lable at ScienceDirect

Soil Biology & Biochemistry 81 (2015) 275e281

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Mechanisms of soil acidification reducing bacterial diversity

Ximei Zhang a, b, Wei Liu c, d, Guangming Zhang e, Lin Jiang b, Xingguo Han a, e, *

a State Key Laboratory of Forest and Soil Ecology, Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang 110016, Chinab School of Biology, Georgia Institute of Technology, Atlanta, GA 30332, USAc State Key Laboratory of Systematic and Evolutionary Botany, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, Chinad Key Laboratory of Plant Resources, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, Chinae State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China

a r t i c l e i n f o

Article history:Received 21 May 2014Received in revised form30 October 2014Accepted 5 November 2014Available online 18 November 2014

Keywords:Microbial diversitySoil acidificationEcological filteringEnvironmental filteringEvolutionary adaptationDispersal

* Corresponding author. State Key Laboratory of Forof Applied Ecology, Chinese Academy of Sciences, Shen139 1131 7831.

E-mail address: xghan@ibcas.ac.cn (X. Han).

http://dx.doi.org/10.1016/j.soilbio.2014.11.0040038-0717/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

A central goal in soil microbial ecology research is to identify the biodiversity patterns and reveal theunderlying mechanisms. Long-term soil acidification is known to reduce soil bacterial diversity, but themechanisms responsible for this pattern have not been well explored. Soil acidification may reducebacterial richness through ecological filtering (EF). In contrast, two types of processes may promote themaintenance of bacterial richness: species may adapt to the acidic pressure through evolution, andendemic species already adapted to the acidic pressure can colonize the acidified soils through dispersal.To identify the relative contribution of EF and evolution/dispersal (ED), we collected soils with a pH rangeof 4e7 from different ecosystems, conducted an acidification experiment with a similar pH range in aneutral soil, and proposed a conceptual framework that could distinguish the three potential types ofmechanism (neither EF nor ED operate; EF operates alone; ED counteracts some effect of EF). We foundthat the entire bacterial domain was driven by the third type of mechanism, with ED counteracting about42.4% (95% confidence interval: 32.7e50.4%) effect of EF. Meanwhile, different bacterial phyla/classeswere governed by different types of mechanisms, and the dominant was the third type. Our resultshighlight the importance of both ecological and evolutionary mechanisms for regulating soil bacterialcommunities under environmental changes.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Soil microbial ecology research is still at the descriptive stage(Horner-Devine et al., 2004; Martiny et al., 2006). Most previousstudies usually aimed to reveal various types of biodiversity pat-terns, but the underlyingmechanismswere often not fully explored(Prosser et al., 2007; Hanson et al., 2012). While precipitation andtemperature are the primary drivers of plant diversity acrossdifferent terrestrial ecosystems, soil pH has been found to be a keyfactor shaping bacterial diversity (Fierer and Jackson, 2006; Lauberet al., 2009; Rousk et al., 2010). For example, Fierer and Jackson(2006) found that soil pH alone accounts for more than 70% vari-ation in bacterial diversity across different terrestrial ecosystems,and other ecological factors, such as precipitation and soil nitrogen

est and Soil Ecology, Instituteyang 110016, China. Tel.: þ86

content, account for only a small part of variation. They also foundthat as soil pH increases (from acidic to alkaline), bacterial diversityfirst increases then declines. The relationship between soil pH andbacterial diversity is arguably one of themost important patterns inmicrobial ecology (Fierer and Jackson, 2006; Jones et al., 2009).

It is generally agreed that all terrestrial ecosystems and theirbiological components are originated from the ocean (Nisbet andSleep, 2001; Martin et al., 2008). Because the pH of the primitiveocean was nearly neutral, the original soil bacterial communitiesshould be adapted to the neutral environment. While the soil pH ofsome terrestrial ecosystems remained nearly neutral, that of someothers gradually acidified/alkalized in the long-term ecosystemdevelopment process (Wu, 1994; Fierer and Jackson, 2006). Thisnatural acidification/alkalization process was driven by the in-teractions among climate, organisms and soil, such as by thedecomposition of the plant detritus (Chapin et al., 2011). Theintracellular pH of most microorganisms is generally within one pHunit of being neutral (Madigan et al., 1997), and the acidification/alkalization process has resulted in declined soil bacterial diversity.

X. Zhang et al. / Soil Biology & Biochemistry 81 (2015) 275e281276

However, the ecological/evolutionary mechanisms of soil pHaffecting bacterial diversity have yet been explored.

