Monodehydroascorbate reductase 2 and dehydroascorbate reductase 5 are crucial for a mutualistic interaction between Piriformospora indica and Arabidopsis

ARTICLE IN PRESS

Journal of Plant Physiology 166 (2009) 1263—1274

0176-1617/$ - sdoi:10.1016/j.

�CorrespondE-mail addr

www.elsevier.de/jplph

Monodehydroascorbate reductase 2 anddehydroascorbate reductase 5 are crucial for amutualistic interaction between Piriformosporaindica and Arabidopsis

Jyothilakshmi Vadasserya, Swati Tripathib, Ram Prasadb, Ajit Varmab,Ralf Oelmullera,�

aFriedrich-Schiller-Universitat Jena, Institut fur Allgemeine Botanik und Pflanzenphysiologie, Dornburger Str. 159,07743 Jena, GermanybAmity Institute of Herbal and Microbial Studies, Sector 125, Noida 201303, UP, India

Received 21 September 2008; accepted 19 December 2008

KEYWORDSAscorbate–

glutathione cycle;Dehydroascorbatereductase;Drought;Monodehydro-ascorbate reductase;Piriformosporaindica

ee front matter & 2009jplph.2008.12.016

ing author.ess: [email protected]

AbstractAscorbate is a major antioxidant and radical scavenger in plants. Monodehydroas-corbate reductase (MDAR) and dehydroascorbate reductase (DHAR) are two enzymesof the ascorbate–glutathione cycle that maintain ascorbate in its reduced state.MDAR2 (At3g09940) and DHAR5 (At1g19570) expression was upregulated in the rootsand shoots of Arabidopsis seedlings co-cultivated with the root-colonizingendophytic fungus Piriformospora indica, or that were exposed to a cell wallextract or a culture filtrate from the fungus. Growth and seed production were notpromoted by Piriformospora indica in mdar2 (SALK_0776335C) and dhar5(SALK_029966C) T-DNA insertion lines, while colonized wild-type plants were largerand produced more seeds compared to the uncolonized controls. After 3 weeks ofdrought stress, growth and seed production were reduced in Piriformospora indica-colonized plants compared to the uncolonized control, and the roots of the drought-stressed insertion lines were colonized more heavily by the fungus than werewild-type plants. Upregulation of the message for the antimicrobial PDF1.2 proteinin drought-stressed insertion lines indicated that MDAR2 and DHAR5 are crucial forproducing sufficient ascorbate to maintain the interaction between Piriformosporaindica and Arabidopsis in a mutualistic state.& 2009 Elsevier GmbH. All rights reserved.

Elsevier GmbH. All rights reserved.

e (R. Oelmuller).

ARTICLE IN PRESS

J. Vadassery et al.1264

Introduction

Piriformospora indica is an endophytic fungus ofthe Sebacinaceae family that colonizes the roots ofmany plant species, including Arabidopsis (cf.Verma et al., 1998; Sahay and Varma, 1999; Varmaet al., 1999, 2001; Pham et al., 2004; Peskan-Berghofer et al., 2004; Oelmuller et al., 2005;Shahollari et al., 2005, 2007; Sherameti et al.,2005, 2008a; Waller et al., 2005). Hyphae can bedetected on the root surface, in the outer celllayers of the roots and within the root cells, but donot form arbuscular structures in the plant cells(Peskan-Berghofer et al., 2004). Root colonizationand association of fungal hyphae with roots resultsin the promotion of plant growth and higher seedyield, especially under non-optimal conditions.Piriformospora indica has been isolated from theroots of plants grown in the Indian desert Thar(Verma et al., 1998), suggesting that the fungusmay confer fitness benefits under drought stressconditions. Because the fungus is associated withdifferent plant species, it is likely that theinteraction is based on general rather than plant-specific interaction mechanisms.

Ascorbate is a major redox buffer in plants(Pignocchi and Foyer, 2003), a cofactor of manyenzymes (Smirnoff and Wheeler, 2000), a regulatorof cell division and growth (Kerk and Feldman,1995), and a molecule for signal transduction(Noctor et al., 2000). In many organisms, theascorbate–glutathione cycle plays a major role inthe protection of the organism against reactiveoxygen species because it maintains high levelsof ascorbate in the different cell compartments(cf. Asada, 1997). In this cycle, H2O2 is reduced toH2O by ascorbate peroxidase using ascorbate,which generates monodehydroascorbate. Monode-hydroascorbate is a radical and reduced back toascorbate by monodehydroascorbate reductase(MDAR). If not reduced rapidly, monodehydroascor-bate is disproportionated into ascorbate anddehydroascorbate. Dehydroascorbate will then bereduced to ascorbate by dehydroascorbate reduc-tase (DHAR) using reduced glutathione. Oxidizedglutathione is, in turn, reduced by glutathionereductase using NADPH. Thus, MDAR and DHAR arethe two enzymes of the ascorbate–glutathionecycle that maintain ascorbate in its reduced state.

Enzymes of the ascorbate–glutathione cycle arefound in the cytosol (Dalton et al., 1993), chlor-oplasts (Hossain and Asada, 1984; Hossain et al.,1984), mitochondria and peroxisomes (Jimenezet al., 1997). The MDAR mRNA level is upregulatedin response to oxidative stress (Yoon et al., 2004).Further, transgenic Arabidopsis plants expressing

the rice DHAR gene are resistant to salt stress(Ushimaru et al., 2006). Yoshida et al. (2006) haveshown that the cytosolic DHAR is important forozone tolerance in Arabidopsis.

We demonstrate here that ascorbate, the MDAR2(At3g09940) and the DHAR5 (At1g19570) mRNAlevels are upregulated in Arabidopsis roots colo-nized by the beneficial endophytic fungus Pirifor-mospora indica. Insertional inactivation of the twogenes showed that they are crucial for maintainingthe interaction between Piriformospora indica andArabidopsis in a mutualistic state, and underdrought stress in particular.

