Forty strains of Lb. plantarum, previously isolated from traditional red wines and available in the culture collection of the DiAAA (Dept. of Agricultural, Environmental and Food Science, University of Molise), were screened in order to assess the cell survival under acid or ethanol stress conditions. The commercial strain Lb. plantarum v22 (Lallemand Inc., Montreal, Canada) and the type strain Lb. plantarum DSMZ 20174 (Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany) were used as controls. Strains, stored at −80°C in Skim Milk (Succi et al., 2007), were propagated twice in MRS broth at 28°C prior their use. Then, 50 mL of each culture, grown in MRS broth (Oxoid, Milan, Italy) at 28°C, were taken in the mid-exponential phase (OD₆₀₀ = 2–3), standardized at an OD₆₀₀ = 2 (corresponding to 1 × 10⁹ CFU/mL) and centrifuged at 7,500 rcf for 15 min at 4°C. The pellet was washed 2 times with 1X phosphate buffer (1X PBS) and resuspended in 50 mL sterile Ringer's solution (RS) (Oxoid, Milan, Italy) containing ethanol (10 or 14%, v/v), or acidified with HCl up to pH 3.5 or 3.0. Inoculated broths were incubated for 2 h at 28°C and the viable count was performed in order to assess the cell survival. The results were expressed as Ln (N/N₀), where N are the CFU/mL after 2 h of incubation and N₀ are the CFU/mL at time 0.

At the end of the screening assay, 10 strains were selected on the basis of their different response to ethanol or acid stress and they were used in the subsequent experiments.

Batch fermentations were carried out at 28°C in Erlenmeyer flaks containing 500 mL of MRS broth containing 2, 4, 8, 10, or 14% (v/v) of ethanol. For this purpose, 1% of each overnight culture was inoculated into sterile MRS broth added with filter-sterilized ethanol (Filter Unit Red 0.22-μm pore size; Schleider & Schuell, Dassel, Germany) at different concentrations. A fermentation in MRS broth without ethanol was performed as control for each strain. Microbial growth was followed over the time by measuring the optical density at 600 nm (OD₆₀₀). The maximum specific growth rate (μ_max) was calculated by linear regression of Ln (OD/OD₀) as a function of the time, where OD₀ is the optical density at the beginning of the exponential growth phase.

Similarly, the effect of low pH on the growth of Lb. plantarum strains was assessed. In short, overnight cultures (1%) were inoculated in Erlenmeyer flaks containing 500 mL of sterile MRS broth acidified with HCl until pH 3.0, 3.5, 4.0, 4.5, 5.0, or 5.5. A fermentation in MRS broth at pH 6.5 was performed as control for each strain. The microbial growth was followed over time as reported above.

Strains of Lb. plantarum, cultivated in MRS broth containing 2 and 8% of ethanol as described above, were transferred (1%) at the beginning of the stationary phase in MRS broth containing 8, 10, or 14% of ethanol. A fermentation in MRS broth without ethanol was performed as control for each strain.

Similarly, strains cultivated in MRS broth at pH 4.0 and 5.0 as described above, were collected by centrifugation at the beginning of the stationary phase, and transferred (1%) in MRS broth acidified at pH 3.5 or 3.0 with HCl. A fermentation in MRS broth at pH 6.5 was performed as control for each strain. Microbial growth was monitored over time by measurement of the optical density at 600 nm (OD₆₀₀) of the cultures. For each experiment, the microbial growth and the maximum specific growth rate (μ_max) were obtained as described previously.

Two strains of Lb. plantarum (GT1 and LT11) were cultivated at 28°C in MRS broth containing 2% of ethanol, in MRS broth at pH 5.0, both prepared as reported above, and in MRS broth containing 2% of ethanol and acidified at pH 5.0 (both conditions, BC). MRS at pH 6.5 without alcohol was used as control. The microbial growth was estimated over time by measuring the optical density at 600 nm (OD₆₀₀) and reported as Ln (OD/OD₀), where OD₀ is the optical density at the beginning of the exponential growth phase.

