Natural Contaminants in Wines: Determination of Biogenic Amines by Chromatographic Techniques

Biogenic amines (BAs) are natural contaminants of wine that originate from decarboxylase microorganisms involved in fermentation processes. The primary relevance of biogenic amines in food could have both toxic effects on consumers’ health (i.e., allergic reactions, nausea, tremors, etc.), if present at high concentrations, and concurrently it can be considered as a remarkable indicator of quality and/or freshness. Therefore, the presence of nine biogenic amines [Tryptamine (TRP), ß-phenylethylamine (ß-PEA), putrescine (PUT), cadaverine (CAD), histamine (HIS), serotonin (SER), tyramine (TYR), spermidine (SPD), and spermine (SPM)] was investigated in red and white wine samples, which differed in the winemaking processes. The qualitative-quantitative determination of BAs was carried out by chromatographic methods (HPLC-UV/Vis and LC-ESI-MS). The analysis showed that both winemaking processes had all the nine BAs considered in the study at different amounts. Data showed that red wines had a higher concentration of PUT (10.52 mg L−1), TYR (7.57 mg L−1), and HIS (6.5 mg L−1), the BAs most involved in food poisoning, compared to white wines, probably related to the different type of fermentation (alcoholic and malolactic).


Introduction
Biogenic amines (BAs) are a class of organic, basic, and low-molecular weight compounds with heterocyclic (histamine, tryptamine), aliphatic (spermine, spermidine, putrescine and cadaverine) and aromatic (tyramine, phenylethylamine) structures [1,2]. They can be endogenous or exogenous in plants, as well as in animal and microorganisms, where they play, at lower concentration, an important role in physiological and metabolic functions-e.g., membrane stabilization, nucleic acid regulation, and protein synthesis [3]. The most amines occurring in food originate from proteolytic processes that make available large quantities of amino acids, which are the ideal substrate for enzymatic decarboxylation reactions. In addition, BAs may also be synthetized from the amination and transamination of aldehydes and ketones by the amino-acetic transaminases. BAs can be synthesized both in perishable and fresh food (i.e., fruits and vegetables, meat, fish, etc.), which are exposed to decarboxylase-positive microorganisms [4], as well as in fermented and/or processed food (i.e., wine, beer, coffee, chocolate, etc.), as a direct consequence of their transformation process-e.g., alcoholic and malolactic fermentation. The primary relevance of biogenic amines in food could have both toxic effects on consumers' health, if present at high concentrations, and concurrently it can be considered as a remarkable indicator of quality and/or freshness. From the toxicological point of view, BAs have been widely investigated as human harmful compounds, since their excessive food-mediated intake and a reduced or absent catabolism may induce symptoms that are similar to those of food poisoning: migraine headaches, gastric disorders, nausea, cardiac palpitations, and psychoactive effects [4,5]. The toxic effect most attributable to the ingestion of BAs is scombroid syndrome (Scombrotoxin Fish Poisoning, SFP), also called "Histamine poisoning", as this amine is the one responsible for this intoxication. However, the other BAs enhance the effects of histamine. Therefore, the toxicity of BAs in food or beverages is mainly due to a synergistic effect of several BAs [5]. Table 1 shows the physiological and pathological effects of BAs on human health. Below physiological conditions, BAs can be metabolized by three different enzymes, present in the gastrointestinal tract [diamino oxidase (DAO), monoamine oxidase (MAO), and histamine-N-methyltransferase (HMT)], which have the function of inactivating BAs by oxidizing the amino groups. Through these three enzymes, the human body has the ability to inactivate BAs that are normally taken with food and beverages [4,6]. Furthermore, in alcoholic beverages, it has been shown that the presence of ethanol and acetaldehyde, its catabolite, inhibits the enzymatic activity of BAs detoxifying enzymes (DAO, MAO, and HMT) and increases the biogenic amines permeability in the gastrointestinal wall, consequently increasing its toxic effects. This aspect is of relevant importance in alcoholic beverages, such as wine, which are eaten on its own or paired with other foods containing high concentrations of BAs (i.e., cheese, fish, meat, etc.) [7].
