Microbial habitat specificity largely affects microbial co-occurrence patterns and functional profiles in wetland soils
Introduction
Soil microorganisms play pivotal roles in biogeochemical cycles. Exploring biogeographic patterns of soil microbial communities and the underlying mechanisms dominating community assembly are essential for understanding ecosystem functions (Falkowski et al., 2008, Graham et al., 2016). Unraveling microbial community assembly mechanisms requires a sophisticated understanding of niche-based and neutral theories (Anderson et al., 2011, Hanson et al., 2012, Leibold, 1995, Stegen et al., 2013, Vellend et al., 2014, Zhou and Ning, 2017). Niche-based theory, corresponding to deterministic processes, hypothesizes that abiotic factors (e.g., pH and salinity) and biological interactions control community structure and dynamics (Faust and Raes, 2012). The biological foundations of both the abiotic and biotic selections depend on the microbial traits, including phenotypic, metabolic and other functional traits. In contrast, neutral theory assumes that stochastic processes (e.g., birth, death, immigration and emigration) maintain community diversity, independent of species traits and environmental factors (Vellend, 2010).
The selection effects imposed by deterministic factors in a community can alter from one species to another, which can be attributed to the optimized niche and niche breadth of species. In the evolution of a species, one trait may be invested at the cost of others theoretically. Niche breadth represents a kind of evolutionary trade-off among traits on resource utility and stress tolerance, which constrains the taxa not to be widely distributed, especially in very heterogeneous habitats (Barberán et al., 2014, Goberna et al., 2015, Morrissey and Franklin, 2015, Treseder et al., 2011). For instance, bacteria with a large genome size are prone to be generalists (species with broad distribution range or high occupancy rate), mainly due to their various metabolic abilities and strong stress tolerance (Barberán et al., 2014). Microbial taxa considered specialists are optimized for particular traits and have a narrow niche breadth (high specificity). The narrow niche breadth of specialists make them dwell in the relatively unique habitat. For instance, Rhodobacter capsulatus has a narrow pH range (6.5–7.5) for the growth in comparison with Rhodobacter sphaeroides (pH range: 6–8.5) (Imhoff, 2015). In terms of the metabolic capabilities, Syntrophobotulus glycolicus is a metabolic specialist as it only grows chemotrophically by fermentative oxidation of glyoxylate (Schink and Friedrich, 2015). Specialists may account for a large percentage of microbial community members, irrespective of the habitats’ characteristics (Mariadassou et al., 2015). Generally, specialists can be affected by more deterministic factors than generalists (Liao et al., 2016, Székely and Langenheder, 2014). Since specialists have narrow habitat range, one specialist may have low existence frequency at large geographic scale and also low co-occurrence frequency with other specialists and generalists. Unlike generalists, specialists are abundant only under appropriate ambient conditions (Fodelianakis et al., 2017). Furthermore, in a community, the ratio of specialists profoundly influences community assembly and dynamics. If one community consists of a larger proportion of specialists than another, their co-occurrence patterns may also be distinct because of the differences of traits and functions between specialists and generalists (Bell and Bell, 2020). Accordingly, it is essential to evaluate the relationships between species habitat specificity and species co-occurrence pattern at community level. In addition, it is largely unknown as to how functional profiles of species shape the extent of their habitat specialization and contribute to the community assembly. Functional differences of species can be vital in deciphering how specialists and generalists function and how they co-exist in communities.
Soil is extremely complicated ecosystem, and exploring the properties of microbial community in it is very difficult with traditional approaches in community ecology. As a systematic approach, network inference has been applied in soil biogeography to explore co-occurrence patterns of microbial community in recent years (Barberan et al., 2012, Bissett et al., 2013, Ma et al., 2016). The topological properties of an ecosystem network, such as modularity, average degree and complexity, can indicate some community properties not easily captured by other statistical approaches (Boccaletti et al., 2006, Deng et al., 2012, Ma et al., 2016, Montoya et al., 2006, Shi et al., 2016). Co-occurrence patterns have been explored in elucidating the relationships that exist between pairs of microbial taxa in diverse ecosystems (Ma et al., 2020a, Williams et al., 2014). Microbial network can be understood as a structure composed of nodes and edges, which represents a temporary or spatial snapshot of an ecosystem (Röttjers and Faust, 2018). Microbial co-occurrence patterns are related to many factors, such as niche similarity and differentiation, functional complementation, predation and competition. This type of approach has proved useful in studying those systems with complex species interactions and allows the inference of gene-regulatory and other complex patterns (Matchado et al., 2021). Thus, the co-occurrence network is an appealing approach that allows a comprehensive analysis and understanding of assembly and functioning of biological communities (Xue et al., 2022). Rationally using network method can help us decipher more ecological patterns on microbial species in biogeographic studies (Goberna and Verdú, 2022). So far, more network construction approaches have been developed to efficiently unveil the associated patterns among taxa based on the abundance data instead of pure correlation network (Matchado et al., 2021, Röttjers and Faust, 2018). For instance, FlashWeave method (Tackmann et al., 2019) is developed using a learning framework based on the probabilistic graphical models (PGM) and has shown good performance on heterogenous datasets to find the direct associations among taxa (Matchado et al., 2021).
