Submitted 30 April 2018Accepted 17 September 2018Published 12 October 2018

Corresponding authorDianxiang Zhang dx-zhangscbgaccn

Academic editorPaolo Giordani

Additional Information andDeclarations can be found onpage 17

DOI 107717peerj5767

Copyright2018 Qian et al

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

Shifts in community composition andco-occurrence patterns of phyllospherefungi inhabiting Mussaenda shikokianaalong an elevation gradientXin Qian12 Liang Chen3 Xiaoming Guo12 Dan He4 Miaomiao Shi1 andDianxiang Zhang1

1Key Laboratory of Plant Resources Conservation and Sustainable Utilization South China Botanical GardenChinese Academy of Sciences Guangzhou Guangdong China

2University of Chinese Academy of Sciences Beijing China3CAS Key Laboratory of Pathogenic Microbiology and Immunology Institute of Microbiology ChineseAcademy of Sciences Beijing China

4Center for Ecological and Environmental Sciences South China Botanical Garden Chinese Academyof Sciences Guangzhou Guangdong China

ABSTRACTThe altitudinal effects on the distributions of phyllosphere fungal assemblages inconspecific plants remain poorly elucidated To address this phyllosphere fungalcommunities associated with Mussaenda shikokiana were investigated at four sitesacross a 350 m elevation gradient in a subtropical forest by employing Illuminametabarcoding of the fungal internal transcribed spacer 2 (ITS2) region Our resultsdemonstrated that phyllosphere fungal assemblages with a single host possessed hightaxonomic diversity and multiple trophic guilds OTU richness was significantlyinfluenced by elevation The elevation gradient also entailed distinct shifts in thecommunity composition of phyllosphere fungi which was significantly related togeographical distance andmean annual temperature (MAT) Additionally comparisonof phyllosphere fungal networks showed reduced connectivitywith increasing elevationOur data provide insights on the distribution and interactions of the phyllosphere fungalcommunity associated with a single host along a short elevation gradient

Subjects Ecology MycologyKeywords Phyllosphere fungi Biodiversity Elevation gradient Mussaenda shikokiana High-throughput sequencing

INTRODUCTIONThe fungal kingdom represents one of the largest and most diverse eukaryotic lineages withestimates of around 51 million extant species (Blackwell 2011 Truong et al 2017) Fungifunction as important ecosystem components that govern global carbon cycling plantpathology development and reproduction (Tedersoo et al 2014) Although fungi directlyimpact ecosystem functions only a fraction of the total fungal species have been described(Truong et al 2017) Therefore it is essential to obtain adequate taxonomic and ecological

How to cite this article Qian et al (2018) Shifts in community composition and co-occurrence patterns of phyllosphere fungi inhabitingMussaenda shikokiana along an elevation gradient PeerJ 6e5767 DOI 107717peerj5767

knowledge of members of the fungal lineage to advance basic ecology and to recognizeand predict the future consequences of environmental changes (Bryant et al 2008 Gemlet al 2014)

The distribution pattern of species communities along altitudinal gradients is afundamental theme of ecology and biogeography mainly because altitudinal gradientsare often characterized by different climatic conditions and complex changes in abioticandor biotic factors over short geographic distances (Bryant et al 2008 Ren et al 2018Yao et al 2017) Although there is a long history of studies that have been conductedon the elevational patterns of vascular plants and animals (Bryant et al 2008) similarstudies have only recently been made possible for microbial communities with theadvent of culture-independent molecular techniques (Datlof et al 2017 Lanzen et al2016 Wainwright et al 2017 Zhang et al 2018c) Several studies of the elevation patternsof fungal communities showed no consistent conclusions In a tropical montane cloudforest Looby Maltz amp Treseder (2016) documented increasing alpha-diversity of soilfungi and a significant shift in the community composition with increasing elevation Asimilar elevational pattern was observed in soil fungal communities in the Italian Alps(Siles amp Margesin 2016) However opposite patterns of species richness were found forectomycorrhizal fungi in the Hyrcanian forests of northern Iran (Bahram et al 2012) andfor arbuscularmycorrhizal fungi in a Tibetan alpine grassland (Liu et al 2015) In additiona unimodal pattern with a mid-elevation peak in diversity was detected in ectomycorrhizalfungal communities on the northwest slopes of the Fuji mountains (Miyamoto et al 2014)In a study conducted in the Cairn Gorm mountains in Scotland Jarvis Woodward ampTaylor (2015) found that elevation did not influence ectomycorrhizal fungal richnessof Scots pine but strongly influenced community composition Along gradients in theChangbai mountains in China neither the species richness nor community compositionof soil fungi were significantly correlated with elevation (Shen et al 2014) These differentdistribution patterns may partly be explained by difference in environmental variablesincluding edaphic qualities (Geml et al 2014 Jarvis Woodward amp Taylor 2015 Shen et al2014 Siles amp Margesin 2016) mean annual temperature and precipitation (Bahram et al2012) However most studies focused only on soil or root associated fungal assemblagesand few studies have investigated the altitudinal effects on the distribution of phyllospherefungi (Cordier et al 2012)

In natural environments microorganisms usually live within complex ecologicalnetworks rather than existing in isolation (Faust amp Raes 2012) Co-occurrence patternswhich describe how species within a community coexist are omnipresent and are crucialfor understanding the structure of microbial communities and for revealing the potentialinteractions between microbial taxa (Kay et al 2018 Ma et al 2016) Network analysiswhich can identify and describe taxon co-occurrence patterns in large datasets has provento be a powerful method that has been recently applied for characterization of bacterialcommunities across a wide variety of habitats (Baldassano amp Bassett 2016) includingsoil (Barberan et al 2012 Jiao et al 2016) ocean (Milici et al 2016) gut (Liggenstofferet al 2010 Zhang et al 2018b Zhang et al 2014) activated sludge (Ju et al 2014) andmammalian milk (Li et al 2017) These studies have uncovered previously unknown

Qian et al (2018) PeerJ DOI 107717peerj5767 227

co-occurrence patterns thereby offering new insight into the distribution of speciesassociations Several processes may be responsible for inter-taxa associations includingniche differentiation shared environmental requirements and phylogenetic relatedness(Kay et al 2018 Zhang et al 2018a) Relative to bacterial communities there have been farfewer studies conducted using high throughput sequencing data to detect the co-occurrencepatterns of fungal communities By analyzing community composition of arbuscularmycorrhizal fungi across Europe Bouffaud et al (2016) suggested that closely-related taxaare more likely to co-occur than distantly-related species MoreoverHe et al (2017) foundgreater linkage in soil fungal networks when the seasonal precipitation was manipulatedsuggesting that inter-species interactions may be intensified to adapt to differences in wateravailability In addition a study conducted on the networks of fungal communities in soilswhere there has been a prolonged potatomonoculture demonstrated scale-free small worldand modular properties (Lu et al 2013) Scale-free network usually share an importantproperty a few fungi have a tremendous number of connections to other fungi whereasmost fungi have just a handful Small world network usually showed small average pathlength and high clustering coefficient which indicated that the fungus in this communityhas close interaction Modularity is one measure of network structure Networks withhigh modularity have dense connections between the nodes within modules but sparseconnections between nodes in different modules

Phyllosphere fungi which occupy leaves superficially or intracellularly without causingvisible symptoms are important factors affecting plant health and ecosystem function(Drake White Jr amp Belanger 2018 Gomes et al 2018 Jumpponen amp Jones 2009) Forexample Phyllosphere fungi can protect host plants from pathogen damage (Arnold etal 2003 Mejia et al 2008) enhance tolerances against herbivores (Albrectsen et al 2010Estrada Wcislo amp Van Bael 2013 Hartley amp Gange 2009) and play an essential role inthe initial decomposition of leaves following senescence (Guerreiro et al 2018 Kembel ampMueller 2014 Voriskova amp Baldrian 2013) Although the elevational distribution patternsof phyllosphere fungal communities have been previously described in the Gave Valleyof the French Pyrenees (Cordier et al 2012) such an examination of the elevationalco-occurrence patterns of phyllosphere fungi in a subtropical climate has not yet beenperformed

Mussaenda shikokiana Makino is a mostly autogamous plant species that is widelydistributed throughout the warmer parts of China and Japan (Chen Luo amp Zhang 2013)The floral diversity andpollination biology ofM shikokianahas been described in a previousstudy (Chen Luo amp Zhang 2013) but little is known about the community structure ofits phyllosphere fungal assemblages To assess this we used Illumina MiSeq sequencingdata of fungal ribosomal internal transcribed spacer 2 (ITS2) to analyze the changes in thecomposition and inter-taxa interactions of the phyllosphere fungal communities inhabitingM shikokiana along an altitudinal gradient in a subtropical forest Specifically we tried toaddress the following questions (1) How does the richness and community compositionof the phyllosphere fungal assemblage change in response to the elevation gradient (2)Does the network topology of the fungal community shift significantly with elevation

