CATALYST-generated Olfactophore Models for Structure-Odor Correlations in Fragrance Research

1 CATALYST®-generated Olfactophore Models for Structure-Odor Correlations in Fragrance Research Nancy Ogihara and Omoshile Clement Accelrys Inc. We provide here a summary of recent work in the area of fragrance research which applied ACCELRYS’ RDD program – CATALYST®. The studies were originally described in a review: “Odds and Trends: Recent Developments in the Chemistry of Odorants” by Philip Kraft, Jerzy A. Bajgrowicz, Caroline Denis, and Georg Fráter, published in Angew. Chem. Int. Ed. 2000, 39, 2980-3010. _____________________________________________________________________________ Abstract A comprehensive review of structure-odor correlations and olfactophore models for the main odor notes of perfume business has been recently published by Kraft et. al.1 The study described structure-odor correlation models generated for fruity, marine, green, floral, spicy, woody, amber, and musky odors. An olfactophore model (related to a pharmacophore model for biological activity of drug candidates) is a representation of a generalized set of molecular features that are critical for a given odor, and correspond to binding properties and steric geometries of the olfactophore-binding receptor active site. The olfactophore models presented are based on the molecular similarity of known odorants. They provide some insight into the molecular parameters responsible for a given odor, and ideally, will be helpful for computer-aided design of new odorants. The olfactophore models generated in the examples described below were derived using Accelrys’ CATALYST® program.2 _____________________________________________________________________________ Olfactophore Model for Fruity Odorants Fruity odorants are very popular in the perfume industry and are present in nearly every feminine fragrance available. Popular fruity notes include peach, coconut, apple, grapefruit, and blackcurrant. Even though esters have been used as the basis for fruity odorants, their odor is still difficult to predict. Kraft et. al. used the CATALYST® program to generate an olfactophore model for pear odorant. This model contained five chemical features -two hydrogen-bond acceptors (HBA) and three hydrophobic features (HYD). The two oxygen atoms of each ester moiety maps the hydrogen-bond acceptors in the model, while the hydrophobic features fit onto two hydrophobic pockets positioned 7.2/7.9 Å and 5.3/5.9 Å away from the HBAs on one side, and a third hydrophobic pocket located on the opposite side of the HBA and positioned 2.5/3.1 Å away. Addition of a ligand-inaccessible zone defined by excluded volume spheres, and positioned 2.9/4.0 Å away accounts for the steric geometry of the receptor binding site residues. Olfactophore models which aligned to ligand conformations within 3 kcal/mol energy range were selected as the best fitting models, and were used to estimate activity of other new odorant candidates based on their fit to the olfactophore features (Figure 1).2 Figure 1: CATALYST®-generated olfactophore model for pear odorants. Top: aligned with 1; Bottom: aligned with 2. An olfactophore model for galbanum notes The constituents responsible for the harsh green odor of galbanum (Ferula galbaniflua Boiss. et Buhse and ferula rubicaulis Boiss) are minor constituents, most prominent of which are (3E, 5Z)-undeca-1,3,5-triene and 2-methoxy-3-isobutylpyrazine. Study of the galbanum odor requirements revealed a family of proposed olfactophores consisting of one non-aromatic hydrophobic function, one oriented HBA, and two customized electron-rich fragments (an ether or an aliphatic double bond). The olfactophore hypotheses thus generated consist of a bulky hydrophobic function with an electron-rich link to the pent-4-en-1-one or the flexible allyloxycarbonyl sidechain of these compounds. The 3D shape of the hydrophobic pocket was hypothesized based on a Conolly surface of the aligned active structures. Small changes such as addition of a single carbon to the aromatic ring, or shifting of methyl substituents by a single position in the ring, can result in complete loss of