Journal of Experimental Botany, Vol. 65, No. 5, pp. 1403–1413, 2014 doi:10.1093/jxb/ert486 Advance Access publication 27 January, 2014 This paper is available online free of all access charges (see http://jxb.oxfordjournals.org/open_access.html for further details)

Research paper The effects of redox controls mediated by glutathione on root architecture in Arabidopsis thaliana

Gisele Passaia1,2, Guillaume Queval1,*, Juan Bai1,3, Marcia Margis-Pinheiro2 and Christine H. Foyer1,† 1 Centre for Plant Sciences, School of Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, UK 2 Depto. Genética, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves 9500, Prédio 43.312, CEP 91501–970 Porto Alegre, RS, Brazil 3 College of Life Science, Northwest A&F University, Shaanxi 712100, China

* Present address: Gregor Mendel Institute of Molecular Plant Biology, Austrian Academy of Sciences, 1030 Vienna, Austria. † To whom correspondence should be addressed. E-mail: [email protected]

Received 5 November 2013; Revised 10 December 2013; Accepted 13 December 2013

Abstract Glutathione peroxidases (GPXs) fulfil important functions in oxidative signalling and protect against the adverse effects of excessive oxidation. However, there has been no systematic characterization of the functions of the different GPX isoforms in plants. The roles of the different members of the Arabidopsis thaliana GPX (AtGPX) family were therefore investigated using , , , , , , and T-DNA insertion mutant lines. The shoot phenotypes were largely similar in all genotypes, with small differences from the wild type observed only in the gpx2, gpx3, gpx7, and gpx8 mutants. In contrast, all the mutants showed altered root phenotypes compared with the wild type. The gpx1, gpx4, gpx6, gpx7, and gpx8 mutants had a significantly greater lateral root density (LRD) than the wild type. Conversely, the gpx2 and gpx3 mutants had significantly lower LRD values than the wild type. Auxin increased the LRD in all genotypes, but the effect of auxin was significantly greater in the gpx1, gpx4, and gpx7 mutants than in the wild type. The application of auxin increased GPX4 and GPX7 transcripts, but not GPX1 mRNAs in the roots of wild-type plants. The synthetic strigolactone GR24 and abscisic acid (ABA) decreased LRD to a similar extent in all genotypes, except gpx6, which showed increased sensitivity to ABA. These data not only demonstrate the importance of redox controls mediated by AtGPXs in the control of root architecture but they also show that the plastid-localized GPX1 and GPX7 isoforms are required for the hormone-mediated control of lateral root development.

Key words: Abscisic acid, auxin, glutathione, oxidative stress, redox regulation, root architecture, strigolactones.

Introduction Plant cells contain a large array of that and DHAR). In parallel, glutathione peroxidases (GPXs), control the metabolism of oxidants such as reactive oxy- glutathione S- (GSTs), and gen species (ROS) and also play a role in redox signalling (PRXs) reduce H2O2 and hydroperoxides by ascorbate-inde- (Foyer and Noctor, 2009, 2011). The induction of antioxi- pendent thiol-mediated pathways (Dietz et al., 2002; Chang dant enzymes is a particularly important component of plant et al. 2009). While many classes of GSTs display stress responses that limit oxidant-induced programmed cell activity (Dixon et al., 2009), in general they are only able to death (PCD) (Wu et al., 2010). Within the antioxidant net- metabolize H2O2 at low rates (Mannervik, 1985). work of plant cells, ascorbate peroxidases reduce H2O2 at the GPXs are non-haem thiol hydroperoxidases, which catalyse expense of ascorbate, which is then regenerated by monode- the reduction of H2O2 or organic hydroperoxides to water or hydroascorbate and dehydroascorbate reductases (MDHAR corresponding alcohols (Herbette et al., 2007). The

