Genetic Resourcesand CropEvolution 42: 97-106, 1995. (~) 1995KluwerAcademicPublishers. Printedin the Netherlands.

97

RFLPs of the rRNA genes in the genus Phaseolus M a r e n Jacob, D o r o t h e a Z i n k & Walter N a g l Department of Biology, and The Biotechnology Program, University of Kaiserslautern, 67653 Kaiserslautern, Germany

Received26 October 1993;accepted25 March1994 Key words: restriction fragment length polymorphism, Phaseolus, rDNA, dendrogram

Abstract

Restriction fragment length polymorphisms of the rRNA genes were investigated among 20 genotypes of P. vulgaris, P. coccineus, P polyanthus, P microcarpus, P glabellus, P. acutifolius, P maculatus, P oligospermus and P lunatus. Detection of the polymorphisms was performed nonradioactively with a digoxigenin-labeled rDNA probe. RFLP-based phylogenetic trees for the rDNA of the species studied were computed with several distance matrix and parsimony methods. The estimated molecular relationships within the genus Phaseolus coincide, on principle, with the classical taxonomical investigations. Hence, RFLP analyses of the rRDA have been proven useful for systematic studies in the genus Phaseotus.

Introduction

Restriction fragment length polymorphisms (RFLPs) provide a powerful tool to study the genetic, systematic, and phylogenetic relationships in many plant taxa. Since the last years, the RFLP technique gained importance in plant breeding as a diagnostic tool (reviewed by Gebhardt & Salamini, 1992; Young, 1992). RFLP markers are co-dominant, phenotypically neutral, environmentally independent and provide more characters for analysis than biochemical, cytological or morphological data (Olsen & Woese, 1993). The genus Phaseolus comprises about 50 species (Debouck, 1988), whereby the species P vulgaris L. (common bean), P. coccineus L. (runner bean), P acutifolius A. Gray (tepary bean) and P. lunatus L. (lima bean) include the most prominent cultivated representatives. Plant ribosomal genes coding for the 18S, 5.8S and 25S cytoplasmatic RNAs represent repeated genes with up to several thousand copies per genome (Appels & Honeycutt, 1986). The genes are tandemly arranged and clustered at one or few chromosome locations, the nucleolus organizer regions. Each repeat unit includes a single rDNA transcription unit as well as an inter-

genic spacer (IGS) that separates the coding regions. The repeat units vary from 7.8 to 18.5 kb, depending on the length of the IGS. The IGS may contain sequences for initiation, termination and enhancement of transcription, and also elements for processing the rRNA precursor molecule. While the coding regions are strongly conserved (Gerbi, 1985), the IGS has been found to be highly variable in sequence, length, and copy number of subrepeats in several plant genera (Rogers & Bendich, 1987; Hemleben, 1990). Differences in the copy number of subrepeats and in the sequence of the IGS could be detected between plant species, subspecies, varieties, and in a few cases among individuals of a population, within an individual plant and within a single rDNA locus (Schaal & Learn, 1988). Hence, the variation of the IGS represents a valuable source for phylogenetic analyses of both wild and cultivated plants (e.g. Zimmer et al., 1988; Molnar et al., 1989; Santoni & Berville, 1992; King et al., 1993). To achieve a better understanding of, and to provide molecular evidence for, the genetic relationship in the genus Phaseolus, we investigated the polymorphism of the rDNA among 20 genotypes.

98 Table 1. Genotypes of Phaseolus analyzed

Taxon Phaseolus acut!fblius A, Gray var. acutifolius Phaseolus acutifolius A. Gray var. latifolius G. Freeman Phaseolus acut!folius A. Gray var. tenuifolius (Woot. & Standl.) A. Gray Phaseolus coccineusL, subsp, coccineus Phaseolus coccineus L. subsp, coccineus ev. Desir6e Phaseolus coccineus L. subsp, coccineus ev. Hammond's

Accession number

Abbreviation

Status

State of origin

NI0576 NI0562

Pal Pa2

wild cultivated

Mexico, Sonora Mexico, Sonora

NI0692 NI0132

Pa3 Pcl Pc2

wild cultivated cultivated

Mexico, Sonora Romania Germany

Pc3 Pc4 Pg H1 P12 P13 P14 P15 Pma Pmi Po Pp 1 Pp2 Pva Pv

cultivated wild wild cultivated cultivated cultivated cultivated wild wild wild wild cultivated cultivated wild cultivated

Germany Mexico Mexico Zaire Brasilia USA, Califomia USA Costa Rica Mexico, Zacatecas Mexico, Durango Mexico, Durango Guatemala, Alta Verapaz Mexico, Puebla Brasilia Netherlands

