Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Bacterial Community Responses to Soils along a Latitudinal and Vegetation Gradient on the Loess Plateau, China

  • Quanchao Zeng,

    Affiliation College of Natural Resources and Environment, Northwest A&F University, Yangling, Shaanxi, P.R. China

  • Yanghong Dong,

    Affiliation College of Natural Resources and Environment, Northwest A&F University, Yangling, Shaanxi, P.R. China

  • Shaoshan An

    Affiliations College of Natural Resources and Environment, Northwest A&F University, Yangling, Shaanxi, P.R. China, State Key Laboratory of Soil Erosion and Dry Land Farming on Loess Plateau, Northwest A&F University, Yangling, Shaanxi, P.R. China

Bacterial Community Responses to Soils along a Latitudinal and Vegetation Gradient on the Loess Plateau, China

  • Quanchao Zeng, 
  • Yanghong Dong, 
  • Shaoshan An


Soil bacterial communities play an important role in nutrient recycling and storage in terrestrial ecosystems. Loess soils are one of the most important soil resources for maintaining the stability of vegetation ecosystems and are mainly distributed in northwest China. Estimating the distributions and affecting factors of soil bacterial communities associated with various types of vegetation will inform our understanding of the effect of vegetation restoration and climate change on these processes. In this study, we collected soil samples from 15 sites from north to south on the Loess Plateau of China that represent different ecosystem types and analyzed the distributions of soil bacterial communities by high-throughput 454 pyrosequencing. The results showed that the 142444 sequences were grouped into 36816 operational taxonomic units (OTUs) based on 97% similarity. The results of the analysis showed that the dominant taxonomic phyla observed in all samples were Actinobacteria, Proteobacteria, Chloroflexi, Acidobacteria and Planctomycetes. Actinobacteria and Proteobacteria were the two most abundant groups in all samples. The relative abundance of Actinobacteria increased from 14.73% to 40.22% as the ecosystem changed from forest to sandy, while the relative abundance of Proteobacteria decreased from 35.35% to 21.40%. Actinobacteria and Proteobacteria had significant correlations with mean annual precipitation (MAP), pH, and soil moisture and nutrients. MAP was significantly correlated with soil chemical and physical properties. The relative abundance of Actinobacteria, Proteobacteria and Planctomycetes correlated significantly with MAP, suggesting that MAP was a key factor that affected the soil bacterial community composition. However, along with the MAP gradient, Chloroflexi, Bacteroidetes and Cyanobacteria had narrow ranges that did not significantly vary with the soil and environmental factors. Overall, we conclude that the edaphic properties and/or vegetation types are driving bacterial community composition. MAP was a key factor that affects the composition of the soil bacteria on the Loess Plateau of China.


The Loess Plateau in China is one of the most eroded areas in the world. Accordingly, the “Grain-for-Green Program” was implemented on a large scale by the central government beginning in 1999 in which vegetation restoration was implemented in this area to remediate the soil degradation problem. Bacteria are a dominant group of soil organisms [1] and play an essential role in ecosystem recovery [2]. Over the past few decades, an increasing number of researchers have been investigating the bacterial diversity in different soils, particularly during the process of vegetation restoration. Bacteria have strong effects on soil processes, and management and land uses have effects on the soil bacterial diversity [3]. However, there is little information regarding the long-term effects of vegetation restoration on the soil bacterial communities on the Loess Plateau in China. Vegetation types have large direct or indirect effects on soil community composition and diversity. The effects of alterations in the vegetation types on the soil physical and chemical properties have been well-studied. Due to the changes in plant species composition, vegetation types can have significant and long-lasting effects on soil carbon and nutrient contents, soil texture, and pH [46]. The effects of the vegetation types and soil properties affect the groups of soil organisms [7].

Previous studies have focused exclusively on plant and animal taxa, showing that macroorganisms such as trees and animals exhibit biogeographical distributions [8]. However, it is not known whether the microorganisms such as fungi and bacteria vary along a latitudinal gradient. Stegan et al. [9] found that the microbial species were widely distributed and that the microbial community composition was governed by ecological Drift and Selection. Fierer et al. [10] suggested that geographic distance should be the best predictor of genetic divergence among communities. Liu et al. [11] used a high-throughput sequencing method to explore the diversity in the black soils in northeast China. The results suggested that a latitudinal diversity gradient of the bacterial communities might be present in the black agriculture soil zone. So far, similar studies on a large scale in disturbed ecosystems in the Loess Plateau are still limited.

Soil is a complex environment, within which the types of microorganisms are associated with the soil properties, latitude, vegetation types, and other factors. Therefore, understanding the diversity of soil microorganisms along a latitudinal gradient has important ecological significance. The development of high-throughput sequencing technologies such as 454 pyrosequencing currently offers an opportunity to effectively sequence DNA fragments, which substantially improves the researcher’s ability to detect non-dominant microbial communities. Next-generation sequencing technologies have made high-throughput sequencing easy and inexpensive to implement [12].

In this study, we selected five vegetation zones (forest, forest-grass, grass, sand and desert ecosystems) as subjects, which represented the vegetation ecosystems from south to north in the Loess Plateau. Our objectives were to determine the effects of edaphic properties and soil properties on the diversity of soil microbial communities using high-throughput 454 pyrosequencing technology.

