the effects of afforestation on soil bacterial communities in … · 2019-01-11 · at each...

Submitted 30 May 2018Accepted 21 November 2018Published 11 January 2019

Corresponding authorsShu-Hong WuwshuhongbjfueducnPei-Chun Liao pcliaontnuedutw

Academic editorMiquel Gonzalez-Meler

Additional Information andDeclarations can be found onpage 18

DOI 107717peerj6147

Copyright2019 Wu et al

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

The effects of afforestation on soilbacterial communities in temperategrassland are modulated by soil chemicalpropertiesShu-Hong Wu1 Bing-Hong Huang2 Jian Gao3 Siqi Wang1 and Pei-Chun Liao2

1 School of Nature Conservation Beijing Forestry University Beijing China2Department of Life Science National Taiwan Normal University Taipei Taiwan3 Faculty of Resources and Environment Baotou Teachersrsquo College Inner Mongolia University of Science andTechnology Inner Mongolia China

ABSTRACTGrassland afforestation dramatically affects the abiotic biotic and ecological functionproperties of the original ecosystems Interference from afforestation might disruptthe stasis of soil physicochemical properties and the dynamic balance of microbiotaSome studies have suggested low sensitivity of soil properties and bacterial communityto afforestation but the apparent lack of a significant relationship is probably dueto the confounding effects of the generalist habitat and rare bacterial communitiesIn this study soil chemical and prokaryotic properties in a 30-year-old Mongoliapine (Pinus sylvestris var mongolica Litv) afforested region and adjacent grasslandin Inner Mongolia were classified and quantified Our results indicate that the highrichness of rare microbes accounts for the alpha-diversity of the soil microbiome FewOTUs of generalist (core bacteria) and habitat-specialist bacteria are present Howeverthe high abundance of this small number of OTUs governs the beta-diversity of thegrassland and afforested land bacterial communities Afforestation has changed thesoil chemical properties thus indirectly affecting the soil bacterial composition ratherthan richness The contents of soil P Ca2+ and Fe3+ account for differentially abundantOTUs such as Planctomycetes and subsequent changes in the ecologically functionalpotential of soil bacterial communities due to grassland afforestation We concludethat grassland afforestation has changed the chemical properties and composition ofthe soil and ecological functions of the soil bacterial community and that these effectsof afforestation on the microbiome have been modulated by changes in soil chemicalproperties

Subjects Agricultural Science Conservation Biology Ecology Microbiology ForestryKeywords Grassland afforestation Microbial composition Microbial ecological function Soilchemical properties

INTRODUCTIONAfter a decline in forest coverage in China from roughly 40ndash60 in remote antiquity (He etal 2008) to only 86 due to war urban construction and reclamation efforts in the last30 years by the Chinese government to promote afforestation have increased forest coverageto nearly 20 (State Forestry Administration of China 2011) However these plantations

How to cite this article Wu S-H Huang B-H Gao J Wang S Liao P-C 2019 The effects of afforestation on soil bacterial communities intemperate grassland are modulated by soil chemical properties PeerJ 7e6147 httpdoiorg107717peerj6147

are not exclusively located at the original sites of deforestation and consequently thesedeforestation and afforestation events have greatly changed the landscape and ecosystemsof China (Ahrends et al 2017) Mongolia pine (Pinus sylvestris var mongolica Litv) anendemic tree in Inner Mongolia including in Honghuarsquoerji is one of the main forestationspecies in temperate regions of China Scots pine (P sylvestris L) a relative of Mongoliapine that is widespread from Western Europe to Eastern Siberia is also an importanttree in forestry (Krakau et al 2013) and is physiologically sensitive to environmentalpollution (Chudzińska amp Diatta 2014) geographic weather variation (Oleksyn et al 2003)and climate change (Hurme et al 1997 Savolainen et al 2004) Studies have demonstratedthat Scots pine and by reasonable extensionMongolia pine have broad plastic adaptabilityin response to environmental heterogeneity supporting the wide use of these trees inafforestation

Grassland afforestation is an artificial and direct change in vegetation that alters boththe above- and underground ecosystems including abiotic changes (Li et al 2016 Penget al 2014 Khamzina Lamers amp Martius 2016 Lafleur et al 2015) biotic changes (Maet al 2013 Maacuterquez et al 2015 Pedley et al 2014 Gunina et al 2017 Šnajdr et al 2013Xiao et al 2017) and ecological function changes (Lu et al 2017 Ren et al 2016 Cibils etal 2015) Land use is a significant determinant of runoff and soil redistribution processes(Arnaacuteez et al 2015) and soil microbial species responses may be sensitive to compositionalchanges in both species and ecological functions (Xiao et al 2017) However theseunderground biological changes are sometimes ascribed to changes in litter amountand chemistry that stimulate the development of the fungal community rather than thebacterial community (Klein et al 1995) The response of the prokaryotic microbiome tochanges in vegetation type is small likely because bacterial assemblies are based mainly onfunctional genes rather than species (Burke et al 2011) and the changes in soil propertiesdue to afforestation are relatively small and occur slowly (Gunina et al 2017 Jangid et al2011) In addition abundant bacterial lsquolsquocore speciesrsquorsquo with highly conserved core functions(Falkowski Fenchel amp Delong 2008) and rarely occurring species (Ai et al 2013) may actas confounders in statistical analyses of the effects of afforestation or soil properties onprokaryoticmicrobiome change Therefore habitat generalists (core species) and specialists(divergent species) in bacterial communities must be classified before analysis particularlywhen using high-throughput sequencing (HTS) technology (eg 16S rRNA metagenomicsequencing) which can generate huge amounts of data to outline the composition andstructure of bacterial communities (Szeacutekely amp Langenheder 2014)

Although the environmental bacterial composition may be sensitive to environmentalvariation the effects of these bacterial composition changes on ecofunction remainunclear A comprehensive review demonstrated that the soil microbial composition variesto alter acquisition metabolism and degradation processes in response to changes insoil phosphorus (P) due to grassland afforestation (Chen Condron amp Xu 2008) Moreimportantly environmental conditions and existing microbial diversity determine theecological function and nutrient transformation efficiency of soil microbiota (Zechmeister-Boltenstern et al 2015) Therefore quantifying the effects of afforestation on changes in soil

Wu et al (2019) PeerJ DOI 107717peerj6147 225

properties bacterial composition and ecofunctions would accelerate the understanding ofplant-soil-microbe interactions

In this study to quantify the effect of afforestation on soil properties and soil bacterialcomposition and functions soil chemical and bacterial compositions were measured in anafforested region and adjacent grassland Based on previous studies that have demonstratedfunctional assembly of bacterial communities (Burke et al 2011) and a small response ofbacterial communities to afforestation with low sensitivity to soil properties (Gunina etal 2017 Jangid et al 2011 Klein et al 1995) we hypothesized the following (1) The soilprokaryotic microbiome is indirectly affected by afforestation mediated by soil chemicalproperties If so soil chemical properties should differ significantly between afforestedand grassland sites and should be correlated with the composition of the soil prokaryoticmicrobiome (2) Soil bacterial ecological functions are more sensitive to changes invegetation type than changes in soil bacterial composition If so significant differentiationof the predicted ecologically functional potential should be observed between afforested andgrassland sites To test these hypotheses the 16S rRNA metagenomes of the afforested andadjacent grassland soil bacterial communities were sequenced and the relative abundancesof bacteria and predicted ecophysiological functions were quantified by multivariate andregression analyses Since root exudates did not significantly affect the bacterial assemblagesin a previous study (Wu et al 2018) we excluded the effect of root exudates in this studyBased on these analyses we provide a possible explanation of the link between soil chemicalproperties and bacterial community change in grassland afforestation

MATERIALS AND METHODSStudy sites and samplingThe study site was located 6 km west of the town of Honghuarsquoerji in an artificial forestproduced by seedling afforestation in a large area of thin grassland (savanna N48257443

E119996448) The annual precipitation is approximately 390 mm and the mean annualtemperature is roughly minus24 C The elevation is approximately 740ndash1100 m Theforest coverage is as high as 698 in the forest regions of the nature reserve whilethe Honghuarsquoerji grassland is a bare grassland ecosystem that is mainly composed of weedyspecies such as Stipa baicalensis Festuca ovina and Carex pediformis (Wen et al 2002) Thefield experiments were approved by the Honghuarsquoerji Nature Reserve (permit number20066150221) The tree ages of the Mongolian pines in this afforestation forest range from27 to 33 years with a diameter at breast height (DBH) of 2121 plusmn 419 cm (1209sim3210cm) and tree height of 1286 plusmn 086 (112sim146 m)

