Blind trials of computer-assisted structure elucidation software

DeepSAT: Learning Molecular Structures from Nuclear Magnetic Resonance Data

The identification of molecular structure is essential for understanding chemical diversity and for developing drug leads from small molecules. Nevertheless, the structure elucidation of small molecules by Nuclear Magnetic Resonance (NMR) experiments is often a long and non-trivial process that relies on years of training. To achieve this process efficiently, several spectral databases have been established to retrieve reference NMR spectra. However, the number of reference NMR spectra available is limited and has mostly facilitated annotation of commercially available derivatives. Here, we introduce DeepSAT, a neural network-based structure annotation and scaffold prediction system that directly extracts the chemical features associated with molecular structures from their NMR spectra. Using only the 1H-13C HSQC spectrum, DeepSAT identifies related known compounds and thus efficiently assists in the identification of molecular structures. DeepSAT is expected to accelerate chemical and biomedical research by accelerating the identification of molecular structures.

Phaeosphspirone (1/1'), a pair of unique polyketide enantiomers with an unusual 6/5/5/6 tetracyclic ring from the desert plant endophytic fungus Phaeosphaeriaceae sp

Phaeosphspirone, an undescribed polyketide with a unique 6/5/5/6-fused tetracyclic system, and two known analogues, herbarin and O-methylherbarin, were purified from the endophytic fungus Phaeosphaeriaceae sp. isolated from the desert plant Bassia dasyphylla. The connectivity and relative configuration of phaeosphspirone was elucidated by comprehensive HR-ESI-MS and NMR analysis together with a computer-assisted structure elucidation (CASE) method. A pair of enantiomers existing in phaeosphspirone were separated by HPLC chromatography after reacting with chiral reagents, from which the absolute configuration of phaeosphspirone was simultaneously determined based on Mosher's rule. This tandem strategy provides a useful approach for the separation and stereochemical determination of enantiomers possessing secondary hydroxyl groups. The structural feature of phaeosphspirone, herbarin and O-methylherbarin together with gene cluster analysis suggested their polyketide biosynthetic origin. Herbarin and O-methylherbarin exhibited moderate cytotoxicity against three cancer cell lines.

UV-adVISor: Attention-Based Recurrent Neural Networks to Predict UV-Vis Spectra

Ultraviolet-visible (UV-Vis) absorption spectra are routinely collected as part of high-performance liquid chromatography (HPLC) analysis systems and can be used to identify chemical reaction products by comparison to the reference spectra. Here, we present UV-adVISor as a new computational tool for predicting the UV-Vis spectra from a molecule's structure alone. UV-Vis prediction was approached as a sequence-to-sequence problem. We utilized Long-Short Term Memory and attention-based neural networks with Extended Connectivity Fingerprint Diameter 6 or molecule SMILES to generate predictive models for the UV spectra. We have produced two spectrum datasets (dataset I, N = 949, and dataset II, N = 2222) using different compound collections and spectrum acquisition methods to train, validate, and test our models. We evaluated the prediction accuracy of the complete spectra by the correspondence of wavelengths of absorbance maxima and with a series of statistical measures (the best test set median model parameters are in parentheses for model II), including RMSE (0.064), R² (0.71), and dynamic time warping (DTW, 0.194) of the entire spectrum curve. Scrambling molecule structures with the experimental spectra during training resulted in a degraded R², confirming the utility of the approaches for prediction. UV-adVISor is able to provide fast and accurate predictions for libraries of compounds.

ACD/Structure Elucidator: 20 Years in the History of Development

The first methods associated with the Computer-Assisted Structure Elucidation (CASE) of small molecules were published over fifty years ago when spectroscopy and computer science were both in their infancy. The incredible leaps in both areas of technology could not have been envisaged at that time, but both have enabled CASE expert systems to achieve performance levels that in their present state can outperform many scientists in terms of speed to solution. The computer-assisted analysis of enormous matrices of data exemplified 1D and 2D high-resolution NMR spectroscopy datasets can easily solve what just a few years ago would have been deemed to be complex structures. While not a panacea, the application of such tools can provide support to even the most skilled spectroscopist. By this point the structures of a great number of molecular skeletons, including hundreds of complex natural products, have been elucidated using such programs. At this juncture, the expert system ACD/Structure Elucidator is likely the most advanced CASE system available and, being a commercial software product, is installed and used in many organizations. This article will provide an overview of the research and development required to pursue the lofty goals set almost two decades ago to facilitate highly automated approaches to solving complex structures from analytical spectroscopy data, using NMR as the primary data-type.

