Kinetic Improvement of Bioactive Compounds Extraction from Red Grape (Vitis vinifera Moldova) Pomace by Ultrasonic Treatment

Grape pomace from a red grape variety (Vitis vinifera Moldova) cultivated in the northeastern region of Romania has been studied as a source for the extraction of total monomeric anthocyanin (TMA) and total phenolic content (TPC) using ultrasonic treatment. The method of extraction described here uses two different solvents, namely 2-propanol and methanol. For each of the extraction solvents, we evaluated the singular influence and the impact of interactions between process parameters (solvent concentration, ultrasonic frequency, temperature, and extraction time) on the extraction yields of anthocyanins and phenolic compounds. Response surface methodology was implemented via a Box–Behnken design to optimize the extraction of TMA and TPC from grape pomace. According to the optimization, in order to achieve the highest yield of TPC (62.487 mg gallic acid equivalent (GAE)/g (d = 1.0)), the following conditions are necessary: solvent—2 propanol, solvent concentration 50%, temperature −50 °C and extraction time 29.6 min.


Introduction
Grapes (genus Vitis) are one of the most widely cultivated fruit crops in the world, with uses in the food industry that range from juice, wine, jam, jelly, raisins, and vinegar production to oil extraction from grape seeds. It is estimated that about 80% of the grape harvest is used in the winemaking industry [1]. Wine production and consumption has traditionally been concentrated in the European continent, and continues to be dominated by three major European producers: Italy, France and Spain. Since its major decline in 2012 [2], global wine production rose to 275.7 MhL in 2015, with a slight increase of 2% compared with the previous year [3]. As can be expected, such a large industry is associated with the production of significant quantities of solid and liquid waste. Solid waste includes grape stalks (2.5-7.5%), grape pomace (up to 25-45%) and grape seeds (3-6%) [4]. In European countries alone, it has been estimated that 14.5 million tons of grape by-products are generated annually by winemaking industry [5,6].
Grape pomace is the main winery waste, and results from the winemaking process during the production of grape must by crushing whole grapes. Grape pomace, which is comprised of skins and seeds, is reported to account for an approximate percentage of between 25% and 30% of the grapes used [7,8]. The reported content of seeds in grape pomace varies from 7-20% to a greater proportion of 38-52% [9,10]. Regarding the chemical composition of grape pomace, several studies reported on the content in protein, fiber, oil, and sugars of the by-product [11][12][13][14]. Due to its high fiber content, grape pomace is being used as a feed additive [15]. Other reuses described for this winery waste are as raw materials in the production of alcoholic beverages and fertilizers, and also in distilleries for obtaining ethanol [16][17][18].

Total Monomeric Anthocyanin Determination
The total monomeric anthocyanin (TMA) content of grape pomace extract was determined using a spectrophotometric method developed by Rabino and Mancinelli [43]. A UV-VIS spectrophotometer (Ocean Optics, Largo, FL, USA) was used for spectral measurements at 657 nm and 530 nm.

Total Phenolic Content
The total phenolic content of crude extracts was measured using the Folin-Ciocalteu (FC) reagent-based colorimetric assay as described by Singleton and Rossi [44]. The TPC values were expressed in mg gallic acid equivalent (GAE)/g [45] in extract wet basis. The calibration curve was done using gallic acid at concentrations of 0.5, 1, 2, 3, 4, 5 and 10 mg/L with a regression coefficient of 0.995.

Raman Spectra Analysis
Spectra of grape pomace extracts were recorded using an i-Raman spectrometer (EZM-A2-785L, B&W TEK Inc., Newark, DE, USA) equipped with a fiber-optic Raman probe; the spectrum was recorded from 2600 − 175 cm −1 . Samples were scanned at an increment of 10 nm. Integration time was 15 s. In order to display the spectra of the extract, Spekwin32-optical spectroscopy software (Version 1.72.2, 2016, http://www.effemm2.de/spekwin/) was used.

Analysis of Variance
Multifactor analysis of variance (ANOVA) was applied to detect differences in the extraction yield of total phenolic content and total monomeric anthocyanins with changes in the process parameters, namely, extraction solvent, solvent concentration, ultrasonic frequency, temperature, and extraction time. Analyses were performed with the Unscrambler X 10.1 (Camo, Oslo, Norway). Differences were considered to be significant at p < 0.05. It was a Box-Behnken design based on a second-order (quadratic) polynomial response surface model using the following equation: where: y is the predicted response (TMA or TPC), x i stands for the coded levels of the design variable (solvent concentration, time, temperature and ultrasonic frequency; Table 1), b 0 is a constant, b i represents the linear effects, b ii represents the quadratic effects and b ij represents the interaction effects. The parameters were optimized in order to achieve the highest concentration of TMA and TPC using the desirability function approach [46].

