Exploring Natural Products for New Taste Sensations

Home , Polygodial, Umbellulone

Nutrition Chemistry in Switzerland CHIMIA 2011, 65, No. 6 407 doi:10.2533/chimia.2011.407 Chimia 65 (2011) 407–410 © Schweizerische Chemische Gesellschaft Exploring Natural Products for New Taste Sensations

Christian Starkenmann*, Isabelle Cayeux, and Anthony A. Birkbeck

Abstract: This paper discusses the discovery of uncommon taste or trigeminal active compounds and their associated sensory analysis using human tasting panels with the aim of enhancing the overall taste experience whilst reducing where possible the sugar and salt content of foods. The first example outlines the discovery of the sensory quality attributes of (R)-2-(carboxymethylamino)propanoic acid, named (R)-strombine, as assessed by a panel of 47 subjects to confirm its contribution to the typical taste of scallop muscle. The second example discusses the pungency and trigeminal effect of polygodial, which is compared with piperine and capsaicin, as well as the elucidation of a new structure eliciting a trigeminal effect, (±)-trans-2,3,3a,7a-tetrahydro-1H-indene- 4-carbaldehyde, discovered in Amomum tsao-ko. Finally, the time intensity trigeminal effect of (–)-menthol is compared with (1R,2RS,4RS)-1-isopropyl-4-methylbicyclo[3.1.0]hexan-2-ol, named dihydroumbellulol, a new cooling compound obtained by hemi-synthesis from umbellulone extracted from Umbellularia californica Nutt. Keywords: (R)-2-(Carboxymethylamino)propanoic acid · Dihydroumbellulols · Flavor design · Polygodial · Trigeminal and taste properties

Introduction sweet and umami tastes, commonly associ- nal effects such as hot, pungent, cooling, ated with the taste of monosodium gluta- or mouthfeel sensations, their sensory as- When we eat food, our brain not only mate.[3–14] This new process complements pects are also assessed by a trained panel. integrates the visual, touch, smell, and, perfectly the ‘classical’ approach, success- The sensory characterization of these new to a smaller extent, noise attributes (for fully applied in our analytical laboratories compounds allows the comparison of their crunchy food), but also many other param- over the past 100 years, which consists in overall performance with benchmark mar- eters relayed by specific sensors located the careful and focused screening of in- ket products. The sensory aspects currently in the mouth such as taste and trigeminal teresting animal and plant species: slow- assessed comprise both quantitative mea- receptors.[1–9] In addition, changes in flavor throughput screening with a high success sures, such as detection thresholds, dose over time are induced by enzymatic and rate as opposed to high-throughput screen- response curves, or intensity profiles over bacterial processing during mastication.[10] ing with a low success rate. This high suc- time, and qualitative measures to define The ever-present innovation within the cess rate is possible only if extractions/pu- the most appropriate terms to describe the food industry means that there is a con- rifications are made on selected materials product. Different types of sensory studies stant need to investigate natural sources having known taste or trigeminal effects. that we use to characterize the performance for novel compounds displaying interest- Materials are selected on the basis of in- of the new compounds are presented with ing trigeminal or taste effects. How can we formation gathered from the botanical lit- each example in the following sections. discover these new molecules? Could the erature, ethnobotanists, personal curiosity, The aim of this paper is to outline some flavor industry follow the pharmaceutical or field work. This approach has been very of the results that we have obtained over industry’s example by heavily investing in successful for the discovery of sweet natu- the past few years in screening natural high-throughput screening?[11–13] ral compounds.[15,16] Screening for taste products to discover molecules and inte- The food and flavor industry has al- modifiers has relied on tasting of purified grating all parameters of chemosensation ready started to exploit this strategy, thanks extracts by human subjects. Although this to help our flavorists be innovative in their to the development by certain companies means very low throughput, the great ad- flavor design. This work was particularly of new cell-based assays with expressed vantage, compared with screening specific motivated by the current high demand in G protein-coupled taste receptors for both receptor interactions, is that we are able the marketplace to reduce salt, monoso- to assess the overall effect in the human dium glutamate (MSG), and/or sugar con- mouth and to avoid wasting any time with tained in highly processed foods. The holy false positives. grail of the food industry has been to find Following the isolation and character- compounds that help magnify taste and the ization of novel compounds that elicit an sensations that go with it.[2] interesting mouth effect, we undertake de- tailed sensory perception tests to evaluate the parameters of this effect. To do this, we Discovery of the Importance of require grams of material that we produce (R)-Strombine in the Taste Profile

