q NIAB 2007 Plant Genetic Resources: Characterization and Utilization 5(1); 27–36

ISSN 1479-2621 DOI: 10.1017/S1479262107390916

Simple sequence repeat (SSR)-based

diversity analysis of groundnut (Arachis
hypogaea L.) germplasm resistant to
bacterial wilt
E. S. Mace1*†, W. Yuejin2†‡, L. Boshou3, H. Upadhyaya1, S. Chandra1 and
J. H. Crouch1§
1
International Crop Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, A.P.
502 324, India,2Institute of Crop Germplasm Resource, Chinese Academy of Agricultural
Sciences (CAAS), Beijing 100081, P.R. China and 3Oil Crops Research Institute (OCRI),
Chinese Academy of Agricultural Sciences (CAAS), Wuhan, Hubei 430062, P.R. China

Received 2 April 2006; Accepted 9 June 2006

Abstract
Groundnut is one of the most important oilseed crops in the world. Bacterial wilt, caused by
Ralstonia solanacearum E. F. Smith, is one of the major biotic constraints to groundnut pro-
duction particularly in South-East Asia and East Africa. Several sources of resistance to bacterial
wilt have been identified through field screening of groundnut germplasm. The aim of the pre-
sent study was to quantify the genetic diversity among selected bacterial wilt-resistant lines, in
comparison with the levels of variation observable within the cultivated A. hypogaea gene
pool. Thirty-two SSR markers were used to assess the degree of molecular polymorphism
between 46 selected genotypes revealing 107 alleles, of which 101 (99.4%) were polymorphic
with gene diversity scores ranging from 0.103 to 0.669, averaging 0.386. Cluster and multidi-
mensional scaling analysis revealed two distinct groups within the germplasm broadly corre-
sponding to the two subspecies (hypogaea and fastigiata) of A. hypogaea. However,
accessions of varieties peruviana and aequatoriana grouped together with the varieties
from subsp. hypogaea, rather than grouping with the other varieties of subsp. fastigiata. Anal-
ysis of molecular variance (AMOVA) revealed that 15% of the total observed variation was
accounted for by disease response groups. This analysis will be useful in the selection of par-
ental genotypes for mapping populations and breeding programmes attempting to broaden
the genetic base of future groundnut cultivars. In particular, this opens up significant opportu-
nities for the development of intraspecific mapping populations that will be highly relevant to
modern groundnut breeding programmes.

Keywords: bacterial wilt; genetic diversity; groundnut; SSRs

* Corresponding author (present address): Hermitage Research Introduction

Station, 604 Yangan Road, Warwick, Qld 4370, Australia.
E-mail: emma.mace@dpi.qld.gov.au Groundnut (Arachis hypogaea L.), also known as peanut,
†
Authors contributed equally. is one of the most important oilseed crops in the world. It
‡
Present address: Department of Biology/Microbiology, South
Dakota State University, Brookings, SD 57 007, USA.
is grown extensively in the Americas, Africa and Asia with
§
Present address: CIMMYT, Apdo. Postal 6-641, 06 600 Mexico, a total annual global area of nearly 24 million hectares
D.F., Mexico. yielding 33.5 million tonnes. Around 53% of the global
28 E. S. Mace et al.

