Research Article Bacterial Community of the Rice Floodwater Using Cultivation-Independent ...

14 Pages·2017·2.47 MB·English

Hindawi
International Journal of Microbiology
Volume 2018, Article ID 6280484, 13 pages
https://doi.org/10.1155/2018/6280484
Research Article
Bacterial Community of the Rice Floodwater Using
Cultivation-Independent Approaches
MichelePittol ,1ErinScully,2DanielMiller,3LisaDurso,3
LidiaMarianaFiuza,1andVictorHugoValiati1
1ProgramadePo´s-Gradua¸ca˜oemBiologia,UniversidadedoValedoRiodosSinos(UNISINOS),950UnisinosAvenue,
Sa˜oLeopoldo,RS,Brazil
2UnitedStatesDepartmentofAgriculture(USDA),AgriculturalResearchService(ARS),CenterforGrainandAnimalHealthResearch,
StoredProductInsectandEngineeringResearchUnit(SPIERU),1515CollegeAve.,Manhattan,KS,USA
3UnitedStatesDepartmentofAgriculture(USDA),AgriculturalResearchService(ARS),
AgroecosystemManagementResearchUnit(AMRU),251FilleyHall,UNLEastCampus,Lincoln,NE,USA
CorrespondenceshouldbeaddressedtoMichelePittol;[email protected]
Received 2 August 2017; Revised 9 December 2017; Accepted 26 December 2017; Published 30 January 2018
AcademicEditor:ClemenciaChaves-Lo´pez
Copyright©2018MichelePittoletal.ThisisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense,
whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
Inagriculturalsystems,interactionsbetweenplantsandmicroorganismsareimportanttomaintainingproductionandprofitability.
Inthisstudy,bacterialcommunitiesinfloodwatersofricefieldsweremonitoredduringthevegetativeandreproductivestages
of rice plant development using 16S amplicon sequencing. The study was conducted in the south of Brazil, during the crop
years2011/12and2012/13.Comparativeanalysesshowedstrongdifferencesbetweenthecommunitiesoffloodwatersassociated
withthetwodevelopmentalstages.Duringthevegetativestage,1551operationaltaxonomicunits(OTUs)weredetected,while
lessthanhalfthatnumber(603)wereidentifiedinthereproductivestage.Thehigherbacterialrichnessobservedinfloodwater
collected during the vegetative stage may have been favored by the higher concentration of nutrients, such as potassium, due
torhizodepositionandfertilizerapplication.Eighteenbacterialphylawereidentifiedinbothsamples.Bothcommunitieswere
dominatedbyGammaproteobacteria.Inthevegetativestage,AlphaproteobacteriaandBetaproteobacteriaweremoreabundant
and,incontrast,BacilliandClostridiawerethemoredominantclassesinthereproductivestage.Themajorbacterialtaxaidentified
havebeenpreviouslyidentifiedasimportantcolonizersofricefields.Therichnessandcompositionofbacterialcommunitiesover
cultivationtimemaycontributetothesustainabilityofthecrop.
1.Introduction varyingphysicalandchemicalpropertiesinthisenvironment
could support the growth of microorganisms possessing
Rice (Oryza sativa L.) is the most important cereal crop in widerangesofmetabolicplasticity,allowingthemtoquickly
theworld,feedingmorethan50%ofthehumanpopulation. adapt to changing environmental conditions [4]. Thus, the
In South America, Brazil is the main producer [1], with rice ecosystem may be a prime habitat for microorganisms
floodedricefieldsaccountingfor50%ofthetotalcroparea adaptedtofluctuatingnutritionallevelsandoxygenandlight
undercultivationinthecountry[2].Infloodedrice,theneed availability.
to maintain adequate water depth throughout most of the The phyla Acidobacteria, Actinobacteria, Bacteroidetes,
cropyearcharacterizestheagriculturalsystemasaquaticin Chloroflexi,andProteobacteriahavebeenpreviouslyfound
nature.Incomparisontootheraquaticenvironments,suchas in soil samples of rice-alfalfa [5] and rice-wheat cropping
lakes, ponds, and swamps, the environmentalconditionsin [6]. In these agroecosystems, rice exudates and nutrients
floodedricefieldsarerelativelyunstableduetophysicaland fromstraw incorporationwere shown to influencethe bac-
chemical and biological characteristics that vary according terial community’s composition. Breidenbach and Conrad
tocurrentagriculturalpracticesandwatersupplies[3].The [7] found a uniform bacterial composition in soil over the
2 InternationalJournalofMicrobiology
rice growing season, with Proteobacteria being the most searchfortherichnessandabundanceofthelocalmicrobiota.
highly abundant phylum, while Firmicutes represented the Afterwards, the 16 subsamples from each stage within a
fifth most abundant phylum. Firmicutes were also present locationwerecombinedintoasinglecompositewatersample
in relatively low abundance upon the introduction of a of 800mL for DNA isolation. The 12 samples representing
maizerotationintoanirrigatedricefield[8].Althoughthese twolifestagesfromthreedifferentlocationswerecentrifuged
microorganismsarecommoninhabitantsofagriculturesoils, at12,000×gfor15minat4∘Candthepelletswerestoredat
thelowcountsofFirmicutesareintriguingconsideringthat −20∘CforsubsequentDNAextractionandanalysis.
thisgroupcontainstheclassesBacilliandClostridia,which
are often highly abundant in rice agricultural soils where 2.2. DNA Extraction and Pyrosequencing. Total DNA was
theydecomposeplantresiduesusingcellulolyticenzymes[9]. extracted from each of the 12 pellets collected from the
Furthermore, the researchers observed higher bacterial 16S floodwater samples from two life stages at three locations
rRNAabundanceinthefloodingstagethanindrainagestage, using the MoBio Power Soil(cid:2) DNA Isolation Kit (MoBio,
whichwasascribedtothericestrawthatremainedinthefield LaboratoriesInc.,Carlsbad,CA,USA)followingthemanu-
[10]. facturer’sinstructions.TheDNAyieldwasmeasuredusinga
In the studies mentioned above, the shifts in microbial NanoDrop(cid:2)ND-1000spectrophotometer(NanoDropTech-
community structure were governed by environmental fac- nologies, Inc. Wilmington, DE, USA) and the DNA quality
tors found in rice fields. These factors can be related to the was verified on a 1.5% agarose gel stained with ethidium
input of nutrients from water sources, rice residues, and −1
bromide(0.1mgmL ).
