Production of Exopolysaccharide by Lactobacillus rhamnosus R and Analysis of Its Enzymatic Degradation during Prolonged Fermentation

Abstract

The potential of Lactobacillus rhamnosus R for producing exopolysaccharide (EPS) when grown on basal minimum medium supplemented with glucose or lactose was investigated. EPS production by L. rhamnosus R is partially growth associated and about 500 mg of EPS per liter was synthesized with both sugars. The product yield coefficient (Y_EPS/S) was 3.15 (0.0315 g of EPS [g of lactose]⁻¹) and 2.88 (0.0288 g of EPS [g of glucose]⁻¹). It was clearly shown that the amount of EPS produced declined upon prolonged fermentation. Degradation of EPS in fermentation processes was also assessed by measuring its molecular weights and viscosities. As these reductions might have a negative effect on the yield and viscosifying properties of EPS, it was essential to examine possible causes related to this breakdown. The decrease in viscosities and molecular weights of EPS withdrawn at different cultivation times permitted us to suspect the presence of a depolymerizing enzyme in the fermentation medium. Our study on enzymatic production profiles showed a large spectrum of glycohydrolases (α-d-glucosidase, β-d-glucosidase, α-d-galactosidase, β-d-galactosidase, β-d-glucuronidase, and some traces of α-l-rhamnosidase). These enzymes were localized, two of them (α-d-glucosidase and β-d-glucuronidase) were partially purified and characterized. When incubated with EPS, these enzymes were capable of lowering the viscosity of the polymer as well as liberating some reducing sugars. Upon prolonged incubation (27 h), the loss of viscosity was increased up to 33%.

Exopolysaccharides (EPS) produced by lactic acid bacteria (LAB) have generated increasing attention among researchers for the last few years. LAB are food-grade organisms, and the EPS that they produce contribute to the specific rheology and texture of fermented milk products. These EPS represent safe additives for novel food formulations and may have applications in nonfood products (8).

There exist three important groups of EPSs produced by LAB: (i) α-glucans, mainly composed of α-1,6- and α-1,3-linked glucose residues, namely, dextrans produced by Leuconostoc mesenteroides subsp. mesenteroides and L. mesenteroides subsp. dextranicum and mutans produced by Streptococcus mutans and S. sobrinus; (ii) fructans, mainly composed of β-2,6-linked fructose molecules, such as levan produced by S. salivarius; (iii) heteropolysaccharides produced by mesophilic (Lactococcus lactis subsp. lactis and L. lactis subsp. cremoris) and thermophilic (Lactobacillus delbrueckii subsp. bulgaricus, L. helveticus, and S. thermophilus) LAB (5). The EPS produced by Lactobacillus rhamnosus belong to third group (14). The sugar composition of the EPS produced by L. rhamnosus R studied in this work, as determined on the hydrolysate by high-pressure liquid chromatography (HPLC) and by gas chromatography of the alditol acetates, is the following: Rha, 4; Glc, 2; and Gal, 1 (M. R. Van Casteren, personal communication).

A considerable variation can be observed in EPS quantifications. The amount of EPS reported varies from 25 to 132 mg/liter for L. lactis subsp. cremoris (26) and L. rhamnosus C83 (14) and from 130 to 250 mg/liter for Lactobacillus casei CG11 (3) and L. delbrueckii subsp. bulgaricus NCFB 2772 (18), respectively. Mozzi et al. (28) reported an EPS production of 488 mg/liter for L. casei. The highest production levels ranged from 1,200 mg/liter (L. rhamnosus 9595M) (13) to 1,375 mg/liter (L. sake 0-1) (34).

Many studies showed a decrease in the total EPS amount when incubation times were increased (4, 5, 7, 8, 9, 11, 15, 16, 28). The decreased EPS level upon prolonged fermentation may be due to an enzymatic degradation (4, 5, 6, 9, 15, 16) or a change in the physical parameters of culture (9, 11, 15, 16). Gancel and Novel (15) suggested some reversible DNA rearrangements leading to different cell types which differ in exopolymer production capabilities. However, the possible relationship between EPS production and the factors contributing to EPS degradation have not been investigated yet.

