INTRODUCTION. Milano, Italy. Italy

Microbiology (2015), 161, 362 373 DOI 10.1099/mic.0.000004 The vintage effect overcomes the terroir effect: a three year survey on the wine yeast biodiversity in Franciacorta and Oltrepò Pavese, two northern Italian vine-growing areas Ileana Vigentini, 1 3 Gabriella De Lorenzis, 2 3 Vincenzo Fabrizio, 1 Federica Valdetara, 1 Monica Faccincani, 3 Carlo Alberto Panont, 4 Claudia Picozzi, 1 Serena Imazio, 5 Osvaldo Failla 2 and Roberto Foschino 1 Correspondence Roberto Foschino roberto.foschino@unimi.it 1 Department of Food, Environmental and Nutrition Sciences, Università degli Studi di Milano, Milano, Italy 2 Department of Agricultural and Environmental Sciences, Università degli Studi di Milano, Milano, Italy 3 Consorzio per la Tutela del Franciacorta, Erbusco, Brescia, Italy 4 ex Consorzio Tutela Vini Oltrepò Pavese, Torrazza Coste, Pavia, Italy 5 Department of Life Sciences, Università degli Studi di Modena e Reggio Emilia, Italy Received 27 September 2014 Accepted 10 November 2014 A three year survey on the dominant yeast populations in samples of air, must and wine in different vineyards and cellars of two northern Italian vine-growing territories (six sites in Franciacorta and eight sites in Oltrepò Pavese areas) was carried out. A total of 505 isolates were ascribed to 31 different species by RFLP analysis of the ITS1 5.8SrRNA ITS2 region and partial sequence analysis of the 26S rrna gene. The most commonly found species were Saccharomyces cerevisiae (frequency, F9558.7 %; incidence, I 553.5 %), Hanseniaspora uvarum (F9514.3 %; I955.3 %), Metschnikowia fructicola (F9511.1 %; I 55.0 %) and Torulaspora delbrueckii (F9510.3 %; I 53.8 %). Among 270 S. cerevisiae new isolates, 156 (57.8 %) revealed a different genetic pattern through polymorphism analysis of the interdelta regions by capillary electrophoresis, while 47 isolates (17.4 %) were clones of starter cultures. By considering the Shannon Wiener index and results of principal component analysis (PCA) analyses, the year of isolation (vintage) proved to be a factor that significantly affected the biodiversity of the yeast species, whereas the geographical site (terroir) was not. Seventy-five per cent of S. cerevisiae isolates gathered in a unique cluster at a similarity level of 82 %, while the remaining 25 % were separated into minor groups without any evident relationship between d-pcr profile and territory, year or source of isolation. However, in six cases a similar strain appeared at the harvesting time both in Franciacorta and Oltrepò Pavese areas, whereas surprisingly no strain was reisolated in the same vineyard or cellar for consecutive years. INTRODUCTION In winemaking, yeasts are essential for the transformation of grape sugars into ethanol and carbon dioxide through alcoholic fermentation; nonetheless, due to their specific 3These authors contributed equally to this paper. Abbreviations: ADY, active dry yeast; CE, capillary electrophoresis; ITS, internal transcribed spacer; LSD, least significant difference; PCA, principal component analysis; UPGMA, unweighted pair group method with arithmetic means. Two supplementary figures are available with the online Supplementary Material. enzymic activities and cell autolysis, they can also generate typical sensorial characteristics in wine, like secondary flavours and smoothness (Romano et al., 2003a). Although selected Saccharomyces strains are usually added by oenologists as starter cultures to control the fermentative process, several micro-organisms enter the must from the vineyard environment, winery facilities and cellar equipment, and these can affect the quality of the end product. Nowadays, for a certain style of wines, the use of the so called autochthonous yeasts is considered essential in providing for the valorization and preservation of the environmental microbial biodiversity (Pretorius, 2000). In 362 000004 G 2015 The Authors Printed in Great Britain

Autochthonous yeasts may not exist in wine environments fact, it has been suggested that the land from which the grapes are grown imparts a unique quality to the wine, especially when spontaneous fermentations are carried out (Csoma et al., 2010; Di Maio et al., 2012). According to an OIV resolution (International Organization of Vine and Wine, 2010), terroir refers to an area in which collective knowledge of the interactions between the identifiable physical and biological environments and applied viticultural and oenological practices develops, giving distinctive characteristics for the products originating from this area. This assumption is at the base of the Appellation of Origin systems around the globe, with a strong impact on the wine market since they drive consumer choices. Many research groups have explored the opportunity to isolate, select and use indigenous strains with technological properties and quality traits as a strategic asset for wine-makers to unequivocally link a wine with its environment of production (Furdíkova et al., 2014; Martínez et al., 2004; Rodríguez- Palero et al., 2013; Settanni et al., 2012; Tosi et al., 2009; Tristezza et al., 2012). This experimental activity might also be interesting for the manufacturers of starter cultures, because it is the basis for genetic improvement and formulation of new products (Giudici et al., 2005; Mannazzu et al., 2002). However, the isolation of new strains with novel oenological properties is infrequent because different populations of yeasts naturally succeed each other during the must transformation (Torija et al., 2001; Vigentini et al., 2014) and they contribute in an unpredictable manner. Moreover, contamination with commercial starter cultures is possible and uncontrollable, since they are used for a long time and spread throughout the territory (Blanco et al., 2011; Cordero-Bueso et al., 2011a). Sometimes the phenomenon of codominance between autochthonous and commercial strains has been observed in spontaneous and controlled fermentations, with huge variation in the frequency of appearance of different yeast population (Beltran et al., 2002; Santamaria et al., 2008; Valero et al., 2005). The available commercial strains are numerous, and the same yeast is often called by different names or acronyms, depending on the manufacturers of the starter culture (Fernández-Espinar et al., 2001; Vigentini et al., 2009). An obvious consequence of this practice is the standardization of sensory characters resulting from the fermentation process, including in quality wines, at the expense of the originality of aromas that could be expressed from the same products by exploiting indigenous micro-organisms (Pretorius, 2000; Romano et al., 2003b). All of these considerations demonstrate that the debate on autochthonous yeasts is still open and, in any case, the development of analytical protocols for the accurate identification of strains is a requisite to demonstrating