J Sci Food Agric 1998, 78, 220È224 Variation in Seed Protein Content in the Annual Wild Cicer Species Bruno Ocampo,* Larry D Robertson and Kharag B Singh1 International Center for Agricultural Research in the Dry Areas (ICARDA), PO Box 5466, Aleppo, Syria (Received 14 March 1997 ; revised version received 9 January 1998 ; accepted 4 February 1998) Abstract : In a search for higher seed protein content than in cultivated chickpea (Cicer arietinum L), the authors evaluated a collection of 228 accessions from the International Center for Agricultural Research in the Dry Areas (ICARDA), representing the eight annual wild species of the genus Cicer, along with 20 cultivated chickpea check lines. Variation in seed protein content ranged from 168 g kg~1 in C cuneatum Hochst ex Rich to 268 g kg~1 in C pinnatiÐdum Jaub & Spach, with an average seed protein content of 207 g kg~1 over the eight wild species. C yamashitae Kitamura had the highest mean (217 g kg~1), while C echinospermum PH Davis had the lowest (192 g kg~1). The mean protein content of the cultivated checks was 188 g kg~1. SigniÐcant variation was present in C judaicum Boiss, C pinnatiÐdum and C reticulatum Ladiz. C pinnatiÐdum had the highest number of accessions with high protein content. Overall, protein content showed negative association with harvest index, but little association with leaf area, days to maturity and canopy width. Although the variation for seed protein content of the collection showed accessions with higher protein content than with the cultivated checks, it falls within the range reported for cultivated chickpea. It is expected that agronomically superior selections from interspeciÐc hybrids involving C arietinum and its annual relatives should not be inferior to the cultigen in protein content. Moreover, as usually occurs in distant hybridisation, unexpected epistatic e†ects could produce positive transgressive segregants, as has already been reported in Cicer. ( 1998 Society of Chemical Industry J Sci Food Agric 78, 220È224 (1998) Key words : breeding ; chickpea ; Cicer arietinum ; germplasm ; legumes ; pulses ; quality ; wild species INTRODUCTION its protein quality is superior to most food legumes (Bressani 1973 ; Meiners and Litzenberger 1973 ; Chavan et al 1986). Therefore, chickpea plays an important role in complementing the biological value of staples based chieÑy on cereals, roots, tubers and plantain crops. There is strong evidence of the possibility of substantial improvement of protein content in food legumes (Moreno 1983). The genus Cicer comprises 43 species, 33 perennials and 9 annuals, including the cultivated chickpea and 1 unspeciÐed (van der Maesen 1987). The International Center for Agricultural Research in the Dry Areas (ICARDA, Aleppo, Syria) maintains the worldÏs largest collection of wild Cicer species (270 accessions) and of the kabuli-type chickpea (9870 accessions). The phenological, agronomic and stress reaction features of the Protein quantity is one of the most serious aspects of food deÐcit in developing countries. Chickpea is a valuable source of protein for diets in developing countries, where about 95% of the crop is produced (FAOSTAT 1995). Seed protein content is the most important quality factor for chickpea breeding programmes. Chickpea seed protein content is low compared with other food legumes (Meiners and Litzenberger 1973) ; however, * Joint publication from ICARDA and ICRISAT (International Crops Research Institute for the Semi-Arid Tropics), Patancheru PO, AP 502, 324, India. 1 Present address : KB Singh, 81 Chander Nagar, Janakpuri, A Block, New Delhi 110058, India. * To whom correspondence should be addressed. 