Genetic control of seed weight and calcium concentration in chickpea seed

Plant Breeding 119, 427Ð431 (2000) # 2000 Blackwell Wissenschafts-Verlag, Berlin ISSN 0179-9541 Genetic control of seed weight and calcium concentration in chickpea seed S. ABBO 1,M.A. GRUSAK 2,T.TZUK 3 and R. REIFEN 3 1 Department of Field Crops, Vegetables and Genetics, The Hebrew University of Jerusalem, Rehovot 76100, Israel; 2 USDA- ARS, Baylor College of Medicine, 1100 Bates Street, Houston, TX 77030±2600, USA; 3 School of Nutritional Sciences, The Hebrew University of Jerusalem, Rehovot 76100, Israel With 2 gures and 2 tables Received December 25, 1999/Accepted April 8, 2000 Communicated by A. Ashri Abstract Chickpea, Cicer arietinum L., is a staple protein source in many Asian and Middle Eastern countries. Hence, the mineral content of its seed, especially that of calcium, is of nutritional importance. lcium is transported through plants and to legume pods almost exclusively via the xylem stream, with accretion in developing seeds resulting primarily from di usion of from the adjoining pod wall. Thus, for seeds of di ering surface-to-mass ratios, concentration is expected to correlate inversely with seed weight. The relationship between seed weight and concentration in chickpea seeds was studied using a range of germplasm and derivatives from crosses between types di ering in seed concentration. Among the cultivars tested, low seed mass was associated with high concentration. However, the study of hybrid progeny indicated that seed content was mainly determined by genetic factors other than grain weight genes. This nding may assist in future breeding of high nutritional quality chickpea cultivars. Key words: Cicer arietinum Ð calcium concentration Ð seed weight Chickpea is the third most important grain legume in the world. It is cultivated over a vast geographical range, from south-east Asia, across the Indian subcontinent, the Near East, southern Europe, the Balkans, the Mediterranean basin and East Africa. As a post-colombian introduction, chickpeas are now grown in North America and Latin America (Ladizinsky 1995). Expansion of area under chickpea cultivation has recently taken place in Australia (Siddique and Sykes 1997). In most of the New World growing areas, chickpea was introduced into cereal-based systems as a rotation crop. Global chickpea production is 19 10 9 kg/year, of which only about 5% are traded on the international market (FAO 2000). The limited trade in chickpea grain is a good indication of its economic and nutritional importance in the rural communities in which it is traditionally produced (Ladizinsky 1995). Chickpea seeds are a good source of protein, carbohydrates, fat and minerals (Williams and Singh 1987). Chickpea is a major source of dietary protein in India, where the vast majority of the Hindu population is vegetarian. In recent years, there has been increasing interest in some western countries in chickpea as a health food, as an additive and as a fast-food product. As such, the mineral content of the seed, among other nutritional criteria, is of great importance. lcium () is an essential nutrient for humans, but is quite often limited in diets of low-income sectors and is of particular concern for pre-school children, adolescents, and pregnant and lactating women. Chickpea seeds contain 103±259 mg /0.1 kg dry weight Ð70% of this is in the seed coat (Williams and Singh 1987) Ð and are therefore a potential source of dietary. Adequate nutrition during childhood has important implications for bone growth and development, and is thought to reduce the incidence of osteoporosis in later life (Johnston et al. 1992). lcium is transported through plants, and in particular to the pods, almost exclusively via the xylem (Ho et al. 1987); only trace amounts of are transported in the phloem system (Raven 1977). As in other legume seeds (Mix and Marschner 1976), is thought to move apoplastically to developing chickpea seeds, following di usion through the pod wall to the surface of the testa. In the absence of transport in the phloem stream, smaller seeds (low seed weight, relatively high surface-to-volume ratio) are expected to contain more per unit of dry weight relative to larger seeds (high seed weight, relatively low surface-to volume ratio). Because seed size (weight) is a major price determinant, especially among Kabuli chickpeas, the relationship between concentration and seed size is of great relevance for Kabuli chickpea consumers and for breeders interested in improving the nutritional