lable at ScienceDirect

Soil Biology & Biochemistry 99 (2016) 1e20

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Assessment of bacterial populations associated with banana tree rootsand development of successful plant probiotics for banana crop

Iris-Esther Marcano a, C�esar-Antonio Díaz-Alc�antara a, Beatriz Urbano b,Fernando Gonz�alez-Andr�es c, *

a Faculty of Agronomic and Veterinary Sciences, Universidad Aut�onoma de Santo Domingo, Dominican Republicb Department of Agricultural and Forestry Engineering, Universidad de Valladolid, Spainc Institute of Environment, Natural Resources and Biodiversity, Universidad de Le�on, Spain

a r t i c l e i n f o

Article history:Received 18 September 2015Received in revised form4 April 2016Accepted 10 April 2016Available online 4 May 2016

Keywords:Plant probiotic microorganismsMycosphaerella fijiensisPGPRMusa sp.Caribbean

* Corresponding author. Institute of Environment,versity, Av. de Portugal, 41, E-24009, Leon, Spain.

E-mail address: fgona@unileon.es (F. Gonz�alez-An

http://dx.doi.org/10.1016/j.soilbio.2016.04.0130038-0717/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

Two hundred and sixty-one bacterial isolates associated with banana tree roots in organic systems wereisolated from 19 farms located in four different provinces of the Dominican Republic. The isolates wereanalysed by means of ARDRA plus RAPD, and as a consequence 114 of them were selected for identifi-cation by means of complete 16S rRNA sequencing and phylogenetic reconstruction. The 114 isolatesbelonged to 20 different genera, with Bacillus, Pseudomonas, Enterobacter and Stenotrophomonas theprevailing genera. Of these, 65 isolates showed more than 99.5% similarity with a type strain, and theywere assigned to 34 different species from 16 genera; 29 isolates could constitute new species and theremaining 20 isolates belonged to groups containing more than one species with identical 16S rRNAgenes, and therefore they could not be assigned to any species. This result showed a higher number ofbacterial taxa associated with banana tree roots than previously described. Additionally, we found sevenbacterial species with significant in vitro plant growth promoting (PGP) activity, which had not beenpreviously described as PGP bacteria in any crop. Field trials with isolates pre-selected based on theirin vitro PGP activity showed that one strain of the species Pseudomonas plecoglossicida improved fruityield and controlled the incidence of the disease black sigatoka caused by Mycosphaerella fijiensis. Thisactivity was tentatively attributed to induced systemic resistance mechanisms. Bacterial diversity wasanalysed among the 261 isolates based on the Shannon index of diversity (H), calculated from the ARDRAprofiles. Interestingly, the majority of the bacterial diversity was found within farms (86% of the total),being higher than the bacterial diversity between farms (14%). Moreover, the differences in the average HIndex between provinces were very low. Consequently the biodiversity of the bacterial communities waslittle influenced by the soil characteristics. These results could work in favour of the efficient adaptationof the bacterial strains selected for use as plant probiotics in a range of soils in the region analysed.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Banana is a key crop for the economy in tropical regions (De laTorre-Guti�errez et al., 2008). Moreover, it is ranked the eighthcrop worldwide for its importance as a foodstuff, and the fourth,after rice, wheat and maize when only developing countries areconsidered (Arias et al., 2004). Banana is a “dual crop” in manytropical countries because it is cultivated as a subsistence crop but

Natural Resources and Biodi-

dr�es).

is also one of the most important crops for export. Banana export isa technological and economic activity, different from its productionas subsistence crop (Arias et al., 2004). The preferences of the in-ternational market make bananas produced in certified organicsystems more competitive than conventionally produced bananas.In the Dominican Republic, organic bananas representedmore than50% of the total banana exports on average during the last fiveyears, and over 85% of the organic bananas produced in DominicanRepublic are exported (CEI-RD, 2013).

Nowadays, there is no doubt about the important role of thebeneficial interactions between microbes and plants in improvingagricultural production. The exploitation of the intrinsic biologicalpotential of rhizosphere processes improves nutrient use efficiency

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e202

(Pii et al., 2015) and plant health by different mechanisms of action(Babalola, 2010; Cummings, 2009; Lugtenberg and Kamilova, 2009;Maheshwari, 2011). In organic systems, where the use of traditionalchemical products is limited and strictly regulated, the exploitationof plantemicrobe interactions to improve the nutritional andhealth status of the plant is especially important.

The use of seeds or soil inoculants consisting of products basedon soil microorganisms called bioprotectants, biofertilisers or bio-stimulants (Viveros et al., 2010; Prashar et al., 2014) ormore generalplant probiotic microorganisms (PPM) (Lugtenberg and Kamilova,2009), is becoming more popular in agriculture. What it is ex-pected from such inoculants is that the crops show a yield increaseand a better defence against pathogens, with reduced or eliminatedchemical inputs (Bhardwaj et al., 2014). These results are a conse-quence of a complex mix of different modes of action involving:improvement of plant nutrition through better nutrient uptake,improved nitrogen fixation, and phosphate, potassium and ironsolubilisation or mineralisation; control of diseases by severalmechanisms; and increase of chlorophyll content and photosyn-thetic activity (Mantelin and Touraine, 2004; Glick et al., 2007;Esitken et al., 2010; Singh et al., 2011; Adesemoye et al., 2009; DeSouza et al., 2013; Sinha et al., 2010).

However, there are still inconsistencies in the performance ofthe inoculants at the field scale (Morrissey et al., 2004). To tacklethis issue, research is being undertaken to learn how to prepare therhizosphere environment for PGPR rhizosphere colonization bymeans of rhizosphere engineering (Ryan et al., 2009). Nevertheless,at present, the best strategy to improve the performance of themicroorganisms at the field scale is to search for region-specificmicrobial strains, to be used as the inocula to achieve the desiredeffects in the crop (Deepa et al., 2010). Knowledge of the nativebacterial population, its characterisation, and identification isrequired to understand the distribution and diversity of indigenousbacteria in the rhizosphere of specific crops in a specific region(Keating et al., 1995; Chahboune et al., 2011).

Therefore, the present work analyses the culturable bacterialcommunities associated with the roots of banana trees in theDominican Republic, with the objective of designing plant pro-biotics based on microorganisms for such crops. As each plantspecies selects the bacteria associated with them by means ofseveral mechanisms, it is important to know the regional bacterialstrains associated with a given crop in order to design successfulinoculants. Consequently the specific objectives were: i) to identifythe cultivable bacterial populations associated with banana treeroots; ii) to test the effect of strains with in vitro PGP traits in fieldconditions; iii) to investigate the biodiversity of the bacteria asso-ciated with the banana tree roots and their determinant factors; iv)to work out practical considerations for the design of successfulplant probiotics based on microorganisms for banana crops.

2. Materials and methods

2.1. Isolation of bacteria

The isolation of bacteria associated with the roots of bananatrees (Musa AAA cv. 'Dwarf Cavendish') was carried out in 19 farmsunder organic management, from the four provinces that are themain producers of banana in the Dominican Republic (Fig. 1). Thefarmers planted the crop from plants produced in vitro. The sam-pling was made in the third, fourth or fifth ratoon depending on thefarm, at the phenological stage of bunch emergence (E). A briefsummary of the soil characteristics is Table 1. The complete soilanalyses are in Table S1. For the isolation of endophytic bacteria, theroots were washed with tap water to eliminate adhered particles ofsoil, following sterilization of the outer part of the roots with 70%

ethanol for 1 min and 7% NaClO for 5 min, and three successivewashings with sterile distilled water. Five grams of the surface-sterilised roots were macerated with 45 ml of sterile saline solu-tion and dilution series were prepared with the saline solution.Aliquots of 100 ml from 10�1 to 10�3 dilutions were plated onto petridishes with TSA medium (SIGMA Cat. No. 22091) supplementedwith 1 mg l�1 cycloheximide to prevent fungal growth. For rhizo-spheric bacteria extraction, roots were brushed with a No. 4 brush,and 1 g of the soil obtained was placed in 9 ml of sterile salinesolution and dilution series were prepared as indicated above, andplated onto petri dishes with the same medium. In both cases, in-dividual colonies were tested for purity in other petri dishes withTSA medium, and used for the next steps.

2.2. Amplified 16S rRNA restriction analysis (amplified ribosomalDNA restriction analysis, ARDRA)

DNA bacterial extraction was carried out as indicated by�Alvarez-Martínez et al. (2009). Amplification of the 16S rRNA genewas performed from bacterial genomic DNA by PCR. The primers1522R (50-AAGGAGGTGATCCANCCRCA-30) and 8F(50AGAGTTTGATCCTGGCTCAG-30) were used to amplify almost thefull length of the 16S rRNA gene. The primers were purchased fromISOGEN, and the kit used for the PCR reaction from Promega Cor-poration (USA) (“Go Taq Flexi DNA Polymerase”). Each PCR mixture(50 ml) contained a reaction mix of 1X buffer, 200 mM of each dNTP,0.12 mM of each primer, 2 mMMgCl2, 0.85 U of the polymerase andapproximately 100 ng template DNA. PCR amplification was per-formed under the following conditions: 5 min at 95 �C for initialdenaturation; 35 cycles of 1 min at 94 �C for denaturation; 30 s at45 �C for annealing; 1 min 30 s at 72 �C for extension; and 7 min at72 �C for a final extension. Five microlitres of the PCR product wasused for electrophoresis in a 1.5% agarose gel in TBE buffer (100mMTris, 83 mM boric acid, 1 mM EDTA, pH 8.5) at 6 V/cm, stained in asolution containing 0.5 g/ml ethidium bromide, and photographedunder UV light. Standard VI (Roche, USA) was used as a size marker.One single band corresponding to the 16S rRNA gene was typicallyvisualised.

After PCR amplification of the 16S rRNA, 6 ml of each product wasdigested separately by the restriction enzymes (RE) BanI, EcoRI,HinfI and StuI (Thermo Fisher Scientific, USA) and Sau3AI (NewEngland Biolabs Inc. USA) following the manufacturer's in-structions, and subjected to electrophoresis on a 2% (w/v) agarosegel prepared in the same way as indicated above. The electropho-resis conditions, the staining procedure and the photography pro-cess were also the same as above.

The band pattern obtained from the ARDRA, was coded in apresence/absence matrix in order to calculate the Shannon index ofdiversity (H index) (Shannon and Weaver, 1949). A principalcomponent analysis (PCA) was carried out for the 18 soil physico-chemical characters analysed (Table S1) plus the H index, for the11 farms (cases) with a complete soil analysis (the rest of the farmshad to be discarded because several parameters of the soil analysiswere missing). The data matrix was log10 transformed for stand-ardisation, and then the variance-covariance matrix was obtainedand the PC (principal components) extracted, and finally rotatedusing the varimax method. A biplot was drawn for the two first PCfor the 11 farms and the 19 characters analysed. All the analyseswere carried out by means of the software SPSS Statistics (IBMCorp. Released, 2010).

Additionally, the band pattern model for each isolate and eachRE was recorded.

Fig. 1. Locations of the isolation of bacteria.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 3

2.3. RAPD fingerprinting

Bacterial DNA isolated as described abovewas used to obtain theRAPD patterns following the procedure described by Rivas et al.(2006) using the primer M13 (50e GAGGGTGGCGGTTCT e30)(2 mM final concentration) purchased from ISOGEN, and the“Dream Taq Green PCR Master Mix” from Thermo Fisher Scientific,USA. PCR conditions were: preheating at 95 �C for 9 min; 35 cycles

of denaturing at 95 �C for 1 min; annealing at 45 �C for 1 min;extension at 75 �C for 2min; and a final extension at 72 �C for 7min.Seventeen microlitres of each PCR product were used in electro-phoresis in a 1.5% agarose gel in TBE buffer and photographed usingthe conditions described above. Standard VI (Roche, USA) was usedas a size marker. For all the isolates included in a restriction group(RG) (see Section 3.2 for definition) the band patterns were codedas presence/absence and Jaccard's similarity coefficient was used to

Table 1Location and summary of the main soil characteristics of the sites where the soil bacteriawere isolated, as well as the same information about the plot for the field experiment.

