Bioprotection in Winemaking

Abstract

Bioprotection in the wine sector is a strategy for protecting grape musts that have been used for a few years now. Bioprotection is intended to be a partial or total alternative to the use of sulfites. The principle of bioprotection consists in providing, from the harvest, on the grapes or on the grape must, yeast biomass, which, by its action, will limit the development of the native microbial flora and consequently avoid microbiological alterations at the early stages of the winemaking process. Most often, the biomasses studied are selected strains of non-Saccharomyces such as Torulaspora delbrueckii or Metschnikowia pulcherrima, but the Saccharomyces cerevisiae species can also be used. We propose to present the results of bioprotection used in white and red wine processes obtained in recent years and to underline the limits of this technique. Finally, a section will be devoted to describing proven or potential mechanisms that may explain how the biomass provided limits the development of native flora. Finally, the perspectives on the use of bioprotection in must and wine will be discussed.

1. Introduction

Bioprotection in food refers to the use of natural or controlled microorganisms, enzymes, or antimicrobial compounds to prevent or inhibit the growth of harmful bacteria, fungi, and other microorganisms in food products [1]. The primary goal of bioprotection is to enhance the safety and extend the shelf life of food without relying on synthetic chemical preservatives.

There are several methods of bioprotection in food. In the context of food, certain probiotic strains can inhibit the growth of pathogenic bacteria and spoilage microorganisms, thus contributing to food safety [1]. Bacteriocins are antimicrobial peptides or proteins produced by certain bacteria that can inhibit the growth of closely related bacteria or other pathogens [2]. Some of these natural compounds can be used as food preservatives to prevent the growth of spoilage and pathogenic bacteria. Certain strains of lactic acid bacteria and other microorganisms are used as protective cultures in food products like meat, fish, and dairy to outcompete harmful bacteria and create an acidic or competitive environment that prevents spoilage [2]. Enzymes can play a role in food preservation by breaking down certain components that promote microbial growth or spoilage. For example, enzymes like lysozyme can inhibit the growth of bacteria by disrupting their cell walls [3].

Bioprotection offers several advantages over traditional chemical preservatives. It involves the use of natural and generally recognized as safe (GRAS) microorganisms or compounds, making it more acceptable to consumers seeking clean label products. Additionally, it can have positive effects on the sensory and nutritional properties of food.

Bioprotection in wine refers to the use of natural, non-chemical methods to protect wine from undesirable spoilage microorganisms, primarily spoilage yeasts and bacteria. The aim of bioprotection is to preserve wine quality, prevent off-flavors, and reduce the need for the addition of chemical preservatives. The main reason why winemakers adopt bioprotection is that it reduces or eliminates the need for sulfites in the pre-fermentation phase.

The main factor that might contribute to the increased success of wine bioprotection is the consumer demand. With the growing awareness of wine safety and the desire for natural and clean-label products, consumers are seeking wine products with fewer synthetic chemical additives. Wine bioprotection offers a natural alternative to traditional chemical preservatives, making it a more appealing option for health-conscious consumers.

The main strategy for bioprotection is the use of selected non-Saccharomyces yeasts (NS yeasts). Non-Saccharomyces yeasts, such as Torulaspora delbrueckii, Metschnikowia pulcherrima and Lachancea thermotolerans, can be added before or alongside the primary fermentation with Saccharomyces cerevisiae (traditional yeast). These non-Saccharomyces yeasts contribute to flavor complexity and can inhibit spoilage microorganisms through competition for nutrients and space.

It is important to note that while bioprotection can be effective in managing microbial populations and preserving wine quality, they also require careful monitoring and control. The success of bioprotection methods may vary depending on factors such as grape variety, winemaking techniques and environmental conditions. Winemakers need to strike a balance between promoting natural fermentation and ensuring stability and consistency in the final product.

In this chapter, we will make a synthesis of the experiments carried out in bioprotection on white, rosé and red grape musts. In the second part, we will be interested in the mechanisms that can explain the protective effect, and finally, we will take stock of the future prospects.

2. Synthesis of bioprotection experiments

The process of bioprotection generally consists of adding one or more microorganisms in sufficient quantity on harvest or at the vatting. Thus, the high proportion of these microorganisms allows them to colonize quickly the medium preventing the development of spoilage yeasts and bacteria present in smaller quantities. Non-Saccharomyces yeasts are the most documented microorganisms today about their potential bioprotection properties.

The use of bioprotection in the wine industry dates back some 10 years, during which time experiments were carried out on an empirical basis. Indeed, the choice of yeast species, the dosage and the timing of addition are all parameters that have been defined without any scientific basis. Optimal bioprotection conditions then emerged from various experiments.

