Protection of Bi dobacteria Encapsulated in Polysaccharide-Protein Gel Beads against Gastric Juice and Bile

Transcription

1 2076 Journal of Food Protection, Vol. 66, No. 11, 2003, Pages Copyright q, International Association for Food Protection Protection of Bi dobacteria Encapsulated in Polysaccharide-Protein Gel Beads against Gastric Juice and Bile DANIEL GUÉRIN, JEAN-CHRISTOPHE VUILLEMARD,* AND MURIEL SUBIRADE Centre de Recherche en Sciences et Technologie du Lait STELA, Faculté des Sciences de l Agriculture et de l Alimentation, Université Laval, Québec, Canada G1K 7P4 MS : Received 24 October 2002/Accepted 4 April 2003 ABSTRACT Bi dobacterium cells were encapsulated in a mixed gel composed of alginate, pectin, and whey proteins. Two kinds of capsules were obtained: gel beads without membranes and gel beads with two membranes formed by the transacylationreaction. In vitro studies were carried out to determine the effects of simulated gastric ph and bile salts on the survival of free and encapsulated Bi dobacterium bi dum. The protective effects of gel beads without membranes and gel beads coated with two membranes formed by the transacylation reaction were evaluated. After 1 h in an acidic solution (ph 2.5), the free-cell counts decreased by 4.75 log units, compared with a,1-log decrease for entrapped cells. The free cells did not survive after 2 h of incubation at ph 2.5, while immobilized-cell counts decreased by about 2 log units. After incubation (1 or 3 h) in 2 and 4% bile salt solutions, the bi dobacterium mortality level for membrane-free gel beads (4 to 7 log units) was higher than that for free cells (2 to 3 log units). However, counts of bi dobacteria immobilized in membrane-coated gel beads decreased by,2 log units. Cell encapsulation in membrane-coated protein-polysaccharide gel beads could be used to increase the survival of healthy probiotic bacteria during their transit through the gastrointestinal tract. The discovery of bi dobacteria dates back to the beginning of the 1900s. Indeed, Henry Tissier of the Pasteur laboratory discovered bi dobacteria in infant feces and named them Bacillus bi dus communis (20). These bacteria were found to be the predominant microorganisms in the intestinal ora of breast-fed infants (20, 38). However, the bi dobacterium distribution changes dramatically in human intestines as humans get older: bi dobacteria account for.99% of the microorganisms in the infant intestinal micro ora, but they are the fourth most prominent type of bacteria in the adult gastrointestinal ora, accounting for,15% of the microorganisms present (45). Many studies have shown that bi dobacteria play an important and bene cial role in the maintenance of human health (3, 20, 27, 45). However, to be bene cial to their host, bi dobacteria must reach the colon at a concentration on the order of 10 7 living bacteria per g of intestinal content (3) or on the order of 10 6 bacteria per ml of chyme at the terminal ileum (33). Therefore, bi dobacteria must survive in large numbers in the food to which they are added, tolerate the extremely low ph of the human stomach, subsist at different bile concentrations in the small intestine, and, nally, reach and colonize the colon in suf cient numbers (8). The survival rate for bi dobacteria as they travel through the human digestive system varies greatly depending on the strain. Indeed, Bi dobacterium bi dum 1900, B. bi dum 1901, B. infantis 1912, B. adolescentis * Author for correspondence. Tel: ; Fax: ; jean-christophe.vuillemard@aln.ulaval.ca. 1920, B. breve 1930, and B. thermophilum display broad and varied log changes after 3 h of incubation at 378C at ph 2.5. Viable counts of these bacteria were found to decrease by 6 to 9.5 log units from an original population level of 10 8 to CFU/g, while for B. longum 1941 and B. pseudolongum 20099, the decline from an initial population level of CFU/g ranged from 6 to 2.5 log units (26). In one study, populations of B. infantis, B. bi dum, and B. adolescentis, which ranged from to CFU/ml before treatment, did not survive after incubation for 12 h at 378C in a 2% bile solution, while the level of B. longum went from to CFU/ ml with the same treatment (8). More resistant and tolerant strains were used in other studies, but such strains did not always produce viability data that were compatible with the expression of probiotic effects (7, 8, 26). Moreover, sensitive strains are more numerous than resistant strains. To minimize the problem associated with the low survival rate for bi dobacteria as they travel through the human digestive system, microencapsulation could be used to protect bi dobacteria and therefore increase their survival (6, 28, 45, 54). In fact, after incubation for 3 h at 378C at ph 1.55, free B. longum KCTC 3128 and B. longum HLC 3742 did not survive for.45 min, while the population of bacteria immobilized in an alginate gel declined by 2 to 3 log units from an initial level of CFU per bead (28). Thus, the aim of the present study was to develop an ef cient method of protecting bi dobacteria against the low ph of the stomach and the bile concentrations in the small intestine.

