Tracing the volatilomic fingerprint of grape pomace as a powerful approach for its valorization

The huge amount of grape pomace (GP) generated every year worldwide, particularly in Europe, creates negative impacts at the economic and environmental levels. As far as we know, scarce research has been done on the volatilomic fingerprint of GP. To meet consumer demand for healthy foods, there is a growing interest in the characterization of particular volatile organic metabolites (VOMS) in GP that can be used for industrial applications, including the food industry. In this study, the volatilomic fingerprint of GP obtained from different Vitis vinifera L. grapes was established by solid phase microextraction (HS-SPME) combined to gas chromatography-mass spectrometry (GC-MS), to explore the properties of most dominant VOMs in a context of its application on marketable products. A total of 52 VOMs belonging to different chemical families were identified. Alcohols, carbonyl compounds, and esters, are the most dominant, representing 38.8, 29.3, and 24.2% of the total volatile profile of the investigated GP, respectively. Esters (e.g., isoamyl acetate, hexyl acetate, ethyl hexanoate) and alcohols (e.g., 3-methyl butan-2-ol, hexan-1-ol) can be used as flavoring agents with potential use in the food industry, and in the cosmetic industry, for fragrances production. In addition, the identified terpenoids (e.g., menthol, ylangene, limonene) exhibit antioxidant, antimicrobial, and anticancer, biological properties, among others, boosting their potential application in the pharmaceutical industry. The obtained results revealed the potential of some VOMs from GP to replace synthetic antioxidants, colorants, and antimicrobials used in the food industry, and in the cosmetic and pharmaceutical industry, meeting the increasing consumer demand for natural alternative compounds.


Introduction
Being one of the most important agricultural activities throughout the world, the winemaking processes generate huge quantities of underutilized by-products and waste, namely GP generated by pressing whole grape bunches during the production of must, constituted of stalks, seeds, skins, and pulp, which account for 25-30% of produced grapes (Ferreira and Santos, 2022), from which grape skins are the major components contributing to half of the GP.This is especially important for Europe since about 65% of the world's wine production is managed by European winegrowers (Commission E., 2013) mainly from France, Italy, Spain, Germany, and Portugal.The improper disposal practices for winery wastes cause serious impacts on the environment, severely damaging the eco-and aqua-systems and, consequently, human and animal health (Câmara et al., 2020;Peixoto et al., 2018).However, GP is characterized by the presence of several high-value macromolecular components such as polysaccharides, proteins, and lipophilic compounds.In addition, the presence of high levels of bioactive compounds, such as phenolic compounds, anthocyanins, tannins, carotenoids, and dietary fibers (Fig. 1), further enhances its potential for industrial (food, pharmaceutic and cosmetic) applications.These compounds exhibit remarkably antimicrobial, anti-inflammatory, anticancer, antifungal, and antioxidant properties, establishing them as a valuable and sustainable bioresource (Ferreira and Santos, 2022;Ilyas et al., 2021;Peixoto et al., 2018) This approach will improve both the economic and environmental sustainability of the winemaking sector regarding its valorization based on the circular bioeconomy concept, constituting a powerful platform for shaping the future roadmap in the field of wine industry sustainability (Antonić et al., 2020;Bordiga et al., 2019;Jimenez-Lopez et al., 2020).
Nonetheless, most of the research on GP valorization has been conducted on phenolic compounds with the purpose of minimizing the waste's environmental burden and recovering extremely added-valued natural compounds.In this sense, Apostolou et al. (2013) verified that low dosages of proanthocyanidins from grape seeds suppress the proliferation of liver cervical (HeLa) and (HepG2) cells using in vitro settings.While Hamza et al. (2018) studied the anti-cancer property of grape seed extract and the data obtained demonstrated that this wine by-product has an anticancer impact by modifying oxidative damage, inducing apoptosis, inhibiting inflammatory response, and inhibiting cell growth.The anti-platelet activity of GP extracts, which plays a crucial role in atherothrombosis was evaluated by Choleva et al. (2019).Despite the significant and varied health benefits, the geometry and large molecular size of phenolic compounds are crucial for their bioavailability and bioaccessibility, making them more troublesome to use as therapeutic agents.In this sense, some VOMs such as terpenoids offer a better alternative for being low-molecular and lipophilic compounds that can penetrate the biological barriers with greater ease (Mahizan et al., 2019).
Nowadays, several headspace extraction processes, including dynamic headspace (DHS), solid phase microextraction (SPME), stir-bar sorptive extraction (SBSE), and gas chromatography (GC) combined with different detectors, have been employed to determine the volatile fingerprint of agri-food by-product (Câmara et al., 2020;Deng et al., 2021).From these, SPME is the most widely used headspace analysis for the extraction of VOMs because it offers a fully automated approach at a lower cost than other headspace techniques such as DHS and SBSE.In addition, SPME compared to traditional extraction procedures (liquid-liquid extraction, solid phase extraction) showed several advantages, like solvent-free sample preparation, a short extraction time, a small amount of sample required, and lower labor (Deng et al., 2021;Song et al., 2019).Generally, the SPME extraction procedure is coupled to gas chromatography with a sensitive mass spectrometry detector (GC-MS) and is highly used for the identification and quantification of several chemical families of VOMs in food-related products (Câmara et al., 2020;Song et al., 2019).In addition, GC-MS provides the identification of VOMs with a low limit of detection and the potential to separate complex mixtures.
As far as we know, scarce research has been done on the volatile fingerprint of GP.To meet consumer demand for healthy foods, there is a growing interest in the characterization of particular VOMs in GP that may be used in the food industry as antioxidant, antimicrobial, and antiinflammatory active agents for the formulation of novel functional foods.In this context, the current study focused on the establishment of the volatile fingerprint of GP obtained from different Vitis vinifera L. grape varieties using HS-SPME followed by gas chromatography-mass spectrometry (GC-MS).To identify valuable VOMs that could be used in the food industry for the development of novel food ingredients and/ or products for human consumption, pharmaceutical and cosmetic formulations, namely as antioxidant, antimicrobial, and anti-inflammatory agents.The discrepancies in the volatile signatures of the GP from different grape varieties under study were established using statistical methods.