Our objectives for this study were (i) to determine the abundance, taxonomic diversity and composition of the bacterial communities across the 600 km geodistance of the Loess Plateau in Shaanxi province; (ii) to directly compare the variability in the bacterial communities across this 600 km geodistance with the variability among soils collected from the wide range of mean annual precipitation from North to South; and (iii) to determine which factors most strongly affect the soil bacterial communities.

Materials and Methods

Experimental design and soil sampling

This study was performed in Yulin City, Jingbian City, Liandaowan Citiy, Ansai City and Fuxian City in the northern Loess Plateau (108°41′40.27′′-109°58′10′′E, 36°03′37.16′′-38°55′9.63′′N), China. These areas repented five vegetation stages after the implementation of “Grain-for-Green Program”. The study areas are characterized by a semi-arid climate and a deeply incised hilly-gully Loess landscape. The slopes are between 0° and 20°. Sites stabilized through vegetation restoration for different vegetation ecosystems provide an ideal opportunity to understand vegetation types in extreme environments.

A sandy ecosystem is located in the city of Jingbian, and a desert ecosystem is located in the city of Yulin. These two regions have similar dominant plant species, including Artemisia ordosica, Hedysarum scoparium and Hedysarum leave and other herb species. The soils in the two ecosystem are sand soil with less vegetation, and the soils in the forest, forest-grass and grass ecosystems are loess-derived. In the forest ecosystem, Platycladus orientalis, Quercus wutaishanica and Robinia pseudoacacia predominate. The grass ecosystem is dominated by Artemisia giraldii, Artemisia vestita and Stipa bungeana. However, the forest-grass ecosystem has trees, shrubs and grasses, including Robinia pseudoacacia, Artemisia vestita and Sophora moorcroftiana.

Along the latitudinal gradient from 36° to 39°, the mean annual temperature (MAT) decreases from 11°C to 8°C and the mean annual precipitation (MAP) decreases from 560 mm to 390 mm. From south to north, there were five types of vegetation ecosystems (forest, forest-grass, grass, sandy and desert ecosystems). The field sites do not contain endangered or protected species and did not require research permission.

In August 2013, when the vegetation community biomass had peaked, 15 sites were selected on the basis of five vegetation ecosystems (forest, forest-grass, grass, sandy and desert ecosystems). These sites were separated by 2 to 2.5 km (Table 1 and Fig 1). At each site, we set up five plots of 20 m×20 m. Plots in both sites were at least 1000 m apart. In each plot, five quadrants (1 m×1 m) were individually chosen, one in each of the four corners and one in the center of the plot. Soil samples were taken at five points in the quadrants of each plot, and soil samples in 0–10 cm soil layer were collected. After carefully removing the surface organic materials and fine roots, each mixed soil sample was divided into two parts. One part of the soil sample was air-dried for the evaluation of the soil properties. The other part was frozen at -80°C (liquid nitrogen) for subsequent high-throughput 454 pyrosequencing analysis. The collected samples were analyzed for physicochemical properties or used for DNA extraction.

Selected soil properties

The soil pH was determined on the air-dried samples (sieved to 1 mm) using a glass electrode meter in a suspension of 1:2.5 soil/water ratio (w/v). Soil moisture was determined gravimetrically in fresh soils at 105°C overnight, and the water content was expressed as a percentage of the dry weight. The soil bulk density (BD) was determined using the auger-hole method [13]. The fumigation-extraction method was used to determine the soil microbial biomass (MBC) [14]. The organic matter (SOM) content was analyzed using dichromate oxidation [15]. Total N (TN) was measured using the Kjeldahl method [16]. Total phosphorus (TP) was measured spectrophotometrically after wet digestion with H2SO4 and HClO4 [17]. Available P (AP) was extracted using the Olsen bicarbonate method. NH4+-N (AN) and NO3-N (NN) were measured using a Seal AutoAnalyzer3.

DNA extraction and sequencing PCR

Soil DNA was extracted from 0.5 g of the mixed soil samples using a Soil DNA kit (OMEGA, Bio-Tek, GA, USA) according to the manufacturer’s instructions. An aliquot of the extracted DNA from each sample was used as a template for amplification. The V1-V3 hypervariable regions of the bacterial 16S rRNA gene were amplified with primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 533R (5′-TTACCGCGGCTGCTGGCAC-3′) containing the A and B sequencing adapters (454 Life Sciences) [11].

The 20 μl PCR reaction system contained 4 μl 5× FastPfu Buffer, 2.5 μl mM dNTPs, 0.8 μl forward and reverse primers (5 μM), 0.4 μl FastPfu polymerase (TransStart Fastpfu DNA Polymerase, TransGen) and 10 ng of the template DNA. The PCR amplification was performed using an ABI GeneAmp 9700 instrument. The PCR conditions used for the amplification of the 16S rRNA gene were as follows: 2 min at 95°C; 25 cycles of 30 seconds at 95°C for denaturation, 30 seconds at 55°C for annealing and 30 seconds at 72°C for extension; and 5 min at 72°C. For each sample, 20 parallel PCRs (including two negative control reactions) were performed, and the PCR products were evaluated by analyzing 2 μl of the product on a 2% agarose gel. Next, all of the PCR products from the same sample were pooled, and PCR products of an approximate size of 400 bp were purified using a PCR Purification kit (Axygen Bio, USA). After purification and quantification using QuantiFluor-ST (Promega, U.S.), a mixture of the amplified products was used for pyrosequencing on a Roche 454 GS FLX+ Titanium platform (Roche 454 Life Sciences, Branford, CT, U.S.) according to standard protocols [18]. Pyrosequencing was carried out on a 454 Life Sciences Genome Sequence FLX (Roche) at the Shanghai Majorbio BioPharm Technology Co., Ltd., Shanghai, China.

