Volatile profile and antioxidant-related measurements of two Sauvignon Blanc vinifications from the Brazilian Cerrado: an exploratory chemical characterization

1 Introduction

Brazilian viticulture began around 1532, during the colonial period, with the introduction of the first grapevines from Madeira Island by the Portuguese. Its consolidation occurred in the 19th century, driven largely by European immigration, particularly Italian immigration, which helped establish the productive base of the Serra Gaúcha region, still the country’s main wine-producing area. In recent decades, Brazilian viticulture has expanded and diversified, with new producing regions emerging across distinct biomes (Flores et al., 2024).

In this context, viticulture has advanced in tropical and subtropical high-altitude areas, particularly within the macroregion of Winter Wines in the Southeast, Central-West, and parts of Northeast Brazil. This production model uses double pruning to separate the vegetative and productive cycles, allowing harvest during the autumn-winter period (Vieira et al., 2021). The Brazilian Cerrado has emerged as a new viticultural frontier because of its high solar radiation, altitude, and marked seasonality (Pereira et al., 2020; Flores et al., 2024). These environmental and management factors may contribute to the distinctive chemical identity of wines produced in the region.

Among the most relevant varieties worldwide, the Sauvignon Blanc cultivar, originating from the Bordeaux and Loire regions of France, stands out as one of the main fine white grape varieties, widely cultivated in so-called “New World” countries such as New Zealand, South Africa, Chile, and the United States (Parr et al., 2010). In Brazil, this variety is distributed across different wine-producing regions, including Campanha Gaúcha, Planalto Catarinense, Chapada Diamantina, and the Sub-middle São Francisco Valley (Guerra and Pereira, 2018). Wines produced from this cultivar exhibit a characteristic aromatic profile, frequently described as herbaceous and fruity, with high acidity and a refreshing character, attributes directly related to their chemical composition and fermentative processes (Filho et al., 2025; Coetzee and Toit, 2012).

From a chemical-sensory perspective, wine aromas are commonly classified as primary, originating from the grape; secondary, formed mainly during fermentation; and tertiary, developed during aging (Carpena et al., 2020). Alcoholic fermentation plays a central role in secondary aroma formation and is predominantly conducted by S. cerevisiae (Swiegers et al., 2005). Non-Saccharomyces yeasts, naturally present on grape skins, have also been investigated for their ability to modulate wine volatile composition (Borren and Tian, 2020). Among them, Hanseniaspora vineae has been associated with the production of acetate esters, particularly 2-phenylethyl acetate. Its behavior in mixed fermentations with Saccharomyces cerevisiae depends on strain, inoculation protocol, must composition, and fermentation conditions (Carrau et al., 2023; Gallo et al., 2024).

Although tropical viticulture has advanced in recent years, Sauvignon Blanc wines produced in the Brazilian Cerrado, particularly under double-pruning management, remain insufficiently characterized. Studies on Brazilian Sauvignon Blanc have focused mainly on other producing regions, including the high-altitude areas of Santa Catarina, where physicochemical parameters, volatile profiles, and cultivar-associated aroma compounds have been evaluated (Filho et al., 2021; Filho et al., 2025). Information remains limited on the chemical characteristics of Sauvignon Blanc wines from the Cerrado and on wines produced there under different inoculation conditions.

Thus, this study aimed to provide an exploratory chemical characterization of two Sauvignon Blanc vinifications produced under tropical conditions in the Brazilian Cerrado: one inoculated with S. cerevisiae and the other co-inoculated with H. vineae and S. cerevisiae. The integrated analytical approach combined physicochemical measurements, spectrophotometric assays, voltammetry, and GC-MS, with emphasis on volatile composition and antioxidant-related measurements. Given the experimental design, the results were interpreted descriptively and in the context of the different post-fermentation evolution of the two wines.

2 Materials and methods

2.1 Raw material and fermentation design

The study comprised two winemaking conditions applied to Sauvignon Blanc grapes obtained from a single harvest lot cultivated under double-pruning management at Quartetto Vinhos and Vinhedos winery in the Goiás Cerrado. T1 was inoculated with S. cerevisiae only, whereas T2 was co-inoculated with H. vineae and S. cerevisiae. Only one fermentation was performed for each condition. Therefore, no independent fermentation replicates were included. Unless otherwise stated, the reported triplicates correspond exclusively to repeated analytical measurements performed on aliquots of the same wine sample and estimate analytical repeatability rather than fermentation-to-fermentation variability. After alcoholic fermentation, both wines were bottled and stored at a controlled temperature of 15 °C.

Ripening monitoring, carried out by Quartetto Vinhos and Vinhedos, included total soluble solids (°Brix), measured using a digital refractometer, and pH, measured using a digital pH meter. Based on weather forecasts and with the producers’ approval, the 2024 harvest for the Sauvignon Blanc cultivar was conducted in July, during which 60 kg of grapes were set aside for the winemaking process.

Harvesting was conducted to ensure adequate representativeness of the vineyard area. Clusters were collected from plants distributed across the cultivation plot. Harvest took place on 3 July 2024, from 18:00 to 00:00, and the grapes were transported under refrigeration to minimize oxidation and preserve fruit quality.

On the same day, the grapes were taken to the facilities of Vinícola Brasília for automatic destemming and simultaneous sulfiting, then the must was transferred to the Laboratory of Simulation and Automation of Bioprocesses (LaSABio), Institute of Chemistry, Federal University of Goiás (UFG), where the other winemaking processes took place.

2.2 Winemaking process

Upon arrival at the Brasília Winery, Sauvignon Blanc grapes were destemmed, and sulfiting was simultaneously performed by adding sulfur dioxide (SO₂) in the form of potassium metabisulfite (K₂S₂O₅) at a final concentration of 50 mg L^-1. The grapes were then pressed to separate skins and seeds from the must, which presented 22 °Brix and a pH of 3.34.

Pre-fermentation clarification (débourbage) was subsequently carried out using a commercial clarifying agent (Fitoclar, Amazon Group) at 300 mg L⁻¹ to promote the sedimentation of suspended solids. The clarified must was then transferred to the Laboratory of Simulation and Automation of Bioprocesses (LaSABio), where an initial racking step was performed to remove the settled lees.