a possible link between wine and territory. The present work assesses the microbial diversity in oenological environments (vineyard air, must and base wine), in terms of yeast species and Saccharomyces sensu stricto group strains, through use of a large-scale sampling plan over a period of three consecutive years in fourteen different vineyards and respective cellars, located in two Lombard vine-growing areas (Franciacorta and Oltrepò Pavese). These territories are traditionally suited for the production of sparkling wines by the Champenoise method with the respective Italian designation of origin Franciacorta D.O.C.G and Oltrepò Pavese Metodo Classico D.O.C.G. METHODS Sampling and yeast isolation. Yeast isolates were collected during the 2009, 2010 and 2011 vintages in the Franciacorta (Brescia, Italy) and Oltrepò Pavese (Pavia, Italy) areas, which are more than 100 km apart and separated by the Po river. Each year, sampling was performed during July (air), August (must) and October (base wine) at the same places. Six vineyards (named A, B, C, D, E and F, with a maximum distance between them of 14 km) prevalently planted with Chardonnay cultivar and their related cellars in Franciacorta area, and eight vineyards (G, H, I, J, K, L, M and N, with a maximum distance among them of 18 km) prevalently planted with Pinot noir cultivar and their related cellars in Oltrepò Pavese area were involved (Fig. 1). In both territories, controlled fermentations were generally performed for the base wine production of sparkling wine by the Champenoise method through the inoculation of active dry yeast (ADY) cultures. Air sampling was carried out in July when the ripening of berries of both cultivars was advanced. In each vineyard, a 1 m 3 volume of air (two aliquots of 500 l in different points of the vineyard) was collected by an air sampler MAS 100 Eco (VWR International PBI) equipped with a support for Petri dishes. The instrument was manually transported between rows in the fields keeping it at a distance of approximately 30 cm from the plants and 1 1.2 m from the ground (at the height of the bunches). Yeast counts and isolation were carried out using YPD medium (10 g yeast extract l 21,20g peptone l 21, 20 g glucose l 21 ) combined with 18 g agar l 21, 200 mg K 2 S 2 O 5 l 21, 100 mg chloramphenicol l 21 and adjusted to ph 3.6 with tartaric acid. The plates were incubated at 25 uc in anaerobic conditions (GasPak system, Mikrobiologie Anaerocult A, Merck) for 5 days. Must sampling was carried out in August, at the wineries that were processing the grapes of the vineyards at which the air had previously been analysed; fresh juices were collected just after crushing, before treatment with sulphur dioxide. Finally, base wine sampling was performed in October, directly from the tanks when alcoholic fermentation of the relevant must was completed. Two 100 ml aliquots of must or wine sample for each winery were transferred to laboratories, storing them at 4 uc. Count and isolation of yeasts were carried out by culture techniques according to Vigentini et al. (2014). In addition, forty samples of commercial starter cultures (ADY) sold in Lombardy by six different brands were analysed in order to isolate and recognize whether starter strains were present in the collection of new isolates (Table 1). The choice of products was made after interviewing the involved winemakers. Yeast isolates were maintained at 280 uc in YPD broth with 20 % (v/v) glycerol. Yeast identification and biodiversity evaluation. The protocol for yeast DNA extraction was described by Vigentini et al. (2012). DNA samples were stored at 220 uc. A preliminary species identification of the isolates was carried out using RFLP analysis of the ribosomal internal transcribed spacer (ITS) 1 5.8S rrna gene region ITS 2 (ITS region) according to Esteve-Zarzoso et al. (1999). A T Gradient Biometra Thermocycler (Biometra) was used for DNA amplification. Approximately 2 mg of each PCR product was digested with Hin6I restriction endonuclease (Fermentas) according to the http://mic.sgmjournals.org 363

I. Vigentini and others Pavia Po River L N I F C A B Iseo Lake E D K J M G H Brescia Fig. 1. Site sampling of the 14 vineyards and related wineries, indicated by the letters A N, examined in 2009, 2010 and 2011 vintages located in the Franciacorta and Oltrepò Pavese areas. Bar, 100 km. supplier s instructions. The restriction fragments were separated and assessed by gel electrophoresis on 1.5 % agarose gels with TAE buffer (0.4M Tris/acetate, 0.01 M Na 2 EDTA, ph 8). Isolates showing the same restriction profile were grouped together and some samples per cluster were used for amplification and partial sequencing of the ITS region, as described above, or of the 26S rdna D1/D2 domain, using primers NL1 and NL4 (Spírek et al., 2003). Amplification products were sequenced by an outside provider (Primm, Milan, Italy). The sequences obtained were identified through the BLAST algorithm by comparison with sequences listed in databases (www.ncbi.nlm.nih.gov). A further analysis was performed for all yeast isolates attributed to Saccharomyces sensu stricto group to discriminate within oenological species, such as Saccharomyces bayanus, Saccharomyces cerevisiae and Saccharomyces pastorianus. A PCR protocol based on species-specific primers targeting the HO (homothallism) gene was used according to the protocol of de Melo Pereira et al. (2010). A measurement of the general diversity found in yeast communities was determined by calculating the richness of species (S ) and the Shannon Wiener index (H ) (Cordero Bueso et al., 2011a). Moreover, the frequency (F9) and the incidence (I ) were calculated according to Tristezza et al. (2013). ANOVA and Fisher s least significant difference (LSD) test were performed to assess the main effects of the investigated factors (Cordero Bueso et al., 2011a) using StatGraphics Centurion XVI statistical package. S. cerevisiae interdelta regions typing by capillary electrophoresis detection. The new isolates and those isolated from commercial starter cultures, which were attributed to S. cerevisiae strains, were subjected to typing by the amplification of interdelta (d- PCR) regions. Amplification was performed according to the protocol described by Legras & Karst (2003). The delta21 primer was 59-dyelabelled with 6-carboxyfluorescein (6-FAM; Primm) in order to detect amplification of the product in capillary electrophoresis (CE) (Tristezza et al., 2009). To determine the quality of d-pcr products, the amplification fragments were separated by electrophoresis in 2.0 % (w/v) agarose gels submitted to 100 V for 1.5 h in 16 TAE buffer and detected by ethidium bromide straining. The d-pcr products were then separated and detected by CE in an ABI Prism 310 Genetic Analyzer (Applied Biosystems Life Technologies) using POP-4 polymer, 310 Genetic Analyzer Buffer with EDTA, and a 47 