PRESSURIZED SOLVENT EXTRACTION OF FLAVONOIDS FROM GRAPE POMACE UTILIZING ORGANIC ACID ADDITIVES

Headnote

ABSTRACT

Flavonoids present in freeze-dried Sunbelt (Vitis labrusca L.) grape pomace were extracted at a pressure of 10.3 MPa and temperatures between 60° and 140°C using pressurized hydroethanolic solvents (0, 40, 80% v/v ethanol in water) acidified with organic acids (formic, acetic, citric and tartaric acid at a pH of 2.5). Acetic acid had a significant effect in maximizing anthocyanin yield from grape pomace when using 80% aqueous ethanol as a solvent at 800C (1028.9 mg/100 g DW). However, organic acids did not have a significant effect on the extraction of flavonols from grape pomace. The maximum amount of flavonols extracted from grape pomace using 80% aqueous ethanol as the solvent at 120°- 1400C was 9.65 mg/100 g DW. Surface response regression analysis of the experimental data indicated the optimum temperature for extraction of a maximum amount of anthocyanins from grape pomace was 85.4°C and that for flavonols was 124°C. These optimized conditions obtained from this study performed on an analytical scale can be used in scale-up and design of a dynamic subcriticai fluid extraction system for extraction of flavonoids from food processing waste, such as grape pomace.

- Key words: anthocyanin, flavonoid, grape pomace, pressurized solvent extraction -

(ProQuest: ... denotes formulae omitted.)

INTRODUCTION

The grape pomace consists of the skins, stems and seeds that are left behind after the pressing of the grapes (MOURE et al, 2001). Studies have estimated that the about 20 wt% of the processed grapes are left behind as pomace (SCHIEBER et al, 2001) with approximately 60-70 wt% of total phenolics concentrated in grape seeds, and the rest in the grape skins (GARCIA-MARINO et al, 2006). The localization of such phenolic compounds may vary with the type of grapes, climate and other environmental conditions as well as the maturity of the grape during cultivation (PINELO et al, 2006). Sunbelt grape (Vitis labrusca L), a juice grape variety designed for growth in the State of Arkansas, is a type of juice grape that contains a large amount of anthocyanin diglucosides, unlike wine grapes which contain larger amounts of anthocyanin monoglucosides (REVILLA et al, 1999).

Flavonoids are a group of polyphenolic compounds that are present in many natural products and have been studied extensively due to their antioxidant, anti-microbial and antiproliferative activities (HARBORNE and WILLIAMS, 2000; LURTON, 2003; BOZIDAR, 1995; HALPERN et al, 1998). Studies have indicated that a greater amount of anthocyanins, flavonols (quercetin) and flavan-3-ols (catechin/epicatechin) are concentrated in the grape skins while the grape seeds contain greater amounts of procyanidins (BRAIDOT et al, 2008). Apart from their health beneficial and nutraceutical value, anthocyanins are also used as food colorants (FRANCIS and MARKAKIS, 1989). Fundamental studies have indicated that at different pH levels, the anthocyanins undergo a structural transformation that is responsible for their color change in solution (HEREDIA et al, 1998). The most stable form for anthocyanin in grape pomace is the flavylium cation which exists at pH less than 4 (CLIFFORD, 2000). Studies have indicated that with an increase in pH, there is rapid decrease in the color and an increase in its antioxidant capacity (LAPIDOT et al, 1998).

SPIGNO and DE FAVERI (2007) and other research groups around the world have proposed the use of agricultural wastes such as grape pomace as a low-cost source for the extraction of antioxidant compounds. Studies have indicated that about 10.5-13.1 million tonnes of grape pomace is produced in the world each year (DE CASTRO et al, 2008) Extraction of flavonoids from grape pomace has been traditionally done using alcoholic solvents such as methanol and ethanol with water as a cosolvent (LAPORNIK et al, 2005; METIVIER et al, 1980).

Usually, a low concentration of acid such as hydrochloric acid (HCl) is added to the solvent to increase the efficiency of the extraction to yield maximum amount of anthocyanins (JU and HOWARD, 2003). The addition of the acid denatures the cellular membranes present in the grape pomace matrix, thereby facilitating flavonoid dissolution in the extraction solvent (JACKMAN et al, 1987). The concentration of the acid in the extraction solvent should be very low since the addition of excess acid can cleave the acylated and sugar groups of the anthocyanins during the extraction and concentration steps, resulting in formation of undesirable components, REVILLA et al (1998) have previously reported on the partial hydrolysis of the aceylated malvidin glucosides when 1% concentration (by weight) of 12N HCl was added to methanol used as a solvent. It has been proposed that weaker organic acids such as formic, acetic, citric and tartaric acids or extremely low concentrations of stronger acids such as HCl and trifluoroacetic acid can be used to solve this problem (DE CASTRO et al, 2008; JU and HOWARD, 2003). Studies performed by BLEVE et al (2008) indicated that the flavonoids extracted from grape pomace using hydroethanolic solvents acidifed with 0.2% trifluoroacetic acid were easily available for further purification compared to those extracted in the absence of the acid. However, studies performed by DOWNEY et al (2007) on the stability of the anthocyanins and flavonols extracted from grape pomace using methanol acidified with HCl, acetic, formic, citric and maleic acids indicated that the flavonoid content decreased rapidly with increasing temperature or residence time of the pomace. Moreover, in the afore-mentioned studies, the total contact time between the solvent and the pomace matrix was very high (~ 1 day or more) .

A number of studies have been performed on the extraction of flavonoids from grape pomace using hydroethanolic solvents. SAVOVA et al (2007) indicated that the maximum amount of flavan-3-ols could be extracted from grapes using 50% ethanol in water at 65°C. Another study on the extraction of phenolic compounds from winery wastes indicated maximum recovery of these phenolic compounds using 50% ethanol in water (LAFKA et al, 2007). Recently, similar studies performed by MONRAD et al (2010a,b) indicated that 50-70% v/v ethanol in water at temperatures between 80° and 1200C was optimal for the recovery of anthocyanins and procyanidins from grape pomace.