Materials and methods

Growth conditions for Piriformospora indicaand Arabidopsis

Piriformospora indica was cultured as describedpreviously (Verma et al., 1998; Peskan-Berghoferet al., 2004) in Petri dishes on a modified Kaefer’smedium (KM; NaNO3, 7.0mM; KCl, 7.0mM; MgSO4,2.1mM; KH2PO4, 9.2mM; ZnSO4, 0.77mM; H3BO4,0.18mM; MnSO4, 0.02mM; CoCl2, 0.007mM; CuSO4,0.0065mM; FeSO4, 0.02mM; EDTA, 0.02mM; ammo-nium molybdate, 0.001mM; thiamine, 0.003mM;gylcine, 0.005mM; nicotinic acid, 0.002mM; pyridox-ine, 0.0004mM; glucose, 110mM; peptone, 2g/l;yeast extract, 1g/l; casein hydrolysate, 1g/l, pH6.5) with 1% (w/v) agar. The plates were inoculatedwith the fungus and kept at room temperature in thedark for 1–2 weeks. Before the co-cultivation experi-ments with Arabidopsis seedlings, Piriformosporaindica was grown in liquid KM media until an OD650of 0.5. 3.5ml of this culture was distributed on top ofa Petri dish containing 20ml of a modified plantnutrient medium (PNM; 5mM KNO3, 2mM MgSO4, 2mMCa(NO3)2, 0.01mM FeSO4, 70mM H3BO3, 14mM MnCl2,0.5mM CuSO4, 1mM ZnSO4, 0.2mM Na2MoO4, 0.01mMCoCl2, 10.5g l

�1 agar, pH 5.6). The plates were thenincubated at 22 1C under continuous illumination(80mmolm�2 s�1) for 72h. During this time, thefungal hyphae began to develop a loan.Arabidopsis seedlings were transferred to the loan(see below).

Arabidopsis wild-type and mutant seeds (ecotypeColumbia) were surface-sterilized and placed onPetri dishes containing MS nutrient medium(Murashige and Skoog, 1962). After cold treat-ment at 4 1C for 48 h, plates were incubatedfor 7 days at 22 1C under continuous illumination(100 mmolm�2 s�1). These seedlings were thentransferred to the fungal loan on PNM medium or

ARTICLE IN PRESS

MDAR2 and DHAR5 in Arabidopsis/P. indica interaction 1265

PNM medium alone as a control. One seedling wasused per Petri dish. The plates were incubated at22 1C under continuous illumination from the side(80 mmolm�2 s�1), until roots or shoots were har-vested or the seedlings were transferred to soil.

For the drought stress experiments, Arabidopsisseedlings were transferred to PNM plates with thefungal loan (or without the loan as control) for 24 h.The lids of the Petri dishes were removed and rootsor shoots were harvested as described in theResults section.

Experiments on soil

24 h after transfer of the seedlings to the fungalloan on PNM medium (or PNM medium without thefungus as control), the seedlings were planted intopots and cultivated in a temperature-controlledgreen house at 2273 1C. For plants grown in thepresence of the fungus, the soil was mixed with0.6% (w/w) air-dried fungal mycelia. The myceliawere obtained from 14-day-old liquid cultures,after centrifugation at 8000 rpm for 10min. Seed-lings were watered every 2nd day. For droughtstress experiments on soil, watering was inter-rupted from the 6th to the 9th week. After 9weeks, the plant material was either harvested forthe analyses (described in the Result section), orfurther cultivated and watered for the determina-tion of the seed yield. Seeds were collected inAracon tubes.

Homozygote mdar2 (At3g09940) and dhar5(At1g19570) insertion lines

Homozygote ko lines were obtained from theNottingham stock center (mdar2, SALK_076335Cand dhar5, SALK_029966C). After propagation onsoil, DNA was isolated from the leaves and theproper insertion of the T-DNA confirmed accordingthe information provided by the stock center. Seedsfrom homozygote lines were collected and used forfurther studies in the next generation. Gene-specific primers were used to confirm the absenceof the mRNAs.

RNA analysis

RNA extraction, reverse transcription and semi-quantitative RT-PCR were described in Shahollari et al.(2007). PCR products were separated on 2.0% agarosegels, stained with ethidium bromide, and visualizedbands were quantified with the ImageMaster Videosystem (Amersham Biosciences). The following primerpairs were used: MDAR2 (At3g09940): CCTCTCCATCT-

GAGTTGGTGC and GCGGCGGTTTCTTAGGGC; DHAR5(At1g19570): CAGTCACCCACTTGTCGTCG and GTCC-GTTCAGCCAACGGGC; PDF1.2 (At5g44420): ATGGT-CAGGGGTTTGCGGAAA and ATGGTCAGGGGTTTGCG-GAAA. Piriformospora indica was monitored with aprimer pair for the translation elongation factor 1(Pitef1; Butehorn et al., 2000): ACCGTCTTGGG-GTTGTATCC and TCGTCGCTGTCAACAAGATG. For theanalysis of root colonization, colonized (and control)roots were removed from the agar plate, rinsed 12times with an excess of sterile water (50ml each) toremove the loosely attached fungal hyphae, andfrozen in liquid nitrogen.

Real-time quantitative PCR

All quantitative data were based on real-timequantitative RT-PCR, performed with an iCycler iQreal-time PCR detection system and the iCyclersoftware version 2.2 (Bio-Rad, Germany). Total RNAwas isolated from three independent replicates, orthe number of replicates indicated in the Resultsection. SE values refer to the replicates. For theamplification of the RT-PCR products, iQ SYBRGreen Supermix was used according to the manu-facturer’s protocol in a final volume of 25 ml. TheiCycler was programmed to 95 1C 2min, 40� (95 1C30 s, 55 1C 40 s, 72 1C 45 s), 72 1C 10min, followedby a melting curve program of 55–95 1C in increas-ing steps of 0.5 1C. All reactions were performed intriplicate. The mRNA levels for each cDNA probewere normalized with respect to the plant actinmRNA level. The mRNA levels of the genes in thefigures are expressed relative to the mRNA levelsfound in control seedlings.

Isolation of a cell wall extract fromPiriformospora indica

The cell wall extract was prepared using theprotocol of Anderson-Prouty and Albersheim (1975)with modifications. Mycelia from 14-day-old liquidcultures were homogenized using a mortar andpestle in 5ml water per gram of mycelia. Thehomogenate was filtered using a coarse sinteredglass funnel. The residue was washed three timeswith water, once with chloroform/methanol (1:1),and finally in acetone. This preparation was air-dried for 2 h, after which the pellet was dissolved in100ml water and autoclaved for 20min at 121 1C.Autoclaving released the active compound(s). Thesuspension was filter-sterilized using a 0.22 mMfilter to remove undissolved substances and fungalspores. The filtrate was then concentrated to halfand used for the assay. An aliquot of the preparation

ARTICLE IN PRESS

80

100

120

140

160

180

200

220as

corb

ate

[(+P

.i./-

P.i.

) x

100]

95

100

te +

x

100

[h with fungus] 126543210

J. Vadassery et al.1266

(50ml) was added to the roots immediately aftertransfer of the seedlings to the PNM medium.Control seedlings were treated with water.

Preparation of a culture filtrate fromPiriformospora indica and application toArabidopsis roots

The culture filtrate was obtained from a 14-day-old liquid culture (1 l). The fungal material wasspun down for 50min at 40,000g and the super-natant was filter-sterilized to remove all fungalhyphae and spores. For the preparation of the solidco-cultivation medium, 25% of the water wasreplaced by the culture filtrate. An equal amountof fresh medium was used as control.

