Crop Science Grow Your Career with CSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 27 May 2005
Published in Crop Sci 45:1320-1328 (2005)
© 2005 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Durán, L. A.
Right arrow Articles by Beaver, J. S.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Durán, L. A.
Right arrow Articles by Beaver, J. S.
Agricola
Right arrow Articles by Durán, L. A.
Right arrow Articles by Beaver, J. S.
Related Collections
Right arrow Other Legumes
Right arrow Germplasm Enhancement

PLANT GENETIC RESOURCES

Morphological and Molecular Characterization of Common Bean Landraces and Cultivars from the Caribbean

L. A. Durána, M. W. Blaira,b, M. C. Giraldoa,b, R. Macchiavellia, E. Prophetea,d, J. C. Nina,e and J. S. Beavera,c,*

a L.A. Duran, Pioneer HiBred International, Inc., Salinas, Puerto Rico 00751
b CIAT, AA6713 Cali, Colombia
c R. Machiavelli, Dep. of Agronomy and Soils, Univ. of Puerto Rico, P.O. Box 9030, Mayaguez, PR 00681
d CRDA, Port-au-Prince, Haiti
e IDIAF, San Juan de la Maguana, Dominican Republic

* Corresponding author (jbeaver{at}uprm.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Little information is available regarding the relationship of Caribbean bean landraces with the Andean and Mesoamerican gene pools of common bean (Phaseolus vulgaris L.) While small-seeded Mesoamerican black beans are prevalent in parts of the Caribbean, many of the regionally preferred red mottled and medium-to-large-seeded bean landraces found there are postulated to belong to the Andean gene pool. To test this hypothesis, we analyzed the morphological characteristics, phenological traits, phaseolin status, and random amplified polymorphic DNA (RAPD) banding patterns of 54 red mottled or large-seeded bean landraces and cultivars from the Caribbean (16 from the Dominican Republic, 14 from Haiti, 1 from Jamaica and 23 from Puerto Rico) compared with 11 Andean bean lines from other regions. To estimate phylogenetic relationships among the lines, distances were calculated and dendrograms constructed for morphological and/or phenological and molecular characteristics by the complete clustering method. The landraces were grouped into two clusters morphologically: one with Mesoamerican characteristics, which included all the red mottled lines from Haiti and three landraces from the Dominican Republic collected near the Haitian border and the other with Andean characteristics, which included all the lines from Puerto Rico and the remaining lines from the Dominican Republic. RAPD and phaseolin polymorphisms identified three groups, one corresponding to the genotypes with Mesoamerican morphologies, one to those with Andean morphologies, and another that had Andean phenotypes but proximity to the Mesoamerican group, suggesting possible introgression between the gene pools. Phaseolin variability agreed with other molecular and morphological data allowing the assignment of genotypes to a Mesoamerican or Andean gene pool. Knowledge of this variability may provide useful information concerning the potential value of Caribbean landraces to Andean bean breeding programs.

Abbreviations: CIAT, Centro Internacional de Agricultura Tropical • CRDA, Centre de Recherche et de Documentation Agricoles • IDIAF, Instituto Dominicano de Investigaciones Agropecuarias y Forestales • MCA, multiple correspondence analysis • RAPD, random amplified polymorphic DNA • UPGMA, unweighted pair group method with arithmetic mean • UPR, University of Puerto Rico


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
THE INTRODUCTION of common beans into the Caribbean from South America was postulated to have occurred along the "Arawak Arc" of the Leeward Islands and the Greater Antilles long before the arrival of Columbus (Rouse, 1992). A second introduction probably occurred from Central America along trade routes from the Yucatan to Cuba and Hispaniola (today the countries of Haiti and Dominican Republic) or around the Caribbean coast (Castiñeiras et al., 1991). Other crops of South American origin produced by the Tainos in the West Indies that may have accompanied common bean on its introduction from South America, include cassava (Manihot esculenta Crantz subsp. esculenta), pineapple [Ananas comosus (L.) Merr.], and peanuts (Arachis hypogaea L.) (Rouse, 1992). Historical references show that beans were well distributed in the Caribbean dating from the voyages of Christopher Columbus in 1493, who observed fields of "beans," unlike those produced in Spain, planted on the northern coast of Cuba (Castiñeiras et al., 1991) and later those of Oviedo and Valdés who in 1535 noted that Native Americans produced beans in Hispanola (Gepts and Debouck, 1991).

