From Paramagnetic Proteins to Field Alignment of Small Molecules
Tools for Structure Determination via NMR
by Niels Ulrich Karschin
Date of Examination:2022-01-26
Date of issue:2022-03-17
Advisor:Prof. Dr. Christian Griesinger
Referee:Prof. Dr. Christian Griesinger
Referee:Prof. Dr. Marina Bennati
Referee:Prof. Dr. Gottfried Otting
Files in this item
This file will be freely accessible after 2023-01-25.
EnglishIn this dissertation we show how to push the boundaries of structure elucidation using tensorial NMR parameters. Possible sources for these effects are paramagnetic centers or molecular alignment, and they contain valuable long-range structural information that more common scalar NMR observables, such as couplings, chemical shifts, or nuclear Overhauser constraints, cannot provide. However, their interpretation can be challenging. We show the application of these tensorial constraints in structural questions both in proteins as well as small molecules. Paramagnetic NMR is particularly well suited to study interdomain dynamics of proteins as the paramagnetic center is localized within one domain. Therefore, the nuclei within this domain will experience the full degree of paramagnetic effects induced by the metal. The other domain, on the other hand, moves with respect to the paramagnetic center, and the measured effects will be a time average over all relative orientations. This way, it is possible to probe the conformational ensemble sampled by the two domains. This approach is not new, and the protein calmodulin and its various complexes have been a popular target for these studies for more than 20 years. For the sake of simplicity, this interdomain motion is often reduced to a scalar quantity called an order parameter, and although this certainly gives some first insight into the motion, we believe that much more information can be extracted from paramagnetic constraints. In this work we report the application of this methodology to the complex of calmodulin with Munc13-1. This complex is interesting due to the unique 1-5-8-26 binding motif of Munc13-1, which is exceptionally long, and the complex formation is expected to restrict the interdomain flexibility of calmodulin much less than in other complexes. To exploit the full potential of paramagnetic NMR, we have acquired four types of RDCs and four types of PCSs for six different metals, leading to a very large number of 2691 constraints for the mobile domain. From this data we have first determined an order parameter of the complex of 0.16. We then go further to find ensembles that give us a very detailed picture of the conformational space sampled by calmodulin, and we find that the interdomain movement is a combination of both rotation and translation. Finally, we believe that this data set could be used in the future to explore more formal motional models and to investigate more strictly which characteristics of the interdomain motion can be distinguished by RDCs and PCSs and which cannot. In natural product chemistry anisotropic NMR is particularly valuable to elucidate molecules with separated stereoclusters. However, we believe that it is useful as a more general tool to corroborate one's findings. Natural products can be very complex molecules, and the unambiguous interpretation of scalar NMR parameters can turn out to be quite difficult. As a result, structural misassignments are more common than they should be. It is therefore always a good idea to apply different approaches for structure elucidation and check whether they all come to the same conclusion. In particular, the combination of spectroscopic data with molecular modeling is a strategy that has seen a lot of development in the more recent past, and many techniques rely on results from DFT and molecular mechanics simulations. Anisotropic NMR is an example of such a combination, and the field-alignment approaches that we developed benefit even more from molecular modeling. Inspired by the magnetic field alignment of paramagnetic proteins, we have shown that it is possible to use diamagnetic alignment on natural products for the elucidation of their structure. Unlike in the case of alignment media, the degree of magnetic alignment can be reliably predicted by DFT, which greatly enhances the discriminating power of this approach. This is expected to be particularly useful for flexible molecules. We have also set out to design a device for the alignment via electric fields. While this shares many of the benefits with magnetic alignment, it may be even superior since the alignment could be turned on and off at the push of a button, eliminating the need for multiple magnetic fields. Additionally, it may be applicable to a larger variety of molecules. This is mostly an engineering challenge, and so far we were not able to induce alignment via electric fields. Nonetheless, we have learned a plethora of lessons in this rather unfamiliar field, and we shall continue to pursue this goal in the future. Finally, we have made an excursus into metrology. In all our NMR experiments it was necessary to set the sample temperature as accurately as possible to avoid or reduce temperature-dependent systematic perturbations of the chemical shift. This was typically done using a standard sample with a pronounced and well-known chemical shift dependence, such as methanol-d4. Since the existing calibration of this NMR thermometer lacked in both range and accuracy, we decided to recreate it. Our new calibration extends the original one by about 100 K to the lower end, and the experimental uncertainty was reduced by almost an order of magnitude.
Keywords: NMR; calmodulin; alignment; paramagnetic NMR; natural products; RDC; PCS; RCSA; interdomain dynamics