fertilizers,weatherconditions,andcroprotation,highlight-
Purified DNA from the six replicates collected during
ing the complex network of elements that govern bacterial
the vegetative stage and the six replicates collected during
structure in this agricultural system. Although microbial
the reproductive stage were pooled together to create a
ecologystudiesinsoil,rhizosphere,andendophyticenviron-
single DNA pool for the vegetative samples and a single
ments have been previously conducted in rice systems [11–
DNA pool for the reproductive samples. These two DNA
14], the dynamics of bacterial communities in floodwaters
pools(oneDNApoolfromvegetativesamplesandoneDNA
from rice fields have not been extensively studied despite
pool from reproductive samples) were used as templates
the important roles of the floodwaters and microbial com-
forampliconpyrosequencinganalysis,whichwasperformed
munities in rice plant nutrition [15]. Because microbes can
by MR DNA Laboratory (Shallow Water, TX, USA), with
contribute strongly to rice nutrition and production, there
the goal of investigating the bacterial community structure
ismuchinterestinsurveyingthemicrobialcommunitiesin
and composition. The V1–V3 variable regions of the 16S
this flooded agroecosystem and determining how agricul- 󸀠
rRNA gene were amplified using primers ill27Fmod (5 -
tural practices influence the composition and structure of 󸀠 󸀠
AGRGTTTGATCMTGGCTCAG-3 ) and ill519Rmod (5 -
these communities [16]. The purpose of the present study 󸀠
GTNTTACNGCGGCKGCTG-3 ) [19]. The PCR reactions
was to investigate the bacterial community structure and werepreparedwith1𝜇LofDNA(5ng𝜇L−1)usingHotStar-
composition in floodwaters associated with vegetative and
TaqPlusMasterMix(Qiagen,Valencia,CA,USA).DNAwas
reproductivestagesofriceinordertoverifyifthereisamicro- ∘
initiallydenaturedat94 Cfor3minfollowedby28cyclesof
bialcommunityrelatedtoaparticularstageofthericesystem. ∘ ∘ ∘
94 C for 30s; 53 C for 40s and 72 C for 1min; and a final
∘
extensionat72 Cfor5min.PCRproductsofapproximately
2.MaterialsandMethods 300bpwerepurifiedusingAgencourtAmpurebeads(Agen-
court Bioscience Corporation, MA, USA). The amplicons
2.1. Study Area and Sampling Protocols. Three isolated
were sequenced using 454 GS FLX Titanium chemistry
floodedricefieldswereselectedforthisstudy.Thericefields
(Roche)followingthemanufacturer’sguidelines.Theraw454
werebetween164and310hainareaandarelocatedinthecity
readsgeneratedduringthecurrentstudyhavebeendeposited
∘ 󸀠 󸀠󸀠 ∘ 󸀠 󸀠󸀠
ofViama˜o(30 04 51 S,51 01 22 W),outercoastalplain,Rio
inNCBI’sSequenceReadArchive(SRA)undertheaccession
GrandedoSul(RS),Brazil,Figure1.
SRP077313andareassociatedwithBioProjectPRJNA326968.
Twelve water samples were collected from the three
locationsaboveintheagriculturalyears2011/12and2012/13, 2.3. Protocols and Equipment Used for the Physicochemical
sixinthevegetativeandsixinthereproductivestages.One Measurements of Water. Water parameters such as temper-
sample from each location and developmental stage was ature,turbidity,andpHwere measuredfromall12 samples
collected during the 2011/12 growing season and another usingtheIntelligentMeterequipment(INSTRUTHERMPH-
sample was collected during the 2012/13 growing season. 1300).Inaddition,nutrientlevels,includingtotalphosphorus
The collections were organized as follows: the vegetative (P)andpotassium(K),weremeasuredfromeachofthesix
(tillering) (October/November) stage was defined by active water samples collected from the two developmental stages
tilleringandcoleoptileandradicle(2-3mm)formationtothe using standard methods [18]. The available P and K were
beginning of panicle differentiation, while the reproductive determined using Mehlich-1 solution by molecular colori-
stagewasdefinedfromtheformationuntilcompletematura- metricandflameemissionmethods,respectively[20,21].
tionofthepanicle(DecembertoMarch)[17].Eachofthe12
individualwatersamplescollectedrepresentedacombination 2.4. Data Analysis. The 454 amplicon data from bacterial
of16subsamplesof50mLcollectedfromthetopsurfaceof communitiessampledfromfloodwatersfromthereproduc-
thefloodwaterto10cmindepth[18],therebymaximizingthe tive(𝑛=1)andvegetative(𝑛=1)stageswereanalyzedusing
InternationalJournalofMicrobiology 3
７ ７ ７ ７ ７ ７
７ ７
0 15 30 45 0 15 30 45
8 6 4 2 1 9 7 5
5 5 5 5 5 4 4 4
∘50 ∘50 ∘50 ∘50 ∘50 ∘50 ∘50 ∘50 80∘00７ 60∘00７ 40∘00７
0∘00 0∘00
29∘5315３
23∘270３ 23∘270３
29∘550３
46∘540３ 46∘540３
29∘5645３
80∘00７60∘00７ 40∘00７
(a)
29∘5830３
56∘500７ 53∘100７ 49∘300７
30∘015３ 27∘300３ 27∘300３
30∘20３
31∘100３ 31∘100３
30∘345３
0 2.500 5.(0m00) 10.000 56∘500７ 53∘100７ 49∘300７
(b)
７ ７ ７ ７ ７ ７ ７ ７
0 5 0 5 0 5 0 5
58 61 43 24 51 91 73 54
∘50 ∘05 ∘05 ∘05 ∘50 ∘04 ∘04 ∘04
5 5 5 5 5 5
(c)
Figure1:(a)LocationofRioGrandedoSulstate,SouthernofBrazil.(b)Viama˜ocity,outercoastalplainofRioGrandedoSulstate;(c)study
areawiththethreesamplingsitesinthefloodedricefields.
