*Article* **Color Stabilization of Apulian Red Wines through the Sequential Inoculation of** *Starmerella bacillaris* **and** *Saccharomyces cerevisiae*

**Matteo Velenosi 1,2,†, Pasquale Crupi 1,†, Rocco Perniola <sup>1</sup> , Antonio Domenico Marsico <sup>1</sup> , Antonella Salerno 1,2 , Hervè Alexandre <sup>3</sup> , Nicoletta Archidiacono <sup>2</sup> , Mario Ventura 2,\* and Maria Francesca Cardone 1,\***


**Abstract:** Mixed fermentation using *Starmerella bacillaris* and *Saccharomyces cerevisiae* has gained attention in recent years due to their ability to modulate the qualitative parameters of enological interest, such as the color intensity and stability of wine. In this study, three of the most important red Apulian varieties were fermented through two pure inoculations of *Saccharomyces cerevisiae* strains or the sequential inoculation of *Saccharomyces cerevisiae* after 48 h from *Starmerella bacillaris*. The evolution of anthocyanin profiles and chromatic characteristics were determined in the produced wines at draining off and after 18 months of bottle aging in order to assess the impact of the different fermentation protocols on the potential color stabilization and shelf-life. The chemical composition analysis showed titratable acidity and ethanol content exhibiting marked differences among wines after fermentation and aging. The 48 h inoculation delay produced wines with higher values of color intensity and color stability. This was ascribed to the increased presence of compounds, such as stable A-type vitisins and reddish/violet ethylidene-bridge flavonol-anthocyanin adducts, in the mixed fermentation. Our results proved that the sequential fermentation of *Starmerella bacillaris* and *Saccharomyces cerevisiae* could enhance the chromatic profile as well as the stability of the red wines, thus improving their organoleptic quality.

**Keywords:** HPLC-UV-ESI-MSn; free anthocyanins; co-pigmented anthocyanins; mixed fermentation; *starmerella bacillaris*; PCA

## **1. Introduction**

Yeast metabolism, during the winemaking process, influences the wine organoleptic properties and, consequently, wine quality. It can directly or indirectly affect the content of several compounds related to both the aroma and color characteristics. Recent studies on mixed starter cultures have proved that the resulting wines differ significantly, concerning both their chemical composition and sensory characteristics. Different yeast species and the ratio of non-*Saccharomyces/Saccharomyces* yeasts determine the organoleptic properties of the final product, and therefore contribute differently to the improvement or depreciations of wine quality [1].

The color is the most important visual attribute of red wines [2], which strongly impresses consumers' purchasing preference [3]. Moreover, it influences the perception

**Citation:** Velenosi, M.; Crupi, P.; Perniola, R.; Marsico, A.D.; Salerno, A.; Alexandre, H.; Archidiacono, N.; Ventura, M.; Cardone, M.F. Color Stabilization of Apulian Red Wines through the Sequential Inoculation of *Starmerella bacillaris* and *Saccharomyces cerevisiae*. *Molecules* **2021**, *26*, 907. https:// doi.org/10.3390/molecules26040907

Academic Editor: Mirella Nardini Received: 18 January 2021 Accepted: 5 February 2021 Published: 9 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of other sensory properties, such as aroma and flavor. Therefore, winemakers have accustomed to adopting suitable practices that improve color extraction and enhance the stability of chromatic characteristics of wine over time [4]. The color of red wines is mainly due to anthocyanins, which are transferred from grape skins into wine throughout the maceration/fermentation process [5]. Whereas, the stability of color during wine aging is affected by the phenolic derivatives which stabilize anthocyanins through co-pigmentation reactions [6,7]. The types and concentrations of polyphenols in wine may depend on the grape variety, the degree of ripening [8], and the vine growing methods employed, specifically the pruning and training system [7,9]. The joining of additives (i.e., enzymes, yeasts, or tannins) during winemaking is also a determinant [4,10,11].

In this aspect, there has been growing interest in the use of non-*Saccharomyces* yeasts due to the positive impact some of their metabolites exert on wine quality [12,13]. Many authors have demonstrated that non-*Saccharomyces* yeasts have a protective effect on wine color [4,10,14]. Among these, *Starmerella bacillaris (S. bacillaris)* [15] has been considered one of the most promising non-*Saccharomyces* yeasts [16–18] (having strong fructophilicity, high tolerance to low temperatures, and ability to grow at an elevated sugar concentration) [19]. However, non-*Saccharomyces* yeasts possess low fermentation ability and cannot carry out the must fermentation alone, due to their ethanol sensitivity [20,21]. Consequently, their use in combination with selected *Saccharomyces cerevisiae (S. cerevisiae)* (Desm. Meyen 1838) strains is necessary for completing the fermentation and taking advantage of their unique features [22].

Recently, a meaningful knowledge has been accumulated about the importance of yeast inoculation density, timing, and combination of strains in improving the organoleptic properties of wines [16,23,24]. The use of *S. bacillaris* during winemaking has allowed increasing the must total acidity and enhancing the color intensity of wine [25,26]. Similarly, this yeast strain has led to a higher production of pyruvic acid, which is involved in the formation of stable pigments (i.e., vitisin A and B), compared to *Saccharomyces* [27]. Thereby, it could be hypothesized that a mixed fermentation (by employing both the yeasts, sequentially) works in improving the color intensity as well as the color stability of wine. This study aimed at comparing the anthocyanin profiles and chromatic characteristics of wines produced through two mono-*S. cerevisiae* fermentations (SCE16 and SCE138, respectively) or the sequential fermentation of *S. bacillaris* and *S. cerevisiae* SCE16/SCE138 inoculated 48 h later. The analyses were conducted on wines produced from the most important red Apulian varieties (Primitivo, Negramaro, and Aleatico) at draining off and after 18 months of bottle aging, to investigate the potential of the color stabilization and shelf life of these wines.

#### **2. Results**

#### *2.1. Interaction between Saccharomyces Yeast Strains and Pilot Scale Fermentation*

In order to evaluate the suitability of the three yeast strains in mixed fermentation, we first evaluated the phytotoxic activity towards each other both on plate and liquid culture assays.

In the experiment performed on the plate assay, the three yeast strains were able to grow independently of the previous growth of the other tested yeast strain on the cellophane disc. Furthermore, the growth curves of each yeast strain are similar regardless of the type of filtered supernatant added (Supplementary Figure S1). Likewise, no inhibition of growth was observed in the liquid culture assay combining two yeast strains together, both considering the interaction of *S. cerevisiae* strains or *S. bacillaris* with each of the *S. cerevisiae* strain. Taking into consideration the absence of any phytoxic activity among the different combinations of yeast strains, we were able to test their effect on wine production in a mixed fermentation where the two *S. cerevisiae* strains (SCE16 and SCE138, 1:1) were added together 48 h after the inoculation of *S. bacillaris* (FA18), and compare this trial with mono-*saccharomyces* fermentation. Moreover, in order to assess the fermentation ability of the chosen yeast combination with respect to mono-fermentation and, in particular, to further verify the absence of any negative interaction in mixed fermentation among the

yeast strains, fermentation kinetics were followed for each trial (Supplementary Figure S2). Mono inoculation SCE16 and SC138 showed a similar or equal consumption in sugar level in every variety considered, thus demonstrating the same fermentation capacity of the two *S. cerevisiae* strains. On the contrary, the mixed FA18 was characterized by a slow start, regarding the sugar consumption, reaching up to 7% in Primitivo, 8% in Negroamaro, and 13% in Aleatico. The higher delay we found in the Primitivo could be ascribed to the sugar concentration effect on the *S. bacillaris* activity, as previously described [28]. Indeed, the sugar and nitrogen composition of the grape must are key factors for the evolution of the alcoholic fermentation and the development of the yeasts [29,30]. Notably, also in the FA18 mixed fermentation, complete sugar consumption was reached around 4 days after the inoculum with the two *S. cerevisiae* strains, thus confirming the absence of a negative interaction between the yeast strains both considering *S. bacillaris* against *S. cerevisiae*, and between the two *S. cerevisiae* strains.

#### *2.2. Basic Oenological Parameters and Chemical Composition*

The chemical composition of Primitivo, Negramaro, and Aleatico wines produced by pure and mixed culture fermentation at draining off and after 18 months of bottle aging were listed in Table 1.

**Table 1.** Chemical analysis and polyphenolic indexes of Primitivo, Negramaro, and Aleatico at draining off (A) and after 18 months of bottle aging (B).


Each value was calculated as means of three independent replicates <sup>±</sup> *#* standard deviation at *p* < 0.05. Different letters on the same line are significantly different at a 5% level (Tukey's HSD post-hoc test). CI: Color intensity; MA: Monomeric anthocyanins; TA: Total anthocyanins; TP: Total polyphenols; A: Total acidity; ET: Alcoholic degree; VA: Volatile acidity; H: Hue; CEI: Color evolution index.

27

Overall, the fermentation type factor influenced the titratable acidity (A) of the wines. Indeed, samples obtained by mixed fermentation generally contained more acids, in particular, Negramaro and Aleatico FA18 wines had a significantly higher A (*p* < 0.01). These differences (ranging from 0.25 to 0.57 g/L) cannot be imputed to the main organic acids (citric, malic, tartaric, and lactic acids) whose values did not significantly change in all the wines (Supplementary Table S1).

Furthermore, pH values were not affected by the different fermentation protocols at drying off (Table 1). Conversely, these findings may be due to the capability of *S. bacillaris* strains to relatively synthesize high concentrations of keto acids either during the early stages of fermentation from sugar metabolism or from the corresponding amino acids (alanine for pyruvic acid and glutamate for α-keto glutaric acid), as previously reported [27,31,32]. On the contrary, we revealed a significantly higher pH value in 18 months aged wines connected to the partial tartaric precipitation that happened during aging in the bottle. However, not surprisingly, the slight decrease of A (total acidity) during the wines aging could also be due to a series of maturation reactions involving pyruvic acid [7].

No significant difference in the alcoholic degree (% *v*/*v*) was registered between pure and mixed fermentation in all the samples (Table 1). Moreover, the volatile acidity was strongly influenced by the fermentation protocol and bottle aging, as well as by the interaction of the two factors (*p* < 0.001), even though all the wines contained <0.40 g/L (Table 1), which cannot be considered detrimental to the sensorial quality of wine as in agreement with literature data [33]. Furthermore, we analyzed the polyphenolic content and we found that monomeric anthocyanins (MA), total anthocyanins (TA), and total polyphenols (TP) values appeared significantly higher in SCE16 and SCE138 than in the FA18 samples, especially for Negramaro and Aleatico (Table 1). Moreover, we detected a decrease of phenolics after 18 months which, was generally more marked in FA18 than SCE16 and SCE138 wines (Table 1).

#### *2.3. HPLC-MS Analysis of Anthocyanin Profile in the Wines*

The color changes during wine maturation are usually attributed to anthocyanin polymerization reactions and the evolution of co-pigments resulting from interactions between anthocyanins and other compounds at the fermentation phase [34,35]. For these reasons, we investigated the anthocyanin profile of the wines by HPLC-MS analyses and the pigments, identified through their retention time (RT), molecular ion (M<sup>+</sup> ), and principal MS/MS fragments, as listed in Table 2.

