© The Authors, published by EDP Sciences 2017

1. Introduction

Warm temperatures due to climate change decrease acidity in grapes and allow new allochthonous parasites to establish. Higher pH values limit the microbicide effect of SO₂ and enable good living conditions for harmful organisms, which may empower the formation of SO₂ binding metabolites such as pyruvate and acetaldehyde, further reducing the microbicide effect of SO₂. Various physical sterilization processes like UV-C treatment are considered as techniques for the inactivation of microorganisms. While UV-C treatment is a well- established method in other industries such as water purification, only a few publications focused on the effect of UV-C on microorganism in must and wine [1, 2]. UV-C induces photochemical reactions such as strand breakage of the DNA. As a result, damaged cells are not able to reproduce. For the inactivation of microorganisms, wavelengths between 250 and 260 nm are particularly suitable, since the DNA absorbs very well in this range. The effect of UV-C in liquids generally depends on the penetration depth of the light waves [3], which are either trapped by the absorption of ingredients or scattered by colloids in the liquid [4]. However, this problem can be circumvented by choosing an appropriate technique such as shown in several UV-C reactors [5]. Application fields for the UV-C technology in wineries are the treatment of musts with microbial contamination, the treatment of young wines to stop fermentation, and treatment of wines for the purpose of stabilization, e.g. prior to bulk wine shipping. The technique might be considered as an alternative for thermal pasteurization. In any case, treated products must not suffer from UV-C irradiation. However, UV-C is known for its ability to form powerful off-favours such as methional or methanethiol [6]. Accordingly, the objective of this study was to strive for the minimal UVC energy to inactivate microorganisms without generating detrimental light-struck flavour.

2. Materials and methods

Vitis vinifera grapes of the variety Riesling were harvested in 2016 from the vineyards of the DLR Rheinpfalz, Neustadt, Germany. Harvesting by hand was conducted in three consecutive weeks on same sites obtaining grapes of different quality and different microflora. Immediately after harvesting, grapes were crushed and pressed obtaining musts which were further processed by means of sedimentation for clarification reasons (< 50 NTU). In order to investigate UV-C efficiency in turbid musts, grape lees were blended into the clarified juice yielding musts with 100, 500 and 900 NTU. UV-C treatment of musts was achieved by a pilot-scale coiled tube reactor designed and assembled at the Max-Rubner- Institut (Karlsruhe, Germany). The main component is a module which consists of a helically wound quartz glass tube with a diameter of 3.0 mm which is assembled around a 30 W low pressure mercury lamp with maximum peak radiation at 254 nm (UVN 30, UV Technik Speziallampen GmbH, Wu¨mbach, Germany). Between lamp and coiled quartz glass, a water cooling system allows for stable must temperatures during treatment. Musts were pumped through the device at a flow rate of 25 L/h by a peristaltic pump (Pump drive Pd 5206, Heidolph, Schwabach, Germany). Flow was adjusted to 25 L/h after preliminary studies taking into account that turbulent flow is necessary to maximize UV-C effect [7]. In this study, various UV C doses were obtained by recirculating must and wine through the device for multiple times. UV C dose in J/L was determined using a potassium iodide/iodate chemical actinometry method previously described [8]. All experiments were conducted in triplicate.

Microbial analysis was carried out in 10−5-dilutions (in 0.9% NaCl) using YPD medium (with chloramphenicol und kanamycin) for yeasts and LB medium (with cycloheximide) for bacteria. Agar plates were incubated at 30 °C for 24 h, then colonies were counted. GC- MS analysis of aroma precursors in Riesling must was carried out after SPE purification and enzymatic hydrolysis using a 30 m × 0.25 mm i.D. fused silica capillary column ZB-Wax (PEG) as described by Schober [9]. Photometric analysis was conducted according to OIV-MA-AS2-11 procedure using a Varian Cary100 Spectrophotometer. In order to investigate the sensory effect of UV-C treatment, musts were processed by means of a standardized winemaking protocol. Microvinification of UV-C treated and control musts was carried out in 15 L batches at 18 °C using a dry active yeast strain of Saccharomyces cerevisiae. After alcoholic fermentation wines were racked off the lees and stored for two months at 15 °C for clarification reasons. Stabilization was achieved by using a commercial solution of ammonium hydrogensulphite yielding 30 mg/L free SO₂. Sensory analysis was performed by 21 judges examining the wines by triangle test according to DIN 10954. Statistical calculations were performed using analysis of variance and Tukey-Kramer post-hoc test, significance level was set at p ≤ 0.05.