the mothur pipeline (version 1.32.1) [22]. In brief, flow- taxonomy for each OTU was determined using the clas-
grams were demultiplexed using the trim.flows command sify.otucommand.Rarefactioncurveswerecalculatedusing
and denoised using the shhh.flows command, which is therarefaction.singlecommandatageneticdistanceof0.03,
themothurimplementationofPyroNoise[23]. In addition, while the summary.single command was used to compute
adapters,barcodes,primers,shortreads,andreadscontain- various community metrics including the Shannon-Wiener
󸀠
ingmorethaneighthomopolymerbaseswereremovedfrom (H )andSimpsonindices,theChao1richnessestimator[26],
the dataset with trim.seqs command. Cleaned sequences which estimates species richness based on the number of
were aligned to the Silva reference database [24] using the rareOTUsdetectedineachcommunity,andtheAbundance
Needlemanalignerwiththealign.seqscommand.Chimeric Coverage-basedEstimator(ACE),whichestimatesthenum-
sequenceswereflaggedandremovedfromthedatasetusing berofundetectedOTUsbasedontheabundanceanddistri-
theuchimealgorithmwiththereference=self-flag[25].Then, butionofrareOTUs.Priortocomputingrarefactioncurves
the remaining high quality sequences were taxonomically andcommunitydiversity/richnessmetrics,thesamenumber
classified using the Wang method of the classify.seqs com- ofreads(7,348)weresubsampledwithoutreplacementfrom
mand and the Silva reference taxonomy files. An 80% con- eachofthelibrariesfornormalizationandtopreventdiffer-
fidence score was required for all taxonomic assignments. encesinlibraryyieldsfromdrivingdifferencesbetweenthe
Ampliconsclassifiedasmitochondrial,chloroplast,Archaeal, twocommunities.ThemothurcommandLibshuffwasused
eukaryotic, or unknown in origin were removed from the to determine whether the bacterial communities from the
dataset. vegetative and reproductive stages have the same structure
Theremainingbacterialsequenceswerecategorizedinto withiterssetto1000.
operationaltaxonomicunits(OTUs)at3%sequencedissimi- Heatmapsandcluster dendrogramsbasedonEuclidean
larity using the average neighbor algorithm. The consensus dissimilarity metrics were built according to the relative
4 InternationalJournalofMicrobiology
−3 −2 −1 0 1 2 3 physicalandchemicalvariationsoffloodwaterofricefields
in vegetative and reproductive stage of two consecutive
0.6 3
agricultural years. The physical and chemical patterns of
floodwater retrieved from the vegetative and reproductive
0.4 Temperature 2 ricestagesareshowninFigure2.
Thefirsttwoprincipalcomponentsexplainedacombined
0.2 K 1 total of 68% of the total variance. Most water samples
%
5 collected from the vegetative stage were separated from the
1.
axis 2 0.0 P 0 rdeipffreorednuccetsivebesttwageeensamlocpaletisonalsonogr tahgeriPcuCl1tuarxails.yeNaorsmwaejorer
2
PC −0.2 −1 noted. In general, vegetative stage samples tended to have
higherlevelsofpH,K,andPcomparedtothereproductive
−0.4 pH −2 samples.Incontrast,thereproductivesamplestendedtobe
Turbidity more turbid and had higher temperatures compared to the
vegetativesamples.Inaddition,somevariabilitywasapparent
−0.6 −3
withinthewatersamplescollectedfromeachdevelopmental
stage,whichislikelyreflectiveofthedynamicnatureofthe
−0.6 −0.4 −0.2 0.0 0.2 0.4 0.6
physicalandchemicalvariablesinricefloodwaters(Figure2).
PC 1 axis 46.5% Temperature and K differed significantly between the two
Vegetative ricedevelopmentstages.Specifically,thetemperaturewason
Reproductive averagethreedegreeswarmerinfloodwatersassociatedwith
thereproductivestage.Kwasalmostfourtimeshigherinthe
Figure2:PCAbiplotofphysicalandchemicalparameterssampled
vegetativecomparedtothereproductivestage(Table1).
fromricefloodwatersfromvegetativeandreproductivestages.
3.2.CompositionoftheBacterialCommunityinFloodwaters
abundancesofthemostcommongenerainthetwocommu- over Two Stages of Rice Growth. The pyrosequencing-based
nitiesusingthe“vegan,”“heatmap.2,”and“cluster”packages analysis generated 25,981 amplicons from the vegetative
[27] in the R statistical computing environment (version stage samples and 18,849 amplicons from the reproductive
3.1.0).Toselectthemostabundantgenera,readscountsfor stage samples. After quality filtering and dereplication of
allOTUsassignedtothesamegenusfrombothsampleswere identicalsequences,8,483readswereretainedfromthewater
summed together, normalized by relative abundance, and samples collected during vegetative stage and 7,348 reads
rankedinascendingorder.Therelativeabundancesofthese wereretainedfromsamplescollectedduringthereproductive
genera in floodwaters associated with each developmental stage. For normalization purposes, 7,348 reads were ran-
stagewerelog2 transformed,𝑧-scalestandardized,andused domlysubsampledwithoutreplacementfromthevegetative
forclusteringandheatmapanalysis.Venndiagramsdepicting librarypriortoOTUanalysisandtaxonomicclassification.