Five mono glucoside anthocyanins, namely delphinidin (3), cyanidin (5), petunidin (6), peonidin (8), and malvidin (9), together with malvidin-3-*O*-acetylglucoside (15), malvidin-3-*O*-caffeoylglucoside (17), cyanidin-3-*O*-(*p*-coumaroyl)glucoside (17), peonidin-3-*O*-trans-(*p*-coumaroyl)glucoside (20), and malvidin-3-*O*-trans-(*p*-coumaroyl)glucoside (21) were revealed in all the samples. Whilst other acyl compounds, such as peonidin-3-*O*-acetylglucoside (14), petunidin-3-*O*-(*p*-coumaroyl)glucoside (18), and malvidin-3-*Ocis*-(*p*-coumaroyl)glucoside (18), also belonging to the group of free-anthocyanins directly extracted from grape skin [36,37], were not detected in Aleatico wines. Four compounds corresponding to carboxy-pyranoanthocyanins derived from the reaction between glucoside anthocyanins and pyruvic acid (A-type vitisins) were also identified (Table 2). In particular, petunidin (7) and malvidin (10) 3-*O*-glucoside pyruvate were present in all the samples, while peonidin (13) and malvidin (14) 3-*O*-(*p*-coumaroyl)glucoside pyruvate were absent in Aleatico wines. Two well resolved chromatographic peaks (11 and 12) referring to isobaric ions with similar MS/MS spectra were achieved for the species with [M]+ at *m*/*z* 809, which were identified as isomers of malvidin-3-*O*-glucoside-8-ethyl-(epi)catechin [38]. Then, other ethylidene-bridged flavanol anthocyanins, namely peonidin-3-*O*-(*p*-coumaroyl)glucoside-8-ethyl-(epi)catechin (19) and malvidin-3-*O*-(*p*-coumaroyl)glucoside-8-ethyl-(epi)catechin (23), were revealed in the wines (Table 2). With regards to vinyl-linked flavanol anthocyanins, also known as flavanol pyranoanthocyanins [33], malvidin-3-*O*-acetylglucoside-4 vinyl-(epi)catechin (16) and malvidin-3-*O*-glucoside-4-vinyl-(epi)catechin (22) were only de-

tected in Primitivo and Negramaro wines, respectively. Finally, three flavanol-anthocyanins derivatives, having molecular ions and fragmentation patterns typical of (epi)-catechinpeonidin (1) or malvidin-3-*O*-glucoside (2 and 4) adducts [7] were found (Table 2).


**Table 2.** Chromatographic and mass spectral data of the identified anthocyanin compounds.

In order to investigate the influence of the fermentation type on the formation and evolution of anthocyanin derived pigments, involved in the color intensity and stability, PCA analyses were performed on Primitivo, Negramaro, and Aleatico wines at draining off and after 18 months of bottle storage. Moreover, the percentage content of the five different classes of pigments were compared among the wines at the two time-points of aging (Figures 1–3). Overall, the mixed fermentation protocol provoked the increasing synthesis of stable pigments in the wines during the vinification process. Indeed, at draining off, the FA18 samples appeared richer in pyranoanthocyanins and ethylidene-bridged compounds, whose content was also enhanced during the bottle aging, thus contributing to the intensity and stability of the color. This was in agreement with the effect of sequential inoculum (delay of 5 days) with *S. bacillaris* CZ1 in the production of wines with a higher level of Atype vitisins [39]. Regarding Primitivo at draining off (Figure 1a), FA18 was characterized by a higher content of vitisin A (10), but also reddish/violet ethylidene-bridged compounds

(11, 12, and 19) and bluish pigment (4). On the contrary, SCE16 (and less SCE138) showed greater amounts of free anthocyanins, especially the compounds 6, 9, 17, 20, and 21 together with pyruvic and vinyl derivatives (7 and 16, respectively). Having λmax > 530 nm [31], the relative predominance of the compounds 4, 10, 11, 12, and 19 could partially explain the slightly higher CI in FA18 than SCE wines (Table 1). *Molecules* **2021**, *26*, x FOR PEER REVIEW 8 of 19

**Figure 1.** PCA—Primitivo. Principal component diagram of anthocynin‐derived red pigments in Primitivo wines SCE138 (green point), SCE 16 (blue point), FA18 (red point), and distribution (percentage) calculated (**a**) at draining off and (**b**) after 18 months of bottle aging; in (**c**), we report the percentage of each pigment as measured by HLPC assays both at draining off and after 18 **Figure 1.** PCA—Primitivo. Principal component diagram of anthocynin-derived red pigments in Primitivo wines SCE138 (green point), SCE 16 (blue point), FA18 (red point), and distribution (percentage) calculated (**a**) at draining off and (**b**) after 18 months of bottle aging; in (**c**), we report the percentage of each pigment as measured by HLPC assays both at draining off and after 18 months of bottle aging. Variables correspond to peaks reported in Table 2.

months of bottle aging. Variables correspond to peaks reported in Table 2. Moreover, Negramaro FA18 wines at draining off (Figure 2a,c) were distinguished for having a higher content of stable pigments 4, 10, and 23, which positively affected their color intensity (Table 1). Whereas, SCE wines were separated on the sore plot since Moreover, Negramaro FA18 wines at draining off (Figure 2a,c) were distinguished for having a higher content of stable pigments 4, 10, and 23, which positively affected their color intensity (Table 1). Whereas, SCE wines were separated on the sore plot since more correlated to the free anthocyanins 3, 5, 6, 9, 14, 15, 17, and 21 showing greater factor loadings (>|0.9|) on PC1 and PC2 (Figure 2a).

more correlated to the free anthocyanins 3, 5, 6, 9, 14, 15, 17, and 21 showing greater fac‐ tor loadings (>|0.9|) on PC1 and PC2 (Figure 2a). Finally, with regards to Aleatico, even though the use of *S. bacillaris* in winemaking partially enhanced the formation of stable conjugated forms (especially vitisin A 10 and compound 23) in wines at draining off (Figure 3a,c), this was not enough to intensify and

Finally, with regards to Aleatico, even though the use of *S. bacillaris* in winemaking partially enhanced the formation of stable conjugated forms (especially vitisin A 10 and

stabilize the color. Indeed, there was no significant variation of CI, H, and CEI among the three wines (Table 1). Furthermore, SCE 18 month‐old wines were less clearly separated from FA18 and their relative percentage of pigment families was very close (Figure 3b).

stabilize the color. Indeed, there was no significant variation of CI, H, and CEI among the three wines (Table 1). Furthermore, SCE 18 month-old wines were less clearly separated from FA18 and their relative percentage of pigment families was very close (Figure 3b).

These findings, coupled with the highest H and CEI values in the aged samples (Table 1), indicated a similar and faster color change from red to orange tone and color loss [4]. A possible explanation for this behavior can be attributed to the very low content of anthocyanins (TA) and polyphenols (TP) extracted from grapes in Aleatico wines during both fermentation types. *Molecules* **2021**, *26*, x FOR PEER REVIEW 9 of 19

**Figure 2.** PCA—Negramaro. Principal component diagram of anthocynin‐derived red pigments in Negramaro wines SCE138 (green point), SCE 16 (blue point), FA18 (red point), and distribution (percentage) calculated (**a**) at draining off and (**b**) after 18 months of bottle aging; in (**c**), we report the percentage of each pigment as measured by HLPC assays both at draining off and after 18 **Figure 2.** PCA—Negramaro. Principal component diagram of anthocynin-derived red pigments in Negramaro wines SCE138 (green point), SCE 16 (blue point), FA18 (red point), and distribution (percentage) calculated (**a**) at draining off and (**b**) after 18 months of bottle aging; in (**c**), we report the percentage of each pigment as measured by HLPC assays both at draining off and after 18 months of bottle aging. Variables correspond to peaks reported in Table 2.

months of bottle aging. Variables correspond to peaks reported in Table 2.

**Figure 3.** PCA—Aleatico. Principal component diagram of anthocynin‐derived red pigments in Aleatico wines SCE138 (green point), SCE 16 (blue point), FA18 (red point), and distribution (per‐ centage) calculated (**a**) at draining off and (**b**) after 18 months of bottle aging; in (**c**), we report the percentage of each pigment as measured by HLPC assays both at draining off and after 18 months **Figure 3.** PCA—Aleatico. Principal component diagram of anthocynin-derived red pigments in Aleatico wines SCE138 (green point), SCE 16 (blue point), FA18 (red point), and distribution (percentage) calculated (**a**) at draining off and (**b**) after 18 months of bottle aging; in (**c**), we report the percentage of each pigment as measured by HLPC assays both at draining off and after 18 months of bottle aging. Variables correspond to peaks reported in Table 2.

#### **3. Discussion**

**3. Discussion**

of bottle aging. Variables correspond to peaks reported in Table 2.

These findings, coupled with the highest H and CEI values in the aged samples (Table 1), indicated a similar and faster color change from red to orange tone and color loss [4]. A possible explanation for this behavior can be attributed to the very low content of anthocyanins (TA) and polyphenols (TP) extracted from grapes in Aleatico wines during both fermentation types. Wine is the result of a complex biochemical process, that starts with grape harvest‐ ing, continues with the alcoholic and malolactic fermentations, wine aging, and bottling [40]. In this process, the diversity and composition of the yeast micro‐population may Wine is the result of a complex biochemical process, that starts with grape harvesting, continues with the alcoholic and malolactic fermentations, wine aging, and bottling [40]. In this process, the diversity and composition of the yeast micro-population may significantly contribute to the organoleptic characteristics of wine, and consequently, those known as terroir. Indeed, modern oenology is increasingly oriented today to the development of technologies and strategies that allow enhancing the typicity and the quality of autochthonous vines. In this regard, one of the most promising ways is the identification of yeasts which are used as a starter in innovative winemaking processes and allow improving the quality of wines. A combination of *S. bacillaris* and *S. cerevisiae* in a sequential fermentation has been described promising to satisfy the modern market and consumer preferences due to its peculiar characteristics [18].

significantly contribute to the organoleptic characteristics of wine, and consequently, those known as terroir. Indeed, modern oenology is increasingly oriented today to the

quality of autochthonous vines. In this regard, one of the most promising ways is the identification of yeasts which are used as a starter in innovative winemaking processes and allow improving the quality of wines. A combination of *S. bacillaris* and *S. cerevisiae*

In the present paper, we investigated how mixed fermentation combining the use of *S. bacillaris* with *S. cerevisiae* might influence the color and its stability during aging, one of the most important organoleptic characteristics in red wine, on three of the most typical and commercially important wines in the South of Italy, Primitivo, Negramaro, and Aleatico. In our trials, we first demonstrated that no killer effect exists of the *S. bacillaris* strain FA18 against the chosen *S. cerevisiae* strains (SCE16 and SCE138), thus confirming their suitability in mixed fermentation. Moreover, the kinetics of fermentation and chemical analysis demonstrated that the two *S. cerevisiae* strains have a similar fermentation capacity on all the three cultivars, thus confirming their suitability of combination in *S. bacillaris.*

Our results revealed that mixed fermentation influences both basic parameters and chemical compounds (i.e., pyranoanathocyanins) specifically related to the co-pigments formation and color stabilization. *S. bacillaris* has been described to affect the chemical composition of the musts and wines by producing various metabolites of enological interest [18].

Among these effects, the reduction of ethanol levels in wines has been described when *S. bacillaris* was used coupling with plus *S. cerevisiae* [16,18,25,26]. However, we did not find any variation in the alcoholic degree (% *v*/*v*) (Table 1). Indeed, no significant differences in the ethanol production have been described between mono-*Saccharomyces* and mixed fermentations with some *S. bacillaris* strains. On the contrary, the significant reduction in ethanol is shown when *S. cerevisiae* is added 48 to 72 h after the *S. bacillaris* inoculation, and oxygen is applied during the fermentation process in order to favor the respiration rather than fermentation [16]. Furthermore, the reduction in ethanol for the sequential fermentation is emphasized when the fermentation occurs in a synthetic must medium rather than the natural grape must [41].

Moreover, our data revealed that the fermentation type significantly affected (*p* < 0.05) MA, TA, and TP in the analyzed wines. Pure fermentations allowed a better extraction of anthocyanins and polyphenols as demonstrated by the significant higher value of MA, TA, and TP in SCE16 and SCE138 than in the FA18 samples, especially for Negramaro and Aleatico (Table 1). Despite the aforementioned non-variation of ethanol in our samples, it is known that mixed fermentation of *S. bacillaris* and *S. cerevisiae* leads to a slower development of ethanol in the early stages of winemaking [25,26], thus reducing the extraction of phenolic compounds during the skin maceration [42]. This could partially explain the reduction in phenolic compounds we observed in FA18. Moreover, we detected an even more evident decrease in MA and TA, as well as in TP during aging which in fact is due to the precipitation and degradation phenomena (both oxidative and reductive), that can involve the less stable and oxidizable forms of red wine (such as cyanidin-3-*O*-glucoside and peonidin-3-*O*-glucoside) already described in literature data [7,43].

Most relevant, substantial differences emerged among our wines considering several compounds playing a critical role in the wine color. Indeed, the evolution of wine color is influenced by a number of factors, such as the amount of tannin and acids, grape variety, alcohol and acetaldehyde concentrations, as well as the winemaking and storage conditions of wine [42,44,45]. In particular, the color changes during wine maturation are usually attributed to anthocyanin polymerization reactions and the evolution of copigments resulting from interactions between anthocyanins and other compounds at the fermentation phase and during aging [34,35].

