ABSTRACT
Semi-arid and arid areas occupy about 33% of terrestrial ecosystems. However, little information is available about microbial diversity in the semi-arid Caatinga, which represents a unique biome that extends to about 11% of the Brazilian territory and is home to extraordinary diversity and high endemism level of species. In this study, we characterized the diversity of microbial genes associated with biomass conversion (carbohydrate-active enzymes, or so-called CAZYmes) in soil and freshwater of the Caatinga. Our results showed distinct CAZYme profiles in the soil and freshwater samples. Glycoside hydrolases and glycosyltransferases were the most abundant CAZYme families, with glycoside hydrolases more dominant in soil (â¼44%) and glycosyltransferases more abundant in freshwater (â¼50%). The abundances of individual glycoside hydrolase, glycosyltransferase, and carbohydrate-binding module subfamilies varied widely between soil and water samples. A predominance of glycoside hydrolases was observed in soil, and a higher contribution of enzymes involved in carbohydrate biosynthesis was observed in freshwater. The main taxa associated with the CAZYme sequences were Planctomycetia (relative abundance in soil, 29%) and Alphaproteobacteria (relative abundance in freshwater, 27%). Approximately 5-7% of CAZYme sequences showed low similarity with sequences deposited in non-redundant databases, suggesting putative homologues. Our findings represent a first attempt to describe specific microbial CAZYme profiles for environmental samples. Characterizing these enzyme groups associated with the conversion of carbohydrates in nature will improve our understanding of the significant roles of enzymes in the carbon cycle. We identified a CAZYme signature that can be used to discriminate between soil and freshwater samples, and this signature may be related to the microbial species adapted to the habitat. The data show the potential ecological roles of the CAZYme repertoire and associated biotechnological applications.