Development of a Stable Extract for Anthocyanins and Flavonols from Grape Skin

Transcription

1 358 Downey et al. Development of a Stable Extract for Anthocyanins and Flavonols from Grape Skin Mark O. Downey, 1 * Marica Mazza, 1 and Mark P. Krstic 1 Abstract: The lability of anthocyanins and flavonols extracted from grape (Vitis vinifera L.) skin with commonly used extraction solvents was examined. Results indicate that both anthocyanins and flavonol glycosides are unstable in even mildly acidic extraction solvents and highly labile in 1% hydrochloric acid in methanol, the most commonly used extraction solvent. While lability of anthocyanins has previously been reported, this is the first report of flavonol glycoside stability in these solvents. As a result of these observations, an extraction protocol was developed that maximized anthocyanin (malvidin-3-o-glucoside, malvidin-3-o-acetylglucoside, and malvidin- 3-O-p-coumaroylglucoside) and flavonol (quercetin-3-o-glucoside and quercetin-3-o-glucuronide) extraction while minimizing degradation in the time between extraction and HPLC analysis. A range of acidified methanol solvents including hydrochloric, formic, acetic, citric, and maleic acids were tested together with acidulated ethanol and several methanol water mixtures ranging from 0% to 100% methanol. The most efficacious solvent system was 50% aqueous methanol, which had the advantage of incorporating both high anthocyanin and flavonol extractability and subsequent stability prior to HPLC analysis. Key words: anthocyanin, flavonol, malvidin-glucoside, quercetin-glucoside, stability, extraction, solvent Extraction of flavonoids from grape berries has gradually evolved to a state where there are separate extraction and analytical techniques for each of the different classes of flavonoids. Anthocyanins are commonly extracted by methods originally applied to the study of flower color in other plant genera. These generally use methanol or ethanol, usually with 1% acid, commonly hydrochloric acid (HCl) (Scott-Moncrieff 1930, Bate-Smith 1948, Harborne 1958). Subsequently, these techniques have been successfully applied to viticultural research (Wulf and Nagel 1978, Metivier et al. 1980, Roggero et al. 1986, Spayd et al. 2002, Downey et al. 2004). Independent extraction methods have also been developed for flavonols, including boiling water, boiling ethanol, acetic acid/methanol, dimethylsulfoxide/methanol, 70% acetone, and 50% petroleum/water (Harborne 1962, Recourt et al. 1992, Lister et al. 1996, Lu et al. 2000, Moriguchi et al. 2002, Vallejo et al. 2004). The methods specifically applied to grapevines include 100% methanol, methanol/hcl, ethanol/hcl, and methanol/water (Cheynier and Rigaud 1986, Price et al. 1995, Hmamouchi et al. 1996, Revilla et al. 1998). Of these, acidulated ethanol is most commonly used in viticultural research (Price et al. 1995, 1 Primary Industries Research Victoria, P.O. Box 905, Mildura, VIC 3502, Australia. *Corresponding author ( mark.downey@dpi.vic.gov.au; fax: ) Acknowledgments: This research was supported by Primary Industries Research Victoria and by the grapegrowers and winemakers of Australia through their investment body the Grape and Wine Research and Development Corporation (GWRDC). Manuscript submitted January 2005; resubmitted November 2006 Copyright 2007 by the American Society for Enology and Viticulture. All rights reserved. Haselgrove et al. 2000, Spayd et al. 2002, Downey et al. 2003b, 2004). In addition, more involved protocols such as supercritical fluid extraction (Palma et al. 2000) and pressurized liquid extraction (Ju et al. 2003) have been developed. These approaches, while academically interesting, are generally inappropriate for processing the large number of samples typically collected during the course of viticultural field trials. As part of investigations into the anthocyanin content and composition of grape skins (Vitis vinifera L.) in response to light exposure (Downey et al. 2004), it became apparent that anthocyanins extracted in methanol/hcl were relatively unstable. The partial hydrolysis of malvidin-3-o-acetylglucoside in solvents containing up to 1% concentration of 12 N HCl has been previously reported (Revilla et al. 1998). While developing a method for simultaneous determination of anthocyanins and flavonols in grape skin extracts, we observed that flavonols were also unstable in methanol/hcl (3 ml concentrated HCl:97 ml methanol) (Harborne 1958). The instability of anthocyanins and flavonols in methanol/hcl was a constraint that required samples to be extracted and run immediately, meaning that anthocyanin and flavonol extracts for highperformance liquid chromatographic (HPLC) analysis could only be prepared immediately before analysis. To increase throughput efficiency, we developed a relatively rapid and simple extraction protocol wherein anthocyanins and flavonols remained stable for at least 12 to 24 hr, allowing a higher number of extracts to be prepared, loaded onto an HPLC autosampler, and analyzed without substantial degradation of the extract. Thus, we sought to identify a stable and efficacious solvent suitable to extract both anthocyanins and flavonols from grape skins for routine analysis of high numbers of grape skin samples by HPLC. 358