In this study, the potential of L. rhamnosus R to produce EPS was investigated in a chemically defined medium supplemented with fermentable sugars. Furthermore, since a breakdown in EPS quantity and viscosifying properties could be observed during prolonged fermentation, we attempted, for the first time, to elucidate the possible linkage between enzyme activities present in cell extracts and the EPS yield. Indeed, it is important for the application of EPS in products and processes to investigate the mechanism involved in EPS degradation during fermentation processes. A similar analysis might lead to an efficient production of EPS with desired properties, in which EPS breakdown can be minimized. Purification and characterization of hydrolytic enzymes isolated from cell extract are also presented.

MATERIALS AND METHODS

Microorganisms and culture conditions.

L. rhamnosus R was obtained from Rosell Institute (Montréal, Québec, Canada). Stock cultures were stored at −40°C in brain heart infusion (Difco) broth in 15% (vol/vol) glycerol. Before experimental use, the cultures were propagated twice in MRS (Difco) at 37°C for 16 h.

Fermentation.

Fermentations were performed at 37°C in 7-liter Chemap fermenters (Chemapec, Inc., Woodbury, N.Y.) containing 6 liters of working volume of the culture medium (BMM) as described previously (27). The pH was controlled at 6.0 with 7 N NH₄OH. The fermenters were maintained by constant stirring at 100 rpm, and no air was added. The culture medium was inoculated with a 16-h active culture at the rate of 1% (vol/vol), and fermentations were allowed to proceed for 72 h. Samples were aseptically withdrawn at different times to determine the EPS yield, the LAB concentration, and the residual lactose and glucose concentrations. Samples were cooled on ice immediately after removing them from the fermenters. Biokinetic parameters such as the maximal specific growth rate (μ_max) and the EPS yield coefficient (Y_EPS/S) were calculated.

Growth determination.

Growth was monitored by measurement of the optical density (OD) at 650 nm. Cell numbers (CFU per milliliter) were estimated by plating the diluted samples on solid MRS medium.

EPS isolation and purification.

EPS were isolated and purified according to the method of Cerning et al. (3). The cultures were heated at 100°C for 15 min to inactivate enzymes potentially capable of polymer degradation, and the cells were removed by centrifugation at 12,785 × g for 30 min at 4°C. The EPS were precipitated with 3 volumes of chilled 95% ethanol. After standing overnight at 4°C, the resultant precipitate was collected by centrifugation (11,325 × g, 20 min). The EPS was dissolved in deionized water and dialyzed against deionized water at 4°C for 24 h and then lyophilized. The lyophilized powder was dissolved in 10% trichloroacetic acid to remove proteins. The supernatant was dialyzed at 4°C against deionized water for 5 days and lyophilized. These preparations were referred as purified EPS and stored at 4°C.

Analytical methods.

All chemicals and reagents were purchased from Sigma Chemical Co. and Pharmacia Biotech. The protein concentrations were measured by bicinchoninic acid protein method (Sigma). The total neutral carbohydrate was determined by the phenol-sulfuric method of Dubois et al. (12). Residual sugar levels and lactic acid concentrations were determined by HPLC on a Waters chromatograph (Milford, Mass.) using a Waters 410 RI detector and a Linear UVIS 200 detector. This system was equipped with an Ion-300 column (Interaction Chromatography, San Jose, Calif.) maintained at 35°C. The mobile phase was 0.02 N H₂SO₄ with a fixed flow rate of 0.4 ml min⁻¹.

Preparation of cell extracts (crude enzymatic preparation).

Cells were harvested by centrifugation at 12,785 × g at 4°C for 20 min, washed twice with 50 mM sodium phosphate buffer (SPB) (pH 7.0), and then resuspended in the same buffer to give a final concentration of 20% (wt/vol) wet weight. The cells were disrupted in a mixer mill at 4°C for 1 h. Cell debris was removed by centrifuging at 51,200 × g and 4°C for 20 min, and the clear solution obtained was used as a crude enzyme extract.

Assay for glycohydrolasic enzymes.

Glycohydrolase activities in regard to α-d-glucosidase, β-d-glucosidase, α-d-galactosidase, β-d-galactosidase, β-d-glucuronidase, and α-l-rhamnosidase were determined by measuring the rate of paranitrophenol (PNP) released from the appropriate p-nitrophenyl sugars. The reaction was stopped by addition of 0.5 ml of 1 M Na₂CO₃, and the amount of PNP released was determined by spectrophotometry at 420 nm. One unit of enzyme activity was defined as the amount of enzyme required to liberate 1 μmol of p-nitrophenol per min.

Ion-exchange chromatography.