At the study site we collected soil pit samples from 20 quadrants including 10 inside theforest (ie forest soils) and 10 in the adjacent grassland (ie grassland soils) Each surfacearea of quadrant was roughly 100 m2 At each location 1-kg soil samples were collectedat a depth of 50 cm in permeable bags of nylon mesh from 10 quadrants located 10 to100 m from each other (Fig S1) This depth was chosen because the major microbiomecomposition and the major transition of soil physicochemical characteristics are locatedat a soil depth of approximately 50 cm (Fierer Schimel amp Holden 2003) All equipment


was sterilized by an autoclave The soil samples were separated into two parts one partwas dried for quantification of soil physicochemical properties and the other was usedto quantify the microbiome The latter samples were stored in RNAguardian stabilizationsolution (MBGEN Biosciences Taipei Taiwan) on ice immediately after collection untiltransfer to the laboratory where the samples were stored at minus20 C before metagenomicDNA extraction

Quantifying soil chemical propertiesOrganic carbon (C) was measured by the external-heat potassium dichromate oxidationmethod and total nitrogen content (N) was measured by the Kjeldahl distillation methodThe inductively coupled plasma (ICP) method was used with soils digested in a mixture ofHFndashHClO4ndashHNO3 to quantify the contents of the soil elements K+ P Ca2+ Fe 3+ Mg2+and Na+ Soil pH was determined using a Sartorius pH meter PB-10 (Germany) with asoilwater ratio of 125 All soil properties were determined and quantified by the StateKey Laboratory of Vegetation and Environmental Change Institute of Botany ChineseAcademy Sciences

16S rRNA metagenome sequencingBacterial metagenomic DNA was extracted with an EZNA Rcopy Soil DNA Kit (QiagenValencia CA USA) and the concentration was adjusted to 50 ng mlminus1 via dilution Themetagenomic DNA was quantified using Qubit Rcopy 20 (Invitrogen Life TechnologiesCA USA) The primers 341F (5prime-CCTACGGGNGGCWGCAG-3prime) and 805R (5prime-GACTACHVGGGTATCTAATCC-3prime) were used to amplify the V3-V4 hypervariable 16SrRNA region (Mizrahi-Man Davenport amp Gilad 2013) and the PCR products were usedto construct a DNA library with the Roche GS FLX Titanium emPCR kit (Roche AppliedScience) The DNA libraries were then sequenced by Sangon Biotech Co (Shanghai China)on an Illumina MiSeq 2X300 The sequencing procedures followed the manufacturerrsquosinstructions

Before analysis the raw HTS data were cleaned by removing sequence fragmentsshorter than 200 bp or with missing barcodes or polyN or polyAT in the RibosomalDatabase Project (RDP) (Cole et al 2007) We also discarded reads with PHRED qualityscores ltQ25 (Ewing amp Green 1998 Ewing et al 1998) The Mothur package was usedto remove non-prokaryotic sequences and de-noise and trim the sequences (Mothurver 1301 httpwwwmothurorgwikiMiSeq_SOP (Schloss et al 2009) Chimericsequences were removed using Uchime (Edgar et al 2011) After quality filteringwe clustered sequences using a criterion of gt97 sequence similarity as operationaltaxonomic units (OTUs) defined as representing the same species by the UPARSEpipeline (httpdrive5comuparse) (Edgar 2013) and closed-reference OTU pickingWe estimated the relative abundances (RAs) of the soil microbiome according to 16SrRNA metagenome sequencing The rarefied OTU table was generated using Qiime(Caporaso et al 2010) and deposited in Mendeley (doi 1017632gjskh8wswz1) All stepswere conducted by Sangon Biotech Co (Shanghai China) according to their pipeline(httpswwwsangoncomservices_ngs_metseqhtml) Each OTU was annotated and


classified according to the RDP classifier and SILVA database The raw sequence datawere deposited in NCBI GenBank under Bioproject PRJNA317430 (accession numberSAMN04607375)

Predictive functional profiling of bacterial communitiesPICRUSt v 100 a functional prediction tool for estimating shared gene content accordingto the corresponding bacterial phylogeny was used to predict the functional potential ofeach sample (Langille et al 2013) PICRUSt generates the composition of gene families foreach metagenome using an extended ancestral-state reconstruction algorithm The onlineversion of PICRUSt implemented in Galaxy (httphuttenhowersphharvardedugalaxy)was used to assist the algorithms The quality-filtered sequences were assigned to anOTU table against the Greengenes v 135 OTU database (DeSantis et al 2006) forPICRUSt prediction implemented in QIIME v 180 (Caporaso et al 2010) Each OTUwas normalized by its copy number The functional contribution of each OTU memberwas reconstructed and predicted by mapping the 16S sequences to their nearest referencegenome lsquoVirtualrsquometagenomeswith gene content abundancewere then generated using theKyoto Encyclopedia of Genes and Genomes (KEGG)Ortholog and Clusters of OrthologousGroups (COGs) databases

Statistical analysesAll statistical analyses were conducted in R (R Core Team 2013) This study includedtwo types of measurements the soil chemical properties and soil bacterial communities(ie relative abundances of bacterial OTUs) and one treatment grassland afforestation(ie vegetation type grassland vs forest) Due to inequality of variance (certain factorssuch as Na+ and Ca2+ deviated from equal variance in Levenersquos test) we first assessed thedifferences in soil properties between vegetation types by the nonparametric MannndashWhitney U (MW) test using R In addition to determine whether the diversity ofprokaryotic microbiome differs after afforestation we calculated diversity indexessuch as the Shannon-Wiener H reciprocal Simpsonrsquos index 1D species richness andPieloursquos evenness for each of the soil bacterial communities and compared them betweenafforested and grassland sites by the MannndashWhitneyU test We also adopted the R packageDESeq2 (Love Huber amp Anders 2014) to identify OTUs that were differentially abundantbetween the afforested and grassland sites A logistic regression model was further usedto determine if each of the soil properties predicted the grassland-afforestation treatmentThe likelihood ratio test (LRT) was performed to identify the best-fitting model of thelogistic regression To test the hypotheses of a stochastic process of bacterial assembly(random distribution model) or resource-governed assembly (niche-based mechanism)we usedRank-AbundanceDominance (RAD) analysis to display logarithmic relative speciesabundances against species rank order (McGill et al 2007) BrayndashCurtis distance-basedredundancy analysis (dbRDA) implemented in the R package vegan (Dixon amp Palmer2003) was used to examine the explanatory proportions of the treatment (vegetation type)for these two measurements (soil properties and microbiomes) Because soil chemicalproperties can also affect the soil bacterial communities (Stutter amp Richards 2012) we


further performed a partial dbRDA to assess the effects of grassland afforestation on thesoil bacterial communities conditioned by soil chemical properties and the effects of the soilchemical properties on the soil bacterial communities conditioned by vegetation type usingthe Bray-Curtis distance The best model were chosen via forward selection by the ordistepfunction implemented in the R package vegan (Dixon amp Palmer 2003) Type II ANOVAwas used to evaluate the fit of the model of each constraint factor We further identifieddivergent bacteria (ie forest and grassland specialists) using the supermajority rule (gt23difference in abundance) with the assistance of the multinomial species classificationmethod (CLAM) test (Chazdon et al 2011) The soil elements that significantly explainedthe soil bacterial communities by partial dbRDA were then used as independent factorsto predict the abundance of bacterial specialists by the generalized linear model (GLM)We also performed hierarchical cluster analysis (H-cluster) and discriminant analysis ofprincipal components (DAPC) to illustrate the grouping patterns between afforested andgrassland sites using the R package adegenet (Jombart 2008) Because the ecophysiologicalfunctions (COGs and KEGG modules) of the soil microbiomes were also predicted alltests were repeated by replacing the soil bacterial OTUs with the relative abundances of theCOGs and KEGG modules All R codes and input files have been deposited as Mendeleydata (httpsdatamendeleycomdatasetsgjskh8wswz1)

RESULTSEffects of afforestation on soil chemical propertiesTo characterize the soil chemical properties the contents of eight elements (organic Ctotal N P K+ Ca2+ Mg2+ Fe3+ and Na+) and the pH value of each soil sample weredetermined (Table 1) Among these soil variables the contents of C P Ca2+ Mg2+ andFe3+ were significantly higher in the forest soils than in the grassland soils (P lt 001)and the pH of the grassland soils was significantly higher than the pH of the forest soils(P = 0031) (Fig 1) To estimate the explanatory proportions of the afforestation effect onthe variance of soil properties we performed dbRDA using vegetation type (ie forest andgrassland) as the independent factor Vegetation type significantly explained 5018 of thevariance of soil properties (P = 0002 Table 2 and Fig S2) indicating that afforestationhas changed the soil properties

