Maintaining elevated Fe2+ concentration in solution culture for the development of a rapid and repeatable screening technique for iron toxicity tolerance in rice (Oryza sativa L.)

1 23

Plant and SoilAn International Journal on Plant-SoilRelationships ISSN 0032-079XVolume 372Combined 1-2 Plant Soil (2013) 372:253-264DOI 10.1007/s11104-013-1739-4

Maintaining elevated Fe2+ concentrationin solution culture for the development ofa rapid and repeatable screening techniquefor iron toxicity tolerance in rice (Oryzasativa L.)Venus Elec, Celsa A. Quimio, RhulyxMendoza, Andres Godwin C. Sajise,Sarah E. J. Beebout, Glenn B. Gregorio &Rakesh Kumar Singh

1 23

Your article is protected by copyright and all

rights are held exclusively by Springer Science

+Business Media Dordrecht. This e-offprint

is for personal use only and shall not be self-

archived in electronic repositories. If you wish

to self-archive your article, please use the

accepted manuscript version for posting on

your own website. You may further deposit

the accepted manuscript version in any

repository, provided it is only made publicly

available 12 months after official publication

or later and provided acknowledgement is

given to the original source of publication

and a link is inserted to the published article

on Springer's website. The link must be

accompanied by the following text: "The final

publication is available at link.springer.com”.

REGULAR ARTICLE

Maintaining elevated Fe2+ concentration in solutionculture for the development of a rapid and repeatablescreening technique for iron toxicity tolerancein rice (Oryza sativa L.)

Venus Elec & Celsa A. Quimio & Rhulyx Mendoza &

Andres Godwin C. Sajise & Sarah E. J. Beebout &Glenn B. Gregorio & Rakesh Kumar Singh

Received: 24 December 2012 /Accepted: 17 April 2013 /Published online: 5 May 2013# Springer Science+Business Media Dordrecht 2013

AbstractBackground and aims Iron toxicity decreases rice(Oryza sativa) grain yield especially in acid soils afterflooding. Our aim was to establish a high-throughputscreening technique using nutrient solution culture foridentifying Fe-toxicity-tolerant genotypes.Methods Varying levels of Fe, pH, and chelators inYoshida nutrient solution culture were tested to maintainsufficient Fe2+ concentration over time to optimize theseverity of Fe toxicity stress for distinguishing betweena tolerant (Azucena) and sensitive (IR64) genotype. Theoptimized solution was tested on 20 diverse genotypesin the greenhouse, with measurement of leaf bronzingscores and plant growth characteristics at the seedlingstage. The same 20 genotypes were grown tomaturity in

a field with natural Fe toxicity stress, with measurementof seedling-stage leaf bronzing scores and grain yield todetermine their inter-relationship.Results Optimized nutrient solution conditions were300 mg L−1 Fe supplied as Fe2+ at pH 4.0 with a1:2 molar ratio of Fe:EDTA, which maintained suffi-cient Fe2+ stress over 5 days. The highest correlationof nutrient solution phenotypic data with field grainyield was found with leaf bronzing scores at 4 weeks,with a Pearson r of 0.628 for simple association and aSpearman corrected r of 0.610 for rank association (P<0.01) using 20 diverse rice genotypes with proven Fetoxicity tolerance reaction. The Leaf bronzing scoresat 4 weeks in nutrient culture solution were also foundhighly correlated with LBS under natural field stressafter 8 weeks that had highest correlation with grainyield under stress.Conclusion This culture solution-based standardizedscreening technique can be used in plant breedingprograms as a high-throughput technique to identifygenotypes tolerant to Fe toxicity.

Keywords Rice . Fe2+ concentration . pH . EDTA .

Iron toxicity . Screening technique

AbbreviationsEDDHA Ethylenediamine

(2-hydroxyphenylacetic) acidEDTA Ethylenediamine tetra-acetic acid

Plant Soil (2013) 372:253–264DOI 10.1007/s11104-013-1739-4

Responsible Editor: Hans Lambers.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s11104-013-1739-4) containssupplementary material, which is available to authorized users.

V. Elec : R. Mendoza :A. G. C. Sajise : S. E. J. Beebout :G. B. Gregorio : R. K. Singh (*)International Rice Research Institute,DAPO Box 7777, Metro Manila, Philippinese-mail: [email protected]

C. A. QuimioCrop Science Cluster, College of Agriculture,University of the Philippines at Los Baños,Los Baños, Philippines

Author's personal copy

HEDTA (2-hydroxyethyl) ethylenediamine triaceticacid

IRRI International Rice Research InstituteLBS Leaf bronzing scoreWAT Weeks after transplanting

Introduction

Iron toxicity is one of the most important abiotic stressesin wetland conditions in different parts of Asia, SouthAmerica, and Africa (Audebert and Sahrawat 2000;Shimizu et al. 2005). It is a complex nutrient disorderbrought about by excessive Fe uptake from flooded soilsunder reducing conditions. The reactivity of this transi-tion metal causes problems at the cellular level, becausemany intracellular reactions use molecular oxygen as anelectron receptor producing superoxide or H2O2, whichcontributes to the generation of an extremely reactivehydroxyl radical using iron as a catalytic agent (Briat2002). Plants with high iron are characterized by stuntedgrowth, rusty leaf spots, stained leaf edges, and a poorlydeveloped root system. In severe cases, this can causeplant death and could contribute to a 12–100 % yieldreduction depending on the intensity of the toxicity andthe tolerance of the rice cultivar (Sahrawat 2004). Arecent yield gap study fromWest Africa clearly showedthat iron toxicity stress reduces rice yields by 16–78 %;however, this large range is due to variations in thetolerance of cultivars, severity of the stress, and man-agement options adopted (Audebert and Fofana 2009).