2 PROTECTION OF NUTRACEUTICAL BACTERIA 2077 MATERIALS AND METHODS Materials. Low-viscosity sodium alginate, high apple methoxyl pectin, and porcine pepsin were obtained from Sigma-Aldrich (Oakville, Ontario, Canada). Bipro (whey protein isolate containing 95% protein) was obtained from Davisco Food International Inc. (Le Sueur, Minn.). Freeze-dried B. bi dum RO71 (10 9 to CFU/g), a strain sensitive to bile salts and low stomach ph, was purchased from Rosell Institute (Montreal, Quebec, Canada). Aseptic conditions were used for all of the procedures outlined below. Preparation of the immobilization matrix. All glassware and solutions used in the protocols were sterilized at 1218C for 15 min. Sodium alginate, pectin, and whey proteins in powder form were sterilized at 1218C for 15 min and then maintained in the autoclave without steam for an additional drying period of 15 min. The immobilization procedure, as outlined by Levy and Edwards-Levy (29), was modi ed and adapted to make it more suitable for a bacterial preparation. The initial aqueous phase was prepared by dissolving 3.25 g of whey proteins in 20 ml of distilled and deionized water. The second aqueous phase was prepared by dissolving 1.0 g of alginate and 1.5 g of pectin in 25 ml of a saline solution (1.8% [wt/vol] NaCl), and the ph was adjusted to 7.0 with 5 M NaOH. The protein solution was then poured into the polysaccharide solution maintained under agitation. Preparation of the gelling solution. An 8% (wt/vol) whey protein solution (sterilized at 1218C for 15 min) was added to a 3.4% (wt/vol) CaCl 2 solution (0.22-mm- lter sterilized) to yield a nal mixture of 1.7% calcium chloride and 4% protein. This solution was used to produce the instant gelation of the immobilizing matrix. Rehydration of freeze-dried bi dobacteria. One gram of freeze-dried bacteria was rehydrated for 10 min at 258C in 4 ml of a solution composed of 0.5% (wt/vol) meat extract (Lab-Lemco Oxoid, Basingstoke, Hampshire, UK), 1.5% (wt/vol) peptone (Difco Laboratories, Detroit, Mich.), and 1.0% (wt/vol) tryptone (Difco). These rehydration conditions were found to lead to high recovery rates (5, 14). The rehydrated bacterial suspension was then added to 45 ml of the polysaccharide-protein solution to yield nal concentrations of 2% (wt/vol) alginate, 3% (wt/vol) pectin, 6.25% (wt/vol) whey proteins, and 0.9% (wt/vol) NaCl. Formation of gel beads. Forty milliliters of the resulting solution was added dropwise to 400 ml of the gelling solution with a syringe equipped with a 20-gauge needle. Magnetic stirring was carried out for 30 min. The resulting beads were rinsed twice with distilled and deionized water. Afterward, the beads were split into two equal quantities: the rst batch was resuspended in 300 ml of a 0.01 M CaCl 2 solution at ph 7 and kept refrigerated overnight, while the second batch was resuspended in 200 ml of saline solution (0.9% [wt/vol] NaCl) with stirring for the further coating step. Formation of membrane-coated gel beads. The membranes were formed by a transacylation reaction. This reaction is based on the interaction between pectin (an esteri ed polysaccharide with.50% esteri ed carboxylic groups) and a protein (12, 29, 36, 37, 40 42). Upon alkalinization, the formation of amide bonds between the carboxylic groups of the polysaccharide and the amino groups of the protein occurs (Fig. 1). The transacylation reaction was started by adding 1,000 ml of 1 M NaOH to the bead suspension (ca. 20 g) in saline solution. The suspension was stirred for 10 min and was then neutralized with 500 ml of 1 M FIGURE 1. Transacylation reaction between pectin and proteins. HCl while magnetic stirring was maintained for 10 min. The resulting membrane-coated beads were collected by ltration, resuspended in 300 ml of a 0.01 M CaCl 2 solution (ph 7.0), and kept refrigerated overnight for further use. Examination of beads. Membrane-coated beads were sliced with a razor blade and resuspended in a 10% (wt/vol) sodium citrate solution for 10 min. Photographs of entire beads were also taken to evaluate their different shapes. A bead s mean diameter was measured with a scale rule. Optic microscopy examination. Optic microscopy was used to evaluate the thickness of the external and internal bead membranes. Membrane-coated beads were sliced with a razor blade and placed