- Email: [email protected]
Using RFLP-mtDNA for the rapid monitoring of the dominant inoculated yeast strain in industrial wine fermentations
International Journal of Food Microbiology 145 (2011) 331–335
Contents lists available at ScienceDirect
International Journal of Food Microbiology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j f o o d m i c r o
Short Communication
Using RFLP-mtDNA for the rapid monitoring of the dominant inoculated yeast strain in industrial wine fermentations María Esther Rodríguez a, Juan José Infante b, Montse Molina c, Laureana Rebordinos a, Jesús Manuel Cantoral a,⁎ a Laboratorio de Microbiología y Genética, CASEM, Departamento de Biomedicina, Biotecnología y Salud Pública, Universidad de Cádiz, Polígono del Río San Pedro s/n, 11510 Puerto Real, Cádiz, Spain b Bioorganic Research and Services, S. L.-Bionaturis, Centro Andaluz de Biología del Desarrollo, Sevilla, Spain c Bodegas Barbadillo S. L. Sanlúcar de Barrameda (Cádiz), Spain
a r t i c l e
i n f o
Article history: Received 9 August 2010 Received in revised form 20 November 2010 Accepted 24 November 2010 Keywords: Inoculated industrial fermentation mtDNA-RFLP PFGE Selected wine yeast
a b s t r a c t The analysis of restriction fragment length polymorphism of mitochondrial DNA (mtDNA-RFLP) has been applied as a test to monitor the abundance of the starter yeast strain during industrial wine fermentations without previous isolation of yeast colonies. For white wine fermentations, we performed a rapid assay consisting in taking a sample of fermenting must, purifying the DNA from harvested cells, and obtaining the restriction patterns by digestion with the endonuclease HinfI. The same protocol, but adding an overnight cultivation step before DNA purification, was also applied to red wine fermentations. The results were compared with those obtained from the subsequent characterisation of strains, for the same samples, by analysis of the electrophoretic karyotype of isolated yeast colonies. In all cases, when the inoculated strain was dominant within the yeast population, the rapid assay anticipated the result by showing the coincidence between the restriction profiles obtained from both total cells and the inoculated strain. The results were obtained at 11 or 23 h after sampling for white- or red-wine fermentations respectively. This method allows a rapid intervention of the wine-producer if the presence of the inoculated yeasts has suffered a sudden decrease in any phase of the fermentation process. © 2010 Elsevier B.V. All rights reserved.
1. Introduction In winemaking, spontaneous fermentation is an ecologically complex process involving successive changes in strains of yeast species, whether Saccharomyces or non-Saccharomyces (reviewed in Fleet, 2008). The diversity of strains can give high-quality and uniqueflavoured wines (Pérez-Coello et al., 1999), thus contributing to the commercial value of products sold in a very competitive market. However, the dynamics of a spontaneous fermentation are often unpredictable. The dominance of specific wild yeast strains might lead to stuck- or slow-fermentations and to inconsistencies in the wine quality (Fleet, 2008). In order to avoid these problems and to control fermentations, winemakers often add carefully selected yeast cultures as fermentation starters. Either commercial active dried wine yeast, ADWY (Manzano et al., 2006) or autochthonous yeast strains (Rodríguez et al., 2010) are used as the fermentation inoculum. The molecular techniques commonly used to differentiate between yeast strains of the same species are Pulsed Field Gel Electrophoresis (PFGE) and Restriction Analysis of the mitochondrial
⁎ Corresponding author. Tel.: +34 956 016156; fax: +34 956 016180. E-mail address: [email protected] (J.M. Cantoral). 0168-1605/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2010.11.035
DNA (mtDNA-RFLP) (González et al., 2007; Rodríguez et al., 2010; Santamaría et al., 2005; Schuller et al., 2004). However, the mtDNARFLP technique is used more frequently because it enables a larger number of strains to be analysed in less time; it is a fast, simple, reliable and economic method, and does not require sophisticated material or specialised personnel (reviewed in Fernández-Espinar et al., 2006). For these reasons, the technique is very suitable for use by industry. The application of these techniques demonstrates that, in many cases, the population of a fermenter is “taken over” by wild yeasts, which relegate the inoculated strains to a minority presence (EsteveZarzoso et al., 2000; Lopes et al., 2007; Raspor et al., 2002). In order to minimise the impact of unwanted ecological evolutions, the industry needs a simple method for rapid diagnosis of the dominance of inoculated strains that could be performed routinely during the fermentation process (Ambrona et al., 2006; López et al., 2003). For this purpose, we have applied mtDNA-RFLP to evaluate the dominance of the inoculated yeast in industrial fermentations, and have obtained the results in about 7 h. In a previous work, we analysed the fermentation strategy of a winery during seven consecutive years by monitoring the proportions of inoculated and wild yeast strains during the fermentation process, using pulsed-field gel electrophoresis (PFGE). These results showed a
332
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
positive correlation between the proportion of the inoculated strain and the quality of the resulting wine (Rodríguez et al., 2010). We also found several examples of real situations that led to a significant decrease in the proportion of the inoculated strain and reduced significantly the quality of the wine. For this reason, it was considered necessary to apply another method for periodical testing of the dominance of the inoculated strains once the must has been inoculated. In this study, we have used RFLP analysis of mtDNA to monitor rapidly the dominance, or otherwise, of inoculated yeast strains in industrial fermentations of white and red wine in a winery of South Western Spain. We apply this technique directly to samples of fermenting wine without previously isolating yeast colonies. Subsequently, we check results using PFGE to validate the previous results obtained from the RFLP test. 2. Materials and methods 2.1. Yeast strains and inoculation of industrial wine fermentations 2.1.1. White wine The fermentations were inoculated with cultures of the autochthonous strain S. cerevisiae P5, previously characterised and selected in our laboratories (Rodríguez et al., 2010). Industrial fermentations were carried out in stainless steel vessels of 400,000 l, and the inoculation was carried out by preparing a starter as described in Rodríguez et al. (2010). In the starter, each scaling-up round was performed when the ºBé reached a value between 1 and 2. In each round, the fermentation volume was increased tenfold to give high initial levels of inoculum (N60 × 106 viable cells/ml) and ensure the correct development of the inoculated strain. The industrial vessels were not filled completely with fresh must and refills were done at irregular intervals. These refills depended on the harvesting and production dynamics of each particular vintage. Moreover, partial volumes of the must fermented in the industrial vessels were used for the inoculation of other 400,000-l fermentations. The samples included in this work were taken 24 h after refills with fresh must, since in these situations fresh inoculums of diverse wild yeast strains were introduced into the fermentation process. The refills were done at either the initial, middle, or final fermentation stages in vessels with wines at different °Bé. 2.1.2. Red wine Fermentations were inoculated separately with different commercial dried yeast strains of S. cerevisiae, Lavin Rhône 2056, Uvaferm VN (Lallemand, Ontario, Canada), AWRI 796 (Maurivin, Toowoomba, Australia), BP 725 (Maurivin, Toowoomba, Australia), and 71B (Lallemand, Ontario, Canada). Industrial fermentations were carried out in stainless steel vessels of 27,000 l, and no refills of must were carried out. In the 2007 vintage, the first four yeasts cited were used. Fermentations were inoculated by hydration of cells following manufacturer's instructions, and the broth was added to the vessel after the cool maceration. In the 2008 vintage, the fermentations studied were inoculated with AWRI 796, BP 725 and 71B yeast by preparing an inoculum in red must, after the hydration of yeast, and allowing it to ferment for 24 h to adapt the yeast to the final conditions of the fermentation.
