Fatty Acid Profiling in Greek Wines by Liquid Chromatography–High-Resolution Mass Spectrometry (LC-HRMS)

Abstract

In recent years, the interest in lipids present in wines has increased, because these natural components, even at low or very low concentrations, play an important role in wine evolution and quality and contribute substantially to the taste and mouthfeel of wines. Herein, we present a liquid chromatography–high-resolution mass spectrometry (LC-HRMS) method for the profiling of free fatty acids (FFAs) in wines. The method is fast and allows the simultaneous determination of twenty-seven saturated and unsaturated FFAs in wine samples, avoiding any prior derivatization step. After validation, a variety of white and rose commercial wine samples from the Greek market, either sparkling or non-sparkling, were analyzed by the present method. The majority of wine FFAs are saturated long aliphatic, in particular palmitic (C16:0) and stearic (C18:0) acids, followed by myristic (C14:0) and pentadecanoic (C15:0) acids, while oleic (C18:1), palmitoleic (C16:1) and linoleic (C18:2) acids were quantified among the unsaturated FAs. The medium-chain C6:0 and the unsaturated C16:1 and C18:2 acids were found at higher concentrations in rose wines compared to white.

1. Introduction

Although the major components of wine are water (86%) and ethanol (12%), a variety of minor components, such as glycerol, polysaccharides, organic acids and volatile compounds, are present at different concentrations [1]. Even at low or very low concentrations, minor components play an important role in wine evolution and quality, contributing substantially to the color, mouthfeel or aromatic properties of wine [2,3]. Lipids, in particular fatty acids (FAs), constitute minor components of wines, and as recently reported [4], their concentration was found to be less than 0.1% in commercial Pinot noir wines. FAs found in wine may originate from the tissues of the grapes used, e.g., grape skin, flesh and seeds [5,6], since lipids are an integral part of the solid grape tissues. However, they may also be generated during alcoholic fermentation, since FAs are a building block in wine yeasts and they are released during this winemaking process [6,7].

The majority of FAs in wines occur as free acids, and they contribute directly to the flavor properties of the wine [6], since FAs are considered to have cheese, rancid and fatty notes [8,9,10]. The production of precursors of aldehydes and alcohols of six carbon atoms from unsaturated FAs may also be an indirect contribution to the wine flavor [1]. In recent years, several reports have discussed the contribution of the medium-chain fatty acids (MCFAs) hexanoic, octanoic and decanoic acids in the aroma of wines and the attempts to modify their levels and subsequently the aroma characteristics [10,11,12,13,14,15].

Lipid analysis and understanding links between lipid profiles and wine sensory perception and provenance is a field of great current interest, which may benefit wine research [16]. Recently, Tomasino and co-workers have demonstrated that lipids have an important impact on the sensory perception of wine [17]. Adding various food-grade lipids, such as triglycerides, mono- and diglycerides and phospholipids, at concentrations of 0.1% (w/v) into a model wine solution, they observed that the wine lipids may alter the taste and mouthfeel perception of wine [17]. Furthermore, Tomasino’s group adopted an untargeted lipidomic approach for Pinot noir wine samples to study 222 lipids from 11 different classes, concluding that wine lipids have a strong potential for the classification of wines by origin [4]. In their most recent studies, they examined the effect of winemaking factors, in particular fermentation temperature and yeast product addition, on the composition of lipids and the sensory characteristics of Pinot noir wines, and employing random forest analysis, they indicated that FAs were one of the most important lipid classes affected by these production factors [18].

The presence of FAs may also influence the foam formation and stability in sparkling wines [19,20,21]. Voce et al. have claimed that appreciable amounts of FAs such as palmitic acid potentially correlate with enhanced foam stability [21]. Culbert et al. have studied sparkling wines produced following different sparkling production methods and found that the levels of FAs and their esters depend on the production method [22].

The determination of lipids in wines is of high current importance [16]. However, analytical methods for the determination of lipids in wines are limited. The analytical methodology usually used for the determination of FAs in wines is gas chromatography–mass spectrometry (GC-MS). Such an approach traditionally requires a prior derivatization to fatty acid methyl esters (FAMEs). Gallart et al. described the determination of FFAs (C6-C18) and their ethyl esters in musts and wine by conversion to FAMEs [19]. Phan and Tomasino employed ultra-performance liquid chromatography–time-of-flight tandem mass spectrometry (UPLC-TOF-MS/MS) to develop an untargeted lipidomic profiling approach in order to correlate the lipid composition, including FFAs, with the origin of commercial Pinot noir wines [4]. Most recently, a new GC-MS method has been reported for the determination of short- and medium-chain FFAs in wines [23].