220 ( 1998 Society of Chemical Industry. J Sci Food Agric 0022È5142/98/$17.50. Printed in Great Britain Seed protein content in annual wild Cicer species 221 TABLE 1 Annual wild Cicer species germplasm used in this study and their country of origin (ICARDA Genetic Resources Unit) Origin Afghanistan Ethiopia Jordan Lebanon Palestine Syria Turkey Total Cicer speciesa ari ret ech jud pin bij cho yam cun 4 2 3 1 2 7 1 È È È È È È 51 È È È È È È 11 2 2 8 17 9 22 7 È 1 È 1 5 8 34 È È È È È 4 33 5 È È È È È È 3 È È È È È È È 4 È È È È È 20 51 11 67 49 37 5 3 4 a ari, arietinum (check) ; ret, reticulatum ; ech, echinospermum ; jud, judaicum ; pin, pinnatiÐdum ; bij, bijugum ; cho, chorassanicum ; yam, yamashitae ; cun, cuneatum. annual wild Cicer species collection of ICARDA have been catalogued (Robertson et al 1995). Wild relatives of chickpea are reported to be promising sources of genes for resistance to the major biotic and abiotic factors which a†ect yield stability in chickpea (Singh et al 1989 ; Weigand and Tahhan 1990 ; Singh and Reddy 1993 ; Kaiser et al 1994 ; Singh and Weigand 1994 ; Di Vito et al 1996). There are many examples of crops improved by means of wide hybridisation (Prescott-Allen and Prescott-Allen 1988). For almost a decade, ICARDA has embarked on a wideranging enterprise of exploiting the annual wild Cicer species for the improvement of the cultigen (Ocampo 1995 ; Singh and Ocampo 1997). The objective of this study was to estimate the variability of seed protein content of a collection of annual wild Cicer species. MATERIALS AND METHODS The material of this study comprised 228 accessions of eight annual wild Cicer species from the ICARDA genebank : C bijugum KH Rech, C chorassanicum (Bunge) M Popov, C cuneatum Hochst ex Rich, C echinospermum PH Davis, C judaicum Boiss, C pinnatiÐdum Jaub & Spach, C reticulatum Ladiz and C yamashitae Kitamura and 20 randomly selected kabuli-chickpea (C arietinum TABLE 2 Statistical distribution of seed protein content (g kg~1, Kjeldahl procedure) in the annual Cicer species collection of ICARDA State Meanc SE (^) SD Range Skewness Kurtosis Cases Cicer speciesa Totalb ari ret ech jud pin bij cho yam cun 188d 0É20 0É88 171È203 0É08 [0É62 20 201bc 0É10 0É69 188È205 [0É10 [0É66 51 192cd 0É25 0É84 174È210 [0É02 3É26 11 210ab 0É13 1É05 190È231 0É09 [0É85 65 210ab 0É21 1É42 194È268 1É92 5É05 47 212ab 0É11 0É66 200È226 0É53 [0É56 35 194c 0É70 0É99 187È201 NEd NE 2 217a 2É19 3É79 174È247 [1É35 NE 3 201bc 1É23 2É46 168È220 [1É81 3É27 4 207 0É08 1É22 168È268 0É710 3É132 217 146 215 185 232 226 215 251 Accession numbers with highest seed protein content within each species 2639 182 179 193 226 178 2758 229 239 173 251 80 6455 233 180 206 221 83 a ari, arietinum (check) ; ret, reticulatum ; ech, echinospermum ; jud, judaicum ; pin, pinnatiÐdum ; bij, bijugum ; cho, chorassanicum ; yam, yamashitae ; cun, cuneatum. b calculated considering only the wild Cicer species. c Means followed by same letters are not signiÐcantly di†erent at P \ 0É05. d Not estimable. B Ocampo, L D Robertson, K B Singh 222 TABLE 3 Analysis of variance for seed protein content (g kg~1) of Cicer species Source of variation df Mean square F-value Accessions within species C arietinum C bijugum C cuneatum C echinospermum C judaicum C pinnatiÐdum C reticulatum C yamashitae 19 34 3 10 64 46 50 2 1É563 0É865 10É734 1É405 2É234 4É142 1É572 8É725 1É92 1É55 6É09 1É10 3É00** 4É96** 3É55** 5É39 7 2É075 5É48* Species * SigniÐcant at P \ 0É05 ; ** signiÐcant at P \ 0É01. L) accessions representing the geographical distribution of the wild Cicer collection. Table 1 shows the countries of origin of the accessions evaluated in this study, along with the number of accessions per species. Seed samples were obtained from plants grown during 1993È94 at Tel Hadya (36¡35@E, 36¡51@N and 284 m asl), Syria, the main research station of ICARDA. Morphological, phenological, agronomic and seed details and stress reactions of these 248 accessions are given in Robertson et al (1995). This site has a semi-arid continental Mediterranean climate. Rainfall during the 1993È94 season was 373 mm. Tel Hadya soils are Ðne clay montmorillonitic, thermic calcixerollic xerochrept with a pH of about 8É0. They are low in organic matter (about 1É0%) and in available phosphorus (about 5 lg g~1 at 0È20 cm, and about 3 lg g~1 at 20È40 cm). The 248 accessions were grown in a randomized complete block layout with two blocks. Each accession was sown in a three-row plot 2 m long with spacing of 45 cm between and 15 cm within rows on 29 November 1993. Forty seeds of each accession were sown in each plot. The plots were fertilised with 50 kg P O ha~1. 