2 Stable Extract for Anthocyanins and Flavonols 359 Materials and Methods Plant material. Skins were collected from Vitis vinifera L. cv. Shiraz (syn. Syrah) at commercial harvest (21.8 Brix), frozen in liquid nitrogen, and stored at -80 C until analyzed. Grapes were collected from a commercial vineyard in the Sunraysia region of Victoria, Australia, during the 2002 to 2003 season. Before analysis, grape skins were ground to a fine powder under liquid nitrogen and 0.10 g aliquots were extracted with 1.0 ml of solvent (as described below) and sonicated (Unisonic, Sydney, Australia) for 20 min at room temperature. Solvent extraction of anthocyanins and flavonols. Anthocyanins and flavonols were extracted from grape berry skins with a range of common solvent systems: 1% HCl in methanol (3 ml concn. HCl:97 ml methanol) (Harborne 1958), methanol:water (50:50) (Markham 1982), and 1% formic acid in methanol by volume (1 ml formic acid:99 ml methanol) (adapted from Wulf and Nagel 1978). Because it was postulated that anthocyanin lability was due to acid catalyzed cleavage of the p-coumarate and/or acetate moieties from the 3-O-glucoside of the anthocyanin, several weaker acids were also tested in 100% methanol: glacial acetic acid (1% v/v), citric acid (1% w/v), and maleic acid (1% w/v). Acidulated ethanol (Price et al. 1995) was also evaluated. Methanol, ethanol, acetic acid, formic acid, and concentrated hydrochloric acid were supplied by Merck (Melbourne, Australia). Citric and maleic acids supplied by Sigma (Sydney, Australia). A separate experiment examined the efficacy of different methanol/water mixtures in extracting flavonols and anthocyanins. These mixtures were 0% methanol (100% deionized water), 10%, 20%, 40%, 50%, 60%, and 80% methanol (v/v), with the balance comprised of deionized water, and 100% methanol. Analyses of the stability of anthocyanins and flavonol glycosides in potential extraction solvent systems using methanol with formic, acetic, citric, and maleic acids and acidulated ethanol were conducted as single analytical replicates. Single replicates were used initially to identify gross differences between extraction solvents. Each injection represented an independent extraction in a separate HPLC autosampler vial such that sample integrity was maintained between injections. All analyses of extraction efficacy in mixtures of methanol and water were conducted in triplicate. The stability of commercial anthocyanin and flavonol standards in 50% methanol/water (v/v) was also determined in triplicate. Anthocyanin and flavonol extracts of grape skin were analyzed by reversed-phase HPLC using a HP1100 system (Agilent Technologies, Santa Clara, CA) with a Wakosil analytical column (150 mm 4.6 mm; 3-μm packing; SGE, Melbourne, Australia). The HPLC separation used a binary solvent gradient where solvent A was 10% formic acid (v/v with water) and solvent B was methanol with 10% formic acid (v/v). The gradient conditions were 0 min, 17% solvent B; 15 min, 35% solvent B; 40 min, 37% solvent B; 42 min, 100% solvent B; 44 min, 100% solvent B; 45 min, 17% solvent B; 48 min, 17% solvent B. The column was maintained at 40ºC and the flow rate was 1.0 ml/min. Anthocyanin and flavonol peaks were identified by comparison of their elution order with published separations (Wulf and Nagel 1978, Cheynier and Rigaud 1986) supported by spectral analysis and comparison of elution time and absorbance spectra with commercial standards of malvidin-3- O-glucoside and quercetin-3-o-glucoside (Extrasynthese, Genay, France). Extracts in the different solvents were repeatedly injected over a 24-hr period. Anthocyanin results are presented only for the glucosides of malvidin, as representing the bulk of anthocyanins in the skins of the major winemaking cultivars Shiraz and Cabernet Sauvignon (Mazza 1995). Flavonol results are presented for the major flavonols detected in Shiraz grape skin, quercetin-3-o-glucoside and quercetin-3-o-glucuronide (Cheynier and Rigaud 1986). Data are presented as HPLC chromatogram peak areas over time and expressed as milli-absorbance units (mau) at 520 nm for anthocyanins and at 353 nm for flavonols (Markham 1982). Concentrations were calculated from calibration curves prepared from commercial standards and expressed as malvidin-3-o-glucoside equivalents for anthocyanins and