Page 1
Theor Appl Genet (2010) 121:731–739
DOI 10.1007/s00122-010-1344-3ORIGINAL PAPER
An interspeciWc linkage map of SSR and intronic polymorphism
markers in tomato
Kenta Shirasawa · Erika Asamizu · Hiroyuki Fukuoka · Akio Ohyama · Shusei Sato ·
Yasukazu Nakamura · Satoshi Tabata · Shigemi Sasamoto · Tsuyuko Wada · Yoshie Kishida ·
Hisano Tsuruoka · Tsunakazu Fujishiro · Manabu Yamada · Sachiko Isobe
Received: 16 December 2009 / Accepted: 11 April 2010 / Published online: 30 April 2010
© The Author(s) 2010. This article is published with open access at Springerlink.com
Abstract Despite the collection and availability of abun-
dant tomato genome sequences, PCR-based markers
adapted to large scale analysis have not been developed in
tomato species. Therefore, using public genome sequence
data in tomato, we developed three types of DNA markers:
expressed sequence tag (EST)-derived simple sequence
repeat (SSR) markers (TES markers), genome-derived SSR
markers (TGS markers) and EST-derived intronic polymor-
phism markers (TEI markers). A total of 2,047 TES, 3,510
TGS and 674 TEI markers were established and used in the
polymorphic analysis of a cultivated tomato (Solanum lyco-
persicum) ‘LA925’ and its wild relative Solanum pennellii
‘LA716’, parents of the Tomato-EXPEN 2000 mapping
population. The polymorphic ratios between parents
revealed by the TES, TGS and TEI markers were 37.3, 22.6
and 80.0%, respectively. Those showing polymorphisms
were used to genotype the Tomato-EXPEN 2000 mapping
population, and a high-density genetic linkage map com-
posed of 1,433 new and 683 existing marker loci was con-
structed on 12 chromosomes, covering 1,503.1 cM. In the
present map, 48% of the mapped TGS loci were located
within heterochromatic regions, while 18 and 21% of TES
and TEI loci, respectively, were located in heterochromatin.
The large number of SSR and SNP markers developed in
this study provide easily handling genomic tools for
molecular breeding in tomato. Information on the DNA
markers developed in this study is available at http://
www.kazusa.or.jp/tomato/.
Introduction
The tomato, Solanum lycopersicum, which originated in
Latin America, is the second most important vegetable crop
and is cultivated throughout the world (Foolad 2007).
Tomato belongs to the family Solanaceae, which consists
of approximately 100 genera and 2,500 species, including
several plants of agronomic importance such as potato, egg-
plant, pepper and tobacco (Olmstead et al. 2008). Tomato
has a relatively compact genome within the Solanaceae
species, characterized by its diploidy (2n = 2X = 24). It is
approximately 950 Mb in size, and is one of the most inten-
sively characterized Solanaceae genomes (Arumuganathan
and Earle 1991). The International Tomato Sequencing
Project, established in 2004 with members from 10 coun-
tries, promotes structural genome analysis in tomato by
sequencing the gene-rich regions of all 12 chromosomes
through the generation of high quality sequences from bac-
terial artiWcial chromosomes (BAC). The data are released
Communicated by I. Paran.
Electronic supplementary material The online version of this
article (doi:10.1007/s00122-010-1344-3) contains supplementary
material, which is available to authorized users.
K. Shirasawa · E. Asamizu · S. Sato · Y. Nakamura · S. Tabata ·
S. Sasamoto · T. Wada · Y. Kishida · H. Tsuruoka · T. Fujishiro ·
M. Yamada · S. Isobe (&)
Kazusa DNA Research Institute, 2-6-7 Kazusa-kamatari,
Kisarazu, Chiba 292-0818, Japan
e-mail: sisobe@kazusa.or.jp
E. Asamizu
Gene Research Center, University of Tsukuba,
Ten-no dai 1-1-1, Tsukuba, Ibaraki 305-8572, Japan
H. Fukuoka · A. Ohyama
National Institute of Vegetable and Tea Science,
360 Kusawa, Ano, Tsu, Mie 514-2392, Japan
Y. Nakamura
National Institute of Genetics, 1111 Yata, Mishima,
Shizuoka 411-8540, Japan123
Page 2
732 Theor Appl Genet (2010) 121:731–739immediately after completion, principally through the SOL
Genomics Network (SGN) website (http://solgenomics.net/
; Mueller et al. 2005). In addition, ESTs and full-length
cDNAs of a miniature tomato cultivar (S. lycopersicum)
named Micro-Tom have been published at MiBASE (http://
www.kazusa.or.jp/jsol/microtom/) and KafTom (http://
www.pgb.kazusa.or.jp/kaftom/) (Yamamoto et al. 2005).