Lb. plantarum GT1 and LT11 were cultivated at 28°C in MRS broth containing 2% of ethanol, in MRS broth at pH 5.0 and in MRS broth BC. MRS at pH 6.5 without alcohol was used as control. In the middle of the exponential growth phase or at the beginning of the stationary phase, cultures were centrifuged (7,500 rcf for 15 min at 4°C), the supernatant was discarded and the pellet was washed twice with a phosphate buffered saline (PBS). Cellular pellet was suspended (about 3.0 × 10⁸ CFU/mL) in RS containing 14% of ethanol or in RS acidified at pH 3.5 and incubated for 24 h at 28°C. At regular time intervals, an aliquot of cultures was recovered and enumerated by plate counts on MRS agar (Oxoid). Plates were incubated at 28°C for 72 h under anaerobic conditions using an anaerobic system (Oxoid). Three replicates were made for each experiment.

The log-transformed survival data were modeled using the linear or the non-linear regression approach. For this purpose, the Geeraerd and Van Impe Inactivation Model Fitting Tool (GInaFiT) was used, comprising nine different models (Geeraerd et al., 2005). In particular, the biphasic-linear model (Cerf, 1977) Equation 1, the Log-linear model with shoulder and tail (Geeraerd et al., 2000) Equation 2, and the double Weibull model (Coroller et al., 2006) Equation 3 are reported in the following formulae.

where N₀ is the initial population (CFU/mL), N is the residual population at time t (CFU/mL), f is the fraction of the initial population in a major sub-population, (1 − f) is the fraction of the initial population in a minor sub-population (which is more resistant than the previous one), and k_max1 and k_max2 (h⁻¹) are the specific inactivation rates of the two populations, respectively.

where N, N₀, and K_max have identical meaning as that reported in Equation 1, N_res is the residual population (CFU/mL), and S_l (h) is the parameter that represents the shoulder length.

where N, N₀, and t have identical meaning as that reported previously, p is an adimensional shape parameter, α is the ratio of the fraction of sub-population 1 to the fraction of the sub-population 2 at time 0, δ₁ is the time (h) required for the first decimal reduction of sub-population 1, δ₂ is the time (h) needed for the first decimal reduction of sub-population 2.

The detection limit (DL) was fixed to 1. In order to assess the goodness of fit of each model, the sum of square error (SSE), the root mean square error (RMSE), the adjusted coefficient of determination (adj-R²) were used.

The strains Lb. plantarum GT1 and LT11 were selected on the basis of their acid and ethanol stress response with the purpose to investigate the effect of acid and ethanol adaptation on the survival and consumption of L-malic acid in a model system at 14% ethanol and pH 3.5 (Synthetic Wine, SW). The synthetic wine was prepared as described by Bravo-Ferrada et al. (2013) and inoculated (about 1 × 10⁸ CFU/mL) with Lb. plantarum GT1 or LT11 preadapted in MRS containing 2% of ethanol or in MRS at pH 5.0 or in MRS BC. Cells were recovered in the mid-exponential phase as well as at the beginning of the stationary phase and then inoculated in SW. Microbial growth, pH and L-malic acid concentration were monitored during the incubation period (15 days at 20°C). Analogous experiments were performed using non-adapted cells as control.

Figure 1 displays the survival of 42 strains of Lactobacillus plantarum evaluated in Ringer's solution (RS) containing 10 or 14% of ethanol (Figure 1A) and in RS acidified at pH 3.5 or 3.0 (Figure 1B). A higher variability among strains resulted from the exposure to low pH than that to ethanol. Specifically, the exposure at 10% of ethanol did not affect the survival, whereas in presence of 14% of ethanol a reduction of about 4 Log CFU/mL was observed for the assayed strains, with the exception of v22 and GT1, which were inhibited to a lower extent. A different scenario was observed when the strains were exposed to low pH (Figure 1B). In this case, Lb. plantarum strains were divided into 4 groups, arbitrarily individuated on the basis of the ability to survive in acid condition: group I collected 23 strains with high susceptibility to the pH; group II gathered 6 strains with a medium-high susceptibility; group III collected 9 strains with a moderate sensitivity; group IV convened 4 strains with the lowest susceptibility. On the basis of their different response to ethanol and acid stress, 10 strains (LM27, TSC11H, LM28, LP6, LT11, v22, LM25, LM29, PCQA, and GT1), randomly chosen from each group, were selected and used in the following experiments.

