Theor Appl Genet
DOI 10.1007/s00122-009-1244-6
ORIGINAL PAPER
A single recessive gene controls cadmium translocation
in the cadmium hyperaccumulating rice cultivar Cho-Ko-Koku
Kouichi Tezuka · Hidenori Miyadate · Kazunao Katou · Ikuko Kodama · Shinichi Matsumoto ·
Tomohiko Kawamoto · Satoshi Masaki · Hideki Satoh · Masayuki Yamaguchi · Kenji Sakurai ·
Hidekazu Takahashi · Namiko Satoh-Nagasawa · Akio Watanabe · Tatsuhito Fujimura · Hiromori Akagi
Received: 18 September 2009 / Accepted: 8 December 2009
© Springer-Verlag 2009
Abstract The heavy metal cadmium (Cd) is highly toxic
to humans and can enter food chains from contaminated
crop Welds. Understanding the molecular mechanisms of Cd
accumulation in crop species will aid production of safe
Cd-free food. Here, we identiWed a single recessive gene
that allowed higher Cd translocation in rice, and also deter-
mined the chromosomal location of the gene. The Cd
hyperaccumulator rice variety Cho-Ko-Koku showed 3.5-
fold greater Cd translocation than the no-accumulating
variety Akita 63 under hydroponics. Analysis of an F2 pop-
ulation derived from these cultivars gave a 1:3 segregation
ratio for high:low Cd translocation. This indicates that a
single recessive gene controls the high Cd translocation
phenotype. A QTL analysis identiWed a single QTL,
qCdT7, located on chromosome 7. On a Cd-contaminated
Weld, Cd accumulation in the F2 population showed contin-
uous variation with considerable transgression. Three QTLs
for Cd accumulation were identiWed and the peak of the
most eVective QTL mapped to the same region as qCdT7.
Our data indicate that Cd translocation mediated by the
gene on qCdT7 plays an important role in Cd accumulation
on contaminated soil.
Communicated by M. Xu.
Electronic supplementary material The online version of this
article (doi:10.1007/s00122-009-1244-6) contains supplementary
material, which is available to authorized users.
K. Tezuka · H. Miyadate · K. Sakurai · H. Takahashi ·
N. Satoh-Nagasawa · A. Watanabe · H. Akagi
Laboratory of Plant Genetics and Breeding,
Department of Biological Production,
Faculty of Bioresource Sciences,
Akita Prefectural University,
Kaidoubata-Nishi 241-438, Shimoshinjyo-Nakano,
Akita 010-0195, Japan
K. Katou · I. Kodama · S. Matsumoto · T. Kawamoto · S. Masaki
Akita Agricultural Experiment Station,
Genpachizawa 34-1, Aikawa, Yuwa,
Akita 010-1231, Japan
H. Satoh · M. Yamaguchi
Department of Paddy Farming,
National Agricultural Research Center for Tohoku Region,
Yotsuya, Daisen 014-0102, Japan
T. Fujimura
Institute of Agricultural and Forest Engineering,
University of Tsukuba, Tennoudai 1-1-1, Tsukuba,
Ibaraki 305-8572, Japan
H. Akagi (&)
Laboratory of Plant Breeding and Genetics,
Department of Biological Production,
Faculty of Bioresource Sciences,
Akita Prefectural University,
Kaidoubata-Nishi 241-438, Shimoshinjyo-Nakano,
Akita 010-0195, Japan
e-mail: akagi@akita-pu.ac.jp
123
Theor Appl Genet
Introduction
Cadmium (Cd) is one of the most toxic heavy metals with
respect to human health, especially in Cd-contaminated
areas, where it can enter the food chain. Cd is often present
in deposits of valuable heavy metals, such as copper, silver,
and zinc, and mining activities can release Cd into the envi-
ronment, where it may remain as a contaminant for a num-
ber of years. The contamination of crop Welds with Cd is an
important problem, particularly for paddy Welds irrigated
with water that passes through mining sites. Crops grown
on Cd-polluted soils take up and accumulate Cd, with
potentially devastating consequences for human health.
Phytoremediation is a promising and environmentally
friendly approach for removing Cd pollution from soils.