production is crushed for edible oil, 32% for confection- for groundnut production should the pathogen overcome
ery consumption and the remaining 15% is used for feed these sources of resistance. In addition, the currently
and seed production. Cultivated groundnut consists of available BW-resistant cultivars are low yielding and
two subspecies, hypogaea and fastigiata, which are have poor tolerance to drought. Identification and utiliz-
further classified into six botanical varieties largely ation of a broad spectrum of genetically diverse sources
based on growth habit, flowering and branching of BW resistance is, therefore, critical for the develop-
patterns, presence of hairs on leaf surfaces and numbers ment of a new generation of broad-based high-yielding
of seeds per pod (Krapovickas and Gregory, 1994) and BW-resistant groundnut cultivars. Limited knowledge
with lesser support by protein and genomic analysis about the genetic diversity of the BW-resistant germplasm
(Smartt and Simmonds, 1995), with a number of recent and deleterious linkage drag have impeded the utilization
genomic studies refuting this classification (He and of a wide spectrum of BW resistance donors.
Prakash, 2001; Raina et al., 2001; Ferguson et al., 2004a; Diversity studies in groundnut have generally revealed
He et al., 2005; Tallury et al., 2005). Botanical varieties extensive phenotypic variation amongst varieties (Upad-
hypogaea (Virginia) and hirsuta (Peruvian) are currently hyaya et al., 2001, 2003) yet limited variation at the mol-
classified under subsp. hypogaea while varieties fasti- ecular level (Halward et al., 1991, 1992; Kochert et al.,
giata (Valencia), peruviana, aequatoriana and vulgaris 1991, Paik-Rao et al., 1992; He and Prakash, 1997;
(Spanish) are assigned within subsp. fastigiata. Subramanian et al., 2000; Moretzsohn et al., 2004). It is
Bacterial wilt (BW), caused by Ralstonia solanacearum hypothesized that this may be due to the selective neu-
E. F. Smith, is a major biotic factor affecting groundnut trality of the molecular markers utilized, while phenoty-
production particularly in South-East Asia and East pic traits have been subjected to intense selection
Africa (Hayward, 1990). It also infects many other crop (He and Prakash, 1997). It has also been suggested
plants including potato (Solanum tuberosum L.), tomato that the lack of molecular polymorphism revealed to
(Lycopersicon esculentum Mill), tobacco (Nicotiana date within the cultivated groundnut gene pool could
spp.), pepper (Capsicum spp.), eggplant (Solanum mel- be due to the inadequacy of the material studied and
ongina L.) and ginger (Zingiber officinale Rosc.). The the range of techniques used (Singh et al., 1998).
bacterial species have been isolated and classified into Microsatellite markers, also known as simple sequence
five races based on host range (Buddenhagen and repeat (SSR) markers, have been reported to detect high
Kelman, 1964; He et al., 1983) and five biovars based levels of polymorphism even amongst closely related cul-
on biochemical characteristics (Hayward, 1964; He et al., tivated germplasm (Gianfranceschi et al., 1998). For this
1983). Race 1 causes wilt in groundnut, in addition to reason, considerable efforts have been recently made to
many other leguminous and solanaceous plants. Biovar develop a large number of SSR markers in groundnut
1 causes wilt in groundnut and predominantly occurs in (Hopkins et al., 1999; He et al., 2003, 2005; Ferguson
America, whereas Biovars 3 and 4 cause wilt of ground- et al., 2004b; Moretzsohn et al., 2004, 2005). It has been
nut in Asia and Africa (Hayward, 1991). demonstrated in previous studies that SSR markers are
Extensive screening of groundnut germplasm, largely more variable within genomes than other marker types
based on field evaluations at disease ‘hot-spots’ in Indo- (e.g. Belaj et al., 2003). Additionally, SSRs have the
nesia and China, has resulted in the identification of advantage of being co-dominant, only requiring very
many BW-resistant lines. Most of these sources of BW small amounts of DNA and hence have been widely
resistance originate from China or Indonesia. Up to a applied in many plant genetics studies, e.g. for evaluating
quarter of germplasm accessions have shown some genetic diversity (Zhebentyayaeva et al., 2003; Fahima
level of resistance to BW (Singh et al., 1997; Pande et al., 1998), genome mapping and gene tagging, e.g.
et al., 1998); resistance has been identified across differ- in rice (Chen et al., 1997), wheat (Röder et al., 1998),
ent botanical types and also in some wild species. barley (e.g. Künzel and Waugh, 2002) and tomato
Despite the broad range of BW-resistant donor geno- (Broun and Tanksley, 1996). The recent development of
types, only a few of the resistant landraces have been groundnut-specific SSRs (Hopkins et al., 1999; He et al.,
successfully used in breeding in China and Indonesia 2003, 2005; Ferguson et al., 2004b; Moretzsohn et al.,
(Liao et al., 1998). The majority of BW-resistant cultivars 2004, 2005) now offers new and exciting opportunities
released in China were derived from just three sources for groundnut genomics.
(Xiekangqing, Taishan Sanlirou or Taishan Zhenzhu) In the present study we have used groundnut-specific
(Liang, 1998; Liao et al., 1998), while a single resistance SSRs to analyse a diverse range of cultivated groundnut
donor source (Schwartz 21) has been the basis of the accessions encompassing all six botanical varieties. The
majority of cultivars released in Indonesia. Thus, there purpose of this study was to investigate the level of mol-
is a progressive narrowing of genetic diversity in BW- ecular polymorphism amongst BW-resistant accessions
resistant breeding programmes that creates a major risk and to compare this with the genetic diversity across
SSR-based diversity analysis of groundnut 29

the cultivated A. hypogaea gene pool. This analysis is (2 18C/cycle) for 1 min and 728C for 1 min 30 s. This
important for the selection of genetically diverse parental was then followed by 20 cycles of 948C for 45 s, 558C
genotypes for mapping populations and BW resistance for 1 min and 728C for 1 min 30 s, followed by a final
breeding programmes aimed at the development of extension step of 10 min (728C).
broad-based cultivars with durable disease resistance.
Electrophoresis and data collection

