RiPPing apart the rules for peptide natural products

Humankind has looked to peptide natural products as a source of valuable therapeutics harboring anti-cancer, immunomodulatory, and antibiotic properties since the discovery of penicillin in the late 1920s. Peptide natural products display a wide array of proteolytically resistant backbone modifications, non-canonical amino acids, and cyclizations that rigidify their structures and enable binding to specific cellular targets. Two distinct biosynthetic routes yield nearly all complex modified peptides currently identified―those incorporating amino acids via the ribosome and those that use multimodular protein complexes, termed non-ribosomal peptide synthetases (NRPSs), to ligate individual amino acids in an assembly-line fashion. Unlike ribosomal synthesis, NRPS-derived peptides are not limited to the canonical 20 amino acids. NRPSs have been shown to incorporate upwards of 500 unique amino acids directly into their peptide scaffolds [1]. The impressive amino acid building block repertoire of NRPS-derived peptides was thought to cover a much larger structural and chemical space compared to their ribosomally-derived counterparts. While ribosomally synthesized and post-translationally modified peptide (RiPP) pathways are currently known to introduce unique lanthionine thioether linkages and lasso peptide-like topologies [2] compared to NRPS pathways, for many years these examples were minor concessions to the breadth of distinct modifications only thought possible for non-ribosomal synthesis. But the (RiPP) tides are turning, and in dramatic fashion.

Culminating with their most recent findings published in Science [3], the Piel group at ETH Zurich has authored a series of papers that are redefining our understanding of what backbone compositions are achievable in ribosomally encoded peptides. First, they elucidated the biosynthetic machinery responsible for the cytotoxic, membrane-spanning β-helical linear peptides, the polytheonamides [4]. This work unveiled a new family of radical S-adenosylmethionine (rSAM) epimerases capable of iteratively introducing a wide variety of d-amino acids into translated peptides and proteins [4,5]. d-Amino acids were largely thought to be a hallmark of NRPS synthesis, with the exception of a few animal-derived peptides and microbial lanthipeptide pathways capable of l-serine to d-alanine conversions [6]. Next, a collaborative effort by Freeman, Piel and Kuenzler [7], followed soon after by Suessmuth [8], provided the first example of a RiPP pathway incorporating α-N-methylations into peptide backbones of the basidiomycete-originating nematotoxins, the omphalotins. Finally, Piel and coworkers have now revealed β-amino acids are not out of reach for peptides synthesized by the ribosome [3].

In a remarkable set of findings, Morinaka et al. discovered that a ‘splicase’ enzyme, PlpX, catalyzes C—C bond cleavage of the peptide backbone to yield α-keto β-amides in Nif11-like proteusin precursor peptides (Fig. 1). Essentially all RiPPs are translated as precursor proteins with an extra N-terminal sequence (leader peptide) that is not found in the natural product. The leader peptide functions as a binding domain to recruit post-translational modifying enzymes to tailor the C-terminal amino acid sequence found in the mature natural product (core peptide) that is later released from the leader peptide. Originating from the cyanobacteria Pleurocapsa sp. PCC 7319, plpX was heterologously coexpressed in Escherichia coli with genes from its putative cyanobacterial RiPP pathway. Interestingly, plpX alone shows no activity on precursors plpA2 or plpA3. A closer look at the gene cluster revealed plpY, a small, unannotated hypothetical gene directly downstream of plpX. PlpY shows weak homology to PqqD from pyrroloquinoline biosynthesis that corresponds to a domain termed the RiPP recognition element (RRE) [9], previously shown to play a crucial role in precursor peptide binding in many RiPP systems. When plpX and plpY were coexpressed with plpA2 and plpA3, net losses of C₈H₉NO from each precursor were detected by liquid chromatography-mass spectrometry. Isotope labeling and NMR spectral analysis confirmed the construction of α-keto β-amides through net excision of tyramine, leaving only the C1 carbonyl of tyrosine of the motif “XYG” in the peptide backbone. Mechanistic details of C—C bond cleavage, such as the possibility of p-cresol as an intermediate, is an area of active interest in the Piel group.

Fig. 1.

Enzymatic transformations recently discovered in RiPP biosynthesis providing peptide scaffolds once thought exclusive to NRPS pathways.

Next, Morinaka et al. looked to gain insight into the prevalence and phylogenetic distribution of this unusual enzymatic strategy. Metagenomic analysis along with newly sequenced cyanobacterial genomes from the Pasteur Institute revealed >50 plpX homologs throughout cyanobacteria, proteobacteria, and select actinobacteria. Using this in silico-generated sequence diversity as a guide, experiments delving into the substrate preferences of PlpXY revealed a considerable amount of tolerance to mutation, deletion, and successful multiplication of the substrate “XYG” motifs. This enzymatic promiscuity was highlighted by the in vivo synthesis of an α-keto β-amide scaffold evocative of the protease inhibitor boceprevir. The outcome of these experiments defined an 11-residue recognition sequence that opens the door for incorporation of α-keto β-amides into proteins. In this vein, the authors utilized the unique reactivity of the α-keto group to ligate a commercially available fluorogenic probe onto the protein precursor PlpA3.

The last decade has seen an explosion of new catalytic strategies and structures discovered from RiPP natural product pathways. Insights concerning post-translational modifications to the peptide backbone have been particularly fruitful, with mechanistic understanding concerning formation of N-to-C macrocycles, isopeptide bonds, thia/oxazoles, thioamides, thioethers, amino acid epimerizations, α-N-methylations, and β-amino acids now advancing in RiPP biosynthesis. Perhaps not surprisingly, rSAM enzymes have recently garnered interest as their capacity to catalyze an increasing number of novel biosynthetic transformations in peptide natural products are being reported. Estimates of the rSAM superfamily suggest less than 1% of their catalytic versatility has been observed [10]. A great strength of RiPP pathways over their NRPS-derived counterparts is the seemingly immense plasticity of these systems to evoke millions of unique peptide scaffolds directly from genomic sequences. For this vast opportunity to reach its true potential, detailed studies concerning protein-protein interactions and protein dynamics in RiPP pathways need to progress. Piel and coworkers have provided yet another inspiring example of the benefits of basic scientific research and natural product discovery to reveal the unexpected mysteries that Nature has to offer.