In this study, we aim to explore themechanisms in the process oflong-term soil acidification reducing bacterial diversity acrossdifferent land ecosystems. Several processes could contribute to therelationship between soil bacterial diversity and the natural acidi-fication pressure. On the one hand, soil acidification will decreasebacterial diversity through ecological filtering (EF), because thespecies not adapted to the acidic pressure will be driven to extinc-tion. On the other hand, some species may adapt to the acidificationpressure through evolutionary processes. Meanwhile, there aresome species already adapted to the acidification pressure in otherhabitats, and they may disperse into the acidified soils and suc-cessfully establish populations (Martiny et al., 2006; Hanson et al.,2012). Overall, both evolution and dispersal (ED) may promote themaintenance of bacterial diversity, counteracting the effect of EF.Therefore, there were three potential types of mechanism fordifferent bacterial groups under the soil acidification pressure: I)neither EF nor ED changes bacterial diversity; II) EF alone reducesbacterial diversity; III) EDcounteracts someeffectof EF. Logically, thebacterial groups driven by the second type of mechanism will bemost vulnerableunder theacidificationpressure, because theyshowno adaptive characteristics. The groups driven by the first and thirdtypes of mechanismwill be relatively resistant to changes in pH.

To identify the relative contribution of EF and ED for various soilbacterial taxonomic groups, we first collected soils with a pH rangeof 4e7 from different terrestrial ecosystems across China. Thisnatural acidification process took place at such a large spatiotem-poral scale that both EF and ED could be responsible for the patternof soil acidification declining bacterial diversity. We also conductedan acidification experiment with a similar pH range from a neutralsoil. This experimental acidification process took place at a smallspatiotemporal scale such that EF alone could be the primary driverof the pattern of bacterial diversity, especially for the diversity ofbacterial OTUs (operational taxonomic units) defined as 16S rRNAgene clusters with larger than 97% sequence similarity(Stackebrandt and Goebel, 1994). The difference between the nat-ural and experimental biodiversity patterns should be primarilycaused by ED. Thus, this conceptual framework allows us todistinguish the three potential types of mechanism by comparingthe patterns observed in the two types of studies. Here we appliedit to identify the mechanism type for the entire bacterial domain aswell as 13 dominant phyla/classes.

2. Materials and methods

2.1. Sampling

To understand the effect of natural soil acidification on bacterialdiversity over a large spatiotemporal scale, we collected soil sam-ples (with pH 4e7) at 7 sites along the latitude of 43�N and at 17sites along the longitude of 116�E in the Mainland of China inAugust 2009 (Fig. S1; Table S1). At each sampling site, we collectedfour soil cores (10 cm depth, 3.5 cm diameter) from four locationsthat were at least 3 m apart. The four soil cores were then thor-oughly mixed.

To examine the effect of soil acidification on bacterial diversityover a small spatiotemporal scale, we performed a ten-year soilacidification experiment at the intersection point of the latitudinaland the longitudinal transects (43�N, 116�E; Fig. S1), with a similarpH gradient (4e7) as observed in the soil samples collected fromthe two transects. The soil at the interaction point was neutral(~6.90). The experimental design has been described elsewhere(Zhang et al., 2011). In brief, the study was conducted in a typicalsteppe ecosystem near the Inner Mongolia Grassland Ecosystem

Research Station in China, which lies between 43�260e44�080 N and116�040e117�050 E at an average elevation of 1200 m. A continentalmiddle temperate semiarid climate dominates the area, and ischaracterized by a cold, dry winter and awarm, moist summer. Theregion is characterized by a dark chest soil. The dominant plantspecies accounting for >80% of the total aboveground plantbiomass in the area are Leymus chinensis (Trin.) Tzvel., Stipa grandisP. Smirn., Agropyron cristatum (L.) Gaertn. and Achnatherum sibir-icum (L.) Keng. The experimental site (400 m� 600m), constructedin 1980, was surrounded by an iron fence to exclude animal grazing.In early July each year from 2000 to 2009, NH4NO3 was addedhomogeneously to plots (5m� 5m)with a 1-m buffer zone at ratesof 0, 1.75, 5.25, 10.5, 17.5, and 28 g N/(m2$yr), respectively. It hasbeen demonstrated that N addition affected soil bacterial com-munities primarily through decreasing soil pH in this steppeecosystem and that the increase in soil nitrogen content only had asmall effect (Zhang et al., 2011, 2013, 2014; Zhang and Han, 2012).Each treatment was replicated in three plots. All 18 plots weredistributed across an area of 55 m � 110 m in a randomized blockdesign. In late August 2009, four soil cores (10 cm depth, 3.5 cmdiameter) were collected at four random locations from each plotand thoroughly mixed.