However, the toxic effects of BAs are dose-dependent, and the severity of the toxicity response is also influenced by personal sensitivity to these compounds. The symptoms, similar to those of food poisoning (nausea, vomiting, respiratory dysfunctions, itching, skin rash, etc.), have a variable duration between 8-12 h and they can therefore occur with more or less serious consequences even in subjects with a correct functioning of the enzymatic activity [4,8]. The daily dose of BAs acceptable to the human body is not yet known, as the toxic effects of the individual BAs are correlated and enhanced by their co-presence in food. Based on this, the EFSA (European Food Safety Authority) has defined that the dose, referred to histamine, for which the human body is able to activate the defense mechanisms is equal to about 25-50 mg and that the poisoning occurs following the intake of BAs equal to about 70-300 mg.
However, although their potential toxicity is known, to date, there is no legislation that allows for limiting the sale of products with high BA content. Currently, European legislation regulates only the presence of histamine in fish and fishery products (Reg. 2073/2005), while, as regards wine, only some European Countries (Germany, France, The Netherlands, Belgium, and Austria), arbitrarily, they proposed limits for histamine, ranging from 2 to 10 mg L −1 [5].
The formation of BAs therefore presupposes the co-presence of three factors: (i) a precursor, that is a specific amino acid for each specific BAs; (ii) contaminating microorganisms with decarboxylase activity; (iii) favorable environmental conditions (i.e., pH, temperature, water activity, etc.). Wine is an excellent substrate for BAs synthesis while ensuring the presence of amino acids, microbial populations with decarboxylating activity and a generally favorable environment for microorganism growth. Therefore, the BAs determination in wines is not only important to safeguard the consumers' health, but also for food quality assessment, because the presence of biogenic amines may influence the organoleptic characteristics of the finished product. The winemaking processes (Figure 1), which occur following different biochemical and metabolic pathways by microorganisms (bacteria, fungi, and yeasts), can lead to the formation of BAs [2,6].
The presence of these compounds in wine occurs at different points in the winemaking process. Especially, it can be influenced by the conservation conditions of grapes bunches, by their degree of ripeness and by the pedoclimatic conditions in which the wine is cultivated. The main step involving the BAs formation is the fermentation phase, in fact, depending on the type of microorganism involved and the type of fermentation (alcoholic and malolactic), not only the concentration but also the type of BAs present can be influenced in the finished product [1]. Table 2 shows the main microorganisms involved and the main BAs synthesized for the two types of fermentation. Altered behavior and neurochemical activities, cognitive decline, muscular inflammation, and immune activation [24,25] Two different winemaking processes, red and white ones, were taken into consideration as they are the most representative wine classes of the Italian market. In this study, to evaluate safety and quality of wines samples, nine biogenic amines were analyzed (TRP, HIS, TYR, β-PEA, CAD, PUT, SER, SPD, and SPM) and the BA profile was studied to identify the difference between red and white wines. The content of BAs has been analyzed by two chromatographic methods (HPLC UV/Vis and LC-ESI-MS).

Standard Solution
For each BAs (TRP, ß-PEA, PUT, CAD, HIS, SER, TYR, SPD and SPM), individual standard solutions were prepared at 1 mg mL −1 in purified water and kept in the dark at 4 ± 1 • C. In addition, the standard solution containing all nine BAs (MIX 9) was obtained with 1 mL of each standard solution of the individual BAs and diluted in 25 mL with purified water. Different aliquots of these standard solutions were applied to obtain the standard solutions necessary for the construction of the calibration curves and for the execution of

Samples
Forty-four wine samples (24 white wines and 20 red wines) produced in different regions of Italy were purchased in local supermarkets. The wine samples were stored at room temperature and protected from light until the day of analysis. The wine samples were chosen to be as representative as possible of the Italian wine market and of the red and white winemaking processes.

Biogenic Amines Extraction
The extraction of biogenic amines from the wine samples was carried out by applying the method described in a previous article [37]. The wine samples were previously filtered using a 0.20 µm Millipore membrane filter. Subsequently, for HPLC-UV/Vis analysis, 10.3 M HClO 4 was added to 25 mL of the filtered wine samples to obtain an acid solution at 0.2 M. Instead, for LC-ESI-MS analysis, HFBA was added to 25 mL of the filtered wine samples to obtain an acid solution at 5 mM. After a second filtration, an aliquot of 50 µL of the wine samples was injected into a chromatographic column.

