Reduction of Catechin Astringency by the Complexation of Gallate-Type Catechins with Pectin

Transcription

1 Biosci. Biotechnol. Biochem., 69 (7), , 2005 Reduction of Catechin Astringency by the Complexation of Gallate-Type Catechins with Pectin Nobuyuki HAYASHI, y Tomomi UJIHARA, and Katsunori KOHATA Department of Physiology and Quality Science, National Institute of Vegetable and Tea Science, 2769 Kanaya, Shimada, Shizuoka , Japan Received February 1, 2005; Accepted April 16, 2005 The reductive effect of pectin on tea catechin astringency was investigated by using a taste sensor system and 1 H-NMR spectroscopy. The sensor analysis revealed that the astringency of gallate-type catechins (EGCg and ECg) was reduced by the addition of pectin, whereas that of non-gallate-type catechins (EGC and EC) hardly changed. Changes in the 1 H-NMR chemical shifts of the catechins and pectin in mixed solutions showed that the gallate-type catechins formed complexes with pectin more closely than the non-gallatetype catechins. These results demonstrate that complexation between the gallate-type catechins and pectin is a factor for reducing catechin astringency. Key words: catechin; pectin; astringent reduction; taste sensor; 1 H-NMR The taste of foodstuffs is an extremely complicated system controlled by various factors such as the synergistic or inhibitory action among tasting substances. With the example of Japanese green tea liquor, ( )- epigallocatecin gallate (EGCg) of 0.82 mm, ( )-epicatechin gallate (ECg) of 1.2 mm, ( )-epigallocatechin (EGC) of 0.12 mm, and ( )-epicatechin (EC) of 0.31 mm (the structures of the catechins are shown in Fig. 1) are each included as astringent polyphenols when tea leaves (3 g) have been infused in hot water (200 ml) at 90 C for 2 min. 1) However, if a solution including the catechins at the same concentrations is drunk, the taste is more astringent than that of the tea infusion itself. This phenomenon suggests that the astringency of catechins is reduced by other chemical components in the tea infusion. Taira et al. have found that persimmon astringency was reduced by a complex formation between pectin and tannins (high-molecular-weight polyphenols). 2) Pectin is certainly also present in the green tea infusion. 3,4) Horie et al. have reported that tea catechin astringency was also mitigated by the addition of pectin. 5) However, the relationship between the structure of tea catechins and the degree of reduced astringency by pectin has not been clarified. Furthermore, it is not clear whether, in the case of low-molecular-weight polyphenols such as the tea catechins, they can form a complex with pectin. In the present study, the reductive effect on catechin astringency by pectin was investigated by means of a taste sensor system and 1 H-NMR spectroscopy. We report here the difference between the gallate-type catechins (gallic acid ester at the C 3 position of the carechins) and the non-gallate-type catechins (without the galloyl group) for both the astringency reduction by pectin and the interaction with pectin. The causality between the reduction of astringency and the complexation is discussed from the results. Materials and Methods Fig. 1. Four Major Catechins in a Tea Infusion. Materials. All chemicals were obtained from commercial suppliers and used without further purification. EGCg and the other catechins were respectively purchased from Wako Pure Chem. Ind. (Osaka, Japan) and Nakahara Kagaku (Gifu, Japan). The pectin from citrus was a comercially available product from Wako Pure Chem. Ind. Measurement of astringency by the taste sensor system. The electrical potential difference corresponding to the astringency of a sample solution was measured by the SA402B taste sensor system (Intelligent Sensor Technology Co., Ltd.), fitted with a sensor probe for y To whom correspondence should be addressed. Tel: ; Fax: ; hayn@affrc.go.jp