© The Author 2014. Published by Oxford University Press on behalf of the Society for Experimental Biology. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 1404 | Passaia et al. of these enzymes is composed of a formed by the harmful effects of uncontrolled oxidation (Gaber et al., /cysteine, glutamine, and tryptophan (Herbette 2012). However, there has been no systematic characterization et al., 2007). In contrast to the animal GPXs, which are sele- of the functions of the different GPX isoforms in Arabidopsis. noproteins containing a selenocysteine at the catalytic site, Moreover, the precise roles of each GPX isoform in leaves and the plant enzymes do not bind (Eshdat et al., 1997; roots remain largely uncharacterized. The following experi- Herbette et al., 2007; Toppo et al., 2008). Moreover, the ani- ments were therefore undertaken to explore the roles of the mal GPXs exclusively use GSH as the reducing substrate but, different AtGPX isoforms using gpx1, gpx2, gpx3, gpx4, gpx6, despite their current nomenclature, the plant GPXs prefer gpx7, and gpx8 T-DNA insertion mutant lines. A preliminary reduced thioredoxin (TRX) as reductant and they have com- characterization of the different mutant lines revealed that the paratively low activities with GSH (Iqbal et al., 2006; Navrot shoot phenotypes were largely similar in all genotypes, regard- et al., 2006; Herbette et al., 2007). They are therefore better less of the GPX composition. Due to the absence of differences described as thiol peroxidases (Navrot et al., 2006). in shoot parameters between the gpx mutants and the wild type, GPXs are considered to play important roles in preventing the focus of this study was on the role of the different GPX oxidant-induced PCD. In Chlamydomonas reinhardtii, for exam- isoforms in the control of root architecture, where significant ple, a HOMOLOGOUS differences were observed between the gpx mutants and the wild (GPXH) gene is highly expressed in cells exposed to high light type. The aim of this study was to characterize the root phe- or singlet oxygen generation in the chloroplasts (Leisinger et al., notypes observed in the gpx mutants in relation to the action 2001; Fischer et al., 2007; Ledford et al., 2007). Constitutive of hormones that control root architecture, particularly auxin, over-expression of GPXH together with GSTS1 enhanced tol- strigolactones (SLs), and abscisic acid (ABA). erance to singlet oxygen, suggesting that GPXH is important in defences against photooxidative stress. GPXs interact with other proteins and hence they are consid- Materials and methods ered to have signalling functions (Delaunay et al., 2002; Miao et al., 2006). Some members of the GPX superfamily can form Identification and isolation of T-DNA insertion mutants tetrameric structures with other proteins in response to oxidants T-DNA insertion lines from SAIL (Sessions et al., 2002), SALK such as H O or hydroperoxides. In yeast, for example, GPX3 (Alonso et al., 2003), and WiscDsLox (Woody et al., 2007) collec- 2 2 tions were obtained. Donor stock numbers SALK_027373 (gpx1; is considered to activate the Yeast Activation Protein 1 (Yap-1) At2g25080), SALK_082445 (gpx2; At2g31570), SAIL_278_E06 transcription factor via an oxidation-induced event in order to (gpx3-1; At2g43350), SALK_071176 (gpx3-2; At2g43350), promote the activation of encoding thiol-reducing and SALK_139870 (gpx4; At2g48150), WISCDSLOX_321_H10 (gpx6; antioxidant proteins (Delaunay et al., 2002). The Arabidopsis At4g11600), SALK_023283 (gpx7; At4g31870), and SALK_127691 AtGPX3 interacts with protein phosphatase type 2C (PP2C) (gpx8; At1g63460) were obtained from the ABRC (Ohio State University) and NASC (Nottingham Arabidopsis Stock Center). All proteins such as ABSCISIC ACID INSENSITIVE (ABI) 1 2+ + T-DNA mutants except gpx3-1 are in the Col-0 background; gpx3- and ABI2, in order to activate plasma membrane Ca and K 1 is in the Col-3 background. Plants were germinated and grown channels that facilitate stomatal closure (Miao et al., 2006). Rice in conventional soil and, to confirm the position and the homozy- plants lacking GPX3 had shorter roots and smaller shoots than gosity of the insertion, PCR was performed with genomic DNA wild-type plants, indicating that GPX3 plays an essential role in using gene-specific oligonucleotides whose sequences are given in Supplementary Table S2 available at JXB online. The position of the root and shoot development (Passaia et al., 2013). Two stromal T-DNA insertion sequence and oligonucleotides for each genotype GPXs (AtGPX1 and AtGPX7) were shown to be important in are shown in Supplementary Fig. S1 available at JXB online. chloroplast functions, particularly light acclimation and also in Upon request, all novel materials described in this publication will plant immune responses (Chang et al., 2009). be made available in a timely manner for non-commercial research In Arabidopsis thaliana, GPXs are encoded by a small gene purposes, subject to the requisite permission from any third-party owners of all or parts of the material. family, named AtGPX1 to AtGPX8. The enzymes products of these genes are considered to be important redox sensors as well as , but the precise functions of the different GPX Growth on soil and rosette measurements isoforms remains poorly characterized. Each GPX member Wild types (Col-0 and Col-3) and gpx1, gpx2, gpx3-1, gpx6, gpx7, localizes to a distinct cellular location. While there is some con- and gpx8 mutant seeds were grown on compost in controlled envi- troversy in the literature regarding the intracellular localizations ronment chambers under optimal growth conditions under either of AtGPX3 and AtGPX6, there is a consensus that AtGPX1 short (8 h) photoperiod or long (16 h) photoperiod conditions for 4 weeks. Rosette diameters were measured every 3 d. Leaf number, and AtGPX7 are localized to the chloroplasts, AtGPX4 and leaf fresh and dry weight measurements, and leaf pigment determi- AtGPX5 are cytosolic, AtGPX2 is localized to the endoplas- nations were made on material harvested at 27/28 d. mic reticulum/cytosol, and AtGPX8 to the cytosol/apoplast (Rodriguez Milla et al., 2003; Margis et al., 2008) or the cytosol and nucleus (AtGPX8; Gaber et al., 2012). Phylogenetic analy- Growth on plates and hormone treatments sis indicates that the AtGPX3 protein is localized to the endo- Seeds of the wild type (Col-0) and gpx1, gpx2, gpx3-2, gpx4, gpx6, plasmic reticulum/cytosol and that AtGPX6 is localized in the gpx7, and gpx8 mutants were surface sterilized for 2 min in 75% ethanol and 5 min in 4% sodium hypochlorite, and washed sev- mitochondria (Margis et al., 2008). Each isoform within a given eral times in sterilized water until the pH was ~7. Sterilized seeds cellular compartment is considered to fulfil distinct functions. were then sown on 12 cm square plates containing half-strength