The pH was measured in 1:2.5 (W/V) suspensions of soil indistilled water. DNA was extracted from 0.5 g of mixed soil usingthe FastDNA SPIN kit for soil (Qbiogene, Carlsbad, CA) according tothe manufacturer's instructions, except that 350 mL of DNA elutionsolution was used to elute the DNA in the tenth procedure insteadof 50 mL. The DNA solution was stored at �20 �C.

2.2. Measurement of bacterial composition and analysis ofpyrosequence data

The method of 454 pyrosequencing was used to measure thebacterial composition of each soil sample. The primers 27F (50-AGAGTT TGA TCC TGG CTC AG-30) and 338R (50-TGC TGC CTC CCG TAGGAG T-30) were used to amplify the fragment of 16S rRNA gene fromsoil DNA. In order to measure bacterial composition of all 42 sam-ples in one run, a unique 10-mer tag for each soil DNA sample wasadded to the 50-end of the primer 338R (Hamady et al., 2008). Each20-ml PCRmixture contained 4 ml FastPfu Buffer (5�; Transgen), 2 mlof 2.5 mM dNTPs, 0.4 ml of each primer (5 mM), 0.8 ml of DNA tem-plate, and 0.4 ml of FastPfu Polymerase (Transgen). The PCR protocolwas 95 �C for 2 min (denature); 25 cycles of 95 �C for 30 s (dena-ture), 55 �C for 30 s (anneal), 72 �C for 30 s (elongate); and 72 �C for5 min. Three PCRs were performed for each sample. The combinedproducts were purified by agarose gel electrophoresis and recov-ered. The recovered products were quantified with PicoGreen usinga TBS-380 Mini-Fluorometer, and equal molar concentrations ofPCR products for each sample were pooled. The pooled productswere sequenced in a Roche 454 Genome Sequencer FLX Titaniumsystem at Shanghai Majorbio Bio-pharm Technology Co., Ltd.

The pyrosequence reads were analyzed using the Mothur soft-ware package (Version 1.19) (Schloss et al., 2009). The reads werefirst assigned to samples according to their tags, and those <150 bpin length or with ambiguous characters were removed. Thechimeric sequences were excluded by the chimera.uchime com-mand with default parameters. The first 150 bp of the remainingreads were aligned to the Silva database (Version 106) (Pruesseet al., 2007), and non-bacterial reads were further removed. Tominimize the influence of unequal sampling on the subsequentlycalculated indexes, 3007 reads were randomly selected from eachsample for analysis. All the 126,294 (3007 � 42) sequences wereclustered into OTUs based on 97% similarity (Stackebrandt andGoebel, 1994). The OTU number of each sample was used torepresent the richness of the entire bacterial domain. A different

X. Zhang et al. / Soil Biology & Biochemistry 81 (2015) 275e281 277

approach was adopted to calculate the OTU richness of the 13dominant bacterial phyla/classes. For each phylum/class, werandomly selected 30 representative sequences from a sample,calculated the OTU number represented by the 30 sequences,repeated this process 1000 times, and used the average OTUnumber to represent the richness. For each of the 14 bacterialtaxonomic groups (including the entire bacterial domain) in bothnatural and experimental soils, we further tested whether therewere linear relationships between OTU richness and soil pH withSPSS software (SPSS 13.0 for WINDOWS).

For each of the 14 bacterial groups, we calculated theBrayeCurtis distance based on the abundance of OTUs to representthe compositional variation among localities (Bray and Curtis,1957). Mantel test was used to analyze whether the latitudinaland longitudinal differences affected the bacterial compositionalvariation (Bonnet and Peer, 2002). None of the 14 taxonomic groupsshowed significant positive response (r > 0 and P < 0.05; Table S2),suggesting that none of themwas dispersal-limited. In other words,bacterial species already adapted to other acidic habitats mighthave colonized these acidified soils through random dispersal.

