GENETIC VARIABILITY AND PEDIGREE ANALYSIS OF BRAZILIAN COMMON BEAN ELITE GENOTYPES

Genetic diversity is essential for any breeding program. However, breeders tend to concentrate on specific genotypes, which combine traits of interest and may be used as progenitors in several breeding programs. Common bean (Phaseolus vulgaris L.) breeding programs are not different in this sense. In this study, the genetic diversity of 21 common bean elite lines from the Bean Regional Trials conducted by the Embrapa Rice and Bean Research Center was evaluated using the Random Amplified Polymorphic DNA (RAPD) and pedigree analyses. Based on genetic dissimilarity, three groups were defined: group I lines 1, 9 and 10, with low genetic distances among them (0.00 to 0.06), originated from 11 Mesoamerican parents; group II 17 lines with genetic distances ranging from 0.03 to 0.33, originated from 50 parents (mostly Mesoamerican); and group III line 21 (PR 93201472), which parents are the Andean cultivar Pompadour and the cultivar Irai (unknown origin). The genetic distances between line 21 and the lines of the other two groups varied from 0.68 to 0.93. Pedigree analyses demonstrated that cultivars Carioca, Cornell 49-242, Jamapa, Tlalnepantla 64, Tara and Veranic 2, all of Mesoamerican origin, were the most widely used parents for developing lines present in group II.


INTRODUCTION
Brazil, Mexico, Argentina and United States are the leading common bean (Phaseolus vulgaris) producers in the world.In Latin America and Africa, beans are major staple food and an important protein source.Brazil is the world leading producer of Phaseolus beans, but its production is destined to internal consumption.The U.S. is the leading exporter marketing 40% (0.4 million metric tons) of all their dry bean production overseas, Scientia Agricola, v.60, n.2, p.283-290, Abr./Jun.2003 while Argentina exports 99.5% of its production of 0.25 million metric tons (Phaseolus Crop Germplasm Committee Report, 1996;FAO, 1998).
Studies on variation patterns of the seed protein phaseolin, alloenzymes and morphological evidences, revealed the existence of a Middle American and an Andean gene pools in Phaseolus vulgaris (Gepts et al., 1986;Singh et al., 1991a;Singh et al., 1991b).Phaseolin S is common among the three Phaseolus races in the Middle American gene pool, while Sb, Sd and B phaseolins are less common.On the other hand, phaseolin T, C and H are present in genotypes from the Andean gene pool.Not only the phaseolin patterns but also the size of the seeds differentiates Middle American from Andean gene pools.Mesoamerican landraces have usually smaller seeds than their Andean counterparts (Gepts, et al., 1986;Singh et al., 1990).
Biochemical studies have shown that the cultivated bean germplasm presents a low level of genetic diversity as compared to their wild ancestors (Gepts et al., 1986;Koenig et al., 1990;Sonnante et al., 1994;Johnson & Gepts, 1998).Progress in bean breeding around the world has been slow, probably because of a limited genetic variability of the parents which have been selected from the same gene pool.
Isozymes and DNA-based markers have been used to evaluate germplasm genetic diversity in Phaseolus species, within both cultivated and wild genetic materials; to analyze crop evolution; and the effect of domestication on genetic diversity (Gepts, 1988;Singh et al., 1990;Koening et al., 1990;Chase et al., 1991;Singh et al., 1991a, b, c;Arraes et al., 1992;Thome et al., 1996;Nienhuis et al., 1996;Alvarez et al., 1998, Vasconcelos, 1995;Beebe et al., 2000).Métais et al. (2000) evaluated the effectiveness of Restriction Fragment Length Polymorphism (RFLP), Direct Amplified Minisatellite DNA-PCR (DAMD-PCR), Inter Simple Sequence Repeat (ISSR) and Random Amplified Polymorphic DNA (RAPD) markers in the assessment of polymorphism and relationships between commercial lines of P. vulgaris.RAPD and RFLP analyses led to the same clustering of the bean lines.All analyzed genotypes could be distinguished by only seven RAPD markers.
The objectives of the present study were: to evaluate the genetic variability within 21 elite cultivars from the Bean Regional Trials coordinated by Embrapa Rice and Bean Research Center using RAPD markers; and to identify the parents involved in the development of these cultivars through pedigree analyses.

MATERIAL AND METHODS
The pedigree of each of the 21 elite lines from Regional Trails used in this study is shown in Table 1.
Each amplification cycle consisted of one denaturation step at 94 o C for 15 s, one annealing step at 35 o C for 30 s, and one extension step at 72 o C for 1 min.After 40 cycles, an extra extension step was performed for 7 min at 72 o C. Amplification products were analyzed on 1.2% agarose gels immersed on TBE buffer (90 µmol L -1 Trisborate; 2 µmol L -1 EDTA, pH 8.0) containing 10 µg of ethidium bromide per mL (Williams et al., 1990).DNA bands were visualized under UV light and photographed with the Eagle Eye II photosystem (Stratagene, La Jolla, CA, USA).
The clustering analysis was based on the genetic distance calculated by the Nei & Li method for binary data, using the GENES computer program (Cruz, 1997).Based on the genetic distances a dispersion graphic on two-dimensional space was obtained (Cruz & Soriano, 1994).

