lable at ScienceDirect

Soil Biology & Biochemistry 68 (2014) 392e401

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lbio

Dynamics of the bacterial community structure in the rhizosphereof a maize cultivar

Xiangzhen Li a,b,1, Junpeng Rui b,1, Yuejian Mao a, Anthony Yannarell c, Roderick Mackie a,d,*

a Energy Biosciences Institute, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USAbKey Laboratory of Environmental and Applied Microbiology, Chengdu Institute of Biology, Chinese Academy of Sciences, Chengdu 610041, ChinacDepartment of Natural Resources and Environmental Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USAdDepartment of Animal Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA

a r t i c l e i n f o

Article history:Received 24 April 2013Received in revised form27 September 2013Accepted 10 October 2013Available online 25 October 2013

Keywords:Maize rhizospherePlant growth stageRoot exudatesBacterial community structurePyrosequencingPredominant bacterial generaDynamics

* Corresponding author. 132 Animal Sciences LaborUrbana, IL 61801, USA. Tel.: þ1 217 2442526; fax: þ1

E-mail address: r-mackie@illinois.edu (R. Mackie).1 These authors contribute equally.

0038-0717/$ e see front matter � 2013 Published byhttp://dx.doi.org/10.1016/j.soilbio.2013.10.017

a b s t r a c t

Rhizosphere bacteria have significant contributions to crop health, productivity and carbon sequestra-tion. As maize (Zea mays) is an important economic crop, its rhizosphere bacterial communities havebeen intensively investigated using various approaches. However, low-resolution profiling methods oftenmake it difficult to understand the complicated rhizosphere bacterial communities and their dynamics.In this study, we analyzed growth-stage related dynamics of bacterial community structures in therhizosphere of maize using the pyrosequencing method, which revealed an assembly of bacteriaenriched in the rhizosphere. Our results revealed that the rhizosphere of maize was preferentiallycolonized by Proteobacteria, Bacteroidetes and Actinobacteria, and each bacterial phylum was repre-sented by one or two dominating subsets of bacterial groups. Dominant genera enriched in the rhizo-sphere included Massilia, Burkholderia, Ralstonia, Dyella, Chitinophaga and Sphingobium. Rhizospherebacterial community structures significantly changed through different growth stages at lower taxo-nomic ranks (family, genus and OTU levels). Genera Massilia, Flavobacterium, Arenimonas and Ohtaek-wangia were relatively abundant at early growth stages, while genera Burkholderia, Ralstonia, Dyella,Chitinophaga, Sphingobium, Bradyrhizobium and Variovorax populations were dominant at later stages.Comparisons of pyrosequencing data collected in Illinois, USA in this study with the available data fromBraunschweig, Germany indicated many common bacterial inhabitants but also many differences in thestructure of bacterial communities, implying that some site-specific factors, such as soil properties, mayplay important roles in shaping the structure of rhizosphere bacterial community.

� 2013 Published by Elsevier Ltd.

1. Introduction

The rhizosphere is directly influenced by root secretions andassociatedmicroorganisms. Soil pH, structure, oxygen and nutritionlevels in the rhizosphere differ from those in the bulk soil (Singhet al., 2004). The unique ecological niche shapes the structure ofrhizosphere bacterial community through the interactions of plantspecies, the chemical nature of root exudates, soil properties, andmany other factors (Savka and Farrand, 1997). While some detri-mental microbes undermine plant health, mutualistic rhizospheremicrobes provide plants with mineral nutrients, phytohormones,and protect the plant against phytopathogens (Mendes et al., 2011;

atory, 1207 W. Gregory Drive,217 3338286.

Elsevier Ltd.

Singh et al., 2004; Vercellino and Gómez, 2013). Thus, to under-stand the suite of bacterial interactions that may occur over thelifetime of a plant, it is important to know the rhizosphere bacterialcommunity and its variation over plant growth stages.

Rhizosphere bacterial communities differ across plant species,soil type, root architecture and growth stage (Berg and Smalla,2009; Marschner et al., 2001, 2004). As maize (Zea mays) is animportant crop, its rhizosphere bacterial community has beenintensively investigated using a variety of approaches (Aira et al.,2010; Castellanos et al., 2009; Chelius and Triplett, 2001;Dohrmann et al., 2013). Previous studies suggested that maize se-lects for a specific bacterial community depending on soil proper-ties (Castellanos et al., 2009), genotypes (Aira et al., 2010), cropmanagement, such as fertilizer (Aira et al., 2010) and growth stages(Cavaglieri et al., 2009; Di Cello et al., 1997). Some studies havesuggested that microbial community composition in the maizerhizosphere is independent of cultivar (Dohrmann et al., 2013;Schmalenberger and Tebbe, 2002), growth stage (Gomes et al.,

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401 393

2001) and genotype (Schmalenberger and Tebbe, 2002). Theapparent disparity of bacterial responses in the rhizosphere couldbe caused partially by differences among plant species and soiltypes, but likely also by differences among the methodologies used,such as denaturing gradient gel electrophoresis (DGGE) and clonelibrary analysis, which have different resolution and detectionlimitation (Bent and Forney, 2008). Low-resolution profilingmethods potentially miss some important information and make itdifficult to detail the phylogenetic composition of a rhizospherebacterial community.

Plant growth stage will influence root physiology, and changethe quality and quantity of root exudates; consequently, thesechanges exert a selection on root-associated microorganisms atdifferent growth stages (Dunfield and Germida, 2003; Houldenet al., 2008). Some studies suggested that seasonal variations inthe activity and relative abundance of rhizosphere bacterial com-munities are plant-dependent (Dunfield and Germida, 2003;Houlden et al., 2008; Mougel et al., 2006; Smalla et al., 2001).Although some research showed dynamic changes of maizerhizosphere bacterial community at different growth stages usingculture-dependent (Cavaglieri et al., 2009; Nacamulli et al., 1997)and fingerprinting methods (Di Cello et al., 1997), detail phyloge-netic compositions cannot be monitored due to low resolution ofthese approaches. It is also unclear at which taxonomic levels suchdynamics can be detected.

In this study, we used DGGE and pyrosequencing methods toinvestigate the growth-stage related dynamics of the bacterialcomposition and diversity in the maize rhizosphere. The objectivesare (i) to analyze the bacterial community structure and theirgrowth-stage related dynamics in the rhizosphere of a maizecultivar, (ii) to compare the data of this study obtained from a fine-silty soil located in Illinois, USA, to those from a similar studyconducted in a silty sand soil in Germany, in order to identifycommon bacterial inhabitants and the difference in the bacterialcommunity structure.

2. Materials and methods

2.1. Field description and sampling

Our sample sitewas located at the Energy Farm of the Universityof Illinois (Champaign, IL, USA, 40�030N, 88�120W, 230m elevation).The soil is a Drummer-Flanagan series (fine-silty, mixed, mesicTypic Endoaquoll, pH 5.8, organic matter 3.8%, total nitrogen 0.18%)formed from deep deposits of loess and silt parent material on thetop of the glacial till and outwash plains. It is an organically rich,highly productive Corn Belt soil that has been rotation croppedwith maize and legumes for over 100 years. The Z. mays cultivar cv34B43 (Pioneer Hi-Bred International, Des Moines, IA, USA) wasplanted in the middle of May, 2011, in a 0.7-ha field and managedfollowing standard agricultural practices in this regionwith annualapplications of 20 g m�2 of mixed urea, ammonia, and nitrate fer-tilizer at planting.

