enona, GPr2 Vilnviro

Article history:Received 3 August 2012Received in revised form 8 November 2012Available online 22 December 2012

Keywords:

root exudates, are, in return, able to colonize plant root systemsand contribute to enhance plant growth through a variety of mech-anisms including direct effects on nutrient uptake (e.g. N2-xation,P-mobilization, iron-chelation), and on root growth through the

2009; Kumar et al., 2011; Lugtenberg and Kamilova, 2009;Richardson et al., 2009; Saharan and Nehra, 2011). They can stim-ulate the growth of several important crops and grasses both ingreenhouse and in eld trials under various soils and climatic con-ditions (Bashan et al., 2004; Shaharoona et al., 2007; Steenhoudtand Vanderleyden, 2000; Veresoglou and Menexes, 2010). How-ever, discrepancies were observed in the outcome of PGPR inocula-tions because plant growth promotion depends on many edaphic,ecological and biotic factors that are quite difcult to manage alto-gether. Among PGPR, Azospirillum is considered as one of the mostimportant rhizobacterial genus for improvement of plant growth

Corresponding author. Address: CNRS, UMR 5557, Ecologie Microbienne,Universit Lyon 1, 43 bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France.Tel.: +33 4 72 44 58 89; fax: +33 4 72 43 12 23.

E-mail address: claire.prigent-combaret@univ-lyon1.fr (C. Prigent-Combaret).

Phytochemistry 87 (2013) 6577

Contents lists available at

Phytoche

lse1 Present address: CIRAD, UMR LSTM, F-34398 Montpellier, France.1. Introduction

Plant development and growth are strongly inuenced by bioticand abiotic factors encountered by roots within soils. Certain soilmicrobial populations can benet the plant by improving itsgrowth, development and health. This is typically the case of plantgrowth-promoting rhizobacteria (PGPR), establishing associativesymbiotic interactions with their host plant (Barea et al., 1998;Bloemberg and Lugtenberg, 2001; Raaijmakers et al., 2009;Richardson et al., 2009). PGPR, whose growth are stimulated by

production of phytohormones (auxins, giberrellins and cytokinins)(Richardson et al., 2009; Shaharoona et al., 2007; Yang et al., 2009).They can also protect plant against root and/or foliar pathogens(Raaijmakers et al., 2009) through antibiotic production, competi-tion for ecological niches, and induction of systemic resistance(Antoun and Prvost, 2005; Lugtenberg and Kamilova, 2009;Rezzonico et al., 2005).

PGPR belong to a wide range of genera like Acetobacter,Azospirillum, Bacillus, Burkholderia, Herbaspirillum, Paenibacillus,Phyllobacterium, Pseudomonas or even rhizobia (Desbrosses et al.,AzospirillumPGPROryza sativaRoot colonizationPhenolic metabolitesSecondary metabolite proling0031-9422/$ - see front matter 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.phytochem.2012.11.009Azospirillum is a plant growth-promoting rhizobacterium (PGPR) able to enhance growth and yield ofcereals such as rice, maize and wheat. The growth-promoting ability of some Azospirillum strains appearsto be highly specic to certain plant species and cultivars. In order to ascertain the specicity of the asso-ciative symbiosis between rice and Azospirillum, the physiological response of two rice cultivars, Nippon-bare and Cigalon, inoculated with two rice-associated Azospirillum was analyzed at two levels: plantgrowth response and plant secondary metabolic response. Each strain of Azospirillum (Azospirillum lipofe-rum 4B isolated from Cigalon and Azospirillum sp. B510 isolated from Nipponbare) preferentiallyincreased growth of the cultivar from which it was isolated. This specic effect is not related to a defectin colonization of host cultivar as each strain colonizes effectively both rice cultivars, either at the rhizo-plane (for 4B and B510) and inside the roots (for B510). The metabolic proling approach showed that, inresponse to PGPR inoculation, proles of rice secondary metabolites were modied, with phenolic com-pounds such as avonoids and hydroxycinnamic derivatives being the main metabolites affected. More-over, plant metabolic changes differed according to Azospirillum strain cultivar combinations; indeed,4B induced major secondary metabolic prole modications only on Cigalon roots, while B510, probablydue to its endophytic feature, induced metabolic variations on shoots and roots of both cultivars, trigger-ing a systemic response. Plant secondary metabolite proling thereby evidences the specic interactionbetween an Azospirillum strain and its original host cultivar.

