Interaction of gymnemic acid with cyclodextrins analyzed by isothermal titration calorimetry, NMR and dynamic light scattering Yusuke Izutani 1 , Kenji Kanaori 2 , Toshiaki Imoto 3 and Masayuki Oda 1 1 Graduate School of Agriculture, Kyoto Prefectural University, Japan 2 Department of Applied Biology, Kyoto Institute of Technology, Japan 3 Faculty of Medicine, Tottori University, Yonago, Japan Gymnemic acid (GA), a saponin of triterpene glyco- side is contained in leaves of Gymnema sylvestre, which is native to India, and has various physiological effects such as an antisweet taste, inhibition of intestinal glu- cose absorption, and lowering of plasma glucose and insulin levels [1–6]. As shown in Fig. 1, GA is not a pure entity, but is composed of several types of homo- logues [7]. Regarding the antisweet effect in humans, when around 1 mm of partially purified GA in water is tasted beforehand, the ability to taste anything sweet is abolished for 30–60 min [2]. Although it is not clear how GA acts as an inhibitor, it is considered that GA binds to the sweet taste receptor, similar to another taste antagonist, lactisole [8]. It should also be noted that strogin, whose structure resembles that of GA, has sweet and sweetness-inducing activity, and seems to bind to the same site on the receptor [9]. The sweet- taste receptor was recently identified to be a hetero- meric dimer of G-protein-coupled receptors (T1R2 and T1R3) expressed in subsets of taste receptor cells on the tongue and palate [10,11]. Interestingly, the antisweet taste effect of GA has been known to be immediately diminished by rinsing the tongue with c-cyclodextrin (c-CD) solution after GA has been held in the mouth. The restorative effect of c-CD on the suppressed sweet taste by the extract of G. sylvestre was first reported by Nagaoka et al. [12] and the same effect has been observed in case of GA [13]. Similarly, the sweet-inducing effect of strogin is also diminished by application of c-CD [9]. Keywords aggregation; cyclodextrin; gymnemic acid; molecular interaction; thermodynamics Correspondence M. Oda, Graduate School of Agriculture, Kyoto Prefectural University, 1-5, Hangi-cho, Shimogamo, Sakyo-ku, Kyoto 606-8522, Japan Fax: +81 75 703 5673 Tel: +81 75 703 5673 E-mail: oda@kpu.ac.jp (Received 30 August 2005, revised 6 October 2005, accepted 11 October 2005) doi:10.1111/j.1742-4658.2005.05014.x The physiological phenomenon that the antisweet taste effect of gymnemic acid (GA) is diminished by application of c-cyclodextrin (c-CD) to the mouth was evaluated at the molecular level using isothermal titration calorimetry, NMR and dynamic light scattering. These analyses showed that GA specifically binds to c-CD. Thermodynamic analysis using isother- mal titration calorimetry revealed that the association constant of GA and c-CD is 10 5 )10 6 m )1 with favorable enthalpy and entropy changes. The heat capacity change was negative and large, despite the change in access- ible surface area upon binding being small. These thermodynamics indicate that the binding is dominated by hydrophobic interactions, which is in agreement with inclusion complex formation of c-CD. In addition, NMR measurements showed that in solution the spectra of GA are broad and sharpened by the addition of c-CD, indicating that unbound GA is in a water-soluble aggregate that is dispersed when it forms a complex with c-CD. Dynamic light scattering showed that the average diameter of unbound GA is > 30 nm and that of GA and c-CD complex is 2.2 nm, similar to unbound c-CD, supporting the aggregate property of GA and the inclusion complexation of GA by c-CD. Abbreviations CD, cyclodextrin; DLS, dynamic light scattering; GA, gymnemic acid; ITC, isothermal titration calorimetry. 