Yeast preparation followed the manufacturers’ recommendations. Rehydration was conducted in preheated water at 30 °C for the non-Saccharomyces strain and 40 °C for S. cerevisiae. After a 15 min rehydration period, 150 mL of must was added to promote acclimation and minimize thermal and osmotic stress. The activated cultures were then inoculated into the respective musts to initiate alcoholic fermentation. Treatment T1 consisted of inoculation with S. cerevisiae (AWRI Rosa) at 20 g hL^-1, supplemented with thiamine (5 mg L^-1). In treatment T2, co-inoculation was performed using H. vineae (HV205) and S. cerevisiae at a 9:1 (m/v) ratio, with a total inoculum of 20 g hL^-1, also supplemented with thiamine (5 mg L^-1), according to the manufacturer’s recommended protocol (Oenobrands, 2022).

Alcoholic fermentation for both treatments was initiated on the same day and conducted at approximately 15 °C under dark conditions. Fermentation progress was monitored daily through density and temperature measurements and was considered complete when density reached 0.990 g mL^-1. Following alcoholic fermentation, the wines were transferred to 750 mL glass bottles, sealed, and stored at 15 °C. Both wines were analyzed after the completion of alcoholic fermentation, rather than after reaching an equivalent malolactic fermentation stage. Yeast population dynamics were not monitored by plate counting, qPCR, sequencing, or another species-specific identification method. Therefore, the persistence, relative abundance, and metabolic activity of H. vineae during fermentation were not determined.

2.3 Physicochemical analyses of Sauvignon Blanc wines

The physicochemical properties of the wines were characterized through the determination of alcohol content (%, v/v), residual sugars (g L^-1), tartaric acid (g L^-1), lactic acid (g L^-1), malic acid (g L^-1), volatile acidity (g L^-1), glycerol (g L^-1), and pH. Analyses were carried out in analytical triplicate using Fourier-transform infrared spectroscopy (FT-IR) with a WineScan SO2 analyzer (FOSS Analytics, Hillerød, Denmark), following the manufacturer’s validated calibration protocols for wine analysis and previously reported FT-IR applications for the physicochemical characterization of grape must and wine-related matrices (Cornehl et al., 2023). Sulfur dioxide was added during vinification as potassium metabisulfite; however, free and total SO₂ concentrations were not independently quantified. Consequently, possible differences in SO₂ availability and their effects on microbial activity, spontaneous malolactic fermentation, oxidation, Folin-Ciocalteu response, and volatile composition could not be evaluated. All measurements were conducted at the Laboratory of Technology of Products of Animal and Plant Origin (TPOA/TPOV), Federal University of Pampa (Unipampa), located in Dom Pedrito, Brazil.

2.4 Phenolic compounds, antioxidant activity, and electrochemical analyses

2.4.1 Folin-Ciocalteu reducing capacity

Folin-Ciocalteu reducing capacity was assessed using a modified colorimetric procedure based on Singleton et al., (1999). Before analysis, wine samples were diluted 1:10 (v/v) in methanol to keep absorbance readings within the instrument’s linear range. Because the Folin-Ciocalteu reagent responds to several reducing substances and is not specific to phenolic compounds, the results were expressed as gallic acid equivalents and interpreted as the overall Folin-reactive reducing capacity of the wines.

An aliquot of 1.0 mL of the diluted sample was transferred to a 25 mL volumetric flask, followed by the addition of 2.5 mL of Folin-Ciocalteu reagent (10%, v/v) and 2.0 mL of sodium carbonate (Na₂CO₃) solution (7.5%, v/v). The final volume was adjusted to 25 mL with MeOH. The reaction mixture was incubated for 60 min in the absence of light, after which absorbance was recorded at 740 nm using a UV–Vis spectrophotometer (BEL, SPECTRO S05 1105). All measurements were performed in analytical triplicate. A calibration curve was prepared using gallic acid as the reference standard at concentrations of 12.5–150 μg mL^-1. The results were expressed as milligrams of gallic acid equivalents per liter (mg GAE L^-1).

2.4.2 Total flavonoid content

Total flavonoid content was quantified using an aluminum chloride (AlCl₃) colorimetric method, based on previously reported procedures with slight modifications (Hosu et al., 2014; Leite et al., 2020). A calibration curve was constructed using quercetin as the reference standard, with concentrations ranging from 0.6 to 4.0 μg mL^-1.

For the analysis, 1.0 mL of the white wine sample was placed in a 25 mL volumetric flask, followed by the addition of 1.0 mL of 5% (w/v) AlCl₃ solution and 2.0 mL of 1 mol L^-1 sodium acetate (CH₃COONa). The final volume was adjusted with distilled water. The reaction mixture was allowed to stand for 30 min prior to absorbance measurement at 440 nm using a UV–Vis spectrophotometer (BEL, SPECTRO S05 1105). All determinations were carried out in analytical triplicate, and results were expressed as milligrams of quercetin equivalents per liter (mg QE L^-1).

2.4.3 Antioxidant activity by DPPH^• assay

A solution of the radical 2,2′-diphenylpicrylhydrazyl (DPPH^•) was prepared by dissolving 6 mg of DPPH^• in 100 mL of methanol (MeOH). Subsequently, a 1 mL aliquot of this solution was diluted with MeOH until the absorbance reached 0.827 at 518 nm, measured using a UV-Vis spectrophotometer (BEL, SPECTRO S05 1105) (Olszowy and Dawidowicz, 2018). All procedures were carried out in the absence of light to prevent radical degradation.

The antioxidant standard used was 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox). A stock solution (1 mg mL^-1) was prepared and used to construct a calibration curve through serial dilution (1:1, v/v). Briefly, 1 mL of the stock solution was transferred to a test tube and mixed with 1 mL of ethanol. This procedure was repeated sequentially to obtain nine concentrations (500, 250, 125, 62.5, 31.25, 15.62, 7.81, 3.90, and 1.95 μg mL^-1).

Wine samples (T1 and T2) were subjected to the same serial dilution procedure (1:1, v/v), starting from 1 mL of each sample diluted in ethanol. For the assay, 100 µL of each dilution was transferred to a light-protected test tube, followed by 1.9 mL of DPPH• solution. The mixtures were vortexed and kept at room temperature for 60 min, after which absorbance was measured at 518 nm. The percentage of DPPH^• inhibition was calculated according to Equation 1:where A₀ is the absorbance of the control and A₁ is the absorbance of the sample.

