Magnetic additives slash measurement times in fluorine protein NMR

Scientists watching a protein at work have long faced an awkward trade-off: the experiments that reveal the most about molecular structure are often the ones that demand the most patience. A team at the Institute of Science and Technology Austria, working with colleagues at the University of Vienna, has now shown that a simple additive borrowed from an established trick in nuclear magnetic resonance can dramatically shorten the waiting times that plague one of the most promising techniques in structural biology. By adding paramagnetic gadolinium compounds to solid protein samples, the researchers accelerated fluorine-detected magic-angle spinning NMR experiments by more than a factor of three, without sacrificing the spectral resolution that makes the method valuable.

The technique in question, magic-angle spinning NMR, spins a solid sample at an angle of 54.74 degrees relative to the magnetic field, the so-called magic angle at which the anisotropic interactions that broaden spectral lines in solids are averaged away. For proteins, this opens a window onto structure and dynamics in states that are difficult to study by other means: microcrystals, membrane proteins embedded in lipid bilayers, and enormous assemblies such as virus capsids. In recent years, researchers have increasingly combined this approach with fluorine labelling, introducing atoms of the isotope fluorine-19 into proteins at chosen positions. Because natural biological molecules contain no fluorine at all, the fluorine signal stands out against a silent background, and the nucleus itself has favourable magnetic properties, including a high sensitivity and a wide chemical shift range that makes individual sites easy to distinguish.

Fluorine labelling also brings spectroscopic bonuses. Pairing fluorine-19 with carbon-13 at the same position in an aromatic ring enables two-dimensional experiments and allows scientists to exploit the TROSY effect, which preserves sharp lines in large molecules. Deuterating the labelled amino acids further reduces unwanted relaxation pathways and the need for proton decoupling. The high chemical shift anisotropy of fluorine and its strong dipolar couplings, once formidable obstacles in solid samples, have been tamed by modern probes capable of spinning samples at more than 55 kilohertz per second, making biomolecular fluorine MAS NMR increasingly practical for measuring long-range distances of up to roughly 20 angstroms and mapping structural features of proteins that would otherwise remain invisible.

Yet the method has a stubborn bottleneck. After each scan, the nuclear spins must relax back to their equilibrium polarisation before the next scan can begin, and for fluorine-19 in proteins this longitudinal relaxation time, T1, is often several seconds long. The optimal recycle delay for maximising the signal-to-noise ratio per unit time is directly proportional to T1, which means that most of the experiment is spent simply waiting for the spins to recover rather than collecting data. Ironically, the very strategy used to sharpen fluorine spectra makes the problem worse: deuterated labelling precursors remove the short-range proton-fluorine dipolar couplings that would otherwise help the spins relax quickly, stretching T1 even further.

The Austrian team, led by Lea Becker and Paul Schanda, attacked this bottleneck with paramagnetic doping, a technique long used to speed up proton-detected solid-state NMR. The idea is to add a chelated paramagnetic ion, typically copper(II) or gadolinium(III) bound to a carrier molecule, to the sample buffer. The unpaired electrons of the metal act as relaxation catalysts: through dipolar interactions with nearby nuclei, they accelerate the return of spin polarisation to equilibrium. Because the effect falls off with the sixth power of the electron-nucleus distance, only nuclei close to the metal are directly relaxed, but in protonated samples the enhancement spreads throughout the protein via proton-proton spin diffusion. The art lies in choosing a concentration that shortens T1 substantially without shortening T2, the transverse relaxation time, which would blur the spectral lines and destroy resolution.

What makes the new study notable is that it extends this approach to fluorine for the first time in a biological solid-state sample. The strength of the paramagnetic relaxation enhancement scales with the square of the gyromagnetic ratio of the observed nucleus, and fluorine-19 has one of the highest gyromagnetic ratios of any stable nucleus. The researchers therefore expected a strong effect at concentrations similar to those used in proton experiments. Their test bed was TET2, a dodecameric aminopeptidase from the hyperthermophilic archaeon Pyrococcus horikoshii, a 468-kilodalton assembly of twelve identical 39-kilodalton subunits that has served as a benchmark system in the Schanda laboratory for years. Each subunit carries four tryptophan residues, which the team labelled with fluorine by feeding the bacteria a specially synthesised deuterated 5-fluoroanthranilic acid precursor that the cells convert into 5-fluorotryptophan.