Province Location Latitude Longitude Soil characteristics

Texture pH OM (%) Phosphorus Olsen (mg kg�1) Potasium (cmol(þ) kg�1) Electric conductivity (dS m�1)

Collecting pointsAzua IDIAF N18 23,405 W70 50,375 Loam 8.58 1.81 <5.44 1.34 0.14Azua Project 2C N18 22,357 W70 50,410 e 8.30 3.89 42.49 6.02 0.41Azua Project Isura N18 25,233 W70 50,254 Clay 8.00 3.01 12.27 2.51 0.61Azua Pueblo Viejo N18 24,006 W70 46,001 e 8.27 8.83 12.42 1.66 0.22Azua Los Tramojos N18 25,397 W70 44,290 e 8.40 2.02 19.24 3.58 0.23Dajab�on Carrizal N19 20,887 W71 37,056 Clay loam 5.08 9.64 <5.44 0.19 0.05Dajab�on Rabizal N19 16,404 W71 33,583 Silt 7.51 0.79 16.44 0.14 0.09Dajab�on La Rosa N19 16,897 W71 33,990 Silt 6.47 2.98 <5.44 0.34 0.04Mao Amina 1 N19 33,440 W71 00,400 Silt loam 8.15 2.32 <5.44 0.51 0.21Mao Amina 2 N19 32,409 W70 59,467 Sandy clay loam 7.88 1.79 <5.44 0.24 0.10Mao Tierra Fría 1 N19 33,048 W70 58,129 Sandy loam 7.90 1.15 <5.44 0.44 0.16Mao Tierra Fría 2 N19 33,130 W70 58,229 Loam 8.10 1.40 5.73 0.46 0.11Mao El Charco N19 36,081 W71 04,779 Silty clay loam 8.12 2.16 5.97 0.98 0.18Montecristi La Pinta N19 39,085 W71 33,300 Loam 6.95 2.35 <5.44 0.22 0.89Montecristi Pepillo

SalcedoN19 39,964 W71 36,669 e 7.56 4.10 45.90 2.44 0.21

Montecristi Gozuela N19 40,261 W71 36,862 Clay loam 7.97 2.85 25.94 2.95 0.20Montecristi Santa María 1 N19 39,878 W71 36,243 Sandy clay loam 8.29 1.64 <5.44 0.52 0.12Montecristi Santa María 2 N19 39,270 W71 35,002 Clay loam 7.96 3.09 <5.44 0.18 0.12Montecristi Clavellina N19 39,864 W71 33,642 Clay loam 7.36 2.96 8.16 0.64 0.24Field experiment plotAzua IDIAF N18 23,608 W70 50,834 Clay loam 7.7 3.45 20 0.84 0.19

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e204

construct a similaritymatrix. For each RG, a dendrogramwas drawnfrom the similarity matrix using unweighted pair group witharithmetic mean (UPGMA) for clustering.

2.4. Identification based on 16S rRNA sequencing

The amplification of the 16S rRNA genewas performed using theprimer pair 27F (50e AGAGTTTGATCMTGGCTCAG e30) and 1492R(50e TACGGYTACCTTGTTACGACTT e30). The sequencing of the genewas carried out using the primers 518F (50e CCAGCAGCCGCGG-TAATACG e30) and 800R (50e TACCAGGGTATCTAATCC e30) (Lane,1991; Weisburg et al., 1991) by Macrogen (The Netherlands) thatyield a product around 1400 bp or more. The obtained sequenceswere assembled using the Seqman program (Madison, Wisconsin,USA) and compared with those from EzTaxon-e server, that con-tains the type strains of all described bacterial species (Kim et al.,2012). The sequences were aligned using Clustal W software(Thompson et al., 1997) and distances were calculated according toKimura's two-parameter model (Kimura, 1980). Phylogenetic treeswere inferred using the neighbour-joining (Saitou and Nei, 1987)and maximum likelihood (Rogers and Swofford, 1998) models, andthe MEGA5 package (Tamura et al., 2011) was used for all analyses.

2.5. Determination of PGP traits in vitro

2.5.1. Determination of IAA productionFor each isolate, a suspension of 3� 108 cfu ml�1 in sterile saline

solution was prepared from a bacterial mass growth obtained in aPetri dish with TSA medium. Five hundred microlitres were trans-ferred into a test tube containing 4 ml of TSB (SIGMA Catalogue No.T8907), and incubated for 3 day at 28 �C with continuous shaking.Subsequently, 500 ml of the suspension was transferred to a testtube containing 4 ml of JMM medium (O'Hara et al., 1989) sup-plemented with 0.167 g l�1 of L-Tryptophan, and incubated for4 day at 28 �Cwith continuous shaking. After the incubation period,the optical density (600 nm) was measured to estimate the bac-terial concentration, and 1.5 ml of suspension was centrifuged(2500 g) every 10 min. One millilitre of Salkowsky agent (FeCl30.01 M and perchloric acid 34.3%) was added to 1 ml of the

supernatant and incubated for 30 min in dark. The IAA productionwas estimated from the OD at 535 nm by comparison with astandard curve prepared for known concentrations of IAA, andexpressed as mg IAA ml�1.

2.5.2. Determination of ACC deaminase activityACC-deaminase activity of the isolated strains was detected on

plates with Dworkin-Foster (DF) minimal medium (Dworkin andFoster, 1958) containing 1-aminocyclopropane-1-carboxylate(ACC) as the sole source of nitrogen (Penrose and Glick, 2003).The plates were incubated at 28 �C in the dark for three days. Thegrowth of the isolates on DF agar medium amended with ACC wastaken as an indicator of potential ACC deaminase activity, as thegrowing bacteria are able to utilize ACC as N source. For quantita-tive determination of the ACC deaminase activity, the protocoldescribed by El-Tarabily (2008) was carried out. After determiningthe amount of protein and a-Ketobutyrate (a-KB), the enzyme ac-tivity was expressed as micromoles of a-KB per milligramme ofprotein per hour of the active isolates.

2.5.3. Siderophore productionSiderophore production was estimated following the Alexander

and Zuberer (1991) methodology, using the chrome azurol S (CAS)agar medium (Schwyn and Neilands, 1987). CAS agar plates wereinoculated with cells and incubated at 28 �C for 72 h. Orange halosaround the colonies on blue agar were indicative of siderophoreexcretion. The result was coded as 0 for the absence of halo, 1 for ahalo of less than 5 mm, 2 for halo between 5 and 10 mm, and 3 forhalo >10 mm.

2.5.4. Growth in N-free mediumThe ability to grow in N-free medium was analysed following

the methodology described by Wertz et al. (2012) and the inocu-lation was carried out following the procedure suggested by E.Vel�azquez (University of Salamanca, Spain, pers. comm.). For eachisolate, a suspension of 3 � 108 cfu ml�1 in sterile saline solutionwas obtained as indicated above. One hundred microlitres wereused to inoculate a test tube containing 5 ml of sterile NorrisGlucose Nitrogen-free medium and another 100 ml was used to

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 5

inoculate another test tube containing 5 ml of the same mediumsupplemented with 2 g l�1 of (NH4)2SO4 as a positive control. Un-inoculated tubes were used as control. The test tubes were staticallyincubated for 3 day at 28 �C. At the end of the period the bacterialconcentration was estimated by measuring the optical density at600 nm. The growth in the N-free medium was expressed as per-centage of the growth observed in the N-supplemented medium.

2.5.5. Phosphate solubilisationInorganic phosphate solubilisation was estimated following the

methodology of Nautiyal (1999). Each pure isolate was grown in apetri dish containing YED-P medium amended with 0.5% of tri-calcium phosphate at 30 �C for 10 days. Four different isolates weregrown per petri dish. The presence of a clearing halo around bac-terial colonies is indicative of phosphate solubilisation. The resultwas coded as 0 for no solubilisation and 1 for the presence of a halo.

2.5.6. Selection of isolates for the field experimentThe selection of the isolates for the field experiment was based

on a PCA of the in vitro PGP traits. A PCA was carried out for thenumerical results of the five PGP traits, and 50 isolates fulfilling thefollowing conditions were selected: endophytic, fully identified onthe basis of 16S rRNA gene, and non-pathogenic for humans orplants. The data matrix was standardised by subtracting the meanvalue of each trait and dividing the result by the standard deviation.For the PCA, a similarity matrix among traits was calculated basedon the Pearson productemoment correlation. The PC wereextracted and then rotated using the varimax method. A biplot wasdrawn for the two first PC for the 50 isolates and the five PGP traits.All the analyses were carried out by means of the software SPSSStatistics (IBM Corp. Released, 2010).

2.6. Field experiment

The field experiment was located in the province of Azua, in theDominican Republic (18� 230 3800 N; 70� 500 4700 W). The chemicalcharacteristics of the soil are shown in Table 1. The soil wasmediumin phosphorous and magnesium content, and high in potassium.The farm had been cultivated with banana trees 20 years previ-ously, and from then until the start of the experiment it wascultivated with a horticultural rotation of tomato, pumpkin, cour-gette, eggplant, peppers, cucumber, and guandul (Cajanus cajan L.).In the six years preceding the start of the experiment it was culti-vated using an organic system.

The statistical pattern of the experiment was a randomisedcomplete block with three replicates. The experimental plot was72 m2, with three rows 3 m apart and a space between plants of2 m, for 12 plants per plot. Banana plants (Musa AAA cv. 'DwarfCavendish') were purchased from the Instituto de Innovaci�on enBiotecnología e Industria (IIBI) (Santo Domingo, Dominican Repub-lic). They were produced in vitro, and grown in individual pots(50 ml) with a peat based substratum. When purchased, the plantshad finished the acclimation process, and they had 4e5 leaves. Thetreatment to be tested was the inoculation (five different isolates)plus the non-inoculated control, which received the same quantityof sterilized culture medium than the treatments. The inoculantswere obtained after five days of incubation of the correspondingisolate in TSB at 28 �C to a final concentration of 6 � 108 cfu ml�1.The inoculation was carried out with 5 ml of the correspondingbacterial suspension, by injection of the inoculant in the substra-tum 24 h before transplanting. Transplanting was carried out onJune 6th 2012.

The crop was fertilised with 5000 kg per ha of vermicompostfrom cattle manure (N 3.1%, P 1.7%, K 1.6%) one month after trans-planting, and drip irrigated to keep the soil humidity near to the

field capacity. The crop was kept free from weeds manually.Flowering started at the beginning of January 2013, and the routinemanagement practices were carried out (deleafing which consistsof the removal of the leaves interfering with the bunch develop-ment, and cutting off male buds). Cropping time started in mid-March 2013.

In order to estimate the effect of the treatments on the crop, thetwo central plants in each plot were subjected to measurements ofthe typical parameters for an agronomic study using a banana crop.At regular time intervals during the vegetative growth the numberof developed leaves was recorded, as was the first necrotic leaf as aconsequence of endemic black sigatoka (Mycosphaerella fijiensisMorelet) infection. In the same plants, at flowering time, the plantheight and the diameter of the pseudostem were also scored. Atcropping time, the number of leaves, the yield and the yield com-ponents (number of hands per bunch, number of fruits per bunch,and average weight per fruit) were also measured. Analysis ofvariance appropriate to randomised complete block design wasperformed using the software SPSS Statistics (IBM Corp. Released,2010).

3. Results

3.1. Number of isolates

From the 261 isolates, 185 were endophytic and the rest rhizo-spheric. Eighty-eight were isolated in the province of Azua (fivefarms), 41 in Dajab�on (three farms), 61 inMao (five farms) and 72 inMonte Cristi (six farms) (Table 2 and Table S2).

