INTRODUCTION
Microorganisms, especially bacteria and fungi, contribute enormously to terrestrial ecosystem services: for example, by playing a vital role in soil nutrient cycling (
1–4). Particularly, the contribution of plant symbiotic microbes in soil nutrient cycling has been well reported. For example, mycorrhizal fungi form symbiotic associations with around 90% of terrestrial plant species and take part in nutrient cycling by mobilizing nitrogen (N) and phosphorus (P) in soils (
5,
6). Similarly, plant-symbiotic bacteria belonging to
Rhizobium and
Frankia can fix nitrogen and thus essentially participate in N cycling (
7). Moreover, at the community level, it is also important to consider the extensive contribution of free-living soil bacteria and fungi to soil nutrient cycling as they constitute a major part of soil microbiota (
8). A few examples include carbon-fixing
Actinobacteria (
9,
10), nitrogen-fixing
Azotobacter (
9,
10), and phosphate-solubilizing
Acidobacteria (
11,
12). Likewise,
Penicillium,
Aspergillus, and
Trichoderma are free-living fungi and known for being actively involved in the decomposition of soil organic compounds (C cycle), nitrification (N cycle), and P solubilization (P cycle), respectively (
13–15).
Soil stoichiometry of nutrients like C/N/P ratios is known to affect the soil microbial communities, depending upon their constituting members’ organismal nutrient stoichiometric ratios (
16,
17). For example, it was reported that high N and P abundances in soil favor the abundance of fast-growing bacteria (i.e., copiotrophic,
r-strategists) like
Actinobacteria and
Alphaproteobacteria while discriminating against slow-growing bacteria (i.e., oligotrophic,
K-strategists) like
Acidobacteria (
18,
19). Also, previous research suggests that ectomycorrhizal fungi (EMF) preferentially associate with soils of high-C/N substrates, whereas saprotrophic fungi prevail in soils with low C/N ratios (
20–22). There has been a surge in recent studies showing the link between microbial diversity, community composition, and soil ecosystem multifunctionality (
23–27). However, there is still a knowledge gap about how the soil microbial communities vary in the stoichiometry of their nutrient cycling genomic potential, which can be the relative combinations of genes coding for different nutrient cycling enzymes. In a study taking a genomic perspective on soil carbon cycling, Hartman et al. (
28) reported links between microbial community composition, the microbe’s C, N, and P substrate utilization potential, and C turnover. This highlights the importance of studying the genomic potential of microbial communities to better understand soil nutrient cycling.
Given the fact that soil C, N, and P cycles are linked, it is essential to study the co-occurring bacterial and fungal communities together for their genomic potential in the cycling of different major nutrients and their combinations (
viz. C, N, P, CN, CP, NP, and CNP). For instance, the ability to decompose soil organic matter (SOM) with various nutrient ratios depends on the composition of soil microbial communities (
29). Subsequently, the decomposed SOM would be available for bacteria and fungi conditioned on their abilities to continue with either N fixation or denitrification (
30,
31) and/or concurrently also be available for P mineralization or solubilization (
32,
33). This linkage between different soil nutrient cycling processes and the different microbes involved can be viewed from a “microbial syntrophy” (microbial metabolic interrelationships) perspective (
34), which is affected by many factors (for example, available nutrient ratios, etc.) but essentially depends on the genomic potential of the members of the microbial communities.
The ecological processes and relationships within a microbial community can cumulatively emerge from the constituting microbial groups/clusters (i.e., taxa that are more strongly associated within that group than with other groups), which are also known as subcommunities (
35,
36). Based on network theory, the study of subcommunities, also known as modules, can provide key insights into the overall functioning of the microbial community, allowing us to assess the metabolic potential based on the single microbes’ functional roles, which otherwise remains a black box. In addition, knowledge of subcommunities also sheds light on the ecological processes that shape and regulate community structure and organization, such as environmental filtering or niche differentiation (
37). For example, recent studies in soil microbial ecology have taken the advantage of subcommunity-based analyses to develop a deeper understanding of environment-specific relationships (
38,
39) and the functional roles of microbial communities (
40–42).
One of the key factors influencing the soil microbial communities in forests is the tree mycorrhizal type (
43), which is also known to impact microbial functional genes (
44) and soil nutrient cycling (
45). In addition, tree diversity has also been reported to affect soil microbial communities (
46–48) and soil nutrient availability (
49). Despite these efforts, there is still a great need to understand how the tree mycorrhizal type and tree diversity affect the co-occurring soil bacterial and fungal communities at the subcommunity level and, in consequence, their genomic functional potential for nutrient cycling. Insight into these processes would provide a broader understanding of the intrinsic characteristics of soil microbial groups operating in ecological processes and the functional potential emerging at the community level. Such in-depth mechanistic understanding would also be the basis for managing forest soil ecosystems to maintain or increase forest multifunctionality.
To fill this knowledge gap, this study was conducted at the BEF-China experimental research platform (
50), using tree species of two mycorrhizal types, namely, ectomycorrhizal (EcM) and arbuscular mycorrhizal (AM), at different tree diversity levels (
43). We employed the fungal-bacterial interkingdom co-occurrence network approach (
51) to derive the microbial subcommunities (here, interchangeably used with “modules”) and used PICRUSt2 (
52) to predict the potential genomic functions with regard to nutrient cycling from the amplicon sequencing data. Our main objective was to understand how the stoichiometry in genomic functional potential of soil microbial communities and their subcommunities with regard to the three major nutrient cycles and their combinations (C, N, P, CN, CP, NP, and CNP) varies in EcM and AM trees at different tree diversity levels. In particular, we asked the following research questions.
1.
How do the EcM and AM tree species pair (TSP) soil bacterial and fungal community co-occurrence network structures differ across tree diversity levels, and which soil characteristics drive the composition of the subcommunities in these networks?
2.
What are the effects of tree diversity and tree mycorrhizal type on the predicted genomic functional potential (in terms of C, N, and P cycles and their combinations) of the co-occurring bacterial and fungal communities?
3.
How do EcM and AM TSPs soil microbial subcommunities differ in their genomic functional abundances in the three nutrient cycles and their combinations within the tree diversity levels, and which microbial taxa drive these differences?