OTUsfoundinfloodwatersfrombothdevelopmentalstages The structures of the bacterial community differed sig-
and and OTUs unique to each developmental stage were nificantlybetweenthevegetativeandreproductivestagesof
constructedusingthevenncommandinmothur. flooded rice (𝑝 = 0,0001) using Libshuff. In the vegetative
The physical and chemical characteristics of the flood- stage, 1,551OTUs were identified, while less than half the
water from the rice fields sampled during two agricultural number of OTUs (603) were detected in water samples
yearswereanalyzedbyprincipalcomponentsanalysis(PCA) collected from the reproductive stage (Table 2). Further,
∘
using the following variables: temperature ( C), turbidity only 88 (4.08%) of the OTUs detected in this analysis were
(NTU), pH, total phosphorus (P) (mgL−1), and potassium commontobothcropstages,possiblyreflectingthedynamic
(K)(mgL−1).Priortoanalysis,allvariableswerelog10trans- structure of the bacterial communities associated with this
formed in order to facilitate comparisons among variables environment(Figure3).
expressedindifferentunits.Thecorrelationmatrixwiththe The sequences were distributed among 18 different bac-
varimaxrotationwassetandonlyautovaluesgreaterthan1 terial phyla. The most species-rich phyla in both the repro-
wereutilizedascriteriaforextractionoftheprincipalcom- ductive and vegetative stages were Proteobacteria, which
ponents. The analysis was performed in the R statistical contained 53.34% of the OTUs detected in both sam-
computing environment (version 3.1.0) using the princomp ples, Bacteroidetes (16.80% of the OTUs), Actinobacteria
andbiplotfunctions.Thedifferencesamongphysicochemical (7.84%),andFirmicutes(5.52%).Intotal,83.50%ofthetotal
parametersbetweenthericestagesweretestedusingStudent’s OTUs detected in these two communities was associated
𝑡-testinSYSTAT12(SystatSoftware,Inc.,CA,USA).𝑝values with these four phyla. Several phyla with lower richness
of<0.05wereconsideredstatisticallysignificant. estimates were also detected in these samples, defined as
phyla containing less than 1% of the total OTUs detected
3.Results in this analysis. Based on this criterion, a total of twelve
rare phyla were detected including Chloroflexi, Deferrib-
3.1.PhysicalandChemicalCharacteristicsofRiceFieldWater. acteres, Deinococcus-Thermus, Fusobacterium, Gemmati-
Principal components analysis was used to monitor the monadetes, OP10, Planctomycetes, SR1, Spirochaetes, TM7,
InternationalJournalofMicrobiology 5
Table1:Meanvaluesandcorrespondingstandarddeviations,forthephysicochemicalparametersmeasuredinthewatersamplesinrelation
tothericestages.
Ricestages Temperature∗(∘C) pH Turbidity(NTU) TotalP(mg/L) TotalK(mg/L)∗
30.20 6.50 4.30 0.65 14.70
24.06 6.21 34.16 0.44 31.75
22.73 7.43 68.83 0.84 12.34
Vegetative 29.40 6.70 39.86 0.01 11.77
23.60 6.50 37.56 0.02 4.29
27.20 7.43 42.20 1.30 11.59
Mean±SD 26.20±3.18 6.80±0.51 37.82±20.60 0.54±0.49 14.41±9.19
28.70 6.17 108.00 0.05 4.50
31.30 6.70 41.80 0.05 4.20
27.30 6.20 253.20 0.01 1.89
Reproductive 29.20 6.43 93.10 NA 6.10
29.60 6.50 27.55 0.17 5.04
30.80 6.50 391.00 0.14 1.62
Mean±SD 29.48±1.44 6.42±0.20 152.44±141.70 0.08±0.06 3.89±1.77
NA:notavailable;SD:standarddeviation;∗meansdifferedsignificantly(𝑝<0.05).
󸀠
Table2:Richnessestimators(Chao1andAce)anddiversityindicesofShannon-Wiener(H)andSimpsonfrombacterialcommunities
sampledfromfloodwatersduringreproductiveandvegetativestagesofriceusing16Samplicons.A3%nucleotidedissimilaritywasusedfor
OTUdiscrimination.
Numberofsequences 3%divergence
Cropstage
󸀠
Rawsequences Analysedsequences Normalization OTUs Chao1 Ace Shannon-Wiener(H) Simpson
Vegetative 25,981 8,483 7,348 1,551 2,282 2,266 5.77 0.03
Reproductive 18,849 7,348 7,348 603 822 842 4.28 0.05
the class level distributions of OTUs unique to floodwaters
sampled from either the vegetative or reproductive stages
Vegetative Reproductive
were drastically different. For example, the community of
OTUs unique to the floodwaters of vegetative stage plants
1463 88 515 wasdominatedbyGammaproteobacteria(34.5%),Alphapro-
teobacteria (18.4%), and Betaproteobacteria (10.8%) (Fig-
ure 4(c)). Of significance, 24.6% of the amplicon reads
derivedfromOTUsexclusivelyassociatedwiththevegetative
stage could not be conclusively assigned to the class level
(Figure 4(c)). In contrast, the community of OTUs unique
Figure 3: Venn diagram of 16S OTUs shared and unique to
floodwater samples collected from vegetative and reproductive tofloodwatersofreproductivestageplantswasdominatedby
stages of rice plants. Sequences were classified into OTUs at 97% Gammaproteobacteria(43.9%oftheampliconreads),Bacilli
similarityusingtheaverageneighboralgorithminmothur. (25.4%), Clostridia (21.8%), and Betaproteobacteria (17.7%)
(Figure4(d)).
Of the 169 genera identified in the two developmental
Tenericutes, and Verrucomicrobia, which were detected in ricestages,22(13%)occurredinbothstages(Figure5).Bei-
both stages of crop development. The phyla Acidobacteria jerinckia, Curvibacter, Pelomonas, and Rhodoferax occurred
(1.94% of the OTUs) and Cyanobacteria (1.25%) were also in high counts in both phenological stages. In general, the
present,butatlowfrequencies(≤2%oftheOTUs). abundancesofBeijerinckia,Pelomonas,andRhodoferaxwere
In total, 27 different bacterial classes were identified in similar in flood waters from both communities; however,
both developmental stages. In general, the class level abun- therelativeabundanceofCurvibacter wasslightlyhigherin
dancesofthe88OTUssharedbetweenthefloodwatersfrom vegetative stage. The abundances of the genera Aeromonas,
thevegetativeandreproductivestageswereverysimilarand Bradyrhizobium,Emticicia,Massilia,Methylibium,andSpiro-
werebothdominatedbyBetaproteobacteria,Actinobacteria, soma were lower and did not differ significantly between
andAlphaproteobacteria(Figures4(a)and4(b)).Incontrast, vegetativeandreproductivestages(Figure5).
