Cp

DEa

b

6c

d

a

ARRA

KEDPIBP

1

g1ptspt1

h0

Applied Soil Ecology 80 (2014) 100–107

Contents lists available at ScienceDirect

Applied Soil Ecology

jo u r n al homep ag e: www.elsev ier .com/ locate /apsoi l

ombined effects of earthworms and IAA-producing rhizobacteria onlant growth and development

iana Edith Castellanos Suareza, Agnès Gigonb, Ruben Puga-Freitasb, Patrick Lavellec,lena Velasquezd, Manuel Blouinb,∗

Facultad de Agronomía, Universidad Nacional de Colombia, Bogotá D.C., ColombiaUniversité Paris-Est Créteil Val de Marne, Institute of Ecology and Environmental Sciences of Paris (UPMC, UPEC, Paris Diderot, CNRS, INR, INRA),1 avenue du Général De Gaulle, 94010 Creteil cedex, FranceCentro Internacional de Agricultura Tropical, AA 6713, Cali, ColombiaUniversidad Nacional de Colombia, Palmira, Colombia

r t i c l e i n f o

rticle history:eceived 23 November 2013eceived in revised form 26 March 2014ccepted 3 April 2014

eywords:arthwormseleterious rhizo-bacteria (DRB)lant growth promoting bacteria (PGPB)ndole acetic acid (IAA)iocontrollant growth and development

a b s t r a c t

Micrococcus luteus is considered a deleterious rhizo-bacteria (DRB) or a plant growth promoting rhi-zobacteria (PGPR) depending on the plant species. This bacteria is known to produce Indole Acetic Acid(IAA), a hormone that can affect plant growth. In this study, we tested the effect of M. luteus on themodel plant Arabidopsis thaliana at three different concentrations (105, 106 and 107 CFU ml−1), to char-acterize its effect on plants according to its concentration, and eventually classify it as DRB or PGPR.These bioassays were conducted in the presence or absence of earthworms (Aporrectodea caliginosa),known to modify A. thaliana growth through the stimulation of IAA-producing bacteria. We wanted totest the ability of earthworms to synergistically enhance potential positive effects of the bacteria throughdispersal or stimulation, or to control the negative effect of the bacteria through grazing or inhibition.This would help determine the potential of using earthworms as a vector or selective filter of bacteriainoculated or present in the field. M. luteus had a negative effect on plant growth proportional to thebacteria inoculum size. In the absence of bacteria, earthworms were responsible for an 85% increase in

belowground biomass, but no effect on aboveground or total biomass was observed. When both organ-isms were present, the positive effect of earthworms progressively decreased with increasing densitiesof M. luteus. M. luteus affected floral organs and root system structure, whereas earthworms affected onlyroot system structure. Therefore, M. luteus is a DRB for A. thaliana, and earthworms cannot be used as abiocontrol agent for this DRB in conditions such as those of our experiment.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Some free living bacteria of the plant rhizosphere, called plantrowth promoting rhizobacteria or PGPR (Bashan and Holguin,998), have the potential to increase plant growth. Among severalroposed mechanisms, PGPR release signal molecules analogouso phytohormones, fix atmospheric nitrogen in the rhizosphere,olubilize nutrients, promote mycorrhizae, or regulate ethyleneroduction in roots; they can also act indirectly on plant growth

hrough the suppression of major plant pathogens (Nehl et al.,996; Persello-Cartieaux et al., 2003).

∗ Corresponding author. Tel.: +33 1 45 17 16 14; fax: +33 1 45 17 16 70.E-mail address: blouin@u-pec.fr (M. Blouin).

ttp://dx.doi.org/10.1016/j.apsoil.2014.04.004929-1393/© 2014 Elsevier B.V. All rights reserved.

On the contrary, other free living rhizobacteria described asdeleterious rhizo-bacteria (DRB) (Suslow and Schroth, 1982) caninhibit root and shoot growth. Since the symptoms caused by thesebacteria are scarce and non-specific (foliar symptoms similar tonutrient deficiencies, discrete root symptoms such as browning anddiscoloration, few recorded cases of mortality), and because theyneed not have obligate physical contact with plant roots, they haveattracted much less scientific attention than physically associatedbacteria (pathogenic or symbiotic) and PGPR.

Nehl et al. (1996) argue that the distinction between PGPR andDRB is probably artificial, since bacteria are classified in one groupon the basis of bioassays, in which their net effect on plant growth

is determined. Positive effects of PGPR and negative effects of DRBmay be host specific at the species and even cultivar levels, andthere is little apparent relationship between taxonomic character-istics of rhizobacteria and their function as DRB or PGPR (Campbell

plied

eaat

moAeabCpwtBmeatIs

ogsrama2peeSmdOtti

tolreaasMiaWa

2

tetla

D.E. Castellanos Suarez et al. / Ap

t al., 1986; Suslow and Schroth, 1982). It is therefore likely that single organism may function alternatively as a DRB or PGPRccording to circumstances, i.e. plant genotype, soil conditions orhe presence of mycorrhizae (Nehl et al., 1996).

The effect of these microorganisms on plants can actually beediated by several kinds of molecules such as nutrients, antibi-

tics, siderophores or signal molecules mimicking plant hormones.mong these hormone-like compounds, auxins, cytokinins andthylene are the most widespread, although other compounds suchs gibberellin and jasmonate-like substances have been reported toe synthesized by soil and rhizosphere microorganisms (Persello-artieaux et al., 2003). Generally, auxins and cytokinins increaselant growth through the modification of plant morphogenesis,hereas ethylene, salicilic acid and jasmonic acid are beneficial due

o their role in the activation of resistance to pathogens (Ping andoland, 2004; van Loon et al., 1998). The concentration of the signalolecule is crucial. For example, the dose response curve of plant to

xogenous auxin application is bell shaped (Taiz and Zeiger, 2010)nd rhizobacteria can induce positive or negative effect accordingo their level of auxin production (Barazani and Friedman, 1999).t is thus of importance to test a gradient of inoculation dose whentudying the effect of a rhizobacteria.