6154 FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS Cyclodextrins (CDs) are cyclic oligosaccharides com- posed of a-(1fi 4)-linked a-d-glucosyl unit, in which the most common and studied CDs are a-, b- and c-CDs, consisting of 6-, 7- and 8-glucosyl units, respectively [14]. CDs can be described as toroidal, hollow, trun- cated cones with a hydrophilic surface and a hydropho- bic pocket, which forms an inclusion complex with an organic compound, known as a host–guest interaction [15,16]. The hydrophobic pocket diameters of a-, b- and c-CDs are 4.7–5.3, 6.0–6.5 and 7.5–8.3 A ˚ , respectively [14]. The unique properties of CDs are utilized in many applications such as the pharmaceutical, food and chemical industries: solubility enhancement, stabiliza- tion of labile drugs, control of volatility and sublimat- ion, physical isolation of incompatible compounds, long-term protection of color, odor and flavor, and sup- pression of hemolysis and the bitter tastes of drugs [6,14,17]. Taking into account the size of the hydropho- bic pockets of CDs and the chemical structure of GA (Fig. 1), it can be speculated that c-CD forms an inclu- sion complex with GA, diminishing the antisweet effect of GA, although there has been no information about their specific molecular interaction to date. In this study, we analyzed the interactions between GA and CDs using isothermal titration calorimetry (ITC), NMR and dynamic light scattering (DLS). ITC measurements provide thermodynamic parameters, not only the binding affinity (K a ), but also the enthalpy change (DH) and entropy change (DS) [18]. The bind- ing experiments of GA to a-CD, b-CD and c-CD showed that GA specifically interacts with c-CD. To explore the recognition mechanism between GA and c–CD, the interaction was further analyzed under dif- ferent conditions of pH, buffer and temperature. In addition, NMR and DLS analyses demonstrated the characteristic properties of GA. Our results of the molecular interaction between GA and c-CD can be correlated with the physiological phenomenon of sweet taste modification. Results ITC analysis Figure 2 shows typical ITC profiles at 25 °C for the interactions between GA and CDs. Exothermic heat was gradually decreased after each injection of GA into c-CD, whereas only heat of dilution was observed after each injection of GA into the experimental buffer and a-CD (Fig. 2A–C). When GA was injected into b-CD, small exothermic and endothermic heats were gradually titrated (Fig. 2D). Although it is difficult to determine thermodynamics for the interaction between GA and b-CD, owing to the small heats, the binding affinity of b-CD is much lower than that of c-CD des- cribed below. These results clearly indicate that GA specifically binds to c-CD. Assuming formation of the inclusion complex for these interactions, it could be deduced that the cavity size of c-CD is suitable for GA binding, but those of a-CD and b-CD are too small. In order to determine the binding site of GA, binding of glucuronic acid to c-CD was also analyzed, although that of gymnemagenin could not be examined due to its low solubility. Because only heat of dilution was observed for the interaction between glucuronic acid and c-CD (data not shown), the aglycone portion of GA should penetrate into the hydrophobic pocket of c-CD. To determine the thermodynamic parameters, the area of each exothermic peak as observed in Fig. 2B was integrated, and the heat of dilution was subtrac- ted from the integrated values. The corrected heat was divided by the number of moles of GA injected, which were calibrated in consideration with the pur- ity of GA, and the resulting values were plotted as a function of the molar ratio. The resultant data were best-fit according to a model for one binding site, using a nonlinear least-squares method (Fig. 2E). The