Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. 2010 Oct; 76(20): 6751–6759.
Published online 2010 Aug 20. doi:  10.1128/AEM.01063-10
PMCID: PMC2953013

Horizon-Specific Bacterial Community Composition of German Grassland Soils, as Revealed by Pyrosequencing-Based Analysis of 16S rRNA Genes


The diversity of bacteria in soil is enormous, and soil bacterial communities can vary greatly in structure. Here, we employed a pyrosequencing-based analysis of the V2-V3 16S rRNA gene region to characterize the overall and horizon-specific (A and B horizons) bacterial community compositions in nine grassland soils, which covered three different land use types. The entire data set comprised 752,838 sequences, 600,544 of which could be classified below the domain level. The average number of sequences per horizon was 41,824. The dominant taxonomic groups present in all samples and horizons were the Acidobacteria, Betaproteobacteria, Actinobacteria, Gammaproteobacteria, Alphaproteobacteria, Deltaproteobacteria, Chloroflexi, Firmicutes, and Bacteroidetes. Despite these overarching dominant taxa, the abundance, diversity, and composition of bacterial communities were horizon specific. In almost all cases, the estimated bacterial diversity (H′) was higher in the A horizons than in the corresponding B horizons. In addition, the H′ was positively correlated with the organic carbon content, the total nitrogen content, and the C-to-N ratio, which decreased with soil depth. It appeared that lower land use intensity results in higher bacterial diversity. The majority of sequences affiliated with the Actinobacteria, Bacteroidetes, Cyanobacteria, Fibrobacteres, Firmicutes, Spirochaetes, Verrucomicrobia, Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria were derived from A horizons, whereas the majority of the sequences related to Acidobacteria, Chloroflexi, Gemmatimonadetes, Nitrospira, TM7, and WS3 originated from B horizons. The distribution of some bacterial phylogenetic groups and subgroups in the different horizons correlated with soil properties such as organic carbon content, total nitrogen content, or microbial biomass.

Soil is probably the most complex microbial environment on Earth with respect to species richness and community size. The microbial richness in soils exceeds that of other environments (44) and is higher by orders of magnitude than the biodiversity of plants and animals. Cultivated soil or grassland soil contains an estimated 2 × 109 prokaryotic cells per gram (12). Soil microbial communities are an important factor of agriculturally managed systems, as they are responsible for most nutrient transformations in soil and influence the above-ground plant diversity and productivity (53).

To analyze the bacterial community in soils, most approaches target the 16S rRNA gene by PCR amplification and subsequent analysis employing sequencing of clone libraries (10, 24), denaturing gradient gel electrophoresis (DGGE) (38), or terminal restriction fragment length polymorphism (T-RFLP) (17, 52). Most of these approaches provided limited insights into the structure of soil bacterial communities, as the survey sizes and the number of compared sampling sites were small with respect to the enormous bacterial diversity present in different soil samples. For example, the reported clone libraries vary considerably in size, but small sample sizes (500 or fewer 16S rRNA gene sequences) are usually analyzed and employed for the theoretical estimation of species richness (39). This provides snapshots of the predominant bacterial community members, but phylogenetic groups that are present in a low abundance and which may possess important ecosystem functions are not assessed (47). In addition, it has been shown that rich sampling (several thousands of clones) of complex bacterial communities is required to perform robust measurements and estimations of community diversity parameters (37). Thus, the detection bias accompanying analyses of small sample sizes can lead to invalidated assumptions. Genetic profiling techniques such as DGGE and T-RFLP have high-throughput capability. These approaches allow researchers to unravel differences in community structure but are limited for assessing diversity (23, 40). To deeply survey the diversity and the composition of the bacterial communities within different soil samples, large-scale pyrosequencing of partial 16S rRNA genes has been employed recently. Previous pyrosequencing-based studies of soil (1, 30, 34, 43) have generated large data sets, which comprised 39,707 (30) to 152,359 (34) 16S rRNA partial gene sequences. Those studies provided comprehensive insights into the biogeography of bacterial soil communities and taxa that were present in a low abundance. However, all those studies focused on the analysis of microbial communities present in topsoil. The subsoil is also known to harbor an important part of the soil microbial biomass (18). It has been shown that the microbial population in the shallow subsurface is impacted by agricultural production to a similar extent as that in topsoil (5).

In this study, we performed large-scale pyrosequencing-based analyses of 16S rRNA genes to assess the bacterial community composition in topsoil and the corresponding subsoil of nine different grassland sites in the Hainich region (Thuringia, Germany). To provide a high level of coverage at the species level (97% genetic distance) and minimize detection bias, we exceeded the above-described numbers of analyzed 16S rRNA gene sequences (752,838 in this study). To examine the impact of land use on bacterial diversity and community composition, the selected grassland sites covered a range of three different land use types, including samples from unfertilized pastures grazed by cattle, fertilized mown pastures grazed by cattle, and fertilized meadows. In many recent studies, surveys were focused on comprehensive analyses of a single soil or a few soil samples (1, 14, 37, 43). This allowed the determination of overall bacterial species richness and community composition, but the assessment of spatial patterns and environmental factors that drive these patterns is hampered by the limited number of examined soils. To assess spatial distribution and the impact of soil edaphic factors and land use on community structure, we used triplicate samples of each land use type from different locations. In addition, composite samples derived from five soil cores after the separation of soil horizons were employed.


Study sites and soil sampling.