quercetin-3-o-glucoside equivalents for flavonols. Results and Discussion Comparative extractions of the anthocyanins malvidin- 3-O-glucoside, malvidin-3-o-acetylglucoside, and malvidin- 3-O-p-coumaroylglucoside in different solvents over a 24- hr period are shown in Figure 1. Compared with the other solvents, 50% methanol/water (v/v) and 1% HCl in methanol were the most effective extraction solvents with the highest absorbance of anthocyanins at zero hours. However, the anthocyanin composition of the sample extracted in 1% HCl in methanol changed substantially during the 24-hr evaluation (Figure 1). Malvidin-3-O-acetylglucoside decreased rapidly by ~50% in the first 5 hr and had decreased by ~75% at 9.5 hr. During this time, there was a commensurate increase in malvidin-3-o-glucoside in the 1% HCl in methanol extract. Thus, while composition changed, the level of total anthocyanins in methanolic HCl (1%) remained relatively constant over 24 hr. This was consistent with our previous experience and with the observations of Revilla et al. (1998). The content and composition of anthocyanins in the other extraction solvents (formic, acetic, citric, and maleic acids) was relatively unchanged throughout the 24-hr observation period. Generally, these were all effective extraction solvents, although the level of anthocyanins extracted, based on peak area, was lower than that achieved with 50% methanol and 1% HCl in methanol. However, in the formic, acetic, citric, and maleic acid extraction systems where the solvent was predominantly methanol, the level of malvidin-3-o-glucoside extracted was lower than the level of the less polar

3 360 Downey et al. malvidin-3-o-p-coumaroylglucoside (Figure 1). This was the reverse of the observation for 50% methanol and 1% HCl in methanol. Comparative extractions of the flavonols quercetin-3-o-glucoside and quercetin-3-o-glucuronide in different solvents over a 24-hr period are shown in Figure 2. As was observed with the anthocyanins, 50% methanol/water (v/v) was the most efficacious extraction solvent, extracting 30 to 50% more quercetin-3-o-glucoside than any of the other solvents. In contrast, the level of quercetin-3-o-glucuronide extracted in 50% methanol was between 5% and 25% lower than that extracted in most other solvents, with exception of 1% HCl in methanol (Figure 2). The levels of quercetin-3-o-glucoside and quercetin-3-o-glucuronide remained stable in the 50% methanol extract over 24 hr, while there was substantial degradation of quercetin-3-o-glucoside in the 1% HCl in methanol extract (Figure 2). The decrease in quercetin-3-o-glucoside in the 1% HCl in methanol extract was ~40% in the first 5 hr and the level had decreased by ~65% at the end of 24 hr. To our knowledge, this is the first report of the lability of flavonol glucosides in acidified extraction solvent. Historical use of such solvents for flavonol analysis may well have resulted in substantial underestimation of the flavonol content of grape and other plant tissues. The level of quercetin glycosides in the remaining extracts was relatively stable over time. There appeared to be an increase in flavonols extracted in acetic and citric acids, or at least an increase in absorbance at 353 nm over time. It is also interesting to note that in 50% methanol and 1% HCl in methanol extractions, proportionally more quercetin-3-oglucoside than quercetin-3-o-glucuronide was extracted, while the reverse was true of the other solvents (Figure 2). The reason for this is uncertain, but may reflect the differential solubility of quercetin glycosides in solvents that were virtually 100% methanol. However, the close retention times of these two compounds in reversed-phase HPLC does not indicate differences in polarity that might account for the relative differences in extraction between solvent systems. These initial analyses were conducted as single analytical replicates to determine Figure 1 Comparative extraction of anthocyanins in different solvents and stability of extracts over 24 hr: methanol/water (50% v/v); methanol/hcl (1% concn); methanol/formic acid (1% v/v); methanol/acetic acid (1% v/v); methanol/citric acid (1% w/v); and methanol/ maleic acid (1% w/v). Data presented for glucosides of malvidin only. All extracts performed in single replicate except 50% methanol/water (n = 3), standard error <5.0% of mean. Figure 2 Comparative extraction of flavonols in different solvents and stability of extracts over 24 hr: methanol/water (50% v/v); methanol/hcl (1% concn); methanol/formic acid (1% v/v); methanol/acetic acid (1% v/v); methanol/citric acid (1% w/v); and methanol/maleic acid (1% w/v). Data presented for quercetin-3-o-glucuronide and quercetin-3-o-glucoside only. All extracts performed in single replicate except 50% methanol/water (n = 3), standard error <5.0% of mean.

