Model free analysis of experimental residual dipolar couplings in small organic compounds.

Residual dipolar couplings (RDCs) contain information on the relative arrangement and dynamics of internuclear spin vectors in chemical compounds. Classically, RDC data is analyzed by fitting to structure models, while model-free approaches (MFA) directly relate RDCs to the corresponding internuclear vectors. The recently introduced software TITANIA implements the MFA and extracts structure and dynamics parameters directly from RDCs to facilitate de novo structure refinement for small organic compounds. Encouraged by our previous results on simulated data, we herein focus on the prerequisites and challenges faced when using purely experimental data for this approach. These concern mainly the fact that not all couplings are accessible in all media, leading to voids in the RDC matrix and the concomitant effects on the structure refinement. It is shown that RDC data sets obtained experimentally from currently available alignment media and measurement methods are of sufficient quality to allow relative configuration determination even when the relative configuration of the analyte is completely unknown.

TITANIA: Model-Free Interpretation of Residual Dipolar Couplings in the Context of Organic Compounds.

Residual dipolar couplings (RDCs) become increasingly important as additional NMR parameters in the structure elucidation of organic compounds but are usually used in fitting procedures to discriminate between (computed) structures that are in accordance with RDCs and others that can be ruled out. Thus, the determination of configurations requires prior structural information. The direct use of RDCs as restraints to construct structures based on RDCs has only recently begun also in organic compounds. No protocol has been published though that uses the vector and dynamics information available in multialignment data sets directly for the joint determination of conformation and configuration of organic compounds. This is proposed in the current study. We show that by employing these data, even a flat or random start structure converges into the correctly configured structure when employing multiple alignment data sets in our iterative procedure. The requirements in terms of the number of RDCs and alignment media necessary are discussed in detail.