Acid adaptation of Lactobacillus helveticus during continuous cultures to improve undissociated lactic acid production

Abstract

This study aims at improving the undissociated lactic acid production by Lactobacillus helveticus using a whey-based fermentation. It first describes the effect of pH on the ability of this bacterium to produce lactic acid, by considering final lactic acid concentration, production rate, volumetric productivity and sugar consumption. As a low performance was achieved at pH 4.3, an adaptive evolution of Lb. helveticus LH-B01 to acidic conditions was performed during continuous cultures of sweet hydrolysed whey. Two mutants have been isolated, which exhibited different characteristics. The mutant Lb. helveticus LH-B01-B4 displayed the higher maximal total lactic acid concentration (37.9 g/L), sugar consumption (82%) and volumetric productivity (0.39 g/L/h), when compared to the parental strain and the mutant Lb. helveticus LH-B01-A4. This performance was explained by the higher critical undissociated lactic acid concentration (10.1 g/L) of Lb. helveticus LH-B01-B4, compared with those of the parental strain (8.7 g/L) and the mutant Lb. helveticus LH-B01-A4 (7.5 g/L). From these results, the mutant strain Lb. helveticus LH-B01-B4 was the most promising option to produce undissociated lactic acid during low pH fermentation, thus making it suitable for industrial use as a descaling agent and biocide in detergents.

Graphical abstract

Introduction

Lactic acid, or 2-hydroxypropanoic acid, is a natural organic acid with a wide range of applications in the food, pharmaceutical, chemical and leather tanning industries (Komesu et al. 2017). In the form of undissociated lactic acid, this compound is an effective alternative for household cleaning, where it acts as a descaling agent and a biocide in detergents (Huang et al. 2010). For these applications, it shall be provided at low pH, where undissociated lactic acid is predominant (Kim et al. 2022).

The vast majority of lactic acid is produced by fermentation using lactic acid bacteria or strains from the Bacillus species, Escherichia coli or Corynebacterium glutamicum (Abdel-Rahman et al. 2013). Lactic acid bacteria are mostly used for industrial lactic acid production, and Lactobacillus helveticus, Lactiplantibacillus plantarum and Lacticaseibacillus casei species are particularly interesting due to their high acid tolerance (Castillo Martinez et al. 2013). Among these species, Lb. helveticus has been used by various authors to study the production of lactic acid using different batch or continuous fermentation processes with immobilised cells or cell recycling (Abdel-Rahman et al. 2013). However, lactic acid production faces many challenges, due to the cost of raw materials such as sugars and nitrogenous supplements to meet nutritional requirements of the bacteria (Alves De Oliveira et al. 2018) and to the inhibition caused by lactic acid accumulation and low pH during the cultures (Singhvi et al. 2018).

Whey of milk is a suitable raw material for lactic acid production as it is cheap and contains all necessary nutrients. Long considered as a waste product, it is however rich in lactose, serum proteins, water-soluble vitamins and minerals (Tsermoula et al. 2021). In addition, the hydrolysis of serum proteins can provide the necessary nitrogenous compounds to meet the nutritional requirements of microorganisms (Ghosh et al. 2017). Among the bacteria capable of producing lactic acid, Lb. helveticus is a thermophilic homofermentative lactic acid bacterium that can convert both glucose and galactose that derive from lactose, into D (−) or L ( +) lactic acid, making it suitable for lactic acid fermentation from whey (Amrane and Prigent 1998; Gatje and Gottschalk 1991).

The inhibition phenomena linked to lactic acid production are more complex to manage. During fermentation, lactic acid bacteria produce lactic acid in the dissociated form lactate (La⁻) and undissociated form acid (HLa), as the intracellular pH is generally higher than 5.5 (Béal et al. 2023). To maintain the lactate balance on both sides of the cell membrane, lactate form is excreted within protons according to a symport mechanism, thus generating ATP and providing a significant energy gain for the cells (25 to 33% per lactate for the cell) (Konings 2002). This phenomenon works well at low lactate concentration, but as the fermentation progresses, lactate accumulates intracellularly, thus leading to a reduction of the intracellular pH (pHi). This is partly counteracted by protons export through the membrane ATPase, thus consuming ATP (Konings 2002). This phenomenon, which is called lactic acid inhibition, is amplified at low pH. In that case, the acid–base balance is modified to favour the undissociated form of lactic acid (HLa), which can move freely through the cellular membrane. This progressively induces a decrease of the intracellular pH that causes metabolism slowdown, growth inhibition and possibly cell death. As a consequence, lactic acid production performance is strongly affected at low pH, as demonstrated by Schepers et al. (2002) who compared fermentations of whey permeate supplemented with yeast extract by Lb. helveticus at pH 5.5 and 4.2. Sugar consumption was only 61% at pH 4.2 instead of 100% at pH 5.5 and lactic acid volumetric productivity was strongly reduced from 1.74 g/L/h at pH 5.5 to 0.32 g/L/h at pH 4.2.

