NMR Characterization of an Unusual 37 kDa Epimerization Domain of Yersinabactin Synthetase

Abstract

Nonribosomal Peptide Synthetases (NRPSs) are endogenous microbial supramolecular (1-4 MDa) assemblies that are responsible for the biosynthesis of important natural products, including antibiotics. Genetic reprogramming of NRPS machinery for the generation of novel-or perhaps “designer”-antibiotics (or other therapeutics) represents the apogee of NRPS research. Our model NRPS system is yersiniabactin synthetase (YS), which produces the virulence factor yersiniabactin. Heterocycles (e.g. thiazoline rings) in natural products, such as yersiniabactin, exhibit unique medicinal properties (antibiotic, anticancer, immunosuppressant). An unusual epimerization domain (EA) in YS catalyzes the stereoconversion of a chiral heterocyclic center. Interestingly, EA itself is embedded within the primary sequence of a so-called adenylation domain (A, recognizes and activates substrates). To understand the overall choreography of catalysis, we require structural models of each catalytic domain in its monomeric state, bound to its substrate(s), and in complex with other relevant domains. NRPS domains can be difficult to crystallize in the biologically relevant conformations and complexes, so we take advantage of NMR spectroscopy, which can offer structural, kinetic, thermodynamic, and dynamical data. However, conventional bimolecular NMR experiments will fail on domains such as EA (37 kDa) largely because of relaxation losses but also spectral crowding. This work discusses how nonuniform sampling allowed us to rescue sensitivity and collect preliminary data on a partially deuterated sample. We also exploit covariance NMR spectroscopy to ease the challenging backbone assignment procedure. Using such data we report ∼70% backbone assignment. We also present a chemical biology strategy that will allow us to synthesize otherwise challenging substrates, which will facilitate measurement of steady state kinetic parameters and inform our structural models.