Impact of Long-Term Bottle Aging on Color Transition, Polymers, and Aromatic Compounds in Mulberry Wine

Abstract

Long-term aging has traditionally been associated with issues such as color fading and oxidation; therefore, it limits grape wine production. Here, we analyzed 90 bottles of mulberry wine aged for various periods (up to 12 years) and observed unique trends in color, flavor, and aroma compounds during prolonged aging. Results from Somers and methylcellulose precipitation (MCP) assays indicated that the tannin and anthocyanin concentrations in newly fermented mulberry wines were 167 to 216 mg/L and 1.04 to 1.37 g/L, respectively. The total phenolics, tannins, and anthocyanin contents exhibited significant negative correlations with aging years, while the non-bleachable pigment content and hue showed positive correlations with aging times. High-performance liquid chromatography–electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS) analysis further revealed a positive correlation between the content of pyranoanthocyanins (including cyanidin-3-O-glucoside-pyruvic acid, cyanidin-3-O-glucoside-acetaldehyde, cyanidin-3-O-glucoside-4-vinocatechol, and cyanidin-3-O-glucoside-4-vinophenol) and aging times, whereas the impacts of aging on the polymeric pigment (cyanidin-3-O-glucoside-epicatechin) were not observed. This suggests that the anthocyanins in mulberry wine primarily transformed into pyranoanthocyanins rather than polymeric pigments during aging. The aging-induced reductions in protein, polysaccharide, and key aroma compounds (contributing to the fruity, sweet and floral odors) remained unaffected by prolonged aging.

1. Introduction

The mulberry (Morus alba L.) belongs to the Moraceae family, and its leaves were historically used to feed the silkworms in China’s sericulture, serving as the primary material for silk production [1]. Today, the mulberry has diversified into over 20 species, cultivated extensively across Asia, Africa, and North America [2]. Mulberry fruits, known for their low caloric content, slight acidity, and enjoyable flavor, are popular in Asia. The fruit is considered a fine material for fruit winemaking since it is rich in organic acids (malic acid, tartaric acid, and critic acid, juice pH = 3.5–4.0 [3,4]), pigments (cyanidin-3-rutinoside and cyanidin-3-glucoside, up to 2057 µg/g berry [5,6]), aromatic compounds responsible for the fruity and fresh sensory characters [4], and other polymeric compounds responsible for fullness (proteins and polysaccharides [1]). In China, Japan, and Korea, drinking mulberry wine has a long history and a dedicated consumer base [2]. Unfortunately, only limited studies in the literature are available for mulberry wines. In particular, information on the mechanical processes involved in mulberry winemaking processes is scarce. This significantly restricts the development of techniques for mulberry winemaking.

Aging, a requisite maturation period in vinification, has long been used to reduce astringency and bitterness, heighten aroma, and stabilize color in wines [7]. No matter what kind of fruit is used as the raw material, the aging process is essential for red-colored wines because it involves reactions such as slow oxidation, esterification, condensation, polymerization, thereby imparting the desired sensory profile [7].

For example, the color of grape red wine evolves ‘from purplish-red to brick red’ over time [8]. It is because the colors of young red wines are primarily determined by the free anthocyanins, while the old red wines are severely impacted by anthocyanin derivates. The pyranoanthocyanins originate from the reactions of anthocyanins and other wine molecules like pyruvic acid and acetaldehyde, and the polymeric pigments are the products of condensation between the anthocyanins and tannins [9]. Both pyranoanthocyanins and polymeric pigments show more stable colors than free anthocyanins.

The polymer compounds in wine, including tannins, proteins, and polysaccharides, undergo significant alterations during aging. Throughout aging, the C–C bonds responsible for connecting the subunits within tannin molecules gradually break, leading to the detachment of these units [10]. The molecular chain lengths of the tannins are reduced after aging, thereby impacted the astringency of final wines. The tannins, also influenced by oxygen in wine, can transform into A-type molecules with intramolecular bonds, or evolve into quinone substances which may deepen the wine’s color [11]. The polysaccharides and proteins, originating from yeast or grapes, undergo tannin-induced binding or self-aggregation and partially become lees during vinification [12,13,14]. The aging process changes both the concentration and composition of these polymers, thereby influencing the stability and taste of final wines.

Based on the comprehensive understanding of grape red wine aging mechanisms, various oenological products and techniques have been developed such as inactive dry yeast, toasted oak pieces, heating-cooling process, micro-oxygen aging, ultrahigh pressure aging, pulsed electric field aging, and ultrasound field aging [15,16]. These products and methods can either accelerate the aging process or significantly improve the quality of aged wines.

Grape red wines typically benefit from the aging process due to their wealth of anthocyanins and other phenolic compounds. However, the influence of the aging process on mulberry wine is not fully understood yet. To explore whether mulberry wine is suitable for long-term aging and to examine the effects of prolonged aging on its color, aroma, and flavor compounds, 90 bottles of mulberry wine (produced from 2012 to 2023 and aged in bottles) were collected. The Somers color analysis, methylcellulose precipitation (MCP) assay, high-performance liquid chromatography–electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS), headspace–solid phase microextraction–gas chromatography–mass spectrometry (HS-SPEM-GC-MS) were used to analyze the composition and concentration of pigments, colors, tannins, volatile compounds, polysaccharides and proteins in aged mulberry wines. Finally, the impacts of aging on the composition of mulberry wine were discussed. The results not only provide information for mulberry wine aging, but also provide hints for the aging of other fruit wines.