The use of strains able to grow at low pH such as 4.7 or less, thus represents a potential strategy to partly overcome lactic acid inhibition. This is however a real challenge due to the sensitivity of lactic acid bacteria to acid stress, and an adaptation of the strains is a solution to overcome this barrier. According to Mavrommati et al. (2022), adaptive laboratory evolution, which consists in exposing microorganisms to stress conditions under a controlled environment, allowed to obtain natural evolution and to select resilient phenotypes. This approach has been implemented using serial batch cultures, where the selective pressure is maintained constant or progressively intensified through successive transfers. A number of studies have been focused on the selection of acid-tolerant strains of lactic acid bacteria by serial batch cultures. Zhang et al. (2012) isolated a Lacticaseibacillus casei mutant by serially transferring the parental strain recovered in mid-exponential phase, with a gradually decreasing pH (from pH 5.5 to pH 4.3) for 70 days. Subsequently, the L. casei Zhang mutant demonstrated superior survival rates at low pH, an increase in the production of lactate and acetate by 14% and 66% respectively, and enhanced intracellular pH and NH4⁺ concentrations, in comparison with the parental strain. Li et al. (2025) applied an innovative pH-shifting adaptive evolution to the engineered Pediococcus acidilactici ZB220 strain, in order to generate low pH tolerant-mutants and to improve the production of lactic acid from xylose and glucose. By alternating acidic pH (pH 4.8 or 4.6) and optimal pH 5.5 during cultures in modified MRS medium, they obtained one mutant that produces 43% more lactic acid at pH 4.6 and 2.1 times more at pH 4.4 than the parental strain. However, adapting strains through a series of batches is cumbersome and may lead to periodic variations in environmental conditions during each batch. Consequently, this approach could complicate the interpretation of strain adaptation phenomena, as the selection pressure combines changes of the culture medium such as carbon and/or nitrogen sources limitations, with the selected stress (Sauer 2001).

Alternatively, adaptive laboratory evolution can be conducted during continuous cultures, which enables better controlled conditions, the maintenance of constant nutrient levels, the elimination of starvation phases, the regulation of stress levels, and the constant application of selection pressure to the strains (Gresham and Dunham 2014). In addition, this approach reduces the duration of the adaptation procedure, as the cells are continuously maintained in exponential phase. Zhang et al. (2020) carried out an adaptive laboratory evolution of Actinobacillus succinogenes NJ113 to improve succinic acid production at low pH (pH 5.8). Throughout the chemostat, an adapted strain was obtained that is able to produce 2.54 times more succinic acid from glucose than the original one, at a lower pH than the optimal pH of 6.8. They ascribe these changes to modifications of the intracellular ATP content and of the membrane composition. Hartono et al. (2023) developed a novel continuous adaptive evolution method, called “stressostat”. This system aims at improving microbial resistance to inhibitory end-products by introducing the inhibitory compound in the feed medium during continuous cultures. It was used to select Lactococcus lactis FM03P lactate resistant strains, during continuous cultures done at increased lactate concentrations (from 45.0 to 60.8 g/L) at pH 6.5. Four clusters of variants among 34 were identified, the moderate resistant one, the high resistant one with a metabolic shift (acetate production), the high resistant one with a slow growth at high lactate concentration and the best variants that simultaneously showed high growth rate and high lactate resistance. However, to our knowledge, the adaptation of Lb. helveticus to acidic pH specifically by continuous fermentation has never been implemented.

Based on this state of the art, a knowledge gap still exists about the ability of Lb. helveticus to efficiently produce undissociated lactic acid at acidic pH, thus requiring the obtention of acid-adapted strains. Present work aims at adapting Lactobacillus helveticus to acidic conditions during continuous fermentations of sweet whey at decreasing pH. This approach aims at obtaining mutants with higher tolerance to undissociated lactic acid, in order to further improve undissociated lactic acid production during batch cultures. A bioreactor was implemented at decreasing pH, to generate acid-tolerant Lb. helveticus mutants. The dilution rate was augmented at the final stage of the chemostat to select cells with increased specific growth rates. Thanks to this original methodology, mutants were isolated and their performances, such as undissociated lactic acid concentration and productivity, were compared to those of the parental strain, and their critical undissociated lactic acid concentration was determined.

Material and methods

Bacterial strain and culture media

Lactobacillus helveticus LH-B01 was obtained from Novonesis (Arpajon, France) and stored at -80 °C. This homofermentative species was selected for its ability to metabolize both glucose and galactose from lactose, to produce a racemic mixture of lactic acid.

Freshly harvested sweet whey from a Reblochon cheese dairy factory (Fromagerie Masson, Juvigny, France) had an average lactose concentration of 47.2 ± 1.2 g/L. It was stored at – 20 °C, then thawed prior to use. A sweet whey powder (Epi Ingredients, Ancenis-Saint-Géréon, France) was also used after rehydration at 70 g/L to obtain a lactose concentration of 45.5 ± 1.7 g/L. The two kinds of whey were hydrolysed with Flavourzyme 1 000 L (Novozymes, Bagsvaerd, Denmark) at 0.1 g/L for 2 h at 42 °C, then sterilized at 110 °C for 10 min. In addition, hydrolysed whey obtained from powder was centrifugated at 17 696 g, 4 °C for 10 min (Avanti J-E, Beckman Coulter, Brea, CA, USA) before use.

Batch fermentations protocol

Batch fermentations were conducted in a 500 mL sterilized bioreactor (Global Process Concept, Périgny, France) filled with 200 mL of fresh hydrolysed and sterile whey. Temperature was maintained at 42 °C and agitation rate at 200 rpm. The pH was left free until the set pH value was reached. It was then maintained constant with 5 M NaOH at different pH values (5.5, 4.7 and 4.3), according to the experimental design. The lowest pH value (pH 4.3) corresponded the inhibition threshold for the parental strain. The time at which NaOH consumption started was normalised at 0 in order to become independent from the physiological state of the inocula. The final fermentation time was determined as the moment when NaOH consumption ceased. All conditions were duplicated.

Continuous fermentations and adaptative evolution

Acid adaptation of Lb. helveticus was performed during two continuous fermentations in a 500 mL bioreactor (Global Process Concept) containing 400 mL of rehydrated and hydrolysed sterile whey powder. The experimental strategy is summarized in Fig. 1. The continuous fermentations were carried out at 42 °C and 200 rpm, at an initial dilution rate of 0.2 h⁻¹ and an initial pH of 6.3 (± 0.1). The pH was left free until pH 4.7, then progressively lowered by steps of 0.1 pH unit from pH 4.7 to pH 3.6. It was controlled, at each step, by adding 5 M NaOH. Between each pH step, at least 3 residence times were ensured. At the final pH, the dilution rate was increased from 0.2 to 0.5 h⁻¹, by steps of 0.1 h⁻¹. Samples were collected each day and stored at -80 °C after adding 13% glycerol.