Promoting microbial communities with dominance of PGPRver DRB is a promising ecological approach to increase plantrowth through an improved plant development. Our hypothe-is is that earthworms might promote this dominance for severaleasons. First, earthworms are ecosystem engineers that influencectivities of other organisms at modifying the physical environ-ent (Jones et al., 1994; Lavelle et al., 1997). Earthworm galleries

llow better water infiltration (Le Bayon et al., 2002; Shuster et al.,002) and drainage (Blouin et al., 2007b) and thus could avoid DRBroliferation which often occurs in waterlogged situations (Chant al., 1989; Mandryck, 1969). Earthworms actually have a positiveffect on plant growth in 80% of reported cases (Brown et al., 1999;cheu, 2003). Several results show that this effect is due to signalolecules emitted in the soil (Puga-Freitas et al., 2012b), probably

ue to the activation of PGPR (Brown et al., 2004; Elmer, 2009).ther indirect effects of earthworms on plant growth are mediated

hrough stimulation of plant defense (Blouin et al., 2005), or con-rol of plant pathogens, which can be associated with an increasen PGPR population (Elmer, 2009).

In this study, we tested the effects of the earthworm Aporrec-odea caliginosa and the rhizospheric bacteria Micrococcus luteusn the growth of the plant Arabidopsis thaliana. The bacteria M.

uteus is responsible for positive effects on the growth of tomato andadish and has been therefore ascribed to the PGPR group (Arguellot al., 2011). However, in another study, this bacterium had neg-tive effects on growth of young lettuce seedlings (Lactuca sativa)nd was thus considered a DRB (Barazani and Friedman, 1999). Ourtudy aimed to answer several questions: (i) What is the effect of. luteus on the growth and development of A. thaliana? (ii) If there

s a positive effect, is there a synergy with earthworms? (iii) If neg-tive, is there a control due to the presence of earthworms? (iv)hat can we learn regarding the mechanisms of action of M. luteus

nd earthworms from such a factorial experiment?

. Materials and methods

The experiment was carried out in the laboratory usinghe model plant A. thaliana cv. Colombia, grown in the pres-

nce/absence of the endogeic earthworm A. caliginosa and withhree different concentrations of viable cells of the bacteria M.uteus selected for its ability to produce large amounts of indolecetic acid (IAA).

Soil Ecology 80 (2014) 100–107 101

2.1. Microcosms

Soils were collected from the top layer (0–20 cm), at the Cen-tre de Recherche en Ecologie Expérimentale et Prédictive—CEREEP(Saint-Pierre-Lès-Nemours, France). It is a poor cambisol support-ing a natural meadow (total organic carbon content: 14.7 g kg−1;total nitrogen content: 1.19 g kg−1; pH, 5.22; CEC: 4.08 cmol kg−1)with a sandy texture (6.9% clay, 19.0% silt, 74.1% sand). This sandytexture is the most convenient to observe an effect of earthwormson plant growth (Brown et al., 1999). Soil was air dried in the shadeat 25 ◦C for several weeks, and sieved with a 2 mm mesh sieve.PVC pots with 13 cm diameter and 12 cm height were filled with1.2 kg of soil. Soil was moistened and kept at 80% field capacity withdeionised H2O (240 g of water per 1200 g dry soil), by weighing potstwice a week. No fertilizer was added.

2.2. Bacterial inoculum

Bacteria were isolated from the rhizosphere of healthy potatoplants (Solanum tuberosum) in an organic cultivation in the savannaof Bogotá (Colombia) (León, 2003). Taxonomic characterizationperformed by 16S rRNA gene sequences identified the bacteria usedin this experiment as M. luteus. M. luteus is a Gram positive coccus(Arguello et al., 2011). Colonies were isolated on agar supplementedwith lysogeny broth (LB) and grown in 100 ml of LB in 250 ml flask,up to a cell concentration of 1.0 × 106 cells ml−1 (Castellanos et al.,2009). Bottles were placed for 48 h in an orbital shaker at 170 rpmat 27 ◦C (Celis Bautista and Gallardo, 2008), for a final concentrationof 1.9 × 108 CFU ml−1. From this culture, we obtained respectivedilutions of 105, 106 and 107 CFU ml−1 by adding sterile deionizedwater. Three applications of bacterial inoculum were carried out attime zero (seeding), 7 and 14 days later. Each application consistedof four additions of 10 ml inoculum in four holes on the microcosmsurface (total of 40 ml), located on a 5 cm diameter circle from thecenter of the microcosm, around the seeds.

2.3. Earthworms

Adult earthworms (Aporrectodea caliginosa, the most abundantspecies in agrosystems, commonly used to assess the effect of earth-worms on plant growth) were collected in a silty and slightly acidicsoil (Dai et al., 2004), in which their density is higher than in thesandy soil used in the experiment, at the Institut de Recherchepour le Developpement station in Bondy (Seine Saint Denis, France).They were purged from their original soil by feeding them in thesandy soil used for the experiment for several weeks. This endogeicearthworm produces horizontal temporary burrows in the soil andcomes rarely to the soil surface (Bouché, 1977; Lee, 1985). They maydaily ingest 1–2 times their own weight of soil and deposit rathercompact casts with high contents of ammonium and available P thatattract roots (Zangerlé et al., 2011). Two weeks after soil set up andwatering, two adult earthworms with a 0.46 g average individualweight (gut content included) were inoculated in each unit, corre-sponding to a density of 151 ind. m−2 and a biomass of 52 g m−2,which falls in the range of 5–344 ind. m−2 and 1–63 g m−2 foundin clayed and sandy soils with organic or mineral amendments inagricultural fields (Lapied et al., 2009).

2.4. Plant growth conditions

Arabidopsis thaliana cv. Columbia was chosen because it is themodel plant in molecular biology, with many tools available for

future studies (CATMA microarrays, mutants, core collections, . . .).Ten seeds of were placed in the center of each microcosm 10days before earthworm introduction. Fifteen days after sowing,pots were thinned to leave one single plant in each microcosm.