Overall, our data highlighted that a 48 h sequential fermentation employing the FA18 *S. bacillaris* in Primitivo, Negramaro, and Aleatico enhances the synthesis of stable anthocyanin pigments, in particular, A-type vitisins and ethylidene-bridge flavonol-anthocyanin adducts, as well as their preservation after 18 months of aging in the bottle. The acidogenic nature of *S. bacillaris*, leading to a more consistent production of pyruvic and acetaldehyde during fermentation, would be responsible for the preferential synthesis of these compounds [25,46]. It is worth pointing out that Primitivo, Negramaro, and Aleatico grapes, used in winemaking, derived from minimal or no canopy management grown vineyards and, thus, were poorer in anthocyanins and polyphenols with respect to conventional

conditions, as previously reported in literature [47]. This could motivate the lack of various pigments, such as pinotins, anthocyanin dimers, and trimers, as well as more different vinyl-linked and ethylidene-bridged compounds, compared to wines analyzed by direct injection [33] or after fractionation [7].

Notably, the pyranic structure of malvidin-3-*O*-glucoside pyruvate (10) is recognized as more resistant to the bleaching effect due to SO<sup>2</sup> than malvidinic free anthocyanins, thereby its presence in wine implies a greater red color stabilization [48]. Furthermore, this vitisin A is resistant to a pH increase [48] and oxidative degradation [49], as well as temperature changes [50]. It is worth noting that, although free anthocyanins more strongly decreased in FA18, mixed fermentation seemed to protect the wine from further non-oxidative degradation reactions. It was confirmed by the relative unstable ethyl linked anthocyanins (11, 12, and 19), whose percentage slightly increased during aging (Figure 1b,c), and the reduced color loss, as proved by the significant lower values of H and CEI than those found in SCE16 and SCE138 after 18 months in the bottle (Table 1). This would be a very important finding from a technological standpoint, since the use of *S. bacillaris* in tandem with *S. cerevisiae* could contribute to mitigate the often-reported rapid change of Primitivo color into orange hue compared to other international wines [7]. In addition, the significant lower values of H and CEI highlighted that Negramaro derived from the mixed inoculum of *S. bacillaris/S. cerevisiae* remained more stable in the color after bottle storage than SCEs (Table 1). This was corroborated by the most pronounced increase in vitisins, ethylidene-bridged pigments, and flavanol-anthocyanin adducts percentage in FA18 aged wines (Figure 2b,c). However, the remarked difference in the color stability of Negramaro wines was less evident respect to Primitivo ones (Table 1), maybe due to the different ethyl linked compounds prevailing in the former (i.e., malvidin-3-*O*-glucoside-8-ethylepicatechin) despite the latter samples (i.e., malvidin-3-*O*-glucoside-8-ethyl-epicatechin isomers and peonidin-3-*O*-pcoumaroyl-glucoside-8-ethyl-epicatechin), as well as their relative concentrations (Figure 2).

Notably, at our knowledge, this is the first evidence that mixed fermentation induced the production of ethylidene-bridge flavonol-anthocyanin adducts. Indeed, these adducts have been previously found unstable and intermediate products formed during winemaking and aging, also using different vinification procedures [51,52] or present only at a low concentration, in addition to their importance has been hypothesized [53]. As a matter of fact, these ethylidene linked pigments are associated to a color increase with a shift towards violet [54,55]. Moreover, these pigments undergo further polymerization phenomena, thus leading to an important reduction in astringency [37] that improve the organoleptic quality of the red wines.

#### **4. Materials and Methods**

#### *4.1. Yeast Strains*

Two *S. cerevisiae* strains and one *S. bacillaris* strain available at the I.U.V.V.—Institut Universitaire de la Vigne et du Vin Jules Guyot of Dijon (France) were inoculated in red vinification experiments. The two *S. cerevisiae* strains were isolated from 'Savigninin Jura' and were coded SCE16 and SCE138, while the *S. bacillaris* strain was isolated from 'Pinot noir' in Burgundy and identified as FA18. These strains were previously isolated in Burgundy, characterized, and then selected based on their oenological performances [24,56]. The 5.8S ITS rDNA sequencing confirmed the pure culture condition of these strains and the correct identity of these species [57].

#### *4.2. Grape Varieties and Vineyard Conduction*

The experiments were carried out in 2017 on three important Apulian *Vitis vinifera* L. red grape varieties: Primitivo, Negramaro, and Aleatico, chosen as used for the most important enological production in Apulia Region, Southern Italy. They were cultivated in an experimental vineyard of the CREA-VE, located in the area around Rutigliano (Bari), Apulia Region, Southern Italy. The vineyards are composed of 13-year-old vines, grafted34

E.M., trained on Gobelet Alberello, and pruned with four spurs of two buds. Plants are planted 1.5 m between rows and 1.0 m in the row. All the vines were cultivated without water supply, chemical inputs, and canopy management. Samples of 130 kg per each variety were hand-harvested at the same time in mid-October, at technical maturity [58]. At harvest, the total soluble solids (TSS) content, A, and pH were as follows: Aleatico: TSS 19.8 ◦Brix, A 5.9 g/L, pH 3.40; Negramaro: TSS 21 ◦Brix, A 6.9 g/L, pH 3.38; Primitivo: TSS 25 ◦Brix, A 7.1 g/L, pH 3.42. The grapes were hand-picked in small pierced plastic crates and immediately crushed and destemmed. After crushing and destemming, 4 g/hL of potassium metabisulphite (the equivalent of 20 mg/L of SO2) was added in the unpasteurized must. Organic and inorganic nitrogen sources were added, as described in the laboratory scale protocol of Nisiotou et al. [59]. The obtained must were directly processed for winemaking.

#### *4.3. Interaction between Saccharomyces Yeast Strains*

In order to test the killer action between the three yeast strains we performed two experiments.

Experiment 1: Cellophane agar layer technique [60]. Sterilized disc of 90 mm diameter of cellophane was laid aseptically over the solidified Yeast Peptone Dextrose Agar (YPDA) medium in culture plates. The plates were laid overnight to allow the excess moisture to evaporate. In addition, 10 <sup>µ</sup>L (at the concentrations of 1.0 <sup>×</sup> <sup>10</sup><sup>7</sup> CFU/mL) of each yeast strain (*S. cerevisiae* SCE16, *S. cerevisiae* SCE138, and *S. bacillaris* FA18) were uniformly distributed on the cellophane disc. For each yeast, six plates were produced. Moreover, 10 µL of sterilized YPDA without yeast were used as a control on nine different plates. After 48 h of incubation at 25 ◦C, the cellophane disc with and without yeast was removed from the plates. On the first three plates previously covered with the cellophane disc with the *S. cerevisiae* strain SCE16, 10 <sup>µ</sup>L (at the concentrations of 1.0 <sup>×</sup> <sup>10</sup><sup>7</sup> CFU/mL) of the *S. cerevisiae* strain SCE138 were uniformly distributed and on the other three plates, <sup>10</sup> <sup>µ</sup>L (at the concentrations of 1.0 <sup>×</sup> <sup>10</sup><sup>7</sup> CFU/mL) of the *S. bacillaris* strain FA18. The same procedure was used for the six plates covered with the cellophane agar with the saccharomyces strain SCE138 and for the six plates covered with the cellophane disc with the *S. bacillaris* strain FA18. On the plates used as a control, the three yeast strains were uniformly distributed. After incubation at 25 ◦C for 48 h, the growth of each yeast strain was recorded.

Experiment 2: Growth in liquid media. Each yeast strain was grown in a tube containing liquid YPD for 24 h at 25 ◦C. Cells were removed by a double centrifugation at 7240 g for 5 min and the supernatant was filtered through a syringe filter (0.22 µm pore size). In addition, 7.5 mL of the filtered supernatant of the *S. cerevisiae* strain SCE138 were placed in two sterilized tubes and added with 7.5 mL of liquid YPD containing the *S. cerevisiae* strain SCE16 or liquid YPD containing the *S. bacillaris* strain FA18, both at the concentration of 1.0 <sup>×</sup> <sup>10</sup><sup>6</sup> CFU/mL. The same procedure was followed using the filtered supernatant of SCE16 and FA18 added with liquid YPD containing living cells of other yeast strains. The tubes were placed on an orbital shaker at 25 ◦C for 48 h. After 18, 24, 42, and 48 h, two aliquots of 1 mL each were aseptically withdrawn from each tube. Using a spectrophotometer (Thermo Scientific NanoDrop2000) the growth of yeast cultures was monitored by measuring the optical density (OD) at 600 nm. For each aliquot, five replicates/lectures have been performed and the average values were used to plot the growth curve of each yeast strain in the presence of the filtered supernatant of another yeast strain.

#### *4.4. Pilot Scale Fermentation Procedure*

Pilot scale vinification trials of 20 kg (equal solid/liquid ratio in each trial) were conducted in stainless steel fermenters. The must obtained, corresponding to 18 Lt from each sample of single variety were fermented separately following a standard red winemaking procedure and three independent replicates for each trial were finally carried out.

Each trial was as follows: (i) Mono-SCE16 inoculation (SCE16), (ii) mono-SCE138 inoculation (SCE138), and (iii) a mixed fermentation where the two *S. cerevisiae* (SCE16 and SCE138, 1:1) were added together 48 h after the inoculation of *S. bacillaris* (FA18). Each yeast strain was inoculated at a starting concentration of about 5 <sup>×</sup> <sup>10</sup><sup>6</sup> CFU/mL. The possible lack of nutrients was avoided through a standard addition of nitrogen nutrients and enzymatic cofactors into the fermenting juice (20 g/hL of organic nitrogen). This was applied when sugar consumption reached 50 gr/Lt in the mono-SCE16 and -SCE138, while it was implemented in the mixed fermentations (FA18) after 48 h, when the two *S. cerevisiae* strains were inoculated, thus to enhance the *Saccharomyces* metabolic activities, avoiding nutrients depletion and preventing *Saccharomyces* growth arrest.

The fermentation proceeded at a constant temperature of 25 ± 0.5 ◦C, performing manual pushing down of the pomace cap three times a day during the first half of the fermentation and two times a day until the end. Fermentation kinetics were measured, checking the level of sugar consumption (◦Babo), utilizing a standard hydrometer. Macerations and fermentations were considered ended when residual sugar levels, measured with a hydrometer (Babo Klosterneuburg Mostimeter), reached 0 ◦Babo (8–10 days). The complete fermented must was pressed (up to 2–3 bar) and kept in the cellar for 2 days before storage in a refrigerated room (4–5 ◦C) to allow the residual solid parts (solid lees) to settle down. The wines were racked after a week to remove the solid lees. Consequently, wines were poured in 0.75 L glass bottles, supplemented with potassium metabisulphite to achieve a final concentration of 80 mg/L of total SO2. The wines were stored at a constant temperature of 15 ◦C and analyzed at draining off and after 18 months of aging to assess the color stabilization and the variation of chromatic characteristics.

#### *4.5. Chemical Analysis*

A chemical analysis on wine and must was performed according to the EEC regulation 2676/90, as reported by the International Organization for Vine and Wine (OIV, 2018: https://www.oiv.int/en/technical-standards-and-documents/methods-of-analysis/ compendium-of-international-methods-of-analysis-of-wines-and-musts-2-vol (accessed on 24 April 2020)). Titratable acidity, A (g/L of tartaric acid) was measured following OIV MA-AS313-01 R2015 par.5.3;pH: OIV MA-AS313-15 R2011; volatile acidity, VA (g/L of acetic acid): OIV MA-AS313-02 R2015; alcoholic degree, ET (% *v*/*v*): OIV MA-AS312-01A R2016 par. 4C. The wine color was assessed by the Glories chromatic parameters [61]: Color intensity (CI) was calculated as the sum of absorbance (λ<sup>420</sup> + λ<sup>520</sup> + λ<sup>620</sup> nm); hue (H) was defined as the ratio λ420/λ<sup>520</sup> nm, while the color evolution index (CEI) was calculated as (λ420−λ<sup>520</sup> nm)/λ<sup>420</sup> nm.

#### *4.6. Phenolic Indexes*

Total polyphenols (TP), total anthocyanins (TA), and monomeric anthocyanins (MA) were measured spectrophotometrically to assess the phenolic wine composition and the overall chromatic characteristics.

