Genetic diversity and conservation of common bean (Phaseolus vulgaris L., Fabaceae) landraces in Nilgiris

Abstract

Genetic diversity was studied among 20 common bean (Phaseolus vulgaris L.) landraces collected from different traditional farming villages of Nilgiris District, Tamil Nadu, India, with Random Amplified Polymorphic DNA (RAPD) markers. Evaluation of genetic diversity is essential for conservation, management and to trace the hybrids. Thirteen RAPD primers were selected from an initial screening with 72 primers. The PCR product revealed 102 bands, out of which 63 were found to be polymorphic (63.5%). Jaccard's pair-wise similarity coefficient (0.50 to 0.95) indicating an intra-specific genetic variation prevails in landraces of common bean in the Nilgiris biosphere reserve. No two accessions had a similarity of one or a distance of zero, showing that there were no duplicate entries. A dendrogram of the relationship of accessions constructed based on Jaccard's coefficients of 102 RAPD markers using the average distance method (UPGMA), separated the accessions into two major clusters, A and B, with Mesoamerican and Andean gene pools respectively. Principal coordinate analysis of the same dataset revealed similar results as those of the dendrogram, with the first two components accounting for 61.8% of the total variation. Among the 20 landraces, seven were Mesoamerican origin, 11 Andean origin, and two were possible recombinants between the two gene pools. A correlation was observed between RAPD dendrogram clustering and seed weight. The common bean population of the Nilgiris is highly diverse and the Nilgiris can be considered as a secondary centre of genetic diversity of common bean. A better knowledge of genetic aspects of common bean will help in genetic improvement and conservation programmes for its endangered landraces in the Nilgiris.

Full text

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
227
*For correspondence. (e-mail: fcjooty@yahoo.com)
Genetic diversity and conservation of common
bean (Phaseolus vulgaris L., Fabaceae)
landraces in Nilgiris
Franklin Charles Jose
1,
*, M. M. Sudheer Mohammed
2
, George Thomas
3
,
George Varghese
3
, N. Selvaraj
4
and M. Dorai
1
1
Department of Plant Biology and Biotechnology, Government Arts College, Stone House Hill P.O., Udhagamandalam 643 002, India
2
Department of Plant Biology and Biotechnology, Government Arts College (Autonomous), Coimbatore 641 018, India
3
Department of Plant Molecular Biology, Rajiv Gandhi Centre for Biotechnology, Thiruvananthapuram 695 014, India
4
Horticultural Research Station, Tamil Nadu Agricultural University, Centenary Rose Park, Udhagamandalam 643 001, India
Genetic diversity was studied among 20 common bean
(Phaseolus vulgaris L.) landraces collected from dif-
ferent traditional farming villages of Nilgiris District,
Tamil Nadu, India, with Random Amplified Polymor-
phic DNA (RAPD) markers. Evaluation of genetic di-
versity is essential for conservation, management and
to trace the hybrids. Thirteen RAPD primers were sele-
cted from an initial screening with 72 primers. The
PCR product revealed 102 bands, out of which 63
were found to be polymorphic (63.5%). Jaccard’s
pair-wise similarity coefficient (0.50 to 0.95) indicating
an intra-specific genetic variation prevails in landraces
of common bean in the Nilgiris biosphere reserve. No
two accessions had a similarity of one or a distance of
zero, showing that there were no duplicate entries. A
dendrogram of the relationship of accessions con-
structed based on Jaccard’s coefficients of 102 RAPD
markers using the average distance method (UPGMA),
separated the accessions into two major clusters, A
and B, with Mesoamerican and Andean gene pools
respectively. Principal coordinate analysis of the same
dataset revealed similar results as those of the dendro-
gram, with the first two components accounting for
61.8% of the total variation. Among the 20 landraces,
seven were Mesoamerican origin, 11 Andean origin, and
two were possible recombinants between the two gene
pools. A correlation was observed between RAPD
dendrogram clustering and seed weight. The common
bean population of the Nilgiris is highly diverse and
the Nilgiris can be considered as a secondary centre of
genetic diversity of common bean. A better knowledge
of genetic aspects of common bean will help in genetic
improvement and conservation programmes for its
endangered landraces in the Nilgiris.
Keywords: Genetic variation, germplasm conservation,
multivariate analysis, Phaseolus vulgaris, RAPD markers.
T
HE common bean (Phaseolus vulgaris L.) was intro-
duced from the Americas into the Nilgiris approximately
400 years ago. It is the most important legume worldwide
for direct human consumption and has the broadest range
of genetic resources
1
. The Blue Mountains or Nilgiri hills
in Tamil Nadu, India, lies between 11°12′N and 11°43′N,
76°14′E and 77°1′E. The region is a veritable paradise
because of the rich and diversified geographical condi-
tions, flora and fauna which have played a pioneering
role in the introduction and cultivation of plants since
colonial times
2
. This region is highly species-rich and is
considered one of the 25 biodiversity hotspots of the
world
3
.
The importance of common bean landraces in Nilgiris
agriculture cannot be neglected. Most of these varieties
are resistant to anthracnose, bean common mosaic virus
and rust, can be grown even if there is scarcity of water.
Most of the landraces are tall plants, and produce pods in
long periods than the short duration commercial cultivars.
The seeds of these landraces are used as sowing material
by the traditional small-scale farmers, who are conserving
these landraces for centuries. The immense genetic diver-
sity of landraces of crops is the most directly useful and
economically valuable part of biodiversity. Unlike high-
yielding varieties, the landraces maintained by farmers
are endowed with tremendous genetic variability, as they
are not subjected to subtle selection over a long period.
