Th

GARa

b

c

ARRAA

KOGRGGO

1

sptetrtdTf(uo

BT

(

0h

Plant Science 208 (2013) 93– 101

Contents lists available at SciVerse ScienceDirect

Plant Science

jo ur nal home p age: www.elsev ier .com/ locate /p lantsc i

he mitochondrial glutathione peroxidase GPX3 is essential for H2O2

omeostasis and root and shoot development in rice

isele Passaiaa, Leila Spagnolo Foninia,b, Andréia Caverzana,b, Douglas Jardim-Messedera,na Paula Christoff a,b, Marcos Letaif Gaetac, Jorge Ernesto de Araujo Mariathc,ogério Margisb, Marcia Margis-Pinheiroa,b,∗

Department of Genetics, Federal University of Rio Grande do Sul, RS, BrazilBiotechnology Center, Federal University of Rio Grande do Sul, RS, BrazilDepartment of Botany, Federal University of Rio Grande do Sul, RS , Brazil

a r t i c l e i n f o

rticle history:eceived 21 January 2013eceived in revised form 11 March 2013ccepted 17 March 2013vailable online 8 April 2013

eywords:

a b s t r a c t

Glutathione (GSH) peroxidases (GPXs: EC 1.11.1.9 and EC1.11.1.12) are non-heme thiol peroxidases thatcatalyze the reduction of H2O2 or organic hydroperoxides to water, and they have been identified inalmost all kingdoms of life. The rice glutathione peroxidase (OsGPX) gene family is comprised of 5members spread throughout a range of sub cellular compartments. The OsGPX gene family is inducedin response to exogenous H2O2 and cold stress. In contrast, they are down regulated in response todrought and UV-B light treatments. Transgenic rice plants have been generated that lack mitochondrial

ryza sativaramineaeaicePXlutathione peroxidasexidative stress

OsGPX3. These GPX3s plants showed shorter roots and shoots compared to non-transformed (NT) plants,and higher amounts of H2O2 mitochondrial release were observed in the roots of these plants cultivatedunder normal conditions. This accumulation of H2O2 is positively associated with shorter root lengthin GPX3s plants compared to NT ones. Moreover, GPX3 promoter analysis indicated that it is mainlyexpressed in root tissue. These results suggest that silencing the mitochondrial OsGPX3 gene impairsnormal plant development and leads to a stress-induced morphogenic response via H2O2 accumulation.

. Introduction

Whereas animals can move away from environmental adver-ities such as water deficit, high temperature or heavy metalollution beyond others, plants are sessile organisms. As a result,he anti-oxidant and repair system of plants is considerablynhanced and diversified compared to bacterial or animal sys-ems because plants also have to cope with a second source ofeactive oxygen species (ROS) from chloroplasts, which is beyondhe mitochondrial electron transport chain [1]. The most abun-ant ROS produced during stress is hydrogen peroxide (H2O2).his compound is not only an oxidant but also a signal generatedrom superoxide ions through the action of superoxide dismutases

SOD) [2]. It is important that H2O2 levels are properly controllednder stress or normal conditions, and this is regulated by therchestrated participation of a range of different enzymatic and

∗ Corresponding author at: Universidade Federal do Rio Grande do Sul, Avenidaento Gonc alves, 9500, CEP: 915910-970, Porto Alegre, RS, Brasil.el.: +55 51 3308 9814; fax: +55 51 3308 7311.

E-mail addresses: marcia.margis@ufrgs.br, marcia.p.margis@gmail.comM. Margis-Pinheiro).

168-9452/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.plantsci.2013.03.017

© 2013 Elsevier Ireland Ltd. All rights reserved.

non-enzymatic antioxidant systems. The non-enzymatic antiox-idant compounds are represented by carotenoids, tocopherols,reduced glutathione (GSH) and ascorbate (AsA), while the enzy-matic systems rely on superoxide dismutases (SOD), catalases(CAT), ascorbate peroxidases (APX), peroxiredoxins (PRXs) and glu-tathione peroxidases (GPXs) [3]. Glutathione (GSH) Peroxidases(GPXs: EC 1.11.1.9 and EC1.11.1.12) are non-heme thiol peroxidasesthat catalyze the reduction of H2O2 or organic hydroperoxides towater, and they have been identified in almost all kingdoms of life[4]. In animals, GPXs are selenoproteins that contain selenocys-teine at the catalytic site, whereas the plant enzymes do not bindselenium. The active site is composed of a catalytic triad formedby selenocysteine/cysteine, glutamine and tryptophan. This differ-ence allows plants to preferentially reduce thioredoxin instead ofGSH, which is opposite the preference of the enzymes in animals[4,5]. The rice glutathione peroxidase (OsGPX) gene family consistsof 5 members. Phylogenic analysis has predicted that GPX is local-ized to the mitochondria (GPX1 and GPX3), chloroplasts (GPX4) andcytosol/endoplasmic reticulum (GPX2 and GPX5) [6].

