Microbial metabolism of dead plant biomass (litter) is an important part of terrestrial carbon cycling, but it is still unclear how climate change affects microbial diversity of litter microbiome, which in turn can influence carbon storage in soils. Here we used a shotgun metagenomics approach (Illumina Hiseq 4000, PE150) to investigate the taxonomic and functional diversity of litter microbial communities across a climate gradient in southern California, representing ecosystems from desert to subalpine forest. Six sites were selected based on elevation and sampled in October of 2015. We characterized the carbohydrate decomposition potential using a trait-based approach to capture the microbial response to the climate gradient. A total of 97 carbohydrate-active enzymes (CAZymes) were chosen as functional traits, targeting cellulose, chitin, xylan, mixed carbohydrates, OPP, OAP, dextran, fructan, and oligosaccharides.
Results/Conclusions
We found that both bacterial and fungal communities (composition and relative abundance) were driven by climate gradients (temperature and precipitation). All CAZymes relevant for the complex carbohydrates decomposition were positively correlated across the litter samples. Further, the dominant genera encoding carbohydrate decomposition were generally similar among ecosystems except the saline lake site. A strong correlation between bacterial:fungal (B:F) ratios of abundance and B:F ratios of CAZymes was detected across the climate gradient. Consistent with the idea of trait conservatism, we found that functional traits involved in carbohydrate decomposition had strong phylogenetic signal, especially the microbial decomposition potential for different carbohydrates were generally identified in the same subset of microbial taxa. For example, the most abundant taxa within Streptomyces (bacteria) and Aspergillus (fungi) encoded genes targeting the decomposition of 10 carbohydrates. We also predicted the frequency of functional traits, targeting cellulose, chitin, and xylan at genus level, using the phylogenetic signal and relative abundance of each climate gradient. Our results indicate that this trait-based method might be useful to predict the response of carbon cycling to climate change.