Ascorbate and dehydroascorbatemeasurements

Roots (0.2 g) were homogenized with a mortarand pestle in 0.4ml of ice-cold 5% (w/v) metapho-sphoric acid. The homogenate was centrifuged at40,000g for 20min at 4 1C, and the supernatant wascollected for analysis. The ascorbate and dehy-droascorbate contents were determined as de-scribed by de Pinto et al. (1999).

Data analysis

Data points represent the mean of three replica-tions (if no otherwise stated). Data were analyzedusing Student’s t-test at 95% confidence limit.

65

70

75

80

85

90

[asc

orba

te/a

scor

abde

hydr

oasc

orba

te]

[h with fungus] 126543210

Figure 1. Ascorbate content (A) and its reduction stateexpressed as ascorbate/ascorbate+dehydroascorbate ra-tios (B) in Arabidopsis roots (closed symbols) and shoots(open symbols). Panel (A) shows the ratios of the valuesfrom seedlings which were transferred to a mycelial loanat t ¼ 0, divided by the values obtained from seedlings,which were grown on the medium without the fungus forthe same time. The ascorbate level at t ¼ 0 in roots(shoots) corresponds to 0.17870.018 (0.20270.019) mMascorbate/g fresh weight. (B) The ascorbate/ascorbate+dehydroascorbate ratios did not change significantly forseedlings grown on medium without the fungus during the12 h (7479). For the determination of the data requiredfor the values given in the y-axis, a detailed protocol isdescribed in de Pinto et al. (1999). Results are based onsix independent measurements. Bars represent SEs.

Results

Piriformospora indica stimulated theaccumulation of ascorbate in roots andshoots of Arabidopsis

SeedlingsTransfer of Arabidopsis seedlings to PNM plates

with a mycelial loan of Piriformospora indicaresulted in a �1.5-fold increase in the totalascorbate level within 1 h in roots and withinapproximately 5 h in shoots (Figure 1A). Further,while approximately 74% of the total ascorbate/dehydroascorbate pool was reduced in the controls,the level increased to 93% in the presence of thefungus (Figure 1B). Again, the effect was faster inthe roots than in the shoots. This indicates that theascorbate level and its redox state is influenced byPiriformospora indica in Arabidopsis, similar toprevious reports for barley (Waller et al., 2005;

Baltruschat et al., 2008), and that the informationis transmitted from the roots to the shoots.

Piriformospora indica and Piriformosporaindica-derived factors stimulated MDAR2 andDHAR5 transcript accumulation

Analysis of microarray data (Shahollari et al.,2007; Vadassery et al., 2009) revealed that fiveMDAR and four DHAR genes were expressed abovethe background level in Arabidopsis roots. How-ever, only the messenger mRNA levels for twoisoforms, MDAR2 (At3g09940) and DHAR5(At1g19570), were upregulated in colonized roots(Table 1). The mRNA levels for the only geneencoding glutathione reductase (At3g24170) andfor eight genes encoding ascorbate peroxidases(At4g32320, At4g08390, At4g35970, At4g3500,At3g09640, At1g07890, At1g77490, At4g09010)

ARTICLE IN PRESS

50100150200250300350400450500

0 1 2 3 4 5 6 [h with fungus] 12

MD

AR

2 m

RN

A

(fol

d in

duct

ion

x 10

0)

50100150200250300350400

DH

AR

5 m

RN

A

(fol

d in

duct

ion

x 10

0)

MDAR2 and DHAR5 in Arabidopsis/P. indica interaction 1267

were not significantly upregulated in response tothe fungus (data not shown, based on threeindependent microarray analyses). We thereforeinvestigated MDAR2 and DHAR5 in greater detail.

The response of MDAR2 and DHAR5 to Pirifor-mospora indica was confirmed by quantitative real-time PCR. 1 h after the transfer of the seedlings tothe fungal mycelium, the MDAR2 and DHAR5 mRNAlevels in the roots showed an approximate 3-foldincrease (Figure 2A and B). Stimulation was alsoobserved in leaves, except that the response wasslower (Figure 2A and B). In addition, the MDAR2and DHAR5 mRNA levels were upregulated if a cellwall extract or a culture filtrate was applied to theseedlings instead of fungal hyphae (Figure 2C). Thissuggests that MDAR2 and DHAR5 are targets ofsignals from the fungus.

0 1 2 3 4 5 6 [h with fungus] 12

0100200300400500

CWE CF CWE root shoot

mR

NA

(f

old

indu

ctio

n x

100)

Figure 2. The induction of the MDAR2 and DHAR5 mRNAlevels in the roots (closed symbols) and shoots (opensymbols) of Arabidopsis seedlings after transfer toPiriformospora indica (A and B), the application of a cellwall extract (CWE) or a culture filtrate (CF) from thefungus (C). Arabidopsis seedlings were transferred to afungal loan at t ¼ 0 (or to the medium without the

Regulation of MDAR2 and DHAR5 underdrought stress in the presence or absence ofPiriformospora indica

Expression of MDAR2 and DHAR5 was upregulatedunder drought stress in leaves (Genevestigator;http://www.genevestigator.ethz.ch). Exposure ofArabidopsis seedlings to mild drought stress for 6 h(by removal of the lids from the Petri dish, cf.Sherameti et al., 2008b) also caused significantupregulation of the MDAR2 and DHAR5 mRNA levelsin roots (increase relative to the control: DHAR3,2.8270.28-fold; DHAR5, 2.5470.49-fold; n ¼ 6).No regulation was observed for the other members

Table 1. Analyses of microarray data for the expressionof MDAR and DHAR genes in Arabidopsis roots, whichwere either mock-treated or co-cultivated with Pirifor-mospora indica for 3 h.

Gene mock-treated + P. indica

At3g09940, MDAR2 659733 14437113At5g03630, MDAR5 711753 759744At1g63940, MDAR1 681759 598772At3g52880, MDAR4 287766 323757At3g27820, MDAR3 221724 196725At1g19570, DHAR5 12707143 19877159At1g75270, DHAR3 187717 201711At5g16710, DHAR1 100726 97712At5g36270, DHAR2 1573 1373

The numbers refer to the signals from three independentmicroarray hybridizations, errors are given as SEs.Although microarray signals for different genes are difficult tocompare, they might give an idea for the relative expressionlevels of these genes. Gene annotations were taken from Mittleret al. (2004).

fungus) and the MDAR2 (A) and DHAR5 (B) mRNA levelswere followed over a period of 48 h, no significantchanges were observed between 12 and 48 h. (C)Stimulation of the MDAR2 (light grey) and DHAR5 (darkgrey) mRNA levels in the roots and shoots after a 1 hexposure of the seedlings to a CWE from Piriformosporaindica or a CF. Data are [mRNA level from seedlings thatwere transferred to a mycelial loan/mRNA level fromseedlings that were grown on the medium without thefungus for the same time]� 100. The results are based onsix independent biological experiments, and bars repre-sent SEs.

