A Sensitive LC-MS Method for Anthocyanins and Comparison of Byproducts and Equivalent Wine Content

Anthocyanins are a group of phenolic compounds with great importance, not only because they play a crucial role in a wine’s quality, but also due to the fact that they can have beneficial effects on human health. In this work, a method was developed for the detection and identification of these compounds in solid wastes of the wine-making industry (red grape skins and pomace), using liquid-liquid extraction (LLE) prior to the liquid chromatography-mass spectrometry technique (LC-MS). The complete process was investigated and optimized, starting from the extraction conditions (extraction solution selection, dried matter-to-solvent volume ratio, water bath extraction duration, and necessary consecutive extraction rounds) and continuing to the mobile phase selection. The extraction solution chosen was a methanol/phosphoric acid solution (95/5, v/v), while three rounds of consecutive extraction were necessary in order to extract the maximum amount of anthocyanins from the byproducts. During the LC-MS analysis, acetonitrile was selected as the organic solvent since, compared with methanol, not only did it exhibit increased elution strength, but it also produced significantly narrower peaks. To enable accurate identification of the analytes and optimization of the developed method, kuromanin chloride and myrtillin chloride were used as standards. Furthermore, the wine variety (Syrah) from which the specific byproducts were produced was analyzed for its anthocyanin content, leading to interesting conclusions about which anthocyanins are transferred from grapes to wine during the vinification procedure, and to what extent. The results of this study showed that the total concentration of anthocyanins estimated in wine byproducts exceeded almost 12 times the equivalent concentration in Syrah wine, while the four categories of detected anthocyanins, simple glucosides, acetyl glucosides, cinnamoyl glucosides, and pyroanthocyanins, were present in different ratios among the two samples, ranging from 18.20 to 1, to 5.83 to 1. These results not only confirmed the potential value of these byproducts, but also indicated the complexity of the anthocyanins’ transfer mechanism between a wine and its byproducts.


Introduction
Nowadays, the number of people who are interested in their physical health and hence opt for a healthier way of living is increasing, and one of the main ways to achieve this is through their nutrition. Phytochemicals, which can be found in vegetables, nuts, fruits, juices, wines and related matrices, have beneficial effects on human health. Various compound groups belong to this category, including flavonols, flavanols, and anthocyanins.  One of the most interesting and distinguishing characteristics of the anthocyanins group is that in different pH conditions, their structure changes, resulting in a change of their color as well. This characteristic property is used to measure the amount of the total monomeric anthocyanin content (TAC) in samples, in an easy, quick, and cost-effective way. This is known as the "pH Differential Method" and was introduced by Sondheimer and Kertesz in 1948. Nowadays, the two pH values used for the above purpose are 1.0 and 4.5, respectively, and in order to avoid any potential interference with the calculation of the TAC, the absorbance is measured both at 520 nm and at 700 nm.
However, in order to evaluate the concentration of individual anthocyanins, other selective analytical techniques must be employed. In the past, capillary electrophoresis and various classic chromatographic methods such as paper and thin-layer chromatography were used [5][6][7][8][9][10][11]. Nowadays, the separation technique which is commonly applied is liquid chromatography using mainly C18 reversed-phase columns. This technique combined with UV-Vis (PDA (Photodiode array)) and/or mass spectrometry detectors has been used for the determination of anthocyanin content in various matrices [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28].  One of the most interesting and distinguishing characteristics of the anthocyanins group is that in different pH conditions, their structure changes, resulting in a change of their color as well. This characteristic property is used to measure the amount of the total monomeric anthocyanin content (TAC) in samples, in an easy, quick, and cost-effective way. This is known as the "pH Differential Method" and was introduced by Sondheimer and Kertesz in 1948. Nowadays, the two pH values used for the above purpose are 1.0 and 4.5, respectively, and in order to avoid any potential interference with the calculation of the TAC, the absorbance is measured both at 520 nm and at 700 nm.
One of the most common sources of anthocyanins in the human diet is red wine. Therefore, the determination of these compounds is of great importance, since, in conjunction with other phenolic compounds [29], inorganic content (such as calcium) [30], added sulfites [31], and some off-flavor substances [32,33], they are known to play a crucial role in defining the sensorial characteristics of wines.
Two papers that study the effect of the presence of lees on the phenolic compounds of wine and apply the HPLC-UV-Vis technique are those published by Fernandez et al. [12] and by Mazauric and Salmon [13]. The same technique was also used in other studies [14][15][16] for the determination of anthocyanins in wines, grape extracts, and red and black currant extracts. HPLC-DAD-MS has been widely used for the analysis of anthocyanins in various samples, such as purple corn cob [17][18][19][20][21][22], while the LC-nano-DESI-MS has been used in red wine analysis [23]. Last but not least, the use of LC-MS/MS has been applied in order to analyze wine grape skins [24], berry skins [22,25], human saliva [26], elderberry extracts [27] and red wine vinegar [28].
In this paper, a sensitive method for the quantitative extraction and determination of anthocyanins from winery byproducts is presented. The specific method investigates and optimizes not only the extraction procedure, but also the liquid chromatography-mass spectrometry (LC-PDA-MS) determination parameters. The purpose of this research was to investigate to what extent the anthocyanins are distributed between the wine (Syrah) and the byproducts that are produced during its vinification procedure, and if the estimated ratio of total anthocyanins between byproducts and wine is the same for every anthocyanin group or not. It is the first time that a comparative study about the concentration of anthocyanins' between a specific wine variety (Syrah) and its byproducts is made and the obtained results could point to a direction towards the possible benefits from the utilization of these "wastes".