in a 10% sodium citrate solution for 10 min without agitation. The microscope used was a Leica LaborLux S equipped with a Sony CCD-Iris/RGB color video camera controlled by a computer (Inspector 1.71 software). Scanning electron microscopy examination. Fresh membrane-coated and membrane-free beads were kept overnight at 48C in a xative solution containing 0.01 M CaCl 2, 2.5% (wt/vol) glutaraldehyde, and 1.5% (wt/vol) paraformaldehyde (ph 7). Beads were then rinsed three times with a 0.01 M CaCl 2 solution for 20 min and post xed in a 1% (wt/vol) osmium tetroxide solution for 2 h at room temperature. After three other washes with 0.01 M CaCl 2 had been carried out, the beads were dehydrated in a succession of solutions with increasing ethanol concentrations (up to 100%) and dried by the CO 2 critical-point drying technique (19). Finally, half and whole beads were xed on a bracket and recovered with a gold layer. The external and internal appearances of beads were observed with the JEOL scanning electron microscope (Model JSM-35CF, JEOL, Tokyo, Japan) at an accelerating voltage of 15.0 kv. Survival of immobilized and free bacteria under gastric conditions. Simulated gastric juice (ph 1.5) was prepared with a buffer mixture composed of 0.2 M HCl solution and 0.2 M KCl solution, to which pepsin was added at a ratio of 10 U/ml (46). Another buffer composed of a 0.2 M HCl solution and a 0.2 M glycin solution was used to simulate gastric juice at ph 2.5 (containing 10 U of pepsin per ml). Samples of fresh beads with and without membranes and samples of rehydrated free bacteria were added to test tubes containing prewarmed (378C) simulated gastric juices at phs of 1.5 and 2.5 (26) and incubated at 378C for 60 and 120 min (45) while orbital agitation at 100 rpm in a controlled environment incubator shaker was maintained. A 1-g bead sample was weighed in a culture tube containing 1 ml of distilled water and 200 ml of a 3.4% (wt/vol) CaCl 2 solution (to prevent bead disruption during simulated digestion tests), and then 7.8 ml of simulated gastric juice was poured into the culture tube containing the beads. For free cells, a 100-ml aliquot was placed in a culture tube, and then

3 2078 GUÉRIN ET AL. CaCl 2 solution. After 1 and 3 h of incubation, 100-ml aliquots were removed, serially diluted, and immediately assayed for bi- dobacterium survival. Peptone water (ph 7.0) was used as a control. The experiment was repeated four times in duplicate for free and immobilized cells as well as for controls (Fig. 2). FIGURE 2. Protocol used for the simulated digestive tests. 9.7 ml of simulated gastric juice and 200 ml of a 3.4% (wt/vol) CaCl 2 solution were added in the test tube. After incubation, the beads were removed by ltration and placed in 4 ml of a 0.2 M sodium phosphate buffer (ph 7.0). For free cells, 100-ml aliquots were removed and immediately assayed for B. bi dum survival. Free and immobilized bacteria subjected to the same treatments in peptone water (ph 7.0) instead of the simulated gastric juice were used as controls. For the determination of viable cell counts, the immobilized bacteria were released from the capsules by disintegration with a high-speed homogenizer (Ultra-Turrax, Model T25-S1, Janke & Kunkel Ika-Labortechnik, Staufen, Germany). This treatment had previously been tested and had been shown not to affect the viability of bi dobacteria (18). One gram of beads was placed in a tube containing 4 ml of 0.2 M sodium phosphate buffer (ph 7.0) and kept in an ice bath for 45 s at 20,000 rpm while the sample disintegrated. Bi dobacterium counts were determined on Reinforced Clostridial Medium agar (AES Laboratoire, Montréal, Canada) by the pour plate technique with incubation under anaerobic conditions at 378C for 96 h. The experiment was repeated four times in duplicate for free and immobilized bi dobacteria and for the controls (Fig. 2). Survival of immobilized and free bacteria in simulated high-bile conditions. A solution of Oxgall (Difco) was prepared by adding 10 g of dry powder base to 90 ml of distilled water (8). This solution was then used to prepare 2.0% (wt/wt) and 4.0% (wt/wt) concentrations of bile. The solutions were adjusted to ph 7.5 with 1 M NaOH and lter sterilized with a 0.22-mm lter. To assay the survival of immobilized and free B. bi dum cells under small