Unraveling the of sweet and

Grant E. DuBois* Corporate Innovation Division, The Coca-Cola Company, P.O. Box 1734, Atlanta, GA 30301

aste is an important part of ev- Class C GPCRs are unique in that they single orthosteric site. Xu et al. (8) dem- eryone’s daily life. Sweet possess very large N-terminal Venus onstrate that the human sweetener re- is particularly important as evi- flytrap-like domains (VFDs). In the ceptor contains a minimum of two denced by the fact that wars case of metabotrophic glutamate recep- orthosteric sites, and thus the Shallen- Thave been fought and people have been tor 1, it has been demonstrated that its ͞ berger Acree model and its improve- enslaved over , the prototypical VFD closes on binding glutamate just as ments, although perhaps correct for sweet stimulus (1). Savory taste, often expected (7). This precedent, and the some sweeteners, are not correct in the referred to as umami (delicious in Japa- fact that the sweetener and umami re- general . nese), is also important, its prototypical ceptors contain the common subunit stimulus being monosodium glutamate T1R3, leads to the expectation that What Do Sweeteners Bind to the (MSG). Despite the importance of these sweeteners likely bind in the VFD of Sweetener Receptor? taste sensations to our daily enjoyment T1R2 and glutamate likely binds in the Many more questions remain to be an- of life, until the late 1980s, the biochem- swered about the sweetener receptor. ical pathways that mediate them were Xu et al. (8) have demonstrated that largely unknown. Then, for sweet taste, The sweetener and bind to the evidence began to accumulate that it VFD of T1R2. Is this the orthosteric site must be by G protein-coupled receptor receptor is the first one for and other (GPCR)-mediated. More specifically, it sweeteners? Kniazeff et al. (12) have was generally believed to be mediated demonstrated to have demonstrated that both VFDs of the by several GPCRs because the findings homodimeric metabotrophic glutamate 5 of biochemical, electrophysiological, and more than one locus must be populated by glutamate to give psychophysical experiments could only of agonist binding. a maximal response. And for the het- be easily explained by a plurality of re- erodimeric ␥-aminobutyric acid ceptors (2, 3). And this expectation was (GABA) type B receptor, Knaizeff et al. supported by the fact that multiple sub- demonstrated that binding of GABA to types of GPCRs commonly exist for VFD of T R . The work of Xu et al. (8) 1 1 only the GABA type B1 receptor pro- other important signal (e.g., in this issue of PNAS probes the funda- vides a maximal response (13). If su- acetylcholine, norepinephrine, dopa- mental question of sweetener binding crose and other carbohydrate sweeteners mine, serotonin, etc.). Then, in the early locus with the finding that, although bind to the VFD of T R , do they also 2000s, a breakthrough occurred, dramat- some sweeteners do bind in the VFD of 1 2 bind to the VFD of T R as might be ically increasing the understanding of T R (i.e., aspartame and neotame), at 1 3 1 2 expected from glutamate binding in both sweet and umami tastes. Nelson et least one sweetener (i.e., cyclamate) al. (4) reported the discovery of the rat does not, but rather it binds within the metabotrophic glutamate 5, or, because sweetener receptor. In a functional as- seven-transmembrane domain (TMD) of the VFD of T1R3 is different from that of T1R2, do they bind only in T1R2?It say, they showed that all substances that T1R3. The sweetener receptor is the first rats generalize to sucrose taste are me- class C GPCR demonstrated to have has been demonstrated in psychophysical diated by a single receptor, which is a more than one locus of agonist binding experiments that carbohydrate sweeten- ers invariably give higher, and appar- heterodimer of two GPCRs, T1R2 and (orthosteric site). ently equivalent, maximal responses, T1R3. Further, they showed that umami taste is also mediated by a heterodimer What Is the Sweetener Pharmacophore? whereas other structural types of sweet- eners give lower, and variable, maximal of the two GPCRs T1R1 and T1R3. During the 1980s and before, the taste responses (14). Could this be because T1R1,T1R2, and T1R3 are members of literature was dominated by discoveries the small family of class C GPCRs. The of novel synthetic sweeteners and struc- the sweetener receptor binds two carbo- most studied members of the class C ture elucidations of natural sweeteners. hydrate sweetener molecules, one in GPCRs are the homodimeric metabo- By the 1980s, chemists had discovered at each VFD, thus leading to a high state trophic glutamate (mGluR), het- least 50 structural classes of sweet- of activation, while it responds to other erodimeric ␥-aminobutyric acid type B tasting organic compounds, and many structural types of sweeteners binding (GABABR), and homodimeric extracel- models had been developed to rational- only a single in T1R2 with a lular calcium receptors, which have re- ize their activities. These were pharma- lower state of activation? Another curi- cently been reviewed (5). Interestingly, cophore models, and the most publicized osity on the sweetener receptor relates the umami and sweetener receptors are of them was the so-called A-H͞B model to its enantioselectivity. I (unpublished 50% identical in that they share the of Shallenberger and Acree (9). They work) and others (15) have observed common subunit T1R3. This rat receptor hypothesized that all sweeteners contain that some enantiomeric pairs of simple discovery was quickly followed by a re- H-bond donor and H-bond acceptor (e.g., D- and L-) port of parallel findings on the human groups separated by not Ͻ2.5 or Ͼ4.0 Å. are identical in taste, whereas others are system by Li et al. (6). Again, it was sur- This model was later elaborated by Kier quite different (e.g., D- and L-). prising to see that the single human het- (10) and, more recently, Tinti and Nofre How can this be explained? erodimeric sweetener receptor, often (11) to rationalize the activities of ͞ written as T1R2 T1R3, responded to all sweeteners more potent than simple car- structural types of sweeteners tested and bohydrates. An assumption implicit in See companion article on page 14258. did so in a manner consistent with ex- all of these models is that sweetness is *E-mail: [email protected]. pectation from sensory experiments. mediated by a single receptor with a © 2004 by The National Academy of Sciences of the USA

