44321-AC4
Solid-State NMR as a Probe of Flavin Interactions, Electronics and Reactivity
The past year has seen a consolidation of skills and people in my group. We have also adopted extremely powerful new technology that complements the conventional solid-state NMR originally proposed, by developing our flavins as endogenous sources of dynamic nuclear polarization. This is an emerging method that yields NMR signal strength enhancements of more than 200-fold for 1H and theoretical signal enhancements of over 2000 for 13C, and over 5000 for 15N. Signal enhancements on this scale will revolutionize the way NMR is conducted, and are a natural next step for studies such as ours where the target molecules have a naturally-ocurring semiquinone state.
I took advantage of my sabbatical to spend the past year at the Massachusetts Institute of Technology in the Francis Bitter Magnet Lab. I worked in the research group of Prof. Robert G. Griffin, an expert in solid-state NMR. In addition to learning more about how to perform and understand solid-state NMR, I took the opportunity to combine their new advances in DNP with our ability to prepare our flavins in their semiquinone state.
NMR spectroscopy is invaluable for studying the structures of molecules, as well as for elucidating their electronic structures. However the method requires substantial quantities of sample because the signals obtained are weak. This is a direct result of the small energetic separation between nuclear spin energy levels, which translates in turn into a very small Boltzmann excess population in the ground state, on the order of only one part in 105. By contrast, the 660-fold larger magnetic moment of the electron spin produces a correspondingly larger population excess in the electron spin ground state (this is the polarization). Thus, EPR spectroscopy produces signals some 660 times stronger than those of NMR. Dynamic nuclear polarization (DNP) effectively transfers to nuclei the large spin polarization of unpaired electrons. Theoretical NMR signal enhancements of up to 660-fold have not yet been realized, but even the 250-fold enhancements that have been reported by the Griffing group can make previously impossible experiments practical, while also greatly accelerating data collection for experiments we already undertake. This will be particularly valuable for biomolecules, which tend to be huge and therefore tend to be present in relatively fewer copies (lower concentration) in a sample, necessitating more signal averaging.
Previous DNP studies worldwide have relied upon unpaired electrons supplied as exogenous free radicals. These are typically added to 5-20 mM, and are assumed to be randomly distributed throughout the sample, providing close-to-uniform enhancement of all NMR signals. I proposed to exploit the fact that many important biological systems in the area of interest of the P.R.F. (eg. electron transfer in energy transduction) contain at least one cofactor that passes through a seminquinone state as part of the normal catalytic cycle. Our flavins are a particularly convenient example of this. Moreover the flavin (or other cofactor) is bound in a well-defined identical manner in all molecules making up the sample. Thus, DNP enhancements may be tunable via manipulation of experimental conditions to optimize DNP over different time and distance scales. I therefore initiated experiments to test the possibility of using flavin semiquinones as bases for strong and site-specific ('smart') NMR signal enhancement via DNP.
My group prepared samples of model flavins, as well as the test flavoprotein flavodoxin. Both were titrated to their semiquinone states and frozen in sapphire rotors for study by solid-state NMR at Å100 K. DNP experiments were performed in collaboration with Dr. Thorsten Maly, a skilled postdoc in Prof. Griffin's lab. By the end of my stay there, I was also collecting data myself. We began by characterizing enhancements obtainable in 1H NMR spectra. These are not resolved into individual signals, but are sufficiently strong as to allow rapid progress as we determined the parameters of the experiment. Experiments involved irradiation of electron spins with microwaves to drive coupled electron-nuclear transitions, followed by 90° excitation of 1H and observation. We characterized the effects of power and duration of microwave irradiation, and the temperature, on the extent of 1H signal enhancement. Despite the fact that many of the flavodoxin 1H sites are probably too distant to benefit from strong enhancement, we obtained an overall 9-fold enhancement of the 1H signal. This could be transferred to 13C by conventional cross-polarization, to obtain a 13C spectrum that was similarly enhanced. Our current enhancements cannot yet be attributed to specific sites because they are based on uniform isotopic incorporation and NMR observation at only 212 MHz. However now that we have demonstrated success at the 'proof-of-concept' level, we will produce selectively 13C and 15N labelled samples and take advantage of the higher-field spectrometers available at the Bitter Magnet Lab. Prof. Griffin has agreed to a continuing collaboration on this work, and I retain visiting scientist status at M.I.T. for this purpose.