Samples
Grapes were harvested upon ripening in the 2022 vintage and the GP was kindly provided by Justine Company (coordinates 32 • 39′ 04″ North latitude and 16 • 51′ 45″ West longitude) as a by-product of winemaking processes.The current research includes 3 red and 5 white V. vinifera L. grape varieties.The white grape varieties include Boal, Malvasia, Terrantez, Verdelho, and Sercial, while the red grape varieties include Malvasia Roxa (MR), Tinta Negra (TN), and Complexa.The GP from the different grape varieties was collected immediately after grapes pressing and transported in cool boxes under refrigeration (ca.2-5 • C) to the laboratory.Afterward, the GP were lyophilized (Telstar, Cryodos, Spain) during 2 h, milled using a laboratory grinder (Grindomix GM200, Rech, Germany) to obtain a fine and homogeneous powder, and finally stored at − 80 • C until analysis.

Determination of total phenolics, total flavonoids, and antioxidant activity of GP extracts
To determine the total phenolic content (TPC), total flavonoids content (TFC) and antioxidant activity of grape pomace under study ethanolic extract were prepared.Briefly, 0.5 g of lyophilized grape pomace were placed in a vial with 3 mL of ethanolic solution 50% v/v.The mixture was then centrifuged for 5 min at 5000 rpm (centrifuge: SIGMA 1-7, max capacity 6 × 15 mL, max.RCF 6153×g) after being subjected to ultrasonic agitation for 5 min in an ultrasonic bath.After that, the supernatant was collected and submitted to spectrometric assays.
The TPC was determined spectrophotometrically using the Folin-Ciocalteu assay, while the TFC was estimated by the AlCl 3 colorimetric assay according to Perestrelo et al. (2018).The results were expressed as milligrams of gallic acid equivalent per liter of sample [mg (GAE)/L] since gallic acid was utilized as a reference standard to plot the calibration curve.The spectrophotometric measurements were done with a UV-Vis spectrophotometer (Lambda 25, PerkinElmer, Waltham, MA, USA).
In vitro assays were carried out to evaluate the antioxidant activity of the free DPPH • radical-scavenging capacity (A AR ) and by the ABTS discoloring assay according to Perestrelo et al. (2018), with slight modifications.The DPPH method is based on the scavenging of DPPH • by the addition of antioxidants able to discolor the DPPH solution which is proportional to the concentration of antioxidant type of molecules.For the DPPH • radical assay, a stock solution of DPPH • radical was prepared in ethanol (400 μM) and kept in the dark at 25 • C. For the reaction, the DPPH • solution was diluted to obtain a working solution with an absorbance of 0.900(±0.030)at 515 nm.Trolox, used as a reference antioxidant, was assembled at concentrations ranging from 5 to 600 mg/L.The percentage of DPPH radical inhibition is calculated considering the absorbance before and after the reaction according to the following equation: %DPPH inhibition = [(A 0 -A 1 /A 1 ) × 100], where A 0 is the absorbance before the reaction, and A 1 the absorbance after the reaction.