1 plied

Pt21

2

(aCwtta

2

awEsCt

dBpvditbaao

2b

3su(ss(a(t��(dste

2

Tam

02 D.E. Castellanos Suarez et al. / Ap

lants were grown for 60 days under controlled conditions in a cul-ure chamber (Adaptis A1000, Conviron, Winnipeg, Canada) with2/20 ◦C day/night temperature, 70/75% day/night humidity and2 h photoperiod with 150 �mol photon m−2 s−1 intensity.

.5. Experimental design

Overall, there were 8 different treatments: 4 with earthwormsE0, E1, E2, E3), and four without earthworms (C0, C1, C2, C3). C0nd E0 were the control treatments without bacteria inoculated.1–E1, C2–E2 and C3–E3 treatments were inoculated respectivelyith 105, 106 and 107 CFU ml−1. Each treatment was replicated 6

imes, for a total of 48 microcosms. Microcosms were randomly dis-ributed in the culture chamber, and their position changed twice

week, after weighing and watering.

.6. Plant measurements

At the end of the experiment, floral stem length (plant height)nd rosette diameter were measured. Leaf surface area of each plantas analyzed with a digital scanner (Epson Expression 10000 XL,

pson America Inc., San Jose, CA, USA) coupled with the WinRHIZOoftware (WinRHIZO, V. 2007 pro, Regent Instruments, Quebec,anada), at 16 p mm−1 (400 dpi) resolution and using the automaticransformation threshold.

For root system morphology, we characterized dry root biomassistribution among diameter classes according to the method oflouin et al. (2007a). This method is based on the method used inedology to assess soil granulometry: roots are dried, cut trans-ersely with a mixer and placed on a column of sieves withecreasing mesh size; after shaking of the sieve column and cross-

ng of sieves by root fragments with a section diameter inferioro the mesh size, biomass distribution is assessed by weighting theiomass recovered in each sieve. Five diameter classes were chosen,ccording to sieve mesh size: 0–100, 100–200, 200–400, 400–600nd >600 �m. After root system analysis, above- and belowgroundrgans were dried for 2 days at 45 ◦C and weighed.

.7. Determination of indole acetic acid production by cultivableacteria

Twenty gram soil samples were retrieved from each microcosm0 and 60 days after sowing, at the four inoculation points. Sub-amples from the same date were homogenized and one gram wassed for measurements. Bacteria were extracted in 9 ml NaCl 0.9%w/v) solution under circular agitation at 40 rpm for 30 min. Theolution was then centrifuged at 1000 rpm for 3 min. One ml ofupernatant was mixed with 6.5 ml of Czapeck culture mediumNienhaus, 1969) and incubated under constant agitation (150 rpmt 26 ◦C). After 24 h, 800 �l of a stock solution of L-tryptophan2 �g �l−1) was added as IAA precursor. Colorimetric quantifica-ion of IAA production was made at 24, 48 and 72 h. Five hundredl were centrifuged for 75 s at 14,000 rpm. Two hundred and fiftyl of supernatant were mixed to 750 �l of Salkowsky reagent

Glickmann and Dessaux, 1995) and incubated for 30 min in theark. Absorbance was measured at 535 nm with the Novaspec IIIpectrophotometer (Amersham Biosciences, USA). IAA concentra-ion was obtained from absorbance according to a standard curvequation (y = 16.636x; R2 = 0.99).

.8. Statistical analyses

All analyses were performed using the R software (Team, 2008).o get an overview of plant response in our data set, we performed

multivariate analysis (Principal Component Analysis) of all treat-ents based on dependent variables (biomass and root and shoot

Soil Ecology 80 (2014) 100–107

morphology). The significance of treatment effect was tested witha Monte Carlo test. We also performed two-way ANOVA, withearthworm and M. luteus as factors, on all dependent variables.In addition, non-linear regressions for all dependent variablesallowed determination of the equation of regression curves for agiven dependent variable as a function of inoculation dose of M.luteus either in the presence or absence of earthworms (two dis-tinct regressions). Because the dose–response curve of the plantto M. luteus could be bell-shaped as in the case of exogenousauxin application, we first tested a second degree polynomialfunction (dependant variable = f(inoculum) = ax2 + bx + c) either inthe absence or presence of earthworms. If the second degreeparameter was not significantly different from zero, we used afirst degree polynomial function (f(x) = ax + b). As bacterial con-centrations corresponded to different powers of ten (105, 106

and 107), log-transformation of the x-axis scale (log(x + 10,000))was needed in order to present the data in a figure. This led todifferent regression equation than those obtained without trans-formation.

3. Results

During this experiment, earthworm number and biomassslightly decreased (from 2 to 1.7 for the number, from 0.46 to 0.39 gfor the biomass per microcosm with earthworms). However, thisdecrease had no significant effect on belowground, aboveground ortotal biomasses (p-value = 0.68; 0.31 and 0.40, respectively). Therewas also a production of cocoons (0.50 cocoon microcosm−1), with-out correlation with plant total biomass (p-value = 0.30).

3.1. Overall effect of treatments on plant parameters

The principal component analysis constrained with the discrim-inating factors “earthworms” and “M. luteus” (“between” analysiswith R) indicated that 50% of the total variance in plant parameterswas explained by the analysis (Monte Carlo test: p = 0.001) (Fig. 1).Axis 1 explained 39% of the total variance explained and axis 2 22%.Axis 1 separated treatments with the highest inoculum concentra-tions on the left (E3, C3) to treatments with the lowest inoculumon the right (E0, C0, C1). Treatments with low inoculum concen-tration had higher shoot and total biomass. From a developmentalviewpoint, plants from these treatments had root systems with ahigh percentage of fine roots (0–100 and 100–200 �m diameterclasses), a longer floral stem and numerous leaves. Treatments withthe highest inoculation rates exhibited plants with the oppositeproperties, especially smaller biomass and coarser roots (400–600and >600 �m diameter classes).