The aim of our work was the development and validation of a fast and simple analytical method for the determination of FFAs in wine samples. Herein, we present a liquid chromatography–high-resolution mass spectrometry (LC-HRMS) method that allows the detection and quantification of FFAs in wine, without the need for any tedious and time-consuming sample preparation or derivatization. The method was applied in the analysis of thirty brand Greek wines (20 white and 10 rose wine samples, including sparkling and non-sparkling samples).

2. Materials and Methods

2.1. Chemicals and Reagents

The following LC-MS analytical-grade solvents were used in the present study: acetonitrile from Carlo Erba (Val De Reuil, France), isopropanol and methanol from Fisher Scientific (Laughborough, UK) and formic acid 98–100% from Chem-Lab (Zedelgem, Belgium). Caproic acid (C6:0, >98%) was purchased from Alfa Aesar (Lancashire, UK). Caprylic acid (C8:0, >99.5%), capric acid (C10:0, >99%), myristic acid (C14:0, >99.5%), myristoleic acid (C14:1, >99%), pentadecanoic acid (C15:0, >99%), margaric acid (C17:0, >98%), linoleic acid (C18:2, >99%), linolenic acid (C18:3, >99%), arachidonic acid (C20:4, >99), behenic acid (C22:0, >99%), cis-7,10,13,16-docosatetraenoic acid (C22:4, >98%) and 4,7,10,13,16,19-cis-docosahexaenoic acid (C22:6, >98%) were purchased from Sigma Aldrich (Steinheim, Germany). Nonanoic acid (C9:0, >98%), 10-Z-heptadecenoic acid (C17:1, >98%), arachidic acid (C20:0, >98%), dihomo-γ-linolenic acid and 5,8,11-eicosatrienoic acid (C20:3, >98%), 7,10,13,16,19-cis-docosapentaenoic acid (C22:5, >98%) and lignoceric acid (C24:0, >98%) were purchased from Cayman Chemical Company (Ann Arbor, MI, USA). Lauric acid (C12:0, >99%) was purchased from Acros Organics (Geel, Belgium). Palmitic acid (C16:0, analytical standard), cis-9-palmitoleic acid (C16:1, analytical standard), stearic acid (C18:0, analytical standard), oleic acid and petroselinic acid (C18:1, analytical standard) and 5,8,11,14,17-Z-eicosapentaenoic acid (C20:5, analytical standard) were purchased from Fluka (Karlsruhe, Germany). A table presenting a detailed summary of the FFAs studied in this work is presented in the Supplementary Material (Table S1, Supplementary Material).

2.2. Stock and Working Solutions

Standard compounds were dissolved in methanol at a concentration of 1000 mg/L, and the stock solutions were stored at 4 °C. For the daily work, the stock solutions were appropriately diluted and used at 500 and 1000 ng/mL concentrations.

2.3. Instrumentation

A micro-LC Eksigent (Eksigent, Darmstadt, Germany), which was equipped with an autosampler set at 5 °C and a thermostated column compartment, was employed for the chromatographic studies. The column used was a Halo C18 one (2.7 μm, 90 Å, 0.5 × 50 mm) from Eksigent, and the mobile phase system consisted of solvent A: H₂O/0.01% and solvent B: acetonitrile/0.01% formic acid/isopropanol 80/20 v/v. A gradient elution program was set up for 10 min at a flow rate of 55 μL/min. More specifically, the program was as follows: 0–0.5 min, 5% B; 0.5–8.0 min, gradually increasing to 98% B; 8.0–8.5 min, 98% B, followed by a 1.5 min equilibration step to the initial conditions prior to the next injection. A 5 μL injection volume was selected for all the experiments.