The cell extract was fractionated in an FPLC Biopilote Q-Sepharose 35/100 (Pharmacia) column, equilibrated with 50 mM Tris-HCl buffer (pH 7.4) and eluted with the same buffer containing 1 M NaCl. Elution was performed at a flow rate of 2.5 ml min⁻¹ with a linear gradient of 0.2 to 1 M NaCl, and 5-ml fractions were collected. These fractions were immediately desalted by using desalting gel PD-10 Sephadex G-25M (Pharmacia Biotech). Active fractions were pooled and then concentrated with Centriprep-10 concentrators (Amicon).

Gel filtration chromatography.

This partially purified enzyme solution was then applied on an FPLC Hi-Load 16/60 Superdex 75 (Pharmacia Biotech) column equilibrated with 100 mM SPB (pH 7.0). Elution was carried out at a flow rate of 0.3 ml min⁻¹, and 1.5-ml fractions were collected.

Determination of optimal pH and temperature, pH, and thermal stability.

The effect of pH on enzyme activities was determined in 100 mM sodium acetate and 100 mM sodium phosphate buffers with a pH ranging from 3.6 to 8.0. To measure pH stability, reaction mixtures containing enzymes and buffers at various pH values were kept at 37°C for 30 min.

The influence of temperature on enzymatic activity was determined by incubating the assay mixture at temperatures from 30 to 80°C. Thermal stability was determined by incubating the enzymes at temperatures from 30 to 80°C for 30 min.

Effect of metal ions, inhibitors, and other substances on enzyme activities.

Stock solutions of CaCl₂, HgCl₂, MgCl₂, CoCl₂, CuSO₄, MnSO₄, FeSO₄, ZnSO₄, NaCl, KCl, and LiCl were prepared in sodium acetate buffer (pH 5.0). Inhibitors such as 2-mercaptoethanol, dithiothreitol (DTT), and urea, as well as compounds such as EDTA, were also tested.

Apparent molecular weights.

The purity of the enzyme-containing fractions was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) after each chromatographic step. The molecular mass of the purified enzyme was estimated by SDS-PAGE with reference to molecular mass standards from 14.4 to 94 kDa (Pharmacia Biotech) and by gel filtration chromatography on a Bio-Sil Sec 250 (Bio-Rad) column.

Viscosity measurement.

A Brookfield digital viscosimeter model DV-II with a spindle 18 was used to measure apparent viscosity. Viscosity was expressed in millipascals. Intrinsic viscosity was determined with a capillary Ubbelohde viscometer no. 1B at 25°C.

Ability of glycohydrolases to degrade EPS.

The ability of the isolated enzymes to degrade EPS was determined by examining changes in viscosities and liberation of reducing sugars. EPS degradation was initiated by the addition of cellular extract or purified enzymes to EPS solution in sodium acetate buffer (pH 5.0) containing 0.02% NaN₃. After different incubation intervals at 37°C, the capillary viscosities were determined with a Ubbelohde viscometer at 25°C, and the liberated reducing sugars were detected by the Nelson-Somoygi method.

Determination of EPS apparent molecular mass.

The apparent molecular mass of the EPS was determined by gel permeation chromatography. Chromatography was performed on a Bio-Rad gradient-module HPLC equipped with an Ultrahydrogel Linear Column of Waters (7.6 by 350 mm) at room temperature. The solvent was 0.1 N NaCl, and the flow rate of mobile phase was fixed at 0.4 ml min⁻¹. The sample size was 20 μl, and polymer standard from Polymer Laboratories (Amherst, Mass.) was used as a standard for the molecular weight determination. A Bio-Rad RI Detector model 1755 was used to detect the EPS.

Detection of lytic activity.

Bacteria were grown on the surface of MRS agar containing 0.2% (wt/vol) autoclaved and lyophilized Micrococcus luteus (Sigma Chemical Co.) cells. Agar plates were incubated at 37°C. Lytic activity was detected as clear zones around colonies in the agar.

RESULTS

Growth and kinetics of EPS production.