Fig. 1.

Scheme of the continuous protocol used to adapt Lactobacillus helveticus LH-B01 to acidic conditions

Strain isolation, production and preservation

Isolates were obtained from samples that were recovered during the continuous cultures, on MRS agar medium (Biokar Diagnostics, Beauvais, France). They were then cultured in batch mode, in 10 mL MRS medium (Biokar Diagnostics) acidified at pH 4.8 with 11.4 M lactic acid. The cultures were incubated under anaerobic conditions at 42 °C for 48 h. They were added with 13% glycerol and stored at -80 °C.

Identification of bacterial species

For DNA extraction, isolates were cultured in MRS (Biokar Diagnostics) for 48 h at 42 °C. Samples were collected, centrifuged at 5 000 g for 5 min at 4 °C, rinsed with physiological water and centrifuged again before being stored at -20 °C. DNA extraction was performed using DNeasy PowerFood Microbial Kit (Qiagen, Courtaboeuf, France), according to the manufacturer’s protocol with modifications. The pellets were incubated in 180 µL of pre-lysis buffer including TE 2X buffer (Tris–HCl 20 mM pH 8, EDTA 2 mM, Sigma-Aldrich, Saint-Louis, MO, USA), 1.2% Triton X100 (Sigma-Aldrich), 20 mg/mL Lysozyme (100 mg/mL, Sigma-Aldrich) and 10 U Mutanolysin (1000 U/mL, Sigma-Aldrich) for 30 min at 37 °C and vortexed every 10 min. The resuspended cells were transferred in a PowerBead Tube (Qiagen) with 450 µL of MBL buffer (DNeasy PowerFood Microbial Kit, Qiagen) and incubated for 10 min at 70 °C. After this incubation step, the tubes were placed into a bead-beater (Precellys® Evolution, Berlin Technologies, Montigny-Le-Bretonneux, France) for two 45 s-cycles at 6 500 rpm. The inhibitor removal step was done according to the manufacturer’s protocol. 10 µL of RNASe A (10 mg/mL, Sigma-Aldrich) was added to the samples, which were incubated at room temperature for 10 min. After this incubation step, the manufacturer’s protocol was applied without any modification. The DNA samples were finally resuspended in 60 µL of elution buffer (10 mM Tris, pH 8, DNeasy PowerFood Microbial Kit).

The gene coding for 16S ribosomal RNA was amplified using FS1A (5′-AGAGTTTGATCCTGGCTCAG-3′) and FS5H (5′-AAGGAGGTGATCCAGCCGCA-3′) universal primers (Paillet et al. 2022). The amplification procedure was carried out in a 50 µL volume comprising DNA extract (20 ng), 0.25 µL GoTaq DNA polymerase (5 U/µL, Promega, Charbonnières-les-Bains, France), 0.2 µM of each primer and 0.8 mM dNTP. The amplification PCR conditions were: 94 °C for 5 min for initial denaturation, followed by 25 cycles of denaturation at 94 °C for 30 s, hybridisation at 57 °C for 30 s and polymerisation at 72 °C for 1.5 min. A final elongation step was carried out at 72 °C for 5 min, followed by cooling to 4 °C. DNA amplicons (expected size of 1.5 kb) were assessed on 1.25% w/v agarose gel and sent for sequencing to Eurofins Genomics (Köln, Germany). Raw sequences were cleaned. Chromatograms were processed by Chromas software (Technelysium Pty Ltd, South Brisbane, Australia) and trimmed sequences were finally compared to NCBI and EZbiocloud databases.

Determination of sugars and lactic acid concentrations

Lactose, glucose, galactose and total lactic acid concentrations were quantified by high performance liquid chromatography (HPLC). Samples were centrifugated at 13 200 g for 5 min at 4 °C and supernatants were filtered (0.22 µm). The Alliance Waters e2695 HPLC (Waters, Guyancourt, France) was equipped with an Aminex HPX-87H 300 mm × 7.8 mm column (Bio-Rad, Hercules, CA, USA). Temperature of the column was set at 35 °C and the mobile phase used was H₂SO₄ 5 mM at a flowrate of 0.4 mL/min. Detection was achieved with refractive index and UV (210 nm) detectors. Data collection was obtained using the Empower V3 software (Waters). All analyses were done in duplicate.

Using these data, the mass balances have been calculated for all experiments. On average, they were equal to 96.5 ± 2.8%.

Determination of dissociated and undissociated lactic acid concentrations

Undissociated lactic acid (, in g/L) and lactate (, in g/L) concentrations were determined thanks to the total lactic acid concentration determined by HPLC (, in g/L), the pH and the lactic acid dissociation constant (), using the Henderson-Hasselbach relationships (Schepers et al. 2002):

1

2

where ; ; for lactic acid = 8.3 × 10^–4.

Indirect measurement ofsugar and lactic acid concentrations

As total sugars (, in mol/L) and total lactic acid (, in mol/L) concentrations were determined intermittently during sampling, and as the NaOH consumption (, in mol/L) was measured continuously, calculations were performed to estimate these concentrations at each time interval (Béal & Corrieu 1995). The following equations were applied at pH 4.3:

3

4

Equations 3 and 4 were used, respectively, for smoothing the experimental data of total sugars and total lactic acid concentrations at pH 4.3. The smoothed concentrations of undissociated lactic acid and lactate were obtained by combining these equations with the Henderson-Hasselbach relationships (Eqs. 1 and 2).

Calculation of lactic acid production rate and volumetric productivity

The lactic acid production rate (, in g/L/h) was defined by the variation of the total lactic acid concentration (, in g/L) during a given time interval (, in h):

5

The lactic acid volumetric productivity (, in g/L/h) is calculated by the ratio of the lactic acid concentration (, in g/L) to the corresponding fermentation duration (, in h):

6

These variables were calculated by using the smoothed data obtained from Eqs. 3 and 4.