Materials and methods PCR amplification products were separated on 6% non-

denaturing polyacrylamide gels and revealed using a
Plant material and DNA extraction silver staining procedure based on ammoniacal solutions
of silver, modified from Kolodny (1984). The size of the
Thirty-one groundnut genotypes from the Oil Crops allele scored was determined through comparison with
Research Institute (OCRI) of the Chinese Academy of Agri- the 100 bp DNA ladder (Amersham) included on all gels.
cultural Sciences (CAAS) and 15 genotypes from the Inter- Estimates of similarity were based on two different
national Crops Research Institute for the Semi-Arid Tropics measurements: (1) Nei and Li’s (1979) definition of simi-
(ICRISAT), consisting of landraces, released cultivars and larity: Sij ¼ 2a/(2a þ b þ c), where Sij is the similarity
intraspecific derivatives representing all six varieties of A. between two individuals, i and j, a is the number of
hypogaea, were selected for the present study (Table 1). bands present both in i and j, b is the number of bands pre-
Total genomic DNA was extracted using a CTAB-based sent in i and absent in j, and c is the number of bands absent
procedure reported previously, with 3% (v/v) b-mercap- in i and present in j; (2) Jaccard’s coefficient (Jaccard, 1908):
toethanol in a 3% (w/v) CTAB buffer (Mace et al., 2003). Sij ¼ a/(a þ b þ c). The similarity matrices were then ana-
The quantity and quality of DNA were determined lysed using the clustering method UPGMA (unweighted
electrophoretically through comparison with known con- pair group method; Sokal and Michener, 1958) using the
centrations of uncut l DNA standards and spectrophoto- NTSYS 2.1 software (Rohlf, 2001). The dendrograms were
metric analysis at 260/280 nm, and subsequently diluted created with the tree program of NTSYS, and the goodness
to 5 ng/ml. of fit of the clustering to the data was calculated using the
COPH and MXCOMP program. Multidimensional scaling
(MDS) (Kruskal and Wish, 1978) was also performed to
SSR amplification confirm whether the observed molecular variation indi-
cated evidence of clustering among accessions, as com-
Thirty-two SSR markers were assayed for their ability to pared to the clustering by UPGMA.
detect polymorphism among the 46 cultivated groundnut An analysis of molecular variance (AMOVA) was
accessions selected (Table 2). The 32 SSRs were selected undertaken to partition genetic variability using Arlequin
on the basis of prescreening approximately 200 ground- software version 2.0 (Schneider et al., 2000), and signifi-
nut SSRs, based on the level of polymorphism revealed cance values assigned to variance components based on
between BW-resistant and -susceptible genotypes and the random permutation (10,000 times) of individuals
the reliability and quality of amplicon detection. assuming no genetic structure. Additionally, the gene
PCR reactions were conducted in 20 ml volumes using a diversity (GD) of each SSR was determined as described
GeneAmp PCR System 9700 (Applied Biosystems). The by Weir (1990). GD ¼ 1 2 SP 2i ; where Pi is the frequency
PCR reaction mixtures contained between 5 and 15 ng of the ith allele in the examined genotypes.
of genomic DNA, 10 –30 pmol of each primer, 100–
125 mM of dNTP, 0.6–1.2 U/ml of Taq DNA polymerase
(Amersham), 1 £ PCR buffer (10 mM Tris–HCl pH 8.3, Results
50 mM KCl) and 0.5–2.5 mM MgCl2.
The fixed-temperature PCR programmes consisted of All 32 SSRs successfully generated at least one allele in the
an initial denaturation step for 2 min at 948C, followed region of the expected size in all 46 cultivated groundnut
by 35 cycles of denaturation for 45 s (948C), annealing genotypes (Fig. 1; Table 2). A total of 107 alleles were
for 1 min (57–648C; see Table 2) and extension for observed following amplification of 29 polymorphic and
1 min 30 s (728C). The PCR products were then incubated three monomorphic (pPGPseq-13B06, pPGPseq-3D09
at 728C for a further 10 min to ensure complete extension. and A1-275) SSR loci, of which 101 (99.4%) were poly-
A second PCR programme using the touchdown morphic. The total number of alleles revealed per poly-
approach was also used for selected SSRs (see Table 2) morphic SSR locus ranged from two (pPGPseq-1B09) to
with the following conditions: initial denaturation for 10 (pPGPseq-7H6) with an average of 3.34 alleles per
2 min at 948C, followed by 10 cycles: 948C for 45 s, 658C locus. The observed allele sizes ranged from 131 bp
30 E. S. Mace et al.
Table 1. Groundnut genotypes included in the current study with different levels of resistance (R)
and susceptibility (S) to bacterial wilt (BW) (botanical variety and geographic origin also indicated)

Accession Botanical variety R/S to BW Origin Code

93-76 (Zhonghua No. 6) vulgaris R China 1
Gouliaozhong hypogaea R China 2
Qingmiaodou hirsuta R China 3
Zao18 vulgaris S China 4
Dayebentianzi hirsuta S China 5
9102 vulgaris R China 6
Yueyou200 vulgaris R China 7
Xiekangqing vulgaris R China 8
Luoao Wanhuasheng hirsuta R China 9
Wuchang Laohuasheng hirsuta R China 10
Zhonghua No. 2 vulgaris R China 11
Changsha Tuzihuasheng hirsuta R China 12
Feilongxiang hirsuta R China 13
Shitang Dahuasheng hirsuta R China 14
Zhonghua 212 vulgaris R China 15
Taishan Zhenzhu vulgaris R China 16
Jiangtianzhong hypogaea R China 17
Qidong Dahuasheng hirsuta R China 18
Nanning Sanjindou hirsuta R China 19
Lingui Make hirsuta R China 20
Chico (China) vulgaris S China 21
Bobai Dahuasheng hirsuta R China 22
Mashanguling hirsuta R China 23
QiongxianXiaohongmao hirsuta R China 24
Ehua No. 5 vulgaris R China 25
Zhongxingchi hirsuta R China 26
Bobai Shiyaodou hirsuta R China 27
Taishan Sanlirou fastigiata R China 28
91-074 vulgaris R China 29
Jiankang (89-15 048) vulgaris R China 30
ICG 1704 peruviana R Peru 31
ICG 7894 peruviana R Peru 32
ICG 5276 vulgaris R Russia 33
ICG 14 159 vulgaris Unknown Vietnam 34
J 11 vulgaris S India 35
Gangapuri fastigiata S India 36
ICG15222-1 hypogaea R China 37
ICG15222-2 hypogaea R China 38
Chico vulgaris S USA 39
ICG 15 208 hirsuta Unknown Mexico 40
ICG 15 213 hirsuta Unknown Mexico 41
ICG 12 625 aequatoriana Unknown Ecuador 42
ICG 12 722 aequatoriana Unknown Ecuador 43
ICG 2381 hypogaea Unknown Brazil 44
ICG 3027 hypogaea Unknown India 45
Zhonghua No.5 vulgaris S China 46

(pPGPseq-8D9) to 531 bp (A1-193). The GD scores of the termed an amphidiploid (Burow et al., 2001). This has
29 polymorphic SSR loci ranged from 0.103 (pPGPseq- significant implications in the appropriate choice of bio-
13E9) to 0.669 (pPGPseq-3A08) (Table 2). metric analysis. In particular, SSRs may not always
retain their codominant nature as defining allelic relation-
ships becomes difficult when three or more alleles are
Genetic diversity analysis of cultivated groundnut detected in a single individual, however, this does not
germplasm occur in any of the polymorphic loci scored in this
study, with the exception of A1-275. For this reason clus-
Groundnut is a complex polyploid with two distinct gen- ter analysis was undertaken based on two different simi-
omes that largely segregate in isolation and is, therefore, larity measures: Jaccard (for dominant datasets) and
 SSR-based diversity analysis of groundnut
Table 2. Details of SSRs used in the groundnut diversity analysis (SSRs were developed by Ferguson et al., 2004a unless indicated otherwise)