Statistical and bioinformatics analysis

Several statistical analyses were performed separately on the soil property datasets using the statistical package for the social sciences (SPSS version 20.0 for Windows), including one-way ANOVA and S-K-N multiple range comparison (P = 0.05). The relationships between soil bacterial composition and the environmental factors were tested using linear regression analyses using SPSS 20.0 for Windows. The relationships between the soil bacterial composition and properties were evaluated using R.

The resulting sequences were processed using Mothur software [19]. Briefly, the raw reads were first assigned to samples according to their tags and then the standard primers and barcodes were trimmed off, after which reads with length less than 200 bp or with ambiguous characters were removed. The chimeric sequences were also excluded by the chimera.uchime command with default parameters. After removing the barcode and primer sequences, the unique sequences were aligned against the reference sequence database (Silva database). The remaining reads were pre-clustered ( and then clustered using uncorrected pairwise algorithm. In addition, Operational taxonomic units (OTUs) were defined as sharing >97% sequence identity using Furthest neighbor method ( An OTU-based analysis was performed to calculate the richness and diversity indices (Ace, Coverage, Chao, Simpson and Shannon) with a cutoff of 3% dissimilarity. Raw sequence data in FASTQ format are accessible from the NCBI SRA study number SRP070625, accession numbers SRX1602057- SRX16072.


Effects of MAP on soil biogeochemical properties

The ecosystem types had significantly different soil properties (Table 2). The soil pH and BD varied from 7.95 to 8.36 and from 0.94 to 1.66 g/cm3, respectively. The soil pH and BD were the lowest in the forest ecosystem and highest in the desert ecosystem. SOM content of the soils differed as a function of the vegetation types. The concentration of SOM in the forest ecosystem were significantly higher than the other ecosystems. TN content in samples from the forest ecosystem was the highest, whereas this value was lowest in the sandy ecosystem. TP content was highest in the forest and forest-grass ecosystems and lowest in the sandy ecosystem. The ratio of C/N in the desert ecosystem was significant higher than that in grass and forest ecosystem, and the ratio of N/P in the forest and forest-grass ecosystems was significant higher than that in other ecosystems. MBC ranged from 552.05 mg/kg in the forest ecosystem to 20.18 mg/kg in the sandy ecosystem, and it consistently decreased as MAP increased. Overall, the forest ecosystem with higher MAP had the highest level of nutrients, while sandy and desert ecosystems had the lowest. Overall, from south to north, along the increasing MAP gradient, soil pH and BD increased and SM, SOM, TN, TP, NN, AN, AP, AK and MBC decreased.

Table 2. Biogeochemical properties of soils under different vegetation types.

Bacterial community composition

Across all of the soil samples, we obtained a total of 36816 OTUs with 1593–2829 OTUs per sample (mean 2454). In the grass ecosystem, Shannon’s diversity index, Chao index and Ace index were highest, and the Simpson diversity index was lowest. In contrast, the Simpson diversity index was the highest in the sandy ecosystem (Table 3). Vegetation ecosystems had no significant effect on Shannon’s diversity index, Chao index and Ace index (P>0.05).

Table 3. Characteristics of soil bacteria richness and diversity indices under different vegetation types (Means ± SD, n = 3).

As shown in Fig 2A, at the phylum level, the structures of the microbial communities differed in terms of both the predominant phylum and the relative abundance of each phylum. Actinobacteria (12.66–42.96%), Proteobacteria (17.45–40.33%), Chloroflexi (9.25–16.25%), Acidobacteria (7.10–20.21%), Planctomycetes (4.70–10.01%), Bacteroidetes (1.52–4.47%), Gemmatimonadetes (1.17–6.29%), Armatimonadetes (1.17–6.29%) and Cyanobacteria (0.32–2.98%) were the 9 most abundant phyla. In all samples, Actinobacteria and Proteobacteria were the most abundant phyla, which together accounted for 46.89–67.23% of all bacterial sequences obtained from all of the soil samples. The relative abundance of Actinobacteria was much higher in the sandy ecosystem than in other ecosystems. Proteobacteria were the second highest bacterial abundance found in all samples. The relative abundance of these bacteria was highest (35.35%) in the forest ecosystem and lowest in the sandy ecosystem (21.40%). The ANOVA test indicated that Actinobacteria, Proteobacteria and Planctomycetes differed significantly under different ecosystems (P<0.05) and that Planctomycetes were significantly more abundant in forest and grass ecosystem soils (P<0.05). Compared with the Actinobacteria and Proteobacteria, Chloroflexi, Plantctomycetes had narrow ranges, which accounted for 10.90–13.86%, 2.05–2.88% and 0.34–1.65%, respectively (Fig 2C). And for the three phyla, there was no significant differences between different vegetation ecosystems (P>0.05).