part of the vessels. Of this sample volume 225 ml was used for analysis of the mtDNA as described below. The remaining 25 ml of each sample volume was harvested and suspended in 20% glycerol in the original media for conservation at −80 °C for subsequent characterisation by electrophoretic karyotype. 2.3. Rapid assay by restriction analysis of the mtDNA The samples were processed differently for the white and red wine fermentations, respectively. For white wine fermentations, a 225-ml sample was centrifuged at 650 × g for 5 min to collect the cells, which were washed three times or more with sterile water to eliminate the must remains and obtaining a clean biomass. For the red wine fermentation, after centrifugation, 300 μl of the samples was spread onto YPD agar plates (1% yeast extract, 2% peptone, 2% glucose and 2% agar), and incubated overnight at 28 °C. After incubation the yeast biomass was collected from the surface of the YPD agar plate, pelleted, washed and then the DNA was extracted for RFLP assays. For white wine, yeast biomass sedimented directly from the wine was used for DNA extraction. The DNA was purified using a method based on that described by Querol et al. (1992) with modifications. The cell pellet was resuspended in 1 ml of buffer 1 (0.1 M EDTA pH 8.0, 0.9 M sorbitol) and transferred to a 2-ml tube containing two 1/4″-ceramic spheres (Q-BIOgene) and 200 μl of 40 mg/ml Lytic Enzyme (Sigma) in buffer 1. The tubes were incubated at 37 °C with shaking for 60 min, after which the cells were pelleted and re-suspended in 500 μl of buffer 2 (50 mM Tris–HCl pH 8.0, 20 mM EDTA) and 200 μl of 10% SDS. The mixture was incubated at 65 °C for 15 min. After adding 200 μl of 5 M potassium acetate, the tubes were kept on ice for 10 min and centrifuged at 9000 × g for 15 min at 4 °C. The supernatant was transferred to a microfuge tube and the DNA was precipitated with 1 vol. of isopropanol for 5 min at room temperature, washed with 70% ethanol, vacuum-dried, and resuspended in 40 μl of sterile water. Five microliters of the DNA sample was digested with 10 units of the endonuclease HinfI (Fermentas) by incubation at 37 °C for 2 h. The restriction fragments were separated by electrophoresis in 1% agarose gels made with 1× TBE buffer, 0.5 μg/ml of ethidium bromide. The image of the gel was digitalised in a Molecular Imager apparatus (GelDoc XR) and analysed using Quantity One 1-D software (Bio-Rad). Our criterion for considering the result of the rapid test to be positive was obtaining restriction patterns of mtDNA that were identical for the total cells and the inoculated strain; in this case, the starter yeast can be taken to be dominant in the fermentation. The results were considered negative when additional bands, or absences of bands, were observed in the patterns; in that case neither the presence nor the dominance of the inoculated yeast strain can be assured. 2.4. Analysis of the electrophoretic karyotype Several dilutions of the frozen stocks were plated on YPD agar and incubated at 28 °C for 2–3 days. After incubation, 20–30 colonies were taken randomly. The karyotypes were obtained by PFGE according to the method described in Rodríguez et al. (2010). This technique was utilised to corroborate the results of the mtDNA-RFLP. 3. Results and discussion 3.1. RFLP analysis of mtDNA
2.2. Sampling For white wine the samples were taken both during the scaling-up process, before each refill, and in the industrial fermentation vessels after partial refills, at the initial (11–6 ºBé), middle (6–3 ºBé) and final stage (3–0 ªBé) of the process. For red wine, samples were taken at the initial, middle and final phases of the alcoholic fermentation. For the molecular analysis, samples of 250 ml were taken from the central
For white wine we began our rapid test in the 2005 vintage. Table 1 shows the results of the total samples tested by applying RFLP-mtDNA directly to total cells of both the starter (S) and the different industrial fermentations vessels (IF) at initial, middle and final phases of the process. For the starter, we analysed 4–7 samples taken from different volumes, for example 3000, 20,000, 85,000 and 150,000 or 300,000 l. The positive results for the starter S indicated the dominance of the
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
333
Table 1 Total samples tested by applying RFLP-mtDNA in successive vintage years in white- and red-wine fermentations. White wine
Red wine
2005 S
a
4+ Total samples analysed
16
2006 b
2007
2008
2009
2007
2008
IF
S
IF
S
IF
S
IF
S
IF
IF
IF
12 +
4+
19 + 14 −
4+
23 + 23 −
7+
30 +
7+
40 + 4−
3+ 12 − 15
17 + 4− 21
37
50
37
51
(+) Positive results which indicate that the restriction pattern of the total cells of the samples and the inoculated strain were identical; (-) negative results which indicate the absence of an identical restriction pattern and the non-dominance of the inoculated strain. a Starter of the inoculated yeast strain P5. b Industrial fermentations.
inoculated strain in each volume analysed before each refill during the scaling-up. Thus the correct course of the fermentation of all the starters is assured before the inoculation of the industrial vessels. For the industrial fermentations, we tested different vessels in at least two different phases of the process, after refills with fresh must and on the completion of the fermentations (data not shown). Fig. 1 (Panel A) shows examples of positive results of the RFLP test, because the restriction patterns of the samples and inoculated strain were identical, for the final fermentation phase in the industrial vessels B, C, G and H. In these cases the inoculated strain, P5, was present in
majority, and this strain was responsible for the entire fermentation process. For the vessels M and N, negative results of the RFLP test for the middle and final fermentation phases respectively are shown (Fig. 1, Panel A). In these vessels, and others where the test was negative, evidence was observed of spontaneous fermentations before the inoculation, due to the volume of must stored. When these vessels were inoculated, the strain P5 was not implanted and we concluded that another yeast population was dominant. After the results of the RFLP tests, the winemaker decided not to use these fermentations for inoculating other vessels.