Sampling was conducted after maize emergence at 2 weeks(June 1, 2011, vegetative stages V2eV3), 5 weeks (June 21, V8eV9),9 weeks (July 18, VT, tasseling) and 12weeks (August 11, R2, blister).Growth stage was defined based on Hanway’s work (Hanway,1963). We selected four plots in the maize field for sampling.Sampling sites for the next sampling were nearby the previouslocations to minimize the effects caused by spatial variation of soilproperties on the bacterial communities. Within each plot, 4rhizosphere soil samples and 4 bulk soil samples were collectedand composited, so that each individual sample was a pool of fourplants or soil cores (1 m away for different plants or soil cores). Intotal, 16 bulk soil and 16 rhizosphere soil samples were obtained

during the growing season (four plots sampled at four time points).To collect rhizosphere soil, plants were carefully excavated using ashovel, and visible soil particles were removed. Fresh roots werecollected into a 50ml centrifuge tube (ca. 50 g). Soil tightly attachedon the root surface was rinsed using phosphate buffer solution, andcentrifuged at 8000 rpm for 10 min. The pellet was collected asrhizosphere soil. Soil cores of 0e30 cm depth between plants werealso collected as bulk soil samples. All the samples were transferredto the lab on dry ice, and pre-processed and stored at �80 �C fordownstream applications.

2.2. DNA extraction

Soil genomic DNA was extracted using the FastDNA SPIN Kit forSoil (MP Biomedicals, Solon, OH, USA). DNA quality was assessed bythe ratio of A260/280 and A260/230 using a nanoDrop spectropho-tometer. DNA concentrations were also quantified with Qubit assay(Invitrogen, CA).

2.3. Denaturing gradient gel electrophoresis (DGGE)

PCR amplification of bacterial 16S rRNA genes from soil genomicDNA for DGGE analysis was done using the bacteria-specific for-ward primer 357F containing a GC-clamp and the reverse primer907R (Muyzer et al., 1993). The PCR mixture (50 ml) contained 1�PCR buffer, 1.5 mM MgCl2, each deoxynucleoside triphosphate at0.4 mM, each primer at 1.0 mMand 1 U of Ex Taq polymerase (TaKaRaBio Inc, Japan) and 20 ng soil genomic DNA. The PCR amplificationprogram included initial denaturation at 94 �C for 3 min, followedby 30 cycles of 94 �C for 40 s, 56 �C for 60 s, and 72 �C for 60 s, and afinal extension at 72 �C for 10 min. PCR products were subjected toelectrophoresis using 1.5% agarose gel. The band with a correct sizewas excised and purified using Wizard� SV Gel and PCR Clean-UpSystem (Promega, USA).

DGGE was performed using a DCode system (Bio-Rad, Califor-nia) as described by Muyzer et al. (1993). An 8% polyacrylamide gelwith a linear denaturant concentration from 30 to 55% (where 100%denaturant contains 7 M urea and 40% formamide [v:v]) was usedto separate the PCR products. Approximately 300 ng of PCR productper lane was loaded onto the DGGE gel. The gel was electro-phoresed for 4 h at 60 �C at a constant voltage of 200 V, stained for1 h with SYBR Green I (Molecular Probes, Eugene Org.), illuminatedon a transilluminator under UV-light. The sharp bands wereexcised. 16S rRNA clone library was constructed using pGEM�-TEasy Vector Systems according to manufacturer’s instruction(Promega, USA). Ten clones from each sample were sequencedusing Sanger method.

DGGE images were analyzed with Quantity One software (Bio-Rad, USA). A matrix was constructed for all lanes, taking into ac-count the presence or absence of the individual bands and therelative contribution of each band to the total intensity of the lane.A Dice coefficient similarity matrix was generated based on abovematrix. Finally, DGGE profiles for all samples were clustered basedon Dice’s similarity coefficient using the complete-linkage method.

2.4. Pyrosequencing

To amplify the V4-V5 hypervariable regions of 16S rRNA genes forpyrosequencing, universal primers 519F (50-CAGCMGCCGCGG-TAATWC-30) and 926R (50-CCGTCAATTCMTTTRAGTT-30)were used inPCR (Baker et al., 2003). The oligonucleotides included 454 Life Sci-ence’s A or B sequencing adapter fused to the 50 end of forward andreverse primers. A unique 10-mer barcode sequence was added be-tween the sequencing adapter and forward primer to differentiatebetween samples. The methods for 16S rRNA gene amplification and

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401394

PCR product purification for pyrosequencing were the same as thosedescribed above. Amplicons of all samples were pooled in an equi-molar library, and then 454 pyrosequencing was carried out using aGS FLX system (454 Life Sciences, Branford, CT) at Roy J. CarverBiotechnology Center at University of Illinois.

2.5. Pyrosequencing data analysis

The raw sequences were sorted based on unique sample tags,and trimmed for sequence quality using RDP Pipeline Initial Process(http://pyro.cme.msu.edu/). Sequences with low quality(length < 300 bp, with ambiguous base ‘N’, and average basequality score < 20) were removed, as were any sequences that didnot perfectly match the proximal PCR primer. The sequences werephylogenetically assigned according to their best matches to se-quences in RDP database using the RDP classifier. Sequences an-notated as chloroplasts or Viridiplantae were removed. Theprocessed pyrosequences were aligned using the fast, secondary-structure aware infernal aligner in RDP pyrosequencing pipeline(Cole et al., 2009). The aligned 16S rRNA gene sequences werescreened for chimeric sequences using the Uchime algorithm(Edgar et al., 2011). Re-sampling to the same sequence depth (2854sequences per sample) was performed using Daisychopper (http://www.genomics.ceh.ac.uk/GeneSwytch/Tools.html) prior to down-stream analysis. Sequences were clustered by the complete-linkageclustering method incorporated in the RDP pyro pipeline. Opera-tional taxonomic units (OTUs) were classified using a 97% nucleo-tide sequence similarity cutoff. Rarefaction, Shannon index andChao1 estimator were calculated in RDP at 97% sequence similarity.The phylogenetic affiliation of each 16S rRNA gene sequence wasanalyzed by RDP Classifier at a confidence level of 80%.

Dohrmann et al. studied bacterial diversity in the rhizosphere ofBt- and conventional maize varieties at the vTI research center inBraunschweig, Germany (52�1800300N, 10�2604000E) (Dohrmannet al., 2013). Soil is para brown earth, silty sand (sand 81.8%, silt8.7%, clay 7.6%, pH 6.4, organic matter 1.67%) (Castellanos et al.,2009). They measured the bacterial community composition inthe samples from different cultivars and root microhabitats (fine orcoarse roots). To our knowledge, these datasets are the onlycurrently available pyrosequencing datasets of maize rhizosphereproduced using the same PCR primers and DNA extraction methodas the current study. Thus, we compared bacterial communitycomposition in these datasets in order to identify the commonbacterial inhabitants and the differences in bacterial communitystructures in maize rhizosphere. The original pyrosequencing dataare available at the European Nucleotide Archive by accessionERP002580 (current study) and ERP001118 (Germany study)(http://www.ebi.ac.uk/ena/data/view/).