2.3. Conceptual framework

To differentiate the three types of mechanism (see the thirdparagraph of the introduction), we proposed a conceptualframework. During the long-term development of various eco-systems along the two sampling transects, soils that were origi-nally neutral were found to have been gradually acidified fromwest to east and from north to south (Fig. S1; Table S1) (Wu,1994). This process has taken place slowly over a long time spanof perhaps tens of millions of years (Chapin et al., 2011), implyingthat bacterial diversity might have been driven by both EF and ED.For the ten-year soil acidification experiment, however, the timescale was negligibly small; although genetic changes might takeplace in some rapidly evolving genes, the 16S rRNA gene was soconservative that evolution could not have had a noticeable effecton the diversity of OTUs based on 16S rRNA gene sequences(Stackebrandt and Goebel, 1994). Meanwhile, the soil acidificationexperiment was conducted in a large steppe ecosystem with onlyneutral soils, and the effect of species dispersal from acidic en-vironments in other ecosystems would be presumably small.Since the effect of ED is therefore negligible, changes in bacterialdiversity due to soil acidification could be mainly due to EF.Meanwhile, it seems reasonable to assume that, if EF occurred inthe experimental soils, it would surely have occurred over thelonger time intervals in the natural soils. If EF did not occur in thenatural soils, it would not occur over the shorter time intervals inthe experimental soils.

Table 1The three types of mechanism of EF (ecological filtering) and ED (evolution/dispersal) co

The type of mechanism Cases Experimentalsamples

EF ED

I 1 e e

II 2 e e

II 3 Y e

II 4 Y e

III 5 e e

III 6 Y e

III 7 Y e

I, II and III represent the three types of mechanism (see the details in the third paragradecreases OTU richness. “[” means that ED promotes OTU richness. The size of “Y” and

The three types of mechanismwill generate different patterns ofbacterial OTU diversity vs. soil pH (Table 1; Fig. 1). The first type ofmechanism means that neither EF nor ED operates in the naturalsoils, and thus they will not operate in the experimental soils,either. In this case (case 1), as soil pH decreases, bacterial OTUrichness does not change in either the experimental or the naturalsoils (Fig. 1a1). The second type of mechanism means that only EFoperates in the natural soils, and there are three possible cases(cases 2, 3 and 4; Table 1). In case 2, EF operates in the natural soilsbut not in the experimental soils, leading to the pattern that soilacidification decreases OTU richness in the natural soils but not inthe experimental soils (Fig. 1b1). In case 3, EF operates in the nat-ural soils as well as in the experimental soils, resulting in soilacidification declining OTU richness equally in both soils (Fig. 1b2).In case 4, EF has a larger effect on OTU richness in the natural soilsthan in the experimental soils, so the OTU richness decreases to alarger extent in the natural soils than in the experimental soils(Fig. 1b3). The third type of mechanism means that ED counteractssome effect of EF in the natural soils, and there are three possiblecases (cases 5, 6 and 7; Table 1). In case 5, while ED counteracts allthe effect of EF in the natural soils, EF does not operate in theexperimental soils, leading to the pattern that OTU richness doesnot change in both soils (Fig. 1c1). In case 6, while ED counteractssome effect of EF in the natural soils, EF operates in the experi-mental soils, generating the pattern that soil acidification reducesOTU richness more slowly in the natural soils than in the experi-mental soils (Fig.1c2). In case 7, ED counteracts all effect of EF in thenatural soils and EF operates in the experimental soils. As soil pHdecreases, OTU richness remains constant in the natural soils butdeclines in the experimental soils (Fig. 1c3).

While cases 1 and 5 produce indistinguishable pattern of OTUrichness vs. soil pH (Fig. 1a1 and c1), all the other five cases can bedistinguished according to their patterns (Fig. 1b1, b2, b3, c2 andc3), enabling us to infer the relative contribution of ED and EF indriving OTU richness under the long-term natural acidificationpressures. slopeE and slopeN were used to represent the slope of thelinear relationship between OTU richness and soil pH in theexperimental and natural soils, respectively. In Fig. 1b1, slopeE

should be non-significant, but slopeN should be significant(P < 0.05). Both slopeE and slopeN should be significant in Fig. 1b2,b3 and c2, and slopeN should be equal to, larger than and smallerthan slopeE, respectively. In Fig. 1c3, slopeE should be significant,but slopeN should be non-significant.

We could further develop a quantitative index to calculate therelative contribution of EF and ED in the natural soils for the fivecases. While slopeE represents the intensity of EF in the experi-mental soils, slopeN represents the balance between EF and ED inthe natural soils. Because EF and ED have opposite effects on OTU

-driving bacterial diversity in the experimental and natural soils.

Natural samples Correspondingpattern in Fig. 1

slopeE�slopeN

slopeE

EF ED

e e a1Y e b1 <0Y e b2 ¼0

e b3 <0

Y [ c1Y c2 0e1

Y [ c3 ¼1

ph of the Introduction). “e” means that EF or ED has no effect. “Y” means that EF“[” represent the intensity of these effects.