If the soil elements are altered after afforestation the contents of soil elements willpredict the vegetation type (ie afforestation effect) To test this hypothesis we comparedthe simple logistic regression (SLR) model using a single soil element as the independentfactor and the multivariate logistic regression (MLR) model using all soil elements asindependent factors (ie the full model M1) to the empty (null) model M0 Most of theSLRmodels andM1 rejected the null model M0 (P lt 0005) except the SLRmodels with K(LRT P = 0084) or Na (LRT P = 0108) as the independent factor We further comparedthe SLR models with each single soil element (C N P Ca2+ Mg2+ Fe3+ and pH) as theindependent factor with the MLR model using all C N P Ca2+ Mg2+ Fe3+ and pH asindependent factors (full model M2) by LRT to determine which individual factors mightbe related to afforestation The SLR models with Mg or Fe as the independent factor could


Table 1 Contents of soil elements (g kgminus1) and pH of the forest and grassland soils reported in format of averageplusmn standard deviation

Group C N P K+ Ca2+ Mg2+ Fe3+ Na+ pH

Forest 12358plusmn 1950 0835plusmn 0116 1469plusmn 0199 26303plusmn 0341 1424plusmn 0222 0456plusmn 0099 8355plusmn 1047 11665plusmn 0531 6216plusmn 0095Grassland 7339plusmn 3025 0627plusmn 0292 0990plusmn 0463 26269plusmn 0563 1183plusmn 0031 0313plusmn 0035 5607plusmn 1308 11960plusmn 0243 6318plusmn 0029

NotesSignificant difference between the forest and grassland soils

Wuetal(2019)PeerJD

OI107717peerj6147

725

Figure 1 MannndashWhitneyU test revealing significant differences in soil properties ie contents of or-ganic C P Ca2+ Mg2+ and Fe2+ and pH between the forest (Fr) and grassland (Gr) soils

Full-size DOI 107717peerj6147fig-1

not be rejected by M2 (P = 0278 and 0130 respectively Fig 2) Although the other SLRmodels with C N P Ca2+ or pH as the independent factor were rejected by M2 noneof the independent factors in M2 could significantly predict the presence of forestation(Z lt 10minus6 P gt 09999 in each term) These results indicated that the contents of Mg2+

(0456 plusmn 0099 and 0313 plusmn 0035 g kgminus1 in forest and grassland soils respectively) andFe3+ (8355 plusmn1047 and 5607 plusmn 1308 g kgminus1 in forest and grassland soils respectively)could singly reflect the changes in soil properties due to afforestation (Table 1 and Fig 2)

Afforestation effects on the relative abundances of soil bacteriaOverall a total of 694993 sequences were obtained after rarefaction analysis (314923for grassland and 380070 for forest) Divergence of the soil microbiome was inferredby significant or marginal differences in the diversity indices (ShannonndashWiener indexH W = 26 P = 00753 Reciprocal Simpsonrsquos index 1D W = 23 P = 00432 Pieloursquosevenness J W = 85 P = 00019 Table 3) However no difference was estimated in thespecies richness of the soil microbiome between forest and grassland (W = 50 P = 1Table 3) Hence we used RA as an indicator to compare the microbiomes of the soilsamples (Fig S3) H-cluster and DAPC (five first PCs of PCA used which conserved70 of the variance of the bacterial RA) presented similar patterns of clear divergence ofsoil microbiomes between the forest and grassland (Fig 3) A total of 350 OTUs with anadjusted P-value lt001 exhibited twofold changes between forest and grassland vegetation(178 OTUs were higher in grassland while 172 were higher in forest) The differentiallyabundant OTUs were only from the most abundant phyla (Proteobacteria Acidobacteriaand Verrucomicrobia) and Planctomycetes These results confirm that the soil bacterialcomposition has changed due to grassland afforestation despite no change in bacterialrichness

Because the close-reference OTUpicking strategy discards samples that do not bin with97 similarity assignments of different OTUs were rarely found among different samplesWe performed the CLAM test (Chazdon et al 2011) to classify these OTUs as four types


Table 2 Summary of the (partial) dbRDA results for the afforestation effect (ie vegetation type) onsoil properties and on the relative abundance (RA) of soil bacteria The effect of soil properties on thebacterial RA was also tested Type II ANOVA was used to test the model fitting for each independent vari-able Significant P values (lt005) are in bold

dbRDA ANOVA

SS Proportion F P

Soil propertysimVegetation typeConstrained 0027 0+ 1813 0002Unconstrained 0027 0498

Microbial RAsimVegetation typeConstrained 0849 0163 3504 0001Unconstrained 4362 0837

Microbial RAsimVegetation type + Condition (Soil property)Conditional 3356 0644Constrained 0194 0037 1053 0466Unconstrained 1661 0319

Microbial RAsim Soil propertyConstrained 3356 0644C 0595 0114 3210 0001P 0524 0100 2823 0001Ca2+ 0467 0090 2516 0001Mg2+ 0368 0071 1986 0003Fe3+ 0546 0105 2940 0001pH 0228 0044 1232 0145N 0249 0048 1343 0070K+ 0188 0036 1012 0438Na+ 0190 0037 1026 0365Unconstrained 1855 0356

Microbial RAsim Soil property + Condition (Vegetation type)Conditional 0849 0163Constrained 2701 0518C 0350 0067 1899 0005P 0493 0095 2673 0001Ca2+ 0457 0088 2477 0001Mg2+ 0350 0067 1896 0004Fe3+ 0283 0054 1534 0009pH 0203 0039 1102 0238N 0183 0035 0994 0490K+ 0190 0037 1031 0362Na+ 0190 0036 1029 0378Unconstrained 1661 0319

of bacteria using the supermajority (23) rule too-rare bacteria generalist bacteria andgrassland- and forest-specialist bacteria In this classification a large proportion of OTUs(9546) belonged to the too-rare bacteria which accounted for 3842 and 3673 ofthe abundance of grassland and forest soil bacteria respectively only 298 of the OTUs


Figure 2 A simple logistic regressionmodel revealing significant prediction of the absence (0) andpresence (1) of forestation based on the contents of the soil elements (AndashH) and pH (I) (A) Organic C(B) N (C) P (D) K (E) Ca (F) Fe (G) Mg (H) Na) (I) pH


Table 3 Diversity indices comparisons of soil bacterial communities

Forest(meanplusmn SD)

Grassland(meanplusmn SD)

KW χ2 P

ShannonndashWiener H 8429plusmn 0244 8612plusmn 0174 3291 0070Reciprocal Simpsonrsquos index 1D 877783plusmn 281326 1190206plusmn 364143 4166 0041Species richness (S) 146850plusmn 484682 136646plusmn 210768 0 1Pieloursquos evenness (J ) 0883plusmn 0017 0905plusmn 0008 9864 0002

were generalists but they accounted for 3555 and 3588 of the abundance of grasslandand forest soil bacteria respectively Only 156 of the OTUs belonged to specialists ofwhich 1061 (07) and 1224 (08) bacterial OTUs were identified as forest and grasslandspecialists respectively The RA of grassland-specialist bacteria was 2284 in grasslandsoils and 284 in forest soils while the RA of forest-specialist bacteria was 319 ingrassland soils and 2455 in forest soils


Figure 3 Divergence of the microbiome between the forest and grassland soils as revealed by (A) hier-archical cluster analysis and (B) discriminant analysis of principal components (DAPC)


We further used the RAs of the bacterial OTUs as the dependent response to access theimpact of afforestation on the soil bacterial communities When vegetation type was usedas the categorical independent factor 163 of the variance of bacterial RA was explainedsignificantly (P = 0001) However the significant explanatory effect of vegetation typewas lost after removing (conditioning on) soil properties (P = 0466 Table 2) When soilproperties were used as a constraint factor 644 of the variance of bacterial RA wasexplained by soil properties in which organic C (114 explanation) P (100) Ca2+

(90) Mg2+ (71) and Fe3+ (105) significantly fit the model by type II ANOVA(Table 2) Similar significant explanations by C (67) P (95) Ca2+ (88) Mg2+

(67) and Fe3+ (54) were obtained when the effect of vegetation type was removed(ie partial dbRDA Table 2) In addition when conducting forward model selection thebacterial RA was best explained by vegetation type Ca2+ Mg2+ Na+ and pH (constrainedproportion 4807) However although pH is important in shaping the soil microbiomeat continental scales (Lauber et al 2009) the pH did not vary among vegetation types(forest 63ndash636 grassland 613ndash635) in our fine-scale study There was no collinearity ofpH with other factors [variance inflation factor (VIF) 1125] Therefore we do not discussthe importance of pH in our study These analyses suggested that afforestation has changedthe soil chemical properties with changes in at least Ca2+ and Mg2+ in all analyses thusindirectly affecting the soil bacterial communities (Figs S2 and S3)

Effects of soil properties on the divergence of bacterial phyla betweenforest and grassland soilsTo identify the divergent soil bacteria the habitat-specialist bacterial phyla were classifiedunder the supermajority rule (ie 23 majority) At the phylum level five phyla werehabitat specialists including one forest specialist (Thaumarchaeota phylum RA 23)and four grassland specialists (Chloroflexi Fibrobacteres Nitrospirae and Parcubacteria