Fig. 1. Rapid RFLP test and PFGE of samples from white- and red-wine fermentations. The fermentations were analysed at initial (i), middle (m) or final (f) phases of the process. Panel A shows RFLP-mtDNA test with HinfI of the inoculated autochthonous strain S. cerevisiae P5 and total cells from white-wine samples taken in 2007 from fermentation vessels B, C, G, H, M and N. Panel B shows karyotype analysis by PFGE of the inoculated strain (P5) and 14 colonies isolated from the sample taken in 2007 from vessel G at the final fermentation phase. Reference chromosomal sizes are indicated. Panel C shows RFLP-mtDNA test of total cells from vessels T1, T3 and T4 of red-wine fermentations (2007 vintage) and inoculated commercial yeast. T1 was inoculated with Rhône 2056, and T3, T4 with BP 725. Reference band sizes are indicated. The results of the PFGE analysis of 14 colonies isolated from vessel T3 at the final fermentation phase in 2007 are shown in the Panel D. Only 3 colonies (white arrows) showed the karyotype of the inoculated strain BP 725. Reference chromosomal sizes are indicated. Panel E shows RFLP-mtDNA with HinfI of total cells from red-wine vessels T2 and T3 of the 2008 vintage, in which the commercial strain S. cerevisiae AWRI 796 was inoculated. MWs are lambda-HindIII.
334
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
For red wine we analysed seven and nine different vessels at initial, middle or final phases of the fermentations in vintages of 2007 and 2008 respectively (data not shown).The total samples analysed by RFLP tests are shown in Table 1. It should be noted that in the 2007 vintage the results of the tests were, in general, negative, and the inoculated strains were not successfully implanted in the fermentations studied. Fig. 1 (Panel C) shows examples of these negative results in the RFLP test for vessels T3, T4 and T1, in which the displacement of the inoculated commercial strains S. cerevisiae BP 725 and Rhône 2056 was observed, probably by a population of wild yeasts; this is concluded because the RFLP patterns were not identical and they showed additional bands. On the other hand, for the 2008 vintage, Fig. 1 (Panel E) shows examples of the dominance of the inoculated yeast strain AWRI 796 in two different vessels, T2 and T3. These results were taken as confirmation that the best method of inoculation in red must was the used in the 2008 vintage; this consisted of preparing an inoculum after the hydration of yeast, and allowing it to ferment for 24 h to enable the yeast cells to adapt to the final conditions of the process (see Materials and methods). In the following vintage of 2009, all the red must fermentations of the winery were carried out using this method (data not shown). 3.2. Corroboration of the RFLP assay applying PFGE
Table 2 Correlation between the results obtained by RFLP tests and karyotype analysis of samples from white-wine fermentations inoculated with the autochthonous yeast S. cerevisiae P5 in 3 different vintages. The letters from B to O identify the particular fermentation vessels sampled after partial refills performed at either the initial (i), middle (m) or final (f) phases of fermentation. Vintage
Vessel
Fermentation phase
RFLP-mtDNA result
% Inoculated strain by PFGE (# of yeast colonies analysed)
2005
E
m f i m f i m m i f m f m f m f m
+ + + + + + + + + + − − + + + + +
100 (20) 65 (20) 100 (20) 90 (20) 95 (20) 100 (9) 100 (20) 95 (20) 64 (14) 100 (20) 60 (20) 15 (20) 100 (20) 100 (20) 100 (20) 100 (20) 100 (20)
H G 2007
2008
B G M N G H S M O
Vintage
Vessel
Fermentation phase
RFLP-mtDNA result
% Inoculated strain by PFGE (# of yeast colonies analysed)
2007
T1
i m f i f m f m f m f m m f i f m
− − − − − − − − − − − + − − + + +
30 Rhône 2056 (20) 10 Rhône 2056 (20) 10 Rhône 2056 (20) 40 AWRI 796 (20) 55 AWRI 796 (20) 5 BP 725 (20) 15 BP 725 (20) 30 BP 725 (20) 40 BP 725 (20) 17 AWRI 796 (18) 30 AWRI 796 (20) 75 VN (20) 20 AWRI 796 (20) 0 AWRI 796 (20) 95 AWRI 796 (19) 100 AWRI 796 (20) 93 AWRI 796 (14)
T2 T3 T4 T5
2008
T7 T1 T2 T3
To validate the results obtained by RFLP, 34 of the tested samples were analysed by electrophoretic karyotype. For white wine, 323 colonies were analysed and the results are presented in Table 2. The comparison with the results obtained by PFGE analysis shows that a positive RFLP test indicated that the inoculated strain is present in majority (Fig. 1, Panel B). In this type of wine, our results indicated that when the RFLP test was positive, the inoculated strain was present in the fermentations at ≥64%. When the RFLP tests were negative, as shown in Fig. 1 (Panel A) for the sample Mm, we confirmed by PFGE that the yeast starter was present at ≤60% (Table 2). Nevertheless, further experiments will be necessary to confirm these correlations because the RFLP assays show qualitative results and we cannot know the actual percentage implantation of starter yeast when the results are positive or negative. For red wine the results of 331 colonies analysed by PFGE are shown in Table 3. In all the cases the results of the RFLP test predicted the results obtained later by applying PFGE. However, we cannot
F
Table 3 Correlation between the results obtained by RFLP tests and karyotype analysis of samples from red-wine fermentations inoculated with commercial starter strains in 2 successive vintages. The references T1 to T7 identify the particular fermentation vessels sampled at the initial (i), middle (m) or final (f) phases of fermentation. In the 2008 vintage only the results for three fermentations inoculated with commercial yeast AWRI 796 are shown.