Soil samples were collected from nine different grassland sites of the Hainich region in Germany, which is located in the west of Thuringia near the border to Hessen (latitude, 51.2167/N 51°13′0″; longitude, 10.45/E 10°27′0″). The Hainich region is one of the three locations investigated within the framework of the German Biodiversity Exploratories initiative (www.biodiversity-exploratories.de). The nine sampling sites encompassed the following three different land use types: fertilized meadow (plots 1 to 3), fertilized mown pasture grazed by cattle (plots 4 to 6), and unfertilized pasture grazed by cattle (plots 7 to 9) (for coordinates, see Table S1 in the supplemental material). Sampling was performed in April and May 2008. At each sampling site, five soil cores (8.3 cm in diameter) were sampled with a motor-driven soil column cylinder at each corner and in the center of the plot within a given area of 20 m by 20 m. The soil was classified using the World Reference Base of Soil Resources (27). The predominant soil type in the studied plots is Stagnosol, which is characterized by a perched water table, strong mottling, and reducing conditions (27). For each soil core, we determined soil horizons according to the Guidelines for Soil Description (28). The horizons were homogenized and pooled into one composite sample per plot and horizon. Coarse roots and stones (>5 mm) were removed from the samples. In the majority of the samples, the horizons Ah (topsoil) and Btg (subsoil) were detected. In plots 2 and 3 the top horizon was a transitional horizon (ABth) that was employed instead of the missing Ah horizon. In plots 7 and 8 only a transitional horizon (ABth) between topsoil and parent rock material was present. This horizon was employed instead of the missing Btg horizon. Throughout the study the topsoil horizon and the subsoil horizon were designated horizon A and horizon B, respectively.

Edaphic properties of the soil samples.

For determinations of organic carbon (OC) content, total nitrogen (N) content, and soil texture, subsamples from the same composite sample were dried at 40°C and sieved to <2 mm. Total carbon and nitrogen were measured after grinding subsamples to a size of <100 μm in a ball mill. The ground samples were analyzed for total carbon and nitrogen by dry combustion with a Vario Max CN analyzer (Elementar Analysensysteme GmbH, Hanau, Germany). Inorganic carbon was quantified by measuring the total amount of carbon after the removal of organic carbon by the ignition of samples at 450°C for 16 h. To determine soil pH, the subsamples were suspended at a soil-to-liquid ratio of 1:2.5 (soil/0.01 M CaCl2). Subsequently, pH was measured in the supernatant with a glass electrode. Soil texture was determined on 30 g soil according to a method described previously by Schlichting and Blume (45). The edaphic properties are depicted in Table Table11.

Physical and geochemical characteristics of the soil samples from two different soil horizons

Determination of microbial biomass.

To determine microbial biomass, we performed phospholipid fatty acid analysis (PLFA) on soil samples from the A and B horizons of the sampling sites. The composite samples were kept frozen at −80°C after sampling and freeze-dried prior to PLFA extractions. PLFA extractions were performed by using a modified Bligh and Dyer (4) method. Briefly, 2 g of freeze-dried sample was extracted twice in a chloroform-methanol-citrate buffer (1:2:0.8), followed by overnight phase separation. Fatty acids in the organic phase were then separated by using a silica-bonded phase column (silica-based solid-phase extraction [SPE-SI] Bond Elut, 3 ml, 500 mg; Varian Inc., Darmstadt, Germany) to remove glycolipids and neutral lipids. The polar lipids were then converted to fatty acid methyl esters by mild alkaline methanolysis. Methyl-esterified fatty acids were analyzed by using a Hewlett-Packard 6890 gas chromatograph equipped with a DB-5MS column (60-m length; Agilent Technologies, Böblingen, Germany) and interfaced with an Agilent 5973 mass selective detector. Peak areas of each lipid were converted to nmol/g soil using internal standards (19:0 nonadecanoic methyl ester). The total nmol lipid/g dry soil (sum of all lipids present, 20 or fewer carbons in length) was used as an index of microbial biomass (19, 25).

DNA extraction, amplification of 16S rRNA genes, and pyrosequencing.

Total microbial community DNA was isolated from approximately 10 g of soil per sample. For this purpose, the MoBio Power Max soil DNA extraction kit (MoBio Laboratories, Carlsbad, CA) was used according to the manufacturer's instruction. To analyze the taxonomic composition of the soil bacterial community, the V2-V3 region of the 16S rRNA gene (Escherichia coli positions 101 to 536) was chosen for the amplification and subsequent pyrosequencing of the PCR products. The V2-V3 region was amplified with the following primer set, containing the Roche 454 pyrosequencing adaptors (underlined): V2for (5′-GCCTCCCTCGCGCCATCAGAGTGGCGGACGGGTGAGTAA-3′) (modified from that described previously by Schmalenberger et al. [48]) and V3rev (5′-GCCTTGCCAGCCCGCTCAGCGTATTACCGCGGCTGCTG-3′) (7).

For each sample, three independent PCRs were performed. The PCR mixture (final volume, 50 μl) contained 5 μl 10-fold reaction buffer (MBI Fermentas GmbH, St. Leon-Rot, Germany), 30 to 70 ng of soil DNA, 0.4 μM each primer, 0.5 U Pfu polymerase (MBI Fermentas), and 800 μM concentration of each of the four deoxynucleoside triphosphates. In some cases, to achieve amplification of 16S rRNA genes, a different DNA polymerase was used as recommended by the manufacturer (PCR Extender system; VWR International, Hannover, Germany). The polymerase was applied to samples derived from the A horizons of plots 2, 3, 4, and 6 and from the B horizons of plots 4, 6, and 8. Negative-control reactions lacked template DNA. The following thermal cycling scheme was used: initial denaturation at 94°C for 2 min and 25 cycles of denaturation at 94°C for 1 min, annealing for 1 min using a temperature gradient ranging from 60.9°C to 68.2°C, and extension at 72°C for 1.25 min, followed by a final extension period at 72°C for 10 min. Subsequently, the three PCR products per soil sample were pooled in equal amounts and purified by employing the peqGOLD gel extraction kit (Peqlab Biotechnologie GmbH, Erlangen, Germany). Quantification of the PCR products was performed by using the Quant-iT dsDNA BR assay kit and a Qubit fluorometer (Invitrogen GmbH, Karlsruhe, Germany) as recommended by the manufacturer. The Göttingen Genomics Laboratory determined the sequences of the partial 16S rRNA genes by employing the Roche GS-FLX 454 pyrosequencer (Roche, Mannheim, Germany) and using picotiter sequencing plates subdivided into 8 parts (1 part per sample). Amplicons were sequenced as recommended in the instructions of the manufacturer for amplicon sequencing.