of the MDAR and DHAR gene families under theseconditions in the roots (data not shown). BecausePiriformospora indica is capable of conferringdrought tolerance to Arabidopsis leaves (Sherametiet al., 2008b), we also tested whether the fungushad an influence on the accumulation of MDAR2and DHAR5 mRNA levels in drought-stressedleaves. MDAR2 transcript accumulation was faster

ARTICLE IN PRESS

80130180230280330

1 2 3 4 5 6 7

[h]

MD

AR

2 m

RN

A

leve

l (%

)

Figure 3. Accumulation of MDAR2 mRNA in the leaves ofArabidopsis seedlings, which were either co-cultivatedwith Piriformospora indica (black squares) or mock-

J. Vadassery et al.1268

in drought-exposed leaves from seedlings culti-vated with Piriformospora indica relative to thecontrol (Figure 3). No significant differences weredetected for DHAR5 transcript accumulation,although the same RNA was used for the quantita-tive real-time RT-PCR analyses (data not shown).Thus, Piriformospora indica induced signals fromthe root controlled MDAR2 transcript accumulationin leaves exposed to drought stress (cf. alsoSherameti et al., 2008b).

treated (open squares) after opening the lids at t ¼ 0.The mRNA level of Piriformospora indica-colonizedseedlings at t ¼ 0 was set as 100% and all other valuesexpressed relative to it. The results are based on sixindependent biological experiments, and bars representSEs.

Growth and seed production were notpromoted by Piriformospora indica in themdar2 and dhar5 lines, and were inhibitedby Piriformospora indica under droughtstress

As reported previously (cf. Shahollari et al.,2007), growth of the wild-type in the presence ofPiriformospora indica resulted in an increase in thefresh weight after 6 weeks on soil. In contrast, theMDAR2 and DHAR5 ko lines (Figure 4A) did notrespond significantly to Piriformospora indica, andthe fresh weights were comparable to or evenlower than (although not significantly) those of theuncolonized controls (Figure 4B). Furthermore,flower development was retarded by approximately2 weeks for the two mutants. The lack of a growth-promoting effect of the mutants to the fungus wasnot caused by developmental differences relativeto the wild-type, since older mutants with the same

Figure 4. The effect of DHAR5 (At1g19570) and MDAR2 (At3gand seed yield. (A) The ko lines did not express the inactivatroots of wild-type (WT) and dhar5 seedlings and of MDAR2 (At3of the PCR products given in the figure correspond to the cproducts, which are 342 bp for DHAR5 and 293 bp for MDA6 weeks after transfer to soil, which were either grown iPiriformospora indica. WT, wild-type; mdar2 and dhar5, ko lwas taken as 100% and the other values expressed relativeexperiments with 80 plants each. Bars represent SEs. (C) Seeplants, which were either grown in the absence (white bars) oyield of the WT plants (without fungus) was taken as 100% abased on three independent biological experiments with 80 pldrought stress on the growth and fresh weight of Arabidopsgrown in the absence (�) or presence (+) of Piriformosporaexposure to drought for 3 weeks. White (black) bars, �(+) Piritype plants grown without the fungus was taken as 100% ( ¼ 2.it. Based on three independent biological experiments withwhich were grown on soil for 6 weeks, then exposed to drougand watered again for 2 weeks for testing their recovery. (dhar5) plants, which were either grown in the absence (whitThe seedlings were exposed to drought stress between the 6thwatering until harvesting of the seeds. The seed yield of theother values expressed relative to it. The results are based oneach. Bars represent SEs.

size as wild-type plants also did not respond to thefungus after 6 weeks on soil (data not shown).Likewise, younger wild-type plants with the same sizeas the mutants, after 6 weeks on soil, responded tothe fungus (data not shown). Further, the seed yieldwas not stimulated by Piriformospora indica for thetwo ko lines, while a significant increase was observedfor the wild-type (Figure 4C). This indicates that bothenzymes are crucial for the fitness of the plants undergreenhouse conditions, and that their absence cannotbe fully replaced by other members of the genefamilies.

A stronger effect of the fungus was observed ongrowth when 6-week-old wild-type or ko plantswere exposed to drought for 3 weeks before further

09940) on growth of Arabidopsis on soil, on drought stressed genes. RT-PCR products for DHAR5 (At1g19570) in theg09940) in the roots of WTandmdar2 seedlings. The sizesDNA and differ from the intron-containing genomic PCRR2 (not shown). (B) Fresh weight of Arabidopsis plants,n the absence (open bars) or presence (black bars) ofines. The fresh weight of the WT plants (without fungus)to it. Results are based on three independent biologicald yield of wild-type (WT) and mutant (mdar2 and dhar5)r presence (black bars) of Piriformospora indica. The seednd the other values expressed relative to it. Results areants each. Bars represent SEs. (D) The effect of 3 weeks ofis wild-type (WT) and mutant (mdar2 and dhar5) plantsindica. Difference in fresh weight before and after theformospora indica. The difference in fresh weight of wild-5770.51 g) and the other values are expressed relative to60 plants each, bars represent SEs. (E) Picture of plants,ht for 3 weeks (to determine the data shown in panel D)F) Seed yield of wild-type (WT) and mutant (mdar2 ande bars) or presence (black bars) of Piriformospora indica.and 9th week, as described under (D) and (E), before re-WT plants (without fungus) was taken as 100% and thethree independent biological experiments with 80 plants

ARTICLE IN PRESS

MDAR2 and DHAR5 in Arabidopsis/P. indica interaction 1269

irrigation. While Piriformospora indica promotedgrowth of wild-type plants, it inhibited growth ofthe mutant plants during the 3 weeks of droughtrelative to the uncolonized controls (Figure 4D).Furthermore, colonized mutants barely recoveredfrom the drought stress within the following2 weeks, while the colonized wild-type plantsrecovered well (Figure 4E). Finally, the seed yieldof the drought-exposed colonized mutants waslower than that of the uncolonized controls, whilethe opposite was observed for the wild-type (Figure4F). Taken together, the data suggest that MDAR2and DHAR5 are crucial for Piriformospora indica-

0

50

100

150

200

WT mdar2

fres

h w

eigh

t (%

)

0

50

100

150

200

WT mdar2

seed

yie

ld (

%)

0

50

100

150

200

250

WT mdar2

incr

ease

in f

resh

wei

ght

duri

ng d

roug

ht s

tres

s (%

)

0

50

100

150

200

WT mdar2 d

seed

yie

ld a

fter

dr

ough

t st

ress

(%

)

mediated growth promotion and seed production,especially under drought.