2.6. Statistical analysis

Overall structural changes of rhizosphere and bulk soil bacterialcommunities were evaluated by principal coordinates analysis(PCoA) in Fast UniFrac (http://bmf.colorado.edu/fastunifrac/)(Lozupone et al., 2011). Representative sequences of OTUs werealigned with MUSCLE (Edgar, 2004), and then used to generate aphylogenetic tree using FastTree (Price et al., 2010). The phyloge-netic tree was then used for weighted UniFrac PCoA analysis. Thestatistical significance among datasets was assessed by PerMA-NOVA using the weighted PCoA scores in PAST (http://folk.uio.no/ohammer/past/). A heatmap of the top 150 OTUs selected from thewhole community was constructed using the R package pheatmap(http://cran.r-project.org/web/packages/pheatmap/index.html).

Differences in relative abundances of taxonomic units betweensamples were tested by one-way-analysis of variance (ANOVA).

OTUs or bacterial groups with a p-value of�0.05 were selected andidentified as the phylotypes or bacterial groups that were signifi-cantly influenced by rhizosphere or growth stages.

3. Results

3.1. DGGE profiles of bacterial communities

DGGE profiling of 16S rRNA fragments from rhizosphere andbulk soil showed high similarity of DGGE patterns for the fourreplicate samples from each sampling time, suggesting a lowvariability caused by DGGE sample preparation (field sampling,DNA extraction, PCR amplification and DGGE analysis). Werandomly selected two of four replicates from each sampling timein order to run in one DGGE gel (Fig. 1A). Cluster analysis showedthat DGGE patterns of rhizosphere samples and bulk soil sampleswere well separated. In addition, DGGE profiles from early growthstages (June 1 and 21) were clustered together and were distinctfrom those of later stages (July 18 and August 11) for both rhizo-sphere and bulk soil samples.

The rhizosphere DGGE pattern contained three strong bandsthat were enriched compared to bulk soil. Band 1 was composed ofseveral bacterial populations, which were not separated by DGGE,including phylotypes affiliated to Massilia, Burkholderia and Ral-stonia. Band 2 contained sequences mainly affiliated to unculturedBradyrhizobium (AJ301632, identity 99%), uncultured Lysobacter(AY074793, 98%) and uncultured Steroidobacter (EU979081, 95%).Band 3 contained sequences affiliated to uncultured Acidobacteriabacterium (HM062194, 97.5%) and plant chloroplast DNAs. Theseresults suggested a strong shift in rhizosphere bacterial communityacross root growth stages. The intensity of band 1 increased at latergrowth stages, while the intensities of band 2 and band 3 decreasedwith growth stage.

3.2. Overall structural changes of bacterial communities

In total, 159,327 high quality and chimera-free reads with anaverage length of 408 bp were obtained (Table SI_1) by pyrose-quencing. Bacterial diversity measures based on OTU, Chao1 rich-ness, Shannon’s diversity and evenness index revealed theenrichment of some phylotypes in rhizosphere compared to bulksoil (Table 1). Among 2854 pyrotags, 1058e1234 and 1324e1401OTUs (defined at 97% sequence similarity) were found in eachrhizosphere and bulk soil samples, respectively (Table 1, Fig. SI_1).PCoA analysis indicated that four randomly collected replicatesusually clustered closely (Fig. 2), underscoring the reproducibilityof these bacterial community profiles. The rhizosphere and bulksoil bacterial communities were clearly separated from each other(Fig. 2), indicating a significant impact of maize roots on rhizo-sphere bacterial community (PerMANOVA, p < 0.05; Table SI_2).

Rhizosphere bacterial communities were clearly separated intotwo clusters: an early growth-stage community (first two samplingpoints) and a later one (the later two sampling points) (PerMA-NOVA, p < 0.05) (Fig. 2, Table SI_2), indicating a significant shift ofrhizosphere bacterial structure along maize growth stages, eventhough the total number of OTUs and Chao1 did not changesignificantly between sampling time points (Table 1). Rhizosphereeffect was the main factor in the first principle coordinate axis(PCo1), which contributed 46.4% of total variation, and samplingtime (growth stage) was the main factor in the second coordinateaxis (PCo2), which contributed 23.1% of variation. The pattern wasrecapitulated by hierarchical clustering based on UniFrac distancematrix (Fig. 2B), which showed similar cluster pattern with DGGEprofiles.

Fig. 1. DGGE patterns of PCR amplified 16S rRNA genes from maize rhizosphere and bulk soil (A); clustering of DGGE profiles based on Dice’s similarity coefficient using complete-linkage method (B). Band 1 included phylotypes affiliated to genusMassilia, Burkholderia, Ralstonia; Band 2: Bradyrhizobium, Lysobacter; Band 3: Acidobacteria_Gp1 and chloroplast.

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401 395

3.3. Taxonomic distributions of bacteria enriched in the rhizosphereof maize

The most abundant phyla were Proteobacteria, Acidobacteria,Actinobacteria, Bacteroidetes and Gemmatimonadetes, accountingfor 86.4e89.6% and 72.1e73.8% of total bacteria in the rhizosphereand bulk soil, respectively. However, the rhizosphere significantlyenriched a subset of phyla (Fig. 3, Table SI_3), which is consistentwith the lower Shannon diversity of the rhizosphere bacterialcommunity (Table 1). Specifically, the rhizosphere enriched Pro-teobacteria, Bacteroidetes and Actinobacteria, which accounted for

Table 1Bacterial diversity indices (at 97% sequence similarity) in the maize rhizosphere and bul

Soil Jun 1

Number of OTUsa Rhb 1234 � 20c

Bk 1324 � 17Chao1 estimator of richnessd Rh 2817 � 99

Bk 2807 � 108Shannon’s diversity index Rh 6.428 � 0.053

Bk 6.709 � 0.039Shannon’s evenness Rh 0.903 � 0.006

Bk 0.933 � 0.004

a All sequence datasets were re-sampled to ensure equal number of reads (2854 readb Rh: rhizosphere, Bk: bulk soil.c Values are means of four replicate pyrosequencing datasets.d The means of Chao1, Shannon’s diversity and evenness indices were significantly dif

June 1. All these indices were not significantly different between different sampling tim

73.2e80.8% of total reads in the rhizosphere and only 46.0e56.2% inbulk soil. As the most abundant phylum, Proteobacteria accountedfor 46.2e48.0% of total bacteria in the rhizosphere at early growthstages, and increased up to 54.7e56.6% at the later growth stages.On the other hand, the abundance of phyla Acidobacteria, Gem-matimonadetes, Chloroflexi, Firmicutes and Nitrospira decreased inrhizosphere in comparison to bulk soil. These phyla accounted for25.7e34.3% of total bacteria in bulk soil, but only 11.1e15.7% in therhizosphere. Much higher proportion of unclassified sequences wasdetected in bulk soil, indicating that it contained more unknownbacterial species.

k soil.