There is increasing evidence that GPXs are able to interactwith other proteins, conferring peroxide-induced oxidation inaddition to their well-established ROS scavenging function. Forexample, in yeast, GPX3 activates the Yeast Activation Protein 1

9 Scien

(taapI

sdpfctraTmmitrpR[Nsptstgddr

2

2

1gft5pvupigpt5it

LLtaCGCGA

4 G. Passaia et al. / Plant

Yap-1) transcription factor via such an oxidation-induced evento promote the activation of genes encoding thiol-reducing andntioxidant proteins [7]. In Arabidopsis thaliana, GPX3 links ABAnd H2O2 signaling during stomatal closure by interacting withrotein phosphatase type 2C (PP2C) proteins, such as Abscisic acid

nsenstive (ABI) 1 and ABI2 [8].Rice roots are typical of semi-aquatic plants in that they have

pecialized tissues that allow for root growth during flooding con-itions [9]. The aerenchyma is a tissue that is specially designed toromote gas exchange from the shoot, and it also acts as a reservoiror oxygen, which is required for root respiration under floodingonditions. Characteristics such as deep, thick, branched root sys-ems are correlated with better survival and crop productivity. Aange of phytohormones, such as abscisic acid (ABA) and auxin,re known to be involved in root development and architecture.he accumulation of ABA in roots under water stress is required foraintaining primary root elongation [10], and auxin transport pro-otes root growth via auxin maxima conditions [11,12]. Moreover,

t was recently shown that ABA accumulation modulates auxinransport in the rice and Arabidopsis root tip to maintain primaryoot elongation under water stress [13]. ABA also triggers H2O2roduction via plasma membrane NADPH oxidases (RbohD andbohF), which increase cytosolic Ca2+, leading to stomata closure14]. On the other hand, auxin promotes ROS production throughADPH oxidase activation, which plays a role in root gravitropism;

pecifically, H2O2 causes root curvature as a downstream com-onent of the auxin-mediated signaling pathway [15]. To date,he functions of rice GPXs have not been reported. The presenttudy was undertaken to assess the functional characterization ofhe GPX gene family in rice, especially GPX3, using a knockdownene silencing approach to abolish GPX3. These transgenic plantsisplayed short shoot and root phenotypes and increased H2O2 pro-uction compared to NT plants, indicating that GPX3 is involved inice development processes in an H2O2-dependent manner.

. Materials and methods

.1. Vectors construction and plant transformation

The target gene sequence from OsGPX3 (LOC Os02g44500;97 bp) was inserted into the pANDA RNAi vector [16] using theateway LR reaction (Invitrogen®). cDNAs were used as a templateor PCR reactions. The following primer pairs were used to amplifyhe sequence: RNAiGPX3: 5′-CACCACCGCCCCGTTCATCTCT-3′ and′-ATGCTGAACAGCGCGAAC-3′. The hairpin RNA produced by theANDA vector is driven by the maize ubiquitin promoter. Toerify the GPX3 promoter expression pattern, a 2 kb fragmentpstream of the start codon was amplified by PCR, cloned intoENTR (Invitrogen®) and recombined with pHGWFS7 [17], produc-

ng the vector proGPX3. The proGPX3 vector contains the Gus andfp reporter genes and also hpt as selection marker gene that conferslant resistance to hygromycin. The following primers were usedo amplify the hpt gene: 5′-CACCCGTTTCTCGATGACATGACG-3′ and′-ATAGCGGTAAGTCCGGTGTG-3′. The plasmids were introduced

nto rice calli via Agrobacterium tumefaciens-mediated transforma-ion, as described by Upadhyaya [18].

Using cDNA from leaves, OsGPX genes (1–5:OC Os04g46960, LOC Os03g24380, LOC Os02g44500,OC Os06g08670, LOC Os11g18170) were amplified usinghe following primers: GPX1: 5′-ATGGCCGCCGCGCCGTCC-3′

nd 5′-AGAGCTCCCAAGCAGCTTCTTGATA-3′; GPX2: 5′-ATGGGGG-GGCGGAATCC-3′ and 5′-ATCCTCGAGCGCCTTCAGGATGTC-3′;

PX3: 5′-ATGGCCACCGCCGCCTCC-3′ and 5′-AGAACTTCCAAGCAG-TTCTTGATG-3′; GPX4: 5′-ATGGCGTCCACCACCACC-3′ and 5′-GGCTGCCTCACCATCAGTA-3′; GPX5: 5′-ATGGCGGCTACAACT-CAAG-3′ and 5′-CGAGGTCCCAAGTAGCTTCT-3′. PCR products

ce 208 (2013) 93– 101

were cloned into the pART7-HA-YFP plasmid [19] using Gatewaytechnology (Invitrogen®), in fusion with the Yellow FluorescentProtein (YFP) coding sequence at its N-terminus, under control of35S promoter of CaMV. The constructs were used to transform riceprotoplasts.

Protoplast isolation was performed as described by Chen et al.[20], with minor modifications, and protoplast transformation wasperformed as described by Tao et al. [21]. After transformation, theprotoplasts were incubated for 24–48 h in the dark at 28 ◦C beforeimaging. To evaluate fluorescence, an Olympus FluoView 1000confocal laser-scanning microscope, equipped with a set of filterscapable of distinguishing between green and yellow fluorescentprotein (EGFP and EYFP, respectively) and plastid autofluorescence,was used. The images were captured with a high-sensitivity pho-tomultipler tube detector.

2.2. Plant growth, abiotic stress and hormone treatments

Rice plants (Oryza sativa L. ssp. Japonica cv. Nipponbare) wereused in this study. The plants were grown in a growth chamberwith supplemental lighting (8 h dark/16 h light; 150 mol m−2 s−1) at28 ◦C or in a greenhouse (sunlight), submerged in H2O. Seeds weregerminated in H2O at 28 ◦C in the dark. Four-day-old seedlings weretransferred to plastic pots containing Furlani’s solution (48 mg/LN–NO3

−; 12 mg/L N–NH4−; 200 mg/L Ca; 200 mg/L K; 40.6 mg/L

Mg; 8.0 mg/L P; 151 mg/L S; 234 mg/L Cl; 4.85 mg/L Fe; 0.67 mg/LMn; 0.36 mg/L B; 0.20 mg/L Zn; 0.05 mg/L Cu; and 0.11 mg/L Mo)and grown at 28 ± 2 ◦C, with a 12 h photoperiod.