Before testing the dopants, the researchers first had to know which spectral peak belonged to which tryptophan. The fluorine spectrum of the labelled protein showed four distinct signals, one per tryptophan site, but assigning them required a clever genetic workaround. The team constructed five mutant proteins in which combinations of tryptophans were replaced by phenylalanines, leaving either a single fluorinated tryptophan or three of the four in place. Comparing the spectra of these mutants pinned down every signal. The measured fluorine T1 values without any dopant ranged from about 1.7 seconds to 6 seconds depending on the site, implying that an optimally sensitive experiment would demand recycle delays of up to 7.6 seconds between scans, an enormous expenditure of instrument time for a multi-hour or multi-day measurement.

The team then compared two gadolinium chelates, Gd(DTPA-BMA), the agent known clinically as Omniscan, and Gd(DTPA), at concentrations from 2 to 16 millimolar. Both compounds accelerated fluorine relaxation, but their behaviour differed strikingly. Gd(DTPA) produced a longitudinal enhancement roughly seven times stronger per millimolar concentration than Gd(DTPA-BMA), and a transverse enhancement more than fourteen times stronger, differences the authors attribute to the distinct physicochemical properties of the two chelates, such as their water exchange rates. That raw power came at a cost: at 8 millimolar, Gd(DTPA) broadened the fluorine lines so severely that the individual peaks merged, and it also altered the crystallisation behaviour of the protein, likely by binding to its surface. Gd(DTPA-BMA) proved the better compromise. At 8 millimolar, it cut the average fluorine T1 from about 3.2 seconds to 0.92 seconds, reducing the optimal recycle delay from roughly 4 seconds to 1.2 seconds, a more than threefold acceleration with no significant loss of resolution.

Perhaps the most intriguing findings concern the residue-by-residue pattern of the enhancement. The four tryptophan sites responded differently to the dopants, and the pattern depended on which relaxation parameter was measured, which compound was used, and whether fluorine or carbon was observed. The team tested whether these differences could be explained by simple structural parameters, such as the solvent accessibility of each tryptophan or the density of surrounding protons, but found no direct correlation. One site, W164, showed an unusually strong longitudinal enhancement relative to its transverse enhancement for both compounds, a signature the authors suggest could reflect a locally altered rotational correlation time of the gadolinium complex, perhaps through weak binding to the protein. Because the longitudinal and transverse enhancements sample the spectral density at different frequencies, they respond differently to such changes, offering a sensitive but complicated probe of the molecular environment.

The implications reach well beyond one protein. Faster fluorine MAS NMR means more signal per unit time for the experiments that benefit most from fluorine labelling: distance measurements, studies of conformational dynamics, and structural work on large assemblies where sensitivity is chronically scarce. The authors anticipate that the approach can be transferred to many sample types and experimental designs, while cautioning that the optimal dopant and concentration will depend on the specific system, since the balance between accelerating T1 and preserving T2 is delicate and compound-dependent. For a field that has watched fluorine NMR swing back into fashion on the strength of new labelling chemistry and faster spinning hardware, the message of this study is that one of the oldest sensitivity tricks in the solid-state NMR toolbox now works for fluorine too, and it may make the difference between an experiment that takes days and one that fits into an overnight run.

Subject of Research: Paramagnetic doping to accelerate fluorine-19 magic-angle spinning NMR of proteins

Article Title: Accelerated 19F biomolecular magic-angle spinning NMR with paramagnetic dopants

Article References: Becker, L. M., Toscano, G., Kapitonova, A., Singh, R., Guillerm, U., Lichtenecker, R. J., & Schanda, P. (2026). Accelerated 19 F biomolecular magic-angle spinning NMR with paramagnetic dopants. Magnetic Resonance, 7(1), 29-37. https://doi.org/10.5194/mr-7-29-2026

Image Credits: AI Generated

DOI: 10.5194/mr-7-29-2026

Keywords: fluorine-19 NMR, magic-angle spinning, paramagnetic doping, gadolinium chelates, protein structure, solid-state NMR, relaxation enhancement, TET2 aminopeptidase, fluorotryptophan labelling, T1 relaxation, structural biology, spectroscopy