Optimization of Microwave-Assisted Extraction and Matrix Solid-Phase Dispersion for the Extraction of Polyphenolic Compounds from Grape Skin

Abstract

Polyphenols are organic compounds that can be found in food, especially in fruits, vegetables, and their products. It was shown that their presence significantly affects the properties of food products and that the consumption of food rich in phenolic compounds has a beneficial effect on human health. The subjects of this research were polyphenols: anthocyanins, flavonols, and flavan-3-ols in the skin of grapevine variety Regent. Polyphenols from grape skins were extracted via microwave-assisted extraction (MAE) and matrix solid-phase dispersion (MSPD) as unconventional and green techniques. Therefore, the main aim of this work was to optimize the conditions for the extraction of polyphenolic compounds from grape skin using MAE and MSPD. The extracts were analyzed using high-performance liquid chromatography with a diode array detector and fluorescence detector. Analyses showed that MAE was a very effective method for extracting polyphenolic compounds from grape skin with 10 mL of 60% ethanol for 5 min at 40 °C. The best results for the MSPD extraction of polyphenolic compounds from grape skin were obtained with phenyl as an MSPD sorbent with 10 mL of acetonitrile:water 50:50 v/v as an elution solvent. This scientific research can be used for the better use of grapes as a basis for obtaining flavonoids for commercial purposes.

1. Introduction

Polyphenols are a group of compounds of plant origin, the structure of which consists of one or more aromatic rings with one or more hydroxyl groups. They are a common component of the human diet and are found in varying amounts in fruits, vegetables, wine, fruit juices, tea, and coffee. There are several reasons for the identification and determination of polyphenols in grapes. In the last ten years, numerous studies showed that foods containing polyphenols have a positive effect on health, mainly due to the antioxidant effect. Many studies showed that the antioxidant action of these substances in grapes protects against atherosclerosis and coronary heart disease. In addition, their anticancer, anti-inflammatory, anti-aging, and antimicrobial activity were demonstrated. Polyphenols in red grapes are found in the skin, seeds, and pulp. The composition and content of individual polyphenols depend on the grape variety, climatic conditions, and canopy management [1,2,3,4,5,6,7].

The specific polyphenolic composition of plant materials is the subject of numerous studies, including in wine technology and analysis, where polyphenols (anthocyanins, flavonols, and flavan-3-ol) are of exceptional importance. Anthocyanins are important because they are responsible for the red color of wine and there is a possibility of using anthocyanin profiles as chemotaxonomic properties for determining differences between grape species that would allow for the control of wine adulteration. Tannins also contribute to the organoleptic properties of wine and have the greatest impact on the acerbity and bitterness of wine [6,7,8,9]. These reasons are a great incentive for the extraction of polyphenols from grapes. For this purpose, extraction methods such as ultrasonic-assisted extraction (UAE), enzyme-assisted extraction (EAE), superheated water extraction, and mostly solid-liquid extraction (SLE) are used [1,2,3,4,5,6,7,10,11,12,13,14,15]. Despite the fact that microwave-assisted extraction (MAE) is not a new extraction technique, its application in the analysis of grape phenolics is rare. This technique can be very beneficial in terms of the usage of the low volume of extraction solvent and the short duration of extraction. The matrix solid-phase dispersion (MSPD) is a well-accepted technique for the extraction of various analytes from different solid samples. One of the most beneficial characteristics of this technique is its very high selectivity and the obtained extracts are almost clean of the interfering substances. Despite this, MSPE was not previously used in the analysis of grape phenolics [16,17,18].

The classic approach to the optimization of extraction methods involves a step-by-step method, which is long-lasting and cannot provide insight into the interaction between different extraction factors, which can be crucial for achieving the highest possible yield of analytes. The response surface methodology is a set of mathematical and statistical techniques based on fitting a polynomial equation to experimental data, which must describe the behavior of a set of data using statistical prediction. There are numerous designs for quadratic response surfaces, of which the most used is the Box-Behnken experimental design (BBD). The goal of applying the BBD is to simultaneously optimize the levels of several factors while achieving the best system efficiency and conducting a small number of experiments. The BBD also takes into account the interaction effects of individual factors. Optimization through a BBD consists of the following steps: the selection of independent factors; conducting experiments following the selected experimental matrix (factors are adjusted to three levels so that the distances between these levels are equal); mathematical-statistical processing of experimentally obtained data using a polynomial function; evaluating the suitability of the obtained model; taking into account the p-value of the model, which must be less than 0.05; finding the p-value of the lack of fit, which must be greater than 0.05; finding the coefficient of determination R², adjusted R², and predicted R², whose values must be as close as possible to 1; and obtaining optimal values for each tested factor. When it is necessary to simultaneously consider several obtained responses and when it is necessary to find an appropriate compromise between all responses, the Derringer function or desirability function is applied [19].