2. Materials and Methods

2.1. Materials

The mulberry wines were purchased from Yayuan Bio tec Inc. (Shantou, China). To minimize the impact from: (1) mulberry variety; (2) soil conditions (texture, slope, elevation); (3) climate (temperature, precipitation); and (4) winemaking processes on the mulberry wines, all mulberries used for winemaking were Hongguo No.2 mulberries obtained from a fixed area of the “Longhu Mulberry Agricultural Park” located in Jinzao District (23°27′0.27″ N, 116°23′14.55″ E). After destemming and crushing, the mulberries were mixed and then fermented in three fermentation tanks (30 tons for each tank). Samples from these tanks (five bottles per tank, fifteen bottles per vintage) were used to eliminate the effects of mulberry variety and soil conditions. To minimize the impact of climatic variations across different vintages, the aging period of the samples was extended to 11 years (2012 to 2023), during which the average temperature and precipitation levels in six vintages (2012, 2015, 2018, 2019, 2020, and 2023) were very similar (as shown in Table S1). To avoid the influence of winemaking processes, all mulberries were harvested at a closely matched maturity level (9.5–10.5°Baumé) and underwent very similar winemaking procedures (as shown in Protocol S1). After winemaking, the mulberry wines were bottled into 750 mL amber wine bottles and sealed with polyurethane-powdered cork stoppers. They were stored in a cellar at 12 °C with 60% humidity across the aging. For each vintage, 15 bottles of samples were collected and underwent analysis.

2.2. Analysis of Pigments and Colors

The mulberry wines were centrifuged (8000× g, 30 min) prior to analysis, each mulberry wine sample was analyzed (in triplicate) using a modified version of the Somers assay [17,18]. Briefly, a 20 µL aliquot of samples was added to 980 µL of 1 M HCl in a centrifuge tube, then mixed. Samples were left to stand in the dark for 1 h. A further 100 µL of sample was added to 900 µL of buffer solution containing 0.5% (w/v) tartaric acid and 12% (v/v) ethanol (pH 3.4), mixed, and analyzed immediately. Other 100 µL mulberry wine samples were, respectively, mixed with 900 µL of the buffer solutions containing 0.1% (v/v) acetaldehyde or 0.375% (w/v) sodium metabisulfite, left to stand in the dark for 1 h, and then analyzed. For spectral analysis of color measures, a 300 µL aliquot was transferred to a 370 µL 96 well UV plate and the 280 nm, 420 nm and 520 nm absorbances measured using a microplate reader, corrected using an appropriate sample blank and a water constant correct function to approximate a 1 cm path length. Estimates of total anthocyanin, non-bleachable (bisulfite-resistant) pigment, color density and hue were determined from spectral data according to the calculations outlined in the published method [18], while total phenolics were calculated referenced with a standard curve made by gallic acid and expressed as gallic acid equivalent (GAE) concentrations.

2.3. Analysis of Tannins

Tannin concentrations were measured in mulberry wines in duplicate using the MCP assay [17]. A 350 μL sample was combined with a 3 mL of 0.04% (w/v) methyl cellulose (Sigma-Aldrich, St. Louis, MO, USA) solution, a 2 mL aliquot of saturated ammonium sulfate solution and 4.65 mL water to give a final volume of 10 mL. A sample control was included where water was added in place of the methyl cellulose solution. Samples were mixed by vortex for 3 min and left to stand for 10 min, then centrifuged for 30 min at 8000× g. An aliquot of the supernatant (2 mL) from the methylcellulose-treated and control samples were transferred into a quartz cuvette; 280 nm absorbance was measured using a UV–Vis spectrometer (UV-1800PC, AOE instrument, Shanghai, China). The difference in 280 nm absorbance between control and MCP-treated samples was measured, and tannin was quantified using (−)-epicatechin as the quantitative standard.

2.4. Analysis of Pyranoanthocyanin and Polymeric Pigments

The pyranoanthocyanin and polymeric pigments were characterized and quantified by HPLC-ESI-MS/MS. as previously described [19]. Briefly, a 3 mL aliquot of the sample was loaded onto an Oasis HLB cartridge (Waters Australia, Rydalmere, NSW, Australia), which was conditioned by methanol, followed by water. The cartridge was then washed with 3 mL of water to remove salts. Then, the sample was recovered with 2 mL of methanol and concentrated under stream nitrogen to about 100 µL, to which 1 mL of acetonitrile solution (formic acid: acetonitrile: water = 5:15:80, by volume) was added for the analysis. A 10 µL of the solution was injected into the HPLC-ESI-MS/MS system (Agilent 1290II-6460, Palo Alto, CA, USA) equipped with a Synergi Hydro-RP (4 u, 80 A, 150 mm × 2 mm) column (Phenomenex, Torrance, CA, USA). A binary gradient was used, where mobile phases A and B contained formic acid–water–acetonitrile in volume ratios of 5:95:0, and 5:15:80, respectively. The flow rate was 300 µL/min, and the linear gradient elution conditions were 0 to 45 min, 0 to 25% solvent B; 45 to 46 min, 25 to 90% solvent B; 46 to 60 min, 90 to 5% solvent B.