Electrospray ionization (ESI) in negative mode was selected for all the HRMS measurements, which were performed employing an ABSciex Triple TOF 4600 (ABSciex, Darmstadt, Germany). The data acquisition method consisted of a TOF-MS full scan m/z 50–850 and several information-dependent acquisition (IDA) TOF-MS/MS product ion scans, using a 40 V collision energy (CE), with a 15 V collision energy spread (CES) used for each candidate ion in each data acquisition cycle (1091). The MS resolution working conditions were as follows: ion energy 1 (IE1), −2.3; vertical steering (VS1), −0.65; horizontal steering (HST), 1.15; and vertical steering 2 (VS2), 0.00.

2.4. Data Processing and Quantification

For the acquisition of the data, MultiQuant 3.0.2 and PeakView 2.1 from ABSciex (Darmstadt, Germany) were applied. Also, MultiQuant 3.0.2 was applied to record EICs, creating the base peak chromatograms for masses achieving a 0.01 Da mass accuracy width. A margin of ±2.5% was selected as the relative tolerance of the retention time. MultiQuant 3.0.2 was also used for the integration of the peak areas, which was carried out as previously reported [24,25], and in all cases, the same integration parameters were applied.

2.5. Sample Preparation

A liquid–liquid extraction protocol, using isopropanol, was applied. In a screw cap glass centrifuge tube containing 50 μL of a wine sample, 950 μL of isopropanol was added. The mixture was vortexed for about 30 s, and the suspension was centrifuged at 4000× g for 10 min. After filtration, 500 μL of the supernatant was mixed with 500 μL of water, and this mixture was used for the LC-MS/MS analytical measurements.

2.6. Method Validation

A mixed solution containing all the standards was used to spike the wine samples at 50, 100 and 300 ng/mL concentration levels. Experiments for each fortification level were carried out in triplicates. For the quantification of FFAs in the wine samples, the recovery of each FFA was used.

2.7. Wine Samples

Thirty brand Greek wine products from different regions and various producers in Greece were collected from the local market in Athens, Greece. They included 20 white wine samples (12 non-sparkling and 8 sparkling) and 10 rose wine samples (8 non-sparkling and 2 sparkling). The wine varieties, production areas and additional characteristics of wines (vintage year, % alcohol by volume (ABV), total acidity and pH, and aroma) are summarized in

Table 1. Code, grape variety, geographical origin and characteristics of the Greek wines studied.