Expected
product Total no. of alleles Gene diversity
Primer name Repeat motif Tm (8C) size (bp) Observed product size(s) (bp) observed (GD)
pPGPseq-1B09 GA 64 282 268; 269 2 0.306
pPGPseq-2B10 TAA 58 259 268; 269 2 0.297
pPGPseq-2D12B TAA 60 265 289; 300; 323; 333 4 0.57
pPGPseq-2E06 GA 60 250 269; 289; 306; 323 4 0.504
pPGPseq-2G03 TAA 64 215 254; 269; 281 3 0.459
pPGPseq-2G04 TAA 60 289 269; 289; 300; 333 4 0.47
pPGPseq-3A01 TAA 64 238 257; 269; 277; 289 4 0.304
pPGPseq-3A08 TAA 64 152 173; 178; 191; 197 4 0.669
pPGPseq-3B06 GA 61 244 157 1 0
pPGPseq-3B08 TAA 56 266 289; 300; 314; 323 4 0.555
pPGPseq-3D09 GA,GT 63 292 281 1 0
pPGPseq-4A06 AT 63 126 167; 173; 177 3 0.402
pPGPseq-7G2 TATC 65 225 223; 239; 246; 250; 257; 262 6 0.61
pPGPseq-7H6 CTT 60 300 308; 310 2 0.297
pPGPseq-8D9 CTT 61 132 131; 135; 146 3 0.468
pPGPseq-8E12 TTG,TAA 59 198 204; 207; 210; 214 4 0.485
pPGPseq-10H1A CTT 58 139 193; 200 2 0.296
pPGPseq-12F7 TAA 57 290 305; 310 2 0.375
pPGPseq-13E9 TAA 59 299 323; 333 2 0.103
pPGPseq-13A7 TAA 58 265 289; 291; 293 3 0.44
pPGPseq-14A7 CTT,CTG 60 173 167; 173; 177 3 0.402
pPGPseq-14F4 TAA 60 163 173; 178; 184 3 0.402
pPGPseq-14H6 GT 59 285 269; 271; 284; 297; 306; 310; 312; 319; 333; 348 10 0.618
pPGPseq-15C10 TAA 64 203 212; 220 2 0.427
pPGPseq-16G8 TAA 60 194 214; 217; 227; 229 4 0.32
pPGPseq-18A5A AT,TAA 60 268 300; 328 2 0.468
Lec-1a AT 65– 55 120,125 243; 250; 261; 281; 300 5 0.473
Ah4-26a CT 65– 55 160 173; 178; 184 3 0.616
A1-041b Unknown 65– 55 230,350 269; 281; 293 3 0.351
A1-193b Unknown 65– 55 460 510; 520; 531 3 0.403
A1-275b Unknown 65– 55 190,300 181; 195; 305; 330 4 0
A1-745b Unknown 65– 55 150,250 224; 226; 236 3 0.57
a
Microsatellites developed by Hopkins et al. (1999).
b
Microsatellites developed by Moretzsohn et al. (2004).

31
32 E. S. Mace et al.

