation of defense-related genes. Real-time PCR (RT-PCR) was used to verify microarray expression ratios. The resistant and susceptible lentil accessions differed not only in the type of genes expressed but also in the time and level of expression in response to A. lentis inoculation (Mustafa et al. 2009).

Tullu et al. (2006) searched for sources of resistance to anthracnose caused by Colletotrichum truncatum in cultivated and wild materials and found resistant cultivars (Indianhead and PI 320937) to race Ct1 but none to race Ct0. Some accessions of wild Lens species are potential sources of genes for resistance to both the races. Transfer of resistance to both races, Ct1 and Ct0, from L. ervoides accession L-01-827A to the cultivar Eston through embryo rescue followed by single-seed descent from F2 individuals to F7:8 RILs was described (Fiala et al. 2009); authors indicated that resistance to race Ct1 and race Ct0 may be conferred by two recessive genes. Resistance to Ct1 race was reported to be under the control of one recessive gene (lct-1) in one cultivar, while two dominant genes designated LCt-2 and LCt-3 were respectively responsible for resistance in two additional cultivars (Buchwaldt et al. 2004). The LCt-2 resistance locus was tagged with two flanking RAPD markers at 6.4 and 10.5 cm (Tullu et al. 2003); in addition, three AFLP markers were also found to be linked to this locus (Tar’an et al. 2003). However, most recently Buchwaldt et al. (2013) have established that resistance to Ct1 race is controlled by two recessive genes (lct-1 and lct-2) and three closely linked dominant genes (LCt-3, LCt-4 and LCt-5).

Pea seed-borne mosaic virus (PSbMV) is a major disease in lentil transmitted through seed as well as aphids. Monogenic recessive inheritance of viral immunity has been confirmed in four crosses, and the gene symbol sbv was proposed to denote PsbMV resistance in lentil (Haddad et al. 1978). Pakistan accession 66013-6 was found to have a high level of resistance to blight, rust, and pea seed-borne mosaic virus (Hussain et al. 2008).

6.8 Interspecific Hybridization in Crop Species

Artificial cross-pollination in a highly self-pollinated crop species, such as lentil, is important to increase genetic variability. Wide crosses to yield interspecific hybrids allow for the introgression of important alleles of agricultural interest from wild species to cultivars, as, for instance, the resistance or tolerance to abiotic and biotic stresses (Erskine et al. 1994; Ocampo et al. 2000; Davis et al. 2007; Cubero et al. 2009; Pérez de la Vega et al. 2012). Lentil is generally described as a strictly self-pollinated species holding cleistogamous flowers (Wilson 1972; Kumar and Singh 1998); nonetheless, recent studies point out that natural outcrossing rates of lentil can be relatively high (Horneburg 2006). Lentil flowers are complete with a typical structure of the subfamily Papilionaceae of the Leguminosae family (Muehlbauer et al. 1980), and as it was previously mentioned, emasculation and artificial crossing in lentil is a demanding task. In the case of interspecific hybridization, crossing efficiencies are much lower than intraspecific ones, for instance, out of a total of 1,707 pollinations, six interspecific hybrids with L. odemensis, two with L. nigricans, and one with L. ervoides were recovered using embryo rescue (Fratini and Ruiz 2006). Therefore, a large number of manual pollinations are needed to recover each single interspecific hybrid.

As has already been discussed previously, intraspecific crosses between cultivated lentils produce viable descendants (Malhorta et al. 1978; Solh et al. 1980; Mera and Erskine 1982; Kumar and Singh 1998; Fratini et al. 2004; Singh et al. 2013b). With regard to inter-subspecific hybrids of lentil, it has been reported that the domesticated lentil is readily crossable with subspecies orientalis (Ladizinsky 1979; Muehlbauer et al. 1989; Vandenberg and Slinkard 1989; Vaillancourt and Slinkard 1992; Fratini et al. 2004), although the fertility of the hybrids depends on the chromosome arrangement of the wild parent (Ladizinsky 1979; Ladizinsky et al. 1984). Lens odemensis produces partially fertile hybrids with both subspecies of L. culinaris depending on the accessions used (Cubero et al. 2009). Interspecific embryos between cultivated lentils and either L. ervoides or L. nigricans abort (Abbo and Ladizinsky 1991, 1994; van Oss et al. 1997; Fratini and Ruiz 2006) and embryo rescue techniques are necessary to recover hybrids (Cohen et al. 1984; Ladizinsky et al. 1985; Fratini and Ruiz 2006; Fiala et al. 2009). Nonetheless, gibberellic acid (GA3) application after pollination aided to develop viable pods and interspecific Lens hybrids were obtained without the need of in vitro embryo culture (Ahmad et al. 1995). To enhance the gene pool and the hybridization range of lentil, hand-pollination followed by embryo rescue is in most cases the only practical method to recover interspecific hybrids of Lens; thus, it is necessary to dispose of an efficient embryo rescue protocol.