L. rhamnosus R was grown on glucose- or lactose-supplemented (20 g/liter) basal minimum medium at 37°C. The pH was maintained at 6.0 by titration with 7 N NH₄OH. Glucose and lactose fermentation profiles were similar, as presented in Fig. 1. The course of biomass and metabolite production showed that both of the fermentations were partially growth associated. Only after 15 h of incubation did the pattern of fermentation show parallels between EPS synthesis and growth rate (CFU and OD). The lag growth phase lasted up to 12 h. The exponential growth phase occurred over a period of approximately 9 h. There was no stationary growth phase since cell numbers dropped quickly after reaching their maximum after 21 h. At the end of the fermentation, the lactic acid produced was about 17 g/liter. L. rhamnosus R was able to grow in glucose BMM with a maximal growth rate of 0.32 h⁻¹ and a final OD of 4.05; producing 438 mg of EPS per liter. On the other hand, in the medium supplemented with lactose, the growth rate was 0.46 h⁻¹ with a final OD of 5.18 and an EPS level of 495 mg/liter. When lactose was used as a carbon source, the product yield coefficient (Y_EPS/S) was 3.15 (ca. 0.0315 g of EPS [g of lactose]⁻¹) compared to 2.88 (ca. 0.0288 g of EPS [g of glucose]⁻¹). Cells entered quickly into the decline phase, probably due to carbon source limitation (glucose or lactose levels dropped to zero after 24 h). In the case of lactose, the EPS was produced mainly during the exponential growth phase and reached its maximum level during the early decline stage of growth. With glucose as a carbon source, the EPS production continued during the decline phase, reaching its maximum at 36 h. Finally, it was observed that EPS quantity decreased during the early decline phase in both cases, and this reduction was more pronounced when lactose was the carbon source. When the incubation period was extended up to 48 h, the reduction in EPS was remarkable (82%).

FIG. 1.

Batch fermentation profile of L. rhamnosus R growth and EPS production at 37°C and constant pH 6.0. Each value represents the average of triple measurements. (A) Growth on lactose. (B) Growth on glucose.

In order to clarify whether this reduction of EPS production was due to enzymatic degradation, as speculated by some authors (3, 5, 9), it was thought pertinent to carry out a study of enzymatic production profiles by L. rhamnosus R.

Glycohydrolase production and localization.

The Api-Zym test showed that cell lysates were strongly active against some glycoside substrates. In an attempt to localize these enzymes, enzyme activities (for enzymes α-d-glucosidase, β-d-glucosidase, α-d-galactosidase, β-d-galactosidase, and β-d-glucuronidase) in extracellular, intracellular, and cell-bound fractions were determined (Table 1). The presence of three enzyme activities (for enzymes α-d-glucosidase, β-d-glucosidase, and β-d-glucuronidase) in the extracellular fraction was indicated. The percentage of glycohydrolase activities linked to the cells was high. The activity of α-glucosidase in the intracellular fraction was highest, showing that the enzyme is mostly intracellular. However, the high activity of this enzyme was also detected in cell-bound fractions. The activity of β-glucosidase was highest in the cell-bound fraction, showing that this enzyme is mostly cell bound rather than intracellular. The activities of α- and β-galactosidases were detected in the cell-bound fraction, as well as intracellularly, but not in the extracellular fraction. These activities in the intracellular fractions dropped to zero in cells harvested after 24 h, however.

TABLE 1.

Specific activity of glycohydrolases produced by Lactobacillus rhamnosus R

Fractions	Time (h)	Sp act (mU/mg of protein)a
		α-Glc	β-Glc	α-Gal	β-Gal	β-GlcAu
Intracellular	24	8.64	0.38	0.14	0.28	4.39
	48	8.39	0.01	ND	ND	4.82
	72	9.33	0.04	ND	ND	4.39
Extracellular	24	0.39	0.11	ND	ND	0.32
	48	0.74	0.02	ND	ND	0.46
	72	4.04	0.01	ND	ND	3.12
Cell-bound	24	4.88	6.00	0.06	0.36	3.07
	48	3.11	6.21	0.08	0.27	2.27
	72	2.39	6.57	0.05	0.34	1.76

^aND, not detected; α-Glc, α-d-glucosidase; β-Glc, β-d-glucosidase; α-Gal, α-d-galactosidase; β-Gal, β-d-galactosidase; β-GlcAu, β-d-glucuronidase. 

Enzymes involved in EPS degradation.

The cell extract was tested for its ability to degrade EPS by incubating polysaccharides with a cell extract referred to as crude enzyme. A sample prepared without active enzymes served as a control. Enzyme action on EPS was determined by the increasing release with time of reducing sugars, as monitored by the Nelson-Somogyi method, and the reduction in viscosity of the EPS solution. Activity against the polymer was relatively slight in terms of reducing sugars (Fig. 2). Reducing sugars were released predominantly over the first 7 h of incubation and continued at a slower rate for up to 24 h. This was accompanied by a gradual decrease in apparent viscosity. In another experiment, changes in capillary viscosity were monitored (Fig. 3). The results showed a loss in viscosity. Upon prolonged incubation (27 h), the viscosity was reduced by one-third.

