We used the following genotypes: (i) 5 grapevine lines (V.vinifera), ‘Hongmeigui’, ‘Cabernet Sauvignon’, ‘Chardonnay’, ‘Pinot Noir’ and ‘Victoria Blanc’, collected from Yangling, Shaanxi Province; (ii) 6 Chinese wild varieties, including three wine grapes, ‘Shuanghong’, ‘Beibinghong’ and ‘Shuangyou’, of V. amurensis from Tonghua, Jilin Province; one variety, ‘Maoputao’, of V. pentagona from Lantian, Shaanxi Province; and two wine grapes, ‘Junzi’ and ‘Baiyu’, of V. davidii from Chongyi, Jiangxi Province. Ideal grape canes with moderate vigor (0.8-1.0 cm diameter) were collected during the 2008 pruning period (Fig. 1). First, all sampled vine shoot samples were frozen in liquid nitrogen. Then, samples were lyophilized under -50 °C in an lyophilizer (VirTis Genesis 25 XL, Gardiner, NY, USA) and ground using a domestic electrical grinder (final particle size <0.5 mm). Powdered grape canes were packed in air-tight bags and stored at -20 °C before use.

Reagents as 2,4,6-tri(2-pyridyl)-s-triazine (TPTZ), 2,2-diphenyl-1-picrylhydrazyl (DPPH), tertiary butylhydroquinone (TBHQ), ferrozine, β-carotene, Folin−Ciocalteu’s phenol regent, 6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (Trolox), linoleic acid, and all the phenolic compounds (purity > 97%) were purchased from Sigma-Aldrich (Shanghai, China). Methanol and acetonitrile were obtained from Kermel Chemical Reagent Co. Ltd. (Tianjin, China) and were of HPLC grade. Water was double distilled and purified through a Milli-Q system (Millipore, Bedford, MA, USA). All other chemicals were obtained from Xi’an Chemical Reagent Co. Ltd. (Xi’an, China) and were of analytical grade.

The procedure used to prepare the extracts/fractions is presented in Fig. 2. Pulverized vine canes of each cultivar (100 g, dw) were extracted with 1000 mL of 1 mol/L HCl/methanol/water (1:80:19, v/v/v) and centrifuged using a high-speed refrigerated centrifuge for 20 min at 12000 rpm and 4 °C. The resulting supernatant was collected and extracted thrice. The combined extracts were evaporated (Büchi RE-111, Switzerland) under 35 °C to remove methanol and were then defatted with petroleum ether (3:1, v/v). The remaining aqueous extract was lyophilized to obtain methanolic extracts (ME), which were re-suspended in distilled water (1:10, w/v) and then successively partitioned with an equal volume of chloroform and ethyl acetate (3:1, v/v) to afford chloroform fractions (CF) and ethyl acetate fractions (EAF), respectively. The remaining extract was freeze-dried and considered to be water fractions (WF). The ME and its three fractions were dissolved in methanol (5 mg/mL) and stored at -20 °C for further analysis. The yields of extracts/fractions were calculated and expressed as percent of the dry plant sample (%, w/w).

The determination of total phenolic contents (TPC) of ME and its fractions was performed according to Folin-Ciocalteu colorimetric method (Singleton et al., 1999). The determination of total flavonoids contents (TFC) was performed according to the aluminum chloride colorimetric method (Chang et al., 2002).

The free radical-scavenging activity of each extract/fraction was measured according to the procedure of Brand-Williams et al. (1995). The results were expressed as the IC50 values (μg/mL).

The ability to reduce Fe²⁺ was measured according to the protocol of Benzie & Strain (1996). The fresh working solution was prepared by mixing 300 mM acetate buffer (pH 3.6) with 10 mM 2,4,6-tri(2-pyridyl)-s-triazine solution (dissolved in 40 mM HCl) and 20 mM FeCl₃ solution (10:1:1, v/v/v). The solution was stored at 37 °C before use with 100 μL distilled water, followed by 2.85 mL of working solution. The absorbance value was measured at 593 nm after a 30-min incubation at 37 °C. The results were calculated using regression equations and expressed as mM Trolox equivalents/g extracts or fractions (mM TE/g).

This test was performed according to the protocol of Višnja et al. (2010). Briefly, the adequately diluted grapevine extract (1 mL) was mixed with methanol (3.7 mL), 2 mM FeCl₂ (0.1 mL) and 5 mM ferrozine (0.2 mL). The mixture was shaken vigorously and left standing at room temperature in the dark for 10 min. The absorbance of the resulting solution was measured spectrophotometrically at 562 nm. A low absorbance of the resulting solution indicated a strong Fe²⁺-chelating ability. The ability to chelate ferrous ion and prevent the formation of ferrous ion-ferrozine complex was calculated using the following equation:

Chelating effect (%)= [1-(A_sample/A_control)]×100,

where A_control was the absorbance of a mixture of methanol (4.7 mL), 2 mM FeCl₂ (0.1 mL) and 5 mM ferrozine (0.2 mL). All analyses were performed in triplicate and averaged. Sample concentration providing 50% inhibition (IC50) was calculated from the graph plotting inhibition percentage against extract concentration.

