Four strains of Oenococcus oeni of diverse origins were initially used: VP41 (Lalvin VP41) from Lallemand Inc. (Montréal, Canada); 1Pw13 (CECT 8893) from our own collection; PSU-1 (ATCC BAA-331), a reference strain with the genome fully annotated; and CH11 (Viniflora CH11), from Chr. Hansen Holding A/S (Hoersholm, Denmark). Before the fermentation assays, cells of these strains were precultured at 28 ºC in an incubator with 10 % (v/v) CO₂ in MRS broth supplemented with D, L-malic acid (4 g/L) and fructose (5 g/L) at pH 5.0 for 72 h at least two times prior to experimental use.

In order to see the influence on O. oeni growth, the above-mentioned precultures in supplemented MRS broth were also assayed with three different succinic acid concentrations (0.5, 1 and 2 g/L),^{and bacterial populations were quantified indirectly by measuring the Abs at 600 nm in a spectrophotometer Polarstar Omega (Biogen, Madrid, Spain).}

MLF were carried out in wine-like media (WLM) (Bordas et al., 2013) containing 2 g/L of L-malic acid, 12 % (v/v) ethanol, and three different succinic acid concentrations (0.5, 1 and 2 g/L),^{as well as the control without succinic acid, at pH 3.5 and 4.0, adjusted with NaOH 1N. All fermentations were performed in triplicate in 250 mL bottles at 20 ºC, inoculated with O. oeni strains at 2 × 107 CFU/mL. The samples were taken every 24 h to evaluate the L-malic acid consumption using an enzymatic kit (BioSystems, Barcelona, Spain) and the cells were harvested by centrifugation (6000 × g, 10 min). They were then frozen in liquid nitrogen and kept at –80 ºC until the RNA extraction.}

To evaluate the influence of the ratio of the two acids on MLF, experiments were carried out in WLM with different concentrations of L-malic acid (0.5, 1, 2 and 3 g/L) and succinic acid (0.5, 1 and 2 g/L). All assays were carried out at pH 3.5 and 4.0, at 20 ºC. Control assays without succinic acid were included for each L-malic acid concentration. All experiments were performed in 50 mL Falcon tubes inoculated with 2 × 10⁷ CFU/mL. O. oeni and samples were taken every 24 h to evaluate L-malic consumption using an enzymatic kit (BioSystems).

O. oeni cells were grown in 250 mL of MRS broth at 28 ºC to the early stationary phase. They were then harvested by centrifugation at 6000 × g for 5 min at room temperature, following Mira de Orduña et al. (2000). Cells were resuspended in resting-cell buffer, which contained per litre of deionised water 7.5 g tartaric acid and 1 mL of a mineral solution with 200 g/L MgSO₄·7H₂O and 50 g/L MnSO₂·4H₂O). Then, resting cells were transferred to 50 mL Falcon tubes containing 2 g/L of L-malic acid (control assay) and 12 % (v/v) of ethanol. The other conditions had the same molar amount of 2 g/L of L-malic acid (14.9 mM) and different proportional molar succinic acid concentrations (3.7 mM, 7.4 mM, 14.9 mM and 29.8 mM) at pH 3.5 and 4.2. This last pH was chosen because it is the pK_a1 for succinic acid. The Falcon tubes were placed in a water bath at 25 ºC and stirred gently. Samples were taken periodically and centrifuged at 6000 g for 5 min and the supernatants were kept at 4 ºC until the L-malic was analysed with an enzymatic kit (BioSystems).

The O. oeni RNA extractions were performed as described by Chomczynski and Sacchi (2006) and then purified with a Roche RNeasy kit according to the manufacturer’s instructions (Roche Diagnostics GmbH, Mannheim, Germany).

cDNA was synthesised from RNA (10 ng/μL) using TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, CA, USA) as recommended. To analyse the expression of four genes related to stress and malolactic activity, four pairs of primers (Table 1) were taken from previous works (Beltramo et al., 2006; Desroche et al., 2005; Olguín et al., 2010). They were about 18-22 bp long, contained 50 % G/C and had a melting temperature (T_M) above 60 ºC. The O. oenigyrA and gyrB genes were used as housekeeping genes (internal control), using the primers described by Desroche et al. (2005).

