Abstract:Caves are extreme environments with permanent darkness and limited nutrients, which serve as natural laboratories to study the subsurface deep biosphere. Although microbial communities have demonstrated strong niche specificity in different caves, the environmental driving mechanism of microbial communities and the ecological processes responsible for community assembly in different niches particularly those in different substrates e. g. solid versus liquid samples were poorly understood.[Objective] Here we aim to explore the environmental driving mechanisms and community assembly process of bacterial communities in different substrates e. g. solid versus liquid samples combined with physicochemical properties.[Methods] To this end, a karst cave, the Luohandu Cave in Guilin city, Guangxi province, locating in the typical karst region in southwestern China, was selected for a systematic investigation of microbial communities. Solid samples (weathered rocks and sediments) and liquid samples (dripping water, pool water and ground river water samples) were collected along the cave and subjected to high-throughput sequencing of 16S rRNA gene. Meanwhile, physicochemical properties of these samples were analyzed.[Results] Results showed strong niche specificity of microbial communities in the Luohandu cave. Actinobacteria and Acidobacteria dominated bacterial communities of weathered rocks and sediments, respectively. In contrast, water samples were dominated by γ-Proteobacteria. Redundancy analysis (RDA) between microbial communities and environmental variables demonstrated that temperature, weathering index and SO42- concentration significantly affect microbial communities in solid samples. USCγ and Pseudomonas positively associated with temperature, while Pseudonocardia, Solirubrobacter and Gemmatimonas negatively correlated with temperature. However, microbial communities in water samples were significantly controlled by electrical conductivity (EC) and dissolved oxygen (DO) as indicated by RDA. The co-occurrence network of bacterial communities was characterized by a good modularity with nodes from liquid samples locating in the same module, indicating a niche preference. Positive links dominated in the network, suggesting a corporative strategy among different microbial groups to survive in caves. Different ecological processes were found to be responsible for bacterial community assembly in different substrates. Deterministic process (48.75%) and stochastic process (51.25%) contributed almost equally to bacterial community assembly in solid samples, whereas stochastic process (64.76%) dominated the microbial community assembly in liquid samples. As for the individual ecological processes, homogenizing dispersal, undominated and dispersal limit in stochastic process contributed 40.42%, 10.46% and 3.13% to community assembly respectively; whereas homogenous selection and heterogeneous selection within deterministic process contributed 26.75% and 21.99%, respectively, in solid samples. In liquid samples, individual processes contributed to community assembly in order from high to low were undominated (28.57%), homogenous selection (25.71%), homogenizing dispersal (24.76%), dispersal limitation (10.48%), and heterogeneous selection (9.52%).[Conclusion] Our results for the first time revealed the different environmental driving mechanisms and different community assembly processes underlining microbial communities living in different substrates (solid versus liquid). This study offers a new window to understand the interactions among microbes and their environments, and interactions among different microbial groups in the subsurface biosphere in caves.
