lable at ScienceDirect

European Journal of Soil Biology 64 (2014) 6e14

Contents lists avai

European Journal of Soil Biology

journal homepage: http: / /www.elsevier .com/locate/ejsobi

Original article

Soil microbial community structure and activity along a montaneelevational gradient on the Tibetan Plateau

Meng Xu a, Xiaoliang Li a, Xiaobu Cai b, Jingping Gai a, Xiaolin Li a, Peter Christie a,Junling Zhang a, *

a College of Resources and Environmental Sciences, China Agricultural University, Key Laboratory of Plant-Soil Interactions, Ministry of Education,Beijing 100193, Chinab Tibet Agricultural and Animal Husbandry College, Tibet University, Linzhi 860000, China

a r t i c l e i n f o

Article history:Received 9 April 2014Received in revised form21 June 2014Accepted 27 June 2014Available online 14 July 2014Handling editor: Bryan Griffiths

Keywords:Elevational gradientCarbon source utilizationPhospholipid fatty acids (PLFAs)Soil microbial communityTibetan Plateau

* Corresponding author.E-mail addresses: mengx@cau.edu.cn (M. Xu), ju

caitw21@sohu.com (X. Cai), gaijp@cau.edu.cn (J. Gdrpeterchristie@aol.com (P. Christie), junlingz@cau.ed

http://dx.doi.org/10.1016/j.ejsobi.2014.06.0021164-5563/© 2014 Elsevier Masson SAS. All rights res

a b s t r a c t

Knowledge about the distribution pattern and function of soil microbial communities is essential forunderstanding their vital role in ecosystem functioning. However, little is known about the elevationalpatterns of microbial communities at high altitudes on the Tibetan Plateau, a region that is very sensitiveto global change. We investigated the microbial community composition and functional patterns alongan elevational gradient (3100e4600 m above sea level) on Mount Segrila using phospholipid fatty acids(PLFAs) and community level physiological profiles (CLPP). Soil microbes were abundant even at higheraltitudes and ranged from 1.52 to 3.53 mmol per g organic carbon (OC) by total PLFAs. Soil microbialbiomass (expressed as Total PLFAs) declined at higher elevations, and the highest abundance wasobserved at the low-elevation site. Fungal to bacterial ratio decreased with increasing elevation. While noconsistent elevational pattern was observed for PLFA profiles, richness and diversity of carbon sourceutilization by the microbial community decreased significantly at higher altitudes. Soil microbes athigher altitudes had the potential to metabolize relatively more recalcitrant carbon components whilemicrobes at lower altitudes tended to utilize labile carbon sources. The results were in line with thechemical composition of soil organic matter. Increasing O-alkyl C and decreasing alkyl C content indicateslow decomposition at higher elevations. Variations in structure and activity of microbial communitywere mainly attributable to MAT, pH, and litter C:N. Our results indicate that pH is a major factoraffecting microbial communities, and the impact of soil pH is closely correlated with temperature andvegetation changes along the elevational gradient. As future climate warming may lead to temperatureincreases and an upward migration of vegetation, our results provide fundamental knowledge of soilmicrobial communities in the Tibetan alpine region and may help to predict responses of the below-ground community to global climate change.

© 2014 Elsevier Masson SAS. All rights reserved.

1. Introduction

Understanding of microbial community structure and its func-tion is essential as soil microorganisms play vital roles in regulatingecosystem function and influence a variety of important ecosystemprocesses related to soil organic matter turnover and biogeo-chemical cycling [34]. In recent years the spatial pattern of micro-bial communities in montane regions has attracted great interestwith the advent of molecular techniques. The dramatic

eshi199777@163.com (X. Li),ai), lixl@cau.edu.cn (X. Li),u.cn (J. Zhang).

erved.

environmental gradients over short distances in montane regionsprovide a unique opportunity to assess the effect of high turnover ofaboveground vegetation, local soil conditions and climate regimeson spatial patterning of microbial communities along elevationalgradients [21]. In addition, as the microbial community is moresensitive than plants and animals to impacts of environmentalchange [4] changes in microbial community composition and ac-tivity with altitude may in turn affect the stability of the ecosystemunder climate change.

So far research on elevational patterns of microbial communityor activity in mountain areas has been conducted below 3500 mabove sea level (a.s.l.). Results have shown contrasting elevationalpatterns of plant and bacterial taxon richness and phylogeneticdiversity in the Colorado Rocky Mountains [5] and bacteria in

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e14 7

eastern Peru (200e3450 m a.s.l.) [10]. Other studies have indicatedthat as altitude increases microbial population [20], bacterialpopulation [15] and fungal diversity [27] decrease. In contrast [21],reported increases in fungal population and relative numbers ofGram-negative bacteria with increasing elevation in the AustrianCentral Alps (1500e2530 m a.s.l.). However [29], did not find anyelevational gradient in soil bacterial richness/diversity on ChangbaiMts., China, while a mid-elevation richness/diversity “peak” wasdetected for both bacteria [31] and archaea [32] on Mt. Fuji. Thecontradictory results are attributed to many factors, including bothbiotic factors such as vegetation and interactions with other mi-croorganisms, and abiotic factors such as elevation, mean annualtemperature (MAT) and precipitation (MAP), soil pH, temperature,and nutrient availability. The contribution of different variables tothe microbial community varies with the target microbial groupsinvestigated and is sometimes site-specific [5,8,10,29,39,42]. Incontrast to diversity/composition changes, studies on microbialactivities along the elevational gradient are very limited. Decreasesin microbial activities such as dehydrogenase activity [21] andother enzyme activities and CO2 evolution [28] with increasingaltitude have been reported. The changes of microbial activitieswere mostly correlated with variations of climatic factors and dif-ferences in microbial community composition.

TheTibetanPlateau is the largest andhighest plateau in theworldwith an average altitude of 4000 m a.s.l. The harsh geographic andclimatic conditions contribute to one of the most challenging envi-ronments for plants andmicroorganisms, and the fragile ecosystemis very sensitive and vulnerable to climate change and anthropo-genic perturbation [37]. Understanding the vertical distributionpatterns of microbial communities can provide fundamentalknowledge aboutmicrobesat highelevational areas,whichmight beof great significance forpredictingmicrobial community response toenvironmental change. Recent Geochip based studies showed theimportance of soilmicrobial communities in Nand C cycles in alpinemeadow [41] and Tibetan grasslands [38]. However, little is knownabout the shift in microbial community structure and activitiesalong an alpine climosequence. The present study aimed to inves-tigate the composition and activity of the microbial communityalong an elevational gradient and to explore the driving forcesaffecting microbial community structure and activity. We selectedMount Segrila to investigate the spatial patterns of microbial com-munity structure and function at high altitudes above 3500 m a.s.l.Mount Segrila situated on the southeastern part of the TibetanPlateau, and is representative of the typical montane frigid-temperate forest of southeast Tibet. Compared to previous studiesinTibetanPlateau,meanannualprecipitation in the studysite ishighand distinct vegetation types are observed along the elevationalgradients. We hypothesized that vegetation and soil organic carbonmight be the main variables affecting elevational patterns of soilmicrobial community structure and activity.