% I n h i b i t i o n = A 0 − A 1 A 0 × 100 (1)

The IC₅₀ values were determined by nonlinear regression of semilogarithmic concentration-response curves, plotting radical inhibition (%) against the logarithm of the nominal wine concentration in the reaction mixture, expressed as µL of wine per mL of reaction mixture (µL wine mL^-1). The analysis followed Martinez-Morales et al. (2020) and was performed using GraphPad Prism 8.0 (San Diego, CA, United States of America), with a 95% confidence interval.

2.4.4 Antioxidant-related response by the ABTS^•+ assay

A 7 mM stock solution of 2,2′-azinobis (3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS) was prepared by dissolving 0.1936 g of the reagent in 50 mL of ethanol. A 140 mM potassium persulfate solution was prepared by dissolving 378.4 mg of potassium persulfate in distilled water and completing the volume to 10 mL. The ABTS radical cation (ABTS^•+) was generated by mixing 5 mL of the 7 mM ABTS stock solution with 88 µL of the 140 mM potassium persulfate solution. The mixture was kept in the dark at room temperature for 16 h. A 1 mL aliquot of this solution was then diluted with ethanol until an absorbance of 0.770 at 730 nm was obtained, measured using a UV-Vis spectrophotometer (BEL, SPECTRO S05 1105) (Olszowy and Dawidowicz, 2018). All procedures were performed in the dark. Trolox was used as the antioxidant standard, and the calibration curve was constructed as described for the DPPH^• assay using serial dilutions (1:1, v/v) to obtain concentrations ranging from 500 to 1.95 μg mL^-1.

Wine samples (T1 and T2) were diluted following the same procedure, starting from 1 mL of sample in ethanol. For the assay, 100 µL of each dilution was mixed with 1.9 mL of ABTS^•+ solution in light-protected test tubes. After vortexing, the mixtures were allowed to react for 6 min at room temperature. Absorbance was measured at 730 nm, and the percentage of inhibition was calculated using Equation 1. IC₅₀ values were determined as described for the DPPH^• assay.

2.4.5 Electrochemical response by voltammetry

Electrochemical measurements were carried out using an AUTOLAB® potentiostat/galvanostat (Eco Chemie), operated with NOVA software (version 4.3). Analyses were performed in a conventional three-electrode electrochemical cell, consisting of a carbon paste working electrode (70% graphite and 30% Nujol®, Ø 2 mm), a platinum spiral as counter electrode, and an Ag/AgCl/KCl_sat reference electrode.

Square-wave voltammetry (SWV) experiments were conducted with a pulse amplitude of 50 mV, frequency of 50 Hz, and a potential increment of 2 mV, resulting in an effective scan rate of 100 mV s^-1. Differential pulse voltammetry (DPV) measurements were performed using a pulse amplitude of 50 mV, pulse width of 0.5 s, and a scan rate of 10 mV s^-1. All voltammetric data were processed and plotted using OriginPro 8 software. The electrochemical index (EI) was determined according to Equation 2:

E I = I p a 1 E p a 1 + I p a 2 E p a 2 + … + I p a n E p a n (2)

In this context, I _*pa* corresponds to the anodic peak current of the nth peak, E _*pa* represents the respective anodic peak potential, and n denotes the total number of anodic peaks considered in the voltammogram. The electrochemical index integrates both thermodynamic contributions, associated with the electron-donating capacity (E _*pa*), and kinetic aspects, related to the concentration of electroactive species at the electrode interface (I _*pa*), as previously reported by Lino et al. (2014).

Prior to analysis, wine samples were diluted in a 0.1 M phosphate buffer solution (pH 5.0) at a volumetric ratio of 0.1:1 (v/v). After each voltammetric measurement, the surface of the carbon paste electrode was renewed mechanically by extruding approximately 0.5 mm of the paste, followed by smoothing with filter paper. All measurements were performed at ambient temperature.

2.5 Analysis of volatile compounds in Sauvignon Blanc wines

Volatile compounds were investigated using two complementary GC-MS approaches. HS-SPME-GC-MS was employed to characterize free volatile compounds in the wine headspace, whereas SPE-GC-MS was used to investigate free and glycosidically bound volatile fractions after solid-phase extraction. The combined application of these methods provided complementary information on the volatile and precursor-derived composition of the wines.

2.5.1 Aroma profile by HS-SPME-GC-MS

Volatile compounds were determined in analytical triplicate using headspace solid-phase microextraction coupled to gas chromatography-mass spectrometry (HS-SPME-GC-MS). Extractions were performed with a 75 µm carboxen/polydimethylsiloxane (CAR/PDMS) StableFlexTM fiber (Supelco®), following previously described procedures with modifications (Whitener et al., 2016; Wang et al., 2022).

For sample preparation, 1.0 mL of wine was mixed with 4.0 mL of distilled water in a headspace vial, and 50 µL of an octan-1-ol solution (1 ppm) was added as internal standard. The vials were immediately sealed and equilibrated at 40 °C for 5 min to establish partitioning between liquid and gas phases. Subsequently, the SPME fiber was exposed to the headspace under the same temperature and time conditions (40 °C for 5 min) to allow adsorption of volatile compounds.

After extraction, the fiber was introduced into the GC injector and thermally desorbed at 250 °C for 2 min. Prior to use, fiber conditioning was carried out at 250 °C for 55 min in accordance with the manufacturer’s instructions.

Chromatographic analyses were performed using a Shimadzu QP2010 Plus gas chromatograph equipped with an AOC-5000 autosampler (Shimadzu Corporation, Kyoto, Japan) and coupled to a mass spectrometer. Helium was employed as the carrier gas at a constant flow rate of 1.0 mL min⁻¹. Separation was achieved on a Restek Rxi-5MS capillary column (30 m × 0.25 mm i. d., 0.25 µm film thickness; 5% phenyl, 95% polydimethylsiloxane), with the injector operating in splitless mode.

The mass spectrometer was operated under electron ionization (EI) at 70 eV, with the interface temperature set at 250 °C. Data acquisition was performed in full-scan mode over the m/z range of 29–350. The oven temperature program was as follows: initial temperature of 40 °C (2 min), ramped at 4 °C min^–1 to 130 °C (2 min hold), followed by an increase at 8 °C min^–1 to 250 °C (4 min hold). The total run time was 45.5 min.