cm650 mm capillary (Applied Biosystems). As reported by Vigentini et al. (2012), the samples were prepared in a solution of 0.9 ml ofd-pcr amplified fragments diluted 1 : 1000, 20 ml of formamide (Applied Biosystems) and 0.75 ml size standard GENESCAN-1200 LIZ (Applied Biosystems). The solution was incubated at 95 uc for 3 min in order to denature DNA fragments and cooled on ice for 5 min. The method used for CE was injection for 20 s at 1.5 kv and migration at 8 kv for 80 min with a run temperature of 60 uc. Data recording was performed using ABI Prism GeneMapper 3.7 software (Applied Biosystems) and fragments between 50 and 1200 bp were scored. Peaks showing a fluorescent intensity value less than 100 were not considered. The discrimination power of the CE protocol was estimated as follows: three independent DNA extractions of four different S. cerevisiae strains were performed and the genomic DNA was used as template in independent d-pcrs. The amplified fragments were then resolved in CE and d-pcr profiles were transformed in binary matrix (15presence, 05absence of amplified fragment) in order to display the similarity among replicates of each sample by cluster analysis. Based on the similarity matrix performed by GenAlEx 6.5 software (Peakall & Smouse 2006) using Dice s coefficient, a UPGMA (unweighted pair group method with arithmetic means) dendrogram was built using MEGA 5.0 software (Tamura et al. 2011). The lowest value of similarity that grouped the replicates of a same S. cerevisiae strain was used to determine the discrimination threshold among the yeast isolates. 364 Microbiology 161

http://mic.sgmjournals.org 365 Table 1. Distribution of the yeast species as number of isolates collected from air (a), must (m) and base wine (w) samples during the 2009, 2010 and 2011 vintages in two Lombard vine-growing areas Yeast species 2009 2010 2011 F I Franciacorta Oltrepò Pavese Franciacorta Oltrepò Pavese Franciacorta Oltrepò Pavese Aureobasidium pullulans 4 (a) 1 (a) 3.2 % 1.0 % Candida diversa 1 (m) 0.8 % 0.2 % Candida glabrata 1 (m) 0.8 % 0.2 % Candida parapsilosis 1 (m) 0.8 % 0.2 % Candida railenensis 14 (m), 4 (w) 1 (m) 9.5 % 3.8 % Candida zemplinina 1 (m) 1 (m) 1 (m) 3 (m) 4.8 % 1.2 % Cryptococcus flavescens 1 (m) 0.8 % 0.2 % Cryptococcus laurentii 4 (a) 1.6 % 0.8 % Hanseniaspora uvarum 1 (m) 13 (m), 2 (w) 4 (m), 2 (w) 1 (m) 5 (m) 14.3 % 5.5 % Hanseniaspora vinae 5 (m) 3.2 % 1.0 % Issatchenkia occidentalis 6 (m) 2 (m) 5 (m), 1 (w) 2 (m) 6.3 % 2.8 % Issatchenkia terricola 1 (a), 1 (m) 5 (m) 1 (a), 6 (m) 7.1 % 2.8 % Lachancea thermotolerans 2 (m) 1 (m) 2.4 % 0.6 % Meyerozyma guillermondii 1 (m) 0.8 % 0.2 % Metschnikowia fructicola 15 (m) 2 (m) 8 (m) 11.1 % 5.0 % Metschnikowia pulcherrima 3 (m) 1.6 % 0.6 % Pichia fermentans 1 (a), 5 (m), 3 (w) 6.3 % 1.8 % Pichia fluxuum 3 (w) 2 (m), 4 (w) 4.0 % 1.8 % Pichia kluyveri 10 (m) 1 (m) 4.0 % 2.2 % Pichia kudriavzevii 1 (w) 2 (m) 1.6 % 0.6 % Pichia manshurica 1 (w) 0.8 % 0.2 % Pichia membranefaciens 9 (w) 4 (w) 1 (m) 3 (w) 1 (m), 1 (w) 9.5 % 3.8 % Rhodotorula glutinis 4 (a) 3.2 % 0.8 % Rhodotorula graminis 2 (m) 0.8 % 0.4 % Rhodotorula nothofagi 1 (a) 0.8 % 0.2 % Saccharomyces cerevisiae 21 (m), 34 (w) 7 (a), 13 (m), 50 (w) 27 (m), 11 (w) 1 (a), 17 (m), 6 (w) 31 (m), 19 (w) 1 (a), 16 (m), 16 (w) 58.7 % 53.5 % Torulaspora delbrueckii 1 (w) 4 (m) 5 (m) 6 (m) 3 (m) 10.3 % 3.8 % Wickerhamomyces 3 (m) 1 (m) 2.4 % 0.8 % anomalus Zygoascus hellenicus 1 (m) 0.8 % 0.2 % Zygoascus bailii 10 (m), 1 (w) 2 (m), 1 (w) 1 (w) 2 (w) 1 (w) 8.7 % 3.6 % Zygoascus parabailii 1 (m), 1 (w) 1 (m) 2.4 % 0.6 % Autochthonous yeasts may not exist in wine environments

I. Vigentini and others Data analysis. To investigate a possible correlation between species diversity and abiotic factors such as year, area or material of isolation, a principal component analysis (PCA; Price et al. 2006) was performed by using GenAlEx 6.5 software. The dataset was created on the frequencies of the single species found in the air and must samples; the isolates collected from wine samples were excluded from the calculation, since S. cerevisiae species is normally added with ADY cultures to control fermentation. As regards the diversity within S. cerevisiae species, the d-pcr profile per each isolate obtained by CE was transformed in binary matrix. To display the genetic similarity among isolates and discriminate the strains, a cluster analysis was carried out, as already described above for the assessment of the discrimination power of the CE protocol. Moreover, the data obtained from the isolates that were not clones of commercial strains were subjected to PCA (Price et al., 2006) by GenAlEx 6.5 software (Peakall & Smouse, 2006), starting with the SSR correlation matrix. The SPSS 21.0 program (SPSS) was used to display a 2D representation of genetic relationship among genotypes. A structure analysis was performed using STRUCTURE 2.3.1 software package (Pritchard et al., 2000). To estimate the K value (K number of ancestral genetic groups), the algorithm was run 10 times per each K value from 1 to 10, using the following parameters: 200 000 burn-in steps (i.e. an initiation phase without result recording) and 1 000 000 steps for data acquisition, assuming admixture. According to Evanno et al. (2005), the K value was chosen estimating the highest value of DK as the mean of the absolute values of L (K) averaged over 10 runs, divided by the SD of L(K), DK5m[L(K+1)22 L(K)+ L(K21)] s 21 [L(K)]. L(K) is an estimate of the posterior probability of the data for a given K; L (K) is the rate of change of the likelihood function with respect to K [L (K)5L(K)2L(K21)]; L (K) is the absolute value of the difference between successive values of L (K). Based on the percentage of membership assignment, each sample is probabilistically assigned to a population (membership values higher than 0.8), or jointly to two or more populations indicating admixture. RESULTS Evaluation of species biodiversity A total of 505 yeast isolates (278 from Franciacorta and 227 from Oltrepò Pavese areas) were ascribed to 31 different species (Table 1). In particular, 26 isolates found in the air samples were attributed to seven yeast species, 297 isolates collected from the must samples belonged to 26 species and 182 isolates retrieved in base wine samples were representative of 12 species. Since the RFLP analysis of the ITS region, and the partial sequencing of ITS region or 26S rdna gene, may be insufficient to discriminate among the oenological species of Saccharomyces sensu stricto group, an investigation on the HO gene by species-specific PCR-based assay was performed. Results revealed that all 270 Saccharomyces isolates could be ascribed to S. cerevisiae, as they generated a single amplicon of approximately 400 bp with ScHO set primers, while no amplification products were obtained with LgHO primer pairs (de Melo Pereira et al., 2010). In general, the yeast species