MDAR2 and DHAR5 restricted rootcolonization of Piriformospora indica underdrought stress and prevented theupregulation of the message level for theantimicrobial protein PDF1.2

We have previously demonstrated that loss ofbenefits for the plants could be caused by a shiftfrom mutualism to parasitism, a phenomenon that

dhar5

dhar5

dhar5

har5

ARTICLE IN PRESS

0

1

2

3

4

5

WT mdar2 dhar5

rela

tive

Pite

f1/a

ctin

mR

NA

Figure 5. The transcript levels of the fungal translationelongation factor 1 mRNA (cPitef1) in the roots ofcolonized Arabidopsis seedlings were compared to theamount of the plant actin mRNA level. RNA was isolatedfrom the roots of wild-type (wt) and mutant (mdar2 anddhar5) plants after co-cultivation of both organisms (+)on soil for 6 weeks+3 weeks without irrigation. (–) mocktreatment without the fungus. (A) After reverse tran-scription, cPitef1 and actin were amplified by real-timePCR. (B) The actin mRNA-normalized Pitef1 transcriptlevels of the two mutant lines are expressed relative tothe level in wild-type roots, which was taken as 1.0.Results are based on three independent biologicalexperiments. For one experiment, the root RNA wasisolated from 60 individual plants. Bars represent SEs.

J. Vadassery et al.1270

occurred due to an uncontrolled growth of fungalhyphae in the roots (Sherameti et al., 2008a). Totest whether MDAR2 and/or DHAR5 participated inthe control of root colonization under droughtstress conditions, we determined the amount of thefungal translation elongation factor 1 (Pitef1)mRNA relative to the plant actin mRNA in the rootsof colonized and uncolonized mutant and wild-typeplants exposed to drought stress for 3 weeks. Asshown previously, the plant actin mRNA level is areliable reference for the root biomass, while thePifef1 mRNA level can be used as a marker for thefungus (Sherameti et al., 2008a). Thus, the Pitef1/actin mRNA ratio represents the degree of rootcolonization. This ratio was significantly higher inthe mdar2 and dhar5 mutant roots compared to thewild-type 3 weeks after the withdrawal of thewater (Figure 5), indicating that the drought-exposed mutants were less protected againstfungal colonization compared to wild-type plants.Furthermore, the message level of the antifungalprotein PDF1.2 (Penninckx et al., 1996), which isnot expressed in uncolonized and colonized wild-type roots, is upregulated when the interactionshifts from mutualism to parasitism (Sherametiet al., 2008a, and unpublished data). The PDF1.2mRNA level was at the detection limit in colonizedwild-type roots exposed to 3 weeks of drought, butupregulated in colonized mdar2 and dhar5 rootsexposed to drought stress (Figure 6). This demon-strates that MDAR2 and DHAR5 also contribute tothe repression of defense gene expression againstPiriformospora indica under drought stress condi-tions.

0

50

100

150

WT mdar2 dhar5

PD

F1.

2 m

RN

A le

vel (

%)

Figure 6. PDF1.2 mRNA levels in Arabidopsis roots,which were grown on soil for 6 weeks+3 weeks withoutirrigation. Plants were either grown in the absence(white bars) or presence (black bars) of P. indica. WT,wild-type; mdar2 and dhar5, ko lines, respectively. Basedon three independent biological experiments. The high-est mRNA level was taken as 100% and the other valuesexpressed relative to it. Bars represent SEs.

Discussion

Ascorbate is important for antioxidative andmetabolic functions. It is oxidized to dehydroas-corbate under antioxidative conditions, and DHARreduces it to ascorbate. An intermediate during theoxidation process is monodehydroascorbate, whichis reduced to ascorbate by MDAR. Thus, bothenzymes are crucial for ascorbate recycling (cf.Ushimaru et al., 2006). Reduced ascorbate levelsand antioxidant enzyme activities have often beenreported when plants are exposed to stress orenemies (Kuzniak and Sklodowska 2001, 2005;Kiddle et al., 2003). They also play crucial roles inbeneficial plant/microbe interactions. Piriformos-pora indica-colonized barley plants containelevated antioxidative capacity due to the activa-tion of the ascorbate–glutathione cycle, and thefungus enhances the ratio of reduced to oxidized

ARTICLE IN PRESS

MDAR2 and DHAR5 in Arabidopsis/P. indica interaction 1271

ascorbate, as well as the DHAR and glutathionereductase activities under salt stress (Walleret al., 2005; Baltruschat et al., 2008).

Salt-stressed tomato plants grow better whencolonized by arbusuclar mycorrhizal fungi (AMF)(He et al., 2007), presumably because they showhigher superoxide dismutase, ascorbate peroxidaseand peroxidase activities. Citrus plants showedenhanced drought tolerance when colonized byAMF due to an increased concentration of antiox-idant enzymes and non-enzymatic antioxidants(Wu et al., 2006). Higher levels of superoxidedismutase activities were also observed in lettuceroots colonized by Glomus (Ruiz-Lozano et al.,1996). Furthermore, appressorium formation ofGlomus mosseae with tobacco roots is accompaniedby a transient increase in catalase and ascorbateperoxidase activities. High levels of ascorbate andglutathione are important for nodule formation andthese metabolites decline as the nodule ages(Puppo et al., 2005); higher catalase activity hasalso been measured in soybean nodules (Porcelet al., 2003). Thus, several reports support the ideathat antioxidant enzymes might have a protectivefunction in the interaction between mutualisticpartners (Foyer et al., 1997; Alguacil et al., 2003).

Our data define a role of MDAR2 and DHAR5 in themutualistic interaction between Piriformosporaindica and Arabidopsis. Growth and seed produc-tion of mdar2 and dhar5 plants is no longerpromoted by the fungus (Figure 4B and C) and themutualistic interaction shifts towards parasitismwhen the mutants are exposed to drought stress(Figure 4D–F). In contrast to wild-type plants,growth of mutants during drought stress isreduced in the presence of Piriformospora indica(Figure 4D), they barely recover from the droughtstress in the presence of the fungus (Figure 4E) andthey produce fewer seeds (Figure 4F). AlthoughMDAR and DHAR are encoded by multigene familiesin Arabidopsis, MDAR2 and DHAR5 cannot bereplaced by the other members of their family.Interestingly, both genes are upregulated byPiriformospora indica, the application of (a)component(s) released for the fungal cell wallextract or the culture filtrate. Thus, it appears thatthe response is induced by (a) soluble factor(s)released from the fungus.