Dwarf Scarlet Phaseolus coccineus L. subsp, purpurascens Phaseolus glabellus Piper Phaseolus lunatus L. var. lunatus cvgrp. Big Lima Phaseolus lunatus L. var. lunatus cvgrp. Potato Phaseolus lunatus L. var. lunatus cvgrp. Sieva Phaseolus lunatus L. var. lunatus cv. Jackson Wonder Phaseolus lunatus L. vat. siivester Bander Phaseolus maculatus Seheele Phaseolus microcarpus Mart. Phaseolus oligospermus Piper Phaseolus polyanthus Greenm. PhaseoluspolyanthusGreenm. Phaseolus vulgaris L. vat. aborigineus (Burk.) Bandet Phaseolus vulgaris L. cv. Tendergreen

NI0722 N11270 NI0018 NI0783 NI0549 NI0583 NI0808 NI0703 Nil 116 NI0913 NI0519 NI0573

NI = accession number of the Jardin botanique national de Belgique, Meise, Belgium

Material and methods P l a n t material

The different genotypes of Phaseolus investigated are listed in Table 1. Plants were grown under a lightbench (13,000 lux) with a 16 h:8 h light/dark rhythm. For D N A extraction, 10 g of young leaves were harvested, frozen with liquid nitrogen and stored at - 7 0 ~C. D N A isolation

Total D N A was extracted using the "maize D N A miniprep" method according to Dellaporta et al. (1984), and further purified by a modified method described by Saghai-Maroof et al. (1984). The "maize D N A miniprep" method was scaled up seven times and followed exactly the protocol until the first precipitation of the crude nucleic acids with isopropanol. The pellet was redissolved in 5 ml of TE buffer (10 mM Tris pH 8, 1 m M EDTA) at 65~ with the addition of

RNase A (100/zg/ml) and proteinase K (500 #g/ml). For D N A purification, 5 ml samples were centrifuged (13,000 • g for 10 min). The supernatant was supplemented with one volume of 2 • CTAB (2% hexadecyltrimethylammonium bromide, 100 m M Tris pH 8, 20 mM EDTA, 1.4 M NaC1), followed by an extraction with chloroforrn/octanol (24:1). For D N A precipitation, 2/3 volumes isopropanol were added, and the precipitated D N A was lifted out with a glasshook, washed twice in 76% ethanol, 10 m M ammonium acetate. The D N A was dissolved in an appropriate volume of TE buffer at 65~ The yield was 2 5 - 8 0 # g D N A per gram tissue, the UV absorbance ratios 260 nm/280 nm and 260 nm/230 nm were at least 1.9 and 2.0, respectively. Restriction digestion and Southern transfer

The D N A samples were digested with B a m H I , DraI, E c o R I , HincII, SacI and X b a I in buffers recommended by the manufacturers (Boehringer, Mannheim; New

99 England Biolabs) at 37~ over night. The restricted DNA (2.5-4 #g) was separated by gel electrophoresis at 1.5 V/cm for 24-36 h in 1% agarose gels. The buffer used for casting and for running the gel was TBE (89 mM Tris, 89 mM boric acid, 2.5 mM EDTA pH 8.3). As molecular size markers, Raoul I (Appligene, France) and DIG III (Boehringer, Mannheim) were used. Completeness of digestion was checked by the pattern of plastid DNA fragments. The digested DNA was transferred to nylon membranes (Hybond N, Amersham) according to the method of Smith & Summers (1980). The DNA was immobilized on the membrane by "baking" as described by the manufacturer.

Source of probe The 3.5 kb BamHI fragment of the genomic clone pTA250, which contains a complete 8.8 kb repeat unit of the rDNA of wheat (Gerlach & Bedbroock, 1979) was used as probe. The fragment carries the complete 18S and 5.8S rDNA and part of the 25S gene. The genomic clone pTA250 was kindly provided by Prof. Dr. N. Blin (University of Ttibingen, Germany). The insert was purified by electroelution with the Biotrap apparatus (Schleicher & SchSll) and was labeled with digoxigenin (Boehringer) by random priming according to Feinberg & Vogelstein (1983).

Hybridization and detection Prehybridization was performed in glass tubes containing 20 ml hybridization solution (5 x SSC, 1% blocking reagent (Boehringer), 0.1% N-lauroylsarcosine, 0.02% SDS) at 68~ for 2 h. Hybridization was carried out with 5 ml hybridization solution (2.5 ml per 100 cm 2 membrane) containing the denatured labeled rDNA probe (5 ng/ml), over night at 68~ Filters were sequentially washed at room temperature in 2 x SSC, 0.5% SDS for 15 min, two times at 68~ in 2 x SSC, 0.5% SDS for each 15 min, and finally by two stringent washes at 68~ in 0.I x SSC, 0.5% SDS for 30 rain (the latter condition corresponds to T~ - 8~ The membranes with DraI- and XbaI-digested DNA were sequentially washed as mentioned above, except the stringent washes, which were performed at 720 C (corresponding to Tr~ - 40 C). After stringent washes the membranes were briefly rinsed in 2 x SSC at room temperature, and subsequently subjected to chemiluminescent signal detection (DIG DNA detection system, Boehringer, Mannheim) according to the

instructions of the manufacturer, except for the final steps. The damp membrane was incubated with 1 ml diluted CSPD (disodium 3-[methoxyspiro(1,2dioxetane-3,2'-tricyclo[3.3.1.1.3,7]decan)-4-yl] phenyl phosphate; Serva, Heidelberg) per 100 cm 2 in a plastic bag. Excess CSPD solution was pressed out with a ruler. The membrane was sealed in a new plastic bag and incubated at 37 ~C for 15 rain, prior to the exposure (30-90 rain) to Hyperfilm MP (Amersham).