Axis 2 separated plants according to the presence of earth-worms, particularly when the inoculum concentration was low(E0 and E1 versus C0 and C1). Treatments with earthworms (inthe top of graph) exhibited higher leaf biomass, diameter andsurface area, whereas treatments without (in the bottom) had ahigher number of small leaves, a big floral stem and a high per-centage of roots of mid-size diameter (100–200 and 200–400 �mclasses).

3.2. Plant biomass

Regardless the presence or absence of earthworms, the two-way ANOVA (Table 1) indicated a significant negative effect of thedose of M. luteus on total biomass (p < 0.001), aboveground biomass

(p < 0.001) and belowground biomass (p = 0.003). When restrictingthe analysis to the four treatments without earthworms (C0, C1, C2and C3), the effect of M. luteus dose was not significant. The decreasein plant production due to M. luteus was significantly higher in

D.E. Castellanos Suarez et al. / Applied Soil Ecology 80 (2014) 100–107 103

Table 1Results of the two-way ANOVA with the factors earthworms, bacterial inoculum and their interaction on plant growth and development parameters.

Aboveground drybiomass

Belowground drybiomass

Total dry biomass Leaf area Plant height IAA concentration

Df F p-Val Df F p-Val Df F p-Val Df F p-Val Df F p-Val Df F p-Val

Earthworm 1 0.402 0.53 1 6.23 0.02 1 3.323 0.08 1 0.203 0.66 1 0.242 0.63 1 10.542 0.002Bacterial inoculum 3 9.797 <0.001 3 5.56 0.003 3 12.394 <0.001 3 0.978 0.41 3 30.224 <0.001 3 2.197 0.1Earthworm × bacterial 3 1.335 0.28 3 1.86 0.15 3 2.263 0.1 3 3.351 0.03 3 0.497 0.69 3 6.493 <0.001

tT1i

FawdtctM

inoculumResiduals 38 38 37

he presence of earthworms than in the control with no worms.otal biomass was equal to 1.56 ± 0.01 g (mean ± S.E.) in C0 and.20 ± 0.07 g in C3, corresponding to a decrease of 23% (p = 0.21)

n the absence of earthworms, whereas in their presence, it was

ig. 1. PCA ordination of the different growth parameters of Arabidopsis thalianaccording to inoculum dose of Microccocus luteus in the absence or presence of earth-orms (Aporrectodea caliginosa). (a) Correlation circle indicating the contribution ofependant variables to the principal components. “t0p”:root biomass proportion inhe 0–100 �m diameter class, etc. “biom”: biomass. (b) Position of the different repli-ates and barycenter for each treatment in the plane defined by the axis 1 and 2 ofhe PCA. C: control without earthworms, E: treatment with earthworms, 0: without. luteus, with an inoculum of 1: 105 CFU ml−1, 2: 106 CFU ml−1, 3: 107 CFU ml−1.

38 32 40

equal to 1.78 ± 0.03 g in E0 and 1.11 ± 0.03 g in E3, correspond-ing to a decrease of 38% (p < 0.001). Belowground biomass wasequal to 0.206 ± 0.012 g in C0 and 0.142 ± 0.011 g in C3, correspond-ing to a decrease of 31% (p = 0.96), in the absence of earthworms,whereas in their presence, it was equal to 0.382 ± 0.138 g in E0and 0.138 ± 0.004 g in E3, corresponding to a decrease of 64%(p = 0.003). Aboveground biomass was equal to 1.38 ± 0.02 g inC0 and 1.06 ± 0.06 g in C3, corresponding to a decrease of 24%(p = 0.16), in the absence of earthworms, whereas in their pres-ence, it was equal to 1.40 ± 0.01 g in E0 and 0.97 ± 0.03 g in E3,corresponding to a decrease of 30% (p = 0.007).

The two-way ANOVA also showed that earthworms had apositive effect on belowground biomass (p = 0.017) but not onaboveground biomass (p = 0.53). Restricting the analysis to thetreatments without Microccocus luteus (C0 and E0), we found a sig-nificant effect of earthworms on total biomass (+14%, p = 0.017),belowground biomass (+85%, p = 0.045), but not on abovegroundbiomass (+1.4%, p = 0.762).

The interaction between M. luteus and earthworms had nosignificant effect on biomass. When analyzing these data in apolynomial regression of biomass as a function of M. luteus dosein presence/absence of earthworms, we found a decrease inbiomass production with an increasing dose of M. luteus inocu-lum, especially for the root system (Fig. 2): (i) for root biomass,equations were y = −0.05x + 0.18 in the absence of earthwormsand y = −0.18x + 0.25 in the presence of earthworms (R2: 0.43, F-statistic: 4.07 on 7 and 38 DF, p-value: 0.002 for the global model);(ii) for shoot biomass, there was no significant difference betweenthe absence and presence of earthworms, and the equation wasy = −0.19x2 − 0.30x + 1.28 (R2: 0.47, F-statistic: 4.83 on 7 and 38DF, p-value: 0.0006 for the global model); (iii) for total biomass,equations were y = −0.20x + 1.53 in the absence of earthwormsand y = −0.50x + 1.53 in the presence of earthworms (R2: 0.56, F-statistic: 6.76 on 7 and 37 DF, p-value: 3.5 × 10−5 for the globalmodel); (iv) for shoot/root ratio, the regression model was notsignificant (multiple R-squared: 0.24, adjusted R-squared: 0.093,F-statistic: 1.64 on 7 and 37 DF, p-value: 0.15 for the global model).Earthworm presence induced higher root biomass in the treatmentwithout M. luteus, but this positive effect disappeared as M. luteusinoculum dose increased.

3.3. Aboveground morphology

M. luteus dose affected plant development, especially reproduc-tive organs, with a negative impact on floral stem height (p < 0.001)(Table 1 and Fig. 3), tissue density (height/biomass) of floral stem(p = 0.003) and the ratio between floral stem biomass and leafbiomass (p < 0.001) (data not shown).

Surface leaf area was affected neither by earthworm nor M.

luteus, but by the interaction between these factors (p = 0.029)(Table 1 and Fig. 4): the dose of M. luteus was responsible foran increase in leaf area in the absence of earthworms, but for adecrease in the presence of earthworms.

