A Rapid Determination Method for Anthocyanin Profiling in Grape Genetic Resources

Transcription

1 J. Japan. Soc. Hort. Sci. 76 (1): Available online at JSHS 007 A Rapid Determination Method for Anthocyanin Profiling in Grape Genetic Resources Mikio Shiraishi*, Masahiko Yamada, Nobuhito Mitani and Toshihito Ueno Department of Grape and Persimmon Research National Institute of Fruit Tree Science, Akitsu, Higashihiroshima , Japan A rapid determination method has been developed to elucidate the anthocyanin profiles in grape genetic resources. It consists of crude extraction of skin anthocyanins with 50% aqueous acetic acid and analysis of reversed-phase HPLC with a visible spectrum detector. A total of 1 anthocyanins could be rapidly identified without the need for multi-step extraction and equipment such as diode array spectroscopy and/or mass spectrometry. Analysis of 17 colored cultivars exhibited diversified anthocyanin profiles in the levels of hydroxylation, methylation, glycosidation, and acylation of aglycones, suggesting that significant inter- and intraspecific variations exist in grape genetic resources. The anthocyanin profiles for a grape cultivar from the same location appeared to be quite similar between vines or years. Key Words: anthocyanin, genetic resources, grape, HPLC, profiling. Introduction The quantity and quality of color in grape berries are crucial factors that influence table grape marketing and wine or juice making. The varietal difference in the color of black and red grapes results from the accumulation of anthocyanidins that are modified by the attachment of glucose moieties to form anthocyanins. Anthocyanins are generally limited to the vacuoles of a few cell layers below the epidermis that affect light scattering or penetration of violet-black or red skin tissue of grape cultivars (Fig. 1). No complex formation between anthocyanins and metal ions or flavonoids is involved in the color expression of grapes (Cheynier and Rigaud, 1986; Kitamura et al., 005a; Moskowitz and Hrazdina, 1981). The total level of anthocyanins increased throughout ripening, whereas composition of the individual anthocyanins did not vary greatly (Boss et al., 1996a; Shiraishi and Watanabe, 1994a). Vitis species or cultivars have a unique set of anthocyanins, and their profiles can be used as finger prints for chemotaxonomic classification (Amerine et al., 198; Boss et al., 1996b; Shiraishi and Watanabe, 1994b). V. vinifera cultivars produce 3-monoglucoside and 3-p-coumarylglucoside derivatives of aglycones based on the structural difference in the B-ring as follows: cyanidin, peonidin, delphinidin, petunidin, and malvidin (Fig. ). Most other Vitis species or their hybrids produce 3,5-diglucosides which are used as a diagnostic tool to determine if a non-vinifera species has been used in wine making (Ribéreau-Gayon, 1974). First done by paper chromatography, and then thinlayer chromatography over several decades, high performance liquid chromatography (HPLC) has become the standard technique because it allows the rapid and complete separation of anthocyanins from different grapes. In a previous report (Fujishima and Shiraishi, 1997), a total of 18 anthocyanins could be identified using a classical HPLC approach with a visible spectrum detector. However, the detection of petunidin derivatives Received; February 10, 006. Accepted; August 10, 006. * Corresponding author ( mikioshi@farc.pref.fukuoka.jp). * Present address: Fukuoka Agriculture Research Center, Chikushino, Fukuoka , Japan. Fig. 1. Varietal difference in the density of cell layers below the epidermis in the violet-black and red skin tissue of grape cultivars. A: Kyoho (violet-black), B: Muscat Hamburg (violet-black), C: Yoho (red), D: Ruby Okuyama (red). 8