For example, AtGPX8 protects the nucleus and cytosol from Murashige and Skoog medium (1/2 MS medium, pH 5.7) with 0.01% Glutathione peroxidase functions in Arabidopsis thaliana | 1405 myo-inositol, 0.05% MES, 1% sucrose, and 0.8% plant agar. Plates Results were stored for 2 d in the cold and dark to synchronize germination and were placed vertically in a plant growth cabinet with a 16 h pho- Relative abundance of AtGPX mRNAs in the gpx1, toperiod and a temperature of 22 °C. After 3 d, the seedlings were gpx2, gpx3, gpx4, gpx6, gpx7, and gpx8 mutants gently transferred using forceps to new plates containing the same compared with the wild type medium plus 2 μM GR24 (synthetic SL) and grown for a further 5 d. Auxin treatments were performed by transferring the seedling As a first step in the characterization of the functions of the dif- after a 5 d growth period to 1/2 MS medium containing 1 μM NAA (1-naphthaleneacetic acid) for 3 d. ABA was added to the media at ferent AtGPX isoforms, the relative abundance of the different concentrations of 0.3 μM or 0.5 μM. Seeds were placed directly on AtGPX transcripts was compared in the roots and shoots of ABA-containing medium, cold treated, and grown as above for 8 d. the gpx1, gpx2, gpx3-2, gpx4, gpx6, gpx7, and gpx8 mutants relative to the wild type (Col-0; Fig. 1; Supplementary Table Root architecture S1 available at JXB online). AtGPX1 and AtGPX2 transcripts were abundant in shoots. The level of AtGPX2 transcripts The root length and number of lateral roots formed were analysed on 8-day-old seedlings. Photos were taken and the root length was was highest in roots, as was the level of AtGPX6 mRNAs measured using ImageJ software. Lateral root density (LRD) was (Fig. 1A). AtGPX4 mRNAs were below the level of detection calculated by dividing the number of visible lateral roots by the pri- in roots and shoots of the wild-type plants (Fig. 1A). AtGPX1 mary root length for each root analysed. mRNAs were more abundant in the shoots of gpx3-2 and gpx4 mutants. The AtGPX2 transcripts were less abundant in Root staging the roots of the gpx7 mutants. AtGPX4 transcripts were less Measurements of the stages of lateral root development were per- abundant only in the shoots of the gpx3-2 mutant. AtGPX5 formed as described previously (Malamy and Benfey, 1997). Roots mRNA levels were higher in the roots of the gpx1 and gpx4 were incubated in 0.24 N HCl and 20% methanol at 62 °C for 20 min. mutants. AtGPX6 mRNA levels were lower in the root of the This solution was then replaced with 7% NaOH with 60% ethanol gpx7 mutant. AtGPX7 transcripts were increased in the shoots and the roots were then incubated for a further 15 min at room tem- perature. Following rehydration in 40% ethanol, 20% ethanol, and of the gpx4 mutants. AtGPX8 transcripts were less abundant 10% ethanol, the roots were infiltrated for 15 min in 5% ethanol in roots of gpx2, gpx3-2, and gpx7 mutants relative to the wild with 25% glycerol. The root samples were then maintained in 50% type. data for each mutant are provided in glycerol until microscopic analysis (M 4000-D, Swift). The stages of Supplementary Table S1 available at JXB online. primordium development were classified as described by Péret et al. (2009) as follows: stage I (single layered primordium composed of up to 10 small cells of equal length formed from individual or a pair Shoot phenotypes in the gpx1, gpx2, gpx3, gpx6, of pericycle founder cells), stage II (periclinal cell division forming gpx7, and gpx8 mutants an inner and an outer layer), stages III, IV, V, VI, and VII (anticli- nal and periclinal divisions create a dome-shaped primordium), and In order to determine the importance of the different AtGPX stage VIII (emergence of the primordium from the parental root). isoforms for shoot parameters, the shoot phenotypes of the different mutants (gpx1, gpx2, gpx3-1, gpx6, gpx7, and gpx8) Gene expression analysis were compared with those of the wild type after 4 weeks RNA was extracted from the shoots and roots of 10-day-old Col-0, growth under long (16 h) photoperiod growth conditions. gpx1, gpx2, gpx3-2, gpx4, gpx6, gpx7, and gpx8 plantlets grown in Shoot biomass (fresh weight per plant) and leaf pigment (chlo- 1/2 MS medium in a growth chamber at 20 °C with a 16 h photo- rophylls and carotene, expressed on a leaf area basis) contents period using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). were similar in all the lines (data not shown). However, the For each genotype, three biological replicates for the shoot and the root were prepared. Reverse transcription of 1 μg of RNA into gpx2 and gpx8 mutants had significantly fewer leaves than the cDNA was performed using the QuantiTect Reverse Transcription wild-type plants at the 28 d growth stage. In contrast, the gpx7 Kit (Qiagen). The qPCR was performed on 20 ng of cDNA with mutants had significantly more leaves than the wild type at a QuantiFast SYBR Green PCR kit (Qiagen) in the presence of this point (Fig. 2). To characterize the roles of the different 0.5 μM primers in a CFX96 thermocycler (Biorad, Hercules, CA, AtGPX isoforms in shoot development further, shoot growth USA). PCR conditions were as follows: 5 min 95 °C, 45 cycles 10 s 95°C and 30 s 60°C. Additionally melting curve analysis was per- was also examined under short photoperiod (8 h) growth con- formed at the end of each run to ensure specificity of the products. ditions. The shoot phenotypes of the mutants were very simi- The mean value of three replicates was normalized using Actin2 as lar to those of their respective wild types when plants were internal control (primer sequences can be found in Supplementary grown under short day conditions (Fig. 3). However, gpx3-1 Table S2 available at JXB online). The quality of amplification and gpx7 had a significantly greater rosette diameter than the curves was checked with the LinRegPCR software (Ramakers et al., 2003) and curves with R >0.995 and efficiencies between 1.9 and wild type after 4 weeks under these conditions (Fig. 3C, E). 2.1 were kept for the following expression value calculations. All expression data analyses were performed after comparative quantifi- Root phenotypes in the gpx1, gpx2, gpx3, gpx4, gpx6, –ΔΔCt cation of amplified products by using the 2 method (Livak and gpx7, and gpx8 mutants Schmittgen, 2001; Schmittgen and Livak, 2008). To explore the roles of the different AtGPX isoforms in the Statistical analysis control of root development, comparisons of the growth of Data represent the mean ±standard error of the mean (SEM). the primary roots, the number of lateral roots, and LRDs were Statistical analysis was performed by Student’s t-test. The values performed on 8-day-old seedlings (Fig. 4). Root phenotypes were considered statically different when P was <0.05. were compared in the different mutants (gpx1, gpx2, gpx3-1, 1406 | Passaia et al.

