We show that selection of drug-resistant bacterial mutants allows the discovery of antibacterial compounds. Mutant strains of a soil-isolated Streptomyces species that does not produce antibacterials synthesize a previously unknown class of antibacterial, which we name piperidamycin. Overall, 6% of non-Streptomyces actinomycetes species and 43% of Streptomyces species that do not produce antibacterials are activated to produce them. The antibacterial-producing mutants all carried mutations in RNA polymerase and/or the ribosomal protein S12.

Natural products are the major source of currently available drugs. However, screening natural product presents several challenges, including the time-consuming and labor-intensive steps required for the isolation of a drug from crude extracts as well as the differences between the activities of compounds in vitro and in vivo. To address these challenges, we used silkworm larvae infected with Aspergillus fumigatus to screen a natural products library for potent drugs to treat invasive aspergillosis. A rationally designed library was constructed using numerous, geographically diverse fungal species and then screened to collect extracts of microorganisms that had detectable anti-Aspergillus activity. We evaluated this library using cultures of A. fumigatus and a silkworm model system of A. fumigatus infection. With this model, we identified the novel antifungal compound ASP2397 that not only cured infected silkworm larvae but also increased the rates of survival of mice infected with A. fumigatus. These findings strongly support the utility of the silkworm screening system for the simple and rapid isolation of antibiotics from natural products libraries.

The novel antifungal agent ASP2397 (Vical's compound ID VL-2397) is produced by the fungal strain MF-347833 that was isolated from Malaysian leaf litter and is identified here as an Acremonium species based on its morphology, physiological properties and 28S ribosomal DNA sequence. Because of its potential importance for producing novel antifungal agents, we determined the taxonomic and biologic properties of MF-347833. We show here that ASP2397 is a cyclic hexapeptide that chelates aluminum ion and is therefore similar to ferrichrome, a hydroxamate siderophore. However, ASP2397 differs structurally from licensed antifungal agents such as amphotericin B, triazoles and echinocandins. To understand the relationship between chemical structure and biological function, we isolated certain ASP2397 derivatives from the culture broth, and we further chemically converted the metal-free form to other derivatives.

WF11899A, B and C, novel antifungal lipopeptide antibiotics were isolated from the culture broth of Coleophoma empetri F-11899. These compounds belong to the echinocandin type of lipopeptides. Of these compounds, WF1 1 899A showed good solubility in water. These three antibiotics possess potent in vitro antifungal activities against Candida spp.

WF11899A, B and C, novel water-soluble lipopeptides related to the echinocandins, possess potent anti-Candida activities. The IC5Osof the compounds against four clinical isolates of Candida albicans ranged from 0.004 to 0.03 jug/ml by microbroth dilution assay. These compounds mildly suppressed the growth of Aspergillusfumigatus and A. niger. WF11899A, B and C showed a potent in vivo anti-Candida activity. Particularly, WF11899A was superior to cilofungin, and equal to fluconazole. l,3-/?-glucan synthase was inhibited by these compounds at the IC5Os of 0.7, 0.7 and 1.8 /ig/ml for WF1 1899A, B and C, respectively. However, they hemolysed mouse red blood cells in vitro at the concentration of 62 fig/ml. WF11899A, B and C are novel echinocandin type of lipopeptide antibiotics produced by Coleophoma empetri F-1 1 8991}. These compoundswere structurally characterized by having a sulfate moiety. Although the other related natural lipopeptides lacked water-solubility, WF11899A possessed excellent watersolubility ascribed to the sulfate moiety.