The plant species selected for use in phytoremediation
accumulate Cd at a high level, especially in their aerial
parts. Two such species are Thlaspi caerulescens and Ara-
bidopsis halleri, which can accumulate more than
100 mg kg¡1 shoot dry weight of Cd under Weld conditions
(Baker et al. 2000; Bert et al. 2002). In addition to their
ability to accumulate Cd, the selected species also need to
show rapid growth and large biomass productivity. Some
rice (Oryza sativa) cultivars are promising candidates for
phytoremediators of Cd contaminated paddy Welds (Ibaraki
et al. 2009; Murakami et al. 2009): they obviously grow on
the same habitat as crop cultivars, can accumulate Cd in
their aerial parts at relatively high concentrations, and can
produce a large biomass (Arao and Ae 2003; Liu et al.
2003; Ishikawa et al. 2005a). The cultivation of rice,
including mechanical planting and harvesting, is obviously
well established.
Understanding the mechanisms of accumulation of Cd in
the aerial part of plants will be indispensable for creating
more eYcient hyperaccumulators. Plants are presumed to
uptake Cd from the soils through their root systems, load it
to the xylem and then transport it to the aerial parts (Clem-
ens 2006). It is therefore likely that several genes involved
in these transport processes control Cd accumulation in
plants. A recent study on the hyperaccumulator A. halleri,
in comparison with the model plant A. thaliana, revealed
that xylem loading is controlled by HMA4, which enables
the hyperaccumulation of Cd in shoots (Courbot et al.
2007; Hanikenne et al. 2008).
The genus Oryza shows considerable genetic diversity
among varieties and also displays large diVerences in abil-
ity to concentrate Cd in shoots and brown rice (Morishita
et al. 1987; Arao and Ae 2003; Liu et al. 2003; Uraguchi
et al. 2009). Such genotypic and phenotypic variations pro-
vide a valuable resource for identifying genes that confer
the ability to hyperaccumulate Cd. Using chromosome seg-
ment substitution lines of Koshihikari and Kasalath, which
accumulate diVerent Cd concentrations in their brown rice,
three chromosome segments were identiWed that enable
accumulation of high Cd concentrations (Ishikawa et al.
2005b). Recently, three putative QTLs controlling shoot Cd
concentration were identiWed in an F2 population between
the Badari Dhan and Shwe War cultivars that have diVerent
abilities to concentrate Cd in their shoots (Ueno et al.
2009). However, the speciWc genes involved were not iden-
tiWed and, therefore, the molecular mechanisms of Cd accu-
mulation in rice plants remained uncertain.
On Cd-contaminated soil, the indica rice cultivar
Cho-Ko-Koku was found to accumulate the metal at a
much higher level than other tested cultivars (Matsumoto
et al. 2005; Uraguchi et al. 2009; Murakami et al. 2009). We
therefore selected this cultivar for use in our study on the
mechanisms of Cd hyperaccumulation in rice grown under
deWned hydroponic conditions. In this study, we identiWed a
single recessive gene (named qCdT7) located on chromo-
some 7 that conferred a high Cd translocation ability to
Cho-Ko-Koku. We also found that the QTL with the greatest
inXuence on Cd accumulation in rice grown in a Cd contami-
nated Weld also mapped to the same location as qCdT7. From
these results, we conclude that the gene located at the qCdT7
locus will be of importance for phytoremediation.
Materials and methods
Plant materials
Two rice (Oryza sativa L) cultivars, Cho-Ko-Koku (indica
cultivar) and Akita 63 (japonica cultivar), were used in this
study. Cho-Ko-Koku and Akita 63 are high and low Cd
accumulation varieties, respectively. An F2 population
between Cho-Ko-Koku and Akita 63 was used for a QTL
analysis of genes associated with Cd accumulation. One F2
plant carrying heterozygous allele for qCdT7 was selected
from the F2 population. As much as 20 seedlings were
selected from the self-pollinated F3 progeny of the F2 plant
by analyzing the genotypes of RM6776 and RM5436 of
their leaf tips and were used for analysis of Cd translocation
rate.
Hydroponic culture and Cd treatment
Rice seeds were surface sterilized with 5% sodium hypo-
chlorite, rinsed with water, and then incubated for 24 h at
32°C. Germinating seeds were transferred onto a nylon
mesh Xoating on a nutrient solution (pH 5.4) containing
0.23 mM MgSO4, 0.18 mM NH4NO3, 0.18 mM CaCl2,
0.14 mM K2SO4, 0.09 mM Na2HPO4, 0.09 mM SiO2,
22.5 �M Fe(III)-EDTA, 9.2 �M H3BO3, 2.3 �M MnSO4,
0.78 �M CuSO4,0.77 �M ZnSO4, and 0.5 �M (NH4)6
Mo7O24. The seedlings were cultured for 10 days in a
123
Theor Appl Genet
growth cabinet (14 h light at 26°C, 10 h dark at 22°C).
