Geographical variation in soil bacterial community structure in tropical forests in Southeast Asia and temperate forests in Japan based on pyrosequencing analysis of 16S rRNA

Geographical variation in soil bacterial community structure in 26 tropical forests in Southeast Asia (Malaysia, Indonesia and Singapore) and two temperate forests in Japan was investigated to elucidate the environmental factors and mechanisms that influence biogeography of soil bacterial diversity and composition. Despite substantial environmental differences, bacterial phyla were represented in similar proportions, with Acidobacteria and Proteobacteria the dominant phyla in all forests except one mangrove forest in Sarawak, although highly significant heterogeneity in frequency of individual phyla was detected among forests. In contrast, species diversity (alpha-diversity) differed to a much greater extent, being nearly six-fold higher in the mangrove forest (Chao' index = 6,862) than in forests in Singapore and Sarawak (similar to 1,250). In addition, natural mixed dipterocarp forests had lower species diversity than acacia and oil palm plantations, indicating that aboveground tree composition does not influence soil bacterial diversity. Shannon and Chaol indices were correlated positively, implying that skewed operational taxonomic unit (OTU) distribution was associated with the abundance of overall and rare (singleton) OTUs. No OTUs were represented in all 28 forests, and forest -specific OTUs accounted for over 70% of all detected OTUs. Forests that were geographically adjacent and/or of the same forest type had similar bacterial species composition, and a positive correlation was detected between species divergence (beta-diversity) and direct distance between forests. Both alpha- and beta-diversities were correlated with soil pH. These results suggest that soil bacterial communities in different forests evolve largely independently of each other and that soil bacterial communities adapt to their local environment, modulated by bacterial dispersal (distance effect) and forest type. Therefore, we conclude that the biogeography of soil bacteria communities described here is non-random, reflecting the influences of contemporary environmental factors and evolutionary history.