Genome Mining and Discovery of Imiditides, a Family of RiPPs with a Class-Defining Aspartimide Modification

Ribosomally synthesized and post-translationally
modified
peptides
(RiPPs) are a large and diverse class of natural products of ribosomal
origin. In the past decade, various sophisticated machine-learning-based
software packages have been established to discover novel RiPPs that
do not resemble the known families. Here, we show that tailoring enzymes
that cluster with various RiPP families can serve as effective bioinformatic
seeds, providing a complementary approach for novel RiPP discovery.
Leveraging the fact that O-methyltransferases homologous
to protein isoaspartyl methyltransferases (PIMTs) are associated with
lasso peptide, graspetide, and lanthipeptide biosynthetic gene clusters
(BGCs), we utilized a C-terminal motif unique to RiPP-associated O-methyltransferases as the search query to discover a novel
family of RiPPs, the imiditides. Our genome-mining algorithm reveals
a total of 670 imiditide BGCs, distributed across Gram-positive bacterial
genomes. In addition, we demonstrate the heterologous production of
the founding member of the imiditide family, mNmaAM, encoded
in the genome of Nonomuraea maritima. In contrast to other RiPP-associated PIMTs that recognize constrained
peptides as substrates, the PIMT homologue in the mNmaAM BGC, NmaM, methylates a specific Asp residue on the linear precursor
peptide, NmaA. The methyl ester is then turned into an aspartimide
spontaneously. Substrate specificity is achieved by extensive charge–charge
interactions between the precursor NmaA and the modifying enzyme NmaM
suggested by both experiments and an AlphaFold model prediction. Our
study shows that PIMT-mediated aspartimide formation is an emerging
backbone modification strategy in the biosynthesis of multiple RiPP
families.