Barcoded Pyrosequencing Reveals a Shift in the Bacterial Community in the Rhizosphere and

17 Pages·2017·1.98 MB·English

International Journal o f
Molecular Sciences
Article
Barcoded Pyrosequencing Reveals a Shift in the
Bacterial Community in the Rhizosphere and
Rhizoplane of Rehmannia glutinosa under
Consecutive Monoculture
LinkunWu1,2,JunChen1,3,ZhigangXiao1,2,XiaochengZhu4,5 ID,JuanyingWang1,2,
HongmiaoWu1,2,YanhongWu1,2,ZhongyiZhang3,5andWenxiongLin2,3,*
1 CollegeofLifeSciences,FujianAgricultureandForestryUniversity,Fuzhou350002,China;
[email protected](L.W.);[email protected](J.C.);[email protected](Z.X.);
[email protected](J.W.);[email protected](H.W.);[email protected](Y.W.)
2 FujianProvincialKeyLaboratoryofAgroecologicalProcessingandSafetyMonitoring,FujianAgriculture
andForestryUniversity,Fuzhou350002,China
3 KeyLaboratoryofCropEcologyandMolecularPhysiology(FujianAgricultureandForestryUniversity),
FujianProvinceUniversity,Fuzhou350002,China;[email protected]
4 GrahamCentreforAgriculturalInnovation(CharlesSturtUniversityandNSWDepartmentofPrimary
Industries),CharlesSturtUniversity,WaggaNSW2678,Australia;[email protected]
5 CollegeofCropScience,FujianAgricultureandForestryUniversity,Fuzhou350002,China
* Correspondence:[email protected];Tel.:[phone]
Received:10February2018;Accepted:13March2018;Published:14March2018
Abstract: The production and quality of Rehmannia glutinosa can be dramatically reduced by
replant disease under consecutive monoculture. The root-associated microbiome, also known
as the second genome of the plant, was investigated to understand its impact on plant health.
Culture-dependentandculture-independentpyrosequencinganalysiswasappliedtoassesstheshifts
in soil bacterial communities in the rhizosphere and rhizoplane under consecutive monoculture.
The results show that the root-associated microbiome (including rhizosphere and rhizoplane
microbiomes) was significantly impacted by rhizocompartments and consecutive monoculture.
Consecutive monoculture of R. glutinosa led to a significant decline in the relative abundance
of the phyla Firmicutes and Actinobacteria in the rhizosphere and rhizoplane. Furthermore, the
familiesFlavobacteriaceae,Sphingomonadaceae,andXanthomonadaceaeenrichedwhilePseudomonadaceae,
Bacillaceae, and Micrococcaceae decreased under consecutive monoculture. At the genus level,
Pseudomonas,Bacillus,andArthrobacterwereprevalentinthenewlyplantedsoil,whichdecreased
inconsecutivemonoculturedsoils. Besides,culture-dependentanalysisconfirmedthewidespread
presenceofPseudomonasspp. andBacillusspp. innewlyplantedsoilandtheirstrongantagonistic
activities against fungal pathogens. In conclusion, R. glutinosa monoculture resulted in distinct
root-associatedmicrobiomevariationwithareductionintheabundanceofbeneficialmicrobes,which
mightcontributetothedeclinedsoilsuppressivenesstofungalpathogensinthemonocultureregime.
Keywords: Rehmannia glutinosa; consecutive monoculture; root-associated microbiome; deep
pyrosequencing;beneficialrhizobacteria
1. Introduction
Rehmanniaglutinosa,aperennialmedicinalplant,belongstothefamilyScrophulariaceaeandhas
verywidespreadmedicinalpurposes. ThetuberousrootofR.glutinosaisoneofthe50fundamental
herbsusedintraditionalChinesemedicine. Itprovidesmultiplepharmacologicalactionsontheblood,
Int.J.Mol.Sci. 2018,19,850;doi:10.3390/ijms19030850 www.mdpi.com/journal/ijms
Int.J.Mol.Sci. 2018,19,850 2of17
immune, endocrine, cardiovascular, and nervous systems [1]. This species is produced mainly in
JiaozuoCity,HenanProvince,centralChina,knownasthegeo-authenticproductionzonewiththe
mostsuitablesoilandclimateconditions. Thequalityofplantsgrownoutsidethegeo-authenticareas
cannot be assured. Meanwhile, consecutive monoculture of this plant in the same field can result
intotalcropfailureofundergroundtubers. ComparedwithnewlyplantedR.glutinosa,atwo-year
consecutivemonocultureledtounexpandedtuberousrootswithlargenumbersofadventitiousfibrous
roots(FigureS1)[2]. Inaddition,consecutivemonocultureresultedinasignificantincreaseinthe
abundanceofsoil-bornefungalpathogens, suchasFusariumoxysporumandAspergillusflavus[2,3].
Therefore,commercialR.glutinosaaretypicallyreplantedevery15–20yearsinthesamefield[4]. Itis
critical tounderstand the mechanismsthat drivethe replant disease of R. glutinosa to increase the
efficiencyandqualityofproduction.
Numerousstudieshaveindicatedthatreplantdiseaseisattributedtoadeficiencyinsoilnutrients
andthedirectallelopathicautotoxicityofrootexudatesorplantresidue[5,6]. However,ourprevious
studyhasindicatedthatR.glutinosaconsecutivemonoculturedidnotleadtodecreasedsoilorganic
matter and available nutrients [7]. Many studies have indicated that plants could produce toxic
compoundsthatinhibittheirowngrowthandreproduction[8,9]. Forexample,allelochemicalsand/or
autotoxiccompoundssuchasβ-cembrenediol,Di-n-hexylPhthalate,andbis(2-propylheptyl)phthalate,
whichsignificantlyreduceseedlinggrowth,werefoundinflue-curedtobacco[9]. Incontrast,indirect
allelopathy is driven by variations in the soil microbial communities that are associated with root
exudates[6,10,11]. Lietal.[10]reportedindirectallelopathyofpeanut(Arachishypogaea)throughthe
exudation of phenolic acids that induced the growth of pathogens and reduced the beneficial soil
microorganisms. ZhouandWu[12]indicatedthatp-coumaricacidinfluencedconsiderablythesoil
microbialcommunitiesandpromotedthegrowthofpathogenicFusariumoxysporumf.sp. cucumerinum
Owen, which then resulted in replant disease of cucumber. Similar examples were also found in
Maluspumila,Radixpseudostellariae,andR.glutinosa[2,13,14].