The effect of different concentrations of ethanol on the growth of 10 strains of Lb. plantarum was tested. Results (Figure 2A) showed that no strain was able to grow in MRS containing 10 or 14% of ethanol, whereas all strains were able to grow up to 8%. However, the maximum specific growth rate (μ_max) resulted significantly reduced (p < 0.05) when strains were cultivated in MRS containing 8% of ethanol, and a significant reduction was also appreciated in presence of 4%. Contrarily, the growth was little affected by the presence of 2% of ethanol, as evidenced by the μ_max (0.55 ± 0.04) which did not vary significantly (p > 0.05) if compared with the control.

The effect of low pH on the growth of the selected Lb. plantarum strains is reported in Figure 2B. No strain was able to grow at pH 3.0, and a slight growth was appreciated at pH 3.5 and 4.0. Starting from pH 4.5, most of the strains showed the ability to grow, even if the corresponding μ_max (0.22) was significantly lower than the control. The kinetic of growth at pH 5.5 or 5.0 showed in both cases a reduction of the μ_max of about 2-fold in comparison with the control (pH 6.5), but between the two conditions (pH 5.5 and pH 5.0) no significant difference was recognized (p > 0.05).

The 10 selected strains of Lb. plantarum, pre-cultivated in MRS broth containing 2 or 8% of ethanol until the stationary phase (long-term adaptation), were subsequently cultivated in the same medium containing 8, 10, or 14% of ethanol to investigate if a non-lethal alcohol concentration was able to improve the cellular growth in presence of a high ethanol concentration. Results showed that the growth in MRS containing 10 or 14% of ethanol was inhibited independently by the preadaptation to 2 or 8% (data non shown). Strains preadapted to 8% of ethanol and then transferred in the same medium with the same alcohol concentration showed identical kinetics of growth with the controls (data not shown). Instead, a higher growth than that obtained in control conditions was observed when the strains were cultivated with 8% of ethanol after a preadaptation to 2%. Results are reported in Figure S1A, where a significant (p < 0.05) higher μ_max of preadapted strains was detected compared with their controls. This datum was particularly evident for the preadapted strains GT1 and LT11, having μ_max values 1.4- and 1.3-fold higher than those of their controls, respectively.

The impact of the long-term adaptation to low pH (5.0 and 4.0) on the growth at pH 3.5 or 3.0 of the selected Lb. plantarum strains was also evaluated. Results showed that the growth in MRS at pH 3.0 was inhibited independently by the preadaptation to pH 4.0 or 5.0 (data non shown). Moreover, strains preadapted to pH 4.0 (data non shown) or 5.0 (Figure S1B) and then transferred in the same medium at pH 3.5 showed similar kinetics of growth then their controls. A sensitive increment of the μ_max was detected only for the strain GT1 preadapted to pH 5.0 and subsequently cultivated in MRS at pH 3.5.

On the basis of previous results, sub-optimal conditions (pH 5.0 or ethanol 2%) and their combination (BC) were chosen to assess the effect on the growth of selected strains GT1 and LT11. Results (Figure S2) highlighted that the two strains had similar behaviors in the assayed growth conditions. Conversely, significant differences in the maximum specific growth rate (μ_max) were detected depending on the different growth conditions. The presence of 2% of ethanol did not affect the growth of both strains (Figures S2A,B), as showed by μ_max values substantially similar to those detected in the respective controls. The growth was instead significantly affected when GT1 and LT11 were cultivated at pH 5.0 or in BC. In fact, both situations caused a significant μ_max decrease (p < 0.05) which assumed values of 1.6- and 2.5-fold lower then the controls, respectively.