Non-transgenic rice plants were used to evaluate OsGPX geneexpression during vegetative and reproductive development. Riceseeds were sterilized and germinated on filter paper, and afteremergence, they were transferred to vermiculite. Leaves and roottissues were sampled 7, 14 and 30 days after germination (DAG).Moreover, panicle samples were taken during reproductive devel-opment (R), based on the following stages [22]: R2, R3, R4, R5, R6,R7, R8 and R9. OsGPX gene expression was evaluated in responseto exogenous H2O2 treatment, cold, UV-B light and drought stress.

To analyze the response of rice plants to exogenous H2O2 treat-ment, plants were germinated and grown in 1/2 Murashige andSkoog (MS) medium. After 4 weeks of growth, the seedlings weretransferred to liquid MS supplemented with 10 �M H2O2 and incu-bated for 2, 4, and 8 h. Control plants were maintained in MSwithout H2O2.

For cold stress, 3-week-old rice plants were exposed to a10 ◦C climate for 24 h. Control plants remained in the conditionsdescribed previously. Shoot samples were taken after this period.

For aluminum treatment, 4-day-old seedlings were grown innutrient solution supplemented with 20 ppm AlCl3, and plantmaterial was sampled at 4 h and 8 h after treatment.

For UV-B light treatment, 3-week-old rice plants were exposedto UV-B irradiation (254 nm) for 4 h. Samples were collected imme-diately after exposure (time 0) and at 6 h after exposure.

For drought stress, seedlings were grown to the four-leaf stageon soil supplemented with water. After this period, the seedlingswere separated into two lots: control and drought stressed. Thecontrol lot continued to receive a normal supply of water, whereaswater was withheld from the stressed seedlings for 6 and 15 days.After this period, plant material was collected from both lots ofplants.

Transgenic rice plants in which OsGPX3 was silenced wereobtained via A. tumefaciens-mediated transformation of riceembryogenic calli (O. sativa L. ssp. Japonica cv. Nipponbare), and

they were induced from seeds and cultivated in NB mediumat 28 ◦C in the dark [23]. T1 rice seeds from OsGPX3s plantsand non-transformed (NT) were germinated in MS medium(Sigma–Aldrich) supplemented with hygromycin under controlled

Scien

c(tas2tapr

2p

GGudsa

2

tBspsrq

2

arcfMbccP1pCGOGaA3tsaTOGuCeue

G. Passaia et al. / Plant

onditions (150 �mol m−2 s−1 photosynthetically active radiationPAR), 25 ◦C, 80% relative humidity and a 12 h photoperiod). Non-ransformed rice and GPx3s seeds were grown in MS media withnd without hygromycin (30 mg/mL), respectively. After 8 days,eedlings were transferred to Furlani’s solution and grown at8 ± 2 ◦C with a 12 h photoperiod for 30 days. The pH of the solu-ion was adjusted to 5.4 using 0.1 N HCl. After 10 days, the shootnd root lengths were measured. Three biological replicates wereerformed for each experiment; a pool of at least three plants pereplicate was used.

.3. Analysis of the Gpx3 promoter expression pattern in ricelants

Transgenic plants were used to assess the expression of theus gene under the control of the GPX3 promoter (promGPX3-US). Leaves, roots, stem and panicles were sampled and analyzedsing the X-Gluc histochemical assay (Fermentas®), as previouslyescribed [24], with minor modifications. After Gus staining, theamples were clarified with graded ethanol series (30–70%) andnalyzed in stereomicroscope and bright field microscopy.

.4. Determination of mitochondrial H2O2 release

The amount of H2O2 released by rice roots was determined byhe Amplex Red oxidation method, as described previously [25].riefly, roots (20 mg) were incubated in 20 mM MES buffer (pH 6.0)upplemented with 10 mM Amplex Red and 5 units/mL horseradisheroxidase. Fluorescence was monitored at excitation and emis-ion wavelengths of 563 nm (slit 5 nm) and 587 nm (slit 5 nm),espectively. Calibration was performed by the addition of knownuantities of H2O2.

.5. Quantitative real-time PCR (RT-qPCR)

Total RNA was isolated using the TRIZOL reagent (Invitrogen®),ccording to the manufacturer’s protocol, and the quality of theesulting RNA was assessed by gel electrophoresis. To eliminateontaminant genomic DNA, samples were treated with RQ1 RNase-ree DNase, according to the manufacturer’s instructions (Promega,

adison, WI, USA). Synthesis of first-strand cDNA was performedy incubating 1 �g of total RNA with the M-MLV Reverse Trans-riptase (Promega) and a 24-polyTV primer (Invitrogen®). AfterDNA synthesis, the samples were diluted 100-fold in sterile water.rimers were designed to produce DNA fragments ranging from80 to 250 bp. PCR amplifications were performed using specificrimers for OsGPX1: 5′-AGCAACCTGCACTTATGCACT-3′ and 5′-AGCAAGGAAATTTATTGACATGA-3′; OsGPX2: 5′-CTGGTTGGGTA-GCACTGTT-3′ and 5′-TGCAAACACAAACCTTACGCTAC-3′;sGPX3: 5′-TTGCATTGAGCACTTGGAAC-3′ and 5′-AGGGGCAAA-TGATGCAGTA-3′; OsGPX4: 5′-CTGTACATATGCCTTGCCTCA-3′