Currently, more than 320,000 ESTs and 461,000 BAC-end
sequences have been registered in the databases.
Genetic diversity in the cultivated tomato is generally
low, due to the occurrence of population bottlenecks during
the domestication and the generation of modern varieties
(Rick 1976). Therefore, during the past quarter century,
several linkage maps of the tomato genome have been
developed with more than 20 mapping populations derived
from crosses between cultivated tomato (S. lycopersicum) and
its wild relatives, such as S. pennellii, S. pimpinellifolium,
S. cheesmaniae, S. neorickii, S. chmielewskii, S. habrocha-
ites and S. peruvianum (reviewed by Foolad 2007). In par-
ticular, an F2 mapping population named Tomato-EXPEN
1992, which was derived from a cross between S. lycopersi-
cum cv. VF36 and the inbred accession of S. pennellii
LA716, facilitated the development of a high-density
restriction fragment length polymorphism (RFLP) linkage
map with 1,030 loci (Tanksley et al. 1992). Subsequently,
other types of DNA markers, such as random ampliWed
polymorphic DNA (RAPD), simple sequence repeats
(SSRs), ampliWed fragment length polymorphisms
(AFLPs), cleaved ampliWed polymorphic sequence (CAPS)
and single nucleotide polymorphisms (SNPs), were devel-
oped and mapped onto the Tomato-EXPEN 2000 mapping
population derived from the cross between S. lycopersicum
LA925 and S. pennellii LA716 (Frary et al. 2005; Fulton
et al. 2002). Currently, the Tomato-EXPEN 2000 map
comprises a total of 2,604 markers including 1,088 CAPS,
1,342 RFLP, 155 SSR and 19 SNP markers (http://solge-
nomics.net/cview/map.pl?map_id=9; Fulton et al. 2002). In
addition, a total of 349 PCR-based markers are now avail-
able at Tomato Mapping Resource Database (http://
www.tomatomap.net), and mapped on interspeciWc crosses
between S. pimpinellifolium and S. lycopersicum.
Most of the loci mapped on the two Tomato-EXPEN
mapping populations consist of RFLP and CAPS markers,
which result in a higher cost for the identiWcation of poly-
morphisms due to their laborious nature and the need for
restriction enzymes. The rapid progress in genome analy-
sis during the past several years has enabled large scale
segregation analysis in molecular genetics. We therefore
considered it essential to develop PCR-based markers
designed to adapt large scale genotyping systems, such as
SSR and SNP markers, for the promotion of genetics and
genomics in tomato species. Recently, Ohyama et al.
(2009) developed SSR markers using BAC-end and
cDNA sequences in tomato and mapped a total of 148
SSR loci onto the EXPEN 2000 map. They investigated a
total of 89,824 cDNA and 310,583 BAC-end sequences to
generate SSR markers. Their results suggested that abun-
dant sequence resources would allow the generation of a
larger number of PCR-based markers in tomato. In this
study, we developed three types of DNA markers, namely
EST-derived SSR markers (TES markers), genome-
derived SSR markers (TGS markers) and EST-derived
intronic polymorphism markers (TEI markers), using the
public genome sequence data. As Ohyama et al. (2009)
reported that most of the SSR markers derived from BAC-
end sequences mapped to the centromeric regions, we
used diVerent polymorphic sources for marker generation.
The developed markers were mapped onto the Tomato-
EXPEN 2000 mapping population to Wll in the gaps in
knowledge within and between tomato genomics and
genetics.
Materials and methods
Plant material
A previously reported F2 mapping population, which was
derived from a cross between S. lycopersicum LA925 and
the inbred accession S. pennellii LA716 (Fulton et al.