Reagents and Solutions
All reagents used were of analytical-reagent grade and deionized water was used for preparation of all aqueous solutions. Methanol (MeOH, Chem-Lab, Zedelgem, Belgium), ethanol (EtOH, Chem-Lab), acetic acid (CH 3 COOH, Panreac, Barcelona, Spain), formic acid (HCOOH, Panreac), and ortho-phosphoric acid (H 3 PO 4 85%, Panreac) were used for the preparation of the extraction solutions tested during the optimization of the extraction procedure. Two buffer solutions were prepared for the application of the pH Differential Method, using potassium chloride (KCl, J.T. Baker, Deventer, The Netherlands), hydrochloric acid (HCl 37%, Carlo Erba Reagents, Milan, Italy), acetic acid (CH 3 COOH, Panreac), and sodium acetate (CH 3

Extraction Procedure
The solid wastes were stored in a freezer (´20˝C). Amount of these wastes were freeze-dried and the resulted powder was collected and stored in the fridge (5˝C). The analogy between the amount before the freeze-drying and the resulting powder was 3 to 1.
After that, a certain quantity of the powder was mixed with the extraction solution. The mixture was vortexed, transferred to an ultrasound bath, and after that to a water bath at 45˝C for 2 h. The next step was the centrifugation of the mixture at 4000 rpm for 10 min and the collection of the supernatant.
The final step, in order to evaluate the effectiveness of the extraction procedure was the application of the pH Differential Method in the collected supernatant fluid. In order to do so, two buffers had previously been prepared, one for pH = 1, HCl-KCl, and one for pH = 4.5, CH 3 COOH-CH 3 COONa.

Optimization of the Extraction Procedure
Four parameters were investigated in order to optimize the extraction procedure of anthocyanins from these byproducts. Those were the selection of the extraction solution, the determination of the optimum analogy between dried matter and the extraction solution, the water bath extraction duration, and the determination of consecutive extraction rounds needed in order to retrieve the maximum amount of anthocyanins from the examined wastes. During the optimization studies of the extraction procedure, the pH differential method was applied for the evaluation of results.

Extraction Solution Selection
Anthocyanins are extracted using organic solvents, and the most commonly used are methanol, ethanol, and acetone. In many cases acidified solutions are used because in this way it becomes easier for the anthocyanins to pass through the cellular membranes and be obtained by the extraction solution. Weak organic acids, or low concentrations of strong acids such as HCl, should be used.
In this work, two organic solvents were investigated, methanol and ethanol, in various combinations with acids and water. In total, 14 extraction solutions were tested (Figure 2), using formic, acetic, and phosphoric acid, each one at two different concentration levels, and water. The results showed that methanol performs better than ethanol as a solvent for the extraction of anthocyanins from the specific substrate, as has also been reported in the paper of Metivier et al. [34]. More specifically, the methanol/phosphoric acid solution (95/5, v/v) was chosen to be used as the extraction solution in this work since it exhibits the best results compared to all other solutions tested. formic, acetic, and phosphoric acid, each one at two different concentration levels, and water. The results showed that methanol performs better than ethanol as a solvent for the extraction of anthocyanins from the specific substrate, as has also been reported in the paper of Metivier et al. [34]. More specifically, the methanol/phosphoric acid solution (95/5, v/v) was chosen to be used as the extraction solution in this work since it exhibits the best results compared to all other solutions tested.