Jun 21 Jul 18 Aug 11

1123 � 33 1058 � 123 1122 � 461372 � 46 1401 � 12 1360 � 242463 � 93 2491 � 271 2448 � 623000 � 127 3002 � 75 2928 � 646.284 � 0.130 5.947 � 0.396 6.292 � 0.1066.774 � 0.067 6.856 � 0.089 6.794 � 0.0490.895 � 0.015 0.854 � 0.042 0.896 � 0.0100.938 � 0.005 0.946 � 0.002 0.942 � 0.005

s) for downstream analysis.

ferent (p < 0.05) between rhizosphere and bulk soil except the samples collected ate in both rhizosphere and bulk soil.

Fig. 2. Overall structural changes of maize rhizosphere (Rh) and bulk soil (Bk) bacterial communities from pyrosequencing data. (A) PCoA score plot based on weighted UniFracmetrics; (B) clustering of soil bacterial communities based on the UniFrac distances calculated in Fast UniFrac. Early stage samples were collected at June 1 and June 21 (sample Aand B in figure B); later stage samples were collected at July 18 and August 11 (sample C and D in figure B). Detail sample information was shown in Table SI-1.

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401396

The most dominant classes in the rhizosphere were Betapro-teobacteria, Actinobacteria, Alphaproteobacteria, Gammaproteo-bacteria and Sphingobacteria, which accounted for 63.1e67.4% oftotal bacteria at the early stages and 76.4e77.2% at the later stages(Table SI_3, Fig. SI_2). Within the rhizosphere Proteobacterialcommunity at early growth stages, there was over-representationof Betaproteobacteria families compared to Alphaproteobacteriaand Gammaproteobacteria. However, the relative proportion ofAlphaproteobacteria was the highest in August samples.

Lower-order taxonomic analysis demonstrated that a specificsubset of Proteobacteria was enriched in the rhizosphere, includingXanthomonadaceae (mainly including genera Dyella and Arenimo-nas), Oxalobacteraceae (mainly genus Massilia), Burkholderiaceae(mainly genera Burkholderia and Ralstonia) and Sphingomonadaceae(genus Sphingobium) (Fig. 3). The family Streptomycetaceae (genusStreptomyces) in Actinobacteria, and the families Chitinophagaceae(genus Chitinophaga) and Flavobacteriaceae (genus Flavobacterium)in Bacteroidetes were also enriched in the rhizosphere.

At the genus level, 333 genera were detected in the rhizosphere.Eighty-nine genera were identified as significantly enriched in therhizosphere (p < 0.05) compared to bulk soil (Table SI_3). Amongthem, 49 genera were significantly enriched at early growth stages,and 67 genera were significantly enriched at the later stages.Twenty-two genera were identified to be dominant genera withrelative abundance> 0.5% of total bacteria reads in the rhizosphere,and these genera occurred in all the rhizosphere and most bulk soilsamples (Table 2). Among them, 18 enriched genera were repre-sented with relative abundance > 1% in the rhizosphere, andtogether they accounted for average 33.5% of total reads in therhizosphere and 3.5% in the bulk soil (Fig. SI_3). These mostabundant genera mainly included Massilia, Burkholderia, Chitino-phaga, Ralstonia, Dyella, Sphingobium and Flavobacterium.

At the OTU level, seventy-five OTUs were identified as signifi-cantly enriched in the rhizosphere (Table SI_3). Among them, 28OTUs were identified as dominant enriched OTUs with relativeabundance > 0.5% in at least one sample (Table 2). The mostabundant 28 rhizosphere OTUs accounted for 22.5% of total readson average. These OTUs were mainly affiliated with the sameabundant genera mentioned above. Among 28 OTUs, 5 OTUs werevery low or not detected in some early stage samples, but theyappeared in the late stage samples with relative high abundance.

These OTUs included OTU27365 (Chitinophaga), OTU5987 (Strep-tomyces), OTU5729 (Sphingobium), OTU3354 (Sphingobium) andOTU1352 (Sphingobium).

3.4. Growth-stage related dynamics of the bacterial communitystructure

The structures of early growth-stage communities were signifi-cant different from those of later growth stages in both rhizosphereand bulk soil (Fig. 2 and Table SI_2). At the phylum level, the relativeabundance of Proteobacteria and Actinobacteria increased signifi-cantly with growth stage in both rhizosphere and bulk soil(Table SI_3), while the relative abundance of Acidobacteria, Bacter-oidetes andGemmatimonadetesdecreased.Moreover, at thephylum,class and order levels, the dynamic patterns of many bacterial com-munities were comparable in rhizosphere and bulk soil (Fig. SI_4).

At the family level, Oxalobacteraceae decreased at later stages inrhizosphere soil, while Burkholderiaceae and Sphingomonadaceaeincreased significantly (p< 0.05) (Fig. 4). Members of these familiesdid not change significantly in the bulk soil (Table 2). At the genuslevel, Massilia, Flavobacterium and Arenimonas dominated in theearly growth-stage bacterial community; Ralstonia, Burkholderia,Ohtaekwangia and Bradyrhizobium dominated in themiddle growthperiod; Chitinophaga, Dyella, Sphingobium and Variovorax domi-nated in the later growth stage (Fig. SI_5). Massilia was the mostabundant genus in the early growth stages, and it sharply decreasedin the late stages. The changes in Massilia proportion were mainlydriven by the shifts in three dominant OTUs (OTU6022, 5806,3337). However, these genera did not change significantly in thebulk soil. In bulk soil, genera Gp4 (Acidobacteria), Nitrospira, Pas-teuria, Ohtaekwangia and Arenimonas decreased significantly at thelater stages, while genera Ktedonobacter and Bradyrhizobiumincreased at the later stages (Table SI_3, Fig. SI_6).

3.5. Comparisons of Illinois datasets with German datasets

PCoA analysis indicated that bacterial communities fromGerman maize rhizosphere clustered together regardless of culti-vars or root microhabitats (fine or coarse roots), and they weresignificantly different from Illinois energy farm samples (p < 0.01by PerMANOVA) (Fig. SI_7). The differences between German and

Fig. 3. Phylogenetic distributions of phyla (% of total reads) present in the rhizosphere and bulk soil (A), and relative abundance (% of individual taxonomic group) of classifiedfamilies in the phyla Actinobacteria (B), Bacteroidetes (C) and Proteobacteria (D) and in class Alphaproteobacteria (E), Betaproteobacteria (F) and Gammaproteobacteria (G). Dis-tribution of classified genera in two dominant Proteobacteria orders of Burkholderiales (H) and Xanthomonadales (I).

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401 397

Illinois rhizosphere samples were larger than the differencesamong the maize cultivars, root microhabitats or growth stages.