thermodynamic parameters at 25 °C are summarized in Table 1. Similar thermodynamic parameters at dif- ferent pH values between 4.5 and 9.5 indicate that the chemical structures of both GA and c–CD and their interactions are little perturbed in this pH range. The similar DH values at pH 7.4 in Tris ⁄ HCl and P i indicate that there is little effect of buffer ionization on complex formation [19]. The binding stoichiometry of GA to c-CD shows that the com- plex forms in an equimolar ratio of respective mole- cules. ITC measurements of the interaction of GA and c-CD were further performed at four different temper- atures, ranging from 20 to 35 °C (Table 2). Through- out the temperature range analyzed, binding is Fig. 1. Chemical structure of GA. Y. Izutani et al. Interaction of gymnemic acid with cyclodextrins FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS 6155 accompanied by favorable changes in enthalpy and entropy, and shows strong temperature dependence for both DH and TDS, which compensate each other to make the Gibbs free energy change (DG) almost insen- sitive to temperature. The temperature dependence of DH yields a heat capacity change (DC p ) ¼ )0.14 kcalÆ mol )1 Æ°C )1 , assuming that DC p is constant within the experimental temperatures. NMR analysis NMR methods were applied for the analysis of inter- actions between GA and CDs. Line widths of 1 H NMR signals of GA alone were much broader in D 2 O than those reported in pyridine-d 5 containing a few drops of D 2 O [20], and those broad signals were unchanged by the addition of a- and b-CDs Fig. 2. Typical ITC profiles of the GA binding to CDs. A 2.2 m M solution of GA was injec- ted 16 times in increments of 10 lL into the experimental buffer (A), and a 0.1 m M solu- tion of c-CD (B), a-CD (C), and b-CD (D) at 25 °C. Titrations were performed over 10 s at intervals of 180 s. All samples were in 50 m M Tris ⁄ HCl buffer (pH 7.4). (E) The data points were obtained by integration of the peaks in (B), corrected for the dilution heat (A), and plotted against the molar ratio (GA ⁄ c-CD). The data were fitted using a nonlinear least-squares method. Table 1. Thermodynamic parameters of the interaction between GA and c-CD at 25 °C. pH n a K a (· 10 5 M )1 ) DG (kcalÆmol )1 ) DH b (kcalÆmol )1 ) TDS (kcalÆmol )1 ) GA injection into c-CD 4.5 c 1.01 3.3 ± 0.4 )7.5 )4.5 3.0 5.5 c 1.01 4.5 ± 0.5 )7.7 )4.5 3.2 7.4 d 1.02 5.5 ± 0.5 )7.8 )4.4 3.4 7.4 e 1.02 5.5 ± 0.5 )7.8 )4.2 3.6 9.5 f 1.03 5.2 ± 0.5 )7.8 )4.7 3.1 c-CD injection into GA 7.4 d 1.01 4.3 ± 0.5 )7.7 )4.5 3.2 a The n-value represents binding stoichiometry of GA to c-CD. The fitting error was < 1%. b The fitting error was < 2%. c In 50 mM sodium acetate buffer, d in 50 mM Tris ⁄ HCl buffer, e in 50 mM phosphate buffer, f in 50 mM glycine ⁄ NaOH buffer. Interaction of gymnemic acid with cyclodextrins Y. Izutani et al. 6156 FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS (Fig. 3A). Because no precipitation was observed in these NMR samples, GA, at least above the concen- tration of 2.0 mm, would form a water-soluble aggre- gate. In the presence of c-CD, however, the GA signals became sharp (Fig. 3A). These results indicate that the aggregated GA in the aqueous solution is dispersed by the addition of c-CD but not a-or b-CDs. The higher dispersion effect of c-CD is poss- ibly related to its high affinity for GA, as shown in the ITC experiments. Thus, the concentration depend- ence of c-CD on the spectral change of GA was examined. As the concentration of c-CD was increased the line width of GA became sharper, and above the equimolar c-CD to GA, the spectra were unchanged (Fig. 3B). This is in accordance with the ITC results that c-CD and GA form a 1 : 1 complex. Sharp singlet signals were observed around 1 p.p.m. in the complex of GA