nd 5′-GTTACAGGGGCCAGATAAGC-3′; OsGPX5: 5′-AAGATTG-GAATGATATCCAGAAGC-3′ and 5′-GCAAACCACATTCTTACGAACA-′. The following primers were used as internal controlso normalize the amount of mRNA present in eachample: OsActina2: 5′-GGACGTACAACTGGTATCGTGTT-3′

nd 5′-GTTCAGCAGTGGTAGTGAAGGAG-3′; OsFdh3: 5′-TTCCAA-GCATTCAAAGCTG-3′ and 5′-CAAAATCAGCTGGTGCTTCTC-3′;seF1˛: 5′-TTTCACTCTTGGTGTGAAGCAGAT-3′ and 5′-ACTTCCTTCACGATTTCATCGTAA-3′. qPCR assays were carried outsing the Step One plus PCR System (Applied Biosystems, Foster

ity, CA, USA). Three biological replicates were performed forach experiment; a pool of at least three plants per replicate wassed. Also, a total of 4 technical replicates were performed forach reaction, and expression data analyses were performed after

ce 208 (2013) 93– 101 95

comparative quantification of amplified products using the 2−��Ct

method [26,27].

2.6. Statistical analysis

The data represent the means ± standard error of the mean ofthe three housekeeping genes (SEM). Statistical analysis was per-formed using Student’s test, and a p value of 0.05 was consideredto be statistically significant.

3. Results

3.1. GPX proteins are mainly localized in plant cell organelles

Margis et al. [6] identified 6 GPX genes in the rice genome. Sub-sequent analysis revealed that OsGPX6 is actually one of the foursplice forms of OsGPX1. In the referenced work, in silico analysispredicted that OsGPX1 and OsGPX3 were localized in the mitochon-dria, OsGPX2 in cytosol, OsGPX4 in chloroplasts and OsGPX5 incytosol anchored at the endoplasmic reticulum. In the presentwork, we experimentally analyzed the sub-cellular localization ofthe rice GPX family. We constructed a translational fusion of differ-ent OsGPX that expressed YFP, driven by the CaMV 35S promoter,in rice protoplasts. Confocal analysis of protoplasts expressing35S-OsGPX1::YFP, 35S-OsGPX2::YFP and 35S-OsGPX3::YFP, 35S-OsGPX4::YFP and 35S-OsGPX5::YFP fusions revealed that YFPfluorescence was localized at the mitochondria (Fig. 1A), cyto-sol and chloroplasts (Fig. 1B), mitochondria (Fig. 1C), chloroplasts(Fig. 1D) and cytosol/endoplasmic reticulum and chloroplasts(Fig. 1E), respectively. Tetramethyl rhodamine ethyl ester per-chlorate (TMRE), a mitochondria-specific fluorescent marker, wasused in rice protoplasts transformed with 35S-OsGPX1::YFP and35S-OsGPX3::YFP constructs to confirm their mitochondrial local-ization. 35S-OsGPX2::YFP and 35S-OsGPX5::YFP were also presentbeyond the cytosol or cytosol/endoplasmic reticulum, respectively,in chloroplasts.

3.2. GPX gene expression is mainly regulated by exogenous H2O2

To verify how exogenous H2O2 modulates rice GPXs genes, riceplants were submitted to 10 �M H2O2 treatment, and RT-qPCRanalysis was performed. All of the GPX genes examined were signif-icantly induced after 2, 4 and 8 h of treatment, except Gpx4, whichwas induced after 8 h of treatment (Fig. 2A). Moreover, to elucidatehow the GPX gene family responds to abiotic stress, the expressionof GPX genes was analyzed after exposure to cold, drought, UV-Blight, and aluminum. Rice GPXs were not modulated by 20 ppm alu-minum after 4 or 8 h in roots (data not shown). On the other hand,GPX1, GPX3 and GPX5 expression were significantly induced whenplants were subjected to a 10 ◦C climate for 24 h (Fig. 2B). Rice plantscultivated without water for 15 days displayed a significant reduc-tion in the expression of GPX4 and GPX5 (Fig. 2C). Regarding UV-Blight treatment, the expression of GPX3 and GPX4 was significantlyreduced immediately after UV-B light exposure and after 6 h, whilethe expression of GPX1, GPX2 and GPX5 was significantly reducedonly after 6 h of UV-B light exposure (Fig. 2D).

GPX4 showed the highest expression in the leaves 14 and 30DAG (days after germination) followed by GPX3 at 14 DAG. In roots,the most highly expressed gene was GPX3 (Suppl. Fig. 1A). Duringreproductive development, all of the GPX genes showed very low

expression levels at the R2 and R3 stages, while GPX5 presented thehighest level of expression at R4, R5 and R6. Moreover, the expres-sion of GPX2 and GPX4 was higher at stages R5 and R6, respectively(Suppl. Fig. 1B).