A comparison of Figures 1 and 2 demonstratesthat the total ascorbate level (Figure 1A) and theamount of reduced ascorbate (Figure 1B) in theroots increase more rapidly after transfer ofArabidopsis seedlings to a fungal loan than themRNA levels of the enzymes (Figure 2). Thissuggests that enzyme activity is upregulated priorto the stimulation of gene expression. Rapid

changes of the activity states of enzymes of theascorbate–glutathione cycle have been reportedfor many systems exposed to diverse stimuli(del Rıo et al., 2003, and references therein; cf.also Baltruschat et al., 2008).

The ascorbate–glutathione cycle enzymes alsofunction as a reactive oxygen species (ROS)-scavenging system (cf. Mittler et al., 2004), andROS generated under stress conditions in particular.ROS production occurs preferentially during earlyphases of beneficial plant/microbe interactions(Salzer et al., 1999; Santos et al., 2001; Ramuet al., 2002; Matamoros et al., 2003; Shaw andLong, 2003; Fester and Hause 2005), and it isgenerally believed that antioxidant enzymes areupregulated with the appearance of ROS during theinfection to maintain a constant redox balance inthe cytoplasm (Arines et al., 1994; Blilou et al.,2000; Lambais et al., 2003). However, ROS pro-duced by the fungus can also be critical formaintaining a mutualistic plant/fungus interaction(Tanaka et al., 2006) and is necessary for rootdevelopment (Foreman et al., 2003). Thus, inacti-vation of either MDAR2 or DHAR5 might disturb thebalance between the ROS levels required for rootdevelopment and a mutualistic interaction withPiriformospora indica on the one hand, and anoxidative burst induced during defense responseson the other hand, in particular when the plants areexposed to stress. Although we attempted tomeasure differences in H2O2 levels under ourexperimental conditions, we never observed asignificant difference relative to the untreatedcontrol (Vadassery et al., 2009).

Antioxidant and antioxidant enzymes repressdefense gene activation. In mycorrhiza, defensegenes are upregulated mainly during early stages ofinteraction (Blilou et al., 2000; Garcıa Garrido andOcampo, 2002), but can also be detected duringarbuscule development (Grunwald et al., 2004). Wenever observed a marked stimulation of defenseresponses by Piriformospora indica in wild-typeroots. However, Figure 6 demonstrates that theantimicrobial protein PDF1.2, which is not upregu-lated in the wild-type by the fungus or by droughtstress, is upregulated in colonized mdar2 and dhar5ko lines exposed to drought stress. It appears thatPDF1.2 is only upregulated in the roots when theinteraction with Piriformospora indica is shiftedfrom mutualism to parasitism (cf. Sherameti et al.,2008a). Thus, MDAR2 and DHAR5 participate in thestabilization of the mutualistic interaction underdrought stress, a hypothesis that is further sup-ported by the observations that the stress-exposedko lines do not grow taller in the presence of thefungus (Figures 4D and E) and produced fewer seeds

ARTICLE IN PRESS

J. Vadassery et al.1272

(Figure 4F). The partial shift from mutualism to aless beneficial interaction might indicate that thecellular environment in the host cell is not optimalfor the invader.

Besides the upregulation of the microbial de-fense gene PDF1.2 (Figure 6), the greater rootcolonization of the mutants by Piriformosporaindica (Figure 5) indicates that the mutualisticinteraction between the two partners is disturbedin the drought-stressed ko lines. This defines a roleof MDAR2 and DHAR5 in protecting the roots againstuncontrolled hyphal growth. Overcolonization ofroots can lead to the activation of defenseresponses (Sherameti et al., 2008a). Upregulationof defense genes in the roots could control, at leastin part, the intraradical fungal growth (Lambais andMehdy, 1995).

This work also demonstrates that antioxidantenzymes are involved in Piriformospora indica-mediated drought tolerance in Arabidopsis seed-lings (Figures 3, 4D–F, 5 and 6), as reportedpreviously for other plant/microbe interactionsystems (cf. Conklin et al., 1996; Ruiz-Lozanoet al., 1996; Badawi et al., 2004a b; Adriano et al.,2005; Huang et al., 2005; Eltayeb et al., 2006).Drought stress is associated with a dramaticdecrease in photosynthesis. This decrease is sub-stantially retarded in the presence of Piriformos-pora indica (Sherameti et al., 2008b), and is knownto influence many metabolic processes, includingrespiration, photosynthetic electron transport andoxidation of glycolate in photorespiration. Sincethe MDAR2 mRNA level was upregulated faster incolonized and drought-stressed leaves compared tothe uncolonized controls, the plants must containan efficient system to transfer the information fromthe roots to the shoots. The availability ofPiriformospora indica responsive genes in theleaves, in combination with the molecular toolsavailable in the Arabidopsis community, will allowthe identification of the signal transducers.

In summary, MDAR2 and DHAR5, two enzymes ofthe ascorbate–glutathione cycle involved in main-taining ascorbate in its reduced state, are crucialfor the mutualistic harmony between Arabidopsisand Piriformospora indica. The roles of the othermembers of these two multigene families and ofthe other enzymes of this cycle remain to beelucidated.

Acknowledgments

This work was supported by the SFB 607, theIMPRS Jena and the DFG.

References

Adriano S, Angelo CT, Bartolomeo D, Cristos X. Influenceof water deficit and rewatering on the components ofthe ascorbate–glutathione cycle in four interspecificPrunus hybrids. Plant Sci 2005;169:403–13.

Alguacil MM, Hernandez JA, Caravaca F, Portillo B, RoldanA. Antioxidant enzyme activities in shoot fromthree mycorrhizal shrub species afforested in adegraded semi-arid soil. Physiol Plant 2003;118:562–70.

Anderson-Prouty AJ, Albersheim P. Host–pathogen inter-actions. VIII. Isolation of a pathogensynthesizedfraction rich in glucan that elicits a defense responsein the pathogens host. Plant Physiol 1975;56:286–91.

Arines J, Quintela M, Vilarino A, Palma JM. Proteinpatterns and superoxide dismutase activity in non-mycorrhizal and arbuscular-mycorrhizal Pisum sati-vum L. plants. Plant Soil 1994;166:37–45.

Asada K. The role of ascorbate peroxidase and mono-dehydroascorbate reductase in H2O2 scavenging inplants. In: Scandalios JG, editor. Oxidative stress andthe molecular bio logy of antioxidant defenses.New York: Cold Spring Harbor Laboratory Press;1997. p. 715–35.

Baltruschat H, Fodor J, Harrach BD, Niemczyk E, Barna B,Gullner G, et al. Salt tolerance of barley induced bythe root endophyte Piriformospora indica is asso-ciated with a strong increase in antioxidants. NewPhytol 2008;180:501–10.

Badawi GH, Kawano N, Yamauchi Y, Shimada E, Sasaki R,Kubo A, et al. Overexpression of ascorbate peroxidasein tobacco chloroplasts enhances the tolerance to saltstress and water deficit. Physiol Plant 2004a;121:131–238.