3.2. Amplified 16S rRNA restriction analysis (amplified ribosomalDNA restriction analysis, ARDRA)

We used five different RE to obtain the ARDRA profiles for the261 strains isolated and each RE produced a different number ofband patterns in this collection (Fig. 2). The results were analysed intwo different ways. Firstly, the band patterns were converted into apresence/absencematrix (matrix not shown) for all the isolates andthe five RE. This was in order to calculate the Shannon index ofdiversity. In this way we could assess the biodiversity of the cul-turable bacteria associated with banana roots in the four differentanalysed provinces of the Dominican Republic. Secondly, for eachisolate, we recorded the model for each RE. Combining the modelsfor each RE in the 261 isolates, we found 40 restriction groups (RG)(Table 3), and we assigned each isolate to a RG. Since this groupingof the isolates was based on the 16S rRNA gene, it was used to selectisolates for further detailed taxonomic study, consisting of thecomplete sequencing of this gene in some representative isolates.

3.3. Biodiversity of the bacteria associated with the roots of bananatrees, estimated from ARDRA analysis

The H index was calculated for 218 isolates from 15 farms (fivefrom Azua, two from Dajab�on, three from Mao and five fromMontecristi), out of the 261 total isolates from 18 farms. The otherthree farms were not included because of the lack of soil analysis.The H index showed values ranging from 0.19 in the farm Clavellina(Montecristi province) to 0.34 in the farm Amina 1 (Mao province)(Table 4). For each province, we calculated the average value of theH index for all the farms in the province, and we also compared theaverage values of the four provinces by ANOVA, obtaining an Fvalue of 0.565 and a p value of 0.65. This indicates that there was nosignificant difference between the provinces in terms of theaverage biodiversity of the bacterial populations associated withthe roots of banana trees.

Table 2Number of endophytic and rhizospheric bacterial isolates from each farm.

Province Farm Code given to the bacteria isolated Number of isolates

Endophytic Rizospheric Total

Azua IDIAF RD_AZIDI 12 8 20Azua Proyecto 2C RD_AZPRC 16 10 26Azua Proyecto Isura RD_AZPIS 9 3 12Azua Pueblo Viejo RD_AZPVI 12 3 15Azua Los Tramojos RD_AZLTR 11 4 15Dajab�on Carrizal RD_DACAR 5 5 10Dajab�on Rabisal RD_DARAB 14 5 19Dajab�on La Rosa RD_DAROS 9 3 12Mao Amina 1 RD_MAAMIA 19 6 25Mao Amina 2 RD_MAAMIB 5 3 8Mao Tierra Fría 1 RD_MATFA 9 2 11Mao Tierra Fría 2 RD_MATFB 4 2 6Mao El Charco RD_MAELC 7 3 10Montecristi La Pinta RD_MOLAP 6 5 11Montecristi Pepillo Salcedo RD_MOPEP 12 6 18Montecristi Gozuela RD_MOGOZ 10 2 12Montecristi Santa María 1 RD_MOSAA 9 5 14Montecristi Santa María 2 RD_MOSAB 4 0 4Montecristi Clavellina RD_MOCLA 12 1 13

Total 185 76 261

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e206

Moreover, the global value of the H index for the collection ofbacteria isolated in the 15 mentioned farms was 0.325, and theaverage value of the H index within the farms (intra-farm averagevalue of the H index) was 0.280, thus it can be deduced that most ofthe global biodiversity (86%) is intra-population diversity (intra-farms) and only 14% is among farms (Table 4).

In order to analyse the influence of the soil properties on thebacterial diversity, we carried out a PCA with the soil parametersand the H index. The first two components accounted for 71.3% ofthe total variability and the first three for 84.5% (Table 5). The pa-rameters with the highest influence in the first principal compo-nent were the lime content, and the contents of Fe, Ca, and K. Forthe second principal component the contents of organic matter, Ca,Mg and Fe had the greatest influence (Table 6). The first componentdistinguished between the farms in the Azua province on the onehand and those in Montecristi and Dajab�on provinces on the other,the main difference being in the lime content (high Azua and low inMontecristi and Dajab�on) (Fig. 3). The second component distin-guished between the farm of Carrizal in Dajab�on on the one handand the farms in Mao, Montecristi and Azua on the other. The soil atCarrizal had a very low Ca content, and was high in Fe and organicmatter, with the other farm from the province of Dajab�on (La Rosa)in an intermediate position. In the biplot representing in the firsttwo PC the farms, the soil parameters and the H index (Fig. 3), thislast was plotted very near to the coordinate origin. This is aconsequence of the small differences in biodiversity betweenfarms, and reveals the small influence of the soil properties on theH value.

3.4. Identification of 114 bacterial strains isolated from therhizosphere of banana trees, based on 16S rRNA gene sequences

The entire collection of 261 isolateswas divided into 40 differentRG (Table 3 and Table S2). In order to establish a correspondencebetween the RG and their taxonomic identity, several representa-tives from each RG were identified by 16S rRNA sequencing. Inorder to select the isolates for sequencing, a RAPD analysis with theprimer M13 was carried out within each RG, and a dendrogramobtained for each RG. The dendrogram was cut at the similaritylevel of 0.30 and one representative of each group or branch wassequenced (Fig. S1 shows as an example the dendrogram obtained

for the RAPD analysis of the RG IV with 33 isolates, in which 11 ofthem were selected). After the RAPD analysis 114 isolates wereselected (Table 3 and Table S2) and the complete 16S rRNA genesequences were obtained for each. The GenBank accession numbers(Benson et al., 2005) are given in Table S2.

The identification of the 114 isolates was based on the sequencesimilarity of the 16S rRNA gene with respect to those of the typestrains deposited in the EzTaxon database and the closest relatedspecies according this database were used for the construction ofthe phylogenetic trees. We obtained four phylogenetic trees(Figs. 4e7), each one corresponding to a different bacterial group:b-proteobacteria, gram positive, g-proteobacteria, and a-proteo-bacteria, respectively. For the construction of the phylogenetictrees, we used the 16S rRNA gene sequences of the isolates plus thesequence of the type strain closest to each isolate, as well as anoutgroup.

The general criterion used to identify the isolates at species levelwas the existence of a similarity of 99.5% or higher with respect to atype strain in the EzTaxon database. The 114 isolates belonged to 20different genera. Sixty five isolates were identified at the specieslevel, and belonged to 34 species from 16 genera (Table 3 andFigs. 4e7). Twenty-nine strains showed a similarity level lowerthan 99.5% with respect to any type strain in the database, andtherefore could represent new species, further research based onthe sequencing of other genes would be needed to clarify theiridentity. The remaining 20 strains belonged to groups in whichidentification at species level is not possible on the basis of the 16SrRNA gene sequences because the sequence is identical amongseveral species. This is the case for the “Bacillus cereus group”,including the species B. cereus, Bacillus thuringiensis and Bacillusanthracis, which contained 16 strains isolated in this study, and forthe “Bacillus pumillus group”, including the species Bacillus safensis,Bacillus altitudinis and Bacillus aerophilus, which contained fourstrains isolated in this study (Fig. 5 and Table 3).

3.5. In vitro assessment of PGP traits of the isolates

The 261 isolates were analysed for five in vitro plant growthpromoting traits: IAA production; ACC deaminase activity; side-rophore production; phosphate solubilisation; and the ability togrow in the absence of a source of combined nitrogen. The values

Fig. 2. Band patterns obtained after the restriction analysis of the 16S rRNA gene with the five restriction enzymes indicated.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 7

obtained for the 261 isolates are given in Table S2. Each trait variedwidely among the isolates. In order to aid the interpretation of thedistribution of this variation, we defined quartiles for each trait.

IAA production ranged from 0.66 to 17.94 mgml�1. The isolates inthe last three quartiles, produced between 0.66 and 3.24 mg ml�1 ofIAA, and those in the first quartile produced 3.24e17.94 mg ml�1. Of

Table 3Taxonomic identification of the bacteria isolated. The restriction groups (RG) are a combination of the models generated with ARDRA using five different restriction enzymes(RE) (see Fig. 2 for the description of the model for each RE). The selection of the 114 isolates for complete 16S rRNA gene sequencing was based on the RAPD patterns (see textfor details). The taxonomic identification is based on the similarity of the complete 16S rRNA sequence with the type strains deposited in the database EzTaxon, and theconstruction of the subsequent phylogenetic tree. Sixty-five isolates were identified at the species level because the similarity with a type strain was higher than 99.5%. Theremaining 49 isolates were only identified at the genus level (see text for details).

Model for each restriction enzyme Restrictiongroup (RG)

Total number ofisolates in each RG

No. of isolates selectedfor the complete 16S rRNAsequencing in each RG(based on RAPD)

Genera includedin the RG

Species identified. The numbercorresponds to the number ofisolates belonging to a givenspecies

BanI EcoRI HinfI Sau3AI StuI

A A A A A I 4 2 Pseudomonas P. otitidis (1)A A A A D II 1 1 Pseudomonas P. aeruginosa (1)A A I E A III 1 1 PseudomonasA A I E C IV 33 11 Pseudomonas P. mosselii (1)

P. plecoglossicida (5)P. taiwanensis (4)

A B A K D V 4 2 Acinetobacter A. pittii (2)A B B B A VI 3 3 Enterobacter E. cancerogenus (1)

Pantoea P. dispersa (1)A B B B B VII 13 4 Enterobacter Leclercia E. asburiae (1)A B B B E VIII 1 1 Serratia S. nematodiphila (1)A B D D B IX 1 1 AchromobacterA B F K D X 6 2 AcinetobacterA B H B B XI 1 1 Citrobacter C. youngae (1)B A A E C XII 1 1 Pseudomonas P. koreensis (1)B A D I A XIII 3 3 Brevundimonas B. olei (3)B B A I A XIV 1 1 Bacillus B. kochii (1)B B B B B XV 3 3 Enterobacter E. cloacae subsp. dissolvens (1)B B C D B XVI 1 1 Comamonas C. aquatica (1)B B D C A XVII 80 23 Bacillus B. marisflavi (2)

B. aryabhattai (2)B B D H A XVIII 4 2 Lysinibacillus L. xylanilyticus (2)B B D M A XIX 1 1 LysinibacillusB B E L B XX 3 2 PantoeaB B J B A XXI 4 2 Bacillus B. amyloliquefaciens subsp.

plantarum (2)B B J B B XXII 1 1 CitrobacterB B J C A XXIII 1 1 BacillusB B J D A XXIV 5 2 Comamonas C. testosteroni (2)B B J D D XXV 1 1 BurkholderiaB B J F A XXVI 23 10 Bacillus B. siamensis (2)

B. subtilis subsp. inaquosorum (3)B B J G A XXVII 1 1 BacillusB B J H A XXVIII 1 1 Bacillus B. circulans (1)B B J I A XXIX 7 4 BacillusB B K D D XXX 1 1 Burkholderia B. diffusa (1)C B A G B XXXI 13 4 Luteimonas;

PseudomonasP. beteli (3)

C B A K B XXXII 15 4 Stenotrophomonas S. maltophilia (4)C B J J D XXXIII 1 1 Cupriavidus C. nec�ator (1)D A A J D XXXIV 3 2 Pseudochrobactrum P. saccharolyticum (2)D A J J D XXXV 3 3 Pseudochrobactrum P. asaccharolyticum (3)D B A J A XXXVI 5 2 Ochrobactrum O. pseudogrignonense (2)D B D K D XXXVII 1 1 Ochrobactrum O. intermedium (1)D B G C D XXXVIII 6 5 Rhizobium R. radiobacter (5)D B G J D XXIX 3 1 Rhizobium R. massiliae (1)E A A G D XL 1 1 PseudoxanthomonasTotal number of isolates 261 114 e 65

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e208

these only five isolates produced more than 10 mg ml�1 andbelonged to the species Pseudomonas otitidis, Pseudochrobactrumasaccharolyticum, Pseudomonas saccarolyticum and Enterobactercancerogenus.