most frequently isolated and with the highest incidence were S. cerevisiae (F9558.7 %, I 553.5 %), Hanseniaspora uvarum (F9514.3 %, I 55.5 %), Metschnikowia fructicola (F9511.1 %, I 55.0 %) and Torulaspora delbrueckii (F9510.3 %, I 53.8 %) (Table 1), in accordance with recent literature (Barata et al., 2012; Zott et al., 2010). The air sampling was carried out in the same vineyards every year after the grape veraison, depending on the cultivar. Following our protocol, 27 samples did not reveal the presence of yeasts in 1 m 3, whereas 15 samples showed a log mean count value of 0.95±0.45 per m 3, with different isolation rates for some species, such as S. cerevisiae (F9511.9 %, I 534.6 %), Aureobasidium pullulans (F95 9.5 %, I 519.2 %), Rhodotorula glutinis (F959.5 %, I 5 15.4 %) and Cryptococcus laurentii (F954.8 %, I 515.4 %). The climatic conditions, rainfall and phytosanitary treatments (Cordero-Bueso et al., 2011a; Milanović et al., 2013) probably affect the presence of fungi in the air, since clear differences were found throughout territories and vintages: in particular, S. cerevisiae was consistently isolated in the air of Oltrepò Pavese vineyards, while no yeast was collected in 2010 and 2011 in the Franciacorta area. The musts were sampled immediately after the drainage of the grape juice and before the addition of sulfur dioxide, in order to collect samples at the harvesting time from a large mass of raw materials, deriving from the vineyard where the air sampling was done. The log mean value of yeast counts was 6.25±0.66 per ml, with most isolates belonging to S. cerevisiae (F9581.0 %, I 542.1 %), H. uvarum (F95 35.7 %, I 58.1 %), Met. fructicola (F9533.3 %, I 58.4 %), T. delbrueckii (F9528.6 %, I 56.1 %) and Candida railenensis species (F9519.0 %, I 55.1 %). The sampling of wines was performed at the end of the alcoholic fermentation from the tanks in which the relevant must, previously analysed, was processed. The log mean count was 3.75±0.93 per ml. The reason for this low value is because the samples were taken from the top of the tanks without mixing the vinous mass to meet the precautions intended by winemakers. As expected, S. cerevisiae was the predominant species (F9583.3 %, I 574.7 %); interestingly, the isolation of minor species like Pichia membranefaciens (F9523.8 %, I 59.3 %), Pichia fluxuum (F959.5 %, I 53.8 %) and Zygosaccharomyces bailii (F959.5 %, I 5 3.3 %) were shown to be detectable at an ethanol level.10.5 %, v/v, and total SO 2.40 mg l 21. The presence of some yeast species was discontinuous along the timeframe studied. Only S. cerevisiae was found during all three investigated years in both territories, whereas Candida zemplinina, H. uvarum, Issatchenkia occidentalis, P. membranefaciens, T. delbrueckii and Z. bailii were detected for at most two years. Excluding the species represented by one isolate, Cryptococcus laurentii, Hanseniaspora vinae, Pichia kudriavzevii and Rhodotorula graminis were recovered only in the Franciacorta area at different vintages, while Metschnikowia pulcherrima, Pichia fermentans and R. glutinis were detected only in the Oltrepò Pavese area, but all of these yeasts were not found in more than one year. Forty-eight isolates were obtained from ADY samples and all of them were ascribed to S. cerevisiae species. 366 Microbiology 161

1 0 1 15.49 % 2 3 Autochthonous yeasts may not exist in wine environments On the basis of the number of identified species deriving from air and must samples, the richness (S) and the Shannon Wiener indexes (H ) were calculated. The isolates collected from base wine samples and those recognized as commercial starter cultures (see next paragraph) were omitted from the calculation, because they would have introduced a bias in the biodiversity assessment. The richness index in the Franciacorta area was higher than that found in Oltrepò Pavese, since 25 vs 20 different yeast species were isolated, respectively. Although data from the Franciacorta territory showed a level of yeast biodiversity (H 52.38) greater than those from Oltrepò Pavese (H 52.08), ANOVA revealed that this difference was not significant (P50.62). Conversely, the vintage proved to be a factor that significantly affected the yeast biodiversity since the values of the Shannon Wiener index determined for each year were different (P,0.05); in particular, the H calculated in 2010 (H 52.38) was different from those of 2009 (H 51.26) and 2011 (H 51.54) vintages, by applying the LSD test (at 95 %). In order to uncover relationships between microbial diversity and vintage or terroir, data describing 30 species frequencies in different years, areas or materials of isolation were processed by PCA. The first three principal components explained over 66 % of the total variability (Fig. 2). In the scatter plot, a distribution of the samples according to the year of isolation was noticed. In particular, the data from the samples collected in 2009 were grouped in the negative quadrant of the first two coordinates, while those observed in 2010 proved to be the most dispersed ones, confirming the results previously pointed out by the Shannon Wiener index. On the contrary, a clear connection among frequencies of yeast species and vine-growing area or material of isolation was not found (data not shown). S. cerevisiae strain diversity by d-pcr typing A preliminary evaluation of the discriminating power of the d-pcr typing technique in our experimental conditions was carried out. The analysis of electrophoretic profiles allowed determination of a size range of 2 bp as a sensibility threshold of the molecular marker used. Three independent d-pcrs per four different S. cerevisiae strains were performed. In the UPGMA dendrogram obtained (Fig. S1, available with the online Supplementary Material), the replicates of each strain were grouped correctly in the same cluster; particularly, similarity values within the cluster ranged from 98.0 % to 100 %. Therefore, the lowest value of similarity grouping the replicates of the same yeast (98.0 %) was assumed to be the threshold limit of the protocol to separate strains. In other words, isolates having a similarity value higher than this limit were considered clones of the same strain. The new isolates (270) and those obtained from starter cultures (48), that were previously identified as S. cerevisiae, were analysed by the d-pcr typing protocol. For each sample, a clear and unambiguous amplification pattern was generated and separated by CE. A total of 220 polymorphic d-pcr products were scored among new and commercial S. Year 2009 2010 2011 1 0 1 2 38.01 % 3 4 0 1 2 1 3 2 12.94 % Fig. 2. Scatter plot obtained from PCA applied to 299 yeast isolates collected from air and must samples during the 2009, 2010 and 2011 vintages in the Franciacorta and Oltrepò Pavese areas. The axes represent the first three principal components (PCs) and the variability explained by each PC is reported as a percentage of the total variability. http://mic.sgmjournals.org 367

