Sebastian Schindler

Local deuteration of leucine sidechains, now extended to the β-position, enhances linewidths and signal intensity to better enable eukaryotic expression of difficult target proteins.

NMR spectroscopy of biomolecules provides atomic level information into their structure, dynamics and interactions with their binding partners. However, signal attenuation from line broadening caused by fast relaxation and signal overlap often limits the application of NMR to large macromolecular systems. Here we leverage the slow relaxation properties of ¹³C nuclei attached to ¹⁹F in aromatic ¹⁹F–¹³C spin pairs as well as the spin–spin coupling between the fluorinated ¹³C nucleus and the hydrogen atom at the meta-position to record two-dimensional ¹H–¹³C_F correlation spectra with transverse relaxation-optimized spectroscopy selection on ¹³C_F. To accomplish this, we synthesized [4-¹⁹F¹³C^ζ; 3,5-²H₂^ε] Phe, engineered for optimal relaxation properties, and adapted a residue-specific route to incorporate this residue globally into proteins and a site-specific 4-¹⁹F Phe encoding strategy. This approach resulted in narrow linewidths for proteins ranging from 30 kDa to 180 kDa, enabling interaction studies with small-molecule ligands without requiring specialized ¹⁹F-compatible probes.

The external control of catalytic activity and substrate specificity of enzymes by light has aroused great interest in the fields of biocatalysis and pharmacology. Going beyond, we have attempted to photocontrol enzyme stereoselectivity on the example of phosphotriesterase (PTE), which is capable of hydrolyzing a wide variety of racemic organophosphorus substrates where one of two enantiomers is often highly toxic. To pursue this goal, the photocaged unnatural amino acid o-nitrobenzyl-l-tyrosine (ONBY) was incorporated by genetic code expansion at the large subsite of the active center, together with additional mutations at the small subsite. The stereoselectivities of the resulting PTE variants were tested with the achiral control substrate paraoxon and four different racemic substrates, which contained a p-nitrophenol leaving group in combination with either methyl-phenyl, ethyl-phenyl, methyl-cyclohexyl, or ethyl-cyclohexyl substituents. Comparison of the enantioselectivities (k _cat/K _M for S_p divided by k _cat/K _M for R_p) before and after decaging of ONBY using irradiation revealed the desired photoinduced inversion of enantioselectivity for three of the variants: PTE_I106A-H257ONBY exhibited a 43-fold stereoselectivity switch for the methyl-phenyl substrate, PTE_I106A-F132A-H257ONBY a 184-fold stereoselectivity switch for the methyl-cyclohexyl substrate, and PTE_I106A-F132A-S308A-H257ONBY a 52-fold and a 57-fold stereoselectivity switch for the methyl-cyclohexyl and the ethyl-cyclohexyl substrates. A computational analysis including molecular dynamics simulations and docking showed that a complicated interplay between steric constraints and specific enzyme-substrate interactions is responsible for the observed effects. Our findings significantly broaden the scope of possibilities for the spatiotemporal control of enantioselective transformations using light in biocatalytic systems.

The current standard method for amino acid signal identification in protein NMR spectra is sequential assignment using triple-resonance experiments. Good software and elaborate heuristics exist, but the process remains laboriously manual. Machine learning does help, but its training databases need millions of samples that cover all relevant physics and every kind of instrumental artifact. In this communication, we offer a solution to this problem. We propose polyadic decompositions to store millions of simulated three-dimensional NMR spectra, on-the-fly generation of artifacts during training, a probabilistic way to incorporate prior and posterior information, and integration with the industry standard CcpNmr software framework. The resulting neural nets take [ ¹ H, ¹³ C] slices of mixed pyruvate–labeled HNCA spectra (different CA signal shapes for different residue types) and return an amino acid probability table. In combination with primary sequence information, backbones of common proteins (GB1, MBP, and INMT) are rapidly assigned from just the HNCA spectrum.