Socio-economical survey of alcoholic beverage producers

From a list of 75 producers of V. tiliifolia alcoholic beverages in Naolinco, Veracruz, Mexico, 25 were arbitrarily selected for an interview, representing 30 % of the producers. The socioeconomic and cultural aspects investigated were: 1) gender (1 = male, 2 = female), 2) age, 3) academic education (1 = primary, 2 = junior high school, 3 = high school, 4 = college), 4) knowledge about the health benefits of V. tiliifolia alcoholic beverage consumption (1 = cardiovascular, 2 = digestive, 3 = circulatory, 4 = others), 5) years of experience making alcoholic beverages, 6) sale price, 7) amount of grapes used per liter of beverage (kg∙L^-1), 8) quantity of alcoholic beverage produced per year (L), 9) price of grapes per kg, 10) place where grape is bought (1 = street vendor, 2 = local market, 3 = no purchase), 11) use of fruit peel (1 = waste, 2 = fertilizer), 12) use of grape seeds (1 = waste, 2 = fertilizer, 3 = others), and 13) problems in marketing alcoholic beverage (1 = a lot of competition, 2 = no competition).

Plant material

The fruits of typical V. tiliifolia used to make the alcoholic beverages were collected or bought in the county of Naolinco, Veracruz, Mexico, in 2016 and 2017. All the wild plants were located between 1,469 masl (19° 38’ 47.9’’ N and 96° 51’ 13.6’’ W) and 1,528 masl (19° 38’ 59.0’’ N and 96° 51’ 37.0’’ W) (^{Cruz-Castillo et al., 2009}). At each site, per year (2016 and 2017), samples were taken for analysis at a depth of 40 cm from the soils where the plants grew.

In September 2016, when the plants had mature fruits, the soil had a pH of 4.6-4.7, density of 1.02-1.29 g·cm^-3, total N of 0.23-0.25 %, P of 4.25-5.61 ppm, K of 200 ppm, Ca of 1.12-1.7 meq·100 g^-1, Mg of 0.023-0.025 meq·100 g^-1, organic matter of 4.73-5.27 %, and electrical conductivity of 0.1353-0.1941 μS. In February 2017, when the plants had leaves but no reproductive growth, the soil had a pH of 5.2-5.6, density of 1.33-1.34 g·cm^-3, total N of 0.20-0.27 %, P of 3.12-3.86 ppm, K of 225 ppm, Ca of 1.58-1.79 meq·100 g^-1, Mg of 0.60 meq·100 g^-1, organic matter of 4.19-5.44 %, and electrical conductivity of 0.342-0.1446 μS. The sampling and analyses of the soil were conducted according to standard NOM-021-RECNAT-2000 (^{Secretaria de Medio Ambiente y Recursos Naturales [SEMARNAT], 2002}).

Chemical attributes of the alcoholic beverages

In 39 different bottles of alcoholic beverages, the pH, total soluble solids (TSS), total sugars, alcohol percentage, and acidity were determined in August 2016. The pH was determined with a potentiometer (pHM82, Radiometer^®, Denmark). The TSS were determined using refractometry by placing a drop of the alcoholic beverage on a digital refractometer (HI96801, Hanna instruments^®), which was calibrated with distilled water. Titratable acidity was determined by the volumetric method and expressed in percentage malic acid (^{Jackson, 2008}) using a titator (HI84532U-01, Hanna instruments^®). Total sugars were evaluated by a modified anthrone method (^{Shakappa, 2015}), and the alcohol percentage by direct distillation (^{Servicio Agrícola y Ganadero [SAG], 2011}).

Total phenols, total flavonoids, anthocyanins, tannins, and antioxidant capacity were evaluated in August 2017 in another five bottles of alcoholic beverages made with V. tiliifolia. One named commercially by the producer as Tinto Salvaje and the others were coded from 1 to 4. All chemical determinations of alcoholic beverages were made in triplicate. The beverages made with V. tiliifolia in Naolinco were locally classified as semi-sweet wines, so they were contrasted with the semi-sweet Lambrusco^® wine (made with V. vinifera cv Lambrusco Salamino in Italy), because its sweet taste resembles that of the drink made with V. tiliifolia. All chemical determinations of the alcoholic beverages were repeated three times. A completely randomized design was used for these evaluations.