4 Stable Extract for Anthocyanins and Flavonols 361 whether any of these approaches was a viable alternative to 1% HCl in methanol. Since the extraction efficiency of methanolic formic acid (1% v/v), acetic acid (1% v/v), citric acid (1% w/v), and maleic acid (1% w/v) was significantly lower than either methanolic HCl (1% v/v) or 50% methanol/water (v/v), these approaches were not pursued further. Data are included for comparison only. Anthocyanins and flavonols extracted in acidulated ethanol were relatively stable in this solvent. Anthocyanins decreased ~10% over 24 hr, while flavonols decreased ~5% (data not shown), comparable to the stability of anthocyanins and flavonols observed in the other solvent systems examined here (methanolic solutions of formic, acetic, citric, and maleic acids). Extraction efficacy of flavonols in acidulated ethanol was comparable to that of 50% methanol/water (v/v); however, the level of anthocyanins extracted was only ~30% of the amount extracted in 50% methanol. Furthermore, it was observed that ethanolbased extraction solvents were not compatible with the HPLC method employed here, which primarily uses methanol-based solvents, resulting in poor peak resolution particularly of the minor anthocyanin and flavonol components of grape skin (data not shown). Consequently, acidulated ethanol was discarded as a potential extraction solvent for routine HPLC analysis of anthocyanins and flavonols in grape skin. Based on the data here and in the work of others (Revilla et al. 1998), it is clear that 1% HCl in methanol is also an unsuitable solvent for extraction of both anthocyanins and flavonols for HPLC analysis, as content and composition changed substantially over 24 hr. In contrast, 50% methanol/water (v/v) generally extracted more anthocyanins and flavonols than any of the other solvents tested, and these extracts remained stable over the 24-hr period. Having determined that 50% methanol/water (v/v) was a better solvent than the other potential extraction solvents, it remained uncertain whether the 50% mixture was the most efficacious composition compared with other methanol/water mixtures. Thus, several different percentage methanol/water mixtures were assessed. The extraction and stability of anthocyanins in mixtures of methanol and water ranging from 0% methanol (100% deionized water) to 100% methanol are shown in Table 1. Cursory inspection of the means indicates that the level of anthocyanins extracted by 60% methanol tended to be slightly greater than that extracted with 50% methanol, with the effectiveness of all mixtures being ranked (most effective to least effective) 60% > 50% > 40% > 80% > 20% > 10% > 100% > 0% methanol. Overall, the mixture of 20% methanol in water (v/v) had the lowest variability in total anthocyanin concentration as indicated by the lowest coefficient of variation (2.38; n = 15), while extraction with 100% water had the highest coefficient of variation (8.03), closely followed by 100% methanol (7.10). Variation was also relatively high in the 40% methanol/water mixture (6.18), which belies the similarity of the mean values for the 40% mixture and should dissuade others from employing this mixture. The extraction and stability of flavonol glycosides in methanol/water mixtures is presented in Table 2. Of all the methanol/water mixtures, 0% methanol (100% deionized water) and 100% methanol were the least effective extraction solvents. While both extracted similar levels of flavonols, this level was ~35% lower than was extracted in 50% methanol/water (Table 2). The efficacy of 10% methanol as an extraction solvent was also lower than 50% methanol/water, but the remaining methanol/water mixtures (20%, 40%, 50%, 60%, and 80%) were equally effective in extracting flavonol glycosides (quercetin-3-o-glucoside and quercetin-3-o-glucuronide). Extraction of flavonols in 20% percent methanol/water (%CV = 3.63) and 50% methanol/water (3.78) showed the lowest variability of all of