To identify functional lignocellulases from cow rumen microbes, we generated 20 Gigabases (Gb) of metagenomic sequence from the microbial community adherent to cow rumen incubated switchgrass. Community composition analysis revealed several microbial groups without cultured representatives that were significantly enriched in the fiber adherent community. A gene-centric discovery strategy was used to identify 765, 108 and 605 full-length β-(1-4)-endoglucanases, cellobiohydrolases and β-glucosidases respectively, thereby increasing the number of urgently needed lignocellulases significantly. To verify our functional prediction and to evaluate the industrial potential of the identified lignocellulases, we cloned randomly selected candidates and tested the corresponding proteins for their cellulolytic activity. Significant cellulose degradation was observed for one-third of the tested candidates. Most interestingly, some of the tested enzymes converted Miscanthus into monomeric glucose, supporting their value for biofuel production from lignocellulosic biomass.
In summary, our deep metagenomic sequencing of biomass adherent microbes followed by computational and functional analyses dramatically expanded the number and diversity of cellulolytic genes able to degrade biomass. Importantly, the technology developed during this study can be easily adapted for other target enzymes, facilitating the large-scale identification of urgently needed biocatalysts with unique physicochemical properties.