In 1986, a pharmaceutical research group in Japan discovered that a certain species of soil bacteria, Streptomyces cattleya, produces an unusual amino acid. This amino acid, β-ethynylserine (βes), structurally resembles canonical amino acid L-threonine but contains a terminal alkyne (C≡H) at the Cγ. Almost 15 years later, Barry Sharpless described the first ‘click’ reaction – the Cu(I)-catalyzed azide-alkyne cycloaddition, which ushered in a new wave of work in manipulating biological systems using synthetic chemistry. Today, the terminal-alkynes remain the most widely used bioorthogonal handle, despite a requirement that both terminal alkyne and azide must be provided exogenously. More recently, interest in genetically encode terminal alkynes has been growing, but only one enzyme family (JamB) had thus far been characterized that is capable of making terminal alkynes on fatty acid chains. A genetically encodable terminal-alkyne amino acid could have wide appeal since it could readily introduced into various biomolecules and natural products.
Limited knowledge of enzymatic terminal acetylenases made the discovery of the βes pathway challenging. Numerous bioinformatics tools for the discovery of gene clusters, which form part of the primary pipeline in natural product work, are trained on known biosynthetic pathways. Indeed, our initial approach at elucidating the genes involved in βes biosynthesis consisted of performing targeted knockouts of fatty acid desaturases in S. cattleya. When this approach failed, we opted to take an agnostic approach to gene discovery.
Of the three known species of Streptomyces that have been observed to make terminal alkynes, two of them (S. cattleyaand S catenulae) have recently had their genome’s sequenced. We narrowed down our genomic search by asking two fundamental questions: 1) Which sets of genes between both these organisms are clustered and 2) Of these gene clusters, which ones are unique among S. cattleyaand S. catenulae. The putative βes was confirmed by knocking out each gene in S. cattleya. Subsequently, pathway intermediates were identified using comparative metabolomics, while each individual step was characterized and reconstituted in vitro.
Characterizing the pathway revealed an unexpected strategy in alkyne biosynthesis. First, a αKG/Fe-dependent enzyme performs a radical halogenation of free L-lysine at the Cγ, generated an unstable halo-amino acid, 4-Cl-L-lysine. Next through an unknown Fe-mediated mechanism, the second enzyme in the pathway catalyzes an oxidative cleavage of the Cε, generated vinyl-halide amino acid, 4-Cl-L-allylglycine (Cl-Alg). We found that in the case of βes, Nature had decided to take a different approach to terminal alkyne synthesis. While JamB uses an iron-radical mediated desaturation, BesB (the acetylenase described in this work), makes the terminal alkyne through the elimination of the Cl in Cl-Alg to generate a terminal alkyne amino acid, L-propargylglycine. Two additional steps, ligation onto glutamate at the α-amine and hydroxylation of the γ-glumayl-propargylglycine dipeptide, form a glutamyl-βes dipeptide. Upon amide-bond hydrolysis, βes is released. This unconventional pathway for enzymatic desaturation reveals a platitude of means by which enzymes can carry out unusual modifications on amino acids, and establishes the first genetically encodable pathway that can be used to make terminal alkyne L-amino acids in vivo.
You can read more about our work in Nature.
This post was co-authored by Jorge A. Marchand and Michelle C. Y. Chang.