2. Materials and methods

2.1. Site description

The study was carried out at Mount Segrila (29�210e29�500 N,94�280e94�510 E) on the southeastern part of the Tibetan Plateau.The mountain is located on the convergence of the east Nyain-qentanglha range and the east Himalaya range and is representa-tive of the typical montane frigid-temperate forest of southeastTibet. The altitude of the peak is 5200 m a.s.l. The climate is humidand frigid with a mean annual temperature (MAT) of �0.73 �C,mean annual precipitation (MAP) of 1134 mm, and annual evapo-ration of 544 mm. Mount Segrila has four distinct vegetation types,namely temperate coniferous and broad-leaved mixed forests

(3000e3500 m), frigid dark coniferous forests (3500e4200 m),alpine subfrigid shrub meadow (4200e4500 m), and alpine frigidmeadow (above 4500 m) [18]. The forest stands between 2700 and4300 m and is characterized by Pinus armandi, Pinus densata, Picealikiangensis var. linzhensis, and Abies georgei var. smithii. Shrubswere abundant at higher altitudes of 4300e4500 m a.s.l., e.g.,Sabina saltuaila, Rhododendron hiale and Rhododendron nying-chiense. Alpine meadow is dominated by Rhododendron bulu, Carexsp., Saussurea spp., Potentilla sp., Polygonum spp. and Poa annua.Vegetation above 4500 m a.s.l. has patchy distribution of Cassiopeselaginodes, Rheum nobile and some Rhodiola spp.

2.2. Soil sampling and soil chemical characterization

We selected five sampling sites in the natural open grasslandsalong the west slope of Mount Segrila at altitudes of 3100e4600 ma.s.l. (Fig. 1). The changes in the dominant plant species along thealtitude gradient are shown in Table 1. Samples of the topsoil(0e15 cm)were collected in July 2011 from the five study sites alongthe elevational gradient, and at each site five sampling quadrats(area approximately 5 � 5 m) were selected. At each quadrat threesoil monoliths (20 � 20 cm) were randomly collected and compos-ited as one replicate and thus pooled toyieldfive composite samplesper altitude. Litters were collected at each sampling quadrat andcomposited into one replicate. The soil was placed in polyethylenebags, stored on ice and transported to the laboratory for furthertreatment. The soil samples were sieved (<2 mm) to remove visiblestones, animals, root fragments and plant material before freeze-drying in a lyophilizer, and were stored at �20 �C prior to lipidextraction. Sub-soil samples for analysis of community level physi-ological profiles were kept at 4 �C. For bulk soil analysis the soilsamples were air-dried and sieved (<2 mm). Soil pH was measuredin 1 M KCl (soil: water ratio of 1:2.5). Soil available phosphorus (AP)was extracted with 0.5 M NaHCO3 (OlseneP). Soil organic carbon(SOC) was measured by wet oxidation followed by titration withferrous ammonium sulfate. The litters were oven-dried at 65 �C,ground and sieved (60 mm). Total carbon (TC) and total nitrogen (TN)of soil and litters were measured with an elemental analyzer(EA1108, Fisons Instruments SpA, Milan, Italy).

2.3. PLFA analysis

The culture-independent method of phospholipid fatty acids(PLFA) was explored to characterize the microbial communitystructure [11]. Lipid extraction and PLFAs were performed accord-ing to [11]. Briefly, 5 g of freeze-dried soil was extracted with 19 mlof a single-phase mix of chloroform: methanol: citrate buffer so-lution (1:2:0.8, v/v/v, pH¼ 4.0) for lipid extraction. After extraction,the collected non-polar phase was fractioned into neutral lipids,glycolipids and phospholipids by sequential elution with chloro-form (6 ml), acetone (6 ml) andmethanol (3 ml), respectively, usingpre-packed silica solid phase extraction columns (500 mg/3 ml,Cleanert™ Silica-SPE, Bonna-Agela Technologies Inc., Wilmington,DE). The phospholipid fraction was then methylated with a meth-anoletoluol (1:1) solution (1 ml) and 0.2 M methanolic KOH (1 ml)to produce fatty acids methyl esters (FAMEs). After addition of fattyacid 19:0 as internal standard, samples were analyzed on an Agilent6850 Series II gas chromatograph (Agilent Technologies Inc., SantaClara, CA) and identified with a microbial identification system,MIDI Sherlock 6.1 (MIDI Inc., Newark, DE).

We used the fatty acid nomenclature described by Ref. [11]. Thetotal sum of extracted PLFAs (total PLFAs, mmol (g OC)�1) was usedto quantify microbial biomass. Gram-positive bacteria were esti-mated by the iso- and anteiso-branched saturated fatty acids (i14:0,i15:0, a15:0, i16:0, i17:0, a17:0). The straight chain fatty acids (14:0,

Fig. 1. Map showing the location of Mount Segrila and the sampling sites.

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e148

15:0, 17:0), the mono-unsaturated fatty acids (16:1u7c and18:1u5c) as well as cyclopropyl (cy17:0 and cy19:0) representedGram-negative bacteria. PLFAs 18:2u6, 9c and 18:1u9c served asfungal biomarkers and 16:1u5c as biomarker for arbuscularmycorrhizal (AM) fungi [23]. The biomass of aerobic and anaerobicbacteria was estimated by 16:1u7c and cy19:0, respectively [35].

2.4. Metabolic diversity

The community level physiological profiles (CLPP) were used toinvestigate the functional diversity of carbon metabolism. Thelatter has been applied successfully inmicrobial community studiesin soils, aquatic and other environmental conditions [13]. The CLPPwere characterized based on carbon source utilization pattern us-ing Biolog EcoPlates™ (Biolog Inc., Hayward, CA). Five replicate

Table 1Climatic conditions and vegetation composition at different study sites along the west slseason temperature; GSP, growing season precipitation; SWC, soil water content.

Site Coordinates Altitude(ma.s.l.)

