Lysine agar

The vitality of the strains after 3 months of bottling was carried out using viable cell counts on WL Nutrient Agar (Oxoid, Hampshire, UK) and Lysine Agar (Oxoid, Hampshire, UK). Lysine Agar is a medium unable to support the growth of S. cerevisiae (Lin, 1975) (link) for the differentiation of non-Saccharomyces yeast from S. cerevisiae strain. The media were incubated at 25 • C for 2-3 days.

From preliminary screening T. delbrueckii DiSVA 254 was selected and used in the pure and mixed fermentations with the S. cerevisiae US-05 starter strain at different S. cerevisiae to T. delbrueckii ratios (i.e., 1:1, 1:10, 1:20, respectively). A batch of 1,500 l of malted barley wort for the production of American Amber Ale was used in this study. Its main analytical characters were: pH 5.47; specific gravity 12.7 °Plato. The fermentation potential of the selected strain was evaluated in fermentation trials carried out at 20 °C in flasks containing 500 ml wort under sterile conditions. The flasks were locked with a Müller valve containing sulphuric acid, to allow only CO 2 to escape from the system. Pre-cultures were grown in 10% malt extract at 20 °C for 48 h. The fermentation kinetics were monitored by measuring the weight loss of the flasks due to the CO 2 evolution, which was followed to the end of the fermentation (i.e., constant weight for 3 consecutive days). The growth kinetics were monitored using viable cell counts on WL Nutrient Agar (Oxoid, Hampshire, UK) and Lysine Agar (Oxoid, Hampshire, UK) a selective medium unable to support the growth of S. cerevisiae (Lin, 1975) , for differentiation of the T. delbrueckii yeast from the S. cerevisiae starter strain. The fermentations were carried out in duplicate trials under static conditions.

The yeast strains used in this study were the commercial S. cerevisiae strain Lalvin EC1118 (Lallemand Inc.) and Z. florentina strain 42 from the Yeast Culture Collection of the Department of Agricultural Biotechnologies (GESAAF, University of Florence, Italy), which were maintained at -80 °C. The Z. florentina strain 42 was identified by D1-D2 domain analysis.
After revitalisation, these two yeast strains were routinely cultured on YPD agar (20 g/L glucose, 20 g/L peptone, 10 g/L yeast extract, 20 g/L agar) at 27 °C for 48 h, and maintained at 4 °C for the duration of study. The media used to evaluate viable cell counts during mixed fermentations were WL nutrient agar (Oxoid, Hampshire, UK) and lysine agar (Oxoid, Hampshire, UK). WL nutrient agar plates were used to differentiate the fermenting yeast from must samples (Pallman et al. 2001) , and lysine agar for the viable counts of the non-Saccharomyces yeast population. The viable yeast cells in the single culture fermentations of S. cerevisiae and Z. florentina (controls)
were determined using YPD agar plates.

One hundred millilitres of grape juices was poured into sterile bottles and left to ferment at 28°C ('fermented musts') until no further effervescence could be seen and no further reduction in glucose concentration was measured using a Keto Diabur Test (Roche, Mannheim, Germany). We plated aliquots of grape juices and fermented musts, eventually serially diluted in duplicate with sterile peptone 1 g l À1 , on WL Nutrient Agar (Oxoid, Basingstoke, England) (Pallmann et al. 2001) and on Lysine Agar (Oxoid) (Fowell 1965) , with Diphenyl 0Á1 g l À1 to slow-down moulds (Kurtzman et al. 2011) .
After 5 days at 28°C, we analysed plates with 20-200 colonies. When the number of colonies from undiluted samples was low, we analysed a high number of plates. We streaked yeasts in single colonies on WL Nutrient Agar to better recognize their morphotypes. Based on colony morphology on WL Nutrient Agar and microscopic analyses (data not shown), morphotypes of 3805 yeast colonies were identified. A number of representative colonies were isolated and purified: 553 yeast isolates were stored at 4°C on Malt Agar (30 g l À1 Malt extract, 15 g l À1 Agar; Oxoid) and used for further investigations.