Each seedling was then transferred to a glass tube
(�35 mm £ 20 cm) covered with aluminum foil and con-
taining 200 mL of the nutrient solution supplemented with
1 �g of CdCl2 and was cultured in the growth cabinet for
20 days or 27 days (for the F3 seedlings). The nutrient solu-
tion was renewed every 2 days. At the end of the culture
period, the roots of the seedlings were washed with deion-
ized water 3 times. The seedlings were divided into shoot
and root tissues and dried at 105°C for 24 h. The shoot and
root tissues were weighed and then digested with 12 mL of
HNO3–HClO4 (2:1 v/v) mix solution. Inductively coupled
plasma-atomic emission spectrometry (ICP-AES) (Nippon-
Jarrell-Ash, Tokyo, Japan) was used to determine the
concentrations of Cd, Fe, Mg, Mn, and Zn in the digest
solutions. The Cd translocation rate was estimated as the
percentage of cadmium in the shoot compared to the whole
plant.
Plants grown on a Cd-polluted paddy Weld
Rice seeds of the Cho-Ko-Koku and Akita 63 cultivars and
of the F2 population were sown on 21st April 2008, and
seedlings were transplanted to the Cd-polluted paddy Weld
located in the Akita prefecture the northern part of Japan
with spacing 15 cm £ 15 cm on 28th May, and the soil in
this paddy Weld classiWed as gray lowland soil (glysol). The
irrigated water was drained on 29th July. The Cd concen-
tration in the soil of the paddy Weld was 1.23 mg kg¡1
(0.1 M HCl). Akita 63 and Cho-Kou-Koku headed on 12th
and 20th August, respectively. Most of F2 individuals
headed within August. We selected F2 plants which had
headed between 9th and 29th August, and harvested on 1st
October. Fertilizer was applied before planting: N,
70 kg ha¡1; P, 100 kg ha¡1; and K, 100 kg ha¡1. The har-
vested plants were divided into shoots (leaves and stem)
and ear parts. After weighing the dried materials, samples
were ground to a powder, then 0.5 g was digested in 12 mL
of HNO3–HClO4 (2:1 v/v) mixture. The Cd concentrations
in the digest solutions were determined with the ICP-AES.
DNA extraction and analysis with SSR markers
Total DNA was extracted from the leaf tips of F2 seedlings
according to the method described previously (Kato et al.
2007). Leaves were dried at 70°C for 2 h and then ground
to powder with a stainless ball (�3 mm) on a Micro Smash
(MS-100 TOMY) vibrator. An extraction buVer (Edwards
et al. 1991) was added to the powder and the mixture was
incubated at room temperature for 1 h. DNA was precipi-
tated from the mixture by adding an equal volume of 2-pro-
panol and used in the PCR analyses.
The PCR was carried out in a 20 �l of reaction volume
consisting of 10 mM Tris–HCl (pH 8.3), 50 mM KCl,
1.5 mM MgCl2, 1 U of TAKARA Taq HS (TAKARA),
4 nmol dNTP, 10 ng of genomic DNA, and 10 pmol of
each set of primers for SSR markers in a Thermal Cycler
9600 or 9700 (Perkin-Elmer, Foster City, Calif.). As much
as 35 PCR cycles, each consisting of 10 s of denaturation at
94°C, 30 s of annealing at 55°C, and 1 min of polymeriza-
tion at 72°C, were performed. Polymorphisms for SSR
markers were identiWed using 2.5% MetaPhor Agarose gels
(FMC, Rockland, Me).
Construction of linkage map and QTLs analysis
Genetic linkage maps for the F2 populations grown either
under hydroponic conditions or in the paddy Weld were con-
structed with 114 SSR markers (Akagi et al. 1996; Tem-
nykh et al 2001; McCouch et al. 2002; International Rice
Genome Sequencing Project 2005) using Mapmaker ver.2
(Lander et al. 1987). QTL analyses for both F2 populations
were performed using the composite interval mapping
method with Windows QTL Cartographer ver. 2.5 (Wang
et al. 2007). An LOD value of 2.5 was used as the threshold
for detection of potential QTLs inXuencing Cd concentra-
tions in the shoot or root, and for the translocation rate
under hydroponic culture conditions.