MAT (�C) GST (�C) GSP (mm) SWC (%) Vegetation type

E1 N 29�34.0200

E 94�30.38903105 4.0 9.4 583.6 23.5 Temperate conife

and deciduous mforests

E2 N 29�33.7070

E 94�34.60203877 �0.6 6.6 949.3 56.0 Frigid dark conife

forests

E3 N 29�36.3080

E 94�36.36504176 �2.4 3.7 751.2 45.1 Frigid dark conife

forests

E4 N 29�37.4430

E 94�37.76504306 �3.2 2.8 776.8 48.1 Alpine subfrigid

and meadow

E5 N 26�36.8530

E 94�38.99004556 �4.6 0.1 795.6 46.8 Alpine frigid mea

samples from each study site were composited as one sample. Soilsuspension was prepared with sterilized 0.85% NaCl solution andthe turbidity of each soil suspension sample was standardized to63% with the same NaCl solution before inoculating into micro-plates. The soil dilution was poured into sterile pipette trays and150 ml was injected into each well of the EcoPlates and then incu-bated at 25 �C. Each Biolog EcoPlate contained different carbonsources that were categorized into six biochemical groups (poly-mers, carbohydrates, carboxylic acids, amino acids, amines andphenolic compounds). As the plate consisted of a triplicate com-bination of 32 response wells (31 wells with sole carbon sourcesand one control well with water), we treated each of these within-plate replicates as if they were plate replicates [7]. A redox tetra-zolium dyewas combined to each substrate. The color developmentinduced by dye reduction after microbial utilization of the

ope of Mount Segrila. Abbreviations: MAT, mean annual temperature; GST, growing

Dominant taxa Understory taxa

rousixed

Pinusarmandi, Pinusdensata,Picealikiangensis var. linzhensis,Salixspp, Quercusauurfolioides

Aster, Cyananthus, Carex,Polygonum, Salix, Carumcarvi,Poaalpina, Saussurea, Primulafarinosa

rous Picealikiangensis var. linzhensis,Abiesgeorgei var. smithii.,Larixgriffithiana, Sinarundinuriasetosa

Salix, Polygonum, Rhododendron,Saussurea, Geranium, Primula,Aster, Cyananthus, Carex, Poa

rous Picealikiangensis var. linzhensis,Abiesgeorgei var. smithii.,Rhododendron nyingchiense

Ranunculus, Polygonum,Rhododendron, Ribes, Potentilla,Spiraea, Primula, Gentiana, Carex

shrub Anemone, Ranunculus, Potentilla,Rhodiola, Primula, Pedicularis,Plantago, Aster, Poaceae

dow Anemone, Ranunculus, Potentilla,Geranium, Viola, Carum, Androsace,Primula, Carduus, Poaceae

Table 2Chemical properties of soil and litter at different study sites along the elevational gradient (mean ± SE). Values within the same column followed by different letters indicatesignificant differences at p < 0.05 level.

Site pH (in KCl) Olsen-P (mg kg�1) SOC (g kg�1) Total C (g kg�1) Total N (g kg�1) Soil C:N Litter C:N

E1 5.4 ± 0.1a 66.4 ± 16.0a 25.5 ± 6.3c 41.2 ± 9.5c 3.4 ± 0.5c 11.4 ± 1.1b 47.4 ± 1.5aE2 4.9 ± 0.1b 45.8 ± 10.7ab 92.4 ± 14.0a 94.4 ± 12.4a 7.3 ± 1.0a 13.2 ± 0.1ab 34.5 ± 3.1bE3 4.4 ± 0.1c 26.7 ± 4.1bc 56.4 ± 7.0b 63.1 ± 6.0bc 5.6 ± 0.6ab 11.3 ± 0.3b 28.2 ± 1.2cE4 4.3 ± 0.1c 16.7 ± 4.9c 63.6 ± 9.8b 67.5 ± 8.8abc 4.8 ± 0.6bc 13.9 ± 0.5a 28.6 ± 0.7cE5 3.9 ± 0.0d 16.6 ± 2.0c 83.3 ± 4.5ab 86.9 ± 8.7ab 6.2 ± 0.5ab 14.0 ± 0.3a 20.9 ± 2.1d

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e14 9

individual carbon substrate in each well was measured as the op-tical density at 590 mm (OD590) every 24 h and thereafter up to144 h with a Biolog plate reader.

Average well color development (AWCD) was calculated ac-cording to [14]; i.e. AWCD ¼ P

(C � R)/n, where C is color devel-opment measured in each treatment well (OD590), R is the opticaldensity in the control well, and n is the number of carbon sub-strates (for EcoPlate™, n ¼ 31). Negative C � R values were set tozero. AWCD values at 96 h were used to assess the ability of the soilmicrobial community to utilize diverse carbon sources included inthe Biolog™ testing protocol [36]. The richness of carbon sourceutilization (S) was the number of wells with C�R values higher than0.25. The ShannoneWiener diversity (H0) was calculated as:H0 ¼ �P

pi � lnpi, where pi denotes the relative color developmentof the ith species at 96 h. For redundancy analysis (RDA), data werefirst transformed by dividing, for each substrate, the difference inoptical density at 120 h, relative to the control well, by the AWCD ofthe plate at 120 h, i.e. (C � R)/AWCD [14].

2.5. Chemical composition of soil organic carbon

The solid-state 13C cross-magic angle spinning nuclear magneticresonance spectroscopy (CP/MS-NMR) was employed to charac-terize the chemical composition of the soil organic matter. Thetechnique has been used to determine the chemical structures ofnatural organic materials and chemical changes during decompo-sition [3]. Numerous studies have shown that the decomposition oforganic material is associated with increases in alkyl C concentra-tions and decreases in O-alkyl C [3]. In addition, alkyl C/O-alkyl Cratio and aromatic C are used as indicators of organic matterdecomposition in forest soils [22]. Five replicate samples weremixed thoroughly to form one composite sample and pre-treatedwith 0.2% HF. The spectra of soil samples were recorded on aBruker AV300 NMR spectrometer (Bruker BioSpin Corporation,Billerica, MA) at a resonance frequency of 75.5MHz using the cross-polarization magic angle spinning technique with a spinning speedof 12 kHz. The recycle delay of the common CPMAS pulse sequencewas set to 3 s. Cross polarization contact time was 1 ms. Indepen-dence of the signal-to-noise ratio, a line broadening between 50and 100 Hz, was used prior to Fourier transformation. The spectrawere divided into four chemical shift regions [3] and the corre-sponding chemical groups divided into alkyl C (�10 to 45 ppm,arising mainly from terminal CH3-groups and CH3 in acetyl groups),O-alkyl C (45e110 ppm, mainly comprising cellulose, hemicelluloseand carbohydrates which originated from deoxygenated carbon ofpolysaccharides), aromatic C (110e160 ppm, representing aromaticcarbonwhich can predominantly be classified as C-atoms of lignin)

Table 3Pearson correlations (r) between soil pH and other environmental properties. Values in btemperature, growing season temperature and growing season precipitation, respectivel

Olsen-P SOC Total C Total N S

r 0.794 �0.348 �0.215 �0.189 �p <0.001 0.089 0.218 0.366

or carbonyl C (160e220 ppm, representing carboxyl groups andprotein derived amide C). The relative contents of the differentchemical shift regions were determined by integration using thespectrometer routine with MestRec v. 4.9.9.6 software (MestrelabResearch Inc., Santiago de Compostela, Spain).

