The 16S ribosomal RNA (rRNA) gene has been the most commonly used sequence to characterize bacterial communities. The classical approach to obtain gene sequences to study bacterial diversity implies cloning amplicons, selecting clones, and Sanger sequencing cloned fragments. A more recent approach is direct sequencing of millions of genes using massive parallel technologies, allowing a large-scale biodiversity analysis of many samples simultaneously. However, currently, this technique is still expensive when applied to few samples; therefore, the classical approach is still used. Recently, we found a community able to remove 50 mg/L Pd(II). In this work, aiming to identify the bacteria potentially involved in Pd(II) removal, the separation of urea/heat-denatured DNA fragments by urea-agarose gel electrophoresis was applied for the first time to select 16S rRNA-cloned amplicons for taxonomic studies. The major raise in the percentage of bacteria belonging to genus Clostridium sensu stricto from undetected to 21 and 41 %, respectively, for cultures without, with 5 and 50 mg/L Pd(II) accompanying Pd(II) removal point to this taxa as a potential key agent for the bio-recovery of this metal. Despite sulfate-reducing bacteria were not detected, the hypothesis of Pd(II) removal by activity of these bacteria cannot be ruled out because a slight decrease of sulfate concentration of the medium was verified and the formation of PbS precipitates seems to occur. This work also contributes with knowledge about suitable partial 16S rRNA gene regions for taxonomic studies and shows that unidirectional sequencing is enough when Sanger sequencing cloned 16S rRNA genes for taxonomic studies to genus level.