96 G. Passaia et al. / Plant Science 208 (2013) 93– 101

Fig. 1. Glutathione peroxidase proteins (OsGPX1, OsGPX2, OsGPX3, OsGPX4 and OsGPX5) are localized in mitochondria, chloroplasts and cytosol. Green signals indicateYFP fluorescence, red signals indicate chlorophyll auto fluorescence, and yellow signals represent the merged image. The OsGPX1 and OsGPX3 proteins are localized in themitochondria of rice protoplasts, as determined by transient expression of the 35S-OsGPX1::YFP and 35S-OsGPX3::YFP cassettes (A, F and K; C, H and M), respectively:cytosol, 35S-OsGPX2::YFP (B, G and L); chloroplast, 35S-OsGPX4::YFP cassette (D, I and N); and cytosol/endoplasmic reticulum, 35S-OsGPX5::YFP cassette (E, J and O). (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

G. Passaia et al. / Plant Science 208 (2013) 93– 101 97

Fig. 2. The relative expression of GPX genes in rice shoots under different abiotic stress conditions: 10 �M H2O2, after 2, 4 and 8 h of treatment (A); 24 h at 10 ◦C (B); 6 and 15d nt the means ± SE of three biological replicates. Asterisks indicate significant differences,p

3

ctGspoaace

3s

gs(c8ptst(shros

Table 1Similarities among the sequences of the hairpin construction and the rice GPX genes.The number of contiguous nucleotides shared between each rice GPX gene and thetarget sequence of the hairpin construct are indicated in parentheses.

RNAi constructs Gpx1 Gpx2 Gpx3 Gpx4 Gpx5

ays without water (C); and 0 and 6 h of UV-B light treatment (D). The data represe < 0.05.

.3. OsGPX3 is mainly expressed in the root tissues of rice plants

Rice calli, transformed with the proGPX3-Gus construct, wereultured with hygromycin. Gus positive calli were selected, andhe plants were regenerated and incubated in X-gluc to visualizePX3 promoter expression patterns. Tissue from the leaves, roots,tems and spikelets of two different lineages were evaluated. TheroGPX3 expression showed Gus staining in the phloem regionf adventitious root of stem primordial basis (Fig. 3A and B) andpices of young roots, specifically in the differentiation zone (Fig. 3Cnd D). Moreover, Gus staining was visualized in three cellular tri-homes in palea/lemma of the spikelet (Fig. 3E and F) and in thembryo of mature seeds (Fig. 3G).

.4. Post-transcriptional knockdown of OsGPX3 affects root andhoot growth

To determine the functional role of mitochondrial (OsGPX3)lutathione peroxidase in rice, the OsGPX3 gene was specificallyilenced using inverse repeat (IR) constructs transcribing dsRNAhairpin). In rice, the nucleotide sequence of GPX genes is wellonserved. Identities among the family members vary from 46 to1%, and genes encoding proteins of the same sub-cellular com-artment share long, identical, contiguous sequence stretches (upo 30 bp in length). Because of the high similarity in the codingequence region between GPX genes, we designed an IR constructranscribing dsRNA for a specific 197 bp region of the OsGPX3 5′UTRTable 1). Rice calli were transformed with the RNAiOsGPX3 con-truct, generating 6 transgenic lines. GPX3s plants showed lower

eight compared to NT (Fig. 4A). The mRNA level of GPX3s waseduced to 5% of the control NT plants (Fig. 4B). Evaluation of thether family members revealed a 40% reduction in OsGPX2 expres-ion in the GPX3s transgenic plants compared to NT plants, but

RNAiGPX3 (197nt) 3% (6nt) 1% (2nt) 100% (197nt) 3% (6nt) 3% (6nt)

none of the other genes were altered significantly (Fig. 4C). Wegerminated three GPX3s line and non-transformed plants in MSmedia, with and without hygromycin, respectively. After 8 days, theseedlings were transferred to a nutrient solution for 10 days, andthe shoot and root lengths were measured. All of the GPX3s plantlines showed shorter shoots (Fig. 4D) and roots (Fig. 4E) comparedto control NT plants.

3.5. Silencing of GPX3 expression leads to higher levels ofmitochondrial H2O2 release

Three lines of GPX3s and non-transformed rice plants weregerminated in MS media for 8 days and then cultivated for 10days in nutrient solution. Then, H2O2 mitochondrial release wasmeasured. GPX3s plants released an average of 20 (lines 6 and8) and 36 (line 7) times more H2O2 in their roots compared towild type plants (Fig. 5A). While NT plants produced 0.00627 pmolH2O2 mg tissue−1 min−1, the GPX3s-6 and GPX3-8 transgenic linesproduced 0.1302 and 0.1208 pmol H2O2 mg tissue−1 min−1, respec-tively, and GPX3s-7 produced 0.2272 pmol H2O2 mg tissue−1 min−1.Moreover, a positive ratio existed between ROS production and

root length (Fig. 5B), where the highest production of H2O2 cor-related with short roots in all transgenic lines. However, the wildtype plants showed longer root length and less H2O2 production.Notably, GPX3s-7 released the highest amount of ROS, and it also

98 G. Passaia et al. / Plant Science 208 (2013) 93– 101

Fig. 3. Analysis of the expression patterns of the rice GPX3 promoter gene merged to the Gus sequence. (A and B) Six-month-old rice plants: phloem of stem and root bases;( elet of

ps

4

lsaiTii

C and D) elongation zone of young root apices; (E–G) young trichomes of rice spik

resented the shortest root length in average between the linestudied (Fig. 5B).