2.1 White must

One of the first experiments carried out and published in a popular journal reported the use of the Torulaspora delbrueckii species in the pre-fermentation phase on white grape juice. The Torulaspora delbrueckii species, naturally present but not in the majority in must, presents interesting characteristics such as low acetic acid production and cryotolerance. This species can therefore establish itself in musts placed at low temperatures (>5°C) without suffering significant population loss. On the other hand, temperatures >15°C are required to trigger alcoholic fermentation. For example, with inoculation of juice after pressing, without triggering alcoholic fermentation on grape solid [4] and using 5 g/hl of Torulaspora delbrueckii in white grape juice just after pressing at 5°C (Table 1), the authors reported a similar effect compared with the sulfite modality with a maximum Saccharomyces population of 1.4 log/ml, while on the nonsulfited modality, this population reached almost 3 log. Another study [5] demonstrated that Torulaspora delbrueckii was successful in limiting the development of spoilage microorganisms (Brettanomyces, acetic acid bacteria and lactic acid bacteria) in Aligoté must. In this experiment, Torulaspora delbrueckii was added at a dose of 5 g/hl after pressing, and microbiological counts were carried out after settling on sulfited or bioprotected batches. It is noteworthy that no significant difference in populations was observed between either modality.

Yeast species	Addition step	Dose	Temperature	Presence of a control	Grape variety	Sensorial analysis	Scale	Reference
Torulaspora delbrueckii	After pressing	5 g/hl (10^6 cells/ml)	5 °C	Yes	Sauvignon	No	Indust.	[4]
Torulaspora delbrueckii	After pressing	5 g/hl (10^6 cells/ml)	15°C	No	Aligoté	No	Indust.	[5]
Metschnikowia pulcherrima	Not precised	10 g/hl	15°C	Yes	Chardonnay	Yes	Lab. and pilot scale	[6]
Metschnikowia pulcherrima	Before skin maceration	20 g/hl	8°C	Yes	Sauvignon	No	Indust	[7]
Metschnikowia pulcherrima	half in the giraffe leading to the press and half in the tray under the press	5 g/hl	From 7 to 18°C	No	Chardonnay	Yes	Pilot scale	[8]
Metschnikowia pulcherrima	Harvest containers	10 g/hl	Non-specified	No	Pinot noir	No	Indust	[9]
Metschnikowia pulcherrima		20 g/hl	22°C	Yes	Muscat grape must + buffer	No	Lab. scale	[10]
Metschnikowia pulcherrima		10 g/l	10°C	Yes	Verdicchio	Yes	Indust	[11]

Table 1.

Synthesis of reported bioprotection experiments in white and rosé musts.

The two previous studies were carried out with the species Torulaspora delbrueckii, but bioprotection can be carried out with the species Metschnikowia pulcherrima. Some Metschnikowia pulcherrima strains have very low fermentation capacity, good implantation at low temperatures, genetic stability, and a high level of resistance. The use of Metschnikowia as a bioprotectant agent on Verdicchio during must clarification during two vintages have shown that without the use of M. pulcherrima, there was an increase of about one log CFU/mL, while the inoculated trial showed a containment of the wild yeast population, which remained almost constant [11]. During another vintage, a significant reduction of indigenous yeast population was shown in both trials with and without the inoculation of M. pulcherrima. The presence of Metschnikowia pulcherrima during cold maceration did not affect significantly the enological parameters. On the other hand, the inoculation of M. pulcherrima at the start of cold clarification generally led to wine with a different volatile profile. Indeed, a significant increase in ethyl hexanoate and monoterpene with relevant high Odor Activity Values in comparison with the control was observed [11]. The sensory analysis highlighted the influence of the use of M. pulcherrima in cold clarification on the sensory profile of aroma wine.

In another study, Gerbaux et al. [6] used this species in a bioprotection trial on Chardonnay; the microbial flora was not determined, the aim of the study is to look at the protection of wines against oxidation, as it was shown that Metschnikowia pulcherima consumed the oxygen in musts and thus prevented oxidation of musts in the absence of oxygen [6]. The sensorial analysis concluded a higher fruity intensity for the bioprotection modality despite the lack of statistical analysis.

The problem of must oxidation when bioprotection is used is just as important as limiting the development of spoilage flora. As we have seen, in white musts, under certain conditions, bioprotection protects against both microbial development and oxidation [5, 6]. A study carried out at a lab scale confirmed the oxygen consumption capacity of Metschnikowia pulcherrima [10]. In this study, it was shown in Muscat juice that the addition of Metschnikowia pulcherrima as a bioprotectant increases the total oxygen consumption capacity.

This problem also exists in rosé musts. In an experiment on rosé must, it was shown by carrying out colorimetric assays associated with chemical analyses of anthocyanins and phenolic compounds that the use of bioprotection alone did not protect the wine from oxidation [9]. An addition of oenological tannins on musts stabilized the color of bioprotected rosé wine in a similar way that SO₂ addition did. Quebracho tannins appeared more efficient than gall nut tannins.

Based on these different experiments, some discrepancies in the results regarding must oxidation could be observed. These differences could be linked to differences in the experimental procedure (temperature of must settling, bioprotectant dose, protection of must from oxygen, etc.). Indeed, controlling these parameters is of utmost importance for successful bioprotection.

Comparing six different maceration routes before alcoholic fermentation (AF) of a bioprotected must, varying the duration and temperature parameters, Simonin et al. [8] have shown that a temperature value ≤12°C was the main factor independent of the duration which allowed good implantation of the bioprotectant. An increase of the maceration duration at 12°C led to browning of the must, without significant effect on the final color of the wine, which was felt as more “floral,” with more length in the mouth [8].