Iron toxicity tends to occur in soil that has beenflooded for an extended period of time, because thedecreasing redox potential causes Fe3+ in soil mineralsto be converted to Fe2+, and the latter is much moresoluble in water, leading to an excess of Fe in the soilsolution, sometimes up to 1,000 mg L−1 (Kirk 2004).Drainage of these soils temporarily relieves the Fe tox-icity stress, but is often not practical because of incom-plete water control. Therefore, it is important to have aplant breeding program focused on developing varietiesthat are tolerant of Fe toxicity.

An easy, quick, and repeatable screening technique isthe prerequisite for any successful breeding program.Several screening methods have been employed for thedevelopment of rice cultivars tolerant of iron toxicity,such as solution culture (Shimizu et al. 2005; Wan et al.2003), pot screening (Olaleye 1998), lysimeter screen-ing, and field screening (Camara 2006; Gridley et al.

2006). Field screening is preferred because it realistical-ly represents the stresses from the complex chemicalenvironment of flooded soil, but its drawbacks includethe difficulty of isolating the effect of Fe toxicity fromother complexities and the expense of finding sufficientexperimental iron-toxic field area to test many germ-plasm accessions. Screening in potted soil requires alarge number of pots and area for experiments, makingthe setup expensive, and also includes the challenge ofmaintaining uniform stress levels across pots. It is there-fore important to develop a simple screening techniquethat could accommodate a large number of germplasmaccessions for reproducible results like those of a fieldscreening.

Solution culture is attractive for abiotic stress toler-ance screening because it is typically convenient formaintaining uniform stress levels for a large number ofplants. However, for iron toxicity screening, maintaininguniform stress levels is difficult due to fluctuations inredox potential that are more problematic in solutionculture than in the field, causing a decrease in Fe2+

concentration due to oxidation to Fe3+. Iron-toxicityscreening methods for rice generally use an Fe2+ sourcerather than Fe3+, which is appropriate for rice since Fetoxicity in soil is caused by excess Fe2+,however, onlyone of the previous methods has verified that the Feremains in the reduced form over time (3 days) (DeDorlodot et al. 2005). The critical limit for the appear-ance of Fe-toxicity symptoms in rice varies with the pHof the soil solution, from about 100 mg L−1 at pH 3.7 to300 mg L−1 or higher at pH 5.0 (Tadano and Yoshida1978). Soil or solution pH greatly influences the Feavailability for plants, because Fe2+ concentration inreduced soil or solution increases sharply with the de-crease in pH, due to a decrease in the rate of Fe2+

oxidation (Stumm and Lee 1961). Therefore, variablesin solution-culture methods include nutrient solution pH,Fe concentration, the use of chelators to maintain Fesolubility, the addition of agar to increase viscosity andslow oxygen diffusion into the solution, and control ofsolution temperature to slow radial oxygen loss fromroots (Li et al. 2001; Kpongkor 2003; Shimizu et al.2005; Ishimaru et al. 2006). Temperature control is ex-pensive on a large scale, and also potentially confoundsthe identification of low-temperature and Fe toxicitytolerance.

Our objective was to improve upon the previousnutrient solution screening techniques in order to designa reliable and inexpensive method for high-throughput

254 Plant Soil (2013) 372:253–264

Author's personal copy

screening that differentiates rice genotypes for iron-toxicity tolerance in a way that is well-correlated withfield screening techniques. Our hypothesis was that pH,iron concentration, and chelator use could be optimizedto maintain a sufficiently high concentration of Fe2+ insolution over time without controlling solutiontemperature.

Materials and methods

Experiment 1. Optimizing nutrient solution chemistry

Effect of pH and presence of chelator on Fe solubilityand oxidation without plants To determine the effect ofpH on the spontaneous oxidation of Fe2+ to Fe3+ and itssubsequent precipitation out of solution as iron oxide inthe absence of plants, 200 mg L−1 Fe2+ (0.544 g ofreagent grade anhydrous FeSO4 in 1 L distilled water)solutions were prepared as follows: (a) Yoshida solutionwith pH 3.0 and (b) Yoshida solution with pH 4.5.Yoshida solution (minus FeCl3) was prepared followingthe procedures described in Yoshida et al. (1976). Thesolutions were transferred to a small tray similar to theones used in plant screening experiments and Fe2+ andtotal Fe content were measured immediately and onceper day for 5 days, following the method of Loeppertand Inskeep (1996). In this method, Fe2+ is determinedby UV/VIS spectrophotometry at 510 nm after colordevelopment with the indicator 1,10-phenanthroline,which is done in the dark to prevent photoreduction ofFe3+ to Fe2+. Total Fe is determined in a parallel sampleby adding hydroxylamine hydrochloride to reduce allFe3+ to Fe2+ prior to color formation, with Fe3+ calcu-lated from the difference between Fe2+ and total Fe. Fivedays was the typical amount of time before replacementof nutrient solution culture with fresh Yoshidasolution to ensure an adequate supply of macronutrientsfor plant growth.

Response of rice plants to different Fe concentration,pH, and chelator in culture solution To optimize nu-trient solution conditions for the purpose of differentiat-ing between tolerant and susceptible genotypes, two rice(Oryza sativa) varieties of known iron toxicity responsewere used in a series of small experiments: Azucena(tolerant) and IR64 (sensitive) (Gregorio et al. 2002;Sahrawat and Singh 1998; Wu et al. 1998). First, thetwo genotypes were grown for 4 weeks in Yoshida

nutrient solution with a factorial combination of threedifferent concentrations of Fe2+ using FeSO4 (200, 300,and 400mg L−1 Fe2+), two pH levels (3 and 4.5), and twoEDTA treatments (with and without EDTA in 1:1 molarratio with Fe2+). Seeds were soaked in water overnightand incubated in petri-plates for 48 h in a growth cham-ber. The germinated seeds were placed on a nylon netattached to the bottom of a polystyrene sheet with 100holes (one seed/hole). The polystyrene sheets were thenfloated in a plastic tray filled with deionized water for thefirst 3 days, and then were exposed to the treatmentsolutions starting on the fourth day. Leaf bronzing score(LBS) was observed once per week, assessed accordingto IRRI’s Standard Evaluation System (SES) for rice, inwhich a score of “0” indicates normal growth and tiller-ing, while the highest score of “9” indicates that almostall leaves are covered with reddish brown spots andmostplants are dead or dying (IRRI 1996). Nutrient solutiontotal Fe and Fe2+ measurements were done as describedabove, in solutions in control trays without plants, toavoid the confounding influence of plant uptake onsolution Fe concentration over time.