Fig. 1. The theoretical patterns of bacterial OTU richness vs. soil pH in experimental (black) and natural (red) soils. a1, b1eb3 and c1ec3 correspond to the mechanism type of I, IIand III, respectively (see Table 1). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

X. Zhang et al. / Soil Biology & Biochemistry 81 (2015) 275e281278

richness, we can use (slopeE � slopeN) to represent the intensity ofED. Then, ((slopeE � slopeN)/slopeE) can indicate the relativecontribution of ED and EF. There was a theoretical interval for eachof the five cases (Table 1), e.g., it should be between 0 and 1 in case6 when ED counteracts only part effect of EF. Overall, the intervalshould be �0 and 0e1 for the II and III type of mechanism,respectively (Table 1). In practice, the 95% confidence interval ofthis index could be estimated with Monte-Carlo simulation.

We used this framework to investigate the relative contributionof EF and ED in driving the OTU richness of 14 bacterial taxonomicgroups, namely the entire bacterial domain, six dominant phyla,and seven dominant classes. We also explained the influence ofother ecological factors (e.g. soil heterogeneity, nutrient contentand vegetation type) on the effectiveness of this framework in thediscussion.

2.4. Accession numbers

The sequence reads for all samples have been deposited in theNational Center for Biotechnology Information Sequence ReadsArchive (accession no. SRA057671).

3. Results

Clustering all 126,294 sequences into OTUs with 97% similarityyielded an average of 1258 OTUs for each sample and a total of26,978 OTUs for all 42 samples.

The OTU richness of the entire bacterial domain showed sig-nificant linear relationship with soil pH in both the experimentaland natural soils, and slopeN (108) was smaller than slopeE (188;Fig. 2a). This pattern was similar to the theoretical pattern inFig. 1c2. The value of ((slopeE � slopeN)/slopeE) of the entire bac-terial domain was 0.42, with the 95% confidence interval of0.33e0.50 (Table 2). This result suggests that the third type ofmechanism was operating (Table 1).

The patterns of Acidobacteria phylum and Deltaproteobacteriaclass (Fig. 2b, n) were also similar to the theoretical pattern inFig. 1c2. The OTU richness of four phyla (Actinobacteria, Bacter-oidetes, Chloroflexi and Planctomycetes) and three classes (Acid-imicrobidae, Actinobacteridae and Gammaproteobacteria) showedsignificant linear relationships with soil pH in the experimentalsoils but non-significant relationships in the natural soils (Fig. 2cef,h, i, m), which was similar to the theoretical pattern in Fig. 1c3. Thevalues of ((slopeE � slopeN)/slopeE) for all the nine phyla/classeswere between 0 and 1 (Table 2), indicating that they were governedby the third mechanism type.

The OTU richness of three bacterial taxonomic groups (theentire Proteobacteria phylum, and the Alpha- and Beta-proteobacteria classes) showed non-significant linear relationshipwith soil pH in the experimental soils but significant relationship inthe natural soils (Fig. 2g, k, l). This pattern was similar to thetheoretical pattern in Fig. 1b1. Meanwhile, the values of((slopeE � slopeN)/slopeE) for all of them were negative (Table 2),showing that the second type of mechanism was operating.

The OTU richness of the Rubrobacteridae class showed non-significant linear relationship with soil pH in both the experi-mental and natural soils (Fig. 2j). This pattern was similar to thetheoretical pattern in Fig. 1a1 and c1, so the first or the third type ofmechanism might be operating (Table 2).

4. Discussion

Overall, our results demonstrated that different bacterial taxo-nomic groups were driven by different types of mechanism.Meanwhile, the third type of mechanism, that ED counteracted atleast some effect of EF, was the dominant mechanism. This type ofmechanism influenced the OTU richness of 10 out of the 14 bac-terial taxonomic groups (Table 2).

The entire bacterial domain was driven by the third type ofmechanism. In particular, the value of ((slopeE � slopeN)/slopeE)

Fig. 2. The actual patterns of OTU richness vs. soil pH in experimental (black) and natural (red) soils for 14 bacterial taxonomic groups. (For interpretation of the references to colorin this figure legend, the reader is referred to the web version of this article.)

X. Zhang et al. / Soil Biology & Biochemistry 81 (2015) 275e281 279

Table 2The relative contribution of EF and ED in driving the OTU richness of the 14 bacterial taxonomic groups.