2 5 Hand-weeding was done three times. The crop was protected from ascochyta blight (Ascochyta rabiei (Pass) Lab) by a periodic spraying of the fungicide chlorothalonil (tetrachloroisopenthalonitrite) at the rate of 0É8 kg a.i. ha~1. Seed protein content was determined according to the macro-Kjeldahl procedure AACC 46-12 (AACC 1983) ; the N to protein factor used was 6É25. The g kg~1 data were based on dry matter using seeds from three randomly-selected plants of the central row of the plots. Analyses of variance were carried out using MSTAT-C (1988) version 2.10. Simple correlation coefÐcients (Pearson coefficients) and frequency distributions of selections were obtained using SPSS for Windows, Release 6.0 (SPSS 1993). Derivatives accessions, previously established based on morphological Fig 1. Frequency distributions of derivatives (y axis) for seed protein content (g kg~1) (x axis). The Ðrst number below the species name refers to the International Legume Wild Cicer accession number of the original accession, while the following number in parenthesis refers to the collection number (number or pedigree identiÐcation code used by the donor). The black vertical line within each chart indicates the original accession seed protein content. grounds, were analysed to determine variation within populations (accessions) for protein content. RESULTS The present evaluation showed wide variation for seed protein content in the annual wild Cicer species (Table 2). Distribution patterns varied from normal to strongly skewed with high kurtosis. Protein content ranged from 168 g kg~1 in C cuneatum to 268 g kg~1 in C pinnatiÐdum. Mean seed protein content of the annual wild Cicer collection was 207 g kg~1. SigniÐcant mean differences (P \ 0É05, Table 3) were recorded among the species, where C yamashitae, C bijugum, C judaicum and Seed protein content in annual wild Cicer species 223 C pinnatiÐdum showed the highest means and C echinospermum the lowest. The mean seed protein content of the cultigen set used in this study was signiÐcantly lower than those of the wild Cicer species, except for C echinospermum (Table 2). SigniÐcant variation was present among accessions of C judaicum, C pinnatiÐdum and C reticulatum (Table 3). Protein content among derivatives of single accessions was variable (Fig 1). The variation was less pronounced in C bijugum compared with C echinospermum, C judaicum, C pinnatiÐdum and C reticulatum. Most derivatives of C judaicum accessions showed higher protein content compared with the original accessions. Linear phenotypic correlations between seed protein content and morphological, phenological and agronomic characters di†ered among species (Table 4). The highest number of signiÐcant associations were recorded in C pinnatiÐdum and the least in C judaicum. Overall, the highest correlations were between seed protein content and harvest index (negative). Less signiÐcant associations were found with 100-seed weight (negative), days to Ñowering (positive), seed yield (negative) and number of seeds per pod (negative and positive correlations) (Table 4). DISCUSSION The variability for seed protein content of the ICARDA collection of annual wild Cicer species was large. The present results di†er substantially from those of Singh U and Pundir (1991), who reported the protein content to range from 265 g kg~1 in C yamashitae to 327 g kg~1 in C bijugum. These di†erences might have