Table 4 Effect of soil properties on the abundance of divergent soil microbial phyla inferred by the generalized linear model (GLM)

Thaumarchaeota Chloroflexi Fibrobacteres Nitrospirae Parcubacteria

t value P t value P t value P t value P t value P

Intercept 0068 0947 minus3280 0005 minus1087 0295 minus1176 0259 minus1590 0134C minus0288 0778 3501 0004 0632 0538 1471 0164 1714 0109P minus0420 0681 4848 00003 3735 0002 4108 0001 3326 0005Ca2+ minus0083 0935 3857 0002 1267 0226 1667 0118 2007 0065Mg2+ 0031 0976 minus3727 0002 minus1174 0260 minus1374 0191 minus1934 0074Fe3+ 2088 0056 minus0332 0745 minus1263 0227 minus2900 0012 minus0575 0574

NotesSignificant P values (lt005) are in bold

93) The forest-specialist phylum Thaumarchaeota belongs to Archaea while the fourgrassland specialists are Eubacteria

We further tested the effects of soil properties on the abundance of these habitatspecialists by GLM Because organic C P Ca2+ Mg2+ and Fe3+ significantly explainedthe RA of soil bacteria in dbRDA these five soil variables were used as independentpredictors in GLM The variances of the bacterial abundance of the sampled soils weresignificantly or marginally significantly greater than the means suggesting overdispersionof these responses (P = 0056 0001 0051 0002 and 0002 alpha = 52675 258711065 23101 and 76612 in Thaumarchaeota Chloroflexi Fibrobacteres Nitrospirae andParcubacteria respectively) Hence we used the quasi-Poisson model for GLM which ledto the same coefficient estimates as the standard Poisson model but with adjustment of thedispersion parameter for overdispersion Under the quasi-Poisson model Fe3+ contentwas marginally correlated with the abundance of the forest specialist Thaumarchaeota forthe grassland specialists the P content was correlated with the abundances of all four phyla(Table 4) In addition C Ca2+ and Mg2+ were significantly correlated with ChloroflexiFe3+ was significantly correlated with Nitrospirae and Ca2+ and Mg2+ were marginallycorrelated with Parcubacteria (Table 4)

Afforestation effects on the soil ecophysiological functions predictedby soil microbiomeTo understand the ecophysiological functions of the soil bacterial communities wepredicted their functional composition using 16S rRNA gene and databases of referencegenomes A total of 4659 clusters of orthologous groups (COGs) and 306 KEGG modules(Level-3 KEGG orthology) were identified Similar to the analyses for testing the effectsof afforestation and soil properties on the soil bacterial communities we used vegetationtype and soil elements as predictors to test the explanatory proportion and significance ofeach predictor on the RAs of the COGs and KEGG modules dbRDA indicated that 220and 212 of the variation of COGs and KEGG modules was significantly explained byvegetation type respectively whereas the explanatory proportion decreased to 34 and45 when conditioning the soil-property effect (Table 5) Soil properties explained 830and 824 of the variation of the COGs and KEGG modules in dbRDA respectively inwhich C P Ca2+ Mg2+ Fe3+ and N significantly or marginally fit the model according


Table 5 Summary of the (partial) dbRDA results for the afforestation effect (ie vegetation type) and soil-property effect on the ecologicalfunction (COGs and KEGGmodules) estimated from the soil bacterial communities Type II ANOVA was used to test the model fitting for eachindependent variable

dbRDA ANOVA dbRDA ANOVA

SS Proportion F P SS Proportion F P

COGssimVegetation type KEGGsimVegetation typeConstrained 0005 0220 5081 0001 0001 0212 4834 0001Unconstrained 0016 0780 0002 0788

COGssimVegetation type + Condition (Soil property) KEGGsimVegetation type + Condition (Soil property)Conditional 0017 0830 0003 0824Constrained 0001 0034 2267 0078 1Endash04 0045 3140 0038Unconstrained 0003 0136 4Endash04 0130

COGssim Soil property KEGGsim Soil propertyConstrained 0017 0830 0003 0824C 0002 0094 5541 0001 3Endash04 0097 55051 0001P 0003 0140 8234 0001 0001 0165 93958 0001Ca 0003 0140 8247 0001 3Endash04 0107 60938 0001Mg 0002 0072 4252 0005 1Endash04 0042 24047 0063Fe 0002 0073 4273 0010 2Endash04 0066 3738 0013pH 5Endash04 0023 1347 0283 9Endash05 0028 15886 0177N 0005 0258 15157 0001 0001 0292 16615 0001K 2Endash04 0010 0575 0653 4Endash05 0012 06817 0580Na 4Endash04 0020 1174 0312 5Endash05 0016 0914 0453Unconstrained 0004 0170 0001 0176

COGssim Soil property + Condition (Vegetation type) KEGGsim Soil property + Condition(Vegetation type)

Conditional 0005 0220 0001 0212Constrained 0013 0644 0002 0658C 0001 0027 1776 0152 7Endash05 0021 1478 0257P 0004 0172 11371 0001 5Endash04 0157 10877 0001Ca 0001 0071 4723 0006 2Endash04 0061 4226 0018Mg 0001 0030 1963 0133 8Endash05 0027 1869 0175Fe 0005 0251 16630 0001 0001 0264 18238 0001pH 0001 0032 2094 0110 1Endash04 0038 2650 0066N 4Endash04 0018 1162 0302 9Endash05 0028 1946 0130K 3Endash04 0014 0928 0406 7Endash05 0022 1510 0223Na 0001 0030 2017 0141 1Endash04 0039 2691 0070Unconstrained 0003 0136 4Endash04 0130


to type II ANOVA (Table 5) When conditioned on vegetation type the explanatoryproportion decreased slightly to 644 and 658 for the COGs and KEGG modulesrespectively and the remaining significant fitting factors were P Ca2+ and Fe3+ (Table 5)

Among all COGs 44 and 21 were identified as forest- and grassland-predominant COGcategories by CLAM analysis respectively Most of the forest-predominant categories


are involved in post-translational regulation such as ubiquitin or E3 ligase while manycategories in grassland are membrane proteins or are involved in cell transport Bycontrast among all KEGG modules only two forest-predominant categories and nograssland-predominant categories were identified Further testing of the correlation ofsoil elements with these specialist COGs and KEGG modules revealed that 50 of the 65COGs and both KEGG modules were significantly correlated with at least one soil elementunder Poisson or quasi-Poisson regression These significant correlations indicate that soilproperties particularly P Ca2+ and Fe3+ account for the changes in ecophysiologicalfunctional due to grassland afforestation

DISCUSSIONThe closure of the tree canopy and increased litter accumulation that accompany theecosystem change from grassland to forest may directly alter the soil environment (Bond ampMidgley 2012 Cunningham et al 2015) The significantly higher contents of soil chemicalfactors (C P Ca 2+ Mg2+ Fe3+ and pH) in the Mongolian pine plantation areas thanin the unplanted region suggest a great influence of grassland afforestation on secondarysalinization Soil mineral elements are usually increased in tree-plantation regions wheregroundwater is insufficient tomeet water requirements (Nosetto et al 2008) High contentsof soil elements in a forest suggest not only a larger amount of litter biomass but also arapid decomposition rate of pine litters compared to other broadleaf flora (Berger et al2015) However despite significant differences in the contents of soil elements betweenforest and grassland these soil elements except Mg2+ and Fe3+ (Fig 2) could not singlypredict the ecosystem change due to afforestation by logistic regression analysis

Significantly high predictable contents of Fe3+ and Mg2+ in forest soils reflect thecharacteristics of litter and humus accumulation in forests (Song et al 2008) Litterdecomposition accelerates the conversion and accumulation of soil non-organic elements(Fenchel King amp Blackburn 2012b) In 2nd-to-3rd-year needle litters of P sylvestris L adecrease in the rate of biomass loss but an increase in the release of Fe3+ and Mg2+ wererecorded (De Marco et al 2007) Consequently we suggest that the high contents of Fe3+

and Mg2+ in the forest soils of our study sites are due to the long steady accumulationand decomposition of needle litter The soil element cycling affects and is affected bythe composition of the soil microbiota (Fenchel King amp Blackburn 2012b) For examplethe iron bacteria family Comamonadaceae which was represented by the genera DelftiaComamonas Acidovorax and Albidiferax in our sampling are able to deposit iron metaloxides under natural conditions These bacteria were present in soil for both vegetationtypes They were slightly more abundant in forest samples and can grow rapidly in iron-richand acidic substrates (Emerson et al 2015 Fenchel King amp Blackburn 2012a) These resultsindicate that alteration of these chemical propertiesmay lead to a change in the compositionof the soil prokaryotic microbiome