Qian et al (2018) PeerJ DOI 107717peerj5767 327

MATERIALS AND METHODSStudy site and samplingThis study was conducted in the Laoshan Nature Reserve of Jingxiu which is located inGuangxi province China This region possesses a subtropical moist climate and abundantvegetation The annual mean rainfall is 1824 mm and the mean annual temperature is170 C with the lowest and highest temperatures in January (138 C) and July (217 C)respectively Here M shikokiana an erect shrub that possesses yellow hermaphroditeflowers and thinly papery leaves occurs at the southern margin of its range at a meanaltitude of 1000 m above sea level (Chen Luo amp Zhang 2013 Chen Luo amp Zhang 2014)The leaves ofM shikokiana contains several biologically active compounds that significantlystimulate immune activity andhas been used as aChinese folkmedicine (Takeda Nishimuraamp Inouye 1977) Sampling was conducted at four different altitudinal stands (838 m 934m 1078 m and 1185 m) each of which contained a high proportion of this speciesMean annual temperature (MAT) and mean annual precipitation (MAP) for each sitewere obtained from the WorldClim global climate data set with a resolution of 30 s(httpworldclimorgversion2 Table S1) At each elevation we selected five individuals(gt10 m apart from each other) that were in the fruiting stage (September 2014) Ninemature and asymptomatic leaves from the middle of three current-year shoots wererandomly collected from each individual plant and immediately placed in sterile plasticbags containing sufficient silica gel for quick drying The samples were then brought backto the laboratory and stored at minus80 C until DNA extraction

Illumina MiSeq sequencing preparationHomogenization and mechanical cell disruption of the freeze-dried samples were carriedout by grinding with liquid nitrogen using sterilized mortars and pestles The resultingpowder was then transferred to tubes with CTAB (2 CTAB 100 mM TrisndashHCl 14 MNaCl 20 mM EDTA pH 80) extraction buffer preheated to 68 C DNA extraction wasperformed following the protocol modified by Cota-Sanchez ReMarchuk amp Ubayasena(2006) Briefly an equal volume of chloroformndashisoamylol mixture was added to eachtube and mixed thoroughly to form an emulsion The mixture was spun at 13000 g for12 min and the aqueous phase was removed into a fresh centrifuge tube The aqueousphase containing DNA was re-extracted using chloroformisoamyl alcohol (241) until nointerface was visible DNA was then precipitated with an equal volume of isopropanolThe resulting genomic DNA was precipitated at 10000 g for 3 min and the DNA pelletwas washed with 70 ethanol three times Finally the DNA pellet was resuspended in100 microL TE buffer A NanoDrop spectrophotometer (Thermo Fisher Scientific WalthamMA USA) was used to quantify the extracted DNA

To amplify the fungal ITS2 region of the nuclear ribosomal repeat unit a semi-nested PCR approach was employed Initial amplification of the entire ITS region wasaccomplished with conventional primers ITS1F (CTTGGTCATTTAGAGGAAGTAA)(Gardes amp Bruns 1993) and ITS4 (TCCTCCGCTTATTGATATGC) (White et al 1990)The first round of amplification was carried out in 25 microL reaction volumes containing finalconcentrations of 02 mM dNTP 04 microM of each primer 25 mM MgSO4 5 ng template

Qian et al (2018) PeerJ DOI 107717peerj5767 427

DNA and 1 unit KOD-Plus-Neo DNA polymerase (TOYOBO Osaka Japan) Thermalcycling conditions for this amplification were as follows 30 cycles of 94 C (1 min) 56 C(50 s) and 68 C (1 min) after an initial denaturation at 94 C (4 min) and followedby a final extension at 68 C for 10 min The PCR products were diluted 50 times and1 microL of each dilution was used as the template for the second amplification of the fungalITS2 region The primer fITS7 (GTGARTCATCGAATCTTTG) (Ihrmark et al 2012) andreverse primer ITS4 extended with a unique sample identification barcode were used forthe second stage of PCR reaction which was performed under the same amplificationconditions Negative PCR controls (in which the template DNA was replaced with sterilewater) were used to evaluate whether there was contamination during the entire extractionand PCR process Each sample was amplified in triplicate to minimize PCR biases andthe three replicate products were mixed in equal volumes into a general sample that waspurified by using an Ultra-Clean PCR cleanup kit (Mo Bio Carlsbad CA USA) Thesamples were pooled with an equal molar amount (100 ng) from each sample and adjustedto 10 ng microlminus1 The library was applied to an IlluminaMiSeq PE250 platform for sequencingusing the paired end ( 2times 250 bp) option

Bioinformatics workflowThe trimming sorting and quality filtering of reads were processed using the QIIMEPipeline (Caporaso et al 2010) All sequences were demultiplexed into samples basedon unique identification barcodes FLASH-128 software (Magoc amp Salzberg 2011) wasused to combine overlapping paired-end reads The ITS2 region from the remainingsequences was identified and extracted using the fungal ITSx software (Bengtsson-Palme etal 2013) Chimeras were detected and eliminated using the UCHIME program (Edgar etal 2011) referenced with the UNITE database ( httpsuniteutee (Koljalg et al 2013)Non-chimera sequences that passed the filtering processes were binned into operationaltaxonomic units (OTUs) at a 97 similarity cutoff using the UPARSE pipeline (Edgar et al2011) All singletons OTU were excluded from the dataset Subsequently the taxonomicidentity of each OTU was determined based on BLAST search against the UNITE referencedatabase and a BLAST e-value lt1eminuswas a criterion for matching Further OTUs occurringin only one sample and OTUs with lt10 reads were removed because those rare sequencesmay have been derived from contaminations or PCRsequencing errors (Toju Tanabeamp Ishii 2016) For downstream analyses each sample was rarefied to the minimumsequencing depth by randomly selecting subsets of reads implemented in the MOTHURprogram (Schloss et al 2009)

Community analysesTo evaluate the comprehensiveness of the sampling strategy the taxa accumulationcurves for each site were calculated by employing the lsquolsquospecaccumrsquorsquo function in the veganR package (Harrison et al 2016 Oksanen et al 2016) Vectors representing horizontalspatial structure were obtained by an analysis of principal coordinates of neighbor matrices(PCNM) (Borcard amp Legendre 2002) based on the geographical coordinates using thelsquolsquoPCNMrsquorsquo package (Legendre et al 2010) Linear regression models were used to test therelationship between fungal OTU richness and elevation

Qian et al (2018) PeerJ DOI 107717peerj5767 527

To identify the effects of elevation PCNM vectors and climate factors on phyllospherefungal OTU richness we first constructed generalized linear mixed models (GLMM)included elevation site as a random effect However we found that random factorsexplaining no variation in the data so we switched to construct generalized linearmodels (GLM) These models were then examined by analysis of variance (ANOVA)tests Moreover the effects of those factors on fungal community structure was assessedby permutational multivariate analysis of variance (PERMANOVA) with site as therandom effect using the vegan R package (9999 random permutations) Non-metricmultidimensional scaling (NMDS) was performed with the BrayndashCurtis distance based onaHellinger-transformedmatrix Subsequently the lsquolsquoenvfitrsquorsquo function of the vegan R packagewas used to fit the elevation variable into theNMDSplot Altitudinal distances geographicaldistance and Bray-Curtis dissimilarity matrices were analyzed for correlations to evaluatethe distance-decay rates and Mantel tests were employed to assess the significance of thedistance-decay regressions from zero (9999 random permutations) (Oono Rasmussenamp Lefevre 2017) The appropriate trophic categories (symbiotroph pathotroph andsaprotroph) of each OTU were inferred by searching against the FUNGuild data base(Nguyen et al 2016b)

To further investigate the phyllosphere fungal taxonomic distribution significanttaxonomic differences between the four elevation sites were identified using LinearDiscriminant Analysis (LDA) effect size (LEfSe) (Segata et al 2011) which introducesthe non-parametric factorial KruskalndashWallis (KW) sum-rank test (α= 005) to identifytaxa having significant differential abundances between four elevation stands (employingone-against-all comparisons) (Bokulich et al 2014) A logarithmic LDA score of 30 wasset as the threshold for discriminative features

To identify the indicator OTUs for each elevation site LEfSe analysis and LegendrersquosINDVAL analysis were combined as per Hubbard et al (2018) Legendrersquos INDVALprocedure was performed in R using the lsquolsquolabdsvrsquorsquo package (Sapkota et al 2015) Thefungal OTUs with an indicator value gt025 and a significant P values (P lt 005) wereconsidered indicator species (Dufrene amp Legendre 1997 Sapkota et al 2015) HoweverOTUs with low abundance hold little indicator information (Rime et al 2015) To avoidrandom effects caused by rare OTUs only those with gt500 reads were defined as true keyplayers (Eusemann et al 2016)

Network analysesTo remove poorly represented OTUs and reduce network complexity only abundantOTUs with a proportion of the total sequences over 002 were kept in the OTU table Theco-occurrence network was inferred based on the Spearmanrsquos Rho between the pairwiseOTUsmatrix constructed by the psych R package (Revelle 2017) The P-values for multipletesting were calculated by using the false discovery rate (FDR) controlling procedure as perBenjamini amp Hochberg (1995) A valid co-occurrence event was considered to be robust ifthe correlation coefficient r gt|06| and if it was statistically significant at P lt 005 (Widderet al 2014) Network images for each elevation site were generated with Gephi (BastianHeymann amp Jacomy 2009) To describe the topology properties of the co-occurrence

Qian et al (2018) PeerJ DOI 107717peerj5767 627

networks the feature set (number of positive correlations number of negative correlationsnumber of vertices average degree average path length centralization betweennessdiameter centralization degree and clustering coefficient) was calculated with the igraphR package (Csardi 2006) Network diameter is a measure of how integrated the networkis (Proulx Promislow amp Phillips 2005) A network that has multiple fractured componentswould have a smaller average diameter (Lawes et al 2017) Spearman correlations andsignificance tests (level at plt 005) between those properties were calculated and plottedusing the corrplot R package (Taiyun amp Viliam 2017)

