Effects of prefermentative cold soak on polyphenols and volatiles of Aglianico, Primitivo and Nero di Troia red wines

Abstract Effectiveness of prefermentative cold soak (PCS) on polyphenols and volatiles extraction during winemaking of three red grape cultivars grown in southern Italy (Aglianico, Primitivo and Nero di Troia) was investigated. Four 200‐L stainless steel horizontal rotary wine fermenters were used. The main result was that PCS improved the extraction of polyphenols and increased the antioxidant activity of wines. Extraction of proanthocyanidins was enhanced (+25%, +14% and +7% for Aglianico, Primitivo and Nero di Troia, respectively) and, consequently, the ratio flavans reactive with vanillin/proanthocyanidin was reduced, thus potentially favoring the chromatic and tannic stabilization of wines. As regards volatiles, PCS increased ester compounds at levels above their odor thresholds, potentially conferring fruity odor to wines. In conclusion, PCS could be favorably introduced in winemaking to enhance the enological potential of Aglianico, Primitivo and Nero di Troia wines.

A common task of winemakers is to increase the extraction of volatile compounds, anthocyanins, and low molecular weight tannins, and limiting the extraction of seed tannins that are tough and bitter.
Aglianico, Nero di Troia, and Primitivo are nonaromatic red grape vines (Vitis vinifera L.) widely grown in southern Italy, from which wines with different phenolic and aromatic characteristics can be produced. The present work aimed to characterize the polyphenol and volatile fractions of wines obtained from these cultivars by application of PCS, and to assess whether this technique allows to better express their enological potential.

| Grapes and wine samples
The  • trial PCS-prefermentative cold soak: preliminary cooling of destemmed grapes at 5°C using cooling jacket, maintenance of the sample at 5°C for 48 hr, then 5 days of maceration at 25°C (7 days of maceration in total, as C).
When maceration was concluded, free-run wine was recovered and pomace was gently pressed for obtaining press-run wine using 43 L traditional cage staves press. The two wine fractions were blended and after 1-week racking was done to eliminate gross lees. Wine was bottled after 6 months, without any treatments, and analyzed.

| Chemical analyses
Representative grape samples were taken as reported in a previous paper (Gambacorta et al., 2017), and the corresponding juices were subjected to analysis of total soluble solids (TSS), pH, and titratable acidity (TA) according to EEC 2676 standard procedure (EEC, 1990).

| Polyphenols analysis
Extraction of polyphenols from grape was carried out according to Gambacorta et al. (2017). The phenolic composition of grape extracts and wines was determined by spectrophotometry as described by Di Stefano and Cravero (1991), whereas the color indices were evaluated according to the method of Glories (1984). Anthocyanin composition was determined by HPLC-DAD according to Coletta et al. (2013), and results were expressed as mg/L of malvidin-3-Oglucoside equivalent. Antioxidant activity (AA) was assessed using ABTS [2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)] assay as reported by Trani, Verrastro, Punzi, Faccia, and Gambacorta (2016), and results were expressed as μM Trolox equivalent antioxidant capacity (TEAC).

| Volatile analysis
Volatiles were extracted by static headspace solid phase micro extraction (HS-SPME) using a preconditioned fiber, 2 cm long 50/30 μm (ThermoFisher Scientific). The detailed operative conditions have been reported in a previous paper (Trani et al., 2016). Volatiles were tentatively identified by comparing them with the linear retention index of pure standard compounds and by comparing the experimental mass spectra with those reported in the NIST Library. The identification of compound was accepted with probability higher than 80%. Volatiles were quantified using relative areas related the 2-heptanone as internal standard.

| Statistical analysis
All measurements were carried out in triplicate, and results were expressed as means ± SD (standard deviation). Statistical analysis was performed using IBM SPSS software v 19. Significant differences between control and cold macerated wines for each cultivar were determined using one-way ANOVA and Tukey's HSD test for multiple comparisons (IBM SPSS software v 1.9). Principal component analysis (PCA) was applied to phenolic and volatile fractions in order to evaluate differences among samples using XlStat (Addinsoft, NY) software. Notes. A-C, Aglianico control; A-PCS, Aglianico prefermentative cold soak; P-C, Primitivo control; P-PCS, Primitivo prefermentative cold soak; NT-C, Nero di Troia control; NT-PCS, Nero di Troia prefermentative cold soak. TP, total polyphenols: as gallic acid; FRV, flavans reactive with vanillin: as (+)-catechin; P, proanthocyanidins: as cyanidin chloride; AA, antioxidant activity; CI, color intensity; T, tonality. In rows, data followed by different letters indicate statistically significant differences at p < 0.05.