The assay was performed in a modified β-carotene/linoleic acid emulsion system. Briefly, an aliquot (10 mL) of 0.2 mg/mL β-carotene chloroform solution was thoroughly mixed with 200 mg of linoleic acid and 2000 mg Tween 20. The mixture was then evaporated at 40 °C to remove chloroform, and the residue was diluted with 1000 mL distilled water to form an emulsion after a vigorous agitation. Each sample (1 mL) diluted in 5 mL of the emulsion was tested at a final concentration of 200 mg/L. The reaction mixture with samples displaced by the same volume of methanol was the control, and for the positive control, tert-butylhydroquinone (TBHQ) was used. After 3 h in a water bath (50 °C), the zero time absorbance was immediately recorded at 470 nm and successively at 180 min against a blank consisting of an emulsion without β-carotene. The antioxidant activity (AA) was calculated using the following equation: AA (%) = [(As₀ − As₁₈₀) / (Ac₀ − Ac₁₈₀)] × 100. As₀ and Ac₀ represent the absorbance values for the samples and control, respectively, which were measured at the initial incubation time (t = 0 min), whereas As₁₈₀ and Ac₁₈₀ are the absorbance values for samples and control, respectively, which were measured at the end of the incubation time (t = 180 min).

The chromatographic analyses were performed using a liquid chromatograph system (Shimadzu LC-2010AHT, Kyoto, Japan) equipped with a quaternary pump, a photodiode array detector (DAD) and a UV detector. The DAD detector was applied to scan phenolic compounds of interest to ascertain their maximal absorbance wavelengths. The variable UV detector was used for quantitative purposes with the external standard method. All standards were dissolved in methanol at a stock concentration of 1 mg/mL. A calibration standard mixture was prepared by appropriate dilutions with methanol from the stock solution. Known amounts of extracts/fractions were dissolved in methanol. All solutions were stored in the dark at -40 °C and filtered through 0.22-µm membranes before injection.

The chromatographic conditions used were similar to those of the method published by Zhang et al. (2007). Briefly, a gradient solvent system was employed, where solvent A was water-acetic acid (97:3, v/v) and solvent B was acetonitrile. The elution profile had the following proportions (v/v) of solvent B: 0.00-5.00 min, 0-8.5%; 5.00-16.50 min, 8.5-2.0%; 16.50-35.00 min, 2.0-18%; 35.00-50.00 min, 18-20%; 50.00-65.00 min, 20-30%; 65.00-70.00 min, 30-0%. The wavelength-switching program was employed. The column was held at 30°C and was flushed at a flow rate of 0.8 mL/min. The UV detector was used for quantitative purposes with the external standard. The injection volume for all solutions was 10 µL, and the procedure was performed in triplicate. The linearity of the method was established by automatic injections of the standard mixture solutions at six calibration levels in three replicates from low to high concentrations. All data were processed using the Shimadzu Workstation CLASS-VP 6.12 software.

All determinations were performed in triplicate, and the results were expressed as the means and standard deviations (SD). The SPSS 17.0 software (SPSS Inc., Chicago, IL, USA) was used to compare the difference between means by Duncan’s t-tests, and p < 0.05 was considered significant. Correlations between antioxidant capacities and phenolic contents were computed as Pearson’s correlation coefficients (r).

Different extraction mediums have significant effects on extraction efficiency, which is an important factor for the recuperation of bioactive compounds. The most widely used solvents for extracting phenolics from solid grape wastes are methanol, acetone, ethanol and their water mixtures. Previous studies have reported that methanol provides relatively higher amounts of extractable compounds compared with other solvents (Zielinski & Kozlowska, 2000). The addition of acids to extraction solutions could enhance the extraction efficiency due to denaturing cellular membranes and facilitating the solubilization of phenolics (Zhang et al., 2001). Therefore, in this study, an acidified methanol solution was employed as a crude extraction medium, and methanolic ME were selected as a basis for further partitioning and purifying by petroleum ether, CF, EAF and WF. Petroleum ether was used to remove non-polar concomitant compounds, such as lipids and chlorophyll, in the crude extracts.