Because of the limitations of morphological and bio-
chemical markers, efforts are being directed to use mole-
cular markers for characterizing germplasm diversity.
Molecular markers have demonstrated a potential to
detect genetic diversity and to aid in the management of
plant genetic resources
4,5
. In contrast to morphological
traits, molecular markers can reveal differences among
genotypes at the DNA level, providing a more direct,
reliable and efficient tool for germplasm characterization,
conservation and management. Several types of mole-
cular markers are available today, including those based
on restriction fragment length polymorphism (RFLP)
6
,
random amplified polymorphic DNA (RAPD)
7,8
, ampli-
fied fragment length polymorphism (AFLP)
9
and simple-
sequence repeats (SSRs)
10
.

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
228
In the present study, we have used RAPD method of
DNA fingerprinting, which is widely used in conservation
biology because of quick results, cost-effectiveness and
reproducibility. The PCR-based RAPD approach using
arbitrary primers requires only nanogram quantities of
template DNA, no radioactive probes, and is relatively
simple compared to other techniques
11
. RAPD is a useful
predictive tool to identify areas of maximum diversity
and may be used to estimate levels of genetic variability
in natural populations. Morphological, biochemical and
molecular analyses have been suggested for the evaluation
of genetic diversity in common bean from America
12–16
,
Galicia
17
, Central Himalaya
18
, Italy
19
and China
20
.
So far, no systematic studies have been conducted to
assess the morphological and genetic variation in the
Nilgiris common bean landraces, because of unavailabi-
lity in outside markets. Tribal agricultural farmers mainly
use these beans to prepare traditional food items for their
own use. The genetic discrimination of an individual is an
important step in investigating the population biology of
any species and a major contribution that conservation
geneticists can make for evaluating population viability.
On the basis of the above, the present study was under-
taken to investigate genetic diversity and origin of P. vul-
garis L. landraces cultivated by tribal agricultural farmers
in the Nilgiris.
Materials and methods
Sample collection
The present study included 20 landraces of common bean
collected from different villages of the Nilgiris. The vil-
lages from where the seeds were collected, its common
name, phesolin type, growth habit, 100 seed weight and
altitude of the region above mean sea level were recorded
(Table 1). Two landraces (LR2 and LR9) were obtained
from the Regional Horticultural Research Station (RHRS)
of Tamil Nadu Agricultural University (TNAU). The
genotypes were selected for genetic analysis after thor-
ough morphological evaluation in the farmer’s field, ac-
cording to the IBPGR descriptor list for P. vulgaris
L.
21
.
Plant height, seed coat colour and seed weight (expressed
in g/100 seeds) were used as the primary indicators of
morphological variation because of the marked difference
in these characters between different landraces. Twenty-
five seeds from each landrace were collected and germi-
nated under laboratory conditions. The healthy, young
leaves collected from these samples were used for RAPD
analysis.
DNA isolation and primer screening
DNA was extracted from young, tender leaves collected
separately from each individual and stored at –70°C in
sealed polythene bags. Genomic DNA was isolated from
100 mg of the tender leaf tissues using Gen Elute Plant
Genomic DNA Purification Kit (Sigma) following the
manufacturer’s instructions. About 2 μl of genomic DNA
isolated from 100 mg of leaf tissue was subjected to elec-
trophoresis on a 0.8% agarose gel containing 500 ng/μl of
ethidium bromide. After electrophoresis, the gel was
viewed over a UV transilluminator and the quality and
quantity of DNA was assessed using undigested λ DNA
as control. Only extracts without RNA smears on agarose
gel and with UV-light absorption ratios (A260/A280)
between 1.8 and 2.0 were used. The genomic DNA was
diluted to 4 ng/μl and stored at 4°C as working solution,
while the stock DNA (undiluted) was stored at –20°C in
aliquots. Initial screening was done with 72 primers
(Operon Technologies Inc., CA, USA) using DNA from
five randomly selected landraces. PCR–RAPD analysis
was repeated at least twice and only primers producing
strong and reproducible bands were used in the final
analysis of all the 20 landraces.
Polymerase chain reaction
Polymerase chain reaction (PCR) was carried out in a
20 μl reaction volume containing 28 ng of genomic DNA,
1 unit of Taq DNA polymerase (Bangalore Genei), 4 μl
primer, 0.2 mM dNTPs, 10 mM Tris-HCl (pH 9.0),
1.5 mM MgCl
2
and 50 mM KCl. Amplification was car-
ried out in a thermocycler (Eppendorf) with an initial
strand separation at 94°C for 4 min, followed by 40
cycles of 1 min at 94°C, 1 min at 37°C and 1.5 min at
72°C. After 40 cycles, there was a final extension step of
5 min at 72°C. Amplification products were resolved on
1.2% agarose gels in 1% TAE buffer at 70 V. The gels
were photographed using a gel documentation system
(SynGene). Size of the amplified products was compared
to λ DNA/EcoRI double-digest marker (Bangalore Genei).