*Correspondence: Dr. C. Starkenmann by relying either on large-scale isolation of Scallops Firmenich SA, Corporate R&D Division from the natural source if possible, or on P.O. Box 239 the unambiguous synthesis of the active The special taste of some seafood such CH-1211 Geneva 8 component. as mussels cannot be imitated by the usual Tel.: +41 22 780 3477 Fax: + 41 22 780 3334 Similarly, when we discover new com- food ingredients, MSG (glutamic acid (1) E-mail: [email protected] pounds having taste modulation or trigemi- at neutral pH), salt (NaCl, KCl), sugars, or fig 1

408 CHIMIA 2011, 65, No. 6 Nutrition Chemistry in Switzerland

ribotides.[17–19] The mouthfeel of seafood is quite different from chicken broth, for NH2 HO OH example, which is fattier, saltier, and more umami. We investigated the taste-active O 1 O compounds in dried scallops by bioassay Dehydrogenase H /Pd/C 5% guided fractionation with cation exchange O 2 HO OH HO OH HO OH chromatography, followed by liquid chro- N O + H2N + H2N H R O O O O matography and multiple tastings, with O O Natural Synthesis the mouth being the ultimate detector. One fraction was clearly described as sweet with 4 2 3 5 6 fullness in the mouth. After careful analysis, we detected the presence of glycine Fig. 1. Synthetic and biological formation pathways of (R)-strombine. (2), (S)-alanine, and (R)-strombine (3) in this interesting fraction. (R)-Strombine (3) is known to be present in mussel muscle as a result of anaerobic metabolism. It be- (R) - StStrombinerombine Glycine longs to the opines, a class of compounds s

se 30

resulting from the condensation of the most es 30 on ns sp o o e abundant amino acids in mussel, in this case r 25 of glycine (2) with pyruvic acid (4, Fig. 1), to resp 25 of 20 form a Schiff’s base, which is then reduced 20 [20,21] Number by opine dehydrogenase. Nothing was Number 15 known in the literature about the taste prop- 15 Fig103 10 erties of (R)-strombine. (R)-Strombine was BY CHANCE METALLIC prepared on a large scale via reductive ami- 5 MOUTHFEEL BY CHANCE SWEET 5 METALLIC BITTER MOUTHFEEL nation of glyoxylic acid (5) and (R)-alanine SWEET 0 ACID UMAMI BITTER [22] 0 ACID (6) to allow us to further evaluate its con- SALTY UMAMI tribution to dried scallop taste. This com- SALTY pound was evaluated by an in-house panel Concentration in (g/L) Concentration in (g/L) of volunteers in accordance with our con- sent guidelines. Sensory characterization of Fig. 2. Sensory attribute assessments of (R)-strombine and glycine, established in water solution, these new compounds allows us to compare by 47 subjects at different concentrations. the robustness of their performance with benchmark products already on the market. O OH To assess the sensory performance of O O O O 3, trained panelists received a 20 mL water O N N solution in a cup (blind test). They were H O HO HO asked to taste the entire sample in their 8 9 mouth. They spat out the sample after 7 O O O O 5 seconds and then answered questions O O H H about it. The sensory characterization was O N O combined with threshold measurements H (ASTM E679 Ascending Concentration O H H H Series Method of Limits), in one session 10 11 12 13 14 per product, with choices of sensory attributes. The panelists were asked to choose Fig. 3. Some natural molecules displaying trigeminal effects. among sweet, salty, bitter, umami, acid, mouthfeel, and metallic or ‘by chance’. Additional comments could also be added in the field ‘others’. from Pipper nigrum L.; capsaicin (8), the that permit the entry of Na+ and Ca++.[9–13] (R)-Strombine is perceived mainly as active component of chili peppers (genus These polymodal nociceptive receptor salty and umami by the majority of panel- Capsicum); and gingerol (9) from ginger, neurons are connected to cranial nerve V in ists. The multisensory aspect of this com- Zingiber officinale (Zingiberaceae)[24,25] the face, which allows us to perceive pain pound is very interesting, especially for (Fig. 3). Spilanthol (10) is a polyunsatu- effects in the mouth and in the nose. savory applications. The taste component rated amide commercially available from Laksa (Persicaria odorata (Lour.) of glycine is unimodal and clearly sweet an oleoresin prepared from Acmella olera- Sojak, a coriander plant commonly used from 1 g/L to 16 g/L. Unfortunately, the cea L. (family Asteraceae) that elicits a tin- in Thai cuisine, displayed a pungency that distribution threshold of (R)-strombine is gling sensation in the mouth. Spilanthol is was quite different than the pungency or centered at 0.5 g/L, which means that it is just one example of a family of polyunsatu- tingling of 7–10; it resembled the pungen- 10 times less potent than MSG[23] (Fig. 2). rated amide analogues that occur in nature cy of galangyl acetate (11) occurring in and display this interesting effect.[9,26,27] Alpinia galangal (family Zingiberaceae). These molecules activate the trigeminal In order to discover the active principle of Investigations of the Pungent receptors, which are quite different from laksa, which was not known in the litera- Effects of some Botanicals the taste receptors because they activate ture, the leaves were extracted with various responses in a subset of nociceptive fi- solvents such as water, ethyl acetate, and The most common commercially ac- bers by changing the activity of a fam- ethanol. Smelling strips were then dipped cessible pungent molecules are piperine (7) ily of temperature-sensitive ion channels into these solutions and, after the solvent Nutrition Chemistry in Switzerland CHIMIA 2011, 65, No. 6 409

