4.2 Divergent drivers and co-occurrence patterns for microbial communities along the chronosequence
In the past, researchers mainly focused on how restoration stages or forest ages influence soil microbial diversity, but rarely reported microbial co-occurrence patterns (especially in the rhizosphere) and the divergent drivers along a chronosequence, for example, in R. pseudoacacia plantations. In the present study, we observed that from 25 years to 45 years of R. pseudoacacia forest age, different clusters of rhizosphere microbial communities formed distinct modules in the microbial network. Modulization indicates habitat heterogeneity, aggregation of phylogenetically related species, niche overlap, and species coevolution (41). Therefore, a similar module means a similar habitat, and species in the same module due to niche overlap have more interactions with each other. Such an interaction may in the form of cooperation or competition for resources. Conversely, there are relatively less interactions between modules, probably due to isolation, which means that the co-occurrence of species is inhibited. Our results suggest that clusters located on the same branch of a phylogenetic tree form unified modules in the microbial network, which indicates that similarity in rhizosphere microbial community structure plays a role in determining the modularized structure of the microbial network.
We observed tight interactions and close aggregations among microbes in Module 1 and Module 2 for both bacteria and fungi in the rhizosphere soils in the R. pseudoacacia plantation. However, Module 1 and Module 2 were not fully separated from Module 3, indicating that forest age was not the only factor driving microbial community module differentiation. In addition, the relationship between rhizosphere soil multinutrient index and bacterial and fungal taxa relative abundance varied in different modules. Consequently, rhizosphere soil functions potentially influenced microbial co-occurrence patterns in different modules; the microbial responses to rhizosphere soil functions varied over time and space. Furthermore, we observed that different forest ages targeted specific fungal lineages, but not bacterial lineages. Not only the stability but also the connectivity of the bacterial network was higher than those of the fungal network.
Notably, the microbial co-occurrence networks illustrated the correlations among different taxa, including real ecological interactions and non-random processes; therefore, the network may not really reflect the direct interactions between different groups (13, 40). Because habitats have complex regulatory mechanisms, microbial networks are useful tools for exploring complex microbial community structures, particularly in rhizosphere microenvironments with complex interactions. Microbial networks may also facilitate the exploration of the interactions among multi-level biological systems in the future. Long-term observations are still required to select microbes with specific functions that could facilitate plant selection of mutually beneficial microbiota, which would enhance plant adaptation without adversely affecting plant growth.