For the ABTS assay, a stock solution of ABTS (20 mM) was prepared in 50 mL of phosphate-buffered saline (PBS, pH 7.4), and 200 μL of 70 mM K 2 S 2 O 8 solution (oxidant agent) was added.The solution was incubated in the dark at 25 ± 1 • C for 16 h.Then, diluted with PBS to obtain a working solution with an absorbance of 0.800 (±0.030) at nm. After, an aliquot (12 μL) of each sample extract was mixed with mL of ABTS •+ solution in 10 mm quartz cells and the decrease of absorption was measured at 734 nm for 30 min.The activity was expressed in terms of Trolox equivalents antioxidant capacity (TRE), for the analyzed extracts, plotting: ln(%ΔA 734 ) vs Trolox concentration (from to 600 mg/L).All determinations were carried out in triplicate.

HS-SPME procedure
The SPME method employed in this study was adapted from earlier research projects conducted in our laboratory (Câmara et al., 2020).
Briefly, 2 g of each lyophilized GP, 0.5 g NaCl (salting-out effect), 5 μL of 3-octanol (IS, 250 μg/L), and 5 mL of H 2 O were added in a 20 mL glass vial containing a stirring bar (2 × 0.5 mm, 600 rpm).After that, the glass vial was capped with PTFE-coated silicone septum and placed in a thermostatic block at 40 • C and the DVB/CAR/PDMS fiber was inserted into the headspace sample vial for 45 min.After the VOMs extraction, the fiber was withdrawn into the needle holder, removed from the vial, and placed in the GC injector at 250 • C for 6 min for thermal desorption of VOMs.All analyses were performed in triplicate (n = 3).Before practice, the SPME fiber was thermally conditioned following the manufacturer's recommendations.Each day, conditioning was performed for 10 min prior to the first extraction to verify the lack of carryover analytes.Moreover, a blank was carried out between samples to confirm the lack of carryover analytes.

GC-MS analysis conditions
Chromatographic separation of the VOMs was done using an Agilent 6890 N (Palo Alto, CA, USA) gas chromatography system equipped with BP-20 fused silica (30 m × 0.25 mm inner diameter × 0.25 μm thick) capillary column purchased from SGE (Darmstadt, Germany).Helium was used as carrier gas with a flow rate of 1 mL/min.The temperature of the injector, GC-qMS interface, and quadrupole detector were 250, 180, and 220 • C, respectively.The GC oven temperature started at 45 • C (1 min), increased to 100 • C at 2 • C/min (3 min), followed by an increase to 130 • C at 5 • C/min (5 min) reaching the final temperature of 220 • C at 7 • C/min (15 min), total run time of 70.35 min.The MS used was an Agilent 5975 quadrupole inert mass selective detector, which operated on electron impact (EI) at 70 eV, maintaining the source temperature at 180 • C. The mass range acquired was from 30 to 300 m/z, at the rate of 1.9 spectra/s, in full scan mode.The VOMs were identified by comparison of the GC retention times (RT), kovat index (KI) and mass spectra with those of the standard, when available; and by comparison of their mass spectra with those in the National Institute of Standards and Technology (NIST) MS 05 spectral database (Gaithersburg, MD, USA) with a matching probability >85%.The van den Dool and Kratz equation was used to determine the KI values, which were then compared to values reported in the literature for equivalent columns (El-Sayed, 2011).
The relative-quantification of VOMs was done using 3-octanol (IS) addition based on the following equation: VOMs concentration = (VOM GC peak area/IS GC peak area) × IS concentration.