. Discussion

Using in silico analysis, we previously predicted the sub-cellularocalization for all members of the OsGPX family [6]. In the presenttudy, our data confirmed those predictions, and found that OsGPX2nd OsGPX5 were located in the chloroplasts of rice protoplasts,

n addition to their expected cytosolic localization (Fig. 1B and E).his unexpected localization was also reported for ascorbate perox-dase, glutathione reductase and monodehidroascorbato reductasen Arabidopsis, that demonstrated dual targeting to mitochondria

6 months; and (H) embryo of a fully expanded rice seed.

and plastids [28]. Additional studies are necessary to confirm thechloroplast localization of OsGPX2 and 5.

Expression analysis showed the strong induction of all OsGPXgenes in response to H2O2 treatment, suggesting that rice GPXsparticipate in maintaining ROS homeostasis and redox signaling(Fig. 2A). In contrast, we did not observe any change in GPX expres-sion in rice roots treated with Al (data not shown). These resultsare not in agreement with what has been reported for rye roots[29], where GPX transcripts were rapidly induced by Al after 4 hand peaked at 8 h, before decreasing to normal levels. OsGPX was

down regulated after drought treatment, in contrast with whathas been observed for OsAPX genes under the same conditions;among the 8 OsAPX members, 5 were induced after 15 days ofdrought [30]. The absence of modulation or repression of OsGPX

G. Passaia et al. / Plant Science 208 (2013) 93– 101 99

Fig. 4. Characterization of GPX3-silenced plants (A–C) for mitochondrial OsGPX3. (A) Non-transformed (NT) and three independent GPX3 line plants were grown undercontrolled conditions. (B) Quantitative determination of OsGPX3 mRNA in the leaves of NT and GPX3s plants grown under control conditions. The transcript level of OsGPX3 inNT plants was used to normalize transcript accumulation in GPX3s and NT plants. (C) Expression analyses of all OsGPX genes in GPX3s plants compared to NT plants. The valuesr ice pls ith an

gtbastttti[doihds

Fde

epresent the means ± SD (N = 3). Shoot (D) and root (E) length (cm) of 10-day-old rignificantly different expression of individual GPX compared to NT is designated w

enes under abiotic stress could be explained by the strength ofhe cell redox state changes during different abiotic stresses. Inarley, GPX was not affected by cold, heat or drought treatment,nd only a slight increase was observed in roots treated withalt and H2O2 [31]. In animals, GPXs function as key enzymes inhe detoxification of ROS, especially H2O2. In plants, detoxifica-ion of ROS is mainly attributed to the catalases and enzymes ofhe ascorbate-glutathione cycle. Moreover, it has been reportedhat during severe and persistent stress, ascorbate peroxidases arenhibited, and GPXs become the main H2O2 scavenging enzymes31]. Similarly, in salt-sensitive citrus cells, ascorbate peroxidasesecreased, whereas GPX activity increased with time [32]. Previ-usly, our group demonstrated that OsGPX expression was inducedn cytosolic ascorbate peroxidase knockdown rice plants under

eat stress and normal conditions [33]. Furthermore, GST and GPXetoxify other toxic organic hydroperoxides, which increase duringevere and permanent stress-induced lipid peroxidation. In the case

0.3

0.2

0.0

GPX3

s-8

GPX3

s-7

GPX3

s-6

*

*

H2

O2

Rele

ase

( pm

ol H

2O

2.m

g t

issu

e-1

.min

-1)

*

NT

0.1

A

ig. 5. ROS generation in rice roots. The effect of OsGPX3 silencing was evaluated in threifferent GPX3s plant lines (L-6, 7, and 8) (A), and calculated the ratio between ROS releaxperiments. Asterisks indicate significant differences, p < 0.05, and the population mean

ants; the values represent the means ± SD (N = 10). Each plant line that promotes a asterisk (*).

of strong stress conditions, where basal antioxidant mechanismsare exhausted, more effective responses are activated by GST/GPXor GPX [31]. In rice, the 5 GPX members appear to play differentroles during plant development and in response to abiotic stress.

The evaluation of OsGPX gene expression during rice vegeta-tive development revealed that OsGPX3 is the predominant isoformexpressed in roots, while OsGPX4 showed the highest expressionin leaves, followed by OsGPX3 (Suppl. Fig. 1A). The expression ofthe OsGPX3 promoter was detected in the adventitious root ofstem primordia basis (Fig. 3A). More specifically, the OsGPX3 pro-moter was found in the phloem region of the provascular meristem(Fig. 3B) and was not detected in the stem primordia itself. Thepromoter activity was also detected in young adventitious roots(Fig. 3C), specifically in the elongation zone (Fig. 3D). This obser-

vation is also agrees with the RiceXPro database analysis for thisgene, which indicated that OsGPX3 is expressed mainly in the elon-gation zone and maturation zone I (Suppl. Fig. 2D), specifically the

NT

0,0 0,1 0,2 0,3

6

8

10

L.7L.6

L.8

Ro

ot

len

gth

(cm

)

H2O2 Re lease

(pMol H2O2.mg tissue1.min-1)

B

e different lines. The flux of H2O2 release was evaluated fluorometrically in threese and root length (B). The values represent the means ± SE of four independents

s are significantly different.