FIG. 2.

Effect of cell extract on liberating of reducing sugars and apparent viscosity of EPS produced by L. rhamnosus R.

FIG. 3.

Effect of cell extract on capillary viscosity of EPS produced by L. rhamnosus R.

Some slight α-l-rhamnosidase activity was also present in bacterial cell extract (data not shown). The two highest activities present in the cell extract are α-glucosidase and β-glucuronidase. Associated with them were β-d-glucosidase, α-d-galactosidase, and β-d-galactosidase activities.

Gel permeation chromatograms showed that EPS produced by L. rhamnosus R at 24 h of fermentation consisted of two main peaks corresponding to high- and low-molecular-weight fractions (Fig. 4). Their relative molecular masses were about 1.4 × 10⁶ and 2.5 × 10⁴ Da, respectively. The large peak corresponding to EPS withdrawn at 24 h of cultivation was eluted at 32.98 min, and the smaller peak representing the low molecular weight was eluted at 41.49 min. At 48 and 72 h of cultivation, the large peak could be observed eluting at 34.66 and 35.03 min, respectively. Thus, the molecular weight of the high-molecular-mass fraction decreased gradually as the cultivation time proceeded from 24 to 48 h and then to 72 h. For EPSs obtained at 48 and 72 h of fermentation, the main peak shifted considerably toward the low-molecular-weight region. This trend confirmed the decrease in viscosity of EPS samples (Fig. 5). The viscosity of EPS present at 24 h was 1,804 ml/g, whereas this value dropped to 1,305 and 967 ml/g for the EPS samples withdrawn at 48 and 72 h of fermentation, respectively.

FIG. 4.

Gel permeation chromatograms of EPS produced by L. rhamnosus R at different cultivation times.

FIG. 5.

Relationship between relative molecular weights and the viscosities of EPS.

Purification of glycohydrolase activities.

To identify the role of each enzyme in EPS degradation, purification and characterization of these enzymes were investigated. The results of a typical purification process are shown in Table 2. During Biopilote Q-Sepharose 35/100 column chromatography, two active enzyme peaks were found (Fig. 6). The major part of the α-glucosidase activity eluted as a small peak (named α-glucosidase) in the wash with the starting buffer, while some trace of α-glucosidase was eluted together with the predominant β-glucuronidase and traces of β-glucosidase as a large peak between 0.45 and 0.64 M NaCl (named fraction beta). The two active fractions (Fig. 6, fractions 8 to 20 and fractions 35 to 44) were pooled, concentrated by using Centriprep 10 (Amicon), and subjected to Hi-Load Superdex 75 gel filtration. For the main α-glucosidase fraction, this step led to obtaining an enzyme found to be homogeneous by SDS-PAGE (Fig. 7). The α-glucosidase was purified 17.6-fold with a 0.46% retention of total activity. The specific activity of the purified enzyme was 125.04 mU/mg of protein (Table 2). The second active fraction showed a partial purification, since a number of bands were still detected by SDS-PAGE (data not shown). β-Glucuronidase was the predominant activity in this fraction. Other activities such as α- and β-glucosidases were also detected.

TABLE 2.

Purification of L. rhamnosus R glycohydrolases

Purification step	Enzymea	Total vol (ml)	Total protein (mg)	Total activity (mU)	Sp act (mU/mg)	Purification (fold)	Activity yield (%)	Protein yield (%)
Crude extract	α-Glc	403	700	4,965.1	7.09	1	100	100
	β-GlcAu			3,758	5.37	1	100
(NH4)2SO4	α-Glc	48	229	1,767	7.72	1.1	35.5	32.92
	β-GlcAu			1,234	5.39	1.01	32.8
Biopilote	α-Glc	1.3	2.51	97.36	38.79	5.5	1.96	3.43
	β-GlcAu	1.6	21.48	143.74	6.69	1.24	3.8
Superdex 75	α-Glc	4.5	0.19	23.13	125.04	17.6	0.46	0.36
	β-GlcAu	2.5	2.31	30.88	13.34	2.5	0.82

^aAbbreviations are as defined in Table 1, footnote a. 

FIG. 6.

Ion-exchange chromatography of cell extract on a FPLC Biopilote Q-Sepharose 35/100 column.

FIG. 7.