The extraction yield, TPC, and TFC of the extracts/fractions from different vine canes are presented in Tables 1 and 2. The results clearly show that ME had the highest extraction yield, ranging from 15.0% for cv. ‘Junzi’ to 25.9% for ‘Cabernet Sauvignon’, whereas EAF exhibited the lowest yield, varying from 0.95% (Junzi) to 2.03% (Cabernet Sauvignon). The high extraction yield of ME could be attributed to the fact that aqueous methanol was effective at extracting polyphenols linked to polar fibrous matrices (Hussein et al., 1990). Conversely, grape canes are a lignocellulosic material, and the main constituents (lignin, cellulose and hemicelluloses) can be hydrolyzed during exposure to low pH values (Spigno et al., 2004). Similar extraction yield orders were observed in other plant materials, such as grape pomace and bamboo shoots (Campos et al., 2008; Park & Jhon, 2010).

The TPC and TFC of extracts/fractions varied from 47.9 to 833 mg gallic acid equivalent (GAE)/g and 7.10 to 589 mg quercetin equivalent (QE)/g, respectively. Among all extracts/fractions, both the highest TPC (accounting for 37.7% of total extractable phenolics) and the highest TFC (41.3% of total extractable flavonoids) were detected in the EAF, whereas the lowest contents were found in WF and CF, respectively. These results suggest that medium polar phenolic compounds might be a major component of vine shoot phenolics and that ethyl acetate is suitable to extract phenolic compounds from grape canes. Ethyl acetate fractions with the highest polyphenolic content were also reported for other plant materials (Campos et al., 2008; Conde et al., 2008). The mean TPC and TFC of all extracts/fractions were in the following orders: EAF > ME > CF ≥ WF and EAF > ME > WF > CF, respectively. The results also indicate that the phenolic level in each fraction was not directly related to the corresponding extraction yield (Table 1). For example, methanolic extracts and their fractions from the varieties of V. davidii, with the lowest extraction yields, had the highest TPC and TFC values. Similar findings were also found in buckwheat extracts by Sun et al. (2005), who reported that the extraction of buckwheat also had the highest TPC and TFC values.

Among all cultivars, ‘Junzi’ had the highest TPC in its ME, EAF and WF (225, 833 and 97.3 mg GAE/g, respectively), whereas ‘Victoria Blanc’ exhibited the lowest CF, EAF and WF contents (52.5, 428 and 47.9 mg GAE/g, respectively) followed by ‘Chardonnay’, with values of 58.5, 465 and 57.5 mg/g, respectively, in the corresponding fractions. The trend of the TFC level was similar to those observed for the TPC of the cultivars analyzed, where ‘Junzi’ had the highest TFC values of 147, 13.2, 589 and 67.5 mg/g QE in its ME, CF, EAF and WF, respectively, followed by ‘Baiyu’, with corresponding values of 110, 9.43, 474 and 53.3 mg QE/g, respectively. In contrast, ‘Victoria Blanc’ had the lowest TFC in ME (33.1 QE/g) and WF (11.9 QE/g), and ‘Shuanghong’ showed the lowest values in CF (7.10 QE/g) and EAF (199 QE/g). It is worth noting that ‘Junzi’ of V. davidii is the varieties with vine shoot extracts that are rich in phenolic compounds. Further statistical analysis in Table 3 indicate good correlations between TPC and TFC for ME (r = 0.92, p < 0.05) and WF (r = 0.84, p < 0.05). Thus, it can be presumed that flavonoids are a major group of phenolics present in vine shoot extracts.

Comparisons of the mean values of TPC and TFC in all extracts/fractions among 4 grape species are depicted in Fig. 3. As expected, considerable variability in TPC or TFC values was observed within the corresponding fractions. For example, significant differences (p < 0.05) in TPC levels were observed among V. amurensis, V.pentagona and V. davidii for ME, EAF and WF (Fig. 3A). TFC levels in EAF of V. amurensis, V. vinifera and V.davidii exhibited significant differences (p < 0.05) (Fig. 3B). Zhang et al. (2011a,b) found that phenolic compounds of grapevine exhibit variation among numerous factors, such as cultivar, environmental conditions and management practice. However, in our case, the considerable differences may depend on the grape species and the distinct habitats in which the vine shoot samples were collected.

Mechanisms of antioxidant action include several aspects, such as inhibiting reactive oxidants generation, scavenging or destroying free radicals to break chain reactions, and binding the transition metal ions (Dinis et al., 1994; Brand-Williams et al., 1995; Jayaprakasha et al., 2001). The specificity and sensitivity of a single method does not typically accurately reflect the complete examination of all antioxidants in complex matrices, such as botanical extracts (Frankel & Meyer, 2000). In this study, a combination of several different test systems, including DPPH radical-scavenging activity, ferric reducing antioxidant power (FRAP), iron-chelating capacity and β-carotene-linoleic acid bleaching assays, were used to provide a reliable antioxidant assessment of grape canes extracts/fractions. The phenolic content of plant extracts is associated with their antioxidant properties. Thus, the relationship between the antioxidant activity and the TPC and TFC was also investigated.