2012 Elsevier Ltd. All rights reserved.a r t i c l e i n f o a b s t r a c tPlant secondary metabolite proling evidin the AzospirillumOryza sativa associati

Amel Chamama, Herv Sanguin a,1, Floriant BellvertFlorence Wisniewski-Dy a, Cdric Bertrand b, ClaireaCNRS, UMR 5557, Ecologie Microbienne, Universit Lyon 1, Universit de Lyon, F-6962bUniversit de Perpignan Via Domitia, Laboratoire de Chimie des Biomolcules et de lE

journal homepage: www.ell rights reserved.ces strain-dependent effect

uillaume Meiffren a, Gilles Comte a,igent-Combaret a,leurbanne, Francennement, EA 4215, F-66860, Perpignan, France

SciVerse ScienceDirect

mistry

vier .com/locate /phytochem

zones, such as the apical root zone (Fig. 1AD), the root hair zone(Fig. 1EH), and sites of lateral root emergence (Fig. 1IL), a coloni-

chemand crop yield worldwide, reaching commercialization in severalcountries like Mexico (Bashan et al., 2004; Jacoud et al., 1998;Okon and Labandera-Gonzalez, 1994).

Plant-benecial effects of Azospirillum strains result mostly inmorphological and physiological changes of the root system (Khalidet al., 2004; Richardson et al., 2009). They are able to increase rootproliferation and elongation, which in turn leads to enhancethe capacity of plants to access to essential nutrients and water(Richardson et al., 2009). Indeed, they stimulate the growth of theirhost plant by producing phytohormones, especially indole-3-aceticacid, and by deaminating 1-aminocyclopropane-1-carboxylate(ACC), the precursor of ethylene, a gaseous plant hormone inhibit-ing root growth (Baldani et al., 1979; Prigent-Combaret et al.,2008; Steenhoudt and Vanderleyden, 2000; Tarrand et al., 1978).

Azospirillum mostly colonized the surface of roots but somestrains, such as Azospirillum brasilense Sp245 and Azospirillum sp.B510, can both colonize the root surface and be recovered fromthe interior of surface-sterilized plant roots or above-ground partsand are thereby dened as endophytes (Elbeltagy et al., 2001;Rothballer et al., 2003; Kaneko et al., 2010; Yasuda et al., 2009).It appears that the extent of internal colonization depends bothon the strain, and the host plant (Elmerich, 2007). Besides theirplant growth promotion effect, endophytes are able to increasethe resistance of the host plant against pathogens and parasites,through induction of disease resistance mechanisms independentof salicylic acid-signaling (Yasuda et al., 2009). Azospirillum endo-phytes have been proposed to be more efcient in promotinggrowth and health when inoculated onto their corresponding hostplant, but no experiment has been designed to address thisquestion.

Despite growing efforts to study Azospirillum plant benecialtraits and plant root colonization patterns, the growth responseof crops to Azospirillum is not completely predictable. On oneside, differential varietal response has been reported for severalcrops, notably for wheat, sorghum, corn and rice, in response toAzospirillum inoculation (Charyulu et al., 1985; Sasaki et al.,2010; Saubidet and Barneix, 1998; Vargas et al., 2012; Veresog-lou and Menexes, 2010). On the other side, differences in yieldwere reported after the inoculation of a single rice or wheat cul-tivar by several Azospirillum lipoferum or A. brasilense strains(Garca de Salamone et al., 2010; Saubidet and Barneix, 1998).These observations have so far not been strengthened by molec-ular studies.