2 astringency (SB2AE1) and a reference probe. The sensor probe consists of a lipid membrane, an Ag/AgCl electrode, and an internal cavity filled with a 3.3 M KCl aqueous solution saturated with AgCl. The lipid membrane is composed of a lipid (tetradodecylammonium bromide), polymer (poly(vinyl chloride)), and dioctyl phenylphosphonate as a plasticizer. The reference probe consists of a liquid junction made with agar saturated with KCl, an Ag/AgCl electrode, and an internal cavity filled with a 3.3 M KCl aqueous solution saturated with AgCl. The electrical potential difference (E ast ) of a solution is recorded as the difference between the potential detected by the sensor electrode and the potential detected by the reference electrode. The astringency of a sample solution is defined as the difference (East sample ) between the potential difference (East sample ) of the sample solution and the potential difference (East KClaq ) of the 10 mm KCl aqueous solution (the standard solution), namely East sample ¼ East sample East KClaq. Nine 10 mm KCl aqueous solutions (10 ml) including the following compounds were prepared: i) EGCg (1.00 mg, 2.18 mmol), ii) ECg (0.964 mg, 2.18 mmol), iii) EGC (0.668 mg, 2.18 mmol), iv) EC (0.633 mg, 2.18 mmol), v) pectin (1.0 mg), vi) EGCg (1.00 mg, 2.18 mmol) and pectin (1.0 mg), vii) ECg (0.964 mg, 2.18 mmol) and pectin (1.0 mg), viii) EGC (0.668 mg, 2.18 mmol) and pectin (1.0 mg), and ix) EC (0.633 mg, 2.18 mmol) and pectin (1.0 mg). These sample solutions were assigned to two groups for measuring convenience: one was composed of the solutions of EGCg, ECg, EGCg/pectin, and ECg/pectin in order of one measurement cycle, and the other was composed of the solutions of EGC, EC, EGC/pectin, EC/pectin, and pectin in order of one measurement cycle. The measurement was automatically carried out for five cycles at 293 K for each group. * The acquisition time for the electrical potential differences (East sample and East KClaq ) was 30 sec, and the interval between successive sample measurements for washing and stabilizing the sensor probe was 7 min. ** The average of three data excluding those in the first and last measuring cycles was adopted as the E sample ast value for each sample solution. 1 H-NMR measurement. Nine sample solutions including the following compounds in deuterium oxide (D 2 O, 0.55 ml) buffered at ph 6 with 0.1 M NaH 2 PO 4 / Na 2 HPO 4 were prepared: i) EGCg (1.25 mg, 2.73 mmol), mol), ii) ECg (1.21 mg, 2.73 mmol), iii) EGC (0.84 mg, 2.73 mmol), iv) EC (0.79 mg, 2.73 mmol), v) EGCg (1.25 mg, 2.73 mmol) and pectin (1.25 mg), vi) ECg Reduction of Catechin Astringency with Pectin 1307 (1.21 mg, 2.73 mmol) and pectin (1.25 mg), vii) EGC (0.84 mg, 2.73 mmol) and pectin (1.25 mg), viii) EC (0.79 mg, 2.73 mmol) and pectin (1.25 mg), and ix) pectin (1.25 mg). NMR spectra were recorded by a JEOL JNM-LA 500 spectrometer. Chemical shift values are reported relative to tetramethyl silane ( ¼ 0:00 ppm) in carbon tetrachloride as an external standard inserted into an NMR tube ( ¼ 5 mm) with a coaxial cell. The digital resolution of the 1 H-NMR spectral data was 0.1 Hz, the measurements being carried out at 301 K. Results and Discussion The degree of reduced astringency by pectin for each catechin (EGCg, ECg, EGC, or EC) was investigated by the taste sensor system which mimics the human response to food taste. This system, which has been developed by the Toko group 6 8) and applied to evaluate the taste of medicines 9 11) and such foodstuffs as beer, 12) mineral water, 13) coffee, 14) and milk, 15,16) has recently also been proved to be able to evaluate the degree of astringency of polyphenols, including some catechins. 17) The system is particularly suitable for expensive samples, because the taste sensor analysis can be performed on a much smaller amount than that required for a human sensory test. As described in the Materials and Methods section, the degree of astringency is indicated by the East sample value which decreases with increasing degree of astringency. The East sample values for the nine sample solutions (EGCg, EGCg/pectin, ECg, ECg/pectin, EGC, EGC/pectin, EC, EC/pectin, and pectin) in a 10 mm KCl aqueous solution were 40:63, 34:21, 37:03, 32:82, 6:36, 6:12, 8:77, 8:37, and 0:70 mv, respectively. These results agree with those in a previous report which described that the gallate-type catechins were organoleptically more astringent than the non-gallate-type catechins. 18) This consistency proves that the taste sensor system could precisely evaluate the astringency of the four tea catechins. As Fig. 2 indicates, * Although measuring the astringency at human body temperature ( K) seemed to be interesting, such an experiment is technically impossible with our present sensor system. ** All measurements for one group involving five cycles was completed within 4 h. No isomerization or decomposition of the catechins under the conditions used was apparent in the 1 H-NMR spectra. Fig. 2. Electrical Potential Differences (East sample ) of Catechin Solutions Measured by the Taste Sensor System. Solid bar, catechin solution without pectin; shaded bar, catechin solution with pectin.