Electrophoresis profile of purified α-glucosidase produced by L. rhamnosus R.

Characterization of glycohydrolase enzymes.

Both fractions (purified α-glucosidase and fraction beta) were characterized. The molecular mass of α-glucosidase was about 41 kDa (Fig. 7). The activity of the enzyme toward p-nitrophenyl-β-d-glucoside (PNPG) was determined at a pH of 3.6 to 8.0 at 37°C. The maximum activity was found at pH 5.0 (Fig. 8A). The enzyme was stable at pH 3.6 to 7.0 for 30 min at 37°C, with 95% activity remaining at pH 3.6 and pH 4.4 and 41.55% activity remaining at pH 8.0. As shown in Fig. 8B, α-glucosidase was most active at 50°C. The purified enzyme in 100 mM sodium acetate buffer (pH 5.0) retained about 60% activity at 50°C for 30 min. Rapid inactivation occurred at 60°C, and the enzyme lost most of its activity after 30 min.

FIG. 8.

Effect of temperature and pH on purified α-glucosidase and partially purified β-glucuronidase.

As the most predominant activity in fraction beta was β-glucuronidase, the characterization was carried out with this activity. The optimum pH for the activity of β-glucuronidase was approximately 4.4 (Fig. 8A). The pH stability of this fraction was tested by incubation at various pH values for 30 min at 37°C. The enzyme was stable at all pH values tested. The optimum temperature for activity was 60°C (Fig. 8B). The temperature stability of the enzyme was tested by incubation of the enzyme at pH 4.4 for 30 min at various temperatures. The enzyme was found to be stable at up to 60°C; it retained 92% of its activity at 70°C and 60.97% of its activity at 80°C.

Effect of metal ions and chemical reagents.

For α-glucosidase, the activity was completely inhibited by Hg²⁺, Mn²⁺, Cu²⁺, and Fe²⁺ and slightly inhibited by Zn²⁺ (Table 3). The enzyme was slightly activated by Li⁺, Na⁺, K⁺, Ca²⁺, Co²⁺, and Mg²⁺. β-Glucuronidase was completely inhibited by Mn²⁺, and it retained 19% of its activity with ion Hg²⁺, 60% of its activity with Fe²⁺, and 72% of its activity with Cu²⁺. The other metal ions had no or little effect on this activity. Furthermore, the α-glucosidase was inhibited by urea and DTT, and showed 95% of activity when incubated with 2-mercaptoethanol. β-Glucuronidase was slightly activated by 2-mercaptoethanol and not affected by urea at all. This enzyme was moderately inhibited by DTT. These two activities were both moderately inhibited by EDTA (a chelating reagent).

TABLE 3.

Effect of cations and reducing agents on purified α-glucosidase and partially purified β-glucuronidase from L. rhamnosus R

Cation or chemical agent (final concn [mM])	Relative activity (%)
	α-Glucosidase	β-Glucuronidase
None	100	100
Cations
Na+ (1)	106	99
K+ (1)	112	104
Li+ (1)	107	92
Ca2+ (2)	103	100
Mg2+ (2)	104	104
Mn2+ (2)	0	0
Zn2+ (2)	88	88
Fe2+ (2)	0	60
Co2+ (2)	110	105
Cu2+ (2)	0	72
Hg2+ (2)	0	19
Reducing agents
DTT (2)	74	87
2-Mercaptoethanol (2)	95	104
Others
EDTA (2)	71	71
Urea (2)	61	99

Kinetic parameters.

Michaelis constant K_m and maximum velocity V_max were measured by using PNPG as a substrate. The V_max value for α-glucosidase (ca. 1,524 mU/mg of protein) was higher than that for β-glucuronidase (225 mU/mg of protein), while the K_m value of α-glucosidase (1.829 mM) was lower than that of β-glucuronidase (2.483 mM).

DISCUSSION

This study reports the first demonstration that L. rhamnosus R produces EPS when cultivated on BMM supplemented with either glucose or lactose. When grown in batch cultures on lactose, the amount of EPS was comparable to that produced by glucose-grown cultures. Zourari et al. (35) suggested that for L. delbrueckii subsp. bulgaricus, lactose uptake takes place via a lactose permease. Inside the cell, β-galactosidase cleaves lactose to form glucose and galactose. The latter is exchanged with lactose via a lactose-galactose antiport system. Our results differ from those of Cerning et al. (3), who showed that, for L. casei CG11, both the yield and the composition of the EPS produced were dependent on the carbon source present in the medium. For L. casei CG11, glucose was the most efficient carbon source for EPS production, whereas lactose was an inefficient carbon source.

