Quantitative Trait Loci Controlling Resistance to Gray Leaf Spot in Maize

Physiologia Plantarum 136: 395–406. 2009 Copyright © Physiologia Plantarum 2009, ISSN 0031-9317

Quantitative trait loci controlling resistance to cadmiumrhizotoxicity in two recombinant inbred populations ofArabidopsis thaliana are partially shared by those forhydrogen peroxide resistanceTanveer Taziba, Takashi Ikkaa, Keishi Kurodaa, Yuriko Kobayashib, Kazuhiko Kimurac, and HiroyukiKoyamaa,*

aLaboratory of Plant Cell Technology, Faculty of Applied Biological Sciences, Gifu University, 501-1193, Gifu, JapanbBioResources Center, RIKEN, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, JapancDepartment of Farm Management, Miyagi University, 981-3298, Sendai, Miyagi, Japan

Correspondence*Corresponding author,e-mail: [email protected]

Received 2 February 2009; revised 12

March 2009

doi: 10.1111/j.1399-3054.2009.01234.x

To understand mechanisms of cadmium (Cd) tolerance variation associatedwith root elongation in Arabidopsis thaliana, quantitative trait loci (QTLs) andepistasis were analyzed using relative root length (RRL: % of the root lengthin +Cd to −Cd) as a tolerant index. Using the composite interval mappingmethod, three major QTLs (P < 0.05) were detected on chromosomes 2, 4and 5 in the recombinant inbred population derived from a cross betweenLandsberg erecta (Ler−0) and Columbia (Col-4). The highest logarithm ofodds (LOD) of 5.6 was detected with the QTL on chromosome 5 (QTL5),which is linked to the genetic marker CDPK9 and explained about 26% of theCd tolerance variation. There was no significant difference in Cd-translocationratio from roots to shoots between tolerant and sensitive recombinant inbreedlines (RILs), while greater accumulations of reactive oxygen species wereobserved in the roots of sensitive RILs. This suggested that accumulation ofROS would explain Cd tolerance variation of the Ler/Col RILs, which is mainlycontrolled by the QTL on chromosome 5. The QTL5 in Ler/Col populationwas also detected as one of the major QTLs controlling tolerances to hydrogenperoxide and to copper, which is another ROS generating rhizotoxic metal.The same chromosomal position was detected as a common major QTL forCd and hydrogen peroxide tolerances in a different recombinant inbreed(RI) population derived from a cross of Col-gl1 and Kashmir (Kas-1). Thesedata, along with a multitraits QTL analysis in both sets of RILs, suggest thatperoxide damage depends on the genotype at a major Cd-tolerant locus onthe upper part of chromosome 5.

Abbreviations – APX, ascorbate peroxidase; Cd, cadmium; CAPS, cleaved amplified polymorphic sequence; CIM, compositeinterval mapping; CV, coefficient of variation; DHE, dihydroethidium; GR, glutathione reductase; GSH, glutathione; GST,glutathione S-transferase; H2DCFDA, 2’,7’-dichlorodihydrofluorescein diacetate; HMA4, heavy metal ATPase 4; HMA5, heavymetal ATPase 5; LOD, logarithm of odds; QTLs, quantitative trait loci; QTL5, QTL on chromosome 5; RI, recombinant inbreed;RIL, recombinant inbreed line; RRL, relative root length; SOD, superoxide dismutase.

Physiol. Plant. 136, 2009 395

Introduction

Cadmium (Cd) is a non-essential heavy metal releasedinto the environment from either anthropogenic or non-anthropogenic sources. It is recognized as a harmfulpollutant because of its high toxicity and large solubilityin water (Pinto et al. 2004). Cd is the fifth most toxicmetal to vertebrates, and the fourth most toxic metal tovascular plants owing to its toxic effects on growth anddevelopment (Benavides et al. 2005). Cd inhibits rootelongation (Arduini et al. 1994) and is also responsiblefor the depletion of glutathione (GSH) (Schutzendubeland Polle 2002). Minimizing Cd concentration in theedible part is the primary target for crop breeding in Cd-contaminated soil, in terms of food security. At the sametime, improvement of tolerance against Cd rhizotoxicitymight be another important target. In this context, it isimportant to understand the mechanism of Cd tolerancein the root, because poor development of the roots mightcause growth inhibition of shoots.

Plants possess diverse resistance mechanisms to heavymetal toxicities, which can be categorized into twogroups, namely, exclusion and compartmentalizationof toxic metals or repair of damage caused by theadverse effects of heavy metals (Sanita di Toppiand Gabbrielli 1999). Pomponi et al. found thatoverexpression of Arabidopsis thaliana phytochelatinsynthase in tobacco plants enhances Cd tolerance andaccumulation (Pomponi et al. 2006). Cd phytotoxicitywas found to induce hydrogen peroxide accumulationand subsequent oxidative stress (Cho and Seo 2005,Metwally et al. 2005). In this case, the resistancemechanisms include enhanced activities of enzymesthat can scavenge reactive oxygen species (Cho andSeo 2005), such as, superoxide dismutase (SOD),ascorbate peroxidase (APX) and glutathione reductase(GR). Quantitative trait locus (QTL) analysis would beone useful approach to understand such a complexsystem at a more molecular level.

Using natural variations is a useful way to studymetal rhizotoxicity, even in case of cadmium (Roosenset al. 2003). QTL analysis is widely used for identifyingchromosomal localization and genes whose functionsinfluence quantitative traits (Ross-Ibarra et al. 2007). Thisapproach has been applied to understand mechanism(s)of toxicity and tolerance to rhizotoxic ions. Several QTLsresponsible for factors affecting root growth [e.g. rootelongation rate (Sergeeva et al. 2006); growth rate andarchitecture (Loudet et al. 2005)] and its interaction forrhizotoxicities [e.g. lanthanides tolerance (Kobayashiet al. 2007b); aluminum tolerance (Hoekenga et al.2003, Kobayashi and Koyama 2002)] have beenidentified. Some of these analyses identified the gene

responsible for the QTL. For example, a QTL study oncopper tolerance of roots showed that a heavy metaltransporting ATPase (heavy metal ATPase 5, HMA5) isinvolved in the mechanism(s) of Cu-tolerance variationin A. thaliana (Kobayashi et al. 2008). QTL analysiswould be a useful approach to understand the complexmolecular mechanisms of tolerance to Cd rhizotoxicity.