14 14 14 Capsaicin 8 Panel Threshold Piperine 7 Panel Threshold Polygodial 12 Panel Threshold 12 12 0.06 ppm 12 5 ppm 2.35 ppm subjects

f 10 10 10 o 8 8 8

Number 6 6 6 4 4 4 2 2 2 0 0 0 0.0175 0.035 0.07 0.14 0.28 1.2 2.4 4.8 9.6 19.2 0.4 0.8 1.6 3.2 6.4 Concentration in ppm (mg/L)

Fig. 4. Distribution thresholds for the three molecules: capsaicin (8), piperine (7), and polygodial (12), established by 25 subjects. evaporated, the strips were placed on the separation (LC-tasting), we found that cooling effects or cooling effects without tongue. The active compound was clearly one fraction had an unusual pungent taste. the smell of mint.[42–48] perceived in the ethyl acetate extract. A After being isolated and characterized by Umbellulone (17) is the major mono- larger quantity of leaves was extracted with comprehensive spectroscopic methods, terpene constituent present in the leaves of ethyl acetate. After concentration of the so- the compound present in this fraction was Umbellularia californica Nutt. When the lution, the oleoresin was purified by chro- found to be either 13 or 14 (Fig. 3). As it leaves were crushed and smelled or tasted, matography on silica gel and all fractions was not possible to unambiguously assign a clear trigeminal effect was perceived. As tasted to localize the pungent molecule. the relative ring junction stereochemistry, a result of this interesting observation, the This led us to the discovery of the known we embarked upon the synthesis of both leaves were extracted and umbellulone iso- polygodial (12), a compound also present cis-14 and trans-13 isomers to confirm the lated as the major constituent. We wished in mountain pepper, Tasmannia lanceolata structure. The synthesis employed a Wit- to hydrogenate the C=C double bond of (Poir.), from the family Winteraceae.[28–30] tig reaction between (E)-(4-methoxy-4- umbellulone and reduce the ketone in or- Polygodial (12) is a good example of oxobut-2-en-1-yl)triphenylphosphonium der to compare the cooling effect of these the potential uses of pungent molecules in bromide and 1-cyclopentene-1-carbalde- molecules with (–)-menthol (15). Catalytic very different applications. It can be used, hyde. By varying the conditions used for hydrogenation (H2, 5% Pd/C) gave dihy- for example, in wasabi paste to enhance the this Wittig reaction, we could synthesize droumbellulone as a mixture of isomers in pungent effect of allyl isothiocyanate with- both cis and trans ring junction isomers 13 good yield. This mixture of dihydroumbel- out the associated rubbery olfactive note, and 14 from the same starting material.[38] lulone isomers was separated and further [31] to enhance the freshness of toothpaste, Finally, by comparison of analytical data, reduced with NaBH4/MeOH to give dihy- or even to improve the sensory acceptabil- we were able to determine that only the droumbellulol (18) and its isomers. With ity of artificial sweeteners.[32] less thermodynamically stable trans com- pure 18 in hand, we wished our sensory The trigeminal effect of polygodial pound 13, present in black cardamom, was panel to compare its cooling strength with (12), piperine (7), and capsaicin (8) was responsible for the trigeminal effect, de- (–)-menthol. determined by a panel of 25 subjects scribed as pungent, tingling, and refresh- The intensity over time is a very impor- (threshold measurements achieved with ing.