Hyatt Grand Cypress, Orlando, FL, USA
Visit our hospitality suite in Poinciana B to attend our workshops, play, and learn!
Sunday, April 24th, 4:00–10:30 PM
Monday, April 25th—Wednesday, April 27th, 6:30–10:30 PM
- Monday: Browser-Based NMR Data Processing: The Future of NMR Software is Here
- Tuesday: How to Make the Most of Your NMR Data with ACD/Labs Software
- Wednesday: 3D Structures and Configuration from 2D Spectra: Measuring and Fitting Residual Dipolar Couplings (RDCs) using ACD/Labs Software
Searching Libraries of Known Structures for Dereplication: Benefits and Requirements
Nadia Laschuk, Dimitris Argyropoulos, Rostislav Pol, Sergey Golotvin
A very important question that natural products chemists face is whether the newly isolated compound they have is truly novel or already known. This process, commonly referred to as dereplication, is also performed by people doing competitive product analysis, drug counterfeit analysis, reaction discovery, etc. There are several methods used to accomplish dereplication, including a comparison of the retention time and observed molecular ion from LC-MS, and a comparison of the NMR spectrum to spectral/structure libraries. When using NMR data in dereplication, 13C NMR spectra are preferred1 in most cases, as they provide a very clear fingerprint of the compound’s carbon skeleton. We have previously explored the benefits of using databases with predicted spectra of compounds. Currently, there are several options in this respect, both commercial2 and freely available.3-4 In this poster, we explore the capabilities and requirements of such systems.
In order to establish a starting point, we selected 56 compounds from the Aldrich library of FT-NMR spectra, with a molecular weight (MW) range of 150-800, where most pharmaceutically active compounds are found. We then searched the library containing the predicted 13C spectra of these compounds using the observed experimental peak frequencies together with the MW.
We explored the search options with respect to inclusion/exclusion of the MW information, together with the requirement to accept or reject hits with extra or missing peaks in the experimental spectrum. We saw that the MW information is essential as it provides a very clear starting point and deals effectively with symmetric compounds. We also saw that as the MW increases, the uncertainty and the number of missing or extra peaks increases as well. However, with careful adjustment of the search parameters, the correct result can be identified within a few seconds.
Detailed results will be presented, alongside an optimized workflow that allows one to unambiguously find the correct structures in the database in all cases.
- R.B. Williams, M. O’Neill-Johnson, A.J. Williams, P. Wheeler, R. Pol, A. Moser. (2015). Dereplication of natural products using minimal NMR data inputs. Org. Biomol. Chem., 13(39), 9957–9962.
- D. Argyropoulos, S. Golotvin, R. Pol, A. Moser, N. Ortel, S. Breinlinger, T. Chilzuk and T. Niedermeyer. (2019, April). Efficient Dereplication of Natural Products Using Predicted 13C Spectra (Presented at 60th ENC, Poster 004).
- H. Kalchhauser, W. Robien. (1985). CSEARCH: a computer program for identification of organic compounds and fully automated assignment of carbon-13 nuclear magnetic resonance spectra. Chem. Inf. Comput. Sci., 25(2), 103–108.
- R. Reher, H. W. Kim, C. Zhang et al. (2020). A Convolutional Neural Network-Based Approach for the Rapid Annotation of Molecularly Diverse Natural Products. J. Am. Chem. Soc., 142(9), 4114–4120.
When does Computer-Assisted Structure Elucidation (CASE) spare you from recording (IN)ADEQUATE?
Dimitris Argyropoulos and Mikhail Elyashberg
The INADEQUATE1 experiment was first reported in the literature in 1980 and was heralded as a unique means of establishing the C-C connectivity of organic molecules. However, it has found limited application mostly due to its inherently low sensitivity. Attempts to improve this pitfall were made in subsequent years. In 1996, the creation of the 1,1-ADEQUATE2 experiment and its variants notably allowed the observation of the C-C bonds through their protons. Since these two experiments are the only means by which a scientist can directly observe the carbon skeleton of a molecule, they have been considered over the years to be the “holy grail” of NMR experiments for elucidation of the structures of organic compounds. Despite the emergence of new, more sensitive NMR hardware like cryogenically cooled probes, they nevertheless remain to be of very low sensitivity, requiring hours and days to run with the amounts of product normally isolated.
During the same period, computer-assisted structure elucidation (CASE) has evolved quite significantly3-5 allowing scientists to elucidate large and complex structures using NMR data. The core of any CASE system is the structure generator engine, which will generate all the possible structures derived from the molecular formula and the imposed restrictions based on the observed NMR correlations. Modern structure generators are extremely efficient and can generate millions of isomeric structures within minutes. After the structures are generated, they are ranked, usually based on the agreement between the predicted 13C chemical shifts and those observed experimentally.
Taking the above into account, in this poster, we investigate how relevant experiments like INADEQUATE and ADEQUATE truly are, given the existence of powerful tools like CASE. To do this, we analysed a series of published examples in which (IN)ADEQUATE information was stated as being vitally necessary for unambiguous structure elucidation. We looked to determine whether using HMBC and COSY data within a CASE system could elucidate the structure(s) in a reasonable amount of time without using (IN)ADEQUATE spectra. We found that using only HMBC and COSY allowed us to get the correct solution in a reasonable time without utilizing time-consuming experiments in a series of examples. However, we also found that in most cases of large hydrogen deficient molecules, structure elucidation requires (IN)ADEQUATE spectra, and CASE only facilitates the structure elucidation.
We will be presenting different examples to illustrate the observations. Moreover, we will be showing examples where even though the correct structure has been generated, it was not possible to clearly identify it as the correct structure because others existed that were ranked similarly. We will be discussing methods to resolve these ambiguities and get the correct result without necessarily involving low-sensitivity experiments.
- A. Bax, R. Freeman, S.P. Kempsell. (1980). Natural abundance carbon-13-carbon-13 coupling observed via double-quantum coherence. J. Am. Chem. Soc., 102(14), 4849–4851.
- B. Reif, M. Kock, R. Kerssebaum, H. Kang, W. Fenical, C. Griesinger. (1996). ADEQUATE, a New Set of Experiments to Determine the Constitution of Small Molecules at Natural Abundance.J. Mag. Reson, series A, 118(2), 282-285.
- M. Elyashberg, D. Argyropoulos. (2019). NMR-Based Computer-Assisted Structure Elucidation (CASE) of Small Organic Molecules in Solution: Recent Advances. eMagRes, 8(3), 239–254.
- M. Elyashberg, D. Argyropoulos. (2020). Computer Assisted Structure Elucidation (CASE): Current and future perspectives. Mag. Reson. Chem., 59, 669–690.
- M. E. Elyashberg, A. J. Williams. (2015). Computer-based Structure Elucidation from Spectral Data. The Art of Solving Problems. Springer.