Extracts/fractions from grape canes indicated different antioxidant capacities in different in vitro testing systems. To determine which constituent(s) is/are the most important active components in different extracts/fractions, HPLC-DAD-UV was used for both qualitative and quantitative analyses. Phenolic compounds were identified and confirmed by comparing their retention times and the spectral characteristics of their peaks with those of authentic standards and by spiking the sample with standards and peak purity detection. The typical chromatographic profiles of phenolic composition of extracts/fractions are presented in Fig. 5. Of note, satisfactory separation with good resolution was achieved by the modified HPLC condition. Seven phenolic compounds, including gallic acid (GA), protocatechuic acid (PA), vanillic acid (VA), syringic acid (SYA), catechin (CAT), epicatechin (EC) and trans-resveratrol (RES), were identified and quantified in different extracts/fractions. These phenolic compounds, which occur widely in lignified organs (tissues) of grapevines, such as stems, shoots, canes and xylem, have already been studied (Luque-Rodríguez et al., 2006; Zhang et al., 2007). The quantitative data of all of the extracts/fractions from grape canes are summarized in Table 5. The results of total individual phenolic contents quantified by HPLC were similar to those obtained using the Folin-Ciocalteu method (Table 1). The composition profiles of ME and the other two fractions were more complex than those of the CF, which was devoid of EC in most cultivars analyzed and of GA in all cultivars. This finding is largely attributed to different extraction solvents with different extracting capacities (Park & Jhon, 2010). The highest contents of phenolic compounds were observed in the EAF, with mean values of 3.50 ± 0.67, 10.7 ± 3.65, 3.98 ± 0.92, 10.4 ± 3.02, 79.9 ± 33.75, 79.9 ± 29.02 and 127 ± 37.9 mg/g for GA, PA, VA, SYA, CAT, EC and RES, respectively. The results indicate that most of the phenolic compounds in ME go to the EAF via solvent-solvent partitioning. This finding may explain why EAF exhibited increased antioxidant activity compared with the other fractions (Table 1). Flavan-3-ols (CAT and EC) and stilbene (RES) were detected as the two main phenolic groups in extracts/fractions, accounting for 92, 83, 90 and 93% of total identified phenolic compounds in the ME, CF, EAF and WF, respectively. Regarding these predominant compounds in the corresponding extract/fraction of different grape canes, varieties from V. davidii exhibited significantly increased amounts of flavan-3-ols compared with V. vinifera and V. amurensis (p < 0.05), whereas V. vinifera and V. amurensis possessed significantly increased trans-resveratrol levels compared with V. davidii (p < 0.05). Furthermore, the extracts/fractions with the highest levels of single compounds did not typically demonstrate the highest antioxidant activities. For example, the ME of cv. ‘Pinot Noir’ and EAF of ‘Baiyu’, with the highest amounts of RES and CAT, respectively, were not the strongest in most antioxidant assays. Thus, the antioxidant activity of extracts/fractions more likely depends on the combination of several or more phenolic compounds. Flavan-3-ols and stilbenes may contribute significantly; however, other yet-unidentified phytochemicals and the possible interactions among them also should be considered in the overall antioxidant effects of extracts/fractions from grape canes.

This study is a first report on the antioxidant activities and phenolic composition of methanolic extract (ME) and its three fractions of vine shoots from V. amurensis, V. davidii, V. pentagona and V. vinifera from a waste utilization perspective. All extracts/fractions have potent antioxidant activity based on the DPPH radical-scavenging, FRAP, Fe²⁺-chelating and β-carotene bleaching assays. The ME of grape canes demonstrated the highest iron-chelating activity, whereas ethyl acetate fractions, which have the highest TPC and TFC, exhibited the highest free-radical scavenging and reducing power activities, as well as the highest anti-lipid peroxidation activity. Among all grape genotypes analyzed, varieties from V. davidii possessed higher antioxidant activities than did those from the three other grape species. Antioxidant activities of extracts/fractions correlated positively with their total phenolic and flavonoids contents in all assays except the metal chelating power test. Qualitative and quantitative analyses of phenolic compounds by HPLC-DAD-UV indicated that catechin, epicatechin and trans-resveratrol were the main phenolic components of extracts/fractions. Further work in our laboratory is in progress to identify and characterize more inherent phytochemicals from different grape extracts and to evaluate their in vivo antioxidant potential.