ESI spectra were recorded from m/z 200 to 2000 with a step mass size of 0.2 and a dwell time of 0.4 msec. For confirmation of the fragments, MS/MS with product ion scans was carried out with a dwell time of 0.5 ms and a step mass size of 0.2, with a 5000 V ESI needle potential; ring potentials varied from 30 to 250 V. The collision energy and orifice potential were varied between 15 and 50 eV to obtain an optimized cleavage. Multiple reaction monitoring (MRM) was carried out with a dwell time of 200 msec for each reaction. Based on the results obtained from multichannel acquisition (MCA) analysis and referenced with previous reports, the ion reactions selected were m/z 517 → 359 (cyanidin-3-O-glucoside-pyruvic acid), m/z 473 → 311 (cyanidin-3-O-glucoside-acetaldehyde), m/z 581 → 419 (cyanidin-3-O-glucoside-4-vinylcatechol), m/z 565 → 403 (cyanidin-3-O-glucoside-4-vinylphenol), m/z 737 → 575 (cyanidin-3-O-glucoside-(epi)catechin) [9,20,21]. Concentrations of the pigments were calculated referencing with a standard curve made by cyanidin-3-O-glucoside (C3G) and expressed as C3G equivalent (C3GE) concentrations.

2.5. Analysis of Polysaccharides and Proteins

For each vintage year, 5 bottles of mulberry wine were randomly chosen and provided for the following measurements. A total of 1 mL of the mulberry wine was combined with 5 mL of absolute ethanol and allowed to precipitate at 4 °C for 18 h. Then, the samples were centrifuged at 8000× g for 5 min to obtain pellets. The pellets were subjected to dialysis against water using a dialysis bag with a molecular weight cutoff of 3.5 kDa. After dialysis, the samples were reconstituted in water, followed by freezing and lyophilization. Subsequently, the lyophilized samples were analyzed for both polysaccharides and proteins.

The polysaccharide analysis was performed in accordance with a previously published protocol [18] and modified as follows. 10 mg sample was taken and dissolved in a 2 mL tetrahydrofuran solution (2 M) prior to hydrolysis at 120 °C for 2 h. The hydrolysates were cooled on ice, concentrated under nitrogen gas, and resuspended in a 0.5 mL of NaOH solution (0.3 M), then mixed with a 0.5 mL of 1-phenyl-3-methyl-5-pyrazolone (PMP) solution (0.3 M) and derivatized at 70 °C for 30 min. The derivatized sample was then neutralized by adding a 1 mL HCl solution (0.15 M), and then subjected to dichloromethane extraction (0.7 mL). The resulting aliquot phase underwent HPLC analysis to quantify the monosaccharides released from the polysaccharides. The HPLC analysis was performed according to a previously described protocol [22] with the following modifications: a 10 µL sample were injected into an LC16 HPLC (SHIMAZU, Kyoto, Japan) equipped with a 5C18-PAQ C-18 column (250 mm × 4.6 mm) column (COSMOSIL, Kyoto, Japan). Separation was achieved with a solvent system of 0.1 M Na2HPO4-NaH2PO4 solution (solvent A) and an acetonitrile (solvent B) at a flow rate of 1 mL/min with an isocratic elution (80% A, 20% B) for 35 min. Residual amino acids were also derivatized according to a previously described protocol [22] and quantified by HPLC. The monosaccharides and amino acids were identified and quantified in accordance with standard samples.

2.6. Analysis of Volatile Compounds

The analysis of volatile compounds in mulberry wine was performed through the HS-SPME-GC-MS analysis. For each vintage year, 5 bottles of mulberry wine were randomly chosen for the HS-SPME-GC-MS measurement. The method for analysis was taken out in accordance with a previous report [23] but modified as follows. A 5 mL of mulberry wine was transferred into a 15 mL glass vial then sealed with a Teflon-lined cap. The vial was then inserted into an SPME workstation (Kongning, Ningbo, China) equipped with a 50/30 μm DVB/CAR/PDMS fiber (Supelco, Philadelphia, PA, USA). The extraction mode, preheating temperature, preheating time, heating temperature, heating time, and stirring speed were set as follows: gaseous phase exposure, 60 °C, 5 min, 60 °C, 30 min and 500 rpm, respectively. After extraction, the fiber was inserted into a QP-2010UItra gas chromatograph–mass spectrometer (Shimadzu, Kyoto, Japan) equipped with a split/splitless injector (230 °C) and desorbed for 6 min; the volatile compounds were directly transferred to the column. The separation was achieved using a J&W DB-WAX 122-7062 fused silica capillary column (60 m × 0.25 mm, Agilent, USA). The GC oven temperature was programmed from 50 to 240 °C at a rate of 2.0 °C/min, then to 280 °C at a rate of 20 °C/min. Helium was used as carrier gas with a 25 kPa inlet pressure. MS scan conditions, including source temperature, interface temperature and energy, were set as 150 °C, 230 °C, and 70 eV. For each analysis, a 500 µL of ethyl butyrate (10 µg/mL) was used as the internal standard and mixed with the mulberry wine sample prior to SPME extraction. The retention indices (RI) of the compounds were calculated and compared with a previous report and are presented in Table S2.