One gene critical for Cd tolerance of the roots wasidentified using recombinant inbred lines made bya cross between A. thaliana and its relative species[i.e. Cd tolerant and hyper-accumulator, Arabidopsishalleri (Courbot et al. 2007)]. In this case, a generegulating the capacity of Cd translocation, heavy metalATPase 4 (HMA4) was identified as a cause of amajor QTL. However, it is possible that other QTLsrelating to different mechanisms would be detectedusing genetically different materials. For example, QTLanalyses for Al tolerance with the Ler/Col (Kobayashi andKoyama 2002) and the Ler/Cvi (Kobayashi et al. 2005)recombinant inbreed lines (RILs) identified a major QTLfor Al tolerance on the upper part of chromosome 1,while another QTL was identified on chromosome 5from an analysis with the Col/Kas RILs (Ikka et al. 2008).The QTL in chromosome 1 regulates one of the major Al-tolerance mechanisms, the Al responsive malate efflux(Hoekenga et al. 2003, Kobayashi et al. 2005), whilethe QTL on chromosome 5 was suggested to regulatedifferent Al-tolerant mechanisms such as the ROS-scavenging capacity (Ikka et al. 2008). A similar QTLapproach using different mapping populations might beuseful to identify Cd-tolerant mechanism in Arabidopsisother than the Cd-translocation capacity.

In the present study, we performed combinations ofQTL analyses to identify the complex Cd tolerance mech-anisms in Arabidopsis other than the Cd-translocationcapacity. First, we identified common QTLs shared bydifferent RI populations, which were obtained fromcrosses between parental accessions possessing simi-lar levels of the Cd-translocation capacity. The sharedchromosome position of the Cd-tolerant QTL was furthercharacterized by QTL pyramiding with ROS tolerance,and several physiological analyses for Cd tolerance.These approaches revealed that a part of the Cd tol-erance in Arabidopsis is linked to its ROS-scavengingcapacity.

Materials and methods

Arabidopsis accessions

RILs of Arabidopsis made by crosses between Landsbergerecta (Ler-0) and Columbia (Col-4) (Lister and Dean1993, were obtained from the Nottingham Arabidopsis

396 Physiol. Plant.136, 2009

Stock Center (NASC), N1899 [designated as ‘set 1′ bythe NASC seed catalogue; http://nasc.life.nott.ca.uk/),which consists of 100 RILs with a dense genetic markerinformation] and N4858 [designated as ‘set 2′ by theNASC, which consists of 200 RILs from the same originwith limited genetic marker information], or Col-gl1and Kashmir (Kas-1) (Li et al. 2006, obtained from theNASC; N84999 http://arabidopsis.info/) and used forQTL analysis, and physiological experiments. Ninety-nine (Ler/Col from the set1) and 87 (Col/Kas) lineswere used for QTL analyses. In case of H2O2 stressin Ler/Col RILs, 19 lines were avoided because oftheir poor germination rates. The Col/Kas RILs weresupplied at the F6 generation and contained higher levelof heterozygous chromosomes (Li et al. 2006). Thus,we made the F8 generation to reduce the frequency ofheterozygous chromosomes and then genotyped thosefor QTL analysis in a previous study (Ikka et al. 2008).The F8 population was used in the current study.

Growth condition, staining and microscopicimaging

Hydroponic culture was performed according toour previous studies (Kobayashi and Koyama 2002,Kobayashi et al. 2005) using culture floats made ofchemically stable materials (i.e. plastics and nylon).The basic culture solution contained 2% MGRLmicronutrients (see Fujiwara et al. 1992) withoutinorganic phosphate and 200 μM CaCl2. As much as3.5 μM CdCl2 was added to the basic culture solutionin order to achieve a toxic level of Cd for bothLer/Col and Col/Kas RILs. In the case of H2O2 testsolutions, 100 μM H2O2 was added to the basic culturesolution for Ler/Col RILs and 200 μM H2O2 was usedfor Col/Kas RILs. In case of the Cu test solution, weadded Japanese Industrial Standard Cu solution foratomic absorption spectrometry [10 mg l−1 Cu(NO3)2in 0.1 M HNO3; Wako Pure Chemical] to the basaltest solution. In all three stress conditions (CdCl2, H2O2

and Cu), the same RILs were used for QTL analyses.The same set of experiments was conducted in thebasic culture solution containing no toxicant as acontrol. The test solutions were renewed every twodays, and the culture was continued for 7 days at 25◦Cwith a 12 h day (approximately 50 μmol m−2 s−1)/12 hnight cycle. Primary root length was measured asaccording to our previous experiment (Koyama et al.2001) using a multiple measure-unit (MC-300, Kenis,Tokyo) equipped with a microscope video camera (PicoScopeman, Kenis).

Ten seedlings of each RI line were grown in threeindividual repeat experiments and all data were batched.

After excluding the longest root-length value, the meanof the first five longest roots among the remaining datawere used for calculating relative root length (RRL). TheRRL (%; of the root length in treatment to the mean ofcontrol) was calculated by dividing the root length withtreatments (i.e. CdCl2, H2O2 or CuSO4) by the mean ofcontrol (i.e. mean of five root-length values obtained inthe control solution). The means of RRL of each treatmentwere used for QTL analyses.

Histochemical analyses of the roots derived fromthe Cd treatment were performed according to Ikkaet al. 2007. Briefly, growing roots (day 3–4) weretransferred to stress condition (3.5 μM of CdCl2)and treated for 30 min, 3 and 24 h. H2O2 accu-mulation caused by Cd stress was visualized using2’,7’-dichlorodihydrofluorescein diacetate (H2DCFDA)(10 μM of H2DCFDA for 10 min) and presumed super-oxide anions (O−

2 ) accumulation caused by Cd stresswas also visualized by dihydroethidium (DHE) (10 μMof DHE for 30 min) (see H2DCFDA, Behl et al. 1994;DHE, Yamamoto et al. 2002). Fluorescence in the roottip was observed using a fluorescent microscope (IMT-2-21-RFL, Olympus, Tokyo) equipped with appropriatedichroic mirror units. Images were photographed usinga digital camera (PMDC α/OL-1, Olympus).