104 D.E. Castellanos Suarez et al. / Applied Soil Ecology 80 (2014) 100–107

Fig. 2. Effect of Microccocus luteus inoculation dose in the absence (C) or presence(E) of earthworms (Aporrectodea caliginosa) on Arabidopsis thaliana (a) shoot (C:r2 = 0.12; E: r2 = 0.64), (b) root (C: r2 = 0.21; E: r2 = 0.38) and (c) total (C: r2 = 0.15; E:r2 = 0.68) biomass production.

Fig. 3. Effect of Microccocus luteus inoculation dose on Arabidopsis thaliana floralstem height, irrespectively of the presence of earthworms (Aporrectodea caliginosa)(r2 = 0.69).

Fig. 4. Relationship between Microccocus luteus inoculation dose and Arabidopsisthaliana leaf area in the absence (C) or presence (E) of earthworms (Aporrectodeacaliginosa) and their respective regression (C: r2 = 0.25; E: r2 = 0.03).

3.4. Belowground morphology

Root biomass distribution in diameter classes was affectedby earthworms and M. luteus. M. luteus affected the absoluteroot biomass in each diameter class: 0–100 (p = 0.04), 100–200(p = 0.01), 200–400 (p = 0.004), 400–600 (p = 0.007), >600 �m(p = 0.001), whereas earthworms affected only the 0–100 �m class(p = 0.02) (Fig. 5a). To detect qualitative changes in root systemstructure regardless of the variations in root system size, we calcu-lated absolute biomass per diameter class divided by the total rootbiomass, to obtain a relative root biomass proportion per diameterclass. M. luteus dose had an effect on the proportion of root biomassin fine (0–100 �m: p = 0.007) and thick (400–600 �m: p = 0.036;>600 �m: p = 0.016) roots. Earthworms only had an impact onmiddle diameter roots (200–400 �m: p = 0.050). The interactionbetween earthworms and M. luteus dose was significant only onthe 400–600 �m class (p = 0.037).

The root biomass distribution profiles for each of the eight treat-ments showed that there were little variations according to theM. luteus dose in the absence of earthworms, but important varia-tions according to M. luteus dose in their presence, for the absolutebiomass (Fig. 5a) as well as for the relative (Fig. 5b).

3.5. IAA production by cultivable bacteria

IAA production by cultivable bacteria was measured in themiddle (30 days after sowing) and in the end (60 DAS) of theexperiment. No relationship between IAA production at 60 daysand factors or dependant variables was observed, so measure-ments at this date will not be discussed. For IAA production at 30days, the two-way ANOVA did not show a clear effect of M. luteusdose on IAA production by cultivable bacteria (p = 0.10). Conversely,earthworms clearly increased IAA production by cultivable bacte-ria (p = 0.002) (Fig. 6a). The interaction between earthworms and M.luteus dose was significant (p = 0.001): the shape of the curve rep-resenting the effect of M. luteus dose in the absence of earthwormswas concave, whereas it was convex in the presence of earthworms(Fig. 6b).

4. Discussion

4.1. Effect of M. luteus on the growth and development of A.thaliana

In our experiment, we observed a negative effect of M. luteus onA. thaliana total, above and below-ground biomass, regardless ofthe dose of bacteria inoculated. M. luteus is therefore a DRB as far

D.E. Castellanos Suarez et al. / Applied Soil Ecology 80 (2014) 100–107 105

Fig. 5. Effect of Microccocus luteus on root profile of Arabidopsis thaliana in the absence (C)distribution in diameter classes. (b) Relative proportion of root biomass in diameter cla106 CFU ml−1, 3: 107 CFU ml−1.

Fig. 6. (a) IAA production by cultivable soil micro-organisms in the absence (C) orpresence (E) of earthworms (Aporrectodea caliginosa). Means ± S.E., n = 24 per treat-ment, different letters indicates a significant difference, Tukey HSD, p-value < 0.05.(b) Relationship between the inoculation dose of Microccocus luteus on IAA pro-duction by cultivable soil micro-organisms in the absence (C) or presence (E) ofearthworm (Aporrectodea caliginosa) and their respective regression (C: r2 = 0.22; E:r2 = 0.39).

or presence (E) of earthworms (Aporrectodea caliginosa). (a) Absolute root biomasssses. Grey intensity indicates increasing doses; 0: no bacteria; 1: 105 CFU ml−1, 2:

as A. thaliana is concerned. Three main hypotheses could explainthis: the dose of M. luteus inoculated, the amount of IAA perceivedby the plant and the plant genotype.

With some exceptions, increasing densities of DRB leads to plantgrowth inhibition (Nehl et al., 1996). For example, a study on theeffect of Pseudomonas thivervalensis (an IAA-producing strain) onA. thaliana root length and branching showed that an inoculum of105 CFU ml−1 induced favorable qualitative changes on colonizedplantlets, whereas above 106 CFU ml−1 the inoculation caused irre-versible damage to plants (Persello-Cartieaux et al., 2001). Ourexperimental inoculation corresponded to a concentration of 105

to 107 CFU ml−1, which is in the common range of PGPR and DRBinoculations. It is difficult to judge the realism of such levels ofinoculation due to the different quantification methods used in thelaboratory and in the field. In natural field conditions, PGPR and DRBare retrieved in the rhizosphere, not in the bulk soil, and the den-sity unit is the number of CFU per g−1 of fresh roots; for example,M. luteus (rarely reported in literature) is found at a concentra-tion of 63 and Burkholderia sp. at 1.5 × 105 CFU g−1 fresh weightroots in the rhizosphere of Chamaecytisus proliferus, with the aver-age abundance for other taxa about 103–104 CFU g−1 fresh weightroots (Donate-Correa et al., 2004). In laboratory experiments, PGPRand DRB are inoculated in the soil, and plant roots meet them dur-ing their growth. To compare between the laboratory and the field,we converted the bacterial concentration in CFU ml−1 inoculated atthe beginning of the experiment to CFU g−1 of fresh root biomassmeasured at the end of the experiment (because root biomass wasquasi null at the beginning), and we got an average concentrationof 3 × 103, 3 × 104 and 3 × 105 CFU g−1 fresh root biomass for the105, 106 and 107 CFU ml−1 treatments respectively. These valuesare probably above the ones encountered by the plant because asignificant proportion has probably died before to meet plant roots.Thus, our laboratory inoculum is close to what can be observed in

the field.