establish the limit found for white wine because from the positive results obtained by applying the RFLP test, the presence of the inoculated strain was greater than 75%, and all the negative results were ≤55%. For example, in Fig. 1 (Panel D) we confirmed the negative result previously obtained by the RFLP test for the sample taken of the final phase of the fermentation in the vessel T3, which shows an example for red wine. In fact, only 15% of the colonies isolated from this sample (Table 3) showed the karyotype of the inoculated strain. The results of the RFLP test were obtained 11 h after taking the sample for white-wine fermentations, and 23 h after taking the sample from red-wine. However, this time can be shortened; for example, the step of incubation at 37 °C (see Section 2.3) could be about 15 min if other methods are used to rupture cells. On the other hand, the time also depends of the number of the samples analysed per day and whether the samples contain a greater amount of must remains. In our work, we have considered the maximum time taken, and have included the time needed to return to the laboratory (about 1 h). Taking these factors into account, the results can be obtained in less time—for white wine, in about 7 h. We were unable to obtain successful endonuclease digestions of DNA in red-wine samples treated with exactly the same protocol as that applied to white-wine samples. Since the red-wine remains were difficult to clear by centrifugation, our interpretation of this finding is that some compounds remaining in the digested DNA sample were inhibitory for the endonuclease. Therefore, we added a step in which a sample of the red-wine must was plated on YPD-agar and incubated overnight at 28 °C (see Materials and methods). Then we used the cleaner biomass obtained from the plate as representative of the yeast population of the sample; this gave good results, although the time required for the assay was somewhat longer. It might be thought that the additional step of cultivation for the samples of red must could lead to the growth of non-Saccharomyces yeast, and this was the reason why the RFLP test was negative. However, when the inoculated yeast is the dominant, the total yeast restriction pattern is identical to the pattern of the inoculated yeast, even at the initial phase of the fermentation, as can be observed in Fig. 1 (Panel E) for the vessels T2 and T3. We think that for red wine, the time taken to obtain the results could also be shortened, like that for white wine, if the clean biomass
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
can be separated from the must remains in a few minutes. To achieve this, further experiments will need to be carried out. In conclusion, the RFLP test for monitoring of the predominance of the inoculated yeast is proposed as a response to one of the major challenges for microbiological control in the wine industry. This study describes real situations taking place during actual wine fermentations, for example spontaneous fermentations before the inoculation, and we offer a test which the winemaker can use to obtain a reliable indication of whether or not wild yeasts are displacing the inoculated strains. Although the characterisation of the inoculated strains and their evolution while competing with wild strains must be performed first by PFGE, this rapid method is presented as a reliable quality control check for use in successive years. If the strategy presented is followed, the wine producer would be able to identify and correct in time the unwanted evolution of the yeast population, usually by reinoculating the selected strains and/or correcting a deviation in temperature or change in some other parameter of the vessel that might have caused the unwanted situation. Acknowledgements We thank R&D Technician Cristina Domínguez and Dr. Juan José Mesa in Bodegas Barbadillo S.L. for their help in sampling the industrial fermentations. This work was supported by the grants PETRI 95-0855 OP from DGICYT of the Ministry of Science and Innovation, and OT 054/174/015/020/114/136/104 from Bodegas Barbadillo S.L. of Sanlúcar de Barrameda, Spain. References Ambrona, J., Vinagre, A., Maqueda, M., Álvarez, M.L., Ramírez, M., 2006. Rhodaminepink as a genetic marker for yeast populations in wine fermentation. Journal of Agricultural and Food Chemistry 54, 2977–2984.
335
Esteve-Zarzoso, B., Gostíncar, A., Bobet, R., Uruburu, F., Querol, A., 2000. Selection and molecular characterization of wine yeasts isolated from the ‘El Penedès’ area (Spain). Food Microbiology 17, 553–562. Fernández-Espinar, M.T., Martorell, P., De Llanos, R., Querol, A., 2006. Molecular methods to identify and characterize yeasts in foods and beverages. In: Querol, A., Fleet, G.H. (Eds.), The Yeast Handbook: Yeast in Food and Beverages. Springer, pp. 55–82. Fleet, G.H., 2008. Wine yeasts for the future. FEMS Yeast Research 8, 979–995. González, S.S., Barrio, E., Querol, A., 2007. Molecular identification and characterization of wine yeast isolated from Tenerife (Canary Island, Spain). Journal of Applied Microbiology 102, 1018–1025. Lopes, C.A., Rodríguez, M.E., Sangorrín, M., Querol, A., Caballero, A.C., 2007. Patagonian wines: implantation of an indigenous strain of Saccharomyces cerevisiae in fermentations conducted in traditional and modern cellars. Journal of Industrial Microbiology & Biotechnology 34, 139–149. López, V., Fernández-Espinar, M.T., Barrio, E., Ramón, D., Querol, A., 2003. A new PCRbased method for monitoring inoculated wine fermentations. International Journal of Food Microbiology 81, 63–71. Manzano, M., Medrala, D., Giusto, C., Bartolomeoli, I., Urso, R., Comi, G., 2006. Classical and molecular analyses to characterise commercial dry yeast used in wine fermentations. Journal of Applied Microbiology 100, 599–607. Pérez-Coello, M.S., Briones Pérez, A.I., Ubeda Iranzo, J.F., Martin Alvarez, P.J., 1999. Characteristics of wines fermented with different Saccharomyces cerevisiae strains isolated from the La Mancha region. Food Microbiology 16, 563–573. Querol, A., Barrio, E., Ramón, D., 1992. A comparative study of different methods of yeast strain characterization. Systematic and Applied Microbiology 15, 439–446. Raspor, P., Cus, F., Povhe Jemec, K., Zagorc, T., Cadez, N., Nemanic, J., 2002. Yeast population dynamics in spontaneous and inoculated alcoholic fermentations of Zametovka must. Food Technology and Biotechnology 40, 95–102. Rodríguez, M.E., Infante, J.J., Molina, M., Domínguez, M., Rebordinos, L., Cantoral, J.M., 2010. Genomic characterization and selection of wine yeast to conduct industrial fermentations of a white wine produced in a SW Spain winery. Journal of Applied Microbiology 108, 1292–1302. Santamaría, P., Garijo, P., López, R., Tenorio, C., Gutiérrez, A.R., 2005. Analysis of yeast population during spontaneous alcoholic fermentation: effect of the age of the cellar and practice of inoculation. International Journal of Food Microbiology 103, 49–56. Schuller, D., Valero, E., Dequin, S., Casal, M., 2004. Survey of molecular methods for the typing of wine yeast strains. FEMS Microbiology Letters 231, 19–26.