3. Results and discussion

3.1. UV-C inactivation of natural yeast flora in grape must

Harvesting the Riesling on Sep. 24th, Oct. 6th and Oct. 17th yielded musts with total yeast counts of 6.8 × 104, 3.2 × 105 and 2.0 × 107 cfu/ml, respectively. Right after pressing, four yeasts could be clearly identified in the fresh grape musts, in detail Saccharomyces cerevisiae, Hanseniaspora uvarum, Schizosaccharomyces pombe and Pichia fermentans. When applying UV-C treatment, cultivable yeast cells decreased immediately in the musts. Regardless of the type of yeast, the inactivation success was strongly depended on the initial cell count. Table 1 shows the inactivation power of UV-C with 2 kJ/l for Schizosaccharomyces pompe in clarified Riesling grape must.

Data in Table 1 clearly reveals that inactivation efficacy of UV-C is dependent on the initial degree of contamination suggesting a shadowing effect of individual germs during treatment. Another factor influencing the inactivation efficacy of UV-C is the type of yeasts present in the grape must. The broken-down view on the four yeast types named earlier allows for detailed assessment of the UV-C sensitivity of different yeasts occurring on grapes. Figure 1 shows the survival rates of naturally occurring yeasts in clarified Riesling must after UV-C treatment with different doses.

Fig. 1. Survival rates of naturally occurring yeasts in clarified Riesling must after UV-C treatment with different doses. The initial number of total yeast germs counted on YPD medium was 3.2 × 105 cfu/ml.

Schizosaccharomyces pombe was found to least sensitive to UV-C treatment, followed by Pichia fer- mentans, Saccharomyces cerevisiae and Hanseniaspora uvarum. While Saccharomyces and Hanseniaspora were completely killed at doses >1 kJ/l, Schizosaccharomyces and Pichia still remained in small cell counts. However, at higher doses <1 kJ/l, all occurring yeasts could be inactivated. The technical option to completely or partially inactivate yeasts in grape musts facilitate new opportunities to control spontaneous fermentations.

3.2. UV-C inactivation kinetics of yeast and bacteria in grape must

While facilitating the targeted inactivation of yeasts with UV-C treatment, the question arises as to whether a similar attempt is possible with parallel existence of bacteria and yeast in must and wine. Therefore, UV-C inactivation kinetics have been investigated in clarified Riesling must, which was co-inoculated with Schizosaccharomyces pompe, Brettanomyces bruxellensis and Lactobacillus plantarum, each at 105 cfu/mL, after must pasteurization (Fig. 2).

Fig. 2. UV-C inactivation kinetics for Schizosaccharomyces pompe, Brettanomyces bruxellensis and Lactobacillus plantarum in clarified Riesling must. All microorganisms were inoculated at 105 cfu/mL after must pasteurization.

Brettanomyces bruxellensis and Lactobacillus plan- tarum revealed similar survival rates due to UV-C treatment and were found to be much more sensitive compared to Schizosaccharomyces pompe. This finding suggests that targeted inactivation is also possible for mixtures of bacteria and yeasts in must or wine allowing for new possibilities in controlling harmful growth of bacteria, e.g. at the end of alcoholic fermentation. The relatively high sensitivity of Brettanomyces bruxellensis towards UV-C might be also considered as an important hint for cellars with Brettanomyces problems. However, it was already noted that UV-C efficacy is influenced by anthocyanins and other phenolics, which play a major role in red wines [1].

3.3. Influence of turbidity on UV-C inactivation of yeast in grape must

Turbidity in must and wine is known to reduce UV-C efficacy [1]. Several technical optimization strategies have been proposed to overcome restrictions in UV-C effect in turbid liquids [5]. When employing state-of-the-art UV-C technology, efficacy of the germicidal effect still depended on the turbidity as shown in Fig. 3.

Fig. 3. Influence of turbidity levels on UV-C inactivation of Saccharomyces cerevisiae in Riesling must

Figure 3 reveals that grape lees attenuate UV-C efficacy with an increasing UV-C dose. At 1 kJ/l, UV-C treatment resulted in an average log₁0 microbial reduction of 0.49, 0.72 and 1.37 for musts with 900, 500 and 100 NTU respectively. While 500 and 900 NTU are unrealistically high turbidity values for fermenting grape musts, these findings indicate that existing UV-C technology can still be improved towards higher efficacy.

3.4. Influence of UV-C treatment on aroma precursors in grape must

The efficacy of UV-C is undisputedly a capable technology to inactivate microorganisms in grape juices. Accordingly, the technology has to be considered as an alternative for thermal pasteurization. Therefore, the effects of UV-C treatment and thermal pasteurization on glycosidically- bound aroma precursors in Riesling must were compared with each other (Fig. 4).