2002), was used for linkage map construction. The map-
ping parents and 83 F2 individuals were kindly provided by
Prof. S. Tanksley from Cornell University. Total DNA was
extracted from the leaves of each plant using the DNeasy
Plant Mini kit (Qiagen, Germany).
Development of TES (tomato EST-SSR) and TGS
(tomato Genome-SSR) markers
Microsatellite or SSR regions were identiWed in tomato
ESTs and BAC-end sequences registered in public dat-
abases, namely MiBASE (http://www.kazusa.or.jp/jsol/
microtom/), KafTom (http://www.pgb.kazusa.or.jp/kaftom/)
and SGN (http://sgn.cornell.edu/). SSRs longer than
14 bases, which contained all possible combinations of
dinucleotide (NN), trinucleotide (NNN) and tetranucleotide
(NNNN) repeats, were identiWed using the FINDPAT-
TERNS module in the GCG software package (Accelrys
Inc., USA). Oligonucleotides for PCR primers were
designed based on the Xanking regions of the identiWed
SSRs using the Primer3 program (Rozen and Skaletsky
2000) in such a way that the ampliWed products ranged
between 90 and 300 bp in length. Markers corresponding to
those previously developed by Fulton et al. (2002) and
Ohyama et al. (2009) were identiWed based on the associ-
ated nucleotide sequences and excluded from our collection123
Page 3
Theor Appl Genet (2010) 121:731–739 733of EST-derived SSR (TES) and BAC-end sequences-
derived SSR (TGS) markers.
PCR was performed using 0.5 ng tomato genomic DNA
in a 5 �l reaction mix containing 1£ PCR buVer (BIO-
LINE, UK), 3 mM MgCl2, 0.04 U BIOTAQTM DNA poly-
merase (BIOLINE, UK), 0.2 mM dNTPs and 0.8 �M of
each primer. A modiWed ‘touchdown PCR’ protocol was
used, as described previously (Sato et al. 2005). PCR prod-
ucts were separated either on a 10% polyacrylamide gel
with TBE buVer according to the standard protocol or in a
Type 3730 DNA fragment analyzer (Applied Biosystems,
USA). In the latter case, the data were analyzed using
GeneMapper software (Applied Biosystems, USA).
Development of TEI (tomato EST-derived intronic
polymorphism) markers
The positions of introns were predicted by comparisons
between KTU2 EST unigene sequences in MiBASE and
Arabidopsis thaliana genomic sequences (http://www.ara-
bidopsis.org/) using the GAP2 program (Huang 1994).
Oligonucleotides for PCR primers were designed based on
the Xanking regions of the predicted intron positions in the
tomato unigenes using the Primer3 program (Rozen and
Skaletsky 2000). PCR was performed as described above.
Polymorphic PCR amplicons were identiWed by high-reso-
lution melting (HRM) analysis (Palais et al. 2005) using the
LightScanner system (Idaho Technology, USA) with 0.5 �l
of 10£ LCGreen Plus + Melting Dye (Idaho Technology)
as the Xuorescent dye and 0.05 ng of genomic DNA of
S. pennellii ‘LA716’ as the control DNA.
Linkage analysis
Linkage analysis was performed using genotypic data
derived from the markers developed in this study along
with data derived from a total of 683 published markers.
These markers included 449 RFLPs, 61 CAPS markers and
173 SSR markers (Fulton et al. 2002; Ohyama et al. 2009),
which were retrieved from the SGN (http://sgn.cornell.edu/)
and VegMarks (http://vegmarks.nivot.affrc.go.jp/) databases.
The genotypic data of the mapping population generated
from the published markers were kindly provided by Prof.
S. Tanksley at Cornell University and Dr. A. Ohyama at the
National Institute of Vegetable and Tea Science of Japan
(NIVTS). Henceforth, we refer to these as Cornell markers
and NIVTS markers, respectively.
Linkage analysis was performed using the JoinMap®
program version 4 (Van Ooijen 2006). The map positions
of the Cornell markers reported by Fulton et al. (2002) were
used as a frame for the linkage analysis. The genotyped
markers were roughly classiWed into 12 linkage groups,
which corresponded to the previous EXPEN 2000 map,
using the grouping module of JoinMap® and based on an
LOD score of 4.0–10.0. Marker order and genetic distance
were calculated using a regression mapping algorithm with
the following parameters: Kosambi’s mapping function,
recombination frequency · 0.35, LOD score ¸ 2.0. Geno-
type probabilities which show possible genotyping and data
entry errors were investigate for all the mapped markers
and presented as ¡log10(P). Euchromatic and heterochro-
matic regions in each linkage group were assumed based on
the previously reported positions of markers anchored to
heterochromatin (Frary et al. 2005; Ohyama et al. 2009;
Tang et al. 2008; Wang et al. 2006).