ACC deaminase activity was estimated by the production of a-Ketobutyrate from 1-aminocyclopropane-1-carboxylic acid, with217 isolates showing some degree of a-KB production, whichranged from 0.01 to 165.69 mM a-KB mg of protein�1 h�1. The iso-lates in the last quartile produced from 0.01 to 1.38 mM a-KB mg ofprotein�1 h�1, and the isolates in the third and second quartilesfrom 1.38 to 13.23 mM a-KB mg of protein�1 h�1. On the other hand,the isolates in the first quartile showed a range of a-KB productionfrom 13.23 to 165.69 mM mg of protein�1 h�1, although only sevenof them exceeded 100 mM mg of protein�1 h�1. They were

Rhizobium radiobacter; Enterobacter absuriae; three Bacillus sp.phylogenetically close to B. sonorensis, B. aerophilus andB. thuringiensis respectively; Pseudomonas taiwanensis and Serratianematodiphila.

The isolates with a significant growth rate in N-free mediumbelonged to the genus Bacillus: B. circulans, B. kochii, B. cereus group,and a Bacillus spp. phylogenetically close to B. gottheilii. Also onePseudomonas beteli strain and a Pseudoxanthomonas spp. straingrew in N-free medium.

One hundred and forty-two isolates produced an orange halo inthe medium indicative of siderophore production, belonging to thegenus Pseudomonas, Bacillus, Burkholderia, Enterobacter, Leclercia,Serratia, Brevundimonas, Pantoea, Stenotrophomonas, Cupriavidus,Pseudochrobactrum, Ochrobactrum, Comamonas and Rhizobium. The

Table 4Shannon Diversity Index (H) for the bacteria isolated in the indicatedfarms, based on the 0/1 matrix from the ARDRA.

Farm H index

Azua-IDIAF 0.2869Azua-Proyecto 2C 0.3297Azua-Proyecto Isura 0.2480Azua-Pueblo Viejo 0.3219Azua-Los Tramojos 0.2518Dajab�on-Carrizal 0.2885Dajab�on-La Rosa 0.3184Mao-Amina 1 0.3360Mao-Tierra Fría 1 0.2369Mao-El Charco 0.2855Montecristi-La Pinta 0.3076Montecristi-Pepillo Salcedo 0.2383Montecristi-Gozuela 0.2425Montecristi-Santa María 1 0.3143Montecristi-Clavellina 0.1909Intra-farm average value 0.2798Global value 0.3253

Table 5Eigenvalues, percentages and cumulative variance corresponding to the Principal CShannon index of diversity (H) based on the 0/1 matrix from the ARDRA, evaluated in

Principal component Eigenvalue

1 0.3822 0.1553 0.0994 0.0675 0.0186 0.0137 0.0108 0.0049 0.00310 0.00111 012 013 014 015 016 017 018 019 0

Table 6Eigenvector values for the four first Principal Components of the 18 physic-chemical soil pthe bacteria were isolated.

Character P.C. 1

H index �0.0001Silt % 0.0908Clay % 0.0215pH 0.0323Lime % 0.4528OM % �0.0490Total N % �0.0049Ratio of organic carbon to total nitrogen �0.0248P Olsen mg/kg 0.0187Ca cmol(þ)/kg 0.1225K cmol(þ)/kg 0.1038Mg cmol(þ)/kg �0.0086Na cmol(þ)/kg 0.0671Cation Exchange Capacity cmol(þ)/kg 0.0052Cu mg/kg 0.0884Fe mg/kg �0.1354Zn mg/kg �0.0605B mg/kg 0.0446Electric Conductivity dS/m 0.0002

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 9

isolates that produced a larger halo belonged to the genera Pseu-domonas, Bacillus and Burkholderia.

Forty-four isolates solubilised tricalcium phosphate in vitro,mainly belonging to the genus Pseudomonas (P.plecoglossicida,P. taiwanensis, P.resinovorans, P. otitidis, Pseudomonas mosselii), andto the species Acinetobacter pittii. Also most of the beta proteo-bacteria isolated, plus two strains of Enterobacter, three of Pantoea,one of Citrobacter, and one of Pseudochrobactrum were able to sol-ubilise tricalcium phosphate.

For each province, we calculated the average value of each of thePGP activities analysed in vitro, comparing the average values forthe four provinces with an ANOVA, obtaining the following F values(p value in parenthesis): IAA production 2.246 (0.083); productionof a-Ketobutyrate from 1-aminocyclopropane-1-carboxylic acid0.982 (0.402); growth in N-free medium 1.467 (0.224); siderophoreproduction 0.961 (0.412); inorganic phosphate solubilisation 5.571(0.001). Therefore, there were no significant differences in theaverage values of the PGP traits between the provinces for four of

omponent Analysis (PCA) of the 18 physico-chemical soil parameters, and thethe 11 farms where the bacteria were isolated.

Percentage of variance Cumulative percentage

50.727 50.72720.603 71.33013.131 84.4618.955 93.4162.383 95.7991.790 97.5891.370 98.9590.588 99.5460.372 99.9180.082 100000000000

arameters plus the Shannon index of diversity (H), evaluated in the 11 farms where

P.C. 2 P.C. 3 P.C. 4

�0.0034 �0.0068 0.00410.0384 0.0186 0.15270.0065 0.1263 0.03320.0410 �0.0048 �0.01430.0472 0.0032 �0.0028

�0.1239 0.0825 0.0571�0.0077 0.0109 0.0062�0.0989 0.0254 0.02550.0519 0.1794 �0.02790.3250 �0.0171 �0.00120.0478 0.1246 �0.02860.2375 0.0862 0.02400.0236 0.0416 �0.00250.0445 0.0057 0.02020.0225 0.0376 0.1157

�0.2020 �0.0419 0.18370.0215 0.0809 0.08800.0236 0.0344 �0.00860.0445 0.0152 0.060

Fig. 3. Biplot obtained after the projection on the two first Principal Components of the 18 physico-chemical soil parameters plus the Shannon index of diversity (H), and the 11farms where the bacteria were isolated.

Fig. 4. Neighbour-joining phylogenetic tree based on 16S rRNA sequences, of the isolated b-proteobacteria strains, and their closest relatives (type strains which GeneBankaccession number are given in parenthesis). Bootstrap values over 50 expressed as percentages of 1000 replications are shown at the branch points. Asterisks denote nodes thatwere also recovered using the maximum likelihood method. Bar 0.02 represents substitutions per nucleotide position. RG means Restriction Group. The codes of the isolated strainsgive information about their origin according to Table 2. For more detailed information about each strain see Table S2.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e2010

the five traits evaluated, in accordance with the lack of differencesin the average genetic biodiversity between provinces. However,there was an exception in the case of phosphate solubilisation. Inthis case, the isolates from Azua province showed higher tricalciumphosphate solubilising activity than those from Dajab�on and Maoprovinces.

3.6. Field trials

The criteria used to select the isolates for the field trials were theresults obtained from the assessment of the five in vitro PGP traits.Only the endophytic bacteria from isolates fully identified bymeansof 16S rRNA sequencing were considered for this trial. Endophyticbacteria likely have better performance from the agronomic

viewpoint than the rhizospheric ones, as they are protected by theroot tissues. From them we eliminated human pathogens: 14 iso-lates closely related with B. anthracis and B. cereus, four isolatesclosely related to Burkholderia metallica, P. otitidis and Steno-trophomonas maltophilia (the last one because it has been describedas an opportunistic pathogen resistant to several antibiotics).Moreover, we also eliminated plant pathogens, five isolates closelyrelated to R. radiobacter. Finally we eliminated five isolates close toB. thuringiensis because it is considered a biocontroller; this can be aproblem for the registration of a bacterial strain as a plant probiotic.Consequently the collection was reduced to 50 isolates. With thevalues obtained for the PGP traits in the 50 pre-selected isolates, aPCA was carried out. Fig. 8 presents the biplot obtained after thePCA showing the five traits and the 50 isolates. The first two

Fig. 5. Neighbour-joining phylogenetic tree based on 16S rRNA sequences, of the isolated gram-positive strains, and their closest relatives (type strains which GeneBank accessionnumber are given in parenthesis). Bootstrap values over 50 expressed as percentages of 1000 replications are shown at the branch points. Asterisks denote nodes that were alsorecovered using the maximum likelihood method. Bar 0.02 represents substitutions per nucleotide position. RG means Restriction Group. The codes of the isolated strains giveinformation about their origin according to Table 2. For more detailed information about each strain see Table S2.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 11

principal components accounted for 52% of the total variability andwith the first three the variability was 75% (data not shown). In thefirst component, the traits with the most influence on the differ-entiation of the isolates were siderophore production and growth

in N-free medium, both being negatively correlated (Fig. 8). In thesecond component, the most important parameters were IAAproduction and ACC deaminase activity, also negatively correlated.Tricalcium phosphate solubilisation was positively correlated with

Fig. 6. Neighbour-joining phylogenetic tree based on 16S rRNA sequences, of the isolated g-proteobacteria strains, and their closest relatives (type strains which GeneBankaccession number are given in parenthesis). Bootstrap values over 50 expressed as percentages of 1000 replications are shown at the branch points. Asterisks denote nodes thatwere also recovered using the maximum likelihood method. Bar 0.02 represents substitutions per nucleotide position. RG means Restriction Group. The codes of the isolated strainsgive information about their origin according to Table 2. For more detailed information about each strain see Table S2.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e2012

Fig. 7. Neighbour-joining phylogenetic tree based on 16S rRNA sequences, of the isolated a-proteobacteria strains, and their closest relatives (type strains which GeneBankaccession number are given in parenthesis). Bootstrap values over 50 expressed as percentages of 1000 replications are shown at the branch points. Asterisks denote nodes thatwere also recovered using the maximum likelihood method. Bar 0.01 represents substitutions per nucleotide position. RG means Restriction Group. The codes of the isolated strainsgive information about their origin according to Table 2. For more detailed information about each strain see Table S2.

Fig. 8. Biplot in the first two Principal Components showing the five plant growth promoting traits analysed in 50 isolated strains (see text for isolates selection). The isolates in boldcharacters were used in the field trail. The codes of the isolated strains give information about their origin according to Table 2. The prefix RD_ of the strain codes has beenintentionally suppressed to save space. For more detailed information about each strain see Table S2.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 13

siderophore production, but had less influence on the separation ofthe isolates. The selection of bacteria was based on the biplot(Fig. 8). We did not consider growth in N-free medium, as it is not a

standardised method for nitrogen fixation estimation, instead weprioritised other traits. Firstly we considered siderophore produc-tion, as it is related to competition with pathogens for Fe ions, but

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e2014

without a direct biocidal effect, and is also related to induced sys-temic resistance (ISR) mechanisms (García Guti�errez et al., 2012;Djavaheri et al., 2012). Other important traits were IAA produc-tion and ACC deaminase activity, but unfortunately they werenegatively correlated, therefore it was not possible to select a singleisolate with a high expression of both characters. Therefore in thepre-selection of isolates, we chose two in the most positive sectionof the first component (AZPVI_05 and MAAMIA_05 P. plecoglossi-cida), which showed the highest expression of siderophore pro-duction, but very low expression of the other traits; plus other threeisolates with the highest possible expression of siderophore pro-duction, but also with IAA production (MOPEP_07 P. mosselii) andACC deaminase activity (DARAB_12 andMAAMIB_01 P. taiwanensis)(Fig. 8).

Banana plants were inoculated with the pre-selected bacteriaone day before transplanting to the field (June 6th, 2012). Duringvegetative growth the number of developed leaves did not showsignificant differences between any of the sampling days (Table 7).It is worth mentioning that on the third sampling day (November30th), there was a dramatic decrease in the number of leaves in allthe treatments, as a consequence of black sigatoka infection. Thisdisease is present in all the banana tree crops throughout thecountry, and therefore it can be considered as another productionfactor. The incidence of black sigatoka showed significant differ-ences between treatments, on four of the five sampling dates. Thiswas observed because the position of the first leaf with necrosissymptoms as a consequence of the black sigatoka attack showedsignificant differences between treatments on the sampling datesOctober 30th (p < 0.05), November 15th (p < 0.05), December 15th(p < 0.01) and December 30th (p < 0.05) (Fig. 9). The isolateRD_MAAMIA_05 (P. plecoglossicida) delayed the black sigatokaattack by between one and four leaves, depending on the samplingdate, with such differences being statistically significant withrespect to the control on the November 15th and December 15thsampling days (Fig. 9). Moreover, the results obtained with theisolate RD_DARAB_12 (P. taiwanensis) did not significantly differfrom the results with RD_MAAMIA_05, which also indicated aneffect against black sigatoka. However the effect of RD_DARAB_12was smaller because the difference with the control was only sta-tistically significant in the sampling done on December 30th.