In a second factorial experiment, the same genotypeswere grown in Yoshida nutrient solution with 300 mgL−1 Fe in all treatments, testing a narrower range of pHvalues (3.5, 4.0, 4.2, 4.5) and additional Fe:EDTAmolarratios (1:1, 1:2, 1:4, 1:6, 1:8, 1:10). Plants were germi-nated, grown, and scored as described above.

Experiment 2. Greenhouse validation of the optimizednutrient solution conditions

The efficiency and repeatability of the method werevalidated using 20 genotypes (Table 1) of known toler-ance of iron toxicity based on yield performance in fieldstudies and reconfirmed results from pot and culturesolution studies (Gregorio et al. 2002; Mendoza et al.2003; Sahrawat and Singh 1998; Wang and Peverly1998). Plants were grown in Yoshida nutrient solutionwith 300 mg L−1 Fe supplied as Fe2SO4 and a 1:2 molarratio of Fe:EDTA, with pH maintained at 4.0 (±0.2) bydaily addition of NaOH or HCl, in 4-L trays in a green-house at IRRI, with 100 plants per tray using fivegerminated seeds of each genotype, one in each hole.Nutrient solution was replaced with a fresh batch at leastonce every 5 days. The 20 genotypes were randomizedwithin each tray, with three replicate trays. The plantswere non-destructively evaluated for degree of leafbronzing after 2 and 4 weeks, using IRRI’s SES as

Plant Soil (2013) 372:253–264 255

Author's personal copy

described above. After 4 weeks, the plants wereharvested for measurement of plant height, root length,and biomass. Total Fe content was analyzed by IRRI’sAnalytical Services Laboratory by inductively-coupledplasma atomic emission spectroscopy (ICP-AES) fol-lowing acid digestion. Only 15 genotypes were ana-lyzed because there was insufficient biomass foranalysis of the other 5 genotypes. For each of the threereplicate trays, the SES and plant growth characteristicsrepresented the average from five plants of one geno-type in one tray (Table 1).

Experiment 3. Field validation of the solution culturescreening procedure

A field trial was conducted in a farmer’s field, using thesame 20 genotypes listed in Table 1, in a randomized

complete block design (RCBD) with three replications.This fieldwas located in San Dionisio, Iloilo, Philippines(11º16′12″ N, 123º5′37″ E), and it had been identified asa field with natural Fe toxicity when flooded, accordingto observed plant growth symptoms. Six soil samplesfrom across the field (each a composite of at least threesub-samples) showed pH 4.2 (1:1 soil:H2O), dithionite-extractable Fe-oxide content (Asami and Kumada 1959)ranging from 0.7 % to 2.6 % (average 1.6 %), anddithionite-extractable Mn-oxide content at or below thedetection limit of 0.002 %. One genotype, IR61923-3B-9-1, was damaged in the nursery; hence, only 19 geno-types were included in the final analysis. Normal agro-nomic practices were followed, including puddling andflooding the soil throughout the season. Low fertilizerapplication at the rate of 40 kgN and 40 kg P/ha (throughdiammonium phosphate and urea) was followed as per

Table 1 List of genotypes with previously known range of sen-sitivity to Fe toxicity and their plant growth characteristics used inthe validation of the nutrient solution culture technique, showingthe comparison between each genotype grown in Fe-toxicity

stressed vs. non-stressed control conditions for 4 weeks as apercent decrease in the measured value. The genotype effect wassignificant for all three parameters (P<0.05)

% Reduction

Variety Origin Tolerance to Iron toxicity Plant height Root length Biomass

Karuna India Sensitive 50 75 45

CN 499-160-3 India Sensitive 51 68 65

Kinandang Patong Philippines Sensitive 49 78 76

IR29 IRRI Sensitive 57 74 50

IR61923-3B-9-1 IRRI Sensitive 54 76 56

IR63262-AC 201 IRRI Sensitive 57 80 51

IR65839-3B-1-2-3 IRRI Sensitive 68 90 91

Bao Thai Vietnam Sensitive 49 74 61

Phalguna India Moderately tolerant 53 78 51

PSBRc 18 IRRI Moderately tolerant 52 83 59

Mahsuri Malaysia Moderately tolerant 48 73 44

WITA 1 AfricaRicea Moderately tolerant 54 84 86

WITA 2 AfricaRice Moderately tolerant 50 79 67

IR61246-3B-15-2-2-3 IRRI Tolerant 62 80 65

IR61612-3B-16-2-2-1 IRRI Tolerant 50 85 45

IR61640-3B-14-3-3-2 IRRI Tolerant 51 81 16

TCA 4 India Tolerant 38 79 -5.7

Suakoko 8 Liberia Tolerant 52 78 28

Azucena Philippines Tolerant 52 60 55

WITA 7 AfricaRice Tolerant 42 75 38

Tukey-Kramer (P<0.05) 15 8.6 49

a Formerly the West Africa Rice Development Association (WARDA)

256 Plant Soil (2013) 372:253–264

Author's personal copy

the local farmers’ practice. No insecticide or pesticidewas used; however, two hand weedings were done at 3and 6 weeks after transplanting. Genotypes were scorednon-destructively at 4 and 8 weeks after transplanting forleaf bronzing using the SES as described above (IRRI1996). Grain yield was measured at maturity from 5-m2

subplots, with area under missing hills subtracted fromharvest area.

Statistical analysis

Statistical analyses were performed with SAS® version9.1. For continuous data, we used analysis of variance(ANOVA) after verifying that the residuals met thecriterion of normal distribution. When comparing upto three pre-determined means, we analyzed differencesbetween means by least significant difference (LSD),

but, when comparing more than three levels, we usedTukey-Kramer means comparisons instead.