Real-time qPCR was performed following Olguín et al. (2010) using a QuantStudio 5 real-time PCR instrument (Thermo Fisher). The results were analysed using the comparative critical threshold (ΔΔCT) method in which the amount of target RNA was adjusted to a reference (internal target RNA) described by Livak and Schmittgen (2001). The relative expression value (RE) was calculated using the Ct values of gyrA and gyrB and the result is the mean of the two results. The analysis was made from biologically duplicated independent assays and for each sample technical triplicates were analysed by qPCR.

The organic acid contents (acetic, citric, succinic, L-malic and L-lactic acids) in the final samples of MLF trials were determined by high-performance liquid chromatography (HPLC) following Zhu et al. (2020). All samples were filtered previously by injecting them through 0.2 μm Captiva filters (Agilent Technologies). The chromatograms were analysed using the Agilent ChemStation Plus software.

The data obtained were submitted to one-way ANOVA using the Tukey test, with a confidence interval of 95 %, obtaining significant results with a p-value of ≤ 0.05, using the XLSTAT 2021 software (Addinsoft, Paris, France).

Table

titre du tableau
Target gene	Description	Forward primer (5´→ 3´)	Reverse primer (5´→ 3´)	Amplicon length (bp)	Reference
gyrA	Gyrase subunit A	CGCCCGACAAACCGCATAAA	CAAGGACTCATAGATTGCCGAA	95	(Desroche et al., 2005)
gyrB	Gyrase subunit B	GAGGATGTCCGAGAAGGAATTA	GCCTGCTGGGCATCTGTATTA	107	(Desroche et al., 2005)
mleP	Malate permease	GTGCTGACTTATTTGACCCGC	ATGTTCCACGACGACCAACC	141	(Augagneur et al., 2007)
mleA	Malolactic enzyme	CCGACAATTGCTGATACAATTGAA	GGCATCAGAAACGACCAGCAG	156	(Beltramo et al., 2006)
hsp18	Heat shock protein	CGGTATCAGGAGTTTTGAGTTC	CGTAGTAACTGCGGGAGTAATTC	102	(Beltramo et al., 2006)
atpB	ATPase F1 F0 β-subunit	ATACTGATCCGGCTCCGGC	CAGCGGGATAAATACCTTG	93	(Beltramo et al., 2006)

To see the effect of succinic acid concentration on the growth of O. oeni strains, previous assays were carried out in the same MRS broth used in precultures of fermentation assays. There was a clear lower growth when succinic acid was present for both strains assayed (Supplementary Figure 1), and this effect was progressively significantly higher with increasing concentration from 0.5 g/L to 2 g/L of succinic acid. As seen, with 1 g/L of succinic acid, the maximum population reached was about 50 % (strain PSU-1) or 70 % (CH11) of the control, and with 2 g/L it was about 30 % for PSU-1 and a mere 20 % of the control for strain CH11.

After the first experiments for comparing four different O. oeni strains, PSU-1 and CH11 were selected and used exclusively throughout the entire study since they were representative of the two behaviours observed in response to adding succinic acid. O. oeni CH11 showed a stronger inhibition of MLF in the presence of 2 g/L of succinic acid than strain PSU-1. This agrees with the above commented stronger growth inhibition of CH11 at 2 g/L of succinic acid added to the MRS medium. The other O. oeni strains VP41 and 1Pw13 showed similar behaviour to CH11 but with longer and slower MLF in all assays (Supplementary Figure 2). Control fermentations in WLM with strain PSU-1, at both pH 3.5 and 4.0, were completed in 120 h, and those with strain CH11 finished in 144 h (Figure 1). The bacterial population in the samples of these WLM fermentations was followed, but no increase was detected in any of them nor in controls (data not shown). It must be taken into account that growth is usually very difficult in the harsh conditions of WLM, similar to wine, with ethanol and low pH, as observed previously (Bordas et al., 2015; Jiang et al., 2018). Nonetheless, the inoculated population can survive and finish the MLF.