At high taxonomic ranks, all the rhizosphere samples enrichedProteobacteria and Actinobacteria. However, Illinois samples con-tained more Bacteroidetes and less Firmicutes. Whenwe comparedthese datasets at lower taxonomic ranks, different patterns wereobserved (Table SI_4). Illinois samples contained more bacteriabelonging to order Burkholderiales, such as Oxalobacteraceae

(genus Massilia, enriched at the early stage) and Burkholderiaceae(genera Burkholderia and Ralstonia, enriched at the late stage), thanGerman samples. In addition, Illinois samples also contained moreXanthomonadaceae (genus Dyella) and Chitinophagaceae (genusChitinophaga), but less Nocardioidaceae, Pseudomonadaceae(genus Rhizobacter) and genus Arthrobacter.

In total 34,456 OTUs were detected in the combining datasetthat contained 518,576 pyrosequencing reads from both Illinois and

Table 2Dominant genera and OTUs significantly enriched or depleted (p < 0.05) in the rhizosphere with relative abundance > 0.5% of total bacterial reads.

Time Relative abundance in Rh (%) Relative abundance in Bk (%)

Jun 1 Jun 21 Jul 18 Aug 11 Jun 1 Jun 21 Jul 18 Aug 11

(a) Enriched genera (22)Ralstonia 0.096 0.858 9.294 1.384 0.009 0.000 0.018 0.000Sphingobium 0.044c 0.184c 4.143b 6.430a 0.018 0.018 0.009 0.044Duganella 0.710b 1.831a 0.832b 0.946b 0.018 0.018 0.018 0.088Massilia 11.361a 10.363a 1.752b 2.094b 0.324a 0.254ab 0.070b 0.377aVariovorax 0.482c 1.051bc 1.314ab 1.901a 0.070 0.018 0.070 0.044Chitinophaga 0.815b 2.461ab 3.968ab 5.019a 0.149ab 0.184ab 0.061b 0.140abBurkholderia 1.384b 1.585b 5.694a 5.168a 0.035c 0.096bc 0.175ab 0.298aStenotrophomonas 0.420 0.149 0.701 1.148 0.105 0.000 0.000 0.009Mucilaginibacter 1.585a 1.612a 0.876ab 0.491b 0.070ab 0.114a 0.000b 0.044abNiastella 0.841b 2.356a 0.596b 0.578b 0.018c 0.070bc 0.175a 0.123abPseudomonas 1.463a 0.333b 0.315b 0.333b 0.044 0.053 0.026 0.096Rhizobium 0.219b 0.307b 1.515a 0.999a 0.044b 0.053b 0.035b 0.158aDyella 1.594 2.120 3.618 3.723 0.193 0.491 0.149 0.736Streptomyces 1.016b 2.356a 1.384b 1.156b 0.175b 0.184bc 0.254b 0.543aFlavobacterium 3.136a 1.454ab 0.745b 0.368b 0.797a 0.324b 0.079b 0.114bLysobacter 0.237b 0.762a 0.473ab 0.631ab 0.175a 0.131ab 0.061b 0.140abSphingomonas 0.272b 0.315b 0.578b 0.893a 0.053b 0.061b 0.035b 0.420aArenimonas 1.892a 1.840a 0.526b 0.806b 0.972a 0.631ab 0.368b 0.368bOhtaekwangia 1.244a 1.515a 0.377b 0.350b 0.543a 0.657a 0.184b 0.298bMarmoricola 0.438 0.569 0.499 0.806 0.175b 0.175b 0.350b 0.622aBradyrhizobium 0.158b 0.359b 1.200a 1.148a 0.149b 0.359b 0.841a 0.745aNocardioides 0.788 0.552 0.797 0.797 0.298b 0.394b 0.491b 1.244a

(b) Depleted generaAcidobacteria Gp1 2.786a 2.032ab 1.472b 1.883ab 3.066ab 3.013ab 3.416a 1.761bRhizomicrobium 0.674ab 0.368b 0.350b 1.034a 1.174a 0.850ab 0.648b 0.745abTerrimonas 0.359 0.447 0.272 0.245 0.771a 0.561ab 0.674ab 0.385bGemmatimonas 2.488a 2.225a 1.130b 1.358b 3.986a 3.784a 3.285a 2.330bAcidobacteria Gp7 0.596a 0.280b 0.298b 0.403ab 0.797 0.736 0.788 0.745Acidobacteria Gp3 0.578 0.552 0.946 0.999 1.296b 1.209b 2.374a 1.621bAcidobacteria Gp6 1.638a 1.323ab 0.850b 0.780b 2.645 3.022 3.127 2.199Acidobacteria Gp10 0.228 0.193 0.105 0.114 0.482 0.438 0.596 0.508Pasteuria 0.298a 0.263a 0.158ab 0.088b 0.893a 1.060a 0.456b 0.429bNitrospira 0.911a 0.657ab 0.683ab 0.403b 3.591a 2.882b 2.225b 1.402cChthonomonas 0.307a 0.140b 0.280ab 0.193ab 1.156a 0.823ab 0.955ab 0.587bAcidobacteria Gp4 2.672a 2.628a 1.165b 1.349b 12.158a 8.847bc 7.665bc 5.019cKtedonobacter 0.026c 0.123bc 0.193ab 0.254a 0.201b 0.412b 0.727b 3.294a

(c) Enriched OTUs (28)OTU27365 (Chitinophaga) 0.000 0.044 1.375 0.000 0.000 0.000 0.000 0.000OTU5987 (Streptomyces) 0.000b 0.210ab 0.701a 0.350ab 0.000 0.000 0.000 0.000OTU5543 (Ralstonia) 0.070 0.657 6.473 0.920 0.000 0.000 0.018 0.000OTU5996 (Ralstonia) 0.026 0.201 2.742 0.438 0.009 0.000 0.000 0.000OTU5729 (Sphingobium) 0.000b 0.026b 1.156a 0.990a 0.000 0.000 0.000 0.009OTU6022 (Massilia) 5.764a 5.808a 0.639b 0.403b 0.009 0.009 0.009 0.026OTU3354 (Sphingobium) 0.000c 0.070c 1.656b 2.593a 0.000 0.009 0.009 0.009OTU5787 (Dyella) 0.044b 0.508ab 1.410a 0.841ab 0.009 0.009 0.000 0.000OTU3337 (Massilia) 1.866a 1.148b 0.403bc 0.710c 0.009 0.000 0.000 0.018OTU1352 (Sphingobium) 0.009c 0.053c 1.218b 2.689a 0.000 0.009 0.000 0.018OTU5682 (Burkholderia) 0.359 0.482 1.349 1.086 0.000 0.000 0.009 0.026OTU5806 (Massilia) 2.645a 2.680a 0.447b 0.622b 0.000 0.026 0.000 0.044OTU6053 (Dyella) 0.044b 0.272b 0.385b 0.745a 0.000 0.000 0.018 0.000OTU5568 (Variovorax) 0.070b 0.131b 0.412b 0.780a 0.009 0.000 0.000 0.009OTU2967 (Chitinophaga) 0.018b 0.166ab 0.438ab 0.674a 0.009 0.000 0.000 0.009OTU3498 (Unclassified Enterobacteriaceae) 0.517ab 0.044b 1.629a 0.333b 0.000 0.009 0.018 0.035OTU5591 (Burkholderia) 0.096b 0.158b 0.972a 0.464ab 0.000 0.000 0.009 0.035OTU1497 (Dyella) 0.718 0.797 1.612 1.918 0.035 0.070 0.009 0.026OTU348 (Duganella) 0.534b 1.568a 0.613b 0.788b 0.000b 0.018b 0.018b 0.070aOTU519 (Burkholderia) 0.315b 0.385b 1.743a 1.691a 0.018 0.053 0.044 0.035OTU3364 (Chitinophaga) 0.070c 0.613b 0.298bc 1.375a 0.009 0.018 0.026 0.044OTU5214 (Rhizobium) 0.105b 0.096b 0.780a 0.508a 0.000b 0.009b 0.009b 0.061aOTU5889 (Niastella) 0.298b 0.937a 0.140b 0.158b 0.009b 0.000b 0.061a 0.026abOTU3131 (Streptomyces) 0.666 0.858 0.526 0.683 0.018 0.018 0.061 0.088OTU165 (Arenimonas) 0.841a 0.631a 0.123b 0.333b 0.307 0.201 0.149 0.140OTU200 (Unclassified Burkholderiales incertae sedis) 0.456ab 0.631a 0.193b 0.219b 0.307a 0.088b 0.088b 0.140bOTU875 (Arenimonas) 0.499a 0.578a 0.184b 0.149b 0.324a 0.210ab 0.114b 0.079bOTU242 (Bradyrhizobium) 0.114b 0.254b 0.841a 0.815a 0.123c 0.228bc 0.526a 0.447ab