intestine conditions without pancreatin, samples of fresh beads or samples of rehydrated free bacteria were added to test tubes containing prewarmed (378C) simulated bile at ph 7.5 and incubated at 378C for 60 and 180 min (26) while orbital agitation at 100 rpm in a controlled-environment incubator shaker was maintained. For immobilized bacteria, 1 g of beads was placed into a culture tube containing 1 ml of distilled water plus 200 ml of a 3.4% (wt/vol) CaCl 2 solution. Afterward, 7.8 ml of simulated bile juice was poured into the culture tube containing the beads. The beads were removed by ltration after their respective 1 and 3 h of incubation and placed in 4 ml of a 0.2 M sodium phosphate buffer (ph 7.0). Bead disruption and viable cell enumeration were carried out as previously described. For free cells, 100 ml of rehydrated cells was placed into a culture tube with 9.7 ml of simulated bile juice and 200 ml of a 3.4% (wt/vol) Diffusion evaluation. The diffusion of simulated bile through beads without membranes was compared with its diffusion through beads with membranes. To evaluate the amount of bile absorbed by the beads, a simple method based on the scienti c principles of photometry (25) was used. Samples of beads were subjected to the same conditions used for bile testing. After incubation, beads were removed by ltration and placed in 4 ml of a buffer mixture (ph 7.0) composed of 0.2 M imidazole solution, 0.2 M HCl solution, and 1 drop of antifoam per 100 ml of buffer mixture. The beads were kept in an ice bath (Ultra- Turrax) at 20,000 rpm for 1 min while they disintegrated. After 20 min of centrifugation at 25,000 3 g, the supernatant was l- tered with a 0.45-mm lter. Absorbance was measured at 421 nm with a Hewlett Packard spectrophotometer (Model 8453). Absorbance results were then expressed as the concentration of bile inside the beads with a standard curve. The standard curve was based on known amounts of bile reported as a function of absorbance at 421 nm. This experiment was conducted in triplicate. Statistical analysis and enumeration. An analysis of variance (ANOVA) with PROC ANOVA of the Statistical Analysis System (SAS Institute Inc., Cary, N.C.) was used to determine signi cant differences (P, 0.05 and P, 0.01). For simulated gastric acidity or bile conditions, a (free cells, capsules with and without membranes 3 three treatments 3 time) factorial model with complete blocks was used. The experiment was replicated four times. For the diffusion test, a (capsules with and without membranes 3 two bile concentrations 3 time) factorial model with complete blocks was used. The experiment was carried out in triplicate. An ANOVA was carried out with PROC ANOVA to determine whether the level of bile diffusion was signi cantly lower for one kind of beads than for the other. RESULTS AND DISCUSSION Immobilization in the mixed gel matrix. The mixed gel matrix was optimized by studying many concentrations of alginate, pectin, or propylene glycol alginate (high esteri ed polysaccharides) and whey protein isolate. The procedure used to sterilize the polysaccharides and the proteins had an effect on their functional properties (2, 4, 10, 15, 39, 48, 51, 52), which had an impact on membrane formation around the beads. Therefore, in preliminary experiments, many combinations did not produce a membrane, while other formulations produced single membranes or double membranes. The presence or absence of a single or double membrane depends either on the types and concentrations of polysaccharides (including the esteri ed polysaccharide) and whey proteins, the form of the sterilization (powder form or liquid solution) used with the matrix components, the time of the transacylation reaction, the amount of 1 M NaOH solution used to start the reaction, or a combination of these factors. Salt composition of the matrix and the gelation solutions. As outlined in the Materials and Methods section, the matrix was prepared with a 0.9% (wt/vol) NaCl