A 0.8 2 ACTIN

0.6

0.4 abundance to

ve 0.2 Relati 0 GPX1 GPX2 GPX3 GPX4 GPX5 GPX6 GPX7 GPX8 B GPX1 GPX2 GPX3 GPX4 GPX5 GPX6 GPX7 GPX8 gpx1 ------gpx2 ------gpx3-2 ------gpx4 ------gpx6 ------gpx7 ------gpx8 ------

Fig. 1. The relative expression of GPX genes in Col-0 (A) and in the different gpx mutants relative to the wild type (B). Black bars and arrows, shoots; white bars and arrows, roots. Data are the mean ±SE of 3–6 biological replicates.

20 in the presence of auxin (1 μM NAA; Fig 4). Auxin treat- ment resulted in a large increase in LRD in all of the geno- 16 * types. However, in contrast to the gpx2 and gpx3-2 mutants, es * * which showed a response to auxin similar to that of the wild 12 type (Fig. 4A–C), the effect of auxin was significantly greater in the gpx1, gpx4, and gpx7 seedlings than in the wild type. 8 Strigolactone (GR24)-dependent inhibition of lateral Number of leav 4 root development

0 The roles of the AtGPX isoforms in the SL-dependent con- Col--0gpx1gpx2gpx6gpx7gpx8CCol-3ol-3gpx3-1 trol of lateral root development were investigated in the gpx1, gpx2, gpx3-2, gpx4, gpx6, gpx7, and gpx8 mutants. Col-0 Fig. 2. The relative number of leaves in 4-week-old gpx mutant rosettes and mutant seedlings were grown either in the absence of relative to the wild types grown under long day conditions. Data are the mean ±SE of three plants. Asterisks indicate significant differences added hormone or in the presence of the synthetic SL, GR24 compared with the wild type, P<0.05. (2 μM; Fig. 4). As predicted from the known action of SLs on root architecture, the addition of GR24 inhibited lateral gpx3-2, gpx4, gpx6, gpx7, and gpx8) and the wild types (Col-0 root formation in all the genotypes studied here. The addi- and Col-3) in control conditions and in the presence of hor- tion of GR24 decreased LRD values to a similar extent in all mones. All the mutant lines showed large differences in LRD the genotypes except for the gpx4 seedlings, which showed a compared with the wild type in control conditions, with the significantly lower response to GR24-induced inhibition of gpx1, gpx4, gpx6, gpx7, and gpx8 mutants having greater LRD than the wild type (Fig. 4C). LRD values relative to the wild type (Fig. 4C). In contrast, gpx2 and gpx3-2 had a lower LRD than Col-0 (Fig. 4C). Abscisic acid-induced inhibition of lateral root development Auxin-dependent increases in lateral root density In order to determine whether any of the AtGPX isoforms were In order to determine whether any of the different AtGPX required for the ABA-dependent control of lateral root devel- isoforms were required for the auxin-dependent control of opment, root phenotypes were compared in gpx1, gpx2, gpx3- lateral root development, root phenotypes were compared in 1, gpx4, gpx6, gpx7, gpx8, Col-0, and Col-3 seedlings grown gpx1, gpx2, gpx3-2, gpx4, gpx6, gpx7, and gpx8 and Col-0 either in the absence of added hormone or in the presence of seedlings grown either in the absence of added hormone or ABA (Fig. 5A–C). Growth in the presence of ABA decreased Glutathione peroxidase functions in Arabidopsis thaliana | 1407

A 4 B 4 Col-0 gpx1 Col-0 gpx2 (cm) (cm) 3 3

2 2

1 * * 1 Rosette diameter Rosette diameter 0 0 10 13 16 19 22 25 28 10 13 16 19 22 25 28 Age of plants (day)Age of plants (day) Col-0 gpx1 Col-0 gpx2

CD5 4 Col-3 gpx3 -1 Col-0 gpx6 * (cm) (cm) 4 3 3 2 2 1 1 Rosette diameter Rosette diameter 0 0 10 13 16 19 22 25 28 10 13 16 19 22 25 28 Age of plants (day)Age of plants (day) Col-3 gpx3-1 Col-0 gpx6

EF4 4 Col-0 gpx7 * Col-0 gpx8 (cm) (cm) 3 3

2 * 2

1 * * 1 Rosette diameter Rosette diameter 0 0 10 13 16 19 22 25 28 10 13 16 19 22 25 28 Age of plants (day)Age of plants (day)

Col-0 gpx7 Col-0 gpx8

Fig. 3. A comparison of rosette phenotypes in 4-week-old gpx mutants relative to the wild types grown under short day conditions. (A) gpx1 mutant; (B) gpx2 mutant; (C) gpx3-1 mutant; (D) gpx6 mutant; (E) gpx7 mutant; (F) gpx8 mutant and the respective wild types. Data are the mean ±SE (n=10). Asterisks indicate significant differences, P<0.05. 1408 | Passaia et al.

A 6

* * * * * § § § 4 * § root length (cm) 2 # # # # # Primar y

0 Col0Col-0 gpx1 gpx2 gpx3-2 gpx4 gpx6 gpx7 gpx8

B 15

s #

root # # 10

# sible lateral * * vi * * 5 * * § * § Number of 0 Col0Col-0 gpx1 gpx2 gpx3-2 gpx4 gpx6 gpx7 gpx8

C 9 # # )

-1 # (c m

y 6

3

Lateral root densit * * * * * * * 0 Col0Col-0 gpx1 gpx2 gpx3-2 gpx4 gpx6 gpx7 gpx8

Fig. 4. A comparison of the effects of auxin and strigolactone on the root phenotypes in the different gpx mutants and the wild type. The primary root length (A), number of visible lateral roots (B), and lateral root density (C) were measured in seedlings of the different gpx mutant lines and the wild type (Col-0) grown in either the absence (black bars) or presence of auxin (1 μM NAA; grey bars) or strigolactone (2 μM GR24, light grey bars). Data are the mean ±SE. Symbols indicate significant differences, P<0.05 in control conditions (*); after auxin treatment (#); and after strigolactone treatment (§) in the mutants compared with the wild type.