The antioxidant capacity of the five different alcoholic beverages made with V. tiliifolia and the Lambrusco wine was determined by DPPH (2,2-diphenyl-1-picrylhydrazyl) (^{Cheng, Moore, & Yu, 2006}), ABTS (2,2′-azino-bis[3-ethylbenzothiazoline-6-sulfonic acid]) (^{Re et al., 1999}), and FRAP (ferric reducing antioxidant power) (^{Benzie & Strain, 1996}) assays, all adapted to microplates. Absorbances were measured in a microplate reader (Synergy 2, Biotek^®, EUA).

To determine antioxidant capacity by the DPPH method, 200 μL of wine sample dilution (1.3 to 13 μmL·mL^-1) and 50 μL of DPPH solution (0.099 mM) were mixed; the mixture was then covered with aluminum foil and incubated for 30 min in the dark. Subsequently, the decrease in absorbance was monitored at 517 nm versus a blank, which consisted of 80 % methanol (250 μL). The control contained 200 μL of 80 % methanol and 50 μL of the DPPH solution (50 μL). Prior to analysis, a standard curve was prepared from a 1 mM Trolox stock solution (6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid) with a linear range of 7.37 to 34.51 µM. Results were expressed as micromoles of Trolox equivalents (TE) per liter of alcoholic beverage (µmol TE·L^-1).

The ABTS method consists of mixing 20 μL of the diluted alcoholic beverage sample and 180 µL of the ABTS·+ solution, which was obtained by reacting 7.4 mM ABTS and 2.6 mM of sodium persulfate. The mixture’s absorbance was measured at 734 nm. Prior to analysis, a curve was prepared from a 1 mM Trolox stock solution with a linear range of 4.99 to 5,993 µM. Results were expressed as micromoles of Trolox equivalents per liter of alcoholic beverage (µmol TE·L^-1).

To determine antioxidant capacity by the FRAP method, 20 μL of the diluted alcoholic beverage, 180 µL of the FRAP solution and 60 µl of distilled water were mixed. Absorbance was measured at 595 nm. The FRAP reagent was prepared from a sodium acetate buffer (300 mM, pH 3.6), TPTZ solution (10 mM) (HCl 40 mM as solvent) and iron (III) chloride solution (20 mM) at a ratio of 10:1:1, respectively. Prior to measurement, a standard curve was prepared from a Trolox stock solution (1 mM) with a linear range of 3.8 to 46 µM. Results were expressed as micromoles of Trolox equivalents per liter of alcoholic beverage (µmol TE·L^-1).

Total phenolic compounds content was determined using the Folin-Ciocalteu reagent (^{Singleton & Rossi, 1965}) adapted to microplates, for which 25 µL of diluted alcoholic beverage, or gallic acid standard, were added to 125 µL of distilled water; subsequently, 20 µL of Folin-Ciocalteu reagent, diluted to 1:10, was added and mixed with 30 µL of 20 % Na₂CO₃ solution. After incubation for 30 min at room temperature, absorbance was measured at 750 nm versus a blank, which consisted of 25 µL of distilled water instead of sample. A gallic acid standard curve with a linear range of 2.5 to 29.0 μg GA·mL^-1 was prepared from 0.5 mg mL^-1 of a gallic acid stock solution. Results were expressed as milligrams of gallic acid equivalents (GAE) per liter of alcoholic beverage (mg GAE·L^-1).