[38] tant parameter, the objective of the sensory the three-alternative forced choice method evaluations being first to confirm the cool- based on the ASTM E679 Ascending Con- ing effect of the dihydroumbellulol (18), centration Series Method of Limits in one Hemi-synthesis of a New Cooling and then to compare its performance with session per product). According to aver- Compound: (1R,2R/S)-1-Isopropyl- that of (–)-menthol.[49] age threshold values (Fig. 4), polygodial 4-methylbicyclo[3.1.0]hex-3-en-2-ol Thirty trained panelists evaluated the is twice as strong as piperine, but 40 times perceived cooling intensity over time of weaker than capsaicin. The wider distribu- The monoterpene reported to elicit the the two samples, dihydroumbellulol and tion range for the perception of polygodial greatest cooling effect in the literature[39,40] (–)-menthol (one sample per session, at than for the other two compounds means is (–)-menthol (15), and cubebol (16) is 50 mg/L). Each subject received a sample that there are more differences between the only sesquiterpene reported to be cool- (30 mL) in a cup and was asked to taste it subjects for this compound. ing[41] (Fig. 5). Many synthetic compounds and to start a timer at the same time. Sub- During a taste trek in China to dis- fig 5with diverse chemical structures have been jects spat out the sample after 5 seconds. cover new interesting materials, we found synthesized to explore new longer lasting Panelists evaluated the time at which they the sensory properties of black cardamom started to perceive the cooling sensation

(Amomum tsao-ko Crevost et Lemarié) (Tbegin), the maximum perceived cooling (Zingiberaceae), widely used as a spice intensity and its corresponding time (Imax in China, to be refreshing because of the and Tmax), the time when the cooling sensa- presence of 1,8-cineol and other known tion started to decrease (Tdec), and, finally, OH constituents. No reported constituent, OH O OH the perceived cooling intensity 3 min after however, could explain the pleasant and they first tasted the sample (Iend) or the time [33–37] refreshing trigeminal effect. 15 16 17 18 when the cooling sensation disappeared, if Guided by the tasting of each frac- it was less than 3 min (Tend). Each evaluation obtained by liquid chromatography Fig. 5. Structure of some cooling compounds. tion was rated on a linear scale from not 410 CHIMIA 2011, 65, No. 6 Nutrition Chemistry in Switzerland

perceived (= 0) to very intense (= 10). Stu- dent t tests (two-tailed, paired) were per- 10 formed on each parameter (Tbegin, Imax, Tmax, 9 (-) Menthol 15 at 50 mg/L T , I , and T ) to compare data obtained dec end end intensity

d 8

from the two cooling compounds. The e Dihydroumbellulol 18 at 50 mgmg/L/L probabilities obtained for each of these Imax (***) ce iv 7 tests indicate whether the compounds have Per been perceived as significantly different or 6 not (significance was defined as p <0.05) for the parameter under consideration. 5 This sensory evaluation (Fig. 6) confirmed 4 the cooling effect of 18 as having a similar type of cooling profile over time compared 3 I 3min (NS) with (–)-menthol (15), but overall, it was 2 less powerful. The Imax values showed that dihydroumbellulol is weaker than (–)-men- 1 Tdec (NS) thol (significant difference for Imax) and its persistent trigeminal effect is shorter, with 0 [49] 0 50 100 150 a more pronounced delay. Tbegin (*) Time (s) Tmax (NS) Tend (***)

