Discovery of Cryptic Natural Products Using High-Throughput Elicitor Screening on Agar Media

Abstract

It is now well-established that microbial genomes carry sparingly expressed biosynthetic gene clusters (BGCs) that need to be induced in order to characterize their products. To do so, we herein subjected two well-known producers, Burkholderia plantarii and Burkholderia gladioli, to high-throughput elicitor screening (HiTES), an emerging approach for accessing the products of these "cryptic" BGCs. Both organisms have previously been examined extensively in liquid cultures. We therefore applied HiTES on agar and found several novel natural products that are only produced in this format and not in liquid cultures. Most notably we found two metabolites, termed burkethyl A and B, that contain an unusual m-ethylbenzoyl group and for which we identified the cognate BGC using bioinformatic and genetic studies. Our results indicate that agar-based HiTES is a promising approach for natural product discovery and are in line with the notion that even "drained" strains remain sources of new metabolites as long as alternative approaches are employed.

Figure 1.

HiTES on solid agar media. (a) Cells are grown in “mini-petri dishes” on agar in the presence of potential elicitors. The induced metabolomes are then extracted and subjected to UPLC-Qtof-MS to identify cryptic metabolites. (b, c) Application of agar-based HiTES to B. plantarii (b) and B. gladioli (c). Metabolites that are overproduced or induced are shown in 3D maps displaying the m/z and intensity for each compound as a function of elicitor.

Figure 2.

Discovery of cryptic burkethyl A and B. (a) 3D map of the secondary metabolome of B. plantarii using agar-based HiTES. HR-MS data were acquired in the m/z range of 200–1000 (Figure 1); a magnified view is shown here focusing on m/z 268 (red box). (b) 2D slice from the 3D plot in panel (a) focusing on 4 with m/z 268. Structures of the best elicitors are shown. (c) (Left) Extracted ion chromatogram (EIC) for 4 upon induction by the elicitor indicated in agar cultures. (Right) EIC of 4 upon treatment of B. plantarii with 1 (or DMSO vehicle control) in liquid or agar cultures. Two independent biological replicates were carried out for each condition; similar production levels of 4 were observed in both replicates. (d) Chemical structures of 4 and 5 and key NMR correlations used to solve the structure of 4.

Figure 3.

Proposed biosynthetic pathway for burkethyl A and B. (a) bet BGC. Genes are color-coded as shown. (b) EIC of burkethyl A in parental wt B. plantarii (red trace) and the betF::tet insertional deletion mutant (black trace). (c) Proposed biosynthesis of burkethyls. See text for details.

Figure 4.

Discovery of gladiobactins B–D. (a) 3D map of the secondary metabolome of B. gladioli using agar-based HiTES, focusing on m/z 1120 and 1266 (red boxes). (b) 2D slice from the 3D plot in panel (a) focusing on 10 with m/z 1120. Fluorocinolone (6), imipenem (7), rifapentine (8), and saquinavir (9) were the best elicitors for gladiobactins B–D. (c) Validation of induction of 10 and 11 by fluocinolone. Shown are EICs for 10/11 (right peak) and 12 (left peak) after growth of B. gladioli with fluocinolone (or vehicle control) in liquid or agar cultures. Two independent biological replicates were carried out for each condition; similar production levels of 10/11 and 12 were observed in both replicates. (d) Chemical structures of gladiobactins B–D (10–12) and NMR correlations used to solve the structure of 12.