Dried Matter-to-Solvent Volume Ratio Optimization
During this stage, four dried matter-to-solvent volume ratios were tested: 1/10, 1/20, 1/30, 1/40 (g/mL) ( Table 3). Statistical analysis of the results was employed for the selection of the optimum dried matter-to-solvent volume ratio, specifically the "Student's t-test" at a 95% confidence level. It was found that the difference between the 1/20 and 1/10 ratios is statistically significant, while among the 1/20, 1/30, and 1/40 ratios there is no statistically significant difference. Consequently, the 1/20 ratio was selected, since the maximum amount of anthocyanins is retrieved from the examined wine byproducts with the least use of the extraction solvent. Table 3. Dry sample mass-to-solvent volume ratio in correlation with total anthocyanins concentration per 100 g of sample.

Dried Matter-to-Solvent Volume Ratio Optimization
During this stage, four dried matter-to-solvent volume ratios were tested: 1/10, 1/20, 1/30, 1/40 (g/mL) ( Table 3). Statistical analysis of the results was employed for the selection of the optimum dried matter-to-solvent volume ratio, specifically the "Student's t-test" at a 95% confidence level. It was found that the difference between the 1/20 and 1/10 ratios is statistically significant, while among the 1/20, 1/30, and 1/40 ratios there is no statistically significant difference. Consequently, the 1/20 ratio was selected, since the maximum amount of anthocyanins is retrieved from the examined wine byproducts with the least use of the extraction solvent. Table 3. Dry sample mass-to-solvent volume ratio in correlation with total anthocyanins concentration per 100 g of sample.

Optimization of Water Bath Extraction Duration
Four different dried matter-extraction solution mixtures were prepared and they remained in the water bath at 45˝C for one, two, three and four hours, respectively (Table 4). Statistical analysis of the results was employed for the selection of the optimum water bath extraction duration. The "Student's t-test" at a 95% confidence level was used again. It was found that the difference between one and two hours is statistically significant, while among two, three, and four hours there is no statistically significant difference. Therefore, two hours was selected as the optimum water bath extraction duration, since the maximum amount of anthocyanins is retrieved in the minimum amount of time.

Consecutive Extraction Rounds Optimization
As illustrated in Table 5, 81% of anthocyanins that can be retrieved from these wastes have been obtained after one extraction round only. However, in order to retrieve the maximum possible amount of anthocyanins from the byproducts, two additional extraction rounds are necessary.

Mobile Phase Selection
Using the same elution program, two organic elution solvents were tested as Solvent B ( Table 2). The first was acetonitrile and the second was methanol (Figure 3). The two compounds are delphinidin-3-O-glucoside and cyanidin-3-O-glucoside (peaks 1 and 2). After the completion of the experiments, acetonitrile was selected as the optimum solvent because the chromatograms obtained with the use of that specific solvent outclassed the ones with methanol in two parameters.
First of all, the peaks' widths were much narrower compared with the methanol-eluted peaks, improving the resolution between these peaks eluting in close elution times. Secondly, using the same elution program, the peaks were eluted much earlier, thus reducing the duration of the analysis. The elution time could also be reduced through the use of methanol as the organic solvent but this would require an increase in the ratio between the organic solvent and the water used, producing an increased volume of organic wastes.  Table 2). Peaks 1 and 2 are delphinidin-3-O-glucoside and cyanidin-3-Oglucoside, respectively. * High-performance liquid chromatography-Photodiode array detector.

Figures of Merit of the Proposed Method
The performance characteristics of the proposed method were derived from calibration using a series of standard samples at different concentration levels of oenin chloride (malvidin-3-O-glucoside chloride) solutions. The results, using linear regression analysis on peak area signals, are listed in Table 6, together with the repeatability, the limit of detection (LOD) and the limit of quantification (LOQ) of the method. The repeatability was calculated from n = 5 measurements at a 10.0 mg•L −1 concentration level. The detection and quantification limits were determined based on the standard deviation of the standard solution at the lowest concentration level. In particular, the limit of detection (LOD) was calculated as three times and the limit of quantitation (LOQ) as 10 times the above standard deviation divided by the slope of the calibration curve. The LOD and LOQ thus were computed as 9.0 ng•mL −1 and 27 ng•mL −1 , respectively. The dynamic range extends between 0.03 and 10.0 mg•L −1 .  Table 2). Peaks 1 and 2 are delphinidin-3-O-glucoside and cyanidin-3-O-glucoside, respectively. * High-performance liquid chromatography-Photodiode array detector.