LRD in all the genotypes; gpx6 showed enhanced sensitivity to stages that are designated I (initiation) to VIII (emergence), ABA-induced inhibition of lateral root development (Fig. 5). based on the cell layers formed (Malamy and Benfey 1997). The detailed comparisons of the different stages of primor- Effects of auxin on the developmental stages of the dia development are shown here only for the gpx2 mutants lateral root primordia and Col-0 (Fig. 6). Auxin treatment significantly increased the number of primordia at developmental stages III–VIII in The initiation and early development of lateral root is regu- Col-0 and gpx2 (as indicated by asterisks), relative to con- lated by auxin. The activation of the lateral root primordium trols grown in the absence of auxin (Fig. 6). The number of (LRP), which occurs in a well-defined spatial order, was com- primordia of stages I–II was not altered by auxin treatment pared in the mutant and wild-type lines. For simplicity, the in either genotype. GPX2 deficiency did not alter the auxin- early steps of LRP development were grouped into different dependent development of LRPs (Fig. 6). Glutathione peroxidase functions in Arabidopsis thaliana | 1409

A 6

* # * * * § * § §§ 4 #

root length (cm) § y 2 Primar

0 Col0Col-0 gpx1 gpx2 gpx4 gpx6 gpx7 gpx8 Col3Col-3 gpx3-1

B 9

roots * * 6 * * * sible lateral vi * 3

# Number of

0 Col0Col-0 gpx1 gpx2 gpx4 gpx6 gpx7 gpx8 Col3Col-3 gpx3-1 C 1.5 * *

) * -1 *

(c m * 1 y

*

0.5 # Lateral root densit

0 Col0Col-0 gpx1 gpx2 gpx4 gpx6 gpx7 gpx8 Col3Col-3 gpx3-1

Fig. 5. A comparison of the effects of different concentrations of abscisic acid (ABA) on the root phenotypes in the different gpx genotypes and their respective wild types. The primary root length (A), number of visible lateral roots (B), and lateral root density (C) were measured in seedlings of the different gpx mutant lines and the respective wild types (Col-0 and Col-3) grown in either the absence (black columns) or the presence of 0.3 μM ABA (grey columns) or 0.5 μM ABA (light grey columns). Data are the average ±SE. Symbols indicate significant differences, P<0.05 in control conditions (*); after 0.3 μM (#); and 0.5 μM (§) ABA treatments between the wild types and the mutants.

Effects of strigolactones and auxin on the abundance roots relative to controls (Fig. 7A, B) and increased the levels of AtGPX transcripts of AtGPX2 and AtGPX7 mRNAs in roots (Fig. 7A, B). The addition of auxin led to a significant increase in the abun- The effects of auxin (NAA) and SL (GR24) on the expression dance of all the AtGPX mRNAs, except for those of AtGPX1 of AtGPX genes was characterized by comparing the effects and AtGPX7 in shoots (Fig. 7C). In contrast, AtGPX1 and of these hormones on the abundance of AtGPX transcripts in AtGPX2 transcript levels were lower in the roots of the auxin- the roots and shoots of wild type seedlings. The addition of treated seedlings. However, AtGPX4 and AtGPX7 mRNA GR24 significantly increasedAtGPX1 mRNAs in shoots and levels were higher in the roots after auxin treatment (Fig. 7D). 1410 | Passaia et al.

A B Col-0 60 gpx2 60 * s * s root root 40 40 lateral lateral # * 20 * * 20 * * * Number of Number of

0 0 I-II III-VI V-VI VII-VIII TOTAL I-II III-VI V-VI VII-VIII TOTAL

Fig. 6. The effects of auxin on the stages of LRP development in the gpx2 mutant relative to the wild type. The different stages of lateral root development were measured in seedlings of the gpx2 mutant line (B) and the wild type (A) grown in either the absence (white columns) or the presence (black columns) of auxin (1 μM NAA). Significant differences (P<0.05) are indicated as follows: asterisks (*) denote differences between treatments in the same genotype, while hashes (#) denote differences between the mutant and wild type as a result of auxin treatment.

Fig. 7. The effects of GR24 (2 μM; A, B) and NAA (1 μM; C, D) on the expression of the GPX genes in Col-0 roots and shoots. Black and white bars denote shoots before and after hormone treatment, respectively; whereas dark and light grey bars denote roots before and after hormone treatment, respectively. Data are the mean ±SE of three biological replicates; asterisks indicate significant differences, P<0.05, between treatment and control.

Discussion growth conditions. Moreover, the gpx7 mutants had a greater rosette diameter than the wild type under shorter photoperiod GPXs are considered to play important roles in oxidative sig- growth conditions and more leaves than the wild type under nalling and the metabolism of hydrogen peroxide, as well as long photoperiod growth conditions. These data suggest that the prevention of oxidant-induced PCD. The data presented GPX7 protein has functions in the pathways that constrain here show that knockout mutations in each of the different shoot development. In this regard, it is interesting to note that GPX proteins had no adverse effects on leaf chlorophyll con- the addition of auxin increased the abundance of all of the tents or rosette growth parameters. This finding would suggest GPX mRNAs, except for those of GPX7 and GPX1 in the that other components of the antioxidant system can com- shoots (Fig. 7). AtGPX7 and AtGPX1 are known to fulfil pensate for the loss of GPX function in leaves in controlled important roles in chloroplasts, contributing to the cross-talk Glutathione peroxidase functions in Arabidopsis thaliana | 1411 between high light stress and responses to pathogens (Chang shows impaired root growth with severe defects in root meris- et al., 2009). A part of this cross-talk involves the control of tem maintenance, suggesting that a high plastid GSH:GSSG growth and development under optimal and stress conditions. ratio is important for root development (Yu et al., 2013). The data presented here suggest a role for AtGPX7 but not Conversely, severe GSH deficiency alone (without changes AtGPX1 in the pathways that control shoot growth. in the GSH:GSSG ratio) led to an arrest of the cell cycle