Despite the fact that MAE is not a new technique, its application in the extraction of phenolic compounds from grape skins is very rare. There are only a few publications in which MAE is applied for the analysis of only one group of flavonoids, namely, anthocyanins [20,21], while in the most recent study, the authors optimized the MAE method, but anthocyanins-3,5-O-diglucosides were not included. Thus, in the present study, we carried out comprehensive optimization of the major flavonoid groups, namely, anthocyanins, flavonols, and flavan-3-ol, contained in grape skins in terms of the type of organic solvent and their content in the extraction solvents and the extraction time and temperature using BBD. To the best of our knowledge, the MSPD technique has not been used for the analysis of grape skin polyphenolics. Considering some benefits of this technique, especially in terms of obtaining a clean extract without interfering compounds, the aim of this study was the optimization of the MSPD method in terms of the type of sorbent, type of organic solvent, and their content in the extraction solvents, elution volume, and sorbent-to-sample ratio as major factors that could influence the extraction yield. The optimization was done via applying the BBD.

2. Materials and Methods

2.1. Chemicals, Grape Samples, and Preparation

Acetonitrile (ACN) of HPLC grade was obtained from J. T. Baker (Deventer, The Netherlands). Formic acid (FA) and 85% orthophosphoric acid were purchased from Sigma (St. Louis, MO, USA). Ethanol (EtOH) was provided by Kemika (Zagreb, Croatia).

Grape samples of Vitis vinifera L. cv. Regent were obtained from the Experimental Station, Jazbina, Faculty of Agriculture, University of Zagreb, Croatia. Grapes were harvested in a state of full ripeness and immediately separated from the stalk. To obtain homogenous samples of the berries at a similar level of ripeness (sugar and flavonoid concentration), a simple flotation method was used with sucrose water solutions of different densities. For further analysis, grape berries with a density range of 1.088 to 1.099 g mL⁻¹ were selected. The berry skins were manually removed from the pulp and freeze-dried. The dry skins were ground (Mini G, SPAX SamplePrep, Metuchen, NJ, USA) and the powder obtained was stored in a glass container (−20 °C).

2.2. HPLC Analysis

The separation, identification, and quantification of flavonoids from grape skin extracts were performed on an Agilent 1100 Series system (Agilent, Waldbronn, Germany). The separation was performed according to the method described by Tomaz and Maslov [3] with a reversed-phase Luna Phenyl-Hexyl column (4.6 × 250 mm; 5 μm particle (Phenomenex, Torrance, CA, USA)) and a Phenyl guard column (4.0 × 3.0 mm) at a set temperature of 50 °C. The solvents were water:phosphoric acid (99.5:0.5 v/v, eluent A) and acetonitrile:water:phosphoric acid (50:49.5:0.5 v/v/v, eluent B), and the flow rate was 0.9 mL min⁻¹. The injection volume for all samples was 20 μL. The DAD was set to an acquisition range of 200 to 700 nm. Using the DAD, flavonol glycosides were detected at 360 nm, and anthocyanins were detected at 518 nm. Using FLD, flavan-3-ols were detected at λex = 225 nm and λem = 320 nm. The resulting contents of the 3-O-glucoside and 3,5-O-diglucoside of delphinidin, cyanidin, peonidin, and malvidin are expressed as their sum (anthocyanin contents); those of 3-O-glycoside quercetin, myricetin, and kaempferol are expressed as their sum (flavonol contents); and those of gallocatechin, procyanidin B1, procyanidin B2, procyanidin B3, procyanidin B4, catechin, epicatechin, and epigallocatechin are expressed as their sum (flavan-3-ol contents) when used as responses (Y, dependent variables).

2.3. Spectrophotometric Analysis

2.3.1. Determination of the Total Phenolics

The total phenolics were determined using the method described by Singleton and Rossi [22]. In brief, 20 μL extract, 6 mL water, and 500 μL Folin-Ciocaulteu reagent were added to the 10 mL volumetric flasks. The mixture was left for 8 min; after that, 2 mL of 20% sodium carbonate solution was added. The solutions were brought to a final volume of 10 mL with water. The obtained reaction mixture was incubated at 20 °C for 2 h. The absorbance was determined at 765 nm. The results were expressed in equivalents of gallic acid (GAE)/mg kg⁻¹, based on the previously constructed calibration curve.

2.3.2. Determination of the Total Anthocyanins

Determination of total anthocyanins was performed using the method described by Ribéreau-Gayon [23]. In brief, 100 μL extract, 100 μL of 0.1% HCl prepared in ethanol, and 2 mL of 2% HCl prepared in water were added to the test tube. The resulting solution was then divided into two test tubes (1 mL in each). To the first test tube (A1), 400 μL of water was added, while in the second test tube (A2), 400 μL of 15% sodium bisulfite was added. The reaction mixture was incubated at room temperature for 20 min. The absorbance was read at 520 nm. The total content of anthocyanins was calculated based on the following equation γ(anthocyanin)/mgL⁻¹ = 640 × (A2 − A1).

