To obtain new amidases of biocatalytic relevance, we used microorganisms indigenous to different types of soil and sediment as a source of DNA for the construction of environmental gene banks, following two different strategies. In one case, DNA was isolated from soil without preceding cultivation to preserve a high degree of (phylo)genetic diversity. Alternatively, DNA samples were obtained from enrichment cultures, which is thought to reduce the number of clones required to find a target enzyme. To selectively sustain the growth of organisms exhibiting amidase activity, cultures were supplied with a single amide or a mixture of different aromatic and non-aromatic acetamide and glycine amide derivatives as the only nitrogen source. Metagenomic DNA was cloned into a high-copy plasmid vector and transferred to E. coli, and the resulting gene banks were searched for positives by growth selection. In this way, we isolated a number of recombinant E. coli strains with a stable phenotype, each expressing an amidase with a distinct substrate profile. One of these clones was found to produce a new and highly active penicillin amidase, a promising biocatalyst that may allow higher yields in the enzymatic synthesis of beta-lactam antibiotics.