Badawi GH, Yamauchi Y, Shimada E, Sasaki R, Kawano N,Tanaka K, et al. Enhanced tolerance to salt stress andwater deficit by overexpressing superoxide dismutasein tobacco (Nicotiana tabacum) chloroplast. Plant Sci2004b;166:919–28.

Blilou I, Bueno P, Ocampo JA, Garcıa-Garrido JM.Induction of catalase and ascorbate peroxidase activ-ities in tobacco roots inoculated with the arbuscularmycorrhizal fungus Glomus mosseae. Mycol Res2000;104:722–5.

Butehorn B, Rhody D, Franken P. Isolation and character-ization of Pitef1 encoding the translation elongationfactor EF-1a of the root endophyte Piriformosporaindica. Plant Biol 2000;2:687–92.

Conklin PL, Williams EH, Last RL. Environmental stresssensitivity of an ascorbic aciddeficient Arabidopsismutant. Proc Natl Acad Sci USA 1996;93:9970–4.

Dalton DA, Baird LM, Langeberg L, Taugher CY, Anyan WR,Vance CP, et al. Subcellular localization of oxygendefense enzymes in soybean (Glycine max [L.] Merr.)root nodules. Plant Physiol 1993;102:481–9.

de Pinto MC, Francis D, De Gara L. The redox state of theascorbate-dehydroascorbate pair as a specific sensorof cell division in tobacco BY-2 cells. Protoplasma1999;209:90–7.

ARTICLE IN PRESS

MDAR2 and DHAR5 in Arabidopsis/P. indica interaction 1273

del Rıo LA, Corpas FJ, Sandalio LM, Palma JM, Barroso JB.Plant peroxisomes, reactive oxygen metabolism andnitric oxide. IUBMB Life 2003;55:71–81.

Eltayeb AE, Kawano N, Badawi GH, Kaminaka H, SanekataT, Morishima I, et al. Enhanced tolerance to ozone anddrought stresses in transgenic tobacco overexpressingdehydroascorbate reductase in cytosol. Physiol Plant2006;127:57–65.

Fester T, Hause G. Accumulation of reactive oxygenspecies in arbuscular mycorrhizal roots. Mycorrhiza2005;15:373–9.

Foreman J, Demidchik V, Bothwell JH, Mylona P, MedermaH, Torres MA, et al. Reactive oxygen species producedby NADPH oxidase regulate plant cell growth. Nature2003;422:442–6.

Foyer CH, Lopez-Delgado H, Dat JE, Scott IM. Hydrogenperoxide and glutathione associated mechanisms ofacclamatory stress tolerance and signaling. PhysiolPlant 1997;100:241–54.

Garcıa Garrido JM, Ocampo JA. Regulation of plantdefence response in arbuscular mycorrhizal simbiosis.J Exp Bot 2002;53:1377–86.

Grunwald U, Nyamsuren O, Tamasloukht M, Lapopin L,Becker A, Mann P, et al. Identification of mycorrhiza-regulated genes with arbuscule development-relatedexpression profile. Plant Mol Biol 2004;55:553–66.

He Z, He C, Zhang Z, Zou Z, Wang H. Changes ofantioxidative enzymes and cell membrane osmosis intomato colonized by arbuscular mycorrhizae underNaCl stress. Colloids Surf B Biointerfaces 2007;59:128–33.

Hossain MA, Asada K. Inactivation of ascorbate perox-idase in spinach chloroplast on dark addition ofhydrogen peroxide: its protection by ascorbate. PlantCell Physiol 1984;25:1285–95.

Hossain MA, Nakano Y, Asada K. Monodehydroascorbatereductase in spinach chloroplasts and its participationin regeneration of ascorbate for scavenging hydrogenperoxide. Plant Cell Physiol 1984;25:385–95.

Huang C, He W, Guo J, Chang X, Su P, Zhang L. Increasedsensitivity to salt stress in an ascorbate-deficientArabidopsis mutant. J Exp Bot 2005;56:3041–9.

Jimenez A, Hernandez JA, del Rıo LA, Sevilla F. Evidencefor the presence of the ascorbateglutathione cycle inmitochondria and peroxisomes of pea leaves. PlantPhysiol 1997;114:275–84.

Kerk NM, Feldman LJ. A biochemical model for theinitiation and maintenance of the quiescent centre:implication for organization of root meristems. Devel-opment 1995;121:2825–33.

Kiddle G, Pastori GM, Bernard S, Pignocchi C, Antoniw J,Verrier PJ, et al. Effects of leaf ascorbate content ondefense and photosynthesis gene expression in Arabi-dopsis thaliana. Antioxid Redox Signal 2003;5:23–32.

Kuzniak E, Sklodowska M. Ascorbate, glutathione andrelated enzymes in chloroplasts of tomato leavesinfected by Botrytis cinerea. Plant Sci 2001;160:723–31.

Kuzniak E, Sklodowska M. Fungal pathogen-inducedchanges in the antioxidant systems of leaf peroxi-

somes from infected tomato plants. Planta 2005;222:192–200.

Lambais MR, Mehdy MC. Differential expression ofdefense-related genes in arbuscular mycorrhiza. CanJ Bot 1995;73:S533–40.

Lambais MR, Rıos-Ruiz WF, Andrade RM. Antioxidantresponses in bean (Phaseolus vulgaris) roots colonizedby arbuscular mycorrhizal fungi. New Phytol 2003;160:421–8.

Matamoros MA, Dalton DA, Ramos J, Clemente MR, RubioMC, Becana M. Biochemistry and molecular biology ofantioxidant s in the rhizobia–legume symbiosis. PlantPhysiol 2003;133:499–509.

Mittler R, Vanderauwera S, Gollery M, Van Breusegem F.Reactive oxygen gene network of plants. Trends PlantSci 2004;9:490–8.

Murashige T, Skoog F. A revised medium for rapid growthand bioassays with tobacco tissue cultures. PhysiolPlant 1962;15:473–97.

Noctor G, Veljovic-Jovanivic S, Foyer CH. Peroxideprocessing in photosynthesis: antioxidant couplingand redox signalling. Philos Trans R Soc Lond 2000;B355:1465–75.

Oelmuller R, Peskan-Berghofer T, Shahollari B, TrebickaA, Sherameti I, Varma A. MATH domain proteinsrepresent a novel protein family in Arabidopsisthaliana, and at least one member is modified inroots during the course of a plant-microbe interaction.Physiol Plant 2005;124:152–66.

Penninckx IA, Eggermont K, Terras FR, Thomma BP, DeSamblanx GW, Buchala A, et al. Pathogen-inducedsystemic activation of a plant defensin gene inArabidopsis follows a salicylic acid-independent path-way. Plant Cell 1996;8:2309–23.