3 1308 N. HAYASHI et al. Fig H-NMR Chemical Shift Change in Catechins ( catechin ) by the Addition of Pectin in D 2 O Buffered at ph 6. catechin ¼ catechin+pectin catechin(free), where catechin+pectin and catechin(free) are the respective chemical shifts in the catechins with pectin and in the free catechins. The x-axes represent the carbon number attached to the proton. the East sample values for EGCg and ECg were clearly increased by the addition of pectin: the amounts of increase were þ6:42 and þ4:42 mv, respectively. On the other hand, in the case of EGC and EC, interestingly, the changes in East sample values by adding pectin were extremely small: they merely increased by þ0:24 and þ0:47 mv, respectively. The taste sensor experiments reveal that the addition of pectin reduced the astringency of the gallate-type catechins, while it hardly depressed the astringency of the non-gallate-type catechins. The increases in the East sample values of the gallate-type catechin/pectin solutions (i.e. the reduction in astringency) did not result from simple additivity of the electrical potentials for the gallate-type catechins and pectin in the solutions, because the East sample value for the pectin solution itself was virtually 0 mv. The increases were probably attributable to intermolecular interaction between the gallate-type catechins and pectin, because, if pectin had directly interacted with the lipid membrane of the sensor probe to prevent the catechin molecules from interacting with the lipid membrane, the East sample values for the non-gallate-type catechin/pectin solutions should also have increased similarly to those for the gallate-type catechin/pectin solutions. These results suggest that the gallate-type catechins formed a complex with pectin more strongly than the non-gallate-type catechins to reduce the astringency. In order to directly erve the intermolecular interaction between the gallate-type catechins and pectin, changes in the 1 H-NMR chemical shifts of the catechins with the addition of pectin were scrutinized. Figure 3 illustrates the difference ( catechin ) between the chemical shift for the catechins with pectin ( catechin+pectin ( catechin(free) catechin+pectin ) and that for the free catechins ) in deuterium oxide, namely, catechin ¼ catechin(free).the catechin values for the gallate-type catechins (0:0061 ppm) were larger than those for the non-gallate-type catechins, whose chemical shift changes were minimal: the shift change was only ppm even for the largest migration (C 2 -H of Fig H-NMR Spectra of Pectin ( ppm) in D 2 O Buffered at ph 6. a, EC/pectin; b, EGC/pectin; c, ECg/pectin; d, EGCg/pectin; e, pectin. EGC). These results support the foregoing presumption that the affinity of the gallate-type catechins for pectin was higher than that of the non-gallate-type catechins, because it is considered that the catechin values are correlated with the binding ability of the catechins for pectin. 19) The 1 H-NMR chemical shift changes for pectin were also investigated. Although the signals for pectin with the catechins were generally slightly shifted upfield by ppm in comparison with those of free pectin, substantial changes were apparent in the area shown in Fig. 4. The signal indicated by the arrow markedly migrated with the gallate-type catechin/pectin solutions further upfield than with the non-gallate-type catechin/pectin solutions: the changes were 0:0055,