4 PROTECTION OF NUTRACEUTICAL BACTERIA 2079 FIGURE 3. Halves of beads with double membranes. Dark arrows indicate the outer membrane, and clear arrows indicate the inner membrane. saline solution. The presence of NaCl in the mixed gel mixture resulted in a competition between Ca2 1 and Na1 ions during the gelation step; this competition slowed down the speed of gelation but led to a more homogenous gel (12, 39, 44, 50). The gelation procedure was carried out with a protein solution containing CaCl2 in order to obtain instant gelation of alginate upon contact with divalent cations and to put an extra protein layer around the beads before the membrane formation process took place, thus enhancing and favoring membrane formation. Membrane formation occurred after the transacylation reaction step. Finally, the two types of beads obtained (i.e., beads with and without membranes) were kept overnight in a 0.01 M CaCl2 solution to obtain a complete and stable gel network (1, 50). Bead characterization: membrane formation. When membrane formation was carried out with a 0.9% NaCl solution to favor the transacylation reaction (12), two distinct membranes were obtained (Fig. 3). Their stability in a 10% sodium citrate solution while the core of the beads was relique ed in this solution (Fig. 3) con rms the presence of covalent bonds. The mean diameter of the beads was 2 to 3 mm (Fig. 3). The rst (outer) membrane was obtained through the transacylation reaction itself, but the second (inner) membrane was formed during the neutralization step for the coating reaction. There are two possible explanations for the formation of the second membrane. First, the ph rapidly decreases during neutralization, which stops the transacylation reaction and separates the outer membrane from the inner membrane as it begins to form. At that time, a ph gradient between the periphery of the beads (outer membrane), which is acidic, and the area near the inner membrane, which is still basic, is brought about, and the formation of a second membrane continues. During the formation of the inner membrane, the ph gradient continues to decrease and nally reaches 5.6 to 5.7, which stops the formation of the second membrane. The formation of the second membrane can also be explained by the isoelectric point of the whey protein. During the neutralization step, the ph decreases to 3.6, which is lower than the isoelectric point of whey proteins, result- FIGURE 4. Bead membranes produced with 1.2% (vol/vol) 1 M NaOH (A) and 1.0% (vol/vol) 1 M NaOH (B). Original magni cation, 340. ing in the gelation of the heat-denatured whey proteins (4, 13, 21, 31, 43, 51). A combination of these two mechanisms could also be involved in the formation of the second membrane during the neutralization step. Membrane thickness. The thicknesses of the inner and the outer membranes is dependent on the NaOH concentration in the saline solution during the transacylation step. Indeed, for a constant reaction time, the penetration of NaOH into the beads increases as NaOH concentrations increase (29). As shown in Figure 4, the thicknesses of the inner and outer membranes are 191 and 433 mm, respectively in a 1.2% (vol/vol) NaOH solution. However, in a 1.0% (vol/vol) NaOH solution, the thicknesses of the inner and outer membranes were 179 and 283 mm, respectively. To minimize the cell death associated with the transacylation step, the beads in the 1.0% (vol/vol) NaOH solution were selected for the in vitro digestion tests. Surface porosity. Figure 5 shows that alginate beads without membrane coatings appeared to be more porous than membrane-coated beads. The beads without membranes had pores with a broad range of sizes, with widths of up to ca. 22 mm and heights of nearly 4 mm, while the range of pore sizes for beads with two membranes was not as wide, with widths of up to ca. 3.5 mm and heights of ca. 6 mm for the largest pores to widths of ca. 1 mm and heights of ca. 1 mm for the smallest pores. Thus, the transacylation reaction enabled the formation of a very tight