The distributions of the soil bacterial abundances in our samples best fit to Zipf andZipfndashMandelbrot rank abundance models (Table 6 and Fig S4 ) The Zipf and ZipfndashMandelbrot rank abundance models belong to the family of random-branching processes


Table 6 Deviance of the species-rank abundance distribution (RAD)models revealing the best fits of the Mandelbrot or Zipf-Mandelbrotmodel for all sampled soil microbial communities

Gr01 Gr02 Gr03 Gr04 Gr05 Gr06 Gr07 Gr08 Gr09 Gr10

Null 234732 244126 208828 216098 296537 239729 161750 212919 251192 314614Preemption 289211 300189 254941 267792 366758 296679 192092 260017 308094 387683Lognormal 95275 115397 97015 101578 129427 105858 78027 98473 106506 132656Zipf 16527 12015 11681 10102 17220 14210 10165 11692 22289 22155Mandelbrot 10571 12015 10108 10102 15252 11656 9193 9608 10502 13389

Fr01 Fr02 Fr03 Fr04 Fr05 Fr06 Fr07 Fr08 Fr09 Fr10


NotesThe lowest Akaike Information Criterion (AIC) values representing the best fit model are shown in bold

these models suggest that individuals are always derived from ancestor individuals (McGillet al 2007) and that microbial community assembly is explained by the niche-basedmechanism (McGill et al 2007 Mendes et al 2014) These models indicate that decadesof grassland afforestation have generated soil properties that provide a divergent but stableresource supply for the soil bacterial community although a strong effect of depth on thestructure of bacterial communities should be recognized (Eilers et al 2012)

Classification by the supermajority rule revealed that generalist bacteria representedmore than 23 of the total counts but lt3 of bacterial OTUs The low richness butrelatively high abundance of generalists suggests that a great proportion of residents utilizebroad resources or are highly tolerant of the environment (Verberk 2011) It has beensuggested that microbes that are present in all or the majority of microbial communitieswith high abundance represent the core set of genes responsible for key elements ofmost metabolic pathways (Falkowski Fenchel amp Delong 2008) Similarly specialist bacteriaexhibited lt1 richness but accounted for approximately 15 to 14 of the RA in thegrassland and forest soils These specialist OTUs with low richness and high abundanceare probably more susceptible than generalists to environmental change Since the originalvegetation was scattered grasses the grassland specialists rarely found in forest soils werethose selected against by the afforestation effect by contrast forest specialists should beenriched after forestation The environmental differences (eg the contents of soil C PCa2+ Mg2+ and Fe3+ Table 2) could result in resource (niche) divergence to differentiatethe bacterial composition descended from the original bacterial communities reflectingthe bacterial abundance distribution in the Zipf and ZipfndashMandelbrot models (Table 6)

In particular the Fe3+ content was significantly correlated with the abundanceof the forest-specialist Archaea phylum Thaumarchaeota and P was correlated withthe four grassland-specialist Eubacteria phyla Chloroflexi Fibrobacteres Nitrospiraeand Parcubacteria (Table 4) suggesting that these two soil elements are key factors


differentiating soil bacterial composition Thaumarchaeota encodes the genes ammoniamonooxygenase A (amoA encoding subunit A of AMO) and amoB which are distantlyrelated to one another and similar to the bacterial amo gene for ammonia oxidation(Stieglmeier Alves amp Schleper 2014) The high abundance of amoA and its transcriptssuggests that ammonia-oxidizing archaea (AOA) are present in higher numbers thanammonia-oxidizing bacteria (Shen et al 2008) and that nitrogen cycling is enhanced in theforest (Konneke et al 2005 Stieglmeier Alves amp Schleper 2014) A high abundance of AOAwith a high content of Fe3+ (eg ferrate an ammonia oxidation reagent) could accelerateammonia oxidation (Sharma Bloom amp Joshi 1998)

Soil P is closely related to plant growth (Shen et al 2011) but microbes that transformP to improve uptake by plants are also influenced by both plant species and soil type (ChenCondron amp Xu 2008) P metabolism is often related to the fungal community (Beever ampBurns 1981 Rodŕıguez amp Fraga 1999) In this study the specialists that were correlatedwith soil P content were primarily not responsible for P transformation but were lightand aerobic thermophiles (eg Chloroflexi) symbionts within ruminant animals (egFibrobacteres) or involved in the nitrogen cycle (eg Nitrospirae) Although not directlylinked to P content these bacterial ecological function reflect the differences betweengrassland and forest ecosystems The major changes in the canopy amounts of litterfallplant composition and root-bacteria interactions due to grassland afforestation mayexplain the differences in soil P content as well as the abundances of these bacterial phyla(Chen Condron amp Xu 2008 Li Lee Allen amp Wollum 2004)

By contrast the significant differences in H 1D and J but not richness between thegrassland and forest soil microbiomes suggest that the change in the soil microbiomeis due to a difference in species RA rather than species number A high proportion oflsquolsquotoo-rarersquorsquo bacterial OTUs (95 richness) accounted for gt13 of the RA of soil bacteriareflecting transient changes in bacteria in the environments This rarity could result fromstochasticity (Ai et al 2013) fitness trade-off (Gobet et al 2012 Gudelj et al 2010) orbiological interactions (Garciacutea-Fernaacutendez De Marsac amp Diez 2004 Narisawa et al 2008Schluter et al 2015) These rare bacteria are still relevant in ecological functions includingbacterial community assembly and function and biogeochemical cycling (Jousset et al2017) The high richness of rare bacteria contributes to the alpha-diversity of the soilbacterial communities

Based on the dbRDA of the relationships among vegetation type soil properties andbacterial composition we concluded that grassland afforestation has affected the soilchemical properties (Table 2) soil microbiome composition and ecologically functionalpotential Post-translational systems may increase the heterogeneity of the soil prokaryoticmicrobiome and facilitate the coexistence of different bacterial species via functionalregulation (Spallek Robatzek amp Goumlhre 2009) In addition these effects of afforestation onthe microbiome were modulated by changes in soil chemical properties (Tables 2 and 5)The changes in chemical properties led to differential abundance of low-taxonomic-levelOTUs For example Planctomycetes which was differentially abundant between forest andgrassland in our data are sensitive to changes in soil management and physicochemicalproperties (Buckley et al 2006) This conclusion was reached because the explanation of


soil microbiome by vegetation type decreased or was even lost when conditioned on soilproperties (Table 2) Albeit indirectly afforestation indeed altered the ecological functionof the soil microbiome (Table 5)

As discussed above soil chemical properties interact with the soil bacterial communitiesBecause the original vegetation before forestation was grassland the changes in the soilproperties are probably attributable to the vegetation changes produced by afforestationSeveral studies have suggested that afforestation can influence biotic and abiotic changesin micro- and macro-ecosystems (Jousset et al 2017 Nosetto et al 2008 Wang Zhu ampChen 2016 Zheng et al 2017) Here we suggest that the underground biotic change wasindirectly affected by forestation mediated by soil property changes (Table 2) especiallythe contents of soil P Ca2+ and Fe3+ which are further related to ecological functionchanges in soil microbiomes for different vegetation types (Table 5 and Table S1 ) Theseresults are similar to the bacterial abundance changes and compositional shifts reportedfor a long-term poplar plantation which were suggested to be highly correlated withthe changes in soil properties caused by afforestation (Zheng et al 2017) Soil bacterialcomposition has been suggested to be more closely related to plant diversity-controlledabiotic soil properties because of the highly resilient characteristics of bacterial communitiesdue to their fast life cycle (De Vries et al 2012 Lange et al 2014)

CONCLUSIONSThe change in vegetation type was linked at least in part to potential ecological functionchanges in the soil bacterial communities despite an indirect impact on bacterialcomposition and the calculation of the ecologically functional potential from the bacterialcomposition Aboveground changes such as abiotic factors and biotic activities may beresponsible for these changes in underground ecological functions According to our studyafforestation may change the soil contents and subsequently affect the ecological functionsand alpha- and beta-diversity of bacterial communities However the relatively smallproportion of bacterial specialists and high proportion of bacterial generalists with respectto ecological function among the bacterial OTUs indicates that the vegetation changepreserved a high proportion of the core functions of the soils (Li Lee Allen amp Wollum2004) This preservation occurred because the core functional genes were distributedwidely across a variety of bacterial taxa However the high proportion of functions thatwere correlated with changes in soil properties indicates that bacterial ecological functionsare highly sensitive to environmental change Future work could focus on the link betweenanthropogenic actions on vegetation and indicator species such as Planctomycetes or ironbacteria such as Comamonadaceae These links could be validated by quantitative analyses(eg qPCR) to determine how changes in the structure of key bacteria lead to ecologicalfunction changes after land cover change