Previous studies using terminal restriction fragment length polymorphism (T-RFLP) and
phospholipidfattyacid(PLFA)analysesdemonstratedthatconsecutivemonocultureofR.glutinosa
ledtosignificantchangesinthesoilmicrobialbiomassandcommunitystructure[2,3]. However,the
resolutionandveracityofsoilmicrobialidentityarelimitedusingT-RFLPandPLFAtechnologies[15].
Thus,amorereliableapproachwithhigherresolutionisrequiredtoimprovetheunderstandingof
thedynamicsofsoilmicrobialcommunities. High-throughputpyrosequencingof16SrRNAgene
fragmentshasproventobeapowerfultoolforin-depthsequencingandparallelanalysisofsoiland
rootmicrobialcommunitiesatareducedsequencingerrorrate[16,17]. Illuminapyrosequencingwas
appliedforthestudyofsoilmicrobialcommunitiesunderR.glutinosamonoculture[18]. However,the
rhizoplaneofplantrootsystemsplaysaselectivegatingroleinmicrobiomerecruitmentfromsoil[19]
dependingontherootphenotypictraits[20]. Thisaspectwasnotstudiedinpreviousstudies[18].
In addition, antagonistic activities of beneficial microbial biomarkers in disease-suppressive soils
are still poorly understood [18]. Even though severe replant disease is observed in the field and
confirmedbythedetectionofspecificfungalpathogens(F.oxysporumandA.flavus)byquantitative
PCR[3],littleisknownabouttheshiftsinroot-associatedbacterialcommunitiesandtheirrelationships
with the increase of some specific soil-borne pathogens (i.e., F. oxysporum and A. flavus) under
consecutivemonoculture. Understandingthecomplexroot-associatedmicrobialcommunitiesfrom
bothrhizosphereandrhizoplaneisimportantandusefulforincreasingsoilhealthandimproving
plantperformance.
Inthisstudy,barcodedpyrosequencingof16SrRNAgenescombinedwithaculture-dependent
approachwasappliedtocharacterizethesoilbacterialcommunitiesunderR.glutinosaconsecutive
monoculture. We hypothesized that consecutive monoculture practice could alter the structure of
bacterialcommunitiesinboththerhizosphereandrhizoplanewithareductionintheabundanceof
beneficial root-associated rhizobacteria showing antagonistic activities against R. glutinosa replant
diseasecausalagents.
Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 3 of 17
Ibnet.nJ.eMficoli.aSlc ir.o2o0t1-8a,s1s9o,8c5ia0ted rhizobacteria showing antagonistic activities against R. glutinosa re3polafn17t
disease causal agents.
2. Results
2. Results
2.1. OperationalTaxonomicUnit(OTU)ClusterandSpeciesAnnotation
2.1. Operational Taxonomic Unit (OTU) Cluster and Species Annotation
Deeppyrosequencingof16SrDNAampliconswasappliedtoassessthechangesintherhizosphere
Deep pyrosequencing of 16S rDNA amplicons was applied to assess the changes in the
andrhizoplanebacterialcommunitiesunderconsecutivemonocultureofR.glutinosa. Atotalof995,947
rhizosphere and rhizoplane bacterial communities under consecutive monoculture of R. glutinosa. A
effectivetagswithspeciesannotationwereobtainedfrom18soilsamples(asshowninTableS1)with
total of 995,947 effective tags with species annotation were obtained from 18 soil samples (as shown
anaverageof55,330effectivetagspersample. Singletonsaccountedfor3.8%(37,683tags)ofthetotal
in Table S1) with an average of 55,330 effective tags per sample. Singletons accounted for 3.8% (37,683
tagsandwereremovedfromthedatasetbeforefurtheranalyses(FigureS2). Alllibrariesrepresented
tags) of the total tags and were removed from the dataset before further analyses (Figure S2). All
thebacterialcommunitieswellastherarefactioncurveswereapproachingplateaus(Figure1). Based
libraries represented the bacterial communities well as the rarefaction curves were approaching
on97%speciessimilarity,584,740,633,649,767,and481operationaltaxonomicunits(OTUs)were
plateaus (Figure 1). Based on 97% species similarity, 584, 740, 633, 649, 767, and 481 operational
obtainedintherhizospheresoilsamplesfromthenewlyplanted(NPR),monocultureduninfected
taxonomic units (OTUs) were obtained in the rhizosphere soil samples from the newly planted (NPR),
(MUR), and monocultured infected (MIR) plants, and the rhizoplane soil samples from the newly
monocultured uninfected (MUR), and monocultured infected (MIR) plants, and the rhizoplane soil
planted(NPP),monocultureduninfected(MUP),andmonoculturedinfected(MIP)plants,respectively
samples from the newly planted (NPP), monocultured uninfected (MUP), and monocultured infected
(FigureS2). Onaverage,99.98%,99.91%,87.95%,86.89%,78.91%,and18.75%ofeffectivesequences
(MIP) plants, respectively (Figure S2). On average, 99.98%, 99.91%, 87.95%, 86.89%, 78.91%, and
fromall18samplescouldbeassignedtothephylum,class,order,family,genus,andspecieslevels,
18.75% of effective sequences from all 18 samples could be assigned to the phylum, class, order,
respectively(TableS2).
family, genus, and species levels, respectively (Table S2).

FFiigguurree 11.. RRaarreeffaaccttiioonn ccuurrvveess ooff bbaacctteerriiaall ccoommmmuunniittiieess bbaasseedd oonn oobbsseerrvveedd ooppeerraattiioonnaall ttaaxxoonnoommiicc uunniittss
((OOTTUUss)) aatt tthhee 9977%% ssiimmiillaarriittyy ccuutt--ooffff lleevveell ffoorr iinnddiivviidduuaall ssaammpplleess.. NNPPRR,, MMUURR,, aanndd MMIIRR rreepprreesseenntt tthhee
rrhhiizzoosspphheerrees osioliflr ofmrotmhe tnheew lnyepwlalyn tepdlaanntdedm oannodc umltuorneodcuunltiunrfeedct eudnainndfeicntfeedc teadndpl ainntfse,crteesdp epctliavnetlsy,.