3. Results and Discussion

3.1. Impact of Aging on the Composition of Phenolic Compounds in Mulberry Wine

Aging, especially long-term aging, is highly reliant on the content of polyphenols, because these compounds work against the negative influence of oxygen during aging [24].

As shown in Figure 1A,B, the 90 bottles of aged mulberry wines showed total phenolics concentrations ranging from 880 to 2594 mg/L. These values were not only significantly higher than grape white wines (200–500 mg/L), but also similar to grape red wines (1–5 g/L) [7]. This indicated that the mulberry wines may have the potential for long-term aging. As for the tannins, their concentrations in mulberry wines ranged 68 to 216 mg/L, which were comparable to those seen in grape white wines but significantly lower than those seen in red wines [25]. These results indicated the compositional differences of phenolics and tannins between aged mulberry wines and red/white wines. Moreover, there was an observed decline in the overall concentrations of total phenolics and tannins with increasing aging times, which is consistent with findings in aged red wines [25].

The impacts of aging on the pigments and colors of the mulberry wines are shown in Figure 1C–E. Prolonged aging led to a decrease in total anthocyanins but increased the content of non-bleachable pigments and hue values. However, no significant association was observed between color density and aging. The impact of aging on total anthocyanins, non-bleachable pigments, hue, and color density is in agreement with the findings observed from aged commercial red wines provided by Cliff et al. [25]. These results suggest that, during aging, the anthocyanins in mulberry wine became anthocyanin derivates that can resist bleaching by SO₂, exhibit more stable characteristics, and cause a shift in the hue [26].

Impressively, when fitting the aging time with the average results of total anthocyanins and non-bleachable pigments, significant first-order linear correlations were observed (R² ≥ 0.95, Figure S1). Such linear relationships often appear in model wines [27,28], but are rarely observed in real wines. To explain this phenomena, compositions of the non-bleachable pigments were further studied and are discussed in the following section.

3.2. Pyranoanthocyanin and Polymeric Pigments

The compositions of the non-bleachable pigments were characterized using ESI-MS/MS (Figure 2). The predominant anthocyanins in mulberries were cyandin-3-glucoside (C3G), as noted previously (Figure S2) [29]. As a consequence, non-bleachable pigments in aged mulberry wines were found to be, basically, derivatives of C3G. As presented in Figure 2A, the precursor ions m/z 517 were identified as a pyruvic acid adduct of the C3G, and the fragment ions at m/z 355 were a result of neutral loss of the glucose moiety (−162 u) [30]. Similarly, the precursor ions m/z 473, m/z 581 and m/z 565, were identified as acetaldehyde adduct, vinocatechol adduct, and vinophenol adduct of the C-3-G, respectively (Figure 2B–D), since their fragment ions at m/z 311, m/z 419 and m/z 403 were the result of neutral loss of the glucose moieties [30]. In accordance with previous research [9], these pyranoanthocyanins were classified as Vitisin A-type (cyanidin-3-O-glucoside-pyruvic acid), Vitisin B-type (cyanidin-3-O-glucoside-acetaldehyde), and Pinotin-type pyranoanthocyanins (cyanidin-3-O-glucoside-4-vinocatechol and cyanidin-3-O-glucoside-4-vinophenol), respectively.

The polymeric pigments formed by anthocyanins and phenolics via direct linkage were also detected (Figure 2E). The precursor ion at m/z 737 exhibited fragment ions at m/z 575, 423, and 285. Based on previous reports, the fragment ions at 575 were speculated to result from neutral loss of the glucose moiety (−162 u), while m/z 423 and 285 were attributed to either cleavage products from the neutral loss of (epi)catechin (−290 u) or the cleavage product of (epi)catechin moiety following Retro–Diels–Alder (RDA) fission (−138 u) [28]. Other polymeric pigments, formed by the anthocyanins and tannins with higher polymerization degrees (such as dimer, trimer etc.) that are commonly seen in red wines [31], were not detected.

Concentrations of the non-bleachable pigments were analyzed by HPLC-ESI-MS/MS, and the results were presented in Figure 3. First, the aging times were found to be positively correlated with the concentrations of pyranoanthocyanins. As the aging time increased, both the mean and median values of these pyranoanthocyanins, including yanidin-3-O-glucoside-pyruvic acid, cyanidin-3-O-glucoside-acetaldehyde, cyanidin-3-O-glucoside-4-vinocatechol, and cyanidin-3-O-glucoside-4-vinophenol, increased. However, for the polymeric pigments (cyanidin-3-O-glucoside-(epi)catechin), their concentrations were impacted by prolonged aging. Specifically, the mulberry wine aged for 12 years (produced in 2012) showed cyanidin-3-O-glucoside-pyruvic acid, cyanidin-3-O-glucoside-acetaldehyde, cyanidin-3-O-glucoside-4-vinocatechol, and cyanidin-3-O-glucoside-4-vinophenol concentrations as 5.08, 1.92, 7.81, and 1.37 mg/L (on average), respectively. While the concentration of cyanidin-3-O-glucoside-(epi)catechin was 0.50 mg/L, which was significantly lower than the pyranoanthocyanins.