Analysis of pyrosequencing-derived data.

Sequences that were shorter than 200 bp in length or of low quality were removed from the pyrosequencing-derived data sets. For taxonomy-based analysis, the Naïve Bayesian rRNA classifier of the Ribosomal Database Project (RDP) (55; http://rdp.cme.msu.edu/) was used. In this way, a rapid taxonomic classification of large partial and full-length rRNA gene sequence data sets according to the new Bergey's bacterial taxonomy (20) was feasible. The bootstrap value was set to ≥80%. Rarefaction curves were calculated by using the tools Aligner, Complete Linkage Clustering, and Rarefaction of the RDP pyrosequencing pipeline (11). We calculated Shannon (50) and Chao1 (8) indices based on the Complete Linkage Clustering data.

Statistical analyses of the pyrosequencing-derived data were carried out with STATISTICA 8.0 (StatSoft Inc., Tulsa, OK). P values of ≤0.05 were considered significant. Significant effects of soil horizon on edaphic soil properties were determined by using the Mann-Whitney U test for nonparametric data. Correlations between phylogenetic groups and soil properties were tested for significance by using Pearson's correlation coefficient. The distributions of phyla between the two soil horizons were calculated by employing the chi-square test. Microbial community compositions were compared to the land use types by using a post hoc Tukey honestly significant difference (HSD) analysis of variance. If the normality test revealed a non-Gaussian distribution, data were transformed.

Nucleotide sequence accession number.

The 16S rRNA gene sequences derived from pyrosequencing have been deposited in the NCBI Sequence Read Archive under accession number SRA020168.1.


General characteristics of the soil samples and the pyrosequencing-derived data set.

In this study, we assessed and compared the compositions of soil bacterial communities present in the A and B horizons of nine different grassland sites of the Hainich region in Germany by a pyrosequencing-based analysis of the 16S rRNA gene sequences. The grassland sites covered a range of the following three different land use types: fertilized meadow (plots 1 to 3), fertilized mown pasture grazed by cattle (plots 4 to 6), and unfertilized pasture grazed by cattle (plots 7 to 9). The soil type of all samples was Stagnosol, except for plot 1, which was a Vertic Cambisol. In addition, further analysis of the Stagnosols revealed that plot 6 was a Luvic Stagnosol, whereas the other plots were Vertic Stagnosols (see Table S1 in the supplemental material).

The two analyzed soil horizons showed significant differences with respect to edaphic properties such as soil pH, OC content, N content, C-to-N ratio, and water content (Table (Table1).1). The pH in the A horizons ranged from 6.03 to 7.20, and the pH in the B horizons ranged from 6.30 to 7.40. In general, the pH value of the B horizon was higher than that of the corresponding A horizon, whereas the water content, the amounts of OC and N, and the C-to-N ratio showed 1.1- to 2.0-fold, 2.0- to 10.9-fold, 1.8- to 7.3-fold, and 1.2- to 1.5-fold decreases with depth, respectively.

The microbial biomass in the B horizons of all samples was lower by 48.7 to 96.3% than that in the corresponding A horizons (Table (Table1).1). A decrease in the total microbial biomass with soil depth was previously reported (5, 9, 16, 18). The total microbial biomass was positively correlated with the concentration of OC (r = 0.88; P < 0.01). This supports the assumption reported previously by Blume et al. (5), that carbon availability is closely associated with microbial biomass. In addition, significant correlations of microbial biomass with the concentration of N (r = 0.84; P < 0.01) and the C-to-N ratio (r = 0.89; P < 0.01) were detected.

The pyrosequencing-based analysis of the V2-V3 region of the 16S rRNA gene was employed for assessments of bacterial community compositions from the A and B horizons of the nine sampling sites. Short pyrosequencing reads assess the microbial diversity almost as reliably as near-full-length sequences when appropriate primers are chosen. Primers derived from V2-V3 region of the 16S rRNA gene were shown previously to be suitable for this purpose (36). In addition, this region is the most effective region for universal genus identification (7, 42). Across all 18 samples, we recovered 752,838 quality sequences with a read length of ≥200 bp. The average read length was 262 bp. The number of sequences per sample ranged from 25,851 to 61,366, with an average of 41,824 (see Table S2 in the supplemental material). We were able to classify 600,544 (79.77%) of the quality sequences below the domain level. The percentage of classified 16S rRNA gene sequences was in the range of those of other pyrosequencing-based studies (35), but the average number of sequences per sample and the total number of analyzed sequences exceeded those of other previously reported studies of pyrosequencing-based determinations of soil bacterial community composition (34, 43).

Bacterial richness and diversity indices.

To determine rarefaction curves, richness, and diversity, we identified operational taxonomic units (OTUs) at sequence divergences of 3% (species level) and 20% (phylum level). The rarefaction analysis of bacterial communities derived from the A and B horizons of the nine sampling sites is depicted in Fig. Fig.1.1. At a 20% genetic distance, almost all curves showed saturation, indicating that the surveying effort covered almost the full extent of taxonomic diversity at this genetic distance. In addition, a comparison of rarefaction analyses with the number of OTUs estimated by the Chao1 richness estimator revealed that 83 to 100% of the estimated taxonomic richness was covered by the sequencing effort. At a 3% genetic distance, the observed richness was 63 to 80% of that predicted by the Chao1 richness estimator (Table (Table2).2). Thus, we did not survey the full extent of taxonomic diversity at the species level. Taking into account that at genetic distances below 5%, rarefaction analyses underestimate the bacterial richness whereas Chao1 estimators overestimate it (43), a substantial fraction of the bacterial diversity at the species level was assessed by the surveying effort. It is important that pyrosequencing provides an unprecedented sampling depth compared to that of traditional Sanger sequencing of 16S rRNA genes (51), but the intrinsic error of pyrosequencing could result in the overestimation of rare phylotypes, since each pyrosequencing read is treated as a unique identifier of a community member and correction by assembly and sequencing depth, which is typically applied during genome projects, is not feasible (26, 32). To ensure per-base error rates lower than that of conventional Sanger sequencing, we used quality filtering of the pyrosequencing-derived data set, such as the removal of reads with atypical lengths (26). In addition, to minimize the overestimation of rare phylotypes, clustering and diversity estimates were performed only at genetic divergences of ≥3% (32).

