Effect of Formic Acid on Exopolysaccharide Production in Skim Milk Fermentation by Lactobacillus delbrueckii subsp. bulgaricus OLL1073R-1.

In yogurt, the formation of formate by Streptococcus thermophilus stimulates the activity of Lactobacillus delbrueckii subsp. bulgaricus (L. bulgaricus). However, there have been no reports how formic acid acts on the exopolysaccharide (EPS) production of L. bulgaricus. Here, the effect of formate on the EPS production in skim milk by L. bulgaricus OLL1073R-1 was investigated. After incubation for 24 hr with 100 mg/l formate, cell proliferation and lactic acid production were accelerated. The viable and total cell numbers were increased about ten- and four-fold, respectively. The amount of EPS in culture with formate (~116 µg/ml) was also four-fold greater than that of the control (~27 µg/ml). Although elongation of cells was observed at 6 hr of cultivation in both cultures, cells cultivated with formate returned to a normal shape after incubation for 24 hr. The sensitivity to cell wall hydrolase and composition of surface layer proteins, as well as the cell membrane fatty acid composition of L. bulgaricus OLL1073R-1, were not influenced by formate. However, differences were observed in intracellular fatty acid compositions and sensitivity to antibiotics. Cell length and surface damage returned to normal in cultures with formate. These observations suggest that formic acid is necessary for normal cell growth of L. bulgaricus OLL1073R-1 and higher EPS production.

INTRODUCTION

The physiological functions of dairy lactic acid bacteria are known, and Lactobacillus delbrueckii subsp. bulgaricus (L. bulgaricus) is one of the most important yogurt starters. Yogurt is generally made from milk using a culture of L. bulgaricus and Streptococcus thermophilus (S. thermophilus). In the fermentation process, L. bulgaricus and S. thermophilus exhibit a symbiotic relationship: S. thermophilus provides formic acid, pyruvic acid, folic acid [1] and carbon dioxide [2, 3] to L. bulgaricus, and then L. bulgaricus leads to S. thermophilus growth by peptides or amino acids production. Suzuki et al. reported that formic acid produced by S. thermophilus is taken into the cells of L. bulgaricus and utilized as a precursor in purine synthesis [4, 5]. S. thermophilus has pyruvate formate lyase, an enzyme that L. bulgaricus lacks [6, 7]; thus, provision of formic acid from S. thermophilus is necessary for the normal proliferation of L. bulgaricus.

Recently, exopolysaccharides (EPS) have been introduced as food emulsifiers/thickeners and functional food ingredients. The bacterial polysaccharides also confer health benefits in humans and livestock. Previously, we characterized EPS produced by L. bulgaricus OLL1073R-1 [8]. For yogurt fermented with this strain, acidic polysaccharides (APS: phosphopolysaccharide) exerted beneficial immunological effects such as the enhancement of lymphocyte mitogenicity, macrophage function [9,10,11], and NK cell activity [12], and the reduction of cold prevalence [13] and reduced influenza virus infections [14]. It is considered that these beneficial activities are attributable to the EPS produced by this strain. At present, yogurt fermented with this strain is marketed in Japan, but its EPS production is lower than that of other EPS-producing strains [15, 16].

In this study, we investigated the effects of several environmental conditions, including differing formic acid concentration in skim milk culture, on EPS production in L. bulgaricus OLL1073R-1.

MATERIALS AND METHODS

Bacterial strain and media

L. bulgaricus OLL1073R-1 was obtained from the cell culture collection of Meiji Co., Ltd. (Tokyo, Japan). Skim milk (10%) with or without formate was prepared at a concentration of 100 mg/l [17, 18] by addition sodium formate (10 mg/100 mL, 1.47 mM) based on formic acid production from S. thermophilus of 40 to 600 mg/l [19, 20]. The media were sterilized by low-temperature long-time (LTLT) pasteurization (63°C for 30 min) to avoid the formation of formic acid [21].

Culture conditions

The strain was propagated twice in 10% (w/v) skim milk (115°C, 15 min). The L. bulgaricus culture was inoculated into 10% skim milk with or without formate and incubated at 37°C for 24 hr.

After cultivation of OLL1073R-1 in skim milk without formate for 24 hr, formate (100 mg/l) was added to the medium and then incubated for a further 24 hr. A part of the same culture was inoculated into another batch of skim milk with formate (100 mg/l) and cultivated for 24 hr.