The results from HPLC-ESI-MS/MS indicate that, during the aging process, anthocyanins in mulberry wine become pyranoanthocyanins and polymeric pigments. However, the pyranoanthocyanins were dominant anthocyanin derivates, since the pyranoanthocyanins showed significantly higher concentrations than the polymeric pigments (Figure 4).

This result in contrast to the phenomenon observed in the red wine aging process, which indicated that the pyranoanthocyanins will gradually transfer to polymeric pigments through condensation with tannins [30]. This is because the mulberry wine showed a very high concentration of anthocyanins (1.04–1.37 g/L, Figure 1C), surpassing those reported in the literature for Shiraz (300–400 mg/L), Pinot noir (170–220 mg/L), and Cabernet Sauvignon (400–750 mg/L) [18,32,33]. However, the tannin concentrations were relatively low (167–216 mg/L, Figure 1B). The anthocyanins, along with the metabolic products of yeast, were primarily converted into pyranoanthocyanins without further condensation with tannins. This phenomenon has been observed in model wines (without tannins) [34], but it is the first time it has been observed in real wines.

3.3. The Impact of Aging on Polysaccharides and Proteins

Wine polysaccharides and proteins originate from either the pulp and skin of the fruit or are produced by yeast. These polymers are responsible for the fullness of the wine body and influence the release of aromatic compounds. As shown in

Table 1. Polysaccharides and proteins in aged mulberry wines *.

Polysaccharide Component	2012	2015	2018	2019	2020	2023
Monosaccharides
Mannose	166.7 ± 1.4 d	168.4 ± 5.9 d	332.2 ± 5.3 a	340.9 ± 2.7 a	285.2 ± 4.8 c	306.7 ± 5.4 b
Rhamnose	78.1 ± 1.3 b	84.4 ± 1.0 a	79.4 ± 1.3 b	71.8 ± 1.2 c	86.0 ± 1.8 a	92.0 ± 1.7 a
Glucuronic acid	29.7 ± 1.0 a	17.6 ± 1.4 c	15.5 ± 0.3 c	24.4 ± 1.2 b	14.2 ± 0.3 c	15.9 ± 1.1 c
Galacturonic acid	78.9 ± 1.3 a	18.8 ± 1.0 b	19.2 ± 1.2 c	12.8 ± 0.2 d	14.3 ± 0.8 d	18.3 ± 0.9 c
Glucose	80.9 ± 2.0 b	82.0 ± 1.4 b	61.3 ± 2.4 c	56.1 ± 1.3 c	89.6 ± 2.5 a	62.5 ± 1.2 c
Galactose	157.0 ± 2.0 b	199.8 ± 2.0 a	74.4 ± 2.4 d	73.1 ± 2.1 d	68.7 ± 1.4 c	74.7 ± 2.1 d
Arabinose	64.5 ± 1.2 c	97.1 ± 1.1 a	71.9 ± 2.2 b	63.1 ± 1.1 c	66.9 ± 1.6 c	71.6 ± 0.8 b
Total polysaccharides	656.0 ± 7.6 a	668.2 ± 8.3 a	654.1 ± 6.8 a	642.4 ± 4.5 b	625.1 ± 7.7 c	641.7 ± 7.2 b
Amino acids
Asparagine	3.6 ± 0.3 c	4.5 ± 0.3 bc	6.6 ± 0.2 a	5.1 ± 0.3 b	1.5 ± 0.0 d	3.7 ± 0.0 c
Glutamic acid	9.0 ± 0.3 b	9.9 ± 0.4 b	11.4 ± 0.2 a	5.7 ± 0.4 c	3.9 ± 0.3 d	11.6 ± 0.8 a
Serine	38.4 ± 0.4 a	32.7 ± 2.2 ab	44.7 ± 3.1 a	39.9 ± 2.1 a	19.2 ± 0.4 c	46.2 ± 1.1 a
Histidine	18.6 ± 0.2 b	24.3 ± 1.0 a	13.2 ± 1.4 c	7.8 ± 0.4 d	11.4 ± 1.2 c	13.3 ± 0.8 c
Glycine	13.2 ± 0.1 a	10.5 ± 1.0 bc	11.7 ± 0.2 b	9.9 ± 0.1 c	8.1 ± 0.2 d	13.0 ± 0.5 a
Threonine	25.8 ± 1.0 a	19.5 ± 1.2 b	19.5 ± 1.4 b	20.4 ± 0.5 b	14.1 ± 0.0 c	20.7 ± 1.2 b
Arginine	8.7 ± 0.2 b	9.3 ± 0.2 a	9.9 ± 0.1 a	9.3 ± 0.3 ab	8.4 ± 0.3 b	9.7 ± 0.3 a
Alanine	10.8 ± 0.1 a	6.9 ± 0.3 c	9.0 ± 0.3 b	8.7 ± 0.2 b	7.2 ± 0.3 c	6.2 ± 0.7 c
Tyrosine	5.1 ± 0.0 b	4.2 ± 0.0 d	5.7 ± 0.1 a	4.5 ± 0.0 c	3.3 ± 0.1 e	4.1 ± 0.1 d
Cysteine	1.2 ± 0.1 bc	1.2 ± 0.0 cd	2.1 ± 0.1 a	1.5 ± 0.0 b	1.2 ± 0.0 c	2.0 ± 0.3 a
Valine	3.1 ± 0.1 a	0.1 ± 0.0 c	1.5 ± 0.1 b	1.2 ± 0.2 b	0.1 ± 0.2 c	0.1 ± 0.0 c
Methionine	6.9 ± 0.4 b	4.8 ± 0.1 c	8.4 ± 0.3 a	9.3 ± 0.5 a	6.3 ± 0.0 b	6.5 ± 0.2 b
Isoleucine	5.4 ± 0.3 bc	4.8 ± 0.2 c	6.9 ± 0.2 a	5.7 ± 0.1 b	3.9 ± 0.1 d	6.7 ± 0.3 a
Phenylaniline	12.6 ± 0.3 ab	10.5 ± 0.3 b	13.8 ± 0.1 a	11.1 ± 0.2 b	12.3 ± 0.1 ab	10.3 ± 0.5 b
Leucine	4.5 ± 0.1 b	7.5 ± 0.1 a	7.8 ± 0.2 a	4.8 ± 0.2 b	3.6 ± 0.1 c	4.7 ± 0.2 b
Lysine	21.1 ± 0.3 a	12.6 ± 0.4 d	18.9 ± 0.1 b	21.0 ± 0.2 a	17.1 ± 0.1 c	12.5 ± 0.1 d
Proline	12.3 ± 0.2 a	11.7 ± 0.2 a	12.6 ± 0.2 a	12.3 ± 0.1 a	12.0 ± 0.2 a	11.9 ± 0.2 a
Total amino acids	200.8 ± 4.3 a	175.4 ± 4.1 b	204.2 ± 7.3 a	178.7 ± 7.1 b	134.2 ± 4.3 c	183.2 ± 6.4 b