Results

Experiment 1. Optimizing nutrient solution chemistry

Effect of pH and presence of chelator withoutplants The first experiment was conducted to determinethe pace of spontaneous chemical conversion of theinitially added reduced form of iron (Fe2+) to the oxi-dized form (Fe3+) in nutrient solution culture during thetypical time between before replenishment of the solu-tion (5 days). The amount of total Fe in solution afterday 5 remained higher at pH 3.0 than at pH 4.5 (Fig. 1),indicating a larger amount of iron oxide precipitation at

Fig. 1 Amount of Fe2+,Fe3+, and total Fe (sumof both) remaining insolution without a chelatorover time after addition of200 mg L−1 Fe2+ to Yoshidanutrient solution witha pH 4.5 and b pH 3.0during 5 days without plants

Plant Soil (2013) 372:253–264 257

Author's personal copy

the higher pH, as confirmed by the presence of a visiblebrown precipitate. On the fifth day, Fe2+ in the culturesolution with pH 3.0 was 120mgL−1, representing 60%of the initial Fe2+ concentration and 90 % of the total Feremaining in solution. The other 10 % of the totalsolution Fe was presumably present as Fe3+. In contrast,at pH 4.5, only 20 mg L−1 Fe2+ remained in solutionafter 5 days, which was only 10 % of the initial Fe2+

concentration and 67 % of the total Fe remaining insolution.

Optimizing Fe concentration, pH, and EDTA concen-tration in culture solution in the presence of plants Wefound that, regardless of EDTA or pH, the 200 mg L−1

Fe2+ solution did not induce severe enough stress.Both Azucena (tolerant) and IR64 (susceptible) werestill very green with very minimal bronzing. On theother hand, at 400 mg L−1 Fe2+, both genotypesshowed severe bronzing, making it difficult to differ-entiate them. Individual brown spots were no longervisible because the high density of spots coalesced toconvert the entire leaf into brown and folded (data notshown). Whereas, at 300 mg L−1, IR64 and Azucenashowed a clear distinction in their leaf-bronzing re-sponse to Fe toxicity stress (Fig. 2). Under this condi-tion, both genotypes showed stunted growth with

some bronzing, but the symptoms were more severefor the susceptible genotype, with brown spots thatoccurred initially at the tip of the leaves and thencovered the entire leaf by the fourth week. The Fetoxicity symptoms were more severe in the presenceof EDTA than without it regardless of pH, suggestingEDTA’s role in exerting more Fe toxicity stress onplants for a relatively longer time (Fig. 2). After ob-serving that 300 mg L−1 Fe2+ produced the appropriatemoderate stress in the presence of EDTA, we furtheroptimized the procedure by testing additional pHvalues and EDTA concentrations using only one Fe2+

concentration (300 mg L−1). When we tested pH var-iations between 3.8 and 4.2, we found no difference inleaf bronzing scores within this pH range (data notshown), and concluded that pH 4.0±0.2 was optimumfor maintaining moderate stress while minimizing ap-parent root damage, providing better Fe toxicity com-parisons than either pH 3.0 or 4.5.

The optimal EDTA concentration for producing visi-ble differences in leaf bronzing symptoms between thetolerant and susceptible genotypes was determined bysetting up another experiment with varying proportionsof Fe2+:EDTAmolar ratio. A lower EDTA concentration(1:1 ratio) could not distinguish the tolerant from thesensitive genotypes after one week of treatment exposure

Fig. 2 Leaf bronzing score of 4-week-old seedlings of ricecultivars IR64 (sensitive) and Azucena (tolerant) grown in nu-trient solution culture with varying Fe2+ concentration, pH, andpresence of EDTA. The shaded box shows the optimal score

range for visual differentiation between genotypes. Error barsindicate standard error of three replicates. For columns with novisible error bar, there was no variation in scores of thereplicates

258 Plant Soil (2013) 372:253–264

Author's personal copy

(Fig. 3). Even a 1:1 ratio of Fe2+:EDTA could havedistinguished Azucena from IR64 based on tolerancebut definitely symptoms would have appeared muchlater than for the 1:2 ratio. This is probably because onlya smaller proportion of Fe2+ was chelated with a 1:1 mo-lar ratio vis-à-vis 1:2 ratio and thus prevented fromoxidation and precipitation. Higher EDTA concentra-tions (1:4 Fe:EDTAmolar ratio and above) bound almostall the Fe2+ ions and kept them available in a reducedtoxic form for a longer duration, imposing more severetoxicity, resulting in a more rapid appearance of leafbronzing symptoms on both tolerant and sensitive plants,so that there was only a small time-windowduringwhichthe genotype differences could be distinguished, afterwhich the toxicity became too severe, causing poor plantgrowth and seedling death in both genotypes. Figure 3 isbased on the appearance of symptoms after 1 week ofstress treatment or about 10 days old plant (treatmentimposed on 4th days after germination). The stresssymptoms appeared very quickly on leaves just after1 week of treatment with 1:4 ratio. By 9–10 days aftertreatment, sensitive genotypes start dying and by 14–15 day after treatment, even the tolerant genotype starteddying. So the screening window was very short becauserate of symptoms appearance after 7 days of treatmentswere not same as on 8th, 9th and 10th day of treatment.Indeed the stress response was much faster after a weekof treatment. However, 1:2 ratio was found as sufficientstress to distinguish the genotypes and allows to keep thesensitive genotypes also surviving after 4 weeks of

treatment. Based on these observations, we found thatthe 1:2 molar ratio (Fe2+:EDTA) worked best fordistinguishing between tolerant and sensitive genotypes(Fig. 3) because it produced a quick response fromgenotypes and also provided a sufficient window toscore and allows sampling of the plants.