RESULTSData characteristicsIn total 451726 raw reads were obtained using Illumina MiSeq sequencing of the ITS2amplicon libraries After a series of filterings 256258 high-quality reads were clusteredinto 792 operational taxonomic units (OTUs) at a 97 sequence similarity cutoff Finallythe sequence number of each sample was rarefied to 4606 resulting in a normalizeddataset that contained 708 fungal OTUs The abundance distribution of the OTUs washighly skewed with plentiful rare and only a small number of very abundant OTUsThe 50 most abundant OTUs accounted for 724 of all sequences The rarefaction andspecies accumulation curves reached a saturation plateau indicating that our samplingand sequencing depths were sufficient to capture the majority of the phyllosphere fungalOTUs at each elevation site (Fig 1A) Fungal communities were strongly dominated bymembers of the phylum Ascomycota (841 of all reads) followed by Basidiomycota(139) The relative abundance of Basidiomycota increased slightly with elevation Atthe class level the most abundant OTUs were assigned to Dothideomycetes (470)Eurotiomycetes (193) Sordariomycetes (123) and Tremellomycetes (80) (Fig 2A)Different community compositions between elevations were observed at this taxonomiclevel (Fig 2B) For example the abundance of the Eurotiomycetes and Tremellomyceteswas higher at the highest elevation than at other sites while Ustilaginomycetes species werehardly found at 1185 m Among the 52 identified orders the three most abundant wereCapnodiales (273) Chaetothyriales (193) and Pleosporales (119) A total of 210genera were classified and Cladosporium (154) was the most abundant genus in the Mshikokiana phyllosphere fungal community The taxonomic assignment of each OTU isprovided in Table S2

At the genus level 719 of OTUs (509 of 708) were assigned to functional groupsin FUNGuild database (Nguyen et al 2016b) Dominant guilds were classified in sevengroups of trophic modes (Fig 2B) Pathotrophic fungi were the highest in abundance(295 of all sequences) followed by the pathotroph-symbiotroph group (198) andthe saprotroph group (170 Fig 2C) Compositions of trophic guilds were also differentbetween elevations (Fig 2D) The relative abundance of the pathotrophs was highest atthe lowest elevation while the symbiotrophs were more abundant at the highest elevationthan at other sites

Qian et al (2018) PeerJ DOI 107717peerj5767 727

0 5000 10000 15000 20000 25000

010

020

030

040

050

0

Number of reads

Num

ber o

f OTU

s

838 m 943 m1078 m1185 m

A

F = 7053 P = 0016

Elevation (m)

100

150

200

250

Fung

al ri

chne

ss

B

800 900 1000 1100 1200

R2 = 0243 P = 0016

Figure 1 Fungal richness along the elevation gradient (A) OTU accumulation curves of the phyllo-sphere fungal communities at four elevation sites (B) OTU richness of phyllosphere fungal communitiesat each site The results from the linear regressions and the one-way analysis of variance (ANOVA) are in-dicated in the upper and lower part of the figure respectively

Full-size DOI 107717peerj5767fig-1

Variation in fungal richness and community compositionFungal OTU richness increased significantly with increasing elevation (R2

= 0243P = 0016 Fig 1B) The generalized linear model (GLM) also supported a significanteffect of elevation on M shikokiana phyllosphere fungal richness (F = 6484 P = 0021Table S3) The nonmetric multidimensional scaling (NMDS) plot revealed a generalclustering according to elevation site and the members of the 1185 m stand were clearlyseparated from the other clusters (Fig 3A) The fit of the variable selected by the lsquolsquoenvfitrsquorsquofunction showed that patterns of community similarity were strongly structured byelevation (R2

= 0736 P = 0001) followed by PCNM vector (R2= 0534 P = 0003) and

MAT (R2= 0739 P = 0001) The significant effects of those factors on the community

structure were further reflected in the results by the permutational multivariate analysis ofvariance (PERMANOVA) assim160 122 and 109 of community variation could beexplained by elevation PCNM andMAT respectively (Table 1) Furthermore Mantel testsindicated that fungal community dissimilarities were significantly positively correlated withaltitudinal distance (r = 0570 P = 0001 Fig 3B) and geographical distance (r = 0503P = 0001 Fig 3C) respectively

The community compositions of phyllosphere fungi between the four elevation siteswere significantly different as is shown by the linear discriminant analysis effect size(LEfSe) cladogram (Fig 4) These results indicate that many groups at different taxonomiclevels can be significantly distinguished along the elevation gradient For exampleSordariomycetes Ustilaginomycetes and Agaricomycetes appeared as main discriminantclades at 838 m 934 m and 1078 m respectively Eurotiomycetes and Lecanoromycetesare the most differentially abundant fungal classes at 1185 m At the taxonomic orderlevel Glomerellales Diaporthales and Chaetothyriales were significantly enriched at838 m Ustilaginales was found to predominate at 934 m Hysteriales Sordariales

Qian et al (2018) PeerJ DOI 107717peerj5767 827

000

025

050

075

100

Rel

ative

abu

ndan

ce

838 943 1078 1185Elevation (m)

000

025

050

075

100

Rel

ative

abu

ndan

ce

838 943 1078 1185

Elevation (m)

PathotrophPathotrophminusSaprotrophPathotrophminusSaprotrophminusSymbiotrophPathotrophminusSymbiotrophSaprotrophSaprotrophminusSymbiotrophSymbiotrophUndefined

295

198170

133

132

4624

03

A

C

AgaricomycetesTremellomycetesUstilaginomycetesAscomycota_cls_sedisDothideomycetesEurotiomycetesLeotiomycetesSordariomycetesOtherUnidentified

470

193

122

80

3938

2015

1411

Basidiomycota

Ascomycota

B

D

Figure 2 Taxonomic composition and trophic guilds of the fungal communities Pie charts showingthe distribution of sequences of (A) fungal classes and (C) fungal trophic guilds Bar charts showing thetaxonomic composition of the phyllosphere fungal communities in four elevation sites at (B) taxonomicclass level and (D) tropic guilds The fungal classes with a relative abundance oflt01 were classified tolsquolsquootherrsquorsquo

Full-size DOI 107717peerj5767fig-2

Microbotryales Agaricales Corticiales and Thelephorales were more abundant at 1078 mand Chaetothyriales Polyporales and Lecanorales presented differential abundances at1185 m

Taken together the LEfSe and INDVAL analyses detected a total of four indicatorOTUs (Mycosphaerella quasiparkii Gibberella fujikuroi Colletotrichum gloeosporioidesand Phyllosticta capitalensis) that were significantly associated with the phyllosphere

Qian et al (2018) PeerJ DOI 107717peerj5767 927

minus10 minus05 00 05 10

minus10

minus05

00

05

10

NMDS1

NM

DS2

934 m

1078 m

838 m

A

B

1185 m

0 50 100 150 200 250 300 350

05

06

07

08

Fung

al c

omm

unity

dis

sim

ilarit

y

Altitudinal distance (m)

ElevationPCNM

00 05 10 15 20 25 30 35

05

06

07

08

C

Fung

al c

omm

unity

dis

sim

ilarit

y

Geographical distance (km)

MAT

Mantel test r = 0570 P = 0001 Mantel test r = 0503 P = 0001

Figure 3 The effects of elevation gradient geographical distance and climate factor on the fungal com-munity composition (A) Nonmetric multidimensional scaling (NMDS) ordination of variation in com-munity structure (BrayndashCurtis dissimilarity) of the phyllosphere fungal assemblages The arrow representssignificant (P lt 001) vectors chosen by the lsquolsquoenvfitrsquorsquo function to fit the fungal communities (B) Correla-tions between altitudinal distance and BrayndashCurtis dissimilarities of the phyllosphere fungal community(C) Correlations between geographical distance and Bray-Curtis dissimilarities of the phyllosphere fungalcommunity Mantel test results are shown on the figures

Full-size DOI 107717peerj5767fig-3

Table 1 Relative importance of elevation and geographical distance on the phyllosphere fungal com-munity composition as revealed by permutational multivariate analysis of variance (PERMANOVA)

Variable DF SS F-statistics R2 P value

Elevation 1 0765 4218 0160

PCNM 1 0584 3219 0122

MAT 1 0519 2861 0109

Residuals 16 2900 0608Total 19 4767 1000

NotesDF degrees of freedom SS sum of squares PCNM principal coordinates of neighbor matrices MAT mean annual tem-perature

P lt 0001

at 838 m five indicator OTUs (Ustilago sp Cladosporium sp Meira sp Hannaella spand Chaetothyriales sp) at 934 m nine indicator OTUs (Plectosphaerella cucumerinaPseudopithomyces maydicus Ramularia rumicis-crispi Bulleribasidium variabile Psiloglo-nium sp Pseudocercospora fici Strelitziana mali Phyllosticta sp and Tetracladium sp) at

Qian et al (2018) PeerJ DOI 107717peerj5767 1027

Sord

ario

myc

etes

Agaricomycetes

Ustilaginom

ycetes

Ascomycota_cls_unclassified

Eurotiomycetes

Lecanoromycetes

838 m934 m1078 m1185 m

je

cbf5f4

f2f0

e8e9

e7

e6e4e0

e2e1d9

d6d5d4

d7

d2

d1 d0c9

c8

c7c4

c3

c2c6

c1c5

c0b9

b2b8

b5b1

b7b4

a9a6

a4

a8 a3

a1

a0z y

x

v

s pml

r oqn

ki h

gf

da

f3f1

w

ut

e5e3

d8d3

Figure 4 Cladogram showing significant discriminative taxa of the phyllosphere fungi associated withMussaenda shikokiana along the elevation gradient The filled circles from inside to out indicate the tax-onomic levels with phylum class order family and genus Circles or nodes shown in color correspondingto that of different elevation sites represented a significantly more abundant group