4 0:0061, 0:0018, and ppm for the EGCg/ pectin, ECg/pectin, EGC/pectin, and EC/pectin solutions, respectively. These results also support the notion that the gallate-type catechins formed a complex with pectin more tightly than the non-gallate-type catechins. The higher affinity of the gallate-type catechins for pectin would be attributable to the high hydrophobicity, 20) the many sites for hydrogen bonding (the hydroxy groups), and the high conformational flexibility *** in comparison with the non-gallate-type catechins. The chemical shift changes shown in Figs. 4-c and -d are alike in the respect that the signals overall migrated upfield by similar behavior. These results suggest that the structure of the EGCg/pectin complex was similar to that of the ECg/pectin complex. Furthermore, the differences ( ECg(pec) EGCg(pec) ) between the ECg+pectin values and the EGCg+pectin values for the protons in the C-ring (C 2 -H, C 3 -H, C 4 -H, and C 4 -H, whose chemical shifts are substantially influenced by the conformational change), the B-ring, and the galloyl group decreased more in comparison with the differences ( ECg(free) EGCg(free) ) between the ECg(free) values and the EGCg(free) values for the corresponding protons. **** These results imply that the conformational difference between EGCg and ECg in the complex with pectin diminished in comparison with that between EGCg and ECg without pectin; that is, EGCg and ECg in the complex would resemble each other in their conformation. On the other hand, in Fig. 3, the profile of the chemical shift changes of EGCg is different from that of ECg. All chemical shifts of the protons in EGCg migrated downfield, whereas those in ECg shifted upfield, except for C 6 0-H and either proton on the A-ring (C 6 -H or C 8 -H; it was difficult to distinguish C 6 -H from C 8 -H in the 1 H-NMR spectra). This phenomenon could be attributed to the difference between EGCg and ECg in their conformations without pectin. Both a reduction in the astringency and migration of the 1 H-NMR chemical shifts occurred in the gallate-type catechin/pectin solution. On the other hand, such *** The 2,3-cis-catechins (EGCg, ECg, EGC, and EC) have pseudoaxial and pseudoequatorial conformers depending on the orientation of the B-ring. The energy barriers between the two conformers of the gallate-type catechins should be smaller than those of the non-gallate-type catechins, because both the 2,3-syn-substituents (the B-ring and the galloyl group) in the gallate-type catechins are bulky. Therefore, deformation of the gallate-type catechin molecule to a stable conformer in the complexation process would occur more easily than that of the non-gallate-type catechin molecule. **** The ECg(pec) ECg(pec) values (ppm) were , , , , , 0:0023, , , and 0:0002 for C 2 -H, C 3 -H, C 4 -H, C 4 -H, C 6=8 -H, C 6=8 -H, C 2 0 -H, C 6 0-H, and C 2 00 þ600 -H, respectively. The corresponding ECg(free) EGCg(free) values (ppm) were , , , , , , , , and Increases in the j ECg(pec) EGCg(pec) j values for C 6=8 -Hs might have been due to some difference between EGCg and ECg in the orientation of the A-rings for the pectin molecule. Reduction of Catechin Astringency with Pectin 1309 phenomena were hardly apparent in the non-gallatetype catechin/pectin solution. These results demonstrate that complexation between the gallate-type catechins and pectin was a factor in reducing the catechin astringency. To the best of our knowledge, the results reported here represent the first direct evidence for correlating the reduction of gallate-type catechin astringency by pectin with complexation of the gallate-type catechins with pectin. Acknowledgment This work was financially supported by NARO Research Project No References 1) Horie, H., Ujihara, T., and Kohata, K., Elution of major tea components in tea infusion. Tea Research Journal (in Japanese), 91, (2001). 