2.5.2 Aroma profile by SPE-GC-MS

The volatile profile was further investigated by solid-phase extraction coupled to gas chromatography-mass spectrometry (SPE-GC-MS), following Boido et al. (2003) with minor modifications. For each analysis, 50 mL of wine was diluted to 100 mL with distilled water, and 0.10 mL of an octan-2-ol solution was added as internal standard.

SPE was performed using Isolute ENV + cartridges [hyper-cross-linked hydroxylated poly (styrene–divinylbenzene) copolymer, IST Ltd., Mid Glamorgan, United Kingdom], which were conditioned with 15 mL of methanol and equilibrated with 20 mL of water. After sample loading, the cartridges were washed with 20 mL of water. The free volatile fraction was eluted with 30 mL of dichloromethane at a controlled flow rate of 4–5 mL min^-1, and the eluate was stored at low temperature to facilitate removal of residual water. The glycosidically bound fraction was subsequently recovered by elution with 30 mL of methanol and stored under refrigeration until hydrolysis.

The free fraction was dried over anhydrous Na₂SO₄, filtered through glass wool, and concentrated using a Vigreux column at 45 °C to approximately 4 mL. Prior to analysis, the extract was further reduced to ∼200 µL under a gentle nitrogen stream using a vial insert.

For the bound fraction, the methanolic extract was evaporated to dryness under reduced pressure and reconstituted with 6.0 mL of citrate buffer (pH 5.0), added in three 2.0 mL portions. Enzymatic hydrolysis was performed by adding 1.0 mL of an enzyme preparation (polygalacturonase/pectinase with β-glucosidase activity), followed by incubation at 40 °C for 16 h in sealed vials. After cooling in an ice bath, 100 µL of octan-2-ol (230 ppm in 50% hydroalcoholic solution) was added, and the hydrolyzed extract was subjected to liquid-liquid extraction using pentane:dichloromethane (2:1, v/v). Extraction involved successive cycles of vortex mixing, phase separation by centrifugation (15 min), and pooling of the organic layers. The combined extracts were dried, when necessary, over Na₂SO₄ and concentrated under nitrogen as described for the free fraction.

GC-MS analyses were carried out on a Shimadzu QP2010 SE gas chromatograph coupled to a QP-2020 mass spectrometer (Shimadzu Corporation, Kyoto, Japan). Separation was achieved on a BP20 fused-silica capillary column (25 m × 0.25 mm i. d., 0.25 µm film thickness; polyethylene glycol phase, SGE, Ringwood, Australia). A 1.0 µL aliquot was injected in split mode (1:40), using hydrogen as the carrier gas at a flow rate of 1.30 mL min⁻¹.

The mass spectrometer operated in electron ionization (EI) mode at 70 eV, with the interface temperature set at 240 °C. Data acquisition was performed in full-scan mode over the m/z range of 40–350. The oven temperature program was: 40 °C (8 min), increased at 3 °C min^–1 to 180 °C, then at 5 °C min^–1 to 220 °C (10 min hold), followed by a ramp of 20 °C min^–1 to 235 °C (15 min hold). The total run time was 88.4 min.

2.5.3 Putative identification and semi-quantitative assessment of volatile compounds

Mass spectral data were compared with established reference libraries (Babushok and Zenkevich, 2009; McLafferty and Stauffer, 1991; Mondello, 2015). Tentative assignments were based on the combined agreement of electron-ionization mass spectra and experimental linear retention indices with library and literature data obtained under similar analytical conditions. Because authentic analyte standards were not used, none of the volatile compound identities was confirmed by co-injection.

Linear retention indices were determined using a homologous series of n-alkane standards (C₇-C₃₀), following the Van den Dool and Kratz (1963) equation (Equation 3):

LRI = 100 n + 100 t r x − t r n t r n + 1 − t r n (3)

In this equation, n represents the number of carbon atoms of the n-alkane eluting immediately before the analyte; t_rx corresponds to the retention time of the analyte; t_rn is the retention time of the preceding n-alkane; and t_r(n+1) refers to the retention time of the subsequent n-alkane. The calculated LRI values were compared with literature data reported by Babushok and Zenkevich (2009) as well as with entries available in the NIST Chemistry WebBook.

For HS-SPME-GC-MS analyses, compound abundance was considered semi-quantitative, as calibration curves and individual response factors were not determined for all analytes. Semi-quantitative estimates were obtained by normalizing the chromatographic peak area of each compound to that of the internal standard. Results were expressed relative to the original wine volume according to Equation 4:

Semi − quantitative estimate µ g L − 1 = A 1 A 2 × C p × V p V i × 1000 (4)

In this expression, A₁ corresponds to the peak area of the analyte, A₂ represents the peak area of the internal standard, C_p denotes the concentration of the internal standard solution (1.0 mg L^-1), V_p is the volume of internal standard solution added (0.050 mL), and V_i is the volume of wine used (1.0 mL). Accordingly, the obtained values were interpreted as semi-quantitative estimates rather than absolute concentrations.

2.6 Statistical analysis

Physicochemical parameters, Folin-Ciocalteu reducing capacity, flavonoid content, antioxidant-related measurements, and volatile compound data were compiled and processed in RStudio version 4.4.2.

As the wines corresponded to aliquots derived from a single fermentation per treatment, the replicates did not constitute independent biological units. Moreover, spontaneous malolactic fermentation (MLF) progressed at different rates between treatments (T1 and T2), and analytical measurements were not fully aligned with a common MLF stage. Consequently, inferential statistical analyses based on significance testing were not performed, and the results were interpreted using a descriptive and exploratory approach. Data obtained from HS-SPME-GC-MS and SPE-GC-MS analyses were further explored through Venn diagrams to identify compounds unique to or shared among treatments.

3 Results

3.1 Physicochemical analyses of wines produced from the Sauvignon Blanc cultivar

The wines fermented with S. cerevisiae alone (T1) and co-inoculated with H. vineae and S. cerevisiae (T2) showed different physicochemical profiles (Table 1).

TABLE 1. Physicochemical characteristics of Sauvignon Blanc wines.