Conclusions Fig. 6. Perceived intensity of cooling over time assessed by 30 subjects (panel average and confidence interval at 5%). We have demonstrated that human tasting of purified compounds isolated from [6] C. S. Zuker, Nature 2002, 416, 27. poration Publication’, Tasmania, N° 99/124, natural sources is a key analytical step in [7] D. E. Clapham, Nature 2003, 426, 517. 1999. discovering molecules contributing to the [8] J. B. Calixto, C. A. L. Kassuya, E. Andre, J. [29] B. J. M. Jansen, A. De Root, Nat. Prod. Rep. taste of delicate foods such as seafood Ferreira, Pharmacol. Therapeutics 2005, 106, 2004, 21, 449. products. Naturally derived compounds 179. [30] C. Starkenmann, L. Luca, Y. Niclass, E. Praz, D. [9] A. Dhaka, V. Viswanath, A. Patapoutian, Annu. Roguet, J. Agric. Food Chem. 2006, 54, 3067. eliciting an interesting taste or mouthfeel Rev. Neurosci. 2006, 29, 135. [31] I. Hisashi, A. S. Suzuki, I. Shigeki, U.S. Patent effect have considerable potential as inno- [10] C. Starkenmann, B. Le Calvé, Y. Niclass, I. 5,523,105, 1996. vative flavors. However, with the large vol- Cayeux, S. Beccucci, M. Troccaz, J. Agric. [32] R. K. L. Kang, L. L. Zyzak, T. Nakasu, U.S. ume of cheap commodity compounds such Food Chem. 2008, 56, 975. Patent 5,948,460, 1999. [11] D. R. Appleton, A. D. Buss, M. S. Butler, [33] B. M. Lawrence, Phytochemistry 1970, 9, 665. as MSG being the market benchmark, the Chimia 2007, 61, 327. [34] ‘VCF Volatile Compounds in Food’: database, principal hurdle becomes the cost of using [12] O. Potterat, M. Hamburger, Curr. Org. Chem. Version 9.2., Ed. L. M. Nijssen, C. A. Ingen- these types of compounds. 2006, 10, 899. Visscher, J. J. H. van Donders, TNO Quality of In terms of actually launching commer- [13] J. Bitzer, B. Köpcke, M. Stadler, V. Hellwig, Y. Life, Zeist, 2007, 1963. cial products, more potential can be seen M. Ju, S. Seip, T. Henkel, Chimia 2007, 61, 332. [35] N. X. Dung, L. K. Bien, P. A. Leclercq, J. [14] X. Nelson, J. Chandrashekar, M. A. Hoon, Essent. Oil Res. 1992, 4, 91. in molecules producing trigeminal effects. L. Feng, G. Zhao, N. J.-P. Ryba, C. S. Zuker, [36] Z. Yi, Q. Qin, Z. Guoying, X. Zhonghua, L. Mouth refreshing products and saltiness Nature 2002, 416, 199. Tingli, Zhongcaoyao 2004, 35, 1225. and sweetness enhancers will always be [15] N. M. Kim, A. D. Kinghorn, Arch. Pharm. Res. [37] L. Zhuqiang, L. Lei, D. Wanghua, H. Rong, Z. molecules of interest. Our recent results 2002, 25, 725. Yuanqing, Yunnan Zhiwu Yanjiu 1998, 20, 119. [16] A. D. Kinghorn, D. D. Soejarto, Med. Res. [38] C. Starkenmann, F. Mayenzet, R. Brauchli, L. have clearly shown that it is still possible Reviews 1989, 9, 91. Wünsche, C. Vial, J. Agric. Food Chem. 2007, to discover new molecules eliciting tri- [17] K. Yoshii, in ‘Umami: A Basic Taste’, Ed. Y. 55, 10902. geminal effects in common spices, which Kawamura, M. R. Kare, Marcel Dekker Inc., [39] R. Eccles, J. Pharm. Pharmacol. 