(d) Depleted OTUsOTU82 (Acidobacteria Gp4) 0.420a 0.403a 0.123b 0.140b 1.437a 1.226a 1.007ab 0.674b

OTU11 (Acidobacteria Gp4) 0.447a 0.272ab 0.158bc 0.061c 1.463a 1.340ab 1.165ab 0.736b

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401398

Table 2 (continued )

Time Relative abundance in Rh (%) Relative abundance in Bk (%)

Jun 1 Jun 21 Jul 18 Aug 11 Jun 1 Jun 21 Jul 18 Aug 11

OTU189 (Unclassified Chitinophagaceae) 0.131 0.096 0.044 0.149 0.692a 0.780a 0.429ab 0.263bOTU26 (Acidobacteria Gp4) 0.166b 0.464a 0.061b 0.105b 1.638a 1.445ab 0.692bc 0.359cOTU230 (Nitrospira) 0.140 0.114 0.053 0.044 0.771a 0.964a 0.272b 0.298b

Values labeled with different letters in a row indicated significant at p < 0.05.

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401 399

German rhizosphere samples. 4800 OTUs (13.9% of total numbers ofOTUs) were shared in both datasets, and they accounted for 70.3%of total reads in Illinois samples and 68.6% in German samples,indicating that most root-associated bacteria were shared in bothsites. In this combined dataset, the average relative abundance of 14OTUs in Illinois samples and 11 OTUs in German samples werehigher than 0.5% of total reads (Table SI_4). These OTUs weremainly affiliated to genera Massilia, Ralstonia, Sphingobium, Bur-kholderia and Duganella in Illinois samples, while Massilia, Arthro-bacter, Rhizobacter and Streptomyces in German samples. Four OTUsoccurred only in Illinois samples, including OTU3070b (Bur-kholderia), 30638b (Ralstonia), 2068b (Dyella) and 6727b (Dyella).Though most of the dominant genera were shared in both Illinoisand German samples, important OTUs in the same genus werequite different, for example, in the genus Massilia, OTU60b wasmore abundant in German samples, while five other OTUs(OTU1234b, 2383b, 298b, 4839b and 3722b) were more abundantin Illinois samples.

4. Discussion

We detected the dominant bacterial taxa enriched in the maizerhizosphere using the high-resolution pyrosequencing technique(Table 2), and we found that changes in the maize rhizospherecommunity composition were strongly linked to the growth stage

Fig. 4. Growth-stage related dynamic changes in relative abundances (% to to

of the plant (Fig. 4). Early growth-stage related communities wereclearly separated from later ones (Fig. 2). Overall, members of phylaProteobacteria (class Betaproteobacteria and Gammaproteobac-teria) and Bacteroidetes (class Sphingobacteria) were the mostenriched bacteria inmaize rhizosphere (Fig. 3). This is considered tobe a carbon-rich environment, and ecological theory suggests thatcopiotrophic behavior or r-selected populations should be welladapted to such nutrient-rich conditions. Many Betaproteobacteriaand Bacteroidetes are copiotrophic soil bacteria, and they becomeabundant when labile substrates are available (Fierer et al., 2007;Goldfarb et al., 2011). Our results confirmed that these organismspredominate in the rhizosphere, presumably because labile organicsubstrates are more available than in bulk soil. In addition, Bur-kholderiaceae, Xanthomonadales and Actinobacteria harbor generaand species with anti-microbial activities (Mendes et al., 2011),which may enable them to operate as interference competitors inthe rhizosphere.

The DGGE profiles in this study revealed a shift in a few domi-nant bacterial populations over the course of maize growth (Fig. 1).Because DGGE showed similar dynamic patterns as those revealedby pyrosequencing, it is a useful technique to give a snapshot of thediversity of the most abundant rhizosphere bacteria (Gomes et al.,2001; Smalla et al., 2001). However, pyrosequencing results pro-vided more specific insight into shifts in bacteria communitycomposition at the levels of family, genus and OTU (Fig. 4). Dynamic

tal reads) of dominant bacterial groups at family, genus and OTU levels.

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401400

changes of many bacterial groups in the rhizospherewere similar tothose in bulk soil at high taxonomic ranks; for example classActinobacteria (mainly order Actinomycetales) and order Rhizo-biales (Fig. SI_2) both displayed similar temporal shifts in therhizosphere and bulk soil. It may be that seasonal climate changeimpacts these bacteria in both rhizosphere and bulk soils.

However, there were several key microbial groups (at the levelof family and below) that displayed different temporal dynamics inthe rhizosphere from that in the bulk soil. In fact, many of thedominant rhizosphere genera (Massilia, Chitinophaga, Sphingobium,etc.) displayed different patterns of temporal changes in therhizosphere as opposed to the bulk soil (Fig. 4), andwe propose thatthe dynamics of these rhizosphere bacteria were largely plant-driven and not due to direct seasonal effects. It can be expectedthat the composition and chemical nature of root exudates willchange over the plant growth cycle (Dunfield and Germida, 2003),and these changes should alter the nutritional environment ofrhizosphere bacteria. Bacteria present in the rhizosphere of youngplants preferentially utilize simple amino acids, whereas bacteriafrom mature plants utilize more complex carbohydrates (Houldenet al., 2008). It is also recognized that copiotrophic behavior or r-selected populations are positively selected in early stages of mi-crobial succession, while oligotrophic or k-selected populations arepositively selected in later stages (Fierer et al., 2007). Our resultssupport the conclusion that dominant bacteria species in the earlygrowth-stage communities are characteristic of fast-growing,copiotrophic bacteria that are competent at utilizing labileorganic substrates.