T1 - Sauvignon Blanc wine fermented with Saccharomyces cerevisiae; T2 - Sauvignon Blanc wine co-inoculated with Hanseniaspora vineae and Saccharomyces cerevisiae. Values are expressed as mean ± standard deviation of repeated analytical measurements. The standard deviation represents analytical variability only; no independent fermentation replicates were available.

Regarding pH, a higher value was observed for T2 (3.46) compared to T1 (3.12). Differences were also observed in the concentrations of malic and lactic acids. T2 exhibited low levels of malic acid (0.10 g L^-1) and higher concentrations of lactic acid (1.90 g L^-1), whereas T1 maintained higher malic acid content (4.50 g L^-1) and lower lactic acid concentration (0.40 g L^-1).

Total acidity was lower in T2 (6.00 g L^-1) compared to T1 (10.90 g L^-1). After alcoholic fermentation, T1 and T2 showed alcohol contents of 12.40% and 12.90% (v/v), respectively. Residual sugar levels remained low in both wines, ranging from 1.40 g L^-1 in T1 to 1.70 g L^-1 in T2.

Volatile acidity was slightly higher in T2 (0.60 g L^-1) than in T1 (0.40 g L^-1). A slight reduction in glycerol concentration was also observed in the co-inoculated wine, with values of 8.10 g L^-1 in T2 and 8.70 g L^-1 in T1.

3.2 Treatment-associated differences in Folin-Ciocalteu reducing capacity, flavonoid content, and antioxidant-related measurements

3.2.1 Folin-Ciocalteu reducing capacity and flavonoid content

The Folin-Ciocalteu reducing capacity and flavonoid content measured in the Sauvignon Blanc wines are shown in Table 2.

TABLE 2. Folin-Ciocalteu reducing capacity and flavonoid content of Sauvignon Blanc wines.

T1 - Sauvignon Blanc wine fermented with Saccharomyces cerevisiae; T2 - Sauvignon Blanc wine co-inoculated with Hanseniaspora vineae and Saccharomyces cerevisiae. Values are expressed as mean ± standard deviation of repeated analytical measurements. The standard deviation represents analytical variability only; no independent fermentation replicates were available. Results are presented descriptively.

The Folin-Ciocalteu reducing capacity was higher in T1 (237.70 mg GAE L^-1) than in T2 (166.50 mg GAE L^-1). Flavonoid content showed a different pattern, with values of 31.80 mg QE L^-1 in T1 and 33.20 mg QE L^-1 in T2. Because the Folin-Ciocalteu assay is not specific to phenolic compounds and free and total SO₂ were not measured, the difference in Folin-Ciocalteu response cannot be attributed solely to phenolic composition.

3.2.2 Antioxidant-related radical-scavenging and electrochemical responses

Radical-scavenging and electrochemical responses were evaluated using complementary spectrophotometric and electroanalytical approaches. The DPPH^• and ABTS^•+ assays were used to estimate radical-scavenging capacity. Results were expressed as IC₅₀, defined as the nominal wine concentration required to inhibit 50% of radical activity. The IC₅₀ values are presented in Table 3, and the corresponding concentration-response curves are shown in Figure 1.

TABLE 3. Antioxidant-related measurements of Sauvignon Blanc wines.

T1 - Sauvignon Blanc wine fermented with Saccharomyces cerevisiae; T2 - Sauvignon Blanc wine co-inoculated with Hanseniaspora vineae and Saccharomyces cerevisiae. Values are expressed as mean ± standard deviation of repeated analytical measurements. The standard deviation represents analytical variability only; no independent fermentation replicates were available. Results are presented descriptively.

Lower IC₅₀ values were measured for T2 than for T1 in both the DPPH^• and ABTS^•+ assays. Because independent fermentation replicates were not available, these differences are reported descriptively and cannot be interpreted as reproducible effects of the inoculation condition.

Differential-pulse voltammograms obtained for T1 and T2 in 0.1 M phosphate buffer (pH 5.0), using a carbon-paste electrode, showed three anodic processes at approximately 0.25, 0.40, and 0.75 V (Figure 2). Freshly prepared orange juice was analyzed under the same voltammetric conditions as a comparative antioxidant-containing reference. Because orange juice and wine are compositionally distinct matrices, the comparison was used only to contextualize the electrochemical response and was not interpreted as evidence of direct equivalence between the samples.

The electrochemical index (EI) was calculated from anodic peak potential (E_pa) and anodic peak current (I_pa), as described in the Methods. The EI values of the two wines were higher than that of the freshly prepared orange juice analyzed under the same conditions. T2 showed a slightly higher EI (10.98 μA mV^-1) than T1 (10.25 μA mV^-1).

The voltammetric profiles further indicated that the anodic processes 1a and 2a exhibited reversible characteristics, as evidenced by the similarity between the currents of the anodic peaks and their corresponding cathodic signals. In contrast, the redox process 3a showed irreversible behavior.

3.3 Treatment-associated differences in the volatile composition of Sauvignon Blanc wines

3.3.1 Tentative identification of free volatile compounds in Sauvignon Blanc wines

The volatile profile of Sauvignon Blanc wines obtained either by S. cerevisiae monoculture or by co-inoculation with H. vineae was investigated using two complementary GC-MS approaches: HS-SPME-GC-MS and SPE-GC-MS.

Tentative compound identification combined electron-ionization mass spectra, experimental linear retention indices (ELRI), and literature information, following the criteria described in Section 2.5.3. Because authentic analyte standards were not used, all assignments are reported as tentative. The HS-SPME-GC-MS data represent semi-quantitative internal-standard-equivalent estimates rather than validated absolute concentrations.

A total of 57 free volatile compounds were tentatively identified across the two Sauvignon Blanc wines (Tables 4; Supplementary Table S1), including esters, lactones, alcohols, ketones, aldehydes, carboxylic acids, phenolic derivatives, sulfur-containing compounds, and terpenoid-related compounds.

TABLE 4. Free volatile compounds tentatively identified by HS-SPME-GC-MS in Sauvignon Blanc wines.

Compounds were tentatively identified based on EI, mass spectral matching and ELRI/LLRI comparison. Values are expressed as mean ± standard deviation of repeated analytical measurements and should be interpreted as semi-quantitative internal-standard-equivalent estimates rather than validated absolute concentrations. *Compounds eluting outside the retention-time window covered by the n-alkane standard series. **Not reported in the literature source used for LLRI comparison. Nd, not detected. ELRI, experimental linear retention index; LLRI, literature linear retention index from the NIST Chemistry WebBook.