Data analysis
Each amplified RAPD marker was treated as a unit char-
acter and was scored as present (1) or absent (0) for each
sample. Ambiguous bands that could not be clearly dis-
tinguished were not scored. The percentage of polymor-
phism was calculated as the proportion of polymorphic
bands over the total number of bands. The 1/0 matrix was
prepared for all fragments scored and the data were used
to generate genetic similarity (GS) based on Jaccard’s
coefficient of similarity as follows: GS(ij) = a/(a + b + c),
where GS(ij) is the measure of genetic similarity between
individuals i and j, a is the number of polymorphic bands
that are shared by i and j, b is the number of bands pre-
sent in i and absent in j, and c is the number of bands pre-
sent in j and absent in i. Jaccard’s coefficients were used
to construct a dendrogram using UPGMA. The Jaccard’s

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
229
Table 1. Phaseolus vulgaris L. landraces in the Nilgiris
Village/place Altitude Growth habit Seed weight
Accession no. Market class of collection m amsl (I, III, IV) (g/100 seeds)
LR1 Cranberry Nanjanadu 2376 I 39.9
LR2 Black turtle RHRS, TNAU*
2376 I 16.5
LR3 Cranberry big Adigaratti 2261 I 53.0
LR4 Pale brown kidney Tumanatti 1920 I 36.7
LR5 White kidney Nanjanadu 2376 I 32.8
LR6 Red kidney Ozahatti 2261 I 46.0
LR7 Black bean Nanjanadu 2376 IV 32.3
LR8 Great northern Ebanad 1930 III 34.3
LR9 White eyed bean RHRS, TNAU*
2376 IV 23.7
LR10 Greenish striped kidney Nundala 2180 IV 67.2
LR11 Ash bean Nanjanadu 2376 IV 66.9
LR12 Large pinto Tuneri 2066 IV 66.0
LR13 Pinto bean Muttinadu 2200 IV 66.1
LR14 Pink bean Sholur 2277 IV 31.8
LR15 Black striped Wellington 2041 IV 68.2
LR16 White round Bikkatti 2532 IV 69.4
LR17 Dark brown Nanjanadu 2376 I 30.8
LR18 Brown kidney Kokkal 2270 IV 63.5
LR19 Dark brown Ebanad 2060 IV 36.6
LR20 Small kidney Muttinadu 2200 III 23.5
*Regional Horticultural Research Station, Tamil Nadu Agricultural University, Udhagamandalam.
similarity matrix was then used as the basis for ordination
by principal coordinates analysis (PCoA), which was
performed to show the distribution of the genotypes in a
scatter plot using the Multivariate Statistical Package
version 3 software.
Results
DNA fingerprinting
DNA extracted from 20 common bean landraces was
examined for the PCR–RAPD patterns. Out of the 72
primers screened, 13 were selected based on robustness
of amplification, reproducibility, scorability of banding
patterns and were used for diversity analysis in all land-
races. The 13 selected decamer oligonucleotide primers
(Table 2) generated 102 amplification products, out of
which 63 bands (63.5%) were polymorphic. The number
of bands per primer ranged from 4 (OPE19) to 13 (OPJ9),
with an average of 7.8 bands per primer. The range of
polymorphic bands per primer was 3 (OPA1, OPA4 and
OPF16) to 7 (OPE6, OPJ9 and OPJ16), with a mean of
4.8 polymeric bands per primer (Table 2). The represen-
tative RAPD patterns generated by primers OPE6, OPA4
and OPE19 for 20 landraces are given in Figure 1. These
three primers are efficient in discriminating landraces
into Andean and Mesoamerican races.
Genetic similarity, cluster analysis and PCoA
The pair-wise Jaccard’s coefficients genetic similarity
matrix was prepared based on RAPD data. The genetic
similarity coefficients among the common bean landraces
varied from 0.50 (between varieties LR3 and LR9) to
0.95 (between varieties LR1 and LR16; Table 3). Cluster
analysis was performed on RAPD data using UPGMA,
which showed overall genetic relationships among the
landraces of common bean (Figure 2). PCoA was carried
out in order to determine the genetic relationships among
the landraces. The landraces were plotted on principal
coordinates 1 and 2, accounting for 53.0 and 8.8% of the
variation respectively, and together explaining 61.8% of
the total variation (Figure 3). UPGMA clustering and
PCoA of RAPD data indicated that the landraces of P.
vulgaris population comprise two major groups, clusters
M and A (Figure 2). The similarity coefficients ranged
from 0.50 to 0.95, indicating that the two clusters did not
show 100% similarity.
Mesoamerican, Andean and hybrid gene pools
Cluster ‘M’ comprised of seven landraces (LR2, LR7,
LR8, LR9, LR14, LR19 and LR20) which are of Meso-
american origin. Cluster ‘A’ comprised of 11 landraces
(LR1, LR16, LR12, LR13, LR18, LR3, LR11, LR10, LR4,
LR15 and LR6) which are of Andean origin. Cluster ‘H’
consists of two landraces (LR5 and LR17) which are hy-
brids. Maximum similarities were observed between LR8
and LR9, with a similarity value of 0.89. LR20 and LR19
formed a separate cluster within cluster M, with a similar-
ity value of 0.81. LR2 formed an operational taxonomic
unit in cluster M. The highest similarity (0.95) was ob-
served among LR1, LR16, LR12 and LR13. Cluster analy-
sis and PCoA of the similarity indices (Figures 2 and 3)

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
230
Table 2. Primers with their sequences used for RAPD analysis of Phaseolus vulgaris L. landraces in the Nilgiris. Total
number of bands, polymorphic bands, and percentage of polymorphism yielded by each primer are also listed
No. of Percentage of
Primer code Primer sequence (5′ → 3′) Total no. of bands polymorphic bands polymorphism
OPA1 CAGGCCCTTC 7 3 42.8
OPA4 AATCGGGCTG 7 3 42.8
OPE1 CCCAAGGTCC 8 4 50.0
OPE6 AAGACCCCTC 10 7 70.0
OPE12 TTATCGCCCC 8 5 62.5
OPE19 ACGGCGTATG 4 4 100
OPF6 GGGAATTCGG 10 6 60
OPF16 GGAGTACTGG 5 3 60
OPJ9 TGAGCCTCAC 13 7 53.8
OPJ16 CTGCTTAGGG 8 7 87.5
OPAC12 GGCGAGTGTG 6 5 83.3
OPAO8 ACTGGCTCTC 9 5 55.5
OPAR1 CCATTCCGAG 7 4 57.1
Total 102 63 63.5
Mean per primer 7.8 4.8
Figure 1. RAPD profile of 20 landraces of Phaseolus vulgaris L. pro-
duced using the random decamer primers OPE6, OPA4 and OPE19. M,
100 bp DNA ladder. Lane numbers correspond to the landtraces LR1 to
LR20 given in Table 1.