Full-size DOI 107717peerj5767fig-4

04 06 08 1 12

Mycosphaerella quasiparkii

Gibberella fujikuroi

Colletotrichum gloeosporioides

Cladosporium sp

Ustilago sp

Pseudocercospora fici

Phyllosticta juglandis

Cryptococcus sp

Periconia sp

Trichomerium foliicola

Chaetothyriales sp

Cyphellophora gamsii

Cryptococcus sp

838 m 934 m 1078 m 1185 m

Mycosphaerella quasiparkiiColletotrichum gloeosporioides

Phyllosticta capitalensisChaetothyriales sp

Hannaella spUstilago sp

Meira spTetracladium sp

Phyllosticta spStrelitziana mali

Pseudocercospora ficiPsiloglonium sp

Bulleribasidium variabileRamularia rumicis-crispi

Pseudopithomyces maydicusPlectosphaerella cucumerina

Chaetothyriales spTrichomerium foliicolaCyphellophora gamsii

Cryptococcus spWojnowiciella cissampeli

Periconia sp

1 2 3 4 5

LDA score (log 10)

838 m 934 m 1078 m 1185 m

Indicator value

A B

Figure 5 Identification of indicator species based on (A) INDVAL analysis and (B) LEfse analyses In-dicator fungi with INDVAL indicator valuegt025 or LDA scoresgt3 at four elevation sites

Full-size DOI 107717peerj5767fig-5

1078 m and eight indicator OTUs (Cryptococcus sp Cyphellophora gamsii Chaetothyrialessp Trichomerium foliicola Periconia sp Cryptococcus sp Phyllosticta juglandis andWojnowiciella cissampeli) at 1185 m (Fig 5)

Qian et al (2018) PeerJ DOI 107717peerj5767 1127

A

B

Agaricomycetes Eurotiomycetes

Sordariomycetes Tremellomycetes UnidentifiedUstilaginomycetes

Ascomycota unclassified Cystobasidiomycetes Dothideomycetes Exobasidiomycetes Leotiomycetes

Microbotryomycetes Pucciniomycetes Saccharomycetes

Pathotroph PathotrophminusSaprotroph PathotrophminusSaprotrophminusSymbiotroph PathotrophminusSymbiotroph Saprotroph SaprotrophminusSymbiotroph

Symbiotroph Undefined

838 m

838 m 934 m

934 m 1078 m

1078 m 1185 m

1185 m

Figure 6 Co-occurrence networks for the phyllosphere fungal communities at each elevation site Thenode color indicates the corresponding taxonomic assignment at class level (A) and trophic mode (B) Thenode sizes are proportional to the OTU abundances and the color of each line reflects positive (yellow) ornegative (blue) associations

Full-size DOI 107717peerj5767fig-6

Key topological properties of the co-occurrence networksCo-occurrence patterns of the phyllosphere fungal communities were constructed foreach elevation site (Fig 6) Statistically significant and strong correlations were visualizedas fungal association networks Topological properties delineated the interaction patternnodes and were used to distinguish differences between the network structures of the fourelevation groups The clustering coefficient which reflects how well nodes are connectedwith neighboring ones decreased significantly with increasing elevation (Fig 7) Thisindicated that the network became more discrete and sparser with increasing elevationThe network diameters of the M shikokiana phyllosphere fungal communities showed asignificantly decrease with elevation (Fig 7) In total over 98 of all correlations werepositive However the negative links were significantly positive related to elevation (Fig7) indicating an increasing mutual exclusion between the fungal species

DISCUSSIONOverall taxonomic diversity and community richnessA taxonomically diverse assemblage of phyllosphere fungi (708 OTUs) was detectedin M shikokiana with taxa from numerous genera (210) and families (134) TheOTU richness was higher than that of the needle mycobiome of Picea glauca at the

Qian et al (2018) PeerJ DOI 107717peerj5767 1227

Elevation

Number of postive correlations

Number of negative correlations

Number of edges

Number of nodes

Connectance

Average degree

Average path length

Diameter

Clustering coefficient

Centralization betweenness

Centralization degree

Elev

ation

Num

ber o

f pos

tive

corre

lation

s

Num

ber o

f neg

ative

corre

lation

s

Num

ber o

f edg

esNu

mbe

r of n

odes

Conn

ecta

nce

Aver

age

degr

eeAv

erag

e pa

th le

ngth

Diam

eter

Clus

terin

g co

effic

ient

Cent

raliz

ation

bet

ween

ness

Cent

raliz

ation

deg

ree

minus1minus09minus08minus07minus06minus05minus04minus03minus02minus0100102030405060708091

Figure 7 Spearman correlations between topological properties of the co-occurrence networks and el-evation gradient lowast indicates P lt 005

Full-size DOI 107717peerj5767fig-7

northern treeline in Alaska (Sapkota et al 2015) but lower than that of the phyllospherefungal community of Fagus sylvatica in the French Pyrenees (Cordier et al 2012) Thephyllosphere fungal communities were largely composed of Ascomycota and dominatedby members of Dothideomycetes as previously reported in foliar fungal studies Forexample Yang et al (2016a) found that the phyllosphere fungi of Betula ermanii in theChangbai mountains were dominated by Ascomycota and in particular by members ofDothidiomycetes Harrison et al (2016) also revealed that the majority of fungal OTUsassociated with the leaves of Sequoia sempervirens in the western United States wereassigned to the Dothideomycetes Similar observations have been documented in the foliarfungal communities of Eucalyptus grandis (Kemler et al 2013) Picea glauca (Eusemannet al 2016) Picea abies (Nguyen et al 2016a) Metrosideros ploymorpha (Zimmermanamp Vitousek 2012) and Hopea ferrea (Izuno et al 2016) However different dominanttaxa were detected in Fagus sylvatica (Cordier et al 2012 Siddique amp Unterseher 2016Unterseher et al 2016) and several bryophytes species (Davey et al 2013) indicatingdifferent environmental filtration and host selection across distinct geographic origins andplant species (Yang et al 2016a Peršoh 2015 Rodriguez et al 2009))

Qian et al (2018) PeerJ DOI 107717peerj5767 1327

Most fungal taxa identified here could be assigned to different functional guildsThe seven main ecological guilds (pathotroph symbiotroph saprotroph pathotroph-saprotroph pathotroph-symbiotroph saprotroph-symbiotroph and saprotroph-pathotroph-saprotroph) were all found in our study Many fungal species belong tomore than one of the ecological categories which is very common in fungal kingdomFor example Fusarium culmorum could be a plant pathogen causing crown rot in cereals(Hollaway et al 2013) but also could be a symbiotic fungal endophyte conferring salttolerance to Leymus mollis (Rodriguez et al 2008) Our result is consistent with a recentstudy on the fungal endophytes of Fagus sylvatica and suggest undeniable advantagesof high-throughput sequencing over traditional cultivation in terms of uncovering agood representation of the major functional guilds as well as rare fungal taxa (SiddiqueKhokon amp Unterseher 2017) In addition to saprotrophic and symbiotrophic taxanumerous pathotrophic fungi (295) were detected on M shikokiana leaves Severalindicator species including Gibberella fujikuroi Ramularia rumicis-crispi Mycosphaerellaquasiparkii Plectosphaerella cucumerina and Strelitziana mali also have been recordedas fungal phytopathogens (De Vos et al 2014 Petriacq Stassen amp Ton 2016 Wardle ampLindahl 2014) In accordance with this Sapkota et al (2015) found that host-specificpathogens lived in a lsquolsquosearsquorsquo of nonspecific fungi in the phyllosphere of six Poaceaespecies Notably several endophytic fungi may be latent pathogens (Vacher et al 2016)Colletotrichum gloeosporioides a significant indicator at 838 m has previously beenreported as an endophytic and symbiotic fungus that serves as a biocontrol agent forTheobroma cacao pathogens (Mejia et al 2008) and is itself also an important plantpathogen causing anthracnose (Phoulivong et al 2010) Alvarez-Loayza et al (2011) foundthat high light environments can trigger the production of H2O2 by Diplodia mutila acommon asymptomatic endophyte in Iriartea deltoidea resulting in pathogenicity due toconstraining niche-space filling of the plant species

Effects of elevation on fungal richness and community compositionOur results show that phyllosphere fungal richness and the community structure of fungiassociated with M shikokiana are affected significantly by elevation These findings agreewith the patterns of fungal endophytes ofB ermanii in theChangbaimountains (Yang et al2016b) and of F sylvatica phyllosphere fungi in the French Pyrenees (Cordier et al 2012)Shifts in community composition along the elevational gradient were significantly related tochanges in temperatureMany studies showed that altitudinal distribution patterns of fungalassemblages often associate with climatic variables and environmental factors that co-varywith elevation The variations in incidence and the abundance of F sylvatica phyllospherefungi were significantly correlated with variation in average temperature (Cordier et al2012) The endophytic community composition ofM ploymorpha can be driven by rainfallalong elevation gradient (Zimmerman amp Vitousek 2012) The ectomycorrhizal fungalrichness on Mount Kinabalu in Malaysia is highest in the elevation zone with the mostavailable water (Geml et al 2017) Jarvis Woodward amp Taylor (2015) also suggested thatsignificant climatic niche partitioning between ectomycorrhizal fungi is associated with asingle host along a short (300 m) elevation gradient Therefore variation in some climatic