bacterial wilt with the exception of Zao 18 (variety vul-

garis). The second subcluster, B(II), consisted of five
landraces originating from South and Central America,
including variety hirsuta, but surprisingly also including
Fig. 1. Polymorphism detected by SSR pPGPseq-8E12 across variety peruviana and variety aequatoriana. The third
46 cultivated groundnut genotypes. subcluster, B(III), consisted of only one genotype, Zhon-
ghua No. 5, belonging to variety vulgaris which is sus-
Dice/Nei and Li (for codominant datasets). The corre- ceptible to BW.
lation coefficients were broadly similar from both Overall there was a clear distinction between the BW-
Jaccard’s similarity coefficient (r ¼ 0.89) and Nei and resistant lines belonging to subsp. fastigiata (cluster A)
Li’s similarity coefficient (r ¼ 0.87). Figure 2 shows the and the BW-resistant lines of subsp. hypogaea (cluster
dendrogram produced by Jaccard’s similarity coefficient B). However, not all the accessions susceptible to BW fol-
using the UPGMA clustering method, with clear evidence lowed this pattern; the two BW-susceptible variety vul-
of two separate clusters (A and B) at a level of approxi- garis accessions belonging to A. hypogaea subsp.
mately 36% similarity. The presence of two primary clus- fastigiata clustered instead with the A. hypogaea subsp.
ters within the data set was also confirmed through MDS hypogaea genotypes in cluster B.
(Fig. 3), where clusters A and B appear clearly separated AMOVA was performed on the data set in order to par-
on two axes. The 46 genotypes were all uniquely ident- tition the total genetic variation within and between three
ified based on the 32 SSR loci, with the exception of Fei- parameters: (i) within and between resistant and suscep-
longxiang and Shitang Dahuashe, both botanical variety tible types, (ii) within and between botanical variety, and
hirsuta, resistant to bacterial wilt. Of the 46 genotypes, (iii) within and between country of origin. The AMOVA
seven exhibit susceptibility to bacterial wilt, 32 show revealed that only 15% of the total variation observed
resistance and the response of the remaining seven gen- was accounted for between BW resistance and suscep-
otypes to bacterial wilt is unknown. The two subspecies tible types, whereas the majority of variation (85%)
of A. hypogaea are equally represented with 23 geno- was observed within each disease response group
types belonging to subsp. hypogaea (17 to variety hirsuta (Table 3A). In contrast, upon partitioning the total genetic
and six to variety hypogaea) and 23 genotypes belonging variation between and within the botanical varieties
to subsp. fastigiata (17 to variety vulgaris and two repre- (Table 3B), 50% of the total variation was accounted for
sentatives each of varieties fastigiata, peruviana and between the different botanical varieties, indicating a
aequatoriana). The 46 genotypes group into two separ- clear differentiation based on botanical variety compared
ate clusters; cluster A containing 19 genotypes, 18 of to BW response groups. Finally, 86% of the variation was
which belong to subsp. fastigiata, and cluster B contain- accounted for within the 11 different countries of origin
ing 27 genotypes, 22 of which belong to subsp. hypo- of the genotypes included in this study, rather than
gaea. The 19 genotypes in cluster A include 15 between the countries of origin.
accessions of variety vulgaris, two of variety fastigiata
and two germplasm lines (Chico and ICG15222-2).
Three subclusters could be further identified within clus- Putative association of SSR loci with BW resistance
ter A at a level of 45% similarity. The first subcluster, A(I),
contained genotypes belonging to only variety vulgaris A locus-by-locus AMOVA was performed to calculate the
and the line ICG15222-2, which were all resistant to contribution of each locus to the differentiation of resist-
BW. The second subcluster, A(II), consisted predomi- ant and susceptible groups within the germplasm tested.
nately of variety vulgaris genotypes, with the exception Six of the 107 alleles (Table 4) were found to contribute
of one accession belonging to variety fastigiata (Taishan significantly to the differentiation between the BW-resist-
Sanlirou) and one variety aequatoriana accession ant and -susceptible genotypes. Furthermore, markers
(ICG12722); of these, six were resistant and four suscep- pPGPseq-16G8 (size: 229 bp) and pPGPseq-12F7 (size:
tible to BW. The third subcluster, A(III), consisted of two 305 bp) contributed 74.91% of the total genetic difference
accessions susceptible to BW; an accession belonging to between the two disease-response groups.
variety fastigiata (Gangapuri) and a germplasm line.
Cluster B contained 27 accessions representing five
botanical varieties, with botanical variety fastigiata unre- Discussion
presented. Three subclusters could be identified within
cluster B, at the level of 42% similarity. The first subclus- SSRs have proven to be powerful tools for the detection of
ter, B(I), consisted of 21 genotypes in total, the majority molecular genetic diversity amongst the cultivated ground-
(15) belonging to variety hirsuta, and all resistant to nut germplasm included in this study, representing all
SSR-based diversity analysis of groundnut 33

six botanical varieties within the A. hypogaea gene pool. A There was some degree of clustering of accessions
clear distinction was observed between the two subspe- from similar geographic origins within subspecies, e.g.
cies, subsp. hypogaea and subsp. fastigiata. Accessions of four landraces and one breeding line from South and
var. hypogaea and hirsuta (subsp. hypogaea) grouped Central America grouped together at a level of approxi-
together in cluster B (Fig. 2), whereas var. vulgaris and fas- mately 50% similarity in cluster B and were clearly differ-
tigiata (subsp. fastigiata) grouped together in cluster entiated from accessions derived from other geographic
A. However, accessions of var. peruviana and var. aequa- regions. Additionally, the BW-resistant accessions Bobai
toriana grouped together with the varieties from subsp. Shiyaodou and Bobai Dahuasheng both originate from
hypogaea in cluster B, rather than grouping with the Bobai County in China and were observed to group
other varieties of subsp. fastigiata in cluster A. These results together at a level of approximately 82% similarity in clus-
support recent conclusions from amplified fragment length ter B. However, AMOVA estimated 86% of the SSR vari-
polymorphism (AFLP) analysis (He and Prakash, 2001) that ation is accounted for within countries. So although
var. aequatoriana and peruviana are much closer to subsp. there may be some influence of geographic isolation on
hypogaea than to subsp. fastigiata. genetic polymorphism, as has been reported previously
Most previous reports of diversity analysis in cultivated (e.g. He and Prakash, 2001), it is likely that breeding
groundnut have readily detected morphological variation selection pressure has had greater impact. However, it
but consistently failed to detect a parallel level of molecu- should be noted that the groundnut accessions studied
lar genetic variability, however recent studies (Ferguson here were selected primarily based on their response to
et al., 2004a; Moretzsohn et al., 2004, 2005) using more BW and thus may not be truly reflective of groundnut
recently developed SSRs have found higher levels of gen- variability in each geographical region.
etic diversity in the six botanical varieties of A. hypogaea Regarding the genetic variation observed within and
than reported previously. Significantly, the SSR screening between the BW-resistant and -susceptible accessions,
reported in this study also detected a substantial level of it was observed that the BW-resistant lines belonging
molecular genetic variation between genotypes of all to subsp. fastigiata and subsp. hypogaea were clearly
botanical varieties. However, AMOVA showed an equal differentiated, and clustered within subspecies. In con-
level of diversity within and between botanical varieties. trast, two BW-susceptible accessions (Zao18 and Zhon-
This may suggest that botanical variety designations gua No. 5) did not fall in the expected cluster (subsp.
(largely based on morphological traits) are not truly fastigiata var. vulgaris), but were grouped in cluster
reflective of gross genetic diversity. B (predominately subsp. hypogaea). This anomaly