Code	Variety	Area of Production	Type	Vintage Year	% ABV a	Acidity of Wine	Aroma
1	Moschofilero	Peloponnese	White dry	2017	11.5	Total acidity: 5.7 g/L pH: 3.4	Strong varietal aroma of flowers and citrus fruit on the nose and palate, with white rose and orange blossom prevailing.
2	Muscat (85%) and Vidiano (15%)	Crete	White dry	2019	12.5	Total acidity: 5.0 g/L pH: 3.5	Nose full of rose petal aromas and notes of mature citrus and stone fruit.
3	Lagorthi 65%, Roditis 25%, Chardonnay 10%	Peloponnese	White dry	2019	12.0	-	Aromas of melon, mango, kiwi and pineapple
4	Muscat	Samos	White dry	2016	13.5	Total acidity: 7.6 g/L pH: 2.9	Floral and fruity aromas to the nose, notably those of citrus, apricot and rose.
5	Savvatiano and roditis	Attica	White dry	2019	11.5	Soft acidity	Citrus fruits, gooseberry and ripe melon.
6	Savvatiano and roditis	Attica	White dry	2019	11.0	-	Aromas of pine resin.
7	Moschofilero	Peloponnese	White dry	2019	12.0	Strong acidity	Aromas of citrus fruits, lemon blossoms and rose.
8	Moschofilero	Peloponnese	White dry	2019	12.0	Total acidity: 5.9 g/L pH: 3.2	Delicate aroma of rose, green apple and lemon blossom.
9	Moschofilero	Peloponnese	White dry	2019	12.0	Total acidity: 5.5 g/L pH: 3.5	Aromas of rose, green apple and lemon blossom.
10	Muscat	Samos	White dry	2016	12.5	Total acidity: 7.1 g/L pH: 3.2	Floral and fruity aromas of muscat (pink roses) along with impressive notes of mainly white fruits, such as melon and peach.
11	Lagorthi	Peloponnese	White dry	2016	12.5	Strong acidity	Aromas of intense mineral character with notes of citrus fruits, ripe pear, Greek wild herbs and freshly cut hay.
12	Moschofilero	Peloponnese	White dry	2019	12.0	Total acidity: 5.7 g/L pH: 3.4	Citrus fruits, lemon and rose.
13	Xinomavro	Florina	Sparkling white semidry	2019	11.5	Total acidity: 5.2 g/L pH: 3.4	Apricot and white floral notes.
14	Xinomavro (85%)–Chardonnay (15%)	Florina	Sparkling white dry	2019	12.0	Total acidity: 5.2 g/L pH: 3.4	Aromas of red berries and citrus, as well as notes of white-fleshed stone fruits and orange blossoms.
15	Moschofilero	Peloponnese	Sparkling white dry	2019	12.0	Total acidity: 5.2 g/L pH: 3.4	Aromas of citrus blossoms, honey, brioche and rose.
16	Athiri	Rodos	Sparkling white dry	2019	13.0	Total acidity: 5.8 g/L pH: 3.1	Aromas of apricot, quince and freshly baked brioche.
17	Vidiano 100%	Crete	Sparkling white dry	2016	12.3	Total acidity: 7.2 g/L pH: 3.2	Aromas of apple, pear, sourdough from the “yeast” of fermentation, on a background of honey, wax and apricot.
18	Assyrtiko	Santorini	Sparkling white dry	2015	11.3	Total acidity: 7.6 g/L pH: 2.8	Citrus fruit, herbs, mineral, toasted.
19	Xinomavro	Florina	Sparkling white dry	2019	12.0	Total acidity: 5.2 g/L pH: 3.4	Aromas of red berries and citrus, as well as notes of white-fleshed stone fruits and orange blossoms.
20	Moschofilero	Peloponnese	Sparkling white dry	2019	12.5	Total acidity: 7.8 g/L pH: 3.1	Fresh and juicy with the classic notes of red cherry, mint and licorice.
1	Agiorgitiko	Attica	Rose dry	2019	12.0		Aromas of summer fruits with intense strawberry tones.
2	Kotsifali (60%) and syrah (40%)	Crete	Rose dry	2019	13.3	Total acidity: 5.6 g/L pH: 3.5	Aroma of strawberry and cherry.
3	Syrah	Attica	Rose dry	2019	12.5		Aromas of strawberry, cherry, apple and orange flowers.
4	Xinomavro	Thessaloniki	Rose dry	2019	13.0	Total acidity: 6.8 g/L pH: 3.8	Aroma of red fruits (strawberry and cherry).
5	Agiorgitiko	Peloponnese	Rose dry	2019	12.0		Red fruit aromas.
6	Xinomavro 100%	Florina	Rose dry	2019	13.0	Total acidity: 6.6 g/L pH: 3.2	Aromas of ripe strawberry, red forest fruits, tomato, and subtle botanical notes in the background.
7	Syrah	Peloponnese	Rose dry	2019	13.5		Delicate aromas of pink grapefruit, strawberry but also notes of cotton candy, white rose, and white pepper.
8	Agiorgitiko	Peloponnese	Rose Semidry	2019	12.0	Medium acidity, pH: 3.51	Aromas of nutmeg and cinnamon.
9	Xinomavro 100%	Florina	Sparkling rose dry	2018	12.0	Total acidity: 6.9 g/L pH: 3.0	Aromas of wild strawberry, white-fleshed cherry, and tomato.
10	Agiorgitiko	Peloponnese	Sparkling rose dry	2018	12.0	-	Aromas of cherry and black fruits.

^a ABV: alcohol by volume.

.

Several different varieties, namely Moschofilero, Vidiano, Muscat, Lagorthi, Savvatiano, Roditis, Xinomavro, Athiri, Assyrtiko, Agiorgitiko, Kotsifali and Syrah, were included in the present study. Among the 12 non-sparkling white wines, 7 originated from Peloponnese, 1 from Crete, 2 from Samos and 2 from Attica. Among the eight sparkling white wines, three were from Florina, two from Peloponnese, one from Rodos, one from Crete and one from Santorini. Among the eight rose wines, two were from Attica, one from Crete, one from Thessaloniki, three from Peloponnese and one from Florina. One of the two sparkling rose wines was from Florina, and the other was from Peloponnese. A map of the origin of samples is depicted in Figure 1.

2.8. Statistical Analysis

The level of significance was estimated using a two-sample Excel t-test assuming unequal variances.