the characteristic green galbanum odor and helps account for galbanum discriminating features. A Muguet Olfactophore Muguet is the delicate floral odor associated with lilies of the valley (Convallaria majalis L.), and is characterized by a hint of greenness accented with a rosy-lemony combination. It has something in common with the scent of rose, jasmine, and lilac blossoms. The muguet odor contains additional different characteristics creating a complex odor that is difficult to reproduce by a single compound. Hydroxycitronellal comes close and therefore stands as a template for muguet odorants. The muguet olfactophore model was generated using a relatively high conformational energy barrier of 15 kcal/mol, and consists of a bifunctional unit with one hydrogen bond donor (HBD), one hydrogen-bond acceptor (HBA), and two hydrophobic functions (HYD). The HBD, HBA, and one of the HYD functions form a quasi equilateral triangle 4.5 Å on a side. The second hydrophobic function sits 2.4 Å away from the first hydrophobic function, 5.7 Å from the HBA, and 8.1 Å from the HBD. A single excluded volume accounts for steric constraints within the receptor binding site. O O O OEt 123 Sandalwood olfactophores The woody scent of East Indian sandalwood oil (santalum album L.) is considered one of the most precious raw materials in perfumery. Because of its rare availability and its complex stereoisomeric composition, it has been difficult to analyze and produce synthetically. Previous studies showed distance constraints between the osmophoric hydroxyl group, which is responsible for orienting the molecule in the receptor binding site, and a bulky lipophilic moiety. The CATALYST®-generated olfactophore model (Figure 2) was obtained using ligand conformations between 10 and 15 kcal/mol, and contained a bulky lipophilic moiety with four hydrophobic regions, one of which sits in a flexible spacer near the osmophoric hydrogen-bond donor (HBD). The best hypothesis was mapped by (Z)-(-)-β-santalol (3) and Javanol (4). Figure 2: Sandalwood olfactophore model mapped by (Ζ)-(−)-β-santanol (cyan, 3), and by Javanol (yellow, 4).3 Amber olfactophore Chemicals considered to have amber-like odor are ethers, ketones, and cyclic ketals derived from natural terpenes, with the cyclic ether, Ambrox (5) being the prototypical example of ambery. The CATALYST®-derived olfactophore model4 consisted of four hydrophobic functions, one oriented hydrogen-bond acceptor (HBA), and six excluded volume spheres, correspondng to the geometric constraints of the putative binding pocket of the amber olfactory receptor. It was shown that the positions of the empirically determined excluded volumes significantly improved the predictability of the presence or absence of the amber odor. Despite the crudeness of this olfactophore model, it was successfully applied to the design and development of a new ambery compound (6) (Figure 3). Figure 3: Ambery olfactophore model mapped by Ambrox (Blue, 5) and the new lead candidate (R,R’ = H, iPr; yellow, 6).4 O O O R R’ 5 6 OHOH 434 Musk olfactophore models Musk odorants belong to structurally diverse chemical classes with different possible features for binding. These range from macrocyclic aliphatic compounds, to polycyclic benzene derivatives, to nitro arenes. A Catalyst-generated olfactophore model was derived based on the threshold data of 13 musk-like thiamacrolide compounds. The model consisted of two HBA’s located 6.9 Å apart, and three hydrophobic functions located 4.1/2.87.1/1.7, and 8.3/5.0 Å, respectively, away from the HBAs. The second HBA appears to account either for intense musk odor, or nearly complete lack of musk odor in some dicarbonyl compounds, depending whether or not this feature is mapped by a ligand. References: 1. P. Kraft, J.A. Bajgrowicz, C. Denis, and G. Fráter. Angew. Chem. Int. Ed., 2000, 39, 2980-3010. 2. CATALYST® 4.0. Accelrys Inc., 9685 Scranton Road, San Diego, CA 92121. 3. J.A. Bajgrowicz and G. Fráter, Enantiomer, 2000, 5, 229-234. 4. J.A. Bajgrowicz and C. Broger. In Proceedings of the 13th International Congress of Flavours, Fragrances and Essential Oils. K.H.C. Baser, ed. AREP Publications