Intrinsic bitterness of flavonoids and isoflavonoids and masking of their taste activity

Many flavonoids and isoflavonoids have been associated with beneficial health effects. Therefore, consumption of (iso)flavonoid-rich food products, and enrichment of foods with (iso)flavonoids is becoming increasingly popular. However, several (iso)flavonoids have been reported as bitter. Consequently, their incorporation in (or fortification of) foods can introduce (or enhance) bitterness. Hence, debittering strategies are demanded. Some (iso)flavonoids have unknown taste properties, as they have never been incorporated in high levels in food products. For other (iso)flavonoids, contradictory findings on bitterness have been made in sensory tests. Therefore, objective tests are necessary to identify which (iso)flavonoids contribute to bitterness of a food product. An objective tool to study bitterness is a cell-based bitter taste receptor assay. Twenty-five different bitter taste receptors (hTAS2Rs) occur on the human tongue, each of which has been introduced in a separate human embryonic kidney (HEK)293 cell line. With these, the “intrinsic bitterness” of a compound can be investigated<em>in vitro</em>. Intrinsic bitterness is the capacity of a compound to activate bitter taste receptors, uncoupled from cross-modal interactions and interactions with salivary proteins and oral mucosa. The aim of this research was to study the intrinsic bitterness of a large set of (iso)flavonoids and to investigate structural requirements for (iso)flavonoids to activate the bitter receptors identified. A subsequent aim was the investigation of different debittering strategies by the use of the bitter receptor assay. <strong>Chapter 1</strong>provides an overview of flavonoids and isoflavonoids with respect to their structural classification, sensorial properties and occurrence as dietary compounds. Taste perception and the mode of action of bitter taste receptors are introduced. The measurement of bitter receptor activation <em>in vitro</em> is explained, as well as strategies to reduce bitter receptor activation, and bitter taste in general. A state-of-the-art overview of all 25 bitter taste receptors is given with respect to known agonists and antagonists. The aim of <strong>Chapter 2</strong>was to identify the bitter receptor(s) that recognize the bitter taste of the soy isoflavone genistein. Screening of all 25 human bitter receptors revealed genistein as agonist of hTAS2R14 and hTAS2R39. Genistein displayed threshold values of 4 and 8 µM on hTAS2R14 and hTAS2R39, and EC₅₀ values of 29 and 49 µM, respectively. Besides, the behavior of structurally similar isoflavonoids was investigated. Although the two receptors are not closely related, the results for hTAS2R14 and hTAS2R39 were similar towards most isoflavonoid aglycones. Glucosylation of isoflavones seemed to inhibit activation of hTAS2R14, whereas four of five glucosylated isoflavones were agonists of hTAS2R39, namely glycitin, genistin, acetyl genistin, and malonyl genistin. A total of three hydroxyl substitutions of the A- and B-rings of the isoflavonoids seemed to be more favorable for receptor activation than less hydroxyl groups. The concentration of the trihydroxylated genistein in several soy foods exceeds the bitter receptor threshold values determined, whereas those of other soy isoflavones are around or below their respective threshold values. Despite its low concentration, genistein might be one of the main contributors to the bitterness of soy products. Furthermore, the bioactive isoflavonoids equol and coumestrol activated both receptors, indicating that their sensory impact should be considered when used as food ingredients. In <strong>Chapter 3</strong>, the intrinstic bitterness of (iso)flavonoids, which can hamper their use as food bioactives, was investigated further.The effect of a large set of structurally similar (iso)flavonoids on the activation of bitter receptors hTAS2R14 and hTAS2R39 was tested, and their structural requirements to activate these receptors were predicted. In total, 68 compounds activated hTAS2R14 and 70 compounds activated hTAS2R39, amongst which 58 ligands were overlapping. Their activation threshold values varied over a range of three log units between 0.12 and 500 μM. Ligand-based 2D-fingerprint and 3D-pharmacophore models were created <em>to detect structure activity relationships</em>. The 2D-models demonstrated excellent predictive power in identifying bitter (iso)flavonoids and discrimination from inactive ones. The structural characteristics for an (iso)flavonoid to activate hTAS2R14 and hTAS2R39 were determined by 3D-pharmacophore models to be composed of two (for hTAS2R14) or three (for hTAS2R39) hydrogen bond donor sites, one hydrogen bond acceptor site, and two aromatic ring structures, of which one had to be hydrophobic. An additional hydrogen bond donor feature for hTAS2R39 ligands indicated the possible presence of another complementary acceptor site in the binding