6.9 Barriers to Interspecific Hybridization

Pollination followed by fertilization normally leads to the production of an embryo which in the intact plant is linked with normal seed development. Crossability was defined by Ladizinsky (1992) as the potential for intercrossing individuals belonging to the same or different taxa to produce embryos or seeds that can give rise to an F1 plant. Crossability is either limited by incompatibility or by incongruity barriers; the sexual barriers that belong to the second aspect have been further divided into pre- and postfertilization barriers (Stebbins 1958). Part of the postfertilization barriers may be overcome by using in vitro embryo rescue methods (Yeung et al. 1981; Fratini and Ruiz 2011), although depending on plant species, the process can entail the culture of ovaries immediately after pollination and/or in ovulo embryo culture and/or embryo culture. Ladizinsky (1992) explained that the success in lentil crosses depends on the interaction between the parental genomes in the hybrid zygote, embryo, and endosperm and between the hybrid tissue and the surrounding maternal tissue. The crossability between the cultivated lentils and the wild Lens relatives has been described to be hampered by pre- as well as by postfertilization barriers (Abbo and Ladizinsky 1991, 1994; Ladizinsky 1993; Ladizinsky and Abbo 1993; van Oss et al. 1997). It is now clear that epigenetic processes during embryo-endosperm development can also influence crossability at a post-zygotic level (Law and Jacobsen 2010).

Cohen et al. (1984) and Ladizinsky et al. (1985) were the first to describe hybrid embryo abortion after interspecific hybridization in lentil. Interspecific hybridization between the cultivated lentil and L. ervoides and L. nigricans resulted initially in pod development followed by an arrest 10–16 days after pollination with a consequent production of shriveled and nonviable seeds. This postfertilization barrier has prevented gene flow from the wild L. ervoides and L. nigricans species into the cultivated lentil, hindering the use of these wild forms for breeding purposes. The anatomical aspects of hybrid embryo abortion in the genus Lens were described by Abbo and Ladizinsky (1991); more recently Fratini and Ruiz (2006) also have provided a description of the appearance and frequency of aborted pods and ovules 14 days after pollination. So far the only study which describes the genetic aspects leading to embryo abortion of wide Lens hybridizations was carried out by Abbo and Ladizinsky (1994).

6.10 Conventional and Contemporary Approaches of Interspecific Gene Transfer

The first lentil embryo rescue protocol published (Cohen et al. 1984) allowed to recover interspecific hybrids between the cultivated lentil and L. ervoides and L. nigricans. This method consisted of a two-step culture procedure to recover interspecific hybrids: first, ovule-embryos were collected 2 weeks after pollination and cultured on MS medium supplemented with 10 % sucrose and 0.2 mg/l IAA + 0.5 mg/l ZEA + 0.5 mg/l GA3 and 0.9 % agar. After 1 week in culture, embryos were excised from the ovular integuments and placed on a second MS medium, supplemented with 3 % sucrose and 0.2 mg/l IAA + 0.2 mg/l ZEA and 0.9 % agar. Later, using the same embryo culture technique, Ladizinsky et al. (1985) again obtained hybrids of the cultivated lentil with L. ervoides.

On the other hand, viable interspecific hybrids were obtained without the use of embryo rescue between lentil cultivars of New Zealand crossed with L. odemensis, L. ervoides, and L. nigricans, applying GA3 after pollination (Ahmad et al. 1995). Fratini and Ruiz (2006) in a first hybridization experiment carried out 630 pollinations with L. odemensis and L. nigricans, performing no embryo rescue, and no hybrid seed was recovered; in a second hybridization experiment, 218 pollinations were carried out with L. odemensis, L. nigricans, and L. ervoides, applying GA3 after fertilization, increasing the percentage of ovules obtained; however, ovules started to turn brown and pods initiated to dry out 18–21 days after pollination (DAP). When ovules of both hybridization experiments were dissected, either no embryos were located or the embryos had dried out. This outcome is in disagreement with the results of Ahmad et al. (1995); nonetheless, it has been described in the literature that certain cultivars and accessions of a species are better combiners to recover interspecific hybrids than other cross combinations (Pitarelli and Stavely 1975; Harlan and de Wet 1977; Williams 1987; Pickersgill 1993). Based on the results of interspecific crosses of Spanish lentil cultivars, Fratini and Ruiz (2006) concluded that embryo rescue is necessary to obtain lentil interspecific hybrids, as had been previously indicated by other authors (Cohen et al. 1984; Ladizinsky et al. 1985; Abbo and Ladizinsky 1991, 1994).