Neither plant height nor the diameter of the pseudostem atflowering time, nor the number of developed leaves at croppingtime showed significant differences between treatments (Table 8).

Regarding yield components (Table 9), the treatments resultedin significant differences (p < 0.01) only in the average weight perfruit, but this also resulted in significant differences (p< 0.05) in theyield per plant. The average fruit weight, and the yield per plant(weight of the bunch) were higher in all of the treatments than inthe control. However the differences with the control were onlystatistically significant for the isolate RD_MAAMIA_05 (P.

Table 7Average number of leaves per banana tree in the field trial during vegetative growth. Treasame letter did not significantly differ (p < 0.05); ns, not significant; D.F., degrees of free

Treatment October 30th Novemb

RD_AZPVI_05 9.8 a 10.7 aUninoculated control 10.3 a 10.8 aRD_MAAMIA_05 10.5 a 10.8 aRD_MOPEP_07 10.6 a 10.8 aRD_DARAB_12 10.6 a 11.3 aRD_MAAMIB_01 10.9 a 11.2 aANOVA (F values and sign. level) D.F.Treatment 5 1.248 ns 0.234 nsBlock 2 2.535 ns 0.483 nsTreatment * Block 10 1.756 ns 1.207 ns

plecoglossicida). In that treatment the average fruit weight was 30%greater and the average yield per plant 29% higher compared withthe control. The treatment with RD_MAAMIB_01 (P. taiwanensis)produced a yield 14.3% higher than the control, but the differencewas not statistically significant.

4. Discussion

Agriculture is concerned with the production of food in envi-ronmentally and socially sustainable ways (Godfray et al., 2010).The maximisation of the coadaptation between plants and rhizo-sphere microbial communities as an alternative to chemical agri-cultural inputs has become a subject of great interest with the aimof achieving sustainability and bio-safety in agriculture (Bhardwajet al., 2014). Therefore, a better management of the soil micro-biota and a parallel reduction in the use of agro-chemicals has beenproposed as the most adequate strategy for a more sustainableagriculture (Singh et al., 2011).

As organic farming manages without chemical inputs, organiccrops are mostly dependent on the natural microflora of the soil(Sinha et al., 2010). As the long term goal of our work is to developbacterial inoculants to be used as plant probiotics in banana in theDominican Republic, we have paid great attention to knowledgeabout the culturable microorganisms associated with the roots oforganically managed banana crops in the Dominican Republic. Thisis because on the one hand development of inoculants based onmicrobes should be based on region-specific microbial strains(Deepa et al., 2010; Mulas et al., 2013; Majeed et al., 2015), and onthe other hand, organic systems are naturally richer in soil biotathan conventional ones (M€ader et al., 2002; Hole et al., 2005;Birkhofer et al., 2008; Henneron et al., 2015). Moreover there issubstantial scientific evidence that plants, through root exudates,select the composition and total population of soil microflora(Chaparro et al., 2012), and therefore the most adequate bacterialspecies to be used as inoculants for a given crop (in this case ba-nana) are those isolated from such a crop.

4.1. Strategy to identify the culturable bacteria associated with theroots of banana trees

A comprehensive study of the soil bacteria associated with agiven crop (i.e. community structure, population index and mo-lecular phylogeny), should be accomplished by means of a meta-genomic analysis of certain genes, typically 16S rRNA (Iwai et al.,2011; Suenaga, 2012; Keshri et al., 2013). However, when the goalof thework is to develop plant probiotics based onmicroorganisms,the efforts must focus on the culturable microorganisms. In such acase a very high number of strains are frequently isolated, whichmust be identified from the taxonomic viewpoint.

Since the phylogenetic classification of the bacteria is currently

tments correspond to inoculations with the indicated strains. Values followed by thedom.

er 15th November 30th December 15th December 30th

5.5 a 8.1 a 9.8 a5.4 a 8.4 a 9.9 a6.8 a 8.8 a 10.2 a6.6 a 8.7 a 9.8 a8.2 a 8.8 a 10.7 a8.7 a 9.4 a 10.7 a

1.047 ns 0.356 ns 0.779 ns0.235 ns 6.889 ns 1.690 ns0.800 ns 0.889 ns 0.738 ns

Fig. 9. Average first necrotic leaf (counted from top to bottom) as a consequence of black sigatoka attack in the field trial, in the non-inoculated control and the different bacterialtreatments. The bars followed by the same letter did not significantly differ (p < 0.05). The codes of the isolated strains give information about their origin according to Table 2. Theprefix RD_ of the strain codes has been intentionally suppressed to save space. For more detailed information about each strain see Table S2.

Table 8Average values for the different parameters in the banana trees corresponding to the field trial, at the stages of flowering and harvesting. Treatments correspond to inoculationswith the indicated strains. Values followed by the same letter did not significantly differ (p < 0.05); ns, not significant; D.F., degrees of freedom.

Treatment Flowering time (1st January 2013) Harvest time (15th March 2013)

No. of leaves Plant heigth (cm) Pseudostem ∅ (cm) No. of leaves

Control 13.2 a 162.7 a 43.3 a 11.5 aRD_DARAB_12 15.2 a 206.0 a 44.7 a 11.2 aRD_AZPVI_05 13.7 a 201.7 a 43.3 a 12.5 aRD_MAAMIA_05 13.3 a 187.7 a 43.3 a 11.0 aRD_MAAMIB_01 13.7 a 183.3 a 41.8 a 11.7 aRD_MOPEP_07 12.2 a 184.2 a 36.0 a 11.2 aANOVA (F values and signification level) D.F.Treatment 5 2.20 ns 0.807 ns 1.072 ns 0.675 nsBlock 2 0.2 ns 0.186 ns 0.295 ns 1.500 nsTreatment * Block 10 0.643 ns 0.927 ns 1.348 ns 0.750 ns

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 15

Table 9Average values for banana yield and yield components, corresponding to the field trial. Treatments correspond to inoculations with the indicated strains. Values followed bythe same letter did not significantly differ (p < 0.05); ns, not significant; D.F., degrees of freedom.

Treatment No. of hands per bunch No. of fruits per bunch Average weight per fruit (g) Bunch weight (yield per plant) (kg)

Control 7.2 100.50 a 151.29 a 15.22RD_AZPVI_05 7.0 96.00 a 169.28 ab 16.04 abRD_DARAB_12 7.3 101.20 a 160.28 a 16.18 abRD_MOPEP_07 7.3 109.20 a 152.39 a 16.70 abRD_MAAMIB_01 7.3 103.70 a 167.34 ab 17.40 abRD_MAAMIA_05 7.5 99.20 a 196.77 b 19.62 bANOVA (F values and signif. level) D.F.Treatment 5 0.800 ns 1.771 ns 5.767** 3.324*Block 2 0.125 ns 0.626 ns 0.114 ns 0.721 nsTreatment * Block 10 1.025 ns 1.965 ns 0.489 ns 1.463 ns

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e2016

based on the 16S rRNA gene (Woese, 2000), methods based on theanalysis of this gene are the most suitable for the identification ofbacterial isolates. Nevertheless, the complete sequencing of thisgene, even with the reduction in the economic costs and effortsexperienced during the last years, is not viable in large collections.For that reason several strategies for pre-grouping of a collection,prior to gene sequencing have been proposed. The ARDRA tech-nique based on 16S rRNA-RFLP patterns was described in 1996 as amethod to cluster bacterial strains corresponding well with knownspecies (Moyer et al., 1996; Heyndrickx et al., 1996). More recently,Sklarz et al. (2009) re-evaluated the predictive power of thismethodology concluding that it can be a suitable tool for genusdifferentiation of environmental isolates, and even for speciesidentification, provided caution is taken in the type and number ofrestriction enzymes (RE) selected. This technique has previouslybeen used for the characterization of PGPR collections (Martínezet al., 2003; Souza et al., 2013; De Souza et al., 2013; Beneduziet al., 2013; Arruda et al., 2013). Although the restriction analysiscan be also applied to housekeeping genes (Martínez et al., 2003)and the nifH gene (Beneduzi et al., 2013; Souza et al., 2013), sincethe 16S rRNA gene is the only one amplified with universal primersand the only one able to allow the identification of bacterial isolatesindependent of the phylogenetic group to which they belong, thisgene is the most appropriate. In our study, this analysis produced40 RG with a different number of isolates within each group,ranging from one in several cases to 80 in the largest group.

Nevertheless, ARDRA has some limitations for analysis of bac-terial biodiversity because bacterial species with identical 16S rRNAgenes have been described. Therefore, it is necessary to use a sec-ondary technique able to detect diversity within each ARDRA groupin order to split each large group into several smaller ones. Theselected technique was RAPD, which as well as ERIC, REP, etc. isstrain dependent (Nocker et al., 2007). Several studies have used a0/1 matrix obtained after the scoring of the RAPD bands patterns tocluster several strains using the UPGMA method, with the condi-tion that the strains are phylogenetically close. In several cases, thedendrograms obtained maintained the taxonomic structure of thestrains coherent with specific or subspecific levels (Mulas et al.,2011; Díaz-Alc�antara et al., 2014). However in other casesdifferent strains from the same species are located in differentbranches of the dendrogram, as for example in the works of DeGoes et al. (2012), Bisht et al. (2013) and in the present study (i.e.Fig. S1).

As on some occasions we sequenced the 16S rRNA gene inseveral isolates per RG, it was possible to conclude that the RGobtained were in general terms genus-specific with some excep-tions. For example three of these groups included different generathat were very close from the phylogenetic viewpoint: RG VI withEnterobacter and Pantoea; RG VII with Enterobacter and Leclerciaand RG XXXI with Enterobacter and Luteimonas (Fig. 6). However,

genera including phylogenetically very diverse species weredistributed among different RG, each of them consisting of phylo-genetically close species. For example, the genus Bacillus wasdistributed in eight different groups (Fig. 5), of which RG XVIIincluded all of the isolates from the “Bacillus cereus group” and alsoB. marisflavi, both located on the same branch of the 16S rRNAphylogenetic tree (Fig. 5). RG XXIX included the “Bacillus pumilusgroup”. Other RGwere bi-specific (XXVI Bacillus subtilis and Bacillussiamensis), and finally several were mono-specific. Therefore, theARDRA techniquewith the five RE used in the present study, provedto be valid for the preliminary identification of the isolates at thegenus level, or even species level in some cases, yielding unam-biguous results.

4.2. Species associated with the rhizosphere of banana trees in theDominican Republic and their determinant factors

The plant drives the most determinant factors in the composi-tion of the bacterial populations associated with a given crop. Theplant species is the most important one (Andreote et al., 2010),although other factors such as the cultivar, the age of the plant, andother characteristics of the roots also play a role in the biodiversityand the predominant species associated with plant roots (Smallaet al., 2001; Macdonald et al., 2004). Root exudates determine themicroorganisms selected by the plant (Chaparro et al., 2012;Prashar et al., 2014). Environmental factors such as the physicaland chemical characteristics of the soil also influence the compo-sition of bacterial populations associated with a given crop (Fanget al., 2005; Ibekwe et al., 2010), but in our research we observedthat the soil features analysed had only a small influence on thebiodiversity of the isolated bacteria. This was made explicit becausethe H index of diversity was the closest trait to the coordinatesorigin in the PCA bi-plot representing the H index and the soilfactors in the different farms analysed. This indicates that thebiodiversity index was the least variable trait between the farms,highlighting that it is not influenced by the soil parameters ana-lysed, which showed high levels of variability. Other studies haveanalysed the influence of soil parameters on the biodiversity of thebacteria isolated in a given crop. Some of them arrived at the sameconclusion as this study, with independence between the H indexand the soil traits analysed (Ambrosini et al., 2012 in sunflowerHelianthus annuus L.; De Souza et al., 2013 in rice Oryza sativa L.).However, other studies have observed that some soil parametersinfluenced the diversity of the bacteria associated with a givencrop: Beneduzi et al. (2013) in sugarcane (Saccharum officinarum L.)related the pH and clay content with the H index; and Arruda et al.(2013) in corn (Zea mays L.) described a positive correlation be-tween H index, clay and organic matter. There is a general opinionin the scientific community in favour of the decisive influence of pHon the composition and richness of the soil microorganisms (Fierer

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 17

and Jackson, 2006), but in our case the variation of the pH acrossthe farms was very small, as the pH was also plotted very near tothe coordinates origin in the PCA biplot. Finally agronomic man-agement factors like irrigation, tillage or fertilisation can also in-fluence the composition of the bacterial populations (Fang et al.,2005; Ibekwe et al., 2010). However, in our research the agro-nomic management practices were fairly similar, as organic bananafarmers usually follow the recommendations of processing andcommercialisation companies.

