Researchers developed a genome mining tool called the Discoverer of Evolutionarily Related Natural Products (DiscERN), which uncovered a silent gene cluster that produces discomycin A, according to a Jul. 15 article in Nature Communications. The study combined multiple measures of biosynthetic similarity to offer a more targeted approach for identifying antibiotic candidates from microbial genomes.
DiscERN integrates four algorithmic approaches—Pfam Vector, BLAST Vector, BLAST Rank, and Structural K-mer Intersection—to classify biosynthetic gene clusters (BGCs) based on domain content and protein sequence similarity. These algorithms were applied to 3,561 Actinomycete genomes processed with the Antibiotics and Secondary Metabolite Analysis Shell (antiSMASH), resulting in 59,236 BGCs. Using four reference antibiotic families as benchmarks, DiscERN identified 688 putative hits for further analysis.
Manual curation classified examined clusters as true positives or false positives when possible. Leave-one-out analysis estimated recall rates for benchmark BGC families; at a confidence threshold requiring support from at least two algorithms (k ≥ 2), DiscERN achieved an F1 score of 0.89, with an estimated recall of 0.95 and precision of 0.84.
Among the BGCs identified as potentially encoding new natural products was one named discomycin (dsc) from Streptomyces kanamyceticus—a strain not previously reported to have calcium-dependent antibiotic activity but containing a regulatory SARP gene suggesting activation potential. Researchers activated this silent cluster by integrating an additional SARP copy under a strong promoter into the chromosome and expressed it in Streptomyces albus Del14, confirming production of discomycin A.
Testing showed that discomycin A exhibited potent calcium-dependent antibacterial activity against Gram-positive bacteria such as Staphylococcus aureus—including methicillin-resistant strains—and Bacillus subtilis but had no effect on Gram-negative bacteria or human colon carcinoma cells under assay conditions. Sub-inhibitory treatment led to accumulation consistent with inhibition of cell wall biosynthesis.
Benchmarking against other genome-mining platforms showed that DiscERN outperformed Genomic Assessment Tool for Orthologous Regions and Gene Clusters (GATOR-GC) and Biosynthetic Gene Similarity Clustering and Prospecting Engine (BiG-SCAPE) in identifying true positives within targeted expansion tasks. The authors said further research is needed to evaluate discomycin A's efficacy in animal models and its pharmacokinetics.