been caused by di†erences in the environment and in the methodology used (decorticated samples) and by the small number of accessions used by Singh and Pundir (21 compared with 248 in the present work), who scanned only a small proportion of the available annual wild Cicer species genetic diversity. The data reported here suggest that the prospects for upgrading the nutritional value (protein content) of chickpea via introgression of wild Cicer species genes into the domesticated are remote, since the seed protein content variability of the wild accessions falls within the range reported for cultivated chickpea (Chavan et al 1986 ; Hulse 1994). A more realistic approach to chickpea improvement using these species in distant hybridisation may be to ensure that the seed protein content of hybrid selections does not fall below that of existing cultivars. This should be easy to do, as most wild species (except C pinnatiÐdum) had mostly non-signiÐcant correlations between protein content and most agronomic traits (Table 4). Furthermore, as usually occurs in distant hybridisation, unexpected epistatic e†ects could produce positive transgressive segregants, as has already been reported in Cicer (Ocampo 1995) and other crops for quantitative traits (Cox et al 1984 ; Frey et al 1984 ; Carpenter and Fehr 1986). Variation for seed protein content among derivatives from a single accession underline the necessity for careful evaluation within accessions discriminated on morphological traits. The variation between species in phenotypic correlations between seed protein content and some morphological, phenological and agronomic descriptors are an TABLE 4 Linear correlations (Pearson coefficients) between seed protein and some morphological, phenological and agronomic descriptors Descriptor Leaf area Days to 50% Ñowering Days to maturity Plant height Canopy width Pods/plant Seeds/pod Biological yield Seed yield Harvest index Seeds/plant 100-seed weight Cicer speciesa ari (20) ret (50) ech (11) jud (65) pin (47) bij (35) 0É104 [0É066 0É128 0É253 [0É152 0É186 0É168 [0É249 [0É383 [0É536* 0É296 [0É450* 0É027 0É088 [0É017 [0É291* [0É106 0É019 [0É117 0É075 [0É094 [0É443** [0É018 [0É120 [0É038 0É654* 0É444 [0É088 [0É142 [0É433 0É237 [0É482 [0É678** [0É756** [0É487 [0É334 0É084 0É048 [0É170 0É097 0É024 [0É102 0É268* [0É038 [0É049 0É014 [0É051 0É028 [0É149 0É432** 0É145 [0É245 [0É254 [0É381** [0É480** [0É324** [0É386** [0É562** [0É410** [0É409** 0É133 0É340* 0É109 0É002 0É171 0É059 [0É138** [0É026 [0É218 [0É490** 0É108 [0É624** a ari, arietinum (check) ; ret, reticulatum ; ech, echinospermum ; jud, judaicum ; pin, pinnatiÐdum ; bij, bijugum. Number of accessions in parenthesis. * SigniÐcant at P\0É05 ; ** signiÐcant at P\0É05. B Ocampo, L D Robertson, K B Singh 224 indication of genetic di†erences in biological pathways among Cicer species. Nevertheless, except for C judaicum, all share a signiÐcant negative association between protein content and harvest index. This association, along with those of smaller magnitude between protein content and seed weight or seed yield (both negative), suggests that selection for desirable protein content would also allow the selection of genotypes with high biomass, low seed yield and low seed weight. This is reported in most legume literature (PAG 1973). It is acknowledged that evaluation of wild genetic resources per se is only a preliminary step in the exploitation of wild relatives for the genetic improvement of crops, because comprehensive estimation of the breeding value of wild accessions is possible only after their introgression into cultivated genotypes. Genetic reshuffling of diverse taxa, along with breeding procedures which enable breakage of undesirable linkages, may produce agronomically suitable genotypes not expected from parental performances. Introgression of C echinospermum and C reticulatum genes into chickpea, with the attainment of positive