The systematic characterization of the functions of the dif- at G2 in the root meristem with no apparent effect on root ferent GPX isoforms in the control of shoot and root phe- meristem maintenance (Schnaubelt et al., 2014). Less severe notypes reported here demonstrates the importance of these GSH deficiency or depletion of only the cytosolic GSH pool enzymes in the control of root architecture, particularly lat- resulted in decreased LRD (Schnaubelt et al., 2013). A lack of eral root development. There is little information in the lit- reduced thioredoxin caused by mutation in the gene encoding erature concerning the functions of GPX in roots. The roots NADPH-thioredoxin reductase (NTR) C led to decreased of the gpx8 mutants are more sensitive to the oxidative stress root elongation and lower rates of lateral root formation induced by the herbicide methyl viologen (paraquat) than (Kirchsteiger et al., 2012). The data presented here show that wild type roots (Gaber et al., 2012). The data reported here LRD values were higher in the gpx1 and gpx7 mutants than suggest that the GPXs are important in the hormone-medi- in the wild-type controls. Moreover, LRD values were more ated regulation of lateral root development. GSH is involved sensitive to auxin in the gpx1 and gpx7 mutants than in the in the interplay between auxin and SL signalling that con- wild type. These results strongly implicate the plastid-local- trols this process (Marquez-Garcia et al., 2013). For example, ized AtGPX1 and AtGPX7 enzymes in the auxin-dependent LRD was significantly decreased in GSH synthesis mutants control of lateral root production. However, while GPX7 (cad2-1, pad2-1, and rax1-1). Moreover, the application of transcripts were significantly increased in roots as a result GR24 increased the root GSH pool in a manner that was of the application of either auxin or SL, GPX1 transcripts dependent on the MORE AXILLARY GROWTH (MAX) 2 were significantly more abundant following auxin treatment signalling protein (Marquez-Garcia et al., 2013). but significantly decreased in response to SL treatment. Auxin regulates the development of the LRP (De Smet Hence, the hormone-dependent regulation of the AtGPX1 et al., 2003). It triggers the first divisions of lateral root and AtGPX7 enzymes is different, at least at the level of gene founder cells in the pericycle tissue of the primary root expression. The gpx1 and gpx7 root phenotypes resemble (Casimiro et al., 2003) and later it accumulates at the tip of the the gstu17 mutant, which exhibits altered auxin-dependent LRP, where it promotes the degradation of Aux/IAA proteins regulation of lateral root development (Jiang et al., 2010). (Péret et al., 2009). Leaf-derived auxins are also important in However, unlike the gstu17 mutants that show a lower sen- the control of lateral root growth after emergence (Bhalerao sitivity to ABA, gpx1 and gpx7 mutants show responses to et al., 2002; Swarup et al., 2008). The action of auxin is mod- ABA similar to those of the wild type. ulated by SLs, which influence auxin signalling pathways on In contrast to GPX1 transcripts, GPX2 mRNA levels were multiple levels (Agusti et al., 2011). SLs inhibit lateral root decreased in roots following the application of auxin but initiation (Kapulnik et al., 2011) by blocking auxin transport increased following the application of SL. However, the roots (Bennett et al., 2006; Crawford et al., 2010; Balla et al., 2011). of the gpx2 mutants showed responses to SLs and to auxin It has previously been shown that NADP-linked thioredoxin very similar to those observed in the wild type, as did mutants and GSH systems are required for auxin transport and signal- deficient in other cytosolic GPX isoforms, such asgpx4 . The ling (Bashandy et al., 2010). Reduced thiols are also required gpx3 mutants had a lower LRD than the wild type but they for other processes that impact on root development. For showed similar responses to the wild type in root architecture example, the gat1 mutant (gfp-arrested trafficking), which to the auxin, SL, and ABA treatments. Like rice plants defi- is defective in thioredoxin m3, accumulates high levels of cient in GPX3, which showed an altered root structure with callose leading to the occlusion of the plasmodesmata in shorter roots than the wild type (Passaia et al., 2013), the gpx3 the root meristem, resulting in arrested root development mutants studied here had shorter primary roots and fewer lat- (Benitez-Alfonso et al., 2009). The root phenotypes associ- eral roots than the wild type. The AtGPX3 protein has been ated with the gpx mutants described here support the concept shown to be important in ABA signalling under drought stress that redox processes involving glutathione and thioredoxins (Miao et al., 2006). ABA suppresses the auxin response and is exert a strong influence on root development. GPXs may be an inhibitor of lateral root development (De Smet et al., 2003). required to mediate GSH and reduced thioredoxin functions The abscisic acid insensitive 4 (abi4) mutant has an increased in roots that impact on lateral root production or growth. number of lateral roots relative to the wild type (Shkolnik- The data presented here demonstrate that GPX1 and GPX7 Inbar and Bar-Zvi, 2010). However, the responses of root are important in the control of root architecture and lateral architecture to ABA may vary between species because leg- root development. Several plastid-localized enzymes that are ume roots were shown to increase LRD in response to ABA important in the control of cellular redox homeostasis have (Liang and Harris, 2005). The data presented here show that also been shown to be required for root development. For roots of the gpx3 mutants had similar responses to the wild example, the miao mutant is defective in the plastid-localized type control to ABA, indicating that AtGPX3 is not required glutathione reductase (GR2) and consequently it has ~50% for ABA-dependent control of root architecture. lower GR activity than the wild type and accumulates glu- The data presented here show that the application of auxin tathione disulphide, GSSG (Yu et al., 2013). The miao mutant increased the abundance of several GPX transcripts. Auxin 1412 | Passaia et al. is known to induce the expression of several glutathione Bennett T, Sieberer T, Willett B, Booker J, Luschnig C, Leyser O. 2006. The Arabidopsis MAX pathway controls shoot branching by S-transferases (GSTs) such as AtGSTF2. AtGSTF2 binds regulating auxin transport. Current Biology 16, 553–563. indole-3-acetic acid (IAA) and the auxin transport inhibi- Bhalerao RP, Eklöf J, Ljung