Common beans are believed to have been domesticated in two centers of origin, one in Middle-America and one in South America giving rise to two separate gene pools, the Mesoamerican and the Andean, that are distinguished at the morphological, physiological, and molecular levels (including phaseolin seed protein, isozyme, and DNA polymorphisms) (Gepts et al., 1986; Koenig and Gepts 1989; Singh et al., 1991c; Lynch et al., 1992; Tohme et al., 1996). On the basis of agro-ecological and morphological data (Singh et al., 1991a), supported by molecular evidence (Singh et al., 1991c; Beebe et al., 2000), the Mesoamerican gene pool includes the races Durango, Jalisco, and Mesoamérica whereas the Andean gene pool includes the races Chile, Nueva Granada, and Peru.

The Andean and Mesoamerican gene pools have marked morphological differences. Andean types generally tend to have larger leaves with ovate or lanceolate central leaflets and lanceolate or triangular bracteoles, longer internodes in the main stem, larger seeds, a stripeless standard on white flower, and a central pod beak position. Mesoamerican types, on the other hand, have striped standards and colored flowers, a placental pod beak position, and cordate or ovate bracteoles. The association between morphological traits and between some molecular markers is due to the autogamous reproductive system of common bean and geographic isolation (Singh et al., 1991a).

Support for the diversity between the Andean and Mesoamerican gene-pools and their various races has been found by various molecular techniques. Diversity in the types of phaseolin, the major seed storage protein in common bean, has been especially useful for classifying beans into Andean and Mesoamerican gene pools since most of the cultivars from one center of domestication possess a certain set of phaseolin types which are not found in cultivars or wild types from the other center of domestication (Gepts et al., 1986; Gepts et al., 1988; Singh, 1989; Gepts, 1990). Notably, phaseolin types C, H, A, and T are commonly found among Andean cultivars while S, Sb, and Sd types are common among Mesoamerican cultivars (Singh, 1988).

Other marker systems that have been applied to common bean diversity studies include isozymes, which have been used to identify races within the gene pools (Singh et al., 1991b, 1991c), as well as RAPD (Johns et al., 1997; Beebe et al., 2000), and amplified fragment-length polymorphism (AFLP) markers (Tohme et al., 1996; Beebe et al., 2001). RAPD analysis has been widely applied to the study of overall bean diversity (Haley et al., 1994) and the diversity of specific germplasm: Johns et al. (1997) found that RAPDs were able to classify bean germplasm from Chile into Andean and Mesoamerican gene pools, while Beebe et al. (2000) found that RAPDs were able to effectively distinguish groups within both Mexican germplasm and the Mesoamerican gene pool as a whole and to allow their assignment to previously determined races.

Little information is available concerning Caribbean landraces and their phylogenetic relationship with the Andean and Mesoamerican gene pools and their respective races. There is a range of small to large seeded landraces in the Caribbean with predominant colors being small black in western Cuba and Haiti and red mottled in eastern/Central Cuba, the Dominican Republic, and highland regions of Haiti (Beaver and Molina, 1997). Castiñeiras et al. (1991) collected over 300 accessions in Cuba half of which were small-seeded black beans and one fifth of which were medium-to-large, red, or red mottled beans, while Lioi et al. (1990) analyzed phaseolin frequency and found two phaseolin types, "S" associated with the small-seeded blacks and "T" associated mostly with the large-seeded landraces. Minor color classes in the Caribbean include small-seeded white or brown beans, medium-seeded tan, pink, or yellow beans, and large-seeded light red kidney beans or pink striped kidney beans (Beaver and Molina, 1997; Voysest, 2000).