transgressive selections for quantitative characters such as seed yield, biomass, plant height and seed weight (Singh and Ocampo 1997), supports this conclusion for the genus Cicer. ACKNOWLEDGEMENT The authors thank Mr Hani Nakkoul for providing technical research assistance. REFERENCES AACC 1983 Approved Methods. American Association of Cereal Chemists, St Paul, MN, USA. Bressani R 1973 Legumes in human diets and how they might be improved. In : Nutritional Improvement of Food L egumes by Breeding. PAG, Rome, Italy pp 15È42. Carpenter J A, Fehr W R 1986 Genetic variability for desirable agronomic traits in populations containing glycine soja germplasm. Crop Sci 26 681È686. Chavan J K, Kadam S S, Salunkhe D K 1986 Biochemistry and technology of chickpea (Cicer arietinum L) seeds. Crit Rev Food Sci Nutr 25 107È157. Cox T S, House L R, Frey K J 1984 Potential of wild germplasm for increasing yield of grain sorghum. Euphytica 33 673È684. Di Vito M, Singh K B, Greco N, Saxena M C 1996 Sources of resistance to cyst nematode in cultivated and wild Cicer species. Genet Res Crop Evol 43 103È107. FAOSTAT 1995 Faostat.PC/FAO Yearbooks on diskettes. FAO, Rome, Italy. Frey K J, Cox T S, Rodgers D M, Bramel-Cox P 1984 Increasing cereal yields with genes from wild and weedy species. In : Genetics : New Frontier, (XV International Congress of Genetics, Vol IV Applied Genetics), ed Chopra V L, Joshi B C, Sharma R P & Bansal H C. Bowker, Epping, UK, pp 51È68. Hulse J H 1994 Nature, composition, and utilization of food legumes. In : Expanding the Production and Use of Cool Season Food L egumes, ed Muehlbauer F J & Kaiser W J. Kluwer Academic, Dordrecht, The Netherlands, pp 77È97. Kaiser W J, Alcala-Jimenez A R, Hervas-Vargas A, TraperoCasas J L, Jimenez-D• az R M 1994 Screening of wild Cicer species for resistance to races 0 and 5 of Fusarium oxysporum f sp ciceris. Plant Dis 78 962È967. Meiners J P, Litzenberger S C 1973 Breeding for nutritional improvement. In : Nutritional Improvement of Food L egumes by Breeding. PAG, Rome, Italy, pp 131È141. Moreno M T 1983 Las leguminosas de grano : una vision de conjunto. In : L eguminosas de Grano, ed Cubero J I & Moreno M T. Ediciones Mundi-Prensa, Madrid, Spain, pp 15È34. MSTAT-C 1988 Michigan State University, MSTAT/Crop and Soil Sciences A87 Plant and Soil Sciences, East Lansing, MI 48824, USA. Ocampo B 1995 Estudio de las especies anuales del genero Cicer L : aplicacion a la mejora genetica de Cicer arietinum L. PhD thesis, University of Cordoba, Cordoba, Spain. PAG 1973 Nutritional Improvement of Food L egumes by Breeding. PAG, Rome, Italy. Prescott-Allen R, Prescott-Allen C 1988 Genes from the W ild. Earthscan, London, UK. Robertson L D, Singh K B, Ocampo B 1995 A Catalog of Annul Cicer Species. ICARDA, Aleppo, Syria. Singh K B, Ocampo B 1997. Exploitation of wild Cicer species for the yield improvement of chickpea. T heor Appl Genet 95 418È423. Singh K B, Reddy M V 1993 Sources of resistance to ascochyta blight in wild Cicer species. Neth J Plant Pathol 99 163È167. Singh K B, Weigand S 1994 IdentiÐcation of resistant sources in Cicer species to L iriomyza cicerina. Genet Res Crop Evol 41 75È79. Singh K B, Malhotra R S, Saxena M C 1989 Chickpea evaluation for cold tolerance under Ðeld conditions. Crop Sci 29 282È285. Singh U, Pundir R P S 1991 Amino acid composition and protein content of chickpea and its wild relatives. Int Chickpea Newslett 25 19È20. SPSS 1993 SPSS for Windows : Base UserÏs Guide, Release 6.0. SPSS, Chicago, IL, USA. van der Maesen L J G 1987 Origin, history and taxonomy of chickpea. In : T he Chickpea, ed Saxena M C & Singh K B. CAB International, Wallingford, UK, pp 11È34. Weigand S, Tahhan O 1990 Chickpea insect pests in the Mediterranean zones and new approaches to their management. In : Chickpea in the Nineties, ed van Rheenen H A & Saxena M C. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, India, pp 169È 175.