TP was determined following the method suggested by Waterhouse et al. [62]. From each sample, 20 µL were collected in separate cuvettes, and mixed with 1.58 mL water and 100 µL of Folin-Ciocalteu reagent. After 5 min, 300 µL Na2CO3 10% were added and the solution was shacked. The absorbance of each solution was read at λ<sup>750</sup> nm against a blank after waiting for 2 h at 20 ◦C. A calibration curve (R2 = 0.9264) was set with a polyphenolic concentration between 0–3000 mg/L of gallic acid, considering the effective range of the assay. Results were reported as mg/L of gallic acid equivalents (GAE).

TA was determined as already reported [38]. Briefly, the samples were diluted in a solution consisting of 70/30/1 (*v*/*v*/*v*) ethanol/water/HCl. The relative absorbance for each sample was measured at λmax of 540 nm. The total anthocyanin content was expressed as mg/L of malvidin-3-*O*-glucoside equivalents.

Finally, MA was measured by the spectrophotometric determination reported by Lee et al. [63]. Briefly, all the dilutions were performed in 50 mL volumetric flasks. At

the beginning, the appropriate dilution factor by diluting the test portion with a pH 1.0 buffer was determined until absorbance at λmax of 520 nm was within the linear range (between 0.2 and 1.4 AU). Using the appropriate dilution factor, two dilutions of each test sample, either for pH 1.0 (potassium chloride, 0.0025 M) or pH 4.5 (sodium acetate, 0.4 M) buffers were prepared. Hence, the determination proceeded through pH 1.0 and 4.5 buffer dilutions of the samples, reading them both at λmax of 520 and 700 nm. The measure at 700 nm was considered a wine haze correction of the reading at 520 nm. The content of anthocyanin pigments was expressed as mg/L of cyanidin-3-*O*-gluoside equivalents.

#### *4.7. Anthocyanin Profile Determined by HPLC-DAD-MS*

An HPLC 1100 equipped with a DAD and XCT-trap Plus mass detector (Agilent Technologies, Palo Alto, CA., USA) coupled with an ESI interface was used. The reversed stationary phase employed was a Zorbax C18 5 µm (250 × 4.6 mm i.d., Agilent Technologies) with a pre-column Gemini C18 5 µm (4 × 2 mm i.d., Phenomenex, Castel Maggiore, Bologna, Italy). The following gradient system was used with water/formic acid (90:10, *v*/*v*) (solvent A) and acetonitrile (solvent B): 0 min, 95% A— 5% B; 10 min, 87% A —13% B; 20 min, 85% A—15% B; 30 min, 78% A—22% B; 50 min 78% A—22% B; 55 min 5% A— 95% B; stop time at 70 min. Finally, the column was re-equilibrated with the initial solvent mixture for 15 min. The flow was maintained at 0.7 mL/min; the sample injection was 5 µL. Wine samples were filtered (0.2 µm RC syringe filters, Phenomenex) before the HPLC analysis. The diode array detection was between 250 and 650 nm, and absorbance was recorded at 520 nm. The positive electrospray mode was used for ionization of the molecules with capillary voltage at 4000 V and skimmer voltage at 30 V. The nebulizer pressure was 40 psi and the nitrogen flow rate was 9 L/min. The temperature of drying gas was 350 ◦C. The monitored mass range was from *m*/*z* 100 to 1200.

Free and co-pigmented anthocyanins were identified by matching the chromatographic elution order, molecular ions, and MS/MS fragments with those reported in the literature [7]. Semi-quantitation was performed using extracted ion chromatograms (EIC): For each compound, the EIC at the corresponding molecular ion was obtained and the relevant peak was integrated (Supplementary Table S2). Subsequently, peak areas were summed with respect to the type of pigment to calculate the percentage content of the different classes determined in the wines.

#### *4.8. Organic Acids Determination by HPLC-UV*

An HPLC 1100 equipped with a VWD detector (Agilent Technologies, Palo Alto, CA, USA) was used. The reversed stationary phase employed was a Synergy Hydro-RP-80A 5 µm (250 × 4.6 mm i.d., Phenomenex, Castel Maggiore, Bologna, Italy) with a pre-column Gemini C18 5 µm (4 × 2 mm i.d., Phenomenex, Castel Maggiore, Bologna, Italy). The separation was conducted in an isocratic mode using water/orthophosphoric acid (0.1%) as the mobile phase. The flow was maintained at 0.7 mL/min and sample injection was 5 µL. Wine samples were 2-folds diluted and filtered (0.2 µm RC syringe filters, Phenomenex) before the HPLC analysis. Absorbance was recorded at 210 nm.

#### *4.9. Statistical Analysis*

Data were analyzed using the R package software (version 3.4.0). Specifically, after testing their normal distribution by the Mardia test, a two-way multivariate analysis of variance (MANOVA) was performed on the chemical composition data in order to evaluate the effect of the factors fermentation type and aging, whose significance was discussed in the text. Tukey's HSD post-hoc test was used to separate the means (*p* < 0.05) when the interaction between the factors was significant (Table 1). Furthermore, the principal component analysis (PCA) of the dataset was performed on semi-quantified HPLC-anthocyanin profiles of each wine at draining off and after 18 months in the bottle to explore qualitative differences. In the PCA, only the first two components were considered accounting for more than 80% of the total variance explained.

#### **5. Conclusions**

In conclusion, the presented results highlighted that the use of *S. bacillaris* in tandem with *S. cerevisiae* has positively contributed to the evolution and stability of the wine color during the aging process. Although preliminary, our data are a further step that highlight the applicative technological potential of mixed fermentations with *S. bacillaris*. [18]. Indeed, our results support the importance of mixed fermentations to enhance the organoleptic characteristics (such as color intensity and stability) and shelf-life of wines that belong to the winemaking tradition. In particular, mixed fermentation with *S. bacillaris* might represent a valuable technological tool for mitigating the often reported rapid change of the color of some mono-varietal wines (such as Primitivo) towards an orange-brown hue. Moreover, we highlighted new clues on the impact of individual components produced in the presence of different starters on the final wine quality. This is a small pilot scale fermentation trial, but as a future perspective, the possibility of testing mixed cultures on different musts while also studying more in-depth yeast interactions, offer the opportunity to evaluate their benefits and limitations in order to select the best starters capable of fully enhancing the qualities of the resulting wines.

**Supplementary Materials:** The following are available online. Table S1: Concentration of organic acids in the studied wines; Table S2: Quantities of the identified anthocyanins into wines; Figure S1: Cell concentration of each yeast strain grown in liquid YPD amended with filtered supernatant obtained from the growth of each yeast strain on YPD. Data are the mean value of two replicates and five lectures/replicates. The time in hours is reported on abscises, while the ordinate axes reported the 10 logarithms of the number of cells per mL; Figure S2: Fermentation kinetics of Primitivo, Negroamaro, and Aleatico in pilot scale conditions: The days of fermentation are reported on abscises, while ◦Babo is reported on the ordinates. SCE16 and SCE138: Mono *S. cerevisiae* fermentations; FA18: Mixed fermentation of *S. bacillaris* FA18 and the two *S. cerevisiae* strains (48 h of delay). Red arrows indicate the addition of the two *S.cerevisiae* strains.

**Author Contributions:** Investigation, validation, methodology, software, data curation, writing original draft, M.V. (Matteo Velenosi) and P.C.; formal analysis, software, validation, R.P. and A.D.M.; formal analysis on the yeast used in the pilot scale fermentation, A.S.; validation, data curation, writing—original draft, H.A.; supervision and resources, reviewing, N.A.; supervision, resources, data curation, validation, writing—reviewing and editing, M.V. (Mario Ventura) and M.F.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding from any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The experiments were funded by local funds from the Consiglio per la ricerca in agricoltura e l'analisi dell'economia agraria-Centro di ricerca Viticoltura ed Enologia (CREA-VE) and the University of Bari Aldo Moro. Matteo Velenosi performed the experiments in the UMR ProcédésAlimentaires et Microbiologiques, EquipeVAlMiS (Vin, Aliment, Microbiologie, Stress), AgroSup Dijon—Université de Bourgogne thanks to GlobalDoc thesis grant.

**Informed Consent Statement:** Informed consent was obtained from all the subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples are available from the authors.

#### **References**


**Roberto Ambra \*, Gianni Pastore and Sabrina Lucchetti**

Council for Agricultural Research and Economics, Research Centre for Food and Nutrition, 00178 Rome, Italy; giovanni.pastore@crea.gov.it (G.P.); sabrina.lucchetti@crea.gov.it (S.L.) **\*** Correspondence: roberto.ambra@crea.gov.it; Tel.: +39-06-51594-570

**Abstract:** This review reports recent knowledge on the role of ingredients (barley, hop and yeasts), including genetic factors, on the final yield of phenolic compounds in beer, and how these molecules generally affect resulting beer attributes, focusing mainly on new attempts at the enrichment of beer phenols, with fruits or cereals other than barley. An entire section is dedicated to health-related effects, analyzing the degree up to which studies, investigating phenols-related health effects of beer, have appropriately considered the contribution of alcohol (pure or spirits) intake. For such purpose, we searched Scopus.com for any kind of experimental model (in vitro, animal, human observational or intervention) using beer and considering phenols. Overall, data reported so far support the existence of the somehow additive or synergistic effects of phenols and ethanol present in beer. However, findings are inconclusive and thus deserve further animal and human studies.

**Keywords:** beer; phenols; alcohol; health

**Citation:** Roberto Ambra, Gianni Pastore and Sabrina Lucchetti The Role of Bioactive Phenolic Compounds on the Impact of Beer on Health. *Molecules* **2021**, *26*, 486. https://doi.org/10.3390/ molecules26020486

Academic Editor: Nardini Mirella; Chemat Farid Received: 30 November 2020 Accepted: 15 January 2021 Published: 18 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Beer is a natural drink and historical evidences indicate a common use since ancient times also for medical and religious purposes [1]. Antique recipes prove widespread production back to 5000 years ago [2]. Beer is actually the most consumed alcoholic beverage in the EU and annual per capita consumption (L/year) has sharply increased in the Czech Republic (141 L), US (50–80 L) and France (33 L) [3]. Such a level of consumption has led some research to focus on the nutritional appropriateness of beer, merely considering health aspects like, for example, the intake of minerals [4] or the ability to prevent dysbiosis [5], properties also present in other beverages. Unfortunately, like wine, beer naturally contains ethanol, a well-known toxic and carcinogenic molecule [6].

Nonetheless, characteristic of beer is the high content in phenolic compounds, which are the focus of this review. The consumption of polyphenol-rich foods, like beer, is a well-accepted factor involved in the prevention of oxidative stress-associated diseases [7]. Traditionally, beer is obtained from as little as four basic ingredients: barley, hop, yeast and water. The first two ingredients naturally contain phenolics, however during beer production, these molecules undergo chemical modifications and new molecules are formed, influencing both the yield and final characteristics of a beer. Aroma, flavors, taste, astringency, body and fullness are the result of the metabolic activity of microbes on raw materials, and scientific evidences suggesting that they are all influenced by phenol content are summarized here. Moreover, this review focusses more deeply on most recent advances on the role of phenolic compounds on affecting human health status, considering how seriously researchers have tackled the effects of alcohol.

#### **2. Main and Minor Beer Phenols**

The polyphenolic composition of beers is considered as one of the quality indicators of beer processing and marketing [8]. In fact, the type and quantity of phenols influence taste, aroma and color, but also colloidal and foam stability, shortening beer's shelf-life

taste (see Section 4, "Phenols and beer attributes"). Several different groups of phenolic compounds have been reported in beer, the main ones being phenolic acids and tannins, and flavones and flavonols [9]. Because of its high concentration, also thanks to high producing yeasts (see Section 5, "The role of barley, yeast and hop genetics on beer phenols"), the simple phenolic alcohol tyrosol is one of the main phenols looked at in beer, present also in alcohol-free beers [10]. Concentration is so high in certain beers, reaching that of red wine [11], that authors have hypothesized that tyrosol could represent an indirect source, through biotransformation, of the more biologically active hydroxytyrosol [12] (see Section 6, "Phenols-related health effects of beer consumption"). In alcoholic beers, both phenols possibly protect yeast from the stress generated by high levels of ethanol, a phenomenon that has been demonstrated for wine's resveratrol [13], indicating that phenols not only undergo changes during brewing, but they also direct it. Accordingly, non-alcoholic beers normally have lower phenolic content [14], supporting the existence of a correlation between phenols and alcohol concentrations. Among minor phenols, those derived from barley, for example alkylresorcinols, are a group of phenolic lipids for which in vitro antioxidant and antigenotoxic [15] and in vivo diet-induced obesity-suppressing [16] activities have been reported. Even if contribution to alkylresorcinols dietary intake appears not significant, higher amounts were reported in stout beer [11]. Other quantitatively minor phenols derived from hop, for example, xanthohumol and other prenylated flavonoids, contribute significantly to beer flavor and aromas and have antibacterial, anti-inflammatory and antioxidant properties, and phytoestrogen activity [17,18]. Prenylflavonoids are of particular interest for beer as, on the one hand, no other food sources other than hop are known and, on the other hand, they are present regardless of the fermentation method, ale or lager, even if higher concentrations were found in stout and India Pale Ale styles [11].