The composition of the culture media used for the isolation ofculturable bacteria also has an unquestionable influence on theisolated taxa. For that reason caution must be employed whencomparing between the results obtained in different studies ifdifferent culture media were used. The most numerous bacterialgenera isolated in banana in our work were Bacillus (45%) followedby Pseudomonas-Luteimonas (20%), Enterobacter-Leclercia-Pantoea(8%), and Stenotrophomonas (6%). Ambrosini et al. (2012) found thatthe most numerous genera in sunflower were Enterobacter, Bur-kholderia and Klebsiela; in rice De Souza et al. (2013) cited thepredominant genera as Agrobacterium, Pseudomonas, Enterobacterand Burkholderia, and Mirzajani et al. (2013) Serratia, Pantoea,Ochrobactrum, Brevundimonas and Enterobacter. In oilseed rape(Brassica napus L.) Farina et al. (2012) isolated Agrobacterium, Bur-kholderia, Enterobacter, Klebsiella and Pseudomonas as the maingenera. For comparison, we must ignore the genus Bacillus becausethe above-mentioned works used a culture media that excludedthis genus. Enterobacter was a major genus in the five mentionedworks; Pseudomonas was a major genus only in three. Burkholderiaand Agrobacterium (Rhizobium) were major genera in other twoworks, and showed a minority representation in our work. How-ever, the main differences in bacterial populations between cropswere found in minority taxa. In total we identified 20 genera. Otherresearchers have isolated five other bacterial genera in banana treeroots that we did not isolate. In Brazil, De Souza et al. (2013) iso-lated bacteria from nine different genera, four of them in commonwith our work plus five different ones (Aneurinibacillus, Klebsiella,Micrococcus, Paenibacillus and Sporolactobacillus). Andrade et al.(2014), also in Brazil, isolated bacteria from six different generaall in common with the work of Souza et al. (2013). Martínez et al.(2003) isolated four genera of free N-fixing bacteria from Colima(Mexico), three in common with our work plus Klebsiella. Surpris-ingly, since they grow well in the media used, some genera such asAneurinibacillus, Klebsiella, Micrococcus and Paenibacillus were notisolated in this work.

We reported for the first time the in vitro mechanisms of PGPfrom P. asaccharolyticum RD_AZPRC_25 and P. saccharolyticumRD_AZIDI_13, with a high IAA production; and for Ochrobactrumpseudogrignonense RD_MAELC_10, also with a significant produc-tion of IAA. On the other hand, other isolates that were not fullyidentified, but that were phylogenetically close to the followingspecies, also showed a high expression of PGP traits, that had notbeen previously described: RD_MAAMIA_18 and RD_MAELC_05(Bacillus sp.), phylogenetically close to B. aerophilus and B. subtilissubsp. Inaquosorum, respectively had a high production of side-rophores; RD_MOLAP_05 (Bacillus sp.), phylogenetically close toB. gottheilii, and RD_MAAMIA_15 (Pseudoxanthomonas sp.), close toP. indica, were both able to grow in vitro in a N-free medium.

4.3. Development of inoculants for banana in the DominicanRepublic

From the broad biodiversity of bacteria isolated in this workfrom banana tree roots, five strains were tested in the field, and theresults are very promising for banana producers because most ofthem showed a tendency to increase the banana yield. This

tendency was statistically significant in the case of RD_MAAMIA_05(P. plecoglossicida) compared with the uninoculated control, with ayield increase of 29%. The field trial with PGPR for banana trees ofKavino et al. (2010) observed a yield increase slightly higher thanours (between 36% and 39%) on inoculation with two chitinoliticPseudomonas fluorescens strains, with the addition of chitin toimprove bacterial survival. On the other hand, Selvaraj et al. (2014)observed an increase in fruit yield under field conditions from 36.5to 45.6%, which was the consequence of the direct biocontrol of twosoil-borne diseases (Fusarium oxysporum f. sp. cubense and thenematode Helicotylenchus multicinctus) with the strain Pf1 (P.fluorescens).

Unexpectedly in our case the yield increase seems to be at leastpartially the consequence of a reduction in the incidence of the leaffungal disease black sigatoka, which is always present throughoutthe country, because the strain RD_MAAMIA_05 delayed theappearance of the first leaf with necrotic symptoms, thus increasingthe number of fully functional leaves. To strengthen this hypothesis,the strain RD_AZPVI_05 had no effect on either the fruit yield or onthe black sigatoka control. Tentatively the phenomenon of thedisease control has been assigned to induced systemic resistance(ISR) mechanisms, although in the future it will be necessary todeepen the study of the production of the enzymes related to theenhancement of plant resistance. The main criteria for the selectionof the bacteria in the field trial was siderophore production, andthis trait has been frequently related with ISR (García-Guti�errezet al., 2012; Djavaheri et al., 2012). There are no references to thecontrol of black sigatoka in banana attributed to ISR and triggeredby rhizospheric bacteria, although there is one study concerningarbuscular mycorrhiza (Oye Anda et al., 2015). On the other handFishal et al. (2010) mentioned positive preliminary results in thecontrol of the banana disease caused by F. oxysporum by ISR pro-cesses triggered by one strain of Pseudomonas sp. and one of Bur-kholderia sp. Peeran et al. (2014) worked with Colletotrichum musaein banana trees and showed evidence of an increase in the activityof defence related enzymes after inoculation with the strain FP7 (P.fluorescens) in field conditions. Moreover Harish et al. (2008, 2009)controlled the Banana Bunch Top Virus (BBTW) with endophytic P.fluorescens and Bacillus spp, also attributed to ISR processes.

The finding that the biodiversity of the bacteria isolated in thiswork from banana trees roots is independent of the soil charac-teristics (and therefore showing small differences across thecountry) is important for the development of plant probiotics basedon microorganisms for banana crops in the Dominican Republic.This could involve an adaptation range of the strains to the differentenvironmental conditions across the island. The lack of differencesin the average bacterial population diversity between provinces inthe Dominican transect of about 150 km, contrasts with the largedifferences encountered by Marasco et al. (2013) in a thousandkilometre transect across the Mediterranean basin in vineyardcrops, from northern Italy to northern Africa. The Island conditionsof the Dominican Republic, which have reduced environmentaldifferences due to the influence of the sea, could explain the lack ofdifference in bacterial diversity and work in favour of the smoothadaptation of a given autochthonous strain to the different pro-ductive regions on the island.

4.4. Conclusions

In conclusion, the biodiversity of the culturable bacteria isolatedin this study from roots of banana trees in organic farming systemsin the Dominican Republic, contains genera most frequently asso-ciated with plant roots such as Bacillus, Pseudomonas and Enter-obacter. However, we found that seven species not previouslydescribed as PGP bacteria showed high expression of PGP traits

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e2018

in vitro and consequently they are candidates to be tested in fieldconditions, for the eventual development of PGPR based inoculantsfor banana crops. It may be possible to describe new species amongthe large number of isolates that could not be fully identified basedon comparison with the EzTaxon database and the subsequentphylogenetic reconstruction. In the field experiment the endo-phytic isolate RD_MAAMIA_05 belonging to P. plecoglossicidasignificantly improved the fruit yield and reduced the incidence ofblack sigatoka. The second best strain, RD_MAAMIB_01 (P. taiwa-nensis) also showed PGPR effects in the field experiment, althoughwith lower efficiency than the first one. The small variations inbacterial biodiversity among the banana producing regions in theDominican Republic indicates the independence of bacterialbiodiversity from the chemical and physical properties of the soil,and opens the door to the possible efficient adaptation of theselected strains to the whole region studied.

Acknowledgements

This work has been financially supported by the Spanish Min-istry of Foreign Affairs and Cooperation (Projects PCI-AECID A/023132/09, PCI-AECID A/030020/10 and PCI-AECID A1/035364/11).I.-E. M. was granted by the Spanish Ministry of Foreign Affairs andCooperation (MAEC-AECID grant 2010e2014). We thank Dr.Encarna Vel�azquez (University of Salamanca, Spain) for criticalscientific reading of the paper. The manuscript has been profes-sionally proofread by PRS.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.soilbio.2016.04.013.

References

Adesemoye, A.O., Torbert, H.A., Kloepper, J.W., 2009. Plant growthpromoting rhi-zobacteria allow reduced application rates of chemical fertilizers. Microb. Ecol.58, 921e929. http://dx.doi.org/10.1007/s00248-009-9531-y.

Alexander, D.B., Zuberer, D.A., 1991. Use of chrome azurol S reagents to evaluatesiderophore production by rhizosphere bacteria. Biol. Fertil. Soils 12, 39e45.http://dx.doi.org/10.1007/BF00369386.

�Alvarez-Martínez, E.R., Valverde, A., Ramírez-Bahena, M.H., García-Fraile, P.,Tejedor, C., Mateos, P.F., Santillana, N., Zú~niga, D., Peix, A., Vel�azquez, E., 2009.The analysis of core and symbiotic genes of rhizobia nodulating Vicia fromdifferent continents reveals their common phylogenetic origin and suggests thedistribution of Rhizobium leguminosarum strains together with Vicia seeds.Arch. Microbiol. 191, 659e668. http://dx.doi.org/10.1007/s00203-009-0495-6.

Ambrosini, A., Beneduzi, A., Stefanski, T., Pinheiro, F., Vargas, L., Passaglia, L.P., 2012.Screening of plant growth promoting rhizobacteria isolated from sunflower(Helianthus annuus L.). Plant Soil 356, 245e264. http://dx.doi.org/10.1007/s11104-011-1079-1.

Oye Anda, C.C., de Boulois, H.D., Declerck, S., 2015. The arbuscular mycorrhizafungus Rhizophagus Irregularis MUCL 41833 decreases disease severity of BlackSigatoka on banana c.v. Grande name, under in vitro culture conditions. Fruits70, 37e46. http://dx.doi.org/10.1051/fruits/2014041.

Andrade, L.F., de Souza, G.L., Nietsche, S., Xavier, A.A., Costa, M.R., Cardoso, A.,Toledo, M.C., Pereira, D.F., 2014. Analysis of the abilities of endophytic bacteriaassociated with banana tree roots to promote plant growth. J. Microbiol. 52,27e34. http://dx.doi.org/10.1007/s12275-014-3019-2.

Andreote, F.D., Nunes da Rocha, U., Luiz Araújo, W., Azevedo, J.L., van Overbeek, L.S.,2010. Effect of bacterial inoculation, plant genotype and developmental stageon root-associated and endophytic bacterial communities in potato (Solanumtuberosum). Antonie Leeuwenhoek 97, 389e399. http://dx.doi.org/10.1007/s10482-010-9421-9.

Arias, P., Dankers, C., Liu, P., Pilkauskas, P., 2004. La economía mundial del banano1985e2002.

Arruda, L., Beneduzi, A., Martins, A., Lisboa, B., Lopes, C., Bertolo, F., Passaglia, L.M.P.,Vargas, L.K., 2013. Screening of rhizobacteria isolated from maize (Zea mays L.)in Rio Grande do Sul State (South Brazil) and analysis of their potential toimprove plant growth. Appl. Soil Ecol. 63, 15e22. http://dx.doi.org/10.1016/j.apsoil.2012.09.001.