A
II

III

II

III

Fig. 2. Dendrogram constructed using Jaccard’s similarity coefficient and UPGMA clustering, for the 46 groundnut geno-
types. Two main clusters (A and B) and subclusters are identified.
34 E. S. Mace et al.

the most powerful tools for revealing genetic variation

B A within the cultivated groundnut gene pool.
Several SSR alleles were also found to be significantly
associated with BW resistance and these represent candi-
dates for marker-assisted selection (MAS) following vali-
dation in traditional, segregating mapping populations.
Molecular breeding offers the potential for improving
the speed, precision and cost of groundnut disease resist-
ance breeding programmes (Dwivedi et al., 2003). In
addition, MAS offers the potential for pyramiding differ-
ent sources of resistance together with resistance to mul-
Fig. 3. Multidimensional scaling (MDS) analysis of SSR data tiple diseases which is difficult or impossible to achieve
across 46 cultivated groundnut genotypes with varying through conventional approaches (Mohan et al., 1997).
levels of resistance to bacterial wilt with two clusters of The paucity of molecular polymorphism previously
genotypes indicated as ‘A’ and ‘B’, corresponding to clusters
revealed in cultivated groundnut posed a considerable
identified in Fig. 2.
obstacle to genetic mapping and MAS. The first and
only currently available genetic linkage map of the tetra-
may be largely explained by the pedigree of these two ploid groundnut genome (Burow et al., 2001) was only
breeding lines which included accessions from subsp. made possible through the use of a synthetic amphidi-
hypogaea. It should be noted that the BW-susceptible ploid, TxAG-6, capturing a high level of genetic diversity
var. vulgaris accessions within cluster B (subclusters from divergent diploid species (Simpson, 1991; Simpson
B(II) and B(III)) grouped only at a level of approxi- et al., 1993). Although SSRs now offer a substantially
mately 60% similarity with the resistant accessions higher level of detectable variation, it is still essential to
belonging to subsp. hypogaea. These results could indi- base the selection of parental genotypes for mapping
cate that the accessions susceptible to BW are more populations on detailed diversity analysis. It is hoped
genetically diverse than their resistant counterparts, that the result of this study will help molecular breeders
suggesting that the selection for adaptation to the bac- in selecting the most appropriate parental genotypes for
terial wilt pathogen may have contributed to reduced mapping BW resistance.
genetic variation in the germplasm which may be due It is now important to distinguish which genotypes
to a combination of pathogen pressure and breeder’s possess different mechanisms of resistance and to
selection pressure. search for additional novel sources of resistance to BW.
Pairwise dissimilarities of up to 64% were observed Landraces and varieties from Indonesia, China and Viet-
between cultivated groundnut genotypes screened in nam (where BW disease pressures are highest) would
this study. This compares with 41% genetic differentiation appear to be the most likely targets for identifying new
revealed between cultivated groundnut genotypes as sources of resistance. On this basis, marker-assisted
revealed by randomly amplified polymorphic DNA gene-pyramiding programmes can aim to develop high-
(RAPDs; Dwivedi et al., 2001) and 52% as revealed by yielding varieties with more durable resistance to this
AFLPs (He and Prakash, 2001). Thus, SSRs are clearly devastating disease.

Table 3. AMOVA (Excoffier et al., 1992) for (A) 39 genotypes of two disease response types, bacterial wilt-resistant
and -susceptible, employing 107 SSR alleles; and for (B) 46 genotypes in six botanical varieties employing 107
microsatellite alleles (nested analysis was carried out on populations grouped as above)

