A proteomics approach to discovering natural products and their biosynthetic pathways

Abstract

Many natural products with antibiotic, anticancer and antifungal properties are synthesized by nonribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs). Although genome sequencing has revealed the diversity of these enzymes, identifying new products and their biosynthetic pathways remains challenging. By taking advantage of the size of these enzymes (often >2,000 amino acids) and unique marker ions derived from their common phosphopantetheinyl cofactor, we adapted mass spectrometry-based proteomics to selectively detect NRPS and PKS gene clusters in microbial proteomes without requiring genome sequence information. We detected known NRPS systems in members of the genera Bacillus and Streptomyces, and screened 22 environmental isolates to uncover production of unknown natural products from the hybrid NRPS-PKS zwittermicin A biosynthetic gene cluster. We also discovered an NRPS cluster that generates a seven-residue lipopeptide. This 'protein-first' strategy complements bioassay- and sequence-based approaches by finding expressed gene clusters that produce new natural products.

Figure 1

The workflow for PrISM. a, Microbial strains are grown in liquid culture. b, The proteome of the strain is subjected to proteomics or in-gel digestion of high molecular weight bands. c, LC-FTMSⁿ is conducted on the resulting peptide mixture, with expressed T domain active site peptides identified by the Ppant ejection assay. d, Peptide sequences are used to generate primers to amplify DNA sequence portions of the expressed gene cluster. e, The gene cluster is identified and sequenced, which informs targeted detection of the natural product produced as depicted in panel f.

Figure 2

Identification of an expressed T domain active site peptide in the NK2018 proteome by the on-line Ppant ejection assay using nanoLC-MS. a, Total ion chromatogram (TIC) from a nanoLC-FTMS analysis of a single SDS-PAGE gel slice containing NK2018 HMWPs. b, Selected ion chromatogram (SIC) for the elution of the Ppant ejection ion of interest (Pant-CAM, m/z 318.1482). c, SIC for the phosphopantetheinylated peptide eluting at 39.3 min and a phosphopantetheinylated peptide observed in nanoLC-MS analysis (inset). See the Supplementary Discussion for detailed information on the identification of this peptide. d, Representative results from proteomic analysis of NK2018. Shown is the MS/MS spectrum for a peptide from C1S2. Inset: The de novo peptide sequence and the degenerate primer designed from the given peptide sequence. e, Representative results of PCR amplification using degenerate primers designed from peptide sequences determined de novo. See Supplementary Table 4 for a complete list of all PCR amplified products and the PCR numbers corresponding to the lanes in Figure 2c.

Figure 3

Identification of new lipoheptapeptides in NK2018. a, Domain organization of cluster #2 based upon the gene sequence in B. cereus AH1134. Amino acid substrates were selected based upon bioinformatic analysis and the structure of the detected peptides. Peptides from the domains in red were identified by nanoLC-MS. b, Base peak chromatogram (top) of a NK2018 culture supernatant sample and a SIC (red, bottom) for the species at m/z 908.4845. The mass spectrum of the 1+ charge state of the ion is shown (inset). c, Putative structure for the lipoheptapeptides detected, including the species detected in Figure 3b. The structures are predicted to differ in the length of the fatty acid tail and in the formation of a lactone ring (table inset). Abbreviations: C – condensation domain, A – adenylation domain, T – thiolation domain, TE – thioesterase.