I. Vigentini and others cerevisiae strains, in the range of 50 1200 bp. The number of fragments detected ranged from 17 to 42 for each isolate, with an average of 21. No monomorphic product was found, while a 460 bp fragment was revealed to be common to 280 isolates out of 318 (88.1 %). In Fig. 3 the UPGMA dendrogram obtained with the d-pcr profiles from all S. cerevisiae isolates is shown. In general, the range of genetic similarity based on the Dice s coefficient ranged between 73 100 %. One-hundred and sixty-two isolates could be considered clones of some strains, since they showed similarity values higher than the determined discrimination limit (98.0 %); on the other hand, 156 amplification patterns were distinctive and were attributed to a single strain. A proportion accounting for 75 % of S. cerevisiae isolates, including 94 % of the starter cultures investigated, grouped in a large unique cluster at a similarity level of 82 %. The remaining 25 % of isolates were separated in minor groups, some of them represented by one strain. No specific clusters could be ascribed to the geographical origin or year of isolation, but samples were casually distributed without any evident relationship between d-pcr profile and territory, year or matrix of isolation. Out of 48 isolates from starter cultures, only 18 were recognized as distinct d-pcr patterns, indicating that the manufacturers of commercial ADY often replicate the same strains or those of their competitors, by changing the names. Furthermore, among the new isolates, 47 (17.4 %) revealed a d-pcr profile undistinguishable from starter yeasts (Table 2). The most common electrophoretic profile was shared by 23 isolates: seven from commercial cultures of different producers, six isolated during the 2009 vintage in Franciacorta, three during the 2009 vintage in Oltrepò Pavese, four during the 2010 vintage in Franciacorta and three during the 2011 vintage in Franciacorta. Frequently, isolates showing an identical d-pcr pattern were collected from samples isolated from the same or neighbouring vineyards/cellars. Nevertheless, an identical amplification profile was identified in the same vintage in Franciacorta and Oltrepò Pavese in six cases (Fig. 3): the isolates 780 (cellar B), 768 (D), 539 (E), 781 (J) and 783 (K) in 2009; the isolates 788 and 790 (C), 787 and 789 (cellar F), 793 and 522 (I), 791 and 510 (K), in 2009; the isolates 813 (B), 814 and 71 (H), 521 (J), 815 and 816 (L), 817-821 and 823 (M) in 2009; the isolates 848 (C) and 849 (N) in 2009; the isolates 637 (F) and 639 (J) in 2010; and the isolates 667 (B), 668 (E), 669 (H) and 670 (G) in 2010 (Table 2). On the other hand, in only two cases an undistinguishable d-pcr pattern appeared in different years within the same vine-growing area (Franciacorta): these were the isolates 782 (B) and 705 (F) present in 2009 and 632 (E) found in 2010; and the isolates 637 (F) present in 2010 and 638 (A) found in 2011. Interestingly, some S. cerevisiae strains collected from the air (the isolates 618 at vineyard K in 2009, 686 at vineyard N in 2010 and 585 at vineyard I in 2011) were then found in must samples picked up at neighbouring cellars in the same vintage (the isolates 620 at cellar G in 2009; 687 and 689 at cellar K, 688 at cellar J in 2010; and 776 at cellar in 2011, respectively). In only one case the same strain was identified both in the air (829 and 830 at vineyard L in 2009) and in wine (828 at cellar I in 2009). Analysis of population structure The genetic structure of 223 S. cerevisiae isolates was analysed using STRUCTURE software. Multiple analyses were performed using a varying number of assumed populations (K51 10). A graphic representation of the most appropriate groups inferred in this dataset is reported in Fig. 4. The maximum resolution was achieved at K52. A proportion accounting for 81.6 % of samples showed an average estimated major membership proportion.0.8. Therefore, according to their largest ancestry membership fraction, the yeast strains could be classified as mainly belonging to one of two distinct genetic groups. The other 18.4 % was considered admixed. The ancestral groups 1 (G1) and 2 (G2) clustered 11.5 % and 88.5 %, respectively, of S. cerevisiae isolates showing a membership proportion higher than 0.8. In G1, samples collected during the 2011 vintage and from air were not grouped. Considering each variable, G1 clustered 3.20 % and 22.50 % of total samples collected from must and wine, respectively, 22.70 % and 2.70 % of yeasts isolated during the 2009 and 2010 vintage, respectively, and 7.30 % and 15.20 % of isolates collected in Franciacorta and Oltrepò Pavese areas were also clustered. The composition of G2 was very dissimilar to G1, grouping the total samples collected during the 2011 vintage and from air, 96.80 % and 77.50 % of samples were collected from must and wine, respectively, 77.30 % and 97.30 % of yeasts were isolated during the 2009 and 2010 vintage, respectively, and 92.70 % and 84.8 % of isolates were collected in Franciacorta and Oltrepò Pavese. As observed for the PCA data (Fig. S2), the structure analysis discriminated the samples into two groups (a major and a minor group), where the first two coordinates, accounting for about 50 % of total variability, grouped isolated strains into a main group independently of their origin, consistent with the observed gene flow between the two geographical areas among matrices (air, must and wine) and vintages (2009, 2010 and 2011). Likewise, the PCA distribution did not cluster samples collected during the 2011 vintage and from air in the minor group. DISCUSSION Some considerations about the experimental approach adopted in this work have to be noted. Firstly, the different protocols applied to sample air, must and base wine, and the plating technique used for the enumeration and isolation of the yeasts, do not allow accurate determination of the actual biodiversity present in the different materials or to make comparisons among them. However, the attention that has been spent maintaining the same places and procedures of sampling corroborates the value of the obtained results and their representativeness as concerns the permanence of species and strains in fourteen different sites for three consecutive years. In particular, the growth conditions of the agar plates to collect micro-organisms in 368 Microbiology 161