QTL analysis

The genotypic data set of the genotypes for the Ler/ColRILs (set 1; N1899) was obtained from NASC. Thegenotype data sets for the F6 generation of Col/KasRILs were obtained from the natural variation database(http://naturalvariation.org/KasCol) (Li et al. 2006), andupgraded for the F8 generation using additional sim-ple sequence length polymorphism (SSLP) and cleavedamplified polymorphic sequence (CAPS) analyses (Ikkaet al. 2008). The genetic linkage map was processedusing MAPMAKER/EXP version 3.0b (Lander et al. 1987).Both the linkage map data and data set of the genotypesand phenotypes (mean RRL derived from three indepen-dent experiments) were then transferred to QTL CARTOG-

RAPHER version 1.13 (Basten et al. 1994, obtained fromhttp://statgen.ncsu.edu/qtlcart/WQTLCact.html). Theproportion of the phenotypic variance explained by eachQTL (R2 value) and other QTL parameters, likelihood,and location, were calculated with model six based onthe composite interval mapping (CIM) method using theZMAPQTL algorithm (Zeng 1993, 1994). To optimize thedetermination of significant QTLs, the thresholds of theLOD for each linkage group were calculated by a per-mutation test method (Churchill and Doerge 1994) with1000 permutations at a permutation significance levelof α = 0.05. The QTLs carrying higher LOD values than

Physiol. Plant.136, 2009 397

threshold were identified as major QTLs. The joint like-lihood ratio score curves for Cd, H2O2 and Cu tolerancewere obtained using the JZMAPQTL algorithm (Jiang andZeng 1995). Epistatic interactions were calculated usinga complete pair-wise search method with a significancethreshold of P < 0.0005 using EPISTAT (Chase et al. 1997).

Statistical analysis

The coefficient of variation (CV; %) for each conditionwas calculated by dividing the standard deviation by themean. Broad sense heritability (h2

b) was calculated fromRRL values (n = 5) using the following formula

h2b = σ 2

g/[(σ 2e/r) + σ 2

g]

where σ 2g is genetic variance, σ 2

e environment variance,and r number of data points employed. Significantdifferences among the means of all accessions wereestimated using one-way ANOVA followed by a leastsignificant difference (LSD) test (P < 0.05).

SSLP and CAPS analysis for confirming theidentified QTL positions for Cd treated Ler/Col RILs

Ten each of the Cd-sensitive and Cd-tolerant RILs ofLer/Col were identified from randomly selected 65 RILsfrom the ‘set 2′ RILs. Genomic DNA was isolated fromthe leaves of individual plants according to Kobayashiet al. (2007b) and genotyped using appropriate geneticmakers linked to the major QTLs for Cd tolerance byCAPS (Konieczny and Ausubel 1993) and SSLP (Belland Ecker 1994) analyses. Specific primers for GPA1,nga1139 and nga139 markers were derived from thesequence information of genetic markers obtained fromthe Arabidopsis information resource database (TAIR,http://www.arabidopsis.org/).

Integration of position of QTLs and knownCd-tolerant genes on the physical map of Colombia

Physical positions of the genetic markers of theLer/Col RIL were obtained through the TAIR databaseusing SEQVIEWER (http:://www.arabidopsis.org/servlets/sv)and Mapviewer (http://www.arabidopsis.org/servlets/mapper), while those of the Col/Kas RIL wereobtained from the website of the Naturalvariation.org(http://naturalvariation.org) according to Li et al (2006).Physical positions of cadmium sensitive (CAD), HMA4and other Cd-tolerant genes were also obtained frominformation at the above database.

Quantification of Cd in selected sensitive andtolerant Ler/Col RILs

Five (Donor number CL54, CL266, CL267, CL279and CL378) Cd tolerant (RRL ≥ 72%) and five (Donornumber CL30, CL62, CL235, CL257 and CL53) Cd-sensitive (RRL ≤ 52%) lines were grown for 10 daysin control solution and then transferred to the Cd(3.5 μM of CdCl2) solution. Fifty seedlings from eachline were grown and the experiments were repeated forthree times. After 48 h (continuous light conditions) ofincubation, the roots were thoroughly washed in distilledwater, and then the root tips (2 mm) were collectedusing a blade. After drying at 600◦C, samples weredigested in Fluon PFA vessels (Asahi Glass, Tokyo, Japan)containing high-purity HNO3 (Wako Pure Chemicals,Tokyo, Japan), using a microwave oven. The amount ofCd was quantified by an individually coupled plasma-mass spectrometer (ELAN 6000; Perkin Elmer Japan,Tokyo, Japan).

Results

QTL and epistatic interactions detected in Ler/Coland Col/Kas RILs under Cd stress

RRL of Ler/Col in Cd toxic solution showed a normaldistribution (Fig. 1A). It suggested that multiple geneticfactors regulating Cd tolerance were segregated amongthe RILs. According to the RRL, Col showed the tolerantand Ler showed the sensitive effect. We detected majorQTLs (α = 0.05) controlling RRL by the CIM method(Table 1, Fig. 2). A major QTL was found on the upperpart of chromosome 5 (Fig. 2), and it explained 26% ofthe phenotypic variation of the RRL in the Cd solution.The other two QTLs were detected on the middle partof chromosome 2 and the lower part of chromosome4, which can explain 12 and 10% of the phenotypicRRL variations of Cd tolerance, respectively. Positiveallelic effects were identified with the Col allele atthe QTLs on chromosomes 4 and 5, while that of theLer allele was found at the QTL on chromosome 2.The negative effect of the Col allele and the positiveeffect of Ler allele of the QTL on chromosome 4 wereconsistent with the transgressive segregation among theRI population. Using a complete pair-wise search, onesignificant (α = 0.0005) epistatic interacting loci pairwas also detected (Table 2), which can also explain thetransgressive segregations.

The comparison of genetic structure among differentRI populations is one approach to identify major locicontrolling target traits (e.g. Ikka et al. 2008). QTLanalysis for Cd tolerance was performed with Col/KasRILs using the same experimental protocols followed for


0

5

10

15

20

25

30

0

5

10

15

20

25

30

0

5

10

15

20

25

30

0

5

10

15

20

25

30

0

5

10

15

20

25

30

0 20 40 60 80 100

0 20 40 60 80 100 0 20 40 60 80 100

0 20 40 60 80 100 0 20 40 60 80 100

Num

ber

of R

I lin

es

Relative root length (%)

D E

CA BRRL (%)LerCol

CVHeritability

54.960.915.30.95

RRL (%)ColLer

CVHeritability

37.649.325.80.93

RRL (%)LerCol

CVHeritability

33.649.817.80.93

RRL (%)KasCol

CVHeritability

37.143.321.50.81

RRL (%)KasCol

CVHeritability

53.464.422.40.91

Fig. 1. Frequency distribution of the relative root length of the RILs. The panels A, B and C indicate the frequency distribution of the RRL of theLer/Col RILs in Cd, H2O2 and Cu stress, respectively, while the panels D and E indicate the frequency distribution of the Col/Kas RILs in Cd and H2O2

stress, respectively. Arrows in the panels indicate the position of parents among the distribution. Relative root length of the parent lines (%), Broadsense heritability (Heritability) and the CV are indicated in each panel.