Total flavonoid content was determined according to ^{Kubola and Siriamornpun (2011)} and adapted to microplates. Briefly, 500 µL of sample or standard (catechin) were mixed with 2.5 mL of distilled deionized water. At time zero, 0.15 mL of 5 % NaNO₂ solution were added and mixed. After 6 min, 10 % AlCl₃·6H₂O solution (0.3 mL) was added and mixed. After 5 min, 5 % NaOH solution (1 mL) was added and mixed. Absorbance of an aliquot (200 µL) of the colored flavonoid-aluminum complex was measured at 510 nm. The blank consisted of 500 µL of 80 % methanol. Prior to analysis, a standard curve was prepared from a catechin stock solution (300 µg·mL^-1) with a linear range of 0.7 to 34.5 µg C·mL^-1. Results were expressed as micrograms of catechin equivalents (CE) per liter of alcoholic beverage (µg CE·L^-1).

Monomeric anthocyanin content of the samples was determined by the spectrophotometric pH differential protocol (^{Giusti & Wrolstad, 2001}), for which 0.5 mL of alcoholic beverage were mixed with 0.5 mL of a pH 1 buffer (KCl 0.025 M). The mixture’s absorbance was measured at 510 and 700 nm. Subsequently, 0.5 mL of alcoholic beverage were mixed with 0.5 mL of a pH 4.5 buffer (CH₃COONa 4.5 M), and the absorbance of this solution was measured at the same wavelengths. The anthocyanin content was calculated based on cyanidin-3-O-glucoside, molar extinction coefficient of 26,900 and a molecular weight of 449.2 g·mol^-1. Results were expressed as milligrams of cyanidin-3-O-glucoside equivalents (CydE) per liter of alcoholic beverage (mg CydE ·L^-1).

Condensed tannins were determined according to ^{Liu, Lin, Wang, Chen, and Yang (2009)}, for which 25 µL of alcoholic beverage were mixed with 750 µL of a 4 % vanillin solution and 375 µL of concentrated HCl. The mixture was vortexed for 20 min. The mixture’s absorbance was measured at 500 nm. Prior to determination, a catechin standard curve was prepared in a linear range of 0.004 to 0.032 mg C·mL^-1 from a catechin stock solution (2 mg·mL^-1) using 80 % methanol as solvent. Results were expressed in milligrams of catechin equivalents per liter of alcoholic beverage (mg CE·L^-1).

Statistical analysis

To obtain information on companies and technological advances, some exploratory studies have analyzed surveys of a few individuals through principal components (^{Corrales-García, Cruz-Castillo, Lozano-López, & Famiani, 2009}), so in this study the responses to the survey were analyzed using this technique. The relevance of PCA in this work was determined by a Kaiser-Meyer-Olkin sample analysis and a Bartlett sphericity test applied to the resulting correlation matrix (^{Álvarez, 1994}).

The 39 alcoholic beverages studied were grouped by cluster analysis with Ward's minimum variance method with Gower distance. For the selection of the number of groups, the cophenetic correlation coefficient was used, with the 0.60 distance value as threshold. The variables used in these analyses were TSS, total sugars, acidity, pH and alcohol percentage. Standard errors were estimated for these variables when groups had two or more replications (Di Rienzo et al., 2008). The concentration data for phenols, flavonoids, anthocyanins, tannins and antioxidant capacity were subjected to an analysis of variance and, subsequently, Tukey’s multiple comparison test (P ≤ 0.05). All statistical analyses were performed with the InfoStat-statistic version 8 software (^{Di Rienzo et al., 2016}).

Socio-economical survey of alcoholic beverage producers

The multivariate analysis yielded four principal components (PCs) greater than 1.0 and one with 0.9, which together accounted for 79 % of the total variability of the data (Table 1), with PC1 explaining the greatest variability (26 %) (Table 1). The main variables which contributed to explaining this variability were the price of alcoholic beverage sales and the final destination of the wastes, represented by fruit peels and seeds (Table 1). Thus, the producers who were able to sell the alcoholic beverages at the highest prices were also able to use the wastes resulting from the processing (peels and seeds). For example, two producers sold their beverages at a high price and used the wastes as food for poultry (Table 2). In contrast, another producer sold his alcoholic beverage at a low price and did not use the fruit peels and seeds (Table 2). In the process of making the beverage, many wastes are produced and the environmental impact of the artisanal production of V. tiliifolia ‘wines’ in Naolinco has not been determined. In this regard, their management is a necessary activity that has to be regulated (^{Devesa-Rey et al., 2011}).