Across all the trials, it seems that sulfiting remains the safest practice for significantly reducing the pre-fermentative microbial population. Bioprotection can contain the development of indigenous populations to a non-detrimental level, but on condition of guaranteeing colonization of the medium by the yeast selected for more than 90% of the total population [12]. Below this proportion, even if the selected yeast is mainly established, it does not prevent the development of an indigenous population at the same level as the sulfite-free modality. The use of non-Saccharomyces fermentative yeast or Saccharomyces cerevisiae is not recommended for white and Rosé vinification. The risk of triggering alcoholic fermentation during settling is too high. It is therefore advisable to use non-fermentative non-Saccharomyces yeasts such as Metschnikowia pulcherrima or Metschnikowia fructicola [12].

All of the tests published and summarized in Table 1 have made it possible to provide recommendations to winemakers wishing to use bioprotection as an alternative to sulfites and to show certain limits of bioprotection. However, many parameters need to be specified. Indeed, the impact of bioprotection on bacterial populations has been little studied, as has the sensory impact. Other species that could have interesting properties in bioprotection should also be tested. For example, Pichia kluyveri has been successfully tested in red must [13], what about white must?

The bioprotection strategy in white or rosé and red is quite different. Indeed, in white or rosé, bioprotection is used to limit the development of indigenous yeasts throughout the phase before settling. Since the settling is then carried out at low temperature, it helps to limit the development of native flora, including spoilage yeasts and bacteria. In red, the bioprotection intervenes early and as there is no settling, the bioprotectant remains in the tank throughout the fermentation and probably contributes more to the sensory profile of the wines than in white.

2.2 Red must

In red winemaking, cold pre-fermentative maceration is a widely used technique throughout the world. The choice of temperature and the dose of sulfite added strongly influence the microbial flora present. The temperature range generally used is 10-15°C. This preferential maceration is often used to improve the color and the fruity aromatic intensity. This increase in fruity character would be linked to the microbial flora present. However, this microbial flora consists of yeasts belonging to the species Hanseniaspora uvarum, which is known for its ability to produce ethyl acetate and acetic acid. Under these conditions, an alternative to sulfites to limit the development of spoilage yeasts and promote aromatic expression is the use of non-fermentative yeasts such as Metschnikowia pulcherrima [14].

The use of Metschnikowia pulcherrima as a bioprotectant agent at vatting on pinot noir at a dose of 5 g/hl (5.10⁵ cell/ml) has been shown to limit the development of spoilage microbiota. A slight increase of B. bruxellensis species was observed in all the wineries and the evolution of acetic bacteria depended on the experimental site.

It is important to note that bioprotection is not always as effective as sulfites in limiting the development of indigenous yeasts and that this depends, among other things, on the initial population level. Trials conducted on Merlot in 2017 and 2018 showed that bioprotection did not limit the growth of Hanseniaspora species compared to sulfites [15]. On the other hand, these authors demonstrate that the use of bioprotective non-Saccharomyces yeast limited the abundance of filamentous fungi that are systematically associated with a decline in grape must quality.

2.2.1 Effect of bioprotection on phenolic compounds, volatile composition and sensoriality

In this study, it was hypothesized that the replacement of sulfites by a bioprotection strain affects proanthocyanidin and anthocyanin levels and leads to a higher degree of polymerization. But bioprotection had no influence on the phenolic compounds protecting musts and wine from oxidation. Wines produced from bioprotected or sulfited musts had different metabolic signatures, probably reflecting the production of specific metabolites by M. pulcherrima or the presence of chemical adducts due to sulfites. The metabolomics approach carried out by FT-ICR-MS analyses revealed statistical discriminations contrary to analyses of volatile compounds and conventional oenological analyses [16]. Bioprotection with Metschnikowia pulcherrima has no impact on the volatile compounds of red wines, although and specific sensory differences were perceived according to the winery. This was probably due to the fact that the low-temperature value of grape must during pre-fermentative maceration (12°C) limits the growth of the Metschnikowia pulcherrima yeast added. The bioprotectant population was 100-fold lower than the Saccharomyces cerevisiae population present in both modalities. The impact of the bioprotection strain was therefore limited on the production of volatile compounds during alcoholic fermentation and, therefore, on the sensory analyses [16].

In a popular science magazine, a synthesis of different experimental assays is presented. These include experiments with Saccharomyces cerevisiae and Torulaspora delbrueckii. Of the six trials conducted on Syrah, Cabernet-Sauvigon, Grenache and Mourvèdre, half showed that the indigenous flora was present in quantities comparable to those of the sulfited modality and in lesser quantities than the non-sulfited control [17]. The same authors reported trials at an industrial scale. The use of Saccharomyces cerevisiae in bioprotection at 30 g of yeast/100 kg of grapes did not reduce the indigenous yeast population compared with the control. However, the sensory profile is more qualitative [18].