Experiment 2. Greenhouse validation of the optimizednutrient solution conditions

Significant variation in response to iron stress was ob-served when genotypes evaluated using optimized cul-ture solution and there was a high percentage ofreduction in plant height, root length and biomass forall 20 genotypes (Table 1). An example of the rootgrowth of tolerant vs. sensitive genotypes is shown inFig. 4 where tolerant variety has the capacity to producenew root growths which is not seen in the sensitivevariety. Although there was a high percentage of reduc-tion in the seedling vigor data (Table 1) upon exposureto Fe toxicity, and there were general trends towardmore severe reduction in the more sensitive lines, thecorrespondence between this data and the previouslyidentified “tolerance” level of each genotype was weak,perhaps due to different methodology used to ascertainthe tolerance of the genotypes in different studies; ordifferent genotypes may have different growth charac-teristics under stress because of their inherent stresstolerance mechanisms. For example, IR61246-3B-15-2-2-3, which was found to be one of the most tolerant

Fig. 3 Leaf bronzing scoresat 1 week after exposure to Fetoxicity, comparing IR64(susceptible) with Azucena(tolerant) varieties exposed tothe same Fe concentration(300 mg L−1) at differentFe:EDTA molar ratios at pH4.2. The shaded box showsthe optimal score range forvisual differentiation betweengenotypes. Error barsindicate standard error ofthree replicates. For columnswith no visible error bar,there was no variation inscores of the replicates

Plant Soil (2013) 372:253–264 259

Author's personal copy

genotypes in this study based on grain yield (OnlineResource 1), and LBS (≤ 5.0 at all different stages andplaces) had reduced their plant growth parameters by60–80 % after 4 weeks of stress in nutrient solution(Table 1). Leaf bronzing is one of the clearly visiblesymptoms of the Fe-toxicity in solution culture or field.It takes about 2 weeks for leaf bronzing symptoms toappear and they are first visible on older leaves. LeafBronzing Scores (LBS) of all the genotypes wererecorded after 2 and 4 weeks after treatment in solutionculture (Online resource 1). Development of leaf bronz-ing was quite conspicuous after 4 week when genotypesranged the LBS score from 3.7 to 9.0 in comparison to1.0 to 7.7 (mostly within range of 1.0 to 3.0) after2 weeks of treatment.

Similarly, Fe concentration in known tolerant geno-types were observed to be higher than those in thesensitive genotypes (Table 2), but the differences be-tween the groups were not large enough as evident fromnon-significant association (0.383) between field toler-ance scores for 15 genotypes and the percent increaseover control in the Fe concentration at 300 mg L−1 Festress in culture solution.

Experiment 3. Field validation of the solution culturescreening procedure

The same 19 genotypes (one damaged in nursery)screened in nutrient solution culture were grown andscreened for iron-toxicity tolerance in a field in San

Dionisio, Iloilo (Philippines), during the 2006wet season.In the field, plants were immediately subjected to irontoxicity upon transplanting, in contrast to plants grown inthe greenhouse, where iron toxicity was imposed 3 daysafter they were established in nutrient solution. Leafbronzing scores at the 4-week stage in the field rangedfrom 3.0 to 5.7 (Online Resource 1). PSBRc 18 and TCA4 had the least bronzing symptoms in the leaf whileBaoThai and two breeding lines (IR63262-AC 201 andIR65839-3B-1-2-3) had the most bronzing symptoms.

Results showed that the leaf bronzing scores obtainedin nutrient solution culture at 4 weeks correlated wellwith field grain yield, and with leaf bronzing scoresobtained in the field at 8 weeks (Table 3). A goodaverage performance as revealed by a low average leafbronzing score demonstrated that 2 weeks of stress insolution culture is not severe enough to clearly distin-guish genotypes based on tolerance of iron toxicity,except to isolate some very sensitive ones such asIR65839-3B-1-2-3. In view of this, our main interestwas to compare the correlations of grain yield in fieldwith the phenotypic leaf bronzing score (genotype per-formance) under three stress conditions: 4 weeks ofstress in solution culture in the greenhouse, 4 weeks ofstress as in a naturally stressed field, and 8 weeks ofstress in a naturally stressed field.When comparing boththe simple association of the measured data and the rankassociation of the relative genotype performance, wefound significant correlations (P<0.01) between green-house leaf bronzing scores at 4 weeks, field leaf

Fig. 4 The standardizedscreening technique (300 mgL−1 Fe as Fe2+, at pH 4, witha 1:2 molar ratio ofFe:EDTA) is able to distin-guish tolerant and sensitiverice genotypes. Treatmentwas imposed on 4th days oldseedlings and continued for4 weeks for this experiment.Inset: Shows poor growth ordeath of roots of a sensitivegenotype while a more toler-ant genotype shows thegrowthof new roots

260 Plant Soil (2013) 372:253–264

Author's personal copy

bronzing scores at 8 weeks, and grain yield in field. Thecorrelations involving leaf bronzing scores in the green-house at 2 weeks were all non-significant (Table 3).

Discussion

This study aimed to develop a solution culture basedsystem for high-throughput screening of rice genotypesfor Fe-toxicity. The main objective was to find out the

best combination of Fe2+ concentration, pH level,EDTA concentration, and the time of replenishment ofFe2+ in solution culture system that can mimic the fieldscreening under Fe-toxic conditions. Genotypes withproven tolerance and sensitivity to Fe- toxicity basedon field trial reports were used for this study.