Minimizing the risk of deducing wrong natural product structures from NMR data

There continues to be a disturbing number of natural products reported in the literature whose structures are incorrect. At least in part, this reflects the fact that many natural product chemists have limited formal nuclear magnetic resonance training. Gaps in training and lack of awareness regarding the challenges and ambiguities associated with two-dimensional nuclear magnetic resonance data interpretation can easily lead to errors in structure elucidation. The purpose of this tutorial is to point out some of these issues, highlight the kinds of errors that have been made and provide specific advice on how to avoid these missteps such that the risk of reporting a wrong structure is minimized.

EBC-232 and 323: A Structural Conundrum Necessitating Unification of Five In Silico Prediction and Elucidation Methods

Structurally unique halimanes EBC-232 and EBC-323, isolated from the Australian rainforest plant Croton insularis, proved considerably difficult to elucidate. The two diastereomers, which consist an unusual oxo-6,7-spiro ring system fused to a dihydrofuran, were solved by unification and consultation of five in silico NMR elucidation and prediction methods [i.e., ACDLabs, olefin strain energy (OSE), DP4, DU8+ and TD DFT CD]. Structure elucidation challenges of this nature are prime test case examples for empowering future AI learning in structure elucidation.

The role of computer-assisted structure elucidation (CASE) programs in the structure elucidation of complex natural products

Computer-assisted structure elucidation can help to determine the structures of complex natural products while minimizing the risk of structure errors.

Unequivocal determination of caulamidines A and B: application and validation of new tools in the structure elucidation tool box

Newly described NMR experimental approaches can provide valuable structural details and a complementary means of structure verification.

Ambiguities and errors in the structural assignment of organic molecules hinder both drug discovery and total synthesis efforts. Newly described NMR experimental approaches can provide valuable structural details and a complementary means of structure verification. The caulamidines are trihalogenated alkaloids from a marine bryozoan with an unprecedented structural scaffold. Their unique carbon and nitrogen framework was deduced by conventional NMR methods supplemented by new experiments that define 2-bond heteronuclear connectivities, reveal very long-range connectivity data, or visualize the ^35,37Cl isotopic effect on chlorinated carbons. Computer-assisted structural elucidation (CASE) analysis of the spectroscopic data for caulamidine A provided only one viable structural alternative. Anisotropic NMR parameters, specifically residual dipolar coupling and residual chemical shift anisotropy data, were measured for caulamidine A and compared to DFT-calculated values for the proposed structure, the CASE-derived alternative structure, and two energetically feasible stereoisomers. Anisotropy-based NMR experiments provide a global, orthogonal means to verify complex structures free from investigator bias. The anisotropic NMR data were fully consistent with the assigned structure and configuration of caulamidine A. Caulamidine B has the same heterocyclic scaffold as A but a different composition and pattern of halogen substitution. Caulamidines A and B inhibited both wild-type and drug-resistant strains of the malaria parasite Plasmodium falciparum at low micromolar concentrations, yet were nontoxic to human cells.

Characterization and Synthesis of Eudistidine C, a Bioactive Marine Alkaloid with an Intriguing Molecular Scaffold

An extract of Eudistoma sp. provided eudistidine C (1), a heterocyclic alkaloid with a novel molecular framework. Eudistidine C (1) is a racemic natural product composed of a tetracyclic core structure further elaborated with a p-methoxyphenyl group and a phenol-substituted aminoimidazole moiety. This compound presented significant structure elucidation challenges due to the large number of heteroatoms and fully substituted carbons. These issues were mitigated by application of a new NMR pulse sequence (LR-HSQMBC) optimized to detect four- and five-bond heteronuclear correlations and the use of computer-assisted structure elucidation software. Synthesis of eudistidine C (1) was accomplished in high yield by treating eudistidine A (2) with 4(2-amino-1H-imidazol-5-yl)phenol (4) in DMSO. Synthesis of eudistidine C (1) confirmed the proposed structure and provided material for further biological characterization. Treatment of 2 with various nitrogen heterocycles and electron-rich arenes provided a series of analogues (5-10) of eudistidine C. Chiral-phase HPLC resolution of epimeric eudistidine C provided (+)-(R)-eudistidine C (1a) and (-)-(S)-eudistidine C (1b). The absolute configuration of these enantiomers was assigned by ECD analysis. (-)-(S)-Eudistidine C (1b) modestly inhibited interaction between the protein binding domains of HIF-1α and p300. Compounds 1, 2, and 6-10 exhibited significant antimalarial activity against Plasmodium falciparum.

