Nuclear magnetic resonance databases

Specinfo, from Chemical Concepts, is a factual database information system for spectroscopic data with more than 660000 digital spectra of 150000 associated structures [24], The database covers nuclear magnetic resonance spectra ( H-, C-, N-, O-, F-, P-NMR), infrared spectra (IR), and mass spectra (MS). In addition, experimental conditions (instrument, solvent, temperature), coupling constants, relaxation time, and bibliographic data are included. The data is cross-linked to CAS Registry, Beilstein, and NUMERIGUIDE. 

Nuclear Magnetic Resonance Spectroscopy. Bmker s database, designed for use with its spectrophotometers, contains 20,000 C-nmr and H-nmr, as weU as a combined nmr-ms database (66). Sadder Laboratories markets a PC-based system that can search its coUection of 30,000 C-nmr spectra by substmcture as weU as by peak assignments and by fiiU spectmm (64). Other databases include one by Varian and a CD-ROM system containing polymer spectra produced by Tsukuba University, Japan. CSEARCH, a system developed at the University of Vieima by Robien, searches a database of almost 16,000 C-nmr. Molecular Design Limited (MDL) has adapted the Robien database to be searched in the MACCS and ISIS graphical display and search environment (63). Projects are under way to link the MDL system with the Sadder Hbrary and its unique search capabiHties. 

Use of an integrated system incorporating CCC separation, PDA detector, and LC-MS proved to be a valuable tool in the rapid identification of known compounds from microbial extracts.6 This collection of analytical data has enabled us to make exploratory use of advanced data analysis methods to enhance the identification process. For example, from the UV absorbance maxima and molecular weight for the active compound(s) present in a fraction, a list of potential structural matches from a natural products database (e.g., Berdy Bioactive Natural Products Database, Dictionary of Natural Products by Chapman and Hall, etc.) can be generated. Subsequently, the identity of metabolite(s) was ascertained by acquiring a proton nuclear magnetic resonance ( H-NMR) spectrum. 

Figure 5.9. Spectral search at Spectral Database Systems (SDBS). The infrared (IR), nuclear magnetic resonance H-NMR and 13C-NMR), electron spin resonance (ESR), and mass (MS) spectra of organic compounds and common biochemical compounds can be viewed/retrieved from SDBS.

Nuclear magnetic resonance spectroscopy is a powerful technique for investigating structure of biomolecules. The 1H- and 13C-NMR spectra of L-a-amino acids have been compiled (Wiithrich, 1986) and can be retrieved from SDBS. Design a database for 1H- or 13C-NMR data that can be used in the identification of amino acids. 

In fact, a tremendous amount of information is available on the structures of biological macromolecules descriptions of structures of proteins and nucleic acids make up major portions of modern textbooks in biochemistry and molecular biology. The Protein Data Bank and the Nucleic Acid Database are online archives that contain sequence and structural data on thousands of specific molecules and complexes of molecules. This structural information comes from in vitro experiments, with structures inferred from the x-ray diffraction patterns of crystallized molecules, spectroscopic measurements using multi-dimensional nuclear magnetic resonance, and a host of other methodologies. 

From the analysis of the data in the LIPID AT database (41), more than 150 different methods and method modifications have been used to collect data related to the lipid phase transitions. Almost 90% of the data is accounted for by less than 10 methods. Differential scaiming calorimetry strongly dominates the field with two thirds of all phase transition records. From the other experimental techniques, various fluorescent methods account for 10% of the information records. X-ray diffraction, nuclear magnetic resonance (NMR), Raman spectroscopy, electron spin resonance (ESR), infrared (IR) spectroscopy, and polarizing microscopy each contribute to about or less than 2-3% of the phase transition data records in the database. Especially useful in gaining insight into the mechanism and kinetics of lipid phase transitions has been time-resolved synchrotron X-ray diffraction (62,78-81). 

In the pharmaceutical industry, the techniques are being used to examine off-target effects particularly for the early identification of toxicity. MOA can be studied through metabolomics and can also be used as a quality control tool for complex mixtures such as foods or herbal medicines. Similarly, the tools and expertise of natural products chemists are essential in metabolomics, particularly in biomarker discovery (see also Volume 9). Biomarker discovery via untargeted metabolomics can lead to metabolite signatures (nuclear magnetic resonance (NMR) spectroscopy, mass spectrometry (MS), etc.) that are not present in current metabolomics databases. This is particularly true for plant secondary metabolism studies and nonmammalian metabolites. Structure elucidation then becomes critical to understanding the metabolomics results and for biomarker development. 

However, other perspectives suggest that there are likely to be as yet unidentified protein folds. As of 2004, the Pfam database (version 10.0) contained 6190 domains only about a third were associated with a protein of known structure. Some of these may be cases in which highly divergent or completely unrelated sequences adopt an already discovered fold. Some may not be amenable to structural characterization, either because they are too large for characterization by nuclear magnetic resonance spectroscopy (NMR), or contain disordered regions that interfere with crystallization. However, there may indeed be unidentified protein domain structures remaining to be discovered. 

---