It should be underlined, that while Simonin et al. [16] reported the absence of differences in aromatic profile between the bioprotection modality and the control, Windholtz et al. [19] observed that the use of non-Saccharomyces yeasts as a bioprotection has a significant impact on the aromatic profile of wines. In a bioprotection trial with a Lachancea thermotolerans and Lactobacillus plantarum pairing on Tempranillo grapes, it was shown that anthocyanin composition was barely altered by the use of bioprotectors. From a sensory point of view, the wines were better evaluated, with fruitier, stewed, peppery, lactic and fresher aromas, bitter and astringent than the control [20].

The experiments summarized in Table 2 show contradictory results in terms of bioprotectant implantation, inhibition of indigenous flora development and sensory profile. However, recommendations can be made to encourage bioprotectant implantation and benefit from bioprotection from both a microbiological and sensory point of view. Bioprotectors should be added as early as possible, temperatures should be controlled (particularly during pre-fermentation), and the bioprotection yeast ratio should be as high as possible.

Yeast species	Addition step	Dose	Temperature	Volatile composition effect/sensorial effect	Grape variety	Scale	Reference
Metschnikowia pulcherrima	Vatting	5 g/hl	12°C	Yes/No	Pinot noir	Industrial	[16]
Torulaspora delbrueckii /Metschnikowia pulcherrima	On grapes at harvest and vatting	5 to 30 g/hl	10°C	No/Yes	Merlot	Industrial	[19]
Torulaspora delbrueckii/Metschnikowia pulcherrima	On grapes at harvest	5 g/hl	10°C	No/No	Merlot	Pilot	[15]
L. plantarum/L. thermotolerans	Non-specified	Non-specified	Non-specified	Yes/Yes	Tempranillo	Pilot	[20]
Torulaspora delbrueckii/Metschnikowia pulcherrima	On grapes at vating	5 g/hl	10°C	Yes/Yes	Merlot	Pilot	[21]

Table 2.

Synthesis of reported bioprotection experiments in red must.

One study focused specifically on the sensory aspect of sulfite-free wines, including wines produced using bioprotection (Table 2) [21]. Sensory evaluation was carried out after 2 years of bottling. The sensory profiles of Merlot wines without sulfites, with and without bioprotection, were very similar to and significantly different from those of wines with sulfites. Four descriptors (“Fresh blackcurrant”, “Cooked black cherries”, “Mint” and “Coolness”) were significantly more intense in wine without sulfites and/or wines with bioprotection treatment.

3. Wine bioprotection during aging

Wines can undergo microbial alterations during aging. The most studied alteration concerns the Brettanomyces bruxellensis yeast responsible for wine depreciation. To prevent the development of this spoilage yeast, we recommend sulfiting wines after malolactic fermentation (MLF). However, the resistance of these yeasts to sulfites is well-known [22]. It has been shown that the use of selected lactic acid bacteria to conduct malolactic fermentation could act as a bioprotector against Brettanomyces [23]. In the absence of stabilization after MLF, the lactic acid bacteria here play a protective role against Brettanomyces contamination. Sensitivity appears to be universal: when exposed to a given biomass of lactic acid bacteria, genetically distinct strains of Brettanomyces, which may exhibit distinct characteristics such as sulfite resistance, are inhibited in a comparable manner. One condition for this protection to be effective is a high level of Oenococcus oeni, which prevents contamination by Brettanomyces. Indeed, without this bioprotection, the yeast can develop within a few weeks and produce volatile phenols. This bioprotection effect is specific to lactic acid bacteria, at least to O. oeni. No protection is obtained by maintaining a living Saccharomyces cerevisiae flora after AF. Maintaining the bacterial flora after MLF is therefore a good way of combating Brettanomyces. This bioprotection phase is favored by good hygiene and fermentation control, which promote low initial volatile acidity [23].

4. How bioprotection works?

It remains difficult to answer this question today, as the fact that oenological conditions imply complex microbial ecosystems including yeasts, bacteria and mold. Hence, many interactions could be set up within native flora and also with microorganisms inoculated during the winemaking process. The efficacy of bioprotection is based on the hypothesis that the added bioprotector induces a negative interaction with the indigenous microbiota, although no mechanism has been clearly identified yet in field trials. The expected effects are of two kinds: either the metabolism of the bioprotector significantly slows down the growth of indigenous microorganisms, allowing them to predominate during pre-fermentation phases, or its metabolism induces the death of these microorganisms. The implied mechanisms could be indirect such as nutrient competition or production of antimicrobial compounds or direct such as cell-cell contact [24, 25], but the main mechanisms put forward to date to explain bioprotective effects appear to be indirect interactions.

