Abstract
Soils harbor the most diverse naturally evolved antibiotic resistomes on Earth that threaten human health, ecosystem processes, and food security. Yet the importance of spatial and temporal variability in shaping the distribution of soil resistomes is not well explored. Here, a total of 319 topsoil samples were collected at a watershed scale during four seasons (spring to winter) and high-throughput quantitative PCR (HT-qPCR) was used to characterize the profiles of soil antibiotic resistance genes (ARGs). A significant and negative correlation was observed between soil ARG profiles and seasonal dissimilarity, which along with seasonally dependent distance-decay relationships highlight the importance of seasonal variability in shaping soil antibiotic resistomes. Significant, though weak, distance-decay relationships were identified in spring, summer and winter, for ARG similarities with geographic distances. There were also strong interactions between specific soil ARGs and Actinobacteria, Firmicutes and Proteobacteria. Moreover, we found that the relative abundance of soil Actinobacteria, Firmicutes and Proteobacteria correlated significantly with annual mean temperature and annual mean precipitation at a watershed scale. A random forest model showed that seasonal change rather than spatial variation was the most important predictor of the composition of soil ARGs. Together, these results constitute an advance in our understanding of the relative importance of spatial and temporal variability in shaping soil ARG profiles, which will provide novel insights allowing us to forecast their distribution under a changing environment.