quality of chickpea seed. In this paper, the relationship between seed weight and concentration in chickpea is analysed, using the cultivars of types Desi and Kabuli, as well as F 2 and F 3 derivatives of Kabuli Desi crosses, to test the existence of factors other than seed-weight genes in determining concentration in chickpea seeds. Materials and Methods Plant materials: The cultivars of chickpea, Cicer arietinum L., selected for this study are listed in Table 1. The list includes Kabuli and Desi types, two lines of the wild progenitor of the crop, Cicer reticulatum Ladiz. and F 3 seeds (seed coat plus embryo), harvested from F 2 plants of the crosses `Hadas' ICC5810, and ICC5810 `Hadas'. In a subsequent season, concentration was determined in the F 4 seed yield of individual F 3 plants. Because 70% of the in chickpea seeds is present in the seed coat, which is a maternal tissue, the data re ect the phenotypic variation of the previous generations. Growth conditions: During the 1996±97 season, three replications of each of the chickpea cultivars and the F 2 populations of the `Hadas' ICC5810 and ICC5810 `Hadas' crosses were grown in an insectproof net-house at the experimental farm of the Faculty of Agriculture, U. S. Copyright Clearance Center Code Statement: 0179±9541/2000/1905±0427 $ 15.00/0

428 ABBO, GRUSAK, TZUK and REIFEN Table 1: Description of the chickpea germplasm used in this study Accession Source Remarks Spharadit Mexico Kabuli type, large seed, ascochyta-sensitive Hadas 123 Israel Kabuli type, high-yielding, ascochyta-tolerant Hadas (95±1) Israel Kabuli type, selection from Hadas 123 A100 Unknown Kabuli type, large seed, early owering Bulgarit Bulgaria Kabuli type, small seed, ascochyta-resistant, late- owering M-2581 Israel Kabuli type, large seed Green Ethiopia Desi type, small green seed ICC11299 ICRISAT Kabuli type, large seed, early owering ICC5810 India Desi type, small seed, ascochyta-sensitive ICC8625 Ethiopia Desi type, small seed, early owering ICC8631 Ethiopia Desi type, small seed, early owering ICC13940 Ethiopia Desi type, small seed, early owering ICC14088 Ethiopia Desi type, small seed, early owering ICC14098 Ethiopia Desi type, small seed, early owering Cr205 Turkey Wild type, small seed, late- owering Cr231 Turkey Wild type, small seed, late- owering Rehovot, Israel (31 55 0 N, 34 50 0 E). The plants were grown in two-row plots along trickle-drip irrigation lines (0.40 m between the rows and 2 m between the irrigation lines), with a spacing of 10 plants/m within the row, on Rhodoxeralf soil. Soil ph at the site ranges between 7.2 and 7.9, and the carbonate content in the soil is around 10%. During the 1997±98 season, the F 3 population of the `Hadas' ICC5810 cross was grown in a replicated eld experiment, including 10 replicates of the two parental lines of six to 10 plants each, on the Kedma farm, about 35 km south of Rehovot on Vertic Haploxeralf soil. The chickpea cultivars analysed in the summers of 1998 and 1999 seasons were grown in a replicated eld experiment in a commercial chickpea eld on the Massuot-Yitzhak farm, about 50 km south of Rehovot, on Vertic Haploxeralf soil. All materials were harvested at full maturity. Soil ph at the Kedma and Massou't-Yitzhak sites ranges between 7.5 and 8.5 and the soil contains about 15% carbonate. The irrigation water at the three sites is from calcareous aquifers and therefore contains considerable levels. Because of the basic soil ph values and the nature of the irrigation water, concentration in the soil solution was not a limiting factor. All net-house-grown materials, including F 2 individuals and cultivars, were hand-harvested, threshed, and analysed on an individual plant basis. Field-grown F 3 individuals were also hand-harvested and threshed on an individual basis, while eld-grown cultivars were combine-harvested. Mean seed weight of all samples was determined after removal of broken and shrivelled seeds, by counting and weighing. For more details on the growth and climatic conditions, see Or et al. (1999). lcium concentration analyses: lcium concentrations of seeds harvested from cultivars and F 2 progeny during the 1996±97 season were determined for each of two subsamples per cultivar or F 2 plant; a subsample consisted of two whole seeds taken randomly from two pods. Seeds were dried at 70 C until a constant weight was attained and dry weights were then determined. Seeds were wet-digested in borosilicate glass tubes and the digestate resuspended as described by Grusak (1994). Resuspended samples were diluted