2.6. Statistical analysis

Statistical analysis was performed using SPSS 18.0 for Windows(SPSS Inc., Chicago, IL). Data were calculated as arithmetic meanswith standard deviations. Duncan's test was used for comparisonbetween pairs of means at the 5% level. Correlations between var-iables were calculated using the Pearson correlation coefficient. Theindividual PLFAs and transformed AWCD data were subjected toredundancy analysis (RDA) based on a covariance matrix usingCANOCO 4.5 for Windows (Microcomputer Power Inc., Ithaca, NY).In RDA as a direct gradient analysis, the ordination axes representaggregates of the environmental data that best explain the PLFAand AWCD data.

3. Results

3.1. Climate, litter quality and soil chemical properties

On Mount Segrila, MAT and growing season temperature (GST)but not precipitation showed an elevation trend. MAT and GSTdecreased with increasing altitude (Table 1). Growing season pre-cipitation (GSP) and soil moisture had the highest and lowestvalues at site E2 and E1 respectively, and mean values at other siteswere within a similar range. Litter C:N ratio decreased significantlywith increasing altitude, changing from 47 to 21 from site E1 to siteE5 (Table 2). The soils were acid with pH values ranging from 3.9 atsite E5 to 5.4 at site E1. Soil pH and available phosphorus declinedmarkedly with increasing altitude. SOC, total C and total N did notshow elevational trends but had higher amounts at sites E2 and E5,moderate (site E3 and E4), and lowest amounts at site E1. Soil C:Nratio fell within the range 11.3e14.0, and the ratios at site E4 and E5were higher than those at lower sites. Soil pH was negativelycorrelated with litter C:N (r ¼ �0.917, p < 0.001), and positivelywith MAT (r ¼ 0.911, p < 0.001), GST (r ¼ 0.916, p < 0.001) and soilavailable P (r ¼ 0.794, p < 0.001) (Table 3).

3.2. Microbial structure along the elevation gradient

There were altogether 66 PLFAs detected in our soil samples.Compared to other sampling sites, nearly all individual PLFAs at siteE1 showed significantly higher abundance. Total PLFAs had the

old indicate significant correlations (p < 0.05). MAT, GST, GSP indicated mean annualy.

oil C:N Litter C:N MAT GST GSP

0.380 �0.917 0.911 0.916 �0.3500.061 <0.001 <0.001 <0.001 0.086

Table

4Microbial

communityco

mposition( mmol

PLFA

(gOC)�

1)at

differentstudysites.Values

within

thesameco

lumnfollo

wed

bydifferentlettersindicatesign

ificantdifferencesat

p<0.05

.

Site

TotalPL

FAs

Gram-n

egativeba

cteria

Gram-positiveba

cteria

Fungi

AMF

Funga

lto

bacterialPL

FA

14:0

15:0

17:0

16:1u7c

18:1u5c

cy17

:0cy

19:0

i14:0

i15:0

a15:0

i16:0

i17:0

a17:0

18:2u6c

18:1u9c

16:1u5c

E13.53

a0.07

1a0.04

3a0.03

8a0.31

a0.05

6a0.12

a0.07

0a0.06

3a0.38

a0.22

a0.15

a0.07

2a0.07

5a0.17

a0.16

a0.13

a0.21

abE2

1.52

b0.03

0b0.01

3b0.01

0b0.19

b0.02

0b0.07

b0.05

8ab

0.02

2b0.17

b0.17

b0.07

c0.03

0b0.03

4b0.06

b0.08

b0.08

b0.18

bE3

1.83

b0.04

0b0.01

8b0.01

3b0.13

bc0.01

5b0.05

bc0.04

4ab

0.02

3b0.17

b0.10

b0.11

b0.03

3b0.03

7b0.05

b0.11

b0.05

b0.19

bE4

1.62

b0.03

4b0.01

1b0.00

9b0.09

c0.01

1b0.04

c0.03

7b0.01

8b0.13

b0.07

b0.06

c0.02

3b0.03

4b0.04

b0.06

bc0.04

b0.25

aE5

1.72

b0.02

8b0.01

7b0.00

8b0.11

bc0.02

0b0.04

c0.05

6ab

0.01

6b0.10

b0.07

b0.10

b0.02

6b0.02

3b0.03

b0.01

c0.05

b0.08

c

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e1410

highest value of 3.53 ± 0.33 mmol (g OC)�1 at site E1 and decreasedsignificantly to 1.52e1.83 mmol (g OC)�1 at the sites above E2(Table 4). Similar patterns were observed for bacterial PLFAs. FungalPLFA (18:2u6 and 18:1u9) decreased significantly with increasingaltitudes. The ratio of fungal-to-bacterial PLFAwas lowest at site E5(Table 4). There was no clear segregation in the redundancy anal-ysis of PLFA profiles in relation to elevation gradient (Fig. 2).However, samples at sites E5 and E2 tend to be separated fromother elevations. The altitude, climate parameter (MAT), soil char-acteristics (pH, SOC and C:N) and litter C:N explained approxi-mately 42.6% of the variability in microbial community structure inthe RDA analysis (F ¼ 3.068, p ¼ 0.002; Fig. 2). The first axis waspositively correlated with SOC and altitude, but negatively withlitter C:N, MAT, and soil pH. The second axis was positively corre-lated with soil pH, litter C:N, MAT and negatively with altitude. SOCwas highly correlated with variation in PLFA 18:1u5c, and fungalPLFAs (18:2u6c and 18:1u9c) were positively correlated with soilpH, C:N and litter C:N (Fig. 2).

3.3. Carbon source utilization by the soil microbial community

The temporal changes in utilization of carbon sources expressedas the average well color development (AWCD) followed sigmoidalcurves (Fig. 3a). The carbon source utilizationwas rapid between 24and 96 h and thereafter increased gradually. Samples at lower al-titudes had higher AWCD values than those at higher altitudesacross all incubation times. The AWCD values at 120 h correlatedsignificantly with pH (r ¼ 0.800, p < 0.001), MAT (r ¼ 0.853,p < 0.001), and litter C:N ratio (r ¼ 0.794, p < 0.001). Similar sig-nificant correlations were also observed at other incubation times(data not shown). The richness and ShannoneWiener index ofcarbon source utilization significantly declined at sites E4 and E5where the aboveground vegetation shifted from forests to shrubsand alpine meadow (Fig. 3b).