Fermentation conditions with formate

Total cell number, cell viability and pH of cultures were examined every three hours during incubation. Total cell numbers were determined according to the method of Sieuwerts et al. [17]. Briefly, turbidity (OD600 nm) was determined after mixing one volume of culture with nine volumes of 0.2% (w/v) NaOH and 0.2% sodium EDTA acid solution. Optical density was monitored using a UVmini-1240 spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Cell viability was determined by pour plating serial dilutions on de Man, Rogosa, and Sharpe (MRS) agar plates (Difco & BBL, Franklin Lakes, NJ, USA; 2% agar), incubating at 37°C for 48 hr, and enumerating the colony-forming units (CFU)/ml. For each culture, 6 CFU counts were performed. The pH was measured with a pH/ION meter (F-24, HORIBA, Tokyo, Japan) in duplicate.

Preparation and quantification of EPS

Cultures were sampled (1 mL) for EPS determination. Proteins were removed by adding 300 µL of 5 M trichloroacetic acid and centrifuging at 750 × g for 10 min at room temperature. The supernatant (500 µL) was collected, adjusted to neutral pH using 2.5 M NaOH (125 µL), and then filtered through an ultrafiltration kit (MW: 10,000 cut off; USY-1, Advantec Toyo Kaisha, Ltd., Tokyo, Japan). After washing five times with distilled water, retentates (EPS) were recovered in 500 µL distilled water. The EPS content as neutral sugars was determined using the phenol-H2SO4 method [22].

Microscopic analysis

After cultures were diluted 1:10 with 0.2% NaOH/EDTA solution, bacterial cells were collected by centrifugation (750 × g, 10 min, 4°C) and washed with sterilized distilled water (SDW). The resuspended cells were dried on a cover glass and fixed with cold acetone. Cells were coated with platinum palladium using a magnetron sputter coater (MSP-1S, Vacuum Device Inc., Mito, Japan) and observed using scanning electron microscopy (SEM; SU8000, Hitachi, Tokyo, Japan).

Sensitivity to cell wall hydrolase

Bacterial sensitivity to cell wall hydrolase was spectrophotometrically determined using a 96-well microplate. After 24 hr of culture, cells were collected by adding nine volumes of 0.2% NaOH/EDTA solution, sterilization in boiling water, washing with SDW, lyophilization and resuspension at 1 mg/ml in SDW. The cell suspensions (95 µL), N-acetylmuramidase SG (from Streptomyces globisporus 1829, Seikagaku Biobusiness, Tokyo, Japan; 5 µL, 1 mg/ml SDW or 5 mg/ml SDW) and 10 mM Tris-maleate-NaOH buffer (containing 4 mM MgCl2, pH 7.0; 100 µL) were mixed in each well and incubated at 37°C for 3 hr. The optical density at 600 nm was monitored with an absorbance microplate reader system (Viento VS, Dainippon Sumitomo Pharma Co., Ltd., Osaka, Japan) and the Gen5 software (Biotek, Winooski, VT, USA).

SDS-PAGE of surface layer protein

Cells after cultivation for 24 hr were collected with 0.2% NaOH/EDTA solution. Surface layer proteins (SLPs) from the cells were suspended in 4 M guanidine hydrochloride and stirred at 37°C for 2 hr. Following precipitation with cold acetone, SLPs were examined by SDS-PAGE according to the method of Laemmli [23] using a 15% (wt/vol) polyacrylamide separation gel and stained with a Rapid CBB kit (Kanto Chemical Co., Inc., Tokyo, Japan).

Fatty acid analysis of the cell membrane using gas chromatography

Cells prepared in 0.2% NaOH/EDTA solution were washed three times with PBS and then lyophilized. The lipids were extracted in 1:2:0.8 (v:v:v) chloroform:methanol:water following the method of Bligh and Dyer [24]. Neutral and polar lipids were separated using a Sep-Pak Plus NH2 Cartridge (Waters Corporation, Milford, MA, USA). Samples were converted to fatty acid methyl esters and analyzed using a Hitachi G-6000 gas chromatograph equipped with flame ionization detectors and a TC-70 column (0.25 mm × 60 m, GL Sciences Inc., Tokyo, Japan) in He carrier gas. The injection and detection port temperature were both 260°C, and the column temperature was held at 140°C for 5 min, increased at 4°C/min to 180°C, held at 180°C for 10 min, increased at 20°C/min to 250°C and then held at 250°C for 10 min. Peaks were identified by comparing the retention times to a FAME standard mix containing 37 components.