Caribbean bean landraces of Andean origin are the source of many useful biotic and abiotic stress resistance traits. For example, the red mottled cultivar PC-50, a selection from Dominican Republic landrace Pompadour Checa, is the source of the Ur-9 and Ur-12 resistance genes to bean rust, caused by Uromyces appendiculatus (Pers.) Unger var. appendiculatus (Jung et al., 1998). Other red mottled landrace cultivars from the Dominican Republic have resistance to common bacterial blight, caused by Xanthomonas axonopodis pv. phaseoli (Beaver et al., 1992), or have been used as the source of the I gene for the small red bean cultivar Catrachita as in the case of G 6616 (Pompadour Checa) (Steve Beebe, Pers. Comm.). The pink-striped kidney landrace Indeterminate Jamaica Red has been used to improve the heat tolerance of kidney beans in the USA (Miklas et al., 2000), while the red mottled cultivar Salagnac 90A has tolerance to aluminum toxicity and acid soils. Despite these advantages, the continued cultivation of the traditional seed classes and landrace varieties in the Caribbean is threatened by high human population density, a general decline in agriculture and replacement by improved bean cultivars or bean imports.

The objective of this research was to characterize the variability of traditional red mottled and medium-to-large-seeded bean landraces from the Caribbean, on the basis of morphological and molecular characteristics, including phaseolin determination and RAPD genotyping. Knowledge of this variability should provide useful information concerning the potential value of these lines to Andean bean breeding programs and should help determine if additional collections should be made within the Caribbean.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Plant Material
A total of 65 bean accessions were included in this study, including 54 red mottled and medium-to-large-seeded bean landraces from the Caribbean (16 from the Dominican Republic, 14 from Haiti, one from Jamaica and 23 from Puerto Rico) as well as 11 Andean bean genotypes from other regions (six advanced lines from the CIAT breeding program, one advanced line from the UPR breeding program, as well as modern cultivars from Colombia (‘ICA Palmar’), Peru (‘Blanco Laran’), and the USA (‘Montcalm’, ‘Redhawk’) (Table 1). All the accessions from Puerto Rico and most of those from the Dominican Republic were named after the site of collection. Accessions from the Dominican Republic, Haiti, and Puerto Rico were collected and are stored by IDIAF, CRDA, and UPR, respectively.


View this table:
[in this window]
[in a new window]
  		Table 1. Origin and description of Andean bean lines planted at Isabela, Puerto Rico, in November 2000 and March 2001.

 
Field Trials
The bean genotypes were planted in field trials at the University of Puerto Rico substation near Isabela, Puerto Rico (18°30' Latitude North, 67°00' Longitude West at 128 m above sea level on a Coto, Tropeptic Haplorthox, clayey, Kaolitic, isohyperthermic soil type) in November 2000 and March 2001. A randomized complete block design with two replications was used. Experimental units consisted of single 1-m row lengths. The spacing was 0.5 m between rows and 6 cm between plants within rows. Each row was harvested to estimate seed yield per plant and seed yield components (number of pods per plant, number of per seedpod, and 100-seed weight).

Morphological and Phenological Traits
A total of nine morphological traits including (i) leaf width in centimeters of three center trifoliolate leaves; (ii) leaf length in centimeters of three center trifoliolate leaves; (iii) leaf shape, recorded as chordate, ovate, rhombohedric, or hastate following the classification developed by Singh et al. (1991a); (iv) growth habit classified by the CIAT (1987) 1-to-4 scale; (v) length of the fifth internode on the main stem measured in centimeters; (vi) bracteole shape classified as cordate, ovate, lanceolate, or triangular; (vii) bracteole size classified as small, medium, or large; (viii) outer base of the standard of the corolla classified as striped or smooth; and (ix) pod beak position classified as either placental or central. Most of the morphological traits were measured at the R1 (initiation of flowering) stage of development (CIAT, 1987). Characteristics were measured on three plants, chosen at random from each row. Two phenological traits, days to flowering and days to harvest maturity, were recorded on the basis of the average of the entire plot's growth stage (CIAT, 1987).

