Isolating antifungals from fungus-growing ant symbionts using a genome-guided chemistry approach.

We describe methods used to isolate and identify antifungal compounds from actinomycete strains associated with the leaf-cutter ant Acromyrmex octospinosus. These ants use antibiotics produced by symbiotic actinomycete bacteria to protect themselves and their fungal cultivar against bacterial and fungal infections. The fungal cultivar serves as the sole food source for the ant colony, which can number up to tens of thousands of individuals. We describe how we isolate bacteria from leaf-cutter ants collected in Trinidad and analyze the antifungal compounds made by two of these strains (Pseudonocardia and Streptomyces spp.), using a combination of genome analysis, mutagenesis, and chemical isolation. These methods should be generalizable to a wide variety of insect-symbiont situations. Although more time consuming than traditional activity-guided fractionation methods, this approach provides a powerful technique for unlocking the complete biosynthetic potential of individual strains and for avoiding the problems of rediscovery of known compounds. We describe the discovery of a novel nystatin compound, named nystatin P1, and identification of the biosynthetic pathway for antimycins, compounds that were first described more than 60 years ago. We also report that disruption of two known antifungal pathways in a single Streptomyces strain has revealed a third, and likely novel, antifungal plus four more pathways with unknown products. This validates our approach, which clearly has the potential to identify numerous new compounds, even from well-characterized actinomycete strains.

Candicidin-producing Streptomyces support leaf-cutting ants to protect their fungus garden against the pathogenic fungus Escovopsis.

Leaf-cutting ants such as Acromyrmex octospinosus live in obligate symbiosis with fungi of the genus Leucoagaricus, which they grow with harvested leaf material. The symbiotic fungi, in turn, serve as a major food source for the ants. This mutualistic relation is disturbed by the specialized pathogenic fungus Escovopsis sp., which can overcome Leucoagaricus sp. and thus destroy the ant colony. Microbial symbionts of leaf-cutting ants have been suggested to protect the fungus garden against Escovopsis by producing antifungal compounds [Currie CR, Scott JA, Summerbell RC, Malloch D (1999) Fungus-growing ants use antibiotic-producing bacteria to control garden parasites. Nature 398:701-704.]. To date, however, the chemical nature of these compounds has remained elusive. We characterized 19 leaf-cutting ant-associated microorganisms (5 Pseudonocardia, 1 Dermacoccus, and 13 Streptomyces) from 3 Acromyrmex species, A. octospinosus, A. echinatior, and A. volcanus, using 16S-rDNA analysis. Because the strain Streptomyces sp. Ao10 proved highly active against the pathogen Escovopsis, we identified the molecular basis of its antifungal activity. Using bioassay-guided fractionation, high-resolution electrospray mass spectrometry (HR-ESI-MS), and UV spectroscopy, and comparing the results with an authentic standard, we were able identify candicidin macrolides. Candicidin macrolides are highly active against Escovopsis but do not significantly affect the growth of the symbiotic fungus. At least one of the microbial isolates from each of the 3 leaf-cutting ant species analyzed produced candicidin macrolides. This suggests that candicidins play an important role in protecting the fungus gardens of leaf-cutting ants against pathogenic fungi.

[Method of partitioning polyene macrolide antibiotics]. / Sposob razdeleniia polienovykh makrolidnykh antibiotikov.

Levorin and mycoheptin, polyenic antibiotics were separated into 3 and 7 heptaenic components respectively by the method of disc electrophoresis in 7.5 per cent polyacryl amide gel prepared with tris-HCl buffer, pH 8.9. It was shown that nystatin had different composition as dependent on the organism producing it. The results of electrophoresis confirmed the data of the counter-current distribution.