In this study, a high-throughput sequencing approach was applied to discover novel biocatalysts for lignocellulose hydrolysis from three dedicated energy crops, Arundo donax, Eucalyptus camaldulensis and Populus nigra, after natural biodegradation. The microbiomes of the three lignocellulosic biomasses were dominated by bacterial species (approximately 90%) with the highest representation by the Streptomyces genus both in the total microbial community composition and in the microbial diversity related to GH families of predicted ORFs. Moreover, the functional clustering of the predicted ORFs showed a prevalence of poorly characterized genes, suggesting these lignocellulosic biomasses are potential sources of as yet unknown genes. 1.2%, 0.6% and 3.4% of the total ORFs detected in A. donax, E. camaldulensis and P. nigra, respectively, were putative Carbohydrate-Active Enzymes (CAZymes). Interestingly, the glycoside hydrolases abundance in P. nigra (1.8%) was higher than that detected in the other biomasses investigated in this study. Moreover, a high percentage of (hemi)cellulases with different activities and accessory enzymes (mannanases, polygalacturonases and feruloyl esterases) was detected, confirming that the three analyzed samples were a reservoir of diversified biocatalysts required for an effective lignocellulose saccharification.