RESULTS AND DISCUSSION
General information on cultivars/lines involved in the breeding program to obtain the elite lines of the Bean Regional Trials is presented in Table 2.The DNA amplification results presented 32 polymorphic and 46 monomorphic bands, among the 21 bean cultivars.Based on the genetic distances a two-dimensional dispersion graph was obtained.In the dispersion graph, it was possible to distinguish three different groups (Table 3, Figure 1).
The first group was formed by three black seed lines with identical genealogies, LR 9115398 (1), AN 9021334 (9) and AN 9021336 (10), derived from 11 Mesoamerican parents.These lines presented low genetic distances between them (0.00 to 0.06) (Figure 1).Lines ( 9) and (10) showed a genetic distance of zero, while line (1) presented a genetic distance of 0.06 in relation to lines (9) and (10) (Table 3).These lines carries anthracnose and rust resistance genes, derived from the cultivar Honduras 35 known in Brazil as Ouro Negro (Faleiro et al., 2000;Alzate-Marin et al., 2001).Table -General information on cultivars/lines involved in the breeding program to obtain the elite lines of the Bean Regional Trials coordinated by Embrapa Rice and Beans.The second group was formed by 17 lines, with different seed colors, originated from 50 parents, with a genetic distance between 0.03 and 0.33 between them (Figure 1).The smallest distance was observed between the cream beige seed lines LM 93204319 (12) and LM 93204328 (13), which have identical genealogies.The highest distance was observed between lines FEB 163 (5) and LR 93201684 (20).Despite the fact that lines LR 93201684 (20) and RAO 33 (6) have Andean cultivar 'Pompadour' as progenitor in their pedigrees, they do not form a separate group.Probably those lines have low percentage of Andean genes in concordance with the small size of their seeds.Almost all lines of groups I and II were originated in Central America and Caribbean (Mexico, Costa Rica, Dominican Republic and Nicaragua), and all have small seeds, characteristic of the Mesoamerican gene pool (Tables 1, 2 and 3).
A third group, was formed only by manteigão line PR 93201472 (21) derived from a cross between the Andean cultivar 'Pompadour' originated in the Dominican Republic (Voysest, 1983) and cultivar 'Irai', of unknown origin.These lines, with medium size seeds pre-sented the highest genetic distances in relation to the other lines tested (0.68 and 0.93) (Tables 1, 2 and 3).
The low genetic variability of the parents used in the original crosses and the frequent use of the same parents for the development of these cultivars may have been the cause of low genetic variability observed in this work.For instance, the Mesoamerican cultivars/lines 'Jamapa', 'Carioca', 'Veranic 2' (selected from 'Jamapa'), 'Tlalnepantla 64' (PI 207.262) and 'Tara' have been used for the development of most of the 17 lines in group II (Vasconcelos, 1995;Gepts et al., 1986;Koenig et al., 1990;Singh et al., 1991b;Voysest, 1983Voysest, , 2000) ) (Tables 1 and 2).Voysest (2000) pointed out that Mesoamerican cultivars 'Jamapa' and'Carioca', released in 1958 and1966, respectively, are among the five cultivars most frequently grown in Latin America.Although these cultivars have been important because of their genetic stability and yield, their use as commercial cultivars and as parents in several breeding programs may have contributed for the reduction of the genetic variability of the new commercial cultivars in Brazil and other of Latin America countries.The cultivar/line 'Tara' and 'Tlalnepantla 64' (PI 207.262) are used as source of resistance to Xanthomonas axonopodis pv.phaseoli.The latter is also used as source of resistance for anthracnose.Cultivar 'Cornell 49-242' carries anthracnose and leaf spot resistance genes and is one of the 12 differential varieties used to classify pathotypes of the pathogens Colletotrichum lindemuthianum and Phaeoisariopsis griseola (Young & Kelly, 1996;Pastor-Corrales, 1992;Pastor-Corrales & Jara, 1995).Lines FEB 163, LM 93204303 and LM 93204453 have in common the same six gene donors described above.Lines TB 94-01, LM 9220225 and LM 93203304 have in their genealogies at least three of the above cultivars used as source of genes (Tables 1 and 2).
Although results show a low genetic variability among 17 bean lines (group II), derived from 50 cultivars (mostly Mesoamerican), there is no doubt that the Mesoamerican germplasm from different bean races (Mesoamerican, Durango and Jalisco) may contain important genetic variability to be incorporated into bean breeding programs.Indeed, Beebe et al. (2000) have demonstrated, by analyzing 269 Mesoamerican bean landraces, that this gene pool is an important source of genetic variability that remains to be explored.
Using 12 RAPD primers, the 21 lines could be classified into 3 groups, variation were detect in 17 lines of group II, duplicate lines were identified in groups I and II.This type of information is essential for germplasm conservation and improvement.Knowledge of the pedigree of a cultivar may be useful for the identification of sources of genes of interest and understand its role in genetic variability.However, the pedigree information is not always available for the breeder.In these cases, molecular markers can be used as an accurate tool to detect similarity/divergence and identificate duplicated bean lines or accessions among bean cultivars.Molecular analyses, in conjunction with morphological and agronomic evaluations of cultivars are recommended, because they provide complementary information and increase the resolving power of genetic diversity analysis (Singh, et al., 1991c).

Table 1 -
Seed color and pedigrees of the 21 elite common bean lines of the Bean Regional Trials coordinated by Embrapa Rice and Bean.

Table 3 -
Genetic distances among pairs of common bean elite cultivars from the Regional Trials coordinated by Embrapa Rice and Beans.