Fig. 4. Influence of UV-C treatment (3 kJ/l) and thermal pasteurization (80 °C, 120 s) on glycosidically bound aroma precursors in Riesling must. The substance classes represent the sum of 2–6 aroma precursor compounds within each class.

Both, UV-C treatment with 3 kJ/l and thermal pasteurization at 80 °C for 120 s yielded sterile musts. While heat resulted in significantly lower concentrations of glycosidically bound C6-alcohols, monoterpens and C13- norisoprenoids, UV-C did not cause major losses of aroma precursors. It may be concluded, that UV-C with 3 kJ/l is more gently than heating at 80 °C for 120 s.

3.5. Use of UV-C treatment to stop alcoholic fermentation

To obtain natural residual sugar the alcoholic fermentation of wines is typically stopped by the addition of high concentrations of SO₂. In order to investigate the efficacy of UV-C to stop alcoholic fermentation, a fermenting Riesling at 17 g/L residual sugar was treated either with UV-C or 100 mg/L SO₂. The log₁0 microbial reduction after these treatments is displayed in Fig. 5.

Fig. 5. Efficacy of UV-C treatment to stop alcoholic fermentation of Riesling at 17 g/L residual sugar compared to the use of 100 mg/L SO2.

Although none of the treatments yielded absolute inactivation of Saccharomyces cerevisiae, alcoholic fer- mentation may be considered as interrupted when cells are <102 cfu/ml. UV-C with 1 kJ/l was not capable to reach that number, however, a treatment with > 2 kJ/l exceeded the effect of 100 mg/L SO₂ by 1.5 log₁₀ counts.

3.6. Sensory and colour changes in wine due to UV-C treatment of grape must

UV-C light is known for its ability to form powerful off- favours such as methional or methanethiol [6]. Increasing concentrations of mercaptans due to UV-C treatment of grape must with < 2 kJ/l could not be determined yet. Sensory triangle testing revealed no differences between wines from control an UV-C treated musts. However, higher energies than 2 kJ/l applied in must were tested significantly different in the wine stadium and even low energies <1 kJ/l applied in wines seem to develop off-flavours reminding of cooked cabbage, popcorn and camphor. Other than aroma changes, chromatic properties of wines made from Riesling musts which were either treated with UV-C or thermal pasteurization have already been measured (Fig. 6). A slightly higher yellow colour was observed for wines obtained from heated musts but not due to UV-C treatment.

Fig. 6. Chromatic properties of wines made from Riesling musts which were either treated with UV-C (3 kJ/L) or thermal pasteurization (80 °C, 120 s).

4. Conclusion and outlook

Outcomes in this ongoing research project support the applicability of UV-C treatment to inactivate microorgan- isms in must and wine. The efficacy towards different germs is much higher than anticipated and the fact, that UV-C is adjustable in its intensity, opens up new vistas in enology. Different sensitivities of yeasts and bacteria may be utilized to inactivate the unwanted and to let the beneficial ones survive. The current focus lies on the potential off-flavour formation by UV-C. GC- PFPD analysis towards low volatile sulphides and GC- MS analysis of 2-aminoacetophenone are about to be completed shortly. Since high concentrations of caftaric acid, riboflavin and dissolved oxygen seem to positively contribute to UV-C induced sensory defects in wines, this finding will be pursued.

References

All Tables

All Figures

	Fig. 1. Survival rates of naturally occurring yeasts in clarified Riesling must after UV-C treatment with different doses. The initial number of total yeast germs counted on YPD medium was 3.2 × 105 cfu/ml.
In the text

	Fig. 2. UV-C inactivation kinetics for Schizosaccharomyces pompe, Brettanomyces bruxellensis and Lactobacillus plantarum in clarified Riesling must. All microorganisms were inoculated at 105 cfu/mL after must pasteurization.
In the text

	Fig. 3. Influence of turbidity levels on UV-C inactivation of Saccharomyces cerevisiae in Riesling must
In the text

	Fig. 4. Influence of UV-C treatment (3 kJ/l) and thermal pasteurization (80 °C, 120 s) on glycosidically bound aroma precursors in Riesling must. The substance classes represent the sum of 2–6 aroma precursor compounds within each class.
In the text

	Fig. 5. Efficacy of UV-C treatment to stop alcoholic fermentation of Riesling at 17 g/L residual sugar compared to the use of 100 mg/L SO2.
In the text

	Fig. 6. Chromatic properties of wines made from Riesling musts which were either treated with UV-C (3 kJ/L) or thermal pasteurization (80 °C, 120 s).
In the text