Despite the fact that prenylated flavonoids can last for 10 years in beer stored at room temperature [19], monophenols and flavonoids show a temperature- and time-dependent decay in beer [20,21]. This phenomenon was initially studied using radioactive isotopes that revealed that almost 65% of molecules belonging to the tannin fraction go through oxidation [22]. Later, other evidences supported the role of oxidation in the time-dependent decay of phenols in beer, also demonstrating the role of the intrinsic haze-forming ability of some phenols [23]. Meanwhile, acetaldehyde was also involved in haze formation, because of its ability to polymerize polyphenols and compromise beer's flavanols level [24]. A resolutive approach to this problem could come from the implementation of dry-conservation. It was recently reported that production of microparticles from beer through high-temperature (up to 180 ◦C) spray-drying, used for the development of functional food with a specific heath objective, yielded a well-accepted beverage, in terms of appearance, taste and color, that kept, up to the entire period of dry-conservation (180 days), the initial amount of total phenols (measured using the Folin–Ciocalteu method) [25]. Even if no qualitative indication of phenols was reported, the study supports the validity of spray-drying in the production of non-alcoholic, high-phenols, beer-flavored beverages (see Section 9, "Phenols in non-alcoholic and isotonic beers").

#### **3. Phenols' Fate during Malting and Brewing**

As mentioned in the introduction, beer content in phenols depends on the type of barley and hops used for production. Even if hops contain a huge amount of phenols (up to 4% of dry matter) compared to barley (up to 0.1%), on average, four fifths of beer's phenols come from malt or other mashed cereals, because of their significantly higher starting amount [26]. Phenols undergo both quantitative and qualitative changes during seed germination and brewing processes [27] (Figure 1). The germination of barley seed, i.e., malting, has been studied deeply and is preceded by seed hydration (steeping), during which phenolic content decreases due to leaching, and followed by seed-drying (kilning), during which the improved crumbliness of the grain enhances the enzymatic release of bound phenolic acids. Kilning can be performed at different temperatures, for example in special malts brewing in order to bring desirable flavors and colors [28]. At temperatures

lower than 80 ◦C, kilning normally induces an increase in the amount of water-soluble total phenolic compounds [29], thanks to a Maillard-enzymatic release of phenols in the matrix [30] and to increased friability and extraction from the grain [31]. According to Leitao and colleagues, total phenolic content of barley (whose antioxidant contribution is mostly for ferulic and sinapic acids) increases four-fold during the transition to malt. Even if final yields depend on the malting procedures, the amount of phenolic compounds present in malt is inversely correlated with the degree of steeping and positively influenced by the germination temperature [32]. More recently, Koren and coworkers reported a 3- to 5-fold increase in the amount of total polyphenols during malting in six barley varieties, independently from the initial amounts [33]. tures lower than 80 °C, kilning normally induces an increase in the amount of water-soluble total phenolic compounds [29], thanks to a Maillard-enzymatic release of phenols in the matrix [30] and to increased friability and extraction from the grain [31]. According to Leitao and colleagues, total phenolic content of barley (whose antioxidant contribution is mostly for ferulic and sinapic acids) increases four-fold during the transition to malt. Even if final yields depend on the malting procedures, the amount of phenolic compounds present in malt is inversely correlated with the degree of steeping and positively influenced by the germination temperature [32]. More recently, Koren and coworkers reported a 3 to 5-fold increase in the amount of total polyphenols during malting in six barley varieties, independently from the initial amounts [33].

seed germination and brewing processes [27] (Figure 1). The germination of barley seed, i.e., malting, has been studied deeply and is preceded by seed hydration (steeping), during which phenolic content decreases due to leaching, and followed by seed-drying (kilning), during which the improved crumbliness of the grain enhances the enzymatic release of bound phenolic acids. Kilning can be performed at different temperatures, for example in special malts brewing in order to bring desirable flavors and colors [28]. At tempera-

*Molecules* **2021**, *26*, x FOR PEER REVIEW 3 of 25

**Figure 1.** Phenolic compounds' fate during the phases of malting and brewing processes: in the phase of mashing, after an initial decrease, total phenolics amount increases 3- to 5-fold; afterwards, phenolics continue to increase throughout mashing and during hop addition, but dramatically decrease during wort boiling, whirpool, fermentation, maturation, stabilization and filtration, so that, during the entire brewing process, about 60% of the malt phenolic content is lost. **Figure 1.** Phenolic compounds' fate during the phases of malting and brewing processes: in the phase of mashing, after an initial decrease, total phenolics amount increases 3- to 5-fold; afterwards, phenolics continue to increase throughout mashing and during hop addition, but dramatically decrease during wort boiling, whirpool, fermentation, maturation, stabilization and filtration, so that, during the entire brewing process, about 60% of the malt phenolic content is lost.

The amount of polyphenols reached in malt then significantly falls during brewing steps, depending on the protocol adopted, with a higher decrease for malt milled in wet conditions [34]. Enzymatic and non-enzymatic solubilization of phenols take place during the first step of mashing (hot hydration), and both are influenced by temperature and time, as well as the separation of wort, during which extraction of phenolic-rich spelt material occurs [9]. A successive increase of total phenolic compounds occurs in the wort separation (lautering) due to the extraction from spelt materials. Brewing is fundamentally ascribable to the metabolic activity of a fermentable carbohydrate source in the absence of oxygen, yielding alcohol and carbon dioxide. Fermentation is normally performed at fixed temperature but can be pushed at higher or lower temperatures. Hops, The amount of polyphenols reached in malt then significantly falls during brewing steps, depending on the protocol adopted, with a higher decrease for malt milled in wet conditions [34]. Enzymatic and non-enzymatic solubilization of phenols take place during the first step of mashing (hot hydration), and both are influenced by temperature and time, as well as the separation of wort, during which extraction of phenolic-rich spelt material occurs [9]. A successive increase of total phenolic compounds occurs in the wort separation (lautering) due to the extraction from spelt materials. Brewing is fundamentally ascribable to the metabolic activity of a fermentable carbohydrate source in the absence of oxygen, yielding alcohol and carbon dioxide. Fermentation is normally performed at fixed temperature but can be pushed at higher or lower temperatures. Hops, which were formerly included in the brewing process mainly for their preserving properties, are then added and wort boiling is started. Hops addition actually has several advantages, improving not only the bitter taste and astringency but giving protection to beer brewing yeasts, thanks to its antibacterial activity, against Gram-positive bacteria, and lowering pH to 4–4.2 [35]. During boiling, hop polyphenols are released and polymerization reactions with proteins occur, yielding precipitated complexes, responsible for the formation of chill

haze, that are then lost in the successive whirpool process and during the final filtration and stabilization. Final processes are critical for polyphenols and include fermentation, warm rest, chill-lagering filtration and clarification [36]. During brewing, around 60% of the malt phenolic content is lost. Decay affects all phenolic compounds, excepting *p*-hydroxybenzoic acid and sinapic acid, whose concentration increases by even four-fold [31]. However, different brewing processes can deeply influence total phenolic compounds, for example bock beers are normally three times richer than dealcoholized beer, with intermediate and decreasing quantities for abbey, ale, wheat, pilsner and lager beers [36]. Recent data also indicate that beer's content in phenols is associated with the production scale. In fact, the lesser characterized craft beers (unpasteurized and unfiltered) [37], whose production scale is limited by law in several countries (200,000 hL/year in Italy), exhibit higher total phenolic compounds' values compared to large-scale beers [38], mainly thanks to the lack of filtration. Finally, the phenolic content of beer is affected negatively by higher temperature pasteurization treatments [39].

#### **4. Phenols and Beer Attributes**

The ability of phenols to influence beer taste has been well known since the early 1960s, when the so-called "sunlight flavor" was ascribed mainly to humulone and lupulone addition after beer fermentation [40]. Phenols' ability to interfere with aroma, instead, was noticed around forty years ago, thanks to a *S. cerevisiae* "killer strain" producing a clove-like aroma [41]. Later, a study clarified that presence of the main phenolic flavors relies on yeasts capability to decarboxylate or reduce phenolic acids: 4-vinylguaiacol and 4-vinylphenol from *S. cerevisiae* and 4-ethylguaiacol and 4-ethylphenol from *Brettanomyces* sp. [42]. More recent data indicate that the ability of phenols to selectively characterize beer's flavors relies on their chemical transformations. For example, thermal decarboxylation of ferulic acid to 4-vinyl guaiacol, occurring during wort boiling and during fermentation, induces a three-orders-of-magnitude increase in its flavor threshold [43]. Unfortunately, some metabolic reactions have side effects, like that involving cinnamic acid and yielding the toxicologically relevant styrene [44]. Moreover, higher concentrations of monophenol can turn spicy or vanilla-like sweet flavor notes to unpleasant medicinal-like flavors [45]. A recent deep analysis of the association between metabolites and sensory characteristics using two-way orthogonal partial least squares indicates that isoferulic acid affects beer's fruity sensory attributes [46], suggesting the possibility to predict to some extent the formation of specific flavors.

With respect to aroma, phenols' protecting properties were found almost 25 years ago: phenols were found to prevent the formation of off-flavors, before and during malting, and the phenomenon was ascribed to their antioxidant activity in barley and malt [47]. More recently, some specific monophenols that confer the typical aroma of some popular beers were identified [48] and recently reviewed [49]. Worthy of interest are Czech beers whose distribution of individual phenolic compounds, that has been brought back to the origin of raw materials and the technology used for processing, is so unique that they have been proposed for authenticity analysis [50,51]. With respect to color, after high-affinity selective removing of tannins, Dadic and Van Gheluwe observed a severe discoloration of beer, demonstrating for the first time the correlation between phenols and beer color [52]. The involvement of monoflavanols' oxidation on beer color was further demonstrated by the recovery of oxidized molecules in polyethylene terephthalate bottle-stored beer [20]. More recently, several works have clearly demonstrated the relationship between phenols and beer color, both in small- and large-scale brewed beers [38].

Barley seeds' phenolic acids, flavonoids and proanthocyanidins influence quality indexes like viscosity, diastatic power and nitrogen content [53], and have an impact on beer turbidity [54], taste, bitterness and aroma [55]. With regard to hop, which was antiquely added in beer especially for its pleasant aroma and bitterness, brewing trials indicate that hop phenols can selectively reduce flavor deterioration during storage [56], specifically the sunstruck off-flavor that is formed in beer upon light exposure [57]. More

recent data indicate a temporal effect. In fact, later addition of hop, just before the end of wort boiling, significantly increases phenolic content [58]. Astringency, bitterness and fullness, which are affected by the boiling time [39], have been linked to different hop phenols fractions [59,60].

#### **5. The Role of Barley, Yeast and Hop Genetics on Beer Phenols**

The yield in phenols of a beer necessarily depends on the genetic background of its raw ingredients, and differences were reported in barley grain [61], hop [62] and yeast [63]. Unfortunately, domestication of barley and hop has reduced phenols' diversity. Nevertheless, total polyphenol content could be linked to specific quantitative trait loci in barley [64] and some specific combinations of phenols in barley can still be attributed to different genotypes. For example, the ratio between barley's main phenolic acids, ferulic acid and *p*-coumaric acid, is genetically determined and combinations can also influence key agronomic traits, such as hull adherence and grain color [65], through functionally related genes [53]. Studies combining genetics and environment on wild barley cultivars, that show a wider genetic diversity in agronomic traits and abiotic stress tolerance, identified some genes involved in phenol accumulation in barley seeds. Such studies are of special relevance as they can give a picture of the loss of genetic variation due to domestication and provide information for the set-up of breeding applications for phenols-related beer improvement. For example, a network analysis of gene expression and secondary metabolites, induced by the well-known stressor drought [66] in developing grains from several different Tibetan wild barley cultivars, recently allowed the identification of genes whose manipulation is believed to help the development of cultivars with specific contents of phenolic compounds [67]. Less data is available for a role of the genetic background on hop phenols. For example, a significant cultivar-dependent role has been recently reported for 2-phenylethyl glucoside [68], but the relevance on final quantities recoverable in beer is still lacking.