Water at temperatures above its boiling point and under sufficient pressure (also referred to as "subcriticai water") has been used as a solvent in recovery of polar compounds from natural products (KING and SRINIVAS, 2009). The enhanced dissolution and diffusion rates of such compounds in water at elevated temperatures is due to a decrease in the hydrogen bonding propensity of water as measured by the total solubility parameter as a function of temperature (SRINIVAS et al, 2009), Subcriticai water at temperatures between 100° and 1500C have been found optimal for recovering maximum amount of anthocyanins from berry substrates (KING et al, 2003), grape skins (JU and HOWARD, 2006) and black currants (CACACE and MAZZA, 2002). However, at these temperatures there is a greater possibility for thermal degradation of the flavonoid compounds during their extraction from grape pomace. However, by optimizing the extraction time, it is possible to minimize the loss of flavonoid recovery due to thermal degradation (MONRAD et al, 2010a). Pressurized liquid extractions can be performed using an accelerated solvent extraction (or ASE). Using this technique, extractions can be performed at higher temperatures (40°-200°C) abetted by pressure using very short extraction times (1-20 min) (PALMA et al, 2001; JU and HOWARD, 2003). Such rapid rate extraction conditions can improve recovery of the polyphenols solutes thereby avoiding thermal degradation of the solutes.

Studies performed on black carrots using water acidified with citric and lactic acids at temperatures between 50° and 1000C and 50 bar pressure yielded the maximum amount of anthocyanins, particularly when lactic acid was used as the acidulant (GIZIR et al, 2008). Greater thermal degradation of the anthocyanin extracts occurred at 1000C along with higher concomitant browning when sulfuric acid was used as an acidulant. In this study, the high temperature extractions were performed in 10 minutes, whereas the sulfuric acid extractions were performed at room temperature over a period of 12 hours.

In this study, an ASE system was used to investigate the effects of temperature, solvent composition and the presence of various organic acids on the extraction of flavonoids (anthocyanins and flavonols) from Sunbelt grape pomace.

EXPERIMENTAL SECTION

Grape and pomace samples

The materials used in this study are similar to those described previously (MONRAD et al 2010a,b). Sunbelt grapes harvested at the University of Arkansas' Agricultural Experimental Station Farm (Fayetteville, AR) in 2006 were crushed and de-stemmed followed by pressing in a 70 L Enrossi bladder press (Enoagricola Rossi s.r.l., Calzolaro, Italy). The recovered pomace was then freeze-dried in ViiTis Genesis freeze-dryer (Gardner, NY) and stored at -700C in a Thermo Scientific Ultra-Low Freezer (Waltham, MA). The freeze-dried pomace was then ground to a homogeneous powder (500 µm) using an Udy Cyclone Sample Mill (Fort Collins, CO) for extraction purposes.

Reagents

Formic acid (CAS# 64-18-6; Lot# K37957664; 98% ACS grade), glacial acetic acid (CAS# 6419-7; Lot# 200733306; ACS grade), citric acid (CAS# 77-92-9; Lot# 10132195; Stock# A10395) as well as L-(+)-tartaric acid (CAS# 87-69-4; Lot# J25S046; Stock# L12804) were obtained from VWR Corporation (Batavia, IL). Absolute ethanol (CAS# 64-17-5; 200 proof; ACS grade) was also obtained from VWR Corporation. HPLC grade methanol was obtained from EMD Chemicals (Gibbstown, NJ). Water (1-5 ppb TOC, 18.2 ?Oat? was purified using a Milli-Q Synthesis AlO system (Millipore, Ballerica, MA, USA).

The anthocyanin HPLC standards used in this study consisted of a mixture of malvidin, delphinidin, peonidin, cyanidin, petunidin and pelargonidin monoglucosides which were purchased from Polyphenol Labs (Sandes, Norway). The flavonols were expressed as quercetin-3-βD-glucoside (CAS# 21637-25-2; Lot# 1373210; Filling code: 11908159; > 90% HPLC) equivalents. The quercetin-3-β-D-glucoside was obtained from Sigma-Aldrich (St. Louis, MO). The HPLC-grade solvents used in the analysis were acquired from VWR (Batavia, IL).

Pressurized acidified solvent extraction of anthocyanins from grape pomace

The extractions of flavonoids from grape pomace were performed using a Dionex ASE-200 extractor system (Dionex Corp., Sunnyvale, CA). 0.5 g of freeze-dried grape pomace was placed in an 1 1 mL extractor cell with a cellulose-loaded filter placed at the bottom of the cell. Solvents consisted of mixtures of ethanol in water (0, 40 and 80% v/v). The solvents were acidified to a pH of 2.5 using the following organic acids: formic, acetic, citric and tartaric acids. The acidification of the solvent was done by adding the organic acids to the solvent while measuring the pH using a Mettler Toledo SevenEasy S-20 pH meter fitted with an InLab Expert Pro "3-in1" pH electrode (Columbus, OH). The initial pH of 0, 40 and 80% ethanol (v/v) solutions in water were found to be 6.85 (±0.5), 5.35 (±0.5) and 4.86 (±0.5) units, respectively.

In the ASE system, heating of the solvent is performed using an aluminum block heater placed in the ASE system. The solvent is allowed to flow into the cell and the cell heated to the set point temperature value within 5-7 minutes. The oven temperature was set at 60°, 80°, 100°, 120° and 1400C for the subsequent extractions. The static extraction time after equilibration to the set experimental time was set at 1 min. An external pressure of 10.3 MPa was applied to the extractor cell during the static extraction period. After the static extraction time, the extract was allowed to flow into a collection vial and the cell flushed using a quantity of solvent equivalent to 70% of the cell's volume, followed by purging the contents of the extraction cell with nitrogen for 110 seconds.

After extraction, the volumes were adjusted to 50 mL with water and centrifuged for 10 minutes at 8,000 rpm using a Beckman GS- 15R centrifuge (Calbiochem, San Diego, CA) to remove any insoluble solids. The centrifuged extracts were then mixed with methanol in 1 : 1 (v/v) ratio and analyzed by HPLC. Triplicate measurements were made.