Results
Development of SSR markers
A total of 7,599 SSR markers were generated by in silico
data mining of 83,785 sequences, as described in the
“Materials and methods”, and designated TES (Tomato
EST-SSR) markers. Of these generated markers, those cor-
responding to the Cornell and NIVTS markers were
excluded from the TES marker group. Of the SSRs in the
TES marker group, 6,043 (80%) were trinucleotide repeats,
while 525 (7%) were dinucleotide repeats and 1,031 (13%)
were tetranucleotide repeats (Table 1). The poly (AAG)n
motif was the most abundant of the trinucleotide repeats
(1,712 SSRs, 28%), followed by poly (ATC)n (919 SSRs,
15%), poly (AGC)n (708 SSRs, 12%) and poly (AAC)n
(641 SSRs, 11%). While three types of dinucleotide repeats
were observed, poly (AT)n and poly (AG)n were the most
abundant and comprised 96% of the dinucleotide repeats.
Among the tetranucleotide repeats, AT-rich motifs, namely
poly (AAAT)n, poly (AAAG)n and poly (AAAC)n, were
more frequently observed than other motifs and when com-
bined, comprised up to 59% of the tetranucleotide repeats.
In addition to the TES markers, genome-derived SSR
markers were generated by extracting SSRs from the
90,763 BAC-end sequences retrieved from the SGN data-
base. A total of 13,501 primer pairs were designed to
amplify the SSRs, which were subsequently named TGS
(Tomato Genome-SSR) markers. Among the TGS markers,
7,005 of the SSRs (52%) were trinucleotide repeats, while
2,338 (17%) were dinucleotide and 4,158 (31%) were tetra-
nucleotide repeats (Table 1). Poly (AAT)n was the most
abundant trinucleotide repeat (2,517 SSRs, 36%), followed
by poly (AAG)n (1,771 SSRs, 25%), poly (AAC)n (1,012
SSRs, 14%) and poly (ATC)n (691 SSRs, 10%). Four types
of dinucleotide repeats were observed in the SSRs of the
BAC-end sequences, and poly (AT)n was the most
frequently observed motif (71% of the identiWed dinucleo-
tide repeats). Similar to the TES markers, AT-rich motifs123
Page 4
734 Theor Appl Genet (2010) 121:731–739were more frequently observed in the tetranucleotide
repeats of the TGS markers. The combination of poly
(AAAT)n, poly (AAAG)n and poly (AATT)n motifs com-
prised 86% of the tetranucleotide repeats.
From all the generated SSR markers, 2,047 TES and
3,510 TGS markers were examined for polymorphisms
between the parents of the EXPEN mapping population.
A total of 451 TES markers (22.0% of all tested mark-
ers) and 229 TGS markers (6.5%) showed co-dominant
polymorphisms between the parents. In addition, 313
TES markers (15.3% of all tested) and 566 TGS markers
(16.1%) showed dominant polymorphisms. Therefore,
41% [313/(451 + 313)] of all TES markers and 71%
[566/(229 + 566)] of all TGS markers were dominant. In
most of the dominant markers, null alleles were
observed in S. pennellii (LA716). The polymorphic
ratios for the di-, tri- and tetranucleotide repeats among
the TES markers were 13.5, 75.7 and 10.9%, respec-
tively, while those for the TGS markers were 40.8, 43.8
and 15.5%, respectively.