The ability of yeasts to adapt to different chemical (sugar, nitrogen) and physical (temperature, pH, oxygen, sulfur dioxide) properties resides in the great genetic diversity that has been exploited by the beer industry, i.e., for the development of strains with distinct flavor profiles. The production of different metabolites, like volatile phenols, is the direct consequence of human influence through wine and beer production. A first evidence testifying the role of the genetic background of yeasts in beer phenols came from the observation, at the beginning of the twentieth century, of volatile "ethereal substances" in English stock ales, during fermentation by Brettanomyces [69]. *Brettanomyces bruxellensis*, the first microorganism to be patented for beer production, was also involved in the spoilage of draught beer [70] and in the clove off-flavor (the ethylphenol 4-vinylguaiacol) [71] but, after being reported together with *Lactobacillus vini* as a contaminant in several ethanolproducing plants [72], was finally isolated from a number of fermented beverages and food, from cider to olives [73]. Spoilage depends on a still not fully identified gene pathway that involves two phenylacrylic acid decarboxylase (PAD) enzymes [74]. Ethylphenols production has been related to strain-dependent PAD amino acid sequence variability [75]. Thanks to their ability to convert ferulic acid to 4-vinylguaiacol, yeasts are believed to have a stronger impact on phenols than thermal processing steps [76]. Yeasts also have a fundamental impact in barrel beer ageing. Barrel-aged beers are sensorially enriched beers obtained by storage of already fermented beers in wood casks or by fermentation of beer's wort directly in wood barrels. Such processes mainly occur because of the spontaneous growth of microbes present in breweries' atmosphere and in barrels [77]. During this fermentative incubation, a bi-directional exchange of different molecules occurs from wood and beer: some beer's molecules are retained by the wood while others are released from wood to the beverage. *Dekkera bruxellensis*, another spoilage-related microbe in wine, is considered the main contributor to the aroma of aged beers, through its ability to convert hydroxycinnamic acids to volatile phenols, and has several advantages, from high ethanol yield to low pH tolerance [78]. Its spontaneous growth is accompanied by some enzymatic

activities that transform wort composition and yield the final chemical and sensory profiles of aged beer.

Aiming at finding optimal conditions for accelerating wort transformations, research is focused at finding optimal chemical conditions to produce beers with specific and preferred bacterial metabolites, normally avoiding those from non-Saccharomyces species, in multistarter cultures. For such purpose, Coelho and coworkers recently found that low glucose or high ethanol conditions favor the yield of *D. bruxellensis*-related metabolites over *S. cerevisiae* ones [79]. Ethanol-resistance and increased dominance towards other *S. cerevisiae* strains were also reported on mixed starter fermentations for the high polyphenols-producing *S. cerevisiae* var. *boulardii* strain [80]. A recent deep genomes/phenomes analysis involving 157 industrial *S. cerevisiae* strains [81] reported that production of 4-vinylguaiacol relies on specific genetic variants able to ferment maltotriose [81]. More recently, next-generation sequencing allowed the identification of a Brettanomyces strain void of phenolic off-flavors, limiting economic losses during production [82], a problem that was bypassed in *S. cerevisiae* by the selection of strains with inactivated alleles and/or functional copies [83]. Worth mentioning is a recent work that, seeking to explain different adaptive abilities, profiled microsatellite markers and ploidy-states of 1488 isolates coming from niches dispersed all over the world [84].

#### **6. Phenols-Related Health Effects of Beer Consumption**

While the serious damages of high alcohol intake are known, the effects of moderate consumption of alcoholic beverages are still a source of heated debate. Moderate beer consumption is believed to be associated with protective cardiovascular function and reduction in the development of neurodegenerative disease. Moreover, there is no evidence that moderate beer consumption can stimulate cancer. Nevertheless, alcohol consumption can become a problem for people at high risk of developing alcohol-related cancer or for those affected by cardiomyopathy, cardiac arrhythmia, depression, liver and pancreatic diseases, and is not recommended for children, adolescents, pregnant women and frail people at risk of alcoholism [85]. Anyway, beer, like wine, contains the already mentioned substances with indubitable protective capacities, not merely anti-inflammatory and antioxidant, as demonstrated by huge in vitro work on single substances [86]. However, the ambitious objective in studying the effects of beer consumption on human health is to analyze it in toto and, in order to understand the single contribution of phenols and alcohol, parallel experiments with similar doses of an equivalent non-alcoholic beer and of alcohol alone are essential. For example, Karatzi and coworkers [87] reported that both non-alcoholic and alcoholic beers improved some arterial biomarkers (reduced aortic stiffness and increased pulse pressure amplification), but the effects were also similar in a parallel vodka intervention, containing the same amount of ethanol as the alcoholic beer. However, as some other effects (wave reflections reduction) were higher in the alcoholic beer intervention compared to alcohol alone (vodka), and the endothelial function was significantly improved only after beer consumption, the authors concluded that the non-alcoholic and the alcoholic fractions of beer could have additive or synergistic effects [87].

We thus thought to analyze the fraction of similar publications that considered, in the search of the health effects of beer containing phenols, also the effects of the presence of alcohol. For such purpose, we used in Scopus.com the search string TITLE-ABS-KEY (beer AND (phenol OR polyphenol OR flavonoid) AND (observational OR administration OR consumption OR drinking OR prospective OR intervention OR crossover OR trial)) AND (LIMIT-TO (DOCTYPE, "ar")). The search was performed on October 2020 and returned 161 documents, including 31 reviews (even if they were already excluded by the string search), 7 not pertinent articles, 9 studies merely evaluating phenols' population intakes, 51 chemical-only reports (papers reporting chemical analyses of phenols of commercial or improved beers) and 22 reports using only single phenols in in vitro or in vivo models. For the remaining 41 (minus one not available even by the authors themselves [88]), experimental models, parameters tested and main findings are summarized in the next

section and sorted chronologically by the most recent, in Tables 1–3, about in vitro and animal models, human intervention and human observational, highlighting the use of alcohol alone (spirits, eventually vodka or gin), as well as non-alcoholic beer.

#### *6.1. In Vitro and Animal Experiments*

As demonstrated by in vitro cancer cell models (Table 1), several cancer types are sensitive to the antiproliferative action of some beer components, including ethanol. For example, epithelial cells' viability was reduced in a similar way by beer or an equivalent amount of ethanol [89]. Unfortunately, the authors did not test an alcohol-free beer. Using single molecules or a matrix containing all beer components, Machado and coworkers showed that phenols' activities are synergic [90]. Unfortunately, in this case, ethanol was not tested. Similarly, a total extract obtained from dark beer conferred higher protection to rat C6 glioma and human SH-SY5Y neuroblastoma cells against an oxidant stressor challenge (hydrogen peroxide) compared total extracts obtained from non-alcoholic and lager beers [91]. Again, neither the phenolic compounds of beers nor an alcoholic reconstituted extract were tested.




**Table 1.** *Cont.*

Wistar rats were used in several experiments with beer. One publication reported that both administration of alcoholic (4%) or lyophilized beer for 4 weeks had low, but statistically significant, beneficial effects on plasma lipidemic and antioxidant markers (total cholesterol, low-density lipoprotein (LDL) cholesterol, triglycerides and lipid peroxides), however alcohol alone was not tested and the authors themselves concluded that minimal effects observed could rely on relatively low alcoholic content of beer [100]. Next, using only a polyphenol-free beer, the same group concluded that lipid effects had to be ascribed to beer proteins, as long as effects were absent in rats fed with polyphenol-free wine [99]. In rats with skin incision-induced wound healing, feeding for 4 weeks with alcoholic beer prevented alcohol-induced markers of inflammation, oxidative stress and angiogenesis [97]. Notably, when beer was enriched with 10 mg of xanthohumol, effects were even more ameliorated. Similar results were obtained using animals with streptozotocin-induced diabetes [96]. On the same streptozotocin-induced diabetes model, hepatic glucolipid metabolism, lipogenic enzymes and glucose transporter 2 levels were tested after 5 weeks

of administration of xanthohumol-enriched alcoholic beer for 5 weeks [95]. Interestingly, beer prevented all the streptozotocin-induced liver catabolic state alterations tested (fibrosis, apoptosis, glycogen depletion, GLUT2 upregulation, lipogenesis reduction) and the effect was not observed in rats fed with normal beer. The authors also tested the effect of ethanol alone but, in none of these last three works were an alcohol-free beer, nor xanthohumol alone, tested, thus it is impossible to distinguish neither the effect of beer components nor of the polyphenol itself. Furthermore, in vitro and in vivo work on xanthohumol metabolites (isoxanthohumol and 8-prenylnaringenin) previously indicated opposite effects on angiogenesis and inflammation processes (pro-angiogenetic for 8-prenylnaringenin and anti-angiogenic and anti-inflammatory for the other two) [101]. Nevertheless, a xanthohumol-fortified alcoholic beer was used again to demonstrate attenuated pharmacologically induced pulmonary vascular remodeling and improved cardiac function [93]. Also, in this case, even if effects were absent in rats fed only with ethanol, no rats were tested with an alcohol-free beer. It is noteworthy that the authors could identify the involvement of extracellular signal-regulated kinase1/2, phosphatidylinositol 3-kinase/protein kinase B and VEGF receptor 2 in the protective properties of beer towards pulmonary arterial hypertension [93]. In a prepubertal rat model, beer with 10% alcohol significantly decreased, after 4 weeks, the levels of sex hormones, compared to ethanolor water-fed rats [92]. Again, even if authors concluded that beer inhibited the ethanolinduced increase of cleaved caspase-3 in Leydig cells, a non-alcoholic beer was not tested. In addition to the works recovered using the Scopus.com search string and mentioned in Table 1, worthy of mention are experiments showing that alcoholic-free beer can decrease the aminooxyacetic acid-induced GABA accumulation in hypertensive animals [98], and prevent brain inflammation and neurodegenerative effects induced by aluminum nitrate [94]. However, while as expected hops administration alone had a beer-overlapping positive effects to some extent, so did silicon administration, reinforcing the need for an appropriate set-up of experimental models.

#### *6.2. Role of Alcohol on Phenols' Metabolism and Beer Antioxidant and Anti-Inflammatory Properties, and on Cardiovascular-Related Effects*

Phenolic acids' absorption, previously reported both in low-alcohol [102] and alcoholic beer [103], is impaired by ethanol removal from beer [104]. The opposite effect of alcohol has been reported for tyrosol metabolization to hydroxytyrosol following beer consumption, as mentioned above. In particular, the administration of a single dose of 250 mL of blonde beer was associated to higher urinary recovery of tyrosol, whilst an identical dose of alcohol-free beer yielded higher urinary recovery of hydroxytyrosol [12]. However, as alcohol consumption proportionally increases hydroxytyrosol excretion through dopamine metabolism [105], hydroxytyrosol bioavailability is hardly attributable only to beer phenols.