The first experiment to determine the pace of sponta-neous chemical conversion was important to find out ifsufficient concentration of the reduced form of iron isretained in soils solution and to determine how muchtime is required to exert toxicity on plants in soil solu-tion. The solubility of Fe3+ in aqueous solutions is muchlower than that of Fe2+ and it decreases as pH increases(Kirk 2004). Thus, Fe3+ is expected to precipitate out ofsolution more readily than Fe2+, forming Fe(OH)3 orother forms of iron oxide, and this is what we observedmore at pH 4.5 rather than pH 3.0. Assuming that theFe3+ was precipitating out of solution soon after itsformation according to its expected chemical equilibria,it can be concluded that the most important reason thelower-pH solution maintained a higher Fe concentrationafter 5 days was that the conversion of Fe3+ to Fe2+ wasdelayed. Our observation of this pH effect on Fe2+ oxi-dation indicated that our nutrient solution system appro-priately simulated this aspect of soil solution chemistry,as, according to Kirk and Bajita (1995), pH is the dom-inant factor in soil solution that affects the time forferrous iron to undergo oxidation to a ferric state. Thelower the pH, the longer the time required for completionof oxidation (Kirk 2004). However, if the pH is too low,it may negatively affect root functionality by causing theroot membranes to leak, thus impairing plant growth(Reid et al. 1980; Tadano and Yoshida 1978).

For a screening method to be effective, it is importantto cause a moderate stress to the plants in order todistinguish between tolerant and susceptible genotypes.If the stress is not severe enough to produce toxicitysymptoms on even the susceptible genotypes, or if it isso severe that all genotypes grow poorly, the differencebetween tolerance and intolerance will not be clear. Byusing the Azucena and IR 64 as proven tolerant andsensitive genotypes (Gregorio et al. 2002; Mendoza etal. 2000; Sahrawat and Singh 1998; Wang and Peverly1998) and series of small multi-factor experiments, theoptimum concentration of Fe2+, pH level and chelatorconcentration was determined. Clear differentiation be-tween tolerant and sensitive genotype with respect toFe-toxicity symptoms and enough time window forphenotyping were the main criteria to choose the

Table 2 Fe concentration in the leaves of 15 genotypes of riceafter 4 weeks of growth in control or Fe-toxic nutrient solution

Fe concentration

Variety Control Fe-toxic % Increaseover controla

mg Fe/kg leaf weight

Karuna 900 9,950 1,000

Kinandang Patong 1,640 9,290 470

IR63262 -AC 201 1,000 6,650 570

IR65839-3B-1-2-3 1,100 8,360 660

Bao Thai 840 6,420 660

Average, susceptible# 1,100 8,130 670

Mashuri 540 4,900 810

WITA 1 1,350 7,670 470

WITA 2 900 5,210 480

Average, moderatelytolerant#

930 5,930 590

IR61246-3B-15-2-2-3 1,320 10,080 660

IR61612-3B-16-2-2-1 1,000 5,730 470

IR61640-3B-14-3-3-2 1,050 9,820 840

TCA4 780 3,310 320

Suakoko 8 1,120 5,620 400

Azucena 1,200 6,880 470

WITA 7 1,100 5,960 440

Average, tolerant# 950 6,770 520

Average, all 990 7,060 610

Tukey-Kramer (P<0.05) 4,800

a Calculated as the Fe‐toxic� controlð Þ control � 100=

#genotypes proven for their degree of tolerance from previousstudies

Plant Soil (2013) 372:253–264 261

Author's personal copy

optimum combination. It was concluded that the optimalnutrient solution conditions for Fe toxicity screeningwere 300 mg L−1 Fe (supplied as Fe2+), pH 4.0 (±0.2),with a 1:2 Fe2+:EDTA molar ratio. We also observedthat the nutrient solution needed to be replaced at leastonce every 5 days along with replenishment of Fe2+

supply in order to maintain sufficient macronutrientsfor plant growth. These conditions were therefore usedfor subsequent validation of the screening method withadditional genotypes.

Screen-house validation using 20 genotypes differingfor their Fe-toxicity tolerance was done using the solu-tion culture screening system that was standardized fol-lowing many small experiments. The standardizedscreening system worked well as far as reduction in theseedling vigor upon exposure to Fe-toxicity wasconcerned. Sensitive genotypes exhibited more stresssymptoms than tolerant ones. The correspondence ofseedling-stage screening of 20 genotypes with theirexpected tolerance level based on field studies was notvery strong, due to some of the genotypes like IR61246-3B-15-2-2-3 which showed different reaction than whatwas expected based on previous studies. This differencecould be probably due to the different inherent tolerancemechanisms operating in different set of genotypes ordue to different methodology used in previous studies.This is substantiated by our small experiments to fix themost appropriate combination (Yoshida nutrient solutionwith 300 mM excess Fe2+ with EDTA in 1:2 molar ratioat pH 4.0±0.2) for high throughput screening, because

by altering just one parameter in our experiments, resultsvaried with respect to behavior of proven genotype.This is the reason that growth reduction percent overnon-stress was not considered as an important or solecriterion to find out the degree of stress tolerance ofgenotypes.

Fe uptake as indicated by Fe concentration in differ-ent genotypes was also studied to find out if there is anyassociation between plant uptake and sensitivity or tol-erance. This parameter did not give any clear trend,hence cannot be used as selection criterion. This couldbe due to the individual genotype characteristics, prob-ably because plant response to Fe-toxicity tolerance de-pends on many mechanisms, such as exclusion, tissuetolerance, initial vigor, and the active oxygen speciesscavenging system (Becker and Asch 2005; Tadano1976). Rice roots diffuse oxygen into the root mediathrough aerenchyma and make the rhizosphere moreoxidized (Marschner 1995), often resulting in the for-mation of Fe3+-oxide plaque on the root surface (Chen2006; Zhang et al. 1998; Zhang et al. 1999).Additionally, exudates from rice roots are also activein the conversion of Fe2+ to Fe3+ (Ahmad and Nye1990). The genotypes showing high Fe accumulationat high biomass production indicate a tissue tolerancemechanism, for example, IR61640-3B-14-3-3-2(Tables 1 and 2). In the case of genotypes accumulatingless Fe, such as TCA4, it is possible that iron is excludedfrom the roots or is sequestered in the roots so that it isnot readily taken up into the shoots.

