Background: Microbial-driven decomposition of plant residues is integral to carbon sequestration in terrestrial ecosystems. Actinobacteria, one of the most widely distributed bacterial phyla in soils, are known for their ability to degrade plant residues in vitro. However, their in situ importance and specific activity across contrasting ecological environments are not known. Here, we conducted three field experiments with buried straw in combination with microcosm experiments with ¹³C-straw in paddy soils under different soil fertility levels to reveal the ecophysiological roles of Actinobacteria in plant residue decomposition. Results: While accounting for only 4.6% of the total bacterial abundance, the Actinobacteria encoded 16% of total abundance of carbohydrate-active enzymes (CAZymes). The taxonomic and functional compositions of the Actinobacteria were, surprisingly, relatively stable during straw decomposition. Slopes of linear regression models between straw chemical composition and Actinobacterial traits were flatter than those for other taxonomic groups at both local and regional scales due to holding genes encoding for full set of CAZymes, nitrogenases, and antibiotic synthetases. Ecological co-occurrence network and ¹³C-based metagenomic analyses both indicated that their importance for straw degradation increased in less fertile soils, as both links between Actinobacteria and other community members and relative abundances of their functional genes increased with decreasing soil fertility. Conclusions: This study provided DNA-based evidence that non-dominant Actinobacteria plays a key ecophysiological role in plant residue decomposition as their members possess high proportions of CAZymes and as a group maintain a relatively stable presence during plant residue decomposition both in terms of taxonomic composition and functional roles. Their importance for decomposition was more pronounced in less fertile soils where their possession functional genes and interspecies interactions stood out more. Our work provides new ecophysiological angles for the understanding of the importance of Actinobacteria in global carbon cycling.

With the continuous expansion of saline soils under climate change, understanding the eco-evolutionary tradeoff between the microbial mitigation of carbon limitation and the maintenance of functional traits in saline soils represents a significant knowledge gap in predicting future soil health and ecological function. Through shotgun metagenomic sequencing of coastal soils along a salinity gradient, we show contrasting eco-evolutionary directions of soil bacteria and archaea that manifest in changes to genome size and the functional potential of the soil microbiome. In salt environments with high carbon requirements, bacteria exhibit reduced genome sizes associated with a depletion of metabolic genes, while archaea display larger genomes and enrichment of salt-resistance, metabolic, and carbon-acquisition genes. This suggests that bacteria conserve energy through genome streamlining when facing salt stress, while archaea invest in carbon-acquisition pathways to broaden their resource usage. These findings suggest divergent directions in eco-evolutionary adaptations to soil saline stress amongst microbial clades and serve as a foundation for understanding the response of soil microbiomes to escalating climate change.

Summary: Ecological assembly processes, by influencing community composition, determine ecosystem functions of microbiomes. However, debate remains on how stochastic versus deterministic assembly processes influence ecosystem functions such as carbon and nutrient cycling. Towards a better understanding, we investigated three types of agroecosystems (the upland, paddy, and flooded) that represent a gradient of stochastic versus deterministic assembly processes. Carbon and nutrient cycling multifunctionality, characterized by nine enzymes associated with soil carbon, nitrogen, phosphorous and sulfur cycling, was evaluated and then associated with microbial assembly processes and co‐occurrence patterns of vital ecological groups. Our results suggest that strong deterministic processes favour microorganisms with convergent functions (as in the upland agroecosystem), while stochasticity‐dominated processes lead to divergent functions (as in the flooded agroecosystem). To benefit agroecosystems services, we speculate that it is critical for a system to maintain balance between its stochastic and deterministic assembly processes (as in the paddy agroecosystem). By doing so, the system can preserve a diverse array of functional traits and also allow for particular traits to flourish. To further confirm this speculation, it is necessary to develop a systematic knowledge beyond merely characterizing general patterns towards the associations among community assembly, composition, and ecosystem functions.