Table 1. Closely linked markers and effects of single QTLs associated with Cd, H2O2 and Cu tolerance detected by the CIM method at an α = 0.05threshold. Chrom., Chromosome number; LOD, additive logarithm of odds value; R2, proportion of the variance explained by each QTL; Allele mean,Genotypic mean of the root length of RI lines at the QTL position; RRL, relative root length; Mb, Megabase.

Allele mean (RRL; %)Approximate position in Additive

RILs Treatment Chrom. Linked marker physical map (Mb) LOD effect R2 Col Ler

Ler/Col Cd 2 g6842 10.9 2.8 −6.4 0.12 54.5 60.64 g8300 15.3 2.2 2.8 0.10 60.9 55.65 CDPK9 7.9 5.6 4.7 0.26 62.8 56.2

H2O2 2 BIO2b 16.3 3.6 −6.5 0.21 41.8 46.44 nga8 5.4 3.6 −4.2 0.12 41.4 46.35 mi138 7.5 2.9 3.4 0.16 48.8 42.5

Cu 3 nga126 3.5 2.4 −6.0 0.09 49.8 45.43 ve021 15.9 4.1 5.0 0.16 42.9 53.25 mi138 7.6 3.9 −4.9 0.13 55.0 44.0

Col KasCol/Kas Cd 5 nga139 8.4 3.8 7.3 0.32 54.0 46.7

H2O2 1 SNP21606928 21.1 2.8 4.1 0.15 35.3 30.94 nga8 5.6 3.8 3.8 0.26 33.7 31.65 nga139 8.4 3.9 4.2 0.33 36.6 30.6

the Ler/Col RILs. According to the RRL, some Col alleleswere expected to show more tolerance than Kas alleles(Fig. 1D). By the CIM method, we identified such apositive allelic effect of Col as a unique QTL controlling

Cd tolerance on the upper part of chromosome 5(Table 1). The QTL explained about 32% of Cd tolerancevariation among RILs, and overlapped with that detectedon chromosome 5 of the Ler/Col RI population (see


LOD

sco

re

Chromosome 1

g6842

Chromosome 2

0

2

Chromosome 3

0

2

Chromosome 4

Position (cM)

CDPK9

0

2

4

6

Col +26%

g8300

0

2

Col +10%

Ler +12%

Chromosome 5

2

0

20015010050 250

Fig. 2. Logarithmic odd (LOD) score curves of CIM analysis for the Cdtolerance RRL of the Ler/Col RI population. Horizontal bars indicate thesignificance thresholds for LOD (α = 0.05) for each chromosome. Thepositive genotypes for the Cd tolerance RRL and their R2 value (%) areindicated at the major QTL peaks, which have greater LOD values thanthe thresholds.

Fig 6). One epistatic interaction linked to the geneticmaker pairs, ATCHIB2 and SNP44607413, was detectedby a complete pair-wise search, which can explain someof the transgressive segregations.

Reliability test of the Cd-tolerant QTLs of theLer/Col RILs via genotyping using subset RI lines

To test the reliability of the QTL for Cd tolerance,we determined linkage of Cd tolerance and geneticmarkers around the detected QTL, using a subset ofRILs (i.e. originating from the same cross by Listerand Dean, but which have not been genotyped). Byscreening for Cd tolerance with the same experimentaldesign, we selected 10 tolerant (RRL � 59 %) and 10sensitive (RRL � 37%) RILs. Genetic markers nga1139and nga139, co-locating near the QTL on chromosomes4 and 5, segregated between tolerant and sensitive RILs(Fig. 3). There were a few exceptions, but almost of allthe genotypes accounted for the allelic effect estimatedby QTL analysis. For example, 15 of 20 genotypes ofnga139 were consistent with expected genotypes (i.e.Col allele has positive effect on Cd tolerance for nga1139and nga139 markers. So tolerant plants had the Col alleleand sensitive plants had the Ler allele). In case of theGPA1 marker, 14 of 20 genotypes were consistent withthe expected genotypes (in this case, the Ler allele hasthe positive effect on Cd tolerance for the GPA1 marker).

Translocation ratio and histochemical stainingamong sensitive and tolerant Ler/Col RILs

To characterize the mechanisms of Cd tolerance of thedetected QTLs, we examined indexes that accountedfor previously identified mechanisms of Cd tolerance.Five tolerant and five sensitive RILs were selected, whichcarried the expected alleles at all QTL positions. First, weexamined Cd-translocation capacity to the shoots, whichis consistent with the function of Cd tolerance QTL ofArabidopsis hellali (Courbot et al. 2007). There was nosignificant difference in the Cd distribution capacity tothe shoot between tolerant and sensitive RILs (Fig. 4).

Table 2. Epistatic interactions between two loci for Cd, H2O2 and Cu tolerance in RILs by a complete pair-wise search at P < 0.005. Significantdifferences among four combinations of each epistatic interacting loci pair estimated by one-way ANOVA (P < 0.05) are represented by different letters(a–c). The asterisk (*) indicates linked epistatic position in Fig. 6. Chrom., chromosome numbers of epistatic loci; LLR, log-likelihood ratio; RRL, relativeroot length; L, Ler allele; C, Col allele; K, Kas allele.