Among first beer intervention studies (Table 2), there is an almost-perfectly set-up randomized acute administration of either 4.5% alcoholic beer (*n* = 14), or dealcoholized beer or 4.5% water solution of ethanol (*n* = 7), for the evaluation of the contribution of beer's alcohol [104]. Results demonstrated that a significant increase in plasma antioxidant capacity (TRAP) could be obtained only following alcoholic beer administration. Unfortunately, no crossover intervention was performed, and the effects were studied only in a temporally limited manner. In another similar, but a crossover, acute intervention of beer or wine (or vodka for the evaluation of the contribution of alcohol) inhibition of oxidative stress induced (by 100% normobaric O<sup>2</sup> breathing) was tested [106]. Analysis of stiffness 3 h after administration showed that only wine prevented oxygen-induced oxidative stress, possibly because of the higher content of polyphenols compared to beer i.e., 2.6 g/L vs. 0.4 g/L gallic acid equivalents (GAE) [106]. No one can say if such a low phenols amount in an equivalent alcohol-free beer could have produced the effects observed with wine. Daily supplementation of breastfeeding mothers (*n* = 30) with 660 mL of non-alcoholic beer was associated with an improvement of mothers' plasma and breastmilk antioxidant capacities, assessed 30 days postpartum, compared to control non-supplemented mothers [107]. For obvious reasons, an alcoholic beer was not tested. Administration of alcohol-free beer

(500 mL) for 45 days to postmenopausal women (*n* = 29) was associated with a reduction of several indicators of early protein oxidation, especially reducing cholesterol levels in subjects with higher than 240 mg/dL [108], supporting the usefulness of long-term alcohol-free beer consumption in fighting low-grade chronic inflammation and preventing metabolic disorders. As an alcoholic beer was not tested, one might speculate that alcohol can abolish the beneficial effect. However, previous work that used a crossover intervention trial (healthy drinkers, *n* = 27) to switch consumption of beers with similar phenolic content (310–330 mg/L) for 4 weeks, from low (0.9%) to high (4.9%) alcohol and vice versa, indicates that while the switch to low-alcohol did not change in vitro LDL oxidizability, the opposite switch did [109]. On the other hand, only non-alcoholic beer daily consumption for one week (17 healthy females, 330 mL) was associated to an increase in the urinary antioxidant capacity, as measured by Trolox equivalents [110], contradicting the results of the study reported at the beginning of this paragraph.

**Table 2.** Intervention studies (*n*, subjects' number; y, age (years)).



#### **Table 2.** *Cont.*


#### **Table 2.** *Cont.*

One observational study (1604 subjects of the IMMIDIET (Dietary Habit Profile in European Communities with Different Risk of Myocardial Infarction: the Impact of Migration as a Model of Gene-Environment Interaction) study, 26–65 years, see Table 3) supports a somewhat interfering property of alcohol on non-alcoholic components of beer. In fact, adjustment of beer intake for alcohol content broke the association between beer consumption and higher plasma and red blood cell omega 3 fatty acids [119]. In the overweight or class 1 obese healthy subjects, the daily consumption of alcoholic beer (but not of alcoholic-free beer with similar amount of total phenols) for four weeks raised HDL levels in subjects with low LDL-lipid profile and facilitated cholesterol efflux from macrophages, without affecting body mass index (BMI), liver and kidney functions, potentially reducing the risk of vessels occlusion by cholesterol deposition [120]. As the consumption of alcohol alone was not tested, it is not possible to exclude that the effects could be at least partially ascribable to alcohol. Similarly, in a crossover study of 28 daily healthy nonsmoking normotensive men consuming alcoholic beer (1125 mL; 41 g alcohol) for 4 weeks, an increase of the awake systolic blood pressure and the asleep heart rate was reported, however the effects were identical in men consuming red wine containing the same amount of alcohol [118], and an alcohol-free beer was not tested. Similarly, analysis of stiffness, 3 h after administration of alcoholic beer or vodka, showed that both protected against oxygen-induced increase in arterial stiffness, making the authors conclude that the observation was probably due to a central vasodilatatory effect of alcohol itself [106]. Again, Gorinstein and coworkers found that alcoholic beer consumption (330 mL daily, containing 510 mg of polyphenols and 20 g of alcohol for 30 days) ameliorated markers of coronary atherosclerosis of hypercholesterolemic in non-drinker males (*n* = 42, 43–71 years) during recovery from coronary bypass surgery [117]. Unfortunately, the control group of the randomized single-blind trial had only water "with minerals of beer", making it impossible to ascribe effects to either phenols or to alcohol. In a double-blind intervention of healthy male runners (*n* = 277), daily consumption of non-alcoholic beer, for 3 weeks before and 2 weeks after a marathon, reduced interleukin-6 immediately after the race, total blood leukocyte counts immediately and 24 h after the race and post-marathon incidence of upper respiratory tract illness [116]. However, like for breastfeeding mothers mentioned above, alcoholic beer was not tested, we guess for similar obvious reasons. Also, other observational studies (Table 3) suffer from this limitation. For example, a significant inverse association between beer consumption (and not for coffee, nuts, tea, olive oil and red or white wine) and hypertension was found by means of food frequency questionnaires submitted to 2044 adults [121], however neither the consumption of alcohol-free beer nor the contribution of pure alcoholic beverages were evaluated.

**Table 3.** Observational studies (*n*, subjects' number; y, age (years)).



#### **Table 3.** *Cont.*

FFQ, food frequency questionnaires; BCQ, beverage consumption questionnaires; ALMICROBHOL, Effects of Alcohol Consumption on Gut Microbiota Composition in Adults; TwinsUK, UK Adult Twin Registry; MEAL, Mediterranean healthy Eating, Ageing, and Lifestyle; CMHS, California Men's Health Study; IMMIDIET, Dietary Habit Profile in European Communities with Different Risk of Myocardial Infarction: the Impact of Migration as a Model of Gene-Environment Interaction; PPSWG, The Prospective Study of Women in Gothenburg; IWHS, Iowa Women's Health Study.

> An open, randomized, crossover, finely set-up controlled intervention trial of 33 highcardiovascular risk males drinking daily, for 4 weeks, a non-alcoholic beer (containing

1243 mg of total polyphenols) or an alcoholic beer (containing 1209 mg of total polyphenol and 30 g of ethanol), was repeatedly used (apparently with the same composition of subjects) by a group of Spanish researchers during the last 6 years, to investigate the possible synergistic effects of beer polyphenols and alcohol, using as control an administration of gin (containing 30 g of ethanol). Firstly, in an attempt to use urinary isoxanthohumol as a marker of beer consumption, a similar amount of the metabolite was recovered following non-alcoholic or alcoholic beer consumption, and no excretion was found following gin administration [115]. Notably, group differences in a female sub-population were found, but only an alcoholic beer was tested. Next, they looked for circulating endothelial progenitor cells (EPC) and reported that non-alcoholic beer consumption increased the number of circulating EPCs by 5 units, while in the alcoholic beer group, the increase was 8-fold. However, even if observations were not statistically significant, alcohol alone (gin) induced a 5-fold decrease in the number of circulating EPCs [114], suggesting the existence of some influencing, maybe genetic, factors. Then, they reported that only non-alcoholic beer consumption reduced leukocyte adhesion molecules and inflammatory biomarkers (decreased homocysteine and increased serum folic acid) [113], suggesting a possible antagonistic effect between alcohol and the non-alcoholic fraction of beer. Importantly, the alcoholic beer improved other plasma lipid and inflammation markers (high-density lipoprotein cholesterol, apolipoproteins A1 and A2, and adiponectin) and decreased fibrinogen and interleukin 5, but the effects were ascribed to alcohol as identical effects were observed following administration of a gin dose containing the same amount of alcohol (30 g). Finally, the group of Spanish researchers applied liquid chromatography-coupled Linear Trap Quadropole-Orbitrap mass spectrometry to discover the urinary metabolites produced in the intervention study. Increased urine excretion of hop α-acids and fermentation products were found following beer consumption with respect to the gin administration, but differences were slight and not completely reported [112].

#### *6.3. Role of Alcohol on Phenols-Related Effects of Beer on Cancer*

A case-control association study (over 14 years) of child acute lymphoblastic leukemia (*n* = 491 + 491) found an inverse relation with maternal moderate consumption (selfreported) of beer (and wine, but not spirits), making authors suggest a protective effect of flavonoids [130]. However, a positive relation was reported also for fathers, which is difficult to explain and minimizes the observation's reliability. In a similar matched case-control study of drinking/smoking habits (over 10 years) of leukoplakia patients (*n* = 187 + 187; 40–65 years), while a role of regular wine consumption was associated with a decreased probability of disease occurrence (compared to that of spirit drinking that was associated to increased risk), no significant effect for moderate beer drinking was found [126]. The authors concluded that weaker effects of beer were probably due to the different composition in substances synergistically or antagonistically, i.e., polyphenols, interacting with ethanol [126]. Nonetheless, using a 20-country wide one-year (2002) evaluation of alcoholic beverages consumption and total deaths for oral cancer, the same authors estimated a lower risk for beer (and wine) consumers compared to heavy alcohol consumption from spirits [128]. Similarly, the consumption of beer (nor liquor) could not be associated with prostate cancer risk, in a population case-control study taking into account the self-reported alcohol consumption (*n* = 753 + 703; 40–64 years), even if the same authors reported a reduced relative risk associated with increasing level of red wine consumption [129]. More recently, lack of association with prostate cancer was reported for beer consumption (but also for wine and liquor) in a bigger prospective study (*n* = 84,170; 45–69 years) [124].

#### *6.4. Role of Alcohol on Phenols-Related Effects of Beer on the Microbiota*

According to a relationship between microbiota, host genes and diet [131], recent work investigated the possibility that alcohol-free beer, acting at the level of gut microbiota, could prevent the metabolic syndrome (MS). In fact, occurrence of MS can be promoted by gut

microbiota dysbiosis, through low-grade inflammation and alteration of lipid metabolism. Gut microbiota dysbiosis can in turn be induced by alteration of the relative abundance of bacterial families [132]. Thus, a daily administration for one month of 355 mL of nonalcoholic beer was found associated to a decrease in fasting blood serum glucose and an increase in functional *β*-cells only, and the effect was not observed with an alcoholic beer containing a similar amount of phenolic compounds [111]. Moreover, the authors observed an enrichment of the microbiota diversity, also with the alcoholic beer, but only alcohol-free beer consumption was associated to a specific microbiota diversity with healthier function, suggesting that alcohol inhibited the positive effects of beer. As β-diversity was observed only after 30 days of treatment, the authors hypothesized that the effect on gut microbiota could depend on polyphenols and phenolic acids [111]. Similar results were obtained in an observational study, especially for higher butyric acid concentration in consumers versus non-consumers of beer [122], but no estimation of phenols intake was performed, nor were consumption of alcohol-free beer nor spirits-only drinkers recorded. On the other hand, another observational study on the microbiota of 916 UK female twins found association only for wine drinkers but not for beer (nor all other alcohols) [123], but also in this case, the consumption of alcohol-free beer was not considered. Figure 2 summarizes gut microbiota changes after beer consumption. *Molecules* **2021**, *26*, x FOR PEER REVIEW 15 of 25

**Figure 2.** Schematic representation of relationship between beer, gut microbiota and metabolic syndrome. Phenolic compounds contained in non-alcoholic beer have a positive effect on the microbiota dysbiosis, one of the main causes of metabolic syndrome, but the effect is prevented by alcohol presence. Non-alcoholic beer consumption also determines a positive modification of some parameters typical of metabolic syndrome such as glycemia and the β-cells' function (AB1, drinkers of 355 mL/day of alcoholic beer; NAB, drinkers of 355 mL/day of non-alcoholic beer [111]). On the other hand, moderate beer consumption can increase the production of butyric acid, a fundamental molecule produced by the microbiota and useful for its healthy implications (AB2, drinkers of 200–600 mL/day; AO, abstainers or occasional consumers of <1.5 alcohol g/day [122]). **Figure 2.** Schematic representation of relationship between beer, gut microbiota and metabolic syndrome. Phenolic compounds contained in non-alcoholic beer have a positive effect on the microbiota dysbiosis, one of the main causes of metabolic syndrome, but the effect is prevented by alcohol presence. Non-alcoholic beer consumption also determines a positive modification of some parameters typical of metabolic syndrome such as glycemia and the β-cells' function (AB1, drinkers of 355 mL/day of alcoholic beer; NAB, drinkers of 355 mL/day of non-alcoholic beer [111]). On the other hand, moderate beer consumption can increase the production of butyric acid, a fundamental molecule produced by the microbiota and useful for its healthy implications (AB2, drinkers of 200–600 mL/day; AO, abstainers or occasional consumers of <1.5 alcohol g/day [122]).