support the above results. In the RAPD dendrogram (Fig-
ure 2) two landraces, LR5 and LR17, formed a separate
group in the Andean gene pool, very closely associated
with the Mesoamerican gene pool. This may be a natural
hybrid gene pool between two major gene pools
22,23
. In
PCoA, the hybrid landraces (LR5 and LR17) formed a
distinct group different from the Mesoamerican and An-
dean gene pools (Figure 3).
RAPD clustering and seed weight
A correlation was observed between seed weight, RAPD
clustering and PCoA of RAPD data. The average 100
seed weight of Mesoamerican landraces ranged from 16.5
to 36.6 g and that of the Andean gene pool ranged from
36.7 to 69.4 g. Therefore, there was a fine separation of
the major gene pools with regard to seed weight, without
any overlapping between them. The hybrid gene pool
comprising two landraces, LR5 and LR17, with 100 seed
weight (32.8 and 30.8 g respectively) and phaseolin types
(‘S’ and ‘T’ respectively) was dwarf plants with kidney-
shaped seeds. Similarly, LR2, the landrace with the low-
est seed weight formed a distinct taxonomical unit in the
Mesoamerican group. Thus, a correlation was observed
between RAPD clustering and seed weight in common
bean landraces. This indicates that RAPD markers are
well suited to determine the genetic diversity and differ-
entiation present in common bean landraces in the
Nilgiris.
Discussion
The present study addresses the utility of RAPD markers
in revealing genetic relationships at the molecular level
among landraces of common bean in the Nilgiris. Sources
of polymorphism in the RAPD assay may be due to base
change within the priming site sequence, deletions of

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
231
Table 3. Similarity matrix for Jaccard’s coefficients for 20 landraces (LR1-20) of Phaseolus vulgaris based on 102 bands obtained with 13 RAPD primers
LR1 LR2 LR3 LR4 LR5 LR6 LR7 LR8 LR9 LR10 LR11 LR12 LR13 LR14 LR15 LR16 LR17 LR18 LR19 LR20
LR1 1.00
LR2 0.55 1.00
LR3 0.89 0.56 1.00
LR4 0.90 0.56 0.84 1.00
LR5 0.71 0.70 0.68 0.70 1.00
LR6 0.90 0.58 0.89 0.88 0.73 1.00
LR7 0.53 0.81 0.54 0.53 0.65 0.53 1.00
LR8 0.54 0.80 0.56 0.53 0.64 0.52 0.90 1.00
LR9 0.53 0.83 0.50 0.52 0.63 0.53 0.88 0.91 1.00
LR10 0.84 0.57 0.90 0.84 0.71 0.89 0.56 0.55 0.53 1.00
LR11 0.88 0.57 0.91 0.83 0.69 0.88 0.55 0.58 0.53 0.89 1.00
LR12 0.87 0.58 0.86 0.82 0.65 0.90 0.54 0.52 0.52 0.86 0.86 1.00
LR13 0.92 0.56 0.89 0.87 0.69 0.93 0.54 0.54 0.53 0.89 0.88 0.95 1.00
LR14 0.58 0.82 0.55 0.57 0.67 0.57 0.85 0.88 0.89 0.57 0.59 0.55 0.57 1.00
LR15 0.84 0.57 0.83 0.77 0.64 0.86 0.51 0.52 0.53 0.80 0.84 0.88 0.84 0.55 1.00
LR16 0.95 0.53 0.91 0.87 0.68 0.90 0.52 0.55 0.51 0.86 0.88 0.87 0.92 0.55 0.84 1.00
LR17 0.76 0.68 0.73 0.73 0.79 0.74 0.66 0.68 0.66 0.71 0.74 0.68 0.70 0.68 0.67 0.74 1.00
LR18 0.89 0.56 0.86 0.82 0.68 0.87 0.57 0.58 0.58 0.85 0.85 0.86 0.91 0.58 0.83 0.89 0.75 1.00
LR19 0.60 0.79 0.57 0.59 0.72 0.58 0.77 0.82 0.81 0.56 0.60 0.54 0.55 0.80 0.56 0.57 0.75 0.59 1.00
LR20 0.54 0.80 0.54 0.52 0.65 0.53 0.78 0.79 0.80 0.53 0.54 0.52 0.51 0.77 0.53 0.55 0.68 0.55 0.82 1.00

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
232
priming site, insertions that render priming sites too dis-
tant to support amplification, and deletions or insertions
that change the size of the DNA fragments which act to
prevent its amplification
8
. RAPD markers were able to
distinguish groups within both the Andean and Meso-
american gene pools
14,15
. The polymorphism revealed by
RAPD has been problematic due to their dominance. As
heterozygotes are not normally detectable, the results are
not readily usable for computing Hardy–Weinberg equi-
librium or Nei’s standard genetic distance
24
. Therefore in
this study RAPD polymorphisms were analysed with a
phonetic distance measure (Jaccard’s coefficient) from
which a dendrogram was constructed (Figure 2), provid-
ing an indication of the diversity present within the land-
races of common bean.