Statistical analyses

The statistical differences between batches, i.e., lactic acid and sugar concentrations, volumetric productivity and fermentation time, were assessed by one-way ANOVA, followed by Tukey & Dunnett multiple comparison tests, at 95% level, using XLSTAT software (Lumivero, Bordeaux, France).

Results

Lactic acid production by Lb. helveticus LH-B01 at different pH

The ability of Lactobacillus helveticus LH-B01 to produce lactic acid and to metabolize sugars from hydrolysed fresh whey was investigated by carrying out batch fermentations in bioreactors at pH 5.5, 4.7 and 4.3. As undissociated lactic acid is of interest for some industrial applications, it has been considered in addition to total lactic acid. Consequently, for each fermentation, the residual sugars concentration as well as the dissociated and undissociated lactic acid concentrations were calculated using Eq. 1 and 2. Additionally, total and undissociated lactic acid volumetric productivities at the end of fermentations were determined using Eq. 6. Results are shown in Figs. 2 and 3.

Fig. 2.

Dissociated and undissociated lactic acid and residual sugars concentrations at the end of fermentations of hydrolysed whey by Lactobacillus helveticus LH-B01 conducted at different pH (n = 2). Letters (a, b, c) indicated significant difference for each variable according to the pH (p < 0.05)

Fig. 3.

Total and undissociated lactic acid volumetric productivity at the end of fermentations of hydrolysed whey by Lactobacillus helveticus LH-B01 conducted at different pH (n = 2). Letters (a, b, c) indicated significant difference for each variable according to the pH (p < 0.05)

From Fig. 2, at the end of the fermentations done at pH 5.5, 43.6 ± 0.2 g/L of total lactic acid were produced while 98% of the sugars were consumed. The fermentation duration, calculated from the first sodium hydroxide consumption to the last, was 40 ± 2 h, thus leading to a high total lactic acid productivity (1.10 ± 0.07 g/L/h). By comparing the results obtained at pH 4.7 and pH 4.3, significant differences were pointed out at 95% level. At pH 4.7, nearly complete sugar conversion was obtained (around 3% residual sugars) with no significant difference with pH 5.5 (p > 0.05). A final amount of total lactic acid of 41.1 ± 0.3 g/L was achieved, which was lower than that obtained at pH 5.5 (p < 0.05). The fermentation time was significantly longer (93 ± 7 h; p < 0.05) and the total lactic acid productivity (0.44 ± 0.04 g/L/h) was strongly lowered comparing to pH 5.5 (Fig. 3). When fermentations were carried out at a pH of 4.3, the residual sugars concentrations remained significantly higher than at pH 5.5 (24%), thus leading to a lower total lactic acid concentration of 32.3 ± 1.2 g/L. The fermentation time was increased to 123 ± 6 h and the total lactic acid productivity (0.26 ± 0.00 g/L/h) was strongly reduced by 86%, compared to pH 5.5. However, as the fermentation pH decreased, the undissociated lactic acid concentration increased, from 0.98 ± 0.00 g/L at pH 5.5 to 8.7 ± 0.09 g/L at pH 4.3. As a consequence of this observation, the undissociated lactic acid volumetric productivity was enhanced to 0.056 ± 0.005 g/L/h at pH 4.7 and to 0.071 ± 0.003 g/L/h at pH 4.3, as compared to that obtained at pH 5.5 (0.025 ± 0.002 g/L/h). The ratios between undissociated lactic acid and total lactic acid volumetric productivities increased when pH decreased, from 2.2% at pH 5.5, to 12.7% at pH 4.7 and to 26.9% at pH 4.3, in agreement with the Henderson-Hasselbach relationship.

As a summary, cultivating Lb. helveticus LH-B01 at pH 4.3 led to the best results. However, it was not possible to obtain simultaneously high undissociated lactic acid concentration and volumetric productivity, together with a low residual sugars content and a moderate fermentation time, with the parental strain Lb. helveticus LH-B01. In order to cope with this limitation, an adaptation of the Lb. helveticus strain will be considered in the next section.

Acid adaptation of Lb. helveticus LH-B01 by continuous fermentation

To increase the ability of Lb. helveticus LH-B01 to produce a high amount of undissociated lactic acid, together with a high productivity and a low final sugar content, a strain adaptation protocol was designed, based on continuous cultures. The cultures were performed at decreasing pH, to progressively select mutants with improved properties. Applying a constant dilution rate together with decreasing the fermentation pH, allowed selecting, in the microbial population, the cells that were able to produce lactic acid at low pH, and to wash out the cells that didn’t resist acidic conditions. Thanks to this approach, acid-tolerant phenotypes were selected at low pH, as undissociated lactic acid proportion increased when pH decreased. At the end of the continuous fermentations, the dilution rate was enhanced, in order to select the cells that showed the highest specific growth rate, and to wash out the other. Two continuous cultures (A and B) were conducted in parallel, by using hydrolysed and sterile rehydrated whey. The time evolutions of pH and total, dissociated and undissociated lactic acid concentrations during these continuous cultures are shown in Fig. 4.

Fig. 4.

Evolution of pH, total, dissociated and undissociated lactic acid concentrations during two continuous cultures (A and B) of Lactobacillus helveticus LH-B01 carried out at decreasing pH and increasing dilution rate

From Fig. 4, the cultures displayed some similarities. An increase of total lactic acid concentration was observed until 100–120 h, i.e., until pH 4.6, reaching a value of 18 g/L for culture A and 21 g/L for culture B. Then, the two cultures evolved differently. In culture B, the lactic acid concentration decreased progressively between pH 4.6 and pH 3.6, whereas in culture A, it declined more sharply between pH 4.6 and pH 4.0. By the end of this observation period, the total lactic acid concentration reached 4.9 g/L in bioreactor A, at pH 4.0 and 4.2 g/L in bioreactor B, at pH 3.6. The dilution rate was then increased from 0.2 to 0.4 h⁻¹, between 480 and 526 h for culture A and from 0.2 to 0.5 h⁻¹ between 415 and 526 h for culture B. As a result, the final total lactic acid concentration in bioreactor B reached 1.9 g/L, whereas it attained 5.0 g/L in bioreactor A.

