Abstract Bacteriocins are antimicrobial peptides expressed by bacteria through ribosomal activity. In this study, we analyzed the diversity of bacteriocin-like genes in the Tucuruí-HPP using a whole-metagenome shotgun sequencing approach. Three layers of the water column were analyzed (photic, aphotic and sediment). Detection of bacteriocin-like genes was performed with blastx using the BAGEL4 database as subject sequences. In order to calculate the abundance of bacteriocin-like genes we also determined the number of 16S rRNA genes using blastn. Taxonomic analysis was performed using RAST server and the metagenome was assembled using IDBA-UD in order to recover the full sequence of a zoocin which had its three-dimensional structure determined. The photic zone presented the highest number of reads affiliated to bacteriocins. The most abundant bacteriocins were sonorensin, Klebicin D , pyocin and colicin. The zoocin model was composed of eight anti-parallel β-sheets and two α-helices with a Zn2+ ion in the active site. This model was considerably stable during 10 ns of molecular dynamics simulation. We observed a high diversity of bacteriocins in the Tucuruí-HPP, demonstrating that the environment is an inexhaustible source for prospecting these molecules. Finally, the zoocin model can be used for further studies of substrate binding and molecular mechanisms involving peptidoglycan degradation.