EPS production by L. rhamnosus R is partially growth associated. During the early exponential phase of growth, there was no EPS biosynthesis. For the two carbon substrates, the production of EPS started only at the end of the exponential phase. This synthesis continued beyond the decline growth phase, as shown in the case of glucose-grown cells, whereas no further polymer production was found after growth had ceased in the medium supplemented with lactose. Polysaccharides should therefore be considered as minor products diverted away from glycolysis rather than as secondary metabolites. It was previously found in other EPS-producing LAB that EPS biosynthesis is growth associated (9, 14, 18, 29, 34). The growth-associated biosynthesis of EPS from S. thermophilus LY03 (9) is supported by the need for an equilibrated carbon/nitrogen ratio and by a direct relationship between optimal growth conditions (temperature, pH, and oxygen tension) and EPS yields. However, both growth-associated and nongrowth-associated production kinetics were observed by Manca de Nadra et al. (25) and Kojic et al. (22). Gassem et al. (17) also found that there was no association between growth rate or acid production and polysaccharide production in different media by LAB (strains CH15, YB57, and YB58 and S. salivarius subsp. thermophilus ST3). Recently, Looijesteijn et al. (23) indicated an uncoupling of growth and EPS production when the production of EPS by L. lactis subsp. cremoris NIZO B40 was investigated. According to them, a possible explanation for this uncoupling is the fact that optimal conditions for EPS production and growth are not the same. It is clear that work will be needed to elucidate the effect of growth conditions and different substrates on the amount and sugar composition of EPS produced by L. rhamnosus R.

The degradation of EPS produced by LAB was observed in a number of studies (4, 5, 6, 13, 15, 16). This trend is not rare with regard to some other microbial EPS such as gellan (21) or sphingan (20). In these cases, the reduced EPS yield is due to the presence of hydrolytic or eliminase enzymes. The majority of EPS-degrading enzymes act through hydrolytic cleavage of the polymers but some enzymes are polysaccharide lyases which act through a β-eliminative mechanism (32). No information about the mechanism of EPS degradation by LAB has been reported. Our results clearly show that the amount of EPS produced by L. rhamnosus R declined upon prolonged fermentation. This reduction is more pronounced in the case of lactose-grown cells than in glucose-grown cells. De Vuyst et al. (9) found that the EPS yield decreased after 12 h of fermentation by S. thermophilus, possibly due to enzymatic degradation. Additionally, these authors found that fermentation temperature and pH influenced EPS degradation. EPS degradation was less pronounced at higher fermentation temperatures and was drastically pronounced at pH 4.9. Gassem et al. (17) observed a reduction of viscosity after 18 h of fermentation for strains CH15, YB57, and YB58 and for S. salivarius subsp. thermophilus ST3. Similar results were obtained by Macura and Townsley (24). They found a decrease in viscosity after 24 h of cell growth. Cerning et al. (6) observed similar results with S. salivarius subsp. thermophilus and suggested that this degradation is due to an enzyme, possibly a glucohydrolase, which progressively destroys the polymer. Polymer degradation has also been reported by other investigators (28, 29) in strains of L. casei and Propionibacterium acidi-propionici. Our results clearly show the presence of different glycohydrolases in cell extracts. Some activities were also detected in extracellular fractions. Moreover, these activities increased in aged cultures (Table 1). This may be due to cell lysis, as indicated in Fig. 9, which shows that L. rhamnosus R possesses lytic activity against M. luteus cells on agar plates. Cell extract was also screened for bacteriolytic activity with M. luteus cells as substrate incorporated in SDS-PAGE. One hydrolysis band is seen on the SDS-PAGE gel (data not shown). This experiment was also carried out with L. rhamnosus ATCC 9595M; this strain produced up to 1,200 mg of EPS per liter (13), with a stable fermentation profile (the degradation was mostly absent in this case). Interestingly, no lytic activity against M. luteus was detected in this strain. This may suggest a lack of glycohydrolases liberated by lysis, therefore leading to an absence of polymer degradation compared to L. rhamnosus R. Further investigation on the eventual linkage between lysis and glycohydrolase activities is now in progress.

FIG. 9.

Lytic zones surrounding L. rhamnosus R grown on MRS agar containing lyophilized M. luteus cells (0.2%).