Resistance activity against antibiotics

Resistance against two antibiotics was evaluated using a plate assay. Vancomycin hydrochloride (1,050 IU/mg) and bacitracin (40 units/mg) were prepared at 5.12 mg/ml, serially diluted with sterilized distilled water and used at the following concentrations: 32, 64 and 128 µg/ml for vancomycin and 32, 64, 128 and 256 µg/ml for bacitracin.

Minimum inhibitory concentration (MIC) was determined using the plate assay method. Cultures in 10% skim milk with or without formate were serially diluted with saline solution, 100 µL of the 102 dilutions were lawn plated on BCP plate count agar (Eiken Chemical Co., Ltd., Tokyo, Japan) into which 6-mm diameter wells were cut and 20 µL of antibiotic solutions were added. The plates were incubated at 37°C for 24 hr. Each sample was assayed in five independent determinations, and the diameter of the clear zone around the well was measured using a caliper.

Statistics

Statistical significance of differences was determined using the Student’s t-test.

RESULTS

Effect of environmental conditions on growth and EPS production of L. bulgaricus OLL1073R-1

The addition of formic acid in 10% skim milk improved the cell number and viability of L. bulgaricus OLL1073R-1 and lowered the pH (Fig. 1). The differences were especially large from the exponential to stationary phase, and the differences were significant for incubation at 6 to 24 hr (p<0.001). The addition of formate increased the amount of EPS more than fourfold compared with the control (116.0 vs. 26.6 µg/ml) incubation for 24 hr (Fig. 2).

Fig. 1.

Culture kinetics of L. bulgaricus OLL1073R-1 in skim milk medium incubated at 37°C for (A) culture pH, (B) cell viability and (C) total cell number of the L. bulgaricus OLL1073R-1 strain in skim milk medium (● and solid line) with or (▲ and dotted line) without formate addition. The average of two pH readings and the mean and standard deviation of six different samples for viability and optical density are presented.

Fig. 2.

EPS concentration after incubation for 24 hr in the (□) absence or (■) presence of formate. Each bar represents the mean and the standard deviation of four different samples. (*** p<0.001).

Effect of formate on morphological alteration of cells

Morphological changes in L. bulgaricus OLL1073R-1 after 6 and 24 hr of incubation are shown in SEM images (Fig. 3). In both cultures, cells were abnormally elongated to over 50 μm in length after 6 hr of incubation (Fig. 3-A, B). After incubation for 24 hr, cells grown with formate changed and became shorter than those grown without formate (Fig. 3-C1, D1). In the culture without formate, cells were easily damaged by the electron irradiation under reduced pressure in the pretreatment for SEM, but very few cells were damaged in the culture with formate (Fig. 3-C2, D2).

Fig. 3.

Scanning electron micrographs of L. bulgaricus OLL1073R-1 grown in skim milk medium. L. bulgaricus OLL 1073R-1 cells were cultured in the (A) absence or (B) presence of formate for 6 hr and (C) absence or (D) presence of formate for 24 hr and are shown at two magnifications: 1, × 5,000; 2, × 30,000. Arrows indicate the part of remarkable cell damage.

After the L. bulgaricus OLL1073R-1 culture without formate (shown in Fig. 3-C1) was inoculated into another batch of skim milk with formate (100 mg/l) and incubated for another 24 hr, the cell shape became almost normal (Fig. 4-B1). The shape of the cells in the culture grown in the presence of formate for two cycles was not changed (Fig. 4-A1). Comparison of the cell surface structure showed that the improvement was observed in both cells (Fig. 4-A2, B2). Inoculation into the medium with formate resulted in the normalization of cell length and reduced damage to the cell surface.

Fig. 4.

Scanning electron micrographs of L. bulgaricus OLL1073R-1 strains grown (A) with successive addition of formate after incubation for 24 hr or (B) without formate for 24 hr and then inoculated into skim milk medium with formate and incubated for 24 hr. Two magnifications are shown: 1, × 2,000; 2, × 10,000. Arrows indicate the part of remarkable cell damage.

