Atolypenes, Tricyclic Bacterial Sesterterpenes Discovered Using a Multiplexed In Vitro Cas9-TAR Gene Cluster Refactoring Approach

Abstract

Most natural product biosynthetic gene clusters identified in bacterial genomic and metagenomic sequencing efforts are silent under laboratory growth conditions. Here, we describe a scalable biosynthetic gene cluster activation method wherein the gene clusters are disassembled at interoperonic regions in vitro using CRISPR/Cas9 and then reassembled with PCR-amplified, short DNAs, carrying synthetic promoters, using transformation assisted recombination (TAR) in yeast. This simple, cost-effective, and scalable method allows for the simultaneous generation of combinatorial libraries of refactored gene clusters, eliminating the need to understand the transcriptional hierarchy of the silent genes. In two test cases, this in vitro disassembly-TAR reassembly method was used to create collections of promoter-replaced gene clusters that were tested in parallel to identify versions that enabled secondary metabolite production. Activation of the atolypene ( ato) gene cluster led to the characterization of two unprecedented bacterial cyclic sesterterpenes, atolypene A (1) and B (2), which are moderately cytotoxic to human cancer cell lines. This streamlined in vitro disassembly- in vivo reassembly method offers a simplified approach for silent gene cluster refactoring that should facilitate the discovery of natural products from silent gene clusters cloned from either metagenomes or cultured bacteria.

Keywords: genome mining; metagenomics; natural products; promoter engineering; sesterterpene.

Figure 1.

Schematic procedure for miCASTAR method including gene cluster disassembly using in vitro Cas9 digestion and reassembly using yeast-mediated TAR to generate a library of differentially refactored gene clusters for use in heterologous expression experiments. The triangles indicate promoter insertion events. Different colored triangles are indicative of different promoter cassettes.

Figure 2.

miCASTAR development. A) Experimental design for assessing the efficiency of using different numbers of prototrophic markers to select for the simultaneous insertion of 2, 4, 6 or 8 promoter cassettes into the tam gene cluster. White boxes indicate marker free bi-directional promoter cassettes. Open arrows indicate bi-directional promoter cassettes with prototrophic markers. B) Efficiency of selecting of the multiplexed, TAR-based insertion of promoter cassettes using between zero and four auxotrophic selections. C) The tam gene cluster was digested at between 1 to 8 interoperonic regions (shown by arrows) using 8 different tam gene cluster specific sgRNAs and Cas9. Fragmented pTARa:tam was visualized by agarose gel electrophoresis. The numbers in the gel indicate the number of interoperonic regions that was targeted in each reaction. Specific cut site patterns are shown E. D) Efficiency of selecting for the correctly refactored construct when single marker miCASTAR was conducted on pTARa:tam using different numbers of promoter (P) cassettes. E) HPLC profiles of ethyl acetate extracts derived from cultures of S. albus transformed with native and refactored tam gene clusters. The production of tetarimycin A is seen upon introduction of four or more promoters. F) The efficiency of using single-marker miCASTAR to correctly engineer BACs of different sizes is shown. BACs used in this study contain either uncharacterized type I polyketide or nonribosomal biosynthetic gene clusters cloned from a soil metagenome.

Figure 3.

Systematic refactoring the silent atolypene gene cluster using miCASTAR. A) Comparison between the brasilicardin A (bra) and ato gene clusters, and TAR assembly of the ato gene cluster. AtoG through atoL were ultimately determined not to be part the ato gene cluster. B) miCASTAR to generate 15 different combinations of promoter refactored constructs. C) Total positive ion chromatogram resulting from UPLC-MS analysis of extracts from S. albus hosting different refactored gene clusters, native gene cluster, and pTARa-Lys empty vector. D) Structures of atolypenes A and B isolated from strains hosting refactored ato gene clusters. This figure shows the relative stereochemistry throughout the tricyclic ring system, and the absolute configures of the amino acid. E) IC₅₀ data for atolypenes A and B against human cell lines. IC₅₀ values were also determined for melittin versus HEK293 (3.71 μM), HeLa (0.91 μM), and A549 (0.39 μM) as positive controls.

Figure 4.

Atolypene A biosynthesis and structure. A) Proposed atolypene A biosynthetic pathway with the table of proposed biosynthetic gene function. B) Geranyl-geranyl pyrophosphate origin brasilicardin A. C) Known tri- and tetra- cyclic sesterterpenes with methyl migration patterns similar to that seen in the atolypenes.