Table 3 Correlation coefficients among simple association andrank association of different screening parameters based on 20genotypes. Greenhouse experiments were done in nutrient solu-tion culture at pH 4.0±0.2 with 300 mg L−1 Fe supplied as Fe2+

(details in text) and 1:2 molar ratio of Fe:EDTA; and fieldexperiments were carried out in a naturally Fe-toxic site at SanDionisio, Iloilo

Association between Pearson coefficient ‘r’for simple association

Spearman corrected ‘r’for rank association

Greenhouse LBSa at 2 weeks vs. field grain yield 0.287 0.186

Greenhouse LBS at 4 weeks vs. field grain yield 0.628** 0.610**

Field LBS at 4 weeks vs. field grain yield 0.530* 0.521*

Field LBS at 8 weeks vs. field grain yield 0.796** 0.779**

Greenhouse LBS at 2 weeks vs. field LBS at 4 weeks 0.399 0.294

Greenhouse LBS at 2 weeks vs. field LBS at 8 weeks 0.452 0.375

Greenhouse LBS at 4 weeks vs. field LBS at 4 weeks 0.608** 0.575*

Greenhouse LBS at 4 weeks vs. field LBS at 8 weeks 0.786** 0.746**

*, ** significant at 5 % and 1 % level, respectivelya LBS leaf bronzing score

262 Plant Soil (2013) 372:253–264

Author's personal copy

Symptoms of leaf bronzing take about 2 weeks toappear on older leaves (Becker and Asch 2005;Fairhurst and Witt 2002; Yamanouchi and Yoshida1981). There are a few genotypic exceptions to theleaf bronzing trends, which could be attributed todifferent tolerance mechanisms. For example,Suakoko 8, which has been reported as a tolerantvariety, showed leaf bronzing symptoms on the olderleaves at a very early stage of stress application, but itwas the one that survived the longest after prolongedstress. This implies that Suakoko 8 might be a goodaccumulator and have better tissue tolerance.

The ultimate step to validate the screening proce-dure efficacy was the field validation. Significant cor-relations in Table 3 support the hypothesis thatscreening of genotypes using a standardized screeningsystem in solution culture can be effectively used tomimic field evaluation and predict relative genotypicperformance under natural field stress conditions. Forhigh-throughput screening of large numbers of geno-types in breeding programs, our data show that leafbronzing scores at 4 weeks in nutrient solution(300 mg L−1 Fe supplied as Fe2+, at pH 4, with a1:2 molar ratio of Fe:EDTA) or at 8 weeks in the fieldcould be used as the evaluation criterion to screengenotypes for Fe-toxicity tolerance instead of growingall the genotypes until maturity for grain yield evalu-ation. This standardized method was employed insolution culture and validated by field screening using20 genotypes and Fig. 4 clearly shows the distinctionbetween the sensitive and tolerant genotype. It wasclearly observed that tolerant genotypes were able todevelop new roots (white roots in Fig. 4 inset) contin-uously when old roots become non-functional due todeposition of Ferric-oxide plaques on old roots surfacewhile sensitive genotypes are unable to produce newroots. Screening in the phytotron can increase the paceof screening in a high-throughput manner during anytime of the year. Where phytotron space is limited ornot available, screening in the field at 8 weeks insteadof waiting until maturity would enable a larger numberof early-generation breeding lines to be tested at least2 times in each season.

Conclusions

The establishment of a reproducible, inexpensive, effi-cient, and reliable high-throughput screening technique

is a prerequisite for a successful breeding program todevelop iron-toxicity-tolerant genotypes. It is necessaryfor the screening procedure to discriminate betweensensitive and tolerant genotypes in a way that is compa-rable to growing the genotypes to maturity in the fieldunder natural stress. Several hydroponics-based culturemethods have been reported but none of them measuredthe toxicity in terms of Fe2+ remaining over time in theculture solutions. In our study, we systematically testednutrient solution chemistry combinations to optimizethe amount of Fe2+ stress using two genotypes, and thenvalidated the procedure with 20 genotypes, comparingthe nutrient solution culture technique with field evalu-ation up to maturity. The optimized screening techniqueused Yoshida nutrient solution culture with 300 mg L−1

Fe supplied as Fe2+ at pH 4.0, in the presence of a1:2 molar ratio of Fe:EDTA, with phenotyping doneby scoring leaf bronzing after 4 weeks. Field screeningfor 8 weeks duration at the Fe-toxic hot spots is alsoa feasible option to screen large number of geno-types in absence of phytotron screening facility.This technique will help molecular breeding effortsfor the identification of robust QTLs or markersusing a quick and repeatable phenotyping throughhigh-throughput screening of Fe-toxicity-tolerantgenotypes under controlled conditions.

Acknowledgments The authors deeply acknowledge the re-search support of the deputy director general for research andhead, Plant Breeding, Genetics, and Biotechnology, at IRRI,who enabled us to conduct this research work.

References

Ahmad AR, Nye PH (1990) Coupled diffusion and oxidation offerrous iron in soils. I. Kinetics of oxygenation of ferrousiron in soil suspension. J Soil Sci 41:395–409. doi:10.1111/j.1365-2389.1990.tb00075.x

Asami T, Kumada K (1959) A new method for determining freeiron in paddy soils. Soil Plant Food 5:141–146

Audebert A, Fofana M (2009) Rice yield gap due to iron toxicityin West Africa. J Agron Crop Sci 195:66–76. doi:10.1111/j.1439-037X.2008.00339.x

Audebert A, Sahrawat KL (2000) Mechanisms for iron toxicitytolerance in lowland rice. J Plant Nutr 23:1877–1885.doi:10.1080/01904160009382150

Becker M, Asch F (2005) Iron toxicity in rice-conditions andmanagement concepts. J Plant Nutr Soil Sci 168:558–573.doi:10.1002/jpln.200520504

Briat J (2002) Metal-ion-mediated oxidative stress and its con-trol. In: Montagu M, Inze D (eds) Oxidative stress inplants. Taylor and Francis, London, pp 171–190

Plant Soil (2013) 372:253–264 263

Author's personal copy

Camara A (2006) Testing and developing tolerant rice varietiesto iron toxicity in lower guinea (cra kilissi and koba). Irontoxicity in rice-based systems in West Africa. Africa RiceCenter (WARDA), Cotonou, pp 64–74