Table 1 Numbers of SSR
motifs in the TES and TGS
markers
Motif TES TGS
Designed (%) Mapped (%) Designed (%) Mapped (%)
Dinucleotide AC 61 (0.8) 4 (0.6) 272 (2.0) 27 (4.3)
AG 229 (3.0) 39 (6.0) 408 (3.0) 35 (5.5)
AT 235 (3.1) 44 (6.8) 1,653 (12.2) 192 (30.3)
GC 0 (0.0) 0 (0.0) 5 (0.0) 0 (0.0)
Trinucleotide AAC 641 (8.4) 56 (8.6) 1,012 (7.5) 38 (6.0)
AAG 1,712 (22.5) 147 (22.7) 1,771 (13.1) 82 (12.9)
AAT 539 (7.1) 84 (13.0) 2,517 (18.6) 117 (18.5)
ACG 121 (1.6) 9 (1.4) 41 (0.3) 1 (0.2)
ACT 188 (2.5) 19 (2.9) 248 (1.8) 11 (1.7)
AGC 708 (9.3) 34 (5.2) 199 (1.5) 5 (0.8)
ATC 919 (12.1) 59 (9.1) 691 (5.1) 17 (2.7)
GGA 491 (6.5) 22 (3.4) 253 (1.9) 5 (0.8)
GGC 230 (3.0) 19 (2.9) 44 (0.3) 1 (0.2)
GGT 494 (6.5) 45 (6.9) 229 (1.7) 5 (0.8)
Tetranucleotide AAAC 160 (2.1) 11 (1.7) 403 (3.0) 9 (1.4)
AAAG 274 (3.6) 12 (1.9) 690 (5.1) 10 (1.6)
AAAT 272 (3.6) 20 (3.1) 1,892 (14.0) 45 (7.1)
AACG 3 (0.0) 0 (0.0) 11 (0.1) 0 (0.0)
AAGC 40 (0.5) 1 (0.2) 44 (0.3) 3 (0.5)
AATC 56 (0.7) 5 (0.8) 165 (1.2) 4 (0.6)
AATG 47 (0.6) 1 (0.2) 194 (1.4) 0 (0.0)
AATT 71 (0.9) 8 (1.2) 607 (4.5) 23 (3.6)
AGGC 2 (0.0) 0 (0.0) 1 (0.0) 0 (0.0)
AGGT 2 (0.0) 0 (0.0) 14 (0.1) 0 (0.0)
GACG 2 (0.0) 0 (0.0) 4 (0.0) 0 (0.0)
GACT 5 (0.1) 1 (0.2) 11 (0.1) 0 (0.0)
GAGC 8 (0.1) 0 (0.0) 6 (0.0) 0 (0.0)
GATC 9 (0.1) 2 (0.3) 4 (0.0) 0 (0.0)
GGAC 2 (0.0) 0 (0.0) 4 (0.0) 0 (0.0)
GGAT 27 (0.4) 0 (0.0) 25 (0.2) 1 (0.2)
GGCA 5 (0.1) 0 (0.0) 6 (0.0) 0 (0.0)
GGCC 2 (0.0) 0 (0.0) 1 (0.0) 0 (0.0)
GGCT 6 (0.1) 1 (0.2) 3 (0.0) 0 (0.0)
GGGA 22 (0.3) 2 (0.3) 36 (0.3) 1 (0.2)
GGGC 0 (0.0) 0 (0.0) 3 (0.0) 1 (0.2)
GGGT 16 (0.2) 3 (0.5) 34 (0.3) 1 (0.2)
Total 7,599 (100) 648 (100) 13,501 (100) 634 (100)123
Page 5
Theor Appl Genet (2010) 121:731–739 735Development of TEI (tomato EST-derived Intronic
polymorphism) markers
A total of 589 unigenes retrieved from a KTU2 unigene set
in the MiBASE database were aligned with A. thaliana
genome sequences using the GAP2 program (Huang 1994).
By taking the GU-AG rule into account, the positions of
1,073 introns were predicted for 206 of the unigene
sequences. The mean number of predicted introns per sin-
gle unigene was 5.2. Primer pairs were designed using
sequences of the Xanking regions of the predicted intron
positions to generate DNA markers named TEI (tomato
EST-derived intronic polymorphism).
A total of 674 TEI markers were tested for polymor-
phisms using HRM analysis, as described in the “Materials
and methods”, and 537 of them (80%) showed polymor-
phisms between the mapping parents. These 537 TEI mark-
ers represented 166 independent unigenes. Therefore, 166
non-redundant TEI markers were selected for subsequent
linkage analysis. Of these selected TEI markers, 148 and 18
showed co-dominant and dominant polymorphisms
between the parents, respectively. Though the SNPs were
identiWed by HRM, it is sometimes not possible to identify
polymorphisms between heterozygous and one parental
speciWc homozygous SNPs, due to the Xanking sequences
of the SNPs. These kinds of SNPs were scored as dominant
SNPs.