and c-CD, probably origin- ating from the methyl groups of the genin moiety [20]. Taking the thermodynamic parameters into con- sideration, the NMR results indicate the specific bind- ing of GA to the pocket of c-CD occurs in the aqueous solution, accompanying dispersion of the self-association of GA. DLS analysis In order to further analyze the aggregated property of GA and the effects of c-CD, size distribution of GA in the absence or presence of c-CD in H 2 O was analyzed by DLS (Fig. 4). Two distributions were observed for a 3.1 mm solution of GA, in which the average radii are 37.1 ± 3.5 and 125.2 ± 34.0 nm, respectively (Fig. 4A). These values are much larger than predicted from the chemical structure of GA, indicating that GA in solution is in the form of a water-soluble aggregate, which is in accordance with the NMR results. The addition of c-CD changed the distribution to much smaller size, 2.2 ± 0.4 nm, which is similar to the size of c-CD itself (Fig. 4B,C), supporting that the aggregate of GA is dispersed when it forms an inclusion complex with c-CD. Table 2. Thermodynamic parameters of the interaction between GA and c-CD at pH 7.4. Temperature (°C) n a K a (· 10 5 M )1 ) DG (kcalÆmol )1 ) DH b (kcalÆmol )1 ) TDS (kcalÆmol )1 ) 20.2 1.03 5.7 ± 0.6 )7.7 )3.7 4.0 25.1 1.02 5.5 ± 0.5 )7.8 )4.4 3.4 30.1 1.01 5.1 ± 0.5 )7.9 )5.3 2.6 35.1 1.03 5.1 ± 0.4 )8.1 )5.8 2.3 All measurements were performed for GA injection into c-CD in 50 m M Tris ⁄ HCl buffer (pH 7.4). a The n-value represents binding stoichiom- etry of GA to c-CD. The fitting error was < 1%. b The fitting error was < 1%. Fig. 3. 1 H NMR spectra (500 MHz) of GA in the absence or pres- ence of CDs. (A) NMR spectra of 2.0 m M GA in the presence of 2.0 m M a-CD, b-CD, and c-CD. (B) The concentration dependence of c-CD on NMR spectra of 2.0 m M GA. The ratios of c-CD to GA are indicated. Y. Izutani et al. Interaction of gymnemic acid with cyclodextrins FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS 6157 Discussion We analyzed the interaction between GA and CDs at a molecular level using ITC, NMR and DLS. These analyses showed that GA specifically binds to c-CD. This is in good agreement with the physiological phe- nomenon in humans, in which the antisweet effect of GA is diminished by the addition of c-CD to the ton- gue. The binding stoichiometry determined by ITC and NMR revealed that GA forms a complex with c-CD in the ratio of 1 : 1. The DLS results that the molecular size of c-CD and GA complex is similar to that of unbound c-CD indicate an inclusion complexa- tion, that is, GA would fit into the hydrophobic pocket of c-CD. Considering that the GA used in this study is a mixture of homologues [7], each homologue of GA would interact with c-CD in similar manner. This is also supported by the NMR results that the broad GA signals were sharpened by the addition of c-CD. Presumably because the small pockets of a-CD and b-CD are not able to form stable complexes with GA, the binding of a-CD to GA is not observed and that of b-CD is much weaker than that of c-CD. The difference in binding affinity toward guest molecules depending on the size of CDs has also been reported for other host–guest interactions [21]. The binding affinities of GA to c-CD were shown to be $ 10 5 )10 6 m )1 under physiological conditions. The binding strength may explain the physiological phe- nomenon that the sweet-suppressing activity of GA is immediately diminished by application of 5 mgÆmL )1 (3.9 mm) c-CD to the mouth after 3 mgÆmL )1 (3.7 mm) GA has been held in the mouth [13]. The specific binding of c-CD would cause dissociation of GA from the sweet taste receptor, resulting in recovery of the sweet taste. Although neither the concentration of sweet taste receptors on the tongue nor the binding