Microbial communities showed significant differences in carbonsource utilization among soil microbial communities from different

Fig. 2. Redundancy analysis (RDA) of PLFA profiles indicating impact of environmentalvariables on microbial community structure at different study sites along the eleva-tional gradient. Abbreviations: litter C:N, C to N ratio of litter; MAT, mean annualtemperature; SOC, soil organic carbon; soil C:N, C to N ratio of the soil.

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e14 11

sites studied (Fig. 4). Soil samples at lower altitude (sites E1, E2 andE3) utilized relatively greater amount of two polymers (glycogenand a-cyclodextrin), D,L-a-glycerol phosphate, two amino acids (L-phenylalanine and L-threonine) and a-ketobutyric acid (Table 5,Fig. 4). Soil microbial communities at sites E4 and E5 utilized a largeproportion of two polymers (Tween 40 and Tween 80), two car-bohydrates (D-cellobiose and i-erythritol), 4-hydroxybenzoic acidand phenylethylamine. Soil samples at site E4 also utilized rela-tively greater proportions of glucose-1-phosphate, glycyl-L-gluta-mic acid and two carboxylic acids (D-galacturonic acid and itaconicacid), while microbes at site E5 utilized more proportions of a-D-lactose and two amino acids (L-serine and L-arginine). The altitude,climatic parameter (MAT), soil characteristics (pH and soil C:N) andlitter C:N explained approximately 45.8% of the variability in mi-crobial community structure in the RDA analysis (F ¼ 2.349,p ¼ 0.004; Fig. 4). The bacteria (both Gram-negative and Gram-positive) biomass was significantly correlated nearly with all car-bon sources except polymers, while fungal biomass was correlatedwith carbohydrates, carboxylic acids and amines (Table 6). Thepolymers were significantly correlated with fungal to bacterialratio.

Fig. 4. Redundancy analysis (RDA) of AWCD date at 120 h from BiologEco Platesindicating impact of environmental variables on microbial carbon utilization patterns

3.4. Spectra of soil organic carbon composition

Representative 13C NMR spectra showed different spectra of soilorganic carbon at different study sites (Table 7). O-alkyl C had the

Fig. 3. (a) Average well color development (AWCD) and (b) carbon source utilizationrichness and diversity index of the microbial communities in soils from different studysites along the elevational gradient. Data are means ± SE (n ¼ 3). Different lettersdenote significant difference at p < 0.05.

at different study sites along the elevational gradient.

highest portion in soil organic matter among all elevation sites. TheO-alkyl C contents increased with increasing elevation (except siteE2). Alkyl-C was lowest at site E5, followed by site E2 and thehighest value was at E3. The aromatic C contents had lower valuesat higher elevation sites. The carbonyl C contents were also low athigher elevation, intermediate at E2 and higher at sites E1 and E3.The ratio of alkyl C/O-alkyl C had the lowest value at site E5, fol-lowed by sites E2 and E4, and the higher values were at sites E3 andE1. Correlation analysis showed that organic carbon contents (O-alkyl C, aromatic C, carbonyl C and alkyl C/O-Alkyl C) were signif-icantly correlated with soil variables (pH, soil C:N, litter C:N),climate parameters (MAT, GSP, SWC) and microbial (bacteria andfungi) biomass (Table 8). The Alkyl-C contents were significantly

Table 5Carbon substrates most heavily loaded on the first two principal components. Thesevalues correspond to the graphical analysis presented in Fig. 4.

PC 1 PC 2

Carbon sources Loading Carbon sources Loading

Polymers CarbohydratesTween 40 �0.728 Glucose-1-phosphate �0.660Tween 80 �0.549 a-D-lactose 0.715Glycogen 0.951 Amino acidsa-cyclodextrin 0.927 glycyl-L-glutamic acid �0.762

Carbohydrates L-serine 0.692D-cellobiose �0.890 L-arginine 0.643i-erythritol �0.658 Carboxylic acidsD,L-a-glycerol phosphate 0.507 D-galacturonic acid �0.514

Amino acids Itaconic acid �0.509L-phenylalanine 0.840L-threonine 0.632

Carboxylic acidsa-ketobutyric acid 0.649

Phenolic compounds4-hydrox benzoic acid �0.519

AminesPhenylethylamine �0.500

Table 6Pearson correlations (r) between utilization of specific carbon substrate groups and microbial groups.

Microbial groups Polymers Carbohydrates Amino acids Carboxylic acids Phenolic compounds Amines

Gram-positive bacteria ns 0.817** 0.647* 0.650* 0.761** 0.658*Gram-negative bacteria ns 0.802** 0.641* 0.612* 0.740** 0.711**Fungi ns 0.811** ns 0.790** ns 0.725**Fungi-to-bacterial ratio �0.549* ns ns ns ns ns

***p < 0.001; **p < 0.01; *p < 0.05; ns non-significant.

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e1412

correlated with soil C:N, GSP, fungi biomass and fungal to bacterialratio.

Table 8Pearson correlations (r) between SOM functional groups with soil, climate andmicrobial parameters.

Alkyl C O-alkyl C Aromatic C Carbonyl C Alkyl C/O-alkyl C

Soil variablespH ns �0.619** 0.692** 0.613** 0.396*

4. Discussion

Climatic regimes at high-elevation mountain areas are oftenassociated with increasing environmental stress, i.e. low air pres-sure and cold climate, unfavorable nutrient conditions and modi-fied vegetation, which in turn may influence microbialcommunities and activities [20]. Accumulating evidence indicatesthat despite being subject to extreme environmental stress, soils athigh-altitude ecosystems have provided habitat for numerous mi-croorganisms such as cyanobacteria and eukaryotic microalgae[25], bacteria [12] and arbuscular mycorrhizal fungi [18]. In thepresent study microbial biomass indicated soil microbes wereabundant even at higher altitudes on Mount Segrila (Table 4). Thetotal PLFAs of beech forest (site E1) and spruce forest (sites E2 andE3) in our study were similar to the values reported by Ref. [8] inthe Austrian Limestone Alps when sampled in September but werelower than those reported by Ref. [16] in Austria (5.3 and 3.3 mmol(g OC)�1 for beech and spruce forest, respectively) during a springsampling from April to June. The total PLFAs of grassland (site E5)were lower than the range reported by Ref. [11] for Swedishgrasslands (3.1 mmol (g OC)�1, sampled in April) but were muchhigher than that found in the European Alps when sampled inOctober (0.2 mmol (g OC)�1, [40]. The results imply that soil mi-crobes may have no strict elevational limits as long as the envi-ronment provides some organic matter and at least short periodswith water [33]. Microbial biomass appeared to coincide with plantactivity, being relatively large during the plant-growing season[19].