The results obtained in the presence of succinic acid varied depending on the pH and strains. Succinic acid inhibited the MLF of the two O. oeni strains at concentrations of 1 and 2 g/L at pH 3.5, whereas at pH 4 only the MLF with 2 g/L of succinic acid with O. oeni CH11 was inhibited in comparison to the control assay (0 g/L succinic acid). At pH 3.5 with 2 g/L of succinic acid, 43 % and 34 % of the L-malic acid was not consumed by CH11 and PSU-1, respectively, at the time that the control had exhausted all the L-malic acid. This shows that O. oeni CH11 was more sensitive to this concentration of succinic acid than PSU-1. The results with 1 and 2 g/L of succinic acid showed that the pH has a relevant influence on the MLF and the inhibitory effect of this organic acid. Meanwhile, at pH 3.5, succinic acid reduced the L-malic acid consumption rate in CH11 and PSU-1; however, at pH 4.0, both strains were able not only to carry out MLF but also to do so in less time than the control assay. Therefore, pH is an important factor because this physicochemical condition affects the dissociated and undissociated forms of succinic and malic acids, as observed by Augagneur et al. (2007).

Nevertheless, both strains CH11 and PSU-1 showed similar results in the 0.5 g/L succinic acid and in all cases, MLF finished about 24 h earlier than in the control assay. This best MLF performance in the presence of 0.5 g/L than without succinic acid was also found for the other two strains, 1Pw13 and VP41 (Supplementary Figure 1). Therefore, we have found that this low concentration of succinic acid could be slightly beneficial for O. oeni and the MLF, while levels of 1 g/L succinic acid or higher are clearly inhibitory. This probable benefit could be related to the above-commented decrease in succinic acid in some MLFs (Taniasuri et al., 2016; Yilmaz and Gökmen, 2021); however, we did not find significant variation in the succinic acid concentration in any case (data not shown).

Table

As seen above, the increasing amounts of succinic acid exerted an inhibition effect on the L-malic acid consumption. To understand this effect in more depth, we carried out assays in the same WLM with different ratios of L-malic and succinic acids. In addition, we aimed to determine whether there was an inhibition threshold of succinic acid depending on the initial L-malic acid concentration. At the same time, these assays simulated the real situation of the diversity of initial L-malic acid levels (from 0.5 to 3 g/L in these assays) in different climatic zones of winemaking.

To determine this influence, L-malic acid consumption rates were calculated for the different assays, shown in Supplementary Table 1. As seen, the main significant inhibition of succinic acid was when L-malic acid was 0.5 or 1 g/L. To visualise these results better, they are summarised in Figure 2.

There is a significantly bigger inhibition of the consumption rate for most conditions when 2 g/L of succinic acid is present, in agreement with the inhibition of MLF kinetics we found and discussed above. In terms of the pH, a slightly greater L-malic acid consumption can be observed at pH 4.0 than at 3.5, as expected since the higher pH is less stressful for O. oeni. The consumption tendencies are similar for both the strains and the succinic concentrations for the two pHs (Figure 2).

Table

Unlike the WLM, where O. oeni can grow, the resting cell buffer provides a way to monitor the malolactic enzymatic activity without interference from nutrients and metabolic changes related to growth. Therefore, in resting cell assays, there is no adaptation phase of growing cells and the effect of succinic is more direct. Cells number was about 5·10⁹/mL as they were harvested from 250 mL at the early stationary phase (10⁹/mL) and resuspended to a final volume of 50 mL, as said in Methods.

The consumption of 2 g/L (14.9 mM) of L-malic acid by resting cells was very quick (around 80–120 min) for both strains and pH levels and at the different succinic acid concentrations (data not shown). Hence, the high number of resting cells present can consume L-malic rapidly. These short times are similar to those found in similar works (Mira de Orduña et al., 2000).

To compare the L-malic acid consumption rate under the different conditions, we considered the amount of L-malic acid for assays with different succinic acid concentrations at the time when control assays had consumed about half of the L-malic acid (Table 2). This middle point of MLF (i.e., about 1 g/L (7.5 mM)) was at about 20 min for the control without succinic acid. Similarly, the data shown are for other sample conditions also taken at about 20 min. The different values of L-malic acid shown for the control at mid-MLF (from 2.08 to 6.82 mM) can be explained because samples were taken every 20 min and the L-malic consumption was very quick. Therefore, these are approximative values for mid-MLF.