Qian et al (2018) PeerJ DOI 107717peerj5767 1427

parameters that we did not record in this studymdashsuch as humidity wind exposure orUV radiation along the elevation gradientmdashmay contribute to the distribution patternsof phyllosphere fungi in M shikokiana The indicator species at the elevation of 1185 minclude the yeasts Cryptococcus spp and the melanized black yeast Cyphellophora gamsiiall of which possess a strong ability to withstand UV radiation (Sapkota et al 2015) whichcorresponds to the fact that solar radiation is higher at higher elevations In addition tothe importance of elevation site parameters in shaping fungal communities abundantresearch has suggested a significant interplay between phyllosphere fungi and their hostplants (Eusemann et al 2016) Whipps et al (2008) summarized that the development ofphyllosphere microbiota must become involved in the phenotypic traits (leaf chemistryphysiology or morphology) exhibited by host plants For instance elevated leaf carbonavailability could significantly enhance the phyllosphere fungal richness of B ermanii alongthe elevation gradient due to broadening of niche space associated with the phyllosphereenvironment (Yang et al 2016b) Furthermore some phyllosphere fungi are dormantsaprotrophs involved in the decomposition of senescent leaves (Vacher et al 2016) Thesignificant indicator species at the highest elevation in our study Cyphellophora gamsii isa saprotrophic fungus usually found in leaf litter (Crous et al 2016 Madrid et al 2016)Leaf litter at higher elevations is usually more recalcitrant because of increased nutrientlimitation and leaf thickness these features may facilitate the presence of specializedfungi capable of breaking down recalcitrant organic matter (Looby Maltz amp Treseder2016) This pattern further supports a greater richness of fungi released to the soiland air through decomposition which may lead to increases in phyllosphere fungalrichness by horizontal transmission (Osono amp Mori 2003) In addition to altitudinaldistance geographical distance also plays an important role in shaping phyllosphere fungalcommunity composition indicating dispersal limitation or environment isolation over alocal scale (Hanson et al 2012)

Co-occurrence networks of phyllosphere fungiBy providing a vivid and simplified version of the complex ecological interactions thatshapemicrobial communities network analyses of the co-occurrence patterns of significanttaxa may help to deepen our understanding of the altitudinal distribution patterns ofphyllosphere fungal assemblages along elevation gradients (Eiler Heinrich amp Bertilsson2011) Our results showed that the phyllosphere fungal network of M shikokiana shiftedbetween different elevation sites Most of the correlations between fungal OTUs (838 m99 934 m 99 1078 m 98 and 1185 m 97) were identified as co-occurrences(positive correlations) Dominant positive associations suggest that most fungal taxa maysynergistically act or share similar ecological niches in the phyllosphere environment(Zhang et al 2018a) A prevalence of positive correlations has also been found in othermicrobial networks (Aschenbrenner et al 2017 Chow et al 2014) although a much lowerproportion was found in soil fungal communities It is possible that nutrient use efficiencyand mechanisms underlying fungal community assembly in soil are markedly differentfrom those in the phyllosphere (Peršoh et al 2018)

Qian et al (2018) PeerJ DOI 107717peerj5767 1527

Although negative correlations which suggest co-exclusion between two taxa weremuchrarer than positive ones the number of negative links increased significantly with elevationThis trend may indicate a more competitive relationship between phyllosphere fungi athigher elevations In phyllosphere microbial communities competition for limited spaceand nutrient resources as well as the production of metabolic compounds and interferencewith hormone signaling pathways are the major mechanisms by which microbialinhabitants inhibit each other (Vorholt 2012) This result agrees with the finding thatnetworks of endophytic communities are almost unconnected inmore northern temperateregions (Guerreiro et al 2018) Interestingly many negative interactions occurred betweenpathogenic members and other taxa a similar finding has been reported in endophyticfungal communities (Peršoh 2013) This may help to maintain community stabilityand plant health Fokkema (1993) suggested that competition between pathogens andsaprotrophs acts as a natural occurring biocontrol process in the phyllosphere This complexmicrobial interaction might possess barrier effects enhancing the plantrsquos resistance topathogenic infections in the phyllosphere (Vorholt 2012) Negative correlations can finallyresult in a less stable network structure (Lawes et al 2017) This increasing tendency ofnegative correlations also could be partly responsible for less coherent and compact networkstructure at higher elevation as shown by clustering coefficients and diameters (Figs 6and 7) Lower elevations may provide more favorable conditions for fungal interactionand niche sharing Higher elevations are often associated with harsher environmentalconditions for plant growth such as stronger ultraviolet exposure and lower temperatures(Pauchard et al 2009) The decrease in connectivity and integrity of the co-occurrencenetworks may indicate a change towards stochastic assembly processes prevention of bioticcommunication or shifts in biogeochemical pathways due to environmental stress (Laweset al 2017)

CONCLUSIONSIn summary this study reports high-throughput sequencing data that suggests that anelevation gradient of only 350 m can entail significant shifts in OTU richness communitycomposition and co-occurrence patterns of phyllosphere fungi in a single host species Thephyllosphere fungal assemblages inhabitingM shikokiana show high diversity in taxonomicclassification and trophic stages Fungal richness showed a pattern of monotonal increasealong an elevation gradient In addition elevation is a critical predictor of the fungalcommunity structure Shifts in community composition along an elevation gradient wereclearly related to MAT which is suggestive of significant temperature niche partitioningamong phyllosphere fungal species A distance-decay pattern was also found in the fungalassemblages Co-occurrence networks became less centralized and more fractured withincreasing elevation However limited information on abiotic and biotic factors thatcovary with elevation demands further investigation to determine the exact drivers andmechanisms of this altitudinal partitioning

Qian et al (2018) PeerJ DOI 107717peerj5767 1627

ACKNOWLEDGEMENTSWe are grateful to Prof Liangdong Guo for field and lab assistance We thank ZhengkaiHe for assistance in sampling

ADDITIONAL INFORMATION AND DECLARATIONS

FundingThis work was supported by the Chinese Academy of Sciences (XDA13020504) Ministryof Science and Technology of China (Grant No 2013FY111200) and National NaturalScience Foundation of China (31570314 and U1603231) The funders had no role in studydesign data collection and analysis decision to publish or preparation of the manuscript

Grant DisclosuresThe following grant information was disclosed by the authorsChinese Academy of Sciences XDA13020504Ministry of Science and Technology of China 2013FY111200National Natural Science Foundation of China 31570314 U1603231

Competing InterestsThe authors declare there are no competing interests

Author Contributionsbull Xin Qian conceived and designed the experiments performed the experiments analyzedthe data prepared figures andor tables authored or reviewed drafts of the paperbull Liang Chen Xiaoming Guo Dan He and Miaomiao Shi contributed reagentsmaterial-sanalysis toolsbull Dianxiang Zhang conceived and designed the experiments authored or reviewed draftsof the paper approved the final draft

DNA DepositionThe following information was supplied regarding the deposition of DNA sequences

Representative sequences of the fungal OTUs are available in European NucleotideArchive (ENA) under accession number PRJEB25945

Data AvailabilityThe following information was supplied regarding data availability

The raw data are provided in the Supplemental Files

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj5767supplemental-information

Qian et al (2018) PeerJ DOI 107717peerj5767 1727

REFERENCESAlbrectsen BR Bjoumlrkeacuten L Varad A Hagner AringWedinM Karlsson J Jansson S 2010

Endophytic fungi in European aspen (Populus tremula) leavesmdashdiversity detectionand a suggested correlation with herbivory resistance Fungal Diversity 4117ndash28DOI 101007s13225-009-0011-y

Alvarez-Loayza PWhite Jr JF Torres MS Balslev H Kristiansen T Svenning JC GilN 2011 Light converts endosymbiotic fungus to pathogen influencing seedlingsurvival and niche-space filling of a common tropical tree Iriartea deltoidea PLOSONE 6e16386 DOI 101371journalpone0016386

Arnold AE Mejia LC Kyllo D Rojas EI Maynard Z Robbins N Herre EA 2003Fungal endophytes limit pathogen damage in a tropical tree Proceedings of theNational Academy of Sciences of the United States of America 10015649ndash15654DOI 101073pnas2533483100

Aschenbrenner IA Cernava T Erlacher A Berg G GrubeM 2017 Differential sharingand distinct co-occurrence networks among spatially close bacterial microbiota ofbark mosses and lichensMolecular Ecology 262826ndash2838 DOI 101111mec14070

BahramM Polme S Koljalg U Zarre S Tedersoo L 2012 Regional and local patternsof ectomycorrhizal fungal diversity and community structure along an altitudinalgradient in the Hyrcanian forests of northern Iran New Phytologist 193465ndash473DOI 101111j1469-8137201103927x

Baldassano SN Bassett DS 2016 Topological distortion and reorganized modularstructure of gut microbial co-occurrence networks in inflammatory bowel diseaseScientific Reports 626087 DOI 101038srep26087

Barberan A Bates ST Casamayor EO Fierer N 2012 Using network analysis to exploreco-occurrence patterns in soil microbial communities The Isme Journal 6343ndash351DOI 101038ismej2011119

BastianM Heymann S JacomyM 2009 Gephi an open source software for exploringand manipulating networks In International conference on weblogs and social media