PC2 accounted for 19 % of the data variability. Gender and grape purchasing site were the main variables (Table 1). Thus, the beverage producer with number 14 was male and harvested the fruit himself. In contrast, a female beverage producer with number 12 bought the grapes from street vendors to make the alcoholic beverage (Table 2).

PC3 explained 14 % of the total variability (Table 1), with the most important contributing variables being the years of experience in beverage making and gender. The producers with more years of experience making the alcoholic beverages were generally male. PC4 explained 11 % of the total variability, and the variable which most contributed to explaining this variability was the knowledge about the health benefits of alcoholic beverage consumption (Table 1). This indicated that some producers are unaware of the health benefits that their beverages could provide to customers. Finally, PC5 explained 9 % of the total variability and the most important variable was the amount of alcoholic beverage produced per year (Table 1). In this regard, producer number 22 had the highest production (1,000 L) in the period evaluated (Table 2), and he or she received no financial support from either the government or any bank for producing the alcoholic beverage.

The alcoholic beverage producers had no difficulties in obtaining the grapes from the wild vines that are overexploited locally. Sometimes unripe fruit was harvested, and the vines did not receive any horticultural management. The beverage makers need training to improve their operations. Wastes generated by winemaking should be considered as a supplementary income source and also strategies to avoid any damage to the local environment should be developed.

Total soluble solids, reducing sugars, acidity, alcohol, and pH

All 39 alcoholic beverages made with V. tiliifolia were classified into seven different groups by cluster analysis. Three alcoholic beverages were not grouped, and one of them showed the lowest TSS (0.26 °Brix) and the lowest alcohol percentage (2 %) (Table 3). In Italy, it was suggested that dealcoholized wines should be marketed not as a wine but as a different type of beverage (^{Stasin,Bimbo, Viscecchia, & Seccia, 2014}). In contrast, another individual alcoholic beverage had the highest TSS (38.16 °Brix) and the highest alcohol percentage (21.94 %) (Table 3), which, according to ^{Jackson (2008)}, can be considered as liquor for the market.

Group number 4 included 51 % of all alcoholic beverages sampled, and these beverages had 14.72 % alcohol and considering their high concentration of sugars (15.97 g·L^-1) and high acidity (0.8 %) (Table 3), they are not classified as dry red wines (^{Matsuhiro et al., 2009}). Sugar commonly needs to be added to the juice from species that do not attain adequate sugar concentration at harvest (~20 %) to develop about 12 % alcohol-content wine (^{Jackson, 2008}). This could be the case for some V. tiliifolia alcoholic beverages in this group.

The second largest group (number 7) had 26 % of all the alcoholic beverages. This group had beverages with 9.29 % alcohol and with intermediate TSS (17.11 °Brix) and sugar (9.61 g·L^-1) values with respect to the other alcoholic beverages (Table 3).

Grape quality and fruit yield per vine have considerable influence on wine quality (^{Poni et al., 2018}; ^{Pozo-Bayon, Polo, Martín-Alvarez, & Pueyo, 2004}). The differences in fruit quality for the making of alcoholic beverages from V. tiliifolia vines has not been determined, and producers of alcoholic beverages use grapes from any wild vine growing in the forest with fruit production that is very different in each one.