Analysis of data The patterns of restriction fragments detected were transformed into a 0/1 matrix. The presence (1) or absence (0) of specific restriction fragments was scored over all genotypes and fragment positions. Extremely strong, not distinct bands were interpreted on xray film after shorter exposure times. The 0/I matrix was used as input for the parsimony programs MIX and DOLLOP (PHYLIP program package version 3.5, developed by Felsenstein, 1989). Furthermore, a computer program converted the 0/1 matrix into a pairwise distance matrix (Dij) (Table 2) using the complement of the Jaquard index (Jackson et al., 1989). The dissimilarity matrix (Dij) was used as input file for the distance matrix programs FITCH and KITSCH. To estimate the variation statistically, the 0/1 matrix was resampled by bootstrap analyses with the program SEQBOOT, which generated 100 data sets from the original matrix. The resulting multiple data sets were computed into the above mentioned programs (MIX, DOLLOP, FITCH and KITSCH, with M = 100). The program CONSENSE then formed strongly supported clades by the majority-rule consensus tree method. The values given at the branchs indicate the number of times the branch occurred among 100 trees. Values of 50 at the forks indicate 95% bootstrap confidence (Felsenstein, 1988).

Results

Effectiveness of endonucleases Inter- and intraspecific polymorphisms could be detected with all six enzymes used. The restriction with SacI, EcoRI and BamHI revealed constant as well as variable bands in all genotypes studied. Digestion with SacI generated two intensive constant bands of about 1.5 kb and 2.4 kb, with EcoRI only one constant intensive fragment about 3.9 kb, and restriction with BamHI

100 r d A B C O E FGH I I K L M N O P Q R S T d r

rdABCDEFGHIIKLMHOPORSTd

rdd 6C OEFG HI J K L M N O P O R $ T d

rABCDEFGfll

IK L H H O P Q R S T r

Figs. 1-4. RFLP patterns of the rDNA locus of different genotypes of Phaseolus. The DNA was digested with SacI (Fig. I), EcoRI (Fig. 2), BamHl (Fig. 3) and HinclI (Fig. 4). Lane r: 750 ng Raoul I marker, Lane d: 200 ng DIG III marker, Lane A: P. vulgaris cv. Tendergreen, Lane B: R vulgaris var. aborigineus, Lane C: P. coccineus N10722, Lane D: P. coccineus cv. Hammond's Dwarf Scarlet, Lane E: R coccineus cv. Desir6e, Lane F: P. coccineus NI 0132, Lane G: P. polyanthus N10913, Lane H: P. polyanthus N10519, Lane I: P. acutifolius var. latifolius, Lane J: P. acutifoIius var. acutifolius, Lane K: P. acutifolius var. tenuifolius, Lane L: P. glabellus, Lane M: P. microcarpus, Lane N: P. oligospermus, Lane O: P. maculatus, Lane P: P. lunatus cv. Jackson Wonder, Lane Q: P. lunatus vat. lunatus cvgrp. Big Lima, Lane R: P. lunatus var. lunatus cvgrp. Potato, Lane S: P. lunatus var. lunatus cvgrp. Sieva, Lane T: R lunatus var. silvester.

generated two constant intensive bands o f about 1.2 kb and 2.7 kb in size (Figs. 1-3). The detected variable bands appeared not as intensive as the constant ones. D i g e s t i o n with H i n c l I (Fig. 4) yielded three intensive fragments o f 2 - 2 . 3 kb in size. These signals h o w e v e r w e r e not present in all genotypes studied. P. acutifolius, P. gIabellus, P. microcarpus and P. maculatus exhibited four bands in the range of 2 . 8 - 4 kb. The fragments obtained in the analyzed Phaseolus geno-

types with DraI ( > 10 kb) and X b a I ( > 5 kb) were almost differently sized and nearly equal in intensity (Figs. 5 a + b ) . C o m m o n bands w e r e detected a m o n g the genotypes ofP. coccineus, P. polyanthus, P. acutifolius and P. lunatus, respectively.