NrePspP,eMctUivPe,lyan. dNMPPIP, rMepUrePs,e natntdh eMrhIiPz orpelpanreesseonitl frthome rthheiznoepwlalnyep lsaonitle dfraonmd mthoen onceuwltluyr epdlaunntiendfe catnedd
amnodnioncfueclttuedrepdl aunntisn,freecstpedec atinvde liyn.fected plants, respectively.
2.2. Alpha Diversity Analysis
2.2. AlphaDiversityAnalysis
The bacterial diversity and richness indices of different soil samples were calculated based on
Thebacterialdiversityandrichnessindicesofdifferentsoilsampleswerecalculatedbasedon
44,256 sequences. In the rhizosphere, MUR showed the highest level of diversity (5.675) and highest
44,256sequences. Intherhizosphere,MURshowedthehighestlevelofdiversity(5.675)andhighest
species richness with observed species, Chao1, and ACE (abundance-based coverage estimator)
speciesrichnesswithobservedspecies,Chao1,andACE(abundance-basedcoverageestimator)indices
indices at 654.7, 718.5, and 742.5 respectively. NPR had the lowest species richness and diversity
at654.7,718.5,and742.5respectively. NPRhadthelowestspeciesrichnessanddiversitycompared
compared with the other two rhizosphere soil groups (Table 1). Most of the richness and diversity
withtheothertworhizospheresoilgroups(Table1). Mostoftherichnessanddiversityindicesshowed
indices showed significant differences among the three soil groups but no significant difference was
significantdifferencesamongthethreesoilgroupsbutnosignificantdifferencewasobservedinthe
observed in the ACE index between NPR and MIR (p > 0.05) (Table 1).
ACEindexbetweenNPRandMIR(p>0.05)(Table1).
Int.J.Mol.Sci. 2018,19,850 4of17
Table1.Speciesrichnessanddiversityindicesindifferentsoilsamples.
Categories ObservedSpecies Chao1 ACE Shannon
Rhizosphere
NPR 495.7c 566.4c 593.7b 5.042c
MUR 654.7a 718.5a 742.5a 5.675a
MIR 555.3b 628.3b 651.9b 5.493b
Rhizoplane
NPP 575.0b 634.3a 652.6ab 5.735a
MUP 696.3a 755.9a 766.7a 5.975a
MIP 431.7c 615.6a 565.8b 2.496b
NPR,MUR,andMIRrepresenttherhizospheresoilfromthenewlyplanted,monocultureduninfectedandinfected
plants,respectively. NPP,MUP,andMIPrepresenttherhizoplanesoilfromthenewlyplanted,monocultured
uninfectedandinfectedplants,respectively.Differentlettersinthecolumnsshowsignificantdifferencesbetweenthe
differenttreatmentsintherhizosphere(NPRvs.MURvs.MIR)orrhizoplane(NPPvs.MUPvs.MIP),determined
byTukey’stest(p<0.05).
Intherhizoplanesoil,therewasnosignificantdifferencebetweenMUPandNPPintermsofthe
Shannondiversityindexandmostofrichnessindices(Chao1andACE).Richnessindices(including
observedspeciesandACE)andShannondiversityindexweresignificantlyhigherinMUPthanin
MIP(p<0.05)(Table1). Inaddition,observedspeciesandShannondiversityindexweresignificantly
higherintherhizoplanesoilthanintherhizospheresoilofNPandMUplants(NPRvs. NPPandMUR
vs. MUP).However,thesetwoindiceswerehigherintherhizospheresoilofMIplants(TableS3).
2.3. BetaDiversityAnalysis
In this study, principal coordinate analysis (PCoA) and unweighted pair-group method with
arithmetic mean (UPGMA) clustering based on unweighted UniFrac metric (UUF) and weighted
UniFracmetric(WUF)wereperformedtoinvestigatebetadiversitypatternsbetweencomplexbacterial
communities(Figure2). BothPCoAandUPGMAclusteringbasedonUUFdistancesindicatedthat
themicrobiomeswereseparatedbyrhizocompartment(rhizospherevs. rhizoplane)andlongevity
ofmonoculture(one-yearvs. two-yearmonoculture)(Figure2A,B).ThefirsttwoaxesPC1andPC2
accounted for 21.38% and 18.42% of total variation, respectively. By contrast, PCoA and UPGMA
clusteringbasedonWUFdistancesshowedadiscernibleseparationamongthenewlyplanted(NP),
two-year monocultured uninfected (MU), and monocultured infected (MI) plants (Figure 2C,D).
The first two axes PC1 and PC2 accounted for 77.78% and 10.07% of total variation, respectively.
Furthermore,non-metricmultidimensionalscaling(NMDS)ordinations(stressvalue=0.101)showed
thattheplotsmonoculturedfordifferentyears(one-yearvs. two-yearmonoculture)separatedacross
thefirstMDSaxisandtherhizocompartments(exceptforMIP1)separatedacrossthesecondMDSaxis
(FigureS3).
Int.J.Mol.Sci. 2018,19,850 5of17
Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 5 of 17

Figure 2. Principal coordinate analysis (PCoA) and unweighted pair-group method with arithmetic
Figure2.Principalcoordinateanalysis(PCoA)andunweightedpair-groupmethodwitharithmetic
mean (UPGMA) dendrogram of bacterial communities based on unweighted (A,B) and weighted
mean(UPGMA)dendrogramofbacterialcommunitiesbasedonunweighted(A,B)andweighted
(C,D) Unifrac distance. NPR, MUR, and MIR represent the rhizosphere soil from the newly planted,
(C,D)Unifracdistance.NPR,MUR,andMIRrepresenttherhizospheresoilfromthenewlyplanted,
monocultured uninfected and infected plants, respectively. NPP, MUP, and MIP represent the
monocultured uninfected and infected plants, respectively. NPP, MUP, and MIP represent the
rhizoplane soil from the newly planted, monocultured uninfected and infected plants, respectively.
rhizoplanesoilfromthenewlyplanted,monocultureduninfectedandinfectedplants,respectively.