In order to decipher the molecular mechanisms governing thespecicity of the associative symbiosis between Azospirillumstrains and plant, a multiple-step project was developed, of whichthe initial objective is to ascertain the specicity of the physiolog-ical response of host cultivars to Azospirillum strains. Thus, the pur-pose of the present work was to compare the growth andmetabolic responses of two cultivars of Oryza sativa japonica, Cig-alon and Nipponbare, to the inoculation with two Azospirillumstrains isolated from each cultivar, A. lipoferum 4B isolated fromthe surface of Cigalon roots in Camargue (Thomas-Bauzon et al.,1982) and Azospirillum sp. B510, a rice endophyte, isolated fromNipponbare surface-sterilized above-ground parts (Elbeltagyet al., 2001). More accurately, the rice root colonization patternof each strain on both cultivars was rst compared. Second, theoverall plant growth promotion effects and root architecture mod-ications induced by each Azospirillum strain on both rice cultivarswas compared. Third, proling of plant secondary metabolitesfrom both root and shoot parts were compared between uninocu-lated and Azospirillum-inoculated plants. This study thus consti-tutes the rst comparative metabolomic analysis of the responses

66 A. Chamam et al. / Phytoof two rice cultivars to bacterial inoculation by Azospirillum differ-ing in their root-colonization strategies and relationships with thehost plant.zation pattern already documented in Azospirillum spp. (Zhu et al.,2002; Pothier et al., 2007; Couillerot et al., 2011). Non-uniformbacterial colonization along the root can be explained by varyingroot rhizodeposition patterns (Compant et al., 2010; Dennis et al.,2010). Rhizodeposits include mucilages, volatiles, and soluble ly-sates and exudates that are released from damaged and intact cells,respectively. The active root zones such as the root cap and themeristematic root regions are major sites of root exudation to-wards which rhizobacteria are chemoattracted (Dennis et al.,2010; Lugtenberg and Kamilova, 2009). Azospirillum cells werefound as single cells as well as small clumps on the root surface(Fig. 1A, C and L). Both strains did not differ from each other intheir pattern of colonization of the surface of rice roots (namely,the rhizoplane colonization) and moreover they displayed thesame colonization pattern when inoculated onto Cigalon (Fig. 1A,B, E, F, I and J) and onto Nipponbare (Fig. 1C, D, G, H, K and L).

In order to check whether the studied Azospirillum strains wereable to colonize the rhizoplane only or both the rhizoplane and theinterior of roots, serial optical sections were made through Cigalonand Nipponbare root samples, and especially at the point of lateralroot emergence (Fig. 2). Whatever the rice cultivar, 4B cells wereobserved only in the rhizoplane (Fig. 2A); B510 cells were foundboth at the surface and inside the root (Fig. 2B) without inductionof visible disease symptoms, conrming its endophytic coloniza-tion of rice. B510 cells were present 20 lm below the root surface,a depth corresponding to the cortex (Marcel et al., 2010).

Thus, while B510 was able to colonize the interior of rice roots,4B only colonized the rhizoplane, a feature observed on both ricevarieties. However, both strains display an equal weak ability todegrade plant cell wall compounds in vitro (Elbeltagy et al., 2001;Wisniewski-Dy et al., 2011). Thus, their differential ability to col-onize internally the host plant might be related to a differentialregulation of the corresponding gene(s) or to the involvement ofother properties for endophytic colonization, such as twitchingmotility (Bhm et al., 2007). According to microscopic observa-tions, no signicant differences were found in the colonizationability of the endophytic strain B510 onto its host cultivar, Nippon-bare, versus its non-host cultivar, Cigalon. The same observationwas made with the non-endophytic strain 4B.