101

dABCOEFGHI JKLMNOPQRSTr

Dwarf Scarlet, although the cultivars differ in their growth habit. Among the P. lunatus species two different RFLP patterns were detected, which were evidently characteristic for the cultivars belonging to either the Mesoamerican gene pool (var. lunatus cvgrp. Sieva and cvgrp. Potato), or the Andean gene pool (var. lunatus cvgrp. Big Lima). The cultivar P. lunatus cv. var. lunatus Jackson Wonder could be assigned to the Mesoamerican gene pool. Within P. acutifolius, the cultivated genotype var. latifolius and the wild form var. tenuifolius showed greater similarity to each other than to the wild form var. acutifolius. Construction o f a phylogenetic tree

Fig. 5. RFLPpatterns of the rDNA locus of different genotypes of Phaseolus. The DNA was digested with XbaI (Fig. 5a) and DraI (Fig. 5b). Lane r: 750 ng Raoul I marker, Lane d: 200 ng DIG III marker, Lane A: P. vulgaris cv. Tendergreen,Lane B: P. vulgaris var. aborigineus, Lane C: P. coccineus NI 0722, Lane D: P. coccineus cv. Hammond's Dwarf Scarlet, Lane E: P. coccineus cv. Desir6e, Lane F: R coccineus NI 0132, Lane G: P. polyanthus NI 0913, Lane H: P. polyanthus NI 0519, Lane I: P. acutifolius var. latifolius, Lane J: P. acutifolius var. acutifolius, Lane K: P. acutifolius var. tenuifolius, Lane L: P. glabellus, Lane M: P. microcarpus, Lane N: P. oligospermus, Lane O: P. maculatus, Lane P: R lunatus cv. Jackson Wonder, Lane Q: P. lunatus var. lunatus cvgrp. Big Lima, Lane R: P. lunatus var. lunatus cvgrp. Potato, Lane S: P. lunatus vat. lunatus cvgrp. Sieva, Lane T: P. lunatus var. silvester.

Inter- and intraspecific variation

The incidence of polymorphism between species was higher than that between varieties and cultivars of a given species. The genotypes of P. vulgaris, P. coccineus and P. p o l y a n t h u s showed great similarity in the hybridization patterns, while P. vulgaris var. aborigineus often differed from the group-specific pattern. Similarities were also revealed between the wild forms P. acutifolius vat. acutifolius, P. glabellus and P. microcarpus, especially after HincII digestion. The wild species P. maculatus and P. oligospermus displayed similarities in hybridization patterns to each other and to P. lunatus. Among P. coccineus genotypes, no differences in hybridization patterns were found between the cv. Desir6e and cv. Hammond's

The rDNA hybridization patterns obtained with 6 endonucleases in 20 genotypes revealed 160 fragment positions in the 0/1 matrix (Table 2). As the reliability of the tree depends on the data (discrete character state or distance data) and the algorithms used, several matrix programs and parsimony programs of the PHYLIP program package were applied (FITCH, MIX, KITSCH, and DOLLOP). As input file the original and multiple distance data set were used. In all obtained phylogenetic trees the genotypes belonging to the same species were grouped together. While the grouping of species was quite similar in the consensus cladograms calculated with FITCH and MIX, different groupings among the Phaseolus species were calculated with the algorithms KITSCH and DOLLOR The consensus cladogram obtained with FITCH (Fig. 6) was most congruent with our hybridization results. Furthermore the grouping of the genotypes was confirmed by the high values at most of the forks. In contrast, the values at the branchings in the consensus cladogram obtained with MIX, carrying out Wagner parsimony (Kluge & Farris, 1969), particularly that of the species grouping, were not significant. The consensus dendrogram calculated by the program KITSCH, which includes the assumption of an evolutionary clock, was not able to reflect the grouping of P. acutifolius and P. lunatus as it was evident from our hybridization results. The consensus tree calculated by the program DOLLOP, which implements the Dollo parsimony method (Farris, 1977), did not illustrate the phylogenetic relationship between the P. vulgaris genotypes and, furthermore, classified P. glabellus as an group separated from all investigated genotypes (like the program KITSCH).

102

c~

c~

0

0

i/~

0

cq

....~ ~.

v--

c-q.

0

0

II

C,

O

0

c~

103 Discussion

p. lun.B~ Lira. ~ ~

The detection of RFLPs by chemiluminescence, as introduced by Daring (1991), proved to be a fast and practical alternative tool to radioactive hybridization. Inter- and intraspecific polymorphisms within the genus Phaseolus could be detected with all six endonucleases used. The digestion with SacI, EcoRI and BamHI resulted in intensive hybridization bands, which were present in almost all Phaseolus genotypes studied (Figs. 1-3). The comparison of these hybridization data with restrictions maps of the rDNA region ofP. coccineus (Maggini et al., 1992) and other plant species (e.g. Hemleben et al., 1988; Benedetelli et al., 1992; Havey, 1992) suggests that these intensive bands contain sequences of the strongly conserved rDNA coding region. The majority of the observed polymorphisms within the rDNA were evidently located in the intergenic spacer (IGS). This interpretation corresponds with results obtained in other plants (Rogers & Bendich, 1987; Hemleben, 1990). These polymorphisms were used for the calculations of phylogenetic trees. The application of RFLP data and the choice of cladistic methods for phylogenetic constructions are still controversally discussed (e.g. Doyle, 1993; Swofford & Olsen, 1990). Therefore, we tested several distance matrix and parsimony methods and compared the results with classical taxonomic classifications. The cladogram obtained with FITCH could best illustrate the relationships of species groups, as it was evident from our hybridization results and the classical taxonomy. Furthermore the high values at the forks confirmed the grouping of the genotypes. This tree of the rDNA displays three species groups. The first group consists of investigated genotypes of the species P. vulgaris, P. coccineus and P. polyanthus, the second group comprises P. acutifolius, P. microcarpus and P. glabellus, and the third group is composed of P. lunatus, P. oligospermus and P. maculatus. This grouping is, on principle, in accordance with the morphological taxonomic classification as published by Mar6chal et al. (1978). Within the group P. vulgaris, R coccineus and P. polyanthus, the taxonomy and phylogeny of P. polyanthus is still under discussion. Investigations on morphological, cytological, and molecular levels by Mar6chal et al. (1978), Delgado Salinas (1988) and Pifiero & Eguiarte (1988) assigned P. polyanthus as a subspecies rank of P. coccineus. In contrast, Hernandez et al. (1959), Miranda-Colin (1967), and Evans (1980)