Conventional extraction

Conventional extraction of anthocyanins and flavonols from grape pomace, as described by HAGER et al (2008), was used to determine the efficiency of the ASE extractions. This method was also used by MONRAD et al. (2010a). In this method, 2 g of ground pomace were homogenized with 20 mL of methanol/water/ formic acid (60:37:3, v/v/v) using an IkaT18 Ultra-Turrax tissuemizer (Wilmington, NC) for 30 s at room temperature conditions. The homogenate was filtered through Miracloth (Calbiochem, San Diego, CA) and the filtrate collected. The residue was collected, and the procedure repeated two more times. The filtrates were combined and adjusted to 100 mL with extraction solvent and centrifuged to remove insoluble solids. The supernatant was collected for analysis and stored at -200C prior to analysis. The conventional extractions were also performed in triplicate.

HPLC analysis

HPLC analysis of the anthocyanins and flavonols in the grape extracts was performed using the method described by CHO et al (2004). A Waters Alliance Model 2695 HPLC system (Waters Corp., MiIf ord, MA) was used. 5% formic acid in water (v/v) (A) and methanol (B) were used as solvents in this HPLC method The HPLC gradient program was executed as follows: 98% A to 40% A in 60 min, 40% A to 98% A in 65 min followed by a 10 min re-equilibration period. The entire HPLC run time was 80 minutes using a flow rate of 1 mL/min.

The resultant grape extracts were filtered before HPLC assay using a 0.5 µm PTFE filter (Varían Inc., Palo Alto, CA), and 100 µL of the filtered extract was injected onto a 250 ? 4.6 mm Waters Symmetry C18 column (Waters Corp., Milford, MA), The anthocyanins were detected using a Waters 2996 Photodiode array detector (Waters Corp., MiIf ord, MA) at a 510 nm, and the peaks were compared with the retention times of a standard grape pomace extract analyzed by HPLC-MS. Calibration curves were obtained using a mixture of anthocyanin monoglucosides. The anthocyanins in the extracts were reported based on these monoglucoside equivalents. The amount of anthocyanins extracted from grape pomace is expressed in terms of mg/ 100 g (of dry weight pomace).

The flavonols extracted from grape pomace were detected at a 364 nm and the peaks were quantified as quercetin-ß-D-glucoside equivalents. The amount of flavonols is also expressed in terms of mg/ 100 g (of dry weight pomace) .

HPLC-MS analysis of anthocyanins in grape pomace

The HPLC-MS analysis of the anthocyanins present in a grape pomace extract was done using the method described by CHO et al (2004). The grape extract was injected into the Waters Symmetry C18 column connected to a Hewlett Packard 110 series HPLC (Agilent Technologies, Wilmington, DE) with an auto sampler, binary HPLC pump, and an UV/Vis detector. The solvents and the grathent program were set as stated above. The anthocyanin peaks were detected at 510 nm and flavonol peaks were detected at 364 nm. The mass spectrophotometric analysis was done using a Bruker Esquire LC-MS (Billerica, MA) ion trap mass spectrometer connected to the Hewlett-Packard 1 100 system. The eluting peaks showing absorptivity at 520 nm were analyzed using the LC-MS system and its attendant software. The mass spectrophotometer parameters set for the anthocyanin analysis were as follows: positive ion elee trospray mode with a capillary voltage of 4 kV; nebulizing pressure of 0.21 MPa; drying gas flow rate of 0.9 mL/min; temperature of 3000C. The flavonols were analyzed in the mass spectrophotometer under same conditions as discussed above but in a negative ion elee trospray mode. The data were collected over the mass range of m/z 50 through 1,000 in full scan mode at 1.0 s/cycle.

Experimental design for statistical analysis

The experimental design was a full factorial (3x5x5) treatment completely randomized design with three replications. The statistical design consists of three solvents (0, 40 and 80% (v/v) ethanol in water), five acid conditions (no acid, formic, acetic, citric, and tartaric acids) and five temperatures (60°, 80°, 100°, 120° and 1400C). The linear statistical model (MONRAD et al, 2010a,b) used for the analysis is as given below:

Y^sub ijkl^= μ + α^sub i^ + β, + γ^sub k^ + (αβ)^sub ij^ + + (βγ)^sub jk^ + (αγ)^sub ik^ + (αβ)^sub ijk^ + ε^sub ijkl^ (1)

where Y^sub ijkl^ is the observed measured response of the 1th replication of the Ith solvent with the jth acid condition on the kth temperature and µ is the overall population average response, α is the 1th solvent main effect (...), β^sub j^ is the 1th acid main effect (...), γ^sub k^is the kth temperature main effect (...), αβ^sub ij^ is the solvent-ac id interaction effect (...), βγ^sub jk^ is the acidtemperature interaction effect (...), αγ^sub ik^ is the solvent-temperature interaction effect (...), αβγ^sub ijk^ is the solvent-acid-temperature interaction effect (...) and ε^sub ijk^ N(O, σ^sup 2^) is the unobserved ijk^ random error effect. The errors are assumed to be independent, identically, and normally distributed with mean zero and common variance a2. The two-way ANOVA representing the aforementioned statistical design is calculated using JMP software (ver. 8; Cary, NC). The Student's t-test was used to analyze the significance of the main and interaction effects evaluated at 95% confidence interval (P<0.05).

In order to assess the variability in the extraction of anthocyanins and flavonols from grape pomace using the pressurized acidified solvents, a surface response regression analysis was used. This method is based on the principle that in a three-component system, if two of the factors can be specified, than the third factor affecting the optimization of the anthocyanin yield can be fixed (TOYOMIZU et al, 1982). The general model for a linear regression equation is given below:

Y = β^sub 0^ + β^sub 1^X^sub 1^+ β^sub 2^ + β^sub 3^β^sub 3^ + β^sub 11^X^sup 2^^sub 1^ + + β^sub 22^X^sup 2^^sub 2^ + β^sup 2^^sub 3^ + β^sub 12^X^sub 1^X^sub 2^ + β^sub 23^X^sub 2^X^sub 3^ + β^sub 13^X^sub 1^X^sub 3^ (2)

where, β^sub 1,6^ are parameter estimates for the best regression equation that yield maximum amount of anthocyanins at the optimized conditions, X1 refers to solvent composition (0.0, 0.4 and 0.8); X2 refers to temperature (60°, 80°, 100°, 120° and 1400C); and X3 refers to acid (1- no acid, 2formic acid, 3- citric acid, 4- tartaric acid, and 5- acetic acid). 'Y' refers to the derived response for the particular model. In this model, X12, X23 and X^sub 13^ refers to the effect due to interactions between the different factors.