2.4. Venn Diagram and Similarity Analysis
2.4. VennDiagramandSimilarityAnalysis
Venn diagram analysis showed that the number of OTUs exclusively found in each group was
VenndiagramanalysisshowedthatthenumberofOTUsexclusivelyfoundineachgroupwas
much lower than that of shared OTUs in both rhizosphere and rhizoplane (Figure S4A,B). Meanwhile,
muchlowerthanthatofsharedOTUsinbothrhizosphereandrhizoplane(FigureS4A,B).Meanwhile,
the number of OTUs exclusively found in the rhizosphere or rhizoplane was much lower than that
thenumberofOTUsexclusivelyfoundintherhizosphereorrhizoplanewasmuchlowerthanthatof
of OTUs shared between the rhizosphere and rhizoplane in the three different groups (Figure S4C–
OTUssharedbetweentherhizosphereandrhizoplaneinthethreedifferentgroups(FigureS4C–E).
E). Permutational multivariate analysis of variance (permutational MANOVA) showed that the
Permutationalmultivariateanalysisofvariance(permutationalMANOVA)showedthatthebacterial
bacterial communities differed significantly (p < 0.05) between the rhizosphere and rhizoplane for all
communities differed significantly (p < 0.05) between the rhizosphere and rhizoplane for all three
three different groups (NPR vs. NPP, MUR vs. MUP, MIR vs. MIP) (Table 2). In both the rhizosphere
different groups (NPR vs. NPP, MUR vs. MUP, MIR vs. MIP) (Table 2). In both the rhizosphere
and rhizoplane, there was a significant difference between the newly planted and the monocultured
andrhizoplane,therewasasignificantdifferencebetweenthenewlyplantedandthemonocultured
infected groups (NPR vs. MIR, NPP vs. MIP). However, no significant difference was found between
infectedgroups(NPRvs. MIR,NPPvs. MIP).However,nosignificantdifferencewasfoundbetween
the newly planted and the monocultured uninfected groups (NPR vs. MUR, NPP vs. MUP) (Table 2).
thenewlyplantedandthemonocultureduninfectedgroups(NPRvs. MUR,NPPvs. MUP)(Table2).
Furthermore, a two-way ANOSIM test of the pyrosequencing data consisting of the relative
Furthermore,atwo-wayANOSIMtestofthepyrosequencingdataconsistingoftherelativeabundance
abundance of genera (logarithmic transformation and standardization, Bray–Curtis similarity)
ofgenera(logarithmictransformationandstandardization,Bray–Curtissimilarity)showedthatthe
showed that the bacterial communities differed significantly among the three different groups
(ANOSIM Global R = 1.0, p = 0.001) and between the rhizosphere and rhizoplane (ANOSIM Global R
= 1.0, p = 0.003). Two-way ANOSIM analysis based on the presence/absence of genera (Bray-Curtis
Int.J.Mol.Sci. 2018,19,850 6of17
bacterialcommunitiesdifferedsignificantlyamongthethreedifferentgroups(ANOSIMGlobalR=1.0,
p=0.001)andbetweentherhizosphereandrhizoplane(ANOSIMGlobalR=1.0,p=0.003). Two-way
ANOSIM analysis based on the presence/absence of genera (Bray-Curtis similarity) also showed
thebacterialcommunitiesweresignificantlyaffectedbythelongevityofthemonoculture(ANOSIM
GlobalR=0.735,p=0.001)andrhizocompartment(ANOSIMGlobalR=0.975,p=0.001).
Table2.Permutationalmultivariateanalysisofvariance(permutationalMANOVA)toinvestigatethe
differencesinbacterialcommunitiesbetweendifferentsoilcategories.
Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 6 of 17
PairwiseComparison R2 p-Value*
similarity) also showed the bacterial communities were significantly affected by the longevity of the
Rhizosphere
monoculture (ANOSIM Global R = 0.735, p = 0.001) and rhizocompartment (ANOSIM Global R =
0.975, p = 0.001). NPR-MUR 0.9778 0.1000
NPR-MIR 0.9735 0.0014
Table 2. PermMutUatRion-Mal mIRultivariate analysis of var0ia.n9c3e 4(p0ermutational MANO0V.A0)0 t1o 4investigate
the differences in bacterial communities between different soil categories.
Rhizoplane
Pairwise Comparison R2 p-Value*
NPP-MUP 0.9158 0.1000
Rhizosphere
NPP-MIP NPR-MUR 00.9.9876787 0.1000 0.0014
MUP-MIP NPR-MIR 00.9.9972359 0.0014 0.0083
MUR-MIR 0.9340 0.0014
Rhizospherevs.rhizoplane
Rhizoplane
NPR-NPP NPP-MUP 00.9.9518580 0.1000 0.0014
MUR-MUP NPP-MIP 00.9.9287676 0.0014 0.0014
MIR-MIP MUP-MIP 00.9.9991293 0.0083 0.0417
Rhizosphere vs. rhizoplane
R2representstheratioofbetween-groupvariancetototalvariance.*Significantdifferencesareindicatedbybold
NPR-NPP 0.9580 0.0014
p-values.
MUR-MUP 0.9276 0.0014
MIR-MIP 0.9913 0.0417
2.5. ShiftsinSoilBaRc2t ererpiraesleCntso mthem rautino iotfy beStwtreuenc-gtruourpe uvanridanecre Cto otontasle vcauritainvcee. M* Siognnioficcaunlt tduifrfeerences are
indicated by bold p-values.
Atthephylumlevel,themajorityoftheOTUsinallsoilcategorieswereassignedtoProteobacteria
2.5. Shifts in Soil Bacterial Community Structure under Consecutive Monoculture
(83.2~97.6%),Actinobacteria(1.4~9.3%),Bacteriodetes(0.5~9.3%),andFirmicutes(0.3~4.5%)(Figure3).
At the phylum level, the majority of the OTUs in all soil categories were assigned to Proteobacteria
However,clear(t8r3e.2n~9d7.s6%in), Avcatirnioabatcitoernia a(1t.4t~h9.e3%p),h Byacltuermiodelteesv (0e.l5~w9.3e%r)e, aonbd sFeirrmviceudtesb (0e.t3w~4.e5%en) (FthigeureN 3P). groupandthe
monoculturedgHroowuevpesr, (cNleaPr trvensd.sM in Uvaraiantiodn Nat tPhev psh.ylMumI )le(vFeli wgeurer eob3se)r.ved between the NP group and the
monocultured groups (NP vs. MU and NP vs. MI) (Figure 3).