p.pdy.0519 ~ p.po~y.0913

p. lun.shnt~ p__ t00 t p. lun.pola~

,--r.

84

/ / p. m|cul.

10Q

49 74

DO

100

55 60 97

p. r / ~

Ham. p. cocr Des,

lO0 p. r 0132 p. cocC.purp.

40

p. ~ln. J

/ I ... p. micro. ~00

p.~sg.~.

P. lCULt ~ , P. ac~t I ~

Fig. 6.

Phylogenetic tree o f the r D N A for the Phaseolus g e n o types studied. The c l a d o g r a m w a s obtained with the p r o g r a m F I T C H (PHYLIP p a c k a g e version 3.5.). The values given at the forks indicate the n u m b e r o f times the b r a n c h o c c u r r e d a m o n g 100 trees. R acut. acut. = P. acut!folius var. acutifolius; E acut. lat. = P. acutifolius var. latifolius; P. acut. ten. = P. acutifolius var. tenuifolius; E coco. 0 1 3 2 = P. coccineus subsp, coceineus NI 0 1 3 2 ; E cocc. Des. = P. coccineus subsp, coccineus cv. Desir6e; E cocc. H a m . = P. coccineus subsp, coccineus cv. H a m m o n d ' s D w a r f Scarlet; E cocc. purp. = P coccineus subsp, purpurascens; E glab. = P. glabellus; E lun. Big L i m a = P. lunatus vat. lunatus cvgrp. Big Lima; E lun. Potato = P. Iunatus var. lunatus cvgrp. Potato; E Inn. Sieva = P. lunatus var. lunatus cvgrp. Sieva; E lun. J a c k s o n W o n d e r = P lunatus var. lunams cv. J a c k s o n Wonder; E lun. silvester = P. lunatus var. silvester; E mac. = P. maculatus; E micro. = P. microcarpus; E oligo. = P. oligospermus; E poly. 0 5 1 9 = P. polyanthus N 1 0 5 1 9 ; E poly. 0913 = P. polyanthus NI 0 9 1 3 ; E vulg. abor. = P. vulgaris var. aborigineus; E vulg. T. = P vulgaris cv. Tendergreen.

proposed P. polyanthus to be a hybrid between P. vulgaris and P. coccineus. This hypothesis is supported by the existence of natural hybrids between P. vulgaris and P. polyanthus, as well as between P. coccineus and P. polyanthus, while hybrids between P. vulgaris and P. coccineus were not found so far (Debouck, 1992). Furthermore, P. polyantus was described as a species rank, closely related but well differentiated from P. vulgaris and P. coccineus, on the basis of morphological, biochemical and molecular biological studies by several authors (Camarena & Baudoin, 1987; Delgado Salinas, 1988; Debouck et al., 1990; Debouck, 1991; Schmit & Debouck 1991). The hybridization results as well as the dissimilarity indexes and the cladogram support the hypothesis that P. polyanthus constitutes a separate, independent species rank. P. vulgaris var. aborigineus was envisaged as the putative ancestral wild form of the common bean (P. vulgaris L.) in South America (Berglund-Brticher & Brticher, 1976; Brticher, 1988). The rDNA hybridization patterns of var. aborigineus exhibit higher similarity to that of P. vulgaris cv. Tendergreen than to those