Gupta and Sharma (2005) used a two-step in vitro method of embryo rescue analogous to Cohen et al. (1984) to overcome the postfertilization interspecific barrier. The hybrid embryos developed multiple shoots which subsequently did not allow induction of root differentiation to recover whole plants. Fratini and Ruiz (2006) applying the rescue protocol of Cohen et al. (1984) observed similar multiple shoot results recovering a single hybrid with L. odemensis out of 1,350 hybridizations. In view of the above, Fratini and Ruiz (2006) developed an interspecific embryo rescue protocol suited for Spanish lentil cultivars consisting of four different stages: (1) in ovule-embryo culture, (2) embryo culture, (3) plantlet development, and finally, (4) the gradual habituation to ex vitro conditions of the recovered interspecific hybrid plantlets (see also Fratini and Ruiz 2011). Ovule-embryos of 18 DAP were cultured on MS salts with 1 % sucrose and 1 μM IAA + 0.8 μM KN; after 2 weeks, embryos were released from the ovular integuments and cultured on the same medium for another 2 weeks in upright position. Afterward, the embryos were transferred to test tubes containing the same medium, and 1 month later interspecific plantlets were obtained (Fratini and Ruiz 2006). Using this rescue protocol, the efficiency of interspecific hybrids with L. odemensis, L. nigricans, and L. ervoides recovered per number of ovules cultured ranged between 50 and 100 % (Fratini and Ruiz 2006). Authors noted that interspecific hybrids were always obtained in the spring season with increasing daylight hours and an average temperature of 20 ºC, albeit an approximate equal number of crosses were carried out in the fall under nearly the same environmental conditions.

The main differences between the embryo rescue protocol of Cohen et al. (1984) and that developed by Fratini and Ruiz (2006) comprise the number of culture media used and the amount of carbohydrate added to the media. Whereas in the case of Cohen et al. (1984) two different media were used for ovule and embryo culture supplemented respectively with 10 and 3 % of sucrose, the single medium used by Fratini and Ruiz (2006) for ovule, embryo, and plantlet culture contained only 1 % sucrose. Embryo rescue protocols in other legumes have also used sucrose concentrations lower than 3 %, such as 2.5 % for Medicago (McCoy and Smith 1986), 2 % for Vicia faba and Vicia narbonensis (Lazaridou et al. 1993), and 2 % for Phaseolus (Mejia-Jimenez et al. 1994). With respect to phytohormones, the embryo rescue procedure of Cohen et al. (1984) compared to that of Fratini and Ruiz (2006) contains approximately equivalent concentrations of auxins and cytokinins, even though ZEA (Cohen et al. 1984) and KN (Fratini and Ruiz 2006) were respectively used.

The most recent embryo rescue technique based on Cohen et al. (1984) for interspecies transfer of anthracnose resistance from L. ervoides to cultivated lentils allowed to recover only one F1 hybrid (Fiala et al. 2009). Some of the interspecific hybrids recovered by Fratini and Ruiz (2006) were cloned in vitro in order to increase the number of segregating descendant seeds obtained to be used for breeding and genetic analysis purposes (Fratini and Ruiz 2008). Contemporary approaches to interspecific gene transfer have been reviewed by Davis et al. (2007) and Fratini and Ruiz (2011).

6.11 Molecular Markers, Genome Mapping, and Genomics as an Adjunct to Breeding

Lentil is a relatively minor crop compared to common bean, pea, and chickpea; as a result, genomic information regarding lentil is still limited by a relatively large genome size together with scarce information available on gene sequences, constituting a major obstacle to undertake genomic studies in lentil. So far for lentil, there exist no descriptions of a bacterial artificial chromosome (BAC) library, BAC-end sequence, or a physical map (Ford et al. 2007, 2009; Varshney et al. 2009; Pérez de la Vega et al. 2012). In comparison to other crop species, the number of Lens data indexed continues to be scarce, although this situation is rapidly changing, in particular for nucleotide-expressed sequence tags (ESTs) (Kaur et al. 2011; Bhadauria et al. 2013; Sharpe et al. 2013; Verma et al. 2013). Reports regarding the characterization of the nucleotide binding site-leucine-rich repeat (NBS-LRR) disease resistance genes of several legume species, including lentils, and on transcriptome sequencing of cool season food legumes using 454 GS-FLX Titanium technology (Penmetsa et al. 2010; Kaur et al. 2010) also reveal a change of tendency.