Chemistry of the cell surface

Changes were detected in the peptide glycan and cell membrane of the cell surface after formate addition (Fig. 5, Fig. 6, 7). Although significant differences were noted in cell lysis between with or without formate (Fig. 5(A), 1 hr and 2 hr: p<0.001), the effect of formic acid addition could not be determined. Surface layer proteins from L. bulgaricus OLL1073R-1 cells cultivated with or without formate showed similar band patterns of surface layer proteins; however, there were some differences in the low molecular weight proteins (MW: 5-20 kDa). The N-terminal amino acid sequences of five bands corresponding to low molecular weight proteins (<20kDa) were analyzed using FASTA. The proteins were considered to have originated from the medium and cytoplasm. Further, determination of the fatty acid composition of the cell membrane and cytoplasm of L. bulgaricus OLL1073R-1 cultivated with or without formate showed the following distribution: 6 polar lipids, 4 of which were identified, and 18 to 20 neutral lipids, 12 of which were identified. The abundance ratio of fatty acids in the polar lipid did not differ between with or without formate. For the neutral lipids, the concentrations of two peaks (C18 and C18:1n9c) were lower in cells cultured with formate than without formate, and three peaks (C18:3n3, C20:1 and C20:2/C22) were higher.

Fig. 5.

Hydrolysis patterns of N-acetylmuramidase SG in L. bulgaricus OLL1073R-1. The cell numbers in the presence of (A) 25 and (B) 125 µg/ml N-acetylmuramidase SG are shown. Solid (●) and dotted (▲) lines indicate cultures with or without N-acetylmuramidase SG, respectively. Optical density was monitored at 600 nm. Data are presented as the mean and standard deviation of six experiments (*** p<0.001).

Fig. 6.

SDS-PAGE (15%) of cell SLPs from L. bulgaricus OLL1073R-1 stained with Coomassie blue. Lane M, molecular weight standard marker (BenchMarkTM protein ladder, Invitrogen); lane 1, SLPs in culture without formate; lane 2, SLPs in culture with formate.

Fig. 7.

Gas chromatograms of fatty acid analysis of the cell membrane and cytoplasm from L. bulgaricus OLL1073R-1 strains grown at 37°C in skim milk medium (A) without formate or (B) with formate. 1, polar lipids; 2, neutral lipids.

Antibiotic sensitivity

The diameters of the inhibition zones for different concentrations of antibiotics against L. bulgaricus OLL1073R-1 on BCP plate count agar are shown in Table 1. Applying a regression formula to the data for the two antibiotics in cultures with or without formate put the MICs of vancomycin and bacitracin in cultures with formate at 0.14 and 1.21 µg/ml, respectively, and at 0.47 and 1.91 µg/ml, respectively, for cultures without formate. Thus, the addition of formate statistically increased (p<0.05) antibiotic resistance in L. bulgaricus OLL1073R-1.

Table 1.

Diameters of inhibition zones around antibiotics at varying concentrations with or without formate

	Without formate	With formate
	Diameter (cm)	Diameter (cm)
Vancomycin
32 μg/ml	2.330 ± 0.054	2.140 ± 0.043***
64 μg/ml	2.578 ± 0.099	2.386 ± 0.088*
128 μg/ml	2.894 ± 0.024	2.538 ± 0.036***
Bacitracin
32 μg/ml	1.882 ± 0.037	1.746 ± 0.055**
64 μg/ml	2.203 ± 0.043	2.052 ± 0.067**
128 μg/ml	2.515 ± 0.044	2.235 ± 0.061***
256 μg/ml	2.830 ± 0.082	2.507 ± 0.085***

The data are shown as means ± standard deviations of five different plates (*p<0.05; **p<0.01; ***p<0.001).

DISCUSSION

In yogurt fermentation, L. bulgaricus utilizes formic acid supplied from S. thermophilus. Here, we made a detailed investigation of the contribution of formate to EPS production of L. bulgaricus OLL1073R-1. Increased cell number, decreased pH and increased EPS production were observed with the addition (100 mg/l) of sodium formate (Fig. 1), and the amount of EPS production increased more than fourfold compared with the control (Fig. 2). These data show that the increase of EPS production is closely related to cell numbers, indicating that EPS production is dependent on cell proliferation rate. Therefore, increasing cell number is important for the higher EPS production in skim milk in yogurt production. Recently, Mende et al. reported that the addition of a mixture of salts including nucleobases, vitamins, and oleic acid and containing sodium formate to a semi-defined medium increased EPS production in L. bulgaricus [25] without affecting cell growth [26]. They suggested that the metabolic pathway of EPS biosynthesis may be activated by a combination of several chemical components, including formic acid.

