Proteic toxin–antitoxin (TA) loci were first identified in bacterial plasmids, and they were regarded as involved in stable plasmid maintenance by a so-called ‘addiction’ mechanism. Later, chromosomally encoded TA loci were identified and their function ascribed to survival mechanisms when bacteria were subjected to stress. In the search for chromosomally encoded TA loci in Gram-positive bacteria, we identified various in the pathogen Streptococcus pneumoniae. Two of these cassettes, sharing homology with the Escherichia coli relBE locus were cloned and tested for their activity. The relBE2Spn locus resulted to be a bona fide TA locus. The toxin exhibited high toxicity towards E. coli and S. pneumoniae, although in the latter, the chromosomal copy of the antitoxin relB2Spn gene had to be inactivated to detect full toxicity. Cell growth arrest caused by expression of the relE2Spn toxin gene could be reverted by expression of the cognate antitoxin, relB2Spn, although prolonged exposition to the toxin led to cell death. The pneumococcal relBE2Spn locus is the first instance of a chromosomally encoded TA system from Gram-positive bacteria characterized in its own host. We have developed a bioluminescence resonance energy transfer (BRET) assay to detect the interactions between the RelB2Spn antitoxin and the RelE2Spn toxin in vivo. This technique has shown to be amenable to a high-throughput screening (HTS), opening new avenues in the search of molecules with potential antibacterial activity able to inhibit TA interactions.