FIG. 1.
Rarefaction curves indicating the observed number of OTUs within the 16S rRNA gene sequences of the A and B horizons, derived from nine German grassland sites. OTUs are shown at the 3 and 20% genetic distance levels. The rarefaction curves were ...
Species richness estimates obtained at genetic distances of 3% and 20%c

Acosta-Martínez et al. (1) postulated previously that in managed soils, the maximum number of OTUs is less than 3,400 at a genetic distance of 3%. This is in contrast to our results, as up to 4,781 and 6,231 OTUs were predicted for fertilized meadows (plot 3, A horizon) and fertilized mown pasture grazed by cattle (plot 5, A horizon), respectively (Table (Table2).2). The differences in the results might be explained by the different surveying efforts. Several studies (13, 37, 43, 46, 56) showed that the number of analyzed sequences per sample has an effect on the predicted number of OTUs. For example, Roesch et al. (43) previously plotted the number of observed OTUs against the sequencing effort using the bacterial community present in a Canadian forest soil sample. The employment of the whole data set (53,632 sequences) revealed 5,500 OTUs at a genetic distance of 3%, whereas the reduction of the same data set to 30,000 sequences yielded 3,500 OTUs. In general, fewer sequences result in lower curve progression and a lower number of predicted OTUs. In addition, the comparison of richness estimates between different surveys might be hampered by the differences in sequence conservation and sequence length of the analyzed 16S rRNA gene regions. Recently, Engelbrektson et al. (15) showed that amplicon length and differences in the analyzed 16S rRNA gene regions markedly influenced estimates of richness and evenness.

The Shannon index of diversity (H′) was determined for all samples (Table (Table2).2). At a genetic distance of 3%, it ranged from 5.65 to 7.16 in the A horizons and from 5.01 to 6.72 in the B horizons. The predicted diversity in the topsoil exceeded that of the corresponding subsoil, except for plot 1 (Table (Table2).2). To our knowledge, no other study assessing bacterial diversity along a soil profile was conducted with a comparable surveying effort. However, a significant decrease of bacterial diversity with soil depth was also recorded by a community analysis employing terminal restriction fragment length polymorphism (33) and phospholipid fatty acid analysis (18) of soil profiles derived from Californian grassland and soil samples of the Sedgwick Reserve (California), respectively.

The bacterial diversity at a genetic distance of 3% was strongly related to the content of OC and N as well as to the C-to-N ratio. Positive correlations between the H′ and the OC content (r = 0.60; P < 0.01), the N content (r = 0.58; P < 0.05), and the C-to-N ratio (r = 0.65; P < 0.01) were observed. Similar correlations were detected by analyzing soil samples from South American grasslands, in which the H′ correlated positively with the microbial biomass C and N (r = 0.53 to 0.58; P = 0.02 to 0.03) (3). Interestingly, the mean H′ was lower in fertilized meadows (plots 1 to 3), with intermediate values in fertilized mown meadows (plots 4 to 6) and the highest values in unfertilized pastures (plots 7 to 9), over both horizons (Table (Table2).2). Thus, a higher bacterial diversity in samples from unfertilized plots, which represent the lowest land use intensity in this study, is indicated. Nevertheless, within the same land use types, strong variations in diversity were observed.

Distribution of taxa and phylotypes across all samples.

The 600,544 classifiable sequences were affiliated with 23 phyla across the entire data set. The dominant phyla across all samples were Acidobacteria, Betaproteobacteria, Actinobacteria, Gammaproteobacteria, Alphaproteobacteria, Deltaproteobacteria, Chloroflexi, Firmicutes, and Bacteroidetes, representing 26.98, 15.76, 11.62, 11.10, 9.69, 5.09, 3.85, 3.22, and 1.45%, respectively, of all sequences that were classified below the domain level (Fig. (Fig.22 and see Table S3 in the supplemental material). In addition, the dominant phyla were present in all samples. These results are in accordance with results from a previously reported meta-analysis of bacterial community composition in soils (29). In addition, the abundances of the five dominant phyla, Acidobacteria, Actinobacteria, Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria, which represented approximately 75% of all classified sequences, corresponded roughly to those found by other studies (21, 34). Thus, despite the different surveying efforts and sampling sites used in the different studies, it is indicated that a variety of soils contain the same dominant bacterial groups.

FIG. 2.
Distribution of phylogenetic groups in the A and B horizons derived from the different grassland sampling sites. Plot numbers are given below the graph. A description of the plots is given in Table Table1.1. A and B indicate the different horizons. ...

The most abundant phylotypes in the A and B horizons were a member of the Alphaproteobacteria (Acetobacteraceae) and a member of acidobacterial subgroup 4, respectively. The corresponding sequences represented 0.74 and 2.86% of all classified sequences in each respective horizon. The most abundant phylotype within one individual sample (plot 3, B horizon) was the above-mentioned member of acidobacterial subgroup 4, representing 8.36% of the sequences from that soil.

Distribution and abundance of the predominant phylum Acidobacteria.

Members of the phylum Acidobacteria were predominant across all samples. This finding is in accordance with findings of other studies of the composition of soil-derived bacterial communities from a variety of environments, such as pristine forest, grassland, and agricultural soils (29). Here, members of the Acidobacteria form a significant fraction (12.68 to 49.86%) of the 16S rRNA gene sequences in both horizons from all land use types (see Tables S3 and S4 in the supplemental material). Correspondingly, members of this phylum have been reported to constitute an average of 20% and a maximum of approximately 50% of bacterial communities derived from various soils (13). Thus, based on their abundance and their presence in various soil types, the Acidobacteria appear to play an important role in the ecosystem function of soils.