5 2080 GUE RIN ET AL. FIGURE 5. Surfaces of beads with membranes (A) and without membranes (B). Arrows indicate pores in the surfaces of the beads. Original magni cation, 32,000. coating network compared with that obtained for the membrane-free beads. Inner structure. The cores of the two types of beads had similar inner gel structures (Fig. 6). The gel network is homogeneous, and the bi dobacteria are uniformly entrapped within the gel. Curiously, the inner gel structures of the uncoated beads (Fig. 6) appear to be signi cantly tighter than the outer structures (Fig. 5). This difference could be due to the treatment applied to the beads before they were observed under the scanning electron microscope or due to the gelation process, which is faster on the surfaces of the beads (34, 35, 39) and slower in the cores of the beads (1, 50). Immobilized bi dobacteria. Figure 7 shows the presence and disposition of bi dobacteria on the surfaces and in the cores of the two types of beads. Many bi dobacteria are present on the surfaces of the membrane-free beads (Fig. 7C), while only a few bacteria are present on the surfaces of membrane-coated beads (Fig. 7A). Consequently, during their transit through the gastric and intestinal uids, the membrane-free beads will release more bacteria than the membrane-coated beads. On the other hand, large bi dobacterium populations seem to be distributed uniformly throughout the membrane-coated (Fig. 7B) and membrane-free (Fig. 7D) beads. In vitro digestive assays: resistance to stomach conditions. Figure 8 shows the living populations of free cells FIGURE 6. Inner structures of beads with surface membranes (A) and without surface membranes (B). Arrows indicate bi dobacterium clusters. Original magni cation, 32,000. and encapsulated cells for the double-membrane-coated and membrane-free beads after simulations of different human gastric conditions. Figure 8 also shows that the free cells and the cells encapsulated in membrane-coated and membrane-free beads did not survive after 1 and 2 h, respectively, at ph 1.5 (1 and 10 CFU/g were the detection limits for the encapsulated and free cells, respectively). However, after 1 h at ph 2.5, the free-cell counts decreased by 4.75 log units, compared with decreases of 0.76 and 0.10 log units for uncoated and coated beads, respectively. The free cells did not survive after 2 h of incubation at ph 2.5, while the encapsulated cell counts remained relatively high: CFU/g of beads (initial population, CFU/g) for the membrane-free beads and CFU/g of beads (initial population, CFU/g) for the membrane-coated beads. After 1 and 2 h of in vitro digestion at ph 1.5 and ph 2.5, the free-cell count was signi cantly different from the encapsulated-cell count (P, ), but no signi cant differences between counts for the membrane-coated beads and the membrane-free beads were observed (P, 0.22 after 1 h and P, 0.31 after 2 h). The reductions in free-cell counts (4.75 and 9.91 log units after 1 and 2 h, respectively) were similar to the values found in the literature (7, 53). The higher survival rates observed for the encapsulated bacteria after 2 h at ph 2.5 ( and CFU/g of beads for the membrane-free and membrane-coated beads, respectively) could be ex-

6 PROTECTION OF NUTRACEUTICAL BACTERIA 2081 FIGURE 7. Bi dobacteria on the surfaces (A, C) and in the cores (B, D) of beads with (A, B) and without (C, D) membrane coatings. Arrows indicate bi dobacteria. plained by the buffering capacity of whey proteins (24). The alginate gel could also have a protective effect in lowph solutions by limiting the direct contact of cells with these solutions, an effect corroborated by other studies (22, 23, 28, 47, 49). Moreover, it was demonstrated that immobilization in alginate gels enhances the tolerance of cells to toxic compounds by modifying their physiology (17). Indeed, entrapment techniques, such as alginate encapsulation, result in changes to the physicochemical properties of the microenvironment, in uencing cell metabolism. These changes to the microenvironment of the immobilized cell include the introduction of ionic charges, altered water activity, modi ed surface tension, and cell con nement. But the main factor likely in uencing cell behavior is the mass transfer limitation, which results in gradients of oxygen, substrate, and product. Consequently, immobilization may in uence cell physiology and performance via a number of mechanisms that are often poorly characterized and possibly operating in opposite directions (17). Moreover, pectin is a polysaccharide that is known to be resistant in an acidic-ph solution (30). Furthermore, the synergistic effect of these three gel components (alginate, whey proteins, and FIGURE 8. Survival of free cells and immobilized cells in gel beads with and without membranes after treatments simulating stomach conditions. Values shown are means 6 standard deviations (n 5 4). Means with different letters are signi cantly different (P, 0.01).