ACKNOWLEDGEMENTSWe thank Dr Dawn Schmidt for English editing of the manuscript


ADDITIONAL INFORMATION AND DECLARATIONS

FundingThis research was financially supported by the Fundamental Research Funds for the CentralUniversities (to Shu-HongWu) and supported by the Ministry of Science and TechnologyTaiwan (MOST 105-2628-B-003-001-MY3 and MOST 105-2628-B-003-002-MY3) (toPei-Chun Liao) This article was also subsidized by the National Taiwan Normal University(NTNU) The funders had no role in study design data collection and analysis decision topublish or preparation of the manuscript

Grant DisclosuresThe following grant information was disclosed by the authorsFundamental Research Funds for the Central UniversitiesMinistry of Science and Technology Taiwan MOST 105-2628-B-003-001-MY3 MOST105-2628-B-003-002-MY3

Competing InterestsThe authors declare there are no competing interests

Author Contributionsbull Shu-Hong Wu conceived and designed the experiments contributed reagentsmaterial-sanalysis tools approved the final draftbull Bing-Hong Huang performed the experiments analyzed the data prepared figuresandor tablesbull Jian Gao and Siqi Wang performed the experiments contributed reagentsmaterials-analysis toolsbull Pei-Chun Liao conceived and designed the experiments analyzed the data preparedfigures andor tables authored or reviewed drafts of the paper approved the final draft

Field Study PermissionsThe following information was supplied relating to field study approvals (ie approvingbody and any reference numbers)

Field experiments were approved by the Honghuaerji Nature Reserve (permit number20066150221)

Data AvailabilityThe following information was supplied regarding data availability

The rarefied OTU table was generated using Qiime (Caporaso et al 2010) All of the Rcode rarefied OTU table and input files are available at Mendeley

Huang Bing-Hong (2018) lsquolsquoAfforestation effects onMongolian Pine soil microbiomesrsquorsquoMendeley Data v1 httpdxdoiorg1017632gjskh8wswz1

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj6147supplemental-information


REFERENCESAhrends A Hollingsworth PM Beckschaumlfer P Chen H Zomer RJ Zhang LWangM

Xu J 2017 Chinarsquos fight to halt tree cover loss Proceedings of the Royal Society BBiological Sciences 284Article 20162559 DOI 101098rspb20162559

Ai D Chu C EllwoodMDF Hou RWang G 2013Migration and niche partitioningsimultaneously increase species richness and rarity Ecological Modelling 25833ndash39DOI 101016jecolmodel201303001

Arnaacuteez J Lana-Renault N Lasanta T Ruiz-Flantildeo P Castroviejo J 2015 Effects offarming terraces on hydrological and geomorphological processes A reviewCATENA 128122ndash134 DOI 101016jcatena201501021

Beever RE Burns DJW 1981 Phosphorus uptake storage and utilization by fungiAdvances in Botanical Research 8127ndash219

Berger TW Duboc O Djukic I Tatzber M GerzabekMH Zehetner F 2015 De-composition of beech (Fagus sylvatica) and pine (Pinus nigra) litter along anAlpine elevation gradient decay and nutrient release Geoderma 251ndash25292ndash104DOI 101016jgeoderma201503024

BondWJ Midgley GF 2012 Carbon dioxide and the uneasy interactions of trees andsavannah grasses Philosophical Transactions of the Royal Society B Biological Sciences367601ndash612 DOI 101098rstb20110182

Buckley DH Huangyutitham V Nelson TA Rumberger A Thies JE 2006 Diversity ofPlanctomycetes in soil in relation to soil history and environmental heterogeneityApplied and Environmental Microbiology 724522ndash4531 DOI 101128AEM00149-06

Burke C Steinberg P Rusch D Kjelleberg S Thomas T 2011 Bacterial commu-nity assembly based on functional genes rather than species Proceedings of theNational Academy of Sciences of the United States of America 10814288ndash14293DOI 101073pnas1101591108

Caporaso JG Kuczynski J Stombaugh J Bittinger K Bushman FD Costello EK FiererN Pena AG Goodrich JK Gordon JI Huttley GA Kelley ST Knights D Koenig JELey RE Lozupone CA McDonald D Muegge BD PirrungM Reeder J SevinskyJR Turnbaugh PJ WaltersWAWidmann J Yatsunenko T Zaneveld J KnightR 2010 QIIME allows analysis of high-throughput community sequencing dataNature Methods 7335ndash336 DOI 101038nmethf303

Chazdon RL Chao A Colwell RK Lin SY Norden N Letcher SG Clark DB FineganB Arroyo JP 2011 A novel statistical method for classifying habitat generalists andspecialists Ecology 921332ndash1343 DOI 10189010-13451

Chen CR Condron LM Xu ZH 2008 Impacts of grassland afforestation with coniferoustrees on soil phosphorus dynamics and associated microbial processes a reviewForest Ecology and Management 255396ndash409 DOI 101016jforeco200710040

Chudzińska E Diatta JB 2014 Adaptation strategies and referencing trial of Scots andblack pine populations subjected to heavy metal pollution Environmental Scienceand Pollution Research 212165ndash2177 DOI 101007s11356-013-2081-3


Cibils L Principe R Maacuterquez J Gari N Albarintildeo R 2015 Functional diversity of algalcommunities from headwater grassland streams how does it change followingafforestation Aquatic Ecology 49453ndash466 DOI 101007s10452-015-9538-z

Cole JR Chai B Farris RJ Wang Q Kulam-Syed-Mohideen AS McGarrell DMBandela AM Cardenas E Garrity GM Tiedje JM 2007 The ribosomal databaseproject (RDP-II) introducing myRDP space and quality controlled public dataNucleic Acids Research 35D169ndashD172 DOI 101093nargkl889

Cunningham SC Mac Nally R Baker PJ Cavagnaro TR Beringer J Thomson JRThompson RM 2015 Balancing the environmental benefits of reforestationin agricultural regions Perspectives in Plant Ecology Evolution and Systematics17301ndash317 DOI 101016jppees201506001

DeMarco A Vittozzi P Rutigliano FA Virzo De Santo A 2007 Nutrient dynamicsduring decomposition of four different pine litters Options Meacutediterraneacuteennes SeacuterieA Seacuteminaires Meacutediterraneacuteens 7573ndash77

De Vries FT Manning P Tallowin JRB Mortimer SR Pilgrim ES Harrison KAHobbs PJ Quirk H Shipley B Cornelissen JHC Kattge J Bardgett RD 2012Abiotic drivers and plant traits explain landscape-scale patterns in soil microbialcommunities Ecology Letters 151230ndash1239 DOI 101111j1461-0248201201844x

DeSantis TZ Hugenholtz P Larsen N Rojas M Brodie EL Keller K Huber T DaleviD Hu P Andersen GL 2006 Greengenes a chimera-checked 16S rRNA genedatabase and workbench compatible with ARB Applied and Environmental Micro-biology 725069ndash5072 DOI 101128AEM03006-05

Dixon P PalmerMW 2003 VEGAN a package of R functions for community ecologyJournal of Vegetation Science 14927ndash930DOI 1016581100-9233(2003)014[0927vaporf]20co2

Edgar RC 2013 UPARSE highly accurate OTU sequences from microbial ampliconreads Nature Methods 10996ndash998 DOI 101038nmeth2604

Edgar RC Haas BJ Clemente JC Quince C Knight R 2011 UCHIME improvessensitivity and speed of chimera detection Bioinformatics 272194ndash2200DOI 101093bioinformaticsbtr381

Eilers KG Debenport S Anderson S Fierer N 2012 Digging deeper to find uniquemicrobial communities the strong effect of depth on the structure of bacte-rial and archaeal communities in soil Soil Biology and Biochemistry 5058ndash65DOI 101016jsoilbio201203011

Emerson D Scott JJ Benes J BowdenWB 2015Microbial iron oxidation in the arctictundra and its implications for biogeochemical cycling Applied and EnvironmentalMicrobiology 818066ndash8075 DOI 101128AEM02832-15

Ewing B Green P 1998 Base-calling of automated sequencer traces using Phred IIerror probabilities Genome Research 8186ndash194 DOI 101101gr83186

Ewing B Hillier L Wendl MC Green P 1998 Base-calling of automated sequencertraces using PhredI Accuracy assessment Genome Research 8175ndash185DOI 101101gr83175


Falkowski PG Fenchel T Delong EF 2008 The microbial engines that drive earthrsquosbiogeochemical cycles Science 3201034ndash1039 DOI 101126science1153213

Fenchel T King GM Blackburn TH 2012a Chapter 1mdashbacterial metabolism InBacterial biogeochemistry Third Edition Boston Academic Press 1ndash34

Fenchel T King GM Blackburn TH 2012b Chapter 6mdashbiogeochemical cycling in soilsIn Bacterial biogeochemistry Third Edition Boston Academic Press 89ndash120