2 J. Japan. Soc. Hort. Sci. 76 (1): Fig.. The B-ring of the anthocyanin biosynthetic pathway with modification of Boss et al. (1996b). Multi-step enzymatic contributions exist in the pathway from dihydroflavonols to cyanidin or delphinidin. F3'H: flavonoid 3'-hydroxylase, F3'5'H: flavonoid 3',5'-hydroxylase, MT: methyltransferase. is not always satisfactory. Malvidin 3-acetylglucoside, which exists in a significant amount among vinifera red wine grapes (Boss et al., 1996b), has not been identified in the above approach. Furthermore, the variations in grape anthocyanin profiles between vines or years have not yet been documented. The objectives of this study were to re-examine the classical HPLC approach and propose a rapid determination method that will allow us in a reasonably short time to extract, identify, and confirm the anthocyanin profiles in grape genetic resources. Materials and Methods A total of 1 anthocyanins were previously isolated and purified from three grape cultivars referring to the chromatographic procedure of Anderson et al. (1970) and Hrazdina (1970): 3,5-diglucosides of malvidin, petunidin, delphinidin, and peonidin, 3,5-diglucosides acylated with p-coumaric acid of malvidin, and peonidin, 3-monoglucoside of malvidin, petunidin, delphinidin, peonidin, and cyanidin, and 3-monoglucoside acylated with p-coumaric acid of malvidin, petunidin, delphinidin, and peonidin in Rubired ; 3,5-diglucosides of cyanidin, 3,5-diglucosides acylated with p-coumaric acid of petunidin, delphinidin, and cyanidin, and 3-monoglucoside acylated with p-coumaric acid of cyanidin in Campbell Early ; and 3-monoglucoside acylated with acetic acid of malvidin in Shiraz. Plant materials consisted of 6 cultivars of Vitis vinifera (V) and 11 cultivars of V. labruscana V. vinifera (LV) grown under open culture in the Department of Grape and Persimmon Research, National Institute of Fruit Tree Science (NIFTS), Higashihiroshima in Japan. One vine of Aki Seedless (LV), Aki Queen (LV), Cabernet Sauvignon (V), Flame Tokay (V), Kyoho (LV), Muscat Hamburg (V), Rizamat (V), Shiraz (V), and Steuben (LV) was used for anthocyanin profiling in 004: ten grape berries were randomly sampled from two to three clusters per vine. Two vines of Aki Queen, Dark Ridge (LV), and Oriental Star (LV) were used for analysis of variations among vines within genotypes in 004: five clusters per vine and ten berries per cluster were taken. One vine of Campbell Early (LV), Delaware (LV), Homan (LV), Kai Noir (LV), Merlot (V), and Yoho (LV) was used to examine yearto-year variations within genotypes between 004 and 005 in which berry sampling was performed in a similar way to anthocyanin profiling. Skin of ca. 0.1 g at the equatorial plane was peeled from each berry. After being semi-dried by kimtowel (Crecia, Japan), 1.0 to 1. g of skins were macerated in 0 ml of 50% aqueous acetic acid (v/v) for 1 h at 4 C in the dark, followed by filtration through a No. filter paper (ADVANTEC, Japan). One ml of filtrate was diluted to 10 ml in 50% aqueous acetic acid, and then measured as the total anthocyanin content by reading the absorbance at 50 nm (A 50 g 1 skin) with a spectrophotometer (UV- 60, Shimadzu, Japan) in a condition of SCALE 0 nm cm 1 and SLIT nm. The concentration of the total anthocyanin content (TAC) was expressed as cyanidin- 3-monoglucoside equivalent based on the following equation where the correlation coefficient was : TAC µg cm skin = A 50 g 1 skin Furthermore, the L* value was calculated