The cell surface structures (Figs. 3 and 4) and antibiotic sensitivity (Table 1) of L. bulgaricus OLL1073R-1 cells were affected by the addition of formate in this study. These findings suggest that formic acid is related to the normal construction of peptidoglycan, including cross-linkage formation in cell wall synthesis. In intermediate lipid biosynthesis during cell wall construction, vancomycin and bacitracin interfere with the reaction of the lipid carrier in the process of peptidoglycan cross-linkage formation. Vancomycin is a kind of glycopeptide and lipoglycopeptide antibiotic and inhibits bacterial cell wall peptidoglycan synthesis [27]. Vancomycin attaches to the terminal D-alanyl-D-alanine group of the disaccharide pentapeptide and blocks the synthesis of peptidoglycans by producing steric hindrance (inhibits transfer of GlcNAc-MurNAc-pentapeptide-(Gly)5 by a lipid carrier (undecaprenyl-PP) toward the cell wall). On the other hand, bacitracin interferes with the dephosphorylation of the lipid carrier during the process of peptidoglycan cross-linkage formation. Namely, tight binding of bacitracin with undecaprenyl pyrophosphate prevents recycling of the sugar carrier, thereby inhibiting cell wall synthesis and forming a tertiary complex in the presence of bivalent ions, which prevents pyrophosphatase from interacting with C-55 isoprenyl pyrophosphate, thus reducing the amount of C-55 isoprenyl phosphate that is available for carrying sugar-peptide units [27]. It is considered that formic acid might bind to the D-Ala-D-Ala terminus of the pentapeptide peptidoglycan precursor, resulting in normalization of cell shape.

A lipid carrier is also necessary for EPS production [28, 29]. The lipid carrier does not compete with cell division and EPS synthesis because cell wall construction and cell division mainly occur in the exponential phase, while EPS synthesis occurs in the stationary phase [19, 29]. Based on these findings, formate is suggested to promote both cell division and EPS synthesis via a lipid carrier.

In this study, some changes were observed in intracellular fatty acid composition. Lipid composition of the cell membrane in L. bulgaricus was reported to be different following exposure to low pH, high or low temperatures or oxidative or osmotic stress conditions [30]. The change of lipid composition reflects adaptation to environment. Thus, the intracellular fatty acid composition may be related to antibiotic sensitivity. In addition, the C20:5n3 peak, which is eicosapentaenoic acid (EPA), was detected in intracellular fatty acids prepared from the culture without formate. Recently, EPA was found to be related to cell division in Shewanella livingstonensis [31]. Although L. bulgaricus and S. livingstonensis are far taxonomically, it seems that the processes of EPA metabolism, formic acid utilization and cell elongation may be related between these two strains.

In cocultivation with L. bulgaricus and S. thermophilus, formic acid is produced by S. thermophilus and is supplied to L. bulgaricus from the early exponential phase [19]. In this experiment, an unusual morphological shape was observed in the exponential phase (after 6 hr of cultivation) of the culture with formate, but the cell length returned to normal when the cultivation continued for 24 hr to the stationary phase. Thus, intake of formic acid is necessary for both cell growth and normal cell division. Suzuki et al. showed that formic acid is a necessary precursor of purine synthesis in DNA reproduction [4, 5] and that cell elongation is induced by having RNA and/or protein that is insufficient for separation [4]. Although it is not found the mechanism that lack of formic acid causes abnormal cell division, formic acid may play a critical role in the biosynthesis of cell division proteins, such as filamentation temperature-sensitive protein Z (FtsZ) [32], or in the other proteins involved in the process of peptide glycan formation.

In this study, we suggest that the control of formic acid levels is important to the regulation of EPS production of L. bulgaricus OLL1073R-1 in yogurt fermentation. The selection and application of strains of S. thermophilus with high formic acid production would be useful for the development of EPS-rich yogurt cultures in the future.