Variance Variation Fixation

Source of variation df SSDa components (%) index FST
(A)
Among populations 1 40.857 2.58 815 Va 15.08*** 0.15 078
Within populations 37 539.348 14.57 699 Vb 84.92 ***
Total 38 580.205 17.16 513 100.00
(B)
Among populations 5 336.005 9.46 528 Va 46.19*** 0.46 193
Within populations 7 407.949 11.02 564 Vb 53.81***
Total 42 743.953 20.49 092 100.00
a
Sum of squared deviations.
SSR-based diversity analysis of groundnut 35
Table 4. List of alleles putatively linked with bacterial wilt haplotypes: application to human mitochondrial DNA
resistance as determined through locus-by-locus AMOVA restriction data. Genetics 131: 479– 491.
Fahima T, Röder MS, Grama A and Nevo E (1998) Microsatellite
Contribution value (%) DNA polymorphism divergence in Triticum dicoccoides
to differentiation accessions highly resistant to yellow rust. Theoretical and
Allele between resistant and Applied Genetics 96: 187– 195.
Primer name size (bp) susceptible groups Ferguson ME, Bramel J and Chandra S (2004a) Gene diversity
pPGPseq-16G8 229 74.91 among botanical varieties in peanut (Arachis hypogaea
pPGPseq-12F7 305 74.91 L.). Crop Science 44: 1847 – 1854.
pPGPseq-14H6 310 62.31 Ferguson ME, Burow MD, Schultz SR, Bramel PJ, Paterson AH,
pPGPseq-14H6 306 60.85 Kresovich S and Mitchell S (2004b) Microsatellite identifi-
pPGPseq-7G2 246 59.07 cation and characterization in peanut (A. hypogaea L.).
A1-745 226 59.07 Theoretical and Applied Genetics 108: 1064 –1070.
Gianfranceschi L, Seglias N, Tarchini R, Komjanc M and Gessler C
(1998) Simple sequence repeats for the genetic analysis of
apple. Theoretical and Applied Genetics 96: 1069 – 1076.
Acknowledgements Halward T, Stalker HT, Larue EA and Kochert G (1991) Genetic
variation detectable with molecular markers among una-
dapted germplasm resources of cultivated peanut and
The authors gratefully acknowledge Mr A. Gafoor for
related wild species. Genome 34: 1013 –1020.
technical assistance, Mr S. Hari Krishna for biometrical Halward T, Stalker HT, Larue EA and Kochert G (1992) Use of
advice and Dr M. Ferguson for early access to SSR mar- single-primer DNA amplification in genetic studies of
kers. This research was undertaken as part of the Asian peanut (Arachis hypogaea L.). Plant Molecular Biology
Development Bank-funded project ‘Rapid Crop Improve- 18: 315– 325.
Hayward AC (1964) Characteristics of Pseudomonas solanacer-
ment for Poor Farmers in the Semi-Arid Tropics of Asia’
arum. Journal of Applied Bacteriology 27: 265–277.
(RETA 5945). Biotechnology research at ICRISAT is grate- Hayward AC (1990) Diagnosis, distribution and status of
ful for unrestricted grants from the governments of Japan groundnut bacterial wilt. ACIAR Proceedings 31: 12– 17.
and the UK and from the European Union. Hayward AC (1991) Biology and epidemiology of bacterial wilt
caused by Pseudomonas solanacerarum. Annual Review of
Phytopathology 29: 65– 87.
He G and Prakash C (1997) Identification of polymorphic DNA
References markers in cultivated peanut (Arachis hypogaea L.). Euphy-
tica 97: 143– 149.
Belaj A, Satovic Z, Cipriani G, Baldoni L, Testolin R, Rallo L and He G and Prakash C (2001) Evaluation of genetic relationships
Trujillo I (2003) Comparative study of the discriminating among botanical varieties of cultivated peanut (Arachis
capacity of RAPD, AFLP and SSR markers and of their effec- hypogaea L) using AFLP markers. Genetic Resources and
tiveness in establishing genetic relationships in olive. Crop Evolution 48: 347– 352.
Theoretical and Applied Genetics 107: 736 – 744. He G, Meng R, Newman M, Gao G, Pittman RN and Prakash CS
Broun P and Tanksley SD (1996) Characterization and genetic (2003) Microsatellites as DNA markers in cultivated peanut
mapping of simple repeat sequences in the tomato (Arachis hypogaea L.). BMC Plant Biology 3: 3– 9.
genome. Molecular and General Genetics 250: 39– 49. He G, Meng R, Gao H, Guo B, Gao G, Newman M, Pittman
Buddenhagen I and Kelman A (1964) Biological and physiologi- RN and Prakash CS (2005) Simple sequence repeat markers
cal aspects of bacterial wilt caused by Pseudomonas sola- for botanical varieties of cultivated peanut (Arachis
nacearum. Annual Review of Phytopathology 2: 203 –230. hypogaea L.). Euphytica 142: 131– 136.
Burow MD, Simpson CE, Starr JL and Paterson AH (2001) Trans- He LY, Sequeria L and Kelman A (1983) Characteristics of strains
mission genetics of chromatin from a synthetic amphiploid of Pseudomonas solanacearum from China. Plant Diseases
in cultivated peanut (A. hypogaea L.): broadening the gene 67: 1357 –1361.
pool of a monophyletic polyploid species. Genetics 159: Hopkins MS, Casa AM, Wang T, Mitchell SE, Dean RE, Kochert
823– 837. GD and Kresovich S (1999) Discovery and characterization
Chen X, Temnykh S, Xu Y, Cho YG and McCouch SR (1997) of polymorphic simple sequence repeats (SSRs) in peanut.
Development of a microsatellite framework map providing Crop Science 39: 1243 –1247.
genome wide coverage in rice (Oryza sativa L.). Theoreti- Jaccard P (1908) Nouvelles recherches sur la distribution florale.
cal and Applied Genetics 95: 553 – 567. Bulletin de la Société Vaudoise des Sciences Naturelles 44:
Dwivedi SL, Gurtu S, Chandra S, Yuejin W and Nigam SN (2001) 223– 270.
Assessment of genetic diversity among selected groundnut Kochert G, Halward T, Branch WD and Simpson CE (1991) RFLP
germplasm I: RAPD analysis. Plant Breeding 120: 345 – 349. variability in peanut (Arachis hypogaea L.) cultivars and
Dwivedi SL, Crouch JH, Nigam SN, Ferguson ME and Paterson wild species. Theoretical and Applied Genetics 81: 565– 570.
AH (2003) Molecular breeding of groundnut for enhanced Kolodny GM (1984) An improved method for increasing the res-
productivity and food security in the semi-arid tropics: olution and sensitivity of silver staining of nucleic acid
opportunities and challenges. Advances in Agronomy 80: bands in polyacrylamide gels. Analytical Biochemistry
153– 221. 138: 66– 67.
Excoffier L, Smouse PE and Quattro JM (1992) Analysis of mol- Krapovickas A and Gregory WC (1994) Taxonomia del genero
ecular variance inferred from metric distances among DNA Arachis (Leguminosae). Bonplandia 8: 1 –186.
36 E. S. Mace et al.