104 of other cultivated species. Although the cultivar P vulgaris cv. Tendergreen and the wild form P vulgaris var. P. vulgaris var. aborigineus are grouped together and evidently belong to the Andean gene pool among P. vulgaris (Gepts & Bliss, 1986; Gepts & Debouck, 1991), our cladogram suggests an early separation of these two genotypes. Among the genotypes of P. coccineus, the dissimilarity value ranged between 0 and 0.28. The index of 0 (identical RFLP patterns) between P coccineus cv. Desir6e and cv. Hammond's Dwarf Scarlet indicated a closer relationship than that between P. coccineus NI 0132 and P. coccineus subsp, purpurascens (dissimilarity value of about 0.20), as it was evident from the calculated cladogram. Furthermore, the positions of the cultivar P coccineus NI 0132 and the wild form P coccineus subsp, purpurascens among P. coccineus were not confirmed, because of the low values at the branching points. Our molecular analyses grouped P. microcarpus and P glabellus near to P acutifolius. In the cladogram, the grouping of P glabetlus and P microcarpus are uncertain, due to the low values at the forks. In contrast to our resulted grouping, investigations of Sullivan & Freytag (1986) and Baudoin & Katanga (1990) suggested that P. acutifolius is a species rather related to P coccineus and P vulgaris. Moreover, Sullivan & Freytag (1986) classified P. microcarpus as a species related to P lunatus. Especially the taxonomy of P glabellus is still unclear. Mar6chal et al. (1978) maintained P. glabellus at the species level within the P. coccineus complex, while Delgado Salinas (1988) established the taxon as a subspecies of P coccineus. Different morphological, biochemical and molecular biological characteristics of P glabellus suggested a revision of this position (Schmit & Debouck, 1990; Schmit et al., 1992). Our RFLP results support the hypothesis that P glabellus should be considered as a species for its own, well separated from the P coccineus complex. In P. lunatus var. lunatus two different RFLP patterns were found, which were characteristic for cultivars belonging to either the Mesoamerican gene pool (cvgrp. Sieva or cvgrp. Potato) or the Andean gene pool (cvgrp. Big Lima). Based on discoveries of wild forms, Bukasov (1931) stated that the center of origin and diversity of the lima beans is to be located in Central America. Findings of ancient beans in Peru (Wittmack, 1888) let Kaplan & Kaplan (1988) suppose that genotypes of the cultivar groups Sieva and Big Lima were domesticated independently but are

conspecific. The existence of two gene pools in Lima beans was also proposed by Debouck et al. (1989) and Maquet et al. (1990) based on geographical collections and on electrophoretic analyses of seed proteins. Our results correspond with this classification. Furthermore the cultivar cv. Jackson Wonder could be classified into the cultivar group Sieva within the Mesoamerican gene pool. On the basis of their rDNA-specific hybridization patterns and our cladogram, the wild species P. maculatus and P. oligospermus are closely related to R Iunatus. While taxonomical data of P. oligospermus are missing, P. maculatus was positioned near P. lunatus on the basis of morphological and pollen characteristics (Mar6chal et al., 1978; Baudoin & Katanga, 1990) as well as cross-breeding experiments (Baudoin et al., 1991). In summary, our investigations indicate that RFLPs of the rDNA are able to detect interspecific and intraspecific polymorphisms. Particularly, the IGS is well suited for evolutionary studies by means of RFLP, due to its high variation (Flavell et al., 1986; Hemleben et al., 1992). Therefore, rDNA analyses may contribute to an understanding of the systematic and phylogeny of the genus Phaseolus.

Acknowledgements The authors thank Prof. Dr. N. Blin (University of Ttibingen, Germany) for providing the subclone pTA250, T. Vanderborght (Jardin Botanique National de Belgique, Meise) for the seed samples, Mario Nenno for help with computer programs, and Simone Molter for technical assistance. Financial support by the Deutsche Forschungsgemeinschaft (DFG) is gratefully acknowledged.

References Appels, R. & R, L. Honeycutt, 1986. rDNA: Evolution overa billion years. In: S. K. Duna (Ed.), DNA systematics, Vol. 2, pp. 81-135, CRC Press, Boca Raton, USA. Baudoin, J. P. & K. Katanga, 1990. Phyletic relationships within the genus Phaseolus on basis of pollen morphology and experimental hybridization. Ann. Rep. Bean Improv. Coop. 33:117-118. Baudoin, J. P., Schmit, V. & B. Wathelet, 1991. Observation on seed protein electrophoretic patterns of some species within the genus Phaseolus. Ann. Rep. Bean Improv. Coop. 34: 85-87. Benedetelli, S., Taurchini, D. & R. D'Ovidio, 1992. Ribosomal DNA structure in Castanea spp. (Fagaceae) and their hybrids. J. Genet. Breed. 46: 57-62.

