Aureobasidin A (AbA) is a cyclic despipeptide antibiotic produced by Aureobasidium pullulans R106. The native compound is a potent inhibitor of inositol phosphoceramide synthase, a key enzyme in the fungal sphingolipid biosynthesis pathway, and consequently cidal to most fungi in vitro. Native AbA is also well-tolerated and has considerably efficacy against Candida and Cryptococci infections in animals. However, significant structural modifications are required for activity against Aspergilli.

We report the characterization of the AbA biosynthetic gene. The identified aba1 comprises an intronless open reading frame of 35 kb, encoding a single nonribosomal peptide (NRP) synthetase with a predicted molecular mass of 1.3 MDa. Consistent with the structure of AbA and the molecular logic of NRP biosynthesis, the biosynthetic template consists of nine modules and shows the expected arrangement of catalytic domains, responsible for substrate recognition and activation, thiolation, peptide-bond formation, and N-methylation, as well as presumed cyclization and product release. A major unexpected finding is a very high degree of shared identity (up to 95%) among the biosynthetic modules in the complex, both at the DNA and the protein levels. Defined areas of low sequence conservation provide valuable insights into regions, potentially involved in substrate recognition. The aba1 sequence information has been used successfully for the construction of an A. pullulans aba1 disruption mutant, further validating the gene’s function. Furthermore, the genetic information is currently being exploited for targeted reprogramming efforts on the AbA biosynthetic template, aimed at generating novel AbA derivatives.