Babalola, O.O., 2010. Beneficial bacteria of agricultural importance. Biotechnol. Lett.32, 1559e1570. http://dx.doi.org/10.1007/s10529-010-0347-0.

Beneduzi, A., Moreira, F., Costa, P.B., Vargas, L.K., Lisboa, B.B., Favreto, R., Baldani, J.I.,

Passaglia, L.M.P., 2013. Diversity and plant growth promoting evaluation abili-ties of bacteria isolated from sugarcane cultivated in the South of Brazil. Appl.soil Ecol. 63, 94e104. http://dx.doi.org/10.1016/j.apsoil.2012.08.010.

Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., Wheeler, D.L., 2005. Gen-Bank. Nucleic Acids Res. 1, D34eD38. http://dx.doi.org/10.1093/nar/gki063.

Bhardwaj, D., Ansari, M.W., Sahoo, R.K., Tuteja, N., 2014. Biofertilizers function askey player in sustainable agriculture by improving soil fertility, plant toleranceand crop productivity. Microb. Cell Factories 13, 66. http://dx.doi.org/10.1186/1475-2859-13-66.

Birkhofer, K., Bezemer, T.M., Bloem, J., et al., 2008. Long-term organic farming fos-ters below and aboveground biota: implications for soil quality, biologicalcontrol and productivity. Soil Biol. Biochem. 40, 2297e2308. http://dx.doi.org/10.1016/j.soilbio.2008.05.007.

Bisht, S.C., Mishra, P.K., Joshi, G.K., 2013. Genetic and functional diversity amongroot-associated psychrotrophic Pseudomonad's isolated from the Himalayanplants. Arch. Microbiol. 195, 605e615. http://dx.doi.org/10.1007/s00203-013-0908-4.

CEI-RD, 2013. Perfil Econ�omico del Banano. Centro de Exportaci�on e Inversi�on de laRepública Dominicana, Santo Domingo. República Dominicana.

Chahboune, R., Barrijal, S., Moreno, S., Bedmar, E.J., 2011. Characterization of Bra-dyrhizobium species isolated from root nodules of Cytisus villosus grown inMorocco. Syst. Appl. Microbiol. 34, 440e445. http://dx.doi.org/10.1016/j.syapm.2011.04.008.

Chaparro, J.M., Sheflin, A.M., Manter, D.K., Vivanco, J.M., 2012. Manipulating the soilmicrobiome to increase soil health and plant fertility. Biol. Fertil. Soils 48,489e499. http://dx.doi.org/10.1007/s00374-012-0691-4.

Cummings, S.P., 2009. The application of plant growth promoting rhizobacteria(PGPR) in low input and organic cultivation of graminaceous crops; potentialand problems. Environ. Biotechnol. 5, 43e50.

De Goes, G.P.K.C., de Castro Fisher, M.L., Cattelan, A.J., Nogueira, M.A., Portela deCarvalho, C.G., Martinez de Oliveira, A.L., 2012. Biochemical and molecularcharacterization of high population density bacteria isolated from sunflower.J. Microbiol. Biotechnol. 22, 437e447. http://dx.doi.org/10.4014/jmb.1109.09007.

De la Torre-Guti�errez, L., Chel-Guerrero, L.A., Betancur-Ancona, D., 2008. Functionalproperties of square banana (Musa balbisiana) starch. Food Chem. 106,1138e1144. http://dx.doi.org/10.1016/j.foodchem.2007.07.044.

De Souza, R., Beneduzi, A., Ambrosini, A., Costa, P., Meyer, J., Vargas, L.K.,Schoenfeld, R., Passaglia, L.P., 2013. The effect of plant growth-promoting rhi-zobacteria on the growth of rice (Oryza sativa L.) cropped in Southern Brazilianfields. Plant Soil 366, 585e603. http://dx.doi.org/10.1007/s11104-012-1430-1.

Deepa, C.K., Dastager, S.G., Pandey, A., 2010. Isolation and characterization of plantgrowth promoting bacteria from non-rhizospheric soil and their effect oncowpea (Vigna unguiculata (L.) Walp.) seedling growth. World J. Microbiol.Biotechnol. 26, 1233e1240. http://dx.doi.org/10.1007/s11274-009-0293-y.

Díaz-Alc�antara, C.A., Ramírez-Bahena, M.H., Mulas, D., García-Fraile, P., G�omez-Moriano, A., Peix, A., Vel�azquez, E., Gonz�alez-Andr�es, F., 2014. Analysis ofrhizobial strains nodulating Phaseolus vulgaris from Hispaniola Island, ageographic bridge between Meso and South America and the first historical linkwith Europe. Syst. Appl. Microbiol. 37, 149e156. http://dx.doi.org/10.1016/j.syapm.2013.09.005.

Djavaheri, M., Mercado-Blanco, J., Versluis, C., Meyer, J.M., Van Loon, L.C.,Bakker, P.A.H.M., 2012. Iron-regulated metabolites produced by Pseudomonasfluorescens WCS374r are not required for eliciting induced systemic resistanceagainst Pseudomonas syringae pv. tomato in Arabidopsis. Microbiol. Open 1,311e325. http://dx.doi.org/10.1002/mbo3.32.

Dworkin, M., Foster, J.W., 1958. Experiments with some microorganisms whichutilize ethane and hydrogen. J. Bacteriol. 75, 592e603.

El-Tarabily, K.A., 2008. Promotion of tomato (Lycopersicon esculentum Mill.) plantgrowth by rhizosphere competent 1-aminocyclopropane-1-carboxylic aciddeaminase-producing streptomycete actinomycetes. Plant Soil 308, 161e174.http://dx.doi.org/10.1007/s11104-008-9616-2.

Esitken, A., Yildiz, H.E., Ercisli, S., Donmez, M.F., Turan, M., Gunes, A., 2010. Effects ofplant growth promoting bacteria (PGPB) on yield, growth and nutrient contentsof organically grown strawberry. Sci. Hortic. 124, 62e66. http://dx.doi.org/10.1016/j.scienta.2009.12.012.

Fang, M., Kreme, R.J., Motavalli, P.P., Davis, G., 2005. Bacterial diversity of rhizo-spheres of non-transgenic and transgenic corn. Appl. Environ. Microbiol. 71,4132e4136. http://dx.doi.org/10.1128/AEM.71.7.4132-4136.2005.

Farina, R., Beneduzi, A., Ambrosini, A., De Campos, S.B., Brito Lisboa, B.,Wendischc, V., Vargasb, L.K., Passaglia, L.M.P., 2012. Diversity of plant growth-promoting rhizobacteria communities associated with the stages of canolagrowth. Appl. Soil Ecol. 55, 44e52. http://dx.doi.org/10.1016/j.apsoil.2011.12.011.

Fierer, N., Jackson, R.B., 2006. The diversity and biogeography of soil bacterialcommunities. Curr. Issue 103, 626e631. http://dx.doi.org/10.1073/pnas.0507535103.

Fishal, E.M.M., Meon, S., Yun, W.M., 2010. Induction of tolerance to Fusarium wiltand defense-related mechanisms in the plantlets of susceptible berangan ba-nana pre-inoculated with Pseudomonas sp. (UPMP3) and Burkholderia sp.(UPMB3). Agric. Sci. China 9, 1140e1149. http://dx.doi.org/10.1016/S1671-2927(09)60201-7.

García-Guti�errez, L., Romero, D., Zeriouh, H., Cazorla, F., Tor�es, J., de Vicente, A.,P�erez-García, A., 2012. Isolation and selection of plant growth-promoting rhi-zobacteria as inducers of systemic resistance in melon. Plant Soil 358, 201e212.http://dx.doi.org/10.1007/s11104-012-1173-z.

Glick, B.R., Cheng, Z., Czarny, J., Duan, J., 2007. Promotion of plant growth by ACC

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e20 19

deaminase-producing soil bacteria. In: Bakker, P.A.H.M., Raaijmakers, J.M.,Bloemberg, G., H€ofte, M., Lemanceau, P., Cooke, B.M. (Eds.), New Perspectivesand Approaches in Plant Growth-promoting Rhizobacteria Research,pp. 329e339. http://dx.doi.org/10.1007/s10658-007-9162-4.

Godfray, H.C., Beddington, J.R., Crute, I.R., Haddad, L., Lawrence, D., Muir, J.F.,Pretty, J., Robinson, S., Thomas, S.M., Toulmin, C., 2010. Food security: thechallenge of feeding 9 billion people. Science 327, 812e818. http://dx.doi.org/10.1126/science.1185383.

Harish, S., Kavino, M., Kumar, N., Saravanakumar, D., Soorianathasundaram, K.,Samiyappan, R., 2008. Biohardening with Plant growth promoting rhizosphereand endophytic bacteria induces systemic resistance against Banana Bunchy TopVirus. Appl. Soil Ecol. 39, 187e200. http://dx.doi.org/10.1016/j.apsoil.2007.12.006.

Harish, S., Kavino, M., Kumar, N., Balasubramanian, P., Samiyappan, R., 2009. In-duction of defense-related proteins by mixtures of plant growth promotingendophytic bacteria against Banana bunchy top virus. Biol. Control 51, 16e25.http://dx.doi.org/10.1016/j.biocontrol.2009.06.002.

Henneron, L., Bernard, L., Hedde, M., Pelosi, C., Villenave, C., Chenu, C., Bertrand, M.,Girardin, C., Blanchart, E., 2015. Fourteen years of evidence for positive effects ofconservation agriculture and organic farming on soil life. Agron. Sustain. Dev.35, 169e181. http://dx.doi.org/10.1007/s13593-014-0215-8.

Heyndrickx, M., Vauterin, L., Vandamme, P., Kersters, K., De Vos, P., 1996. Applica-bility of combined amplified ribosomal DNA restriction analysis (ARDRA) pat-terns in bacterial phylogeny and taxonomy. J. Microbiol. Methods 26, 247e259.http://dx.doi.org/10.1016/0167-7012(96)00916-5.

Hole, D.G., Perkins, A.J., Wilson, J.D., Alexander, I.H., Grice, F., Evans, A.D., 2005. Doesorganic farming benefit biodiversity? Biol. Conserv. 122, 113e130. http://dx.doi.org/10.1016/j.biocon.2004.07.018.

Ibekwe, A.M., Poss, J.A., Grattan, S.R., Grieve, C.M., Suarez, D., 2010. Bacterial di-versity in cucumber (Cucumis sativus) rhizosphere in response to salinity, soilpH, and boron. Soil Biol. Biochem. 42, 567e575. http://dx.doi.org/10.1016/j.soilbio.2009.11.033.

Iwai, S., Chai, B., Jesus, E.d.C., Penton, C.R., Lee, T.K., Cole, J.R., Tiedje, J.M., 2011. Gene-targeted metagenomics (GT Metagenomics) to explore the extensive diversityof genes of interest in microbial communities. In: Bruijn, F.J.d. (Ed.), Handbookof Molecular Microbial Ecology I: Metagenomics and Complementary Ap-proaches. John Wiley & Sons, Inc., Hoboken, NJ, USA, pp. 235e244. http://dx.doi.org/10.1002/9781118010518.ch28.

Kavino, M., Harish, S., Kumar, N., Saravanakumar, D., Samiyappan, R., 2010. Effect ofchitinolytic PGPR on growth, yield and physiological attributes of banana (Musaspp.) under field conditions. Appl. Soil Ecol. 45, 71e77. http://dx.doi.org/10.1016/j.apsoil.2010.02.003.

Keating, J.D., Beck, L., Materon, A., Yurtsever, N., Karuc, K., Altuntas, S., 1995. The roleof D.P rhizobial diversity in legume crops productivity in the west AsianHighlands. Exp. Agric. 31, 473e483.

Keshri, J., Mishra, A., Jha, B., 2013. Microbial population index and communitystructure in salineealkaline soil using gene targeted metagenomics. Microbiol.Res. 168, 165e173. http://dx.doi.org/10.1016/j.micres.2012.09.005.