Results
Characteristics of Cd accumulation in the Cd
hyperaccumulator variety Cho-Ko-Koku
Analysis of Cd uptake in rice grown in a paddy Weld with
moderate contamination by Cd (1.33 mg kg¡1) showed that
shoots of Cho-Ko-Koku averaged 0.88 mg plant¡1 com-
pared to 0.17 mg plant¡1 for Akita 63. This diVerence in
uptake was mirrored by a diVerence in Cd concentrations
between the cultivars: Cho-Ko-Koku, 30.6 mg kg¡1 DW;
Akita 63, 4.03 mg kg¡1 DW. Our results indicate that
Cho-Ko-Koku has a higher rate of Cd translocation to the
shoots and accumulates Cd at a higher concentration (more
than 5-fold greater) than Akita 63.
No diVerence was observed in total amount of Cd uptake
by the seedlings of Cho-Ko-Koku and Akita 63 after
20 days growth under hydroponic conditions (Fig. 1). How-
ever, the shoots of Cho-Ko-Koku seedlings accumulated
4.99 �g of Cd compared to 1.41 �g in Akita 63 (Fig. 1). In
contrast, Cd accumulation in the roots of Akita 63 seedlings
was 2.6-fold higher than those of Cho-Ko-Koku (Fig. 1).
Thus, Cho-Ko-Koku seedlings were characterized by a high
Cd translocation ability from the root to shoot not by a
123
Theor Appl Genet
high Cd uptake ability when grown under hydroponic
conditions.
Next, we compared the rates of translocation of metals
including Cd in the seedlings (Fig. 2). Akita 63 seedlings
retained 80% of the Cd in the root but Cho-Ko-Koku seed-
lings translocated 70% of the Cd to the shoot (Fig. 2). In
contrast, both cultivars translocated between 60 and 90% of
Fe, Mg, Mn, and Zn to the shoot and there were no obvious
diVerences between the two cultivars (Fig. 2). These Wnd-
ings suggest that Akita 63 seedlings have a speciWc mecha-
nism that causes retention of Cd in the root.
Segregation of Cd translocation rates in the F2 population
We investigated the genetic control of the diVerent rates of
accumulation of Cd in the two cultivars by an analysis of an
F2 population between Cho-Ko-Koku and Akita 63. In the
F2 population, the Cd accumulation rate in the shoot was
highly correlated with the Cd translocation rate (r2 = 0.96).
Moreover, the F2 population could be divided into higher
and lower accumulation rate groups. Of the 144 F2 individ-
uals, 36 showed the higher (55–79%) Cd translocation rates
similar to Cho-Ko-Koku, whereas 102 showed lower (12–
48%) Cd translocation rates similar to Akita 63 (Fig. 3).
Segregation rate was signiWcantly Wt to 1:3 ratio (high:low)
(�2 = 0.084 for 1:3, P = 0.768). A similar frequency distri-
bution with two peaks was observed for Cd accumulation
and concentration in the shoot (Figs. S1, S2). In contrast to
shoots, the Cd concentration in the roots showed continu-
ous variation in the F2 population (Fig. S2). The 1:3 Men-
delian inheritance seen in the F2 population indicates that
the higher Cd translocation present in Cho-Ko-Koku is
controlled by a single recessive gene.
Analysis of the gene increasing Cd translocation
The chromosomal location of the recessive gene for Cd
translocation was determined by a QTL analysis of the F2
population derived from Cho-Ko-Koku and Akita 63. A
QTL with a LOD value of 68.6 was detected on chromo-
some 7 (Table 1). At the same position, QTLs were also
detected for Cd accumulation and concentration both in
shoot and in root (Table 1). The QTL controlling Cd trans-
location explained 88% of the phenotypic variation. The
dominance eVect of the QTL indicated that only the homo-
zygous genotype for the Cho-Ko-Koku allele increased the
rate of Cd translocation (Table 1). Since the QTL character-
istics were consistent with the results of the segregation
analysis, there is a gene responsible for Cd translocation
between RM6776 and RM5436 on chromosome 7, and the
locus of the gene was named as qCdT7 (Table 1, Fig. 7).
