Genome mining-based discovery of an atypical fungal non-reducing polyketide synthase encoding dimeric alkylresorcinol biosynthesis.

Genome mining is a powerful strategy for finding biosynthetic gene clusters (BGCs) for unprecedented natural products and their biosynthetic mechanisms. In this study, to obtain novel type polyketides, we performed genome mining focusing on three perspectives for novelty: 1) amino acid sequence-based classification, 2) unique domain architectures of polyketide synthase, and 3) predicted structural features from protein modeling. As a result, we discovered a BGC consisting of a highly reducing polyketide synthase and a non-reducing polyketide synthase (NR-PKS), which harbors noncanonical tandem acyl carrier protein (ACP) domains and a structurally characteristic thioesterase (TE) domain. Heterologous expression revealed that the cyrl cluster produces a novel dimer of long-chain alkylresorcinolic acid (1). Furthermore, site-directed mutagenesis of each ACP domain revealed that both domains are essential for efficient dimerization. This study provides the first example of a fungal dimer-forming NR-PKS in which the tandem ACP domains work "in-series" with nonredundant roles to construct a dimeric long-chain alkylresorcinolic acid. The in silico analysis suggested that the TE domain led the substrate to a dimerizable form. Since 1 is structurally related to integracins with HIV-1 integrase inhibition, our findings provide insight into integracins biosynthesis and offer a basis for generating new integracin derivatives.