Among the four bioactive cationic homo-poly(amino acids) discovered in nature, two are mirror-image isomers of poly(2,4diaminobutyric acid) (poly-Dab) whose biosynthesis has long been unexplained. Their structural analogy plausibly suggested that they could share a common biosynthetic pathway utilizing ε-poly(L-lysine) synthetase-like enzymology but with an unprecedented process for enantiomeric inversion of polymer building blocks. To investigate this possibility, we comparatively explored the biosynthesis of poly-L-Dab and its mirror-image isomer poly-D-Dab in Streptomyces celluloflavus USE31 and Streptoalloteichus hindustanus NBRC15115, respectively, through genome mining, genetic inactivation, and heterologous expression combined with biochemical assays. While they shared the same biosynthetic pathway, the poly-D-Dab biosynthetic gene cluster additionally harbored the racemase gene. The critical finding that poly-D-Dab synthetase, in contrast to the synthetase generating the L-isomer, selectively activated D-Dab through adenylation conclusively demonstrated that free diffusible D-Dab preactivationally generated by the racemase is directly activated to be incorporated into the polymer. Our study thus represents the first demonstration of the stereoselective biosynthesis of a nonribosomal peptide governed by adenylation activity for a D-amino acid other than alanine. In silico sequence comparison between poly-Dab synthetases allowed us to identify amino acid residues potentially responsible for the discrimination of Dab enantiomers. Our results will provide significant insight not only for the future discovery of novel bioactive cationic poly(amino acids) but also for the creation of designer nonribosomal peptides with D-configuration.

While the genome mining approach has enabled the rational exploration of untapped bioactive natural products, in silico identifications of their biosynthetic genes are often unconnected to the actual production of the corresponding molecules in native strains due to the genetic dormancy. We report here the rational discovery of an unexplored cationic homo polyamino acid (CHPA) antibiotic by potential producer prioritization-guided genome mining. Mining the genome of γ-poly-D-diaminobutyric acid (poly-D-Dab)-producing Streptoalloteichus hindustanus NBRC 15115, which was selected based on the finding that the known CHPAs are universally co-produced in pairs, identified a putative CHPA synthetase, PblA, as a potential candidate being expressed actively. Bioinformatic and biochemical analyses of PblA provided the critical clue that its polymer product could be an unusual CHPA consisting of L-β-lysine. Instrumental analyses of the metabolites from S. hindastanus indeed revealed the production of an unprecedented linear CHPA, ε-poly-L-β-lysine, concomitantly with poly-D-Dab. The CHPA we discovered exerted excellent antimicrobial activity against a broad spectrum of microorganisms, including bacteria and fungi, and was revealed to show resistance against nonspecific proteolytic enzymes. This study marks the first report of the efficacy of the strain prioritization-guided genome mining strategy for the discovery of bioactive CHPAs.