by the formula of Watanabe and Shiraishi (1991): L* = Log TAC. For HPLC analysis, the aforementioned filtrate was passed through a 0.45 µm filter (ADVANTEC). The injection volume was 1 to 10 µl for violet or black grapes and 0 to 30 µl for red ones. A 5-µm C18 Inertsil ODS- (50 mm 6.0 mm i.d., GL Sciences, Japan) and LC-10A (Shimadzu) equipment consisting of LC-10AD pumps, a CTO-10A column oven, a SCL-10A system controller, a SPD-10AV detector, and a C-R7A chromatopak were used. The column was operated at 35ºC, and the mobile phase flowed at 0.8 ml/ min. Solvent A was 1.5% phosphoric acid (v/v), and solvent B was 1.5% phosphoric acid (v/v), 0% acetic acid (v/v), and 5% acetonitrile (v/v). The initial condition of solvent B was 5%, which linearly increased to 85% in 40 min. The following conditions were used for the analysis of peak areas in the detector (absorbance 50 nm; range 0.1AUFS) and the chromatopack (width 3 s; slope 10 µv min 1 ; drift 0 µv min 1 ; min area 100; t.dbl 0 min; atten. mv; method 0; idf 0; window 5%; and spl.wt 100). Individual anthocyanin data were determined by comparing the size of the individual peak areas and expressing this as a percentage of the total area of all peaks. Anthocyanin data of variations among vines within genotypes were analyzed by the unpaired t-test. Anthocyanin profiles shown in the percentage data were arc-sin transformed before statistical analysis. Year-to-year variations of TAC within genotypes were analyzed by two-factor ANOVA: the sources of variation are the

3 30 M. Shiraishi, M. Yamada, N. Mitani and T. Ueno genotype, year, and error (genotype year). The similarity of anthocyanin profiles was compared according to the pattern similarity index (Shiraishi and Shiraishi, 00), PSI (OA, OB), which is defined as the cosine of the angle θ between vector OA (a 1, a,, a n ) and OB (b 1, b,, b n ) in n dimensional space: PSI( OA, OB) = cosθ= i= 1 a i b i Results and Discussion One % hydrochloric acid in methanol (1%HCl-MeOH) has been used as an extraction solvent for the analysis of grape anthocyanins. However, Yamaguchi and Terahara (1985) pointed out the instability of carnation anthocyanins in the solvent of 1%HCl-MeOH, especially acylated ones. They also recommend that acetic acid was suitable for the extraction solvent. We used 50% aqueous acetic acid as the extraction solvent instead of 1%HCl-MeOH to prevent grape anthocyanins from deacylating. The skin extracts for HPLC analysis using 50% aqueous acetic acid could be stored at 4 C in the dark for 6 months. The total anthocyanin content, TAC, varied from 16.4 in Flame Tokay (bright red) to 6.1 µg cm skin of Shiraz (blue-black). The TAC of Aki Queen (red), Rizamat (red-violet), Muscat Hamburg (violet-black), Cabernet Sauvignon (blue-black), Steuben (violetblack), Aki Seedless (blue-black), and Kyoho (violetblack) was 34.0, 58.0, 117., 194.0, 195., 01.6, and 09. µg cm skin, respectively. A 5- to 16-fold difference in the TAC was present between red and blue-black grapes, which is similar to data reported previously (Boss et al., 1996b). Figure 3 shows an HPLC chromatogram derived from a mixture of skin extract of Rubired : Campbell Early : Shiraz = 1 : 4 : 4 (v/v). The injection volume was 3 µl. A small injection volume of below 5 µl was satisfactory for the qualitative determination of anthocyanins present in black skinned-grapes. Of the 30 peaks detected, a total of 1 peaks were identified by co-chromatography of the known anthocyanin standards in the following retention time (min) and elution order: 9.61, Dp3,5G (delphinidin- 3,5-diglucoside); 1.95, Cy3,5G (cyanidin-3,5- diglucoside); 