Contents lists available at ScienceDirect
International Journal of Food Microbiology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j f o o d m i c r o
Short Communication
Using RFLP-mtDNA for the rapid monitoring of the dominant inoculated yeast strain in industrial wine fermentations María Esther Rodríguez a, Juan José Infante b, Montse Molina c, Laureana Rebordinos a, Jesús Manuel Cantoral a,⁎ a Laboratorio de Microbiología y Genética, CASEM, Departamento de Biomedicina, Biotecnología y Salud Pública, Universidad de Cádiz, Polígono del Río San Pedro s/n, 11510 Puerto Real, Cádiz, Spain b Bioorganic Research and Services, S. L.-Bionaturis, Centro Andaluz de Biología del Desarrollo, Sevilla, Spain c Bodegas Barbadillo S. L. Sanlúcar de Barrameda (Cádiz), Spain
a r t i c l e
i n f o
Article history: Received 9 August 2010 Received in revised form 20 November 2010 Accepted 24 November 2010 Keywords: Inoculated industrial fermentation mtDNA-RFLP PFGE Selected wine yeast
a b s t r a c t The analysis of restriction fragment length polymorphism of mitochondrial DNA (mtDNA-RFLP) has been applied as a test to monitor the abundance of the starter yeast strain during industrial wine fermentations without previous isolation of yeast colonies. For white wine fermentations, we performed a rapid assay consisting in taking a sample of fermenting must, purifying the DNA from harvested cells, and obtaining the restriction patterns by digestion with the endonuclease HinfI. The same protocol, but adding an overnight cultivation step before DNA purification, was also applied to red wine fermentations. The results were compared with those obtained from the subsequent characterisation of strains, for the same samples, by analysis of the electrophoretic karyotype of isolated yeast colonies. In all cases, when the inoculated strain was dominant within the yeast population, the rapid assay anticipated the result by showing the coincidence between the restriction profiles obtained from both total cells and the inoculated strain. The results were obtained at 11 or 23 h after sampling for white- or red-wine fermentations respectively. This method allows a rapid intervention of the wine-producer if the presence of the inoculated yeasts has suffered a sudden decrease in any phase of the fermentation process. © 2010 Elsevier B.V. All rights reserved.
1. Introduction In winemaking, spontaneous fermentation is an ecologically complex process involving successive changes in strains of yeast species, whether Saccharomyces or non-Saccharomyces (reviewed in Fleet, 2008). The diversity of strains can give high-quality and uniqueflavoured wines (Pérez-Coello et al., 1999), thus contributing to the commercial value of products sold in a very competitive market. However, the dynamics of a spontaneous fermentation are often unpredictable. The dominance of specific wild yeast strains might lead to stuck- or slow-fermentations and to inconsistencies in the wine quality (Fleet, 2008). In order to avoid these problems and to control fermentations, winemakers often add carefully selected yeast cultures as fermentation starters. Either commercial active dried wine yeast, ADWY (Manzano et al., 2006) or autochthonous yeast strains (Rodríguez et al., 2010) are used as the fermentation inoculum. The molecular techniques commonly used to differentiate between yeast strains of the same species are Pulsed Field Gel Electrophoresis (PFGE) and Restriction Analysis of the mitochondrial
⁎ Corresponding author. Tel.: +34 956 016156; fax: +34 956 016180. E-mail address: [email protected] (J.M. Cantoral). 0168-1605/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2010.11.035
DNA (mtDNA-RFLP) (González et al., 2007; Rodríguez et al., 2010; Santamaría et al., 2005; Schuller et al., 2004). However, the mtDNARFLP technique is used more frequently because it enables a larger number of strains to be analysed in less time; it is a fast, simple, reliable and economic method, and does not require sophisticated material or specialised personnel (reviewed in Fernández-Espinar et al., 2006). For these reasons, the technique is very suitable for use by industry. The application of these techniques demonstrates that, in many cases, the population of a fermenter is “taken over” by wild yeasts, which relegate the inoculated strains to a minority presence (EsteveZarzoso et al., 2000; Lopes et al., 2007; Raspor et al., 2002). In order to minimise the impact of unwanted ecological evolutions, the industry needs a simple method for rapid diagnosis of the dominance of inoculated strains that could be performed routinely during the fermentation process (Ambrona et al., 2006; López et al., 2003). For this purpose, we have applied mtDNA-RFLP to evaluate the dominance of the inoculated yeast in industrial fermentations, and have obtained the results in about 7 h. In a previous work, we analysed the fermentation strategy of a winery during seven consecutive years by monitoring the proportions of inoculated and wild yeast strains during the fermentation process, using pulsed-field gel electrophoresis (PFGE). These results showed a
332
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
positive correlation between the proportion of the inoculated strain and the quality of the resulting wine (Rodríguez et al., 2010). We also found several examples of real situations that led to a significant decrease in the proportion of the inoculated strain and reduced significantly the quality of the wine. For this reason, it was considered necessary to apply another method for periodical testing of the dominance of the inoculated strains once the must has been inoculated. In this study, we have used RFLP analysis of mtDNA to monitor rapidly the dominance, or otherwise, of inoculated yeast strains in industrial fermentations of white and red wine in a winery of South Western Spain. We apply this technique directly to samples of fermenting wine without previously isolating yeast colonies. Subsequently, we check results using PFGE to validate the previous results obtained from the RFLP test. 2. Materials and methods 2.1. Yeast strains and inoculation of industrial wine fermentations 2.1.1. White wine The fermentations were inoculated with cultures of the autochthonous strain S. cerevisiae P5, previously characterised and selected in our laboratories (Rodríguez et al., 2010). Industrial fermentations were carried out in stainless steel vessels of 400,000 l, and the inoculation was carried out by preparing a starter as described in Rodríguez et al. (2010). In the starter, each scaling-up round was performed when the ºBé reached a value between 1 and 2. In each round, the fermentation volume was increased tenfold to give high initial levels of inoculum (N60 × 106 viable cells/ml) and ensure the correct development of the inoculated strain. The industrial vessels were not filled completely with fresh must and refills were done at irregular intervals. These refills depended on the harvesting and production dynamics of each particular vintage. Moreover, partial volumes of the must fermented in the industrial vessels were used for the inoculation of other 400,000-l fermentations. The samples included in this work were taken 24 h after refills with fresh must, since in these situations fresh inoculums of diverse wild yeast strains were introduced into the fermentation process. The refills were done at either the initial, middle, or final fermentation stages in vessels with wines at different °Bé. 2.1.2. Red wine Fermentations were inoculated separately with different commercial dried yeast strains of S. cerevisiae, Lavin Rhône 2056, Uvaferm VN (Lallemand, Ontario, Canada), AWRI 796 (Maurivin, Toowoomba, Australia), BP 725 (Maurivin, Toowoomba, Australia), and 71B (Lallemand, Ontario, Canada). Industrial fermentations were carried out in stainless steel vessels of 27,000 l, and no refills of must were carried out. In the 2007 vintage, the first four yeasts cited were used. Fermentations were inoculated by hydration of cells following manufacturer's instructions, and the broth was added to the vessel after the cool maceration. In the 2008 vintage, the fermentations studied were inoculated with AWRI 796, BP 725 and 71B yeast by preparing an inoculum in red must, after the hydration of yeast, and allowing it to ferment for 24 h to adapt the yeast to the final conditions of the fermentation.