Another hypothesis to explain the negative effect on plantgrowth concerns the amount of IAA perceived by the plant, whichis a combination of the amount of IAA produced by the bacteria

1 plied

arSngEisrIrEptpo1f(oipcOlt

ieCoAtoasesonpFdbs(ettM

4e

gTareMeAlpr

06 D.E. Castellanos Suarez et al. / Ap

nd the life time and diffusion of IAA in the soil. Up to 80% ofhizobacteria can synthesize indole-3-acetic acid (IAA) (Loper andchroth, 1986), but the dose response curve of plants to exoge-ous auxin application is bell shaped (Taiz and Zeiger, 2010). Rootrowth promotion by free living PGPR, e.g., Alcaligenes faecalis,nterobacter cloacae, Acetobacter diazotrophicus, species of Azospir-llum, Pseudomonas, and Xanthomonas, as well as by symbionts,uch as Bradyrhizobium japonicum and Rhizobium spp., has beenelated to low levels of IAA production (Patten and Click, 1996).n contrast, the inhibitory effect of some deleterious rhizobacte-ia (DRB) has been related to high amounts of IAA production, e.g.,nterobacter taylorae (Sarwar and Kremer, 1995) and Pseudomonasutida (Xie et al., 1996). For instance, bacteria responsible for let-uce root growth suppression (M. luteus, Streptoverticillium sp., P.utida, and Gluconobacter sp.) were shown to produce high levelsf IAA (an average of 76.6 �M for the four bacteria, and a value of95.1 ± 0.2 �M after 84 h of culture at 27 ◦C for M. luteus), whereasour isolates responsible for a positive effect on lettuce growthAgrobacterium sp., Alcaligenes piechaudii, and two different strainsf Comamonas acidovorans) secreted lower levels of IAA (16.4 �Mn average) (Barazani and Friedman, 1999). For a given level of IAAroduction, soil properties can influence the amount of IAA per-eived by the plant by changing IAA life time and diffusion speed.ur soil is poor in organic matter and clay, so IAA could have been

ess adsorbed than in silty or clayed soil, diffuse more rapidly andhus be perceived in higher amount by the plant.

The third hypothesis is concerning plant genotype: plant speciess also important in explaining the direction and extent of theffect of rhizobacteria on plant growth (Nehl et al., 1996; Persello-artieaux et al., 2003). As with all pathogens, DRB are pathogenicn some hosts and not others (Schippers et al., 1987). For example,zotobacter paspali has almost absolute specificity for a single eco-ype of Paspalum notatum Aüggé (Dobereiner and Pedrosa, 1981). Inur case, the strain used in this experiment has been selected from

rhizospheric soil for its positive effect on red radish (Raphanusativus) and tomato (Lycopersicon esculentum) biomass (Arguellot al., 2011), but here, we found a negative effect on A. thaliana. Theensitivity of the plant species to a given amount of IAA, which reliesn genetic and physiological factors, is likely to be a strong determi-ant of the net effect of rhizobacteria on plant growth. Moreover, alant can actively change the amount of IAA production by bacteria.or example, A. thaliana can reduce by 43% the amount of IAA pro-uced by cultivable bacteria (Puga-Freitas et al., 2012a). This coulde due to the limitation of tryptophan release in the soil by the plant,ince soil tryptophan is a precursor of IAA production by bacteriaPersello-Cartieaux et al., 2003). This regulatory mechanism couldxplain the relatively low level of IAA production by cultivable bac-eria in our experiment (i.e. 45 ng ml−1 for 24 h), approximately 200imes lower than observed by Barazani and Friedman (1999) with. luteus.

.2. Is there a biocontrol of M. luteus effects in the presence ofarthworms?

In our experiment, earthworms had a positive effect on plantrowth (especially roots) in the absence of M. luteus inoculation.his positive effect however decreased with the dose of M. luteus,nd finally disappeared at 107 CFU ml−1 (Fig. 2). Therefore, theegression slope for root biomass was more negative in the pres-nce than absence of earthworms (Fig. 2b). Instead of a control of. luteus effect by earthworms, we observed a disappearance of

arthworm positive effect with increasing inoculum of M. luteus.

. caliginosa is thus not relevant as a biocontrol agent against M.

uteus for A. thaliana in our experimental conditions. As far as IAAroduction by soil bacteria is concerned, the complexity of ouresults (Fig. 6b) shows the difficulty of predicting the effect of an

Soil Ecology 80 (2014) 100–107

assemblage of two organisms known for their involvement in theproduction of signal molecules such as auxine.

4.3. How to explain the combined effect of M. luteus andearthworms on plant growth?

Five mechanisms are potentially responsible for the positiveeffect of earthworms observed on plant production (Brown et al.,2004; Scheu, 2003): (i) increased mineralization of soil organic mat-ter, which increases nutrient availability; (ii) modification of soilporosity and aggregation, which induces changes in water and oxy-gen availability to plants; (iii) bio-control of pests and parasites; (iv)production of plant growth regulators through the stimulation ofmicrobial activity and (v) stimulation of symbionts. Recent studieson mechanisms iv and v have demonstrated that A. thaliana tran-scriptomic response to the presence of A. caliginosa is very close toits response to IAA producing rhizobacteria or exogenous applica-tions of IAA (Puga-Freitas et al., 2012b). Moreover, it is likely thatIAA produced in the presence of earthworms is in fact produced bybacteria specifically activated by earthworms (Elmer, 2009; Puga-Freitas et al., 2012a). The disappearance of the positive effect ofearthworms along the M. luteus dose gradient in our experimentcould be explained by the competitive exclusion, by M. luteus, ofIAA-producing bacteria activated by earthworms, either throughresource acquisition or antimicrobial compounds. This exclusionmight be stronger with increasing inoculum size of M. luteus, up to atotal inhibition of the positive effect of earthworms at 107 CFU ml−1

