The biosynthetic origins of the unusual glycopeptide suomilide and genome mining of aeruginosin family proteases

Natural products have enormous structural and chemical diversity and are either the source or direct inspiration for many drugs in use today. Cyanobacteria are prolific producers of complex natural products with serine protease inhibiting activity. Many of these natural products are the product of non-ribosomal peptide synthetase (NRPS) modular enzyme complexes. Suomilide is a complex tetrapeptide produced by strains of the benthic cyanobacterium Nodularia sphaerocarpa. It has a highly complicated structure and contains an unusual azabicyclononane moiety, a methylglyceric acid, a xylose unit with hexanoic acids and a terminal 1-amidino-3-(2-aminoethyl)-3-pyrroline moiety. Suomilide inhibits thrombin, plasmin and trypsin in low micro-molar concentrations. The biosynthetic of this unusual glycoside remain unclear. However, suomilide is long predicted to be part to the aeruginosin family of protease inhibitors. A 5.4 Mb draft genome of Nodularia sphaerocarpa HKVV was obtained in order to identify the suomilide biosynthetic. The 43.7 kb suomilide gene cluster was identified on a single contig by performing tBLASTn searches on the draft genome of Nodularia HKVV using aerDEF genes from aeruginosins gene cluster as query. This gene cluster encodes 27 genes including two complex NRPS enzymes and a set of tailoring enzymes for the assembly of suomilide. The suomilide gene cluster shares extensive homology to known aeruginosin gene clusters including two aerB and aerG genes encoding NRPS enzymes, 12 genes (aerC, aerD, aerE, aerF, aerI, aerK, two copies of aerN and four copies of aerH) encoding for the enzymes responsible for synthesis of precursor non-proteinogenic amino acids and 13 other tailoring enzymes. The suomilide gene cluster was much larger and encoded a greater number of biosynthetic enzymes reflecting the structural complexity of suomilide. We identified 10 aeruginosin gene clusters and 2 suomilide gene clusters from 12 strains of cyanobacteria by genome mining. Bioinformatics analyses suggested these gene clusters encoded an unanticipated chemical diversity of aeruginosins and suomilides. LC-MS and Q-TOF analysis detected aeruginosins or suomilide variants from 12 of the 15 strains. Surprisingly, inhibition assays with the crude extracts using all three isoforms of human trypsin suggest that these compounds may have potent and selective inhibition of human trypsin isoforms. Further work is required to prove that suomilide alone can carry out selective inhibition of trypsin isoforms or is it a result of synergism between the compounds produce by cyanobacteria. Phylogenetic analysis demonstrated that the aeruginosin evolved through the acquisition of multiple loading mechanisms and tailoring enzymes through horizontal gene transfers. Our results support the hypothesis that suomilides are a part of aeruginosin family as they are made through the same genetic pathway, however have gained a greater degree of structural diversity due to the acquisition of tailoring enzymes. These results together suggest that cyanobacteria produce an unexpected wealth of complex natural products belonging to the aeruginosin family and that some of these may be potent and selective inhibitors of isoforms of human trypsin