2.4. Microwave-Assisted Extraction

All experiments were performed in a microwave extractor with a microwave power of 600 W. After being extracted, the mixture was centrifuged at 3600 rpm for 15 min. After the extractions were completed, supernatants were collected and concentrated under a vacuum to remove the organic modifier (40 °C) on a Hei-Vap Advantage G3 rotary evaporator (Heidolph, Schwabach, Germany) and brought to a final volume of 10 mL with eluent A.

2.4.1. Effect of the Extraction Time on the Recovery

A mixture of ACN:H₂O:FA = 20:79:1 v/v/v was used as the extraction solvent to determine the optimal extraction time range. The ratio of the weight of the sample and the volume of the extraction solvent (phase ratio) was 1:80 g mL⁻¹, i.e., the weight of the sample was 125 mg with a solvent volume of 10 mL, and all experiments were performed at 50 °C in one extraction step. The extraction times were 5, 10, 30, 45, and 60 min.

2.4.2. Determination of the Phase Ratio

All extractions were performed in one extraction step with a mixture of ACN:H₂O:FA = 20:79:1 v/v/v as an extraction solvent at a temperature of 50 °C for 15 min, while the phase ratios (solid-to-solvent ratios) were changed (

Table 1. Phase ratios, extraction solvent volumes, and sample weights used.

Phase Ratio (g mL−1)	Volume of Extraction Solvent (mL)	Sample Weight (mg)
1:40	10	250
1:60	10	165
1:80	10	125

).

2.4.3. Optimization of the Microwave-Assisted Extraction Using a Box-Behnken Experimental Design

To optimize the MAE method, the following extraction conditions were used, which were constant in all performed experiments: sample weight of 125 mg, the volume of extraction solvent of 10 mL, and one extraction step. Besides water and a certain organic solvent in given proportions, the extraction solvents also contained 1% formic acid. The effect of three numerical factors (organic solvent content in the extraction solvent, extraction temperature, and extraction time) and one categorical factor (type of organic solvent) on the content of anthocyanins, flavonols, and flavan-3-ol was examined using a BBD. These three independent numerical factors were tested at three levels, while the categorical factor was tested at two levels (

Table 2. Independent factors and their levels that were used in the BBD to optimize the MAE method.

Factors	Factor Levels
Coded levels	−1	0	1
A: Organic solvent content (%)	20	50	80
B: Extraction temperature (°C)	40	50	60
C: Extraction time (min)	5	10	15
D: Type of organic solvent	ACN	EtOH

); thus, 30 experiments with different extraction conditions were performed.

2.5. Matrix Solid-Phase Dispersion

2.5.1. Solid Phase Selection

Sea sand, Amberlite XAD-2, and Sepra Phenyl materials with different properties were used as solid phases (

Table 3. Properties of the solid phases used.

Solid Phase	Composition	Particle Size (µm)	Pore Size (Å)
Sea sand	Silica	100 to 315
Amberlite XAD-2	Styrene-divinylbenzene	250 to 841	90
Sepra Phenyl	Phenyl silica	50	65

).

2.5.2. Optimization of the Matrix Solid-Phase Dispersion Using a Box-Behnken Experimental Design

In all the experiments performed, the weight of the powder sample of grape skin was 125 mg, which was mixed with a certain weight of a certain solid phase in an agate mortar. After homogenization, the mixture was quantitatively transferred to a column, at the bottom of which a porous polyethylene disk was previously placed. After filling the column, the same porous disk was placed on top and then lightly pressed with a plastic syringe plunger. The prepared columns were placed on a vacuum extraction station while the flow of elution solvent was adjusted to 1 mL min⁻¹. Regardless of the total volume of elution solvent, elution was first performed with 2 mL of solvent and the system was left for 5 min to soak the layer. Thereafter, elution was performed with the remaining volume of solvent. After the MSPD procedure, supernatants were collected and concentrated under a vacuum to remove the organic modifier (40 °C) on a Hei-Vap Advantage G3 rotary evaporator (Heidolph, Schwabach, Germany) and brought to a final volume of 10 mL with eluent A. To optimize the MSPD method, the elution solvents, in addition to water and a certain organic solvent in given proportions, also contained 1% formic acid. Using a BBD, the effect of three numerical factors (organic phase content in the elution solvent, the volume of elution solvent, and sorbent to sample ratio) and one categorical factor (type of organic solvent) for all three used solid phases were used. These three independent numerical factors were tested at three levels, while the categorical factor was tested at two levels (

Table 4. Independent factors used and their levels in the BBD to optimize the MSPD method.

Factors	Factor Levels
Coded levels	−1	0	1
A: Organic solvent content (%)	20	50	80
B: Volume of elution solvent (mL)	4	7	10
C: Sorbent to sample ratio	3	4	5
D: Type of organic solvent	ACN	EtOH

), with a total of 90 experiments, i.e., 30 experiments for each solid phase with different extraction conditions.