2) Taira, S., Ono, M., and Matsumoto, N., Reduction of persimmon astringency by complex formation between pectin and tannins. Postharvest Biol. Technol., 12, (1997). 3) Nakagawa, M., Pectic substance in green tea and its quantitative and qualitative change by heating. J. Jpn. Soc. Food Sci. Technol. (in Japanese), 23, (1976). 4) Takeo, T., Unno, T., Kinugasa, H., Yayabe, F., and Motoyama, M., The chemical properties and functional effects of polysaccharides dissolved in green tea infusion. J. Jpn. Soc. Food Sci. Technol. (in Japanese), 45, (1998). 5) Horie, H., and Kohata, K., New candidate substances related to the taste of green tea. Jpn. J. Taste and Smell Research (in Japanese), 6, (1999). 6) Toko, K., Biomimetic Sensor Technology, Cambridge University Press, Cambridge, pp (2000). 7) Toko, K., A taste sensor. Meas. Sci. Technol., 9, (1998). 8) Toko, K., Taste sensor. Sensors and Actuators B, 64, (2000). 9) Miyanaga, Y., Kobayaashi, Y., Ikezaki, H., Taniguchi, A., and Uchida, T., Bitterness prediction or bitterness suppression in human medicines using a taste sensor. Sensors and Mater., 14, (2002). 10) Uchida, T., Tanigake, A., Miyanaga, Y., Matsuyama, K., Kunitomo, M., Kobayashi, Y., Ikezaki, H., and Taniguchi, A., Evaluation of the bitterness of antibiotics using a taste sensor. J. Pharm. Pharmacol., 55, (2003). 11) Tanigake, A., Miyanaga, Y., Nakamura, T., Tsuji, E., Matsuyama, K., Kunimoto, M., and Uchida, T., The bitterness intensity of clarithromycin evaluated by a taste sensor. Chem. Pharm. Bull., 51, (2003). 12) Toko, K., Murata, T., Matsuno, T., Kikkawa, Y., and Yamafuji, K., Taste map of beer by a multichannel taste sensor. Sensors and Mater., 4, (1992). 13) Iiyama, S., Yahiro, M., and Toko, K., Quantitative sensing of mineral water with multichannel taste sensor. Sensors and Mater., 7, (1995).

5 1310 N. HAYASHI et al. 14) Fukunaga, T., Toko, K., Mori, S., Nakabayashi, Y., and Kanda, M., Quantification of taste of coffee using sensor with global selectivity. Sensors and Mater., 8, (1996). 15) Toko, K., Iyota, T., Mizota, Y., Matsuno, T., Yoshioka, T., Doi, T., Iiyama, S., Kato, T., Yamafuji, K., and Watanabe, R., Heat effect on the taste of milk studied using a taste sensor. Jpn. J. Appl. Phys., 34, (1995). 16) Yamada, H., Mizota, Y., Toko, K., and Doi, T., Highly sensitive discrimination of milk with homogenization treatment using a taste sensor. Mater. Sci. Eng. C, 5, (1997). 17) Tokubo, R., Sato, K., Ikezaki, H., Taniguchi, A., and Toko, K., Evaluation of the strength of astringency using multichannel taste sensor. Jpn. J. Taste and Smell Research (in Japanese), 7, (2003). 18) Nakagawa, M., The relationship between astringency and the reaction aspects of astringents for protein. J. Jpn. Soc. Food Sci. Technol. (in Japanese), 19, (1972). 19) Hayashi, N., Ujihara, T., and Kohata, K., Binding energy of tea catechin/caffeine complexes in water evaluated by titration experiments with 1 H-NMR. Biosci. Biotechnol. Biochem., 68, (2004). 20) Hashimto, T., Kumazawa, S., Nanjo, F., Hara, Y., and Nakayama, T., Interaction of tea catechins with lipid bilayers investigated with liposome systems. Biosci. Biotechnol. Biochem., 63, (1999).