3. Results

3.1. Sample Preparation and Method Validation

A simple sample preparation procedure involving the addition of isopropanol was followed (Figure S1, Supplementary Material). The supernatant obtained after centrifugation was used for the analysis. The guidelines of the EU Commission decision 202/657/EC were followed for the verification of accuracy and precision. Satisfactory recoveries indicate the accuracy of the proposed method. The recoveries ranged from 81 to 104, 83 to 103 and 75 to 106 for the low, medium and high spike levels, respectively (Table S2, Supplementary Material). The relative standard deviation (%RSD) values, indicating the precision, that were obtained for intra-day (RSD_r) and inter-day (RSD_R) variations ranged from 0.01 to 8.28 and from 0.04 to 9.31, respectively, depending on the FA (Table S2, Supplementary Material).

3.2. Analysis of Samples

A rapid LC-HRMS method that allows the simultaneous determination of a variety of FFAs (twenty-seven) in wine samples in a 10-min run was developed. More specifically, FAs C6:0, C8:0, C9:0, C10:0, C12:0, C14:0, C14:1, C15:0, C16:0, C16:1, C17:0, C17:1, C18:0, C18:1, C18:2, C18:3, C20:0, C20:3, C20:4, C20:5, C22:0, C22:4, C22:5, C22:6 and C24:0 were analyzed. The exact masses [M-H]- of all analytes along with their chromatographic retention times R_t are summarized in Table S1 (Supplementary Material). The limits of detection (LODs) and quantification (LOQs), including some data from our previous reports [26,27], are summarized in Table S1 (Supplementary Material). The extracted ion chromatograms (EICs) of FFAs in a white (A), sparkling white (B), rose (C) and sparkling rose (D) wine sample are presented in Figure 2 and Figure S2 (Supplementary Material).

In the case of two couples of isobaric FAs (C18:1 and C20:3), the present chromatographic method provides a satisfactory chromatographic separation, as shown in Figure 3. Both oleic acid (C18:1 cis-9) and petroselinic acid (C18:1 cis-6) share the same molecular formula and exact mass (C₁₈H₃₄O₂, 232.2559). However, they are eluted at different R_ts, as depicted in Figure 3A (EIC of a standard solution). Oleic acid was found to be present in both white and rose wine samples, while petroselinic acid was absent (Figure 3B,C). Similarly, a satisfactory separation was observed for isomers C20:3 5,8,11-eicosatrienoic acid and bishomo-γ-linolenic acid (molecular formula C₂₀H₃₄O₂, exact mass 306.2559), as shown in Figure 3A. None of these FAs was detected in either white or rose wine samples (Figure 3B,C).

Twenty white wine samples (eight of them sparkling wines) and ten rose wine samples (two of them sparkling wines) have been analyzed in the present study. The contents of FFAs in wine samples (μg/mL wine) are summarized in

Table 2. Contents of FFAs in white and sparkling white wine samples (μg/mL wine).