RAPD Analysis
DNA was isolated from young trifoliate leaves of a single plant per accession by the method described by Miklas et al. (1993). DNA was quantified with a Hoefer TKO 100 DNA Fluorometer (Hoefer Scientific Instruments, San Francisco, CA). Twenty-six RAPD primers were used for DNA amplification which had been reported to generate polymorphic DNA fragments on Andean beans and to be good indicators of gene pool origin by Johns et al. (1997). PCR reactions were mixed in a total volume of 25 µL that included 2.5 µL of genomic DNA at a concentration of 10 ng/µL, 2.5 µL of RAPD primer at a concentration of 10 µM, 5.0 µL of 25 mM MgCl2, 4.0 µL of 4 mM dNTPs, 0.2 µL (1 unit) of Taq polymerase (Sigma, St. Louis, MO), and 2.5 µL of reaction buffer (10x). Amplifications were performed on a Model 480 thermal cycler (PerkinElmer, Foster City, CA), which was programmed with two initial cycles of 60 s at 91°C, 15 s at 42°C, and 70 s at 72°C; this was followed by 38 cycles of at 15 s at 91°C, 15 s at 42°C, and 70 s at 72°C and a final extension period of 4 min at 72°C. A total of 15 µL of the amplified DNA from each sample were mixed with 1/10th volume of loading buffer and were run on 1% (w/v) agarose gels (100 mL of TE buffer and 4 µL of 10 mg mL–1 ethidium bromide) for 5 h at approximately 75V in 1x TAE buffer (0.04 M Tris-acetate, and 0.001 M EDTA). The DNA bands were viewed under ultraviolet light and photographed with Polaroid film for permanent record. Only heavily staining polymorphic bands were considered for subsequent statistical analysis. Eighty-three percent of the primers (20 in total) effectively amplified the DNA extracted from the bean genotypes, generating 39 polymorphic bands out of 181 produced. The number of bands per primer varied from five to 15 with an average of 9.05. The average number of polymorphic bands per primer was 1.93.

Statistical Analysis
Analysis of variance and Pearson product-moment correlations were performed for all quantitative traits. Least significant differences were used to compare the means at a confidence level of 95%. Numerical values were assigned to the following categorical traits: leaf shape (0 = chordate, 1 = ovate, 2 = rhombohedric and 3 = hastate), flower pigmentation (0 = white, 1 = lavender), base of standard (0 = striped and 1 = smooth), bracteole size (0 = small, 1 = medium, 2 = large), bracteole shape (0 = cordate, 1 = lanceolate, 2 = ovate and 3 = triangular), and pod beak position (1 = central, 2 = placental). Average cluster analysis was performed on the morphological data by SAS CLUSTER procedure and Euclidean distance as the dissimilarity index, and a dendrogram was constructed with the results. RAPD bands were scored for presence or absence and this dataset was used to estimate genetic distances between accessions on the basis of complete clustering and was also used to construct a dendrogram showing the relationships between accessions.