part of the vessels. Of this sample volume 225 ml was used for analysis of the mtDNA as described below. The remaining 25 ml of each sample volume was harvested and suspended in 20% glycerol in the original media for conservation at −80 °C for subsequent characterisation by electrophoretic karyotype. 2.3. Rapid assay by restriction analysis of the mtDNA The samples were processed differently for the white and red wine fermentations, respectively. For white wine fermentations, a 225-ml sample was centrifuged at 650 × g for 5 min to collect the cells, which were washed three times or more with sterile water to eliminate the must remains and obtaining a clean biomass. For the red wine fermentation, after centrifugation, 300 μl of the samples was spread onto YPD agar plates (1% yeast extract, 2% peptone, 2% glucose and 2% agar), and incubated overnight at 28 °C. After incubation the yeast biomass was collected from the surface of the YPD agar plate, pelleted, washed and then the DNA was extracted for RFLP assays. For white wine, yeast biomass sedimented directly from the wine was used for DNA extraction. The DNA was purified using a method based on that described by Querol et al. (1992) with modifications. The cell pellet was resuspended in 1 ml of buffer 1 (0.1 M EDTA pH 8.0, 0.9 M sorbitol) and transferred to a 2-ml tube containing two 1/4″-ceramic spheres (Q-BIOgene) and 200 μl of 40 mg/ml Lytic Enzyme (Sigma) in buffer 1. The tubes were incubated at 37 °C with shaking for 60 min, after which the cells were pelleted and re-suspended in 500 μl of buffer 2 (50 mM Tris–HCl pH 8.0, 20 mM EDTA) and 200 μl of 10% SDS. The mixture was incubated at 65 °C for 15 min. After adding 200 μl of 5 M potassium acetate, the tubes were kept on ice for 10 min and centrifuged at 9000 × g for 15 min at 4 °C. The supernatant was transferred to a microfuge tube and the DNA was precipitated with 1 vol. of isopropanol for 5 min at room temperature, washed with 70% ethanol, vacuum-dried, and resuspended in 40 μl of sterile water. Five microliters of the DNA sample was digested with 10 units of the endonuclease HinfI (Fermentas) by incubation at 37 °C for 2 h. The restriction fragments were separated by electrophoresis in 1% agarose gels made with 1× TBE buffer, 0.5 μg/ml of ethidium bromide. The image of the gel was digitalised in a Molecular Imager apparatus (GelDoc XR) and analysed using Quantity One 1-D software (Bio-Rad). Our criterion for considering the result of the rapid test to be positive was obtaining restriction patterns of mtDNA that were identical for the total cells and the inoculated strain; in this case, the starter yeast can be taken to be dominant in the fermentation. The results were considered negative when additional bands, or absences of bands, were observed in the patterns; in that case neither the presence nor the dominance of the inoculated yeast strain can be assured. 2.4. Analysis of the electrophoretic karyotype Several dilutions of the frozen stocks were plated on YPD agar and incubated at 28 °C for 2–3 days. After incubation, 20–30 colonies were taken randomly. The karyotypes were obtained by PFGE according to the method described in Rodríguez et al. (2010). This technique was utilised to corroborate the results of the mtDNA-RFLP. 3. Results and discussion 3.1. RFLP analysis of mtDNA
2.2. Sampling For white wine the samples were taken both during the scaling-up process, before each refill, and in the industrial fermentation vessels after partial refills, at the initial (11–6 ºBé), middle (6–3 ºBé) and final stage (3–0 ªBé) of the process. For red wine, samples were taken at the initial, middle and final phases of the alcoholic fermentation. For the molecular analysis, samples of 250 ml were taken from the central
For white wine we began our rapid test in the 2005 vintage. Table 1 shows the results of the total samples tested by applying RFLP-mtDNA directly to total cells of both the starter (S) and the different industrial fermentations vessels (IF) at initial, middle and final phases of the process. For the starter, we analysed 4–7 samples taken from different volumes, for example 3000, 20,000, 85,000 and 150,000 or 300,000 l. The positive results for the starter S indicated the dominance of the
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
333
Table 1 Total samples tested by applying RFLP-mtDNA in successive vintage years in white- and red-wine fermentations. White wine
Red wine
2005 S
a
4+ Total samples analysed
16
2006 b
2007
2008
2009
2007
2008
IF
S
IF
S
IF
S
IF
S
IF
IF
IF
12 +
4+
19 + 14 −
4+
23 + 23 −
7+
30 +
7+
40 + 4−
3+ 12 − 15
17 + 4− 21
37
50
37
51
(+) Positive results which indicate that the restriction pattern of the total cells of the samples and the inoculated strain were identical; (-) negative results which indicate the absence of an identical restriction pattern and the non-dominance of the inoculated strain. a Starter of the inoculated yeast strain P5. b Industrial fermentations.
inoculated strain in each volume analysed before each refill during the scaling-up. Thus the correct course of the fermentation of all the starters is assured before the inoculation of the industrial vessels. For the industrial fermentations, we tested different vessels in at least two different phases of the process, after refills with fresh must and on the completion of the fermentations (data not shown). Fig. 1 (Panel A) shows examples of positive results of the RFLP test, because the restriction patterns of the samples and inoculated strain were identical, for the final fermentation phase in the industrial vessels B, C, G and H. In these cases the inoculated strain, P5, was present in
majority, and this strain was responsible for the entire fermentation process. For the vessels M and N, negative results of the RFLP test for the middle and final fermentation phases respectively are shown (Fig. 1, Panel A). In these vessels, and others where the test was negative, evidence was observed of spontaneous fermentations before the inoculation, due to the volume of must stored. When these vessels were inoculated, the strain P5 was not implanted and we concluded that another yeast population was dominant. After the results of the RFLP tests, the winemaker decided not to use these fermentations for inoculating other vessels.
