Soil microbiome network structure is highly related to nutrient cycling and crop growth in agroecosystems. However, how agricultural practices influence soil microbial network structure and interact with co-varying soil properties, crop yield, and other microbial properties is still unclear. In the present study, soil physicochemical properties, microbial biomass and enzyme activity, and agronomic traits of tobacco were measured across seven fertilization treatments (Conventional Fertilization [CF], Incremental Fertilization [IF] 10 %, Reduced Fertilization [RF] 10 %, RF 20 %, CF combined with Rice Straw [CF + RS], RF 10 % combined with Green Manure [RF 10 % + GM], and No Fertilization [NF]) and two cropping treatments (continuous tobacco and corn-tobacco rotation). Microbial diversity and network structure were also characterized using Illumina MiSeq sequencing and network analysis. According to the results, cropping and fertilization treatments significantly affected twelve soil physicochemical properties and thirteen microbial properties. Among the microbial properties, the proportion of module hubs was most strongly correlated with tobacco yield (Pearson r = 0.613). Soil pH, cation exchange capacity (CEC), soil bacterial richness and module hubs constituted the best combination of variables that cumulatively explained 63.1 % of the variation in tobacco yield, among which module hubs individually explained 18.5 % of the variation in yield. Furthermore, the piecewise structural equation model showed that module hubs in microbial network were significantly affected by agricultural practices, and they indirectly drove tobacco yield by regulating soil properties. Overall, the results of the present study suggest that microbial properties (particularly network-associated keystone species) can mediate the effects of soil properties on crop yield, although the main influence or limiting factor for yield is still soil properties.

中文翻译：

---

施肥和轮作通过调节土壤物理化学和微生物特性来提高烟草产量


土壤微生物组网络结构与农业生态系统中的养分循环和作物生长高度相关。然而，农业实践如何影响土壤微生物网络结构并与协同变化的土壤特性、作物产量和其他微生物特性相互作用仍不清楚。在本研究中，通过七种施肥处理（常规施肥 [CF]、增量施肥 [IF] 10%、减施肥 [RF] 10%、RF 20 %、CF 结合稻草 [CF + RS]、RF 10 % 结合绿肥 [RF 10 % + GM] 和不施肥 [NF]）和两种种植处理（连续烟草和玉米-烟草轮作）测量了土壤理化特性、微生物生物量和酶活性以及烟草的农艺特性。还使用 Illumina MiSeq 测序和网络分析对微生物多样性和网络结构进行了表征。结果表明，种植施肥处理显著影响了 12 种土壤理化性质和 13 种微生物性质。在微生物特性中，模块枢纽的比例与烟草产量最密切相关 （Pearson r = 0.613）。土壤 pH 值、阳离子交换能力 （CEC）、土壤细菌丰富度和模块中心构成了累积解释 63.1% 的烟草产量变化的变量的最佳组合，其中模块中心单独解释了 18.5% 的产量变化。此外，分段结构方程模型表明，微生物网络中的模块枢纽受到农业实践的显着影响，它们通过调节土壤特性间接驱动烟草产量。 总体而言，本研究的结果表明，微生物特性（特别是与网络相关的关键物种）可以介导土壤特性对作物产量的影响，尽管产量的主要影响或限制因素仍然是土壤特性。