The phylum Acidobacteria is divided into 26 subgroups, but only little is known with respect to the physiological and metabolic capabilities of the different subgroups (2). We detected 18 and 22 of these subgroups in the A and B horizons, respectively. Most abundant in both horizons were subgroups 6, 4, 16, and 7 (see Tables S4 and S5 in the supplemental material). In the A horizons, these subgroups were represented by 47, 21, 13, and 7% of all acidobacterial sequences, respectively, and in the B horizons, these subgroups were represented by 26, 38, 5, and 14% of all acidobacterial sequences, respectively (see Table S5 in the supplemental material). These results are in contrast to a previous study by Hansel et al. (21) of samples from a continuous watershed soil profile (Oak Ridge), which is the only other report of acidobacterial diversity with respect to soil horizon. In the A horizon, those researchers detected primarily subgroups 3 (21%), 4 (29%), and 6 (29%), whereas in our samples, these subgroups were represented by 5, 21, and 47%, respectively. In the B horizon, Hansel et al. (21) detected primarily subgroups 1 (32%) and 2 (61%), which were represented by less than 1% of all acidobacterial sequences derived from our soil samples. The predominant subgroups in the B horizons from the Hainich region were subgroups 4, 5, and 7. The major differences in the occurrences of acidobacterial subgroups in the B horizon might be due to the dissimilar pH values of the samples used in both studies. The pH in our subsoil samples ranged from 6.30 to 7.40 (Table (Table1)1) whereas the pH of the soil samples studied by Hansel et al. (21) was 4.5. It was reported previously that the abundance of the phylum Acidobacteria correlates with the soil pH (22, 30). Lauber et al. (34) showed previously that acidobacterial subgroups 1 and 2 were most abundant in acidic soils and decreased with the increase of the pH. Here, no significant correlations of changes in the abundance of the dominant acidobacterial subgroups and other phylogenetic groups with pH were observed. A reason for this finding might be that the sampling effort in most of the other studies was much less than that of this study. Another possibility is that almost all the pH values of our samples were near neutral. Correspondingly, a relatively small pH range was covered by our soil samples (Table (Table1),1), so there is simply a lower pH range from which to determine correlations. Nevertheless, we observed negative correlations between the abundant acidobacterial subgroup 4 and the OC content (r = −0.84; P < 0.01), N content (r = −0.83; P < 0.01), or C-to-N ratio (r = −0.77; P < 0.05) in the A horizons (see Fig. S1 in the supplemental material). In the B horizons, subgroup 4 also correlated negatively with the C-to-N ratio (r = −0.70; P < 0.05), whereas the relative abundance of subgroup 6 showed a positive correlation with the C-to-N ratio (r = 0.70; P < 0.05). Thus, the subgroup distribution varied with respect to the soil profile (horizon) and soil properties and provided some insights into the conditions that are required by the different subgroups. For example, low nutrient/OC conditions (B horizons) appear to favor subgroups 4 and 7, whereas higher nutrient/OC conditions (A horizons) favor subgroup 16. Interestingly, for members of subdivision 6, a high tolerance to nutrient/OC availability was indicated, as they constituted a substantial fraction in the A horizon and the B horizon (8.87 and 9.73% of all classified sequences, respectively).

Taxonomic compositions in A horizons and the corresponding B horizons.

The compositions of the bacterial community and the distributions of the phyla varied between A and B soil horizons. The most frequently present phyla in the A horizons were Acidobacteria (13 to 23%), Betaproteobacteria (14 to 23%), Gammaproteobacteria (10 to 26%), Actinobacteria (5 to 17%), and Alphaproteobacteria (9 to 14%). The most abundant phyla in the B horizons were Acidobacteria (28 to 50%), Betaproteobacteria (10 to 18%), Actinobacteria (4 to 15%), Chloroflexi (3 to 12%), and Alphaproteobacteria (5 to 10%) (Fig. (Fig.22).

We analyzed the respective abundances of the 15 most represented phyla in the A and B horizons. For almost all phyla and land use types, a significant (P < 0.00001) difference in distribution between the two horizons was apparent (see Table S6 in the supplemental material). The distribution of selected phyla in the two horizons is shown in Fig. Fig.33 (see also Fig. S2 in the supplemental material). The majority of sequences affiliated with the Actinobacteria, Bacteroidetes, Cyanobacteria, Fibrobacteres, Firmicutes, Spirochaetes, Verrucomicrobia, Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria were derived from A horizons, whereas the majority of the sequences related to Acidobacteria, Chloroflexi, Gemmatimonadetes, Nitrospira, TM7, and WS3 originated from B horizons. In many other reports, a pH gradient was identified as a major factor for changes in soil community structure, but as mentioned above, this trend was not observed in this study. Therefore, other factors appear to control the distribution of the phyla along the soil profile. As stated above, the total biomass decreased with soil depth (Table (Table1).1). A significant correlation between the total microbial biomass and the occurrence of several phyla was recorded. The relative abundance of the Acidobacteria, Chloroflexi, and Nitrospira, which increased with depth, correlated negatively with total biomass (r = −0.53 to −0.79; P < 0.05). The relative abundances of the Actinobacteria, Bacteroidetes, Verrucomicrobia, Alphaproteobacteria, and Gammaproteobacteria, which decreased with depth, showed a positive correlation with biomass (r = 0.69 to 0.84; P < 0.01). A positive correlation of some of the dominant acidobacterial groups (see above) and the alphaproteobacterial order Rhizobiales with the concentration of OC (r = 0.77; P < 0.05), the concentration of N (r = 0.73; P < 0.05), and the C-to-N ratio (r = 0.87; P < 0.01) was detectable in the B horizon (see Fig. S3 in the supplemental material). A statistically significant positive correlation of the Rhizobiales with the C-to-N ratio (r = 0.80; P < 0.01) was also detected in the A horizon. Taking into account that the Rhizobiales include the genera Rhizobium and Bradyrhizobium, which comprise members that are able to fix nitrogen and are associated with roots of legumes, a positive correlation was expected (49). In addition, land use and management regimens seem to have an impact on the Rhizobiales, as the fertilized plots cluster and the plots with cattle cluster (data not shown). An impact of fertilization on the structure and diversity of rhizobial populations was observed previously in other studies (6, 41). For example, rhizobial populations differed between cultivated and uncultivated Mexican soils. In addition, the affinity of host cultivars for different members of the Rhizobiales influenced the composition of rhizobial populations (54).