Differences between treatments were observed both qualitatively and in semi-quantitative terms. In total, 31 compounds were common to both wines, while each treatment showed a smaller set of compounds detected only under one condition. The Venn diagram shown in Figure 3 provides an exploratory visualization of shared and treatment-associated free volatile compounds.

The clearest semi-quantitative difference involved 2-phenylethanol and 2-phenylethyl acetate. The internal-standard-equivalent estimate of 2-phenylethanol was 43.88 μg L^-1 in T1 and 16.07 μg L^-1 in T2, whereas the estimate of 2-phenylethyl acetate was 1.53 μg L^-1 in T1 and 28.57 μg L^-1 in T2. These values are semi-quantitative and should not be interpreted as validated absolute concentrations or as evidence of a reproducible effect of the inoculation condition.

Other ester compounds detected in the wines included ethyl butanoate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, 3-methylbutyl acetate, and hexyl acetate. Alcohols were also abundant, including 3-methylbutan-1-ol, 2-methylpropan-1-ol, butan-1-ol, hexan-1-ol, and 2-phenylethanol.

Aldehydes such as octanal, nonanal, decanal, dodecanal, and benzaldehyde were detected. Ketones and furanic compounds, including acetoin, 2-octanone, furfural, and 5-(hydroxymethyl)furan-2-carbaldehyde, were also observed.

Carboxylic acids, including butanoic, hexanoic, octanoic, dec-9-enoic, and phenylacetic acids, were mainly identified by SPE-GC-MS. Sulfur-containing compounds, such as 3-(methylsulfanyl)propan-1-ol and 4,5-dimethyl-1,3-oxathiolane-2-thione, were also detected.

3.3.2 Tentative identification of glycosylated volatile compounds in Sauvignon Blanc wines

After SPE extraction and enzymatic hydrolysis, 19 compounds were tentatively identified by GC-MS in the glycosidically bound fraction, with two compounds detected only in T2 (Supplementary Table S2). As with the free volatile fraction, these assignments are provisional because they were based on EI mass spectral matching and retention-index comparison rather than confirmation with authentic standards.

A relevant feature of the volatile compounds released from the glycosidically bound fraction was the presence of terpenoid- and norisoprenoid-related compounds, including trans-linalool oxide isomers, geraniol, citronellol, p-menth-1-ene-7,8-diol, perilla alcohol (E,E)-geranyl linalool, 3-oxo-alpha-ionol, and 14-hydroxy-4,5-dihydro-beta-caryophyllene.

Alcohols such as 3-methyl-3-buten-1-ol, 3-methyl-2-buten-1-ol, pentan-1-ol, and nonan-1-ol were also detected in the hydrolyzed fraction.

The detection of methyl salicylate and methyl 2,4-dihydroxybenzoate indicated the presence of benzenoid-related compounds in the bound volatile fraction. In parallel, acetic acid and 2-oxoglutaric acid were identified. Ketone- and furanone-related compounds, including 4-hydroxy-4-methyl-2-pentanone and 3,5-dihydroxy-6-methyl-2,3-dihydro-4H-pyran-4-one, were also observed.

4 Discussion

Physicochemical characterization is essential for assessing wine composition, technological status, and compliance with regulatory standards. In the present study, the physicochemical differences observed between T1 and T2 were mainly associated with acidity-related parameters. However, only one fermentation was conducted for each condition, and spontaneous malolactic fermentation (MLF) progressed differently between the two wines. Therefore, the findings should be interpreted as a descriptive and exploratory comparison of two individual Sauvignon Blanc vinifications, rather than as evidence of an effect caused by H. vineae. Spontaneous malolactic fermentation was a major confounding factor in this comparison. Because its progression differed substantially between the wines, differences in physicochemical profile cannot be separated from possible effects associated with the inoculation condition. The results therefore represent the combined outcomes of two individual vinifications.

The higher pH and lower total acidity measured in T2 were consistent with its more advanced malolactic fermentation, as indicated by the lower malic acid and higher lactic acid concentrations. The conversion of L-malic acid to L-lactic acid reduces titratable acidity and generally increases wine pH, thereby altering the chemical balance of the wine (Knoll et al., 2011). Physicochemical characteristics of Sauvignon Blanc wines produced in other Brazilian regions have also been reported to vary according to vineyard conditions, harvest maturity, and winemaking practices (Filho et al., 2021; Filho et al., 2025). It is important to note that malolactic fermentation is not typically desired in the production of fresh white wines, such as Sauvignon Blanc, in which the preservation of acidity is a key sensory attribute (Knoll et al., 2011). However, the co-inoculated treatment showed clear evidence of advanced malic acid conversion. Therefore, the results should be interpreted as an exploratory chemical comparison between wines, not as direct evidence that H. vineae alone induced this effect.

The present results do not establish why spontaneous malolactic fermentation progressed to a greater extent in T2. Interactions between non-Saccharomyces yeasts and lactic acid bacteria have been reported, but this possibility was not investigated because bacterial populations, yeast persistence, and relevant metabolites were not monitored (Virdis et al., 2021). Total acidity followed the same pattern, being lower in T2 than in T1, consistent with the greater conversion of malic acid to lactic acid. Such a change may alter perceived acidity; however, no sensory analysis was performed, and its perceptual relevance in these wines was not evaluated.

The alcohol contents of T1 and T2 were compatible with the initial soluble-solids content of the must (22 °Brix), and residual sugar levels were low in both wines. These results indicate that alcoholic fermentation reached a similar dry endpoint in the two vinifications. Volatile acidity was slightly higher in T2 than in T1. Because no sensory analysis was conducted, the presence or absence of perceptible volatile-acidity-related attributes cannot be inferred from these values. Glycerol concentrations were similar in the two wines, with a slightly lower value in T2. Given the absence of independent fermentation replicates and the different post-fermentation evolution of the wines, this small difference cannot be assigned to the inoculation condition.