Distinct elevational patterns of soil microbial communities havebeen reported in European alpine areas [20,21], in the RockyMountains in Colorado [5] and on Mt. Halla in South Korea [30].Although vegetation types differ markedly along elevational gra-dients on Mount Segrila, individual soil microbial populationsbased on PLFA portraits did not show a consistent elevationalpattern except that fungal biomass tended to decrease withincreasing elevation (Table 4, Fig. 2). Our results are consistent withthose of [8]. They found no consistent elevational trend in wholemicrobial community size at an elevation range of 900e1900 ma.s.l., whereas fungal biomass declined with increasing elevation.

Table 7Soil organic carbon composition divided by integral of the 13C NMR chemical shiftregions at different study sites along the elevational gradient.

Site Alkyl C O-alkyl C Aromatic C Carbonyl C Alkyl C/O-alkyl C

[�10e45 ppm] [45e110ppm]

[110e160ppm]

[160e220ppm]

E1 21.6 39.7 14.9 23.7 0.54E2 10.6 68.2 13.4 7.7 0.16E3 27.4 42.7 10.4 19.5 0.64E4 22.9 72.0 2.5 2.6 0.32E5 5.2 82.4 7.9 4.5 0.06

The PLFA method is based primarily on the measurement of lipidcontents of specific microbial groups. Since our sampling was car-ried out during the active plant growth period, microbial growthmight be stimulated as plants can provide copious carbon sourcesvia exudates in rhizosphere soil. In addition, higher resolutionmolecular techniques might help to analyze in more detail aboutthe elevational patterns of microbial communities.

The carbon source utilization pattern displayed a decreasingtrend with altitude. Our results suggested that the microbes athigher elevations (sites E4 and E5) showed lower decompositionpotential than those at lower altitudes. In addition, soil microbesshowed the potential to decompose different carbon substratesbetween higher and lower altitudes (Table 5, Fig. 4). The carbonsource substrates at higher altitudes include carbohydrates (e.g. D-cellobiose derived from cellulose hydrolysis and i-erythritol whichis distributed widely among fungi and mosses), and aromaticcompounds (e.g. 4-hydroxybenzoic acid derived from root exudatesand phenylethylamine). Microbes at lower altitudes tended todecompose more labile carbon sources such as D,L-a-glycerolphosphate which is an intermediate product of glycerol degrada-tion, amino acids (e.g. L-phenylalanine and L-threonine) and a-ketobutyric acid derived from amino acid metabolism. Microbialutilization of carbon sources has been shown to change withdifferent types of ecosystems [13] and shifted with ecosystemprocesses, for instance restoration of forests [36] and peatlands [1].Alteration of climatic conditions, vegetation composition and soilphysicochemical conditions results in changes in the quality andquantity of carbon sources input to the soil, and thusmight lead to ashift in carbon utilization patterns. Further analysis of the compo-sition of soil organic matter with solid-state 13C NMR spectroscopyprovided indirect evidence for different carbon utilization atdifferent altitudes (Table 7). The O-alkyl C region decreased asaltitude declined and had the highest intensity at the highestelevation site, and alkyl C/O-Alkyl C decreased greatly at sites E4

Soil C:N �0.446* 0.689** �0.483* �0.687** �0.615**Litter C:N ns �0.690** 0.675** 0.685** 0.480*

Climate parametersMAT ns �0.728** 0.744** 0.750** 0.499*GSP �0.511** 0.634** ns �0.681** �0.645**SWC ns 0.664** �0.403* �0.749** �0.570**

Microbial biomassGram-positivebacteria

ns �0.678** 0.591** 0.741** 0.482*

Gram-negativebacteria

ns �0.562** 0.586** 0.643** ns

Fungi 0.506* �0.762** 0.533* 0.736** 0.616**Fungi-to-bacterial ratio

0.621** ns ns ns 0.426*

***p < 0.001; **p < 0.01; *p < 0.05; ns non-significant.

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e14 13

and E5 where aromatic C and carbonyl C contents were also low(Table 7). The O-alkyl C content usually indicates un-decomposedplant litter [17] and increasing decomposition rates are oftencharacterized by an increase in the functional group alkyl C and adecrease in the group O-alkyl [3]. In the present study, soil C:N ratiodid not change significantly along the elevational gradient whilelitter C:N ratios decreased with altitude (Table 2). As lower litterC:N ratios indicate high litter quality and a high transfer rate of Cfrom litterfall tomineral soils, the decreasing decomposition is thuscorrelated with low temperatures and high precipitation and pHdecrease at high altitudes (Table 8). Similarly, climatic factors(temperature and precipitation) decreased the decomposition rateat high altitude sites and thus led to lower alkyl C and higher O-alkyl C contents on Mountain Lushan in subtropical China [9]. Onthe other hand, one of themajor drawbacks of using EcoPlates is thebiased selection toward fast-growing bacteria and fungi and theabsence of slow-growing fungi that have more affinity for complexcarbon compounds e.g. lignin degradation [24]. Moreover, as theincubation was conducted at a constant temperature (25 �C in thisstudy), the results might have underestimated the microbial ac-tivity at higher altitudes. Compared with subalpine soils dehydro-genase activity in alpine soils was shown to be significantly higherin the low-temperature range [21]. Other methods evaluating mi-crobial functional diversity such as Micro-Resp™ or multiple SIR(substrate-induced respiration) would be worth considering infuture studies.

Soil and vegetation attributes [6,30] and climatic parameterssuch as precipitation [39] contribute significantly to the microbialcommunity structure in alpine ecosystems. Our results are consis-tent with previous studies. In the present study, besides the mar-ginal effect of SOC on microbial community composition, we foundthat soil pH, MAT, and litter C:N explained approximately 42.6% ofthe variability in microbial groups in the RDA analysis (Fig. 2), andthese three attributes had strong correlations with AWCD values.The predominant impact of soil pH on microbial communitycomposition, particularly fungal or bacterial dominance, has beenwell demonstrated for forests [2], arable soils [26] and mountainsurface soils [29,30]. In our study significant change in soil pH (1.9unit) was detected. Soil pH had a great impact on soil microbialcommunity composition (Fig. 2) and carbon utilization patterns(Fig. 4). In addition, soil pH was positively correlated with fungal(r¼ 0.505, p < 0.05) and AMF PLFAs (r¼ 0.428, p< 0.05) and aerobicbacterial PLFA (r ¼ 0.480, p < 0.05). A recent study conducted in analpine grassland in Nyainqentanglha Mountains on the central Ti-betan Plateau also reported that soil pH had major impacts onbacterial communities [39]. Soil pH is a complex parameter that isrelated to parent material, climate, vegetation, and mineralweathering [39]. As the parent material and time of weathering aresimilar for all the sites studied, distinctive vegetation changesmight be themajor cause as suggested by the significant correlationbetween soil pH and litter C:N (Table 3). In addition, as the climatein the study region is humid and the mean annual precipitation ishigh (1134 mm), higher leaching of basic cations may alsocontribute to soil pH change.