Kruskal JB and Wish M (1978) Multidimensional Scaling. cation, and phylogenetic relationships in peanut (Arachis
Newbury Park, CA: Sage. hypogaea) cultivars and wild species. Genome 44: 763– 772.
Künzel G and Waugh R (2002) Integration of microsatellite mar- Röder MS, Korzun V, Wendehake K, Plaschke J, Tixier M-H,
kers into the translocation-based physical RFLP map of Leroy P and Ganal MW (1998) A microsatellite map of
barley chromosome 3H. Theoretical and Applied Genetics wheat. Genetics 149: 2007 –2023.
105: 660–665. Rohlf FJ (2001) Numerical Taxonomy and Multivariate Analysis
Liang XQ (1998) Present status of groundnut bacterial wilt in System. New York: Applied Biostatistics.
south China. In: Pande S, Liao BS, Hong NX, Johansen C Schneider S, Roessli D and Excoffier L (2000) Arlequin: A Soft-
and Gowda CLL (eds) Groundnut Bacterial Wilt in Asia. ware for Population Genetics Data Analysis. version
Proceedings of the Fourth Working Group Meeting. 11–13 2.000. Geneva: Genetics and Biometry Laboratory, Depart-
May 1998. Hanoi, Vietnam: Vietnam Agricultural Science ment of Anthropology, University of Geneva.
Institute, pp. 23– 26. Simpson CE (1991) Pathways for introgression of pest resistance
Liao B, Duan N, Jiang H, Liang X and Gao G (1998) Germplasm into Arachis hypogaea L. Peanut Science 18: 22 –26.
screening and breeding for resistance to bacterial wilt in Simpson CE, Nelson SC, Starr JL, Woodard KE and Smith OD
China. In: Pande S, Liao BS, Hong NX, Johansen C and (1993) Registration of TxAG-6 and TxAG-7 peanut germ-
Gowda CLL (eds) Groundnut Bacterial Wilt in Asia. Pro- plasm lines. Crop Science 33: 1418.
ceedings of the Fourth Working Group Meeting. 11– 13 Singh AK, Mehan VK and Nigam SN (1997) Sources of Resistance
May 1998. Hanoi, Vietnam: Vietnam Agricultural Science to Groundnut Fungal and Bacterial Diseases, an Update
Institute, pp. 75– 81. and Appraisal. ICRISAT Bulletin No. 50. Patancheru,
Mace ES, Buhariwalla HK and Crouch JH (2003) A high India: ICRISAT.
throughput DNA extraction protocol for tropical molecular Singh AK, Smartt J, Simpson CE and Raina SN (1998) Genetic
breeding programs. Plant Molecular Biology Reporter 21: variation vis-à-vis molecular polymorphism in groundnut,
459a – 459h. Arachis hypogaea L. Genetic Resources and Crop Evolution
Mohan M, Nair S, Bhagwat A, Krishna TG, Yano M, Bhatia CR 45: 119– 126.
and Sasaki T (1997) Genome mapping, molecular markers Smartt J and Simmonds NW (1995) Evolution of Crop Plants.
and marker-assisted selection in crop plants. Molecular Harlow: Longman Scientific and Technical.
Breeding 3: 87– 103. Sokal RR and Michener CD (1958) A statistical method for eval-
Moretzsohn MC, Hopkins MS, Mitchell SE, Kresovich S, Valls JFM uating systematic relationships. University of Kansas Scien-
and Ferreira ME (2004) Genetic diversity of peanut (Ara- tific Bulletin 38: 1409 –1438.
chis hypogaea L.) and its wild relatives based on the anal- Subramanian V, Gurtu S, Nageswara Rao RC and Nigam SN
ysis of hypervariable regions of the genome. BMC Plant (2000) Identification of DNA polymorphism in cultivated
Biology 4: 11. groundnut using random amplified polymorphic DNA
Moretzsohn MC, Leoi L, Proite K, Guimãres PM, Leal-Bertioli (RAPD) assay. Genome 43: 656– 660.
SCM, Gimenes MA, Martins WS, Valls JFM, Grattapaglia D Tallury SP, Hilu KW, Milla SR, Friend SA, Alsaghir M, Stalker
and Bertoli DJ (2005) A microsatellite-based, gene-rich HT and Quandt D (2005) Genomic affinities in Arachis
linkage map for the AA genome of Arachis (Fabaceae). section Arachis (Fabaceae): molecular and cytogenetic
Theoretical and Applied Genetics 111: 1060 –1071. evidence. Theoretical and Applied Genetics 111:
Nei M and Li WH (1979) Mathematical model for studying genetic 1229 – 1237.
variation in terms of restriction end nucleases. Proceedings of Upadhyaya HD, Ferguson ME and Bramel PJ (2001) Status of the
the National Academy of Sciences USA 76: 5269– 5273. Arachis germplasm collection at ICRISAT. Peanut Science
Paik-Rao OG, Smith RL and Knauft DA (1992) Restriction frag- 28: 89– 96.
ment length polymorphism evaluation of six peanut Upadhyaya HD, Ortiz R, Bramel PJ and Singh S (2003) Develop-
species within the Arachis section. Theoretical and Applied ment of a groundnut core collection using taxonomical,
Genetics 84: 201– 208. geographical and morphological descriptors. Genetic
Pande S; Liao BS; Hong NX; Johansen C; Gowda CLL, (eds) Resources and Crop Evolution 50: 139–148.
(1998) Groundnut Bacterial Wilt in Asia. Proceedings of Weir BS (1990) Genetic Data Analysis. Methods for Discrete Gen-
the Fourth Working Group Meeting. 11 –13 May 1998. etic Data. Sunderland, MA: Sinauer Associates.
Hanoi, Vietnam: Vietnam Agricultural Science Institute. Zhebentyayaeva TN, Reighard GL, Gorina V and Abbott AG
Raina SN, Kojima T, Ogihara Y, Singh KP and Devarumath RM (2003) Simple sequence repeat (SSR) analysis for assess-
(2001) RAPD and ISSR fingerprints as useful genetic ment of genetic variability in apricot germplasm. Theoreti-
markers for analysis of genetic diversity, varietal identifi- cal and Applied Genetics 106: 435–444.