The phenolic composition of wines derives mainly from compounds extracted from grape skins, with a smaller contribution from seeds. These constituents are closely associated with wine quality, influencing color, astringency, oxidative behavior, aroma-related properties, and overall sensory perception (Morata et al., 2019a; Morata et al., 2019b). Phenolic compounds in wines are commonly classified into flavonoids and non-flavonoids. Flavonoids include flavanols, flavonols, anthocyanins, flavones, and flavanones, whereas non-flavonoids comprise mainly phenolic acids and related derivatives, such as hydroxycinnamic and hydroxybenzoic acids (Zhang et al., 2021).

A higher Folin-Ciocalteu response was measured in T1 than in T2. The assay reflects Folin-reactive reducing substances and is not specific to phenolic compounds; other reducing constituents, including SO₂, may contribute to the signal. Because free and total SO₂ were not determined and no independent fermentations were conducted, the observed difference should not be interpreted as a specific change in phenolic composition caused by the inoculation condition.

Flavonoid content was similar in the two wines, with a slightly higher analytical value in T2 than in T1. Given the small difference and the absence of independent fermentation replicates, no treatment effect can be inferred. In white wines, flavonoids may influence bitterness, mouthfeel, oxidative stability, and color development even at lower concentrations than in red wines (He et al., 2010).

The DPPH^• and ABTS⁺• assays are widely used to estimate the radical-scavenging capacity of wine samples, particularly in relation to phenolic compounds with reducing properties (Granato et al., 2018). In both assays, antioxidant compounds promote the reduction of stable radical species, leading to a decrease in absorbance proportional to the radical-scavenging response. Since the results were expressed as IC₅₀, lower IC₅₀ values indicate higher antioxidant capacity. Electrochemical analysis provides a complementary perspective when evaluating the redox behavior of electroactive compounds present in wines. Since antioxidant activity is closely related to electron-transfer processes, voltammetric techniques can provide information on the ease of oxidation and the relative contribution of redox-active constituents to the overall antioxidant profile (Lino et al., 2014; Benbouguerra et al., 2020).

The partial overlap between peaks 1a′ and 1a’’ suggests a higher concentration and diversity of oxidized electroactive species within this potential range, as observed in Figure 2. E _*pa* values reflect thermodynamic aspects related to the ability of electroactive species to undergo oxidation and participate in electron-transfer reactions, whereas I _*pa* values are associated with the concentration of these species and kinetic aspects of electron transfer. Thus, EI provides a useful estimate of the antioxidant-related redox potential of the samples (Macêdo et al., 2017).

T2 showed lower DPPH^• and ABTS^•+ IC₅₀ values and a slightly higher EI than T1. These measurements describe the two individual wines but cannot be interpreted as a reproducible effect of the inoculation condition. The assays probe different aspects of reducing capacity and electrochemical behavior and therefore need not follow the same pattern as the Folin-Ciocalteu response.

In the two wines evaluated, a higher Folin-Ciocalteu response did not coincide with lower radical-scavenging IC₅₀ values or a higher electrochemical index. This lack of concordance is compatible with the different chemical principles of the assays and with differences in the reactivity of the compounds present, but no causal or general conclusion can be drawn from this pair of samples (Eder et al., 2023).

Beyond these antioxidant-related responses, the wines also differed in their volatile profiles. Wine aroma is strongly shaped by volatile metabolites originating from grape precursors, fermentative activity of yeasts, and chemical reactions occurring during post-fermentation evolution (Whitener et al., 2016; Moreira et al., 2011; Tofalo et al., 2016). In the present study, 57 free volatile compounds were tentatively identified across the Sauvignon Blanc wines, including esters, lactones, alcohols, ketones, aldehydes, carboxylic acids, phenolic derivatives, sulfur-containing compounds, and terpenoid-related compounds. Among these classes, esters and alcohols were particularly relevant, in agreement with previous studies showing that these compounds are major contributors to fruity, floral, and fermentative notes in wines (Barbará et al., 2020; Sánchez-Palomo et al., 2017; Ubeda et al., 2022; Zhao et al., 2017). It should be noted that no sensory analysis was performed in the present study. Therefore, the odor descriptors associated with the tentatively identified compounds should not be interpreted as proven sensory attributes of the wines evaluated.

Several compounds were detected in both wines, indicating a common volatile profile associated with the grape matrix and general fermentation conditions. Differences in selected compounds, particularly within the ester fraction, may reflect differences in precursor availability, microbial metabolism, and post-fermentation evolution. The most pronounced difference involved 2-phenylethanol and 2-phenylethyl acetate. T2 showed a higher semi-quantitative estimate of 2-phenylethyl acetate than T1. Previous studies have associated this compound with floral and fruity odor descriptors. This pattern is consistent with previous studies of H. vineae fermentations, which have reported enhanced 2-phenylethyl acetate (Carrau et al., 2023; Fresno et al., 2020; 2021; Giorello et al., 2019; Lleixà et al., 2016; Valera et al., 2021; 2024; Viana et al., 2008; 2009; 2011). However, the compound was only tentatively identified and semi-quantitatively assessed, and no sensory difference was demonstrated in the present wines.

From a metabolic perspective, the pattern is compatible with a difference in phenylalanine-derived higher-alcohol and ester metabolism. Through the Ehrlich pathway, phenylalanine may be converted to phenylpyruvate, phenylacetaldehyde, and then 2-phenylethanol, which can be acetylated through acetyl-CoA-dependent reactions to form 2-phenylethyl acetate (Hazelwood et al., 2008). The lower estimate of 2-phenylethanol and the higher estimate of 2-phenylethyl acetate in T2 are therefore biologically plausible, but the data do not demonstrate direct metabolic conversion or causation.

Because only one fermentation was conducted per condition, malolactic fermentation progressed differently, and yeast population dynamics were not monitored, the observed pattern should be considered specific to the two wines evaluated. Precursor consumption, enzyme activity, and metabolic flux were not measured.

The reproducibility of this pattern and its relationship to H. vineae should be tested in replicated fermentations with controlled malolactic fermentation and microbial monitoring.

Other ester compounds also contributed to the volatile profile of the wines, including ethyl butanoate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, 3-methylbutyl acetate, and hexyl acetate, which are commonly associated with fruity, apple-like, banana-like, and tropical notes. These esters are formed mainly through yeast-mediated reactions between alcohols and acyl-CoA derivatives, catalyzed by alcohol acetyltransferases such as Atf1 and Atf2 (Sumby et al., 2010). Their presence reinforces the importance of fermentative metabolism in shaping the aroma profile of these Sauvignon Blanc wines.