Fig 2. Relative abundance in the bacterial community composition at the phylum level (A and C) and at the class level (B and D) within the different ecosystems.

Values designated by different lowercase letters showed significant differences under different vegetation ecosystems in the same phyla, the same as followings (P < 0.05).

At the class level, Actinobacteria, Alphaproteobacteria, Acidobacteria, Chloroflexia and Betaproteobacteria were dominant communities. Vegetation ecosystems had significant effects on soil bacterial communities (P<0.05). The relative abundance of Actinobacteria for sandy ecosystem was significant higher than grass, forest-grass and forest systems (Fig 2B). No significant differences for Alphaprotebacteria, Acidobacteria, Ktedonobacteria, Anaerolineae and Gemmatimonadetes were detected in different ecosystems.

Of all of the edaphic and climate variables examined, only altitude and MAT had no significant association with soil bacterial communities (Table 4). There was a significant negative correlation of Actinobacteria with SM, MAP, SOM, TN, TP, NN, AN, AP, AK, MBC, C/P and N/P and a significant positive correlation between pH and BD. However, Proteobacteria had the opposite correlations with the edaphic and climate variables. Meanwhile, there were no significant correlations between Chloroflex, Bacteroidetes and Cyanobacteria and edaphic and climate variables (Table 4).

Table 4. The Pearson correlations between the relative abundances of the 9 dominant bacterial phyla and environment factors.

When the edaphic and climate variables were used to constrain the ordination of the soil bacterial (at the phylum level) with the RDA (Fig 3), there were distinct differences in bacterial community compositions under different ecosystems. We found that MAP, total P, SOM, TN, SM and C/P were the key factors that affected the soil bacterial compositions. The RDA analysis showed that there were clear soil bacteria community differences between sample sites, with desert and sandy ecosystems clusters differing from the grass and forest vegetation ecosystems (Fig 3). In desert and sandy ecosystems, Actinobacteria was the dominant phylum and affected by soil pH. In Grass and forest-grass ecosystems, Proteobacteria were the most abundant and the soil had the highest soil nutrients and MAP.

Fig 3. Ordination plots of the results from the RDA to examine the relationship between bacterial populations (shown in the blue) and selected environment factors (shown in red), including pH, TP, TN, MAP, MAT, MBC, SOM, SM and C/P in the five different ecosystems.

Site 1, 2 and 3 were in the desert ecosystem; Site 4, 5 and 6 were in the sandy ecosystem; site 7, 8 and 9 were in the grass ecosystem; site 10, 11 and 12 were in the forest- grass ecosystem; site 13, 14 and 15 were in the forest ecosystem.


Soil properties affecting soil bacteria compositions under different vegetation ecosystems

Soil chemical properties, including total C, total N, total P, and available N and K, decreased from south to north in the soils of the Loess Plateau in northwest China (Table 2). The changes in soil moisture demonstrated the same trend. However, soil pH and BD had the opposite trends, that is, they increased with increases in the MAP gradient. MAP was significantly associated with soil pH, BD, SM, SOM, TN, TP, NN, AN, AK, MBC, C/N, C/P and N/P, suggesting that the MAP diversity gradient was the key factor affecting the chemical and physical properties of the soil. This finding was similar to the biogeographic distribution of black soil properties from north to south in the northwest China [20].

Soil pH and water content also correlated with the bacterial community patterns. Although all of the samples were alkaline, the pH of the soils from the sandy and desert ecosystems was higher. The relative abundances of major bacterial classes in each soil type are consistent with predictions based on this difference in soil pH [21] in that there was a significant negative correlation of soil pH with Proteobacteria and a significant positive correlation with Actinobacteria. Acidobacteria were also sensitive to soil pH [22]. Our results showed that the strong influence of pH was also observed with respect to the major phyla as indicated by the observation that the relative abundances of Actinobacteria, Proteobacteria and Armatimonadetes across all sites changed significantly along the soil pH gradient (Table 4). Other environmental factors, such as temperature, might also be determinants of the soil bacterial community across this climate gradient in northwest China. Soil types also reflected differences in the biochemical composition of the litter, in terms of structural organic carbon and mineral nutrients, but the factors that are the best predictors of the composition and structure of soil bacterial communities remain to be identified [23]. In addition, microbial communities have been reported to respond to alterations in air temperature and humidity [2425].

Sowerby et al. [26] observed a variation in the soil microbial extracellular enzymes in heath land ecosystems across geographical and climatic gradients from northern to southern Europe. With respect to the RDA, the soil moisture, SOM, TP, TN and C/P correlated significantly with soil bacterial communities (Fig 3). In this study, significant correlations were observed among of soil SOM, TN and TP. We therefore discuss only the relationship between the soil TP and composition of the bacterial communities in this paper. Soil microorganisms are the key regulators of the biogeochemical phosphorus cycle. Liu et al. [27] found that addition of P significantly increased the microbial biomass and altered the microbial community composition in old-growth forests, but there was no effect in pine forests. Cleveland et al. [28] found that the microbial activity was limited by P availability in tropical rain forests. Ehlers et al. [29] also observed a P limitation of the microbial growth after a combined addition of C and N to a native soil from western Kenya. Wang et al. [30] showed that P availability is vital to a priming effect in the subtropical forest soils and that P addition alone decreased soil respiration by 7.1%. In temperate forest ecosystems, the available pools of P strongly correlate with the microbial composition, and these authors found that from a microbial perspective soil, pH might be less important overall than P availability [31]. In our study, TP was significantly correlated with the relative abundance of Actinobacteria, Proteobacteria, Planctomycetes and Gemmatimonadetes. Thus, our results further emphasized that soil total P plays a key role in changing soil bacterial community compositions in the Loess Plateau.