2.2. Azospirillum strains promote growth and modify root architectureof their host rice cultivar

The effects of A. lipoferum 4B and Azospirillum sp. B510 on plantgrowth and root development of Cigalon and Nipponbare werecompared after 7 days of contact, under gnotobiotic conditions.In presence of 4B, shoot dry weight was signicantly enhancedby 25% and 45% for respectively Nipponbare and Cigalon, com-pared to uninoculated plants (Table 1). Similarly, dry weight ofNipponbare and Cigalon root systems signicantly increased by2. Results and discussion

2.1. A. lipoferum 4B colonizes the outer surface of rice epidermal cellswhereas Azospirillum sp. B510 also colonizes the interior of roots

Root colonization of both rice cultivars was analyzed 7 daysafter inoculation with A. lipoferum 4B or Azospirillum sp. B510.Confocal laser scanning microscopic (CLSM) observations showedthat Enhanced Green Fluorescence Protein (EGFP)-labeled cells ofboth strains preferentially colonized the rhizoplane of active root

istry 87 (2013) 657725% and 34% when inoculated by 4B, respectively (Table 1). Agreater effect of 4B was observed on Cigalon (from which it wasisolated) compared to Nipponbare. While it had no signicant

chemCigalonA. lipoferum 4B

Api

cal r

oot z

on

e

apex x

A

cc

am

bc

x

rc

B

oot h

air

zon

e

rhE

rs

x

rh

F

rs

x

Azospirillum sp. B510

A. Chamam et al. / Phytoeffect on Cigalon, B510 had a signicant effect on shoots and rootsof Nipponbare, from which it was isolated, with respectively 29%and 32% of dry weight increases (Table 1).

Extensive root architecture analyses were also conducted usingthe WinRHIZO software to investigate the effect of Azospirilluminoculation. Compared to uninoculated plants, root systems under-went major architectural changes when inoculated with A. lipofe-rum 4B whereas minor effects were observed with Azospirillumsp. B510 (Table 2). Signicant enhancements were observed with4B inoculation on the total root length, root surface and numberof roots of both cultivars, total root volume of Cigalon, and lengthof the principal root of Nipponbare. A better promoting effect onCigalon was observed compared to Nipponbare, with an increaseof the total root length, root surface and number of roots, by 58%,58%, and 80% respectively, for Cigalon, against 43%, 31%, and 57%for Nipponbare. On the contrary, limited effects were observedwith B510 inoculation on the root system architecture of both cul-tivars, with an increase of the length of the principal root of Nip-ponbare while it decreased for Cigalon and an increase of theroot diameter of Cigalon (Table 2).

Thus, depending on the Azospirillum strain rice cultivar com-bination, the type and number of root parameters signicantly af-fected by bacterial inoculation differed and the phytostimulatoryeffects of A. lipoferum 4B and Azospirillum sp. B510 were higher

RLa

tera

l roo

t em

erge

nce

lr

I x rs

Fig. 1. Confocal laser scanning microscope images of A. lipoferum 4B (A, C, E, G, I, K) anNipponbare (C, D, G, H, K, L) at 7 days after bacterial inoculation. Cells expressing EGtransmitted light. Studied Azospirillum strains preferentially colonized the apical root zonrepresentative of the analysis of at least 10 images per condition. am, apical meristem;emergence; rc, root cap; rh, root hairs.NipponbareA. lipoferum 4B

rc

amxC

cc

ezx

D

rhG

rh

H

rs

x

Azospirillum sp. B510

istry 87 (2013) 6577 67on their respective original rice cultivar. However, the endophyticstrain B510 appeared to have little effect on root morphology onboth rice cultivars compared to the non-endophytic strain in thestudied conditions. It has been often suggested that endophytesare more likely to promote plant growth compared to bacteriaexclusively colonizing the rhizosphere because they achieve closecontact with the plant tissues and escape competition with the rhi-zosphere microbiome (Chanway et al., 2000; Conn et al., 1997;Dbereiner, 1992). In gnotobiotic conditions, despite weak effectson root morphology, B510 induced the highest increase of shootand root biomass on its host cultivar, Nipponbare, conrming eldand greenhouse experiments performed by Isawa et al. (2010) inJapan where B510 was shown to enhance the growth of leavesand the production of seeds.