1 Scien

eroedtegipm

drcwrco2irfih[akgbitaapdcddaonnlTiuc

A

vAI(PYR

A

f2

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

00 G. Passaia et al. / Plant

ndodermis, pericycle and stele (Suppl. Fig. 2E). The activity of GPX-elated genes, in response to stress, has been found in trichomesf Nicotiana tabacum, indicating the role of these structures in thelimination of ROS [34]. Interestingly, strong Gus staining was alsoetected in the embryo of mature rice seed, which could be impor-ant for radicle and coleoptile emergence [22]. In all cases whichxpression was detected in OsGPX3-GUS represent tissues under-oing intense cell division and differentiation. This observation isn agreement with Herbette et al., [35], which suggest that GPX canrotect dividing cells from oxidative stress, helping to maintain theitotic activity.To understand the role of mitochondrial OsGPX3 in rice antioxi-

ant metabolism, this gene was silenced in transgenic lines. Theeduction of OsGPX3 expression by approximately 95% did nothange the response of other OsGPX members, except OsGPX2,hich was down regulated (Fig. 4C). OsGPX3 knockdown effectively

educed the shoot and root lengths of 18-day-old T1 rice plantsompared to wild type plants (Fig. 4D and E, respectively). More-ver, the flux of H2O2 release from rice roots was approximately0 times more intense compared to NT plants (Fig. 5A). Interest-

ngly, the highest H2O2 production also correlated with reducedoot length (Fig. 5B). Thus, H2O2 fine-tuning seems to be crucialor the determination of cell proliferation/elongation or cell arrest-ng. For example, auxin induces H2O2 in the root tip and the apicalalf of the root elongation zone, driving root gravitropism in maize15]. OsGPX3 promoter analysis showed Gus staining in the apicesnd elongation zones of young roots, which suggests that OsGPX3nockdown leads to increased production of H2O2, impairing rootrowth. Moreover, it was recently suggested that root growth inhi-ition was a consequence of redox homeostasis disturbance and

s a general stress-induced morphogenic response (SIMR) of rootso different stresses [36]. As OsGPX3 knockdown plants produce

high level of H2O2, this theory could explain their shorter rootnd shoot phenotype. Key elements in the ontogenesis of the SIMRhenotype appear to be stress-affected gradients of ROS, antioxi-ants, auxin and ethylene. These gradients are integrated on theellular level, affecting cell division, cell elongation and/or cellifferentiation [36]. Additionally, a moderate increase in ROS pro-uction caused root growth inhibition and the induction of GPXctivity in cells through cell wall crosslinking [37]. Taken together,ur results suggest that mitochondrial OsGPX3 is important forormal rice development and that the shorter root and shoot phe-otype of GPX3s plants could be explained by the increased H2O2

evels and conversion to a stress-induced morphogenic response.he functional analysis of genes belonging to the GPX gene fam-ly in rice, which are an important monocot species, could aid ournderstanding of the stress response processes and other uniqueharacteristics of this group.

cknowledgments

This work was supported by Conselho Nacional de Desen-olvimento Científico e Tecnológico (CNPq), Coordenac ão deperfeic oamento de Pessoal de Nível Superior (CAPES) and the

nternational Centre for Genetic Engineering and BiotechnologyICGEB). We are grateful to Universidade Federal do Ceará (UFC),rofessor Joaquim Albenísio da Silveira, Fabrício E.L. Carvalho andugo Lima Melo for rice seed propagation and Carolina Werneribeiro for useful discussions.

ppendix A. Supplementary data

Supplementary data associated with this article can beound, in the online version, at http://dx.doi.org/10.1016/j.plantsci.013.03.017.

[

ce 208 (2013) 93– 101

References

[1] N. Rouhier, et al., Redox based anti-oxidant systems in plants: biochemical andstructural analyses, Biochim. Biophys. Acta 1780 (2008) 1249–1260.

[2] C. Laloi, K. Apel, A. Danon, Reactive oxygen signalling: the latest news, Curr.Opin. Plant Biol. 7 (2004) 323–328.

[3] C.S. Koh, et al., Crystal structures of a poplar thioredoxin peroxidase thatexhibits the structure of glutathione peroxidases: insights into redox-drivenconformational changes, J. Mol. Biol. 370 (2007) 512–529.

[4] S. Herbette, P. Roeckel-Drevet, J.R. Drevet, Seleno-independent glutathioneperoxidases. More than simple antioxidant scavengers, FEBS J. 274 (2007)2163–2180.

[5] S. Toppo, S. Vanin, V. Bosello, S.C.E. Tosatto, Evolutionary and structural insightsinto the multifaceted glutathione peroxidase (Gpx) superfamily, Antioxid.Redox Signal. 10 (2008) 1501–1514.

[6] R. Margis, C. Dunand, F.K. Teixeira, M. Margis-Pinheiro, Glutathione peroxidasefamily – an evolutionary overview, FEBS J. 275 (2008) 3959–3970.

[7] A. Delaunay, D. Pflieger, M.B. Barrault, J. Vinh, M.B. Toledano, A thiol peroxidaseis an H2O2 receptor and redox-transducer in gene activation, Cell 111 (2002)471–481.

[8] Y. Miao, et al., An Arabidopsis glutathione peroxidase functions as both a redoxtransducer and a scavenger in abscisic acid and drought stress responses, PlantCell 18 (2006) 2749–2766.

[9] J. Rebouillat, et al., Molecular genetics of rice root development, Rice 2 (2008)15–34.

10] D. Sengupta, M. Kannan, A.R. Reddy, A root proteomics-based insightreveals dynamic regulation of root proteins under progressive droughtstress and recovery in Vigna radiata (L.) Wilczek, Planta 233 (2011)1111–1127.

11] V.a. Grieneisen, J. Xu, A.F.M. Marée, P. Hogeweg, B. Scheres, Auxin transport issufficient to generate a maximum and gradient guiding root growth, Nature449 (2007) 1008–1013.