Fierer N Schimel JP Holden PA 2003 Variations in microbial community compo-sition through two soil depth profiles Soil Biology and Biochemistry 35167ndash176DOI 101016S0038-0717(02)00251-1

Garciacutea-Fernaacutendez JM DeMarsac NT Diez J 2004 Streamlined regulation and gene lossas adaptive mechanisms in Prochlorococcus for optimized nitrogen utilization inoligotrophic environmentsMicrobiology and Molecular Biology Reviews 68630ndash638DOI 101128MMBR684630-6382004

Gobet A Boer SI Huse SM Van Beusekom JEE Quince C SoginML Boetius A Ram-ette A 2012 Diversity and dynamics of rare and of resident bacterial populations incoastal sands ISME Journal 6542ndash553 DOI 101038ismej2011132

Gudelj I Weitz JS Ferenci T Claire Horner-Devine MMarx CJ Meyer JR FordeSE 2010 An integrative approach to understanding microbial diversity fromintracellular mechanisms to community structure Ecology Letters 131073ndash1084DOI 101111j1461-0248201001507x

Gunina A Smith AR Godbold DL Jones DL Kuzyakov Y 2017 Response of soilmicrobial community to afforestation with pure and mixed species Plant and Soil412357ndash368 DOI 101007s11104-016-3073-0

He F Ge Q Dai J Rao Y 2008 Forest change of China in recent 300 years Journal ofGeographical Sciences 1859ndash72 DOI 101007s11442-008-0059-8

Hurme P Repo T Savolainen O Paumlaumlkkoumlnen T 1997 Climatic adaptation of bud set andfrost hardiness in Scots pine (Pinus sylvestris) Canadian Journal of Forest Research27716ndash723 DOI 101139x97-052

Jangid KWilliamsMA Franzluebbers AJ Schmidt TM Coleman DCWhitmanWB 2011 Land-use history has a stronger impact on soil microbial communitycomposition than aboveground vegetation and soil properties Soil Biology andBiochemistry 432184ndash2193 DOI 101016jsoilbio201106022

Jombart T 2008 adegenet a R package for the multivariate analysis of genetic markersBioinformatics 241403ndash1405 DOI 101093bioinformaticsbtn129

Jousset A Bienhold C Chatzinotas A Gallien L Gobet A KurmV Kusel K RilligMC Rivett DW Salles JF Van der HeijdenMGA Youssef NH Zhang XWei ZHolWHG 2017Where less may be more how the rare biosphere pulls ecosystemsstrings ISME Journal 11853ndash862 DOI 101038ismej2016174

Khamzina A Lamers JPA Martius C 2016 Above- and belowground litter stocks anddecay at a multi-species afforestation site on arid saline soil Nutrient Cycling inAgroecosystems 104187ndash199 DOI 101007s10705-016-9766-1


Klein DA McLendon T PaschkeMW Redente EF 1995 Saprophytic fungal-bacterialbiomass variations in successional communities of a semi-arid steppe ecosystemBiology and Fertility of Soils 19253ndash256 DOI 101007BF00336168

KonnekeM Bernhard AE De la Torre JRWalker CBWaterbury JB Stahl DA2005 Isolation of an autotrophic ammonia-oxidizing marine archaeon Nature437543ndash546 DOI 101038nature03911

Krakau U-K LiesebachM Aronen T Lelu-Walter M-A Schneck V 2013 Scots Pine(Pinus sylvestris L) In Pacircques LE ed Forest tree breeding in europe current state-of-the-art and perspectives Dordrecht Springer Netherlands 267ndash323

Lafleur B LabrecqueM Arnold AA Beacutelanger N 2015 Organic carbon accumulationin topsoil following afforestation with willow emphasis on leaf litter decompositionand soil organic matter quality Forests 6769ndash793 DOI 103390f6030769

LangeM Habekost M Eisenhauer N Roscher C Bessler H Engels C Oelmann YScheu SWilckeW Schulze E-D Gleixner G 2014 Biotic and abiotic propertiesmediating plant diversity effects on soil microbial communities in an experimentalgrassland PLOS ONE 9e96182 DOI 101371journalpone0096182

Langille MGI Zaneveld J Caporaso JG McDonald D Knights D Reyes JA ClementeJC Burkepile DE Vega Thurber RL Knight R Beiko RG Huttenhower C 2013Predictive functional profiling of microbial communities using 16S rRNA markergene sequences Nature Biotechnology 31814ndash821 DOI 101038nbt2676

Lauber CL HamadyM Knight R Fierer N 2009 Pyrosequencing-based assessment ofsoil pH as a predictor of soil bacterial community structure at the continental scaleApplied and Environmental Microbiology 755111ndash5120 DOI 101128aem00335-09

Li Q Lee Allen HWollum AG 2004Microbial biomass and bacterial functional diver-sity in forest soils effects of organic matter removal compaction and vegetationcontrol Soil Biology and Biochemistry 36571ndash579 DOI 101016jsoilbio200312001

Li Y ZhaoMMildrexler DJ Motesharrei S Mu Q Kalnay E Zhao F Li S Wang K2016 Potential and actual impacts of deforestation and afforestation on land surfacetemperature Journal of Geophysical Research Atmospheres 12114372ndash314386DOI 1010022016JD024969

LoveMI HuberW Anders S 2014Moderated estimation of fold change anddispersion for RNA-seq data with DESeq2 Genome Biology 15Article 550DOI 101186s13059-014-0550-8

Lu S Zhang Y Chen C Xu Z Guo X 2017 Plantndashsoil interaction affects the mineral-ization of soil organic carbon evidence from 73-year-old plantations with threeconiferous tree species in subtropical Australia Journal of Soils and Sediments17985ndash995 DOI 101007s11368-016-1602-3

Ma E Liu A Li XWu F Zhan J 2013 Impacts of vegetation change on the regionalsurface climate a scenario-based analysis of afforestation in Jiangxi Province ChinaAdvances in Meteorology 2013Article 796163 DOI 1011552013796163

Maacuterquez JA Cibils L Principe RE Albarintildeo RJ 2015 Stream macroinvertebratecommunities change with grassland afforestation in central Argentina LimnologicamdashEcology and Management of Inland Waters 5317ndash25 DOI 101016jlimno201505002


McGill BJ Etienne RS Gray JS Alonso D AndersonMJ Benecha HK Dornelas MEnquist BJ Green JL He F Hurlbert AH Magurran AE Marquet PA Maurer BAOstling A Soykan CU Ugland KI White EP 2007 Species abundance distribu-tions moving beyond single prediction theories to integration within an ecologicalframework Ecology Letters 10995ndash1015 DOI 101111j1461-0248200701094x

Mendes LW Kuramae EE Navarrete AA Van Veen JA Tsai SM 2014 Taxonomicaland functional microbial community selection in soybean rhizosphere ISME Journal81577ndash1587 DOI 101038ismej201417

Mizrahi-Man O Davenport ER Gilad Y 2013 Taxonomic classification of bacterial 16SrRNA genes using short sequencing reads evaluation of effective study designs PLOSONE 8e53608 DOI 101371journalpone0053608

Narisawa N Haruta S Arai H Ishii M Igarashi Y 2008 Coexistence of antibiotic-producing and antibiotic-sensitive bacteria in biofilms is mediated by re-sistant bacteria Applied and Environmental Microbiology 743887ndash3894DOI 101128AEM02497-07

Nosetto MD Jobbaacutegy EG Toacuteth T Jackson RB 2008 Regional patterns and controls ofecosystem salinization with grassland afforestation along a rainfall gradient GlobalBiogeochemical Cycles 22Article GB2015 DOI 1010292007GB003000

Oleksyn J Reich PB Zytkowiak R Karolewski P Tjoelker MG 2003 Nutrient con-servation increases with latitude of origin in European Pinus sylvestris populationsOecologia 136220ndash235 DOI 101007s00442-003-1265-9

Pedley SMMartin RD Oxbrough A Irwin S Kelly TC OrsquoHalloran J 2014 Com-mercial spruce plantations support a limited canopy fauna evidence from a multitaxa comparison of native and plantation forests Forest Ecology and Management314172ndash182 DOI 101016jforeco201312010

Peng S-S Piao S Zeng Z Ciais P Zhou L Li LZX Myneni RB Yin Y Zeng H 2014Afforestation in China cools local land surface temperature Proceedings of theNational Academy of Sciences of the United States of America 1112915ndash2919DOI 101073pnas1315126111

R Core Team 2013 R a language and environment for statistical computing Vienna RFoundation for Statistical Computing Available at httpswwwR-projectorg

Ren C Sun P Kang D Zhao F Feng Y Ren G Han X Yang G 2016 Responsiveness ofsoil nitrogen fractions and bacterial communities to afforestation in the Loess HillyRegion (LHR) of China Scientific Reports 628469 DOI 101038srep28469