Mesoamerican group
This group is represented by 35% of the total population,
with seven landraces (LR2, LR7, LR8, LR9, LR14, LR19
and LR20) in cluster M (Figure 2). In this group, one
dwarf (LR2), two tall non-woody (LR20 and LR8) and
four tall woody types (LR19, LR14, LR9 and LR7) were
described. The average weight of 100 seeds ranged from
16.5 to 36.6 g. As suggested by Evans
25,26
, the Meso-
american gene pool has both small (<25 g per 100 seed
weight) and medium (25–40 g per 100 seed weight)
seeds. The lowest seed weight (16.5 g per 100 seeds) was
observed in LR2. Zizumbo-Villarreal et al.
27
reported that
the average weedy seed weight ranged from 19 to
21 g/100 seeds and that of wild seeds ranged from 4 to
7 g/100 seeds. Moreover, this variety is less common in
cultivation and least preferred by farmers due to its small
black seeds. In the dendrogram cluster (Figure 2), this
formed an operational taxonomic unit very distant from
other Mesoamerican races, probably due to its unique
agro-morphological characteristics and the presence of
Figure 2. Genetic similarity dendrogram based on 102 RAPD mark-
ers in 20 common bean landtraces (LR1–LR20). M, Mesoamerican; H,
Hybrid; A, Andean.
some rare alleles. Such RAPD cluster with only one
accession in common bean landraces is common
14,15,18
.
DNA analysis with RAPD markers confirmed the exi-
stence of these three races among the climbing beans of
Mesoamerican origin in Guatemala and neighbouring
countries
15
.
As evident from the dendrogram (Figure 2), the Meso-
american landraces in the Nilgiris are highly diverse with
a similarity coefficient lower compared to their Andean
counterpart. It has been calculated that more than 60% of
the world production derives from domesticates of Meso-
american origin
15
. Nevertheless, in the Nilgiris it repre-
sents only 35% of the total population and is used as both
snap and dry beans. One reason might be the preference
of farmers for the large-seeded Andean beans, at least for
dry beans, over the small-seeded Mesoamerican beans.
As a result, the distribution and use of Mesoamerican
beans are less in the Nilgiris compared to the Andean
races. Moreover, in recent years breeding activities have
focused on only a few highly priced, large-seeded market
classes that could have contributed to a reduction in ge-
netic diversity within the small-seeded, less desirable cul-
tivars. Regarding the landraces of common bean in the
Nilgiris, it is not true that farmers always select varieties
with big seeds; instead, they select plants for their special
requirements. For example, LR20 is small-seeded and
widely grown for its unique uses in certain recipes and
good market value. In spite of these hurdles, the Meso-
american germplasm in the Nilgiris is genetically more
diverse than the Andean race, which is evident from the
distribution of its landraces in the RAPD dendrogram
(Figure 2).
Andean group
This group consists of landraces in cluster B. This cluster
is constituted by 11 landraces (55% of the total), out of
which four are dwarf and seven are tall (Figure 2 and
Table 1). In this cluster, two dwarf varieties (LR5 and
LR17) formed a sub-cluster with a similarity value of
0.79 (Figure 2) and are considered to be hybrids between
the Mesoamerican and Andean gene pools. The Andean
race is characterized by large seeds, >40 g/100 seed
weight
24,25
and T type phaseolin
28
. The average 100 seed
weight in Andean race ranged from 36.7 to 69.4 g (Table
1). The T type phaseolin was significantly correlated with
higher seed weight, increased overall seed protein and an
increase in the percentage of phaseolin compared to S
type phaseolin observed in Italian landraces
29
. A correla-
tion between plant height and seed weight in the Nilgiris
Andean gene pool was evident. The four dwarf plants had
seed weight <60.0 g, while all the tall landraces had
>60.0 g (Table 1).
In the Nilgiris, the majority of common bean under
cultivation was from the Andean gene pool. The same

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
233
Figure 3. Principal coordinate analysis of 20 landraces of common bean (LR1–LR20) based on Jaccard’s similarity coefficients.
phenomenon was noticed in other common bean-growing
regions such as China
20
, Italy
29
and Argentina
30
. The high
preference for Andean landraces may be due to some de-
sirable traits such as taste, growth habit, seed size and
colour. In the Nilgiris, LR4 is the major snap bean and
LR11 is the most preferred and highly priced dry bean.
This tall bean with ash coloured seed is widely cultivated
in almost all the villages. Tribal farmers prepare a special
dish out of these beans during special functions and festi-
vals. However, RAPD analysis revealed that this landrace
is closely associated with LR3 (Figure 2), a small-seeded
dwarf plant commonly used as dry beans.
The genetic diversity of the Andean race is a matter of
controversy. Many researchers are of the opinion that in
spite of vast diversity in plant and grain morphology,
generally Andean landraces of common bean proved to
have a very narrow genetic base. Beebe et al.
31
revealed a
narrow genetic base for the Andean race based on AFLP
analysis and the need for widening the genetic basis of
cultivated Andean landraces was suggested
13,32,33
. Our
findings in the Nilgiri beans throw light upon the fact that
the Andean races are genetically less diverse than the
Mesoamerican gene landraces (Figures 1–3). However,
the hybrid gene pool (LR5 and LR17) contributes signifi-
cantly to the genetic diversity of the Andean beans. Con-
trary to these findings, in a study in Argentina
30
, very
high diversity in four complex primitive races of Andean
beans was observed. In China, one of the secondary cen-
tres of common bean diversity, Xiaoyan et al.