(Fig. 6b).Finally, we cannot exclude other hypotheses. In this study, we

only measured IAA production by microorganisms at 30 days aftersowing and not the kinetics of IAA production. A survey of IAA pro-duction during the whole experiment would probably improve ourunderstanding of biotic interactions through signal molecules. Itwould appear that the diversity of molecules in the auxine fam-ily (e.g. IAA, IALD, ILA, . . .) (Sarwar and Kremer, 1995; Glickmannand Dessaux, 1995), the diversity of signal molecules in generaland/or the coupling between these signal molecules and resourceavailability must be considered to understand plant growth promo-tion or inhibition in a given situation. Further research is necessaryto understand the consequence of these biochemical pathways ofplant-bacteria-earthworm interactions.

5. Conclusion

M. luteus, a bacteria selected for its high IAA production anda positive effect on red radish and tomato, was responsible for adensity-dependant negative effect on A. thaliana biomass produc-tion. The earthworm A. caliginosa positively affected plant biomass,but this effect disappeared with increasing M. luteus inoculum. Theexact mechanisms involved in earthworm and M. luteus effectsmust still be unraveled. Signal molecules such as IAA have been pro-posed as important for both agents, but our experiment stresses onthe need for a temporal analysis of IAA production, and the devel-opment of screening methods to identify the diversity of signalmolecules produced in the presence of micro and macro-organisms.A better understanding of signal molecules and their functioning incomplex soil systems could be of major help in developing alterna-tive agricultural practices through soil ecological engineering.

Acknowledgements

We are grateful to University Paris-Est Créteil and the programECOS Sud C09A01, AMAZ Agreg for supporting this research. Wethank Samir Abbad for advices regarding bacterial culture and IAAproduction.

plied

R

A

B

B

B

B

B

B

B

B

C

C

C

C

D

D

D

E

G

J

L

D.E. Castellanos Suarez et al. / Ap

eferences

rguello, H., Rincón, J., Castellanos, D., 2011. Obtención y evaluación de un biofer-tilizante enriquecido que contribuya con el incremento de la productividad decultivos de hortalizas. Universidad Nacional de Colombia, Bogotá, p. 209.

arazani, O., Friedman, J., 1999. Is IAA the major root growth factor secreted fromplant-growth-mediating bacteria? J. Chem. Ecol. 25, 2397–2406.

ashan, Y., Holguin, G., 1998. Proposal for the division of plant growth-promoting rhizobacteria into two classifications: biocontrol-PGPB (plant-growth-promoting bacteria) and PGPB. Soil Biol. Biochem. 30, 1225–1228.

louin, M., Barot, S., Roumet, C., 2007a. A quick method to determine root biomassdistribution in diameter classes. Plant Soil 290, 371–381.

louin, M., Lavelle, P., Laffray, D., 2007b. Drought stress in rice (Oryza sativa L.)is enhanced in the presence of the compacting earthworm Millsonia anomala.Environ. Exp. Bot. 60, 352–359.

louin, M., Zuily-Fodil, Y., Pham-Thi, A.T., Laffray, D., Reversat, G., Pando, A., Tondoh,J., Lavelle, P., 2005. Belowground organism activities affect plant abovegroundphenotype, inducing plant tolerance to parasites. Ecol. Lett. 8, 202–208.

ouché, M.B., 1977. Stratégies lombriciennes, in: Lohm, U., Persson, T. (Eds.), SoilOrganisms as Components of Ecosystems. Ecology Bulletins No. 25, Stockholm,pp. 122–132.

rown, G.G., Edwards, C.A., Brussaard, L., 2004. How earthworms affect plant growth:burrowing into the mechanisms. In: Edwards, C.A. (Ed.), Earthworm Ecology. ,2nd ed. CRC Press, Boca Raton, USA, pp. 13–49.

rown, G.G., Pashanasi, B., Villenave, C., Patron, J.C., Senapati, B.K., Giri, S., Barois,I., Lavelle, P., Blanchart, E., Blakemore, R.J., Spain, A.V., Boyer, J., 1999. Effectsof earthworms on plant production in the tropics. In: Lavelle, P., Brussaard, L.,Hendrix, P. (Eds.), Earthworm Management in Tropical Agroecosystems. CABInternational, Wallingford, pp. 87–148.

ampbell, J.N., Conn, K., Sorlic, L., Cook, F.O., 1986. Inhibition of growth in canolaseedlings caused by an opportunislic Pseudomonas sp. under taboralory and fieldconditions. Can. J. Microbiol. 32, 201–207.

astellanos, D., Cubillos, R., Argüello, H., 2009. Selección de microorganismos pro-motores de crecimiento vegetal (ácido indol acético) a partir de muestras desuelo rizosferico, como primera etapa en el desarrollo de un biofertilizante. Rev.Bras. de Agroecologia 4, 1720–1723.

elis Bautista, L.X., Gallardo, I.R., 2008. Estandarización de métodos de detecciónpara promotores de crecimiento vegetal (ácido indol acético y giberelinas) encultivos microbianos, Facultad de Ciencias Microbiología Agrícola y Veterinaria.Pontificia Universidad Javeriana, Bogotá, p. 140.

han, K.Y., Mead, J.A., Roberts, W.P., Wong, P.T.W., 1989. The effect of soil compactionand fumigation on poor early growth of wheat under direct drilling. Aust. J. Agric.Res. 42, 221–228.

ai, J., Becquer, T., Rouiller, J.H., Reversat, G., Bernhard-Reversat, F., Nahmani, J.,Lavelle, P., 2004. Heavy metal accumulation by two earthworm species and itsrelationship to total and DTPA-extractable metals in soils. Soil Biol. Biochem. 36,91–98.