RESULTS AND DISCUSSION

Pressurized liquid extraction of grape pomace by ASE uses pressures ranging between 3.45 MPa and 20.68 MPa. Studies have indicated that pressures sufficient to maintain the solvent in its liquid state at temperatures above its boiling point (i.e. equal to or above its vapor pressure at a particular temperature) are required for performing such pressurized liquid extractions (SHARIF et al, 2010). In case of subcriticai water, this pressure requirement would be < 2 MPa at temperatures between 100° and 1500C. Studies performed by LUTHRIA (2008) and MUKHOPADHYAY et al (2006) have indicated that pressure does not have a significant effect on the extraction of phenolic compounds from natural products. Hence, in this study, a constant pressure of 10.3 MPa was used for extracting anthocyanins and flavonols from grape pomace using the described subcriticai fluids,

A typical chromatogram obtained from the HPLC analysis of the anthocyanins and flavonols present in the Sunbelt grape pomace is presented in Figs. Ia and b. The anthocyanin peaks are numbered from 1-12 and flavonol peaks are numbered from 1-4 as detected using the HPLCMS system. Mass spectrophotometric data are given in Table 1 .

Peaks 2 [Malvidin-3, 5-O-diglucoside] and 8 [Peonidin-3-(6-0-coumaroyl)-5-0-diglucoside + Malvidin-3-(6-0-p-coumaroyl)-5-0-diglucoside] are the most prominent anthocyanins present in Sunbelt grape pomace extract (Fig. Ia). An elaborate discussion on the identification of the anthocyanins in Sunbelt grape pomace can be found in our previous work (MONRAD et al, 20 1 Oa) . It can also be seen that the anthocyanins can form acyl derivatives with acetic acid and coumaric acids with the coumaroyl derivatives predominating in Sunbelt grape pomace. Studies have indicated that the anthocyanin acyl derivatives can be by-products of hydrolysis of the anthocyanins due to the presence of acylating acids which include ferulic, caffeic, malic, succinic and tartaric acids (ANDERSON et al, 1970).

The HPLC profile for the flavonols detected at 364 nm in Sunbelt grape pomace is shown in Fig. Ib. There is no data in the literature on the mass spectrophotometric analysis of flavonols in Sunbelt grape pomace, but the identification of flavonols in Sunbelt grape pomace was comparable with that reported for various grape genotypes (CHO et al 2004). Although, a number of peaks were detected by HPLC-MS, only 4 major peaks: myricetin-3-glucoside, quercetin-3glucuronide, quercetin-3-glucoside and quercetin-3-rutinoside were positively identified. It should also be noted that quercetin-3-glucoside and quercetin-3-galactoside have identical m/z values and co-eluted using this method. Even though quercetin-3-glucuronide and quercetin3-glucoside/galactoside have similar retention times, their mass spectra vary slightly (Table 1). Studies performed by SOUQUET et al (2000) indicate that quercetin-3-glucuronide is a major phenolic component in grape stems and is easily assimilated by the human body compared to corresponding glucosides/galactosides. In order to avoid any confusion, the total amount of flavonols discussed in this study refers to the sum of the 4 identified peaks and the unidentified flavonols have not been taken into account.

Solvent and temperature optimization

Optimization of the solvent and temperature conditions to maximize anthocyanin yield from grape pomace using subcriticai fluid extraction is based on the sum of all the anthocyanins detected in the grape extracts as previously discussed by MONRAD et al (2010a). The statistical analysis performed using the general linear model given in Equation 1 indicated that there was a significant effect of the solvent (or ethanol composition in water) (P<0.001) and temperature (P=O. 0457) on the amount of total anthocyanins extracted from the grape pomace. The data, as previously discussed, were analyzed with a 95% confidence interval. There was, however, an insignificant effect of the interaction between the solvent and the temperature (P=O. 35 19) on the anthocyanin yield.

A simple Student's t-test at the 95% confidence level indicated that 80% (v/v) ethanol in water was marginally better in extracting maximum number of anthocyanins from grape pomace when compared to 40% (v/v) aqueous ethanol. This hydroethanolic solvent mixture, however, extracted a significantly greater amount of anthocyanins from grape pomace relative to using neat water as a solvent. Similar trends were noted by MONRAD et al (2010a), i.e., significant amount of anthocyanins were extracted using 50 - 70% aqueous ethanol as solvent.

The total amount of anthocyanins extracted by hydroethanolic solvents from Sunbelt grape pomace are shown in Table 2. The standard deviations for the anthocyanin data as presented in Table 2 were calculated using Equation 3.

... (3)

where C^sub g,l^ = amount of total anthocyanins (mg/ 100 g DW) extracted from Sunbelt grape pomace for ith replicate at a particular temperature; = mean of the total anthocyanin (mg/ 100 g DW) extracted from sunbelt grape pomace at a particular temperature for all replicates and; n = number of replicates.

It can be seen from Table 2 that peaks 4 (Pnd3-O-monoglucoside + Ptd-3-0-(6-acetyl)-5-0-diglucoside) and 10 (Cyd-3-0-(6-0-p-coumaroyl) monoglucoside), though identified through HPLC-MS, were not found in the ASE extracts except when acetic acid was added to the hydroethanolic solvent. This can be due to the acylation of the anthocyanins in presence of the acidified solvents or the reduction in polarity of solvent which facilitated extraction of these compounds that might already be present in grape pomace.

The flavonols extracted from Sunbelt grape pomace as a function of solvent composition and temperature is shown in Table 3. The standard deviations for the flavonol data were also calculated using Equation 3. Statistical analysis indicated that there was a significant solvent (P<0.001) and temperature (P<0.001) effect on the extraction of flavonols from the grape pomace, while, the solvent-temperature interaction effect was insignificant (P=O. 1889). It can be seen from Table 3 that almost an equal amount of myricetin and quercetin glycosides were extracted from the grape pomace. A Student's t-test performed at 95% confidence level showed that the efficacy of the solvent in extracting flavonols from grape pomace is as follows: 80% (v/v) aqueous ethanol > 40% (v/v) aqueous ethanol > water. Comparing the results for anthocyanin and flavonol extraction from grape pomace, it can be seen that 80% (v/v) aqueous ethanol is the best solvent composition for extracting flavonoids from Sunbelt grape pomace.

