We are developing spectroscopic methods to detect nuclear spins with enhanced sensitivity using two different approaches. In one case, we detect nuclear magnetic moments via the larger magnetic moment of a paramagnetic molecule, a method called electron-nuclear double resonance (ENDOR). Specifically, we have introduced ENDOR detection of 19F or 17O nuclei to obtain structural information in biomolecules. The dipolar coupling of 19F with a paramagnetic spin label turned out to be sensitive to intermolecular distances in the angstrom to nanometer scale (≲ 2 nm), allowing investigation of active site structures in proteins and RNAs with high sensitivity and precision. For instance, in the enzyme ribonucleotide reductase, a 250 kDa macromolecular complex, we could use 19F ENDOR to measure the distance between two redox-active tyrosines across two protein subunits, shining light on the role of interface residues to controlling communication between these two subunits. These experiments would have not been feasible by direct 19F NMR detection, due to the much lower sensitivity. Also, 17O has offered unique opportunities to detect water molecules in the active RNR enzyme or to study the hydration structure of radical species involved in catalytic reactions.
In a second approach, we have implemented a new instrumental design for NMR signal enhancements using Overhauser dynamic nuclear polarization (DNP). This method relies on pumping with microwaves a paramagnetic molecule, called polarizer, which is mixed to the sample of interest. In analogy to an ENDOR experiment, magnetization is transferred between electron and nuclear spins. At an external magnetic field of 9.4 Tesla, the pumping frequency, to be on resonant with the electron Larmor frequency, requires sub-THz irradiation. The latter is a major bottleneck for liquids, where microwave penetration in the large NMR samples is strongly attenuated and leads to sample heating. In our work we could show a solution for these technical challenges. Empowered by a thin-layer setup in a cylindric NMR tube and microwave engineering, our design allows for hyperpolarized liquid-state NMR spectra in one and two dimensions with a reasonable sample volume (a few µL). We anticipate that our approach provides a starting point for a broader implementation of DNP signal enhancement in liquid-state NMR. The reached sensitivity opens up perspectives for structural determination of natural products or drugs, available in small quantities.