13972–13973 ͉ PNAS ͉ September 28, 2004 ͉ vol. 101 ͉ no. 39 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0405991101 Downloaded by guest on October 2, 2021 COMMENTARY

What Is the Basis for Sweetness tagonist has been reported that is selec- A final curiosity about sweetener re- Synergy? tive in inhibition of sweetness of the ceptor function relates to findings in The biochemical basis for the phenome- various structural types of sweeteners: cross-adaptation psychophysical experi- ϩ non of sweetness synergy is also not Zn2 ion (17). In a study of 15 sweeten- ments. For example, if one adapts the known. It has long been known that as- ers from eight different structural sensory system to the sweetness of as- partame and cyclamate are synergistic in classes, Zn salts inhibited all of them, partame, sweetness perceived from D- sensory experiments (16). Xu et al. (8) except cyclamate. The observations on tryptophan is strongly suppressed, while at the same time, the sweetness of many have demonstrated that the sweetener sweetness antagonism now can be ratio- other sweeteners including nalized with the new information pro- receptor has separate orthosteric sites and cyclamate are unaffected (18). How for aspartame and cyclamate, and thus it vided by Xu et al. (8). It seems most can this be understood in terms of a seems reasonable that a cooperative plausible that , binding in the ͞ single sweetener receptor? binding effect may explain aspartame seven-TMD of T1R3, is a negative allo- cyclamate synergy. It is noteworthy that steric modulator and thus is able to What Is the Mechanism of Action for there are many other synergistic sweet- allosterically inhibit the activities of Enhancers of the Umami Receptor? ener combinations. Exemplary are sac- sweeteners that bind in the VFD and The work of Xu et al. (8) also provides ͞ charin and cyclamate. Does saccharin perhaps competitively inhibit cyclamate insight into the biochemistry of umami cyclamate synergism require different because it also binds in the T1R3 seven- taste. A longstanding mystery in the orthosteric sites? Saccharin and aspar- TMD. And it seems quite plausible that sensory field is the striking synergistic 2ϩ tame are also synergistic, however, and Zn ion may bind in the VFD of T1R2, effect of IMP on the taste of MSG. Xu if binding cooperativity is the mecha- thereby competitively inhibiting activi- et al. report that cyclamate, although nism mediating sweetness synergy, then ties of sweeteners acting at this orthos- exhibiting no activation of the umami the sweetener receptor must have at teric site, while having no effect on receptor by itself, enhances the activity least three orthosteric sites to explain of MSG. And because they have estab- cyclamate that acts in the T1R3 seven- the synergism observed with just these TMD. lished the locus of binding for cyclamate three sweeteners. Sweetness synergy has to be in the seven-TMD of T1R3 and been observed in other combinations of How Can Sweet Water Be because MSG’s orthosteric site is pre- sweeteners as well, which suggests that Explained? sumed to be in the VFD of T1R1,it seems that cyclamate may be a positive the number of orthosteric sites on the A curious observation for several sweet- sweetener receptor may be more than allosteric modulator of the umami re- ener receptor antagonists is that they ceptor. But still, questions remain. How three. Much more work must be done to † exhibit ‘‘sweet water aftertaste.’’ This does IMP enhance the activity of MSG? map the many known sweeteners to effect is particularly dramatic for lacti- their orthosteric sites. Another anomaly Does IMP also act in the seven-TMD of sole. If one tastes a solution of lactisole T R ? And, of course, is the VFD of is that, although significant synergism is 1 3 at a concentration known to give strong T R the orthosteric site for MSG? And common among pairs of noncarbohy- 1 1 sweetness inhibition, expectorates the finally, must MSG populate the VFDs drate sweeteners, synergism is weak or solution, and then rinses the mouth with of both T1R1 and T1R3 to provide a nonexistent in carbohydrate͞noncarbo- water, the water tastes strikingly sweet. maximal response? hydrate sweetener pairs (e.g., aspar- ͞ It is tempting to rationalize this phe- Great progress has been made over tame sucrose). Why should synergy be the last 2 years in understanding the limited to noncarbohydrate sweetener nomenon by a mechanism involving in- activation of a constitutively active biochemical pathways that mediate pairs? sweet and umami tastes. Xu et al. (8), sweetener receptor by an inverse ago- ͞ nist. Is this the operative mechanism? using chimeric rat human receptors, Are Selective Sweetener Antagonists have shed light on several important Known? questions. However, there is no shortage A number of antagonists of the sweet- †DuBois, G. E. & D’Angelo, L., XXIth Association of Chemo- of additional questions to be resolved ener receptor are known, of which the reception Sciences Meeting, April 16, 1999, Sarasota, FL, before we really understand sweet and most studied is lactisole. Only one an- poster. umami tastes.