affinity of GA to the receptor has been determined, the need for c-CD in the mm range would correlate with the binding affinity between GA and the receptor. The thermodynamic parameters obtained are in the range of those of other interactions with c-CD [15]. The interaction of GA with c-CD is accompanied by favorable DS values together with favorable DH values (Tables 1 and 2). Because a large decrease in configu- rational entropy has been reported for CD complexa- tion, a dehydration effect upon binding would contribute to the favorable DS value observed with the interaction of GA and c-CD [21,22]. In addition, the DC p value, )0.14 kcalÆmol )1 Æ°C )1 , is large and negat- ive, which is in the largest range among those of other interactions with CDs [15]. Because negative values for DC p are believed to arise from hydrophobic interac- tions, which release the structured water surrounding the nonpolar groups on the surface of uncomplexed protein [23], the GA and c-CD association would be dominated by hydrophobic interactions, which contrib- ute to the favorable DS observed in this association. The DC p value, )0.14 kcalÆmol )1 Æ°C )1 , would be larger than predicted by the correlation between DC p and accessible surface area upon binding [24,25], suggesting that other effects such as salt dehydration might con- tribute to the DC p value determined for the interaction of GA and c-CD [22,26]. Analyses using NMR and DLS showed the aggrega- ted property of GA. Aggregation should be due to the hydrophobic property of GA, which is also the main driving force for the interaction with c-CD as described above. For the durability of antisweet effect of GA on a human tongue, this water-soluble aggregation might be important for its function of sweet antagonism. It is hypothesized that the GA molecules in the aggregated form can simultaneously bind to several sweet taste receptors. This may increase the durability of antisweet effect, particular because of the slow dissociation rate, which is also seen in antigen–antibody interactions as the avidity effect [27]. The dispersion of GA aggregate by c-CD may help to dissociate from the receptor. In order to elucidate the complex of GA and c-CD in detail, NMR assignment is under way [28]. Several Fig. 4. Particle size distribution of GA in the absence or presence of c-CD. (A) Relative frequency of 3.1 mM GA. (B) Relative frequency of 39 m M c-CD. DLS of c-CD at low concentration such as 3.1 mM was difficult to detect. (C) Relative frequency of 3.1 mM GA in the presence of equimolar concentration of c-CD. Relative frequency of molecule number was shown against the logarithm of molecule diameter. Interaction of gymnemic acid with cyclodextrins Y. Izutani et al. 6158 FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS intra- and intermolecular NOE peaks were observed for the complex (unpublished results). Together with the NMR results, an ITC study indicated that the aglycone moiety of GA penetrates into the c-CD pocket. When the complex structure has been deter- mined at atomic resolution, it will be interesting to analyze the contribution of each interaction, such as hydrogen bond and hydrophobic contacts and hydra- tion upon binding on thermodynamics and its correla- tion with structure. These analyses may help us to generally understand the molecular recognition mech- anism of GA, and to design rationally new materials to control the sense of taste. Experimental procedures Materials Thirty percent ethanol extract of G. sylvestre was kindly provided by Dai-Nippon Meiji Sugar Co., Ltd (Tokyo, Japan). GA was purified from the extract as described pre- viously [29]. The GA sample obtained was a mixture of homologues and its purity was $ 70%. Because very intri- cate and tedious steps are required to purify each