The scarcity of enzymes having an optimal activity in lignocellulose deconstruction is an obstacle for industrial-scale conversion of cellulosic biomass into biofuels. With the aim of mining novel lignocellulolytic enzymes, a ~9 Gb metagenome of bacteria in Vietnamese native goats' rumen was sequenced by Illumina platform. From the data, 821 ORFs encoding carbohydrate esterases (CEs) and polysaccharide lyases (PLs) serving for lignocellulose pre-treatment, 816 ORFs encoding 11 glycoside hydrolase families (GHs) of cellulases, and 2252 ORFs encoding 22 GHs of hemicellulases, were mined. The carbohydrate binding module (CBM) was also abundant with 763 ORFs, of which 480 ORFs are located with lignocellulolytic enzymes. The enzyme modularity analysis showed that CBMs are usually present in endoglucanase, endo 1,3-beta-D-glucosidase, and endoxylanase, whereas fibronectin 3-like module (FN3) mainly represents in GH3 and immunoglobulin-like domain (Ig) was located in GH9 only. Every domain located in each ORF was analyzed in detail to contribute enzymes' modularity which is valuable for modelling, to study the structure, and for recombinant production. With the aim of confirming the annotated results, a mined ORF encoding CBM63 was highly expressed in E. coli in soluble form. The purified recombinant CBM63 exhibited no cellulase activity, but enhanced a commercial cellulase activity in the destruction of a paper filter.
Keywords: Illumina de novo sequencing; Vietnamese native goat; carbohydrate binding model; cellulase; hemicellulase; metagenome; pretreatment.