Structure of soil bacterial communities in relation to environmental variables in a semi-arid region of Mongolia

Abstract

Patterns in soil bacterial communities, and the factors that determine them, have been little explored in arid and semi-arid environments. It is unclear to what extent the diversity and community composition of arid-land soil bacterial communities follow vegetation habitats, or conversely other relatively independent soil variables. It is also unclear whether the factors (e.g. pH) that contribute to variation in bacterial communities in some moister environments also operate on a local scale in semi-arid environments. To identify the main factors in shaping bacterial community structure in semi-arid environments, we sampled a mosaic of habitats under different vegetation, landscape and edaphic conditions in central Mongolia, including steppe, forest-steppe, and abandoned wheat field. Soil DNA was extracted and pyrosequenced for 16S rRNA gene identification. NMDS results showed that bacterial community structures are slightly different from one habitat to another. However, the similarity between communities both within and between habitats is determined more strongly by soil texture than by vegetation type and drainage conditions. Moreover, the relative abundances of certain phyla are correlated with specific soil properties such as salinity and soil texture, in ways that have not previously been found in semi-arid environments. Actinobacteria, for example, show a negative correlation with salinity and Bacteroidetes display a positive relationship with percentage silt and clay. It also appears that the most important environmental variables (soil texture and salinity) affecting the bacterial community within this semi-arid environment are different from those found in moister environments, with no detectable effect of pH.

Introduction

The description and understanding of bacterial diversity in nature is currently in its early stages. Until about ten years ago, only around 9747 species had been identified through culturing (http://www.bacterio.cict.fr/number.html). Culture-independent approaches using molecular techniques, such as DGGE, T-RFLP, cloning, etc., have revealed many thousands of previously unknown species in various environments around the world for decades. Now, metagenetic analysis together with massive parallel sequencing techniques have further revolutionized microbial ecology by making it possible to identify an order of magnitude more microbial taxa (Costello et al., 2009; Delmotte et al., 2009; Sogin et al., 2006)

As taxonomic resolution has increased due to the rapid development of molecular techniques, microbiologists have started to explore patterns in the community structure and diversity of microbial biota. Despite several preliminary reviews, these patterns are still poorly understood (Martiny et al., 2006; Ramette and Tiedje, 2007). In particular, it is not clear to what extent microbial communities are variable within and between habitats, whether there is a spatial pattern in these communities, and what environmental factors explain the greatest part of the observed variation in community structure. One possibility is that the composition of the plant community itself might primarily determine the soil bacterial community. It has been reported that different plant species have distinctive bacterial communities in their rhizosphere (Garbeva et al., 2008), and different vegetation types are known to harbor distinct sets of species of other groups of organisms such as birds (Borges, 2004), insects (Sugiura et al., 2008) and macrofungi (Zhang et al., 2010). Hence there are reasons for expecting that the same could be true of soil bacteria in semi-arid vegetation. However, it has been reported in studies of other biomes that combinations of biotic and abiotic factors, i.e. vegetation, land-use, geographic distance, and soil characteristics, affect bacterial species diversity and community composition (Fierer and Jackson, 2006; Hansel et al., 2008; Yergeau et al., 2007). Recently, it has been clearly shown that soil pH is one of the major drivers controlling soil bacterial community structure on a regional scale (Fierer and Jackson, 2006; Lauber et al., 2009; Tripathi et al., 2012). Soil texture has also been noted as having a strong association with below-ground microbial communities (Girvan et al., 2003; Lauber et al., 2008). Despite such findings, sampling across the world's land environments is still patchy and incomplete, so it is generally unclear to what extent soil texture, pH or other factors influence the composition of soil bacterial communities.

Arid and semi-arid regions occupy about 41% of the total land surface on Earth (Reynolds et al., 2007). Bacterial community structure in arid lands is relatively poorly understood compared to moister environments despite their vast areal extent. In dry environments in other parts of the world outside Mongolia, several environmental variables have previously been reported to be important in shaping soil microbial communities. Precipitation, vegetation cover, and pH have so far been found to be main factors in controlling bacterial community structure in arid and semi-arid environments (Angel et al., 2010; Fierer and Jackson, 2006). Microbial community composition also tends to be sensitive to alterations in local water content and N (Liu et al., 2009; Zhang et al., 2008). We set out here to study whether the same identifiable patterns hold true in Mongolia.

Outwardly, Mongolian steppe resembles semi-arid grasslands in other parts of the world. Grazing effects on microbial communities and physicochemical soil properties have been a principal area of study in the inner Mongolian steppe (Qi et al., 2011; Steffens et al., 2008; Zhao et al., 2009). However, little information is available on the detailed patterns of the soil bacterial community structure and diversity in either Mongolia or Inner Mongolia. Up until recently, correlations between environmental variables and bacterial community structure were explored at a broader empirical level – selected culturable species or major phyla – due to limitations in the analytical precision available. Due to next generation sequencing technologies, it is now possible to investigate bacterial community composition at a much finer scale of taxonomic resolution (Claesson et al., 2010).

In this paper, we set out to assess the patterns of bacterial diversity and community composition in a region that is previously unexplored from a metagenetic viewpoint: the semi-arid steppe region of the state of Mongolia. In particular, we assessed the overall pattern of bacterial diversity on both a within-habitat (alpha) and between-habitat (beta) scale. We then used these patterns of diversity to determine first, whether there is evidence of habitat specificity in the bacterial community within this environment, and second, to what extent it follows identifiable environmental gradients, at both the total bacterial community and different taxonomic levels such as phylum, class, and order.