Data treatment and statistical analysis
A chemical network containing all VOMs (except the ones common to all samples) was constructed by Gephi (version 0.9.7), a networkbuilding tool.Excel was used to transform and import the data to Gephi.This data is comprised of two parts, the nodes, and edges.The nodes correspond to the samples and VOMs, while the edges establish a connection between the nodes (Adam et al., 2021).The antioxidant test values were normalized in an interval between 30 and 150 to better T. Abreu et al. visualize the sample differences in Gephi according to Polatgil et al. (Polatgil, 2022).This interval was attained by trial and error.TPC and TFC were normalized together (sharing the same max(x) and min(x)), while ABTS and DPPH values were normalized independently since they have a different order of magnitude.
The statistical analysis was carried out using the MetaboAnalyst 5.0 web-based tool (Pang et al., 2021).The raw GC-MS data were pre-processed to eliminate VOMs with miss values and then normalized by Log transformation (base 10) and auto-scaling (mean-centered and divided by the standard deviation of each variable).The dataset was subjected to a one-way analysis of variance (ANOVA) followed by Fisher's test for post-hoc multiple comparisons of means of GP from white and red grape varieties data at p value < 0.05 to recognize statistically significant.Principal component analysis (PCA) was used to reduce the dimensionality of the dataset and to obtain a representation from the data projection.Partial least squares-discriminant analysis (PLS-DA), a supervised method based on PLS regression, was applied to the GP volatile fingerprint dataset to provide insights into the separations between GP from white and red grape varieties and to identify VOMs that contribute to differentiate among sample sets, therefore useful to discriminate variable selection.

Assessment of the antioxidant potential of grape pomace
The antioxidant potential of the GP from studied grape varieties was evaluated through spectrophotometric assays (TPC, TFC, DPPH, ABTS).Due to their redox characteristics, phenolic compounds are known to have strong antioxidant activity and play significant roles in absorbing and neutralizing free radicals (Perestrelo et al., 2018).Fig. 2 displays the TPC and TFC patterns of GP from white (Boal, Malvasia, Terrantez, Verdelho, and Sercial) and red (Tinta Negra, Malvasia Roxa, Complexa) V. vinifera L. grape varieties.
The antioxidant potential of the GP under investigation was assessed spectrophotometrically through DPPH and ABTS assays.The results are shown in Fig. 3.The GP from Tinta Negra (524 ± 16.4 and 6446 ± 31.6 mg/L for ABTS and DPPH assays, respectively) and Complexa (516 ± 24.3 and 5695 ± 18.8 mg/L) showed antioxidant activity significantly higher than the obtained for the other grape varieties, while Malvasia Roxa (42.1 ± 4.84 and 708 ± 10.6 mg/L) is the grape variety with the lowest antioxidant capacity, considering the GP from grape varieties under study.In addition, no significant difference (p < 0.05) was observed in the antioxidant capacity of GP from Sercial (256.3 ± 10.5 and 3665 ± 34.8 mg/L for ABTS and DPPH assays, respectively) and Verdelho (257.4 ± 13.2 and 3307 ± 31.7 mg/L).
In sum, the antioxidant potential is highest in GP from red grapes compared to white grape varieties, except for Malvasia Roxa.Although Malvasia Roxa is a red V. Vinifera L. grape variety, its antioxidant capacity is close to that determined in white V. Vinifera L. grape varieties.