Phaseolin Determination
Phaseolin storage protein was analyzed by a procedure used at the Genetic Resource Unit of CIAT, namely: total seed proteins were extracted from 0.10 g of peeled, finely ground, oven-dried seed in 100 mL of extraction buffer developed by Brown et al. (1981) [0.036 M TRIS-HCl pH: 8.5, 1% (w/v) NaCl]. Forty microliters each of the protein extracts were mixed with equal volume of cracking buffer [0.625M TRIS-HCl pH: 6.8, 40% (w/v) sucrose, 1% (v/v) mercaptoethanol, 0.01% (v/v) Bromphenol blue, 2% (w/v) SDS]. This protein mixture was fractionated by boiling for 5 min and loading one microliter onto 6% (w/v) stacking/12% (w/v) running SDS-PAGE (polyacrylamide) minigels run for 50 min and stained with Coomasie Blue dye for 1 h followed by d-staining in 20% (v/v) methanol solution, as described by Ma and Bliss (1978). The phaseolin pattern was compared with known standards provided by Cesar Ocampo of the Genetic Resource Unit of CIAT.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
Agronomic and Morphological Traits
Yields were higher on average for the lines from Haiti than for those from the Dominican Republic, which in turn were higher than those from other regions and those from Puerto Rico in both the 2001 planting and in the combined analysis (Table 1). On the other hand, in the 2000 season, the yields were highest for the lines from the Dominican Republic, with the highest yields registered by three Dominican lines (Pomjor 17, Hondo Valle 24, Vasón 25) each with yields >1700 kg ha–1. In 2001, three genotypes from Haiti (‘35-7’, ‘779-1’ and Salagnac 90A) and two genotypes from the Dominican Republic (‘La Chula’ and ‘Pacasas 21’) produced mean seed yields >1000 kg ha–1, while none of the lines from Puerto Rico or other regions produced mean seed yields >1000 kg ha–1 in this environment. All the bean lines from Haiti and some of the lines from the Dominican Republic and Puerto Rico had indeterminate (Type III) growth habit (Table 1) as well as significantly smaller seeds, greater number of pods per plant, and seeds per pod (Table 2). The selection by Haitian farmers of red mottled bean landraces with Mesoamerican traits may have contributed to greater seed yield stability in these lines as has been observed previously (Beaver et al., 1996).


View this table:
[in this window]
[in a new window]
  		Table 2. Means of yield components of Andean bean lines planted at Isabela, Puerto Rico, in November 2000 and March 2001.

 
The majority of the genotypes collected and evaluated in this study had red mottled seed type (Table 1). Indeed, all the genotypes evaluated from Haiti and the Dominican Republic had either red or purple mottled grain types. This seed class is often referred to as the "Pompadour" class, although this name is often delimited to purple mottled and round beans. The collection from Puerto Rico, on the other hand, included a range of other colors and shapes including one white, two yellow, and nine pink-striped landraces, most of which were kidney shaped. The single landrace from Jamaica also had pink striped grain and this seed class is often referred to as the "Miss Kelly" class. The landraces varied in terms of seed size from a high of 58.7 g 100–1 seed to a low of 22.3 g 100–1 seed for grain harvested over the two environments. The average seed weight was highest for the Dominican landraces in 2000 and for the Andean checks in 2001, although in both seasons these two groups had similar sized seed (Table 2). The next largest average seed was found in Puerto Rico followed by Haiti, which had the smallest seed on average across both seasons.

In the cluster analysis for morphological traits (Table 3), two major groups were observed, one having predominantly Mesoamerican characteristics and the other having Andean characteristics (Fig. 1). The red mottled bean landraces from Haiti along with three similar red mottled bean landraces (Chijor 35, Vasón 25, and Hondo Valle 24) collected in Dominican Republic near the Haitian border formed a homogenous group that had more Mesoamerican gene pool characteristics than Andean gene pool characteristics. In addition to the growth habit and small seed described above, Haitian bean lines also had a placental pod beak position and shorter fifth internode lengths and smaller cordate leaves characteristic of the Mesoamerican gene pool. Meanwhile, the second group was made up of the bean genotypes from Puerto Rico, the majority of genotypes from the Dominican Republic and all of the Andean control genotypes from other regions. All of these genotypes had typical Andean morphologies of larger leaf size and central pod beak position. Most also had determinate (Type I) growth habit, earlier maturity, and consequently lower seed yield potential. However, some Type II and Type III growth habits were found at least for one line from Puerto Rico and seven Andean group genotypes from the Dominican Republic (Table 2). Within the two major groups, similar genotypes with similar seed colors or those from the same collection site tended to cluster together in the dendrogram as was observed for Colorado del País 1 and Colorado del País 2, both of which were striped pink kidneys; Gurabo 2, Gurabo 3, and Gurabo 6, all of which were striped pink beans; José Beta and Maguana, both large red mottled beans from Dominican Republic; Naranjito 1 and Naranjito 2, which were red and purple mottled respectively; Orocovis 1C and Orocovis 2, both of which were purple mottled; and Redhawk and Montcalm, both dark red kidneys.