In this step, the survival of strains GT1 and LT11 in Ringer's solution (RS) containing 14% of ethanol after a long-term adaptation to ethanol 2% or to BC (pH 5.0 and ethanol 2%) was evaluated. The survival of preadapted cells was compared with the survival of non-preadapted cells. Moreover, the impact of the physiologic state (exponential phase vs stationary phase) of cells on the survival in ethanol 14% was evaluated (Figure 3). Analogously, the survival of GT1 and LT11 in RS at pH 3.5 after a long-term adaptation to pH 5.0 or to BC was compared with the survival of non-preadapted cells used in both exponential and stationary phase (Figure 4).

In all the analyzed conditions, the physiological state had a strong influence on the survival of the two strains. In fact, preadapted cells showed the same behavior of non-preadapted cells only when the cultures were recovered at the beginning of the stationary phase, and this fact concerned both ethanol and pH tests (Figures 3, 4). In detail, regardless to the ethanol sensitivity, the microbial load of the strains GT1 and LT11 withdrawn in the stationary phase did not decrease after 24 h of incubation in RS containing ethanol 14% (Figure 3). On the other hand, when strains collected in the stationary phase were tested at pH 3.5 (Figure 4), our results showed a decrease of about 1 Log CFU/mL only for LT11 (Figure 4B).

Conversely, cells recovered in the exponential phase showed a survival decay, which was more marked in the case of non-preadapted cells or cells preadapted to BC. Additionally, a different trend in the survival curves was appreciated between the two strains preadapted in different conditions. In fact, the exposure to 14% of ethanol or to pH 3.5 caused the lowest cell decrease (about 1 Log CFU/mL in both conditions) for GT1 preadapted in ethanol 2% or at pH 5, respectively, whilst the corresponding non-preadapted strain and that preadapted to BC showed a marked reduction in 14% ethanol (about 4 Log CFU/mL) and a minor decrease at pH 3.5 (about 2.5 Log CFU/mL) (Figures 3A, 4A).

The strain LT11 withdrawn in the exponential phase and exposed to 14% of ethanol or to pH 3.5 showed a different behavior compared to GT1. In detail, cells of LT11 had an overall microbial reduction of about 2.5 Log CFU/mL when preadapted in ethanol 2% and exposed to 14% of alcohol, and the highest drop was appreciated in the first 3 h of incubation (Figure 3B). LT11 preadapted to pH 5 and exposed to pH 3.5 had an overall microbial drop of about 4 Log CFU/mL (Figure 4B).

LT11 non-preadapted or preadapted to BC was strongly injured by the presence of high alcohol concentration or by the low pH when collected in the exponential phase, as highlighted by a decrease of about 7.5 or 5.5 Log CFU/mL registered at the end of the incubation with 14% of ethanol or at pH 3.5, respectively (Figures 3B, 4B).

The survival data were modeled using GInaFiT tool (Tables 1, 2). Out of 9 inactivation models fitted with the GInaFiT tool, 3 showed the finest statistical parameters (SSE, RMSE, and adj-R²). Specifically, considering the survival in the medium containing 14% of ethanol, the Log-linear model with shoulder and tail and the double Weibull model were the two mathematical functions that better described the survival of strains GT1 (Figure 3A) and LT11 (Figure 3B), respectively. Regarding to the survival in the medium at pH 3.5, the Log-linear model with shoulder and tail and the biphasic linear model were individuated as the best functions to describe the behavior of GT1 (Figure 4A) and LT11 (Figure 4B), respectively.

The kinetic parameters obtained by the models highlighted different responses of the two strains to 14% of ethanol (Table 1). In detail, during the first hours of exposure to ethanol, a population fraction of non-preadapted cells of LT11 was reduced of 4 Log units (4D) in 5.5 h, about 1.5-fold more than GT1 (3.8 h). A further reduction of LT11 was observable when the exposure was prolonged. In fact, after 24 h of exposure to 14% of ethanol, the strain LT11 was no more detectable, whereas the microbial load of GT1 was about 4 Log CFU/mL.