Figure 3. Relative abundances of the top 10 bacterial phyla in six different soil categories. NPR, MUR,
Figure3.Relatiavned aMbIRu nredpraesnecnet tsheo rfhtizhoespthoerpe s1o0il bfraomct tehrei anelwplyh pyllaantiend asnidx mdoinffoecurletunrteds ounilincfeactteedg aonrdi es.NPR,MUR,
andMIRrepresinefencttetdh pelarnhts,i zreosspepchtiveerlye. sNoPiPl, fMroUPm, atnhd eMnIPe wreplryespenlta tnhet erdhizaopnladnem sooiln forocmu lthtue rneewdlyu ninfectedand
planted and monocultured uninfected and infected plants, respectively.
infectedplants,respectively.NPP,MUP,andMIPrepresenttherhizoplanesoilfromthenewlyplanted
andmonocultuFreirdmiucunteisn fweacst edsigannifdicainntlfye c(tpe d< p0l.a05n)t sd,ercerseapseedct iwvehlilye. Bacteroidetes, Gemmatimonadetes,
Acidobacteria, and Verrucomicrobia were significantly (p < 0.05) increased in both rhizosphere and
rhizoplane under consecutive monoculture (NPR vs. MUR, NPP vs. MUP, Figure 3). In addition, a
Firmicutes was significantly (p < 0.05) decreased while Bacteroidetes, Gemmatimonadetes,

Acidobacteria, and Verrucomicrobia were significantly (p < 0.05) increased in both rhizosphere and
rhizoplaneunderconsecutivemonoculture(NPRvs. MUR,NPPvs. MUP,Figure3). Inaddition,a
higherabundanceofProteobacteriawasdetectedinthemonoculturedrhizoplane(MUP)comparedto
NPP(Figure3).
Int.J.Mol.Sci. 2018,19,850 7of17
Both Actinobacteria and Firmicutes were significantly (p < 0.05) reduced in monocultured
infected plants than in the newly planted plants (NPR vs. MIR, NPP vs. MIP) regardless of the
rhizocompartment. More precisely, Proteobacteria, Actinobacteria, and Firmicutes were significantly
(p < 0.05) reduced in the rhizosphere of the MI plant group (MIR) while Bacteroidetes and
Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 7 of 17
Gemmatimonadetes significantly (p < 0.05) increased in MIR compared to NPR (Figure 3). In the
higher abundance of Proteobacteria was detected in the monocultured rhizoplane (MUP) compared to
rhizoplaneoftheMIplantgroup(MIP),Actinobacteria,Firmicutes,Gemmatimonadetes,andAcidobacteria
NPP (Figure 3).
weresignificantly(Bpot<h A0c.t0in5o)balcotewriae arndw FhirimliecuPters owteeroeb saigcntiefirciaantwly a(ps <s 0i.g05n) irefidcuacendt liny m(pon<ocu0l.t0ur5e)d hinifgechteedr incomparison
plants than in the newly planted plants (NPR vs. MIR, NPP vs. MIP) regardless of the
withNPP(Figure3). Interestingly,thephylumFirmicuteswassignificantly(p<0.05)reducedinboth
rhizocompartment. More precisely, Proteobacteria, Actinobacteria, and Firmicutes were significantly (p
consecutivemo<n o0.c05u) ltruedruecegdr oinu pthse (MrhizUospahnedre Mof I)three gMaI rdpllaenst sgorofutph e(MrIhRi) zwohciolem Bpacaterrtomideetens tan(Fd igure3).
Gemmatimonadetes significantly (p < 0.05) increased in MIR compared to NPR (Figure 3). In the
Furthermore,thelineardiscriminantanalysiseffectsize(LEfSe)methodwasappliedtoanalyzethe
rhizoplane of the MI plant group (MIP), Actinobacteria, Firmicutes, Gemmatimonadetes, and
differentialabuAncdidaobnaccteerioa wfesroe isligbnaifcictaenrtliya (lp t<a 0x.0a5) alotwtehr ewhfailem Pirloyteoabanctderiga wenasu ssignleifvicaenltslyb (pe t<w 0.0e5e) nhigthheer newlyplanted
andconsecutivien lcyommpaorinsoonc wuilthtu NrPePd (Fgigruoreu 3p).s I,ntaersessthinoglwy, tnheb pyhytlhume Fliirnmeicaurtesd wisasc sriigmnifiincaanntlyt (ap n< a0.l0y5)s is(LDA)score
reduced in both consecutive monoculture groups (MU and MI) regardless of the rhizocompartment
inFigure4. TheresultsindicatedthatmicrobialdysbiosiswasprevalentinR.glutinosarhizosphere
(Figure 3).