Bacterial group Relativeabundance (%)a

Actual patternin Fig. 2

SlopeE SlopeN slopeE�slopeN

slopeE

(95% confidence interval)

Theoreticalpattern in Fig. 1

Correspondingmechanism type

Domain Bacteria 100 a 188b 108b 0.42 (0.33e0.50) c2 III

Phylum Acidobacteria 10.2 b 3.02b 0.79b 0.74 (0.69e0.78) c2 IIIActinobacteria 38.4 c 0.63b 0.34 0.46 (0.35e0.55) c3 IIIBacteroidetes 3.1 d 6.29b 0.81 0.87 (0.85e0.89) c3 IIIChloroflexi 7.0 e 1.12b 0.47 0.58 (0.50e0.65) c3 IIIPlanctomycetes 3.3 f 1.66b 0.0068 0.996 (0.995e0.997) c3 IIIProteobacteria 23.0 g 0.95 1.66b �0.75 (�1.06~�0.46) b1 II

Class Acidimicrobidae 2.1 h 2.44b 0.89 0.64 (0.57e0.69) c3 IIIActinobacteridae 21.3 i 0.54b 0.23 0.57 (0.49e0.65) c3 IIIRubrobacteridae 12.8 j a1 or c1 I or IIIAlphaproteobacteria 13.8 k 0.38 1.48b �2.88 (�3.55~�2.19) b1 IIBetaproteobacteria 3.3 l 0.86 1.39b �0.62 (�0.91~�0.36) b1 IIGammaproteobacteria 2.5 m 3.73b 0.10 0.973 (0.968e0.978) c3 IIIDeltaproteobacteria 2.9 n 2.08b 0.79b 0.62 (0.56e0.68) c2 III

a Represents themean relative abundance in all the experimental and natural samples. slopeE and slopeN represent the slopes of experimental and natural soils, respectively(see the details in Fig. 2).

b Means that the linear relationship was statistically significant (P < 0.05).

X. Zhang et al. / Soil Biology & Biochemistry 81 (2015) 275e281280

was 0.42 (Table 2), suggesting that nearly a half of the effect of EFwas counteracted by ED. However, this result did not mean that therelative contribution of EF and ED was the same for all the phylawithin the bacterial domain. In fact, the Proteobacteria phylumwasdriven by EF alone (the second mechanism type). Although thephyla of Bacteroidetes and Planctomycetes were also driven by thethird mechanism type, they had large ((slopeE � slopeN)/slopeE)values (0.87 and 0.996, respectively; Table 2), meaning that almostall the effect of EF was counteracted by ED. Similarly, while theentire Proteobacteria phylum was governed by the second type ofmechanism, Gamma- and Delta-proteobacteria was driven by thethird type of mechanism (Table 2).

These bacterial taxonomic groups were driven by different typesof mechanism, suggesting that they had different fitness under thenatural acidification pressure. Unfortunately, soil acidification wasfurther intensified by current acidic rain and N deposition (Clackand Tilman, 2008; Zhang et al., 2011). The groups being driven bythe third type of mechanism were adaptive evolutionarily or theyhad a high dispersal, so they would be resistant to the acidic se-lective pressure. In contrast, the groups being driven by the secondtype of mechanism (e.g. the Alpha- and Beta-proteobacteria) werenot adaptive and would be more vulnerable to acidification (Lauberet al., 2009; Rousk et al., 2010). Betaproteobacteria play an impor-tant role in oxidizing ammonium (Horz et al., 2004; Martiny et al.,2011), and our result implies that this ecosystem function isvulnerable in the acidic soils. In contrast, when the acidified soilswere neutralized by added lime, the groups being driven by thethird type of mechanism would be vulnerable and the groups bythe second type would be favored.

We have identified the relative contribution of EF and ED indriving soil microbial diversity using a simple conceptual frame-work, which we hope would stimulate the investigation of theirrelative contribution to processes of other ecological factors (e.g.precipitation and temperature) driving soil microbial diversity.However, it is important to note that our conceptual frameworkfocuses on acidification as the main driver of soil bacterial diversity,and that other factors might affect the effectiveness of the frame-work. Firstly, the current neutral soil community was used torepresent the primitive neutral soil community originated from theocean, but the long-term ecological and evolutionary processesmight have caused the former to be more diverse than the latter.This diversification process was primarily due to the high hetero-geneity of soil environment compared to the ocean environment

(Rainey and Travisano, 1998; Horner-Devine et al., 2004; Ranjardet al., 2013), and this process would happen in soils with all types ofpH, regardless of being neutral or acidic. Soil heterogeneity wouldincrease the value of the y-intercept in the linear relationship ofbacterial diversity vs. soil pH. However, in this research, we aimedto quantify the effect of ED induced by soil acidification rather thanby soil heterogeneity, and the effect caused by soil acidificationwasexhibited in the slope (rather than the y-intercept) of the linearrelationship of bacterial diversity vs. soil pH (Fig. 1). Therefore, theincreased environmental heterogeneity would not have significantinfluence on our results.