Volatile fingerprint of grape pomace
To valorize the GP from red and white grape varieties, a valuable source of bioactive compounds with several biological properties (e.g., antioxidant, antimicrobial, antiproliferative), a volatilomic-based analytical approach was carried out.The goal was to establish the volatile fingerprint of GP from different V. vinifera L. grapes, aiming to explore and identify some VOMs with potential industrial applications in the fields of pharmaceuticals, cosmetics, and food.A total of 52 VOMs were identified, from which 18 were common to all analyzed samples.The identified VOMs belong to different chemical families including alcohols ( 14), esters ( 14), carbonyl compounds (10), terpenoids (11), acids (2), and hydrocarbon (1) (Table 1).
Carbonyl compounds (ketones and aldehydes) are the second most abundant chemical family in the analyzed GP.Their contribution to the total volatile fingerprint of GP from the different grape varieties is significantly different.On average, the highest GC total peak area was determined in Malvasia Roxa (51.9%), followed by Verdelho (47.0%),Complexa (40.1%),Boal (36.5%),Sercial (35.2%),Malvasia (15.4%),Tinta Negra (6.25%) and Terrantez (2.16%).Hexanal and 2-hexenal are the most dominant aldehydes in the GP understudy, accounting for 82.9 and 87.9% of the total fraction of carbonyl compounds in GP from red and white grapes varieties under investigation.
The contribution of esters to the total volatilomic fingerprint ranges from 8.05 (Malvasia Roxa) to 69.7% (Terrantez).The highest esters contribution for GP of Terrantez (69.7%) and Tinta Negra (43.1%) grape varieties is explained by the high relative concentration of ethyl acetate, 21.1 ± 0.20 and 47.3 ± 5.65 μg/L, respectively.Nevertheless, the esters' contributions in GP of Malvasia Roxa (8.046%) and Complexa (8.6%) are quite similar.The esters contribute positively with aromas and flavors common to fruits, such as bananas, strawberries, pineapple, raspberries, cherries, and even citrusy and floral odors.In general, esters have low odor thresholds (trace μg/L) depending on the matrix (Perestrelo et al., 2020).The presence of esters in GP is scarce since the vast majority of esters that occur in wine are produced during fermentation.Isoamyl acetate, hexyl acetate, ethyl hexanoate, and ethyl octanoate are, on average, the most abundant esters identified in GP from investigated grape varieties.It should be noted that the GC peak area of isoamyl acetate and hexyl acetate are, on average, 1.87 and 2.66 times higher in GP from white grapes than from red varieties, while ethyl hexanoate and ethyl octanoate are, on average, 1.20 and 4.88 times higher in GP from red grapes than white grapes.

Bioactive potential of volatile organic metabolites identified in grape pomace and potential applications
Several VOMs identified in GP from red and white grape V. vinifera L. varieties under study have been previously reported in food and foodrelated products with a remarkable number of biological properties, namely as antioxidant, antimicrobial, anti-inflammatory, antiproliferative, antidiabetic, antitumor activities, boosting their applications in different fields including in the food industry, as preservatives, additives, and flavoring agents, in the pharmaceutical industry as antiinflammatory antiproliferative, and antidiabetic agents, and also in the cosmetic industry as flavor and coloring agents (Alam et al., 2019;Assmann et al., 2018;Dȃnilȃ et al., 2018;Elkady and Ayoub, 2018).
As previously discussed, alcohols and esters are the chemical families that most contribute to the total volatile profile of grape pomace, independently of grape varieties.Esters (e.g., isoamyl acetate, hexyl acetate, ethyl hexanoate) and alcohols (e.g., benzyl alcohol, 2-phenyl ethanol) are commonly used as flavoring agents in various food products such as candies, baked goods, and beverages; and as ingredients in perfumes, soaps, and some others personal care products (De Melo Pereira et al., 2019).Moreover, isoamyl acetate, 3-methyl butan-2-ol, benzyl alcohol, and 1-octen-3-ol have also been shown to have antimicrobial activity against several microorganisms (Ando et al., 2015;Xiong et al., 2017).In addition, Inamdar et al. (2014) explored the signaling pathways involved in 1-octen-3-ol-mediated dopamine neurotoxicity, and the obtained data showed that 1-octen-3-ol exposure was associated with activation of caspase 3, a lysosomal enzyme (endoprotease) involved in the apoptotic pathway.In another study, 2-ethyl hexan-1-ol and 2-pheny ethanol have been reported in fish feeds (Giogios et al., 2009), whereas a mixture of 3-hexenol and 2-hexenal induce light stress-alleviating effects on dog (Carlone et al., 2018).Regarding carbonyl compounds, special attention has been devoted to hexanal and benzaldehyde, which are powerful antimicrobial constituents because of their capacity to interact with the sulfhydryl and amino groups of protein moieties in the microbial cytoplasmic membrane, causing the membrane transport function to be disrupted and ultimately leading to cell death.In addition, these additives are classified as Fig. 3. DPPH scavenging and ABTS assays of grape pomace from different V. vinifera L. grapes.Different superscript letters suggest significant differences (p < 0.05) among GP under study.