Peskan-Berghofer T, Shahollari B, Giang PH, Hehl S,Markert C, Blanke V, et al. Association of Piriformos-pora indica with Arabidopsis thaliana roots representsa novel system to study beneficial plant–microbeinteractions and involves early plant protein modifica-tions in the endoplasmatic reticulum and at theplasma membrane. Physiol Plant 2004;122:465–77.

Pham GH, Kumari R, Singh A, Sachdev M, Prasad R,Kaldorf M, et al. Axenic cultures of Piriformosporaindica. In: Varma A, Abbott L, Werner D, Hampp R,editors. Plant surface microbiology. Germany: Spring-er; 2004. p. 593–616.

Pignocchi C, Foyer CH. Apoplastic ascorbate metabolismand its role in the regulation of cell signalling. CurrOpin Plant Biol 2003;6:379–89.

Porcel R, Barea JM, Ruiz-Lozano JM. Antioxidant activ-ities in mycorrhizal soybean plants under droughtstress and their possible relationship to the process ofnodule senescence. New Phytol 2003;157:135–43.

Puppo A, Groten K, Bastian F, Carzaniga R, Soussi M,Lucas MM, et al. Legume nodule senescence: roles forredox and hormone signaling in the orchestration ofthe natural aging process. New Phytol 2005;165:683–701.

Ramu SK, Peng H-M, Cook DR. Nod factor induction ofreactive oxygen species production is correlated with

ARTICLE IN PRESS

J. Vadassery et al.1274

expression of the early nodulin gene rip1 in Medicagotruncatula. Mol Plant Microb Interact 2002;15:522–8.

Ruiz-Lozano JM, Azcon R, Palma JM. Superoxide dismu-tase activity in arbuscular mycorrhizal Lactuca sativaplants subjected to drought stress. New Phytol1996;134:327–33.

Sahay NS, Varma A. Piriformospora indica: a newbiological hardening tool for micropropagated plants.FEMS Microbiol Lett 1999;181:297–302.

Salzer P, Corbiere H, Boller T. Hydrogen peroxideaccumulation in Medicago truncatula roots colonizedby the arbuscular mycorrhiza-forming fungus Glomusintraradices. Planta 1999;208:319–25.

Santos R, Herouart D, Sigaud S, Touati D, Puppo A.Oxidative burst in alfalfa–Sinorhizobium melilotisymbiotic interaction. Mol Plant Microb Interact2001;14:86–9.

Shahollari B, Varma A, Oelmuller R. Expression of areceptor kinase in Arabidopsis roots is stimulated bythe basidiomycete Piriformospora indica and theprotein accumulates in Triton X-100 insoluble plasmamembrane microdomains. J Plant Physiol 2005;162:945–58.

Shahollari B, Vadassery J, Varma A, Oelmuller R. Aleucine-rich repeat protein is required for growthpromotion and enhanced seed production mediated bythe endophytic fungus Piriformospora indica in Arabi-dopsis thaliana. Plant J 2007;50:1–13.

Shaw SL, Long SR. Nod factor inhibition of reactiveoxygen efflux in a host legume. Plant Physiol 2003;132:2196–204.

Sherameti I, Shahollari B, Venus Y, Altschmied L, VarmaA, Oelmuller R. The endophytic fungus Piriformosporaindica stimulates the expression of nitrate reductaseand the starch-degrading enzyme glucan–water diki-nase in tobacco and Arabidopsis roots through ahomeodomain transcription factor which binds to aconserved motif in their promoters. J Biol Chem2005;280:2641–7.

Sherameti I, Venus Y, Drzewiecki C, Tripathi S, Dan VM,Nitz I, et al. PYK10, a beta-glucosidase located in theendoplasmatic reticulum, is crucial for the beneficialinteraction between Arabidopsis thaliana and theendophytic fungus Piriformospora indica. Plant J2008a;54:428–39.

Sherameti I, Tripathi S, Varma A, Oelmuller R. The root-colonizing endophyte Pirifo rmospora indica confersdrought tolerance in Arabidopsis by stimulating the

expression of drought stress-related genes in leaves.Mol Plant Microbe Interact 2008b;21:799–807.

Smirnoff N, Wheeler GL. Ascorbic acid in plants:biosynthesis and function. Crit Rev Plant Sci 2000;19:267–90.

Tanaka A, Christensen MJ, Takemoto D, Park P, Scott B.Reactive oxygen species play a role in regulating afungus–perennial ryegrass mutualistic interaction.Plant Cell 2006;18:1052–66.

Ushimaru T, Nakagawa T, Fujioka Y, Daicho K, Naito M,Yamauchi Y, et al. Transgenic Arabidopsis plantsexpressing the rice dehydroascorbate reductase geneare resistant to salt stress. J Plant Physiol 2006;163:1179–84.

Vadassery J, Ranf S, Mithofer A, Mazars C, Scheel D, LeeJ, et al. A cell wall extract from the endophytic fungusPiriformospora indica promotes growth of Arabidopsisseedlings and induces intracellular calcium elevationin roots. Plant J 2009 [in press].

Verma SA, Varma A, Rexer K-H, Hassel A, Kost G, SarbhoyA, et al. Piriformospora indica, gen. et sp. nov., a newroot-colonizing fungus. Mycologia 1998;90:898–905.

Varma A, Verma S, Sudha Sahay NS, Butehorn B, FrankenP. Piriformospora indica, a cultivable plant growthpromoting root endophyte. Appl Environ Microbiol1999;65:2741–4.

Varma A, Singh A, Sudha Sahay N, Sharma J, Roy A,Kumari M, et al. Piriformospora indica: a cultivablemycorrhiza-like endosymbiotic fungus. In: Mycota IX.Germany: Springer Series; 2001. p. 123–50.

Waller F, Achatz B, Baltruschat H, Fodor J, Becker K,Fischer M, et al. The endophytic fungus Piriformos-pora indica reprograms barley to salt-stress tolerance,disease resistance, and higher yield. Proc Nat Acad SciUSA 2005;102:13386–91.

Wu QS, Xia RX, Zou YN. Reactive oxygen metabolism inmycorrhizal and non-mycorrhizal citrus (Poncirustrifoliata) seedlings subjected to water stress. J PlantPhysiol 2006;163:1101–10.

Yoon H, Lee H, Lee I, Kim K, Jo J. Molecular cloning ofmonodehydroascorbate reductase gene from Brassicacampestris and analysis of its mRNA level in responseto oxidative stress. Biochem Biophys Acta 2004;1658:181–6.

Yoshida S, Tamaoki M, Shikano T, Nakajima N, Ogawa D,Ioki M, et al. Cytosolic dehydroascorbate reductase isimportant for ozone tolerance in Arabidopsis thaliana.Plant Cell Physiol 2006;47:304–8.