View this table:
[in this window]
[in a new window]
  		Table 3. Frequency of morphological traits of Andean bean lines planted at Isabela, Puerto Rico in November 2000 and March 2001.

 


View larger version (35K):
[in this window]
[in a new window]
  		Fig. 1. Grouping of Caribbean bean landraces and Andean lines on the basis of morphological traits. Gene pool (Andean, Mesoamerican) identification as explained in the text.

 
There were positive and significant phenotypic correlations between seed yield per plant and both number of pods per plant (r = 0.79) and relative maturity (r = 0.28) (Table 4). The phenotypic correlations were negative and significant between 100-seed weight and both number of pods per plant (r = –0.41) and number of per seedpod (r = –0.73) and positive and significant between 100-seed weight and both fifth internode length (r = –0.43) and relative maturity (r = 0.31) (Table 4). The highest yielding landraces had many Mesoamerican characteristics including indeterminate growth habit, a smaller seed size, a greater number of pods per plant and more seed per pod. Other characteristics were not correlated.


View this table:
[in this window]
[in a new window]
  		Table 4. Phenotypic correlations between seed yield components and morphological traits of Andean bean lines planted at Isabela, Puerto Rico in November 2000 and March 2001.

 
RAPD and Phaseolin Cluster Analysis
RAPD-based clustering detected three main groups at a Eculidean distance <1.4 as shown in the dendrogram for this analysis (Fig. 2). Group I consisted in genotypes of a mixed origin, all large-seeded genotypes which had typical Andean morphologies but also had traits that suggested introgression from the Mesoamerican gene pool as shown by their RAPD profiles. Within this group two subgroups could be distinguished, one included the red kidney lines AFR285, Redhawk, Montcalm and PR9443-4, while the other group contained the red mottled lines A36, AFR619, AFR699, and Pacasas 21. Redhawk and Montcalm, two closely related cultivars from the USA, could not be distinguished. Interestingly, all the lines in Group I, except for Pacasas 21, which is a landrace from the Dominican Republic, were from breeding programs in CIAT (A36, AFR285, AFR619, AFR699), Puerto Rico (PR9443-4), and the USA (Redhawk and Montcalm). This might be expected because of the directed nature of introgression practiced in breeding programs where Mesoamerican genotypes have been used to improve specific characteristics of Andean beans (Beaver, 1999). Most of the Andean breeding lines with the S phaseolin allele are derived from crosses between the Mesoamerican and Andean gene pools. The Mesoamerican bean line GN #1 sel. 27, which has been used as a source of resistance to common bacterial blight, is in the pedigrees of Montcalm, Redhawk, and Africa 285. Redhawk also has the Mesoamerican gene Co-2 for anthracnose resistance that is derived from the line Cornell 49242. DOR303, a CIAT breeding line derived from crosses between gene pools, was the source of Bean golden yellow mosaic virus (BGYM) resistance of PR9943-4.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 2. Grouping of Caribbean bean landraces and Andean lines on the basis of RAPDs and Phaseolin data. Cluster numbering as explained in text.

 
In comparison, introgression appeared to be less frequent or less intense in the Caribbean landraces given that only one landrace pertained to this group. It will be interesting to confirm whether Pacasas 21 is a result of introgression between gene pools with additional markers and comparison to Mesoamerican checks (Amirul Islam et al., 2004).