The differences may be explained by an irregular pH control in bioreactor A, due to the formation of clusters of cells around the pH probe. For this reason, the pH control was altered and the pH reached 3.8 at the end of the culture in bioreactor A, instead of pH 3.6 in bioreactor B. In bioreactor B, an erratic pH control was observed later, after 486 h of fermentation.

By considering the dissociated lactic acid concentrations, it was observed that they strongly decreased during the continuous fermentations (from a maximum of 15.1 g/L to 1.2 g/L in bioreactor A and from 16.9 to 1.1 g/L in bioreactor B), whereas undissociated lactic acid concentrations remained comprised between 0.1 and 4.3 g/L in bioreactor A or 5.8 g/L in bioreactor B. Consequently, the relative proportion of undissociated lactic acid increased during the cultures, from 12% (at pH 4.7) to 53% (at pH 3.8) in bioreactor A and 63% (at pH 3.6) in bioreactor B.

Based on these results, samples were collected at the end of the continuous cultures, after 526 h in bioreactor A and 486 h in bioreactor B. Two isolates were selected from bioreactors A (named LH-B01-A4) and B (named LH-B01-B4) and stored on frozen form. The two mutants were confirmed to belong Lactobacillus helveticus species by genome sequencing.

Comparison of lactic acid production and lactose consumption during batch cultures at pH 4.3 of parental strain and mutants

The two mutants Lb. helveticus LH-B01-A4 and Lb. helveticus LH-B01-B4 were tested for their capacity to produce lactic acid and to consume lactose during batch cultures of hydrolysed whey at pH 4.3. The results were compared to those obtained with the parental strain Lb. helveticus LH-B01, under identical conditions. They are shown in Fig. 5, which displays smoothed data obtained by indirect measurements of total sugars and lactic acid concentrations from NaOH consumption, obtained from Eqs. 3 and 4.

Fig. 5.

Time-course of total lactic acid (A) and total sugars concentrations (B) during batch fermentations of hydrolysed whey at pH 4.3 with parental strain Lb. helveticus LH-B01 and mutants Lb. helveticus LH-B01-A4 and Lb. helveticus LH-B01-B4 (n = 2). Time zero corresponds to the point at which the pH set point (pH 4.3) was reached

Batch fermentations started at the pH of hydrolysed whey (pH 6.0 ± 0.1). Once pH 4.3 was reached, pH control was automatically started. Latency phase duration differed among the three strains, due to different initial physiological states of the bacteria. To ensure comparability between strains, the time corresponding to the first injection of sodium hydroxide to regulate the pH was considered to be equal to zero in Fig. 5. The lactic acid production of the parental and mutants’ strains displayed similar profiles in the early production phase, followed by a progressive slowdown that differed among the strains. After about 20 h of fermentation, significant changes emerged between the strains, by considering final lactic acid concentration (p < 0.05) and fermentation duration (p < 0.05). The two batches of Lb. helveticus LH-B01-A4 exhibited a similar behaviour to that observed with the parental strain during the first 60 h of fermentation. However, final lactic acid concentrations were significantly (p < 0.05) lower (28.2 ± 0.6 g/L) for LH-B01-A4 than for LH-B01 (32.3 ± 1.2 g/L). Fermentation durations also differed (p < 0.01), as they reached 65 ± 17 h for LH-B01-A4, instead of 123 ± 6 h for LH-B01. Lb. helveticus LH-B01-B4 strain was able to achieve significantly higher (p < 0.05) final lactic acid concentrations (37.9 ± 0.9 g/L). Culture duration (100 ± 15 h) was not significantly different from those observed with strains LH-B01 and LH-B01-A4 (p > 0.05). As the cultures were stopped when no more NaOH was consumed to control the pH, i.e., when no more lactic acid was produced in the medium, these differences indicated that the three strains may resist differently to high lactic acid concentrations at pH 4.3.

The initial concentration of total sugars (48.6 ± 1.1 g/L) differed among the fermentations, reflecting some variability in the liquid whey composition. From Fig. 5, sugar consumption was similar for the three strains during the first 40 h of the fermentations and the residual sugar concentrations at the end of the fermentations did not differ significantly among the strains (12.5 ± 3.9 g/L; p > 0.05). The sugars were not totally consumed by the parental strain, as 12.7 ± 3.2 g/L of residual sugars (74% of consumed sugars) remained at the end of the fermentations. With Lb. helveticus LH-B01-A4, residual sugars concentration attained 15.9 ± 1.6 g/L (66% of consumed sugars). Finally, the residual sugars concentration observed with Lb. helveticus LH-B01-B4 were slightly lower (8.8 ± 3.6 g/L, 82% of consumed sugars) than with the other strains (p > 0.05).

By considering the undissociated lactic acid concentrations, differences were observed among the three strains. At pH 4.3, 73.4% of the lactic acid was present on the dissociated form and 26.6% on the undissociated form, according to the Henderson-Hasselbach relationship (Eq. 1). Consequently, the parental strain Lb. helveticus LH-B01 produced 8.7 ± 0.1 g/L of undissociated lactic acid. Strain LH-B01-B4 achieved a higher final undissociated lactic acid concentration of 10.1 ± 0.2 g/L, while strain LH-B01-A4 was less productive, resulting in a lower final concentration of undissociated lactic acid (7.5 ± 0.1 g/L). These results indicated that the metabolism of the three strains stopped at a critical undissociated lactic acid concentration that was significantly different among strains (p < 0.01). This critical undissociated lactic acid concentration represents the higher value of undissociated lactic acid concentration above which the cells were completely inhibited. Consequently, the metabolism of Lb. helveticus LH-B01-B4 was active until about 10.1 g/L of undissociated lactic acid, whereas that of strains LH-B01 and LH-B01-A4 was active until 8.7 g/L and 7.5 g/L, respectively.