* Data are expressed as mean of 3 replicates; data were adjusted through box-cox transformation to fix the heterogeneity of variance prior to a one-way ANOVA and Tukey’s post hoc test, where different letters indicate a significant difference (p < 0.05) for a given structural unit across the whole sample set.

, the polysaccharides in mulberry wine were composed of mannose, rhamnose, glucuronic acid, galacturonic acid, glucose, galactose, and arabinose.

The monosaccharide showed a similar composition to the polysaccharides in red and white wine. In addition, the mannose showed the highest relative content, followed by glucose, galactose, and arabinose. This result indicated that the polysaccharides in mulberry wines are mainly mannoproteins, which are derived from yeast metabolism. Furthermore, the total polysaccharide content ranged from 625 to 668 mg/L, which is similar to that of grape red and white wines [35]. The district relation between polysaccharide concentrations and aging time were not observed. The protein amino acid composition of proteins in all mulberry wine samples were relatively high in serine, glycine, threonine, lysine, alanine, and proline, which is similar to the results of previous studies on the protein composition of mulberry wine and red grape wine [35]. The total protein content did not show a significant relationship with aging time, but it is notable that the protein concentration in mulberry wine was 134–204 mg/L, similar to that of grape red and white wines [35].

This phenomenon was because, during the aging process of wine, the proteins gradually polymerize with tannins to form precipitates, while polysaccharides stabilize the wine body. The tannin contents in mulberry wines are significantly lower than in red wine, but the polysaccharide content were similar to those seen in red and white wines. Hence, aging has a relatively minor impact on proteins and polysaccharides, resulting in a more stable wine body.

3.4. Impact of Aging on Volatile Compounds

The volatile compounds in aged mulberry wine were analyzed using HS-SPME-GC-MS, and the results, along with their odor descriptions [23,24,36,37,38,39,40], are shown in

Table 2. The volatile compounds in mulberry wines aged for different years *.