Higher alcohols can contribute to aromatic complexity and mouthfeel, although their sensory impact depends strongly on concentration and matrix interactions. Hexan-1-ol and (Z)-hex-3-en-1-ol are relevant for Sauvignon Blanc because they are associated with green, grassy, and herbaceous notes, which are part of the typical sensory identity of this cultivar.

Aldehydes such as octanal, nonanal, decanal, dodecanal, and benzaldehyde may contribute citrus, fatty, green, waxy, and almond-like nuances, while ketones and furanic compounds, including acetoin, 2-octanone, furfural, and 5-(hydroxymethyl)furan-2-carbaldehyde, suggest contributions from fermentative and oxidative pathways.

Carboxylic acids, including butanoic, hexanoic, octanoic, dec-9-enoic, and phenylacetic acids, may contribute to fatty, acidic, cheesy, or honey-like nuances depending on their concentrations, while also acting as precursors for ester formation. Sulfur-containing compounds, such as 3-(methylsulfanyl)propan-1-ol and 4,5-dimethyl-1,3-oxathiolane-2-thione, suggest the contribution of sulfur metabolism to the aroma profile. Although varietal thiols typically associated with Sauvignon Blanc aroma were not specifically confirmed in this study, the presence of sulfur-containing volatiles indicates that sulfur-derived pathways may have contributed to aromatic complexity.

Glycosylated aroma precursors constitute an important latent reservoir of varietal aroma compounds in grapes and wines. During fermentation and aging, enzymatic or acid-catalyzed hydrolysis may release volatile aglycones, contributing to the development of wine aroma complexity over time (Liu et al., 2017). The presence of terpenoid- and norisoprenoid-related compounds in the glycosidically bound fraction is relevant because glycosidically bound compounds constitute an important non-volatile aroma reservoir in grapes and wines. Although odorless in their conjugated form, they may release volatile aglycones through enzymatic or acid hydrolysis and during wine evolution, thereby contributing to aroma complexity.

In this context, the detection of linalool oxides, geraniol, citronellol, and ionone-related derivatives suggests the presence of terpene- and norisoprenoid-derived precursors in the Sauvignon Blanc wines. Although Sauvignon Blanc is more commonly associated with volatile thiols and methoxypyrazines, other volatile families, including esters, alcohols, terpenoids, and oxidation-related compounds, may also contribute to its aroma profile depending on grape composition, fermentation conditions, and wine evolution (Coetzee and Toit, 2012). Therefore, the occurrence of terpene-derived compounds in the hydrolyzed fraction may indicate a secondary contribution to floral, citrus-like, resinous, and balsamic nuances, particularly after precursor release.

Norisoprenoid-related compounds, such as 3-oxo-alpha-ionol, are also relevant because they originate from carotenoid degradation and may occur as glycoconjugated precursors, with potential sensory relevance due to their low odor thresholds (Mendes-Pinto, 2009).

Alcohols such as 3-methyl-3-buten-1-ol, 3-methyl-2-buten-1-ol, pentan-1-ol, and nonan-1-ol may reflect the release of bound alcohols or secondary transformations occurring during hydrolysis and post-fermentation evolution. Although they are not primary markers of Sauvignon Blanc aroma, they may contribute green, alcoholic, fatty, or slightly fruity nuances depending on concentration and matrix interactions. Their occurrence should therefore be interpreted as part of the broader volatile precursor pool rather than as cultivar-specific markers.

The detection of methyl salicylate and methyl 2,4-dihydroxybenzoate indicates the presence of benzenoid-related compounds in the bound volatile fraction. Methyl salicylate may contribute sweet, minty, balsamic, or medicinal notes, whereas hydroxybenzoate derivatives may reflect phenolic-related precursor chemistry. In parallel, acetic acid and 2-oxoglutaric acid were identified, although their direct aromatic relevance is limited compared with terpenoids, norisoprenoids, and benzenoids. These compounds may instead reflect fermentative metabolism, organic acid composition, or co-extracted compounds detected after hydrolysis.

Ketone- and furanone-related compounds, including 4-hydroxy-4-methyl-2-pentanone and 3,5-dihydroxy-6-methyl-2,3-dihydro-4H-pyran-4-one, should be interpreted cautiously, since highly oxygenated compounds may arise from transformation reactions during sample preparation, hydrolysis, or wine evolution. Nevertheless, their presence may indicate the contribution of oxidative or transformation pathways to the pool of compounds detected in the hydrolyzed fraction.

Non-Saccharomyces yeasts may contribute to aroma modulation through enzymatic activities involved in precursor hydrolysis, particularly β-glucosidase-related pathways capable of releasing terpene- and norisoprenoid-derived aglycones from glycosylated grape precursors (Padilla et al., 2016). Qualitative differences were observed between the two wines in the hydrolyzed fraction. However, because no independent fermentations were conducted and enzymatic or microbial mechanisms were not directly investigated, these differences cannot be attributed specifically to the inoculation condition.

5 Conclusion

This study provides an exploratory chemical characterization of two Sauvignon Blanc vinifications produced in the Brazilian Cerrado from grapes cultivated under a double-pruning system. The wines differed in physicochemical profile, Folin-Ciocalteu reducing capacity, flavonoid content, radical-scavenging and electrochemical responses, and volatile composition. T2 showed lower malic acid and higher lactic acid, indicating more advanced spontaneous malolactic fermentation at the time of analysis.

The semi-quantitative volatile analysis showed a marked difference in the balance between 2-phenylethanol and 2-phenylethyl acetate. This pattern is compatible with a possible difference in phenylalanine-derived higher-alcohol and ester metabolism, but the underlying mechanism was not directly investigated.

Because the study included one fermentation per condition and the wines differed in malolactic fermentation progression, the observed patterns cannot be attributed specifically to H. vineae. The findings are therefore preliminary and hypothesis-generating. Replicated fermentations with controlled malolactic fermentation, microbial monitoring, quantitative validation of key volatile compounds, and sensory analysis will be required to test the reproducibility and sensory relevance of the observed patterns.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frfst.2026.1914854/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Digambar Kavitake, Pondicherry University, India

Reviewed by: Pham Tri Nhut, Seoul National University, Republic of Korea

Luis Enrique Trujillo, University of the Armed Forces, Ecuador