Soil moisture was another of the most important factors that affected the bacterial communities. Soil moisture has strong effects on the soil properties and vegetation growth [32], consistent with our results. Soil moisture was the limiting factor for the microbial activities of 10 major phyla. Specifically, the relative abundances of Actinobacteria, Proteobacteria, Acidobacteria, Planctomycetes and Armatimonadetes were significantly correlated with soil moisture, especially for Actinobacteria and Proteobacteria. Soil moisture improves the microbes’ access to nutrients and also enhances their motility [33]. Soil moisture changed belowground C allocation. Drought reduced belowground C allocation and weakened the link between plant and bacterial. As a consequence, the more favorable soil moisture conditions in the soils could have allowed the microbes to maintain their activities, so soil moisture was a key factor for the soil microbes.

Bacterial community composition in the soils of different vegetation ecosystems

Across all soil samples, the dominant groups (relative abundance >5%) in the Loess Plateau were Actinobacteria, Proteobacteria, Chloroflexi, Acidobacteria and Planctomycetes, consistent with previous studies on the soils of the Tibetan Plateau [34] and the black soils of northeast China [35]. Lauber et al. [36] found that Acidobacteria, Actinobacteria, Proteobacteria and Bacteroidetes are the dominant phyla in soils from North to South America [36], a result that is consistent with the observations of 29 soils from the Arctic region [37].

To the best of our knowledge, this study is the first to use a pyrosequencing approach to provide information concerning the bacterial diversity (richness and composition) in soils affected by various vegetation types after restoration in the Loess Plateau of China. Using this approach, the differences in the bacterial diversity among the five ecosystems were studied. Bacteria are the most abundant and diverse group of soil organisms [3839]. Despite the key role of bacteria in soil processes, there is still a lack of information regarding the effects of restoration of vegetation ecosystems on bacterial diversity. We obtained OTUs to generate predictive rarefaction models and found differences in the richness (diversity) indices Ace and Chao1 as affected by the vegetation ecosystems evaluated. Generally, the soils of the grass ecosystem showed the highest diversity indices (Ace and Chao1) compared with other ecosystems.

Compared with the forest-grass, grass, sandy and desert ecosystems, the differences observed in bacterial diversity of the soils studied under the forest ecosystem is of ecological significance because this system takes more time to recover, and the soil has more nutrients and higher soil moisture. Due to the more diverse types of litter and rooting systems, the soils in the forest ecosystem have a higher biomass and soil total N, soil total P and soil total organic matter. In addition, our data are consistent with Chim et al. [39], who reported that the effect of the quality of soil organic materials associated with distinct vegetation types (as indicated by the C/N ratio) was stronger than their quantity effects on the structure of the soil bacterial community. The carbon and nutrient levels of the soils of different systems also varied considerably, and pronounced differences among the bacterial community structures among the vegetation types were observed.

In addition to fungal degradation of structural organic carbon, bacteria including Actinobacteria can solubilize complex carbon substrates such as lignin [40]. However, much less is known with respect to the degradation of structural organic carbon by Acidobacteria because few cultured representatives of this phylum are available for physiological studies. Recent Acidobacterial isolates from peat land soils are capable of degrading cellulose, albeit at a slow rate [41]. The plant litter quality among forest, shrub and grass ecosystems also differs markedly. Lignin is the main component of the organic carbon of tree wood debris, but grasses consist mostly of cellulose. In general, lignin has a higher C/N ratio than cellulose, which is consistent with the C/N ratios observed among the five vegetation systems on the Loess Plateau of China. Thus, forest, shrub and grass vary considerably in their substrate composition, as do the microbial communities that are involved in the degradation of the respective types of litter [42]. The roots of trees and grasses are additional major sources of organic matter to the soil, as well as sites of plant and soil microorganism interactions. Root exudates affect the soil bacterial community structure [43], and their quality and quantity may differ in different vegetation ecosystems. A correlation has been found between soil bacterial communities and carbon and nutrient contents, consistent with the findings obtained by Chan et al. [44]. In this study, the carbon and nitrogen levels of the forest ecosystem soils were the highest and could provide enough nutrients for the growth of bacteria.