647 732781 731 730 729 Autochthonous yeasts may not exist in wine environments 681 601 603 849 848 743 885 876 875 799 901 830 829 828 841 625 755 727 728 614 602 756 825 662 800 1044 802 792 623 695 620 618 733 617 616 612 611 851 519 607 606 808 807 806 673 676 902 773 701 700 747 746 624 812 657 889 888 887 886 884 724 578 716 525 694 683 672 655 654 741 742 740 844 748 891 865 867 866 675 674 659 896 893 863 515 852 822 615 530 581 846 709 622 708 604 702 634 836 703 633 837 704 559 706 726 707 839 862 758 874 847 833 864 835 660 533 813 834 814 840 815 737 816 795 794 838 817 818 819 900 820 821 521 71 823 787 788 789 790 791 510 522 793 685 850 768 539 5 780 783 868 646 665 627 628 803 682 796 771 690 691 899 770 898 778 853 590 697 763 696 762 752 677 751 678 845 797 834 679 824 680 735 544 609 648 608 719 599 720 869 861 870 860 871 632 872 782 873 705 684 811 775 810 663 809 658 895 671 621 692 670 693 669 668 667 641 750 642 749 643 637 644 638 645 710 639 713 714 711 712 698 699 753 754 759 761 785 511 805 786 784 798 666 769 801 883 777 760 736 881 882 738 739 879 880 718 757 877 878 717 715 626 723 832 842 843 890 892 661 721 722 656 894 767 651 652 653 779 772 587 664 629 640 649 650 831 858 859 857 856 631 855 585 776 600 610 619 630 744 745 686 687 688 689 636 897 635 765 774 804 764 766 854 Fig. 3. UPGMA dendrogram obtained by cluster analysis of interdelta region profiles of 270 S. cerevisiae isolates collected from air, must and wine samples during the 2009, 2010 and 2011 vintages in the Franciacorta and Oltrepò Pavese areas, and of 48 S. cerevisiae isolated from commercial strains. The 98.0 % value (dashed line) represents the discrimination threshold (scale bar corresponds to 5 % similarity). The empty triangles represent the wild isolates; the filled circles represent the strains isolated from starter cultures. air were highly selective, so that an underestimation of the actual abundance of yeasts may have occurred. These selective conditions were preferred because preliminary tests had shown the invasive growth of moulds on the plates, which did not allow detection of yeast colonies. Nevertheless, contrary to what described by Garijo et al. (2011), S. cerevisiae species were isolated from the air of vineyards for the first time by adopting this protocol. Secondly, as shown by previous ecological studies on wine yeast biodiversity, the discrimination within the non- Saccharomyces taxa is limited at the species level, whereas it would be interesting to deepen knowledge at the strain level, as has been recently explored (Pfliegler et al. 2014; Vigentini et al., 2012). Regarding the interdelta regions typing by CE to detect differences in S. cerevisiae strains, the protocol set up by Tristezza et al. (2009) was confirmed http://mic.sgmjournals.org 369

370 Microbiology 161 Table 2. Distribution of S. cerevisiae isolates collected from air, must and wine samples during 2009, 2010 and 2011 vintages in two Lombard vine-growing areas The isolates underlined are those recognized as commercial starter strains by d-pcr typing technique. Year of isolation Isolates in Franciacorta area Isolates in Oltrepò Pavese area Vineyard/cellar Air Must Wine Vineyard/cellar Air Must Wine 2009 A 610, 619, 672, 525, 797, 798, 825, 902 694, 748, 795, 796, 837, 839, 844, 847, 850 G 620, 621, 623 800, 606, 607, 611, 617, 622, 624 B 782, 661, 726, 833, 836 780, 813, 737, 559 H 511 807, 808, 809, 810, 811, 814, 71, 612 C 640, 650, 790, 824, 831, 758, 788, 799, 843, 604, I 842, 793, 522 599, 832, 828 848 530 D 900, 901 683, 768, 794, 838, 841 J 792, 802, 804 781, 521, 625 E 812, 834, 835 539 K 601, 515, 615, 618 791, 510, 608 783 F 705 629, 787, 789, 533, 840, 603, L 829, 830 609 784, 785, 806, 815, 816 M 1044 716, 820, 519 600, 786, 817, 818, 819, 821, 823 N 616 822, 803, 845, 846, 849, 851 2010 A 662, 677 654, 655, 656, 657, 678 G 670 B 544, 660, 667 H 646, 675, 692 669, 691 C 627, 631, 658, 659, 663, 681 I 628, 643, 644, 645, 685 665, 671, 682 664, 693 D 635, 652, 653, 679, 680, J 634, 639, 647, 688, 633 690, 695 E 632, 641, 651, 666, 668, K 687, 689 673, 684 F 637, 642, 648, 649, 676 630, 636 L 805 801, 602, 626 M 614 N 686 674 2011 A 638, 696, 697, 729, 730, 578, 724, 727, 728, 755 G 777 734, 590, 763 744, 745, 746, 752 B 710, 731, 732, 751 723, 756, 581 H 772, 774, 778 740, 741 C 700, 701, 750 717, 718, 757 I 585 767 736, 743 D 713, 719, 722, 754 703, 704, 706, 715 J 770, 771 742, 769 E 698, 711, 747, 749 702, 709 K 733, 775, 779 762, 766 F 699, 712, 714, 720, 721, 753 707, 708, L 735 759 M 773, 776, 587 739, 761 N 764, 765 738, 760 Starter Pascal Biotech, AEB group 899, 856, 857, 858, 859, 860, 861, 865, 868, 871, 876, 877, 879, 880, 883, 892 cultures Anchor Yeast, Intec 862, 864, 866, 869, 881, 884, 886, 889, 891, 894 I. Vigentini and others

Autochthonous yeasts may not exist in wine environments 1.0 0.8 0.6 0.4 0.2 0 Table 2. cont. Isolates in Franciacorta area Isolates in Oltrepò Pavese area Year of isolation Vineyard/cellar Air Must Wine Vineyard/cellar Air Must Wine DSM, Corimpex 898, 867, 870, 878, 882, 887, 888, 890 Laffort 895, 896 Lallemand 863, 874, 885, 897 Institut Oenologique de Champagne, Perdomini 852, 853, 854, 855, 872, 873, 875, 893 G1 Fig. 4. Inferred population structure for K52 obtained by d-pcr profiles of 223 S. cerevisiae strains using the model-based program STRUCTURE. G1 and G2 indicate two estimated ancestral genetic clusters. The y-axis shows the estimated membership proportion. Each sample is displayed as a vertical line divided into differently coloured segments representing the estimated membership proportions in the two ancestral genetic clusters, where G1 is red and G2 is green. to be reliable and robust, even if the index of discrimination obtained from our results was lower (98.7 %) than that measured by the original authors (99.8 %). Commercial yeasts are easily traceable, thanks to the application of this technique. The exploitation of the indigenous microbiota is a current topic in food quality management, in particular when spontaneous fermentations are involved in wine processing (Di Maio et al., 2012; Settanni et al., 2012). Such a trend supports safeguarding of the diversity of the local products, and has a noteworthy impact in trade, especially by those consumers seeking typical foods. In this context, the recognition and preservation of micro-organisms present in vineyards and cellars is a strategic activity for the oenologist, who aims to express their own terroir in their products. Nevertheless, what exactly is meant by autochthonous yeast is still a controversial question among researchers: it is objectively assumed that a strain should originate in a site and persist in it for a certain period. However, where should we place the boundary line in determining the membership of a strain to a territory? Shall we look at the walls of the cellar, the hill planted with the vines, or at the land included in a valley or an island? Additionally, how long does it take to consider a strain as a native micro-organism of that area? Finally, the term indigenous refers to the ability to isolate the same strain identified in the course of different vintages. Recent works that monitored yeast populations for at least three years, and in which molecular techniques suitable for discrimination at strain level were used, are limited. Analysing the results presented in the work of Versavaud et al. (1995) on the distribution of S. cerevisiae strains in the wineproducing area of Charentes, it is possible to point out that several dominant strains frequently appeared in a specific year for different sites, and only a few of them could be retrieved in different vintages. Torija et al. (2001) reported that yeast populations were different every year, though some Saccharomyces strains were found over three consecutive years in the same cellar, indicating that a G2 http://mic.sgmjournals.org 371