Influence of Extraction Parameters on TMA and TPC Yield
The targeted compounds were subsequently extracted by ultrasonication with two different solvents, namely, 2-propanol and methanol. The influence of extraction solvent selection on the extraction yield was further studied by varying solvent concentration and recording the changes in the solubility of phenolics and anthocyanins. Similarly, the effect of increasing extraction temperature and extending the sonication time was also investigated. The influence of the parameters of TMA and TPC are presented in Table 2, and the interactions between the extraction parameters are presented in Table 3.

Solvent Type and Concentration
Solvent type is one of the main factors affecting the extraction efficiency of phenolic compounds from plant materials. Due to the polar nature of phenolic compounds, the extraction is usually performed in polar protic media, in which these compounds are easily solubilized. Two polar solvents, namely 2-propanol and methanol, were used in this study, and their ability to extract anthocyanins and phenolic compounds was compared. As Table 2 shows, 2-propanol gave higher extraction yields of total monomeric anthocyanin (average concentration of 5.37 mg/g grape pomace) in comparison to methanol (average of 4.58 mg/g). As a result, the choice of extraction solvent had a significant effect on the extraction yield of TMA (p < 0.001). Comparable extraction yields of anthocyanins were obtained by Corrales et al. [37]. Applying an ultrasonic treatment to grape pomace dispersed in a mixture of ethanol and water (50:50) at a frequency of 35 kHz at 70 • C resulted in extraction yields of 7.76 mg/g for total anthocyanins and 4.85 mg/g for total anthocyanin monoglucosides. Through the conventional extraction of dehydrated grape pomace from the red grape variety Benitaka (São João do Piauí, Piui, Northeastern Brazil) in a solution of ethanol 85% and HCl (1.5 N) for 24 h under refrigeration and absence of light, Sousa et al. [29] determined the content of anthocyanins to be 131 mg/100 g. Compared to the content of anthocyanins measured in the current study, the levels reported by the above-mentioned authors are approximately 4 times lower; the differences are justified by the variation in anthocyanin content between varieties [47] and the choice of extraction solvent, and can be also attributed to a degradation of these compounds due to the high temperature (60 • C) applied in the drying process of pomace, as it was previously indicated that high temperature (maximum 35 • C) causes a reduction in total anthocyanin content [48].
In the case of total phenolic content, both extraction solvents ensured an efficient extraction, with methanol being particularly remarkable for its high extraction yield (average of 46.53 mg/g). Similar to total monomeric anthocyanin extraction, the solvent type had a significant effect on the extraction yield (p < 0.001). The highest levels of phenolic compounds using methanol as an extraction solvent were also obtained by Casazza et al. [49], who determined a total content of phenolics of 90 mg/g in dried Pinot Noir skins when using UAE. The affinity of phenolic compounds for methanol may be ascribed to the dielectric constant (32.60), which is greater than that of 2-propanol (19.90) [50], or to the lower surface tension of methanol (22.1 mJ/cm 2 ) when compared to 2-propanol (23.3 mJ/cm2) [51]. For the same red grape variety (Vitis vinifera Moldova) Budiul and Albulescu [52] reported a total phenol content of 1164.3 mg/L for Soxhlet extraction with methanol and of 1562.75 mg/L for extraction with a mixture of MeOH:H 2 O:HCl (90:9.5:0.5, v/v). Drosou et al. [53] previously showed that ultrasonic-assisted extraction of grape pomace leads to greater extraction yields of phenolic compounds than Soxhlet extraction in ethanol or water. Therefore, a great increase in the extraction yield should be expected when using ultrasonic treatment.
The increase in solvent concentration above 50% caused a decrease in both total monomeric anthocyanin and total phenolic content extraction. The decrease in the extraction yield was particularly pronounced in the case of TMA. Also, solvent concentration had a significant effect on the extraction yield of both TPC and TMA (p < 0.001). Several studies describe the use of solvent/water mixtures for the extraction of grape by-products; the addition of water increases the permeability of cell tissue, enabling a better mass transfer by molecular diffusion, as well as the recovery of water-soluble compounds [28,54]. Benmeziane et al. [55] studied the extraction of phenolic compounds from fresh table grapes with distilled water, acetone, methanol, and ethanol. For acetone, methanol and ethanol, the concentration was varied as follows: 20%, 40%, 60%, and 80%. The authors noted an increase in total phenol extraction with the increase in solvent concentration. As in the current study the effect of solvent concentration was opposite, the global effect of other extraction parameters should be considered. In regard to the interactions between solvent concentration and other process parameters, a significant effect of the interactions between solvent concentration and ultrasonic frequency, and concentration and temperature was highlighted (Table 3).