	White (n = 12), Triplicates					Sparkling White (n = 8), Triplicates
Free Fatty Acid	Minimum Value (μg/mL)	Maximum Value (μg/mL)	Mean Value ± SD (μg/mL)	α	Minimum Value (μg/mL)	Maximum Value (μg/mL)	Mean Value ± SD (μg/mL)	α
C6:0	0.55	1.48	0.97 ± 0.20	***	0.58	4.01	1.29 ± 0.62	***
C8:0	0.40	1.09	0.73 ± 0.20	***	0.34	1.68	0.72 ± 0.20	***
C9:0	0.14	0.37	0.25 ± 0.06	***	-	0.40	0.16 ± 0.09	***
C10:0	0.09	0.24	0.15 ± 0.05	***	-	0.14	0.08 ± 0.02	***
C12:0	0.07	0.41	0.30 ± 0.09	**	0.15	2.94	0.54 ± 0.64	**
C14:0	0.73	3.35	2.37 ± 0.90	***	0.65	12.17	2.75 ± 1.1	***
C14:1	-	-	-		-	-	-
C15:0	0.41	2.26	1.53 ± 0.10	***	0.54	5.80	1.62 ± 0.71	***
C16:0	5.54	25.22	17.64 ± 2.1	***	7.04	27.07	16.92 ± 2.7	***
C16:1	<LOQ a	0.05	0.03 ± 0.01 b	**	0.02	0.33	0.08 ± 0.03	**
C17:0	0.23	0.75	0.51 ± 0.09	***	0.13	1.46	0.51 ± 0.21	***
C17:1	-	-	-		-	-	-
C18:0	2.00	15.76	11.98 ± 2.6	***	5.01	19.55	10.83 ± 2.5	***
C18:1 Oleic acid	0.15	2.34	0.90 ± 0.30	***	0.29	9.27	1.52 ± 0.41	***
C18:1 Petroselinic acid	-	-	-		-	-	-
C18:2	0.03	0.12	0.08 ± 0.02	**	0.10	0.92	0.20 ± 0.09	**
C18:3	-	-	-		-	-	-
C20:0	<LOQ c	0.24	0.13 ± 0.04 b	**	<LOQ d	0.52	0.08 ± 0.06 b	**
C20:3 Bishomo-γ-linolenic	-	-	-		-	-	-
C20:3 5,8,11-eicosatrienoic	-	-	-		-	-	-
C20:4	-	-	-		-	-	-
C20:5	-	-	-		-	-	-
C22:0	0.06	0.27	0.14 ± 0.05	***	0.06	0.58	0.17 ± 0.09	***
C22:4	-	-	-		-	-	-
C22:5	-	-	-		-	-	-
C22:6	-	-	-		-	-	-
C24:0	0.42	1.29	0.79 ± 0.26	***	<LOQ d	1.41	0.60 ± 0.31 b	***

Content lower than LOQ in ^a 2, ^c 7, ^d 1 samples; ^b the mean value was determined using medium-bound approach; <LOQ: lower of limit of quantification; SD: standard deviation; α: level of significance; ** p < 0.01, *** p < 0.001.

and

Table 3. Contents of FFAs in rose and sparkling rose wine samples (μg/mL wine).

	Rose (n = 8), Triplicates					Sparkling Rose (n = 2), Triplicates
Free Fatty Acid	Minimum Value (μg/mL)	Maximum Value (μg/mL)	Mean Value ± SD (μg/mL)	α	Minimum Value (μg/mL)	Maximum Value (μg/mL)	Mean Value ± SD (μg/mL)	α
C6:0	1.00	1.61	1.32 ± 0.20	***	0.90	1.00	0.95 ± 0.07	***
C8:0	0.62	1.21	0.89 ± 0.20	***	0.59	0.60	0.60 ± 0.01	***
C9:0	0.18	0.24	0.21 ± 0.02	***	0.11	0.14	0.13 ± 0.02	***
C10:0	0.15	0.26	0.19 ± 0.04	***	0.08	0.10	0.09 ± 0.01	***
C12:0	0.28	0.35	0.32 ± 0.02	***	0.17	0.19	0.18 ± 0.01	***
C14:0	1.65	2.38	1.93 ± 0.09	***	1.07	1.12	1.10 ± 0.04	***
C14:1	-	-	-		-	-	-
C15:0	1.26	1.55	1.42 ± 0.11	***	0.81	0.85	0.83 ± 0.03	***
C16:0	16.74	29.32	20.75 ± 2.21	***	15.16	17.38	16.27 ± 1.57	***
C16:1	0.05	0.06	0.06 ± 0.01	***	0.04	0.05	0.05 ± 0.01	***
C17:0	0.39	0.60	0.46 ± 0.06	***	0.17	0.21	0.19 ± 0.03	***
C17:1	-	-	-		-	-	-
C18:0	9.13	24.39	15.39 ± 1.9	***	10.66	12.43	11.55 ± 1.25	***
C18:1 Oleic acid	0.49	7.64	1.61 ± 1.5	**	0.34	0.47	0.41 ± 0.09	**
C18:1 Petroselinic acid	-	-	-		-	-	-
C18:2	0.07	0.14	0.11 ± 0.03	***	0.15	0.18	0.17 ± 0.02	***
C18:3	-	-	-
C20:0	-	-	-		-	-	-
C20:3 Bishomo-γ-linolenic	-	-	-		-	-	-
C20:3 5,8,11-eicosatrienoic	-	-	-		-	-	-
C20:4	-	-	-		-	-	-
C20:5	-	-	-		-	-	-
C22:0	0.11	0.17	0.14 ± 0.02	***	0.05	0.08	0.07 ± 0.02	***
C22:4	-	-	-		-	-	-
C22:5	-	-	-		-	-	-
C22:6	-	-	-		-	-	-
C24:0	0.58	0.86	0.70 ± 0.07	***	0.40	0.66	0.53 ± 0.18	***

SD: standard deviation; α: level of significance; ** p < 0.01, *** p < 0.001.