2.3. Azospirillum inoculation alters rice secondary metabolite proles

To determine rice metabolomic responses to Azospirillum inoc-ulation, at the early stages of interaction, metabolite prolingwas performed on 10-day-old rice plant methanolic extracts (fromshoots and roots) by Reversed-Phase Liquid Chromatography MassSpectrometry (RP-HPLCMS). Chromatographic analyses weredone at 280 nm, a wavelength allowing the detection of a widerange of secondary metabolites and particularly phenolic

lr

K lrL

cc

d Azospirillum sp. B510 (B, D, F, H, J, L) on roots of rice Cigalon (A, B, E, F, I, J) andFP are green and gray backgrounds correspond to the root image formed by thee (AD), the root hair zone (EH) and sites of lateral root emergence (IL). Images arebc, border cells; cc, cell clumps; ez, elongation zone; lr, lateral root; lre, lateral root

20 m

lr

lr A 4B

4B 4B

B

lr

B510

B510

B510

28 m

Fig. 2. Orthogonal optical sections of lateral roots of Nipponbare inoculated withA. lipoferum 4B (A) and Azospirillum sp. B510 (B). Cells expressing EGFP are greenand red backgrounds correspond to the root image stained with propidium iodide.The central views framed in blue show xy focal plans from the z-stacks. The redand green dashed lines represent vertical optical cuts through the z-stacks, whichresult in the side images framed in red and green, respectively. In these side views,the blue dashed lines mark the position where the central view images are locatedwithin the z-stacks. Images were all taken at the emergence of lateral roots. Strain4B was only found at the surface of roots (A) whereas strain B510 was found both atthe surface and inside roots (B). Endophytic colonization of B510 at the emergenceof a lateral root is indicated by a white circle (B). lr, lateral root.

Table 1Effect of Azospirillum inoculations on rice shoot and root dry weights.

Treatments Shoot dry weight(mg/plant)

Root dry weight(mg/plant)

Cv. CigalonControl 1.36 0.08 b 1.31 0.07 bA. lipoferum 4B 1.97 0.14 a 1.76 0.14 aAzospirillum sp. B510 1.14 0.11 b 1.02 0.07 b

Cv. NipponbareControl 4.20 0.77 b 2.12 0.38 bA. lipoferum 4B 5.23 0.21 a 2.66 0.11 aAzospirillum sp. B510 5.43 0.31 a 2.80 0.19 a

Data shown are mean standard error, n = 4. For each data row and rice cultivar,statistical differences between treatments are indicated with lowercase letters(analysis of variance and Fishers least signicant difference test; P < 0.05). Cv.:cultivar.

68 A. Chamam et al. / Phytochemcompounds. Chromatographic proles of shoot methanolic ex-tracts from uninoculated plants were composed of 45 integratedpeaks for Cigalon and 39 for Nipponbare, 37 peaks being in com-mon between the two cultivars. In the same way, chromatographicproles of root methanolic extracts were composed of 74 peaks forCigalon and 70 for Nipponbare, with 63 common peaks.

After 7 days of contact with Azospirillum strains, secondary met-abolic proles of rice shoot and root extracts underwent variations,mostly in the relative intensity of peaks rather than in the appear-ance/disappearance of peaks. Chromatographic proles were fur-ther compared using Principal Component Analysis (PCA) andHierarchical Cluster Analysis (HCA) methods. First, using retentiontimes and peak areas at 280 nm, two data matrices (for shoot (datanot shown) and root extracts) were built in order to carry out HCA(Fig. S1). At a bootstrap value of 100%, rice root extracts grouped intwo principal clusters (i.e. clusters Cig and Nip) corresponding toeach rice cultivar (Fig. S1). This analysis shows the disparity ofplant secondary metabolism between rice cultivars, a feature pre-viously reported on 3000 rice accessions in the course of a study ofallelopathic potential of rice varieties (Kong, 2002).