2.6. Experimental Design and Statistical Analysis

Three numerical factors (organic solvent content, the extraction temperature, and the extraction time) and one categorical factor (the type of organic solvent) were investigated in the case of MAE optimization, while in the case of MSPD optimization, three numerical factors (organic solvent content, the volume of elution solvent, and sorbent-to-sample ratio), and one categorical factor (the type of organic solvent) were investigated. In both optimization cases, these factors’ effects on the resulting contents of the 3-O-glucoside and 3,5-O-diglucoside of delphinidin, cyanidin, peonidin, and malvidin were expressed as their sum (anthocyanin contents); those of 3-O-glycoside quercetin, myricetin, and kaempferol were expressed as their sum (flavonol glycoside contents); and those of gallocatechin, procyanidin B1, procyanidin B2, procyanidin B3, procyanidin B4, catechin, epicatechin, and epigallocatechin were expressed as their sum (flavan-3-ol contents). These independent factors were investigated at three different coded levels (Table 2 and Table 3). The results of the BBD experiments were analyzed using non-linear multiple regression with backward elimination to fit the following second-order equation to the dependent Y variables:

Y = β₀ + Σβ_ix_i + Σβ_ijx_ix_j + Σβ_iix_i²   (i = 1, 2…k)

(1)

β₀, β_i, β_ii, and β_ij are a constant and coefficients for linear, quadratic, and interaction effects, respectively, while x_i and x_j are the levels of independent factors in coded values. Coefficients were interpreted using an F-test. To establish the optimal conditions for individual anthocyanin, a flavonol glycoside, and flavan-3-ol contents, analysis of variance (ANOVA), regression analysis, and plotting of the response surface plot were conducted. For optimization, a multicriteria methodology (Derringer function or desirability function) was used. This methodology is applied when various responses must be considered at the same time and it is necessary to find optimal compromises between the total number of considered responses. The analysis of the experimental design and calculation of the predicted data was completed using the Design Expert software (Stat-Ease Inc., Minneapolis, MN, USA).

3. Results and Discussion

Despite the applied extraction technique, 21 individual flavonoids were identified in total. The most characteristic chromatograms of the analyzed extracts are presented in Figure 1.

3.1. Microwave-Assisted Extraction Optimization

Due to the very low efficiency of the SLE and UAE extraction methods when using acetone as the organic phase and the formation of new compounds, only ethanol and acetonitrile were used as organic phases in the extraction solvent in this series of experiments [4,19]. The optimal range of extraction time, as well as the solid-to-solvent ratio, was determined via the method of studying the effect of one factor on the final result while the other factors were held constant.

3.1.1. Extraction Time Range Selection

The duration of the extraction process is an important factor that significantly affects the content of flavonoids (Figure 2).

In the period between 5 min and 10 min, a significant increase in the content of all examined groups of flavonoids was observed. In the period between 10 min and 60 min, the decrease in the content of flavonols and flavan-3-ol was small, while the decrease in the content of anthocyanins was extremely large, especially after 30 min, which could be attributed to their thermal instability. To implement the optimization of the method using an experimental design, a time range from 5 min to 15 min was used.

3.1.2. Influence of the Phase Ratio on the Recovery of Flavonoids

Figure 3 shows the dependence of the content of anthocyanins, flavonols, and flavan-3-ol on the phase ratio.

The increase in the phase ratio led to a significant increase in the content of the tested flavonoids and they were the largest in the extracts obtained under conditions when the phase ratio was 1:80 g mL⁻¹.

3.1.3. Optimization of the MAE Method Using BBD

The type of organic solvent and its content in the extraction solvent, as well as the time and extraction temperature, were used as extraction factors (Table 2). As part of the BBD optimization procedure, 30 experiments were performed, whose results were processed, and it was found that the obtained content of anthocyanin, flavonol, and flavan-3-ol could best be described by a quadratic polynomial equation.

Table 5. Coefficients of the second-order polynomial equation and regression response coefficients and ANOVA parameters for the obtained models.