Approximate RRL mean (%; 3.5 μM CdCl2/control)Epistatic physical

RILs Treatment marker interaction * Chrom. position (Mb) LLR L × L L × C C × L C × C

Ler/Col Cd mi72× mi32 A 1 × 4 14.9 × 10.0 7.8 54.3c 63.8a 59.1b 56.8bc

H2O2 mi62× CDPK9 B 1 × 5 9.5 × 7.9 10.1 41.9b 51.6a 44.0b 42.3b

g3843× mi194 C 4 × 5 1.2 × 18.5 8.9 45.4b 43.5bc 39.9c 51.5a

Cu G6842× mi74b D 2 × 3 10.8 × 1.0 8.4 50.7a 48.5a 35.2b 48.2a

mi142× ve021 E 3 × 3 8.0 × 15.9 15.6 53.4a 33.9c 52.9a 46.0b

K × K K × C C × K C × CCol/Kas Cd ATCHIB2× SNP44607413 F 3 × 5 4.1 × 25.1 7.1 44.8b 55.3a 54.1a 47.2b

H2O2 SNP21607030× LFY3 G 1 × 5 28.1 × 24.8 7.5 35.3a 30.3b 30.3b 36.7a


RRL (%)

QTL Marker

Positive effect allele C

olLe

r

CL1

0

CL7

8

CL9

1

CL9

2C

L60

CL6

4C

L42

CL2

2C

L100

CL3

8

CL2

36

CL2

44

CL3

96

CL8

0C

L21

CL9

9C

L159

CL1

16C

L148

CL3

07

65 6062 60 6061 616772 66 35 32 26353637 36 27 2633

Chr. 5 nga139 Col

Chr. 2 GPA1 Ler

Chr. 4 nga1139 Col

Tolerant Sensitive

5561

Fig. 3. Genotyping of Ler/Col RILs with contrasting Cd tolerances that were selected from a subset of RILs (See Materials and methods section).Ten tolerant and 10 sensitive RILs were selected from a subset of 65 Ler/Col RILs. CAPS (GPA1; locates 2.13 cM from the QTL position marker g6842in Lister and Dean RI map) and SSLP (nga1139; locates 2.19 cM from the QTL position marker g8300 and nga139; located 5.93 cM from the QTLmarker CDPK9 in Lister and Dean RI map) markers are closely linked to the position of the major Cd tolerance QTLs on chromosome 2, 4 and 5respectively. Gel images of parent accessions (Ler and Col) and RILs (tolerant and sensitive) and RRL (%) for Cd tolerance RRL are also shown.

Relative root length in Cd solution (%)

Tra

nslo

catio

n ra

tio o

f Cd

(sho

ot/to

tal)

R2 = 0.006

0

0.2

0.4

0.6

0.8

1.0

0 20 40 60 80 100

Fig. 4. Translocation ratio of Cd in the tolerant and the sensitiveLer/Col RILs. The roots of pre-grown seedlings of tolerant (RRL � 72%)and sensitive (RRL � 52%) were incubated in Cd (3.5 μM) solution for48 h and then Cd content in the roots and shoots were quantified.Mean ± SE of translocation ratio (Cd content in the shoot/total) fromthree biologically independent replications are shown (n = 3).

We then determined the difference of accumulation ofROS in cells of the roots between tolerant and sensitiveRILs. Brighter fluorescence was always determined in thesensitive RILs with H2DCFDA staining at an early stageof exposure (30 min and 3 h after the exposure). At theseperiods, the roots of sensitive RILs showed brighter greenfluorescence than that of the tolerant RILs (Fig. 5). After24 h exposure, fluorescence became weaker in the rootof both RILs, possibly because of decreasing viability orbecause of the higher concentration of Cd than in thegrowth experiment. On the other hand, there were nodifferences in fluorescence with DHE staining, whichvisualizes O−

2 accumulation, at all time periods. Takentogether, these observations suggested that Cd toleranceof RILs was not accounted for by the difference of Cd-translocation capacities but might explained by othermechanisms that resulted in accumulating less H2O2

than sensitive RILs.

QTL analyses with H2O2 resistance in Ler/Col andCol/Kas RILs

The QTL on chromosome 5 was shared with two RIpopulations, and shared the same chromosome regionwith the previously identified QTL for the tolerance toother rhizotoxicities (La tolerance in Ler/Col; Kobayashiet al., 2007b, Al tolerance in Col/Kas; Ikka et al. 2008).These results suggest that the QTL on chromosome 5is a common QTL controlling Cd resistance in differentRI populations, and might be related to the commontolerant mechanism to various rhizotoxic stressors, suchas resistance to ROS stress. To determine whether theQTL on chromosome 5 is linked to the tolerance toROS, we performed QTL analyses of H2O2 resistanceof the roots with two RI populations. Three major QTLson chromosome 2, 4 and 5 were detected from theLer/Col RILs, while those on chromosomes 1, 4 and5 were detected from Kas/Col RI population (Table 1,Fig. 6). Although the QTL on chromosome 2 had thebiggest R2 value among the major QTLs from Ler/Col,the QTL on the chromosome 5 was shared by bothmapping populations and overlapped those for the Cd-resistance QTL. The major QTL on chromosome 5 canexplain 16 and 33% of the phenotypic RRL variation ofH2O2 resistance in Ler/Col and Col/Kas RI populations,respectively. In both sets of RILs, Col showed the positiveallelic effect of the QTL on the upper part of chromosome5 for H2O2 resistance.

Copper tolerance QTL in Ler/Col RILs

Recently, we identified that a dysfunctional Cu-transporting ATPase (HMA5), involved in Cu-transloca-tion process, is a cause of a major Cu-tolerance QTLin Ler/Cvi population (Kobayashi et al. 2008). In thatstudy, we found that the Cu-translocation capacities ofCol and Ler were very similar, and had the same HMA5


ROS formation

Tolerant group Sensitive group

CL266 CL279 CL54 CL378 CL267 CL30 CL235 CL257 CL62 CL53

30 min

3 h

81 80 74 73 72 39 50 50 51 52RRL (%)

Superoxide anion accumulation

30 min

24 h

Fig. 5. Histochemical analyses of the root damage caused by Cd stress. The same five tolerant and five sensitive Ler/Col RILs used in Fig. 4 weregrown for 3–4 days in the control growth medium and then transferred to Cd toxic solution (3.5 μM of CdCl2, pH 5.0). The roots were stainedby H2DCFDA or DHE to visualize the accumulation of H2O2 (greenish color) and O−

2 accumulation (reddish color), respectively. White bar indicates50 μm.

protein. In addition, it suggests that Cu tolerance of theLer/Col population is regulated by other Cu-tolerancemechanism(s), including the ROS-scavenging capacity(Babu et al. 2003). Therefore, we performed QTLanalysis of Cu tolerance in the Ler/Col RI population.Using the CIM method, we identified three QTLs onchromosome 3 and 5, which could explain 38% of Cu-tolerance variation (Table 1). Among these QTLs, theQTL on chromosome 5 overlapped the QTL for Cdtolerance (Fig. 6).