#### *6.5. Role of Alcohol on Other Phenols-Related Effects of Beer 6.5. Role of Alcohol on Other Phenols-Related Effects of Beer*

In an observational follow-up prospective study (34 years) of the association between alcoholic beverage consumption (using repeated surveys) and dementia (*n* = 1462 women, 38–60 years), beer consumption was associated to reduced dementia risk compared to subjects consuming only spirits [125]. Unfortunately, the consumption of alcohol-free beer was not taken into account. Moreover, in a case-control study of the association, in postmenopausal women (*n* = 35,816; 55–69 years), between specific self-reported drinking/smoking habits (over 20 years) and diabetes, a reduction of risk was observed for moderate consumption of either beer, red or white wines, but also for liquor, making the authors disprove the hypothesis that flavonoids could protect from diabetes onset [127]. In an observational follow-up prospective study (34 years) of the association between alcoholic beverage consumption (using repeated surveys) and dementia (*n* = 1462 women, 38–60 years), beer consumption was associated to reduced dementia risk compared to subjects consuming only spirits [125]. Unfortunately, the consumption of alcohol-free beer was not taken into account. Moreover, in a case-control study of the association, in postmenopausal women (*n* = 35,816; 55–69 years), between specific self-reported drinking/smoking habits (over 20 years) and diabetes, a reduction of risk was observed for moderate consumption of either beer, red or white wines, but also for liquor, making the authors disprove the hypothesis that flavonoids could protect from diabetes onset [127].

or bioactive compounds. Beer has high market opportunities because of an already high acceptancy of new organoleptic characteristics, due to widespread and previous diffusion of craft beers. Several ingredients have been added such as wheat, corn, rice and fruits. The phenolic profiles of several commercialized beers enriched with ingredients have already been reported and reviewed [8,133]. Studies agree that fruits' refermentation and maturation within beer production is associated to a significant increase of flavors and bioactive compounds supporting benefits of fruit contribution to beer's consumer acceptance. Both qualitative and quantitative increases in phenols have been reported in beers enriched with whole fruits during fermentation and works mainly focused on the role of the technological processes applied. However, rarely did a study report more than one fruit supplement. An exception is a recent report that compared individual phenols amounts in commercial beers enriched with cherry, raspberry, peach, apricot, grape, plum, orange or apple, and respective contribution to the antioxidant activity [134]. Im-

**7. Fruit-Based Enrichment of Beer Phenols** 

#### **7. Fruit-Based Enrichment of Beer Phenols**

Beer is considered a promising beverage in the context of functional foods, which are food items with, in theory, health benefits, due to the enrichment with specific ingredients or bioactive compounds. Beer has high market opportunities because of an already high acceptancy of new organoleptic characteristics, due to widespread and previous diffusion of craft beers. Several ingredients have been added such as wheat, corn, rice and fruits. The phenolic profiles of several commercialized beers enriched with ingredients have already been reported and reviewed [8,133]. Studies agree that fruits' refermentation and maturation within beer production is associated to a significant increase of flavors and bioactive compounds supporting benefits of fruit contribution to beer's consumer acceptance. Both qualitative and quantitative increases in phenols have been reported in beers enriched with whole fruits during fermentation and works mainly focused on the role of the technological processes applied. However, rarely did a study report more than one fruit supplement. An exception is a recent report that compared individual phenols amounts in commercial beers enriched with cherry, raspberry, peach, apricot, grape, plum, orange or apple, and respective contribution to the antioxidant activity [134]. Importantly, this work demonstrates that fruit beers may be enriched with bioactive compounds (catechin, rutin, myricetin, quercetin and resveratrol) that are undetectable in conventional beers at identical extraction conditions and indicates enrichment with peels to be very promising because of the highest amount polyphenols and flavonoids content and antioxidant activity. Notably, resveratrol was found in beers enriched with all fruits except one (plum), with the highest level being measured in grape beer [134].

Other recent beer-added ingredients are quince fruit, mango, sweet potato and olive leaves. Because of organoleptic characteristics, quince fruit is specifically appreciated as a processed food. Many studies have shown that quince fruit lends itself as an affordable and good source of phenolic acids and flavonoids; in particular, in vitro assays have shown that phenols are the main compounds responsible for fruit's hydrophilic antioxidant activity [135]. Quince fruit phenols have been extensively studied [136] and recent data indicate that the addition of different quince cultivars, with different sensory attributes or antioxidant content, can selectively modulate the final content in specific phenols and related sensory descriptors attributes [137]. The addition of quince increased the total polyphenol content, the total hydroxicinnamic acids, concentration of main volatile compounds related with fruity sensory descriptors, and led to higher intensities of floral and fruity sensory attributes [137]. The addition of mango fruit, naturally reach in phenols [138], yielded beers with higher polyphenol content and aroma than traditional beer, especially if the fruit was homogenized before addition, on the condition that no thermal treatment was performed [139].

The addition of dried flakes of sweet potato, naturally rich in phenols [140], before beer brewing increased both total phenols (about 10%) and flavonoids (about 20%) content without changing physicochemical and sensory parameters of beer, which also benefited from an important increase in *β*-carotene [141].

Also, dried olives or resulting extracts, that contain not only common phenols, but also the olive tree family-exclusive secoiridoids [142], were added to beer, and the resulting beer had positive flavor and aroma, but low colloidal stability and showed increasing haze formation during storage due to very high polyphenols content [143]. Similar increase in colloidal haze was reported also for beers with added omija fruits, questioning the validity of increasing the phenolic content of beers too much [144]. Beers enriched with lignans from wood chips or extracts displayed excessive bitter taste and unusual resin aroma, indicating the need for technological approaches to avoid significant changes to the characteristics of beer. A possible solution could come from the use of hot water as a unique solvent, already applied for the removal of resins from wood chips or lignan extracts from the knots of spruce trees (*Picea abies*), a strategy that yielded beer with as much as 100 mg/L of lignans.

#### **8. Cereal-Based Enrichment of Beer Phenols**

Apart from barley, other malted cereals have been used since antiquity for the development of fermented beverages, in a somehow geographical way, for example, rice in India [145], millet in Nigeria [146], sorghum in South Africa [147] and Corn in Mexico [148]. Regarding the latter, a pulque-fermented drink known as "Sendechó" was antiquely prepared by the Mazahuas population in the Valley of Mexico, using chili and pigmented corn varieties with high content of phenolic compounds, mainly anthocyanins [149], which are completely absent in barley beer. In the attempt of developing a beer with traditional ingredients, pulque was substituted with hop and brewer's yeast in an ale fermentation process performed with guajillo chili and blue corn malt. The result was a beer with total polyphenols concentration up to 560 mg GAE/L and of total anthocyanins up to 19.4 mg cyaniding-3-glucoside/L [150]. More recently, the same laboratory obtained blue or red corn malt blended beers with even higher total phenols amount (up to 849.5 mg GAE/L) and identified anthocyanins responsible for the final color yield of red and blue corn beers (pelargonidin-3-glucoside and cyanidin-3-glucoside) [151]. Authors also identified, in corn beers both previously reported and unreported, volatile phenols conferring desirable aromas to beers. Such results are promising with respect to previous reports of lower content of phenols in corn-added beers [152]. Nevertheless, to our knowledge, no consumer acceptability of such beers has been evaluated, and this aspect is crucial especially for the high content of phenols that can contribute to high spicy perception and for astringency of anthocyanins [151]. Among other gluten-free beers, those obtained from oat [153], sorghum [147], teff [154], millet [146], buckwheat [155] and quinoa [156] are in theory valid alternatives in terms of phenols considering the grain natural content, even if very little is available on the phenolic content of such beers, i.e., only a sum of aromatic alcohols was reported for millet [157]. A noteworthy emerging exception is represented by indigenous beer-like fermented beverages "ikigage" [158], "burukutu" and "pito" [159], drinks for which 4-vinylphenols quantities have been reported. Another exception is that of rice-based alcoholic beverages of Assam, India, with total polyphenol content up to 631.33 mg GAE/L [160].

One frequent issue of non-barley cereals is the low diastatic power, that traps phenols, making necessary the combination with other cereals or the addition of exogenous enzymes [161]. Addition to the mashing process of recombinant ferulic acid esterase [162] was recently proven as a valid remedy also for the low amount of the desirable phenol 4-vinylguaiacol (derived from ferulic acid by enzymic decarboxylation), a common issue of top-fermented wheat beer [163]. More recently, the strategy was further implemented by producing yeasts expressing bacterial ferulic acid decarboxylase [164].

#### **9. Phenols in Non-Alcoholic and Isotonic Beers**

Driving laws and a healthier lifestyle have increased the popularity of non-alcoholic beers. In order to not exceed the limit of 0.5% (*v*/*v*) alcohol or to produce beer with a limited alcohol content, two approaches are exploited. The first one consists in limiting the fermentation process, and hence the alcohol production, using low-alcohol yeasts or producing a wort with low degrees Plato and low diastatic power in order to obtain more dextrins than fermentable sugars. The alternative approach involves physical methods to remove the alcohol at the end of brewing, for example by vacuum evaporation or reverse osmosis treatments. Unfortunately, limiting the fermentation process can bring about inadequate conversion of wort to beer and, on the other hand, physical methods for alcohol removal can deteriorate beer composition [165]. Osmotic distillation using a membrane contactor was recently shown to be able to maintain the total phenols content in a lowalcohol top-fermented beer [166]. On the other hand, using the fermentation interruption approach, De Fusco and coworkers recently obtained a low-alcohol isotonic beer with an amount of total phenolic compounds similar to that of Pilsen beer and sport drinks [167]. Isotonic beers are an improvement on low-alcohol beers with similar rehydration potential of sports drinks [168] (beverages with specific osmolality and carbohydrate content [169])

and with the advantage of containing bioactive molecules. Notably, De Fusco and coworkers found that fermentation interruption did not significantly affect total phenols level [167]. Nevertheless, experiments are needed to test the shelf-life of low-alcohol isotonic beers and specifically to test if phenols' antimicrobial activity is adequate in such low-alcoholic and carbohydrate-containing beverages [170].

#### **10. Future Directions and Conclusions**

Here, we attempted to review the more recent findings on beer phenols and their role in human health. Particular attention was dedicated to the role of genetic factors and to the enrichment with phenolic compounds by cereals different from barley or fruits naturally rich in phenols. In this respect, it would be interesting to investigate to what extent fruit addition also increases the alcoholic content, which has health and consumer acceptance consequences that are not negligible. One other interesting question regards the huge amount of debris produced by beer production, especially in terms of phenolic compounds (1% in by-product spent grain [171]) that could be recycled for beer enrichment itself. Recent reports indicate that the recovery of phenols can be improved using the fungi *Rhizopus oligosporus* as a fermenting organism [172].

Even if the Scopus.com search string we used is arbitrary and may not entirely represent the research on health effects of beer ascribable to the presence of phenolic compounds, less than of 25% (40 out of 161) of entries we retrieved were reports on the in vivo (human or animal) or in vitro effects of phenolic compounds within in toto beer. Most of the research is focused on evaluating the effects of single phenolic compounds of beer, which, however, can give rise to partial conclusions that need further experiments performed in physiological conditions. For example, as beer contains a mean amount of xanthohumol around 0.2 mg/L, what is the rationale for supplementing volunteers with an enriched drink containing a daily dose of 12 mg of xanthohumol [173], corresponding to the amount found in 60 L of beer? Indeed, several techniques have been used, starting from almost 20 years ago and featuring a patented addition of an enriched hop product [174], in order to increase the amount of this compound in beer up to 10 mg/L [175–177]. However, even if studies with enriched beers are helpful for assessing the metabolic fate of phenolic compounds, in order to correctly evaluate healthy effects of beer consumption, researchers should consider, besides the side effects of alcohol (where ethically possible), also those possibly due to yet uncharacterized molecules, i.e., those resulting from the addition of the enriched hop product, and those due to a non-physiological consumption of a single flavonoid for which pro-apoptotic effects are already known [178]. In fact, the only health claims authorized for phenolics by the European Food Safety Authority regard, at the moment, olive oil hydroxytyrosol and cocoa flavanols, with high daily amounts (5 and 200 mg, respectively) that can, however, be easily consumed in the context of a balanced diet [179,180]. Among retrieved reports, only six investigated the effects of phenols in the presence and absence of alcohol, thus also considering the effects of alcohol alone. Actually, four publications belong to the same Spanish research group [112–115], and thus probably refer to the same and unique small population of 33 high-cardiovascular-risk males.

In conclusion, studies applying a parallel administration of non-alcoholic beer or/and alcohol alone, in both animal and human intervention studies, support the existence of somehow interfering effects of phenols and ethanol. However, in order to better highlight additive or synergistic effects, further correctly set-up human interventional crossover or observational, or at least animal, studies are required. From this point of view, even insect models could deserve more attention. In fact, using *D. melanogaster* fed with a total beer extract, Merinas-Amo and colleagues were able to demonstrate the synergic interaction between different molecules contained in beer [181].

**Author Contributions:** Scopus.com investigation, R.A.; writing—review and editing, R.A., G.P. and S.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