Vegetation type affects the quantity and quality of carbon andnutrient availability through litter and exudates [4] and thusdramatically influences structure of bacterial and fungal commu-nity [6,16]. A general conclusion from previous studies is thatvegetation type may indirectly affect bacterial community distri-bution along elevational gradients through altering the soilnutrient availability, mainly the C and N status [6,29]. In our studythe five selected study sites represented four typical vegetationtypes along the elevational gradient. We found that litter C:N(Table 2) and chemical composition of soil organic matter (Table 7)changed significantly along the elevational gradient but not soil

C:N. Carbon availability (both quantity and quality of carbon sour-ces) is thus important in affecting soil microbial communitystructure and activity (Tables 6 and 8). In addition, it is also highlypossible that strong pH change (1.9 units) has overrided the impactof soil C:N. A recent study showed that a very narrow range of pH(3.67e4.95) might cause the absence of soil pH influence on soilbacterial community structure along elevational transects on Mt.Halla, South Korea [30].

Climatic attributes such as MAT, MAP have been shown to havestrong impact in shaping elevational patterns of soil microbialcommunities [30,39]. Our study site is located in the southeast ofthe Tibetan Plateauwhere annual precipitation is much higher thanin other parts of the Tibetan Plateau. As our sampling was con-ducted in July when the rainy season starts, precipitation might notbe as important as in the study of bacterial community of [39] atgrasslands in the Nyainqentanglha Mountains on the central Ti-betan Plateau with much lower precipitation (227e420 mm). AsMATwas closely related to soil pH (Table 3) and litter C:N, it is likelythat climatic conditions indirectly affect soil microbial communitythrough changes in vegetation type and soil pH.

In conclusion, during the active plant growth period in summerwe found that soil microbes were abundant even at high-elevationmountain areas of 3100e4600 m a.s.l. Microbial community atdifferent study sites along the selected elevational gradientexhibited distinctive carbon source utilization patterns and wasconsistent with NMR results of carbon composition analysis. Largevariation in soil pH reveals that soil pH is the predominant factordetermining the structure and activity of microbial communitiesacross the elevational gradient. In addition, climate factors (e.g.MAT) and vegetation type (e.g. litter C:N) might be closely relatedto soil pH and substrate quality. Our results provide strong evidencefor the importance of soil microbial communities in carbon cyclingat alpine mountain regions on the Tibetan Plateau. As climatewarming will likely cause an upward migration of vegetation zonesin alpine areas, comprehensive studies revealing the diversity ofthe whole microbial communities and their function in the region,particularly under future climate change conditions, deservefurther investigation.

Acknowledgments

This work was funded by the National Natural Science Foun-dation of China (31071872, 31272251, 41161043 and the InnovativeGroup Grant 31121062) and European Union within the projectsEcoFINDERS (FP7-264465). We thank two anonymous reviews fortheir comments. We also thank Professor John Klironomos, Uni-versity of British ColumbiaeOkanagan, Canada, for initiating theproject.

References

[1] R. Andersen, L. Grasset, M.N. Thormann, L. Rochefort, A.-J. Francez, Changes inmicrobial community structure and function following Sphagnum peatlandrestoration, Soil. Biol. Biochem. 42 (2010) 291e301.

[2] E. Bååth, T.H. Anderson, Comparison of soil fungal/bacterial ratios in a pHgradient using physiological and PLFA-based techniques, Soil. Biol. Biochem.35 (2003) 955e963.

[3] J.A. Baldock, J.M. Oades, P.N. Nelson, T.M. Skene, A. Golchin, P. Clarke,Assessing the extent of decomposition of natural organic materials usingsolid-state 13C NMR spectroscopy, Aust. J. Soil. Res. 35 (1997) 1061e1083.

[4] R.D. Bardgett, D.A. Wardle, Aboveground-belowground Linkages: Biotic In-teractions, Ecosystem Processes, and Global Change, Oxford University Press,New York, 2010.

[5] J.A. Bryant, C. Lamanna, H. Morlon, A.J. Kerkhoff, B.J. Enquist, J.L. Green, Mi-crobes on mountainsides: contrasting elevational patterns of bacterial andplant diversity, Proc. Natl. Acad. Sci. USA 105 (2008) 11505e11511.

[6] H. Chu, J.D. Neufeld, V.K. Walker, P. Grogan, The influence of vegetation typeon the dominant soil bacteria, archaea, and fungi in a low Arctic tundralandscape, Soil. Sci. Soc. Am. J. 75 (2011) 1756e1765.

M. Xu et al. / European Journal of Soil Biology 64 (2014) 6e1414

[7] A.T. Classen, S.I. Boyle, K.E. Haskins, S.T. Overby, S.C. Hart, Community-levelphysiological profiles of bacteria and fungi: plate type and incubation tem-perature influences on contrasting soils, FEMS Microbiol. Ecol. 44 (2003)319e328.

[8] I. Djukic, F. Zehetner, A. Mentler, M.H. Gerzabek, Microbial communitycomposition and activity in different Alpine vegetation zones, Soil. Biol. Bio-chem. 42 (2010) 155e161.

[9] B. Du, H. Kang, J. Pumpanen, P. Zhu, S. Yin, Q. Zou, Z. Wang, F. Kong, C. Liu, Soilorganic carbon stock and chemical composition along an altitude gradient inthe Lushan Mountain, subtropical China, Ecol. Res. 29 (2014) 433e439.

[10] N. Fierer, C.M. Mccain, P. Meir, M. Zimmermann, J.M. Rapp, M.R. Silman,R. Knight, Microbes do not follow the elevational diversity patterns of plantsand animals, Ecology 92 (2011) 797e804.

[11] A. Frostegård, E. Bååth, The use of phospholipid fatty acid analysis to estimatebacterial and fungal biomass in soil, Biol. Fertil. Soils 22 (1996) 59e65.

[12] P. Gangwar, S.I. Alam, S. Bansod, L. Singh, Bacterial diversity of soil samplesfrom the western Himalayas, India, Can. J. Microbiol. 55 (2009) 564e577.

[13] J.L. Garland, Analysis and interpretation of community-level physiologicalprofiles in microbial ecology, FEMS Microbiol. Ecol. 24 (1997) 289e300.

[14] J.L. Garland, A.L. Mills, Classification and characterization of heterotrophicmicrobial communities on the basis of patterns of community-level sole-carbon-source utilization, Appl. Environ. Microbiol. 57 (1991) 2351e2359.