FIG. 3.
Comparison of the overall distribution of selected phyla within the A and B horizons. The black bar represents the sum of all members of a phylum in the A horizon, while the gray bar represents the sum of all members in the B horizon. Box-and-whisker ...

Striking is the distribution of the phyla Fibrobacteres and Nitrospira, which occurred almost exclusively in the A or the B horizon, respectively. Members of the Fibrobacteres are part of the microbial community in the first stomach of ruminant animals and degrade plant-based cellulose (31). Taking into account that members of the Fibrobacteres hardly occurred in grassland samples without the presence of cattle (i.e., plots 1 to 3) and almost exclusively in the topsoil, it can be assumed that members of this phylum were introduced into the samples by cattle. Members of the Nitrospira are found in interspace soils and rarely in the rhizosphere (13). In the latter environment, heterotrophic root-associated microorganisms suppress the growth of autotrophic Nitrospira. This might explain why in our samples, members of the Nitrospira occurred in the subsoil, with less rooting than in the topsoil. In addition, the concentrations of OC and N decreased with soil depth (Table (Table1).1). Thus, chemolithoautotrophic organisms adapted to darkness, like Nitrospira, have a selective advantage in subsoil samples.


Although we recovered an average of 41,824 sequences per sample, we did not survey the full extent of bacterial richness at the species level within an individual soil or horizon. Thus, an increase in surveying efforts would probably result in the identification of more bacterial taxa, which are present in a low abundance. In most cases, the B horizons showed a lower estimated bacterial diversity than the corresponding A horizons (Table (Table2).2). Correspondingly, a greater coverage of the bacterial community in the B horizons can be achieved by using the same surveying effort. In addition, the identification of bacterial taxa at the finest level of taxonomic resolution is currently not feasible by applying large-scale pyrosequencing. However, the advancement of the technology will result in an increase of the read length, and this limitation will become less relevant in the near future.

To provide a robust assessment of the impact of land use, soil factors, or soil depth on bacterial diversity, distribution, and community composition, we used triplicate samples of each land use type from different locations. Overall, the abundance, composition, and diversity of the bacterial communities were strongly depth dependent. The Shannon index of diversity along with the nutrient content (N and OC), water content, and biomass decreased with depth (Tables (Tables11 and and2).2). Based on the sharp decrease of the OC content (up to 10.9-fold) in the B horizons compared to the corresponding A horizons (Table (Table1)1) and other surveys (18, 33, 57), the concentration of OC appears to be the major driver for the diversity and structure of bacterial communities along the soil profile at near-neutral pH values. Nevertheless, we observed a variability of bacterial communities within an individual land use type, and exceptions to the above-mentioned general results were found; i.e., a slightly higher Shannon index was recorded for the B horizon of plot 1. Thus, it is advisable to survey as many soil samples as possible for the identification of general patterns and comparison of the results with those of other soil surveys. One caveat of the latter, however, is the limited comparability of different surveys, as sampling strategy, survey effort, number and type of soil factors measured, and approaches used to analyze the sequence data vary considerably (37). To take full advantage of the increasing number of data sets on soil bacterial communities, minimal requirements for sampling and the set of analyzed soil factors as well as rules for sequence analysis and phylogenetic assignment should be defined.

Supplementary Material

[Supplemental material]


We thank Rudolf-Josef Fischer (Department of Medical Informatics and Biomathematics, Westphalian Wilhelms University, Münster, Germany) for supporting the statistical analyses. We thank the local implementation team of the Hainich exploratory and the BEO (Biodiversity Exploratories Office) for organizing the coordinated soil sampling campaign.

The work has been funded by the DFG Priority Program 1374, Infrastructure-Biodiversity-Exploratories (DA 374/4-1).

Field work permits were given by the responsible state environmental offices of Thüringen.


Published ahead of print on 20 August 2010.

Supplemental material for this article may be found at http://aem.asm.org/.