14.0, Dp3G (delphinidin-3- monoglucoside); 14.49, Pt3,5G (petunidin-3,5- diglucoside); 17.0, Cy3G (cyanidin-3-monoglucoside); 17.30, Pn3,5G (peonidin-3,5-diglucoside); 18.3, Mv3,5G (malvidin-3,5-diglucoside); 18.56, Pt3G (petunidin-3-monoglucoside); 1.6, Pn3G (peonidin-3- monoglucoside);.77, Mv3G (malvidin-3- monoglucoside); 8.44, Dp3pG5G (delphinidin-3-(pcoumarylglucoside)-5-glucoside); 31.13, Cy3pG5G (cyanidin-3-(p-coumarylglucoside)-5-glucoside); 31.95, n n b i n a i i= 1 i= 1 Pt3pG5G (petunidin-3-(p-coumarylglucoside)-5- glucoside); 3.84, Dp3pG (delphinidin-3-pcoumarylglucoside); 33.4, Mv3aG (malvidin-3- acetylglucoside); 34.99, Pn3pG5G (peonidin-3-(pcoumarylglucoside)-5-glucoside); 35.4, Mv3pG5G (malvidin-3-(p-coumarylglucoside)-5-glucoside); 35.99, Cy3pG (cyanidin-3-p-coumarylglucoside); 36.85, Pt3pG (petunidin-3-p-coumarylglucoside); 40.3, Pn3pG (peonidin-3-p-coumarylglucoside); 40.71, Mv3pG (malvidin-3-p-coumarylglucoside). Peaks of Pt3,5G, Pt3pG5G, and Mv3aG were additionally identified in this study compared with a previous report of Fujishima and Shiraishi (1997). It is well known that the elution order of anthocyanins in reversed-phase HPLC using a C 18 type column is closely related to their polarity, with the more polar compounds eluting first (Wulf and Nagel, 1978): the anthocyanins become more polar as the number of hydroxyl groups in the B ring increases and more apolar as the number of methoxyl groups in it increases. Thus, delphinidin (Dp) elutes first, followed in order by cyanidin (Cy), petunidin (Pt), peonidin (Pn), and malvidin (Mv) in spite of the difference in analytical conditions such as column size, mobile phase, and flow rate. Furthermore, the diglucosides are more polar than the monoglucosides, and acylation involves a loss of polarity of the corresponding non-acylated compounds, resulting in an increase in their retention times. The elution order of anthocyanins using classical HPLC in this study was consistent with recent HPLC approaches using diode array spectroscopy and/or mass spectrometry, except for Mv3aG (Hebrero et al., 1989; Kitamura et al., 005a). The peak of Mv3aG, eluted before Dp3pG, which is in contrast to the report of Boss et al. (1996a). However, the relative percent to the total anthocyanin peak of Mv3aG was calculated as 8.1% for Cabernet Sauvignon and 15.4% for Shiraz, agreeing with the results of Boss et al. (1996b). Furthermore, respective anthocyanin profiles in both cultivars were almost identical to those using their results (data not shown). For the remaining 9 peaks, UIP 1 to 9 on the chromatogram, were not identified, but peaks eluting in 4.70 (UIP 4), 5.87 (UIP5), 6.85 (UIP6), and 9.14 (UIP9) min may be 3-acetyglucosides of Dp, Cy, Pt, and Pn, respectively, compared with the result of Boss et al. (1996a). Anthocyanin profiles in the typical colored table cultivars are illustrated in Figure 4. TAC increased in accordance with the color change, and the L* value decreased correspondingly, ranging from 8.3 in Flame Tokay to 1.5 in Kyoho. However, there was little difference in the TAC and L* values among black skinned-cultivars. A decline in the L* value indicates a decrease in the lightness of berry skin, resulting in a bathochromic shift in the skin color. These results are in agreement with previous reports (Gao and Cahoon, 1994; Kitamura et al., 005a; Watanabe and Shiraishi, 1991).

4 J. Japan. Soc. Hort. Sci. 76 (1): Fig. 4. Graphical plot of the skin anthocyanin profiles in six table grape cultivars. Abbreviations of each anthocyanin is described in Fig. 3. TAC: total anthocyanin content µg cm, L*: lightness of fruit skin.