Ultrasonic Frequency
In addition to extraction parameters such as solvent, solvent concentration, solid/liquid ratio, extraction temperature and time, which are common to a wide range of extraction techniques, and have therefore been thoroughly studied, in the case of ultrasound assisted extraction the effect of parameters related to ultrasound wave distribution should be considered. In other words, parameters intrinsically related to the ultrasonic devices (such as the frequency, wavelength and amplitude of the wave), the ultrasonic power (in kWh/L), and consequently, intensity, also have an effect on the extraction [56], and their influence on extraction yield needs to be evaluated. In this study, we recorded the variation in the composition of extracts in phenolic compounds and anthocyanins caused by the change of ultrasonic frequency from 12.5 kHz to 25 KHz, and finally to 37.5 kHz. As can be observed in Table 2, regardless of the compounds targeted by the extraction, an increase of ultrasonic frequency from 12.5 kHz to 25 kHz led to a better solubilization of bioactive compounds in the solvent. In the case of total monomeric anthocyanin, the extraction yield increased by approximately 18% when increasing ultrasonic frequency from 12.5 kHz to 25 kHz. Similarly, the total phenolic content of the extract treated with ultrasounds at a frequency of 25 kHz increased by approximately 11%. A further increase of ultrasonic frequency from 25 kHz to 37.5 kHz resulted in a significant reduction of TMA and TPC content. Consequently, to reach the maximum extraction yield, 25 kHz was the most suitable frequency. Other studies on the extraction of phenolic compounds from grape have reported the application of ultrasonic frequencies of 24 kHz on samples obtained from whole berries and 40 kHz on samples of powdered grape peel [57,58].

Temperature and Time
Extraction temperature and time are important parameters to be optimized in order to achieve high recoveries of anthocyanins and total phenolic compounds, and also to minimize the energy cost of the extraction process. The extraction of TMA and TPC was carried out at 50, 60 and 70 • C. According to the data presented in Table 2, the increase in working temperature from 50 • C to 60 • C had a positive effect on the extraction of anthocyanins from grape pomace. The average concentration of total monomeric anthocyanins measured in the extracts at 60 • C was 5.17 mg/g. However, the positive effects of high temperatures on TMA extraction did not extend to temperatures above 60 • C, as the stability of these compounds decreased leading to lower extraction yields. This reduction of anthocyanin recovery is in agreement with findings of previous studies that demonstrated negative effects of high temperatures (>60 • C) on anthocyanin stability [59,60]. While the change in temperature had both positive and negative effects on TMA, in the case of TPC a continuous negative influence of the increase in this process parameter was observed. For the ultrasound assisted extraction of natural antioxidants from Jatropha integerrima, Xu et al. [61] also reported a negative influence of increasing temperature from 40 • C to 80 • C. The reduction in TPC of extracts could be attributed to the degradation of thermolabile and easily oxidizable phenolics caused by simultaneous effects of high temperatures and increased pressures generated when the cavitation bubbles induced by ultrasound reach a critical size [62].
As mentioned before, time is another critical factor in the extraction of phenolic compounds. In this study, the extraction time was gradually increased as follows: 15 min, 30 min and 45 min. In this way, the extraction period needed for achieving maximum yields of total monomeric anthocyanins and total phenolic compounds from grape pomace was about 30 min. The increase in the extraction of targeted compounds at prolonged sonication times has been noted by numerous researchers, and was attributed to the fact that ultrasonic waves affect the mass transfer rate mainly in the solvent penetration stage [63,64]. González-Centeno et al. [65] studied the aqueous ultrasound assisted extraction of antioxidants from grape pomace to investigate the effect of acoustic frequency, ultrasonic power density and extraction time. Regarding the extraction time, researchers reported that the optimal conditions implied a sonication of 25 min. Ghafoor et al. [66] optimized the process variables of UAE of anthocyanins, phenolic compounds and antioxidants from grape seeds, obtaining for the extraction time levels around 30 min, as follows: 29.03 min for phenolic compounds, 30.58 min for antioxidants, and 29.49 for anthocyanins.
Apart from studying the influence of process parameters on extraction yields, statistical analysis was also used for the evaluation of the interactions between factors (Table 3) on extraction yields of TPC and TMA. As shown in Table 3, in the case of TPC, the interactions of solvent type with ultrasonic frequency and of ultrasonic frequency with solvent concentration, and temperature, respectively, had a great impact on extraction efficiency. The impact of these interactions on the extraction yield of phenolic compounds was indicated by the F-values and p-values given by the analysis of variance, as follows: solvent type with ultrasonic frequency-F-value of 27.74, p < 0.001; solvent concentration with ultrasonic frequency-F-value of 12.80, p < 0.001; and ultrasonic frequency with temperature-F-value of 5.82, p < 0.001. For TMA extraction, the yield was significantly influenced by the interactions of solvent type with ultrasonic frequency and temperature, respectively, and the interaction between solvent concentration and temperature. Multifactor ANOVA gave the following results for these interactions: solvent type with ultrasonic frequency-F-value of 71.76, p < 0.001; solvent type with temperature-F-value of 21.54, p < 0.001; and solvent concentration with temperature-F-value of 6.58, p < 0.001.