Activity-Independent Discovery of Secondary Metabolites Using Chemical Elicitation and Cheminformatic Inference

Most existing antibiotics were discovered through screens of environmental microbes, particularly the streptomycetes, for the capacity to prevent the growth of pathogenic bacteria. This "activity-guided screening" method has been largely abandoned because it repeatedly rediscovers those compounds that are highly expressed during laboratory culture. Most of these metabolites have already been biochemically characterized. However, the sequencing of streptomycete genomes has revealed a large number of "cryptic" secondary metabolic genes that are either poorly expressed in the laboratory or that have biological activities that cannot be discovered through standard activity-guided screens. Methods that reveal these uncharacterized compounds, particularly methods that are not biased in favor of the highly expressed metabolites, would provide direct access to a large number of potentially useful biologically active small molecules. To address this need, we have devised a discovery method in which a chemical elicitor called Cl-ARC is used to elevate the expression of cryptic biosynthetic genes. We show that the resulting change in product yield permits the direct discovery of secondary metabolites without requiring knowledge of their biological activity. We used this approach to identify three rare secondary metabolites and find that two of them target eukaryotic cells and not bacterial cells. In parallel, we report the first paired use of cheminformatic inference and chemical genetic epistasis in yeast to identify the target. In this way, we demonstrate that oxohygrolidin, one of the eukaryote-active compounds we identified through activity-independent screening, targets the V1 ATPase in yeast and human cells and secondarily HSP90.

Turning Spiroketals Inside Out: A Rearrangement Triggered by an Enol Ether Epoxidation

Spiroketals organize small molecule structures into well-defined, three-dimensional configurations that make them good ligands of proteins. We recently discovered a tandem cycloisomerization-dimerization reaction of alkynyl hemiketals that delivered polycyclic, enol-ether-containing spiroketals. Here we describe rearrangements of those compounds, triggered by epoxidation of their enol ethers that completely remodel their structures, essentially turning them "inside out". Due to the high level of substitution on the carbon skeletons of the substrates and products, characterization resorted to X-ray crystallography and advanced computation and NMR techniques to solve the structures of representative compounds. In particular, a new proton-detected ADEQUATE NMR experiment (1,1-HD-ADEQUATE) enabled the unequivocal assignment of the carbon skeleton of one of the new compounds. Solution of the structures of the representative compounds allowed for the assignment of product structures for the other compounds in two separate series. Both the rearrangement and the methods used for structural determination of the products are valuable tools for the preparation of characterization of new small molecule compounds.

Computer-Assisted Structure Elucidation of Black Chokeberry (Aronia melanocarpa) Fruit Juice Isolates with a New Fused Pentacyclic Flavonoid Skeleton

Melanodiol 4″-O-protocatechuate (1) and melanodiol (2) represent novel flavonoid derivatives isolated from a botanical dietary supplement ingredient, dried black chokeberry (Aronia melanocarpa) fruit juice. These noncrystalline compounds possess an unprecedented fused pentacyclic core with two contiguous hemiketals. Due to having significant hydrogen deficiency indices, their structures were determined using computer-assisted structure elucidation software. The in vitro hydroxyl radical-scavenging and quinone reductase-inducing activity of each compound are reported, and a plausible biogenetic scheme is proposed.

Using NMR to identify and characterize natural products

Over the past 28 years there have been several thousand publications describing the use of 2D NMR to identify and characterize natural products. During this time period, the amount of sample needed for this purpose has decreased from the 20-50 mg range to under 1 mg. This has been due to both improvements in NMR hardware and methodology. This review will focus on mainly methodology improvements, particularly in pulse sequences, acquisition and processing methods which are particularly relevant to natural product research, with lesser discussion of hardware improvements.