K, Marchant A, Bennett M, Sandberg G. tor 1-N-naphthylphthalamic acid (NPA), and so enhanced 2002. Shoot-derived auxin is essential for early lateral root emergence in expression of GSTF2 might be linked to stress-mediated Arabidopsis seedlings. The Plant Journal 29, 325–332. growth responses (Smith et al., 2003). Mutants lacking Casimiro I, Beeckman T, Graham N, Bhalerao R, Zhang H, Casero another auxin-inducible GST, AtGSTU17, were less sensitive P, Sandberg G, Bennett MJ. 2003. Dissecting Arabidopsis lateral root development. Trends in Plant Science 8, 165–171. to auxin and had lower numbers of lateral roots in the pres- Chang CCC, Slesak I, Jordá L, Sotnikov A, Melzer M, Miszalski ence of auxin (Jiang et al., 2010). The gstu17 mutants were Z, Mullineaux PM, Parker JE, Karpinska B, Karpinski S. 2009. also less sensitive to ABA-mediated inhibition of root devel- Arabidopsis chloroplastic glutathione peroxidases play a role in cross talk opment (Jiang et al., 2010). between photooxidative stress and immune responses. Plant Physiology 150, 670–683. In conclusion, the data presented here demonstrate that the Crawford S, Shinohara N, Sieberer T, Williamson L, George GPX proteins are important in the control of root architec- G, Hepworth J, Müller D, Domagalska MA, Leyser O. 2010. ture and that loss of any of the GPX isoforms exerts an influ- Strigolactones enhance competition between shoot branches by ence on LRD. The plastid isoform AtGPX7 appears to be dampening auxin transport. Development 137, 2905–2913. important in the control of shoot and root development by De Smet I, Signora L, Beeckman T, Inzé D, Foyer CH, Zhang H. 2003. An abscisic acid-sensitive checkpoint in lateral root development of redox-dependent processes. Arabidopsis. The Plant Journal 33, 543–555. Dietz KJ, Horling F, Konig J, Baier M. 2002. The function of the chloroplast 2-cysteine in peroxide detoxification and its Supplementary data regulation. Journal of Experimental Botany 53, 1321–1329. Dixon DP, Hawkins T, Hussey PJ, Edwards R. 2009. Supplementary data are available at JXB online. activities and subcellular localization of members of the Arabidopsis glutathione superfamily. Journal of Experimental Botany 60, Figure S1. The positions of T-DNA insertions (blue 1207–1218. arrows) in the gpx mutants and the positions of the primers Delaunay A, Pflieger D, Barrault MB, Vinh J, Toledano MB. 2002. used to detect the presence of the insert (black arrows). A thiol peroxidase is an H2O2 receptor and redox-transducer in gene Table S1. The relative expression of GPX genes in different activation. Cell 111, 471–481. genotypes. Eshdat Y, Holland D, Faltin Z, Ben-Hayyim G. 1997. Plant glutathione peroxidases. Physiologia Plantarum 100, 234–240. Table S2. Sequences of the primers used for the isolation of Fischer BB, Krieger-Liszkay A, Hideg E, Snyrychova I, T-DNA insertion mutants (PCR; A and B) and for the gene Wiesendanger M, Eggen RIL. 2007. Role of singlet oxygen in expression analysis (qPCR; C). chloroplast to nucleus retrograde signaling in Chlamydomonas reinhardtii. FEBS Letters 581, 5555–5560 Foyer CH, Noctor G. 2009. Redox regulation in photosynthetic organisms: signaling, acclimation and practical implications. Antioxidants Acknowledgements and Redox Signaling 11, 862–905. GP was funded by Coordenação de Aperfeiçoamento de Pessoal de Nível Foyer CH, Noctor G. 2011. Ascorbate and glutathione: the heart of the Superior–CAPES and Conselho Nacional de Desenvolvimento Científico redox hub. Plant Physiology 155, 2–18. e Tecnológico–CNPq. GQ was funded by Marie Curie Individual Gaber A, Ogata T, Maruta T, Yoshimura K, Tamoi M, Shigeoka S. Fellowship: PIEF-GA-2009–252927: ROXNP. The authors thank the 2012. The involvement of Arabidopsis glutathione peroxidase 8 in the Salk Institute, Syngenta, and the University of Wisconsin for the produc- suppression of oxidative damage in the nucleus and cytosol. Plant and Cell tion of the T-DNA insertion lines, and the Ohio State University and Physiology 53, 1596–1606. the University of Nottingham for the distribution of the seeds. We are grateful to Barbara Karpinska and Belen Marquez-Garcia for useful Herbette S, Roeckel-Drevet P, Drevet JR. 2007. Seleno-independent discussions. glutathione peroxidases. More than simple antioxidant scavengers. FEBS Journal 274, 2163–2180. Iqbal A, Yabuta Y, Takeda T, Nakano Y, Shigeoka S. 2006. Hydroperoxide reduction by thioredoxin-specific glutathione References peroxidase isoenzymes of Arabidopsis thaliana. FEBS Journal 273, 5589–5597. Agusti J, Herold S, Schwarz M, et al. 2011. Strigolactone signaling Jiang H-W, Liu M-J, Chen I-C, Huang C-H, Chao L-Y, Hsieh H-L. is required for auxin-dependent stimulation of secondary growth in 2010. A glutathione S-transferase regulated by light and hormones plants. Proceedings of the National Academy of Sciences, USA 108, participates in the modulation of Arabidopsis seedling development. Plant 20242–20247. Physiology 154, 1646–1658. Alonso JM, Stepanova AN, Leisse TJ, et al. 2003. Genome-wide Kapulnik Y, Delaux P-M, Resnick N, et al. 2011. Strigolactones affect insertional mutagenesis of Arabidopsis thaliana. Science 301, 653–657. lateral root formation and root-hair elongation in Arabidopsis. Planta 233, Balla J, Kalousek P, Reinöhl V, Friml J, Procházka S. 2011. 209–216. Competitive canalization of PIN-dependent auxin flow from axillary buds Kirchsteiger K, Ferrández J, Pascual MB, González M, Cejudo controls pea bud outgrowth. The Plant Journal 65, 571–577. FJ. 2012. NADPH thioredoxin reductase C is localized in plastids of Bashandy T, Guilleminot J, Vernoux T, Caparros-Ruiz D, Ljung K, photosynthetic and nonphotosynthetic tissues and is involved in lateral Meyer Y, Reichheld J-P. 2010. Interplay between the NADP-linked root formation in Arabidopsis. The Plant Cell 24, 1534–1548. thioredoxin and glutathione systems in Arabidopsis auxin signaling. The Ledford HK, Chin BL, Niyogi KK. 2007. Acclimation to singlet oxygen Plant Cell 22, 376–391. stress in Chlamydomonas reinhardtii. Eukaryot Cell 6, 919–930. Benitez-Alfonso Y, Cilia M, San Roman A, Thomas C, Maule Leisinger U, Rufenacht K, Fischer B, Pesaro M, Spengler A, Hearn S, Jackson D. 2009. Control of Arabidopsis meristem A, Zehnder AJB, Eggen RIL. 2001. The glutathione peroxidase development by thioredoxin-dependent regulation of intercellular transport. homologous gene from Chlamydomonas reinhardtii is transcriptionally Proceedings of the National Academy of Sciences, USA 106, 3615–3620. up-regulated by singlet oxygen. Plant Molecular Biology 46, 395–408. Glutathione peroxidase functions in Arabidopsis thaliana | 1413