1. Acosta-Martínez, V., S. Dowd, Y. Sun, and V. Allen. 2008. Tag-encoded pyrosequencing analysis of bacterial diversity in a single soil type as affected by management and land use. Soil Biol. Biochem. 40:2762-2770.
2. Barns, S., E. Cain, L. Sommerville, and C. Kuske. 2007. Acidobacteria phylum sequences in uranium-contaminated subsurface sediments greatly expand the known diversity within the phylum. Appl. Environ. Microbiol. 73:3113-3116. [PMC free article] [PubMed]
3. Berthrong, S., C. Schadt, G. Piñeiro, and R. Jackson. 2009. Afforestation alters the composition of functional genes in soil and biogeochemical processes in South American grasslands. Appl. Environ. Microbiol. 75:6240-6248. [PMC free article] [PubMed]
4. Bligh, E., and W. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37:911-917. [PubMed]
5. Blume, E., M. Bischoff, J. M. Reichert, T. Moorman, A. Konopka, and R. F. Turco. 2002. Surface and subsurface microbial biomass, community structure and metabolic activity as a function of soil depth and season. Appl. Soil Ecol. 20:171-181.
6. Caballero-Mellado, J., and E. Martínez-Romero. 1999. Soil fertilization limits the genetic diversity of Rhizobium in bean nodules. Symbiosis 26:111-121.
7. Chakravorty, S., D. Helb, M. Burday, N. Connell, and D. Alland. 2007. A detailed analysis of 16S ribosomal RNA gene segments for the diagnosis of pathogenic bacteria. J. Microbiol. Methods 69:330-339. [PMC free article] [PubMed]
8. Chao, A., and J. Bunge. 2002. Estimating the number of species in a stochastic abundance model. Biometrics 58:531-539. [PubMed]
9. Cho, S., S. Tsai, A. Ravindran, A. Selvam, and S. Yang. 2008. Seasonal variation of microbial populations and biomass in Tatachia grassland soils of Taiwan. Environ. Geochem. Health 30:255-272. [PubMed]
10. Chow, M., C. Radomski, J. McDermott, J. Davies, and P. Axelrood. 2002. Molecular characterization of bacterial diversity in Lodgepole pine (Pinus contorta) rhizosphere soils from British Columbia forest soils differing in disturbance and geographic source. FEMS Microbiol. Ecol. 42:347-357. [PubMed]
11. Cole, J., Q. Wang, E. Cardenas, J. Fish, B. Chai, R. Farris, A. Kulam-Syed-Mohideen, D. McGarrell, T. Marsh, G. Garrity, and J. Tiedje. 2009. The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res. 37:D141-D145. [PMC free article] [PubMed]
12. Daniel, R. 2005. The metagenomics of soil. Nat. Rev. Microbiol. 3:470-478. [PubMed]
13. Dunbar, J., S. Barns, L. Ticknor, and C. Kuske. 2002. Empirical and theoretical bacterial diversity in four Arizona soils. Appl. Environ. Microbiol. 68:3035-3045. [PMC free article] [PubMed]
14. Elshahed, M., N. Youssef, A. Spain, C. Sheik, F. Najar, L. Sukharnikov, B. Roe, J. Davis, P. Schloss, V. Bailey, and L. Krumholz. 2008. Novelty and uniqueness patterns of rare members of the soil biosphere. Appl. Environ. Microbiol. 74:5422-5428. [PMC free article] [PubMed]
15. Engelbrektson, A., V. Kunin, K. Wrighton, N. Zvenigorodsky, F. Chen, H. Ochman, and P. Hugenholtz. 2010. Experimental factors affecting PCR-based estimates of microbial species richness and evenness. ISME J. 4:642-647. [PubMed]
16. Federle, T., D. Dobbins, J. Thornton-Manning, and D. Jones. 1986. Microbial biomass, activity, and community structure in subsurface soils. Ground Water 24:365-374.
17. Fierer, N., and R. Jackson. 2006. The diversity and biogeography of soil bacterial communities. Proc. Natl. Acad. Sci. U. S. A. 103:626-631. [PMC free article] [PubMed]
18. Fierer, N., J. P. Schimel, and P. A. Holden. 2003. Variations in microbial community composition through two soil depth profiles. Soil Biol. Biochem. 35:167-176.
19. Frostegård, A., and E. Bååth. 1996. The use of fatty acid analysis to estimate bacterial and fungal biomass in soil. Biol. Fertil. Soils 22:59-65.
20. Garrity, G. M., J. A. Bell, and T. G. Lilburn. 2004. Taxonomic outline of the prokaryotes, p. 4-23. In G. M. Garrity (ed.), Bergey's manual of systematic bacteriology, 2nd ed., vol. 2. Springer-Verlag, New York, NY.
21. Hansel, C., S. Fendorf, P. Jardine, and C. Francis. 2008. Changes in bacterial and archaeal community structure and functional diversity along a geochemically variable soil profile. Appl. Environ. Microbiol. 74:1620-1633. [PMC free article] [PubMed]
22. Hartman, W., C. Richardson, R. Vilgalys, and G. Bruland. 2008. Environmental and anthropogenic controls over bacterial communities in wetland soils. Proc. Natl. Acad. Sci. U. S. A. 105:17842-17847. [PMC free article] [PubMed]
23. Hartmann, M., and F. Widmer. 2006. Community structure analyses are more sensitive to differences in soil bacterial communities than anonymous diversity indices. Appl. Environ. Microbiol. 72:7804-7812. [PMC free article] [PubMed]
24. He, J., Z. Xu, and J. Hughes. 2006. Molecular bacterial diversity of a forest soil under residue management regimes in subtropical Australia. FEMS Microbiol. Ecol. 55:38-47. [PubMed]
25. Hill, T., E. McPherson, J. Harris, and P. Birch. 1993. Microbial biomass estimated by phospholipid phosphate in soils with diverse microbial communities. Soil Biol. Biochem. 25:1779-1786.
26. Huse, S., J. Huber, H. Morrison, M. Sogin, and D. Welch. 2007. Accuracy and quality of massively parallel DNA pyrosequencing. Genome Biol. 8:R143. [PMC free article] [PubMed]
27. IUSS Working Group. 2006. World reference base for soil resources 2006, 2nd ed. World soil resources reports no. 103. FAO, Rome, Italy.
28. Jahn, R., H.-P. Blume, V. B. Asio, O. Spaargaren, and P. Schad. 2006. Guidelines for soil description, 4th ed. FAO, Rome, Italy.
29. Janssen, P. 2006. Identifying the dominant soil bacterial taxa in libraries of 16S rRNA and 16S rRNA genes. Appl. Environ. Microbiol. 72:1719-1728. [PMC free article] [PubMed]
30. Jones, R., M. Robeson, C. Lauber, M. Hamady, R. Knight, and N. Fierer. 2009. A comprehensive survey of soil acidobacterial diversity using pyrosequencing and clone library analyses. ISME J. 3:442-453. [PMC free article] [PubMed]