We can observe (Table 2) that for most conditions, when succinic is present, the remaining L-malic acid is significantly higher than the control. Hence, succinic acid slows down the consumption of L-malic acid. Nevertheless, when succinic acid is 3.7 mM (0.5 g/L), the quantity of L-malic acid is significantly lower than the control at mid-MLF, indicating a quicker consumption rate at the lowest concentration of succinic acid. This tendency was not shown for PSU-1 at pH 4.2.

Table

titre du tableau
pH	Succinic acid (mM)	CH11		PSU-1
3.5	0	2.08 ± 0.226 bc		6.82 ± 0.197 d
	3.7	0.29 ± 0.185 d		4.59 ± 0.161 e
	7.4	1.54 ± 0.185 c		8.05 ± 0.161 c
	14.9	2.71 ± 0.185 b		9.00 ± 0.161 b
	29.8	4.55 ± 0.185 a		10.74 ± 0.161 a
4.2	0	4.28 ± 0.583 cd		3.73 ± 0.248 d
	3.7	1.98 ± 0.476 d		5.24 ± 0.203 c
	7.4	5.54 ± 0.476 bc		8.67± 0.203 ab
	14.9	7.01 ± 0.476 ab		8.97 ± 0.203 a
	29.8	8.57 ± 0.476 a		9.30 ± 0.203 b

^a–e Values for different succinic acid concentrations for the same pH and strain are significantly different at p ≤ 0.05, according to the Tukey test.

Overall, these results agree with those obtained previously in WLM (Figure 2), where the succinic acid inhibition of MLF was higher for concentrations higher than 1 g/L, that is, more than 7.4 mM of succinic acid.

To evaluate the possible inhibition of succinic acid in relation to the transcription of relevant genes associated with MLF development, we measured the relative expression of the following genes: i) mleA, encoding the malolactic enzyme, and mleP, encoding the L-malic acid transporter; ii) atpB, encoding the β-subunit of the F₁F₀ ATPase, associated with MLF activity and acid stress response (Cox and Henick-Kling, 1989; Fortier et al., 2003) and iii) hsp18, encoding for one of the most relevant stress proteins of O. oeni with a chaperone function, this gene is known to be activated in response to multiple stresses and has been proposed to be a stress adaptation indicator (Coucheney et al., 2005; Beltramo et al., 2006). The analysis was carried out with samples from the assay performed in WLM with the most inhibitory concentration of succinic acid (2 g/L). All samples were taken when the L-malic acid concentration was approximately 1 g/L in each case, at mid-MLF. Samples from cultures without added succinic acid were also analysed as control conditions.

In O. oeni CH11, the strain that showed the highest sensitivity to succinic acid, the gene hsp18 was clearly up-regulated at pH 3.5 and 4.0 (Figure 3). In contrast, the PSU-1 strain did not show overexpression of the hsp18 gene in any of the assayed conditions. These data suggest that the CH11 strain adopted this strategy to outcome the succinic acid inhibitory effect. However, O. oeni PSU-1, the strain least affected by succinic acid, might respond to stress differently, as seen in previous studies (Olguín et al. 2010).

No changes in expression were observed for atpB in any of the conditions studied. The transient activation of this gene has been described in response to acid shock (Fortier et al., 2003). In this study, there was no difference in the transcriptional levels of atpB at mid-MLF with or without succinic acid.

Surprisingly, O. oeni PSU-1 showed a 3.5 and 2.5-fold up-regulation of the genes mleP and mleA at pH 4 (Figure 3). This could be related to the faster MLF observed under these conditions (2 g/L succinic acid) with respect to the control condition at the same pH. However, the reason for the enhanced MLF due to the presence of succinic acid remains unclear and needs further research.

The main conclusion of the transcriptional study is that succinic acid does not inhibit the transcription of the mleA and mleP genes in any of the conditions. Therefore, the inhibition of MLF caused by this compound might be due to the intracellular acidification and possible substrate competition with L-malic acid at the enzymatic level, as previously discussed.

Table

This work was supported by grants AGL2015-70378-R and PGC2018-101852-B-I00, funded by Spanish MCIN/AEI/10.13039/501100011033 and, when appropriate, by ERDF "A way of making Europe", by the European Union or by the European Union Next Generation EU/PRTR. RT-G is grateful for the predoctoral fellowship from Fundación Carolina, which includes a partial stipend from the Mexican Government and University Rovira i Virgili.