with 0.5% (w/w) lanthanum chloride and concentration (on a dry weight basis) was determined using atomic absorption spectrophotometry (model 2100; Perkin Elmer, Norwalk, CI, USA). The instrument was calibrated using known standards. Values obtained for the two subsamples were averaged to yield a single value for each cultivar or F 2 plant. Seeds harvested from cultivars and F 3 progeny grown during the 1997±98 season were prepared for analysis by microwave-assisted digestion. Analyses were conducted on solutions vs. known standards. was determined in the solutions tested by inductively coupled plasma atomic emission spectrometry (Spectro ame±spectro, ICP- AES, Kleve, Germany). Statistical analyses: Correlation analyses and t-tests were performed using the JMP package (Sall and Lehman 1996). In the following text asterisks *, **, *** indicate levels of signi cance at P ˆ 0.05, P ˆ 0.01 and P ˆ 0.001, respectively. Results Germplasm screening The germplasm tested exhibited considerable variation in both seed weight and concentration (Table 2). In the three growing seasons, the general relationship between mean grain concentration and mean seed weight in the chickpea collection was negative. The correlation coe cient between the cultivar means for the two traits in the 1996±97 season was r ˆ 0.7** using 10 entries, and in the 1997±98 season 0.87*** using 12 cultivars. In summer 1999, only four cultivars were grown with an r ˆ 0.7 ns between seed weight and concentration. Seed weight was a very stable trait across the years with r > 0.92** between the 1997 and 1998, and 1998 and 1999 values. concentration values were less stable, with r ˆ0.42 ns between the 1997 and 1998 values and r ˆ0.88 (P(r) ˆ 0.12) between the 1998 and 1999 values (only two cultivars were grown in both 1997 and in 1999). Segregating progeny The possible impact of genetic factors determining seed concentration (irrespective of seed-size loci) was studied in segregating populations derived from the crosses `Hadas' ICC5810 and ICC5810 `Hadas'. The parental lines of these crosses di er in their seed weights as well as in their seed concentrations (Table 2). No correlation between seed weight and seed concentration was found in the ICC5810 `Hadas' F 2 population (r ˆ 0.17 ns). A negative r-value ( 0.38, P ˆ 0.016) between concentration and seed weight was found in the second population (`Hadas' ICC5810). In order to corroborate the F 2 data, eld-grown F 3 progeny of the `Hadas' ICC5810 were analysed during the 1997±98 season for their seed concentration (the reciprocal population was not tested for a second season owing to budget constraints). The concentrations of the seed harvested from the F 3 progeny ranged between 0.95 and 2.56 (mg /g dry weight) with a mean of 1.73 + 0.06. Mean grain weight values of the F 3 pro-

Genetic control of seed weight and calcium concentration in chickpea seed 429 Table 2: Mean dry seed weight (, mg/seed + SE) and calcium concentration in dry seed (mg /g dry weight + SE) in diverse chickpea cultivars grown in two consecutive seasons in Israel 1 1996±97 Season 1997±98 Season 1998±99 Season Accession Spharadit 680 + 10 0.96 + 0.02 570 + 7 1.4 + 0.025 nd nd Hadas 123 450 + 7 1.6 + 0.06 nd nd nd nd Hadas (95-1) 450 + 5 0.98 + 0.06 440 + 15 1.36 + 0.1 nd nd A100 430 + 10 1.45 + 0.07 490 + 40 1.68 + 0.06 456 + 17 1.07 + 0.05 Bulgarit 350 + 8 1.5 + 0.03 240 + 5 1.92 + 0.04 266 + 7 0.87 + 0.05 M-2581 nd nd 318 + 9 1.4 + 0.1 nd nd Green 165 + 17 1.6 + 0.17 nd nd nd nd ICC11299 nd nd 474 + 9 1.475 + 0.175 534 + 10 0.84 + 0.03 ICC5810 160 + 5 2.06 + 0.11 140 + 6 2.48 + 0.2 nd nd ICC8625 nd nd 140 + 2 2.64 + 0.17 144 + 3 1.48 + 0.08 ICC8631 nd nd 110 + 2 3.08 + 0.07 nd nd ICC13940 nd nd 126 + 2 3.63 + 0.25 nd nd ICC14088 nd nd 128 + 10 3.04 + 0.12 nd nd ICC14098 130 + 11 1.62 + 0.06 150 + 4 3.14 + 0.18 nd nd Cr205 120 + 7 2.5 + 0.10 nd nd nd nd Cr231 140 + 6 2.07 + 0.05 nd nd nd nd 1 nd, Not determined. geny analysed ranged between 140 and 400 mg, with a mean of 246 + 12 mg. No correlation was observed between grain weight and concentration of the seed harvested from the F 3 plants of the `Hadas' ICC5810 population (r ˆ 0.19 ns). In order to obtain an estimate for the heritability of concentration in chickpea seed, the relationship between the concentrations in the F 2 and their sib F 3 from the `Hadas' ICC5810 populations was analysed. Surprisingly, an extremely low correlation of r ˆ0.09 ns was found between the F 2 (F 3 seed yield) concentration values and the respective values in their F 3 progeny (F 4 