20
found
that the level of diversity for Chinese landraces of
Andean origin was higher than that of Chinese landraces
of Mesoamerican origin due to the presence of more
infrequent alleles.
In this context, broadening the genetic base of the
Andean landraces must be considered. Previous studies
revealed that the Andean cultivars are more difficult to
improve
34
. The introgression of additional genetic diver-
sity into the Andean domesticated gene pool may acquire
added importance in the light of genetic bottlenecks
induced by domestication in common bean
32
. One solu-
tion to this problem is the use of natural inter-gene pool
hybrids (LR5 and LR17) for breeding purposes.
Inter-gene pool hybrids
Landraces that possess phaseolin typical of one gene pool
and many morphological and allozyme traits of the other
gene pool are classified as inter-gene pool recombi-
nants
22
. Two accessions, LR5 and LR17, formed a sepa-
rate group in PCoA (Figure 3). Both have 100 seed
weight 32.8 g and 30.8 g respectively, a characteristic
feature of the Mesoamerican race. The lack of correlation
between phaseolin type and RAPD data in these two
landraces might be due to a genetic recombination
between the Andean and the Mesoamerican gene pools. It
represents inter-gene pool introgression that occurred
naturally and selected by farmers in the Nilgiris. The
occurrence of recombinants in the Mesoamerican gene
pool was revealed
27
and an introgression of genes from
the Andean race was observed in Mexican landraces
when analysed with AFLP markers
15,23
. Rodino et al.
22,35
observed inter-gene pool recombinants in a core collec-
tion of common bean landraces from the Iberian penin-
sula, and evaluated them using morphological, agronomic
and biochemical markers. Xiaoyan et al.
20
identified
Mesoamerican introgression in Chinese common bean
landraces when assessed with SSR markers.
Common bean is generally considered an autogamous
species, but out-crossing rates as high as 60–70% have
been observed
36
. Francisco et al.
37
reported exceptionally
high rates of out-crossing in some common bean cultivars
in California. In addition, it seems that even the lowest

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
234
rates of out-crossing reported are sufficient to generate
broad variability over hundreds or thousands of years
33
.
The inter-gene pool hybrids in the Nilgiris could have
been the result of a high frequency of pollinating insects
such as honeybees, short distances between plants and co-
cultivation of landraces from different gene pools
27
. Cul-
tivation of a mixture of Andean and Middle American
beans in close proximity in home gardens combined with
occasional out-crossing may have further facilitated
introgression between the two gene pools. On the other
hand, the genetic variation observed in these landraces
might have resulted during the long cultivation history of
the species (500 years), as an adaptation to the local agro-
climatic conditions. Once these adaptive variations are
fixed in the genotypes, subsequently they could have
been passed onto the next generation. In the end, these
could have resulted in locally adapted genotypes.
We cannot neglect the significance of this hybrid gene
pool in common bean improvement in the Nilgiris. The
genetically narrow Andean races can improve by crossing
with Mesoamerican germplasm. Nevertheless, such crosses
often produce poor progeny due to hybrid weakness
38,39
.
In this context, one alternative is to use landraces that
display the introgression of Mesoamerican genes, such as
those forming a sub-cluster in the Andean gene pool
(Figures 2 and 3). These inter-gene pool recombinants
may be of interest to breeders and geneticists because
they could constitute bridging germplasm that may aid in
the transfer of useful genes between the two gene pools.
Thus, these accessions may merit further studies. These
two landraces are dwarf, with small seeds most preferred
by farmers for their white seeds and snap beans respec-
tively. Therefore, the Nilgiris’ common bean germplasm
is more complex and contains additional diversity that
remains to be explored for genetic and breeding purposes.
RAPD and seed weight
Our study shows a correlation between RAPD clustering
and 100 seed weight (Table 1 and Figure 2). The landraces
varied in terms of seed size from a high of 69.4 g per 100
seed to a low of 16.5 g per 100 seed. A correlation bet-
ween RAPD banding pattern and seed size was observed
in Italian common bean landraces
29
. However, no correla-
tion was observed between RAPD branching pattern and
morphological data for landraces of common bean col-
lected from Central Himalaya
18
. In our study, with regard
to seed weight, no overlapping was observed between the
gene pools. Similarly, the small-seeded intergene pool
hybrids (LR5 and LR17) with seed weight 32.8 g per 100
seed and 30.8 g per 100 seed formed a sub-cluster in the
large-seeded Andean genepool. Duarte et al.
40
found a
close association between phenology and genetic analysis,
and suggested the loci that control molecular and morpho-
logical characteristics are closely associated.
Conclusion
These results indicate that the level of genetic variation
has not eroded since the introduction of the common bean
from the American centres of domestication into the
Nilgiris. The level of polymorphism observed in the pre-
sent study was moderately high, indicating a wide and
diverse genetic base for the common bean landraces in
the Nilgiris. The 63.5% RAPD polymorphic bands and
67% average genetic differentiation coefficient suggest
that the P. vulgaris landraces maintain a higher intra-
specific genetic diversity in the Nilgiris. There exist clear-
cut signs of introgression between the two gene pools as
indicated by the hybrids, which merit further investiga-
tion. The occurrence of both the gene pools in the Nilgiris
is highly significant for further breeding strategies like
interracial hybridization and production of ideal geno-
types of common bean. Such inter-gene pool and inter-
racial crosses will also facilitate broadening the genetic
base of cultivars and maximizing gains from selection for
plant type, adaptation, yield and resistance to high and
low temperature. The predominance of large-seeded
Andean germplasm in the Nilgiris is due to the preference
given by farmers and consumers over the small and
medium seeded Mesoamerican varieties. The present-day
common bean germplasm in the Nilgiris might be the last
remains of a vast introduction during colonial times.