Bengtsson-Palme J RybergM HartmannM Branco S Wang Z Godhe A DeWitP Sanchez-Garcia M Ebersberger I De Sousa F Amend AS Jumpponen AUnterseher M Kristiansson E Abarenkov K Bertrand YJK Sanli K Eriksson KMVik U Veldre V Nilsson RH 2013 Improved software detection and extractionof ITS1 and ITS2 from ribosomal ITS sequences of fungi and other eukaryotesfor analysis of environmental sequencing dataMethods in Ecology and Evolution4914ndash919 DOI 1011112041-210x12073

Benjamini Y Hochberg Y 1995 Controlling the false discovery ratemdasha practical andpowerful approach to multiple testing Journal of the Royal Statistical Society SeriesB-Methodological 57289ndash300

Blackwell M 2011 The fungi 1 2 3 51 million species American Journal of Botany98426ndash438 DOI 103732ajb1000298

Bokulich NA Thorngate JH Richardson PMMills DA 2014Microbial biogeographyof wine grapes is conditioned by cultivar vintage and climate Proceedings of

Qian et al (2018) PeerJ DOI 107717peerj5767 1827

the National Academy of Sciences of the United States of America 111E139ndashE148DOI 101073pnas1317377110

Borcard D Legendre P 2002 All-scale spatial analysis of ecological data by meansof principal coordinates of neighbour matrices Ecological Modelling 15351ndash68DOI 101016S0304-3800(01)00501-4

BouffaudML Creamer RE Stone D Plassart P Van Tuinen D Lemanceau PWipfD Redecker D 2016 Indicator species and co-occurrence in communities ofarbuscular mycorrhizal fungi at the European scale Soil Biology amp Biochemistry103464ndash470 DOI 101016jsoilbio201609022

Bryant JA Lamanna C Morlon H Kerkhoff AJ Enquist BJ Green JL 2008 Col-loquium paper microbes on mountainsides contrasting elevational patterns ofbacterial and plant diversity Proceedings of the National Academy of Sciences of theUnited States of America 105(Suppl 1)11505ndash11511 DOI 101073pnas0801920105

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

Chen S Luo ZL Zhang DX 2013 Self-pollination in buds and homostyly inMussaendashikokiana(Rubiaceae) a monomorphic species in a distylous clade Journal ofSystematics and Evolution 51731ndash742 DOI 101111jse12031

Chen S Luo Z Zhang D 2014 Pre- and post-zygotic reproductive isolation between co-occurring Mussaenda pubescens var alba and M shikokiana (Rubiaceae) Journal ofIntegrative Plant Biology 56411ndash419 DOI 101111jipb12140

Chow CE KimDY Sachdeva R Caron DA Fuhrman JA 2014 Top-down controlson bacterial community structure microbial network analysis of bacteria T4-likeviruses and protists The Isme Journal 8816ndash829 DOI 101038ismej2013199

Cordier T Robin C Capdevielle X Fabreguettes O Desprez-LoustauML VacherC 2012 The composition of phyllosphere fungal assemblages of European beech(Fagus sylvatica) varies significantly along an elevation gradient New Phytologist196510ndash519 DOI 101111j1469-8137201204284x

Cota-Sanchez JH ReMarchuk K Ubayasena K 2006 Ready-to-use DNA extracted witha CTAB method adapted for herbarium specimens and mucilaginous plant tissuePlant Molecular Biology Reporter 24161ndash167 DOI 101007Bf02914055

Crous PWWingfield MJ Richardson DM Le Roux JJ Strasberg D Edwards JRoets F Hubka V Taylor PW HeykoopMMartinMP Moreno G Sutton DAWiederhold NP Barnes CW Carlavilla JR Gene J Giraldo A Guarnaccia VGuarro J Hernandez-RestrepoM Kolarik M Manjon JL Pascoe IG Popov ESSandoval-Denis MWoudenberg JH Acharya K Alexandrova AV Alvarado PBarbosa RN Baseia IG Blanchette RA Boekhout T Burgess TI Cano-Lira JFCmokova A Dimitrov RA DyakovMY Duenas M Dutta AK Esteve-Raventos FFedosova AG Fournier J Gamboa P Gouliamova DE Grebenc T GroenewaldM

Qian et al (2018) PeerJ DOI 107717peerj5767 1927

Hanse B Hardy GE Held BW Jurjevic Z Kaewgrajang T Latha KP Lombard LLuangsa-Ard JJ Lyskova P Mallatova N Manimohan P Miller AN MirabolfathyMMorozova OV Obodai M Oliveira NT OrdonezME Otto EC Paloi SPeterson SW Phosri C Roux J SalazarWA Sanchez A Sarria GA Shin HD SilvaBD Silva GA SmithMT Souza-Motta CM Stchigel AM Stoilova-DishevaMMSulzbacher MA Telleria MT Toapanta C Traba JM Valenzuela-Lopez NWatlingR Groenewald JZ 2016 Fungal planet description sheets 400ndash468 Persoonia36316ndash458 DOI 103767003158516X692185

Csardi G 2006 The igraph software package for complex network research InterjournalComplex Systems 1695

Datlof EM Amend AS Earl K Hayward J Morden CWWade R Zahn G Hynson NA2017 Uncovering unseen fungal diversity from plant DNA banks PeerJ 5e3730DOI 107717peerj3730

DaveyML Heegaard E Halvorsen R Kauserud H OhlsonM 2013 Amplicon-pyrosequencing-based detection of compositional shifts in bryophyte-associatedfungal communities along an elevation gradientMolecular Ecology 22368ndash383DOI 101111mec12122

De Vos L Steenkamp ET Martin SH Santana QC Fourie G Van der Merwe NAWingfield MJWingfield BD 2014 Genome-wide macrosynteny among Fusariumspecies in the Gibberella fujikuroi complex revealed by amplified fragment lengthpolymorphisms PLOS ONE 9e114682 DOI 101371journalpone0114682

Drake I White Jr JF Belanger FC 2018 Identification of the fungal endophyte ofAmmophila breviligulata (American beachgrass) as Epichloeuml amarillans PeerJ6e4300 DOI 107717peerj4300

DufreneM Legendre P 1997 Species assemblages and indicator species the needfor a flexible asymmetrical approach Ecological Monographs 67345ndash366DOI 1018900012-9615(1997)067[0345Saaist]20Co2

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

Eiler A Heinrich F Bertilsson S 2011 Coherent dynamics and association networksamong lake bacterioplankton taxa The Isme Journal 6330DOI 101038ismej2011113

Estrada CWcisloWT Van Bael SA 2013 Symbiotic fungi alter plant chemistry thatdiscourages leaf-cutting ants New Phytologist 198241ndash251 DOI 101111nph12140

Eusemann P Schnittler M Nilsson RH Jumpponen A Dahl MBWurth DG BurasAWilmkingM Unterseher M 2016Habitat conditions and phenological treetraits overrule the influence of tree genotype in the needle mycobiome-Piceaglauca system at an arctic treeline ecotone New Phytologist 2111221ndash1231DOI 101111nph13988

Faust K Raes J 2012Microbial interactions from networks to models Nature ReviewsMicrobiology 10538ndash550 DOI 101038nrmicro2832

Qian et al (2018) PeerJ DOI 107717peerj5767 2027

Fokkema NJ 1993 Opportunities and problems of control of foliar pathogens withmicro-organisms Pesticide Science 37411ndash416 DOI 101002ps2780370416

Gardes M Bruns TD 1993 ITS primers with enhanced specificity for basidiomycetesmdashapplication to the identification of mycorrhizae and rustsMolecular Ecology2113ndash118 DOI 101111j1365-294X1993tb00005x

Geml J Morgado LN Semenova-Nelsen TA SchilthuizenM 2017 Changes inrichness and community composition of ectomycorrhizal fungi among altitudinalvegetation types on Mount Kinabalu in Borneo New Phytologist 215454ndash468DOI 101111nph14566

Geml J Pastor N Fernandez L Pacheco S Semenova TA Becerra AGWicaksono CYNouhra ER 2014 Large-scale fungal diversity assessment in the Andean Yungasforests reveals strong community turnover among forest types along an altitudinalgradientMolecular Ecology 232452ndash2472 DOI 101111mec12765

Gomes T Pereira JA Benhadi J Lino-Neto T Baptista P 2018 Endophytic andEpiphytic phyllosphere fungal communities are shaped by different environ-mental factors in a Mediterranean ecosystemMicrobial Ecology 76(3)668ndash679DOI 101007s00248-018-1161-9

Guerreiro MA Brachmann A Begerow D Peršoh D 2018 Transient leaf endophytesare the most active fungi in 1-year-old beech leaf litter Fungal Diversity 89237ndash251DOI 101007s13225-017-0390-4

Hanson CA Fuhrman JA Horner-Devine MC Martiny JB 2012 Beyond biogeographicpatterns processes shaping the microbial landscape Nature Reviews Microbiology10497ndash506 DOI 101038nrmicro2795

Harrison JG Forister ML Parchman TL Koch GW 2016 Vertical stratification ofthe foliar fungal community in the worldrsquos tallest trees American Journal of Botany1032087ndash2095 DOI 103732ajb1600277

Hartley SE Gange AC 2009 Impacts of plant symbiotic fungi on insect herbivoresmutualism in a multitrophic context Annual Review of Entomology 54323ndash342DOI 101146annurevento54110807090614

HeD ShenWJ Eberwein J Zhao Q Ren LJ WuQLL 2017 Diversity and co-occurrence network of soil fungi are more responsive than those of bacteria to shiftsin precipitation seasonality in a subtropical forest Soil Biology and Biochemistry115499ndash510 DOI 101016jsoilbio201709023

