Bacterial Community Composition and Physicochemical Properties of Soils in Created and Natural Wetlands in Virginia

Abstract

Soil properties are often studied along with vegetation to examine the status of ecosystem development in created wetlands. Soil bacterial communities are essential to the biogeochemical processes in wetlands, yet are rarely examined when assessing ecosystem development in created wetlands. We used two molecular methods, amplicon length heterogeneity polymerase chain reaction (LH-PCR) and multi-tag pyrosequencing (MTPS), of 16S ribosomal DNA to characterize the bacterial communities of soils collected from two created (i.e., LC and BR) and two natural wetlands (i.e., BN and BP) during the growing and non-growing season in the Piedmont physiographic province of Virginia, USA. Soil physicochemical attributes [i.e., percent moisture, pH, soil organic matter (SOM) (%), total organic carbon (TOC) (%), total nitrogen (TN) (%), and C:N ratio were also investigated. The measures of both bacterial communities and physicochemistry of soils were tested for any association or relationship, and examined within and between these wetlands, in terms of age and hydrologic connectivity to a surrounding fluvial system. Soil moisture was significantly higher during the growing season compared to non-growing season in the wetlands, primarily due to the natural wetlands (i.e., BN and BP), and an older and hydrologically better connected created wetland (i.e., BR) holding consistently higher soil moisture over two seasons. Soil pH of the wetlands ranged between 4.2 and 5.8, typical of the acidic soils of the Piedmont. Soil organic matter content were approximately 3 to 6 % with one of the LC wetland sites (LC1) and BN being consistently higher than in the other wetland sites. TOC and TN content followed the same pattern as SOM with no major seasonal differences. Soil bacterial community patterns were more clearly distinguished with MTPS (Ribosomal Database Project level 6: genus) method compared to LH-PCR, showing the merits of the new sequencing method. Based on the MTPS results there were many significant differences found in bacterial community patterns between the wetland sites (ANOSIM R > 0.5, p < 0.05), but with no differences between created and natural wetlands. However, one natural wetland (i.e., BP) with a lower soil pH seemed to be associated with least diverse bacterial communities. Bray Curtis dissimilarities of bacterial communities from each wetland site were correlated with C:N ratio (ρ = 0.43, p < 0.01) during the non-growing season, and with pH and SOM (ρ = 0.40, p < 0.01) during the growing season. In addition, relative abundance and distribution of major bacterial taxa varied both between the wetlands and between the seasons with certain taxa more sensitively responding than the others. The results also show phyla level correlations between certain major bacterial groups with Acidobacteria negatively correlated with Bacteroides (r = -0.70, Bootstrap adjusted p < 0.05), Firmicutes (r = -0.78, Bootstrap adjusted p < 0.05) and Actinobacteria (r = -0.48, Bootstrap adjusted p<0.05). When all plots of the wetlands studied were classified into three hydrologic connectivity (HC) settings the group with the higher HC (i.e., BR3) seemed to support higher bacterial community diversity. However, BN, the natural wetland with low HC, showed as high a bacterial community diversity measure as that of BR3. This might be attributed to the comparable soil moisture content in BR3 and BN plots, which were higher than the rest of the wetland sites/plots. Further investigation is needed on the relationship between bacterial community measures and physicochemical attributes in wetland soils, which may help us develop an useful microbial community indicator that can be used to track the functional progress of wetland soils in created wetlands.