The genotypes with Mesoamerican morphologies, namely all the Haitian red mottled beans and the three Dominican landraces mentioned previously (Chijor 35, Hondo Valle 24, Vason 25) all clustered together in Group II as seen in the dendrogram (Fig. 2). Group III was made up of all the Caribbean genotypes with Andean morphologies and included 12 landraces from the Dominican Republic, one from Jamaica, and 23 from Puerto Rico as well as the Andean control genotypes (CAL96, DRK57, Blanco Laran, and ICA Palmar). Within Group III, there were two subgroups, designated IIIa and IIIb, that were distinguished at a Euclidean distance of <1.2 (Fig. 2). Group IIIa was associated with the Andean controls CAL96 and DRK57, while Group IIIb was associated with Blanco Laran and ICA Palmar.

Within each group, the Puerto Rican genotypes tended to cluster separately from the Dominican landraces, while some genotypes with similar seed colors or those from the same collection site clustered together, as seen for Jose Beta and Maguana; Orocovis 1C, ‘Orocovis 2B’, and Naranjito 2; ‘Naranjito 6’ and Gurabo 6; Gurabo 2 and ‘Coamo 13’; ‘Gurabo 4’, ‘Gurabo 5’ and ‘Coamo 14’; and ‘Gurabo 1’ and ‘Coamo 11’. Some of these same clusters were observed in the dendrogram on the basis of the morphological data; however, in this dendrogram, adjacent clusters did not always have the same seed color or grain type. In this clustering, it was surprising to find ‘PC50’ and Orocovis 2 as well as ‘JB178’ and Orocovis 1C together in the dendrogram given that in each pair, one genotype was from Puerto Rico and one was from Dominican Republic. However, since all of these genotypes were red or purple mottled grain types, this may be taken as evidence of gene flow between the two islands.

It was also notable that while improved cultivars of red kidney beans had all fallen in the introgressed category of Group I as described above, the striped kidneys native to the Caribbean were found in Group III instead. This may indicate that the short-striped kidney grains typical of the Caribbean are substantially different than the longer, more bullet-shaped kidney grains found in Montcalm and Red Hawk that are the standards of international commercial trade (Kelly et al., 1998). Variability in the shade of pink or red grain color also differentiates these two groups.

Only three phaseolin patterns were found among the Caribbean landraces: the most common being the T allele typical of many bush Andean beans, the second most common being the S allele typical of Mesoamerican beans, while a third pattern, the "C" variant, was found for only a single traditional variety from Puerto Rico (Naranjito 1). The C pattern is thought to be a hybrid of S and T phaseolins (Paredes and Gepts, 1995), but the hybrid origin of this accession could not be confirmed without the use of codominant markers.

Phaseolin genotyping confirmed the gene pool assignments made by the RAPD-based phylogenetic analysis described above: all the Haitian landraces were observed to have the Mesoamerican S phaseolin pattern except for the accession ‘66-1’, which had a mixture of S and T phaseolin patterns associated with the mixture of medium and large grains in this accession; the three Dominican landraces (Chijor 35, Hondo Valle 24 and Vasón 25), which had been grouped with the Haitian landraces, on the basis of morphological traits and RAPD profiling, also had Mesoamerican S phaseolin type; the majority of the Puerto Rican and Dominican landraces that were grouped as Andeans according to the dendrograms had Andean T phaseolin (Table 1). Among the Puerto Rican landraces, one genotype, ‘Orocovis 1A’, had the Mesoamerican S phaseolin type, while another, Naranjito 1, had the hybrid Mesoamerica/Andean C phaseolin. Five Andean lines from other regions (Montcalm, Redhawk, PR9443-4 AFR285 and Blanco Laran) also had S phaseolin, which could explain the intergene pool status of the first four of these genotypes in the RAPD cluster analysis. Meanwhile, Orocovis 1A and Naranjito 1, like the control genotype Blanco Laran, all occurred in Group IIIb and may have represented intergene pool introgression that occurred naturally and that was potentially selected for by farmers in Puerto Rico.