While temperatures between 80° and 120°C were found to be optimum for extracting anthocyanins from grape pomace, higher temperatures (120°- 1400C) were better for extracting the maximum amount of flavonols from Sunbelt grape pomace. The difference in the temperature conditions required for maximum anthocyanin versus flavonol yield can be due to lower solubility of flavonols such as quercetin at lower temperatures (SRINIVAS et al, 2010). Solubility measurements made by SRINIVAS et al (2010) indicated that there was an exponential increase in the solubility of quercetin aglycone at temperatures above the boiling point of water. Another reason for this recorded difference could relate to a difference in the relative thermal degradability of flavonols and anthocyanins. Studies performed by TANCHEV and JONCHEVA (1973) have indicated that degradation of flavonoid glucosides increased with an increase in temperature and the thermal degradation kinetics of the flavonoids is also dependent on the pH of the solvent. PETERSSON et al (2010) have shown that flavonol glycosides in red onions started to degrade at around 8 minutes at 1 100C in a static batch extractor. The thermal degradation of flavonoid glycosides was found to result in breaking of the sugar -flavonoid bonds, thereby, forming aglycones and monoglucosides (KIM et al, 1991).

Effect of organic acids on subcriticai fluid extraction of flavonoids from grape pomace

The statistical analysis of the anthocyanin extraction data indicated that there is a significant effect of the solvent, temperature and acid (P<0.0001) on anthocyanin and flavonol yield from grape pomace. It was observed that twoway interaction and the three-way interaction effects of solvent, temperature and acid on the extraction of anthocyanins from Sunbelt grape pomace can also be significant. Similar results were obtained on the statistical analysis of the results on the flavonols extracted from grape pomace in this study.

The total anthocyanins and flavonols in mg/100 g dry weight (DW) of grape pomace extracted by the acidified solvent as a function of temperature are given in Table 4. A standardized Student's t-test on the effect of organic acids on the subcriticai fluid extraction of anthocyanins from grape pomace indicated that acetic acid addition provided a significantly higher total anthocyanin yield relative to using the other acids. However, acetic acid was less efficient in extracting flavonols from grape pomace, and the addition of any acid to the pressurized fluid, actually decreased the extraction efficiency of flavonols. Similar trends were seen by DOWNEY et al (2003) where 50-60% methanol (v/v) in water was found to be a better solvent for extracting flavonols, such as quercetin glycosides, from grape pomace, in the absence of acid addition. DOWNEY et al (2003) also indicated that a mineral acid (such as HCl) had a marginally greater benefit followed by organic acid addition, to hydromethanolic solvent used for the extraction of anthocyanins from grape pomace.

Examining the two-way and three-way interaction effects mentioned previously, 80% (v/v) aqueous ethanol acidified with acetic acid at 800C extracted a significantly higher amount of anthocyanins from grape pomace. Additionally, 80% (v/v) aqueous ethanol was an optimal solvent composition for the extraction of flavonols from grape pomace in the temperature range of 80° and 1400C and in the absence of organic acids. However when acetic acid and tartaric acid were added to 80% (v/v) aqueous ethanol a significantly greater amount of flavonols was extracted from grape pomace at 80° and 1200C respectively.

A negative effect of acids on the extraction of flavonols from grape pomace was also noticed by DOWNEY et al (2003) for 50% (v/v) aqueous methanol extraction of flavonol glycosides in absence of acid. Moreover, these studies also indicated that formic and acetic acids increased the amount of flavonols extracted with increased time of extraction. This negative effect of acid addition on extraction of flavonol glycosides from grape pomace can be related to the effect of pH on thermal degradation of the flavonols. Studies performed by BUCHNER et al (2006) indicated that flavonols such as quercetin and rutin, in aqueous acidified solution at pH=5.0 (with HCl) showed excessive thermal degradation leading to a loss of about 75% relative to the original total concentration at 5 hours for 1000C. Similar studies performed by GIZIR et al (2008) showed thermal degradation of anthocyanins extracted from black carrots when acetic, sulfuric and lactic acids were added to subcriticai water used as a solvent at temperatures greater than 1000C. In this study, the total amount of anthocyanins and flavonols extracted from Sunbelt grape pomace was found to not decrease until a temperature of 1400C was reached, which suggests the conditions used in this study are preferably to those cited above.

The total anthocyanins and flavonols extracted from Sunbelt grape pomace using acidified hydroethanolic mixtures were compared with that extracted using a conventional extraction solvent consisting of (60:37:3 (v/v) methanol:water:formic acid). The amount of anthocyanins and flavonols extracted with this solvent is assumed to be a baseline value in order to determine the efficiency of the ASE extractions. It was found that hydroethanolic mixtures (either in presence or absence of organic acids) extracted only about 12-14% of the total flavonols present in grape pomace when compared to the conventional solvent extraction method. On comparing the results using 0, 40 and 80% (v/v) aqueous ethanol to the conventional solvent extraction results, approximately 68.0, 90.5 and 95.8% of total anthocyanins were recovered respectively.

Studies performed by MONRAD et al (2010a) indicated that 50-70% (v/v) aqueous ethanol extracted about 103-105% of total anthocyanins present in grape pomace when compared with the conventional solvent extraction method. In this present study, when formic, acetic, citric and tartaric acids were added to 80% (v/v) aqueous ethanol, these solvent compositions extracted approximately 92.3, 120, 96.7 and 97.7% of total anthocyanins relative to the conventional solvent extraction method, respectively. Under optimal conditions, acetic acid added to 80% (v/v) aqueous ethanol at 80°C, actually extracted approximately 158% of the total anthocyanins compared to conventional solvent extraction method, indicating that the conventional solvent failed to extract all of the anthocyanins from grape pomace.

Surface regression models on the anthocyanin and flavonols extractions performed on the grape pomace were performed as discussed in the Experimental section. The surface response regression method used in this study is purely empirical, since the acids were assigned numerical values in performing the linear regressions. Hence, this regression model is restricted to this study and should be used only to estimate the optimal conditions to maximize anthocyanin and flavonol yield under these extraction conditions for this grape pomace.