Rodŕıguez H Fraga R 1999 Phosphate solubilizing bacteria and their role in plantgrowth promotion Biotechnology Advances 17319ndash339DOI 101016S0734-9750(99)00014-2

Savolainen O Bokma F Garćıa-Gil R Komulainen P Repo T 2004 Genetic variationin cessation of growth and frost hardiness and consequences for adaptation ofPinus sylvestris to climatic changes Forest Ecology and Management 19779ndash89DOI 101016jforeco200405006

Schloss PDWestcott SL Ryabin T Hall JR HartmannM Hollister EB LesniewskiRA Oakley BB Parks DH Robinson CJ Sahl JW Stres B Thallinger GG


VanHorn DJ Weber CF 2009 Introducing mothur open-source platform-independent community-supported software for describing and comparingmicrobial communities Applied and Environmental Microbiology 757537ndash7541DOI 101128AEM01541-09

Schluter J Nadell CD Bassler BL Foster KR 2015 Adhesion as a weapon in microbialcompetition ISME Journal 9139ndash149 DOI 101038ismej2014174

Sharma VK Bloom JT Joshi VN 1998 Oxidation of ammonia by ferrate(vi) Journal ofEnvironmental Science and Health Part A 33635ndash650DOI 10108010934529809376752

Shen J Yuan L Zhang J Li H Bai Z Chen X ZhangW Zhang F 2011 Phosphorus dy-namics from soil to plant Plant Physiology 156997ndash1005 DOI 101104pp111175232

Shen J-P Zhang L-M Zhu Y-G Zhang J-B He J-Z 2008 Abundance and compo-sition of ammonia-oxidizing bacteria and ammonia-oxidizing archaea com-munities of an alkaline sandy loam Environmental Microbiology 101601ndash1611DOI 101111j1462-2920200801578x

Šnajdr J Dobiaacutešovaacute P UrbanovaacuteM PetraacutenkovaacuteM Cajthaml T Frouz J BaldrianP 2013 Dominant trees affect microbial community composition and activ-ity in post-mining afforested soils Soil Biology and Biochemistry 56105ndash115DOI 101016jsoilbio201205004

Song B Yin X Zhang Y DongW 2008 Dynamics and relationships of Ca Mg Fein litter soil fauna and soil in Pinus koraiensis-broadleaf mixed forest ChineseGeographical Science 18284ndash290 DOI 101007s11769-008-0284-1

Spallek T Robatzek S Goumlhre V 2009How microbes utilize host ubiquitination CellularMicrobiology 111425ndash1434 DOI 101111j1462-5822200901346x

State Forestry Administration of China 2011 In State Forestry Administration edChina National Progress report to the UNFF Secretariat on the implementation ofNLBI and other relevant resolutions Beijing S F A China

Stieglmeier M Alves RJE Schleper C 2014 The phylum Thaumarchaeota In Rosen-berg E DeLong EF Lory S Stackebrandt E Thompson F eds The Prokaryotesother major lineages of bacteria and the archaea Berlin Springer Berlin Heidelberg347ndash362

Stutter MI Richards S 2012 Relationships between soil physicochemical microbiolog-ical properties and nutrient release in buffer soils compared to field soils Journal ofEnvironmental Quality 41400ndash409 DOI 102134jeq20100456

Szeacutekely AJ Langenheder S 2014 The importance of species sorting differs betweenhabitat generalists and specialists in bacterial communities FEMS MicrobiologyEcology 87102ndash112 DOI 1011111574-694112195

VerberkWCEP 2011 Explaining general patterns in species abundance and distribu-tions Nature Education Knowledge 3Article 38

Wang F ZhuW Chen H 2016 Changes of soil C stocks and stability after 70-year af-forestation in the Northeast USA Plant and Soil 401319ndash329DOI 101007s11104-015-2755-3


WenWMaruyama N Bo-wen L Morimoto H Zhong-xin G 2002 Relationshipsbetween bird communities and vegetation structure in Honghuarsquoerji northern innerMongolia Journal of Forestry Research 13294ndash298 DOI 101007BF02860095

Wu S-H Huang B-H Huang C-L Li G Liao P-C 2018 The Aboveground veg-etation type and underground soil property mediate the divergence of soilmicrobiomes and the biological interactionsMicrobial Ecology 75434ndash446DOI 101007s00248-017-1050-7

Xiao H Li Z Dong Y Chang X Deng L Huang J Nie X Liu C Liu LWang D LiuQ Zhang Y 2017 Changes in microbial communities and respiration followingthe revegetation of eroded soil Agriculture Ecosystems amp Environment 24630ndash37DOI 101016jagee201705026

Zechmeister-Boltenstern S Keiblinger KMMooshammerM Pentildeuelas J Richter ASardans J WanekW 2015 The application of ecological stoichiometry to plantndashmicrobialndashsoil organic matter transformations Ecological Monographs 85133ndash155DOI 10189014-07771

Zheng J Chen J Pan GWang G Liu X Zhang X Li L Bian R Cheng K Zheng J 2017A long-term hybrid poplar plantation on cropland reduces soil organic carbonmineralization and shifts microbial community abundance and composition AppliedSoil Ecology 11194ndash104 DOI 101016japsoil201611017


are not exclusively located at the original sites of deforestation and consequently thesedeforestation and afforestation events have greatly changed the landscape and ecosystemsof China (Ahrends et al 2017) Mongolia pine (Pinus sylvestris var mongolica Litv) anendemic tree in Inner Mongolia including in Honghuarsquoerji is one of the main forestationspecies in temperate regions of China Scots pine (P sylvestris L) a relative of Mongoliapine that is widespread from Western Europe to Eastern Siberia is also an importanttree in forestry (Krakau et al 2013) and is physiologically sensitive to environmentalpollution (Chudzińska amp Diatta 2014) geographic weather variation (Oleksyn et al 2003)and climate change (Hurme et al 1997 Savolainen et al 2004) Studies have demonstratedthat Scots pine and by reasonable extensionMongolia pine have broad plastic adaptabilityin response to environmental heterogeneity supporting the wide use of these trees inafforestation

Grassland afforestation is an artificial and direct change in vegetation that alters boththe above- and underground ecosystems including abiotic changes (Li et al 2016 Penget al 2014 Khamzina Lamers amp Martius 2016 Lafleur et al 2015) biotic changes (Maet al 2013 Maacuterquez et al 2015 Pedley et al 2014 Gunina et al 2017 Šnajdr et al 2013Xiao et al 2017) and ecological function changes (Lu et al 2017 Ren et al 2016 Cibils etal 2015) Land use is a significant determinant of runoff and soil redistribution processes(Arnaacuteez et al 2015) and soil microbial species responses may be sensitive to compositionalchanges in both species and ecological functions (Xiao et al 2017) However theseunderground biological changes are sometimes ascribed to changes in litter amountand chemistry that stimulate the development of the fungal community rather than thebacterial community (Klein et al 1995) The response of the prokaryotic microbiome tochanges in vegetation type is small likely because bacterial assemblies are based mainly onfunctional genes rather than species (Burke et al 2011) and the changes in soil propertiesdue to afforestation are relatively small and occur slowly (Gunina et al 2017 Jangid et al2011) In addition abundant bacterial lsquolsquocore speciesrsquorsquo with highly conserved core functions(Falkowski Fenchel amp Delong 2008) and rarely occurring species (Ai et al 2013) may actas confounders in statistical analyses of the effects of afforestation or soil properties onprokaryoticmicrobiome change Therefore habitat generalists (core species) and specialists(divergent species) in bacterial communities must be classified before analysis particularlywhen using high-throughput sequencing (HTS) technology (eg 16S rRNA metagenomicsequencing) which can generate huge amounts of data to outline the composition andstructure of bacterial communities (Szeacutekely amp Langenheder 2014)

Although the environmental bacterial composition may be sensitive to environmentalvariation the effects of these bacterial composition changes on ecofunction remainunclear A comprehensive review demonstrated that the soil microbial composition variesto alter acquisition metabolism and degradation processes in response to changes insoil phosphorus (P) due to grassland afforestation (Chen Condron amp Xu 2008) Moreimportantly environmental conditions and existing microbial diversity determine theecological function and nutrient transformation efficiency of soil microbiota (Zechmeister-Boltenstern et al 2015) Therefore quantifying the effects of afforestation on changes in soil

























Wuetal(2019)PeerJD

OI107717peerj6147

725









dbRDA ANOVA

SS Proportion F P













KW χ2 P

























































































































































































Wuetal(2019)PeerJD

OI107717peerj6147

725









dbRDA ANOVA

SS Proportion F P













KW χ2 P










































































































































































dbRDA ANOVA

SS Proportion F P













KW χ2 P







































































































































































KW χ2 P


































































































































































the effects of afforestation on soil bacterial communities in … · 2019-01-11 · at each...

Documents