The first report on the use of the microarray technology to study gene expression in lentil was published by Mustafa et al. (2009). The highly resistant (ILL7537) and highly susceptible (ILL6002) lentil germplasm accessions were inoculated with Ascochyta lentis; a total of 90 genes were differentially expressed in ILL7537 compared to 95 genes differentially expressed in ILL6002. The expression profiles of the two accessions showed substantial difference in the type of genes and the times of expression in response to the pathogen, whereas the resistant variety showed an early upregulation of PR4 and ten proteins plus other defense-related genes; the susceptible genotype showed an early downregulation of defense-related genes. Real-time PCR (RT-PCR) was used to verify the microarray expression ratios. The resistant and susceptible lentil accessions differed not only in the type of genes expressed but also in the time and level of expression in response to A. lentis inoculation.

Lentil proteomics is also still underdeveloped. Two-dimensional electrophoretic maps of proteins from mature lentil seed have been described (Scippa et al. 2008, 2010). Among the hundreds of protein species, 122 were further identified by MALDI-TOF PMF and/or LC-ESI-LIT-MS/MS. Map comparisons revealed that 103 protein spots were differentially expressed within and between populations (Scippa et al. 2010).

Comparative mapping was able to show extensive macrosynteny between Medicago truncatula and L. culinaris (Phan et al. 2007). These authors generated the first gene-based genetic linkage map of lentil (L. culinaris ssp. culinaris) using an F5 population derived from a cross between cultivar Digger (ILL5722) and Northfield (ILL5588) composed of 79 ITAP and 18 genomic SSR markers. Evidence for the macrosynteny between lentil and M. truncatula was found as 66 out of 71 gene-based markers, which had previously been assigned to the M. truncatula genetic and physical maps, were found in regions syntenic between the L. c. ssp. culinaris and M. truncatula genomes. However, moderate evidence of chromosomal rearrangements was found, which may account for the difference in chromosome numbers between these two legume species. Comparative mapping showed a direct and simple relationship between the M. truncatula and L. culinaris ssp. culinaris chromosomes, with complete homology evident between LGs. Thus, lentil LG 1, 2, 3, 4, 5, 6, and 7 were proposed to be syntenic to M. truncatula LG 4 + 7, 8, 1 + 6, 2, 5, 3, and 3, respectively; next, considering lentil or pea LGs syntenic to the same M. truncatula or M. sativa LGs, lentil LG 1, 2, 3, 4, 5, 6, and 7 were proposed to be syntenic to pea LG VII + V, IV, II + III + VI, III + IV, I, III, and III, respectively (see Pérez de la Vega et al. 2012).

Eighteen common SSR markers were used by Phan et al. (2007) to connect their map to another comprehensive map of lentil (Hamwieh et al. 2005), providing a syntenic context for four important domestication traits. Ellwood et al. (2008) used 151 markers out of 796 ITAP markers screened to construct a comparative genetic map between L. culinaris and Vicia faba. A simple and direct macrosyntenic relationship between faba bean and M. truncatula was evident, while faba bean and lentil shared a common rearrangement relative to M. truncatula. This study also provided a preliminary indication for a finer scale macrosynteny between M. truncatula, lentil, and faba bean. Markers originally designed from genes on the same M. truncatula bacterial artificial chromosomes (BACs) were found to be grouped together in corresponding syntenic areas in lentil and faba bean. Previous research had already demonstrated marker and/or gene order conservation between species closely related to lentil such as Pisum sativum and Medicago sativa (Kalo et al. 2004), P. sativum and M. truncatula (Aubert et al. 2006), and P. sativum, M. truncatula, M. sativa, Vigna radiata, Glycine max, and Phaseolus vulgaris (Choi et al. 2004).

Tar’an et al. (2003) using 156 RILs were able to pyramid into lentil breeding lines two genes for resistance to Ascochyta blight caused by A. lentis, together with the gene for resistance to anthracnose caused by Colletotrichum truncatum. Genetic analysis of the disease reactions demonstrated that the observed segregation ratios of resistant versus susceptible fitted a two gene model for resistance to Ascochyta blight and a single gene model for resistance to anthracnose. Using molecular markers linked to the two genes conferring resistance to A. lentis (UBC 2271290) to ral1 and (RB18680) to AbR1, in conjunction with the major gene for resistance to anthracnose (OPO61250), they found 11 RILs that retained all of the three resistance genes providing a practical evidence of progress toward pyramiding resistance in a lentil population. Table 6.1 displays other molecular markers linked to agronomic traits of interest useful for marker-assisted selection; most recently a polymorphic microsatellite in the first intron of the TFL1a gene involved in the transition from vegetative to flowering stages was detected (de la Puente et al. 2013).