Hollaway GJ Evans MLWallwork H Dyson CB McKay AC 2013 Yield loss in cerealscaused by Fusarium culmorum and F pseudograminearum is related to fungalDNA in soil prior to planting rainfall and cereal type Plant Disease 97977ndash982DOI 101094PDIS-09-12-0867-RE

Hubbard CJ BrockMT Van Diepen LT Maignien L Ewers BEWeinig C 2018 Theplant circadian clock influences rhizosphere community structure and function TheIsme Journal 12400ndash410 DOI 101038ismej2017172

Ihrmark K Bodeker IT Cruz-Martinez K Friberg H Kubartova A Schenck JStrid Y Stenlid J Brandstrom-DurlingM Clemmensen KE Lindahl BD 2012New primers to amplify the fungal ITS2 regionmdashevaluation by 454-sequencing

Qian et al (2018) PeerJ DOI 107717peerj5767 2127

of artificial and natural communities FEMS Microbiology Ecology 82666ndash677DOI 101111j1574-6941201201437x

Izuno A Kanzaki M Artchawakom TWachrinrat C Isagi Y 2016 Verticalstructure of phyllosphere fungal communities in a tropical forest in Thai-land uncovered by high-throughput sequencing PLOS ONE 11e0166669DOI 101371journalpone0166669

Jarvis SGWoodward S Taylor AF 2015 Strong altitudinal partitioning in the distri-butions of ectomycorrhizal fungi along a short (300 m) elevation gradient NewPhytologist 2061145ndash1155 DOI 101111nph13315

Jiao S Liu ZS Lin YB Yang J ChenWMWei GH 2016 Bacterial communities inoil contaminated soils biogeography and co-occurrence patterns Soil Biology andBiochemistry 9864ndash73 DOI 101016jsoilbio201604005

Ju F Xia Y Guo FWang Z Zhang T 2014 Taxonomic relatedness shapes bacterialassembly in activated sludge of globally distributed wastewater treatment plantsEnvironmental Microbiology 162421ndash2432 DOI 1011111462-292012355

Jumpponen A Jones KL 2009Massively parallel 454 sequencing indicates hyperdiversefungal communities in temperate Quercus macrocarpa phyllosphere New Phytologist184438ndash448 DOI 101111j1469-8137200902990x

Kay GM Tulloch A Barton PS Cunningham SA Driscoll DA Lindenmayer DB 2018Species co-occurrence networks show reptile community reorganization underagricultural transformation Ecography 41113ndash125 DOI 101111ecog03079

Kembel SWMueller RC 2014 Plant traits and taxonomy drive host associ-ations in tropical phyllosphere fungal communities Botany 92303ndash311DOI 101139cjb-2013-0194

Kemler M Garnas J Wingfield MJ Gryzenhout M Pillay KA Slippers B 2013 Iontorrent PGM as tool for fungal community analysis a case study of Endophytesin Eucalyptus grandis reveals high taxonomic diversity PLOS ONE 8e81718DOI 101371journalpone0081718

Koljalg U Nilsson RH Abarenkov K Tedersoo L Taylor AF BahramM Bates STBruns TD Bengtsson-Palme J Callaghan TM Douglas B Drenkhan T EberhardtU Duenas M Grebenc T Griffith GW HartmannM Kirk PM Kohout P LarssonE Lindahl BD Lucking R Martin MP Matheny PB Nguyen NH Niskanen TOja J Peay KG Peintner U PetersonM Poldmaa K Saag L Saar I Schussler AScott JA Senes C SmithME Suija A Taylor DL Telleria MTWeiss M LarssonKH 2013 Towards a unified paradigm for sequence-based identification of fungiMolecular Ecology 225271ndash5277 DOI 101111mec12481

Lanzen A Epelde L Blanco F Martin I Artetxe U Garbisu C 2016Multi-targetedmetagenetic analysis of the influence of climate and environmental parameters onsoil microbial communities along an elevational gradient Scientific Reports 628257DOI 101038srep28257

Lawes JC Dafforn KA Clark GF BrownMV Johnston EL 2017Multiple stressors insediments impact adjacent hard substrate habitats and across biological domainsScience of the Total Environment 592295ndash305 DOI 101016jscitotenv201703083

Qian et al (2018) PeerJ DOI 107717peerj5767 2227

Legendre P Borcard D Blanchet G Dray S 2010 PCNM PCNM spatial eigenfunctionand principal coordinate analyses R package version 21r82

Li ZWright AG Yang Y Si H Li G 2017 Unique bacteria community compositionand co-occurrence in the milk of different ruminants Scientific Reports 740950DOI 101038srep40950

Liggenstoffer AS Youssef NH Couger MB ElshahedMS 2010 Phylogenetic diversityand community structure of anaerobic gut fungi (phylum Neocallimastigomy-cota) in ruminant and non-ruminant herbivores The Isme Journal 41225DOI 101038ismej201049

Liu L Hart MM Zhang J Cai X Gai J Christie P Li X Klironomos JN 2015 Altitu-dinal distribution patterns of AM fungal assemblages in a Tibetan alpine grasslandFEMS Microbiology Ecology 91fiv078ndashfiv078 DOI 101093femsecfiv078

Looby CI Maltz MR Treseder KK 2016 Belowground responses to elevation in achanging cloud forest Ecology and Evolution 61996ndash2009 DOI 101002ece32025

Lu LH Yin SX Liu X ZhangWM Gu TY Shen QR Qiu HZ 2013 Fungal networks inyield-invigorating and -debilitating soils induced by prolonged potato monocultureSoil Biology amp Biochemistry 65186ndash194 DOI 101016jsoilbio201305025

Ma BWang H DsouzaM Lou J He Y Dai Z Brookes PC Xu J Gilbert JA 2016Geographic patterns of co-occurrence network topological features for soil mi-crobiota at continental scale in eastern China The Isme Journal 101891ndash1901DOI 101038ismej2015261

Madrid H Hernandez-RestrepoM Gene J Cano J Guarro J Silva V 2016 New andinteresting chaetothyrialean fungi from SpainMycological Progress 151179ndash1201DOI 101007s11557-016-1239-z

Magoc T Salzberg SL 2011 FLASH fast length adjustment of short reads to improvegenome assemblies Bioinformatics 272957ndash2963 DOI 101093bioinformaticsbtr507

Mejia LC Rojas EI Maynard Z Van Bael S Arnold AE Hebbar P Samuels GJ RobbinsN Herre EA 2008 Endophytic fungi as biocontrol agents of Theobroma cacaopathogens Biological Control 464ndash14 DOI 101016jbiocontrol200801012

Milici M Deng ZL Tomasch J Decelle J Wos-Oxley MLWang H Jauregui RPlumeier I Giebel HA Badewien THWurst M Pieper DH SimonMWagner-Dobler I 2016 Co-occurrence analysis of microbial taxa in the atlantic ocean revealshigh connectivity in the free-living bacterioplankton Frontiers in Microbiology 7649DOI 103389fmicb201600649

Miyamoto Y Nakano T Hattori M Nara K 2014 The mid-domain effect in ectomyc-orrhizal fungi range overlap along an elevation gradient on Mount Fuji Japan TheIsme Journal 81739ndash1746 DOI 101038ismej201434

Nguyen D Boberg J Ihrmark K Stenstroumlm E Stenlid J 2016a Do foliar fungalcommunities of Norway spruce shift along a tree species diversity gradient in matureEuropean forests Fungal Ecology 2397ndash108 DOI 101016jFuneco201607003

Nguyen NH Song ZW Bates ST Branco S Tedersoo L Menke J Schilling JS KennedyPG 2016b FUNGuild an open annotation tool for parsing fungal community

Qian et al (2018) PeerJ DOI 107717peerj5767 2327

datasets by ecological guild Fungal Ecology 20241ndash248DOI 101016jFuneco201506006

Oksanen J Blanchet FG Friendly M Kindt R Legendre P McGlinn D et al 2016vegan community ecology package R package version 24-1 Available at httpsCRANR-projectorgpackage=vegan

Oono R Rasmussen A Lefevre E 2017 Distance decay relationships in foliar fungalendophytes are driven by rare taxa Environmental Microbiology 192794ndash2805DOI 1011111462-292013799

Osono T Mori A 2003 Colonization of Japanese beech leaves by phyllosphere fungiMycoscience 44437ndash441 DOI 101007S10267-003-0135-Y

Pauchard A Kueffer C Dietz H Daehler CC Alexander J Edwards PJ Arevalo JRCavieres LA Guisan A Haider S Jakobs G McDougall K Millar CI Naylor BJParks CG Rew LJ Seipel T 2009 Ainrsquot no mountain high enough plant invasionsreaching new elevations Frontiers in Ecology and the Environment 7479ndash486DOI 101890080072

Peršoh D 2013 Factors shaping community structure of endophytic fungimdashevidencefrom the Pinus-Viscum-system Fungal Diversity 6055ndash69DOI 101007s13225-013-0225-x

Peršoh D 2015 Plant-associated fungal communities in the light of metarsquoomics FungalDiversity 751ndash25 DOI 101007s13225-015-0334-9

Peršoh D Stolle N Brachmann A Begerow D Rambold G 2018 Fungal guildsare evenly distributed along a vertical spruce forest soil profile while individualfungi show pronounced niche partitioningMycological Progress 17925ndash939DOI 101007s11557-018-1405-6