| Qualitative characteristics of grapes
TA B L E 4 Anthocyanin composition of wines (mg/L, mean values ± SD)

| Chemical characteristics of wines
The chemical characteristics of wines are reported in Table 2 TA B L E 5 Mean concentration and relative standard deviations (n = 6) of free volatile compounds (mg/L) of wines

| Phenolic characteristics of wines
Phenolic composition, AA, and color indices of wines are reported in Table 3. TP and proanthocyanidins (P) showed significant increase (p < 0.05) in cold soak samples. In particular, the increment of TP was 5%, 18% and 14% with respect to the controls for Aglianico, Primitivo and Nero di Troia, respectively, whereas the increase in proanthocyanidins was 25%, 14%, and 7%. As a consequence, also AA was higher in the cold soak samples than in the relative controls (+13%, +15% and +6%). The higher values of P in cold soak wines led to reduction in flavans reagent with vanillin/proanthocyanidins ratio (FRV/P). This indicates that PCS applied to the investigated cultivars could favor the chromatic and tannic stabilization of wines (Suriano, Alba, Tarricone, & Di Gennaro, 2015). With regard to the color indices, color intensity (CI) significantly increased in Primitivo and Nero di Troia (+13% and +9%, respectively), whereas it did not change in Aglianico that was the most colored wine. The hue, or color tonality T, showed significant differences only in  Tominaga, Murat, and Dubourdieu (1998).
TA B L E 6 Linear retention index, odor threshold and odor description of volatile compounds of wines Nero di Troia samples. This parameter is computed by dividing the absorbance of wine at 420 nm with the absorbance measured at 520 nm. It increases during aging of red wine due to the reduction in anthocyanin forms (responsible for the absorbance at 520 nm) and the concomitant increase in brown pigments (accountable for the absorbance at 420 nm). The lower tonality value found in the Nero di Troia PCS sample suggests that PCS could exert an influence on the polymeric pigments production between tannins and anthocyanins. These latter seem to be preserved from precipitation, favoring color stabilization of wine. The fact that this phenomenon was observed only in Nero di Troia could be explained by the differences found in the anthocyanin composition (Table 4) Table 5 shows the volatile compounds grouped into chemical classes (alcohols, esters, acids, and aldehydes). The linear retention indices, odor threshold, and odor description for each compound are reported in Table 6. The SPME/GC-MS method allowed identification and quantification of 23 volatiles. Compounds present at a trace level (<0.005 mg/L) or identified with lower than 80% of probability have not been considered. All the identified volatile compounds derived from yeasts activity during fermentation, and thus they belong to the class of secondary aromas. It is well known that their concentration is influenced by the yeast strain, sugar content, fermentation temperature, and degree of aeration.

| Volatile characteristics of wines
The total content of volatile compounds ranged from 8.48 mg/L in NT-C to 10.98 in P-PCS. These values are about one-half of those found in Primitivo wine in the 2013 season, and produced by using a different yeast strain (Trani et al., 2016). In comparison with control, cold soak increased the total volatiles content in all wines (+21.3%, +17.2% and +16.1% in Nero di Troia, Aglianico and Primitivo, respectively). The alcohols were the most abundant compounds in all samples, ranging from 54.2% in A-PCS to 65.3% in NT-C, followed by esters, ranging from 34.0% in NT-C to 37.0% in NT-PCS, and by acids and aldehydes with percentage lower than 0.1%. Cold soak caused slight increase in alcohols by 11.1% in Aglianico and Nero di Troia, and 6.0% in Primitivo. Alcohols are recognizable by their strong and pungent smell and taste (Kotseridis & Baumes, 2000). Among them, 1-butanol-3-methyl (isoamyl alcohol), which is responsible for alcohol and fused notes, showed the highest increase. However, the concentrations found were always below the odor threshold (30 mg/L), and it should not contribute

| PCA analysis
The application of PCA to phenol and volatile fractions is reported in

| CON CLUS IONS
The results of this study demonstrated that PCS applied to Aglianico, Primitivo, and Nero di Troia grapes improves the extraction of proanthocyanidins, decreasing the FRV/P ratio. As a consequence, it may be recommended for these cultivars, in order to favor wine color stabilization. This enological practice positively affected the extraction of free anthocyanins in Aglianico and Primitivo, suggesting that the results obtained were cultivardependent. Concerning the volatile compounds, the treatment increased extraction of esters that potentially contribute to the enrichment of fruity odors of wines.

ACK N OWLED G EM ENT
The authors gratefully acknowledge Industrie Fracchiolla for providing the "Gioiello" pilot plants used in this research, and Giuseppe

E TH I C A L S TATE M E NT A N D CO N FLI C T O F I NTE R E S T
I testify on behalf of all coauthors that the research in this article did not involve human experimentation and/or animal testing. All authors have read, approved and are fully conversant with the manuscript; all authors were also aware of its submission to Food Science & Nutrition and declared no conflict of interest.