.

As shown in Table 2 and Table 3, the long-chain saturated C16:0 and C18:0 acids were found at the highest concentrations (higher than 10 μg/mL) in all wine samples of this study, irrespectively white or rose, sparkling or non-sparkling. In white wine samples, C16:0 was found to be the most abundant FFA, ranging from 5.54 to 25.22 μg/mL with a mean value of 17.64 μg/mL, while in white sparkling wine samples, its content ranged from 7.04 to 27.07 μg/mL with a mean value of 16.92 μg/mL (Table 2). It was followed by C18:0, ranging in white wine from 2.00 to 15.76 μg/mL with a mean value of 11.98 μg/mL and in white sparkling wine from 5.01 to 19.55 with a mean value of 10.83 μg/mL (Table 2). Similarly, in rose wine samples, the content of C16:0 ranged from 16.74 to 29.32 μg/mL with a mean value of 20.75 μg/mL, and in the sparkling rose wine samples, from 15.16 to 17.38 μg/mL with a mean value of 16.27 μg/mL (Table 3). The content of C18:0 in rose wine ranged from 9.13 to 24.39 μg/mL with a mean value of 15.39 μg/mL, and that in sparkling rose wine ranged from 10.66 to 12.43 μg/mL with a mean value of 11.55 μg/mL (Table 3).

In white wine samples, the saturated C14:0 and C15:0 FAs were estimated at concentrations higher than 1.0 μg/mL (2.37 μg/mL and 1.53 μg/mL, respectively), followed by C6:0 and the unsaturated C18:1 oleic acid (0.97 μg/mL and 0.90 μg/mL, respectively) (Table 2). In white sparkling wine, all these four FFAs (saturated C14:0, C15:0, C6:0 and unsaturated C18:1 oleic acid) were found at concentrations higher than 1.0 μg/mL (2.75 μg/mL, 1.62 μg/mL, 1.29 μg/mL and 1.52 μg/mL, respectively). In both white wine samples and white sparkling samples, ten FFAs were found at concentrations lower than 1.0 μg/mL (C8:0, C9:0, C10:0, C12:0, C16:1, C17:0, C18:2, C20:0, C22:0 and C24:0), and eleven FFAs were not detected at all (C14:1, C17:1, C18:1 petroselinic acid, C18:3, C20:3, C20:4, C20:5, C22:4, C22:5 and C22:6) (Table 2).

In rose wine samples, four FFAs (saturated C14:0, C15:0, C6:0 and unsaturated C18:1) were found at concentrations higher than 1.0 μg/mL (1.93 μg/mL, 1.42 μg/mL, 1.32 μg/mL and 1.61 μg/mL, respectively) (Table 3). In contrast, in rose sparkling wines, decreased contents of these four FFAs were estimated, and only C14:0 was found at a concentration higher than 1.0 μg/mL (1.10 μg/mL) (Table 3). In addition, in both rose and rose sparkling wines, nine FFAs were found at concentrations lower than 1.0 μg/mL (C8:0, C9:0, C10:0, C12:0, C16:1, C17:0, C18:2, C22:0 and C24:0), and twelve FFAs were not detected at all (C14:1, C17:1, C18:1 petroselinic acid, C18:3, C20:0, C20:3, C20:4, C20:5, C22:4, C22:5 and C22:6) (Table 3).

4. Discussion

The lipid composition of wines, including FFA contents, is highly affected by the origin of grapes, the amount of lipids in musts, the winemaking process and the yeasts used for fermentation. The alterations of the FFA levels in white, sparkling white, rose and sparkling rose wine samples studied in this work are better shown in Figure 4. In the present study, the long-chain saturated FAs C16:0 and C18:0 were found to be the predominant FFAs in all wines, white or rose, sparkling or non-sparkling. Furthermore, the long-chain saturated C14:0 and C15:0 acids were found at notable concentrations in both white and rose wines. The long-chain C24:0 and C22:0 acids were also detected and quantified in both white and rose wines. C20:0 was detected and quantified in some white wines, while it was absent in red wines. As for the unsaturated FAs, C18:1 oleic acid was found as the predominant unsaturated FFA and, in addition, C18:2 was detected and quantified in both white and rose wines.