105 Berglund-B~cher, O. & H. Briicher, 1976. The South American wild bean (Phaseolus aborigineus BURK.) as ancestor of the common bean. Econ. Bot. 30: 257-272. Briicher, H., 1988. The wild ancestor of Phaseolus vulgaris in South American. In: P. Gepts (Ed.), Genetic resource of Phaseolus beans, pp. 185-214, Kluwer Academic Publishers, Dordrecht, The Netherlands. Bukasov, S. M., 1931. The cultivated plants of Mexico, Guatemala and Columbia. Bull. Appl. Bot. Genet. Plant Breed., Leningrad 47: 450-553. Camarena, M. E & J. P. Baudoin, 1987. Obtention des premiers hybrides intersptcifiques entre Phaseolus vulgaris et Phaseolus polyanthus avec le cytoplasme de cette demi~re forme. Bull. Rech. Agron. Gembloux 22: 43-45. Debouck, D. G., 1988. Phaseolus germplasm exploration. In: P. Gepts (Ed.), Genetic resources of Phaseolus beans, pp. 3-29, Kluwer Academic Publishers, Dordrecht, The Netherlands. Debouck, D. G., 1991. Systematics and morphology. In: A. van Schoonhoven & O. Voysest (Eds.), Common beans: research for crop improvement, pp. 55-118, Commonwealth Agricultural Bureaux International, Wallingford, United Kingdom. Debouck, D. G., 1992. Views on variability in Phaseolus beans. Ann. Rep. Bean Improv. Coop. 35: 9-10. Debouck, D. G., Maquet, A. & C. E. Posso, 1989. Biochemical evidence for two different gene pools in Lima beans, Phaseolus lunatus L. Ann. Rep. Bean Improv. Coop. 32: 58-59. Debouck, D. G., Schmit, V., Libreros, D. & H. Ramirez, 1990. Biochemicalevidence for a fifth cultigen within the genus Phaseolus. Ann. Rep. Bean Improv. Coop. 33: 106-107. Delgado Salinas, A., 1988. Variation, taxonomy, domestication and germplasm potentialities in P. coccineus. In: E Gepts (Ed.), Genetic resources of Phaseolusbeans, pp. 441-463, Kluwer Academic Publishers, Dordrecht, The Netherlands. Dellaporta, S. L., Wood, J. & J. B. Hicks, 1984. Maize-miniprep. DNA isolation. In: R. Malmberg, J. Messing & I. Sussex (Eds.), Molecular biology of plants - a laboratory course manual, pp. 36-37, Cold Spring Harbor Laboratory, New York, USA. Doyle, J. J., 1993. DNA, phylogeny, and the flowering of plant systematics. BioScience 43: 380-389. Dfiring, K., 1991. Ultrasensitive chemiluminescent and colorigenic detection of DNA, RNA, and proteins in plant molecular biology. Anal. Biochem. 196: 433-438. Evans, A. M., 1980. Structure, variation, evolution and classification in Phaseolus. In: R. J. Summerfield & A. H. Bunting (Eds.), Advances in legume science, pp. 337-347. Royal Botanic Gardens, Kew, England. Farris, J. S., 1977. Phylogenetic analysis under Dollo's law. Syst. Zool. 26: 77-88. Feinberg, A. P. & B. Vogelstein, 1983. A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132: 6-18. Felsenstein, J., 1988. Phylogenies from molecular sequences: inference and reliability. Annu. Rev. Genet. 22: 521-565. Felsenstein, J., I989. PHYLIP (Phylogeny Inference Package) Version 3.2. Cladistics 5: 164-166. Flavell, R. B., O'Dell, M., Sharp, E, Nevo, E. & A. Beiles, 1986. Variation in the intergenic spacer of ribosomal DNA of wild wheat Triticum dicoccoides, in Israel. Mol. Biol. Evol. 3: 547-558. Gebhardt, C. & E Salamini, 1992. Restriction fragment length polymorphism analysis of plant genomes and its application to plant breeding. Intern. Rev. Cytol. 135: 201-237. Gepts, E & E A. Bliss, 1986. Phaseolin variability among wild and cultivated common beans (Phaseolus vulgaris) in Colombia. Econ. Bot. 40: 469-478.