Extraction Modeling
The response surface methodology was implemented via a three-block Box-Behnken experimental design (Table 4) in order to optimize the ultrasonic-assisted extraction of TMA and TPC from grape pomace.

Total Monomeric Anthocyanin
Experimental data for TMA was fitted to quadratic equations using the RSM, and the equations are presented in Table 5. Statistical parameters, namely, sum of squares, mean square, F-ratio, R-squared and lack of fit values, were calculated for each model. Based on the statistical parameters presented in Table 6, all the models proposed are significant (p < 0.001). The regression analysis revealed that for both extraction solvents (2-propanol and methanol), the process parameters had significant positive linear effects on TMA. Figures 1 and 2 present the 3D evolution of TMA extraction as a function of the parameters employed.  Table 5. Polynomial equations describing the effect of the independent variables considered on the responses (TMA and TPC), for each extraction solvent (2-propanol and methanol).

Total Phenolic Content
The RSM was used for fitting the experimental data with quadratic equations, which are presented in Table 5. According to the values presented in Table 6, all the models proposed are significant (p < 0.001). The regression analysis revealed that for both extraction solvents, 2-propanol and methanol, the above-mentioned process parameters had significant positive linear effects on the total phenol content. The 3D graphs illustrating the evolution of TPC extraction as a function of the parameters employed are presented in Figures 3 and 4.

Raman Spectra Analysis
Raman spectroscopy is a technique that can be used in the analysis of samples from the wine industry, including grape must, wine and by-products, because it is able to provide an estimation of the sample composition in a matter of only a few seconds. Reported applications of Raman spectroscopy in winemaking include the determination of water, ethanol and sucrose content in white wines, and the rapid measurement of phenolic compounds in red wines [67,68]. In this study, To achieve the maximum extraction yield of TPC, the levels of ultrasonic frequency, temperature, solvent concentration and extraction time were optimized. According to the optimization, in order to achieve the highest yield of TPC (62.487 mg gallic acid equivalent (GAE)/g (d = 1.0)), the following conditions are required: solvent-2 propanol, solvent concentration 50%, temperature−50 • C, and extraction time 29.6 min.