Phaseolin alleles have been correlated with seed size in traditional cultivars in the Caribbean (Castiñeiras et al., 1994) with S types generally being small seeded and T types being large seeded. In addition, the phaseolin locus has been associated with a QTL underlying seed size in genetic studies (Gepts et al., 1988). However, as shown by the Andean controls and Caribbean landraces used in this study, medium-to-large-seeded beans have been developed by breeding programs or by farmer selection with S phaseolin. In both cases, this is evidence of introgression between the Mesoamerican and Andean gene pools. Among the landraces, although S phaseolin was found in several medium-seeded landraces, all the largest seeded landraces did have the T phaseolin. On the other hand, all the smaller seeded landraces did have S phaseolin despite their typical "Andean" red-mottled seed coat coloring, further evidence of intergene pool introgression. It was also interesting to note that introgression of phaseolin or seed color from one gene pool to another occurred only in the red mottled genotypes, not in the cream striped, yellow, pink striped, or purple mottled seed classes, where all the genotypes had typical Andean seed coloration and T phaseolin as expected.

Gepts et al. (1988) suggested that the Caribbean was a natural geographic bridge between the two recognized centers of origin for P. vulgaris. Thus, it would be logical not only to find types from both centers but also to follow some sort of gradient across the region for their prevalence. Indeed, in western Cuba, common bean production is almost entirely composed of black-seeded beans of Mesoamerican origin, while red mottled and light red kidney beans are produced in the mountains of eastern Cuba (Castiñeiras et al., 1991). Farther east in Haiti, the predominant market class is black-seeded bean of Mesoamerican origin although red mottled beans are the preferred seed type on the central plateau (Voysest, 2000). In the Dominican Republic, the red mottled bean is the predominant seed type, whereas a striped light red kidney bean is the preferred Andean bean in Puerto Rico (Beaver and Molina, 1997; Voysest 2000).

The phylogenetic analysis conducted in this study agrees with the hypothesis of a west-to-east increase in the proportion of Andean beans within the Caribbean. For example, from the phaseolin genotyping, it was notable that in Haiti 30 out of 31 entries (96.7%) had S phaseolin and in the Dominican Republic, three out of 16 cultivars (18.8%) had the S allele, while in Puerto Rico, only one out of 23 cultivars (4.3%) had S phaseolin. Conversely, the frequency of the Andean T phaseolin allele increased in direct proportion to the reduction in Mesoamerican S phaseolin allele.

In summary, this study has found that medium-to-large-seeded Caribbean landraces of common bean are diverse and are divided into two distinct groups one with a Mesoamerican phenotype and the other with an Andean phenotype. However, genotyping with RAPDs and with phaseolin showed that some introgression has potentially occurred between these two groups or gene pools, although this has been limited. Because pyramiding Mesoamerican and Andean disease resistance genes may provide broader and/or more durable resistance (Kelly et al., 2003), we postulate that the recombination between Andean and Mesoamerican traits found in Caribbean germplasm may provide a unique source of gene combinations useful to breeding programs and that Caribbean germplasm is worthy of further collection.


   ACKNOWLEDGMENTS
 
We are grateful to Carlos German Munóz for assistance in evaluating genotypes in field experiments, to O. Toro and C. Ocampo for advice on Phaseolin determination and to James Kelly and Shree Singh as reviewers. This research was supported in part by a grant from the USAID Bean/Cowpea Collaborative Research Support Program.

Received for publication August 19, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Durán, L. A.
Right arrow Articles by Beaver, J. S.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Durán, L. A.
Right arrow Articles by Beaver, J. S.
Agricola
Right arrow Articles by Durán, L. A.
Right arrow Articles by Beaver, J. S.
Related Collections
Right arrow Other Legumes
Right arrow Germplasm Enhancement

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome