To investigate the impact of temperature on B. bruxellensis, the growth of three strains (2OT13_02, L0463, and 1961-Mx-M1 E2) was compared at either 20 °C or 28 °C in SW supplemented with varying concentrations of p-coumaric acid (0 µM, 170 µM, and 370 µM).

The growth parameters were calculated from the growth dynamics over time (Figures S1–S3) and summarised in Figure 1 and Table 2, along with the physiological state of the cells at the end of the experiment.

nc, not calculated; ***, p-value < 0.001; **, p-value < 0.01; *, p-value < 0.05.

The B. bruxellensis strains displayed variable responses to temperature, with strain-dependent differences in growth and mortality under the tested conditions (Figure 1 and Table 2).

When focusing on strain 2OT13_02 in the absence of p-coumaric acid, reduced growth rate was observed at 28 °C compared with 20 °C, with values decreasing from 0.32 d ^-1 to 0.19 d ^-1 (Figures 1 and S1, Table 2). Similarly, for strain L0463, the growth rate dropped from 0.27 d ^-1 at 20 °C to 0.18 d ^-1 at 28 °C, while the lag phase was substantially prolonged by 211 h (Table 2).

Under the 170 µM p-coumaric acid condition, strains 2OT13_02 and L0463 exhibited reduced growth rates as temperature increased; rates fell from 0.30 d⁻¹ and 0.26 d ^-1 at 20 °C to 0.11 d ^-1 and 0.10 d ^-1 at 28 °C, respectively. Lag times were also affected, increasing by 21 h and 282 h for strains 2OT13_02 and L0463, respectively (Table 2).

Under the 370 µM p-coumaric acid condition, the temperature effect on growth rate was further exacerbated, with values decreasing from 0.23 d ^-1 and 0.20 d ^-1 to 0.10 d ^-1 and 0.05 d ^-1 for strains 2OT13_02 and L0463, respectively. The data also showed marked increases in lag times, which rose by 209 h and 337 h, respectively (Table 2).

Analysis of mortality percentages at the end of the experiment revealed a substantial effect of the temperature increase (Figure 1). At 20 °C, final mortality rates in the absence of p-coumaric acid were 27.2 %, 10.2 %, and 19.1 % for strains 2OT13_02, L0463, and 1961-Mx-M1 E2, respectively. At 28 °C, mortality increased significantly, reaching 82.8 %, 74.9 %, and 99.7 %, respectively, for the same strains at the end of the experiment ( p-value < 0.001). This significant increase in mortality between the two temperature conditions was also observed in the presence of p-coumaric acid, at both initial concentrations tested, 170 µM and 370 µM, and for all strains.

However, within each temperature condition, p-coumaric acid concentration had a limited effect on mortality. At 28 °C, strain 2OT13_02 showed some variability among the three p-coumaric acid conditions, with mortality rates of 82.8 %, 81.3 %, and 72.6 %, respectively. These differences were statistically significant ( p-value < 0.05) (Figure 1). Similarly, mortality rates for strain L0463 ranged from 67.6 % to 83.1 % across the three conditions, with significant differences observed ( p-value < 0.01) (Figure 1).

At 20 °C, the growth of strain 1961-Mx-M1 E2 was similar regardless of the p-coumaric acid concentration, though an increase in the lag phase was observed. The impact of the 28 °C condition was particularly pronounced for this strain, which exhibited a near-zero growth rate under all tested conditions (Figures 1 and S3). Consequently, the lag phase could not be estimated under these conditions. At 28 °C, strain 1961-Mx-M1 E2 exhibited very high mortality rates at the end of the experiment across all tested conditions, reaching 99.7 %, 99.8 %, and 99.5 % under the control, 170 µM, and 370 µM conditions, respectively (Figure 1).

Overall, these results highlight the combined effect of temperature and p-coumaric acid toxicity on growth impairment, as shown by both reduced growth rates and increased lag times ( p-value < 0.01) (Table 2). These findings are consistent with previous observations showing the toxicity of p-coumaric acid towards these strains under wine-like conditions (Bodin et al., 2026). However, the data suggest that temperature had a stronger influence than p-coumaric acid under these conditions. In this experimental context, the increase in temperature had a greater impact on yeast physiology, particularly on total cell populations, than p-coumaric acid concentration. For strain 2OT13_02, total cell populations decreased by 24 % in response to p-coumaric acid, whereas temperature alone induced a 59 % reduction. A similar pattern was observed for strains 1961-Mx-M1 E2 and L0463, with reductions of 9 % and 18 % linked to p-coumaric acid, respectively, compared with stronger decreases of 85 % and 57 % associated with temperature alone. However, this interpretation should be considered within the limits of the experimental conditions tested. A smaller increase in temperature combined with a higher p-coumaric acid concentration could lead to different effects and potentially reveal a stronger impact of p-coumaric acid stress than temperature on yeast physiology. This suggests that the relative contribution of each stress factor may depend on their respective intensity and on the balance between thermal stress and hydroxycinnamic acid-induced toxicity.

The consumption of p-coumaric acid and the production of VP and EP were investigated for the three strains at the two temperatures and the two levels of p-coumaric addition (Figure 2).

At 20 °C, all strains consumed nearly all of the initial p-coumaric acid, regardless of the concentration tested ( p-value > 0.05). However, the fate of the consumed acid differed between strains. Strains 2OT13_02 and 1961-Mx-M1 E2 converted p-coumaric acid almost exclusively into EP (81.4 % and 93.2 %, respectively), whereas strain L0463 displayed a mixed conversion profile, with both EP and VP accumulation (68.2 % and 13.8 %, respectively). In this latter strain, VP reached 14.25 µM and 45.87 µM under the 170 µM and 370 µM conditions, respectively, corresponding to a higher proportion of VP at 370 µM than at 170 µM (Figure 2). For L0463, the sum of VP and EP concentrations was lower than the amount of p-coumaric acid consumed, indicating that approximately 18 % of the consumed p-coumaric acid was not recovered as VP or EP.

Overall, the temperature increase affected p-coumaric acid consumption and conversion, although the magnitude of this effect was strain-dependent. The conversion of p-coumaric acid by 1961-Mx-M1 E2 was almost completely abolished, consistent with the strong sensitivity of this strain to 28 °C (Figure 2). In contrast, 2OT13_02 appeared to be the least affected strain, especially at 170 µM, where the impact of temperature on p-coumaric acid consumption and metabolite production remained limited ( p-value > 0.05) (Figure 2). However, at 370 µM, the temperature effect was more pronounced, with a 40 % decrease in EP production at 28 °C compared with 20 °C. This was also associated with the detection of VP at 28 °C, reaching 31.1 µM and representing 15.7 % of the total volatile phenols produced, suggesting a partial limitation of the VP-to-EP conversion step under this condition. Strain L0463 showed an intermediate but clearly temperature-sensitive profile, as EP production was significantly reduced at 28 °C for both p-coumaric acid concentrations ( p-value < 0.001). At 170 µM, EP production decreased by 77.8 %, while VP accumulation increased from 14.2 µM at 20 °C to 46.5 µM at 28 °C, indicating a marked impairment of VP reduction into EP while p-coumaric acid consumption was maintained. At 28 °C and 370 µM p-coumaric acid, the response was further amplified: p-coumaric acid consumption was drastically decreased by 86.9 %, while EP production was completely abolished.

Together, these results show that the effect of increasing temperature was amplified by the initial hydroxycinnamic acid concentration, particularly in strain L0463. Temperature and p-coumaric acid did not appear to act as independent constraints, but rather as combined stress factors affecting B. bruxellensis physiology. From an applied perspective, this cumulative effect is particularly relevant, as wine represents a multi-stress environment in which low pH, ethanol, nutrient limitation, phenolic compounds, and temperature may jointly impair yeast performance ( Smith & Divol, 2016). Here, the first detectable impact of this stress combination was observed in growth dynamics, through increased lag times and reduced viability. This suggests that cell adaptation and survival are the primary targets of the combined stress, before any marked effect on p-coumaric acid catabolism can be observed. This hierarchy of responses may explain why volatile phenol production was more affected than p-coumaric acid consumption under the harshest conditions (370 µM). Indeed, while the three strains remained able to consume p-coumaric acid at 20 °C, the increase to 28 °C strongly impaired growth and viability, which likely limited the metabolic capacity required for efficient conversion into volatile phenols. Thus, under cumulative stress, cellular physiology appears to be affected progressively: first through altered latency and survival, then through reduced conversion efficiency, and finally through impaired substrate consumption only when physiological stress becomes more severe.

The strain-dependent response observed in this work supports this interpretation. The three strains belong to distinct genetic backgrounds and differ in ploidy, which may contribute to differences in stress tolerance and metabolic behaviour, as previously reported for SO ₂ resistance (Avramova et al., 2018b). However, the genetic background alone does not appear sufficient to predict the response to temperature or p-coumaric acid. For example, strain 2OT13_02 was also previously identified as one of the most tolerant strains to p-coumaric acid under wine-like conditions (Bodin et al., 2026), which may partly explain its enhanced ability to maintain growth and volatile phenol production under combined stresses. Conversely, strains that performed well at 20 °C did not necessarily show higher resistance at 28 °C, indicating that tolerance to wine-like conditions at moderate temperatures does not directly predict tolerance to higher temperatures.

Compared with previous results for p-coumaric acid consumption obtained at 20 °C with an initial concentration of 700 µM (Bodin et al., submitted), the three strains remained strong p-coumaric acid consumers and volatile phenol producers under permissive conditions (≤ 370 µM). However, differences in VP and EP profiles, particularly the higher EP accumulation observed for the strains 2OT13_02 and 1961-Mx-M1 E2, may reflect the combined influence of temperature, strain physiology, and the lower initial p-coumaric acid concentration used in this experiment. One possible explanation is that lower substrate concentrations may favour a more complete conversion towards EP when growth is maintained. Nevertheless, this hypothesis should be considered with caution, as temperature-induced growth inhibition can also restrict pathway activity, and similar mortality levels did not always lead to comparable volatile phenol production (Bodin et al., submitted).

From an oenological perspective, 28 °C should not be considered as a thermal treatment, but rather as a model of transient warm conditions that may occur during wine alcoholic fermentation or the early post-fermentation period. Indeed, 28 °C has previously been used as a warm fermentation regime in red winemaking studies and remains below the temperatures required for reliable thermal inactivation of B. bruxellensis (Couto et al., 2005; Fabrizio et al., 2015). Previous studies showed that B. bruxellensis can grow and produce volatile phenols in wine at 15, 20 and 25 °C, whereas complete loss of viability was observed at 36 °C for the strains tested (Barata et al., 2008). Under otherwise identical conditions, lower temperatures also appear to favour higher final concentrations of 4-ethylphenol. Therefore, the inhibition observed at 28 °C in the present study should be interpreted as a strain- and environment-dependent response resulting from the interaction between temperature and p-coumaric acid, rather than as a general inhibitory effect of moderate heating.

Studies conducted under wine-ageing conditions have shown that very low storage temperatures, generally below 12 °C, may be required to substantially limit the development of B. bruxellensis, often in combination with oenological preservatives such as SO₂ (Cibrario et al., 2020; Oswald & Edwards, 2017; Zuehlke & Edwards, 2013). Nevertheless, such conditions do not necessarily prevent volatile phenol production or ensure complete population elimination, as strain-dependent persistence and subsequent recovery may remain possible. Low temperatures appear to act primarily by extending the lag phase and slowing cellular activity, rather than by immediately suppressing metabolism. This observation is consistent with the present results obtained at higher temperatures, in which changes in growth dynamics and viability preceded the reduction in volatile phenol production. In both cases, temperature mainly modulated the kinetics of B. bruxellensis development, while complete inhibition of volatile phenol production was rarely achieved. However, temperatures below 12 °C are seldom maintained during active fermentation and the early post-fermentation stages, notably in meridional areas.

The results obtained at 28 °C are therefore particularly relevant to the post-alcoholic fermentation and pre-malolactic fermentation stages of red winemaking, during which B. bruxellensis may develop when the activity of other microbial populations is limited. The temperatures reached during these stages may contribute to reducing volatile phenol production without completely preventing it. Early population monitoring during warm post-fermentation periods, followed by controlled low-temperature storage, may therefore help to restrict subsequent proliferation. However, the magnitude of this effect remains dependent on several interacting factors, including the contaminating strain, pH and hydroxycinnamic acid concentration.

Overall, the key outcome is not a direct relationship between mortality and substrate conversion, but rather a strain-dependent physiological response to cumulative stresses. Under the conditions tested, the lag phase and cell viability appeared to be affected first, followed by volatile phenol production and, under more severe stress conditions, hydroxycinnamic acid consumption.