[15] D.D. Giri, P.N. Shukla, S. Kashyap, P. Singh, A.K. Kashyap, K.D. Pandey, Varia-tion in methanotrophic bacterial population along an altitude gradient at twoslopes in tropical dry deciduous forest, Soil. Biol. Biochem. 39 (2007)2424e2426.

[16] E. Hackl, M. Pfeffer, C. Donat, G. Bachmann, S. Zechmeister-Boltenstern,Composition of the microbial communities in the mineral soil under differenttypes of natural forest, Soil. Biol. Biochem. 37 (2005) 661e671.

[17] Y. Kavdir, H. Ekinci, O. Yüksel, A.R. Mermut, Soil aggregate stability and 13C CP/MAS-NMR assessment of organic matter in soils influenced by forest wildfiresin Çanakkale, Turkey, Geoderma 129 (2005) 219e229.

[18] X. Li, J. Gai, X. Cai, X. Li, P. Christie, F. Zhang, J. Zhang, Molecular diversity ofarbuscular mycorrhizal fungi associated with two co-occurring perennialplant species on a Tibetan altitudinal gradient, Mycorrhiza 24 (2014) 95e107.

[19] D.A. Lipson, S.K. Schmidt, Seasonal changes in an alpine soil bacterial com-munity in the Colorado Rocky Mountains, Appl. Environ. Microbiol. 70 (2004)2867e2879.

[20] X. Ma, T. Chen, G. Zhang, R. Wang, Microbial community structure along analtitude gradient in three different localities, Folia. Microbiol. 49 (2004)105e111.

[21] R. Margesin, M. Jud, D. Tscherko, F. Schinner, Microbial communities andactivities inalpine and subalpine soils, FEMS Microbiol. Ecol. 67 (2009)208e218.

[22] M.A. Mu~noz García, A. Faz Cano, Soil organic matter stocks and quality at highaltitude grasslands of Apolobamba, Bolivia, Catena 94 (2012) 26e35.

[23] P.A. Olsson, Signature fatty acids provide tools for determination of the dis-tribution and interactions of mycorrhizal fungi in soil, FEMS Microbiol. Ecol.29 (1999) 303e310.

[24] J. Preston-Mafham, L. Boddy, P.F. Randerson, Analysis of microbial communityfunctional diversity using sole-carbon-source utilization profiles e a critique,FEMS Microbiol. Ecol. 42 (2002) 1e14.

[25] K. �Reh�akov�a, Z. Chlumsk�a, J. Dole�zal, Soil cyanobacteria and microalgal di-versity in dry mountains of Ladakh, NW Himalaya, as related to site, altitude,and vegetation, Microb. Ecol. 62 (2011) 337e346.

[26] J. Rousk, P.C. Brookes, E. Bååth, Contrasting soil pH effects on fungal andbacterial growth suggest functional redundancy in carbon mineralization,Appl. Environ. Microbiol. 75 (2009) 1589e1596.

[27] F. Schinner, G. Gstraunthaler, Adaptation of microbial activities to the envi-ronmental conditions in alpine soils, Oecologia 50 (1981) 113e116.

[28] F. Schinner, Soil microbial activities and litter decomposition related to alti-tude, Plant Soil. 65 (1982) 87e94.

[29] C. Shen, J. Xiong, H. Zhang, Y. Feng, X. Lin, X. Li, W. Liang, H. Chu, Soil pH drivesthe spatial distribution of bacterial communities along elevation on ChangbaiMountain, Soil. Biol. Biochem. 57 (2013) 204e211.

[30] D. Singh, L. Lee-Cruz, W.-S. Kim, D. Kerfahi, J.-H. Chun, J.M. Adams, Strongelevational trends in soil bacterial community composition on Mt. Halla,South Korea, Soil. Biol. Biochem. 68 (2014) 140e149.

[31] D. Singh, K. Takahashi, M. Kim, J. Chun, J.M. Adams, A hump-backed trend inbacterial diversity with elevation on Mount Fuji, Japan, Microb. Ecol. 63(2012a) 429e437.

[32] D. Singh, K. Takahashi, J.M. Adams, Elevational patterns in archaeal diversityon Mt. Fuji, PLoS One 7 (2012b) e44494, http://dx.doi.org/10.1371/journal.pone.0044494.

[33] L.W. Swan, The Aeolian biome, ecosystems of the earth’s extremes, Bioscience42 (1992) 262e270.

[34] M.G.A. van de Heijden, R.D. Bardgett, N.M. van Straalen, The unseen majority:soil microbes as drivers of plant diversity and productivity in terrestrialecosystems, Ecol. Lett. 11 (2008) 296e310.

[35] R.D. Vestal, D.C. White, Lipid analysis in microbial ecology: quantitative ap-proaches to the studyofmicrobial communities, Bioscience39 (1989) 535e541.

[36] Y. Wang, Z. Ouyang, H. Zheng, X. Wang, F. Chen, J. Zeng, Carbon metabolism ofsoil microbial communities of restored forests in Southern China, J. Soil.Sediment. 11 (2011) 789e799.

[37] H. Yang, Q. Lu, B. Wu, J. Zhang, Y. Lin, Vegetation diversity and its applicationin sandy desert revegetation on Tibetan Plateau, J. Arid. Environ. 65 (2006)619e631.

[38] Y. Yang, L. Wu, Q. Lin, M. Yuan, D. Xu, H. Yu, Y. Hu, J. Duan, X. Li, Z. He, K. Xue,J. van Nostrand, S. Wang, J. Zhou, Responses of the functional structure of soilmicrobial community to livestock grazing in the Tibetan alpine grassland,Glob. Change Biol. 19 (2013) 637e648.

[39] Y. Yuan, G. Si, J. Wang, T. Luo, G. Zhang, Bacterial community in alpinegrasslands along an altitudinal gradient on the Tibetan Plateau, FEMSMicrobiol. Ecol. 87 (2014) 121e132.

[40] V. Zeller, R.D. Bardgett, U. Tappeiner, Site and management effects on soilmicrobial properties of subalpine meadows: a study of land abandonmentalong a north-south gradient in the European Alps, Soil. Biol. Biochem. 33(2001) 639e649.

[41] Y. Zhang, Z. Lu, S. Liu, Y. Yang, Z. He, Z. Ren, J. Zhou, D. Li, Geochip-basedanalysis of microbial communities in alpine meadow soils in the Qinghai-Tibetan plateau, BMC Microbiol. 13 (2013) 72.

[42] L. Zinger, D.P.H. Lejon, F. Baptist, A. Bouasria, S. Aubert, R.A. Geremia, P. Choler,Contrasting diversity patterns of crenarchaeal, bacterial and fungal soilcommunities in an alpine landscape, PLoS One 6 (2011) e19950, http://dx.doi.org/10.1371/journal.pone.0019950.