Construction of a linkage map
Segregation data were generated for a total of 1,725 mark-
ers (764 TES, 795 TGS and 166 TEI markers) in the
Tomato-EXPEN 2000 mapping population. These data
were combined with data derived from 547 Cornell and 136
NIVTS markers and subjected to linkage analysis. The map
positions of the 547 Cornell markers reported by Fulton
et al. (2002) were used as a frame for the construction of
the linkage map. Subsequently, 2,116 loci, including 634
TGS, 648 TES, 151 TEI, 36 NIVTS and 547 Cornell loci,
were mapped onto 12 linkage groups corresponding to the
12 chromosomes (Chrs), while 292 loci, including 161
TGS, 116 TES and 15 TEI loci, were excluded from the
analysis. The total length of the linkage groups was
1,503.1 cM, as shown in Table 2, Fig. 1 and the Supple-
mentary Table. The total number of newly mapped loci was
1,433, ranging from 105 to 160 in each linkage group. The
average distance between two loci was 0.71 cM ranging
from 0.59 cM (Chr 9) to 0.88 cM (Chr 5).
Of the 2,116 loci mapped in this study, 1,481 and 635
loci were assumed to be located in euchromatic and hetero-
chromatic regions, respectively, according to the position
of markers anchored to heterochromatin reported in previ-
ous studies (Frary et al. 2005; Ohyama et al. 2009; Tang
et al. 2008; Wang et al. 2006). The markers used as anchor
markers of euchromatic and heterochromatic regions are
listed in the Supplemental Table. In the present map, 48%
(305) of the mapped TGS loci were predicted to be present
in the heterochromatic regions, while only 18% (119) and
21% (32) of the TES and the TEI loci were in the same
regions.
Segregation distortion was observed for 38.2% of the
mapped marker loci (Fig. 1, Table 2). The distortion ratios
varied from chromosome to chromosome; Chr 4, 5, 8, 9 and
12 showed distortions for less than 10% of the mapped loci,
while Chr 1, 10 and 11 showed distortions for more than
70% of the loci. Heterochromatin-speciWc segregation
distortion was observed for Chr 2 and 7. The ‘LA716’
(S. pennellii) genotypes were more frequently observed in
the distorted loci mapped to Chr 1, 2, 3, 6, 10 and 11, while
more ‘LA925’ (S. lycopersicum) genotypes were identiWed
in the distorted loci mapped to Chr 7.
Discussion
In this study, we developed three types of DNA markers
designated TES, TGS and TEI. The TES markers that con-
tained trinucleotide repeats exhibited a higher ratio of poly-
morphisms than those containing di- and tetranucleotide
repeats. On the other hand, a signiWcant diVerence was not
observed in the frequency of polymorphisms between di-
and trinucleotide repeats in TGS markers. Dinucleotide
repeats in coding regions often causes critical changes, such
as frame-shift mutations. The decreased number of poly-
morphisms present in the dinucleotide repeats in TES
markers in comparison to TGS markers suggested that the
coding region contained higher sequence conservation than
the intergenic regions of S. lycopersicum and S. pennellii.
Among the polymorphic markers, the percentage of domi-
nant TES markers was 41% [313/(451 + 313)], while that
of dominant TGS markers was 71% [566/(229 + 566)]. In
most of the dominant markers, null alleles were observed in
S. pennellii. As the primers were designed from the S. lyco-
persicum genome and EST sequences, sequence divergence
between the two species might cause poor annealing of the
primers to the S. pennellii genomic DNA. The results also
indicate that sequence conservation is higher in the intra-
genic regions than the intergenic regions.
A segregation distortion ratio was observed for more
than 70% of loci mapped to Chr 1, 10 and 11, while in less
than 10% of loci mapped to Chr 4, 5, 8, 9, and 12. Segrega-
tion distortion or transmission ratio distortion (TRD) in a
hybrid population can result from various factors, such as
hybrid lethality or sterility of gametophytic competition
(Harushima et al. 2001). The bias of segregation distortion
can be explained by the presence of more TRD factors on123
End of preview.