Moreover, the common bean landraces in the Nilgiris
face serious threat of extinction with the introduction of
short-duration commercial cultivars. A study of molecu-
lar profiles to obtain DNA fingerprints in order to estab-
lish the molecular identity of common bean in the
Nilgiris for documenting the germplasm seems a worth-
while endeavour. Such DNA fingerprinting pattern would
also help monitor genetic stability in the common bean.
Our understanding of the genetic diversity of the P. vul-
garis population can contribute valuable guidelines for
conservation strategies and should therefore be an essen-
tial part of proper conservation management.
1. Singh, S. P., Improvement of small-seeded race Mesoamerican
cultivars. In Common Bean Improvement in the Twenty-First Cen-
tury (ed. Singh, S. P.), Kluwer, Dordrecht, 1999, pp. 225–274.
2. Hastings, K. B., The relationship between the Indian Botanic Gar-
den, Howrah and the Royal Botanic Gardens, Kew. Econ. Bot.
Bull. Bot. Surv. India, 1988, 28, 1–12.
3. Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G.
A. B. and Kent, J., Biodiversity hotspots for conservation priori-
ties. Nature, 2000, 403, 853–858.
4. Virk, P. S., Newbury, H. J., Jackson, M. T. and Ford-Lloyd, B. V.,
Are mapped or anonymous markers more useful for assessing
genetic diversity? Theor. Appl. Genet., 2000, 100, 607–613.
5. Song, Z. P., Xu, X., Wang, B., Chen, J. K. and Lu, B. R., Genetic
diversity in the northernmost Oryza rufipogon populations estimated
by SSR markers. Theor. Appl. Genet., 2003, 107, 1492–1499.
6. Botstein, D., White, R. L., Skolnick, M. and Davis, R. W., Construc-
tion of a genetic linkage map in man using restriction fragment length
polymorphism. Am. J. Hum. Genet., 1980, 32, 314–331.

RESEARCH ARTICLES
CURRENT SCIENCE, VOL. 97, NO. 2, 25 JULY 2009
235
7. Welsh, J. and McClelland, M., Fingerprinting genomes using PCR
with arbitrary primers. Nucleic Acids Res., 1990, 18, 7213–7218.
8. Williams, J. G. K., Kubelik, A. R., Livak, K. J., Rafalski, J. A. and
Tingey, S. V., DNA polymorphism amplified by arbitrary primers
as useful genetic markers. Nucleic Acids Res., 1990, 18, 6531–
6535.
9. Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T. and
Hornes, M., AFLP: a new technique for DNA fingerprinting.
Nucleic Acids Res., 1995, 23, 4407–4414.
10. Tautz, D., Hypervariability of simple sequences as a general
source for polymorphic DNA markers. Nucleic Acids Res., 1989,
17, 6463–6471.
11. Williams, J. G. K., Hanafey, M. K., Rafalski, J. A. and Tingey,
S. V., Genetic analysis using random amplified polymorphic DNA
markers. Methods Enzymol., 1993, 218, 704–740.
12. Singh, S. P., Nodari, R. and Gepts, P., Genetic diversity in culti-
vated common bean: I. Allozymes. Crop Sci., 1991, 31, 19–23.
13. Singh, S. P., Gutierrez, J. A., Molina, A., Urrea, C. and Gepts, P.,
Genetic diversity in cultivated common bean: II. Marker-based
analysis of morphological and agronomic traits. Crop Sci., 1991,
31, 23–29.
14. Haley, S. D., Miklas, P. N., Afanador, L. and Kelly, J. D., Random
amplified polymorphic DNA (RAPD) marker variability between
and within gene pools of common bean. J. Am. Soc. Hortic. Sci.,
1994, 119, 122–125.
15. Beebe, S., Skroch, P. W., Tohme, J., Duque, M. C., Pedraza, F.
and Nienhuis, J., Structure of genetic diversity among common
bean landraces of Middle American origin based on correspon-
dence analysis of RAPD. Crop Sci., 2000, 40, 264–273.
16. Metais, I. C., Aubry, B., Hamon, R. J. and Peltier, V., Description
and analysis of genetic diversity between commercial bean lines
(Phaseolus vulgaris L.). Theor. Appl. Genet., 2000, 101, 1207–
1214.
17. Escribano, M. R., Santalla, M., Casquero, P. A. and de Ron, A.
M., Patterns of genetic diversity in landraces of common bean
(Phaseolus vulgaris L.) from Galicia. Plant Breed., 2006, 117,
49–56.
18. Manoj, T., Singh, N. K., Meenal, R. and Narendra, K., RAPD
markers in the analysis of genetic diversity among common bean
germplasm from Central Himalaya. Genet. Res. Crop Evol., 2005,
52, 315–324.
19. Lioi, L., Piergiovanni, A. R., Pignone, D., Puglisi, S., Santantonio,
M. and Sonnante, G., Genetic diversity of some surviving on-farm
Italian common bean (Phaseolus vulgaris L.) landraces. Plant
Breed., 2005, 124, 576–581.
20. Xiaoyan, Z., Matthew, W. B. and Shumin, W., Genetic diversity
of Chinese common bean (Phaseolus vulgaris L.) landraces
assessed with simple sequence repeat markers. Theor. Appl.