From these results, it can be concluded that the mutant strain LH-B01-B4 was more productive by considering total and undissociated lactic acid production. In contrast, the mutant strain LH-B01-A4 was less efficient than the other ones according to these properties.

Comparison of lactic acid production rate and volumetric productivity during batch cultures at pH 4.3 of parental strain and mutants

The lactic acid production rates were calculated for the three strains, by using smoothed data for concentrations. Figure 6 shows the variation of the total lactic acid production rate as a function of the undissociated lactic acid concentration for the three strains.

Fig. 6.

Influence of undissociated lactic acid concentration on total lactic acid production rate during fermentations of hydrolysed whey at pH 4.3 with the parental strain Lb. helveticus LH-B01 and the mutants Lb. helveticus LH-B01-A4 and Lb. helveticus LH-B01-B4 (n = 2)

For all three strains, the rate of lactic acid production decreased continuously as the concentration of undissociated lactic acid increased. In addition, it declined sharply and cancelled out at the critical undissociated lactic acid concentration of each strain, in agreement with our previous results. Despite some differences appeared between repeated cultures, the lactic acid production rate of mutant Lb. helveticus LH-B01-B4 was always higher than those of the parental strain LH-B01 and the mutant LH-B01-A4. Moreover, it was maintained at higher undissociated lactic acid concentrations. On the contrary, the lactic acid production rate of mutant LH-B01-A4 was equivalent to that of the parental strain LH-B01, until undissociated lactic acid concentration reached 7.5 g/L, which characterized its critical undissociated lactic acid concentration.

The time evolutions of total lactic acid volumetric productivity during the fermentations done with the three strains are shown on Fig. 7. They allowed identifying the maximal lactic acid volumetric productivity that characterizes each strain at low lactic acid concentration and the final lactic acid volumetric productivity that provides information on strain’s ability to perform at high lactic acid concentration.

Fig. 7.

Time evolution of total lactic acid volumetric productivity during fermentations of hydrolysed whey at pH 4.3 with the parental strain Lb. helveticus LH-B01 and the mutants Lb. helveticus LH-B01-A4 and Lb. helveticus LH-B01-B4 (n = 2)

From Fig. 7, for all the batches of the parental and mutant strains, the same behaviour was observed. Lactic acid volumetric productivity increased during the first 15 to 25 h, with a correct reproducibility between batches. The parental strain Lb. helveticus LH-B01 achieved a maximal productivity value of 0.71 ± 0.08 g/L/h, followed by the strain LH-B01-B4 (0.62 ± 0.06 g/L/h), then by the strain LH-B01-A4 (0.46 ± 0.04 g/L/h). Next, a decrease of lactic acid productivity occurred, until reaching the critical undissociated lactic acid concentration of each strain. The final volumetric productivity values were calculated when lactic acid was no longer produced. They were not significantly different among the strains (p > 0.05): they were equal to 0.26 ± 0.00 g/L/h for the parental strain Lb. helveticus LH-B01, 0.45 ± 0.11 g/L/h for Lb. helveticus LH-B01-A4 and 0.39 ± 0.07 g/L/h for Lb. helveticus LH-B01-B4. However, they were achieved at different times, respectively 123 ± 6 h for Lb. helveticus LH-B01, 65 ± 17 h for Lb. helveticus LH-B01-A4 and 100 ± 15 h for Lb. helveticus LH-B01-B4, thus demonstrating differentiated performances, which can have an impact on industrial efficiency.

Selection of a highly undissociated lactic acid producing strain

In order to identify the best Lb. helveticus strain that simultaneously produced high lactic acid concentration at low pH (pH 4.3), together with a reduced fermentation time and a low residual sugar content, a new variable has been searched. It has been defined as the ratio between the lactic acid volumetric productivity and the residual sugars concentration and has been calculated at each time of the fermentations from smoothed data. By combining the two variables ‘volumetric productivity’ and ‘sugar consumption’, it helps identifying the strain that performs best. It is displayed as a function of undissociated lactic acid concentration in Fig. 8. Since liquid whey composition was variable, only the fermentations with similar initial sugars contents were represented.

Fig. 8.

Evolution of the ratio between lactic acid volumetric productivity and residual sugars concentration as a function of the undissociated lactic acid concentration during fermentations of hydrolysed whey at pH 4.3 with the parental strain Lb. helveticus LH-B01 and the mutants Lb. helveticus LH-B01-A4 and Lb. helveticus LH-B01-B4

At undissociated lactic acid concentrations lower than 6.5 g/L, which corresponded to total lactic acid concentration values lower than 25 g/L, the ratios were comprised between 0.031 and 0.036 g_{lactic acid}/g_sugars/h, whatever the considered strain (Fig. 8). During this phase, Lb. helveticus consumed the sugars efficiently and no significant inhibition due to undissociated lactic acid was observed. At higher lactic acid concentrations, the three strains displayed different behaviours. A sharp decline of this ratio was observed for Lb. helveticus LH-B01-A4 that stopped producing lactic acid for an undissociated lactic acid concentration above 7.5 g/L. For the parental strain Lb. helveticus LH-B01, this decline occurred at an undissociated lactic acid concentration of 8.7 g/L, and for the strain LH-B01-B4, at an undissociated lactic acid concentration of 10.1 g/L. These results supported our previous observations concerning the critical undissociated lactic acid concentration values of the three strains. In addition, they confirmed the interest of the mutant Lb. helveticus LH-B01-B4 that displayed high lactic acid concentration, production rate and volumetric productivity, together with low residual sugars concentration as compared to the other strains. This improved behaviour of strain Lb. helveticus LH-B01-B4 can be linked to enhanced intracellular pH homeostasis, membrane robustness, proton pump activity, or altered lactate export.