4.1 Oxygen competition

At the beginning of the fermentation process, the dissolved oxygen in must is about 8 mg/L (at 20°C) and some practices, as crushing, pressing or pumping can add significative quantities of oxygen into the grape must [26, 27]. Taking as the reference the alcoholic fermentation yeast S. cerevisiae, Visser et al. had demonstrated a higher consumption of oxygen by NS yeasts [28]. Among the species recommended for bioprotection, M. pulcherrima remains one that has the most important oxygen needs [29]. Recent work by Windholtz et al. determined the rates of dissolved oxygen consumption (DOC) on grape juice in 47 strains belonging to six different genera and species [30]. The highest OCR values were obtained for M. pulcherrima strains, compared to the values obtained for strains of Saccharomyces genus, confirming the data of Quirós et al. [29]. In contrast, T. delbrueckii and L. thermotolerans strains showed consumption classified as intermediate, with a high heterogeneity of the values obtained within both genera and species. These measurements, which were also carried out on eight strains of Hanseniaspora uvarum, the species most commonly found on grape must, showed interestingly oxygen consumption rates comparable to those obtained in L. thermotolerans or T. delbrueckii. Under oenological conditions, the difference in oxygen consumption and fermentative capacity can be partly explained by the Crabtree effect. In the presence of a high concentration of sugars (as found in must), the respiratory metabolism of Crabtree-positive yeasts is inhibited from redirecting the central carbonaceous metabolism to a fermentative metabolism [31, 32]. The impact of this catabolic repression depends on the yeast species, which will lead to more or less important oxygen consumption. T. delbrueckii as L. thermotolerans are Crabtree-positive yeasts [33], while M. pulcherrima and H. uvarum are mainly reported as Crabtree-negative yeasts, with high respiration capacities [34, 35]. These data suggest that there may be strong competition for dissolved oxygen during the grape must bioprotection, particularly through the application of M. pulcherrima strains as bioprotectors. However, the results obtained by Windholtz et al. [30] and earlier by Visser et al. [28], demonstrating that strains of Hanseniaspora strains (sp. uvarum and valbyensis) were able to grow in anaerobic conditions, suggest that competition for oxygen to explain the effectiveness of bioprotection appears to highly depend on the biodiversity of the Hanseniaspora strains present on the grape must.

4.2 Nitrogen compounds competition

It is well known that the quantity and quality of nitrogen resources are essential for yeast anabolism including the production of compounds that are inextricably linked to wine quality, such as higher alcohols and their esters. In the case of non-Saccharomyces yeasts, literature data remain still scarce, and few data are available not only on their minimum nitrogen requirements for growth but also on the nitrogen resources preferentially consumed. Preferential amino acid uptakes are found to be highly variable between yeast species. But, bibliographic data comparison appears complex due to the wide disparities of experimental conditions including nitrogen composition of media and growth parameters (temperature, oxygenation, etc.). The work of Kemsawasd et al. established the specific N-source influence on growth (in synthetic medium; 25°C) among five wine-related yeast species [36]. The N-sources that more particularly improved performance parameters were arginine, asparagine, glutamine and isoleucine for T. delbrueckii, with mainly same N-source influence profiles as S. cerevisiae. These N-sources were alanine and asparagine or serine, for M. pulcherrima and L. thermotolerans, respectively. H. uvarum was characterized by the most different N-sources influence pattern, with alanine as a booster for its growth.

In comparison with the previous data and focusing on M. pulcherrima, the study of Gobert et al. realized that grape juice showed preferential assimilation of isoleucine, leucine, lysine, methionine, glutamine, and cysteine at 28°C and alanine, cysteine, glutamine, histidine, lysin and tryptophan at 20°C [37]. The preferential assimilation of lysine has been confirmed on a wider panel of M. pulcherrima strains, as the good assimilation of branched-chain amino acids, including leucine and isoleucine [38, 39, 40], hardly suggesting that this assimilation pattern is characteristic of this species.

Concerning H. uvarum, Roca-Mesa reported that leucine and isoleucine were among the highest nitrogen sources consumed in synthetic must, with an early depletion of those amino acids (less than 48 h at 22°C) [41].

The highlighting of common N-sources consumed preferentially by both Hanseniaspora strains and bioprotectors could result from competition phenomena to the predominance of the bioprotective strain over the indigenous microbiota of a grape must. Indeed, a comparison of assimilation kinetics on preferential amino acids would help to confirm this hypothesis. Much broader studies conducted specifically within the framework of possible interactions between genera and species generated specifically by the practice of bioprotection (taking into account nitrogen composition of the grape must, the initial indigenous population level with the biodiversity of the genera and species, the temperature applied and the duration of pre-fermentative stage) will implement the data needed to refine the knowledge of the mechanics behind these nutrients competitions. In addition, the regulatory pathways implied in nitrogen consumption are still unknown among NS yeasts, unlike S. cerevisiae where, depending on the concentration of assimilable nitrogen in the medium, two main systems are involved: Nitrogen Catabolic System (NCR) and regulation involving a specific sensor of the plasma membrane [42].

4.3 Antimicrobial compounds production

4.3.1 Killer toxins

The production of these antimicrobial proteins was described in T. delbruecki by Villalba et al. [43]. The toxin Tdkt, presenting a molecular mass of around 30 kDa, exhibits a broad spectrum against wine spoilage yeasts, more particularly on yeasts belonging to Brettanomyces and Hanseniaspora genera [44]. Tdkt possesses glucanase and chitinase activities, and its putative mode of action includes binding to β 1-6 glucan and chitin with sensitive cells, with a potential degradation of these polysaccharides, cell wall disruption and finally, cell death initially by necrosis, then by apoptosis [43]. Most recently, another smaller killer toxin (TK) with a molecular mass of 15 kDa was purified in T. delbrueckii [45].