5 3 M. Shiraishi, M. Yamada, N. Mitani and T. Ueno There were significant variations in the proportions of anthocyanins specific to species parentage or cultivar. Cultivars of V. vinifera ( Flame Tokay and Rizamat ) had only 3-monoglucoside and its acylated derivatives, whereas those of V. labruscana V. vinifera (LV) ( Aki Queen, Steuben, Aki Seedless, Kyoho ) had both 3- and 3,5-diglucosides with their acylated derivatives. Flame Tokay had predominantly non-methylated forms of Cy3G at 67.5%, with minimal amounts of other anthocyanins. In Aki Queen, Pn and Cy derivatives constituted 61% of the total, with relatively large amounts of Mv derivatives in %. Rizamat and Muscat Hamburg had primarily Pn3G and Mv3G, which are methylated in the B-ring of the molecule, accounting for 65.5% of the total. By contrast, Steuben contained mainly Cy and Dp derivatives, which are hydroxylated in the B-ring of the molecule, accounting for 95.1% of the total. Dp derivatives that trihydroxylated in the B-ring of the molecule were significantly present in Aki Seedless at 67.8% of the total. The major anthocyanins of Kyoho were Mv derivatives (67.4%), especially Mv3pG and Mv3pG5G, followed by Pt derivatives at 10.7%. These results are consistent with those reported by Shiriahi et al. (1986) and Shiraishi and Watanabe (1994b). With respect to Aki Queen, Kitamura et al. (005a) determined 13 peaks, whereas an additional 4 peaks (Cy3pG5G, Dp3pG, Dp3pG5G, and Mv3aG) were identified in this study. However, the relative proportion of major anthocyanins, the latter value by Kitamura et al. (005a): Cy3pG in 8.7% vs. 8.0%, Pn3pG5G in 10.1% vs. 8.7%, Pn3pG in 16.3% vs. 16.3%, and Mv3G in 7% vs. 6.4%, were similar to our results except for Cy3G + Pn3,5G in 10.8% vs. 17.3%. Commercial cultivars that lack the ability to produce higher hydroxylated or methylated aglycones and acylated forms appear to include a higher proportion of light-colored berries (Singleton and Esau, 1969; Watanabe and Shiraishi, 1991), which is partly confirmed by the change in the skin color of red cultivars. Comparisons of the anthocyanin profiles between two vines of three table cultivars in 004 are listed in Table 1. There were no significant differences in TAC and L* between vines within cultivars. Unpaired t-tests show that all cultivars had very similar levels of anthocyanins with negligible differences within cultivars. The similarity of anthocyanin profiles between vines can be explained by Table 1. Comparison of the anthocyanin profiles between two vines of three grape cultivars in 004. Source of variation Aki Queen Oriental Star Dark Ridge Vine 1 v Vine v t-test w Vine 1 v Vine v t-test w Vine 1 v Vine v t-test w TAC z 3.8 ± ± 1.75 NS 17.0 ± ± NS ± ± 7.30 NS L* y 6.4 ± ± 0.15 NS.8 ± ± 0.1 NS 1.8 ± ± 0.10 NS Anthocyanins x Mv3pG5G 3.0 ± ± 0.07 NS 9.3 ± ± 0.5 NS Mv3,5G 3.8 ± ± 0.3 NS 10.0 ± ± 0.13 * Mv3aG 1.5 ± ± 0.08 NS 1.6 ± ± 0.1 * 1. ± ± 0.09 NS Mv3pG 10.6 ± ± 0.94 NS 1.8 ± ± 0.19 NS 9.5 ± ± 0.33 NS Mv3G 7.1 ± ± 0.4 NS 8.3 ± ± 0.77 NS 3. ± ± 0.8 NS Pt3pG5G 3.7 ± ± 0.09 * Pt3,5G 0.3 ± ± 0.03 NS Pt3pG 5.5 ± ± 0.4 NS 0.6 ± ± 0.06 NS 5.4 ± ± 0.17 NS Pt3G 1.9 ± ± 0.1 NS 9.6 ± ± 0.46 NS 5.0 ± ± 0.06 NS Dp3pG5G 0.8 ± ± 0.13 NS 1.9 ± ± 0.09 NS Dp3pG 4.3 ± ± 0.6 NS 1.8 ± ± 0.16 NS 6.5 ± ± 0.18 NS Dp3G 0.9 ± ± 0.11 NS 5.5 ± ± 0.5 NS 0.9 ± ± 0.07 NS Pn3pG5G 8.9 ± ± 0.45 NS 4.9 ± ± 0.4 NS Pn3,5G 4. ± ± 0.1 NS 3.7 ± ± 0.31 NS Pn3pG 15.3 ± ± 0.56 NS.1 ± ± 0.4 NS.0 ± ± 0.09 NS Pn3G 1.0 ± ± 0.56 NS 36.4 ± ± 0.51 NS 1.4 ± ± 0.15 NS Cy3pG5G 1.8 ± ± 0.09 NS.9 ± ± 0.13 NS Cy3pG 8.1 ± ± 0.1 NS 0.8 ± ± 0.09 NS Cy3G 4.5 ± ± 0.81 NS 9.5 ± ± 0.40 *.1 ± ± 0.09 NS unidentified 5.7 ± ± 0. NS.0 ± ± 0. NS 6. ± ± 0.5 NS z Total anthocyanin content (µg cm cyanidin-3-monoglucoside equivalent). y Lightness of berry skin: Log TAC. x See Fig. 3: Data shown as a percentage. w NS, * Non significant or significant at 5% levels, respectively by the unpaired t-test based on the arc-sin transformed data of anthocyanins (n = 5). v Mean ± SE (n = 5).