Second, four matrices were built separately for each cultivarand each plant extract. PCA performed on chromatographic prolesof root and shoot extracts showed signicant differences accordingto biological treatments, with percentages of the rst axes higherthan 9.3% for each of the four PCA (Fig. 3). The data variability ex-plained by the rst and second axes of the PCA represents 25.8%and 46.1% of the total variability for Cigalon shoot and root ex-tracts, respectively, and 15.2% and 34.1% for Nipponbare shootand root extracts, respectively (Fig. 3). Discriminant analysis of sec-ondary metabolite proles from Cigalon root extracts showed sig-nicant differences between all three treatments (uninoculated, A.lipoferum 4B, Azospirillum sp. B510) (Fig. 3C) unlike proles fromNipponbare root extracts (Fig. 3D). HCA on root chromatographicdata conrmed PCA results (Fig. S1). In a lesser extent, signicantdifferences were observed in case of shoot extracts for both culti-vars between the B510 treatment and the two others (Fig. 3Aand B). As previously observed with maize (Walker et al., 2011),the associative symbiosis between Azospirillum PGPR and the hostplant had a signicant impact on plant secondary metabolism. Sec-ondary metabolite proling is thus effective in revealing effects ofinoculation on plant physiology. These results suggest that hostplant recognition of Azospirillum might be strain specic, and trig-gers a specic physiological response by the host plant.

Once inoculated onto Nipponbare, A. lipoferum 4B induced thehighest plant growth promoting effects compared to Azospirillumsp. B510, whereas B510 leads to the highest effect on secondarymetabolism. It is worth noting that major secondary metabolismresponses could arise on host plant, althoughmorphological effectson root architecture were hardly visible. Such a discrepancy be-tween the inoculation effects on plant growth and plant secondarymetabolism has already been reported (Walker et al., 2011). Thissuggests that these bacteria can have effects on the secondarymetabolism of plants without inuencing the physiological func-tions involved in primary metabolism and plant development. Azo-spirillum sp. B510 is the only studied strain able to inducesignicant modications of secondary metabolism on rice shoots,suggesting systemic responses of rice cultivars to B510 inoculation.As this strain has the ability to enter inside the cortex of rice roots(as evidenced in this work) and to colonize the interior of shoots(Elbeltagy et al., 2001), it may severely affect the cellular physiol-ogy of the host plant, triggering plant systemic responses and pro-found changes of secondary metabolism. Accordingly, severalstudies have reported that even if endophytic PGPR promote plant

istry 87 (2013) 6577growth, they nevertheless induce stress and defense responses,causing changes in plant metabolites that lead to a ne controlof bacterial populations inside plant tissues (De Matos Nogueira

rfac

0.32

chemTable 2Effect of Azospirillum inoculations on rice root architecture.

Treatments Principal root length(cm/plant)

Total root length (cm/plant)

Root suplant)

Cv. CigalonControl 7.03 0.42 a 23.29 3.34 b 2.55

A. Chamam et al. / Phytoet al., 2001; Mich et al., 2006; Rosenblueth and Martinez-Romero,2006). Moreover, rice inoculated with the endophyte Azospirillumsp. B510 was shown to display signicant resistance against riceblast disease caused by the fungus Magnaporthe oryzae and riceblight disease caused by the bacterium Xanthomonas oryzae pv.oryzae (Yasuda et al., 2009). Interestingly, pathogenesis-relatedgenes and salicylic acid accumulation were not induced, suggestingthat the priming effect of the endophyte Azospirillum sp. B510 may

A. lipoferum 4B 8.25 0.33 a 36.85 3.02 a 4.04 0.35Azospirillum sp.B510

5.18 0.34 b 22.28 4.41 b 2.93 0.63

Cv. NipponbareControl 9.12 0.43 b 46.35 5.65 b 5.37 0.57A. lipoferum 4B 10.57 0.21 a 66.33 6.04 a 7.08 0.38Azospirillum sp.B510

10.12 0.37 a 54.84 4.21 b 6.21 0.47

Data shown are mean standard error, n = 4. For each root system parameter and rilowercase letters (analysis of variance and Fishers least signicant difference test; P