The datum regarding the preadaptation effect in 2% of ethanol on the survival of LT11 and GT1 was particularly interesting. The results (Table 1) showed that when the strain LT11 was preadapted to ethanol, the 4D value was > 24 h, and the kinetic parameters δ₁ and δ₂ were 2-fold higher than those of non-preadapted cells. These parameters represent the time to obtain a decimal reduction of the first (δ₁) and second (δ₂) fraction of the microbial population. The most important improvement of the kinetic parameters was also observed in the model representing the survival of GT1 after ethanol-adaptation. Among all, the Log(N_res) parameter, that represents the “tail” of model, resulted about 2-fold higher than that of non-preadapted cells. Moreover, the microbial load corresponding to the tail was reduced of 1 Log unit compared with the microbial load at the beginning of the exposure (Log_N₀).

The long-term adaptation to BC did not positively affect the kinetic parameters. In fact, the 4D value and the Log(N_res) parameters for both strains were similar or even lower than those observed for non-preadapted cells.

Considering the survival at low pH, the results regarding the kinetic parameters (Table 2) highlighted that non-preadapted LT11 had a decay of the survival with a k_max1 of 2.63 and a k_max2 of 0.25. These parameters were substantially reduced (−1.5-fold) for cells preadapted to pH 5.0. A similar behavior was observed for the strain GT1. In detail, when the cells adapted to pH 5.0 where exposed to pH 3.5, the kinetic parameter representing the shoulder (S_l) of the curve increased of 1.3 units compared to that recorded for non-preadapted cells. Moreover, the k_max value was reduced of about 1.5 units for the preadapted cells compared to non-preadapted ones. The kinetic parameter regarding the residual population (Log_Nres) after 24 h of exposure to pH 3.5 was also noticeable. In this case, the microbial load was higher than 7 Log CFU/mL for sub-optimal pH preadapted cells and lower than 6 Log CFU/mL for non-preadapted cells. Finally, no improvement was observed when the strains were preadatpted to BC.

In this final step, a synthetic wine (SW) at 14% ethanol and pH 3.5 was used to assess the behavior of the strains GT1 and LT11 preadapted in MRS containing ethanol 2%, or acidified at pH 5.0, or BC. Cells were collected in the exponential phase or at the beginning of the stationary phase and incubated in SW for 15 days at 20°C. As expected, both strains recovered in the exponential phase showed a very high survival decay in SW, regardless of the preadaptation conditions (data not shown). Better performances were appreciated when the cells were collected at the beginning of the stationary phase (Figure 5). In detail, considering the strain GT1 collected at the beginning of the stationary phase, the preadaptation to pH 5 significantly improved the survival in SW (Figure 5A). In fact, after 15 days of incubation, the microbial load was substantially unaffected (about 10⁸ Log CFU/mL). However, an improvement of the survival in comparison with non-preadapted cells was also observed for cells preadapted to BC, whereas cells preadapted to ethanol 2% resulted more sensitive than the control. Considering the tests carried out on LT11, the strain highlighted a complete decay in spite of preadapation conditions (Figure 5B).

The trend of L-malic acid concentration and pH of the SW inoculated with GT1 recovered at the beginning of the stationary phase and preadapted to ethanol 2%, or to pH 5.0, or to BC is displayed in Figure 6A. The results revealed that the L-malic acid concentration and the pH values remained unaffected when the SW was inoculated with non-preadapted GT1, with GT1 cells preadapted to ethanol 2% or with cells preadapted to BC. On the contrary, the L-malic acid resulted completely degraded after 15 days of incubation with acid-preadapted GT1. Moreover, the pH value increased of about 0.3 units only when the SW was inoculated with acid-preadapted cells.

Instead, no changes were observed in L-malic acid concentration and in pH values when preadapted or non-preadapted LT11 cells were inoculated in SW (Figure 6B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.