6 InternationalJournalofMicrobiology
Vegetative
Reproductive
Gammaproteobacteria
Flavobacteria 0.1% Holophagae Gammaproteobacteria Holophagae
Deltaproteobacteria 1.2% 0.09% Sphingobacteria Flavobacteria 0.1% 0.2%
0.7% 1.1% 3.5% Sphingobacteria
Deltaproteobacteria 3.7%
Cyanobacteria 0.2% Actinobacteria
0.02% Cyanobacteria 17.8%
0.07%
Actinobacteria Alpha-
18.8%
proteobacteria
7.0%
Alpha-
proteobacteria
12.1%
Betaproteobacteria
65.5%
Betaproteobacteria
63.7%
(a) (b)
Reproductive
Alphaproteobacteria Sphingobacteria
Vegetative 2.8% 2.7% Actinobacteria
Deltapr3o.t3e%obacteria Bac3te.2r%oidiaO8.t9h%er Unc3la.5s%sified 1.7% O2.t8h%er Gammap4r3o.9te%obacteria
Sphingobacteria Flavobacteria
4.6% Unclassified 3.8%
24.6% Erysipelotrichia
Flavobacteria 5.0%
8.0%
Actinobacteria
8.3%
Betaproteobacteria
Betaproteobacteria 17.7%
10.8%
Alphaproteobacteria
18.4% Clostridia
Gammaproteobacteria 21.8% Bacilli
34.5% 25.4%
(c) (d)
Figure 4: Relative abundances of class level assignments for the 88 shared OTUs in the vegetative stage (a) and reproductive stage (b)
andrelativeabundancesoftheclasslevelassignmentsofthe1463OTUsuniquetothevegetativestage(c)andthe515OTUsuniquetothe
reproductivestage(d).OTUsweretaxonomicallyclassifiedbytheWangmethodinmothurusingan80%confidencethresholdforclasslevel
assignments.
Figure5showsacomparativeanalysisofthetwenty-two isprobablydueinparttothelowerrichnessassociatedwith
mostabundantbacterialgeneradetectedinthereproductive the reproductive stage, the higher number of rare OTUs
andvegetativestagesoffloodedrice. foundinthevegetativestage,andtheoverabundanceofthree
Polynucleobacter (29.5%),Curvibacter (23%),andRhod- genera, Alcaligenes, Curvibacter, and Flavobacterium in the
oferax (10.3%) represented the most abundant genera in reproductivestage.TheChao1richnessestimatorpredicted
the vegetative stage, while Curvibacter (18.3%), Alcaligenes atotalof2,282OTUsinthevegetativebacterialcommunity
(11.8%),andFlavobacterium(11.3%)weremoreabundantin (32%moreOTUsthanwhatweredetected),while822OTUs
thereproductivestage(Figure6). werepredictedinthereproductivebacterialcommunity(27%
󸀠
TheShannon-Wiener(H)indexwasapproximately25% more OTUs than what were detected). Considering these
higher in the water sample collected from the vegetative results, it is clear that a larger number of rare species were
stage compared to the reproductive stage, suggesting that associated with water collected from the vegetative stage
the vegetative community is richer. In addition, the Simp- comparedtothereproductivestage(Table2).
son index, which represents the probability that two reads Reflecting the higher richness and higher numbers of
selected at random belong to the same OTU, was higher in rare OTUs, the rarefaction curve of the vegetative stage
thesamplecollectedfromthereproductivestage.Thisfinding did not reach complete saturation considering the number
InternationalJournalofMicrobiology 7
−2 −1 0 1 2
Column Z-score Vegetative Reproductive
Flavobacterium
Arcicella
Aquitalea
Alcaligenes
Polynucleobacter
Leptothrix
Geobacter
Novosphingobium
Sediminibacterium
Brevundimonas
Spirosoma
Pelomonas
Curvibacter
Bradyrhizobium
Roseomonas
Aeromonas
Rhodoferax
Emticicia
Methylibium
Massilia
Geothrix
Beijerinckia
Figure5:Heatmapandcorrelationanalysisofthe22bacterialgenerafoundinboththevegetativeandreproductivestagesoffloodedrice.
Therelativeabundancesofthesegenerainfloodwatersassociatedwitheachdevelopmentalstagewerelog2transformed,scaledby𝑧-score,
andusedforclusteringandheatmapanalysis.Colorintensityiscorrelatedwiththerelativeabundanceofeachgenusinfloodwatersassociated
withthetwodevelopmentalstages,withblueindicatinghigherrelativeabundanceandyellowindicatinglowerrelativeabundance.
Vegetative Reproductive
Sedimin0i.8b%acteriumGe0o.3th%rAixr0c.i8c%ella Spi0ro.8s%oma Em0t.8ic%iciaBrBaedijy5e2r.r4h.i4%ni%zcokbiaium GeothrixE m0.7ti%cicAiar c1i.c0e%lla 11.1% Sedimin0i.2b%acteriuSmpirosomBar 0a.d5y%1rh.9i%zobBieuimjerinckia 7.5%
Flavobacterium Brevundimonas 3.0% Aeromonas 0.5% Brevundimonas 1.7%
3.3% Aeromonas Novosphingobium 1.1% Novosp0h.2in%gobium
0.5% Roseomonas 0.3% Roseomonas 1.2%
Ge2o.b2a%cter Alc0a.l5ig%enes Flavobacterium 11.3% Alcaligenes 11.8%
Rhodoferax Aquitalea Geobacter 1.0%
10.6% 0.3%
Curvibacter Rhodoferax 10.3% Aquitalea 7.0%
23.0%
Polynucleobacter Leptothrix Polynucleobacter Curvibacter 18.3%
29.5% Massilia 4.6% 4.6%
0.3% Pelomonas 7.2%
Pelomonas Methylibium Massilia 0.5% Leptothrix 1.2%
9.2% 0.3% Methylibium 0.5%
(a) (b)
Figure6:RelativeabundancesofgenuslevelassignmentsfortheOTUssharedbetweenvegetativestage(a)andreproductivestage(b).
