Genetic Diversity in Cultivated Common Bean: I. Allozymes

Previous tudies using phaseolin seed protein as a marker have revealed that cultivated common bean (Phaseolus vulgaris L.) resulted from multiple domestications in Mesoamerica and in Andean South America. Because these studies were based on variation at a single locus, confirmation was sought by analyzing patterns of diversity at nine polymorphic allozyme loci, all unlinked to the phaseolin locus: ribulose bisphosphate carboxylase, shiklmate dehydrogenase, cathodal peroxidase, malic enzyme, malate dehydrogenase (two loci), diaphurase (two loci), and leucine aminopeptidase. A total of 227 landraces representing geographical regions from Mexico to Argentina nd Chile were analyzed for these enzyme systems. A crude homogenate of primary leaf or root tissue (depending on enzymes assayed) from five seedlings of each landrace grown in vermiculite was used for starch gel electrophoresis. A cluster analysis based on Nei’s genetic distance (D) was performed according to the unweighted paired group method of Sneath and Sokal. Our esults confirm the existence of two major groups in cultivated common bean, Mesoamerican vs. Andean American; provide indications of gene flow from wild to cultivated beans; and suggest at least five subgroups within Mesoamerican and four within Andean cultivar groups. S IN CROP EVOLUTION have traditionally attempted to determine the crop’s ancestry based on morphological similarities and the production of viable and fertile hybrids between wild ancestor and cultivated descendant. Recently, the use of molecular markers has helped identification of the actual ancestral populations in maize (Zea mays L.) (Doebley al., 1987) and pea (Pisum sativum L.) (Palmer et al., 1985), and determination of the effect of domestication on genetic diversity in barley (Hordeum vulgare L.) (Clegg et al., 1984; Jana and Pietrzak, 1988) pearl millet (Pennisetum glaucum [L.] R. Br.) (Gepts and Clegg, 1989). The common bean is an annual, diploid (2n = 2x = 22) species that originated in the Americas and consists of wild and cultivated forms. The wild forms are distributed from northern Mexico to northern Argentina (Brticher, 1988; Delgado Salinas et al., 1988). They are morphologically similar to the cultivars and yield viable and fertile progenies when crossed with them (Br0cher, 1988; Gentry, 1969). Because of the extensive distribution of the common bean’s wild ancestor, the exact siting of its domestication has been subject to speculation. Phaseolin diversity data suggest that cultivated common bean arose from multiple domestications along this extended distribution (Gepts and Bliss, 1986; Gepts et al., 1986). In particular, two major domestications appear to have given rise to Mesoamerican and southern Andean cultivars, S.P. Singh, Centro Int. de Agric. Tropical (CIAT), Apartado Aereo 6713, Cali, Colombia; R. Nodari and P. Gepts, Dep. of Agron. and Range Sci., Univ of California, Davis, CA 95616. Received 15 Dec. 1989. *Corresponding author. Published in Crop Sci. 31:19-23 (1991). respectively. A third and minor domestication may have taken place in Colombia or Central America (Gepts and Bliss, 1986; Koenig et al., 1990). Allozyme analyses of wild forms have confirmed the existence of these two major groups and clarified their geographic boundaries: the Mesoamerican forms include wild populations from northern Mexico to Colombia, whereas the Andean types include populations from Peru and Argentina (Koenig and Gepts, 1989b). In this article, we report on allozyme analyses of 227 cultivated landraces representing a geographical distribution extending from Mexico to Argentina and Chile. The objectives were to determine whether cultivated common bean displays a geographic pattern of allozyme diversity similar to its wild ancestors and to identify subgroups within the large groups of Mesoamerican and Andean cultivars. MATERIALS AND METHODS The collection of 227 landraces analyzed in this study was obtained from the Phaseolus world collection at the International Center for Tropical Agriculture (CIAT), Cali, Colombia, and the Western Regional Plant Introduction Station, Pullman, WA. It is representative of major geographical and ecological regions in the area of origin of common bean, which extends from Mexico to Argentina. Among Mesoamerican and Caribbean materials, 66 originated in Mexico, 7 in Guatemala, 5 in E1 Salvador, 2 each in Nicaragua and Costa Riea, and 1 in the Dominican Republic. Among Andean South American materials, 54 landraees came from Colombia, 36 from Ecuador, 19 from Peru, 14 from Chile, 3 from Bolivia, and 1 from Argentina. From lowland South America, 17 Brazilian landraees were included. The wild P. vulgaris accessions included in the cluster analysis were those studied by Koenig and Gepts (1989b). Allozyme analyses of a crude homogenate of primary leaf or root tissue (depending on the enzyme system assayed) from an average of five seedlings (I0 d old) grown in vermiculite were performed as described by Koenig and Gepts (1989b). A total of nine enzyme systems showing polymorphism were assayed: ribulose bisphosphate earboxylase (small subunit; RBCS), shikimate dehydrogenase (SKDH), peroxidase (PRX), malic enzyme (ME), malate dehydrogenase (MDH), diaphorase (DIAP), and leueine aminopeptidase (LAP). The MDH and DIAP enzyme systems each had two independent loci. Seven additional enzyme systems did not reveal polymorphisms among eultivars: aspartate aminotransferase, fruetokinase, giucose-6-phosphate dehydrogenase, glucose phosphate isomerase, peptidase, 6phosphogiuconate dehydrogenase, and triose phosphate isomerase. Allozyme loci and alleles were designated as in Koenig and Gepts (1989a). In the nine systems studied, the most common allele was designated 100 and all other allozymes were measured in millimeters from the standard. In order to compare allozyme diversity in the wild ancestor and the cultivated descendant, data from cultivated landraces were analyzed jointly with data from 83 wild forms obtained earlier (Koenig and Gepts, 1989b). A dendrogram based on Nei’s (1973) genetic distance (D) was constructed according to the unweighted paired group method of Sheath and Sokal (1973) with a statistical package provided by Dr. K. Ritland, Univ. of Toronto.