Wetlands are carbon (C) sinks and play an important role in the global C-cycle and other elements cycles (Zhou et al., 2021). Flooded or saturated conditions in these ecosystems limit the availability of oxygen to soil microbes, and decomposition of organic matter proceeds slowly under the waterlogged conditions (Keller, 2011). Soils at the junction of land and water are susceptible to water levels, and microbial communities may be frequently affected by changes in redox conditions and nutrients and energy availability. To study prokaryotic community structure, we first used the 16S rRNA gene amplicon sequencing dataset of a previous research (An et al., 2019). We also performed shotgun metagenomic sequencing to examine the gene and metabolic pathway profiles of communities in Chinese wetland soils (Fig. S1). Then, we calculated species habitat specificity and constructed a co-occurrence network. To establish the relationships between species habitat specificity and co-occurrence patterns at the community level, we calculated community specificity by averaging the specificity of species in a community and community connectivity to represent the complexity of species co-occurrence patterns in the community. Finally, we analyzed whether abiotic factors (e.g., conductivity and pH), functional guilds, and metabolic pathways were related to community structure. We had two hypotheses (Fig. 1): 1. Communities with more specialists have weaker community connectivity compared to those with fewer specialists. 2. Abiotic factors and functional profiles are both important in shaping community specificity-connectivity relationships in wetland soils.
Section snippets
Soil sampling and soil physicochemical characteristics
Soil samples were collected under waterlogged conditions at the junction of land and water from 49 wetland areas in China from June to October 2013 (An et al., 2019). Detailed soil sampling information was provided in An et al. (2019) (An et al., 2019). Briefly, samples were taken from the topsoil layer (soil core diameter: 5 cm, depth: 10 cm). A total of 200 samples were used, with at least three samples from each site. Based on GPS information, the mean annual temperature (MAT) and mean annual
Co-occurrence network
The co-occurrence network based on probabilistic graphical models (PGM-network) was first constructed to explore the community structure in prokaryotic communities. Total 200 samples and 2203 OTUs (mean abundance > 0.01%) were used for the network construction. There were 3978 positive and 429 negative linkages among 2128 OTUs from different phyla or within the same phylum (Fig. 2a and Fig. S2). Proteobacteria had the largest number of linkages to other phyla, including 1391 positive and 229
Co-occurrence patterns reveal important ecosystem roles of both specialists and generalists
In this study, we first evaluated OTU’s specificity and their co-occurrence network, and we found that some specialists and generalists both hold important roles in the co-occurrence network. For example, in the generalists that were classified into module or network hubs (Fig. 2b), OTU_1112 and OTU_496 were both classified into Acidovorax genus, the nitrate-dependent Fe oxidizing bacteria which could be crucial denitrifiers in the wetland ecosystems (Eugene Rosenberg and DeLong, 2013). The
Conclusions
This study revealed that more specialists in the microbial community correspond to more simple species co-occurrence patterns under a large biogeographic scale. This pattern was closely associated with the biological mechanisms, including the differences of species genes, traits and survival strategies. Our conclusions provide new ideas for simplifying the predictions of species co-occurrence patterns based on the species niche breadth, since the microbial niche breadth is relatively constant
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (42077206, 32071548); and China Biodiversity Observation Networks (Sino BON). We thank the two anonymous reviewers and the editor for their helpful suggestions on the manuscript. We thank Naqu station of Tibet University and Institute of Tibetan Plateau Research (Chinese Academy of Sciences, CAS), and Inner Mongolia Grassland Research Station, Institute of Botany (CAS), for the help with sample collection. We thank
References (106)
- et al.
Soil bacterial community structure in Chinese wetlands
Geoderma
(2019) - et al.
Complex networks: structure and dynamics
Phys. Rep.
(2006) Conservation evaluation and phylogenetic diversity
Biol. Conserv.
(1992)- et al.
Cautionary notes on the use of co-occurrence networks in soil ecology
Soil Biol. Biochem.
(2022) - et al.
Combining centrality indices: Maximizing the predictability of keystone species in food webs
Ecol. Indic.
(2021) - et al.
Network analysis methods for studying microbial communities: A mini review
Comput. Struct. Biotechnol. J.
(2021) - et al.
Wetland management using microbial indicators
Ecol. Eng.
(2017) - et al.
Land-use affects soil microbial co-occurrence networks and their putative functions
Appl. Soil Ecol.
(2022) - et al.
Navigating the multiple meanings of beta diversity: a roadmap for the practicing ecologist
Ecol. Lett.
(2011) - et al.
Using network analysis to explore co-occurrence patterns in soil microbial communities
ISME J.
(2012)