Bacterial community composition is affected by MAP

Many authors have reported that fungal diversity varies with the latitude gradient [4547], and soil bacteria may show the same trend. In this study, we documented the overall pattern of the soil bacterial communities along an increasing latitudinal MAP from south to north on the Loess Plateau. Our results showed that soil properties had significant correlations with MAP, and soil bacterial community compositions correlated closely with the geographical location and the changes in MAP (Fig 4). The Actinobacteria phylum was poorly represented in the forest ecosystem with high MAP but predominated in the sandy ecosystem MAP. The Proteobacteria were the most abundant phylum in the forest ecosystem, and their abundance was lowest in the sandy ecosystem. Along the increasing MAP gradient, Actinobacteria and Armatimonadetes significantly increased, but Proteobacteria, and Planctomycetes significantly decreased. However, the relative abundances of Chloroflexi, Acidobacteria, Bacteroidetes, Gemmatimonadetes and Cyanobacteria were not significantly correlated with MAP (P>0.05). Hillebrand and Azovsky [48] hypothesized that the strength of the latitudinal gradient was positively correlated with body size. In contrast, Fuhrman et al. [49] reported that the diversity of the marine bacterial community decreased with increases in latitude. Fierer and Jackson [50] used a ribosomal DNA-fingerprinting method to investigate the relationship between the soil bacterial communities and conclude that edaphic variables (namely pH) mainly control bacterial biogeography. This is exactly in line with what this study finds. These results were consistent with those of Vries et al. [51]. The observed shift in life strategy between communities with a different precipitation history was accompanied by changes in the relative abundance of bacterial species which employed different strategies [52]. Microbial community composition and function are both sensitive to changes in rainfall [52]. Drought affects plants and soil microorganisms, especially for bacteria, not to fungi [53]. Variation in soil bacterial communities was explained by abiotic factors like climate (MAP), pH and soil properties.

Fig 4. The relationship between latitude and the relative abundance within the soil bacterial community compositions (8 dominant phyla).


In this study, we analyzed the soil bacteria as a function of a MAP gradient on the Loess Plateau. Across all samples, the dominant groups were Actinobacteria, Proteobacteria, Chloroflexi, Acidobacteria, Planctomycetes, Bacteroidetes, Gemmatimonadetes, Armatimonadetes and Cyanobacteria. Actinobacteria and Proteobacteria dominated the five vegetation ecosystems (>14%) and were significantly correlated with soil moisture, C/P, MAP, and total P. Vegetation ecosystems had significant effect on soil bacteria composition. Ours is the first study to document the overall pattern of soil bacteria diversity along a MAP gradient on the Loess Plateau, and our results indicate that soil bacteria (Actinobacteria, Proteobacteria and Planctomycetes) follow the MAP diversity pattern of macroorganisms. Meanwhile, significant correlations between the relative abundance of soil bacteria (Actinobacteria and Proteobacteria) and pH were found in this study, suggesting that pH strongly influences soil bacterial communities. Furthermore, our data provide a starting point for establishing soil bacterial databases in the Loess Plateau, as well as for the plants associated with the vegetation restoration. These databases would provide a basis against which the impacts of future climate changes and ecosystem management practices can be assessed.

Supporting Information

S1 Table. Relevant data underlying the findings described in manuscript.



This study was supported by the National Natural Sciences Foundation of China (41171226, 41101254), the Program for New Century Excellent Talents in University (NCET-12-0479) and the Demonstration of Key Technologies for Soil and Water Conservation and Restoring and Eco-restructing for Damaged Regional Ecosystem of the Shanxi-Shaanxi-Inner Mongolia energy source base (KZCX2-XB3-13).

Author Contributions

Conceived and designed the experiments: SA. Performed the experiments: QZ YD. Analyzed the data: QZ YD. Contributed reagents/materials/analysis tools: QZ YD. Wrote the paper: SA QZ YD.