(Figure4A)andrhFiuzrothperlmaonree, t(hFei lginuearre d4isBcri)muinnandt earnaclyosnis sefefeccut stiizve e(LmEfSoe)n moecthuoldt uwraes .apIpnlietdh teo aNnaPlyzreh izosphere,the
the differential abundance of soil bacterial taxa at the family and genus levels between the newly
dominantmicrobialtaxaincludedthefamiliesofEnterobacteriaceae,Pseudomonadaceae,Micrococcaceae,
planted and consecutively monocultured groups, as shown by the linear discriminant analysis (LDA)
and Bacillaceaescaonred int hFiegurgee 4n. eTrhae rPessueluts diondmicoanteads t,haBt amciiclrloubsia,l adnysdbioAsisr twharso bpraecvtaelern.t inIn R.c goluntitnroasas t, the families
rhizosphere (Figure 4A) and rhizoplane (Figure 4B) under consecutive monoculture. In the NP
Comamonadaceae(inMUR)andFlavobacteriaceae,Sphingomonadaceae,andXanthomonadaceae(inMIR)
rhizosphere, the dominant microbial taxa included the families of Enterobacteriaceae,
wereenrichedinPsetuhdeomrohniazdaocseape,h Meriecroocofctcahceeaet,w aond-y eBaacrillmaceoaen oancdu ltthuer egdenperlaa nPtsseu.dGomeonnaes,r aBRacihlliuzso, bainudm (inMUR)and
ChryseobacteriumAr,thSropbhacitnerg. obIniu mco,natrnasdt, Sttheen oftarmoipliheso mCoomnaamso(niandacMeaeI R()inw eMrUeRe) nraincdh eFdlavionbatcthereiacrehaei,z osphereofthe
Sphingomonadaceae, and Xanthomonadaceae (in MIR) were enriched in the rhizosphere of the two-year
two-yearmonoculturedplants(Figure4A).Intherhizoplane,thefamiliesRhizobiaceaeandBacillaceae
monocultured plants. Genera Rhizobium (in MUR) and Chryseobacterium, Sphingobium, and
andgeneraRhiSztoenboituromphoamnondasB (ianc MilIlRu)s wweree reenrsicihgend iifin cthaen rthliyzosmphoerree oaf bthue ntwdoa-ynetari nmotnhoecunltuerwedl yplapnltsa ntedwhilethe
(Figure 4A). In the rhizoplane, the families Rhizobiaceae and Bacillaceae and genera Rhizobium and
familiesComamonadaceae,Xanthomonadaceae,andAlcaligenaceaeandgeneraLysobacter,Variovorax,and
Bacillus were significantly more abundant in the newly planted while the families Comamonadaceae,
BordetellawereXsiagntnhoimfiocnaadnactelaye, manodr eAalcbaluigenndacaeane tainnd thgeenetrwa oL-yysoebaacrterc, oVnasrieovcouratxi,v aenmd oBonrdoecteullal tuwreere (Figure4B).
significantly more abundant in the two-year consecutive monoculture (Figure 4B).

Figure 4. The linear discriminant analysis effect size (LEfSe) identified the most differentially
abundanttaxabetweenthenewlyplantedandthetwo-yearmonoculturedintherhizosphere(A)
andrhizoplane(B).Onlytaxameetingalineardiscriminantanalysis(LDA)significantthreshold>4
areshown.Theprefixes‘f_’and‘g_’meanfamilylevelandgenuslevel,respectively.NPR,MUR,and
MIRrepresenttherhizospheresoilfromthenewlyplantedandmonocultureduninfectedandinfected
plants,respectively. NPP,MUP,andMIPrepresenttherhizoplanesoilfromthenewlyplantedand
monocultureduninfectedandinfectedplants,respectively.
Int. J. Mol. Sci. 2018, 19, x FOR PEER REVIEW 8 of 17
Figure 4. The linear discriminant analysis effect size (LEfSe) identified the most differentially
abundant taxa between the newly planted and the two-year monocultured in the rhizosphere (A) and
rhizoplane (B). Only taxa meeting a linear discriminant analysis (LDA) significant threshold >4 are
shown. The prefixes ‘f_’ and ‘g_’ mean family level and genus level, respectively. NPR, MUR, and
MIR represent the rhizosphere soil from the newly planted and monocultured uninfected and infected
plants, respectively. NPP, MUP, and MIP represent the rhizoplane soil from the newly planted and
Int.J.Mmolo.nSocic.u2lt0u18re,1d9 u,8n5i0nfected and infected plants, respectively. 8of17
2.6. Isolation of Pseudomonas and Bacillus and Their Antagonistic Activities Assessment
2.6. IsolationofPseudomonasandBacillusandTheirAntagonisticActivitiesAssessment
Both Pseudomonas and Bacillus, commonly used as biocontrol agents of plant pathogens, were
BothPseudomonasandBacillus,commonlyusedasbiocontrolagentsofplantpathogens,were
enriched in the newly planted soil through LEfSe analysis. Therefore, Pseudomonas spp. and Bacillus
enrichedinthenewlyplantedsoilthroughLEfSeanalysis.Therefore,Pseudomonasspp.andBacillusspp.
spp. were isolated from the newly planted soil using two selective screening agar to evaluate their
wereisolatedfromthenewlyplantedsoilusingtwoselectivescreeningagartoevaluatetheirrolesin
roles in maintaining soil health. Pseudomonas spp. and Bacillus spp. were co-cultured with two main
maintainingsoilhealth. Pseudomonasspp. andBacillusspp. wereco-culturedwithtwomaincausal
causal agents of R. glutinosa root rot disease including Fusarium oxysporum and Aspergillus flavus [3].
agentsofR.glutinosarootrotdiseaseincludingFusariumoxysporumandAspergillusflavus[3]. Strong
Strong antagonistic activities were detected in many Pseudomonas spp. and Bacillus spp. (Figure 5A,B),
antagonisticactivitiesweredetectedinmanyPseudomonasspp. andBacillusspp. (Figure5A,B),which
which suggest these two genera are critical for maintaining soil health. The identities of the
suggestthesetwogeneraarecriticalformaintainingsoilhealth. Theidentitiesoftheantagonistic
antagonistic isolates were further confirmed with Sanger sequencing on the 16S rRNA gene region
isolateswerefurtherconfirmedwithSangersequencingonthe16SrRNAgeneregion(Figure5C).
(Figure 5C).

Figure 5. Assessment of antagonistic activity (A,B) and hierarchical clustering of 16S rDNA genes (C)
Figure5. Assessmentofantagonisticactivity(A,B)andhierarchicalclusteringof16SrDNAgenes
of Pseudomonas spp. and Bacillus spp. isolated from the newly planted soil. (A) Strain Z9 (GenBank:
(C) of Pseudomonas spp. and Bacillus spp. isolated from the newly planted soil. (A) Strain Z9
MH037322) and Z12 (GenBank: MH037323) but not Z11 have antagonistic activity against F.