Discussion

Lactic acid production by Lb. helveticus LH-B01 at different pH

In the present study, the initial investigation was focused on the production of lactic acid by Lb. helveticus LH-B01 at three different pH levels, utilising a hydrolysed whey medium during batch fermentations. The final lactic acid concentrations obtained during the duplicated fermentations at pH 5.5 (43.7 ± 0.1 g/L) were similar to those reported in the literature with the same bacterial species, by considering fermentations done at pH 5.4 or 5.5 (Aeschlimann and Von Stockar 1989; Fitzpatrick and O’Keeffe 2001; Schepers et al. 2002). Depending on these authors, they varied between 42.3 and 47.3 g/L. However, due to longer fermentation time, the final lactic acid productivity obtained with Lb. helveticus LH-B01 (1.10 ± 0.07 g/L/h) was lower than in these studies. Aeschlimann and Von Stockar (1989), Fitzpatrick and O’Keeffe (2001), Schepers et al. (2002) reported comparable levels of lactic acid productivities, reaching respectively 1.76 and 1.74 g/L/h, by using whey permeate supplemented with yeast extract at pH 5.5. Similarly, Fitzpatrick and O’Keeffe (2001) obtained a productivity of 1.63 g/L/h in a medium composed of whey supplemented with whey protein hydrolysate at pH 5.4. These differences may be ascribed to the use of hydrolysed whey in our study instead of complemented whey. This led to a less rich medium than those used by the other authors. The use of another strain of Lb. helveticus can also explain the differences.

The use of a lower fermentation pH during the batch cultures of Lb. helveticus LH-B01 was associated to a decrease in lactic acid production, an augmentation of the fermentation duration and an increase of the residual sugars content. Whereas there was no significant difference in residual sugars concentration at the end of the fermentations done at pH 5.5 and 4.7, a significant decrease of 24% in sugar consumption was observed at pH 4.3. As a consequence, the final lactic acid concentration was significantly lower when the cultures were done at pH 4.7 or 4.3, as compared to pH 5.5. In the same time, the undissociated lactic acid concentration was increased to 8.7 g/L, a concentration that could hinder the metabolism, as the undissociated form of lactic acid (HLa) is favored at low pH (Schepers et al. 2002). In addition, the final lactic acid volumetric productivity was strongly reduced at pH 4.7 (by a factor of 2.5) and at pH 4.3 (by a factor 4.1), as compared to pH 5.5. This reduction was the consequence of longer fermentation times and lower final lactic acid concentrations at these low pHs. A comparable outcome was obtained by Schepers et al. (2002) who investigated the effect of pH on the growth of Lb. helveticus during fermentation of whey permeate medium enriched with yeast extract. By calculating the lactic acid volumetric productivity from their experimental data, a value of 0.33 g/L/h was found at pH 4.2. It was 5.4 times lower than that obtained at pH 5.5 by the same authors. In addition, this value was comparable to that obtained in our study (0.26 g/L/h).

As expected, carrying out batch cultures of Lb. helveticus LH-B01 at acidic pH led to an increase of the proportion of undissociated lactic acid in the medium, from 2.2% at pH 5.5 to 26.6% at pH 4.3. Notably, at pH 4.3, lactic acid production stopped when the concentration of undissociated lactic acid reached 8.7 ± 0.1 g/L. These results agreed with those obtained by Schepers et al. (2002), who indicated that the inhibitory effect of undissociated lactic acid on bacterial growth was stronger than that of dissociated lactic acid. These authors showed that the specific growth rate of Lb. helveticus during a fermentation at pH 4.5 was halved when the undissociated lactic acid concentration of was equal to 2.0 g/L and tended towards zero when it became greater than 9.0 g/L. In contrast, when the fermentation was done at pH 6.0, the specific growth rate was nullified when dissociated lactic acid concentration became higher than 50 g/L, which corresponded to an undissociated lactic acid concentration of 0.36 g/L. This greater inhibitory effect of undissociated lactic acid has been explained by its ability to freely permeate the cell membrane (Konings 2002), thus decreasing the intracellular pH. It also explicated the increase of the residual sugars concentration and fermentation time at low pH. Finally, the performance degradation observed at low pH can be explained by the inhibitory effect of high undissociated lactic acid concentration.

Acid adaptation of Lb. helveticus LH-B01 during continuous cultures

One major finding of this study concerned the obtention of mutants of Lb. helveticus LH-B01 with improved resistance to acidic conditions, by using continuous cultures. Two mutants, named LH-B01-A4 and LH-B01-B4 were isolated, each from a bioreactor, at the end of 22 days-cultures conducted at decreasing pH and increasing dilution rate. They displayed differentiated properties by considering their ability to produce lactic acid at low pH as compared to the parental strain. By considering final total and undissociated lactic acid concentrations, production rates and volumetric productivities, the mutant LH-B01-B4 was shown to be more resistant to acidic conditions. It also displayed the higher critical undissociated lactic acid concentration.

Only a few studies have been carried out on the adaptive evolution of microorganisms to acidic conditions using continuous cultures. Except for the work of Zhang et al. (2020) that focused on adaptation of Actinobacillus succinogenes NJ113 to improve succinic acid production at low pH, to our knowledge, nothing has been published by considering acid adaptation of lactic acid bacteria. Our approach was therefore original as it allowed selecting mutants that showed improved properties thanks to the progressive application of decreasing pH and increasing dilution rate during the continuous cultures.