Table 1
Relative concentration (μg/L) and relative standard deviation (in brackets) of volatile organic metabolites identified in grape pomace.generally recognized as safe (GRAS) status, being used to add flavor to foods, and a potent inhibitor of phospholipase D activity (Jash et al., 2018;Ranjan et al., 2020).Hexanal application produced positive results by extending the shelf life of a diversity of fruits (Anusuya et al., 2016;Debysingh et al., 2018;Ranjan et al., 2020).On the other hand, terpenoids are present in the GP of some studied grapes but in lesser amounts.However, their total GC peak areas are higher in GP from white grapes than in red grapes.Menthol and ylangene are the most dominant terpenoids.Some studies have reported that menthol possesses anti-inflammatory, antibacterial, antitumor, antiviral, scolicidal, immunomodulatory, neuroprotective, antifatigue, and antioxidant activities (Zhao et al., 2022), while ylangene is related to cytotoxic and anti-inflammatory properties (Xio et al., 2013).In addition, terpene-mediated actions have been found to be protective against tuberculosis by Sieniawska et al. (2018).Nevertheless, the most noteworthy bioactivity of terpenoids is related to their anticancer potential, acting at different stages of tumor development and in different mechanisms of action (inhibition, regulation of intracellular signaling pathways) (Ansari and Akhtar, 2019;Cho et al., 2017).Numerous studies have proven limonene's chemopreventive and chemotherapeutic effects against several human malignancies (Chebet et al., 2021;Zhou et al., 2021).

Conclusions
The HS-SPME/GC-MS analytical methodology was used in the current research with the purpose to identify VOMs in GP from different V. vinifera L. grape varieties as a useful strategy to investigate the potential properties of the dominant VOMs to suggest possible industrial applications of this wine by-product to boost its use in a context of circular bioeconomy.52 VOMs belonging to different chemical families were identified, namely alcohols (14), esters ( 14), carbonyl compounds (10), terpenoids (11), acids (2), and hydrocarbon (1), 18 of which were common to all GP analyzed.The contribution of esters and terpenoids to the total volatile profile is higher in GP from white grapes, whereas the  1).
Fig. (a) and (b) displayed the PCA score plot and loading plot from the GP respectively.

Fig. 4 .
Fig. 4. The total peak area of chemical families identified in GP of investigated grape varieties.MR: Malvasia Roxa, CC: Carbonyl compounds.

Fig. 6 .
Fig. 6.PCA of the volatile fingerprint of GP.(a) PC1 × PC2 score scatter plot and (b) loading weight plot (attribution of the peak number is pointed out in Table1).

Table 1
(continued ) a Kovats index relative n-alkanes (C8 to C20) on a BP-20 capillary column.bKovatsindexrelativereported in literature for equivalent capillary column[15].cIdentificationconfirmed by comparing mass spectra and retention time with those of authentic standard (S), and by NIST library (matching probability %).-: not detected.T.Abreu et al.