Multitrait QTL analysis on the upper part ofchromosome 5

We detected major QTLs on the upper part ofchromosome 5 for both Cd and H2O2 stress in bothsets of RILs (Table 1, Figs 2 and 6). To evaluate anypleiotropic effect of the QTL, we performed a multitraitsQTL analysis for the QTL on chromosome 5 with multipleQTL datasets. In both sets of RILs, the joint likelihoodratio score curve showed a peak higher than the peaksfor the individual stress suggesting that the locus couldbe responsible for a common pleiotropic effect for allstressors (Fig. 7A, B).

Discussion

QTL analysis is a useful approach for understandingcomplex biological systems. For example, one of themajor mechanisms of Cd tolerance was identifiedby a QTL analysis, followed by molecular biological

characterization. These analyses revealed that the Cd-translocation capacity, regulated by Cd-transportingATPase (HMA4), is the major cause of Cd tolerancein Arabidopsis halleri (Cd tolerant and accumulator)(Courbot et al. 2007). On the other hand, an Arabidopsismutant study showed that accumulation of Cd-bindingmolecules regulated by phytochelatin (PC) synthase(designated as CAD1) is critical for Cd tolerance(Howden et al. 1995). In the present study, we detecteda QTL on chromosome 5 as the major genetic factorregulating Cd tolerance in the Ler/Col population(Table 1, Figs 2, 3 and 6), but it was not co-located withpreviously identified Cd-tolerant genes (i.e. HMA4 andCAD1). This is consistent with complex mechanismsof Cd tolerance, which were suspected by previousphysiological studies. In fact, the Cd-translocationcapacities of the tolerant RILs were similar to that ofthe sensitive RILs(Fig. 4). The QTL on chromosome5 was identified in a second RI population, namelyCol/Kas, suggesting that the locus contains importantgene(s) regulating Cd tolerance in Arabidopsis, but is notdependent on Cd-translocation capacity.

Several groups reported that ROS damage is atypical symptom of Cd cytotoxicity (e.g. Hsu and Kao2004, Metwally et al. 2005). Histochemical analysisin the Ler/Col RILs using fluorescent probes showedgreater ROS accumulation in sensitive RILs than intolerant RILs (Fig. 5). In this analysis, we used a setof fluorescent probes with different specificity to ROS.Weak fluorescence with DHE (relatively specific for O−

2 )indicated little accumulation of O−

2 by Cd treatment.


15%

30%

Ler/Col RILs

Col/Kas RILs

CdH2O2

Cu

H2O2

Cd

0 5 10 15 20 30.4 Mb25Chromosome 1

A Gmi72

B

D

0 5 10 15 19.7 Mb

HMA4

Chromosome 2

GPA1

0 5 10 15 20 23.4 Mb

Chromosome 3

DE

EFATCHIB2

Chromosome 40 5 10 15 18.5 Mb

A nga1139mi32C

0 5 10 15 20 25 27.1 MbChromosome 5

B

CCDPK9

G

F

nga139CAD1

SNP44607413

Fig. 6. Physical positions of major QTLs and epistatic interacting loci pairs associated with relative root length of Ler/Col and Col/Kas RI populationswith Cd and other rhizotoxic stressors. The positions and effects of major QTLs detected by CIM method are indicated by arrows. Upward arrowsindicate positive effects of Col allele in both RI populations. Sizes of the arrows represent the percentage of phenotypic variation explained by eachQTL. Epistatic interacting regions are indicated by rectangles.

On the other hand, bright fluorescence with H2DCFDA(relatively specific for H2O2) was induced early inCd treatment (30 min), suggesting that H2O2 is thedominant ROS accumulated in the sensitive RILs. This isconsistent with previously reported ROS damage patternin Arabidopsis by Cho and Seo (2005). In addition,this possibility is further supported by comparison ofthe genetic architecture of H2O2 resistance to Cdtolerance. The QTL on chromosome 5 was commonto Cd resistance and H2O2 resistance in both Ler/Coland Col/Kas RI populations. There was no difference ofCd-translocation capacity between Col and Kas RILs(Fig. 4). In addition, the same chromosomal regionof the upper part of chromosome 5 was detectedas major QTL for both Cd and H2O2 tolerances ofboth RI populations (Fig. 6). These results suggest thatsame mechanism (i.e. reduction of damage by H2O2

accumulation) regulates Cd tolerance in the Ler/Col andthe Col/Kas RI populations.

The QTL on chromosome 5 (QTL5), which was sharedby both RI populations with Cd and H2O2 resistance,might have an important role in reducing ROS-induceddamage. If this model is correct, the chromosomal regionmight be detected as a QTL regulating tolerance to otherrhizotoxins producing ROS. In fact, in our previousstudy, QTL5 was detected as the major QTL for Altolerance with the same allelic effect of Col (Ikka et al.2008). In this case, Al tolerance is not linked to Al-activated malate release, while it can be explainedby ROS-scavenging capacity. On the other hand, theQTL on chromosome 5 was not detected with theLer/Col and Ler/Cvi RI populations, which show Alexclusion (i.e. Al-activated malate release via AtALMT1;see Hoekenga et al. 2003) as the major Al-tolerancemechanism, and the major QTL is linked to chromosome1 (Kobayashi and Koyama 2002, Kobayashi et al. 2005).This would account for the current understandingof mechanisms of Al tolerance in Arabidopsis. In


Testing position (cM)

A

B

LR te

st s

tatis

tic

H2O2

Cd

CuJoint

CdH2O2

Joint

50 100 150 200 250 3000

50 100 150 200 250 30000

5

10

20

25

15

30

0

5

10

20

25

15

30

Fig. 7. Joint QTL map on chromosome 5 of Ler/Col and Col/Kas RIpopulations by a multitraits CIM analyses. The simple CIM methodwas used for the calculation of likelihood ratios for each trait (brokenline). The black solid line indicates the joint LOD curve of the multipletraits. Horizontal boxes on the x-axis are the overlapping QTL positionsfor stress factors detected by the CIM method. Panel A, joint map ofchromosome 5 using Cd, H2O2 and Cu stress on Ler/Col RILs. Panel B,joint map of chromosome 5 using Cd and H2O2 stress on Col/Kas RILs.