Among the yeasts used for bioprotection, T. delbrueckii is not the only one with the potential to produce killer toxins. The work of Büyüksırıt-Bedir & Kuleaşan [46] have demonstrated the production of a killer toxin by one strain of M. pulcherrima strain isolated from grapevine. This toxin is a peptide with a molecular weight of 10.3 kDa. The amino-sequence analysis showed that a part of the sequence was similar to the KHR killer toxin of S. cerevisiae previously characterized by Goto et al. [47]. Antimicrobial peptides production (molecular mass around 10 kDa) has also been demonstrated in M. pulcherrima by the work of Hicks et al. with strains showing killer phenotypes against pathogenic bacteria [48].

If, as in the case of the Tdkt toxin where the mode of action involves specific enzymatic activities targeting the degradation of parietal compounds of undesirable yeasts [43], it cannot be ruled out that enzymatic activities highly expressed in M. pulcherrima may also be involved. In fact, this species is characterized by a production of a very diverse range of enzymes, more particularly lipase, chitinase and β-1,3-glucanase [49, 50, 51, 52, 53, 54], which may also be involved in killer phenotypes.

The known modes of action of killer toxins produced by yeast are summarized in Figure 1.

Figure 1.

Known toxin-killer mechanisms in yeasts. A: Cell wall damage by glucan hydrolysis through enzymatic activity of killer toxin or by synthesis inhibition of cell wall component. B: Plasmic membrane perturbations leading to iron and other metabolites leakage or inhibition of calcium uptakes. C: Cell-cycle perturbation by locking up cell between G1/S phase or inhibiting the completion of G1 phase. D: RNA cleavage especially for 18S and 25S RNA and tRNA (adapted from Mannazzu et al. [44]).

4.3.2 Production of pulcherriminic acid and pulcherrimin

Firstly characterized in the bacteria Bacillus [55, 56, 57], the pulcherriminic acid biosynthesis pathway implies two leucyl-tRNA that are cyclized in cyclo-(L-Leu-L-Leu). This intermediate compound is oxidized in pulcherriminic acid inside the cell and then excreted to the extracellular medium. Pulcherriminic acid chelates iron ion (Fe³⁺) by a non-enzymatic reaction to form the red pigment named pulcherrimin [58, 59] (Figure 2). In Metschnikowia pulcherrima, as well as in other pulcherrimin-producers Metschnikowia species [60], pulcherriminic acid appears to be the main antimicrobial compound. Its action is not direct as it is due to the chelation of iron (Fe³⁺), which becomes unavailable for yeasts whose growth requires this ionic form of iron.

Figure 2.

Metabolic pathway of pulcherrimin production in M. pulcherrima (from [56, 57]; metabolic pathways, sigma-Aldrich).

This mechanism could explain the inhibitory effect of M. pulcherrima observed on agar plate and in grape juice medium on spoilage yeasts such as apiculate yeast and Brettanomyces bruxellensis [61]. However, pulcherriminic acid production is dependent of the initial iron concentration in the medium [58]. Studies conducted in vitro on agar plates showed that the intensity of the pigmented halo around Metschnikowia, due to pulcherrimin production, increased with the iron concentration in the medium in contrast to the zone of inhibition [50, 62, 63, 64]. These results suggest that the chelation of iron by pulcherriminic acid, as an antagonistic mechanism toward target yeasts such as B. bruxellensis, can only be effective in grape musts containing a very limited concentration of ferric iron.

4.3.3 Antimicrobial effect of L-lactic acid?

This question may be asked with regard to the negative effects on B. bruxellensis observed during malolactic fermentation by Gerbaux et al. [23]. The study of intraspecific diversity among the species B. bruxellensis [64, 65] has enabled to clearly identify different genetic groups, which may explain the difference in sensitivity to sulfites observed within this species. The absence of L-lactate dehydrogenase in some strains of B. bruxellensis (genetic group 1 – data not published) could explain the inhibitory effect of L-lactate production during MLF on their growth. In the same vein, when applying L. thermotolerans for grape must acidification thanks to the production of L-lactate from sugars [66, 67], it cannot be ruled out that this yeast plays simultaneously a bioprotective role against B. bruxellensis.

5. Future perspectives

Numerous field trials have already demonstrated the effectiveness of bioprotection in oenological conditions [15, 19, 68]. However, there is still little information on mechanisms that bioprotection uses to limit the development of indigenous flora. In literature, numerous interactions have been identified in the oenological context [24, 25], but none have been studied in the bioprotection context, leaving many possibilities unexplored.

The main hypothesis on the bioprotection implantation and efficiency are based on nutrient competition between indigenous microorganisms and bioprotection, especially nitrogen resources. It has been shown that nitrogen is an essential resource for yeast’s growth. In the winemaking context, nitrogen deficit could lead to stuck or sluggish fermentation [69]. Nitrogen needs and mechanisms regulating nitrogen consumption are well-known for Saccharomyces cerevisiae [42], but still poorly understood for non-Saccharomyces (NS) yeasts.