Although the role of glycohydrolases in EPS degradation is still not well defined, we have shown a possible relationship between these activities and the decrease in viscosity of the EPS solution (Fig. 2 and 3). Glycohydrolases seem capable of lowering EPS viscosity. The slow rate of reduction in viscosity suggests that the mode of enzymatic action does not involve endo-type mechanisms. The reduction of EPS viscosity in the presence of these enzymes suggests a cleavage of polymer molecular mass. The enzymes should act in an exo-fashion, successively splitting glycosidic linkages giving a polymer with lower molecular weight. These in vitro results seem to be in accord with the molecular weight decrease of EPS withdrawn at different times of fermentation as established from the gel permeation chromatograms (Fig. 4).

EPS degradation usually involves a complex of different enzymes (32). Ross et al. (30) showed that a kiwi fruit preparation of β-galactosidase was implicated in the loosening of the cell wall. This β-galactosidase was able to degrade a number of well-characterized, natural, fruit cell wall polysaccharides. The enzyme acts in an exo-fashion, cleaving monomers from the nonreducing termini of β-linked galactose chains. In Bacillus sp. strain GL1 cells (19), gellan is, at first, converted to tetrasaccharide by extracellular gellan lyase and then hydrolyzed to monosaccharides by the intracellular exoglycosidases, α-rhamnosidase and β-glucosidase. α-l-Rhamnosidase of Sphingomonas sp. strain R1 (20) in the mixed culture degrading gellan may be responsible for the degradation of the rhamnosyl-glucose released from gellan. Sutherland and Kennedy (33) reported that a number of Sphingomonas strains capable of synthesizing the bacterial EPS gellan possessed constitutive gellanase activity. These bacteria also produced intracellular glycosidases. In a previous study (1), Berg et al. reported that the vast spectrum of glycohydrolases produced by B. fragilis might be linked to its ability to hydrolyze and ferment polysaccharides from various sources.

The relative molecular weights of EPS produced by L. rhamnosus R decreased as a function of the fermentation time, suggesting that degradation of the polymer had occurred in older cultures. Similar results have been obtained by Conti et al. (7), who found that Pseudomonas fluorescens and P. putida produced polysaccharides for which the maximum M_r was obtained at 48 h of growth, decreasing with increasing fermentation time. This was ascribed to the degrading activity of alginate lyases which were detected intracellularly in both species and were presumably released by cell lysis.

Our study is the first report on the purification and characterization of glycohydrolases from L. rhamnosus R. The glycoside hydrolases appeared at the end of the logarithmic growth phase and in the decline phase of L. rhamnosus R cultures, mainly in association with the cells but also to a small extent in the culture supernatant. Two active glycohydrolase peaks were found during ion-exchange chromatography. The minor peak contained α-glucosidase, whereas the major peak contained β-glucuronidase and traces of α- and β-glucosidase activities. The α-glucosidase activity from L. rhamnosus R was acidic. The maximal activity of purified α-glucosidase was observed at pH 5.0 and 50°C. The temperature profile of α-glucosidase was sharper compared with the broader range for β-glucuronidase. α-Glucosidase from Brettanomyes lambicus (31) exhibited optimum activity at 39°C and pH 6.2. α-Glucosidase from L. rhamnosus R was strongly inhibited by the sulfhydryl oxidant metals. This result suggested the presence of thiol groups at the catalytic site. Similar findings were reported by Shantha Kumara et al. (31). When working with B. lambicus, these authors found a total loss in the enzyme activity at 5 mM HgCl₂. Studies on L. brevis (10) revealed a total inactivation of the enzyme at 5 mM HgCl₂. Univalent-ion activation in L. rhamnosus R α-glucosidase might be due to competition of these ions with H⁺ for some catalytically important prototropic groups which are more active when complexed with the metal ions (2). The addition of urea, DTT, and 2-mercaptoethanol affected the α-glucosidase activity, indicating that sulfhydryl groups are involved in the active site of the enzyme. The two activities were moderately inhibited by the presence of EDTA, suggesting that metal cations are required for these enzymes.

The complete elucidation of the degradation mechanism requires further work. In fact, studying the accessibility of EPS to glycohydrolases as well as identifying the degraded products would provide valuable information on the possible mechanism of EPS breakdown. Moreover, it would be interesting to investigate whether the glycohydrolases are inducible. If this is the case, it would help explain the difference in the degree of degradation when bacteria are grown with either glucose or lactose as a carbon source.