8 InternationalJournalofMicrobiology
1600 stagesoffloodedrice,withthediversityandrichnessvalues
1400 associatedwiththevegetativestageexceedingthoseobserved
Us1200 in the reproductive stage. Previous research on rice crops
OT1000 reportedthattherichnessanddiversityofbacterialcommu-
of 800 nities can change throughout the rice cultivation period in
ber 600 aquaticenvironments[31–33]andinthesoilandrhizosphere
m
u 400 environments [34–36]. Although previous studies did not
N
200 elucidate the environmental factors responsible for these
0 changesincommunitystructure,floodwatercharacteristics,
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 such as pH, temperature, changes in nutrient profiles, and
0 0 0 0 0 0 0 0 0 0 0 0 0 0
5 10 15 20 25 30 35 40 45 50 55 60 65 70 seasonalvariationsinfertilizerapplication,havebeenrelated
Number of amplicons toshiftsinmicrobialcommunitystructureandcomposition
Vegetative over crop development [32]. Additional factors related to
Reproductive plantgrowthcanalsocauseshiftsinmicrobialcommunities.
Forexample,Okabeandcoworkers[31]determinedthatplant
Figure7:Rarefactioncurvesgeneratedfrom16Sampliconssampled
growthvarieswiththeincidenceofsunlight,which,inturn,
from floodwaters from vegetative and reproductive rice stages.
canimpactthewatertemperatureandalgalgrowth.Higher
Rarefaction curves were generated using the rarefaction.single
commandinmothurwithfreq=100anditers=10000. photosynthesis rates associated with algal proliferation and
alsoplantgrowth[34]canultimatelyleadtoincreasedlevels
of photosynthates in the water, altering the water pH and
of sequences sampled, indicating that additional OTUs will potentiallyalteringthebacterialcommunity.
likelybedetectedwithadditionalsequencing.Ontheother Atotalof2,066OTUsat3%dissimilaritywereidentified
hand,rarefactioncurveofthereproductivestageshowedsta- infloodwaterscollectedfromthereproductiveandvegetative
bilization,indicatingthatthesequencingeffortwassufficient developmental stages of rice. In flooded rice systems, the
tocapturethemajorityoftheOTUsassociatedwiththiscom- short spacing between the plants, the low depth of the
munity(Figure7). floodwaters (10cm deep), and soil management regimes
can increase the levels of suspended soil particles in the
4.Discussion floodwaters [37], which is in contrast to other more static
aquatic ecosystems. Due to the high turbidity and the
Infloodedrice,thewatersourceusedforcultivationisakey presence of soil in the floodwaters, it is expected that
component in the management of the crop. Nutrient levels bacterialgroupsassociatedwithsoil,suchasProteobacteria,
in these water sources can have significant effects on plant Bacteroidetes,Actinobacteria,andFirmicutes,willbepresent
nutrition[28]whilevaryingphysicalandchemicalproperties in the floodwaters of rice fields, contributing to the total
inthesewatersourcescanalsoinfluencethestructure,diver- bacterialdiversity[38].Supportingthishypothesis,themajor
sity, and taxonomic composition of microbial communities phyla identified in the floodwaters associated with both
inthefloodwaters,whichinturncanalsoprofoundlyimpact developmental stages were Proteobacteria, Bacteroidetes,
plant nutrition. In the rice field plots used for this study, Actinobacteria, and Firmicutes, while Acidobacteria and
the primary water source contains a high proportion of Cyanobacteria were less frequent. The same phyla were
wastewaterinputfrombothurbanandagriculturalareas[29]. highlyabundantinotherricehabitats[7],andspecificallyin
Therefore,avarietyofchemicalelementsenterthefloodplain water samples collected from flooded rice crops [3, 32, 39].
and the temporal fluctuations in the abundance and types Furthermore, Proteobacteria (alpha- and beta-subdivision),
of chemical elements entering the fields can have strong Bacteroidetes,andActinobacteriawerealsocitedasthemajor
influences on the microbial community in rice floodwaters groups of bacteria found in freshwater environments [39].
[30]. Inunstableenvironments,suchasricefields,Proteobacteria
In the current study, the water samples retrieved from may contribute to the growth, development, and physiol-
vegetativeandreproductivestagesshowedvariationsinphys- ogy of rice plants due to the presence of plant growth-
ical and chemical parameters. The abiotic parameters most promoting bacteria, which are often able to induce callus
stronglyassociatedwiththevegetativestagewerepH,P,andK in rice and produce phytohormones [40]. According to
andinthereproductivestage,theparametersweretempera- Mhuantongetal.[16],thepresenceofthesebacterialgroups
tureandturbidity(Table1).Oftheseparameters,temperature in a variety of freshwater ecosystems is related to their
(higher in floodwaters from the reproductive stage) and K metabolicplasticitytodecomposeanassortmentoforganic
(higher in floodwaters from the vegetative stage) differed matter.