Fig. 1. Rapid RFLP test and PFGE of samples from white- and red-wine fermentations. The fermentations were analysed at initial (i), middle (m) or final (f) phases of the process. Panel A shows RFLP-mtDNA test with HinfI of the inoculated autochthonous strain S. cerevisiae P5 and total cells from white-wine samples taken in 2007 from fermentation vessels B, C, G, H, M and N. Panel B shows karyotype analysis by PFGE of the inoculated strain (P5) and 14 colonies isolated from the sample taken in 2007 from vessel G at the final fermentation phase. Reference chromosomal sizes are indicated. Panel C shows RFLP-mtDNA test of total cells from vessels T1, T3 and T4 of red-wine fermentations (2007 vintage) and inoculated commercial yeast. T1 was inoculated with Rhône 2056, and T3, T4 with BP 725. Reference band sizes are indicated. The results of the PFGE analysis of 14 colonies isolated from vessel T3 at the final fermentation phase in 2007 are shown in the Panel D. Only 3 colonies (white arrows) showed the karyotype of the inoculated strain BP 725. Reference chromosomal sizes are indicated. Panel E shows RFLP-mtDNA with HinfI of total cells from red-wine vessels T2 and T3 of the 2008 vintage, in which the commercial strain S. cerevisiae AWRI 796 was inoculated. MWs are lambda-HindIII.
334
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
For red wine we analysed seven and nine different vessels at initial, middle or final phases of the fermentations in vintages of 2007 and 2008 respectively (data not shown).The total samples analysed by RFLP tests are shown in Table 1. It should be noted that in the 2007 vintage the results of the tests were, in general, negative, and the inoculated strains were not successfully implanted in the fermentations studied. Fig. 1 (Panel C) shows examples of these negative results in the RFLP test for vessels T3, T4 and T1, in which the displacement of the inoculated commercial strains S. cerevisiae BP 725 and Rhône 2056 was observed, probably by a population of wild yeasts; this is concluded because the RFLP patterns were not identical and they showed additional bands. On the other hand, for the 2008 vintage, Fig. 1 (Panel E) shows examples of the dominance of the inoculated yeast strain AWRI 796 in two different vessels, T2 and T3. These results were taken as confirmation that the best method of inoculation in red must was the used in the 2008 vintage; this consisted of preparing an inoculum after the hydration of yeast, and allowing it to ferment for 24 h to enable the yeast cells to adapt to the final conditions of the process (see Materials and methods). In the following vintage of 2009, all the red must fermentations of the winery were carried out using this method (data not shown). 3.2. Corroboration of the RFLP assay applying PFGE
Table 2 Correlation between the results obtained by RFLP tests and karyotype analysis of samples from white-wine fermentations inoculated with the autochthonous yeast S. cerevisiae P5 in 3 different vintages. The letters from B to O identify the particular fermentation vessels sampled after partial refills performed at either the initial (i), middle (m) or final (f) phases of fermentation. Vintage
Vessel
Fermentation phase
RFLP-mtDNA result
% Inoculated strain by PFGE (# of yeast colonies analysed)
2005
E
m f i m f i m m i f m f m f m f m
+ + + + + + + + + + − − + + + + +
100 (20) 65 (20) 100 (20) 90 (20) 95 (20) 100 (9) 100 (20) 95 (20) 64 (14) 100 (20) 60 (20) 15 (20) 100 (20) 100 (20) 100 (20) 100 (20) 100 (20)
H G 2007
2008
B G M N G H S M O
Vintage
Vessel
Fermentation phase
RFLP-mtDNA result
% Inoculated strain by PFGE (# of yeast colonies analysed)
2007
T1
i m f i f m f m f m f m m f i f m
− − − − − − − − − − − + − − + + +
30 Rhône 2056 (20) 10 Rhône 2056 (20) 10 Rhône 2056 (20) 40 AWRI 796 (20) 55 AWRI 796 (20) 5 BP 725 (20) 15 BP 725 (20) 30 BP 725 (20) 40 BP 725 (20) 17 AWRI 796 (18) 30 AWRI 796 (20) 75 VN (20) 20 AWRI 796 (20) 0 AWRI 796 (20) 95 AWRI 796 (19) 100 AWRI 796 (20) 93 AWRI 796 (14)
T2 T3 T4 T5
2008
T7 T1 T2 T3
To validate the results obtained by RFLP, 34 of the tested samples were analysed by electrophoretic karyotype. For white wine, 323 colonies were analysed and the results are presented in Table 2. The comparison with the results obtained by PFGE analysis shows that a positive RFLP test indicated that the inoculated strain is present in majority (Fig. 1, Panel B). In this type of wine, our results indicated that when the RFLP test was positive, the inoculated strain was present in the fermentations at ≥64%. When the RFLP tests were negative, as shown in Fig. 1 (Panel A) for the sample Mm, we confirmed by PFGE that the yeast starter was present at ≤60% (Table 2). Nevertheless, further experiments will be necessary to confirm these correlations because the RFLP assays show qualitative results and we cannot know the actual percentage implantation of starter yeast when the results are positive or negative. For red wine the results of 331 colonies analysed by PFGE are shown in Table 3. In all the cases the results of the RFLP test predicted the results obtained later by applying PFGE. However, we cannot
F
Table 3 Correlation between the results obtained by RFLP tests and karyotype analysis of samples from red-wine fermentations inoculated with commercial starter strains in 2 successive vintages. The references T1 to T7 identify the particular fermentation vessels sampled at the initial (i), middle (m) or final (f) phases of fermentation. In the 2008 vintage only the results for three fermentations inoculated with commercial yeast AWRI 796 are shown.