31. Kobayashi, Y., T. Shinkai, and S. Koike. 2008. Ecological and physiological characterization shows that Fibrobacter succinogenes is important in rumen fiber digestion—review. Folia Microbiol. (Praha) 53:195-200. [PubMed]
32. Kunin, V., A. Engelbrektson, H. Ochman, and P. Hugenholtz. 2010. Wrinkles in the rare biosphere: pyrosequencing errors can lead to artificial inflation of diversity estimates. Environ. Microbiol. 12:118-123. [PubMed]
33. LaMontagne, M., J. Schimel, and P. Holden. 2003. Comparison of subsurface and surface soil bacterial communities in California grassland as assessed by terminal restriction fragment length polymorphisms of PCR-amplified 16S rRNA genes. Microb. Ecol. 46:216-227. [PubMed]
34. Lauber, C., M. Hamady, R. Knight, and N. Fierer. 2009. Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl. Environ. Microbiol. 75:5111-5120. [PMC free article] [PubMed]
35. Lazarevic, V., K. Whiteson, S. Huse, D. Hernandez, L. Farinelli, M. Østerås, J. Schrenzel, and P. François. 2009. Metagenomic study of the oral microbiota by Illumina high-throughput sequencing. J. Microbiol. Methods 79:266-271. [PMC free article] [PubMed]
36. Liu, Z., C. Lozupone, M. Hamady, F. Bushman, and R. Knight. 2007. Short pyrosequencing reads suffice for accurate microbial community analysis. Nucleic Acids Res. 35:120-130. [PMC free article] [PubMed]
37. Morales, S., T. Cosart, J. Johnson, and W. Holben. 2009. Extensive phylogenetic analysis of a soil bacterial community illustrates extreme taxon evenness and the effects of amplicon length, degree of coverage, and DNA fractionation on classification and ecological parameters. Appl. Environ. Microbiol. 75:668-675. [PMC free article] [PubMed]
38. Muyzer, G., E. de Waal, and A. Uitterlinden. 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59:695-700. [PMC free article] [PubMed]
39. Narang, R., and J. Dunbar. 2004. Modeling bacterial species abundance from small community surveys. Microb. Ecol. 47:396-406. [PubMed]
40. Osborn, A., E. Moore, and K. Timmis. 2000. An evaluation of terminal-restriction fragment length polymorphism (T-RFLP) analysis for the study of microbial community structure and dynamics. Environ. Microbiol. 2:39-50. [PubMed]
41. Palmer, K., and J. Young. 2000. Higher diversity of Rhizobium leguminosarum biovar viciae populations in arable soils than in grass soils. Appl. Environ. Microbiol. 66:2445-2450. [PMC free article] [PubMed]
42. Petrosino, J., S. Highlander, R. Luna, R. Gibbs, and J. Versalovic. 2009. Metagenomic pyrosequencing and microbial identification. Clin. Chem. 55:856-866. [PMC free article] [PubMed]
43. Roesch, L., R. Fulthorpe, A. Riva, G. Casella, A. Hadwin, A. Kent, S. Daroub, F. Camargo, W. Farmerie, and E. Triplett. 2007. Pyrosequencing enumerates and contrasts soil microbial diversity. ISME J. 1:283-290. [PMC free article] [PubMed]
44. Rosselló-Mora, R., and R. Amann. 2001. The species concept for prokaryotes. FEMS Microbiol. Rev. 25:39-67. [PubMed]
45. Schlichting, E., and H.-P. Blume. 1966. Bodenkundliches Praktikum: eine Einführung in pedologisches Arbeiten für Ökologen, insbesondere Land- und Forstwirte und für Geowissenschaftler. Paul Parey, Hamburg, Germany.
46. Schloss, P., and J. Handelsman. 2005. Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl. Environ. Microbiol. 71:1501-1506. [PMC free article] [PubMed]
47. Schloss, P., and J. Handelsman. 2006. Toward a census of bacteria in soil. PLoS Comput. Biol. 2:e92. [PMC free article] [PubMed]
48. Schmalenberger, A., F. Schwieger, and C. Tebbe. 2001. Effect of primers hybridizing to different evolutionarily conserved regions of the small-subunit rRNA gene in PCR-based microbial community analyses and genetic profiling. Appl. Environ. Microbiol. 67:3557-3563. [PMC free article] [PubMed]
49. Sessitsch, A., J. G. Howieson, X. Perret, H. Antoun, and E. Martínez-Romero. 2002. Advances in Rhizobium research. Crit. Rev. Plant Sci. 21:323-378.
50. Shannon, C. 1997. The mathematical theory of communication. 1963. MD Comput. 14:306-317. [PubMed]
51. Sogin, M., H. Morrison, J. Huber, D. Mark Welch, S. Huse, P. Neal, J. Arrieta, and G. Herndl. 2006. Microbial diversity in the deep sea and the underexplored “rare biosphere.” Proc. Natl. Acad. Sci. U. S. A. 103:12115-12120. [PMC free article] [PubMed]
52. Tiedje, J. M., S. Asuming-Brempong, K. Nüsslein, T. L. Marsh, and S. J. Flynn. 1999. Opening the black box of soil microbial diversity. Appl. Soil Ecol. 13:109-122.
53. van der Heijden, M., R. Bardgett, and N. van Straalen. 2008. The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecol. Lett. 11:296-310. [PubMed]
54. Wang, E., J. Martínez-Romero, and E. Martínez-Romero. 1999. Genetic diversity of rhizobia from Leucaena leucocephala nodules in Mexican soils. Mol. Ecol. 8:711-724.
55. Wang, Q., G. Garrity, J. Tiedje, and J. Cole. 2007. Naïve Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73:5261-5267. [PMC free article] [PubMed]
56. Youssef, N., and M. Elshahed. 2008. Species richness in soil bacterial communities: a proposed approach to overcome sample size bias. J. Microbiol. Methods 75:86-91. [PubMed]
57. Zhou, J., B. Xia, D. Treves, L. Wu, T. Marsh, R. O'Neill, A. Palumbo, and J. Tiedje. 2002. Spatial and resource factors influencing high microbial diversity in soil. Appl. Environ. Microbiol. 68:326-334. [PMC free article] [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem chemical compound records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records. Multiple substance records may contribute to the PubChem compound record.
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • SRA
    Massively-parallel sequencing project data in the Short Read Archive (SRA) that are reported in the current articles.
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...