Volatile Compound	Odor Description	Odor Perception Threshold (µg/L)	Produced Year
			2012	2015	2018	2019	2020	2023
Acetate esters
Isoamyl acetate	Banana, fruity, sweet	160	177.8 ± 6.4 d	164.0 ± 6.59 d	194.7 ± 7.3 c	243.5 ± 4.1 b	266.7 ± 7.9 a	269.1 ± 10.2 a
phenethyl acetate	Roses, floral, honey	250	188.9 ± 2.5 c	197.3 ± 2.1 c	212.6 ± 15.0 c	234.6 ± 2.0 b	265.7 ± 1.0 a	282.3 ± 2.5 a
Ethyl acetate	Pineapple, fruity, balsamic	12,000	2370.0 ± 23.5 e	2793.0 ± 17.3 d	3090.0 ± 20.0 c	3460.0 ± 24.8 b	4570.0 ± 42.3 a	5894.6 ± 57.3 a
Hexyl acetate	Green, floral	1500	139.0 ± 8.7 a	115.9 ± 8.3 b	150.3 ± 7.5 a	71.0 ± 5.2 c	0.0	67.5 ± 5.4 c
Ethyl ester
Ethyl lactate	Acid, medicine	150,000	9337.0 ± 85.2 b	7168.0 ± 24.1 d	9005.0 ± 52.2 c	8920.0 ± 25.5 c	9644.0 ± 40.8 a	12,400.3 a
2-methylbutyrate	Fruity, apple aroma	—	22.9 ± 2.3 b	41.1 ± 1.0 b	31.3 ± 12.0 b	31.8 ± 3.1 b	52.4 ± 2.7 a	346.5 ± 11.3 a
Ethyl hexanoate	Fruity, green apple, banana	80	297.4 ± 41.4 b	299.1 ± 37.3 b	285.0 ± 11.6 b	288.0 ± 5.6 b	343.5 ± 7.3 a	1952.7 ± 18.4 a
Diethyl succinate	Cheese, earthy, spicy	1400	786.0 ± 13.3 b	769.0 ± 8.2 b	615.8 ± 33.8 b	690.5 ± 19.7 b	1932.0 ± 26.3 a	1921.9 ± 22.1 a
Ethyl octanoate	Sweet, floral, banana, pear	240	95.2 ± 6.1 b	75.2 ± 4.1 b	53.0 ± 6.1 b	64.4 ± 5.2 b	107.0 ± 3.4 a	65.1 ± 4.7 c
Ethyl dodecanoate	Fruity, apple aroma	1500	11,200 ± 118 b	88,300 ± 170 a	11,100 ± 80 c	64,400 ± 120 c	13,400 ± 200 c	66,523 ± 314 a
Ethyl palmitate	Fruity, sweet, cream	1000	54.8 ± 12.1 b	14.0 ± 2.2 c	33.1 ± 4.1 b	36.4 ± 5.2 b	62.8 ± 3.1 a	18.3 ± 2.6 ab
Ethyl 3-hydroxypropionate	—	—	13.9 ± 4.2 ab	14.5 ± 6.8 a	9.3 ± 4.7 ab	21.3 ± 0.7 a	16.7 ± 1.7 ab	8.2 ± 5.0 a
Ethyl 2-methylbutanoate	Fruity	18,000	6.7 ± 2.4 a	9.8 ± 1.1 a	7.9 ± 5.0 a	10.4 ± 1.7 a	ND	15.6 ± 3.1 a
Ethyl pentanoate	Fruity, apple	5	28.1 ± 8.1 a	17.3 ± 5.7 a	10.4 ± 3.5 a	17.8 ± 7.4 a	ND	ND
Ethyl sorbate	—	—	25.8 ± 8.2 a	13.1 ± 3.2 a	21.3 ± 8.0 a	20.1 ± 2.6 a	ND	ND
Ethyl octadecanoate	—	—	11.8 ± 2.2 b	7.0 ± 2.0 ab	18.3 ± 2.0 ab	20.1 ± 4.9 a	ND	ND
Ethyl propionate	Sweet, fruity, pear-like	2100	13.0 ± 2.1 a	6.1 ± 2.6 a	14.2 ± 8.9 a	7.3 ± 4.8 a	0.0	ND
Diethyl dodecanoate	—-	—	14.9 ± 8.1 a	7.3 ± 0.7 b	19.0 ± 1.7 a	18.4 ± 0.8 a	ND
Higher alcohols
2-Phenylethanol	Flowery, rose, honey	10,000	32,800 ± 199 b	21,200 ± 123 b	55,100 ± 122 a	18,100 ± 180 b	28,000 ± 100 b	35,700 ± 100 b
Nonanol	Fatty, mild, green, melon	6000	332 ± 15.6 bc	1176.7 ± 5.7 a	747.9 ± 1.1 ab	128.5 ± 1.2 c	444.2 ± 1.6 bc	126.7 ± 3.3 c
2-Methylpropanol	Medicinal, wine-like	150,000	92.1 ± 29.6 a	89.0 ± 31.4 a	458.1 ± 103.2 a	463.1 ± 205.6 a	69.6 ± 10.4 a	372.6 ± 10.3 a
Hexanol	Herbaceous, grass, woody	8000	39.5 ± 8.0 a	32.8 ± 1.4 a	67.5 ± 17.8 a	58.8 ± 10.5 a	69.6 ± 4.1 a	71.3 ± 5.1 a
1-Pentanol	Fruity	64,000	63.4 ± 6.1 a	76.3 ± 5.7 a	69.1 ± 9.6 a	74.5 ± 3.4 a	76.7 ± 5.9 a	78.4 ± 2.1 a
2,5-Dimethyl phenylethanol	—	—	4.3 ± 0.1 a	3.6 ± 0.1 a	6.1 ± 1.4 a	11.0 ± 4.7 a	ND	ND
1,10-Decadiol	—	—	10.6 ± 1.9 a	9.0 ± 1.0 a	5.7 ± 0.5 a	7.7 ± 0.1 a	ND	ND
Pentaethylene glycol	—	—	50.1 ± 7.6 a	70.9 ± 46.8 a	6.1 ± 1.4 a	2.9 ± 0.2 a	ND	ND
Pentadiol	—	—	48.3 ± 14.0 a	11.4 ± 0.8 b	32.3 ± 7.5 ab	2.1 ± 0.0 b	ND	ND
2,3-butanediol	Fruity	150,000	34.8 ± 5.5 a	15.8 ± 3.1 b	8.2 ± 4.0 b	16.2 ± 1.6 b	ND	ND
2-propanol	Alcohol-like, ripe fruit	306,000	24.4 ± 11.6 b	19.5 ± 2.1 b	67.4 ± 7.1 a	69.7 ± 15.3 a	ND	ND
Volatile acids
Formic acid	Vinegar, pungent	200,000	17.9 ± 1.1 a	27.1 ± 7.1 a	18.2 ± 0.1 a	24.1 ± 1.6 a	23.0 ± 7.4 a	18.3 ± 4.1 a
2,6-dihydroxy benzoic acid	—	—	11.5 ± 0.1 c	12.6 ± 5.5 c	9.1 ± 1.3 c	19.1 ± 1.2 b	77.8 ± 11.9 a	23.6 ± 4.2 b
Palmitic acid	—	—	23.8 ± 7.6 b	15.6 ± 2.6 b	58.1 ± 4.7 ab	45.5 ± 12.4 b	98.2 ± 10.9 a	43.2 ± 8.4 b
Octadecanoic acid	—	—	8.2 ± 1.9 a	7.6 ± 3.3 a	22.9 ± 4.4 a	21.5 ± 8.5 a	53.0 ± 7.1 a	26.7 ± 4.1 a