The alcohol percentage recorded across all the alcoholic beverages was between 2 and 21.94 %, and this shows the lack of regulations for making alcoholic beverages from V. tiliifolia in Naolinco. Also, the wide-ranging values for reducing sugars, pH, acidity and TSS in the alcoholic beverages indicate different stages of maturity in the harvested grapes. To determine the date of harvest in V. vinifera, it is necessary to consider the TSS of the fruit. For example, as a general rule for making a good-quality table wine, the wine grapes should have a minimum of ~20 °Brix at harvest (^{Poni et al., 2018}). The TSS values reported in mature fruits of V. tiliifolia are about 13 °Brix for the pulp and 8 °Brix for the peel (^{Jiménez, Juárez, Jiménez-Fernández, Monribot-Villanueva, & Guerrero-Analco, 2018}), and the TSS in the beverages studied had values between 0.26 and 38.16 %. The makers of alcoholic beverages of Naolinco do not take into account the TSS as an index for harvest and sometimes the harvested fruit is not ripe or overripe, thereby lowering the quality of the alcoholic beverage.

The V. tiliifolia alcoholic beverages were more acidic (0.5-4.7 %) (Table 3) than the Sangiovese (6.5-6.6 %) (^{Palliotti et al., 2017}), Cabernet Sauvignon (7.83 %) and Merlot (6.90 %) wines (^{de la Cruz-Aquino, Martínez-Peniche, Becerril-Román, & Chávaro-Ortíz, 2012}). This is probably because V. tiliifolia vines in the wild have acidic juice with titratable acidity values of 3 % (Jimémez et al., 2018), while the juice of V. vinifera cultivar grapes used for making commercial red wine is less acidic (^{Sadras, Petrie, & Moran, 2012}; ^{Soyer, Koca, & Karadeniz, 2003}). Therefore, the high acidity and low pH (1.17-3.31) found in alcoholic beverages made with V. tiliifolia (Table 3) would be related to the acidic characteristic of the V. tiliifolia grapes that grow naturally in the tropical forest.

The seven different groups of V. tiliifolia alcoholic beverages had different chemical characteristics, making it difficult to compare them with dry red wines. However, the taste of alcoholic beverages made with V. tiliifolia is similar to that of semi-sweet red wines that are used in Italy to accompany dessert.

Phenols, anthocyanins, flavonoids, tannins and antioxidant capacity

The semi-sweet Lambrusco Italian wine was superior to all red alcoholic beverages of V. tiliifolia in total phenols, anthocyanins, tannin contents, and antioxidant capacity (ABTS, DPPH, and FRAP) (Table 4). Regarding the flavonoid concentration, the Tinto Salvaje ‘wine’, made with V. tiliifolia grapes, surpassed the Lambrusco wine (Table 4). Tinto Salvaje had greater antioxidant capacity than the other alcoholic beverages made with V. tiliifolia (Table 4).

Total phenol and anthocyanin levels determined in the Lambrusco wine (Table 4) were similar to those reported by other researchers for this wine (^{Tassoni, Tango, & Ferri, 2014}). The antioxidant capacity (DPPH, FRAP and ABTS) of the V. tiliifolia alcoholic beverages resembled that determined in V. vinifera rosé wines (^{Lino et al., 2014}). Moderate consumption of V. tiliifolia alcoholic beverages with adequate concentrations of flavonoids could help to regulate immune responses that decrease the inflammation observed in many diseases such as Alzheimer's and dementia diseases (^{Magrone & Jirillo, 2010}).

The alcoholic beverages made from V. tiliifolia are valued by certain consumers in Mexico and Cuba. This species is not commercially cultivated in Mexico, but with a selection of varieties, along with horticultural management, the quality of its fruits (increasing TSS at maturity) can be enhanced and better alcoholic beverages produced. The artisanal processes used in the making of alcoholic beverages of V. tiliifolia were not documented in the present study, but with greater technological knowledge in the production of this beverage, the quality will be improved. The alcohol percentage of the beverages varied from 5.25 to 14.73 %. The Tinto Salvaje had the lowest value (Table 4) and this can be classified as a dealcoholized ‘wine’ (Stasi et al., 2014) or a beverage made from V. tiliifolia. The rest of the alcoholic beverages are within the alcohol-content range for wines (Table 4).