  1. 1. Gans J, Wolinsky M, Dunbar J. Computational improvements reveal great bacterial diversity and high metal toxicity in soil. Science. 2005; 309(5739):1387–90. pmid:16123304
  2. 2. Hart SC, DeLuca TH, Newman GS, MacKenzie MD, Boyle SI. Post-fire vegetative dynamics as drivers of microbial community structure and function in forest soils. Forest Ecol Manag. 2005; 220(1):166–84.
  3. 3. Acosta-Martínez V, Dowd S, Sun Y, Allen V. Tag-encoded pyrosequencing analysis of bacterial diversity in a single soil type as affected by management and land use. Soil Biol Biochem. 2008; 40(11):2762–70.
  4. 4. Deng L, Liu G-B, Shangguan Z-P. Land use conversion and changing soil carbon stocks in China's ‘Grain-for-Green’ Program: a synthesis. Global Change Biology. 2014; 20(11): 3544–3556. pmid:24357470
  5. 5. Clemmensen K, Bahr A, Ovaskainen O, Dahlberg A, Ekblad A, Wallander H, et al. Roots and associated fungi drive long-term carbon sequestration in boreal forest. Science. 2013; 339(6127):1615–8. pmid:23539604
  6. 6. Murty D, Kirschbaum MU, Mcmurtrie RE, Mcgilvray H. Does conversion of forest to agricultural land change soil carbon and nitrogen? A review of the literature. Global Change Biology. 2002; 8(2):105–23.
  7. 7. Nielsen UN, Osler GH, Campbell CD, Burslem DF, van der Wal R. The influence of vegetation type, soil properties and precipitation on the composition of soil mite and microbial communities at the landscape scale. Journal of biogeography. 2010; 37(7):1317–28.
  8. 8. Levin SA. The problem of pattern and scale in ecology: Springer; 1995.
  9. 9. Stegen JC, Lin X, Fredrickson JK, Chen X, Kennedy DW, Murray CJ, et al. Quantifying community assembly processes and identifying features that impose them. The ISME journal. 2013; 7(11):2069–79. pmid:23739053
  10. 10. Fierer N. Microbial biogeography: patterns in microbial diversity across space and time. Accessing uncultivated microorganisms: from the environment to organisms and genomes and back. 2008:95–115.
  11. 11. Liu J, Sui Y, Yu Z, Shi Y, Chu H, Jin J, et al. High throughput sequencing analysis of biogeographical distribution of bacterial communities in the black soils of northeast China. Soil Biol Biochem. 2014; 70:113–22.
  12. 12. Ansorge WJ. Next-generation DNA sequencing techniques. New Biotechnol. 2009; 25(4):195–203.
  13. 13. O'Connell D. The measurement of apparent specific gravity of soils and its relationship to mechanical composition and plant root growth. Technical Bulletin, Ministry of Agriculture, Fisheries and Food. 1975.
  14. 14. Vance ED, Brookes PC, Jenkinson DS. An Extraction Method for Measuring Soil Microbial Biomass-C. Soil Biol Biochem. 1987; 19(6):703–7. WOS:A1987K675600009.
  15. 15. Kalembasa SJ, Jenkinson DS. A comparative study of titrimetric and gravimetric methods for the determination of organic carbon in soil. J Sci Food Agr. 1973; 24(9):1085–90.
  16. 16. Sparks DL, Page A, Helmke P, Loeppert R, Soltanpour P, Tabatabai M, et al. Methods of soil analysis. Part 3-Chemical methods: Soil Science Society of America Inc.; 1996.
  17. 17. Parkinson J, Allen S. A wet oxidation procedure suitable for the determination of nitrogen and mineral nutrients in biological material. Communications in Soil Science & Plant Analysis. 1975; 6(1):1–11.
  18. 18. Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature. 2005; 437(7057):376–80. pmid:16056220
  19. 19. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB et al.(2009). Introducing mothur: open-source, platform-independent, communitysupported software for describing and comparing microbial communities. Appl Environ Microbiol 75: 7537–7541. pmid:19801464
  20. 20. Zhang X-Y, Sui Y-Y, Zhang X-D, Meng K, Herbert S. Spatial variability of nutrient properties in black soil of northeast China. Pedosphere. 2007; 17(1):19–29.
  21. 21. Lauber CL, Hamady M, Knight R, Fierer N. Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl Environ Microbiol. 2009;75(15):5111–20. pmid:19502440; PubMed Central PMCID: PMC2725504.
  22. 22. Jones RT, Robeson MS, Lauber CL, Hamady M, Knight R, Fierer N. A comprehensive survey of soil acidobacterial diversity using pyrosequencing and clone library analyses. ISME J. 2009; 3(4):442–53. pmid:19129864
  23. 23. Russo SE, Legge R, Weber KA, Brodie EL, Goldfarb KC, Benson AK, et al. Bacterial community structure of contrasting soils underlying Bornean rain forests: Inferences from microarray and next-generation sequencing methods. Soil Biol Biochem. 2012; 55:48–59.
  24. 24. Papatheodorou EM, Stamou GP, Giannotaki A. Response of soil chemical and biological variables to small and large scale changes in climatic factors. Pedobiologia. 2004; 48(4):329–38.
  25. 25. Zhang W, Parker K, Luo Y, Wan S, Wallace L, Hu S. Soil microbial responses to experimental warming and clipping in a tallgrass prairie. Global Change Biology. 2005; 11(2):266–77.
  26. 26. Sowerby A, Emmett B, Beier C, Tietema A, Penuelas J, Estiarte M, et al. Microbial community changes in heathland soil communities along a geographical gradient: interaction with climate change manipulations. Soil Biol Biochem. 2005; 37(10):1805–13.
  27. 27. Liu L, Gundersen P, Zhang T, Mo J. Effects of phosphorus addition on soil microbial biomass and community composition in three forest types in tropical China. Soil Biol Biochem. 2012; 44(1):31–8.