(GenBank:MH037322)andZ12(GenBank:MH037323)butnotZ11haveantagonisticactivityagainst
oxysporum (FON). (B) Strain Z34 (GenBank: MH037324) has antagonistic activity against A. flavus
F.oxysporum(FON).(B)StrainZ34(GenBank:MH037324)hasantagonisticactivityagainstA.flavus
(AF). (C) Different Pseudomonas and Bacillus species were used as references. Eight clusters were
(AF). (C) Different Pseudomonas and Bacillus species were used as references. Eight clusters were
identified for both Pseudomonas (PS-I to PS-VIII) and Bacillus (BA-I to BA-VIII). Numbers of isolates of
identifiedforbothPseudomonas(PS-ItoPS-VIII)andBacillus(BA-ItoBA-VIII).Numbersofisolatesof
each cluster are indicated in parentheses. Three isolates (Z9, Z12 and Z34) with antagonistic activities
eachclusterareindicatedinparentheses.Threeisolates(Z9,Z12andZ34)withantagonisticactivities
in A and B are presented in the phylogenetic tree with blue font.
inAandBarepresentedinthephylogenetictreewithbluefont.
3. Discussion
3. Discussion
Replant disease is a very common but complicated phenomenon in the agricultural production
Replantdiseaseisaverycommonbutcomplicatedphenomenonintheagriculturalproductionof
of many crops, especially for Chinese medicinal plants. Our field experiment showed that
manycrops,especiallyforChinesemedicinalplants. Ourfieldexperimentshowedthatconsecutive
consecutive monoculture practice has a profound effect on the growth and health of R. glutinosa,
monoculturepracticehasaprofoundeffectonthegrowthandhealthofR.glutinosa,whichresults
which results in poor plant performance and serious root rot disease (Figure S1). Recently, numerous
inpoorplantperformanceandseriousrootrotdisease(FigureS1). Recently,numerousstudieshave
studies have indicated that indirect allelopathy through modifications in soil microbial community
indicated that indirect allelopathy through modifications in soil microbial community is closely
is closely related to replant disease in agriculture and horticulture [10,21]. The complex plant-
related to replant disease in agriculture and horticulture [10,21]. The complex plant-associated

microbialcommunityiscrucialforsoilhealthandplantgrowth,whichisalsoreferredtoasasecond
genometoaplant[22,23]. Therhizosphereisthebattlefieldforsoil-bornepathogensandbeneficial
microorganisms[24],whiletherhizoplane,orroottissuesurface,playsaselectivegatingroleinroot
microbiomeassemblyfromthesoil[19]. Inthisstudy,parallelpyrosequencingwasappliedtoexplore
theshiftsinbacterialcommunitiesintherhizosphereandrhizoplaneofR.glutinosaandhighlighted
Int.J.Mol.Sci. 2018,19,850 9of17
thatconsecutivemonocultureofR.glutinosasignificantlyrestructuresthebacterialcommunitiesin
boththerhizosphereandtherhizoplane. PCoAanalysisbasedonweightedandunweightedUniFrac
distanceshighlighteddistinctdifferencesinbacterialcommunitystructureamongthesoilssampled
from the newly planted, the monocultured uninfected and infected plants (Figure 2). Moreover,
PCoAalsosuggestedsignificantdifferencesinsoilbacterialcommunitiesbetweenrhizocompartments
(Figure2),whichiswellalignedwithpreviousstudiesonOryzasativaandSolanumlycopersicum[19,25].
Ahighlevelofmicrobialdiversityisgenerallyconsideredasanimportantparameterforsoil
health[26]. Forexample,areductioninspeciesdiversitywiththeextendedmonocultureperiodwas
reportedinSolanumtuberosum[27]. However,thisstudysuggestedmostofrichnessanddiversity
indicesweresignificantlyhigherinthetwo-yearmonoculturedsoilthaninthenewlyplantedsoilinthe
rhizosphere(NPRvs. MUR;NPRvs. MIR)andwassimilarintherhizoplane(NPPvs. MUP)(Table1).
ThisresultisalignedwithpreviousresearchonVanillaplanifolia[28]. Six-yearmonocultureresulted
inasignificantlyhigherspeciesdiversityintherhizosphere[28]. Inaddition,Xiongetal.[28]also
suggestedthatalowbacterialdiversityandhighfungaldiversitycouldassociatewiththesuppression
ofvanillaFusariumwiltdiseaseindisease-suppressivesoil. Therefore,thediscrepancybetweenthese
different studies might be due not only to the longevity of monoculture but also to the different
soilenvironmentalconditions,hosttypes,rhizocompartment,etc. Morerecentresearchsuggested
thatthekeycomponentofsoilecosystemfunctionsisnottaxonomicdiversitybutratherfunctional
diversity[29,30].
Taxonomicanalysisshowedthatinboththerhizosphereandrhizoplane,therelativeabundance
of Actinobacteria was significantly lower in MI and Firmicutes was significantly lower in MU and
MI when compared with NP (Figure 3). This is in line with previous findings where members
of Firmicutes and Actinobacteria were consistently associated with pathogen antagonism and soil
diseasesuppression[31–33]. Xiongetal.[28]comparedmicrobialcommunitiesinsuppressive-and
conducive-soilsassociatedwithFusariumwiltdiseaseandfoundthatActinobacteriaandFirmicutes
were strongly enriched in the suppressive soil in a vanilla long-term continuous cropping system.
Yangetal.[18]indicatedthatsomeprobioticbacteriabelongingtoActinobacteriasignificantlydecreased
withtheextendedmonocultureofR.glutinosa. Inaddition,diseasesuppressionthatisinducedby
bioorganic fertilizer application in Nicotiana tabacum and Persea americana was associated with an
increaseinmultiplebeneficialstrainssuchasActinobacteria,Firmicutes,etc.[34,35]. Apreviousstudy
onbananaPanamadiseasecausedbyFusariumoxysporumf. sp. cubenseshowedthatFirmicutesand
Actinobacteria were more abundant in disease-suppressive soil while Verrucomicrobia, Bacteroidetes,
Proteobacteria, and Gemmatimonadetes were more abundant in diseased soil [32]. In this study, we
alsofoundBacteroidetes,Gemmatimonadetes,Verrucomicrobia,etc. tobesignificantlyincreasedwhile
FirmicutesandActinobacteriaweredecreasedunderconsecutivemonoculture(Figure3).