homo- logue, the mixture of homologues was used as GA in this study, similar to most of other investigations that analyze the functions of GA. Gymnemagenin and Diaion HP20 were purchased from Maruzen Pharmaceuticals Co., Ltd. (Onomichi, Japan) and Mitsubishi Chemical Co., Ltd. (Tokyo, Japan), respectively. All other reagents were pur- chased from Nacalai Tesque, Inc. (Kyoto, Japan). ITC measurements MCS-ITC (Microcal, Northampton, USA) was used for thermodynamic analysis of the interaction between GA and CDs. All samples dissolved in buffers were filtrated through a 0.45 lm filter and degassed before the titrations, using the equipment provided with the instrument. GA or CD solution (2.2 mm) was titrated into CD or GA solution (0.1 mm) using a 250-lL syringe. Each titration consisted of an initial injection (2.5 lL) followed by 15 main injec- tions (10 lL). Measurement data were analyzed by microcal origin version 2.9. The resultant data was best-fit, according to a model for one binding site using a nonlinear least-squares method. The binding stoichiometry (n), K a and DH, were obtained from the fitted curve. The values of DG and DS were calculated from the equation, DG ¼ÀRT lnK a ¼ DH À T DS ð1Þ where R is the gas constant, and T is absolute temperature. The heat capacity change, DC p , was calculated from the linear fitting to the D H values measured at various temperatures, assuming that DC p is constant within the experimental temperatures, DC p ¼ðDH=DTÞð2Þ NMR measurements For the acquisition of NMR spectra, GA was dissolved in D 2 O at a concentration of 2.0 mm. The pH of the solution was 4.6 (meter reading of glass electrode without correction to pD) where the binding manner between GA and c-CD is identical to that at neutral pH. 1 H NMR spectra were measured on a Bruker ARX-500 at 30 °C. 1 H chemical shifts were referred to internal sodium 3-(trimethyl- silyl)propionate-2,2,3,3-d 4 . DLS measurements DLS-7000 (Otsuka Electronics Co., Ltd) was used to meas- ure DLS to estimate the diameters of GA, c-CD, and their mixture in H 2 Oat25°C. The sample was filtrated through a 0.2 lm filter. The measurement was performed using a laser beam of 488 nm at an angle of 90°. The molecule number of diameter distribution was obtained by the histo- gram analysis. Acknowledgements The authors thank Mr Shoichi Nakamura of Otsuka Electronics Co., Ltd. for support of DLS measurements. References 1 Warren RM & Pfaffmann C (1959) Suppression of sweet sensitivity by potassium gymnemate. J Appl Phy- siol 14, 40–42. 2 Kurihara Y (1969) Antisweet activity of gymnemic acid A1 and its derivatives. Life Sci 8, 537–543. 3 Yoshioka S (1986) Inhibitory effects of gymnemic acid and an extract from the leaves of Zyzyphus jujuba on glucose absorption in the rat small intestine. J Yonago Med Assoc 37, 142–145 (in Japanese). 4 Hirata S, Abe T & Imoto T (1992) Effects of crude gym- nemic acid on the oral glucose tolerance test in human beings. J Yonago Med Assoc 43, 392–396 (in Japanese). 5 Nakamura Y, Tsumura Y, Tonogai Y & Shibata T (1999) Fecal steroid excretion is increased in rats by oral administration of gymnemic acids contained in Gym- nema sylvestre leaves. J Nutr 129, 1214–1222. 6 Porchezhian E & Dobriyal RM (2003) An overview on the advances of Gymnema sylvestre: chemistry, pharma- cology and patents. Pharmazie 58, 5–12. 7 Suttisri R, Lee IS & Kinghorn AD (1995) Plant-derived triterpenoid sweetness inhibitors. J Ethnopharmacol 47, 9–26. Y. Izutani et al. Interaction of gymnemic acid with cyclodextrins FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS 6159 8 Jiang P, Cui M, Zhao B, Liu Z, Snyder LA, Benard LM, Osman R, Margolskee RF & Max M (2005) Lacti- sole interacts with the transmembrane domains of human T1R3 to inhibit sweet taste. J Biol Chem 280, 15238–15246. 