Genet., 2008, 117, 629–640.
21. IBPGR, Descriptor list for Phaseolus vulgaris L. International
Board for Plant Genetic Resources, Rome, 1982.
22. Rodino, A. P., Santalla, M., Gonzalez, A. V., Ron, A. M. De and
Singh, S. P., Novel genetic variation in common bean from the
Iberian Peninsula. Crop Sci., 2006, 46, 2540–2546.
23. Rosales-Serna, R., Hernandez-Delgado, S., Gonzalez-Paz, M.,
Acosta-Gallegos, J. A. and Mayek-Perez, N., Genetic relationships
and diversity revealed by AFLP markers in Mexican common
bean bred cultivars. Crop Sci., 2005, 45, 1951–1957.
24. Lynch, M. and Milligan, B. G., Analysis of population genetic
structure with RAPD markers. Mol. Ecol., 1994, 3, 91–99.
25. Evans, A. M., Plant architecture and physiological efficiency in
the field bean. In Potentials of Field Bean and other Food Leg-
umes in Latin America (ed. Wall, D.), CIAT, Cali, Colombia,
1973, pp. 279–284.
26. Evans, A. M., Structure, variation, evolution, and classification in
Phaseolus. In Advances in Legume Science (eds Summerfield, R.
J. and Bunting, A. H.), Royal Botanic Gardens, Kew, UK, 1980,
pp. 337–347.
27. Zizumbo-Villarreal, D., Colunga-Garcia-Marin, P., Payro de la
Cruz, E., Delgado-Valerio, P. and Gepts, P., Population structure
and evolutionary dynamics of wild–weedy–domesticated com-
plexes of common bean in a Mesoamerican region. Crop Sci.,
2005, 45, 1073–1083.
28. Gepts, P., Osborn, T. C., Rashka, K. and Bliss, F. A., Phaseolin
protein variability in wild forms and landraces of the common
bean (Phaseolus vulgaris): evidence for multiple centers of
domestication. Econ. Bot., 1986, 40, 451–468.
29. Ilaria, M., Alessandra, B., Maurizio, M., Pietro, C. and Giovenni,
D., Characterization of some Italian common bean (Phaseolus
vulgaris L.) landraces by RAPD, semi-random and ISSR mole-
cular markers. Genet. Res. Crop Evol., 2007, 54, 175–188.
30. Ron, A. M., De, Menéndez, M. C. and Santalla, M., Variation in
primitive landraces of common bean (Phaseolus vulgaris L.) from
Argentina. Genet. Res. Crop Evol., 2004, 51, 883–894.
31. Beebe, S., Rengifo, J., Gaitan, E., Duque, M. C. and Tohme, J.,
Diversity and origin of Andean landraces of common bean. Crop
Sci., 2001, 41, 854–862.
32. Sonnante, G., Stocklom, T., Nodari, R. O., Becerra Velasquez, V.
L. and Gepts, P., Evolution of genetic diversity during the domes-
tication of common bean (Phaseolus vulgaris L.). Theor. Appl.
Genet., 1994, 89, 629–635.
33. Beebe, S., Toro, O., Gonzalez, A. V., Chacon, M. I. and Debouck,
D. G., Wild–weed–crop complexes of common bean (Phaseolus
vulgaris L., Fabaceae) in the Andes of Peru and Colombia, and
their implications for conservation and breeding. Genet. Res. Crop
Evol., 1997, 44, 73–91.
34. White, J., Kornegay, J., Castillo, J., Molano, C. H., Cajiao, C. and
Tejada, G., Effect of growth habit on yield of large-seeded bush
cultivars of common bean. Field Crops Res., 1992, 29, 151–161.
35. Rodino, A. P., Santalla, M., Ron, A. M. De and Singh, S. P., A
core collection of common bean from the Iberian peninsula.
Euphytica, 2003, 131, 165–175.
36. Wells, W. C., Isom, W. H. and Waines, J. G., Outcrossing rates of
six common bean lines. Crop Sci., 1988, 28, 177–178.
37. Ibarra-Perez, F. J., Ehdaie, B. and Wainess, G. J., Estimation of
out-crossing rate in common bean.
Crop Sci., 1997, 37, 60–65.
38. Gepts, P. and Bliss, F. A., F1 hybrid weakness in the common
bean: differential geographic origin suggests two gene pools in
cultivated bean germplasm. J. Hered., 1985, 76, 447–450.
39. Singh, S. P., Selection for seed yield in Middle American versus
Andean × Middle American interracial common bean populations.
Plant Breed., 1995, 114, 269–271.
40. Duarte, J. M., Santos, J. B. and Melo, L. C., Genetic divergence
among common bean cultivars from different races based on
RAPD markers. Genet. Mol. Biol., 1999, 22, 419–426.
ACKNOWLEDGEMENTS. We are grateful to Paul Gepts, Roberto
Bollini, Steve Beebe and Cesar Ocampo for their valuable suggestions,
the traditional ‘Badaga’ farmers of Nilgiri District who are conserving
and cultivating this germplasm for many centuries, and Dr M. Radha-
krishna Pillai, Director, RGCB, Thiruvananthapuram, for providing
necessary facilities to carry out RAPD analysis. This study was made
possible through a grant from the South Eastern Regional Office
(SERO), of the University Grants Commission, Hyderabad, in the form
of Minor Research Project to F.C.J. We thank the reviewers for their
in-depth review and valuable comments to improve this manuscript.
Received 10 December 2008; revised accepted 1 June 2009