Comparing the FFA contents of white with rose wine samples, a statistically significant increase in rose wines was observed for 3 FFAs, namely C6:0, C16:1 and C18:2 (Figure 4). The varieties of the white wines studied were Moschofilero, Muscat, Vidiano, Lagorthi, Savvatiano and Roditis, totally different from those of the rose wines, namely Agiorgitiko, Kotsifali, Syrah and Xinomavro. Thus, the variations observed for C6:0, C16:1 and C18:2 may be attributed to the different wine varieties. Comparing sparkling with non-sparkling wines, as shown in Figure 4, the contents of all FFAs, apart from C18:2, were found decreased in rose sparkling wines in comparison to non-sparkling rose wines. In particular, a notable decrease was observed for the contents of C9:0, C10:0, C12:0 and C15:0. The sparkling rose wines belong to Agiorgitiko and Xinomavro varieties, while the non-sparkling rose wines to Agiorgitiko, Xinomavro, Syrah and Kotsifali. Because of the overlap of varieties, the differences observed for C9:0, C10:0, C12:0 and C15:0, most likely, should not be attributed to varieties, but to winemaking processes and/or yeasts used. As for the white wines, a statistically significant decrease of C10:0 was observed comparing sparkling with not sparkling (Figure 4). Again, this variation could be a result of the winemaking processes and/or the yeast used rather than the wine variety, since the varieties of the white non-sparkling wines were Moschofilero, Muscat, Vidiano, Lagorthi, Savvatiano and Roditis, in part overlapped with those of the sparkling white wines, namely Moschofilero, Vidiano, Xinomavro, Athiri and Assyrtiko.

The MCFAs have recently attracted special attention in wines, because they may contribute to the aroma of the wine [10,11,12,13,14,15]. According to a recent report [10], C6:0 and C8:0, having a cheese, acidic aroma description, and a sweat and cheese aroma description, respectively, contribute to the aroma of aged Chinese rice wine. In our study, we found that C6:0 was the predominant MCFA in both white and rose wines at concentrations ranging from 0.97 μg/mL to 1.29 μg/mL (white wine) and from 1.32 μg/mL to 0.95 μg/mL (rose wine) (Table 2 and Table 3). It was followed by C8:0, whose content ranged from 0.73 μg/mL to 0.72 μg/mL (white wine) and from 0.89 μg/mL to 0.60 μg/mL (rose wine) (Table 2 and Table 3). Given that the aroma thresholds for C6:0 and C8:0 are reported to be 0.42 mg/L and 0.50 mg/L, respectively [10], both these MCFAs contribute to the aroma of the wines studied in this work, whose aromas are summarized in Table 1.

Differentiation of wines based on chemical descriptors and correlation of the contents of distinct chemical components of wines with geographical origin is of high interest and lipids may serve as such chemical descriptors [16]. A number of diverse lipids, including FFA C10:0 and C12:0, were used by Phan and Tomasino [4] for predicting the origin of Pinot noir wines. Future research applying the developed and validated LC-HRMS method is worthy to be carried out for linking FFAs contents with geographical origin of wines.

5. Conclusions

In the present work, we describe the development and the validation of a convenient and fast LC-HRMS analytical method, which allows the simultaneous determination of a variety of FFAs in wine samples. The method employs a simple sample preparation and does not require any prior derivatization step. Twenty-seven saturated and unsaturated FFAs were analyzed in thirty samples of Greek white and rose, sparkling and non-sparkling, wines. The saturated FAs, in particular palmitic and stearic acids, constitute the major part of wine FFAs, while the oleic acid was the predominant unsaturated, found at significantly lower concentrations. The medium-chain C6:0 and the unsaturated C16:1 and C18:2 acids were found at higher concentrations in rose wines compared to white. Among MCFAs that may contribute to the aroma of the wine, caproic acid was the predominant one in both white and rose wines, followed by caprylic acid. The method developed and validated in the present study may find applications in future metabolomics studies on wines.