significantlybetweenthetworicedevelopmentstages.These Despitethepresenceofseveralcommonphylainflood-
results reflect differences in microenvironments between waters sampled from both developmental stages, only 88
the rice plant-phenological stages that may have affected (4.08%)oftheOTUswerecommonbetweenvegetativeand
microbial growth and persistence. Supporting the possible reproductive stages, suggesting that the dynamics of the
impactoftheseenvironmentalconditionsonbacterialcom- bacterialcommunitiesinfloodwaterschangethroughoutthe
munities,significantdifferencesincommunitycomposition growingseason.Thesechangescouldreflecttemporalvaria-
were observed between the vegetative and reproductive tionsincarbonavailability.Forexample,inapreviousstudy,
InternationalJournalofMicrobiology 9
Aulakh and coworkers [41] documented higher exudation Beijerinckia, Curvibacter, Pelomonas, and Rhodoferax
rates over the course of rice plant development, in which were the four most abundant genera in both stages of the
sugars in the exudates were eventually replaced by organic crop. These bacterial groups are often associated with high
acidsinlaterdevelopmentalstages.Bothsugarsandorganic availabilitiesofsugar(e.g.,glucose)[52].Inagriculturaland
acidsareessentialforbacterialnutritionandcontributetothe freshwaterenvironments,Beijerinckiaarereportedasgrowth
process of microbial colonization [42]. Furthermore, when promoters in plants [53], while the genomes of Curvibacter
soluble carbon is released into the rhizosphere, microbes spp.oftencodeforlargenumbersofsugartransporters[54],
in the soil quickly mineralize the available carbon [43]. enhancing its ability to explore nutrients from the environ-
Thereby,theamountofrootexudatesinthesoilcouldaffect ment. Other genera, such as Pelomonas sp. and Rhodoferax
the relative amounts of carbon and nitrogen available for sp.,werefoundinapproximatelythesameproportioninboth
microbial growth [44], promoting transitions in the size of rice stages. Previous rDNA 16S amplicon analysis showed
populations and structure of microbial communities in the that Rhodoferax spp. are common in diverse freshwater
rhizosphere[45].Nevertheless,thecompositionandquantity systems [55]. Some species of the genera Rhodoferax are
ofexudatesaredynamicintimeandspace,andthereforeit nitrate-reducers [56], while Pelomonas spp. often code for
becomesdifficulttoresolvetheroleofasinglecomponentin machinery required for nitrogen fixation (nifH gene) [57].
thestructureofmicrobialcommunity[46]. Thus, the OTUs assigned to these genera in the current
The richest bacterial classes found in the vegetative studymaycontributetonitrogenfluxinfloodedricefields.
stage were Gammaproteobacteria (276OTUs), Alphapro- In contrast, the abundance of the genus Polynucleobacter
teobacteria(237OTUs),andBetaproteobacteria(211OTUs), differedsignificantlyinfloodwatersfromthetwolifestages.
while richness was the highest in Gammaproteobacteria Itsabundancewassignificantlyhigherinthevegetativestage
(24OTUs), followed by Bacilli (46OTUs) and Clostridia (29.5%) compared to reproductive stage (4.6%). This result
(18OTUs) during the reproductive stage. In flooded plains, could suggest that the OTUs derived from this genus are
theadherenceofBetaproteobacteriarepresentativestoparti- notwelladaptedtoenvironmentalconditionsassociatedwith
clescanplayanimportantroleintheirtransportfromtheter- floodwaterscollectedfromreproductivestageplants.Polynu-
restrialenvironmenttofloodwater[47],whilethericephyl- cleobacter members often dominate planktonic freshwater
losphere is greatly colonized by Alphaproteobacteria class communities [58], where pH, conductivity, and dissolved
[48].Thus,Beta-andAlphaproteobacteriarepresentrelevant organiccarbonconcentrationplayanimportantroleincolo-
bacteria groups in the soil whose abundances are strongly nization[59].Also,flagellategrazingandtheconsumptionof
linkedtocarbonsupply[49].Infloodedricesystems,these lowmolecularweightsubstratesarerelatedtothesurvivalof
bacterial groups decompose organic matter derived from Polynucleobacterinfreshwaterenvironments[60,61].
rice straw and other fermentable crop residues along with Rises in temperature can increase the abundance of
fertilizerwhereorganicmolecules(humicsubstances)canact prokaryotes of the natural communities by enhancing their
asanelectrondonorduringmicrobialrespiration[47]. growthrates[62].TheabundanceofthephylumFirmicutes
Gammaproteobacteriawerepresentinhighrichnessand (5.52%) found in the reproductive stage is consistent with
abundanceinfloodwatersassociatedwithbothricedevelop- the abundance of this phylum commonly found in the rice
mentalstages.Overall,membersfromthephylumProteobac- fields [63]. However, the presence of these spore-forming
teriaareoftenthemostabundantfreshwaterprokaryotes,but populations in the reproductive stage may be related to the
Gammaproteobacteriaareusuallypresentinlowortransient characteristicsoftheorganicmatteravailableassociatedwith
abundances [38], although they were previously found in this developmental stage [63] which may vary from straw
high amounts in the rice phyllosphere [40]. This finding residues and root residues [64]. Species belonging to the
suggests that specific environmental factors associated with class Clostridia are able to express enzymes that degrade
the rice agroecosystem, such as the source of water used the cellulose and sugar transport systems to quickly uptake
to irrigate the fields and the turbidity of the floodwaters, sugars released from plant cell walls [65]. In rice field soil,
provided conditions for the survival and persistence of these organisms were responsible for the decomposition of
∘
members of this phylum with the entrance of nutrients. strawresidueat15and30 C[64].Fromthisbasis,conditions
ThisclasshoststhegeneraEscherichia,Salmonella,Yersinia, found in reproductive stage such as the high temperature
Vibrio, and Pseudomonas characterized by their metabolic andcarbonavailabletobedegradedmayhavepromotedthe
plasticity that can be adapted to varying levels of temper- proliferation of OTUs belonging to Gammaproteobacteria
ature, oxygen supply, and nutritional requirements. Hence, andFirmicutes.
membersoftheclassGammaproteobacteriaoftendominate Rhizodepositionoccurscontinuouslyduringthelifetime
microbial assemblages after nutrient enrichment [50]. In ofplantsintheformofwatersolubleexudates,secretions,and
addition, the members of this group contain pathways for plantdeadcells.Thisprocesscontributesasaprimarysource
degradationofcarbohydrates,aminoacids,andxenobiotics ofenergyandnutrientsformicrobesinthefloodwaters[66].
thatmayprovidecompetitiveadvantagesundercertaineco- In future studies, the release of carbon compounds from
logicalcondition[51].Furthermore,theinteractionbetween rice roots should be considered in the investigation of the
Gammaproteobacteria members and rice plants may be factors responsible for shaping the structure of the bacte-
beneficial to both parties, with the leaf exudates providing rial communities in floodwaters of rice systems [34]. With
nutrientsthatsupportmicrobialgrowthandmicroorganisms respect to environmental and nutritional conditions during
assistingbetterriceyield[40]. the vegetative stage, approximately 20% of the absorbed
10 InternationalJournalofMicrobiology
photosynthetic carbon is released through rhizodeposition regarding the composition and quantification of root exu-
[67]. Moreover, during the reproductive stage, the plants dates and the impacts of rice exudates on microbial com-
releaseloweramountsofcarboncompoundsintheirexudates munities in rice fields are essential to decipher the linkage
becausetheydevotethemajorityoftheirenergytoseedpro- between the microbial community dynamics and rice plant
duction [68, 69]. Breidenbach and Conrad [7] investigated overcultivationtime.