The F3 plants were classiWed into three types based on
their genotypes of qCdT7 using markers, RM6776 and
RM5436. The homozygotes for the Cho-Ko-Koku allele
showed signiWcantly higher Cd translocation rate
(68.3 § 7.97) than either the heterozygotes (40.4 § 5.19)
Fig. 1 Cadmium contents in shoots and roots of Cho-Ko-Koku and
Akita 63 seedlings grown under hydroponic culture conditions. A rel-
atively low level of CdCl2 (5 �g L¡1) was applied for 20 days. Black
and white bars indicate Cho-Ko-Koku and Akita 63, respectively. Bars
represent standard deviations of 6 seedlings. ** shows the level of
signiWcance at P < 0.01
Fig. 2 Metal translocation rate from roots to shoots. Rice seedlings
were grown in a nutrient solution containing (5 �g L¡1) of CdCl2 for
20 days. The metal translocation rate was estimated as the percentage
of cadmium in the shoot compared to the whole plant. Black and white
bars indicate Cho-Ko-Koku and Akita 63, respectively. Bars represent
standard deviations of 6 seedlings. ** and * shows the levels of
signiWcance at P < 0.01 and 0.05, respectively
Fig. 3 Segregation pattern of Cd translocation rates in the F2 popula-
tion. The 144 F2 seedlings were grown in a nutrient solution with
5 �g L¡1 of CdCl2 for 20 days. Means and standard deviations for the
parental translocation rates are also included. CKK and A63 represent
Cho-Ko-Koku and Akita 63, respectively
123
Theor Appl Genet
or homozygotes for the Akita 63 allele (33.2 § 2.55; P = 0;
Fig. 4). No signiWcant diVerence for Cd translocation rate
was observed between F3 plants heterozygous and homozy-
gous for the Akita 63 allele, indicating that Cho-Ko-Koku
allele for qCdT7 increase Cd translocation *** as a reces-
sive manner (Fig. 4). This result conWrmed the existence of
a gene increasing Cd translocation from root shoot at the
locus of qCdT7 on chromosome 7.
Another QTL aVecting Cd concentration in the root was
detected on chromosome 1; in this instance, the Cho-Ko-
Koku allele decreased the concentration of Cd. Because the
Cd concentration varied only in roots continuously under
the hydroponic condition but not in shoots (Figs. S1, S2),
this QTL promoted the root Cd concentration quantitatively
in addition to environmental eVects (Table 1, Fig. S2).
Contribution of qCdT7 to Cd accumulation in plants
grown in a Cd-polluted paddy Weld
We performed a QTL analysis for Cd accumulation in an F2
population grown on a Cd-polluted paddy Weld to determine
the eVect of qCdT7 under normal culture conditions. In
shoots of the F2 population, Cd accumulation showed a
continuous and wide variation ranging from 0.04 to
2.14 mg plant¡1 (Fig. 5). A similar frequency distribution
was also observed for shoot Cd concentration and shoot dry
weight (Fig. S3, Fig. 4). As much as 43 segregants showed
greater accumulation of Cd than the parental cultivar
Cho-Ko-Koku, one of these segregants accumulated Cd at a
2.5-fold greater rate than Cho-Ko-Koku (Fig. 5). The Cd
accumulation in shoot was highly correlated with Cd concen-
tration in shoot in the F2 population grown on Cd-polluted
paddy Weld (r = 0.80) (Fig. 6). The low correlation coeYcient
(r = 0.26) between Cd accumulation and dry weight in shoot
was observed. In this population, either Cd concentration or
dry weight in shoots produced higher Cd accumulating plants
than Cho-Ko-Koku (Fig. 6). The transgressive segregants in
Cd accumulation should be feasible to breed rice varieties
with a Cd accumulation performance that exceeds Cho-Ko-
Koku which accumulated Cd at a much higher level in rice.
Three QTLs for Cd accumulation in shoots were detected
on chromosomes 4, 7, and 12 (Table 2). The Cho-Ko-Koku
alleles of the QTLs on chromosomes 7 and 12 increased shoot
Cd accumulation, as did the Akita 63 allele on chromosome 4
Table 1 Quantitative trait loci for cadmium accumulation and translocation
AE and DE indicate additive and dominance eVects of the Cho-Ko-Koku allele, respectiv