The parameters for the generalized linear response model obtained from JMP are given in Table 5. It can be seen from the R2 value for the goodness of fit given in Table 5 that the predicted equation is in good agreement with the experimental values (70. 1% for anthocyanins and 85.3% for flavonols). However, a regression (R2) value greater than 90% is usually sought to indicate a very good correlation between the surface regression model and the experimental data. From the interaction profiler plots for optimizing the effect of solvent, temperature and acid on the extraction of anthocyanins (Fig. 2a) and flavonols (Fig. 2b) from grape pomace, it can be seen that 80% (v/v) aqueous ethanol was the most desired solvent. Moreover, while maximum anthocyanin yield was obtained at 85.4°C using acetic acid as additive, the maximum flavonol yield was obtained at 124°C in absence of any organic acid added to the solvent. The optimized conditions were reported at three significant digits due to the error reported in the prediction model.

Similar studies performed by MONRAD et al (2010a) indicated the optimal conditions for the extraction of anthocyanins from Sunbelt grape pomace using hydroethanolic solvent mixtures without the addition of acids was 103,70C and 70% (v/v) aqueous ethanol as solvent.

Extraction of acylated and non-acylated flavonoids from grape pomace using pressurized solvents

Statistical analysis on the anthocyanin and flavonol extraction data revealed that water was the least desirable solvent for the optimal extraction of flavonoids from grape pomace when compared to the hydroethanolic solvents. Similar trends were also noticed by MONRAD et al (2010a) and JU and HOWARD (2003; 2005). In the studies performed by JU and HOWARD (2005), it was found that sulfurea water was a better solvent than neat, subcriticai water in extracting phenolic compounds from red grape skins and this was attributed to the possible formation of a sodium bisulfate complex which increased the stability of the phenolic compound. It can be seen from Fig. 3a, that addition of organic acids had a significant effect on extraction of anthocyanins from Sunbelt grape pomace while there was no significant effect of acid additions to subcriticai water on extraction of flavonols from grape pomace (Fig. 3b). Moreover, there was no significant difference between the total amounts of anthocyanins extracted from grape pomace using subcriticai water acidified with the different organic acids. While acetic acid added to subcriticai water decreased the concentration of total anthocyanins extracted above 100°C, citric and tartaric acids showed a minimum at 120°C followed by a slight increase in total anthocyanins extracted at 140°C.

Studies performed by JU and HOWARD (2003) indicated that acidified water extracted an equivalent amount of anthocyanins from grape pomace at 80°- 1000C when compared to acidified 60% (v/v) aqueous methanol at around 1200C. They also reported that there was a greater recovery of anthocyanin mono- and di-glucosides from grape pomace when acidified water was used as extraction solvent. Studies have indicated that while non-acylated anthocyanin glycosides are water-soluble and easily assimilated by the human body (MARKAKIS, 1992), the more non-polar acylated anthocyanins are more stable to pH changes, heat treatment and light exposure (DANGLES et al, 1993; FRANCIS, 1992). This high stability of the acylated anthocyanin glycosides was attributed to higher inter- and intra-molecular copigmentation due to glycosylation and acylation, metal complexing and presence of inorganic salts (GIUSTI and WROLSTAD, 2003). Studies by KURILICH et al (2005) indicated that while acylated anthocyanins have relatively higher stability, their bioavailability is less compared to non-acylated polar anthocyanins. The study performed on human urine and plasma of the intake of acylated and non-acylated anthocyanins from carrots indicated that the polar non-acylated anthocyanins were easily absorbed in the human plasma while the acylated anthocyanins were easily excreted through human urinary system (KURILICH et al, 2005). These studies illustrate the significance and applications of the both acylated and non-acylated anthocyanins in nutraceutical and food industries.

The total amount of acylated and non-acylated anthocyanins extracted from Sunbelt grape pomace as a function of solvent, temperature and acid added to the solvent is given in Table 6. The non-acylated anthocyanins represent the sum of peaks 1-3 while the acylated anthocyanins represents the sum of all the other peaks 4-12 (as given in Table 1). The data given in Table 6 that water had a greater tendency to extract the more polar non-acylated anthocyanins from Sunbelt grape pomace. The ratio of the non-acylated anthocyanins to acylated anthocyanins extracted from Sunbelt grape pomace using water as a solvent varied between 1.5 and 3 while that by hydroethanolic solvents was less than or equal to 1.5. Additionally, the amount of acylated anthocyanins extracted from grape pomace using water as solvent increased with an increase in the extraction temperature, thereby reducing the ratio of non-acylated to acylated anthocyanins. This tendency of subcriticai water (either in presence or absence of organic acids) in extracting more polar non-acylated an

thocyanin glycosides from Sunbelt grape pomace makes it a desirable extraction solvent for nutraceutical applications.

A similar trend, i.e., increase in the amount of non-acylated anthocyanins extracted from grape pomace with an increase in the extraction temperature was also seen by JU and HOWARD (2003) and was attributed to a decrease in polarity of subcriticai water resulting in the extraction of the more non-polar anthocyanins. From Table 6, it can also be seen that the addition of organic acid to subcriticai water also increased the total amount of acylated anthocyanins extracted from grape pomace. Of all the organic acids added to the subcriticai fluids, citric and tartaric acids extracted a larger amount of acylated anthocyanins relative to the amount of non-acylated anthocyanins, thereby exhibiting a lower non-acylated to acylated anthocyanin ratio as a function of solvent and temperature when compared to extraction using nonacidified solvents.

When 40% (v/v) aqueous ethanol was used as a solvent, there was no significant difference between the non-acylated to acylated anthocyanin ratio using different organic acid and temperature conditions. The acidified hydroethanolic mixtures, however, are given greater preference by the food industry due to their tendency to extract greater amounts of acylated anthocyanins from natural products which have greater color, stability to temperature and light (GIUSTI and WROLSTAD, 2003) and can be used as natural colorants.

CONCLUSIONS

In this study, the effect of ethanol as co-solvent, temperature and organic acids on the subcritical water extraction of anthocyanins were analyzed. The solvent pH was maintained at 2.5 units by using formic, acetic, citric and tartaric acids. 80% (v/v) ethanol/water was found to be the most optimal solvent for extracting a larger amount of anthocyanins (in presence of acetic acid) and flavonols (in absence of organic acids added to the solvent) from Sunbelt grape pomace. Surface regression analysis of the anthocyanin and flavonol data indicated that the optimum extraction temperature for maximum anthocyanin yield was 85.4°C and that for maximum flavonol yield as 124°C. Under these optimal conditions, approximately 158% of the total anthocyanins and 14% of total flavonols were extracted from Sunbelt grape pomace when compared to the conventional solvent extraction method (60:37:3 (v/v) methanol:water:formic acid). Subcriticai water, in the presence or absence of organic acids, extracted a greater amount of polar non-acylated anthocyanins from grape pomace when compared to acidified hydroethanolic mixtures. The hydroethanolic mixtures (especially acidified with formic or acetic acid) extracted greater amounts of acylated anthocyanins from Sunbelt grape pomace.