Term	Anthocyanins		Flavonols		Flavan-3-ol		Total Anthocyanins		Total Phenolics
	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value
Model		<0.0001		0.0002		0.0004		<0.0001		0.0002
Lack of fit		0.2654		0.8269		0.6455		0.3746		0.2036
Intercept	11,872.21		1453.65		254.4		13,021.57		16,990.54
A: Organic solvent content	−310.37	0.0105	53.07	0.0018	−1.13	0.853	365.75	0.0018	452.3	0.0048
B: Extraction temperature	−263.62	0.0169	−29.11	0.0433	27.87	0.0001	−279.87	0.009	−329.31	0.0196
C: Extraction time	25.7	0.7954	−0.97	0.9431	−4.46	0.4416	42.5	0.6588	28.05	0.8257
D: Type of organic solvent	−115.29	0.1238	−6.02	0.5432	12.03	0.0093	−124.62	0.0899	−139.28	0.1461
AB	486.32	0.0049	36.31	0.0838	−30.01	0.0022	514.73	0.002	621.52	0.0051
AC	−59.36	0.6898	34.8	0.0997	12.1	0.1677			−29.89	0.8755
AD	75.96	0.4822	23.15	0.1218	14.04	0.0313	107.89	0.2952	134.94	0.3353
BC	−15.07	0.9086	−15.29	0.4023					−22.4	0.8944
BD	−438.88	0.0233					−262.29	0.0134	−322.12	0.022
CD	−167.94	0.1062	−14.33	0.2989			−170.5	0.0885	−233.44	0.0829
A2	−843.68	<0.0001	−44.45	0.0353	−7.52	0.3736	−919.19	<0.0001	−1134.46	<0.0001
B2	41.63	0.7768	19.18	0.3457	13.68	0.1208			95.71	0.6133
C2	−101.92	0.4909							−125.53	0.5089
R2	0.8626		0.8272		0.7642		0.8562		0.8729
Adapt R2	0.8351		0.8097		0.7462		0.7843		0.7548
Precision	10.73		7.89		9.01		14.432		11.179

shows the obtained coefficients for individual members of the quadratic equation, together with the parameters of the performed analysis of variance.

The type of organic solvent in the extraction solvent was a significant factor only in the case of flavan-3-ol. The proportion of organic phase in the extraction solvent was a factor that significantly affected the content of anthocyanins and flavonols, while its effect on the content of flavan-3-ol was negligible. The best organic solvent for the extraction of anthocyanins and flavan-3-ol was ethanol, while the highest content of flavonols was observed in the extracts obtained with acetonitrile as the organic modifier. This result may have been due to the different dielectric properties of aqueous solutions of ethanol and acetonitrile and the different positions of the examined groups of flavonoids within the cells of grape skins. Unlike flavonols, anthocyanins can be contained not only in vacuoles but also in anthocyanin inclusion vacuoles located in cell vacuoles; therefore, for their excretion from the cell, it is necessary to break down two lipid membranes, as well as the cell wall. Flavan-3-ol can be an integral part of cell walls; therefore, during their extraction, it is necessary to destroy the structure of the cell wall. The efficiency of a particular solvent during the implementation of MAE is determined by its dissipation factor. A higher value of this factor indicates a better ability to convert microwave energy into heat. In general, ethanol and its aqueous solutions have significantly higher values of dissipation factors compared with acetonitrile and its aqueous solutions and, therefore, more efficiently translate microwave energy into heat [24,25]. A local temperature rise has a positive effect on the denaturation and degradation of cell membranes, as well as cell walls, which can ultimately lead to the easier extraction of anthocyanins and flavan-3-ol using aqueous ethanol solutions.

The extraction temperature had a significant effect on the extraction efficiency of all examined groups of flavonoids. The highest content of anthocyanins and flavonols was determined in extracts obtained at lower temperatures. This observation may have been due to their thermal instability at temperatures above 50 °C, as well as the fact that when applying microwave radiation inside the pores of the solid sample soaked in the extraction solvent, there may be significantly higher local temperatures than in the surrounding solvent. An increase in temperature led to an increase in the content of flavan-3-ol.

The significant interaction effects of individual extraction factors are shown in Figure 4. Increasing the organic solvent content to 50%, regardless of whether it was acetonitrile or ethanol, with a simultaneous decrease in temperature led to a significant increase in the content of anthocyanins (Figure 4a,b). A similar trend was followed by flavonols, but to achieve higher content, it is necessary to increase the organic solvent content to 65% (Figure 4g,h). Decreasing the organic solvent content with a simultaneous increase in temperature had a positive effect on the extraction efficiency of flavan-3-ol (Figure 3m,n). To achieve a higher content of anthocyanins and flavonols, extraction should be performed at low temperatures (40 °C) for 10 min and 15 min, respectively (Figure 4e,f,k,l). This observation can be explained by the thermal instability of these flavonoids. Despite the positive effect of elevated temperature on solvent viscosity and an increase in the diffusion rate, high temperatures during microwave energy application had a markedly negative effect on anthocyanin and flavonol contents, which may have been due to a local temperature rise.

The obtained optimal extraction conditions using the Derringer function are shown in

Table 6. Optimal MAE extraction conditions with their predicted and experimentally obtained values for individual groups of flavonoids.