1994, 46, 618. motivates us to continue the search, despite New York, 1987, p. 219. [40] J. C. Leffingwell, http://www.leffingwell.com/ this having being the subject of intense re- [18] Y. Komata, Food Rev. Int. 1990, 457. cooler_than_menthol.htm (accessed May 28, [19] T. Hayashi, K. Yamaguchi, S. Konosu, J. Food 2010). search over many years. Sci. 1981, 46, 479. [41] M. I. Velazco, L. Wünsche, P. Deladoey, U.S. [20] J. Thompson, J. Donkersloot, Annu. Rev. Bio Patent 6,214,788 (prior. 31.03.1999), 1999; Acknowledgments chem. 1992, 61, 517. Chem. Abstr. 2000, 133, 265959. We thank: Mrs. Fabienne Mayenzet, Mr. [21] P. H. Yancey, M. E. Clark, S. C. Hand, R. D. [42] K. Bauer, J. Brüning, H. Grüb, German Patent Yvan Niclass, Dr. Christian Margot, Dr. Eric Bowlus, G. N. Somero, Science 1982, 217, DE 2608226A1, 1977. Frérot, Dr. Laurent Wünsche for synthesis and 1214. [43] E. Frerot, N. Van Beem, U.S. Patent scientific discussions, also Mr. Yves Lottaz, [22] C. Starkenmann, I. Cayeux, E. Decorzant, E. 6,359,168A1, 2002. Xiao-Hua Yang, Y. Niclass, L. Nicolas, J. Agric. [44] R. Pelzer, H. Sommer, R. Hopp, European and Mr. Eric Praz, flavorists for inspirations, Food Chem. 2009, 57, 7938. Patent EP 0583651A1, 1994. as also Mrs. Bénédicte Le Calve, Mrs. Nadine [23] L. J. V. Gemert, ‘Flavor Thresholds’, Oliemans [45] H. Grub, R. Pelzer, R. Hopp, R. Emberger, H.-J. Gaudreau and Mrs. Emma Yang for organizing Punters & Partners, Utrecht, 2003. Bertram, U.S. Patent 5,266,592, 1993. sensory panels. [24] A. Szallasi, P. M. Blumberg, Pharmacol. [46] A. Amano, M. Moroe, T. Yoshida, U.S. Patent Reviews 1999, 51, 159. 4,459,425, 1984. Received: April 1, 2011 [25] H. Rentmeister-Bryant, B. G. Green, Chemical [47] C. B. Green, T. Nakatsu, T Ishizaki, A. T. Lupo, Senses 1997, 22, 257. European Patent EP 1122233A1, 2001. [1] G. M. Shepherd, Nature 2006, 444, 316. [26] J. B. Calixto, C. A. L. Kassuya, E. André, J. [48] C. Fuganti, D. Joulain, F. Maggloni, L. Mal [2] M. Wenner, Scientific American 2008, August, Ferreira, Pharmacol. Therapeutics 2005, 106, pezzi, S. Serra, A. Vecchione, Tetrahedron: 96. 179. Asymmetry 2008, 19, 2425. [3] J. Chandrashekar, M. A. Hoon, N. J. P. Ryba, C. [27] D. M. Bautista, Y. M. Sigal, A. D. Milstein, J. [49] C. Starkenmann, I. Cayeux, R. Brauchli, F. S. Zuker, Nature 2006, 444, 288. L. Garrisob, J. A. Zorn, P. R. Tsuruda, R. A. Mayenzet, J. Agric. Food Chem. 2011, 59, 677. [4] X. Li, G. Servant, ACS Symp. Series. 2008, 23, Nicoll, D. Julius, Nature Neurosci. 2008, 11, 368. 772. [5] B. G. Green, Trends in Food Sci. Technol. 1996, [28] R. C. Menary, V. A. Dragar, S. M. Garland, 7, 415. ‘Rural Industries Research Development Cor