Table 6.1 Molecular markers closely linked to desirable lentil breeding characteristics for use in marker-assisted selection
Full size table

Detailed information regarding molecular markers, genome mapping, and genomics as an adjunct to lentil breeding can be found in Ford et al. (2007, 2009), and more recently Pérez de la Vega et al. (2012). Lastly, the development of a deep and diverse transcriptome resource for lentil using next-generation sequencing technology allowed to generate data in multiple-cultivated (L. culinaris) and wild (L. ervoides) genotypes, which together with the use of a bioinformatics workflow enabled for the identification of a large collection of SNPs and SSR markers for the subsequent development of a genotyping platform that was used to establish the first comprehensive genetic map of the L. culinaris genome, comprising seven linkage groups corresponding to the number of chromosome pairs of lentil (Sharpe et al. 2013). Extensive collinearity with M. truncatula was evident on the basis of sequence homology between mapped markers and the model genome, and large translocations and inversions relative to M. truncatula were identified. Synonymous changes that were only observed between L. culinaris and L. ervoides enabled to estimate the time of divergence to 677,000 years ago, whereas synonymous changes observed between L. culinaris and M. truncatula led to estimate divergence to 38 million years ago (Sharpe et al. 2013).

6.12 Conclusions

The exact taxonomy of the genus Lens still needs to be established, especially with regard to L. tomentosus and L. lamottei; further hybridization experiments need to be undertaken to establish if these wild species belong to the secondary or tertiary gene pool of Lens. With regard to wild germplasm, further recollections are needed as, for instance, only four L. tomentosus and 39 L. odemensis accessions are maintained in gene banks worldwide. More variable and valuable alleles of cultivated landraces must be conserved in ex situ collections for genotyping and phenotyping to discover new and useful variants. One example is the recent finding of a high genetic variability among accessions from Azerbaijan which suggests that this gene pool needs to be increased through additional accessions. Another example, Chinese land races are not represented in the ICARDA nor USDA core collections; however, evidence from other Chinese pulse landraces strongly indicates that collected Chinese landraces of lentil, from west and central China, will constitute a very interesting germplasm for trait and allelic variation assessment.

Wild and cultivated accessions need to be further evaluated with regard to biotic and abiotic stresses, placing emphasis on developing more accurate screening methods. In addition, further inheritance studies of resistance/tolerance to biotic and abiotic stresses are required for a better understanding of the respective genetic systems controlling these responses. This knowledge will be useful to design adequate breeding programs based on regional requirements. There is a need to include more morphological and molecular markers and to develop a comprehensive consensus genetic linkage map in lentil, allowing for molecular tagging of resistance genes against biotic and abiotic stresses in order to exploit them in breeding with increased selection efficiency. However, many morphological markers so far considered of monogenic inheritance need to be reassessed as they most likely better fit an oligogenic inheritance.

The development of fine genetic maps that include direct gene markers (functional markers) is expected to revolutionize the use of lentil genetic resources. Microsatellite markers have been developed and deployed to characterize composite and core collections at ICARDA; 109 accessions of lentil including culinaris (57), orientalis (30), tomentosus (4), and odemensis (18) were genotyped using 14 microsatellite (SSR) markers; the study of Hamwieh et al. (2009) revealed that the wild accessions were rich in alleles (151 alleles) compared to cultigens (114 alleles). New molecular tools will increase the speed and precision of introgression of these newly identified alleles from both the adapted and wild lentil species and subspecies into the advanced breeding populations. For example, a lentil pyrosequencing and SNP discovery project has recently been concluded at the University of Saskatchewan (Sharpe et al. 2013). The successful completion of this project has led to a dense linkage map of SSR and SNP in seven linkage groups constituting the seven chromosome pairs of lentil (Sharpe et al. 2013), thus allowing for a future great reduction of gene/QTL discovery timelines. The high throughput genotyping conducted by the CGIAR Challenge Program and other national programs will characterize the world’s ex situ germplasm resources, leading to an understanding of the population structure from a statistical genetics perspective (Furman et al. 2009). This information combined with genome sequencing, SNP variation studies (haplotype mapping), and detailed phenotyping of the lentil germplasm will lead to successful genome-wide association studies. The understanding of allele values from the adapted and wild lentil gene pool will dramatically increase both the efficiency and worth of germplasm utilization in lentil breeding programs (Coyne et al. 2011).