Group	Organic Solvent	Organic Solvent Content (%)	Temperature (°C)	Time (min)	Estimated Value (mg kg−1)	Obtained Value (n = 3) (mg kg−1)
Anthocyanins	EtOH	50	40	7	12,349.40	12,331.64 ± 21.14
Flavonols	ACN	60	40	15	1564.15	1557.32 ± 19.29
Flavan-3-ol	EtOH	20	60	5	333.09	228.20 ± 8.66
Optimal conditions	EtOH	60	40	5
Anthocyanins					12,228.08	12,545.19 ± 162.51
Flavonols					1514.27	1514.06 ± 20.70
Flavan-3-ol					270.26	266.83 ± 24.47
Total anthocyanins	EtOH	50	40	7	13,567.60	13,501.31 ± 16.11
Total phenolics	EtOH	50	40	7	17,665.10	17,600.14 ± 19.36

. For individual groups of flavonoids, the optimal extraction conditions differed significantly. When optimizing the method for all three groups of flavonoids, it is necessary to make significant compromises, and therefore, it is to be expected that the content of individual flavonoids in the case of final conditions is significantly lower than in the case of optimal conditions obtained for individual cases.

The accuracy and suitability of the obtained models were confirmed by comparing the predicted models with those obtained experimentally. The optimal conditions obtained for anthocyanins as determined using HPLC and the spectrophotometric method were the same, as well as those obtained using the Folin-Ciocaulteu method.

3.2. Optimization of Matrix Solid-Phase Dispersion

3.2.1. Solid Phase Selection

Regarding the solid phase, those that contain silica as a carrier to which various groups, such as octyl, octadecyl, and phenyl groups are attached, are most often used. Recently, an increasing number of studies indicated the high efficiency of the application of phases composed of phenyl silica in the separation of different groups of phenols. The separation of analytes possessing aromatic rings with phenyl groups of the solid phase is based on the formation of π-π interactions and dispersive hydrophobic forces, and therefore, such phases are more efficient materials for separating phenolic compounds than conventional phases used in a reverse phase system, such as octadecyl silica or octa silica [26,27]. In addition to the above materials, various polymeric materials composed of polystyrene-divinylbenzene are very often used for phenol separation. Such a group also includes various Amberlite XAD resins. According to a European Commission regulation, the use of these solid phases is allowed in the food, pharmaceutical, and cosmetic industries [28,29,30]. The big disadvantage of solid phases with bound phases and polymer phases is their high price. Sea sand is a very inexpensive material that has a wide range of applications in destroying plant specimens. Although it is considered not to act as an analyte solvent during the mixing process, due to its structure in the form of very sharp edges and a rough surface, it allows for shearing during the mechanical mixing of samples and the solid phase, making it a very effective means of destroying solid sample cells. Some studies showed that sea sand is a more efficient solid phase for phenol extraction compared with various C18 materials [31,32,33]. Based on the above findings, Sepra Phenyl, Amberlite XAD-2, and sea sand were used as solid phases in optimizing the MSPD process.

3.2.2. Optimizing the MSPD Method Using the BBD

During the optimization process of the MSPD method, the type of solid phase, type and volume content of organic phase in the elution solvent, volume of elution solvent, and sorbent-to-sample ratio with the levels given in Table 4 were used as extraction factors, with a total of 90 experiments. Statistical processing of the obtained content of anthocyanins, flavonols, and flavan-3-ol showed that they can best be described using a quadratic polynomial equation, whose coefficients, together with other parameters determined using an analysis of variance, are shown in

Table 7. Coefficients of the second-order polynomial equation and regression coefficients of the response and ANOVA parameters for the obtained models.

Term	Anthocyanins		Flavonols		Flavan-3-ol		Total Anthocyanins		Total Phenolics
	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value
Model		<0.0001		<0.0001		<0.0001		<0.0001		<0.0001
Lack of fit		0.5051		0.3275		0.3090		0.5551		0.4449
Intercept	9187.87		1368.69		128.41		10,106.65		11,622.65
A: Organic solvent content	214.58	0.1513	38.50	0.0320	−1.31	0.4109	236.04	0.1513	271.45	0.161
B: Volume of elution solvent	549.74	0.0005	109.51	<0.0001	4.96	0.0019	604.72	0.0005	693.88	0.0007
C: Sorbent-to-solvent ratio	−685.93	<0.0001	−27.46	0.1171	−3.79	0.0207	−754.53	<0.0001	−869.25	<0.0001
D: Type of organic solvent	381.67	0.0007	4.60	0.6613	3.95	0.0009	419.84	0.0007	481.99	0.0009
E [1]: Solid phase	2237.67	<0.0001	130.11	<0.0001	25.43	<0.0001	2461.43	<0.0001	2831.47	<0.0001
E [2]: Solid phase	−4666.75	<0.0001	−239.25	<0.0001	−49.38	<0.0001	2669.79	<0.0001	3068.62	<0.0001
AB									−102.79	0.7054
AC	411.02	0.0534					452.12	0.0534	519.94	0.0592
AD [1]	775.52	<0.0001	96.58	<0.0001	2.15	0.1809	853.08	<0.0001	981.04	<0.0001
AE [1]	70.09	0.0013	−5.32	<0.0001	−0.47	0.0023	77.1	0.0013	88.66	0.0017
AE [2]	−727.73	0.0013	−102.39	<0.0001	−6.82	0.0023	723.41	0.0012	831.92	0.0017
BC									−35.53	0.899
BD [1]					−1.93	0.2354			105.21	0.59
BE [1]	125.23	0.2595	−45.65	0.0280	−5.28	<0.0001	137.75	0.2595	159.96	0.2796
BE [2]	−346.10		67.00	0.0280	10.65	<0.0001	242.95	0.2588	276.32	0.2633
CD [1]			−13.21						−127.16	0.5151
CE [1]	613.00	<0.0001	37.01	0.0252	5.92	0.0093	674.3	<0.0001	776.98	<0.0001
CE [2]	−1333.85	<0.0001	−69.66	0.0252	−6.36	0.0093	792.93	<0.0001	908.8	<0.0001
DE [1]	−482.74	<0.0001	−11.68	<0.0001	−2.20	<0.0001	−531.02	<0.0001	−609.85	<0.0001
DE [2]	913.01	<0.0001	92.42	<0.0001	8.87	<0.0001	−473.29	<0.0001	−545.93	<0.0001
A2	−1210.99	<0.0001	−199.48	<0.0001	−10.92	<0.0001	−1332.09	<0.0001	−1530.36	<0.0001
B2	−361.33	0.1036			−4.44	0.0621	−397.46	0.1036	−458.62	0.1108
C2	280.14	0.2052					308.15	0.2052	352.84	0.2181
R2	0.9425		0.9392		0.9368		0.9425		0.9433
Adapt R2	0.9266		0.9035		0.9205		0.9266		0.9232
Precision	27.44		21.60		26.09		27.442		24.099