the mixtures. The most variation was observed in the extract using 100% methanol (11.62). All extracts exhibited a slight increase in absorbance at 353 nm over time, which cannot reasonably be accredited to an increasing concentration of flavonols. Although the possibility that modification of other flavonol glycosides in the extraction yields quercetin-3-o-glucoside cannot be discounted, it was beyond the scope of this investigation to explore that possibility. Sample carryover could also explain the increase in absorbance over time, but validation Table 1 Comparative extraction of total anthocyanins in different methanol/water mixtures (v/v) and stability of extracts over a 20-hr period: 0% methanol (100% deionized water) and 10%, 20%, 40%, 50%, 60%, 80%, and 100% methanol. Data represents mean values for the sum of malvidin glucosides ± standard error (n = 3). Coefficient of variation (%CV) of all replicates for each methanol/water mixture also shown (n = 15). Total anthocyanins (mg/g fresh wt of skin) Mixture (%) Time 0 5 hr 10 hr 15 hr 20 hr %CV ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

5 362 Downey et al. of the HPLC analytical method and the absence of a trend on a similar scale in the anthocyanin data (Table 1) eliminate that as a possible cause. It is more likely that oxidative reactions occurring among the many phenolic components in these extracts may generate products that absorb light at 353 nm. When the levels of extracted anthocyanins and flavonol glycosides are considered together, the 50% methanol and 60% methanol/water mixtures were the most effective. It has been previously reported that optimum extraction of anthocyanins was achieved with 60% methanol/water (1% HCl) in a pressurized liquid extraction (Ju et al. 2003). While the data presented here tend to support that report, Ju and colleagues compared 60% methanol/water (1% HCl) with 60% ethanol/water (1% HCl), 70% methanol (7% acetic acid), 70% methanol (0.1% TFA), methanol:acetone:water (40:40:20; 1% HCl), and 0.1% HCl in water, but not with other methanol/water mixtures. Results here have demonstrated the extractability and stability of anthocyanin and flavonol glucosides in various mixtures of methanol and water. While it would appear that the two mixtures are equivalent, there is more variability associated with the system using the 60% methanol mixture, as evidenced by the generally larger standard errors and greater coefficient of variation among samples in each mixture (Tables 1 and 2). This greater variability, and the apparently greater efficacy of extraction solvents containing more than 50% methanol observed, are certainly artifacts of the chromatographic methods used here. As the percentage of methanol in the extraction solvent increased above 50%, peak resolution decreased. For example, in extraction mixtures of 50% methanol and less (40%, 20%, 0%), baseline separation of malvidin and peonidin glucosides was possible. Above 50% methanol in the extraction solvent, it was no longer possible to differentiate assuredly between these peaks and there is a high likelihood that peak areas for malvidin glucosides are artificially inflated by the coelution of peonidin glucosides. Stability evaluation of a single replicate of the 50% methanol mixture was continued, beyond 20 hr, at intervals, for over 40 hr (Table 3). For both anthocyanins and flavonols the totals remained relatively constant over this longer time, which is encouraging if, for whatever reason, there is a delay between extraction and analysis. As we have observed, the absorbance at 353 nm slowly increases over time, resulting in what appears to be a slight increase in total flavonols. Overall this increase from the replicate with the lowest reading at time zero ( mau) to the highest reading at 40.1 hr ( mau) was ~18.5%. Thus, the accuracy required from the flavonol analysis should be considered before leaving samples for lengthy periods between extraction and analysis. As