Chen RF (2006) Response of rice (Oryza sativa) with rootsurface iron plaque under aluminium stress. Ann Bot98:389–395. doi:10.1093/aob/mcl110

De Dorlodot S, Lutts S, Bertin P (2005) Effects of ferrous irontoxicity on the growth and mineral composition of an interspe-cific rice. J Plant Nutr 28:1–20. doi:10.1081/pln-200042144

Fairhurst TH, Witt C (2002) Rice: A practical guide to nutrientmanagement. International Rice Research Institute (IRRI),Manila

Gregorio GB, Senadhira D, Mendoza RD, Manigbas NL, RoxasJP, Guerta CQ (2002) Progress in breeding for salinitytolerance and associated abiotic stresses in rice. FieldCrop Res 76:91–101. doi:10.1016/s0378-4290(02)00031-x

Gridley HE, Efisue A, Tolou B, Bakayako T (2006) Breedingfor tolerance to iron toxicity at WARDA. Iron toxicity inrice-based systems in West Africa. Africa Rice Center(WARDA), Cotonou, pp 96–111

IRRI (1996) Standard evaluation system for rice. InternationalRice Research Institute (IRRI), Manila

Ishimaru Y, Suzuki M, Tsukamoto T, Suzuki K, Nakazono M,Kobayashi T, Wada Y, Watanabe S, Matsuhashi S, TakahashiM, Nakanishi H, Mori S, Nishizawa NK (2006) Rice plantstake up iron as an Fe3+-phytosiderophore and as Fe2+. Plant J.45:335–346. doi:10.1111/j.1365-313X.2005.02624.x

Kirk GJD (2004) The biogeochemistry of submerged soils. JohnWiley & Sons, Ltd., Chichester

Kirk GJD, Bajita JB (1995) Root-induced iron oxidation, pHchanges and zinc solubilization in the rhizosphere of low-land rice. New Phytol 131:129–137. doi:10.1111/j.1469-8137.1995.tb03062.x

Kpongkor DS (2003) Developing a standardized procedure toscreen lowland rice (Oryza sativa) seedlings for toleranceto iron toxicity. MS Dissertation, University of Bonn

Li H, Yang X, Luo AC (2001) Ameliorating effect of potassiumon iron toxicity in hybrid rice. J Plant Nutr 24:1849–1860

Loeppert RH, Inskeep WP (1996) Iron. In: Sparks DL (ed)Methods of soil analysis part 3 chemical methods. SoilScience Society of America, Madison, pp 639–664

Marschner H (1995) Mineral nutrition of higher plants.Academic, San Diego

Mendoza RD, Molinawe JA, Gregorio GB, Guerta CQ, Brar DS(2000) Genetic variability of tolerance for iron toxicity indifferent species of Oryza and their derivatives. In: KhushGS, Brar DS, Hardy B (eds) Advances in rice genetics.Supplement to rice genetics IV. Proceedings of the Fourth

International Rice Genetics Symposium. International RiceResearch Institute, Los Baños, pp 154–157

Olaleye AO (1998) Characterization, evaluation, nutrient dy-namics, and rice yields of selected wetland soils inNigeria. Dissertation, University of Ibadan

Reid MS, Paul JL, Young RE (1980) Effects of pH and ethephonon betacyanin leakage from beet root disks. Plant Physiol66:1015–1016. doi:10.1104/pp. 66.5.1015

Sahrawat KL (2004) Iron toxicity in wetland rice and the role ofother nutrients. J Plant Nutr 27:1471–1504. doi:10.1081/pln-200025869

Sahrawat KL, Singh BN (1998) Seasonal differences in irontoxicity tolerance of lowland rice cultivars. Int Rice ResNotes 23:18–19

Shimizu A, Guerta CQ, Gregorio GB, Ikehashi H (2005)Improved mass screening of tolerance to iron toxicity inrice by lowering temperature of culture solution. J PlantNutr 28:1481–1493. doi:10.1080/01904160500201352

Stumm W, Lee GF (1961) Oxygenation of ferrous iron. Ind EngChem 53:143–146

Tadano T (1976) Devices of rice roots to tolerate high iron con-centration in growth media. Japan Ag Res Quart 9:34–39

Tadano T, Yoshida S (1978) Chemical changes in submergedsoils and their effect on rice growth. Soil and rice.International Rice Research Institute (IRRI), Los Baños

Wan JL, Zhai HQ, Wan JM, Ikehashi H (2003) Detection andanalysis of QTLs for ferrous iron toxicity tolerance in rice,Oryza sativa L. Euphytica 131:201–206

Wang T, Peverly JH (1998) Screening a selective chelator pairfor simultaneous determination of iron(II) and iron(III).Soil Sci Soc Am J 62:611–617

Wu P, Hu B, Liao CY, Zhu JM, Wu YR, Senadhira D, PatersonAH (1998) Characterization of tissue tolerance to iron bymolecular markers in different lines of rice. Plant Soil203:217–226

Yamanouchi M, Yoshida S (1981) Physiological mechanisms ofrice’s tolerance for iron toxicity. Paper presented at IRRISaturday Seminar Series, Los Baños. International RiceResearch Institute (IRRI), Los Baños

Yoshida S, Forno DA, Cock JH, Gomez KA (1976) Laboratorymanual for physiological studies of rice. International RiceResearch Institute (IRRI), Los Baños

Zhang X, Zhang F, Mao D (1998) Effect of iron plaque outsideroots on nutrient uptake by rice (Oryza sativa L.). Zincuptake by Fe-deficient rice. Plant Soil 202:33–39

Zhang XK, Zhang FS, Mao DR (1999) Effect of iron plaqueoutside roots on nutrient uptake by rice (Oryza sativa L.):phosphorus uptake. Plant Soil 209:187–192. doi:10.1023/a:1004505431879

264 Plant Soil (2013) 372:253–264

Author's personal copy