establish the limit found for white wine because from the positive results obtained by applying the RFLP test, the presence of the inoculated strain was greater than 75%, and all the negative results were ≤55%. For example, in Fig. 1 (Panel D) we confirmed the negative result previously obtained by the RFLP test for the sample taken of the final phase of the fermentation in the vessel T3, which shows an example for red wine. In fact, only 15% of the colonies isolated from this sample (Table 3) showed the karyotype of the inoculated strain. The results of the RFLP test were obtained 11 h after taking the sample for white-wine fermentations, and 23 h after taking the sample from red-wine. However, this time can be shortened; for example, the step of incubation at 37 °C (see Section 2.3) could be about 15 min if other methods are used to rupture cells. On the other hand, the time also depends of the number of the samples analysed per day and whether the samples contain a greater amount of must remains. In our work, we have considered the maximum time taken, and have included the time needed to return to the laboratory (about 1 h). Taking these factors into account, the results can be obtained in less time—for white wine, in about 7 h. We were unable to obtain successful endonuclease digestions of DNA in red-wine samples treated with exactly the same protocol as that applied to white-wine samples. Since the red-wine remains were difficult to clear by centrifugation, our interpretation of this finding is that some compounds remaining in the digested DNA sample were inhibitory for the endonuclease. Therefore, we added a step in which a sample of the red-wine must was plated on YPD-agar and incubated overnight at 28 °C (see Materials and methods). Then we used the cleaner biomass obtained from the plate as representative of the yeast population of the sample; this gave good results, although the time required for the assay was somewhat longer. It might be thought that the additional step of cultivation for the samples of red must could lead to the growth of non-Saccharomyces yeast, and this was the reason why the RFLP test was negative. However, when the inoculated yeast is the dominant, the total yeast restriction pattern is identical to the pattern of the inoculated yeast, even at the initial phase of the fermentation, as can be observed in Fig. 1 (Panel E) for the vessels T2 and T3. We think that for red wine, the time taken to obtain the results could also be shortened, like that for white wine, if the clean biomass
M.E. Rodríguez et al. / International Journal of Food Microbiology 145 (2011) 331–335
can be separated from the must remains in a few minutes. To achieve this, further experiments will need to be carried out. In conclusion, the RFLP test for monitoring of the predominance of the inoculated yeast is proposed as a response to one of the major challenges for microbiological control in the wine industry. This study describes real situations taking place during actual wine fermentations, for example spontaneous fermentations before the inoculation, and we offer a test which the winemaker can use to obtain a reliable indication of whether or not wild yeasts are displacing the inoculated strains. Although the characterisation of the inoculated strains and their evolution while competing with wild strains must be performed first by PFGE, this rapid method is presented as a reliable quality control check for use in successive years. If the strategy presented is followed, the wine producer would be able to identify and correct in time the unwanted evolution of the yeast population, usually by reinoculating the selected strains and/or correcting a deviation in temperature or change in some other parameter of the vessel that might have caused the unwanted situation. Acknowledgements We thank R&D Technician Cristina Domínguez and Dr. Juan José Mesa in Bodegas Barbadillo S.L. for their help in sampling the industrial fermentations. This work was supported by the grants PETRI 95-0855 OP from DGICYT of the Ministry of Science and Innovation, and OT 054/174/015/020/114/136/104 from Bodegas Barbadillo S.L. of Sanlúcar de Barrameda, Spain. References Ambrona, J., Vinagre, A., Maqueda, M., Álvarez, M.L., Ramírez, M., 2006. Rhodaminepink as a genetic marker for yeast populations in wine fermentation. Journal of Agricultural and Food Chemistry 54, 2977–2984.
335
Esteve-Zarzoso, B., Gostíncar, A., Bobet, R., Uruburu, F., Querol, A., 2000. Selection and molecular characterization of wine yeasts isolated from the ‘El Penedès’ area (Spain). Food Microbiology 17, 553–562. Fernández-Espinar, M.T., Martorell, P., De Llanos, R., Querol, A., 2006. Molecular methods to identify and characterize yeasts in foods and beverages. In: Querol, A., Fleet, G.H. (Eds.), The Yeast Handbook: Yeast in Food and Beverages. Springer, pp. 55–82. Fleet, G.H., 2008. Wine yeasts for the future. FEMS Yeast Research 8, 979–995. González, S.S., Barrio, E., Querol, A., 2007. Molecular identification and characterization of wine yeast isolated from Tenerife (Canary Island, Spain). Journal of Applied Microbiology 102, 1018–1025. Lopes, C.A., Rodríguez, M.E., Sangorrín, M., Querol, A., Caballero, A.C., 2007. Patagonian wines: implantation of an indigenous strain of Saccharomyces cerevisiae in fermentations conducted in traditional and modern cellars. Journal of Industrial Microbiology & Biotechnology 34, 139–149. López, V., Fernández-Espinar, M.T., Barrio, E., Ramón, D., Querol, A., 2003. A new PCRbased method for monitoring inoculated wine fermentations. International Journal of Food Microbiology 81, 63–71. Manzano, M., Medrala, D., Giusto, C., Bartolomeoli, I., Urso, R., Comi, G., 2006. Classical and molecular analyses to characterise commercial dry yeast used in wine fermentations. Journal of Applied Microbiology 100, 599–607. Pérez-Coello, M.S., Briones Pérez, A.I., Ubeda Iranzo, J.F., Martin Alvarez, P.J., 1999. Characteristics of wines fermented with different Saccharomyces cerevisiae strains isolated from the La Mancha region. Food Microbiology 16, 563–573. Querol, A., Barrio, E., Ramón, D., 1992. A comparative study of different methods of yeast strain characterization. Systematic and Applied Microbiology 15, 439–446. Raspor, P., Cus, F., Povhe Jemec, K., Zagorc, T., Cadez, N., Nemanic, J., 2002. Yeast population dynamics in spontaneous and inoculated alcoholic fermentations of Zametovka must. Food Technology and Biotechnology 40, 95–102. Rodríguez, M.E., Infante, J.J., Molina, M., Domínguez, M., Rebordinos, L., Cantoral, J.M., 2010. Genomic characterization and selection of wine yeast to conduct industrial fermentations of a white wine produced in a SW Spain winery. Journal of Applied Microbiology 108, 1292–1302. Santamaría, P., Garijo, P., López, R., Tenorio, C., Gutiérrez, A.R., 2005. Analysis of yeast population during spontaneous alcoholic fermentation: effect of the age of the cellar and practice of inoculation. International Journal of Food Microbiology 103, 49–56. Schuller, D., Valero, E., Dequin, S., Casal, M., 2004. Survey of molecular methods for the typing of wine yeast strains. FEMS Microbiology Letters 231, 19–26.