Secondly, although soil pH alone accounted for a large propor-tion of the variation in soil bacterial diversity across differentterrestrial ecosystems and it integrated the effect of many ecolog-ical factors (e.g. nutrient availability, vegetation type and soilmoisture (Fierer and Jackson, 2006)), other factors (e.g. traceelement content) might still have some effect on soil bacterial di-versity. If this is indeed found to be the case, the linear relationshipbetween bacterial diversity and soil pH is likely to be weakened,with soil pH playing a lesser role (in terms of R2) in the linearrelationship. However, the contribution of EF and EDwas quantifiedfrom the slope of bacterial diversity vs. soil pH rather than the R2

value. Thus, these factors would have limited influence on ourresults.

Finally, the relationship between OTU diversity and soil pH inthe experimental soils was assumed to be mainly driven by EF inour conceptual framework. However, some evolution and dispersalmight still have occurred over the ten-year time scale. Meanwhile,there might be rare acid-tolerant bacterial species present in theneutral experimental soils, which would become common underacidification. All these factors would partly counteract the effect ofEF in the experimental soils. In other words, when slopeE was usedto represent the effect of EF in our conceptual framework, the actualeffect of EF and thus the ((slopeE � slopeN)/slopeE) value may havebeen underestimated. However, given the difficulty of quantifyingin-situ evolution and dispersal, our approach provides a firstapproximation of how EF and ED combine to regulate soil bacterialcommunities.

Acknowledgments

We thank Dengyun Wu, Yaxing Xie and Yalin Xie for help insampling; Yongfei Bai, Lixia Zhang and many others for help in

X. Zhang et al. / Soil Biology & Biochemistry 81 (2015) 275e281 281

conducting the soil acidification experiment; Qiuping Hu, Guo Yuand many others in Shanghai Majorbio Bio-pharm Technology Co.,Ltd. for help in pyrosequencing; Professor Frederick Cohan ofWesleyan University, Professor Kostas T. Konstantinidis and Dr. JiaqiTan of Georgia Institute of Technology for making comments on theearly drafts. This work was supported by the National Natural Sci-ence Foundation of China (31300431), the “Strategic PriorityResearch Program” of the Chinese Academy of Sciences(XDB15010404), State Key Laboratory of Forest and Soil Ecology ofChina (Grant No. LFSE2013-15), and National Science Foundation ofUSA (Grant No. DEB-1257858 and DEB-1342754).

Appendix A. Supplementary data

Supplementary data associated with this article can be found inthe online version, at http://dx.doi.org/10.1016/j.soilbio.2014.11.004.

References

Bonnet, E., Peer, Y.V., 2002. zt: a software tool for simple and partial Mantel tests.Journal of Statistical Software 7, 1e12.

Bray, J.R., Curtis, J.T., 1957. An ordination of the upland forest communities ofSouthern Wisconsin. Ecological Monograph 27, 325e349.

Chapin III, F.S., Matson, P.A., Vitousek, P.M., 2011. Principles of Terrestrial EcosystemEcology, second ed. Springer, New York.

Clack, C.M., Tilman, D., 2008. Loss of plant species after chronic low-level nitrogendeposition to prairie grassland. Nature 451, 712e715.

Fierer, N., Jackson, R.B., 2006. The diversity and biogeography of soil bacterialcommunities. Proceedings of the National Academy of Sciences of the UnitedStates of America 103, 626e631.

Hamady, M., Walker, J.J., Harris, J.K., Gold, N.J., Knight, R., 2008. Error-correctingbarcoded primers for pyrosequencing hundreds of samples in multiplex. NatureMethods 5, 235e237.

Hanson, C.A., Fuhrman, J.A., Horner-Devine, M.C., Martiny, J.B., 2012. Beyondbiogeographic patterns: processes shaping the microbial landscape. NatureReviews Microbiology 10, 497e506.

Horner-Devine, M.C., Carney, K.M., Bohannan, B.J.M., 2004. An ecological perspec-tive on bacterial biodiversity. Proceedings of the Royal Society B: BiologicalSciences 271, 113e122.