ACKNOWLEDGEMENTS

This study was supported by the U.S. Department of Agriculture (USDA) grant (# 2006-35503-17618) under the CSREES National Research Initiative. We would like to thank Dr Jackson Lay and Dr Rohana Liyanage for HPLC-MS analysis of the grape extracts at the University of Arkansas Statewide Mass Spectroscopy Laboratory. The guidance of Dr Andy Mauromoustakos, Agricultural Sciences Laboratory at University of Arkansas, in the statistical analysis of the flavonoid data is also greatly appreciated.

References

REFERENCES

Anderson D.W., Gueffory D.E., Webb A.D. and Kepner RE. 1970. Identification of acetic acid as an acylating agent of anthocyanin pigments in grapes. Phyto chemistry 9: 1579.

Bleve M., Ciurlia L., Erroi E., Lionetto G., Longo L., Rescio L., Schettino T. and Vasapollo G. 2008. An innovative method for the purification of anthocyanins from grape skin extracts by using liquid and sub-critical carbon dioxide. Sep. Purif. Technol. 64: 192.

Borges G., Degeneve A., Mullen W. and Crozier A. 2010. Identification of flavonoid and phenolic antioxidants in black currants, blueberries, raspberries, red currants and cranberries. J. Agric. Food Chem. 58: 3901.

Bozidar S. 1995. The role of polyphenols as chemopreventers. Polyphenol Actualities 13: 24.

Braidot E., Zancani M., Petrusa E., Peresson C, Bertolini A., Patui S., Macri F. and Vianello A. 2008. Transport and accumulation of flavonoids in grapevine (Vitts vinifera L.). Plant Signal Behav. 3: 626.

Buchner N., Krumbein A., Rohn S. and Kroh L.W. 2006. Effect of thermal processing on the flavonoids rutin and quercetin. Rapid Commun. Mass Spectrom. 20: 3229.

Cacace J. E. and Mazza G. 2002. Extraction of anthocyanins and other phenolics from black currants with sulfured water. J. Agric. Food Chem. 50: 5939.

Cho M.J., Howard L.R., Prior R.L. and Clark J.R. 2004. Flavonoid glycosides and antioxidant capacity of various blackberry, blueberry and red grape genotypes determined by high-performance liquid chromatography /mass spectrometry. J. Sci. Food Agrie. 84: 1771.

Clifford M.N. 2000. Anthocyanins - nature, occurrence and dietary burden. J. Sci. Food Agrie. 80: 1063.

Dangles O., Saito N. and Brouillard R. 1993. Anthocyanin intramolecular copigment effect. Phytochemistry 34: 119.

de Castro M. D. L., Luque-Rodriguez M. L. and Japon-Lujan R. 2008. Exploitation of residues from vineyards, olive groves, and wine and oil production to obtain phenolic compounds of high-added value. In: "Methods of Analysis for Functional Foods and Nutraceuticals" Hurst W. J. (Ed.) p. 207. CRC Press, Boca Raton, FL.

Downey M.O., Mazza M. and Krstic P. 2007. Development of a stable extract for anthocyanins and flavonols from grape skin. Am. J. Enol. Vitic. 58: 3.

Francis F.J. 1992. A new group of food colorants. Trends Food Sci. Technol. 3: 27.

Francis J. and Markakis P.C. 1989. Food colorants: anthocyanins. Crit. Rev. Food Sci. Nutr. 28: 273.

Garcia-Marino M., Rivas -Gonzalo J.C., Ibanez E. and Garcia-Moreno C. 2006. Recovery of catechins and proanthocyanidins from winery by-products using subcriticai water extraction. Anal. Chem. Acta. 563: 44.

Giusti M. M. and Wrolstad R. E. 2003. Acylated anthocyanins from edible sources and their applications in food systems. Biochem. Eng. J. 14: 217.

Gizir A.M., Turker N. and Artuvan E. 2008. Pressurized acidified water extraction of black carrot Daucus carota ssp. Sativus vor. artrombens Alef.) anthocyanins. Eur. Food Res. Technol. 226: 363.

Hager T.J. , Howard L.R. and Prior R.L. 2008. Processing and storage effects on monomeric anthocyanins, percent polymeric color, and antioxidant capacity of pressed blackberry products. J. Agric. Food Chem. 56: 689.

Halpern M.J., Dahlgren A-L., Leakso L, Seppanen-Laakso T. , Dahlgren J. and McAnulty P.A. 1998. Red -wine polyphenols and inhibition of platelet aggregation: possible mechanisms, and potential use in health promotion and disease prevention. J. Intl. Med. Res. 26: 171.

Harborne J. B. and Williams C. 2000. Advances in flavonoid research since 1992. Phytochem. 55: 481.

Heredia, F.J., Francia-Aricha, E. M., Rivas-Gonzalo, J. C, Vicario, LM. and Santos -Buelga C. 1998. Chromatic characterization of anthocyanins from red grapes I. pH effect. Food Chem. 63: 491.

Jackman R.L., Yada R. Y., Tung M.A. and Speers R. A. 1987. Anthocyanins as food colorants - a review. J. Food Biochem. 11: 201.

Ju Z. and Howard L.R. 2006. Sub critical water and sulfurea water extraction of anthocyanins and other phenolic s from dried red grape skin. J. Food Sci. 70: S270.

Ju Z. Y. and Howard L.R. 2003. Effect of solvent and temperature on pressurized liquid extraction of anthocyanins and total phenolics from dried red grape skin. J. Agric. Food Chem. 51: 5207.

Kim Y. C., Higuchi R., Kitamura Y. and Komori T. 1991. Thermal degradation of glycosides, IV. Degradation of flavonoid glycosides. Liebigs Annalen der Chemie, 1991: 1285

King J.W. and Srinivas K. 2009. Multiple unit processing using sub- and supercritical fluids. J. Supercrit. Fluids. 47: 598.