the bacterial community of the soil from rice fields using
16SrDNAampliconanalysisandalsoprofiledtheexpression Disclosure
levelofvarioustaxathroughoutthevegetative,reproductive,
and maturation stages. According to quantitative PCR, the The U.S. Department of Agriculture, Agricultural Research
authorsdetected elevatedbacterialabundancesinsoilfrom Service,isanequalopportunity/affirmativeactionemployer
the reproductive stage, but there were no major differences andallagencyservicesareavailablewithoutdiscrimination.
between the bacterial community compositions during the Mention of commercial products and organizations in this
twogrowthstages.Changesinbacterialabundancebetween manuscriptissolelytoprovidespecificinformation.Itdoes
soil collected from the vegetative and reproductive stages notconstituteendorsementbyUSDA-ARSoverotherprod-
werestronglyattributedtodecreasedlevelsofrootexudates uctsandorganizationsnotmentioned.
by plants nearing maturity. In general, the abundance of
ConflictsofInterest
bacteriainthesoilwasdynamicthroughoutthedevelopment
of rice plants; however, no microbial taxa appeared to be
The authors declare that there are no conflicts of interest
exclusivelyassociatedwithaparticularstageofdevelopment.
regardingthepublicationofthispaper.
Heterotrophic microorganisms, which represent a large
portionofthericefieldmicrobiome[70,71],aredependent
Acknowledgments
oncarbonsubstratesfortheirenergysupply.Thevegetative
stage (tillering) is the most vigorous growth period of the
The authors would like to thank CAPES (Coordenac¸a˜o
rice plant [72]. As plant biomass and growth increase,
de Aperfeic¸oamento de Pessoal de N´ıvel Superior) for the
intensifications of organic carbon secretions can also occur
Sandwich Ph.D. Program [Grant no. 260213-0] and USDA
[73]. Thus, considering that the higher carbon levels could
(United States Department of Agriculture) for the financial
support the persistence of a larger diversity of microbes,
support.
the higher levels of exudates could contributeto the higher
richnessobservedinthisstudyinfloodwaterscollectedfrom
References
vegetativestage.Moreover,theuseofureainfertilizers,which
is applied mainly during the vegetative stage, can affect the [1] IRRI-InternationalRiceResearchInstitute,“RiceinAmerica
+
concentrationsofavailablenutrients,especiallyNH4 ,which Latine,past,present,andpromisingfuture,”RiceToday,vol.14,
isthemainformofavailableNinthesoil[74].Thesefactors, pp.1–23,2015.
aloneorincombination,mayhavehelpedtoshapethebac- [2] GRiSP-GlobalRiceSciencePartnership,“Ricealmanac,”Los
terialcommunities,increasingitsdiversityinthevegetative Ban˜os,Philippines:InternationalRiceResearchInstitute,2012.
stageofcrop.Inthatsense,thereleaseofrootexudatescan [3] T.Shibagaki-Shimizu,N.Nakayama,Y.Nakajima,K.Matsuya,
influencethedynamicsofbacterialcommunitiesduringcrop M. Kimura, and S. Asakawa, “Phylogenetic study on a bac-
developmentandshouldbeinvestigatedasacontributorto terial community in the floodwater of a Japanese paddy field
bacterialcommunitydynamicsinricefloodwaters. estimatedbysequencing16SrDNAfragmentsafterdenaturing
gradientgelelectrophoresis,”BiologyandFertilityofSoils,vol.
42,no.4,pp.362–365,2006.
5.Conclusion
[4] A. E. Bernhard, D. Colbert, J. McManus, and K. G. Field,
Thisstudyrevealedthatthecommunitiesassociatedwithrice “Microbial community dynamics based on 16S rRNA gene
profilesinaPacificNorthwestestuaryanditstributaries,”FEMS
floodwaters were variable and differed between floodwaters
MicrobiologyEcology,vol.52,no.1,pp.115–128,2005.
from the vegetative and reproductive stages of rice. The
[5] A.R.Lopes,C.M.Manaia,andO.C.Nunes,“Bacterialcom-
higherbacterialrichnessobservedinfloodwatersassociated
munity variations in an alfalfa-rice rotation system revealed
withvegetativeplantsmayhavebeen favoredby thehigher
by 16S rRNA gene 454-pyrosequencing,” FEMS Microbiology
concentration of nutrients (for example, phosphorus and
Ecology,vol.87,no.3,pp.650–663,2014.
potassium) due to changes in rhizodeposition and crop
[6] J.Zhao,T.Ni,W.Xunetal.,“Influenceofstrawincorporation
management. OTUs from the phylum Proteobacteria were
withandwithoutstrawdecomposeronsoilbacterialcommu-
present in floodwaters collected from both developmental
nitystructureandfunctioninarice-wheatcroppingsystem,”
stages, indicating the persistence of this phylum in flooded
Applied Microbiology and Biotechnology, vol. 101, no. 11, pp.
rice ecosystems. Moreover, the predominance of the class 4761–4773,2017.
Gammaproteobacteria and the occurrence of the phylum
[7] B.BreidenbachandR.Conrad,“Seasonaldynamicsofbacterial
Firmicutes in the reproductive stage demonstrated that the andarchaealmethanogeniccommunitiesinfloodedricefields
environment was favorable to microbes that can persist andeffectofdrainage,”FrontiersinMicrobiology,vol.5,article
longer in paddy fields where they are able to utilize the 752,2014.
degradablefractionoforganicmaterialsandsurviveinwater- [8] B.Breidenbach,M.B.Blaser,M.Klose,andR.Conrad,“Crop
logged and dry conditions. However, further investigations rotationoffloodedricewithuplandmaizeimpactstheresident