.

The most significant extraction factor that influenced the content of all tested groups of compounds was the applied solid phase. The most effective solid phase was phenyl silica, which had a dual role during application in the MSPD process. It served as an abrasive to destroy the cell wall structure of grape skins, but it also acted as a kind of solvent for flavonoids. Bound-phase phenyl and flavonoids possess aromatic rings in their structures and, in addition to the usual hydrophobic interactions, can also form strong π-π interactions, which ultimately result in high efficiency during the application of the MSPD process. The content of anthocyanins, flavonols, and flavan-3-ol obtained by using sea sand as the solid phase was slightly lower compared with those obtained by using phenyl silica. Due to its structure, the sand was very effective at destroying the cell wall structures of grape skins. Comprehensive destruction of cell walls results in easier access of the elution solvent to the cell components, and thus, an increase in extraction. The content of individual groups of flavonoids obtained using XAD-2 as a solid phase was significantly lower compared with those obtained using phenyl silica and sand. The largest differences were observed in the case of anthocyanins, while the least was observed regarding the content of flavonols. Such a result is in line with previous studies that found that this type of solid phase has the highest affinity for flavonols and an extremely low affinity for anthocyanins, while the affinity for flavan-3-oils was moderate [34,35].

The choice of elution solvent plays an important role during the MSPD process. It has the role of a common mobile phase in the chromatographic system during the MSPD, but it also has a role as a solvent for the desired analytes. The type of organic phase contained in the elution solvent is an important extraction factor that significantly affects the content of individual groups of flavonoids. Anthocyanins are the most abundant group of flavonoids in grape skins, and ethanol must be used for their most efficient extraction using MSPD. Methanol, in contrast to acetonitrile, is thought to enhance the π-π interactions between the anthocyanin and the phenyl-bound phase, resulting in better separation. Since ethanol and methanol have very similar physicochemical properties, this property of methanol can probably be applied to ethanol as well [36]. To achieve the maximum content of flavonol and flavan-3-ol, it is necessary to use an elution solvent containing acetonitrile.

The volume of the elution solvent was an important factor that influenced the extraction efficiency of anthocyanin, flavonol, and flavan-3-ol. The volume of the elution solvent depended on the weight of the sorbent and the content of the analyte. The use of larger volumes of elution solvent ensured complete elution of the desired analytes.

The content of anthocyanin and flavan-3-ol was a function of the sorbent-to-solid ratio. Increasing this ratio resulted in a significant increase in the content of anthocyanins, while decreasing this ratio had a positive effect on flavan-3-ol extraction. Such an observation may have been due to the content of these groups of compounds. Anthocyanins were present in significantly higher contents; therefore, their dissolution in the bound phase required a higher weight because the application of smaller weights led to supersaturation, and thus, the solvent could not further dissolve these analytes. Flavan-3-ols were contained in much smaller contents; therefore, the application of a larger weight in the solid phase could have a negative effect on their extraction because they could lead to their stretching in the solid phase, and thus, to loss in the final extract.

The effect of interactions between individual extraction factors on the content of anthocyanins, flavonols, and flavan-3-ol was determined from contour graphs (Figure 4, Figure 5 and Figure 6).