  28. 28. Cleveland CC, Townsend AR, Schmidt SK. Phosphorus limitation of microbial processes in moist tropical forests: evidence from short-term laboratory incubations and field studies. Ecosystems. 2002; 5(7):0680–91.
  29. 29. Ehlers K, Bakken LR, Frostegård Å, Frossard E, Bünemann EK. Phosphorus limitation in a Ferralsol: impact on microbial activity and cell internal P pools. Soil Biol Biochem. 2010; 42(4):558–66.
  30. 30. Wang Q, Wang S, He T, Liu L, Wu J. Response of organic carbon mineralization and microbial community to leaf litter and nutrient additions in subtropical forest soils. Soil Biol Biochem. 2014; 71:13–20.
  31. 31. DeForest JL, Smemo KA, Burke DJ, Elliott HL, Becker JC. Soil microbial responses to elevated phosphorus and pH in acidic temperate deciduous forests. Biogeochemistry. 2012; 109(1–3):189–202.
  32. 32. Wu X, Zhao L, Chen M, Fang H, Yue G, Chen J, et al. Soil Organic Carbon and Its Relationship to Vegetation Communities and Soil Properties in Permafrost Areas of the Central Western Qinghai—Tibet Plateau, China. Permafrost Periglac. 2012; 23(2):162–9.
  33. 33. Long T, Or D. Dynamics of Microbial Growth and Coexistence on Variably Saturated Rough Surfaces. Microb Ecol. 2009; 58(2):262–75. WOS:000268019000004. pmid:19352771
  34. 34. Zhang X, Zhao L, Xu S, Liu Y, Liu H, Cheng G. Soil moisture effect on bacterial and fungal community in Beilu River (Tibetan Plateau) permafrost soils with different vegetation types. J Appl Microbiol. 2013; 114(4):1054–65. pmid:23241008
  35. 35. Zhang X, Xu S, Li C, Zhao L, Feng H, Yue G, et al. The soil carbon/nitrogen ratio and moisture affect microbial community structures in alkaline permafrost-affected soils with different vegetation types on the Tibetan plateau. Res Microbiol. 2014; 165(2):128–39. pmid:24463013
  36. 36. Lauber CL, Hamady M, Knight R, Fierer N. Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl Environ Microb. 2009; 75(15):5111–20.
  37. 37. Chu H, Fierer N, Lauber CL, Caporaso J, Knight R, Grogan P. Soil bacterial diversity in the Arctic is not fundamentally different from that found in other biomes. Environ Microbiol. 2010; 12(11):2998–3006. pmid:20561020
  38. 38. Vestal JR, White DC. Lipid analysis in microbial ecology. Bioscience. 1989:535–41. pmid:11542183
  39. 39. Chim Chan O, Casper P, Sha LQ, Feng ZL, Fu Y, Yang XD, et al. Vegetation cover of forest, shrub and pasture strongly influences soil bacterial community structure as revealed by 16S rRNA gene T—RFLP analysis. Fems Microbiol Ecol. 2008; 64(3):449–58. pmid:18430004
  40. 40. Lynd LR, Weimer PJ, van Zyl WH, Pretorius IS. Microbial cellulose utilization: Fundamentals and biotechnology. Microbiol Mol Biol R. 2002; 66(3):506–577. WOS:000177857300008.
  41. 41. Pankratov TA, Ivanova AO, Dedysh SN, Liesack W. Bacterial populations and environmental factors controlling cellulose degradation in an acidic Sphagnum peat. Environ Microbiol. 2011; 13(7):1800–14. WOS:000292308500011. pmid:21564458
  42. 42. Aneja MK, Sharma S, Munch JC, Schloter M. RNA fingerprinting—a new method to screen for differences in plant litter degrading microbial communities. J Microbiol Meth. 2004; 59(2):223–31.
  43. 43. Baudoin E, Benizri E, Guckert A. Impact of artificial root exudates on the bacterial community structure in bulk soil and maize rhizosphere. Soil Biol and Biochem. 2003; 35(9):1183–92.
  44. 44. Chan OC, Yang X, Fu Y, Feng Z, Sha L, Casper P, et al. 16S rRNA gene analyses of bacterial community structures in the soils of evergreen broad—leaved forests in south-west China. Fems Microbiol Ecol. 2006; 58(2):247–59. pmid:17064266
  45. 45. Terhonen E, Marco T, Sun H, Jalkanen R, Kasanen R, Vuorinen M, et al. The effect of latitude, season and needle-age on the mycota of Scots Pine (Pinus sylvestris) in Finland. Silva Fenn. 2011; 45:301–17.
  46. 46. Hawksworth DL. The magnitude of fungal diversity: the 1.5 million species estimate revisited. Mycol Res. 2001; 105:1422–32. WOS:000173613200005.
  47. 47. Higgins KL, Arnold AE, Miadlikowska J, Sarvate SD, Lutzoni F. Phylogenetic relationships, host affinity, and geographic structure of boreal and arctic endophytes from three major plant lineages. Mol Phylogenet Evol. 2007; 42(2):543–55. pmid:17005421
  48. 48. Hillebrand H, Azovsky AI. Body size determines the strength of the latitudinal diversity gradient. Ecography. 2001; 24(3):251–6.
  49. 49. Fuhrman JA, Steele JA, Hewson I, Schwalbach MS, Brown MV, Green JL, et al. A latitudinal diversity gradient in planktonic marine bacteria. P Natl Acad Sci USA. 2008; 105(22):7774–8.
  50. 50. Fierer N, Jackson RB. The diversity and biogeography of soil bacterial communities. P Natl Acad Sci USA. 2006; 103(3):626–31.
  51. 51. Vries FT, Manning P, Tallowin JR, Mortimer SR, Pilgrim ES, Harrison KA, et al. Abiotic drivers and plant traits explain landscape—scale patterns in soil microbial communities. Ecol Lett. 2012; 15(11):1230–9. pmid:22882451
  52. 52. Evans SE, Wallenstein MD. Climate change alters ecological strategies of soil bacteria. Ecol Lett, 2014, 17(2): 155–164. pmid:24261594
  53. 53. Fuchslueger L, Bahn M, Fritz K, Hasibeder R, Richter A. Experimental drought reduces the transfer of recently fixed plant carbon to soil microbes and alters the bacterial community composition in a mountain meadow. New Phytol, 2014, 201(3): 916–927. pmid:24171922