Raman Spectra Analysis
Raman spectroscopy is a technique that can be used in the analysis of samples from the wine industry, including grape must, wine and by-products, because it is able to provide an estimation of the sample composition in a matter of only a few seconds. Reported applications of Raman spectroscopy in winemaking include the determination of water, ethanol and sucrose content in white wines, and the rapid measurement of phenolic compounds in red wines [67,68]. In this study, Raman spectroscopy was used to characterize the phenolic compounds contained in two extracts from grape pomace, one with methanol as the extraction solvent, and the other with 2-propanol as the extraction solvent. The levels of other process parameters employed in the extraction of these samples were: solvent concentration of 50%, 37.5 kHz ultrasound frequency, temperature of 60 • C, and 30 min extraction time. The Raman spectroscopy was used to confirm the presence of the phenolics in the extracts.
The Raman spectra of 2-propanol and methanol extracts from grape pomace are shown in Figure 5 (A-extract with 2-propanol, B-methanolic extract). The spectra of the extract with 2-propanol presented main bands in the wavelength regions 2010-1330 cm −1 and 1000-340 cm −1 of the electromagnetic spectrum. In the first region, the peak at 2008 cm −1 is assigned to the stretching vibration of the hydroxyl (O-H) in A, B, and C ring of anthocyanins. The peak at 1638 cm −1 corresponds to the stretching of aromatic C=C, indicating once more the presence of anthocyanins in the extract. The moderate intensity peak at 1378 cm −1 could be attributed to the bending of −CH 3 , while the low intensity peak at 1338 cm −1 marks the presence of flavonol quercetin, as it is assigned to in-plane O-H bends of the C3−OH group in C ring [69]. Flavonoids are also indicated by the stretch of C−C in the aromatic ring, resulting in the peak at 997 cm −1 . The high-intensity peak at 972 cm −1 is attributed to the stretching of C=C in conjugated C−C=C−C groups, specific to hydroxycinnamic acid, resveratrol and various other phenolic compounds. The bands at 862 cm −1 and in the range 800-700 cm −1 are assigned to aromatic C−H out-of-plane bending vibrations, especially in ring A. The peak of moderate intensity at 628 cm −1 indicates C−C deformation. Lastly, the peak at 395 cm −1 corresponds to the bending of C−OH, while the peak at 344 cm −1 indicates C−O stretching in all aromatic rings.
The spectra obtained for the methanolic extract from grape pomace ( Figure 5B) presented main bands in the wavelength region 1340-340 cm −1 . The peak of reduced intensity displayed at 1335 cm −1 and the high-intensity peak at 1226 cm −1 are attributed to C−C stretching in aromatic rings, the latter being specific to the C ring of flavones. The band at 1149 cm −1 is assigned to deformation in CH 2 group [70], while the peak at 970 cm −1 is attributed to C−O stretching. The peak at 790 cm −1 corresponds to stretching of aromatic C=C. The high-intensity band at 714 cm −1 and the moderate intensity bands at 627 cm −1 and 528 cm −1 are assigned to in-plane deformation in rings A, B and C of flavonoids. The high-intensity peak at 495 cm −1 indicates C−C deformation, and the peak at 452 cm −1 is particularly assigned to C−C deformation in ring C of flavones. The last two notable peaks of the spectra correspond to vibrations in aromatic rings, as follows: 406 cm −1 is assigned to ring deformation, while 345 cm −1 indicates ring bending and can also be attributed to in-plane C−O bending. Based on the interpretation of both spectra, it can be concluded that Raman analysis confirms the overall efficiency of methanol in the extraction of total phenolic compounds, and the higher affinity of anthocyanins toward 2-propanol.

Conclusions
This study focused on the ultrasound-assisted extraction of total monomeric anthocyanis (TMA) and total phenolic content (TPC) from grape pomace resulting as a by-product of processing red grapes from a variety cultivated in the northeastern area of Romania (Vitis vinifera Moldova). Two different extraction solvents were used in the extraction, namely, 2-propanol and methanol, and the following process parameters were varied: solvent concentration, ultrasonic frequency, temperature and time. The influence of these experimental parameters on TMA and TPC yields was studied. In TMA extraction, all parameters with the exception of solvent concentration had a positive influence on the extraction yield. In the case of TPC, on the other hand, solvent concentration and temperature had a negative impact on the extraction yield, while the other parameters contributed to the enhancement of the extraction of targeted compounds. For the purpose of reaching the maximum extraction yields of TMA and TPC ultrasonic frequency, temperature, solvent concentration and extraction time were optimized using the RSM methodology. The model predicted that under optimal conditions maximum yields of 7.727 mg/g for TMA and 62.487 mg gallic acid equivalent (GAE)/g for TPC could be achieved. The optimal conditions obtained are based on the design parameters used, but there could be some differences with respect to extraction efficiency if the range of the extraction parameters were changed, taking

Conclusions
This study focused on the ultrasound-assisted extraction of total monomeric anthocyanis (TMA) and total phenolic content (TPC) from grape pomace resulting as a by-product of processing red grapes from a variety cultivated in the northeastern area of Romania (Vitis vinifera Moldova). Two different extraction solvents were used in the extraction, namely, 2-propanol and methanol, and the following process parameters were varied: solvent concentration, ultrasonic frequency, temperature and time. The influence of these experimental parameters on TMA and TPC yields was studied. In TMA extraction, all parameters with the exception of solvent concentration had a positive influence on the extraction yield. In the case of TPC, on the other hand, solvent concentration and temperature had a negative impact on the extraction yield, while the other parameters contributed to the enhancement of the extraction of targeted compounds. For the purpose of reaching the maximum extraction yields of TMA and TPC ultrasonic frequency, temperature, solvent concentration and extraction time were optimized using the RSM methodology. The model predicted that under optimal conditions maximum yields of 7.727 mg/g for TMA and 62.487 mg gallic acid equivalent (GAE)/g for TPC could be achieved. The optimal conditions obtained are based on the design parameters used, but there could be some differences with respect to extraction efficiency if the range of the extraction parameters were changed, taking into account that for some parameters (e.g., temperature, solvent concentration), high concentrations of TPC or TMA could be achieved that were beyond the range studied.