HPLC-UV/Vis Method
Before performing the HPLC-UV/Vis analysis, the samples were subjected to derivatization, obtained through the use of dansyl-chloride [5-(dimethylamino)

LC-ESI-MS Method
During the LC-ESI-MS analyses, the ESI unit operated at 4.0 kV, the capillary was heated to 200 • C and as desolvation gas (300 L/h), as well as for the nebulizer (5 L/now), nitrogen was used. The ESI-MS system was set up to operate in positive ionization (PI) mode. Diagnostic fragment ions were obtained by in-source collisop-induced dissociation (CID) of the protonated molecule [M + H] + after optimization of the skimmer cone voltage. Selected Ion Monitoring (SIM) was applied for scheduled analyte recording. The mobile phase solvents, which were applied for LC-ESI-MS analysis, were (A) methanol (10 mM heptafluorobutyric acid) and (B) water (10 mM heptafluorobutyric acid), respectively. The mobile phase flow was set at a flow rate of 1 mL/min. The column was kept at room temperature and the eluted analytes using an initial linear gradient program from 10% solvent A to 85% in 15 min, then going from 85% solvent A to 100% in 1 min., followed by a 100% isocratic elution of A for 3 min. An additional 10 min have been added to reach initial conditions. The injected volume, both of standard solutions and of samples, was 50 µL. The data acquisition parameters are shown in Table 3.

Descriptive Analysis
All measurements were conducted in triplicate. The data obtained were analyzed mathematically and graphically using Microsoft Excel (Microsoft, Redmond, DC, USA). Data were expressed as mean ± standard deviation (SD). ANOVA tests were performed, and significantly different means were compared with the Turkey's pairwise test (p < 0.05) for all data collected. In addition, PCA (Principal Component Analysis) was applied to highlight a natural grouping of winemaking samples depending on their BA amounts. All the computations were performed using R-based Chemometrics Software (http://www.gruppochemiometria.it/index.php/software/19-download-the-r-basedchemometric-software, accessed on 15 July 2021).

Optimization and Performance Characteristics of the HPLC-UV/Vis Method
Before BA determination, the derivatization conditions were optimized as reported in Vinci et al. [38] and Vinci et al. [39]. Three different parameters have been optimized: pH, temperature, and time of reaction. The method optimization conditions are reported in Supplementary Materials. The performance characteristics of the HPLC-UV/Vis method are shown in Table S1. In Figure S1A, a chromatographic plot of the BA standard solution is shown. Calibration curves of nine biogenic amines are shown in Figure S1B.

Optimization and Performances of the LC-ESI-MS Method
Initially, the LC-ESI-MS method was performed to investigate the fragmentation behavior of the nine BAs, based on their mass/charge ratio. In order to carry out this evaluation, standard solutions of the single column-less BAs were injected in full scan mode. It was shown that these analytes, having a low relative molecular mass, split into a very small number of fragments [38,39]. The optimized LC-ESI-MS conditions to obtain the maximum fragments are summarized in Figure S2A,B.
To evaluate the performance of the method, linearity was taken into consideration, which was evaluated using standard solutions of the 9 BAs in 5mM HFBA acid solution [38,39]. The test results are summarized in Table S2.