I. Vigentini and others natural microbiota was consistently found in the wineries. Interestingly, the appearance of new strains within a same vintage occurred in different cellars located in regions geographically distant by about 50 km. Beltran et al. (2002), in a six year follow-up study, discovered that most of the new strains of S. cerevisiae were isolated in the same year, and that the predominance of a specific strain was more influenced by the vintage than by the grape variety or wine making process. Demuyter et al. (2004), which studied the dominance of S. uvarum during spontaneous fermentation in an Alsatian winery, also stated that the yeast populations on grapes changed from year to year, and that new non-resident strains appeared in the must during the winemaking process. Schuller & Casal (2007) demonstrated that the genetic differentiation among S. cerevisiae populations in the same vineyard in consecutive years showed a similar magnitude to the differences found among the different vineyards of the same vintage. The results obtained in our survey are consistent with those of the preceding works, confuting the current opinion that there are autochthonous strains that permanently reside in vineyards or cellars. In the samples collected during three consecutive years in two different Lombard territories, a high yeast biodiversity is recognized, either at intraspecific and at interspecific level, in the case of S cerevisiae species. The effect of the vintage overcomes the effect of the geographical area of isolation, since dominant species found in air and must change each year. Moreover, the analysis of yeast populations provided evidence of significant population structure within the sample, showing that the diversity of microbial communities could be affected by vintage. In addition, new strains of S. cerevisiae appear simultaneously in both territories although they are more than 100 km away from each other. In this regard, Stefanini et al. (2012) have recently highlighted the specific role of social wasps in the environmental dispersion of yeast cells and in the evolution of S. cerevisiae populations. Surprisingly, no strain was isolated in the same vineyard or cellar during different years, challenging the idea that yeast populations are stable in a wine environment over time (Ciani et al., 2004; Cordero-Bueso et al., 2011b). This outcome is also corroborated by the work of Bokulich et al. (2013), which observed seasonal fluctuations of the surface microbial communities in a winery equipment. Actually, compliance to the hygiene control system based on the Hazard Analysis Critical Control Points principles, through the rigorous application of cleaning and disinfection procedures, should counteract the permanence of microorganisms in oenological plants. ACKNOWLEDGEMENTS The authors gratefully thank Dr Rossana Tonesi for helpful advice, and the personnel of the wineries involved in this work for their precious collaboration, particularly the oenologists Giacomo Barbero, Stefano Capelli, Alice Colombo, Guido Gandossi, Michele Ferrari, Silvia Filisetti, Simone Fiori, Andrea Rossi, Luca Rossi, Alessandro Schiavi, Silvia Uberti, Giuseppe Vezzoli, Raffaello Vezzoli and Daniele Zangelmi, and the students Andrea Barbieri, Shirley Barrera Cardenas, Fabrizio Cadei and Federico Prevadini. This work was funded by Regione Lombardia through the program Piano per la ricerca e lo sviluppo, 2009, Direzione Generale Agricoltura, Regione Lombardia, Italy, research project no. 1315. REFERENCES Barata, A., Malfeito-Ferreira, M. & Loureiro, V. (2012). The microbial ecology of wine grape berries. Int J Food Microbiol 153, 243 259. Beltran, G., Torija, M. J., Novo, M., Ferrer, N., Poblet, M., Guillamón, J. M., Rozès, N. & Mas, A. (2002). Analysis of yeast populations during alcoholic fermentation: a six year follow-up study. Syst Appl Microbiol 25, 287 293. Blanco, P., Orriols, I. & Losada, A. (2011). Survival of commercial yeasts in the winery environment and their prevalence during spontaneous fermentations. J Ind Microbiol Biotechnol 38, 235 239. Bokulich, N. A., Ohta, M., Richardson, P. M. & Mills, D. A. (2013). Monitoring seasonal changes in winery-resident microbiota. PLoS ONE 8, e66437. Ciani, M., Mannazzu, I., Marinangeli, P., Clementi, F. & Martini, A. (2004). Contribution of winery-resident Saccharomyces cerevisiae strains to spontaneous grape must fermentation. Antonie van Leeuwenhoek 85, 159 164. Cordero-Bueso, G., Arroyo, T., Serrano, A., Tello, J., Aporta, I., Vélez, M. D. & Valero, E. (2011a). Influence of the farming system and vine variety on yeast communities associated with grape berries. Int J Food Microbiol 145, 132 139. Cordero-Bueso, G., Arroyo, T., Serrano, A. & Valero, E. (2011b). Remanence and survival of commercial yeast in different ecological niches of the vineyard. FEMS Microbiol Ecol 77, 429 437. Csoma, H., Zakany, N., Capece, A., Romano, P. & Sipiczki, M. (2010). Biological diversity of Saccharomyces yeasts of spontaneously fermenting wines in four wine regions: comparative genotypic and phenotypic analysis. Int J Food Microbiol 140, 239 248. de Melo Pereira, G. V., Ramos, C. L., Galvão, C., Souza Dias, E. & Schwan, R. F. (2010). Use of specific PCR primers to identify three important industrial species of Saccharomyces genus: Saccharomyces cerevisiae, Saccharomyces bayanus and Saccharomyces pastorianus. Lett Appl Microbiol 51, 131 137. Demuyter, C., Lollier, M., Legras, J.-L. & Le Jeune, C. (2004). Predominance of Saccharomyces uvarum during spontaneous alcoholic fermentation, for three consecutive years, in an Alsatian winery. J Appl Microbiol 97, 1140 1148. Di Maio, S., Polizzotto, G., Di Gangi, E., Foresta, G., Genna, G., Verzera, A., Scacco, A., Amore, G. & Oliva, D. (2012). Biodiversity of indigenous Saccharomyces populations from old wineries of southeastern Sicily (Italy): preservation and economic potential. PLoS ONE 7, e30428. Esteve-Zarzoso, B., Belloch, C., Uruburu, F. & Querol, A. (1999). Identification of yeasts by RFLP analysis of the 5.8S rrna gene and the two ribosomal internal transcribed spacers. Int J Syst Bacteriol 49, 329 337. Evanno, G., Regnaut, S. & Goudet, J. (2005). Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol 14, 2611 2620. Fernández-Espinar, M. T., López, V., Ramón, D., Bartra, E. & Querol, A. (2001). Study of the authenticity of commercial wine yeast strains by molecular techniques. Int J Food Microbiol 70, 1 10. 372 Microbiology 161