6 J. Japan. Soc. Hort. Sci. 76 (1): the pattern similarity index (PSI) (Shiraishi and Shiraishi, 00). Relative values of PSI in each cultivar were calculated on the basis of the reference pattern of vine 1. Dark Ridge had the highest similarity of 0.999, followed by Oriental Star of 0.997, whereas Aki Queen had the lowest PSI of Table shows the anthocyanin profiles of 6 grape cultivars collected from the same location in two consecutive years. It can be seen that TAC and L* values tended to be variable. However, there was no significant difference in TAC between the two years (Table 3). When the total variance (σ T ) for TAC is defined as σ T = σ g (genotype) + σ y (year) + σ (error), the respective value of σ g, σ y and σ was 9610., 6.5, and 194.3, resulting in the year effect being very small, accounting for 0.3% of the total variance. The genotypic effect (σ g ) is so large that there were large differences for TAC between red ( Yoho and Delaware ) and black ( Homan, Campbell Early, Merlot and Kai Noir ) skinned-cultivars. The patterns of anthocyanins in each cultivar for two years looked quite similar because the PSI of Yoho, Delaware, Homan, Campbell Early, Merlot, and Kai Noir was 0.999, 0.993, 0.995, 0.986, 0.994, and 0.993, respectively. There are some possible explanations for these high stabilities in the anthocyanin profiles: (i) cultural practice such as canopy management, fertilization, and crop loading is uniformly performed in our experimental field every year; (ii) berry sampling was undertaken at approximately the same degrees of maturity; and (iii) the proportions of each anthocyanins are cultivar-related, reflecting genetic differences among cultivars. For a grape cultivar from the same location in different vines or years, the anthocyanin profiles are considered to have similar patterns, although there may be significant variations in the levels of some anthocyanins. Singleton and Esau (1969) indicated that the anthocyanin chromatographic pattern obtained with a given grape cultivar is evidently rather reproducible with careful techniques, and that there are likely to be considerable differences between cultivars. However, it should be noted that the anthocyanin profiles possibly changed according to differences in the growing area (Mori et al., 000; Ribéreau-Gayon, 1958), high night temperatures (Mori et al., 005), crop load (Kitamura et al., 005b), and cluster shading (Gao and Cahoon, 1994; Okamoto et al., 1995). Boss et al. (1996b) postulated that the anthocyanin profiles of colored cultivars result from the complexity of Table. Comparison of the anthocyanin profiles of 6 grape cultivars between 004 and 005. Descriptors Yoho Delaware Houman Campbell Early Merlot Kai Noir TAC z L* y Anthocyanins x Mv3pG5G Mv3,5G Mv3aG Mv3pG Mv3G Pt3pG5G Pt3,5G Pt3pG Pt3G Dp3pG5G Dp3pG Dp3G Pn3pG5G Pn3,5G Pn3pG Pn3G Cy3pG5G Cy3,5G Cy3pG Cy3G unidentified z Total anthocyanin content (µg cm cyanidin-3-monoglucoside equivalent). y Lightness of berry skin: LogTAC. x See Fig. 3: Data shown as a percentage.