this plant species, both Al exclusion (regulated byAtALMT1, Kobayashi et al. 2007a) and Al translocation(regulated by ALS3, a putative Al translocator, Larsenet al. 2005) are the major mechanisms of Al tolerance.However, the ROS-scavenging capacity might alsobe important for Al tolerance (see Yamamoto et al.2002). They suggested that the QTL5 has a greatercontribution to Al tolerance when Al exclusion andAl translocation do not account for the mechanismof Al tolerance among mapping populations. On theother hand, a similar pattern of greater contributionof QTL5 has been identified by Cu-tolerance QTLstudies. In this study, the QTL on chromosome 5was detected as a Cu-tolerance QTL in Ler/Col RIpopulation (Table 1). In the case of Ler/Col, the Cu-translocation capacity could not explain Cu toleranceamong parental accessions (Kobayashi et al. 2008).Conversely, the QTL on chromosome 5 is not detected asa major QTL for Cu tolerance in the Ler/Cvi population,whose Cu tolerance was regulated by a major QTLcontrolling Cu-translocation capacity (an amino acid

polymorphism in the Cu-transporting ATPase, AtHMA5;(Kobayashi et al. 2008). Taken together, our analysessuggest that the QTL on chromosome 5 has pleiotropiceffects on tolerance to multiple rhizotoxic ions, possiblyvia ROS-related mechanism(s). This factor would beimportant if tolerance could not be explained byaccumulation of particular ions in the roots. This isconsistent with the phenotypes of transgenic plantsoverexpressing ROS-scavenging enzymes. For example,the overexpression of glutathione S-transferase (GST)and peroxidase in Arabidopsis slightly, but significantly,conferred tolerance to multiple rhizotoxic ions (Roxaset al. 1997).

Pleiotropic loci have been identified by QTL-pyramiding approaches. For example, a QTL onchromosome 1 was identified as a shared QTLcontrolling multiple traits relating to the life historyof the Ler/Cvi population (Alonso-Blanco et al. 1999).On the other hand, a QTL for various Ca-channelantagonists (e.g. La, Gd and verapamil) was pyramidedon chromosome 2 of Ler/Col (Kobayashi et al. 2007b).These studies showed that some common mechanismsshared with the multiple traits are regulated by asingle locus. In the present study, we applied thesame concept to study the mechanism of Cd tolerance.Our results might provide genetic evidence of theinteraction of tolerance to multiple rhizotoxic stressorswith the ROS-scavenging capacity. Further efforts, suchas development and characterization of a near isogenicline for this locus or fine mapping approaches, mightidentify the gene(s) accounting for the QTL. Anotherremaining question for the future research would beidentification of mechanisms of epistatic interactionsdetected by various traits (Table 1, Fig 6). For example,some epistasis was shared with various rhizotoxictreatments. Further research for the epistatic interactingloci pairs would be important to understand complextolerance mechanisms to rhizotoxic treatments, whichare regulated by the interaction of multiple geneproducts.

Acknowledgements – This work was supported by theJapan Society for the Promotion of Science and the Ministryof Economy, Trade and Industry, Japan for HK. We thankthe Nottingham Arabidopsis Stock Center and the RIKENBioResource Center for providing the Arabidopsis seeds.

References

Alonso-Blanco C, Blankestijn-de Vries H, Hanhart CJ,Koornneef M (1999) Natural allelic variation at seed sizeloci in relation to other life history traits of Arabidopsisthaliana. Proc Natl Acad Sci USA 96: 4710–4717


Arduini I, Godbold DL, Onnis A (1994) Cadmium andcopper change root growth and morphology of Pinuspinea and Pinus pinaster seedlings. Physiol Plant 92:675–680

Babu TS, Akhtar TA, Lampi MA, Tripuranthakam S,Dixon DG, Greenberg BM (2003) Similar stressresponses are elicited by copper and ultravioletradiation in the aquatic plant Lemna gibba: implicationof reactive oxygen species as common signals. PlantCell Physiol 44: 1320–1329

Basten C, Weir B, Zeng Z (1994) Zmap-a QTLcartographer In: Smith C, Gavora JS, Benkel B,Chesnais J, Fairfull W, Gibson JP, Kennedy BW,Burnside EB (eds) Proceedings of the 5th WorldCongress on Genetics Applied to Livestock Production:Computing Strategies and Software Vol. 22. Universityof Guelph Press, Guelph, ON, Canada, pp 65–66

Behl C, Davis JB, Lesley R, Schubert D (1994) Hydrogenperoxide mediates amyloid beta protein toxicity. Cell77: 817–827

Bell CJ, Ecker JR (1994) Assignment of 30 microsatelliteloci to the linkage map of Arabidopsis. Genomics 19:137–144

Benavides MP, Gallego SM, Tomaro ML (2005) Cadmiumtoxicity in plants. Braz J Plant Physiol 17: 21–34

Chase K, Adler F, Lark K (1997) Epistat: a computerprogram for identifying and testing interactions betweenpairs of quantitative trait loci. Theor Appl Genet 94:724–730

Cho UH, Seo NH (2005) Oxidative stress in Arabidopsisthaliana exposed to cadmium is due to hydrogenperoxide accumulation. Plant Sci 168: 113–120

Churchill GA, Doerge RW (1994) Empirical thresholdvalues for quantitative trait mapping. Genetics 138:963–971

Courbot M, Willems G, Motte P, Arvidsson S, Roosens N,Saumitou-Laprade P, Verbruggen N (2007) A majorquantitative trait locus for cadmium tolerance inArabidopsis halleri colocalizes with HMA4, a geneencoding a heavy metal ATPase. Plant Physiol 144:1052–1065

Fujiwara T, Hirai MY, Chino M, Komeda Y, Naito S (1992)Effects of sulfur nutrition on expression of the soybeanseed storage protein genes in transgenic petunia. PlantPhysiol 99: 263–268

Hoekenga OA, Vision TJ, Shaff JE, Monforte AJ, Lee GP,Howell SH, Kochian LV (2003) Identification andcharacterization of aluminum tolerance loci inArabidopsis (Landsberg erecta x Columbia) byquantitative trait locus mapping: a physiologicallysimple but genetically complex trait. Plant Physiol 132:936–948

Howden R, Goldsbrough P, Andersen C, Cobbett C (1995)Cadmium-sensitive, cad1 mutants of Arabidopsis

thaliana are phytochelatin deficient. Plant Physiol 107:1059–1066

Hsu YT, Kao CH (2004) Cadmium toxicity is reduced bynitric oxide in rice leaves. Plant Growth Regul 42:227–238