In this study, we found that significant growth-related dy-namic changes in bacterial community structure were mainlyassociated with phylum Proteobacteria or Bacteroidetes, espe-cially order Burkholderiales (mainly genera Massilia, Burkholderiaand Variovorax) (Fig. 4 and Fig. SI_2). Members of Burkholderialeswere enriched in the rhizosphere, possibly due to their versatileabilities to utilize root metabolites, degrade aromatic compounds(Goldfarb et al., 2011; Vandenkoornhuyse et al., 2007) and pro-duce anti-microbial substances (Coenye and Vandamme, 2003). Inthis order, genera Massilia predominated at the early stages ofsuccession, while Burkholderia and Variovorax increased at latergrowth stages (Fig. 4, Fig. SI_5). In addition to Massilia, the generaFlavobacterium, Arenimonas and Ohtaekwangia also appeared atearly stages of succession, indicating that they may be copio-trophic and fast-growth bacteria, able to exploit a transient nichesfor nutrition and others at the early growth stage. Ofek et al. alsoreported that the dominance of Massilia may be attributed to itshigh proliferation rates when sufficient carbon and energy sour-ces are present (Ofek et al., 2012). The decline of Massilia at laterstages may be explained by the reduction of quantity and qualityof organic carbon, or interspecies competition for resources. Therelative abundance of genera Chitinophaga and Sphingobiumsharply increased at later stages in similar proportion to thegenus Burkholderia. Members of genus Chitinophaga are chitino-lytic (Kim and Jung, 2007), while Sphingobium and Burkholderiacan typically degrade lignin-derived aromatic compounds(Takeuchi et al., 2001; Vandenkoornhuyse et al., 2007). The abilityof Burkholderia, Chitinophaga and Sphingobium to use polymersmay enable them to be more competitive for colonization ofmature roots.

Although bulk soil is recognized to be highly stable in commu-nity structure and function (Houlden et al., 2008), our pyrose-quencing results also revealed significant seasonal changes of somesoil genera and OTUs (Fig. 4, Fig. SI_6). For example, the relativeabundance of genera Gp4 (Acidobacteria), Nitrospira, Pasteuria,Ohtaekwangia and Arenimonas decreased late in the growth season,while Ktedonobacter and Bradyrhizobium increased.

In addition to changes in rhizosphere community compositionover time, we also documented variability that may be attributableto environmental conditions. The comparison of our results, ob-tained from a fine-silty soil in Illinois, USA, to those from a silty sandsoil in Germany (Dohrmann et al., 2013) showed significantlydifferent bacterial communities at the level of family and below(Table SI_4). The number of reads belonging to shared OTUsaccounted for about 70% of total bacterial reads in samples fromboth sites, and this suggests that the maize rhizosphere generallyselects for similar dominant members independent of variety, ageor soil type. However, the number of shared OTUs only accountedfor 13.9% of total observed OTUs, and this implies that most of site-specific OTUs are rare species. The differences in bacterial com-munity structure of maize rhizosphere between Illinois and Ger-many were larger than the differences between cultivars (Germansamples), root microhabitats (German samples) or growth stages(Illinois; Fig. SI_7). Another study conducted in the carbonate-richsoil (pH 8.5) using 16S rRNA clone library analysis showed thatPseudomonas and Lysobacter genera constituted 45% of the totalbacterial abundance in the maize rhizosphere (García-Salamancaet al., 2013). Though exact mechanisms for such differences arenot completely understood, because these studies were all con-ducted on different soil types, it is likely that soil typemay be one ofthe most important environmental variables influencing thestructure of the maize root-inhabiting bacterial community at lowtaxonomic levels (Berg and Smalla, 2009; Castellanos et al., 2009;Garbeva et al., 2004). Growth-stage related dynamics and the dif-ferences in rhizosphere bacterial community structure fromdifferent soil types further support the general concept that plantand soil type cooperatively shape the structure of microbial com-munity in the rhizosphere (Berg and Smalla, 2009).

In conclusion, the pyrosequencing surveys in this studyrevealed, in detail, the composition and dynamics of bacterialcommunity structure in the maize rhizosphere. These communitiesare dominated by members of the genera Massilia, Burkholderia,Ralstonia, Dyella, Chitinophaga and Sphingobium. Many of theserhizosphere bacteria, especially order Burkholderiales, showedactive, dynamic changes with root growth stage, and there appearsto be, in general, an early growth stage community that is distinctfrom those of later stages. The comparisons of bacterial communitystructures from Illinois soil and Germany soil supported the notionthat soil types may be very important environmental variablesinfluencing the structure of the maize root-inhabiting bacterialcommunity. Future investigation is warranted into whether thepattern of growth-related dynamics in rhizosphere bacteria com-munity is specific with particular cultivars, root segments and/orsoil types.

Acknowledgments

This work was funded by the Energy Biosciences Institute,Environmental Impact and Sustainability of Feedstock ProductionProgram at the University of Illinois, Urbana. We thank Dr. BenliChai for the help in data treatment.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.soilbio.2013.10.017.

References

Aira, M., Gomez-Brandon, M., Lazcano, C., Baath, E., Dominguez, J., 2010. Plant ge-notype strongly modifies the structure and growth of maize rhizosphere mi-crobial communities. Soil Biol. Biochem. 42, 2276e2281.

X. Li et al. / Soil Biology & Biochemistry 68 (2014) 392e401 401

Baker, G.C., Smith, J.J., Cowan, D.A., 2003. Review and re-analysis of domain-specific16S primers. J. Microbiol. Methods 55, 541e555.

Bent, S.J., Forney, L.J., 2008. The tragedy of the uncommon: understanding limita-tions in the analysis of microbial diversity. ISME J. 2, 689e695.

Berg, G., Smalla, K., 2009. Plant species and soil type cooperatively shape thestructure and function of microbial communities in the rhizosphere. FEMSMicrobiol. Ecol. 68, 1e13.

Castellanos, T., Dohrmann, A.B., Imfeld, G., Baumgarte, S., Tebbe, C.C., 2009.Search of environmental descriptors to explain the variability of the bacterialdiversity from maize rhizospheres across a regional scale. Eur. J. Soil Biol. 45,383e393.

Cavaglieri, L., Orlando, J., Etcheverry, M., 2009. Rhizosphere microbial communitystructure at different maize plant growth stages and root locations. Microbiol.Res. 164, 391e399.

Chelius, M.K., Triplett, E.W., 2001. The diversity of archaea and bacteria in associ-ation with the roots of Zea mays L. Microb. Ecol. 41, 252e263.

Coenye, T., Vandamme, P., 2003. Diversity and significance of Burkholderia speciesoccupying diverse ecological niches. Environ. Microbiol. 5, 719e729.

Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., Kulam-Syed-Mohideen, A.S., McGarrell, D.M., Marsh, T., Garrity, G.M., Tiedje, J.M., 2009. TheRibosomal Database Project: improved alignments and new tools for rRNAanalysis. Nucleic Acids Res. 37, D141eD145.

Di Cello, F., Bevivino, A., Chiarini, L., Fani, R., Paffetti, D., Tabacchioni, S., Dalmastri, C.,1997. Biodiversity of a Burkholderia cepacia population isolated from the maizerhizosphere at different plant growth stages. Appl. Environ. Microbiol. 63,4485e4493.

Dohrmann, A.B., Kuting, M., Junemann, S., Jaenicke, S., Schluter, A., Tebbe, C.C., 2013.Importance of rare taxa for bacterial diversity in the rhizosphere of Bt- andconventional maize varieties. ISME J. 7, 37e49.

Dunfield, K.E., Germida, J.J., 2003. Seasonal changes in the rhizosphere microbialcommunities associated with field-grown genetically modified canola (Brassicanapus). Appl. Environ. Microbiol. 69, 7310e7318.

Edgar, R.C., 2004. MUSCLE: a multiple sequence alignment with high accuracy andhigh throughput. Nucleic Acids Res. 32, 1792e1797.

Edgar, R.C., Haas, B.J., Clemente, J.C., Quince, C., Knight, R., 2011. UCHIME improvessensitivity and speed of chimera detection. Bioinformatics 27, 2194e2200.

Fierer, N., Bradford, M.A., Jackson, R.B., 2007. Toward an ecological classification ofsoil bacteria. Ecology 88, 1354e1364.

Garbeva, P., van Veen, J.A., van Elsas, J.D., 2004. Microbial diversity in soil: selectionmicrobial populations by plant and soil type and implications for diseasesuppressiveness. Annu. Rev. Phytopathol. 42, 243e270.

García-Salamanca, A., Molina-Henares, M.A., van Dillewijn, P., Solano, J., Pizarro-Tobías, P., Roca, A., Duque, E., Ramos, J.L., 2013. Bacterial diversity in therhizosphere of maize and the surrounding carbonate-rich bulk soil. Microb.Biotechnol. 6, 36e44.

Goldfarb, K.C., Karaoz, U., Hanson, C.A., Santee, C.A., Bradford, M.A., Treseder, K.K.,Wallenstein, M.D., Brodie, E.L., 2011. Differential growth responses of soil bac-terial taxa to carbon substrates of varying chemical recalcitrance. Front.Microbiol. 2, 94.

Gomes, N.C.M., Heuer, H., Schönfeld, J., Costa, R., Mendonça-Hagler, L., Smalla, K.,2001. Bacterial diversity of the rhizosphere of maize (Zea mays) grown intropical soil studied by temperature gradient gel electrophoresis. Plant Soil 232,167e180.

Hanway, J.J., 1963. Growth stages of corn (Zea mays, L.). Agron. J. 55, 487e492.Houlden, A., Timms-Wilson, T.M., Day, M.J., Bailey, M.J., 2008. Influence

of plant developmental stage on microbial community structure and

activity in the rhizosphere of three field crops. FEMS Microbiol. Ecol. 65,193e201.

Kim, M.K., Jung, H.-Y., 2007. Chitinophaga terrae sp. nov., isolated from soil. Int. J.Syst. Evol. Microbiol. 57, 1721e1724.

Lozupone, C., Lladser, M.E., Knights, D., Stombaugh, J., Knight, R., 2011. UniFrac: aneffective distance metric for microbial community comparison. ISME J. 5,169e172.

Marschner, P., Crowley, D., Yang, C.H., 2004. Development of specific rhizospherebacterial communities in relation to plant species, nutrition and soil type. PlantSoil 261, 199e208.

Marschner, P., Yang, C.H., Lieberei, R., Crowley, D.E., 2001. Soil and plant specificeffects on bacterial community composition in the rhizosphere. Soil Biol. Bio-chem. 33, 1437e1445.

Mendes, R., Kruijt, M., de Bruijn, I., Dekkers, E., van der Voort, M., Schneider, J.H.M.,Piceno, Y.M., DeSantis, T.Z., Andersen, G.L., Bakker, P.A.H.M., Raaijmakers, J.M.,2011. Deciphering the rhizosphere microbiome for disease-suppressive bacte-ria. Science 332, 1097e1100.

Mougel, C., Offre, P., Ranjard, L., Corberand, T., Gamalero, E., Robin, C., Lemanceau, P.,2006. Dynamic of the genetic structure of bacterial and fungal communities atdifferent developmental stages of Medicago truncatula Gaertn. cv. Jemalong lineJ5. New Phytol. 170, 165e175.

Muyzer, G., de Waal, E.C., Uitterlinden, A.G., 1993. Profiling of complex microbialpopulations by denaturing gradient gel electrophoresis analysis of polymerasechain reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol.59, 695e700.

Nacamulli, C., Bevivino, A., Dalmastri, C., Tabacchioni, S., Chiarini, L., 1997. Pertur-bation of maize rhizosphere microflora following seed bacterization with Bur-kholderia cepacia MCI 7. FEMS Microbiol. Ecol. 23, 183e193.

Ofek, M., Hadar, Y., Minz, D., 2012. Ecology of root colonizing Massilia (Oxalo-bacteraceae). PLoS One 7, e40117.

Price, M.N., Dehal, P.S., Arkin, A.P., 2010. FastTree 2 e approximately maximum-likelihood trees for large alignments. PLoS One 5, e9490.

Savka, M.A., Farrand, S.K., 1997. Modification of rhizobacterial populations by en-gineering bacterium utilization of a novel plant-produced resource. Nat. Bio-technol. 15, 363e368.

Schmalenberger, A., Tebbe, C.C., 2002. Bacterial community composition in therhizosphere of a transgenic, herbicide-resistant maize (Zea mays) and com-parison to its non-transgenic cultivar Bosphore. FEMS Microbiol. Ecol. 40,29e37.

Singh, B.K., Millard, P., Whiteley, A.S., Murrell, J.C., 2004. Unravelling rhizosphere-microbial interactions: opportunities and limitations. Trends Microbiol. 12,386e393.

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.

Takeuchi, M., Hamana, K., Hiraishi, A., 2001. Proposal of the genus Sphingomonassensu stricto and three new genera, Sphingobium, Novosphingobium andSphingopyxis, on the basis of phylogenetic and chemotaxonomic analyses. Int. J.Syst. Evol. Microbiol. 51, 1405e1417.

Vandenkoornhuyse, P., Mahé, S., Ineson, P., Staddon, P., Ostle, N., Cliquet, J.-B.,Francez, A.-J., Fitter, A.H., Young, J.P.W., 2007. Active root-inhabiting microbesidentified by rapid incorporation of plant-derived carbon into RNA. Proc. Natl.Acad. Sci. U.S.A. 104, 16970e16975.

Vercellino, M., Gómez, M., 2013. Denitrifying capacity of rhizobial strains ofArgentine soils and herbicide sensitivity. Ann. Microbiol. 63, 1e8.