Figures of Merit of the Proposed Method
The performance characteristics of the proposed method were derived from calibration using a series of standard samples at different concentration levels of oenin chloride (malvidin-3-O-glucoside chloride) solutions. The results, using linear regression analysis on peak area signals, are listed in Table 6, together with the repeatability, the limit of detection (LOD) and the limit of quantification (LOQ) of the method. The repeatability was calculated from n = 5 measurements at a 10.0 mg¨L´1 concentration level. The detection and quantification limits were determined based on the standard deviation of the standard solution at the lowest concentration level. In particular, the limit of detection (LOD) was calculated as three times and the limit of quantitation (LOQ) as 10 times the above standard deviation divided by the slope of the calibration curve. The LOD and LOQ thus were computed as 9.0 ng¨mL´1 and 27 ng¨mL´1, respectively. The dynamic range extends between 0.03 and 10.0 mg¨L´1.

Identification and Quantification of Anthocyanins in Wine Variety (Syrah) and Its Byproducts
The proposed method was applied to two real samples, to the wine variety Syrah, and to the byproducts that were produced during the vinification procedure of the specific wine, and there were four main remarks after the analysis.
In total, 23 anthocyanins were identified. All of them were present in the wine, while 21 of them were detected in the wine byproducts (Figures 4 and 5). The two anthocyanins that were detected in the wine, even in small amounts, but not in the wine byproducts were the cyanidin-3-O-glucoside and the delphinidin-glucoside-pyruvate derivative.
During the preliminary studies the samples' mass spectra were acquired using both the scan and SIM (Selected-ion monitoring) mode for specific anthocyanins. However, during the last steps of the determination procedure, all of the detected anthocyanins were quantified using the SIM mode.
The analyzed samples, both the byproducts' extract and the wine, apart from anthocyanins, were also very rich in many other phenolic compounds that could co-elute with some of the compounds of interest. By combining the use of the SIM mode at specific m/z (mass-to-charge ratio) values with the use of standard compounds and elution order data, there were no interferences problems that occurred during the quantification procedure due to matrix's complexity.
Furthermore, after the analysis of the samples, it was found that the four categories of the detected anthocyanins, simple glucosides, acetyl glucosides, cinnamoyl glucosides, and pyroanthocyanins, were not present in a standard ratio among the two samples, wine byproducts and wine ( Table 7). The simple and the acetyl glucosides exhibited a ratio of almost 10 to 1, while cinnamoyl glucosides had a ratio of 18.20 to 1, and pyroanthocyanins had a ratio of 5.83 to 1.    Table 8. The total concentration of anthocyanins in wine byproducts was estimated to be almost 12 times higher than the equivalent concentration in Syrah wine (Table 8). This is extremely important since it indicates the potential that these "wastes" could have as a raw material for the recovery of that kind of compound, thus revealing their true value.  Table 8.   Table 8. The total concentration of anthocyanins in wine byproducts was estimated to be almost 12 times higher than the equivalent concentration in Syrah wine (Table 8). This is extremely important since it indicates the potential that these "wastes" could have as a raw material for the recovery of that kind of compound, thus revealing their true value.  Table 8. The five-digit numbers at the upper left corner of the figure correspond to the characteristic m/z (M + ) of detected compounds, and the number in the parenthesis indicates a multiplying factor of each peak in the chromatogram.
All the above clearly show that the transfer mechanism of anthocyanins between a wine (Syrah) and its byproducts is not a simple procedure that functions and performs in the same way for every anthocyanin, indicating that it is probably affected by the chemical structure of every compound.
The total concentration of anthocyanins in wine byproducts was estimated to be almost 12 times higher than the equivalent concentration in Syrah wine (Table 8). This is extremely important since it these results confirm the potential value of these byproducts and indicate a direction that should be followed towards their utilization.