Ikka T, Kobayashi Y, Iuchi S, Sakurai N, Shibata D,Kobayashi M, Koyama H (2007) Natural variation ofArabidopsis thaliana reveals that aluminum resistanceand proton resistance are controlled by different geneticfactors. Theor Appl Genet 115: 709–719

Ikka T, Kobayashi Y, Tazib T, Koyama H (2008)Aluminum-tolerance QTL in Columbia/Kashmir inbredpopulation of Arabidopsis thaliana is not associatedwith aluminum-responsive malate excretion. Plant Sci175: 533–538

Jiang C, Zeng ZB (1995) Multiple trait analysis of geneticmapping for quantitative trait loci. Genetics 140:1111–1127

Kobayashi Y, Koyama H (2002) QTL analysis of Altolerance in recombinant inbred lines of Arabidopsisthaliana. Plant Cell Physiol 43: 1526–1533

Kobayashi Y, Furuta Y, Ohno T, Hara T, Koyama H (2005)Quantitative trait loci controlling aluminium tolerancein two accessions of Arabidopsis thaliana (Landsbergerecta and Cape Verde Islands). Plant Cell Environ 28:1516–1524

Kobayashi Y, Hoekenga OA, Itoh H, Nakashima M,Saito S, Shaff JE, Maron LG, Pineros MA, Kochian LV,Koyama H (2007a) Characterization of AtALMT1expression in aluminum-inducible malate release and itsrole for rhizotoxic stress tolerance in Arabidopsis. PlantPhysiol 145: 843–852

Kobayashi Y, Ikka T, Kimura K, Yasuda O, Koyama H(2007b) Characterisation of lanthanum toxicity for rootgrowth of Arabidopsis thaliana from the aspect ofnatural genetic variation. Funct Plant Biol 34: 984–994

Kobayashi Y, Kuroda K, Kimura K, Southron-Francis JL,Furuzawa A, Kimura K, Iuchi S, Kobayashi M, Taylor GJ,Koyama H (2008) Amino acid polymorphisms in strictlyconserved domains of a P-type ATPase HMA5 areinvolved in the mechanism of copper tolerancevariation in Arabidopsis. Plant Physiol 148: 969–980

Konieczny A, Ausubel FM (1993) A procedure for mappingArabidopsis mutations using co-dominantecotype-specific PCR-based markers. Plant J 4:403–410

Koyama H, Toda T, Hara T (2001) Brief exposure tolow-pH stress causes irreversible damage to the growingroot in Arabidopsis thaliana: pectin-Ca interaction mayplay an important role in proton rhizotoxicity. J Exp Bot52: 361–368

Lander ES, Green P, Abrahamson J, Barlow A, Daly MJ,Lincoln SE, Newburg L (1987) MAPMAKER: aninteractive computer package for constructing primary


genetic linkage maps of experimental and naturalpopulations. Genomics 1: 174–181

Larsen PB, Geisler MJ, Jones CA, Williams KM, Cancel JD(2005) ALS3 encodes a phloem-localized ABCtransporter-like protein that is required for aluminumtolerance in Arabidopsis. Plant J 41: 353–363

Li Y, Roycewicz P, Smith E, Borevitz JO (2006) Genetics oflocal adaptation in the laboratory: flowering timequantitative trait loci under geographic and seasonalconditions in Arabidopsis. PLoS One 1: e105

Lister C, Dean C (1993) Recombinant inbred lines formapping RFLP and phenotypic markers in Arabidopsisthaliana. Plant J 4: 745–750

Loudet O, Gaudon V, Trubuil A, Daniel-Vedele F (2005)Quantitative trait loci controlling root growth andarchitecture in Arabidopsis thaliana confirmed byheterogeneous inbred family. Theor Appl Genet 110:742–753

Metwally A, Safronova VI, Belimov AA, Dietz KJ (2005)Genotypic variation of the response to cadmium toxicityin Pisum sativum L. J Exp Bot 56: 167–178

Pinto A, Mota A, de Varennes A, Pinto F (2004) Influenceof organic matter on the uptake of cadmium, zinc,copper and iron by sorghum plants. Sci Total Environ326: 239–247

Pomponi M, Censi V, Di Girolamo V, De Paolis A, diToppi LS, Aromolo R, Costantino P, Cardarelli M (2006)Overexpression of Arabidopsis phytochelatin synthasein tobacco plants enhances Cd2+ tolerance andaccumulation but not translocation to the shoot. Planta223: 180–190

Roosens N, Verbruggen N, Meerts P, Ximenez-Embun P,Smith J (2003) Natural variation in cadmium toleranceand its relationship to metal hyperaccumulation for

seven populations of Thlaspi caerulescens from westernEurope. Plant Cell Environ 26: 1657–1672

Ross-Ibarra J, Morrell PL, Gaut BS (2007) ColloquiumPapers: Plant domestication, a unique opportunity toidentify the genetic basis of adaptation. Proc Natl AcadSci USA 104: 8641–8648

Roxas V, Smith R Jr, Allen E, Allen R (1997)Overexpression of glutathione S-transferase/glutathioneperoxidase enhances the growth of transgenic tobaccoseedlings during stress. Nat Biotechnol 15: 988–991

Sanita di Toppi L, Gabbrielli R (1999) Response tocadmium in higher plants. Environ Exp Bot 41:105–130

Schutzendubel A, Polle A (2002) Plant responses to abioticstresses: heavy metal-induced oxidative stress andprotection by mycorrhization. J Exp Bot 53: 1351–1365

Sergeeva LI, Keurentjes JJB, Bentsink L, Vonk J, van derPlas LHW, Koornneef M, Vreugdenhil D (2006)Vacuolar invertase regulates elongation of Arabidopsisthaliana roots as revealed by QTL and mutant analysis.Proc Natl Acad Sci USA 103: 2994–2999

Yamamoto Y, Kobayashi Y, Devi SR, Rikiishi S,Matsumoto H (2002) Aluminum toxicity is associatedwith mitochondrial dysfunction and the production ofreactive oxygen species in plant cells. Plant Physiol 128:63–72

Zeng ZB (1993) Theoretical basis for separation of multiplelinked gene effects in mapping quantitative trait loci.Proc Natl Acad Sci USA 90: 10972–10976

Zeng ZB (1994) Precision mapping of quantitative traitloci. Genetics 136: 1457–1468

Edited by K. -J. Dietz


Quantitative Trait Loci Controlling Resistance to Gray Leaf Spot in Maize

Documents

Transcript of Quantitative Trait Loci Controlling Resistance to Gray Leaf Spot in Maize