1. Mintz, S. W. (1985) Sweetness and Power: The 7. Kunishima, N., Shimada, Y., Tsuji, Y., Sato, T., 13. Kniazeff, J., Galvez, T., Labesse, G. & Pin, J.-P. Place of Sugar in Modern History (Penquin, New Yamamoto, M., Kumasaka, T., Nakanishi, S., (2002) J. Neurosci. 22, 7352–7361. York). Jingami, H. & Morikawa, K. (2000) Nature 407, 14. DuBois, G. E., Walters, D. E., Schiffman, S. S., 2. Faurion, A. (1987) in Progress in Sensory Physiol- 971–977. Warwick, Z. S., Booth, B. J., Pecore, S. D., Gibes, ogy (Springer, Berlin), Vol. 8, pp. 129–201. 8. Xu, H., Staszewski, L., Tang, H., Adler, E., Zoller, K., Carr, B. T. & Brands, L. M. (1990) in Sweet- 3. DuBois, G. E. (1997) in Olfaction and Taste: M. & Li, X. (2004) Proc. Natl. Acad. Sci. USA 101, eners: Discovery, Molecular Design, and Chemore- A Century for the , ed. Salvadori, G. 14258–14263. ception, ACS Symposium Series 450, eds. Walters, 9. Shallenberger, R. S. & Acree, T. E. (1967) Nature (Allured Publishing, Carol Stream, IL), pp. D. E., Orthoefer, F. T. & DuBois, G. E. (Am. 216, 480–482. Chem. Soc., Washington, DC), pp. 261–276. 32–95. 10. Kier, L. B. (1972) J. Pharm. Sci. 61, 1394–1396. 15. Shallenberger, R. S., Acree, T. E. & Lee, C. Y. 4. Nelson, G., Hoon, M. A., Chandrashekar, J., 11. Tinti, J.-M. & Nofre, C. (1991) in Sweeteners: (1969) Nature 221, 555–556. Zhang, Y., Ryba, N. J. P. & Zuker, C. S. (2001) Discovery, Molecular Design, and Chemoreception, 16. Schiffman, S. S., Booth, B. J., Carr, B. T., Losee, Cell 106, 381–390. ACS Symposium Series 450, eds. Walters, D. E., M. L., Sattely-Miller, E. A. & Graham, B. G. 5. Pin, J.-P., Galvez, T. & Prezeau, L. (2003) Phar- Orthoefer, F. T. & DuBois, G. E. (Am. Chem. (1995) Brain Res. Bull. 38, 105–120. macol. Ther. 98, 325–354. Soc., Washington, DC), pp. 206–213. 17. Keast, R. S. J., Canty, T. M. & Breslin, P. A. S. 6. Li, X., Staszewski, L., Xu, H., Durick, K., Zoller, 12. Kniazeff, J., Bessis, A.-S., Maurel, D., Ansanay, (2004) Chem. Senses 29, 513–521. M. & Adler, E. (2002) Proc. Natl. Acad. Sci. USA H., Prezeau, L. & Pin, J.-P. (2004) Nat. Struct. Mol. 18. Schiffman, S. S., Cahn, H. & Lindley, M. G. (1981) 99, 4692–4696. Biol. 11, 706–713. Pharmacol. Biochem. Behav. 15, 377–388.

DuBois PNAS ͉ September 28, 2004 ͉ vol. 101 ͉ no. 39 ͉ 13973 Downloaded by guest on October 2, 2021