9 Sugita D, Inoue R & Kurihara Y (1998) Sweet and sweetness-inducing activities of new triterpene glyco- sides, strogins. Chem Senses 23, 93–97. 10 Nelson G, Hoon MA, Chandrashekar J, Zhang Y, Ryba NJ & Zuker CS (2001) Mammalian sweet taste receptors. Cell 106, 381–390. 11 Zhao GQ, Zhang Y, Hoon MA, Chandrashekar J, Erlenbach I, Ryba NJ & Zuker CS (2003) The receptors for mammalian sweet and umami taste. Cell 115, 255–266. 12 Nagaoka T, Hane H, Yamashita H & Kensho I (1990) Taste improvement of extracts from the leaves of Gym- nema sylvestre. Seitou Gijutsu Kenkyukai-Shi 38, 61–70 (in Japanese). 13 Imoto T (1990) Sugar discrimination and plant-origin- ated glycosides, gymnemic acids. Seibutsu Butsuri 30, 146–150 (in Japanese). 14 Szejtli J (1998) Introduction and general overview of cyclodextrin chemistry. Chem Rev 98, 1743–1753. 15 Rekharsky MV & Inoue Y (1998) Complexation thermodynamics of cyclodextrins. Chem Rev 98, 1875– 1918. 16 Wimmer R, Aachmann FL, Larsen KL & Petersen SB (2002) NMR diffusion as a novel tool for measuring the association constant between cyclodextrin and guest molecules. Carbohydr Res 337, 841–849. 17 Davis ME & Brewster ME (2004) Cyclodextrin-based pharmaceutics: past, present and future. Nat Rev Drug Discov 3, 1023–1035. 18 Wiseman T, Williston S, Brandts JF & Lin L-N (1989) Rapid measurements of binding constants and heats of binding using a new titration calorimeter. Anal Biochem 179, 131–137. 19 Fukada H & Takahashi K (1998) Enthalpy and heat capacity changes for the proton dissociation of various buffer components in 0.1 m potassium chloride. Proteins 33, 159–166. 20 Liu HM, Kikuchi F & Tsuda Y (1992) Isolation and structure elucidation of gymnemic acids, antisweet prin- ciples of Gymnema sylvestre. Chem Pharm Bull 40, 1366–1375. 21 Chen W, Chang CE & Gilson MK (2004) Calculation of cyclodextrin binding affinities: energy, entropy, and implications for drug design. Biophys J 87, 3035–3049. 22 Harries D, Rau DC & Parsegian VA (2005) Solutes probe hydration in specific association of cyclodextrin and adamantane. J Am Chem Soc 127, 2184–2190. 23 Sturtevant JM (1977) Heat capacity and entropy chan- ges in processes involving proteins. Proc Natl Acad Sci USA 74, 2236–2240. 24 Murphy KP & Freire E (1992) Thermodynamics of structural stability and cooperative folding behavior in proteins. Adv Protein Chem 43, 313–361. 25 Spolar RS, Livingstone JR & Record MT Jr (1992) Use of liquid hydrocarbon and amide transfer data to esti- mate contributions to thermodynamic functions of pro- tein folding from the removal of nonpolar and polar surface from water. Biochemistry 31, 3947–3955. 26 Oda M, Furukawa K, Ogata K, Sarai A & Nakamura H (1998) Thermodynamics of specific and non-specific DNA binding by the c-Myb DNA-binding domain. J Mol Biol 276, 571–590. 27 Oda M & Azuma T (2000) Reevaluation of stoichio- metry and affinity ⁄ avidity in interactions between anti- hapten antibodies and mono- or multi-valent antigens. Mol Immunol 37, 1111–1122. 28 Schneider HJ, Hacket F, Rudiger V & Ikeda H (1998) NMR studies of cyclodextrins and cyclodextrin complexes. Chem Rev 98, 1755–1786. 29 Izutani Y, Murai T, Imoto T, Ohnishi M, Oda M & Ishijima S (2005) Gymnemic acids inhibit rabbit glycer- aldehyde-3-phosphate dehydrogenase and induce a smearing of its electrophoretic band and dephosphoryla- tion. FEBS Lett 579, 4333–4336. Interaction of gymnemic acid with cyclodextrins Y. Izutani et al. 6160 FEBS Journal 272 (2005) 6154–6160 ª 2005 FEBS . Interaction of gymnemic acid with cyclodextrins analyzed by isothermal titration calorimetry, NMR and dynamic light scattering Yusuke. of GA and the inclusion complexation of GA by c-CD. Abbreviations CD, cyclodextrin; DLS, dynamic light scattering; GA, gymnemic acid; ITC, isothermal titration