obereiner, J., Pedrosa, F.A., 1981. Nitrogen Fixing Bacteria in Non-leguminousPlants. Science Tech, Madison, WI, USA.

onate-Correa, J., Leon-Barrios, M., Perez-Galdona, R., 2004. Screening for plantgrowth-promoting rhizobacteria in Chamaecytisus proliferus (tagasaste), a foragetree-shrub legume endemic to the Canary Islands. Plant Soil 266, 261–272.

lmer, W.H., 2009. Influence of earthworm activity on soil microbes and soilbornediseases of vegetables. Plant Dis. 93, 175–179.

lickmann, E., Dessaux, Y., 1995. A critical examination of the specificity of thesalkowski reagent for indolic compounds produced by phytopathogenic bac-

teria. Appl. Environ. Microbiol. 61, 793–796.

ones, C.G., Lawton, J.H., Shachak, M., 1994. Organisms as ecosystem engineers. Oikos69, 373–386.

apied, E., Nahmani, J., Rousseau, G.X., 2009. Influence of texture and amendmentson soil properties and earthworm communities. Appl. Soil Ecol. 43, 241–249.

Soil Ecology 80 (2014) 100–107 107

Lavelle, P., Bignell, D., Lepage, M., Wolters, V., Roger, P., Ineson, P., Heal, O.W., Dhillion,S., 1997. Soil function in a changing world: the role of invertebrate ecosystemengineers. Eur. J. Soil Biol. 33, 159–193.

Le Bayon, R.C., Moreau, S., Gascuel-Odoux, C., Binet, F., 2002. Annual variations inearthworm surface-casting activity and soil transport by water runoff under atemperate maize agroecosytem. Geoderma 106, 121–135.

Lee, K.E., 1985. Earthworms. Their Ecology and Relationships with Soils and LandUse. Sydney, Academic Press.

León, S.T., 2003. Resistencia sistémica inducida con caldos microbianos de rizosferapara control de gota (Phytophthora infestans (Mont.) de Bary) en agroecosis-temas de papa. Escuela Técnica Superior de Ingenieros Agrónomos. UniversidadPolitécnica de Madrid, Madrid, p. 256.

Loper, J.E., Schroth, M.N., 1986. Influence of bacterial sources of indole-3-acetic-acidon root elongation of sugar beet. Phytopathology 76, 386–389.

Mandryck, M., 1969. Frenching of tobacoo in Australian soils and in soil leachates.Aust. J. Agric. Res. 20, 709–717.

Nehl, D.B., Allen, S.J., Brown, J.F., 1996. Deleterious rhizosphere bacteria: an inte-grating perspective. Appl. Soil Ecol. 5, 1–20.

Nienhaus, F., 1969. Phytopathologisches Praktikum. Parey, Berlin.Patten, C.L., Click, B.R., 1996. Bacterial biosynthesis of indole-3-acetic acid. Can. J.

Microbiol. 42, 207–220.Persello-Cartieaux, F., David, P., Sarrobert, C., Thibaud, M.C., Achouak, W., Robaglia,

C., Nussaume, L., 2001. Utilization of mutants to analyze the interaction betweenArabidopsis thaliana and its naturally root-associated Pseudomonas. Planta 212,190–198.

Persello-Cartieaux, F., Nussaume, L., Robaglia, C., 2003. Tales from the underground:molecular plant-rhizobacteria interactions. Plant Cell Environ. 26, 189–199.

Ping, L., Boland, W., 2004. Signals from the underground: bacterial volatiles promotegrowth in Arabidopsis. Trends Plant Sci. 9, 263–266.

Puga-Freitas, R., Abbad, S., Gigon, A., Garnier-Zarli, E., Blouin, M., 2012a. Con-trol of cultivable IAA-producing bacteria by the plant Arabidopsis thaliana andthe earthworm Aporrectodea caliginosa. Appl. Environ. Soil Sci. 2012, ArticleID307415.

Puga-Freitas, R., Barot, S., Taconnat, L., Renou, J.-P., Blouin, M., 2012b. Signalmolecules mediate the impact of the earthworm Aporrectodea caliginosa ongrowth, development and defence of the plant Arabidopsis thaliana. PloS One7, e49504.

Sarwar, M., Kremer, R.J., 1995. Enhanced suppression of plant-growth through pro-duction of L-tryptophan derived compounds by deleterious rhizobacteria. PlantSoil 172, 261–269.

Scheu, S., 2003. Effects of earthworms on plant growth: patterns and perspectives.Pedobiologia 47, 846–856.

Schippers, B., Bakker, A.W., Bakker, P., 1987. Interactions of deleterious and beneficialrhizosphere microorganisms and the effect of cropping practices. Annu. Rev.Phytopathol. 25, 339–358.

Shuster, M., McCartney, P., Studer, S., 2002. Nitrogen source and earthworm abun-dance affected runoff volume and nutrient loss in a tilled-corn agroecosystem.Biol. Fertil. Soils 35, 320–327.

Suslow, T.V., Schroth, M.N., 1982. Role of deleterious rhizobacteria as minorpathogens in reducing crop growth. Phytopathology 72, 111–115.

Taiz, L., Zeiger, E., 2010. Plant Physiology, third ed. Sinauer Associates Inc., Sunder-land, USA.

Team, R.D.C., 2008. R: A Language and Environment for Statistical Computing. RFoundation for Statistical Computing, Vienna, Austria.

van Loon, L.C., Bakker, P.A.H.M., Pieterse, C.M.J., 1998. Systemic resistance inducedby rhizosphere bacteria. Annu. Rev. Phytopathol. 36, 453–483.

Xie, H., Pasternak, J.J., Glick, B.R., 1996. Isolation and characterization of mutants ofthe plant growth promoting rhizobacterium Pseudomonas putida GR12-2 thatoverproduce indoleacetic-acid. Curr. Microbiol. 32, 67–71.

Zangerlé, A., Pando, A., Lavelle, P., 2011. Do earthworms and roots cooperate to buildsoil macroaggregates? A microcosm experiment. Geoderma 167–168, 303–309.