seed yield). However, an r-value of 0.55*** was obtained between the seed content (mg / seed) values of the two generations similar to the r-value obtained between seed weight in the two generations (r ˆ0.56***; r 2 ˆ 0.3). This means that only 30% of the F 3 phenotypic variation in grain weight and content (per seed) may be accounted for by the respective F 2 values. Discussion The observed variation and the phenotypic correlation between concentration and grain weight among the chickpea cultivars during the three years may suggest a possible association (linkage or pleiotropy) between seed weight (size) and its concentration. Because is transported into the seed mainly through di usion from the pod walls (Mix and Marschner 1976), such a relationship would appear to have a sound biological basis. A similar inverse relationship between grain weight and concentration has been noted for two cultivars of common bean, Phaseolus vulgaris (Moraghan and Grafton 1997). However, careful inspection of the 1996±97 season values in Table 2 shows that, over a narrow range of seed weights, considerable variation in concentration can be achieved. For example, among the small-seeded wild and Desi lines (ranging from 120 to 160 mg dry weight/seed), concentration varied between 1.6 and 2.5 mg /gr dry weight. Similarly, for cultivars with comparable concentrations (e.g. 1.3±1.6 mg /gr dry weights), a vefold range of grain weights (130±680 mg dry weight/seed) was observed. The r-values, of 0.7 to 0.86, suggest that grain weight can explain from 50 to 70% of the phenotypic variation of content. The rest of the observed content variation may be attributed to both physiological and environmental factors, such as di erences in root -uptake dynamics, whole-plant partitioning (Grusak et al. 1996), or the water status and temperature during seed maturation. In addition, genetic factors independent of loci a ecting grain weight may also be involved in determining seed content. Branch and Gaines (1983) found no association between seed size and seed concentration in peanut, Arachis hypogaea, which may be taken as support for the existence of independent genetic factors determining seed concentration per se. The lower stability of seed concentration values relative to the grain weight values suggests that concentration is more prone to genotype±environment interaction. In such a comparison between cultivars, the di erent genetic background of the cultivars may a ect the traits analysed di erently. As a result the phenotypic correlations between the traits studied may not be indicative of linkage or pleiotropy. To minimize the genetic background e ect on the traits studied, segregating progeny from a cross between high seed weight/low and a low seed weight/high types were investigated. The F 2 populations employed here have shown reciprocal di erences in seed weight (Or et al. 1999). Mean seed weight of the ICC5810 `Hadas' populations was 240 mg; this value differed signi cantly (P < 0.001) from the mean seed weight of the `Hadas' ICC5810 population (270 mg). Under the assumption that seed weight (size) is a major determinant of concentration, one would expect that in the population with the high mean seed weight (`Hadas' ICC5810), mean seed concentration would be lower compared with the population with the low mean seed weight (ICC5810 `Hadas'). The frequency distributions of the concentration values in the two populations (F 2 and F 3 of the `Hadas' ICC5810, and F 2 of ICC5810 `Hadas' crosses) are presented in Figs 1 and 2. Surprisingly, the mean concentration of the seed harvested

430 ABBO, GRUSAK, TZUK and REIFEN Fig. 1: Frequency distribution of calcium concentration values in dry seed among the hybrid progeny of the cross `Hadas' ICC5810. a. F 2 generation; b. F 3 generation from the F 2 plants of the `Hadas' ICC5810 population was higher (1.8 + 0.049 mg /g dry weight) than that of the ICC5810 `Hadas' F 2 population (1.5 + 0.045 mg /g dry weight). A t-test indicated that the two means di ered at a probability of 0.0001. The F 2 F 3 correlations of both seed content and seed weight were rather low (1 0.5). These r-values are not a parent o spring correlations but rather `parent sib-o spring' correlations because the determination of content is a destructive measurement. The value of each F 2 individual was determined from its digested F 3 seed sample, while the F 3 values were determined from progeny of their sibs (F 4 seeds harvested from sib F 3 plants). Heterozygosity at seed-mass gene loci in the F 2 (50%) and the F 3 (25%) generations and the fact that the calculated values were not obtained from strict parent o spring correlations may partly account for the relatively low (F 2 F 3 ) correlation of content and seed weight in the material studied. In addition, environmental conditions and non-additive gene action may also a ect concentration and seed mass. In our view, despite the above limitations, the `parent sib-o spring' r-value of concentration and the frequency distribution in the F 2 and F 3 (Figs 1, 2) equate with concentration being a polygenic trait. Because the parental lines involved in the crosses di ered for both seed weight and concentration, the alleles of the two traits have segregated in the F 2 progeny. If pleiotropic or linkage relationships exist between the two traits, then the correlation values are expected to di er from zero. The stronger the pleiotropic e ect of seed weight on seed (or the linkage between pertinent loci), the lower the r-value expected. The low (non-signi cant) correlation values of seed weight and concentration obtained in the two populations analysed suggest that, in our segregating populations, concentration loci are mostly independent of seed weight loci. For breeding, a pleiotropic relationship between seed weight and seed concentration would be an obstacle. The implication is that it would be di cult to recover high--concentration types among progeny of crosses between low-, largeseeded genotypes and high-, small-seeded ones. However, the study of the segregating populations, suggests that most of the seed concentration genes segregate independently of the seed weight genes. This nding indicates the feasibility of breeding high nutritional quality chickpea cultivars to suit both consumer demand for large seeds and the nutritional needs of large population sectors in the Middle East and world-wide. Acknowledgements This project was funded in part by federal funds from the US Department of Agriculture, Agricultural Research Service under Cooperative Agreement number 58±6250±6-001 to M. A. G., and by the Hebrew University Intramural Research Fund Basic Project Awards grant to S. A. The contents of this publication do not necessarily re ect the views or policies of the US Department of Agriculture, nor does the mention of trade names, commercial products, or organizations imply endorsement by the US government. Fig. 2: Frequency distribution of calcium concentration values in dry seed of the F 2 progeny of the cross ICC5810 `Hadas' References Branch, W. D., and T. P. Gaines, 1983: Seed mineral composition of diverse peanut germplasm. Peanut Sci. 10, 5Ð12. FAO, 2000: http://www.fao.org/ Grusak, M. A., 1994: Iron transport to developing ovules of Pisum sativum. I. Seed import characteristics and phloem iron-loading capacity of source regions. Plant Physiol. 104, 649Ð655.

Genetic control of seed weight and calcium concentration in chickpea seed 431 Grusak, M. A., B. W. Stephens, and D. J. Merhaut, 1996: In uence of whole-plant net calcium in ux and partitioning on calcium concentration in snap bean pods. J. Am. Soc. Hort. Sci. 21, 656Ð659. Ho, L. C., R. I. Grange, and A. J. Picken, 1987: An analysis of the accumulation of water and dry matter in tomato fruit. Plant Cell Env. 10, 157Ð162. Johnston Jr., C. C., J. Z. Miller, C. W. Slemenda, T. K. Reister, S. Hui, J. C. Christian, and M. Peacock, 1992: lcium supplementation and increases in bone mineral density in children. New Engl. J. Med. 327, 82Ð87. Ladizinsky, G., 1995: Chickpea. In: J. Smartt, and N. W. Simmonds (eds), The Evolution of Crop Plants. 258Ð261. Longman, London. Mix, G. P., and H. Marschner, 1976: lcium-umlagerung in BohnenfruÈ chten waè hrend des Samenwachstums. Z. P anzenphysiol. 80, 354Ð366. Moraghan, J. T., and K. Grafton, 1997: Accumulation of calcium in bean cultivars di ering in seed size. J. Sci. Food Agric. 74, 251Ð256. Or, E., R. Hovav, and S. Abbo, 1999: A major gene for owering time in chickpea. Crop Sci. 39, 315Ð322. Raven, J. A., 1977: H and 2 in phloem and symplast: relation of relative immobility of the ions to the cytoplasmic nature of the transport paths. New Phytol. 79, 465Ð480. Sall, J., and A. Lehman, 1996: JMP Start Statistics. Duxbury Press, Belmont. Siddique, K. H. M., and J. Sykes, 1997: Pulse production in Australia past, present and future. Aust. J. Exp. Agric. 37, 103Ð111. Williams, P. C., and U. Singh, 1987: The chickpea Ð nutritional quality and evaluation of quality in breeding programs. In: M. C. Saxena, and K. B. Singh (eds), The Chickpea, 329Ð356. CAB Int., Wallingford.