The LEfSe analysis indicated a concomitant increase in the families of Flavobacteriaceae
and Sphingomonadaceae with the consecutive monoculture of R. glutinosa (Figure 4). Similarly,
Flavobacteriaceae and Sphingomonadaceae were significantly more abundant in 4- and 15-year
monocropped cotton rhizosphere soil compared to fallow soil [36]. Members of Flavobacteriaceae
andSphingomonadaceaefrequentlysynthesizecellulose-,pectin-,xylan-,andchitin-degradingenzymes
todecomposedeadplantsresultingfromfungalinvasion[36–38]. Therefore,theconcomitantincrease
inthemembersofFlavobacteriaceaeandSphingomonadaceaeinthesoilcouldbeimportantindicatorsof
rootdamagebysoilfungalpathogens[36,38]. Wuetal.[3]reportedthehighabundanceofsoilfungal
pathogensFusariumoxysporumandAspergillusflavusunderconsecutivemonocultureofR.glutinosa.
Thus,theincreaseinFlavobacteriaceaeandSphingomonadaceaedetectedinthisstudymightbecaused
bytheFusariumoxysporumandAspergillusflavusinvasioninR.glutinosa. Furthermore,itwasfound
thatthegenusStenotrophomonaswasenrichedintheconsecutivelymonoculturedsoil,especiallyinthe
rhizosphereofinfectedplants(MIR)(Figure4). Tanetal.[39]alsofoundrelativelyhigherproportions
ofgenerasuchasStenotrophomonas,Sphingobium,andErwinia,etc. intherhizosphereofroot-rotPanax
notoginseng under consecutive monoculture. In contrast, it was reported that several members of
Int.J.Mol.Sci. 2018,19,850 10of17
Stenotrophomonascouldreducetheecotoxicityofcontaminantsinthesoil,promoteplantgrowth,orbe
biologicalcontrolagentsofplantpathogens[40–42]. Therefore,furtherworkisneededtostudythe
rolesofStenotrophomonasintherhizosphereinR.glutinosamonocultureregimes.
Besides, LEfSe analysis indicated that Enterobacteriaceae, Pseudomonadaceae, Bacillaceae,
Micrococcaceae, Pseudomonas, Bacillus, and Arthrobacter were enriched in the newly planted group
(Figure 4), possibly serving as important indicators for soil disease suppression. Many members
oftheEnterobacteriaceae,PseudomonadaceaeandBacillaceaefamilieshavebeenreportedasbiocontrol
agentscapableofsuppressingfungalandinsectgrowth[43–46]. Mendesetal.[31]decipheredthe
rhizospheremicrobiomefordisease-suppressivebacteriaandindicatedthattherelativeabundances
ofPseudomonadaceaeandPseudomonasspecieswerepositivelycorrelatedwithdiseasesuppressionof
sugarbeet. Inaddition,theabundanceofthegeneraPseudomonas,Gp5,etc. werepositively(p<0.05)
correlatedwithFusarium-wiltsuppressionofbanana[47]. Similarly,beneficialbacterialpopulations
Pseudomonas and Bacillus in the rhizosphere were significantly reduced after apple replanting [48].
Furthermore, Xue et al. [32] highlighted that Bacillus was the top contributor to the community
differences between disease-suppressive soil and diseased soil, and the family of Bacillaceae was
reported as the dominant component of the ‘core microbiome’ from the disease-suppressive soil.
Xiong et al. [49] found that the abundance of beneficial bacteria (i.e., Bacillus, etc.) significantly
decreasedalongtheyearsofV.planifoliamonoculture, whichmightleadtothesoilweaknessand
vanilla stem wilt disease under long-term monoculture. Our study (Figure 4) also coincides with
previous studies, which suggested the important roles of the members of Micrococcaceae, such as
Arthrobacter species, in soil disease suppression [34,43,50,51]. Despite these interpretations, there
is a great need to explore the interactions between soil pathogenic and beneficial microbes in situ
underR.glutinosamonoculturebecausemicrobialinteractionsinthephytospherearecrucialforplant
health[46].
MembersofPseudomonasandBacillushavebeenextensivelyusedaspotentialantagonisticagents
againstvariousplantpathogens[52]. Groschetal.[53]foundthatPseudomonasfluorescensB1couldbe
selectedasaneffectivebiologicalcontrolagentagainstRhizoctoniasolaniinlettuceandpotatoplants.
Zhangetal.[54]demonstratedthatBacillussubtilisN11-fortifiedbioorganicfertilizercaneffectively
controlbananawilt. ThisstudyhighlightedthatmanyPseudomonasspp. andBacillusspp. isolates
fromthenewlyplantedsoilshowedstrongantagonisticactivitiesagainsttwomainfungalpathogens
(i.e., F. oxysporum and A. flavus) of R. glutinosa root rot disease (Figure 5). Moreover, a biocontrol
testinapreviousstudyshowedthatantagonisticPseudomonasspp. isolatedfromtherhizosphereof
R.glutinosacouldeffectivelyprotectR.glutinosaplantsagainstF.oxysporuminfection[2]. Ourprevious
studyalsoshowedthattheabundanceofantagonisticPseudomonasspp. againstF.oxysporumgradually
decreased[2],whilethefungalpathogenssuchasF.oxysporumandA.flavusincreasedunderR.glutinosa
consecutivemonoculture[2,3]. Therefore,thedecreaseinbeneficialbacteria(i.e.,Pseudomonasand
Bacillus) with the antagonistic activities against pathogens might contribute to the declined soil
suppressivenesstofungalinvasioninconsecutivemonocultureregimes. Theisolationandbiocontrol
potentialassessmentofmembersofCommamonadaceaeandRhizobiaceaearenowunderway.
4. MaterialsandMethods
4.1. FieldExperimentandSoilSampling
TheR.glutinosacultivar“85-5”iswidelyplantedonalarge-scaleinthegeo-authenticproduction
zonesandwasselectedastheexperimentalmaterial. TheexperimentwasconductedatJiaozuoCity,
HenanProvince,China(34◦56(cid:48)N,112◦58’E),whichhasanannualmeantemperatureof14.3◦Candan
annualmeanprecipitationof552mm. Theexperimentwasconductedinasinglefieldsitethathadnot
beenplantedwithR.glutinosaformorethan10yearstoexcludetheeffectsofprecedingresiduesofthis
plantandtoensuretheuniformgrowthconditionamongtreatments. Thephysicochemicalproperties
ofthesoilwereasfollows: sandyloamsoil,soilorganicmatter12.52g·kg−1,pH7.43,totalnitrogen