with total flavonols, there was also a slight increase in absorbance of anthocyanins (520 nm) over time. However, unlike the increase at 353 nm, from the lowest reading at time zero ( mau) to the highest reading at 41.2 hr ( mau) represents an increase in absorbance at 520 nm of only 5.2%. The levels of anthocyanins reported in these investigations were lower than previously observed in Shiraz grapes (Downey et al. 2004), while the levels of flavonols glycosides were generally higher than earlier reports (Haselgrove et al. 2000, Downey et al. 2003b). These observations Table 3 Stability of total anthocyanins and total flavonols over 40 hr. Values for 0.0, 5.1, 10.1, 15.2, and 20.3 hr performed in triplicate and presented as mean ± SE (n = 3). Subsequent values to 41.2 hr represent a single analytical replicate. Total anthocyanins Total flavonols Time (hr) (mg/g fwt skin) (mg/g fwt skin) ± ± ± ± ± ± ± ± ± ± Table 2 Comparative extraction of flavonols in different methanol/water mixtures (v/v) and stability of extracts over a 20-hr period: 0% methanol (100% deionized water) and 10%, 20%, 40%, 50%, 60%, 80%, and 100% methanol. Data represents mean values for the sum of quercetin glucosides ± standard error (n = 3). Coefficient of variation (%CV) of all replicates for each methanol/water mixture also shown (n = 15). Flavonols (mg/g fresh wt of skin) Mixture (%) Time 0 5 hr 10 hr 15 hr 20 hr %CV ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

6 Stable Extract for Anthocyanins and Flavonols 363 Table 4 Stability of commercial standards of malvidin-3-oglucoside and quercetin-3-o-glucoside. Values expressed as mean absorbance represented by integrated peak area (mau) ± standard error (n = 3) for 0.25 mg/ml standard solutions in 50% methanol/water. Relative stability of the standard solutions is demonstrated by the percent increase in integrated peak area (mau) relative to peak area at time zero for each interval. Time (hr) Mean ± SE % increase Malvidin-3-O-glucoside ± a ± ± ± ± ± Quercetin-3-O-glucoside ± 10.6 b ± ± ± ± ± a Integrated peak area (520 nm). b Integrated peak area (353 nm). are consistent with the fruit in these investigations being from warm-irrigated vineyards (mean January temp.: 23.3 C) (Gladstones 1992). High temperatures have been previously reported to decrease the anthocyanin content of grape berries (Bergqvist et al. 2001, Downey et al. 2004), while highly exposed fruit have reputedly higher levels of flavonols consistent with the role of flavonols as UV-protectants (Smith and Markham 1998, Downey et al. 2003b). In addition to the stability of anthocyanins and flavonols present in grape skin extracts, the stability of commercial standards (Extrasynthese, Genay, France) of both malvidin-3-o-glucoside and quercetin-3-o-glucoside was examined (Table 4). Malvidin-3-O-glucoside and quercetin- 3-O-glucoside were dissolved in 50% methanol/water (v/v) to a concentration of 0.25 mg/ml and injected at regular intervals over 30 hr. The integrated peak areas of malvidin-3-o-glucoside (520 nm) and quercetin-3-o-glucoside (353 nm) (presented as the mean of three replicates) increased over time. Malvidin-3-O-glucoside absorbance (520 nm) increased by ~6% over 30 hr, while absorbance at 353 nm showed and apparent increase in quercetin-3-o-glucoside of 6.6% over the same period. Given that sample carryover was not observed in blank runs between standards, increased absorbance in the standard solutions is likely due to the oxidation processes postulated to account for increasing absorbance of grape skin extracts at 520 nm and 353 nm (Tables 1 and 2). The proportion of total anthocyanins and flavonols extracted by 50% methanol/water from grape skin was also tested (Table 5). A 0.10 g aliquot of grape skin was reextracted twice in 1.0 ml 50% methanol/water (v/v) and the anthocyanin and flavonol content and composition determined by HPLC. A second extraction yielded less than 10% of the amount of anthocyanins and flavonols in