12] H.S. Robert, J. Friml, Auxin and other signals on the move in plants, Nat. Chem.Biol. 5 (2009) 325–332.

13] W. Xu, et al., Abscisic acid accumulation modulates auxin transport in the roottip to enhance proton secretion for maintaining root growth under moderatewater stress, New Phytol. (2012), http://dx.doi.org/10.1111/nph.12004.

14] J.M. Kwak, V. Nguyen, J.I. Schroeder, The role of reactive oxygen species inhormonal responses, Plant Physiol. 141 (2006) 323–329.

15] J.H. Joo, Y.S. Bae, J.S. Lee, Role of auxin-induced reactive oxygen species in rootgravitropism, Plant Physiol. 126 (2001) 1055–1060.

16] D. Miki, K. Shimamoto, Simple RNAi vectors for stable and transient suppressionof gene function in rice, Plant Cell Physiol. 45 (2004) 490–495.

17] M. Karimi, B. De Meyer, P. Hilson, Modular cloning in plant cells, Trends PlantSci. 10 (2005) 103–105.

18] N.M. Upadhyaya, X.-R. Zhou, Q.-H. Zhu, A. Eamens, M.-B. Wang, P.M. Water-house, E.S. Dennis, Transgenic Rice, in: L. O’Brien, R.J. Henry (Eds.), TransgenicCereals, AACC, Minnesota, 2002, pp. 28–87.

19] C.S. Galván-Ampudia, R. Offringa, Plant evolution: AGC kinases tell the auxintale, Trends Plant Sci. 12 (2007) 541–547.

20] S. Chen, L. Tao, L. Zeng, M.E. Vega-Sanchez, W.G.-L. Kenji Umemura, A highlyefficient transient protoplast system for analyzing defence gene expres-sion and protein – protein interactions in rice, Mol. Plant Pathol. 7 (2006)417–427.

21] L. Tao, A.Y. Cheung, H. Wu, Plant rac-like GTPases are activated byauxin and mediate auxin-responsive gene expression, Plant Cell 14 (2002)2745–2760.

22] P.a. Counce, T.C. Keisling, A.J. Mitchell, A uniform, objective, and adaptive sys-tem for expressing rice development, Crop Sci. 40 (2000) 436.

23] Y. Hiei, S. Ohta, T. Komari, T. Kumashiro, Efficient transformation of rice (Oryzasativa L.) mediated by Agrobacterium and sequence analysis of the boundariesof the T-DNA, Plant J. Cell Mol. Biol. 6 (1994) 271–282.

24] R.A. Jefferson, T.A. Kavanagh, M.W. Bevan, GUS fusions: beta-glucuronidase asa sensitive and versatile gene fusion marker in higher plants, EMBO J. 6 (1987)3901–3907.

25] A.M.O. Smith, R.G. Ratcliffe, L.J. Sweetlove, Activation and function ofmitochondrial uncoupling protein in plants, J. Biol. Chem. 279 (2004)51944–51952.

26] T.D. Schmittgen, K.J. Livak, Analyzing real-time PCR data by the comparative CTmethod, Nat. Protoc. 3 (2008) 1101–1108.

27] K.J. Livak, T.D. Schmittgen, Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method, Methods (San Diego,Calif.) 25 (2001) 402–408.

28] O. Chew, J. Whelan, a.H. Millar, Molecular definition of the ascorbate-glutathione cycle in Arabidopsis mitochondria reveals dual targetingof antioxidant defenses in plants, J. Biol. Chem. 278 (2003)46869–46877.

29] M.A.R. Milla, et al., Expressed sequence tag-based gene expression analysisunder aluminum stress in rye, Plant Physiol. 130 (2002) 1706–1716.

30] S.B. Rosa, et al., Cytosolic APx knockdown indicates an ambiguous redoxresponses in rice, Phytochemistry 71 (2010) 548–558.

31] L. Halusková, K. Valentovicová, J. Huttová, L.T. Igor Mistrík, Effectof abiotic stresses on glutathione peroxidase and glutathione S-transferase activity in barley root tips, Plant Physiol. Biochem. 47 (2009)1069–1074.

Scien

[

[

[

[

[

G. Passaia et al. / Plant

32] Y. Gueta-Dahan, Z. Yaniv, B.a. Zilinskas, G. Ben-Hayyim, Salt and oxidativestress: similar and specific responses and their relation to salt tolerance incitrus, Planta 203 (1997) 460–469.

33] A. Bonifacio, et al., Role of peroxidases in the compensation of cytosolic ascor-

bate peroxidase knockdown in rice plants under abiotic stress, Plant CellEnviron. 34 (2011) 1705–1722.

34] E. Harada, et al., Expression profiling of tobacco leaf trichomes identi-fies genes for biotic and abiotic stresses, Plant Cell Physiol. 51 (2010)1627–1637.

[

ce 208 (2013) 93– 101 101

35] S. Herbette, et al., Immunolocalization of a plant glutathione peroxidase-likeprotein, Planta 219 (2004) 784–789.

36] G. Potters, T.P. Pasternak, Y. Guisez, M.a.K. Jansen, Different stresses, similarmorphogenic responses: integrating a plethora of pathways, Plant Cell Environ.

32 (2009) 158–169.

37] M.C. Córdoba-Pedregosa, J.M. Villalba, F. Córdoba, J.a. González-Reyes, Changesin growth pattern, enzymatic activities related to ascorbate metabolism, andhydrogen peroxide in onion roots growing under experimentally increasedascorbate content, J. Plant Growth Regul. 26 (2007) 341–350.