King J.W., Gabriel R.D. and Wightman J.D. 2003. Subcritical water extraction of anthocyanins from fruit berry substrates. Proceedings of the 6th International Symposium on Supercritical Fluids, Tome 1, April 28-30, Versailles, France, p. 409.

Kurilich A.C., Clevidence B.A., Britz S.J., Simon P. W. and Novotny J. A. 2005. Plasma and urine responses are lower for acylated vs nonacylated anthocyanins from raw and cooked purple carrots. J. Agric. Food Chem. 53: 6537.

Lafka T-L, Sinanoglou V and Lazos E.S. 2007. On the extraction and antioxidant activity of phenolic compounds from winery wastes. Food Chem. 104: 1206.

Lapidot T, Harel S., Akiri B., Granit R. and Kanner J. 1998. pH-dependent forms of red wine anthocyanins as anti-oxidants. J. Agric. Food Chem. 47: 67.

Lapornik B., ProsekM. andWondraAG. 2005. Comparison of extracts prepared from plant by-products using different solvents and extraction time. J. Food Sci. 71: 214.

Lurton L. 2003. Grape polyphenols: a new powerful health ingrethents. Innovations Food Tech. 18: 28.

Luthria D. L. 2008. Influence of experimental conditions on the extraction of phenolic compounds from parsley Petroselinium crispum) flakes using a pressurized liquid extractor. Food Chem. 107: 745.

Markakis P. 1992. Stability of anthocyanins in foods. In "Anthocyanins as Food Colors". Markakis P. (Ed.). Chapter 6. Academic Press, New York.

Metivier R.P., Francis F.J. and Clydesdale F.M. 1980. Solvent extraction of anthocyanins from wine pomace. J. Food Sci. 45: 1099.

Monrad J.K., Howard L. R., King J. W., Srinivas K. and Mauromoustakos A. 2010a. Subcriticai solvent extraction of anthocyanins from dried red grape pomace. J. Agric. Food Chem. 58: 2862.

Monrad J. K., Howard L. R., King J. W., Srinivas K. and Mauromoustakos A. 2010b. Subcriticai solvent extraction of procyanidins from dried red grape pomace. J. Agric. Food Chem. 58: 4014.

Moure A., Cruz J. M., Franco D., Domínguez J. M., Sineiro J., Dominguez H., Jose Nunez M. and Parajo C. 2001. Natural antioxidants from residual sources. Food Chem. 72: 145.

Mukhopadhyay S. , Luthria D.L. and Robbins R. J. 2006. Optimization of extraction process for phenolic acids from black cohosh Cimicifuga racemosa) by pressurized liquid extraction. J. Sci. Food Agric. 86: 156.

Murthy K.N.C. , Singh R.P. and Jayaprakasha G.K. 2002. Antioxidant activities of grape ( Vitis vinifera) pomace extracts. J. Agric. Food Chem. 50: 5909.

Palma M., Pineiro Z. and Barroso CG. 2001. Stability of phenolic compounds during extraction with superheated solvents. J. Chromatogr. A. 921: 169.

Petersson E.V., Liu J., Sjoberg P.J. R., Danielsson R., and Turner C. 2010. Pressurized hot water extraction of anthocyanins from red onion: a study on extraction and degradation rates. Anal. Chim. Acta 663: 27.

Pinelo M., Arnous A. and Meyer A. S. 2006. Upgrading of grape skins: significance of plant cell-wall structural components and extraction techniques for phenol release. Trends Food Sci. Tech. 17: 579.

Revilla E., Ryan J-M. and Martin-Ortega G. 1998. Comparison of several procedures used for the extraction of anthocyanins from red grapes. J. Agrie. Food Chem. 46: 4592.

Ruberto G., Renda A., Daquino C., Amico V., Spatafora C., Tringali C. and De Tommasi N. 2007. Polyphenol constituents and antioxidant activity of grape pomace extracts from five Sicilian red grape cultivars. Food Chem. 100: 203.

Savova M., Kolusheva T., Stourza A. and Seikova I. 2007. The use of group contribution method for predicting the solubility of seed polyphenols of Vitis Vinifera L. within a wide polarity range in solvent mixtures. J. Univ. Chem. Technol. Metallur. 42: 295.

Schieber A., Stintzing F.C. and Carle R. 2001. By-products of plant food processing as a source of functional compounds - recent developments. Trends Food Sci Technol. 12: 401.

Sharif A., Saim N., Jasmani H. and Ahmad W.Y.W. 2010. Effects of solvent and temperature on the extraction of colorant from onion Allium cepa) skin using pressurized liquid extraction. Asian J. Appl. Sci. 3: 262.

Souquet J-M., Labarbe B., Le Guerneve C, Cheynier V and Moutounet M. 2000. Phenolic composition of grape stems. J. Agric. Food Chem. 48: 2076.

Spigno G. and De Faveri D.M. 2007. Antioxidants from grape stalks and marc: influence of extraction procedure on yield, purity and antioxidant power of the extracts. J. Food Eng. 78: 793.

Srinivas K., King J.W., Howard L.R. and Monrad J.K. 2010. Determination of solubility of querce tin and quercetin dihydrate in sub critical water using a dynamic flow technique. J. Food Eng. 100: 208.

Srinivas K., King J.W., Monrad J.K., Howard L.R. and Hansen C.M. 2009. Optimization of sub critical fluid extraction of bioactive compounds using Hansen solubility parameters. J. Food Sci. 74: E342.

Tanchev S.S. and Joncheva N. 1973. Kinetics of the thermal degradation of cyaniding-3-rutinoside and peonidin-3-rutinoside. Z. Lebensm. Unters-Forsch 153: 37.

Toyomizu M., AAkiba Y., Horiguchi M. and Matsumoto T. 1982. Multiple regression and response surface analyses of the effects of dietary protein, fat and carbohydrate on the body protein and fat gains in growing chicks. J. Nutr. 112: 886.

Wang H., Cao G. and Prior R. L. 1997. Oxygen radical absorbing capacity of anthocyanins. J. Agric. Food Chem. 45: 304.

Paper received September 1, 2010 Accepted December 16, 2010