Autochthonous yeasts may not exist in wine environments Furdíkova, K., Makysova, K. & Durcanska, K. (2014). Influence of yeast strain on aromatic profile of Gewürztraminer wine. LWT - Food Sci Technol 59, 256 262. Garijo, P., López, R., Santamaría, P., Ocon, E., Olarte, C., Sanz, S. & Gutiérrez, A. R. (2011). Presence of enological microorganisms in the grapes and the air of a vineyard during the ripening period. Eur Food Res Technol 233, 359 365. Giudici, P., Solieri, L., Pulvirenti, A. M. & Cassanelli, S. (2005). Strategies and perspectives for genetic improvement of wine yeasts. Appl Microbiol Biotechnol 66, 622 628. International Organization of Vine and Wine (2010). Definition of vitivinicultural terroir. OIV/VITI 333/2010 Resolution, Tbilisi, 25th June 2010. Legras, J.-L. & Karst, F. (2003). Optimisation of interdelta analysis for Saccharomyces cerevisiae strain characterisation. FEMS Microbiol Lett 221, 249 255. Mannazzu, I., Clementi, F. & Ciani, M. (2002). Strategies and criteria for the isolation and selection of autochthonous starters. In Biodiversity and Biotechnology of Wine Yeasts, pp. 19 33. Edited by M. Ciani. Kerala: Research Signpost. Martínez, C., Gac, S., Lavín, A. & Ganga, M. (2004). Genomic characterization of Saccharomyces cerevisiae strains isolated from wine-producing areas in South America. J Appl Microbiol 96, 1161 1168. Milanović, V., Comitini, F. & Ciani, M. (2013). Grape berry yeast communities: influence of fungicide treatments. Int J Food Microbiol 161, 240 246. Peakall, R. & Smouse, P. E. (2006). GENALEX 6: Genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6, 288 295. Pfliegler, W. P., Horváth, E., Kállai, Z. & Sipiczki, M. (2014). Diversity of Candida zemplinina isolates inferred from RAPD, micro/minisatellite and physiological analysis. Microbiol Res 169, 402 410. Pretorius, I. S. (2000). Tailoring wine yeast for the new millennium: novel approaches to the ancient art of winemaking. Yeast 16, 675 729. Price, A. L., Patterson, N. J., Plenge, R. M., Weinblatt, M. E., Shadick, N. A. & Reich, D. (2006). Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet 38, 904 909. Pritchard, J. K., Stephens, M. & Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics 155, 945 959. Rodríguez-Palero, M. J., Fierro-Risco, J., Codón, A. C., Benítez, T. & Valcárcel, M. J. (2013). Selection of an autochthonous Saccharomyces strain starter for alcoholic fermentation of Sherry base wines. J Ind Microbiol Biotechnol 40, 613 623. Romano, P., Fiore, C., Paraggio, M., Caruso, M. & Capece, A. (2003a). Function of yeast species and strains in wine flavour. Int J Food Microbiol 86, 169 180. Romano, P., Caruso, M., Capece, A., Lipani, G., Paraggio, M. & Fiore, C. (2003b). Metabolic diversity of Saccharomyces cerevisiae strains from spontaneously fermented grape musts. World J Microbiol Biotechnol 19, 311 315. Santamaria, P., López, R., López, E., Garijo, P. & Gutiérrez, A. R. (2008). Permanence of yeast inocula in the winery ecosystem and presence in spontaneous fermentations. Eur Food Res Technol 227, 1563 1567. Schuller, D. & Casal, M. (2007). The genetic structure of fermentative vineyard-associated Saccharomyces cerevisiae populations revealed by microsatellite analysis. Antonie van Leeuwenhoek 91, 137 150. Settanni, L., Sannino, C., Francesca, N., Guarcello, R. & Moschetti, G. (2012). Yeast ecology of vineyards within Marsala wine area (western Sicily) in two consecutive vintages and selection of autochthonous Saccharomyces cerevisiae strains. J Biosci Bioeng 114, 606 614. Spírek, M., Yang, J., Groth, C., Petersen, R. F., Langkjaer, R. B., Naumova, E. S., Sulo, P., Naumov, G. I. & Piskur, J. (2003). High-rate evolution of Saccharomyces sensu lato chromosomes. FEMS Yeast Res 3, 363 373. Stefanini, I., Dapporto, L., Legras, J.-L., Calabretta, A., Di Paola, M., De Filippo, C., Viola, R., Capretti, P., Polsinelli, M. & other authors (2012). Role of social wasps in Saccharomyces cerevisiae ecology and evolution. Proc Natl Acad Sci U S A 109, 13398 13403. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S. (2011). MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28, 2731 2739. Torija, M. J., Rozès, N., Poblet, M., Guillamón, J. M. & Mas, A. (2001). Yeast population dynamics in spontaneous fermentations: comparison between two different wine-producing areas over a period of three years. Antonie van Leeuwenhoek 79, 345 352. Tosi, E., Azzolini, M., Guzzo, F. & Zapparoli, G. (2009). Evidence of different fermentation behaviours of two indigenous strains of Saccharomyces cerevisiae and Saccharomyces uvarum isolated from Amarone wine. J Appl Microbiol 107, 210 218. Tristezza, M., Gerardi, C., Logrieco, A. & Grieco, F. (2009). An optimized protocol for the production of interdelta markers in Saccharomyces cerevisiae by using capillary electrophoresis. J Microbiol Methods 78, 286 291. Tristezza, M., Vetrano, C., Bleve, G., Grieco, F., Tufariello, M., Quarta, A., Mita, G., Spano, G. & Grieco, F. (2012). Autochthonous fermentation starters for the industrial production of Negroamaro wines. J Ind Microbiol Biotechnol 39, 81 92. Tristezza, M., Vetrano, C., Bleve, G., Spano, G., Capozzi, V., Logrieco, A., Mita, G. & Grieco, F. (2013). Biodiversity and safety aspects of yeast strains characterized from vineyards and spontaneous fermentations in the Apulia Region, Italy. Food Microbiol 36, 335 342. Valero, E., Schuller, D., Cambon, B., Casal, M. & Dequin, S. (2005). Dissemination and survival of commercial wine yeast in the vineyard: a large-scale, three-years study. FEMS Yeast Res 5, 959 969. Versavaud, A., Courcoux, P., Roulland, C., Dulau, L. & Hallet, J.-N. (1995). Genetic diversity and geographical distribution of wild Saccharomyces cerevisiae strains from the wine-producing area of Charentes, France. Appl Environ Microbiol 61, 3521 3529. Vigentini, I., Fracassetti, D., Picozzi, C. & Foschino, R. (2009). Polymorphisms of Saccharomyces cerevisiae genes involved in wine production. Curr Microbiol 58, 211 218. Vigentini, I., De Lorenzis, G., Picozzi, C., Imazio, S., Merico, A., Galafassi, S., Piškur, J. & Foschino, R. (2012). Intraspecific variations of the Intron Splice Site in Dekkera/Brettanomyces bruxellensis genome studied by capillary electrophoresis separation. Int J Food Microbiol 157, 6 15. Vigentini, I., Fabrizio, V., Faccincani, M., Picozzi, C., Comasio, A. & Foschino, R. (2014). Dynamics of Saccharomyces cerevisiae populations in controlled and spontaneous fermentations for Franciacorta D.O.C.G. base wine production. Ann Microbiol 64, 639 651. Zott, K., Claisse, O., Lucas, P., Coulon, J., Lonvaud-Funel, A. & Masneuf-Pomarede, I. (2010). Characterization of the yeast ecosystem in grape must and wine using real-time PCR. Food Microbiol 27, 559 567. Edited by: R. Oliver http://mic.sgmjournals.org 373