Despite the two continuous cultures were performed in the same conditions, differences were observed between the fermentations. Indeed, some clusters of cells were formed in bioreactor A, which altered the pH control of the medium. This situation induced a specific behaviour of the culture, thus leading to the mutant LH-B01-A4 that was less effective than the parental strain, by considering its performance within acidic conditions. In contrast, pH control was better achieved in bioreactor B, thus leading to reach a lower final pH and a higher dilution rate and to obtain the mutant Lb. helveticus LH-B01-B4.

Comparison of lactic production by parental strain and mutants at pH 4.3

The performances of the mutants Lb. helveticus LH-B01-A4 and LH-B01-B4 have been compared to those of the parental strain Lb. helveticus LH-B01, in a context of lactic acid production in batch cultures at pH 4.3. As all experiments have been conducted at the same pH, the relative content of undissociated lactic acid was equivalent in all experiments (26.6%). The analyses regarding the undissociated lactic acid content were then equivalent to those concerning total lactic acid concentrations.

The total and undissociated final lactic acid concentrations obtained with the mutant Lb. helveticus LH-B01-B4 were significantly higher, as compared to those of the parental strain and Lb. helveticus LH-B01-A4. This result has to be linked to the critical undissociated lactic acid concentration values that differed among the strains. With a value of 10.1 g/L, Lb. helveticus LH-B01-B4 showed the highest lactic acid tolerance, thus allowing the culture to consume more sugars (the residual sugars content was the lower with this strain). In contrast, the mutant Lb. helveticus LH-B01-A4 that showed the lowest critical undissociated lactic acid concentration (7.5 g/L) induced degraded results, as it produced less lactic acid and consumed less sugars. This low critical undissociated lactic acid concentration of Lb. helveticus LH-B01-A4 explained why the fermentation stopped early, thus leading to a short fermentation time together with a high residual sugars concentration. Finally, the characteristics of the parental strain Lb. helveticus LH-B01 were close to those of strain Lb. helveticus LH-B01-A4 in terms of lactic acid concentrations and residual sugar contents. However, as the fermentation time was much longer, the performances of this strain were of poor quality at pH 4.3.

To help calculating production rates and volumetric productivities, experimental data have been smoothed, according to Eqs. 3 and 4. As the coefficients of determination were higher than 0.99, the regressions were considered as good. Coefficients of proportionality were determined at pH 4.3: they were equal to 1.26 mol of lactic acid produced per mol of NaOH and to 0.37 mol of total sugars consumed per mol of NaOH. These values were higher than those expected from the stoichiometry of homofermentative metabolism: 1.0 mol of lactic acid per mol of NaOH and 0.25 mol of total consumed sugars per mol of NaOH (Béal and Corrieu 1995). This difference can be explained by the higher buffering capacity of the medium at low pH (pH 4.3) as compared to pH 5.5, as demonstrated by Hill et al. (1985) and Wolfschoon Pombo et al. (2017). This buffering action minimizes changes in pH when the base is added to the medium. By considering this effect of buffering capacity, these values were in agreement with the homofermentative character of Lb. helveticus (Roy et al. 1986).

For each strain, the rate of lactic acid production decreased as the undissociated lactic acid concentration increased. It was abruptly cancelled at the critical concentration of undissociated lactic acid, whereas sugars were not completely consumed. This result confirmed that the lactic acid fermentation stopped because of inhibition by high undissociated lactic acid levels, as observed by Schepers et al. (2002). From the results obtained by these authors, a total lactic acid volumetric productivity of 0.32 g/L/h has been calculated. This value was slightly higher than that obtained in our study (0.26 g/L/h). This difference may be explained by the use of a richer medium by these authors (whey complemented with 10 g/L of yeast extract) as compared to hydrolysed whey in this work.

Finally, our results established the interest of the mutant Lb. helveticus LH-B01-B4 to produce undissociated lactic acid. This strain was able to reach high total and undissociated lactic acid concentrations, together with acceptable sugars consumption levels and good volumetric productivities, which were achieved thanks to the high undissociated lactic acid tolerance of this adapted strain.

Conclusion

The production of lactic acid is challenging, mainly because of the inhibition of the bacteria by lactic acid they produce, especially in its undissociated form. In this study, an adaptive laboratory evolution approach was designed to enhance lactic acid production by lactic acid bacteria under inhibitory conditions. By performing continuous cultures at decreasing pH and increasing dilution rate, it allowed to select two mutants, which exhibited divergent phenotypes. When cultivated in batch cultures at pH 4.3, the mutant Lactobacillus helveticus LH-B01-B4 displayed higher lactic acid concentrations, together with lower residual sugar contents and quite good volumetric lactic acid productivities. These characteristics were explained by the better tolerance to undissociated lactic acid of this strain, which exhibited a higher critical undissociated lactic acid concentration than the other strains. In a context of industrial applications, the use of the mutant Lb. helveticus LH-B01-B4 could enable the production of lactic acid at low pH, which is more suitable for use as a descaler and biocide in detergents. In the future, the comparison of the genomes of the two mutants with that of the parental strain may help identifying the genes that have been modified during the adaptation protocol and understanding the mutations that induced these differences. In addition, new experiments may be implemented in order to select strains with increased maximum specific growth rate, in order to further reduce the fermentation time and to increase the final volumetric productivity.

Author contributions

Lauranne Collet: Methodology, Investigation, Formal analysis, Writing—Original Draft; Catherine Béal: Conceptualization, Investigation, Formal analysis, Writing—Review & Editing, Supervision; Jérôme Delettre: Investigation; Violaine Athès: Writing—Review & Editing; Caroline Pénicaud: Writing—Review & Editing. All authors have read and approved the published version of the manuscript.

Funding

This work was carried out with the help of the financial support of ADEME (Agence De l'Environnement et de la Maîtrise de l'Energie) and ANRT (Association Nationale Recherche Technologie).

Data availability

Data available on request from the authors.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

Not applicable.