the initial extraction, while the third extraction yielded ~0.5% of the initial extraction. The applicability of this extraction protocol to other classes of grape phenolics is likely to be of interest to researchers and industry alike, as multiple analyses from a single sample preparation is informative. Because the HPLC method for analysis of anthocyanins and flavonols described here was not appropriate for separating flavan- 3-ols, proanthocyanidins, or hydroxycinnamates, the extractability of these classes of phenolics was not examined in this study. However, based on their low solubility in water (according to the Merck Index), a mixture of 50% methanol and water is likely to prove an efficacious solvent for extraction of hydroxycinnamates before HPLC analysis and is the solvent used in this laboratory for such work (Nagel 1985). Depending on the nature of the chromatographic conditions used to analyze hydroxycinnamates, lower concentrations of methanol might be more appropriate for maintaining peak resolution. As regards flavan-3-ols and proanthocyanidins, a survey of several different solvents (Kallithraka et al. 1995) showed that methanolic extracts yielded proportionally more monomers than oligomers or polymers, while ethanol tended to extract gallic acid preferentially. Thus, a mixture of 50% methanol and water would likely prove a useful extraction solvent for flavan-3-ol analysis. Aqueous methanol (50%) has previously been used in proanthocyanidin analysis (Constantinides and Fownes 1994); however; aqueous acetone was a more effective extraction solvent (Cork and Krockenberger 1991). Mixtures ranging from 50 to 80% have been widely reported, with the most frequent composition being 70% acetone. This solvent has been used extensively in the analysis of grape proanthocyanidins (Kallithraka et al. 1995, Souquet et al. 1996, Kennedy et al. 2001, Downey et al. 2003a). Table 5 Efficacy of methanol/water (50% v/v) as an extraction solvent showing percent yield of individual malvidin and quercetin glycosides in second and third extractions relative to the initial extraction as determined by HPLC. Malvidin-3- Quercetin-3- glucoside acetylglucoside coumaroylglucoside glucuronide glucoside 2 nd extraction 7.70 a a rd extraction a % of initial extraction.

7 364 Downey et al. Conclusion Results have shown that a 50% mixture of methanol and water (v/v) is the most efficacious solvent for the extraction of anthocyanins and flavonols from grape skin for subsequent HPLC analysis. Furthermore, anthocyanins and flavonols extracted in this solvent remained stable for more than 20 hours. Literature Cited Bate-Smith, E.C Paper chromatography of anthocyanins and related substances in petal extracts. Nature 161: Bergqvist, J., N. Dokoozlian, and N. Ebisuda Sunlight exposure and temperature effects on berry growth and composition of Cabernet Sauvignon and Grenache in the central San Joaquin Valley of California. Am. J. Enol. Vitic. 52:1-7. Cheynier, V., and J. Rigaud HPLC separation and characterization of flavonols in the skins of Vitis vinifera var. Cinsault. Am. J. Enol. Vitic. 37: Constantinides, M., and J.H. Fownes Tissue-to-solvent ratio and other factors affecting determination of soluble polyphenols in tropical leaves. Commun. Soil Sci. Plant Anal. 25: Cork, S.J., and A.K. Krockenberger Methods and pitfalls of extracting tannins and other phenolics from plants: Insights from investigations on Eucalyptus leaves. J. Chem. Ecol. 17: Downey, M.O., J.S. Harvey, and S.P. Robinson. 2003a. Analysis of tannins in seeds and skins of Shiraz grapes throughout berry development. Aust. J. Grape Wine Res. 9: Downey, M.O., J.S. Harvey, and S.P. Robinson. 