7 34 M. Shiraishi, M. Yamada, N. Mitani and T. Ueno Table 3. Analysis of variance of the total anthocyanin content (TAC) using 6 genotypes with one vine per genotype for years. Source of variation df MS F z Expected MS y Genotype ** σ + σ g Year NS σ + 6σ y Error σ z NS, ** Non significant or significant at 1% levels, respectively. y Total variance, σ T, is defined as follows: σ T = σ g + σ y + σ. the genetics of the biosynthetic pathway. The anthocyanin profiles listed in Tables 1 and suggest that great genotypic diversity among the cultivars occurs in the levels of hydroxylation, methylation, glycosidation, and acylation of aglycones. It is thus of great interest to elucidate the anthocyanin profiles in grape genetic resources both in genetic and enological research. The reexamined classical HPLC approach using a visible spectrum detector could permit rapid identification of the different anthocyanins without the need for multi-step extraction and equipment such as diode array spectroscopy and mass spectrometry. In our laboratory (NIFTS), a total of 48 colored grape genetic resources including cultivars, selections, breeding stocks, Vitis species, and their interspecific hybrids (rootstock) have been preserved, and evaluated for anthocyanin profiles by this rapid determination method since 004. Acknowledgment The authors would like to thank Dr. Masaatsu Yamaguchi (University of Minamikyushu, Japan) for excellent technical assistance with respect to the identification of grape anthocyanin profiling. Literature Cited Amerine, M. A., H. W. Berg, R. E. Kunkee, C. S. Ough, V. L. Singleton and A. D. Webb The composition of grapes. p In: The technology of wine making 4th ed. AVI Pub. Co. Westport. Anderson, D. W., D. E. Gueffroy, A. D. Webb and R. E. Kepner Identification of acetic acid as an acylating agent of anthocyanin pigments in grapes. Phytochemistry 9: Boss, P. K., C. Davies and S. P. Robinson. 1996a. Analysis of the expression of anthocyanin pathway genes in developing Vitis vinifera L. cv Shiraz grape berries and the implications for pathway regulation. Plant Physiol. 111: Boss, P. K., C. Davies and S. P. Robinson. 1996b. Anthocyanin composition and anthocyanin pathway gene expression in grapevine sports differing in berry skin colour. Aust. J. Grape Wine Res. : Cheynier, V. and J. Rigaud HPLC separation and characterization of flavonols in the skins of Vitis vinifera var. Cinsault. Amer. J. Enol. Viticult. 37: Fujishima, H. and M. Shiraishi Studies on the evaluation of genetic resources in Vitis. 7. A rapid screening test of skin anthocyanin. J. Japan. Soc. Hort. Sci. 66 (Suppl. ): (In Japanese). Gao, Y. and G. A. Cahoon Cluster shading effects on fruit quality, fruit skin color, and anthocyanin content and composition in Reliance (Vitis hybrid). Vitis 33: Hebrero, E., C. Garcia-Rodriguez, C. Santos-Buelga and J. C. Rivas- Gonzalo Analysis of anthocyanins by high performance liquid chromatography-diode array spectroscopy in a hybrid grape variety (Vitis vinifera Vitis berlandieri 41B). Amer. J. Enol. Viticult. 