Biogenic Amines Determination in Wine Samples
Forty-four wine samples (24 white and 20 red) were analyzed using both HPLC-UV and LC-ESI-MS methods under the selected experimental conditions. Three replicates were performed for each determination. Table 4 shows the BA amounts and their total concentration obtained for each wine sample. By comparing the reported values of the total BA amounts, it resulted in the total concentrations of BAs being much higher in red wines than in white wines. Data, showed in Figure 3A,B, can be explained by the fact that red wines are generally less acidic than white wines, and it is known in literature that BAs are produced in high quantities at high pH [27]. Furthermore, high values of BAs in wines are not only related to high pH values but also to the complexity of the bacterial microflora. Optimal growth conditions and greater bacterial diversity are mainly observed in red wines, which, therefore, show a higher content of BAs [8]. The significant differences observed in the content of BAs reported for the samples (Table 4) are probably attributable to the fact that the presence of BAs in wines is strongly dependent on different winemaking processes, which are characterized by different pH values of wines, the duration of fermentation, the aging time, and the soil and climatic conditions under which the wines are grown [30]. The data obtained show a high PUT content in both white (nd-4.22 mg L −1 ) and red (nd-10.52 mg L −1 ) wines. In particular, it was shown that red wines also had high concentrations of HIS (nd-7.57 mg L −1 ) and TYR (nd-6.59 mg L −1 ). Although these two amines have physiological functions, their excessive intake can cause food poisoning in consumers. Furthermore, their toxic effect is enhanced by the simultaneous intake of ethanol and its catabolites present in wines. Therefore, it is essential to determine HIS and TYR simultaneously, as they present a high risk of causing toxic effects due to their vasoactive and psychoactive properties [14]. Red wines also have a higher content of TRP (nd-2.50 mg L −1 ) and SER (nd-3.80 mg L −1 ), compared to white wines which are almost absent. This is probably due to the fermentation processes to which the grapes are subjected; in fact, it has been seen that wines that also undergo malolactic fermentation have higher concentrations of BAs [27]. The analysis of variance (ANOVA) showed significant differences (p < 0.05) among individual and total BA values. For this reason, all the biogenic amines are considered for multivariate analysis.  PCA analysis on the samples was performed to view the dataset in a reduced size and to evaluate the data matrix to highlight natural sample grouping. Figure S3A shows the loading plot of the nine amines and the total concentration of BAs in the samples, while the score plot ( Figure S3B) highlights the similarities and differences between the different wine samples taken into consideration. After autoscaling, two significant components were identified equal to 37.5% and 16.1% of the variance respectively for PC1 and PC2.
To better underline which biogenic amine mostly influenced the two categories of wine, the Biplot was carried out (Figure 4). It results in white wines being grouped in the negative quadrants (on the left) compared to PC1 while red wines are grouped mainly in the positive quadrants; this is explained by the fact that red wines weigh most on the presence of HIS, TYR, and PUT. This distinction that occurs in the two types of wine highlighted how red wines are the category of wine that can pose more risks to human health, as the combined presence of these BAs can lead to the risk of food poisoning. For this reason, and even though no official limit has yet been decided, some Countries, to protect the health of consumers, have established legal or recommended limits for histamine concentrations in wine [5].

Conclusions
The determination of nine biogenic amines in the white and red wine samples was carried out by applying two chromatographic methods. The LC-ESI-MS analysis offers the advantage of being a fast and reliable method for the qualitative-quantitative analysis of non-derivatized BA. This allows for identifying the presence of BA in wine samples more quickly. This is essential to quickly identify the presence of HIS and TYR in samples, since, if ingested with food, they are responsible for the main negative effects on human health (i.e., nausea, cramps, headaches, hypertension, tremors, etc.) [1]. Furthermore, the toxic effects of amines in wine can be enhanced by the synergistic effect of ethanol and acetaldehyde, which inactivates the enzymes responsible for the catabolism of BAs and increases their absorption in the gastro-intestinal wall [14]. The study highlighted the presence of all nine BA considered in the wines. Furthermore, differences in concentration were highlighted between the content of PUT, HIS, and TYR, which in red wines reached higher values, respectively of 10.52, 7.57, and 6.59 mg L −1 , while in white wines a lower content of 4.22 was found, 4.42 and 3.71 mg L −1 , respectively. This could probably be related to multiple factors: pH of wines, oenological processes, and hygienic conditions especially for fermentation processes, in relation to the microorganisms that are involved in alcoholic and malolactic fermentation [27][28][29][30][31][32][33][34][35][36][37][38][39]. Today, it is perhaps very difficult to obtain wines without BA, which keep all their organoleptic properties unaltered, even if one could act by controlling the critical technological factors, in particular the microorganisms involved in the fermentation processes; in this way, there would be the possibility of producing wines with low or moderate levels of BA, not dangerous for the health of consumers.  Table S1: Performance characteristics of HPLC-UV/Vis method. Table S2