Multiple co-occurring environmental changes are affecting soil nitrogen cycling processes, which are mainly mediated by microbes. While it is likely that various nitrogen-cycling functional groups will respond differently to such environmental changes, very little is known about their relative responsiveness. Here we conducted four long-term experiments in a steppe ecosystem by removing plant functional groups, mowing, adding nitrogen, adding phosphorus, watering, warming, and manipulating some of their combinations. We quantified the abundance of seven nitrogen-cycling genes, including those for fixation (nifH), mineralization (chiA), nitrification (amoA of ammonia-oxidizing bacteria (AOB) or archaea (AOA)), and denitrification (nirS, nirK and nosZ). First, for each gene, we compared its sensitivities to different environmental changes and found that the abundances of various genes were sensitive to distinct and different factors. Overall, the abundances of nearly all genes were sensitive to nitrogen enrichment. In addition, the abundances of the chiA and nosZ genes were sensitive to plant functional group removal, the AOB-amoA gene abundance to phosphorus enrichment when nitrogen was added simultaneously, and the nirS and nirK gene abundances responded to watering. Second, for each single- or multi-factorial environmental change, we compared the sensitivities of the abundances of different genes and found that different environmental changes primarily affected different gene abundances. Overall, AOB-amoA gene abundance was most responsive, followed by the two denitrifying genes nosZ and nirS, while the other genes were less sensitive. These results provide, for the first time, systematic insights into how the abundance of each type of nitrogen-cycling gene and the equilibrium state of all these nitrogen-cycling gene abundances would shift under each single- or multi-factorial global change.

Identifying soil microbial feedbacks to increasing temperatures and moisture alterations is critical for predicting how terrestrial ecosystems will respond to climate change. We performed a 5-year field experiment manipulating warming, watering and their combination in a semiarid temperate steppe in northern China. Warming stimulated the abundance of genes responsible for degrading recalcitrant soil organic matter (SOM) and reduced SOM content by 13%. Watering, and warming plus watering also increased the abundance of recalcitrant SOM catabolism pathways, but concurrently promoted plant growth and increased labile SOM content, which somewhat offset SOM loss. The treatments also increased microbial biomass, community complexity and metabolic potential for nitrogen and sulfur assimilation. Both microbial and plant community composition shifted with the treatment conditions, and the sample-to-sample compositional variations of the two communities (pairwise β-diversity distances) were significantly correlated. In particular, microbial community composition was substantially correlated with the dominant plant species (~0.54 Spearman correlation coefficient), much more than with measured soil indices, affirming a tight coupling between both biological communities. Collectively, our study revealed the direction and underlying mechanisms of microbial feedbacks to warming and suggested that semiarid regions of northern steppes could act as a net carbon source under increased temperatures, unless precipitation increases concurrently.