40: Hrazdina, G Column chromatographic isolation of the anthocyanidin-3,5-diglucosides from grapes. J. Agr. Food Chem. 18: Kitamura, H., M. Nakayama, Y. Nishikawa, H. Kondou, M. Koshioka and S. Hiratsuka. 005a. Relationship between the deep coloration and anthocyanins in the skin of Aki Queen berries. J. Japan. Soc. Hort. Sci. 74: (In Japanese with English summary). Kitamura, H., M. Nakayama, Y. Nishikawa, H. Kondou, M. Koshioka and S. Hiratsuka. 005b. Effect of crop load on the expression of deep color in Aki Queen grape berry. Hort. Res. (Japan) 4: (In Japanese with English summary). Moskowitz, A. H. and G. Hrazdina Vacuolar contents of fruit subepidermal cells from Vitis species. Plant. Physiol. 68: Mori, K., H. Saito, N. Goto-Yamamoto, M. Kitayama, S. Kobayashi, S. Sugaya, H. Gemma and K. Hashizume Effects of abscisic acid treatment and night temperatures on anthocyanin composition in Pinot noir grapes. Vitis 44: Mori, K., S. Sugaya, H. Gemma and S. Iwahori Comparison of coloration of Kyoho grape berries grown at different locations. J. Japan. Soc. Hort. Sci. 69 (Suppl. ): 80 (In Japanese). Okamoto, S., I. Kataoka and H. Tamura Anthocyanin constituents of grape berries ripened under different growing conditions. J. Japan. Soc. Hort. Sci. 64 (Suppl. 1): (In Japanese). Ribéreau-Gayon, P Formation et évolution des anthocyannes au cours de la maturation du raisin. Compt. Rend. 46: Ribéreau-Gayon, P The chemistry of red wine color. p In: Chemistry of wine making. Amer. Chem. Soc. Washington, D. C. Shiraishi, M. and S. Shiraishi. 00. Similarity of amino acid composition of grape berries produced in different years and locations. J. Japan. Soc. Hort. Sci. 71: Shiraishi, S. and Y. Watanabe. 1994a. Fruit ripes and anthocyanin pigments in the skin of grapes (Vitis spp.). Sci. Bull. Fac. Agr. Kyushu Univ. 48: (In Japanese with English summary). Shiraishi, S. and Y. Watanabe. 1994b. Anthocyanin pigments in the grape skins of cultivars (Vitis spp.). Sci. Bull. Fac. Agr. Kyushu Univ. 48: 55 6 (In Japanese with English summary). Shiraishi, S., Y. Watanabe, H. Okubo and S. Uemoto Anthocyanin pigments of black-purple grapes related to variety Kyoho (Vitis vinifera L. V. labrusca L.). J. Japan. Soc. Hort. Sci. 55: Singleton, V. L. and P. Esau Qualitative phenolic composition of grapes and wines. p In: Phenolic substances in grapes and wine, and their significance. Academic press, New York. Watanabe, Y. and S. Shiraishi Relationship between skin color and anthocyanin of grapes. Sci. Bull. Fac. Agr. Kyushu Univ. 46: (In Japanese with English summary). Wulf, L. W. and C. W. Nagel High-pressure liquid chromatography of anthocyanins of Vitis vinifera. Amer. J. Enol. Viticult. 9: 4 49.

8 J. Japan. Soc. Hort. Sci. 76 (1): Yamaguchi, M. and N. Terahara, Instability of carnation anthocyanins in solvent of hydrochloric acid in methanol. Fac. Hort. Minamikyushu Univ. Bull. 15: (In Japanese with English summary).