Horz, H.P., Barbrook, A., Field, C.B., Bohannan, B.J.M., 2004. Ammonia-oxidizingbacteria respond to multifactorial global change. Proceedings of the NationalAcademy of Sciences of the United States of America 101, 15136e15141.

Jones, R.T., Robeson, M.S., Lauber, C.L., Hamady, M., Knight, R., Fierer, N., 2009.A comprehensive survey of soil acidobacterial diversity using pyrosequencingand clone library analyses. The ISME Journal 17, 1e12.

Lauber, C.L., Hamady, M., Knight, R., Fierer, N., 2009. Pyrosequencing-basedassessment soil pH as a predictor of soil bacterial community structure at thecontinental scale. Applied and Environmental Microbiology 75, 5111e5120.

Madigan, M., Martinko, J., Parker, J., 1997. Brock Biology of Microorganisms. PrenticeHall, Upper Saddle River, NJ.

Martin, W., Baross, J., Kelley, D., Russell, M.J., 2008. Hydrothermal vents and theorigin of life. Nature Reviews Microbiology 6, 805e814.

Martiny, J.B., Bohannan, B.J., Brown, J.H., Colwell, R.K., Fuhrman, et al., 2006. Mi-crobial biogeography: putting microorganisms on the map. Nature ReviewsMicrobiology 4, 102e112.

Martiny, J.B., Eisen, J.A., Penn, K., Allision, S.D., Horner-Devine, M.C., 2011. Drivers ofbacterial beta-diversity depend on spatial scale. Proceedings of the NationalAcademy of Sciences of the United States of America 108, 7850e7854.

Nisbet, E.G., Sleep, N.H., 2001. The habitat and nature of early life. Nature 409,1083e1091.

Prosser, J.I., Bohannan, B.J.M., Curtis, T.P., Ellis, R.J., Firestone, M.K., Freckleton, R.P.,Green, J.L., Green, L.E., Killham, K., Lennon, J.J., Osborn, A.M., Solan, M., van derGast, C.J., Young, J.P.W., 2007. The role of ecological theory in microbial ecology.Nature 5, 384e392.

Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., et al., 2007. SILVA: acomprehensive online resource for quality checked and aligned ribosomal RNAsequence data compatible with ARB. Nucleic Acids Research 35, 7188e7196.

Rainey, P.B., Travisano, M., 1998. Adaptive radiation in a heterogeneous environ-ment. Nature 394, 69e72.

Ranjard, L., Dequiedt, S., Pr�evost-Bour�e, N.C., Thioulouse, J., Saby, N.P.A., et al., 2013.Turnover of soil bacterial diversity driven by wide-scale environmental het-erogeneity. Nature Communications 4, 1434.

Rousk, J., Bååth, E., Brookes, P.C., Lauber, C.L., Lozupone, C., et al., 2010. Soil bacterialand fungal communities across a pH gradient in an arable soil. The ISME Journal4, 1340e1351.

Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., et al., 2009. Intro-ducing mothur: open-source, platform-independent, community-supportedsoftware for describing and comparing microbial communities. Applied andEnvironmental Microbiology 75, 7537e7541.

Stackebrandt, E., Goebel, B.M., 1994. Taxonomic note: a place for DNA-DNA reas-sociation and 16S rRNA sequence analysis in the present species definition inbacteriology. International Journal of Systematic Bacteriology 44, 846e849.

Wu, D.S., 1994. The Atlas of Soil Environmental Background Value in the People'sRepublic of China. China Environmental Science Press, Beijing.

Zhang, X.M., Chen, Q.S., Han, X.G., 2013. Soil bacterial communities respond tomowing and nutrient addition in a steppe ecosystem. PLoS ONE 9, e89711.

Zhang, X.M., Han, X.G., 2012. Nitrogen deposition alters soil chemical properties andbacterial communities in the Inner Mongolia grassland. Journal of Environ-mental Sciences 24, 1483e1491.

Zhang, X.M., Liu, W., Bai, Y.F., Zhang, G.M., Han, X.G., 2011. Nitrogen depositionmediates the effects and importance of chance in changing biodiversity. Mo-lecular Ecology 20, 429e438.

Zhang, X.M., Wei, H.W., Chen, Q.S., Han, X.G., 2014. The counteractive effects ofnitrogen addition and watering on soil bacterial communities in a steppeecosystem. Soil Biology and Biochemistry 72, 26e34.