pocket, compared to hTAS2R14. Hydrophobic interaction of the aromatic feature with the binding site might be of higher importance in hTAS2R14 than in hTAS2R39. Together, this might explain why OH-rich compounds showed different behavior towards the two bitter receptors. The combination of <em>in vitro</em> data and different <em>in silico</em> methods created a good insight in activation of hTAS2R14 and hTAS2R39 by (iso)flavonoids and provided a powerful tool in prediction of their potential bitterness. By understanding the “bitter motif”, introduction of bitter taste in functional foods enriched in (iso)flavonoid bioactives might be avoided. Bitter receptor hTAS2R39 is activated by many different classes of bitter compounds, amongst which (iso)flavonoids. Nevertheless, several flavanones are known to mask bitter taste sensorially, andnot all flavanones reported in <strong>Chapter 3</strong> activated hTAS2R39. For that reason, in <strong>Chapter 4</strong>, fourteen flavanones were investigated for their potential to reduce activation of hTAS2R39 by epicatechin gallate (ECG), one of the main bitter compounds present in green tea.Three compounds showed inhibitory behavior towards the activation of hTAS2R39 by ECG: 4’-fluoro-6-methoxyflavanone, 6,3’-dimethoxyflavanone, and 6-methoxyflavanone (in order of decreasing potency). The 6-methoxyflavanones also inhibited activation of hTAS2R14 (another bitter receptor activated by ECG), though to a lesser extent. Dose-response curves of ECG at various concentrations of the most potent antagonist 4’-fluoro-6-methoxyflavanone and wash-out experiments indicated reversible insurmountable antagonism. The same effect was observed for the structurally different agonist denatonium benzoate, suggesting a non-competitive orthosteric mechanism. The bitter receptor blockers identified might not be applicable to food products. Nevertheless, they create insight into structural requirements, which might lead to other, more suitable, blockers. <strong>Chapter 5</strong>investigates another strategy to reduce bitterness, namely complexation of bitter flavonoids with food proteins. The binding characteristics of the bitter tea compound epigallocatechin gallate (EGCG) to purified food proteins, and their equivalent food-grade preparations, were related to their effects on reducing bitter receptor activation by EGCG <em>in vitro </em>and their bitter-masking potential <em>in vivo</em>. β-Casein, in particular, and several gelatins, are known as strong binders of EGCG, contrary to β-lactoglobulin. Also in the bitter receptor assay, β-casein showed the strongest effect, with a maximum reduction of hTAS2R39 activation of about 93%. A similar potency was observed for Na-caseinate, which was applied as food-grade alternative for β-casein. β-Lactoglobulin had little effect on bitter receptor activation, as expected based on its low binding affinity for EGCG. The bitter-masking potential of Na-caseinate was confirmed <em>in vivo</em> using a trained sensory panel. β-Lactoglobulin also slightly reduced EGCG bitter perception, which could not be directly related to its binding capacity. The bitter receptor assay appeared to be a valid tool to evaluate <em>in vitro </em>the efficacy of food proteins as complexing agents for bitter-masking. <strong>Chapter 6</strong>discusses the findings presented in this thesis, addresses prospects and limitations of the bitter receptor cell assay, presents additional results on testing (iso)flavonoids for possible antagonistic properties, and compares taste evaluation by sensory tests, receptor assays and modeling. Furthermore, it evaluates strategies for bitter taste reduction, and applies the findings to soy products and tea. The systematic investigation of (iso)flavonoid aglycones showed that the substitution pattern of (iso)flavonoids is of higher importance for bitter receptor activation than the backbone structure. In case of bitter receptor antagonists, the substitution pattern as well as backbone structure revealed to be crucial for functionality. The bitter receptor assay was shown to be an appropriate tool not only for identification of bitter receptor agonists and antagonists, but also for identification of reduced receptor activation by complexing agents. Based on the findings of this thesis, it was concluded that complexation with food proteins is the most promising strategy to reduce bitter taste of flavonoids in tea. On the other hand, for soybean isoflavones, debittering by use of bitter receptor blockers seemed to be a promising debittering strategy. Alternatively to the use of receptor blockers, processing conditions (leading to low isoflavone aglycone formation) or raw material choice (i.e. cultivars low in genistein forms) were recommended. In conclusion, the choice of debittering strategies depends on the molecular structure of the bitter food compounds, as exemplified for soybean products and tea. Therefore, each food product seems to require its own tailor-made debittering solution.