Liang Y, Harris JM. 2005 Response of root branching to abscisic acid regulated by abiotic stresses through diverse signaling pathways. The is correlated with nodule formation both in legumes and nonlegumes. Plant Journal 36, 602–615. American Journal of Botany 92, 1675–1683. Schmittgen TD, Livak KJ. 2008. Analyzing real-time PCR data by the Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression comparative CT method. Nature Protocols 3, 1101–1108. data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Schnaubelt D, Queval G, Dong Y, Diaz-Vivancos P, Makgopa ME, Methods 25, 402–408. Howell G, De Simone A, Bai J, Hannah MA, Foyer CH. 2014. Low Malamy JE, Benfey PN. 1997. Organization and cell differentiation in glutathione regulates gene expression and the redox potentials of the lateral roots of Arabidopsis thaliana. Development 124, 33–44. nucleus and cytosol in Arabidopsis thaliana Plant, Cell and Environment Mannervik B. 1985. Glutathione peroxidases. Methods in Enzymology (in press). 113, 490–495. Schnaubelt D, Schulz P, Hannah M A, Yocgo RE, Foyer CH. 2013. Margis R, Dunand C, Teixeira FK, Margis-Pinheiro M. 2008. Perturbations in cellular GSH homeostasis influence shoot and root Glutathione peroxidase family—an evolutionary overview. FEBS Journal development but have no negative effects on abiotic stress tolerance in 275, 3959–3970. Arabidopsis thaliana. Frontiers in Plant Physiology 4, 416. Marquez-Garcia B, Njo M, Beeckman T, Goormachtig S, Foyer CH. Sessions A, Burke E, Presting G, et al. 2002. A high-throughput 2013. A new role for glutathione in the regulation of root architecture linked Arabidopsis reverse genetics system. The Plant Cell 14, 2985–2994. to strigolactones. Plant, Cell and Environment (in press). Shkolnik-Inbar D, Bar-Zvi D. 2010. ABI4 mediates abscisic acid Miao Y, Lv D, Wang P, Wang X-C, Chen J, Miao C, Song C-P. and cytokinin inhibition of lateral root formation by reducing polar auxin 2006. An Arabidopsis glutathione peroxidase functions as both a transport in Arabidopsis. The Plant Cell 22, 3560–3573. redox transducer and a scavenger in abscisic acid and drought stress Smith AP, Nourizadeh SD, Peer WA, Xu J, Bandyopadhyay A, responses. The Plant Cell 18, 2749–2766. Murphy AS, Goldsbrough PB. 2003. Arabidopsis AtGSTF2 is regulated Navrot N, Collin V, Gualberto J, Gelhaye E, Hirasawa M, Rey P, by ethylene and auxin, and encodes a glutathione S-transferase that Knaff DB, Issakidis E, Jacquot J-P, Rouhier N. 2006. Plant glutathione interacts with flavonoids. The Plant Journal 36, 433–442. peroxidases are functional peroxiredoxins distributed in several subcellular Swarup K, Benková E, Swarup R, et al. 2008. The auxin influx carrier compartments and regulated during biotic and abiotic stresses. Plant LAX3 promotes lateral root emergence. Nature Cell Biology 10, 946–954. Physiology 142, 1364–1379. Toppo S, Vanin S, Bosello V, Tosatto SCE. 2008. Evolutionary and Passaia G, Spagnolo Fonini L, Caverzan A, Jardim Messeder D, structural insights into the multifaceted glutathione peroxidase (Gpx) Christoff AP, Gaeta ML, de Araujo Mariath JE, Margis R, Margis- superfamily. Antioxidants and Redox Signaling 10, 1501–1514. Pinheiro M. 2013. The mitochondrial glutathione peroxidase GPX3 is Woody ST, Austin-Phillips S, Amasino RM, Krysan PJ. 2007. The essential for H2O2 homeostasis and root and shoot development in rice. WiscDsLox T-DNA collection: an arabidopsis community resource Plant Science 208, 93–101. generated by using an improved high-throughput T-DNA sequencing Péret B, De Rybel B, Casimiro I, Benková E, Swarup R, Laplaze L, pipeline. Journal of Plant Research 120, 157–165. Beeckman T, Bennett MJ. 2009. Arabidopsis lateral root development: Wu M, Huang J, Xu S, Ling T, Xie Y, Shen W. 2010. Haem oxygenase an emerging story. Trends in Plant Science 14, 399–408. delays programmed cell death in wheat aleurone layers by modulation Ramakers C, Ruijter JM, Deprez RHL, Moorman AF. 2003. of hydrogen peroxide metabolism. Journal of Experimental Botany 62, Assumption-free analysis of quantitative real-time polymerase chain 235–248. reaction (PCR) data. Neuroscience Letters 339, 62–66. Yu X, Pasternak T, Eiblmeier M, et al. 2013. Plastid-localized Rodriguez Milla MA, Maurer A, Rodriguez Huete A, Gustafson JP. glutathione reductase 2-regulated glutathione redox status is essential for 2003. Glutathione peroxidase genes in Arabidopsis are ubiquitous and Arabidopsis root apical meristem maintenance. The Plant Cell (in press).