GK1.5 treatment (Table 3). Although nearly all thymocytes express L3T4 (5), the percentage of L3T4+ thymocytes was not significantly changed in treated mice (8). The mechanism of the therapeutic ac- tions of GKI .5 administration is still unknown. The observed decrease in the numbers of peripheral L3T4+ helper (in- ducer) or effector T cells may be suffi- cient to account for the amelioration of the disease. Alternatively, the ability of the antibody to bind L3T4 and block the functional activities of this molecule could also play a role. In vivo adminis- tration of monoclonal antibodies to the I- A subregion of the major histocompati- bility complex can prevent and reverse EAE (9). It is possible that antibody to L3T4 and antibodies to I-A effect EAE because both antibodies block T-cell ac- tivation, either at the site of the T-cell receptor for I-A [in the case of treatment with antibody to L3T4 (5, 10)] or at the site of I-A on antigen-presenting cells (in the case of treatment with antibody to I- A). In addition, if there is a reduction in the number of L3T4+ cells or if their activation is blocked (or both), there may be a shift in the balance between regulatory helper and suppressor T cells to an alternate stable state (11) in which suppression becomes the dominant re- sponse to the encephalitogenic compo- nents of MSCH. Regardless of the mechanism of action of GKI.5 treatment, we have shown that in vivo administration of GKI.5 has a dramatic effect on the course of EAE. The successful treatment of murine EAE with GKI.5 suggests that therapy with monoclonal antibodies to the human equivalent of the murine L3T4+ subset might prove effective in the treatment of MS and possibly in other diseases where this T-cell subset plays a central role. MATTHEW K. WALDOR* Departments of Genetics and Neurology, Stanford University, Stanford, California 94305 SUBRAMANIAN SRIRAM Department of Neurology, Stanford University RICHARD HARDY LEONORE A. HERZENBERG LEONARD A. HERZENBERG Department of Genetics, Stanford University LEWIS LANIER Becton Dickinson Monoclonal Antibody Center, Mountain View, California 94043 MAE LIM LAWRENCE STEINMAN Departments of Neurology and Pediatrics, Stanford University 25 JANUARY 1985 References and Notes 1. E. C. Alvord, M. W. Kies, A. J. Suckling, Experimental Allergic Encephalomyelitis, A Useful Modelfor Multiple Sclerosis (Liss, New York, 1984); P. Y. Paterson, in Autoimmunity, N. Talal, Ed. (Academic Press, New York, 1977), pp. 644-692. 2. B. G. Arnason et al., J. Exp. Med. 116, 177 (1962); N. K. Gonatas and J. C. Howard, Sci- ence 186, 839 (1974); L. Ortiz-Ortiz and W. 0. Weigle, J. Exp. Med. 144, 604 (1976). 3. U. Traugott, E. L. Reinherz, C. S. Raine, Science 219, 308 (1982). 4. C. B. Pettinelli and D. E. McFarlin, J. Immunol. 127, 1420 (1981); Z. Lando and A. Ben-Nun, Clin. Immunol. Immunopathol. 30, 290 (1984); J. Trotter and L. Steinman, J. Immunol., in press. 5. D. P. Dialynas et al., Immunol. Rev. 74, 29 (1983). 6. E. L. Reinherz and S. F. Schlossman, N. Engl. J. Med. 303, 370 (1980); E. L. Reinherz et al., ibid., p. 125. Because of such factors as the caloric and cariogenic potential of sucrose, al- ternative means are being sought to sat- isfy the human habit of consuming sweets. Synthetic substances such as saccharin, cyclamate, and aspartame have been or are being used in the Unit- ed States for this purpose, while natural- ly occurring substances such as phyllo- dulcin, stevioside and glycyrrhizin are used in Japan (1, 2). None of these substances is ideal, however, either be- cause of taste characteristics that are generally perceived as unpleasant, ques- tionable safety, chemical instability, or because of the relatively high cost of production (1, 3). The discovery of practically every in- tensely sweet, prototype molecule has been fortuitous; there is still insufficient knowledge to design sweet compounds that are structurally unrelated to existing sweeteners (3). In addition, large-scale screening of potential sweeteners is im- practical because of the lack of conve- nient bioassay systems (4). Thus, as part of a program to develop new sweetening agents and to provide compounds that could aid in understanding the relation between moleclular structure and sweet- ness, the Mexican ethnobotanical litera- ture was searched with the express pur- pose of uncovering records of intensely sweet-tasting plants. While examining a monograph entitled Natural History of New Spain, written 7. C. Bernard and I. R. MacKay, J. Neuroim- munol. 4, 61 (1983). 8. M. K. Waldor et al., unpublished, data. 9. L. Steinman et al., Proc. Natl. Acad. Sci. U.S.A. 78, 7111 (1981); S. Sriram and L. Stein- man, J. Exp. Med. 158, 1362 (1983). 10. P. Marrack et al., J. Exp. Med. 158, 1077 (1983). 11. L. A. Herzenberg, S. J. Black, L. A. Herzen- berg, Eur. J. Immunol. 10, 1 (1980). 12. J. A. Ledbetter and L. A. Herzenberg, Immu- nol. Rev. 47, 63 (1979). 13. K. Hayakawa, R. R. Hardy, D. R. Parks, L. A. Herzenberg, J. Exp. Med. 157, 202 (1983). 14. D. R. Parks, R. R. Hardy, L. A. Herzenberg, Immwnol. Today 4, 145 (1983). 15. Supported by NIH grants NS-571, NS-18235, GM-17367, HD-01287, and CA-04681; a grant from the Multiple Sclerosis Society 1440-A-i; and a grant from the Kroc Foundation. * To whom correspondence should be addressed at Department of Neurology, C-338, Stanford University, Stanford, Calif. 94305. 27 August 1984; accepted 5 November 1984 between 1570 and 1576 by the Spanish physician Francisco Hernandez (5), our attention was drawn to a remarkably sweet plant known to the Aztec people by the Nahuatl name Tzonpelic xihuitl. The literal translation of these words is "sweet herb." The accurate description and illustration of the plant provided by Hernandez combined with another men- tion (6), enabled the assignment of Tzon- pelic xihuitl as Lippia dulcis Trev. (Ver- benaceae). The constituents of this plant do not appear to have been studied in recent years, but investigators have sug- gested that the sweet principle is volatile (7) and distinguishable from the sweet glycoside glycyrrhizin (8). The principal sweet component of L. dulcis (9), which was present mainly in the leaves and flowers, was isolated in pure form after solvent partition and chromatographic fractionation. This compound, 1, which has been named 0 OH 7 H a, 261 2' 4~ 6' 7 hernandulcin in honor of Francisco Her- nandez, was obtained as a colorless oil [a]" +1090 (c, 0.11 g/l00ml in ethanol). Its molecular formula is C,5H2402, based on a high-resolution mass spectral mea- surement of the molecular ion, 236.18005 amu. Analysis of the 'H- and "3C-NMR 417 Hernandulcin: An Intensely Sweet Compound Discovered by Review of Ancient Literature Abstract. Ancient Mexican botanical literature was systematically searched for new plant sources of intensely sweet substances. Lippia dulcis Trev., a sweet plant, emerged as a candidateforfractionation studies, and hernandulcin, a sesquiterpene, was isolated and judged by a human taste panel as more than 1000 times sweeter than sucrose. The structure of the sesquiterpene was determined spectroscopically and confirmed by chemical synthesis. Hernandulcin was nontoxic when adminis- tered orally to mice, and it did not induce bacterial mutation.