2003b. Synthesis of flavonols and expression of flavonol synthase genes in developing grape berries of Shiraz and Chardonnay (Vitis vinifera L.). Aust. J. Grape Wine Res. 9: Downey, M.O., J.S. Harvey, and S.P. Robinson The effect of bunch shading on berry development and flavonoid accumulation in Shiraz grapes. Aust. J. Grape Wine Res. 10: Gladstones, J. (Ed.) Viticulture and Environment. Winetitles, Adelaide. Harborne, J.B Spectral methods of characterizing anthocyanins. Biochem. J. 70: Harborne, J.B Plant polyphenols. VI. The flavonol glycosides of wild and cultivated potatoes. Biochem. J. 84: Haselgrove, L., D. Botting, R. van Heeswijck, P.B. Høj, P.R. Dry, C. Ford, and P.G. Iland Canopy microclimate and berry composition: The effect of bunch exposure on the phenolic composition of Vitis vinifera L cv. Shiraz grape berries. Aust. J. Grape Wine Res. 6: Hmamouchi, M., N. Es-Safi, M. Lahrichi, A. Fruchier, and E.M. Essassi Flavones and flavonols in leaves of some Moroccan Vitis vinifera cultivars. Am. J. Enol. Vitic. 47: Ju, Z.Y., and L.R. Howard Effects of solvent and temperature on pressurized liquid extraction of anthocyanins from dried red grape skin. J. Agric. Food Chem. 51: Kallithraka, S., C. Garcia-Viguera, P. Bridle, and J. Bakker Survey of solvents for the extractions of grape seed phenolics. Phytochem. Anal. 6: Kennedy, J.A., Y. Hayasaka, S. Vidal, E.J. Waters, and G.P. Jones Composition of grape skin proanthocyanidins at different stages of berry development. J. Agric. Food Chem. 49: Lister, C.E., J.E. Lancaster, and J.R.L. Walker Phenylalanine ammonia-lyase (PAL) activity and its relationship to anthocyanin and flavonoid levels in New Zealand-grown apple cultivars. J. Am. Soc. Hortic. Sci. 121: Lu, Y., Y. Sun, L.Y. Foo, W.C. McNabb, and A.L. Molan Phenolic glycosides of forage legume Onobrychis viciifolia. Phytochemistry 55: Markham, K.R Techniques of Flavonoid Identification. Academic Press, London. Mazza, G Anthocyanins in grapes and grape products. CRC Crit. Rev. Food Sci. Nutr. 35: Metivier, R.P., F.J. Francis, and F.M. Clydesdale Solvent extraction of anthocyanins from wine pomace. J. Food Sci. 45: Moriguchi, T., M. Kita, K. Ogawa, Y. Tomono, T. Endo, and M. Omura Flavonol synthase gene expression during citrus fruit development. Physiol. Plant. 114: Nagel, C.W Application of high performance liquid chromatography to analysis of flavonoids and phenyl propenoids. Cereal Chem. 62: Palma, M., T. Taylor, B.W. Zoecklein, and L.S. Douglas Superficial fluid extraction of grape glycosides. J. Agric. Food Chem. 48: Price, S.F., P.J. Breen, M. Valladao, and B.T. Watson Cluster sun exposure and quercetin in Pinot noir grapes and wine. Am. J. Enol. Vitic. 46: Recourt, K., M. Verkerke, J. Schripsema, A.A.N. van Brussel, B.J.J. Lugtenberg, and J.W. Kijne Major flavonoids in uninoculated and inoculated roots of Vicia sativa subsp. nigra are four conjugates of the nodulation gene inhibitor kaempferol. Plant Mol. Biol. 18: Revilla, E., J.M. Ryan, and G. Martin-Ortega Comparison of several procedures used for the extraction of anthocyanins from red grapes. J. Agric. Food Chem. 46: Roggero, J.P., S. Coen, and B. Ragonnet High performance liquid chromatography survey on changes in pigment content in ripening grapes of Syrah. An approach to anthocyanin metabolism. Am. J. Enol. Vitic. 37: Scott-Moncrieff, R Natural anthocyanin pigments. II. The magenta flower pigment of Primula polyanthus. Biochem. J. 24: Smith, G.J., and K.R. Markham Tautomerism of flavonol glucosides: Relevance to plant UV protection and flower colour. Photochem. Photobiol. A 118: Souquet, J.M., V. Cheynier, F. Brossaud, and M. Moutounet Polymeric proanthocyanidins from grape skins. Phytochemistry 43: Spayd, S.E., J.M. Tarara, D.L. Mee, and J.C. Ferguson Separation of sunlight and temperature effects on the composition of Vitis vinifera cv. Merlot berries. Am. J. Enol. Vitic. 53: Vallejo, F., F.A. Tomás-Barberán, and F. Ferreres Characterisation of flavonols in broccoli (Brassica oleracea L. var. italica) by liquid chromatography-uv diode-array detectionelectrospray ionisation mass spectrometry. J. Chromatogr., A 1054: Wulf, L.W., and C.W. Nagel High-pressure liquid chromatographic separation of anthocyanins of Vitis vinifera. Am. J. Enol. Vitic. 29:42-49.