The content of anthocyanin, regardless of the applied solid phase and the elution solvent, was a function of the interaction between the sorbent-to-sample ratio and the proportion of the organic phase or the volume of the elution solvent. Increasing the proportion of the organic phase with the application of a smaller sorbent-to-sample ratio led to a significant increase in the content of this group of compounds. The highest content of anthocyanins was observed in the extracts obtained by applying a small sorbent-to-sample ratio and a large elution volume. The elution solvent acted as a solvent that competed with the bound phase for the analytes, and therefore, a larger volume of solvent compensated for the small weight of the bound phase (Figure 5).

The content of flavonols, regardless of the elution solvent used and the type of solid phase was influenced by the interactions between the organic phase content in the elution solvent and the sorbent-to-sample ratio or volume. The applied volume of solvent did not significantly depend on the applied solid phase, and its increase led to an increase in the content of flavonols (Figure 6).

As in the case of anthocyanins, the content of flavan-3-ol depended on the interaction between the sorbent-to-sample ratio and the volume of the elution solvent or its content. The proportion of the organic phase in the extraction solvent, as well as its volume, depended on the organic solvent used (Figure 7).

The optimization of the MSPD procedure was performed using the Derringer function. The obtained optimal conditions for the extraction of all the individual flavonoids are listed in

Table 8. Optimal conditions of the MSPD-predicted and experimentally obtained values for individual groups of flavonoids.

Group	Solid Phase	Organic Phase	Organic Phase Content (%)	Volume of Elution Solvent (mL)	Sorbents to Sample Ratio	Estimated Value (mg kg−1)	Obtained Value (n = 3) (mg kg−1)
Anthocyanins	Phenyl silica	EtOH	80	9	5:1	13,098.32	13,093.54 ± 93.84
Flavonols	Phenyl silica	ACN	50	10	5:1	1632.24	1643.38 ± 1.42
Flavan-3-ol	Phenyl silica	ACN	50	7	3:1	159.03	151.45 ± 2.65
Optimal conditions	Phenyl silica	ACN	50	10	3:1
Anthocyanins						12,296.37	12,870.70 ± 15.70
Flavonols						1615.81	1571.31 ± 6.07
Flavan-3-ol						157.54	160.33 ± 9.54
Total anthocyanins	Phenyl silica	EtOH	80	9	5:1	14,088.80	14,107.36 ± 89.46
Total phenolics	Phenyl silica	EtOH	80	10	5:1	16,269.30	16,300.18 ± 100.66
Optimal conditions	Sand	EtOH	60	10	5:1
Anthocyanins						12,201.20	12,175.03 ± 19.97
Flavonols						1586.29	1591.87 ± 9.12
Flavan-3-ol						149.94	150.96 ± 7.30

. Since the use of sea sand as a solid phase and ethanol as an organic solvent is allowed in the food, pharmaceutical, and cosmetic industries, optimal conditions were determined in the case of the use of these extraction conditions.

By conducting experiments under the obtained optimal conditions and comparing the values predicted by the model with those obtained experimentally, the suitability of the obtained equations for individual models was determined. The differences between the experimentally obtained values and the values predicted by the model were very small, thus confirming that the obtained models were reliable and accurate. The optimal conditions obtained for anthocyanins determined using HPLC and the spectrophotometric method were the same, as well as those obtained using the Folin-Ciocaulteu method because the anthocyanins represented a great majority of the analyzed flavonoids.

The SLE technique is the most applied technique for the extraction of polyphenolic compounds from grape skins. Methanol, EtOH, acetone, ethyl acetate, and their aqueous solutions are the most frequently applied extraction solvents for the recovery of polyphenolics from grapes. The extraction time can be in the range from 8 min to 48 h, while the phase ratio can be in the range from 1:1.5 up to 1:125 g mL⁻¹. The application of these newly optimized methods, namely, MAE and MSPD, enables the use of smaller volumes of solvent compared with SLE, as well as a shorter duration of the extraction procedure.

4. Conclusions

The MAE and MSPD are well-established extraction techniques, and applications in the analysis of grape skin polyphenols are not common. These methods were shown to be fast and efficient for the analysis of polyphenolic compounds from grape skins. A total of 21 flavonoids were analyzed and identified using MAE and MSPD methods. Since these techniques are very sensitive to experimental conditions, the MAE and MSPD were optimized by employing a BBD. For the MAE, the optimal extraction conditions were an extraction temperature of 40 °C, an extraction time of 5 min, and a solvent composition of ethanol:water 60:40 v/v. The optimal MSPD conditions were phenyl silica as the solid phase, a solid-to-sample weight ratio of 3:1, a solvent composition of acetonitrile:water 50:50 v/v, and an elution volume of 10 mL. The application of the optimized methods provided a powerful tool for the establishment of the phenolic profile of grape skins. The optimized methods were conducted in a single pot, which greatly reduced sample loss and improved the RSD. Furthermore, the application of these methods was simple and cheap and required a small amount of sample and extraction solvents.