Temperature Shapes Ecological Dynamics in Mixed Culture Fermentations Driven by Two Species of the Saccharomyces Genus

Abstract

Mixed culture wine fermentations combining species within the Saccharomyces genus have the potential to produce new market tailored wines. They may also contribute to alleviating the effects of climate change in winemaking. Species, such as S. kudriavzevii, show good fermentative properties at low temperatures and produce wines with lower alcohol content, higher glycerol amounts and good aroma. However, the design of mixed culture fermentations combining S. cerevisiae and S. kudriavzevii species requires investigating their ecological interactions under cold temperature regimes. Here, we derived the first ecological model to predict individual and mixed yeast dynamics in cold fermentations. The optimal model combines the Gilpin-Ayala modification to the Lotka-Volterra competitive model with saturable competition and secondary models that account for the role of temperature. The nullcline analysis of the proposed model revealed how temperature shapes ecological dynamics in mixed co-inoculated cold fermentations. For this particular medium and species, successful mixed cultures can be achieved only at specific temperature ranges or by sequential inoculation. The proposed ecological model can be calibrated for different species and provide valuable insights into the functioning of alternative mixed wine fermentations.

Keywords: Gilpin-Ayala; Lotka-Volterra; Saccharomyces; ecological modeling; mixed culture fermentations.

Figure 1

Formulation of candidate models and model selection. (A) presents the different mechanisms included in the models, namely, growth, lag-phase, intra- and interspecific competition. Different definitions for intra- and interspecific competition result in different candidate models: Lotka-Volterra (LV), Gilpin-Ayala (GA), and Gilpin-Ayala with saturable competition (SGA). (B) presents the comparison of the best-fitted models. SGA is the best model; its fit to the multi-experiment data set results in better cAIC and the minimum training and validation errors in cross-validation. (C) shows the worst and best fit to the data for the multi-experiment model calibration.

Figure 2

Comparison of strain behavior in single and mixed cultures using the best model. (A) presents the specific growth rate μ_i(T) as a function of the temperature for both strains. (B) shows the carrying capacity as a function of the temperature (K_i(T)). (C) shows the biomass yield after 10 days in mixed co-inoculated (50/50) fermentations. (D) shows the effect of the lag-phase. In the case of S. cerevisiae, the time required to achieve the maximum specific growth rate depends on the temperature. The lag-phase is significantly longer at lower temperatures. (E) presents the relative decrease of the mean per-capita growth rate (over time) in mixed culture. The per-capita growth rate for Sk in mixed cultures (R_m,Sk) is up to a 15.4% lower than its value in single culture (R_s,Sk). For the case of Sc, the effect of temperature on the per-capita growth rate is higher than the presence of another cell. (F) presents the 2x and 10x times for single and mixed cultures. (G) shows the relative increase in 2x and 10x times in mixed culture. The presence of Sc has a significant effect on the 10x time for Sk. Figures were obtained for a 2 × 10⁶(CFU/ml), 50% of each species (50/50).

Figure 3

Characteristics of mixed cultures. (A) presents the projection of the model nullclines over the temperature and the final density of Sk; the black line shows the intersection between the nullclines indicating the range of temperatures for which both cells coexist independently of the initial inoculation. Black and red squares mark two different examples: Ex. 1 at T = 24.5°C in which Sc excludes Sk and Ex.2 at T = 16°C in which both strains coexist. (B) shows the model nullclines for the two examples and the corresponding mixed culture dynamics. (C) presents the density in % for both cells in mixed culture after around 10 days, as a function of temperature and considering two different co-inoculation conditions. The percentage of Sk exceeds that of Sc at cold temperatures. While the opposite behavior is found for T > 12°C. For T > 24°C Sk is excluded. (D) presents an illustrative example of sequential inoculation at 25°C showing how both species may coexist.