Kim, O.S., Cho, Y.J., Lee, K., Yoon, S.H., Kim, M., Na, H., Park, S.C., Jeon, Y.S., Lee, J.H.,Yi, H., Won, S., Chun, J., 2012. Introducing EzTaxon-e: a prokaryotic 16S rRNAgene sequence database with phylotypes that represent uncultured species. Int.J. Syst. Evol. Microbiol. 62, 716e721. http://dx.doi.org/10.1099/ijs.0.038075-0.

Kimura, M., 1980. A simple method for estimating evolutionary rates of base sub-stitutions through comparative studies of nucleotide-sequences. J. Mol. Evol. 16,111e120. PMID: 7463489.

Lane, D.J., 1991. 16S/23S rRNA sequencing. In: Stackebrandt, E., Goodfellow, M.(Eds.), Nucleic Acid Techniques in Bacterial Systematics. John Wiley and Sons,New York, NY, pp. 115e175.

Lugtenberg, B., Kamilova, F., 2009. Plant-growth-promoting rhizobacteria. Annu.Rev. Microbiol. 63, 541e556. http://dx.doi.org/10.1146/annurev.micro.62.081307.162918.

Macdonald, L.M., Paterson, E., Dawsonm, L.A., McDonald, A.J.S., 2004. Short-termeffects of defoliation on the soil microbial community associated with twocontrasting Lolium perenne cultivars. Soil Biol. Biochem. 36, 489e498. http://dx.doi.org/10.1016/j.soilbio.2003.11.001.

M€ader, P., Fliessbach, A., Dubois, D., Gunst, L., Fried, P., Niggli, U., 2002. Soil fertilityand biodiversity in organic farming. Science 296, 1694e1697. http://dx.doi.org/10.1126/science.1071148.

Maheshwari, D.K., 2011. Plant growth and health promoting bacteria. In: Microbi-ology Monographs. Springer, Munster. http://dx.doi.org/10.1007/978-3-642-13612-2.

Majeed, A., Abbasi, M.K., Hameed, S., Imran, A., Rahim, N., 2015. Isolation andcharacterization of plant growth-promoting rhizobacteria from wheat rhizo-sphere and their effect on plant growth promotion. Front. Microbiol. 6, 198.http://dx.doi.org/10.3389/fmicb.2015.00198.

Mantelin, S., Touraine, B., 2004. Plant growth-promoting bacteria and nitrateavailability: impacts on root development and nitrate uptake. J. Exp. Bot. 55,27e34. http://dx.doi.org/10.1093/jxb/erh010.

Marasco, R., Rolli, E., Fusi, M., Cherif, A., Abou-Hadid, A., El-Bahairy, U., Borin, S.,Sorlini, C., Daffonchio, D., 2013. Plant growth promotion potential is equallyrepresented in diverse grapevine root-associated bacterial communities fromdifferent biopedoclimatic Environments. BioMed. Res. Int. 2013, 1e17. http://dx.doi.org/10.1155/2013/491091.

Martínez, L., Caballero-Mellado, J., Orozco, J., Martínez-Romero, E., 2003. Diazo-trophic bacteria associated with banana (Musa spp.). Plant Soil 257, 35e47.

http://dx.doi.org/10.1023/A:1026283311770.Mirzajani, F., Askari, H., Hamzelou, S., Farzaneh, M., Ghassempour, A., 2013. Effect of

silver nanoparticles on Oryza sativa L. and its rhizosphere bacteria. Ecotoxicol.Environ. Saf. 88, 48e54. http://dx.doi.org/10.1016/j.ecoenv.2012.10.018.

Morrissey, J.P., Dow, J.M., Mark, G.L., O'Gara, F., 2004. Are microbes at the root of asolution to world food production? Rational exploitation of interactions be-tween microbes and plants can help to transform agriculture. EMBO Rep. 5,922e926. http://dx.doi.org/10.1038/sj.embor.7400263.

Moyer, C.L., Tiedje, J.M., Dobbs, F.C., Karl, D.M., 1996. A computer-simulated re-striction fragment length polymorphism analysis of bacterial small-subunitrRNA genes: efficacy of selected tetrameric REs for studies of microbial di-versity in nature. Appl. Environ. Microbiol. 62, 2501e2507. Doi: PMC168032.

Mulas, D., García-Fraile, P., Carro, L., Ramírez-Bahena, M.H., Casquero, P.,Vel�azquez, E., Gonz�alez-Andr�es, F., 2011. Distribution and efficiency of Rhizo-bium leguminosarum strains nodulating Phaseolus vulgaris in Northern Spanishsoils: selection of native strains that replace conventional N fertilization. SoilBiol. Biochem. 43, 2283e2293. http://dx.doi.org/10.1016/j.soilbio.2011.07.018.

Mulas, D., Díaz-Alc�antara, C.A., Mulas, R., Marcano, I., Barquero, M., Serrano, P.,Gonz�alez-Andr�es, F., 2013. Inoculants bases in authochthonous microorganims,a strategy to optimize agronomic perfomance of biofertilizers. In: RodelasGonz�alez, M.B., Gonz�alez-L�opez, J. (Eds.), Benefical Plant-microbial Interactions:Ecology and Applications, pp. 300e328.

Nautiyal, C.S., 1999. An efficient microbiological growth medium for screeningphosphate solubilizing microorganisms. FEMS Microbiol. Lett. 170, 265e270.http://dx.doi.org/10.1111/j.1574-6968.1999.tb13383.x.

Nocker, A., Burr, M., Camper, A.K., 2007. Genotypic microbial community profiling: acritical technical review. Microb. Ecol. 54, 276e289. http://dx.doi.org/10.1007/s00248-006-9199-5.

O'Hara, G.W., Goss, T.J., Dilworth, M.J., Glenn, A.R., 1989. Maintenance of intracel-lular pH and acid tolerance in Rhizobium meliloti. Microbiol. Appl. Environ.Microbiol. 55, 1870e1876. PMC202972.

Peeran, M.F., Nagendran, K., Gandhi, K., Raguchander, T., Prabakar, K., 2014. Water inoil based PGPR formulation of Pseudomonas fluorescens (FP7) showed enhancedresistance against Colletotrichum musae. Crop Prot. 65, 186e193. http://dx.doi.org/10.1016/j.cropro.2014.07.010.

Penrose, D.M., Glick, B.R., 2003. Methods for isolating and characterizing ACCdeaminase-containing plant growth-promoting rhizobacteria. Physiol. Plant.118, 10e15. http://dx.doi.org/10.1034/j.1399-3054.2003.00086.x.

Pii, Y., Mimmo, T., Tomasi, N., Terzano, R., Cesco, S., Crecchio, C., 2015. Microbialinteractions in the rhizosphere: beneficial influences of plant growth-promoting rhizobacteria on nutrient acquisition process. A review. Biol. Fertil.Soils 51, 403e415. http://dx.doi.org/10.1007/s00374-015-0996-1.

Prashar, P., Kapoor, N., Sachdeva, S., 2014. Rhizosphere: its structure, bacterial di-versity and significance. Rev. Environ. Sci. Biotechnol. 13, 63e77. http://dx.doi.org/10.1007/s11157-013-9317-z.

Rivas, R., Peix, A., Mateos, P.F., Trujillo, M.E., Martínez-Molina, E., Vel�azquez, E.,2006. Biodiversity of populations of phosphate solubilizing rhizobia that nod-ulates chickpea in different Spanish soils. Plant Soil 287, 23e33. http://dx.doi.org/10.1007/s11104-006-9062-y.

Rogers, J.S., Swofford, D.L., 1998. A fast method for approximating maximum like-lihoods of phylogenetic trees from nucleotide sequences. Syst. Biol. 47, 77e89.

Ryan, P.R., Dessaux, Y., Thomashow, L.S., Weller, D.M., 2009. Rhizosphere engi-neering and management for sustainable agriculture. Plant Soil 321, 363e383.http://dx.doi.org/10.1007/s11104-009-0001-6.

Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method for recon-structing phylogenetic trees. Mol. Biol. Evol. 4, 406e425. PMID 3447015.

Schwyn, B., Neilands, J.B., 1987. Universal chemical assay for the detection anddetermination of siderophores. Anal. Biochem. 160, 47e56. PMID 2952030.

Selvaraj, S., Ganeshamoorthi, P., Anand, T., Raguchander, T., Seenivasan, N.,Samiyappan, R., 2014. Evaluation of a liquid formulation of Pseudomonas fluo-rescens against Fusarium oxysporum f. sp. cubense and Helicotylenchus multi-cinctus in banana plantation. BioControl 59, 345e355. http://dx.doi.org/10.1007/s10526-014-9569-8.

Shannon, C.E., Weaver, W., 1949. The Mathematical Theory of Communication.University of Illinois Press, Urbana.

Singh, J.S., Pandey, V.C., Singh, D.P., 2011. Efficient Soil microorganisms: a newdimension for sustainable agriculture and environmental development. Agric.Ecosyst. Environ. 140, 339e353. http://dx.doi.org/10.1016/j.agee.2011.01.017.

Sinha, R.K., Valani, D., Chauhan, K., Agarwal, S., 2010. Embarking on a second greenrevolution for sustainable agriculture by vermiculture biotechnology usingearthworms: reviving the dreams of Sir Charles Darwin. J. Agric. Biotechnol.Sustain. Dev. 2, 113e128.

Sklarz, M.Y., Angel, R., Gillor, O., Soares, M.I., 2009. Evaluating amplified rDNA re-striction analysis assay for identification of bacterial communities. AntonieLeeuwenhoek 96, 659e664. http://dx.doi.org/10.1007/s10482-009-9380-1.

Smalla, K., Wieland, G., Buchner, A., Zock, A., Parzy, J., Kaiser, S., Roskot, N., Heuer, H.,Berg, G., 2001. Bulk and rhizosphere soil bacterial communities studied bydenaturing gradient gel electrophoresis: plant-dependent enrichment andseasonal shifts revealed. Appl. Environ. Microbiol. 67, 4742e4751. http://dx.doi.org/10.1128/AEM.67.10.4742-4751.2001.

Souza, S.A., Xavier, A.A., Costa, M.R., Cardoso, A.M.S., Pereira, M.C.T., Nietsche, S.,2013. Endophytic bacterial diversity in banana “Prata An~a” (Musa spp.) roots.Genet. Mol. Biol. 36, 252e264. PMID: 23885208.

Suenaga, H., 2012. Targeted metagenomics: a high-resolution metagenomicsapproach for specific gene clusters in complex microbial communities. Environ.

I.-E. Marcano et al. / Soil Biology & Biochemistry 99 (2016) 1e2020

Microbiol. 14, 13e22. PMID: 21366818.Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., Kumar, S., 2011. MEGA5:

molecular evolutionary genetics analysis using maximum likelihood, evolu-tionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28,2731e2739. PMID: 21546353.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. TheCLUSTAL_X windows interface: flexible strategies for multiple sequence align-ment aided by quality analysis tools. Nucleic Acids Res. 25, 4876e4882. PMID:9396791.

Viveros, O.M., Jorquera, M.A., Crowley, D.E., Gajardo, G., Mora, M.L., 2010.

Mechanisms and practical considerations involved in plant growth promotionby rhizobacteria. J. Soil Sci. Plant Nutr. 10, 293e319.

Weisburg, W.G., Barns, S.M., Pelletier, D.A., Lane, D.J., 1991. 16S Ribosomal DNAamplification for phylogenetic study. J. Bacteriol. 173, 697e703.

Wertz, J.T., Kim, E., Breznak, J.A., Schmidt, T.M., Rodrigues, J.L.M., 2012. Genomic andPhysiological Characterization of the verrucomicrobia isolate Diplosphaera col-itermitum gen. nov., sp nov., reveals microaerophily and nitrogen fixation genes.Appl. Environ. Microbiol. 78, 1544e1555.

Woese, C.R., 2000. Interpreting the universal phylogenetic tree. Proc. Natl. Acad. Sci.U. S. A. 97, 8392e8396.