In the last decades, more studies have started to investigate NS nitrogen requirement [36, 37, 38, 39, 41], but those studies investigate mainly yeast genera such as Lachancea, Torulaspora, or Metschnikowia corresponding mainly to NS genera used in winemaking including in bioprotection strategy. Those results give important information about the nitrogen requirement of important yeasts used in winemaking, but the temperature and medium used in those studies were not always close to pre-fermentative conditions. Nitrogen consumption of yeast is highly impacted by the initial nitrogen concentration and the temperature, in order to obtain a better characterization of the nitrogen needs of bioprotective yeasts, it is important to study their consumption on conditions closer to those of winemaking steps where bioprotection is used. However, the nitrogen requirements on NS yeasts associated with main microbial alteration (such as Hanseniaspora yeasts genera) are very limited [36, 41], and to determine if a possible competition could be involved in inhibitory mechanisms, it is also important to characterize the nitrogen needs of spoilage flora. It would also be important to monitor nitrogen consumption during the growth of bioprotective yeasts in coculture with target native flora. If some nitrogen resources appear to be preferentially consumed by bioprotective yeasts and spoilage microorganism, a supplementation in these resources could help to determine the involvement of this competition in the inhibitory effect.

The same reasoning could be applied to other nutrient found in must, such as vitamins and lipids. These latter are not produced by yeasts under anaerobic conditions, so yeasts are dependent on lipids naturally present on grapes and musts for their metabolism and membrane integrity. Some vitamin deficiencies have also been found to be highly deleterious to yeast growth [70].

At the beginning of the winemaking process, oxygen was found in the mud in significative concentration. Information about the respiro-fermentative mechanisms of NS yeast in enology is still scarce. Studying these mechanisms, as well as the oxygen requirement of M. pulcherrima under oenological conditions, could allow a better understanding of the conditions supporting its establishment, as well as the interactions between M. pulcherrima and the other yeasts. Indeed, it was previously shown that the presence of M. pulcherrima has a strong influence on the respiro-fermentative metabolism of S. cerevisiae. After 3 h of coculture, the presence of M. pulcherrima induced a strong repression of genes involved in the respiratory metabolism of S. cerevisiae, as well as an over expression of genes involved in the fermentative pathway [35]. Sadoudi et al. have also shown that the presence of M. pulcherrima in coculture with S. cerevisiae impacts the metabolism and expression level of genes implied in PDH-pathway in S. cerevisiae [71]. In order to better understand interactions set up by this species during its use in bioprotection, it could be interesting to apply this strategy for a fine analyze of the effects of M. pulcherrima on the metabolism of other NS yeasts, more particularly H. uvarum or B. bruxellensis.

Strong competition for some resources essential to the growth of indigenous flora could explain the observed bioprotective effect, but the low temperatures used in winemaking condition induce a limited nutrient competition. The implication of nutrient competition in bioprotective efficiency remains to be proven.

Other, less passive, interactions have also been reported in the literature, such as the production of antimicrobial compounds. Among them, we can found the pulcherriminic acid production by yeasts of the Metschnikowia genera, and in particular by M. pulcherrima species. Many experiments have already shown the antimicrobial efficiency of these compounds by its capability to chelate iron in the medium [48, 61, 62, 63, 72, 73]. However, these experiments were generally carried out on petri dishes or in vivo on fruit (apples, grapes and oranges). To date, there are only few data available on the effect of pulcherriminic acid in oenological conditions [61]. Teams have set up protocols to quantify pulcherriminic acid [58], making it possible to follow pulcherriminic acid production during M. pulcherrima growth, and possibly to see whether the pulcherriminic acid concentration of different M. pulcherrima strains correlates with cell death of sensitive flora in liquid medium. In addition, other teams have created mutant strains which exhibited a lower or higher pulcherriminic acid production, but their antimicrobial efficiency was investigated in fruits or on agar plate [62, 63]. Studying the antimicrobial efficiency of these mutants under oenological conditions (synthetic must or real grape must) would also be a promising way to determine the impact of these compounds on the bioprotective effect of this yeast.

Pulcherriminic acid, produced by the Metschnikowia genera, is not the only toxic compound produced by NS species used in bioprotection. Numerous NS yeasts have been reported in the literature to produce killer toxins, including Torulaspora delbrueckii and Metschnikowia pulcherrima, which are the main yeasts genera commercialized. Most of the studies were carried out on synthetic liquid medium or in agar plate, and few of them were conducted in must/grape juice or grapes [74, 75, 76, 77]. Furthermore, a lot of studies investigate the “killer phenotype” on plate but did not correlate the inhibition area to the secretion of peptidic-type compounds. It would be interesting to investigate the killer toxin production in marketed bioprotectors and to characterize those toxins as well as to obtain their sequences in order to improve protein databases.