* Data are expressed as mean of 3 replicates; data were adjusted through box-cox transformation to fix the heterogeneity of variance prior to a one-way ANOVA and Tukey’s post hoc test, where different letters indicate a significant difference (p < 0.05) for a given structural unit across the whole sample set, ND means not detected.

. In summary, 35 volatile compounds were identified in mulberry wine, categorized into: acetate esters, ethyl esters, higher alcohols, and organic acids.

Acetate esters, especially isoamyl acetate, phenethyl acetate, and ethyl acetate, are typically formed during fermentation because of the interaction between ethanol and acetic acid, facilitated by yeast enzymes [24,38]. The presence of ethyl acetate contributes fruity and sweet characteristics to the overall aroma profile of the wine [36]. The formation of ethyl lactate, ethyl hexanoate, and diethyl succinate involves esterification with various organic acids (lactic acid, hexanoic acid, and succinic acid) and ethanol, contributing to the fruity and floral aromas of the wine. The higher alcohols, comprising both aliphatic and aromatic types, constituted the primary group of volatile compounds in mulberry wine, in agreement with previously reported findings [23]. Notably, the alcohols with the highest concentrations were 2-phenylethanol, nonanol, 2-methylpropanol, 1-pentanol, hexanol, and 1-pentanol. These alcohols predominantly arise during the fermentation process, and are characterized by their intense and robust aroma and taste, often associated with floral, fruity, and herbal notes [23].

Based on the ratio of volatile compound concentrations to their odor perception threshold, the odor activity values (OAV) were calculated [23]. The results are presented in Table S3. As indicated, while many of the 35 compounds detected in mulberry wine exhibited decreasing concentrations with extended aging, compounds with high OAV values (OAV ≥ 2), such as ethyl hexanoate, ethyl decanoate, ethyl dodecanoate, and ethyl pentanoate, did not show a notable decline with aging. Importantly, these compounds primarily exhibited fruity and sweet characteristics.

Referencing previously published studies [23,24,36,37,38,39,40], the volatile compounds were classified into fruity, balsamic, solvent, floral, herbaceous, sweet, fatty, green, and spicy odorant series. Then, the total OAVs of each series were obtained by combining the OAV values of the corresponding compounds and presented them as odor profiles (Figure 5). Among the series, the fruity, sweet and floral odor characters made the highest contributions to the aged mulberry wines, which was similar to the observation made by Butkhup et al. [23]. Furthermore, the mulberry wines produced in 2015, 2019, and 2023 showed higher fruity characters than the other samples, while the mulberry wine produced in 2018 and 2023 showed the highest sweet character.

4. Conclusions

In newly made mulberry wine, substantial amounts of anthocyanins (1.04–1.37 g/L) were present, whereas the reaction substrates for polymeric pigments and tannins were limited (only 167–216 mg/L). Consequently, during prolonged aging, cyanidin-3-O-glucoside primarily transformed into anthocyanidin derivatives, including Vitisin A-type (cyanidin-3-O-glucoside-pyruvic acid), Vitisin B-type (cyanidin-3-O-glucoside-acetaldehyde), and Pinotin-type pyranoanthocyanins (cyanidin-3-O-glucoside-4-vinocatechol and cyanidin-3-O-glucoside-4-vinophenol). A small portion of anthocyanins transformed into polymeric pigments, cyanidin-3-O-glucoside-(epi)catechin. This illustrated a significant difference in the transformation of anthocyanins in grape wine. Additionally, no significant decrease in polysaccharides and proteins induced by tannins with aging time was observed. Some aroma compounds gradually decreased with extended aging, but compounds with sweet and floral odor characters (OAV ≥ 1) did not show a significant relationship with aging. The impact of active components in elderberry wine of different vintages on its sensory characteristics will be reported in our subsequent studies.