Petriacq P Stassen JH Ton J 2016 Spore density determines infection strategyby the plant pathogenic fungus Plectosphaerella cucumerina Plant Physiology1702325ndash2339 DOI 101104pp1500551

Phoulivong S Cai L Chen H McKenzie EHC Abdelsalam K Chukeatirote E HydeKD 2010 Colletotrichum gloeosporioides is not a common pathogen on tropicalfruits Fungal Diversity 4433ndash43 DOI 101007s13225-010-0046-0

Proulx SR Promislow DE Phillips PC 2005 Network thinking in ecology and evolu-tion Trends in Ecology amp Evolution 20345ndash353 DOI 101016jtree200504004

Ren C ZhangW Zhong Z Han X Yang G Feng Y Ren G 2018 Differential responsesof soil microbial biomass diversity and compositions to altitudinal gradients dependon plant and soil characteristics Science of the Total Environment 610ndash611750ndash758DOI 101016jscitotenv201708110

RevelleW 2017 psych procedures for personality and psychological research Version= 178 Evanston Northwestern University Available at httpsCRANR-projectorgpackage=psych

Rime T HartmannM Brunner I Widmer F Zeyer J Frey B 2015 Vertical distributionof the soil microbiota along a successional gradient in a glacier forefieldMolecularEcology 241091ndash1108 DOI 101111mec13051

Qian et al (2018) PeerJ DOI 107717peerj5767 2427

Rodriguez RJ Henson J Van Volkenburgh E HoyMWright L Beckwith F Kim Y-ORedman RS 2008 Stress tolerance in plants via habitat-adapted symbiosis The IsmeJournal 2404 DOI 101038ismej2007106

Rodriguez RWhite Jr J Arnold A Redman R 2009 Fungal endophytes diversity andfunctional roles New Phytologist 182314ndash330 DOI 101111j1469-8137200902773x

Sapkota R Knorr K Jorgensen LN OrsquoHanlon KA NicolaisenM 2015Host genotypeis an important determinant of the cereal phyllosphere mycobiome New Phytologist2071134ndash1144 DOI 101111nph13418

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

Segata N Izard J Waldron L Gevers D Miropolsky L Garrett WS Huttenhower C2011Metagenomic biomarker discovery and explanation Genome Biology 12R60DOI 101186gb-2011-12-6-r60

Shen CC LiangWJ Shi Y Lin XG Zhang HYWu X Xie G Chain P Grogan PChu HY 2014 Contrasting elevational diversity patterns between eukaryotic soilmicrobes and plants Ecology 953190ndash3202 DOI 10189014-03101

Siddique AB Khokon AM Unterseher M 2017What do we learn from cultures inthe omics age High-throughput sequencing and cultivation of leaf-inhabitingendophytes from beech (Fagus sylvatica L) revealed complementary communitycomposition but similar correlations with local habitat conditionsMycoKeys201ndash16 DOI 103897mycokeys2011265

Siddique AB Unterseher M 2016 A cost-effective and efficient strategy for Illuminasequencing of fungal communities a case study of beech endophytes identifiedelevation as main explanatory factor for diversity and community compositionFungal Ecology 20175ndash185 DOI 101016jFuneco201512009

Siles JA Margesin R 2016 Abundance and diversity of bacterial archaeal and fungalcommunities along an altitudinal gradient in alpine forest soils what are the drivingfactorsMicrobial Ecology 72207ndash220 DOI 101007s00248-016-0748-2

TaiyunW Viliam S 2017 R package corrplot visualization of a correlation matrix(Version 084) Available at https githubcom taiyun corrplot

Takeda Y Nishimura H Inouye H 1977 Two new iridoid glucosides from Mus-saenda parviflora and Mussaenda shikokiana Phytochemistry 161401ndash1404DOI 101016S0031-9422(00)88791-9

Tedersoo L BahramM Polme S Koljalg U Yorou NSWijesundera R Villarreal RuizL Vasco-Palacios AM Thu PQ Suija A SmithME Sharp C Saluveer E SaittaA Rosas M Riit T Ratkowsky D Pritsch K Poldmaa K PiepenbringM PhosriC PetersonM Parts K Partel K Otsing E Nouhra E Njouonkou AL NilssonRHMorgado LN Mayor J May TWMajuakim L Lodge DJ Lee SS LarssonKH Kohout P Hosaka K Hiiesalu I Henkel TW Harend H Guo LD Greslebin

Qian et al (2018) PeerJ DOI 107717peerj5767 2527

A Grelet G Geml J Gates G DunstanW Dunk C Drenkhan R Dearnaley JDe Kesel A Dang T Chen X Buegger F Brearley FQ Bonito G Anslan S AbellS Abarenkov K 2014 Fungal biogeography Global diversity and geography of soilfungi Science 3461256688 DOI 101126science1256688

Toju H Tanabe AS Ishii HS 2016 Ericaceous plant-fungus network in a harsh alpine-subalpine environmentMolecular Ecology 253242ndash3257 DOI 101111mec13680

Truong C Mujic AB Healy R Kuhar F Furci G Torres D Niskanen T Sandoval-LeivaPA Fernandez N Escobar JM Moretto A Palfner G Pfister D Nouhra E SwenieR Sanchez-Garcia M Matheny PB SmithME 2017How to know the fungicombining field inventories and DNA-barcoding to document fungal diversity NewPhytologist 214913ndash919 DOI 101111nph14509

Unterseher M Siddique AB Brachmann A Persoh D 2016 Diversity and com-position of the leaf mycobiome of beech (Fagus sylvatica) are affected bylocal habitat conditions and leaf biochemistry PLOS ONE 11e0152878DOI 101371journalpone0152878

Vacher C Hampe A Porte AJ Sauer U Compant S Morris CE 2016 The Phyllospheremicrobial Jungle at the Plant-Climate Interface Annual Review of Ecology Evolutionand Systematics Vol 47 471ndash24 DOI 101146annurev-ecolsys-121415-032238

Vorholt JA 2012Microbial life in the phyllosphere Nature Reviews Microbiology10828ndash840 DOI 101038nrmicro2910

Voriskova J Baldrian P 2013 Fungal community on decomposing leaf litter undergoesrapid successional changes The Isme Journal 7477ndash486 DOI 101038ismej2012116

Wainwright BJ Zahn GL Spalding HL Sherwood AR Smith CM Amend AS 2017Fungi associated with mesophotic macroalgae from the lsquoAulsquoau Channel westMaui are differentiated by host and overlap terrestrial communities PeerJ 5e3532DOI 107717peerj3532

Wardle DA Lindahl BD 2014 Ecology Disentangling global soil fungal diversityScience 3461052ndash1053 DOI 101126scienceaaa1185

Whipps JM Hand P Pink D Bending GD 2008 Phyllosphere microbiology withspecial reference to diversity and plant genotype Journal of Applied Microbiology1051744ndash1755 DOI 101111j1365-2672200803906x

White TJ Bruns T Lee S Taylor J 1990 Amplification and direct sequencing of fungalribosomal RNA genes for phylogenetics In PCR protocols a guide to methods andapplications San Diego Academic Press 315ndash322

Widder S Besemer K Singer GA Ceola S Bertuzzo E Quince C SloanWT Rinaldo ABattin TJ 2014 Fluvial network organization imprints on microbial co-occurrencenetworks Proceedings of the National Academy of Sciences of the United States ofAmerica 11112799ndash12804 DOI 101073pnas1411723111

Yang T Sun H Shen C Chu H 2016a Fungal assemblages in different habitats in an Er-manrsquos Birch forest Frontiers in Microbiology 71368 DOI 103389fmicb201601368

Yang TWeisenhorn P Gilbert JA Ni Y Sun R Shi Y Chu H 2016b Carbon constrainsfungal endophyte assemblages along the timberline Environmental Microbiology182455ndash2469 DOI 1011111462-292013153

Qian et al (2018) PeerJ DOI 107717peerj5767 2627

Yao F Yang S Wang ZWang X Ye J Wang X DeBruyn JM Feng X Jiang Y Li H2017Microbial taxa distribution is associated with ecological trophic cascades alongan elevation gradient Frontiers in Microbiology 82071 DOI 103389fmicb201702071

Zhang B Zhang J Liu Y Shi P Wei G 2018a Co-occurrence patterns of soybean rhizo-sphere microbiome at a continental scale Soil Biology and Biochemistry 118178ndash186DOI 101016jsoilbio201712011

Zhang LWuW Lee YK Xie J Zhang H 2018b Spatial heterogeneity and co-occurrenceof mucosal and luminal microbiome across swine intestinal tract Frontiers inMicrobiology 948 DOI 103389fmicb201800048

Zhang Z Geng J Tang X Fan H Xu J Wen X Ma ZS Shi P 2014 Spatial heterogeneityand co-occurrence patterns of human mucosal-associated intestinal microbiota TheIsme Journal 8881ndash893 DOI 101038ismej2013185

Zhang Z Luo L Tan X Kong X Yang J Wang D Zhang D Jin D Liu Y 2018cPumpkin powdery mildew disease severity influences the fungal diversity of thephyllosphere PeerJ 6e4559 DOI 107717peerj4559

Zimmerman NB Vitousek PM 2012 Fungal endophyte communities reflect en-vironmental structuring across a Hawaiian landscape Proceedings of the Na-tional Academy of Sciences of the United States of America 10913022ndash13027DOI 101073pnas1209872109

Qian et al (2018) PeerJ DOI 107717peerj5767 2727