Gepts, P. & D. G. Debouck, 1991. Origin, domestication and evolution of the common bean (P. vulgaris L.). In: A. van Schoonhoven & O. Voysest (Eds.), Common beans: research for crop improvement, pp. 7-53, Commonwealth Agricultural Bureaux International, Wallingford, United Kingdom. Gerbi, S. A., 1985. Evolution of ribosomal RNA. In: R. J. Mclntyre (Ed.), Molecular evolutionary genetics, pp. 419-518, Plenum Press, New York, USA. Gerlach, W. L. & J. R. Bedbrook, 1979. Cloning and characterisation of ribosomal RNA genes from wheat and barley. Nucl. Acid. Res. 7: 1869-1883. Havey, M. J., 1992. Restriction enzyme analysis of the nuclear 45S ribosomal DNA of six cultivted Alliums (Alliaceae). P1. Syst. Evol. 181: 45-55. Hemleben, V., 1990. Molekularbiologie der Pflanzen, Gustav Fischer Verlag, Stuttgart, Germany. Hemleben, V., Ganal, M., Gerstner, J., Schiebel, K. & R. A. Torres, 1988. Organisation and length heterogeneity of plant ribosomal RNA genes. In: G. Kahl (Ed.), Architecture of eucaryotic genes, pp. 371-383, VCH Verlagsgesellschaft, Weinheim, Germany. Hernleben, V., Zentgraf, U., King, K., Borisjuk, N. & G. Schweizer, 1992. Middle repetitive and highly repetitive sequences detect polymorphisms in plants. In: G. Kahl, H. Appelhans, J. K6mpf& A. J. Driesel (Eds.), DNA-polymorphisms in eucaryotic genomes, pp. 157-170, Hiithig Buch Verlag, Heidelberg, Germany. Hernandez, X., Miranda-Colin, S. & C. Prwyer, 1959. E1 origin de Phaseolus coccineus L. darwinianus Hemandez X. & Miranda C. subspecies nova. Rev. Soc. Mex. Hist. Nat. 20: 99-121. Jackson, D. A., Somers, K. M. & H. H. Harvey, 1989. Similarity coefficients: measures of co-occurence and association or simply measures of occurrence. Am. Nat. 133: 436-453. Kaplan, L. & L. N. Kaptan, 1988. Phaseolus in archaeology. In: E Gepts (Ed.), Genetic resources of Phaseolus beans, pp. 125-142, Kluwer Academic Publishers, Dordrecht, The Netherlands. King, K., Torres, R. A., Zentgraf, U. & V. Hemleben, 1993. Molecular evolution of the intergenic spacer in the nuclear ribosomal RNA ofCucurbitaceae. J. Mol. Evol. 36: 144-152. Kluge, A. G. & J. S. Farris, 1969. Quantitative phyletics and evolufion of anurans. Syst. Zool. 18: 1-32. Maggini, E, Tucci, G., Demartis, A., Gelati, M. T. & S. Avanzi, 1992. Ribosomal RNA genes of Phaseolus coccineus. I. Plant Mol. Biol. 18: 1073-1082. Maquet, A., Gutierrez, A. & D. G. Debouck, 1990. Further biochemical evidence for the existence of two gene pools in Lima beans. Ann. Rep. Bean Improv. Coop. 33: 128-129. Mar6chal, R., Mascherpa, J. M. & E Stainer, 1978. t~tude taxonomique d'un groupe complexe d'esp~ces des genres Phaseolus et Vigna (Papilionaceae) sur la base de donn~es morphologiques et polliniques, traittes par l'analyse informatique. Boissiera 28: 1-278. Miranda-Colin, S., 1967. Infiltraci6n genttica entre Phaseolus coccineus L. Colegio de Postgraduados, Escuela National de Agricultura, Chapingo, M~xico. Serie de Investigaci6n 9: 1-48. Molnar, S. J., Gupta, E K., Fedak, G. & R. Wheatcroft, 1989. Ribosomal DNA repeat unit polymorphism in 25 Hordeum species. Theor. Appl. Genet. 78: 387-392. Olsen, G. J. & C. R. Woese, 1993. Ribosomal RNA: a key to phylogeny? FASEB J. 7:113-123. Pifiero, D. & L. Eguiarte, 1988. The origin and biosystematic status of Phaseolus coccineus ssp. polyanthus: electrophoretic evidence. Euphytica 37: 199-203. Rogers, S. O. & A. J. Bendich, 1987. Ribosomal RNA genes in plants: variability in copy number and in the intergenic spacer. Plant Mol. Biol. 9: 509-520.

106 Saghai-Maroof, M. A., Soliman, K. M., Jorgensen, R. A. & R. W. Allard, 1984. Ribosomal spacer-length polymorphismsin barley: Mendelian inheritance, chromosomal location, and population dynamics. Proc. Natl. Acad. Sci. USA 81: 8014-8018. Santoni, S. & A. Berville, 1992. Characterization of the nuclear ribosomal DNA units and phylogeny of Beta L. wild forms and cultivated beets. Theor. Appl. Genet. 83: 533-542. Schaal, B. A. & G. H. Learn, 1988. Ribosomal DNA variation within and among populations. Ann. Missouri Bot. Gard. 75: 1207-1216. Schmit, V. & D. G. Debouck, 1990. Phaseolus glabellus PIPER, a noteworthy variant of the P. coccineus complex? Ann. Rep. Bean Improv. Coop. 33: 124-125. Schmit, V. & D. G. Debouck, 1991. Observations on the origin of Phaseolus polyanthus Greenman. Econ. Bot. 45: 345-364. Schmit, V., Du Jardin, P., Baudoin, J. P. & D. G. Debouck, 1992. Diversity studies of some Phaseolus taxa using chloroplast DNA as molecular marker. Ann. Rep. Bean Improv. Coop. 35: 213214.

Smith, G. E. & M. D. Summers, 1980. The bidirectional transfer of DNA and RNA to nitrocellulose or diazobenzyloxymethyl-paper. Anal. Biochem. 109: 123-129. Sullivan, J. G. & G. Freytag, 1986. Predicting interspecific compatibilities in beans (Phaseolus) by seed protein electrophoresis. Euphytica 35: 201-209. Swofford, D. L. & G. J. Olsen, 1990. Phylogeny reconstruction. In: D. M. Hillis & C. Moritz (Eds.), Molecular systematics, pp. 41-501, Sinauer Associates, Inc. Publishers, Sunderland, Massachusetts, USA. Wittmack, L., 1988. Die Heimath der Bohnen und der Kiirbisse. Ber. Deutsch. Bot. Ges. 6: 374-380. Young, N. D., 1992. Restriction fragment length polymorphisms (RFLPs) and crop improvement. Exp. Agricult. 28: 385-397. Zimmer, E. A., Jupe, E. R. & V. Walbot, 1988. Ribosomal gene structure, variation, and its inheritance in maize and its ancestors. Genetics 120:1125-1136.