In addition, Kemsawasd et al. [78] have shown that the inhibition found between L. thermotolerans and S. cerevisiae involves the production of a killer toxin, but that this efficiency is also due to direct contact between cells. Indeed, when growing in a membrane-separated fermenter, L. thermotolerans mortality is lower than when growing in coculture. Cell-cell interactions are still poorly understood in NS yeasts and need more investigation. A study has already referred to M. pulcherrima ability to form biofilms [54], thus demonstrating the potential for cells to aggregate with others. The ability to form cell adhesion would not necessarily mean that yeasts will be capable to form cell-cell adhesion with other yeast species. In order to study the direct interaction between species, it could be possible to create genetically modified fluorescent mutants of commercialized bioprotection strains in order to follow by fluorescent microscopy their possible aggregation with spoilage microorganisms in mixed culture. Impact of cell-cell contact interaction could also be studied by liquid culture in double-compartment fermenters with and without spatial separation [78]. Furthermore, some genes responsible for cell adhesion are already known in S. cerevisiae, it could be interesting to sequence more bioprotective strains and search for homologous genes to those involved in attachment in S. cerevisiae. Those genes could thereafter be studied by mutagenesis (targeted or un-targeted) in order to confirm their implication in cell adhesion.

Indirect interactions also include quorum sensing (QS), which is a way of communication mediated by the production of signal molecules that lead to phenotypic change at the population, not cell, level. This mode of communication has been extensively studied in bacteria, and signal molecules and phenotype changes are well known. However, the yeast Candida albicans has been found in the literature to have a quorum-sensing system capable of mediating certain reactions such as hyphae morphology by higher alcohol production [79, 80]. Quorum sensing molecules involved in QS of C. albicans have also been found produced by S. cerevisiae [81, 82, 83], but their role in a QS mechanisms in S. cerevisiae remain controversial [84, 85]. NS yeasts were also reported in literature as able to produce some of the higher alcohol involved in QS in C. albicans [38, 86]. However, the involvement of these compounds in QS still needs to be demonstrated. Identification and quantification of ARO genes, which are involved in the production of higher alcohols, have to be investigated in bioprotective yeasts by molecular biology. Furthermore, it is still important to correlate their production to possible inhibition against spoilage flora in must. In a larger scale, metabolomic analyses are increasingly used in order to explore yeasts interaction in must and wine [87, 88, 89, 90, 91]. Implementation and improvement of databases could be helpful in characterizing compounds involved in interactions between bioprotection and spoilage microorganisms.

Bioprotection was firstly investigated in order to reduce the SO₂ addition during pre-fermentative steps of winemaking by using NS yeasts to protect most against microbial spoilage. But another positive effect of the use of some bioprotective yeasts could also be pointed. For example, the use of L. thermotolerans leads to the acidification of the medium by sugar transformation into lactic acid. On preliminary results, this lactic acid production was reported in a French review for wine industry professionals as responsible for the potential inhibition of B. bruxellensis growth and so limits its possible production of off-flavors, but more investigations are needed [92]. Furthermore, only a fraction of the total SO₂ is active against microbial spoilage, and the proportion of each SO₂ form is influenced by must and wine pH [93]. The acidification of the medium led to a higher proportion of active SO₂ [93], and so to a lower use of SO₂ to obtain the necessary part of active SO₂. The use of L. thermotolerans could help to protect most against B. bruxellensis by lactic acid production as well as enhancing the aromatic profile of wines and reducing SO₂ addition. Yeasts could also influence wine color by metabolite production such as pyruvic acid or acetaldehyde which are able to interact with anthocyanins to produce more stable pigments [94, 95, 96, 97, 98]. More investigations are needed in order to determine the capability of marketed bioprotective strains to modify and stabilize wine color. Selection of yeast strains that are able to limit indigenous yeast proliferation as well as protect wine color, especially for sensitive matrix such as rosé wine, could help to replace SO₂ and limit its use during winemaking.

Bioprotection was used at the early steps of winemaking (harvest or after pressing), but microbial alteration could also occur after alcoholic fermentation, especially by yeasts B. bruxellensis, which could alter wine organoleptic properties after malolactic fermentation and wine aging. In order to avoid this problem, SO₂ was traditionally added, but another strategy could be investigated. Some lactic acid bacteria have been reported in literature as bacteriocins producers able to inhibit fungi proliferation [2]. Some lactic acid bacteria isolated from wines have also proven their ability to produce toxic compounds in vitro, but their efficiency have not yet been proven in oenological context [99]. The use of lactic acids bacteria strain producing bacteriocin could be helpful to achieve malolactic fermentation and protect wine against potential microbial spoilage. Furthermore, the lactic acid bacterium O. oeni, have demonstrated some capabilities to inhibit B. bruxellensis growth after malolactic fermentation as bioprotector. It is a newly developed concept in enology, and more investigations are needed in order to understand and develop the use of O. oeni in the bioprotection context.

Even if we already have some evidence and clues of the ability of microorganisms to protect must against spoilage microorganisms and to influence wine color at the end of the fermentations, there are still a lack of knowledge about the medium- and long-term effect of the use of bioprotection to reduce SO₂ during wine aging, as well as diverse interactions and mechanisms set up by bioprotection.

Moreover, the cost of bioprotection is still high today, compared with the cost of sulfiting. In fact, bioprotection yeast preparations cost more than active dry yeasts (ADYs), starters of alcoholic fermentation. The development of production and stabilization strategies better adapted for these NS yeasts should make it possible to reduce the cost of the bioprotection strategy in the future.