7 2082 GUE RIN ET AL. FIGURE 9. Survival of freeze-dried free cells, immobilized cells in mixed gel capsules with membranes and immobilized cells in mixed gel capsules without membranes after treatments simulating bile concentrations of 2 and 4% at 378C for 1 and 3 h. Values shown are means 6 standard deviations (n 5 4). Means with different letters are signi cantly different (P, 0.01).. pectin) should also be taken into consideration in explaining the higher survival rate for immobilized bi dobacteria. Resistance to gastrointestinal conditions. Figure 9 illustrates the survival of free and immobilized B. bi dum RO71 after different treatments simulating the human gastrointestinal tract. The free-cell count decreased by 2.76 log units, the immobilized-cell count for membrane-free beads decreased by 6.3 log units, and the immobilized-cell count for membrane-coated beads decreased by 2.16 log units after 1 h of incubation in a 2% bile solution. In a 4% bile solution, the free-cell count decreased by 2.53 log units, the immobilized-cell count for membrane-free beads decreased by 6.91 log units, and the immobilized-cell count for membrane-coated beads decreased by 1.18 log units. The differences between reductions for free and immobilized cells as well as between reductions for membrane-coated and membrane-free immobilized cells were signi cant (P, 0.01). After 3 h of incubation at 378C, the death rate was slightly lower for the free cells and the cells immobilized in the membrane-coated beads, while for the cells immobilized in the membrane-free beads, the loss of viability after 3 h was more extensive than the loss of viability after 1 h. Indeed, free-cell counts decreased by 2.58 log units for a 2% bile solution and by 2.07 log units for a bile concentration of 4% after 3 h, whereas counts of these cells de- creased by 2.76 and 2.53 log units after 1 h in 2 and 4% bile solutions, respectively. Counts of immobilized cells in membrane-free beads decreased by 7.14 and 4.00 log units after 3 h of incubation and by 6.30 and 6.91 log units after 1 h of incubation for 2 and 4% bile concentrations, respectively. Finally, counts of immobilized cells in membranecoated beads decreased by 1.10 and 0.65 log units after 3 h of incubation and by 2.16 log and 1.18 log units after 1 h of incubation for 2 and 4% bile concentrations, respectively. The differences between the reductions in viable populations of free cells and the reductions in viable populations of immobilized cells, as well as those between reductions for the membrane-coated beads and reductions for the membrane-free beads, were statistically signi cant (P, 0.01). The free-cell survival rate was higher in 2 and 4% bile solutions after 3 h of incubation than after 1 h of incubation, which seems surprising. However, this phenomenon is consistent with results obtained in other investigations in which a slight increase in cell populations after 12 h of incubation in 2% bile solution (49) and the resistance of free-cell populations after treatments in bile solutions (11) were observed. This outcome could be explained by the ability of some bi dobacterium strains, such as B. bi dum, to grow in bile solutions, which contain proteins and glucose (9), and their activity against conjugated bile acids (11, TABLE 1. Amounts of bile diffusing inside beads after treatments simulating concentrations of bile in the human small intestinea Bile concn (%) in beads after 1 h Bead type Membrane free Membrane coated a Bile concn (%) in beads after 3 h 2% bile solution 4% bile solution 2% bile solution 4% bile solution A A B B A A Values shown are means 6 standard errors (n 5 3). Means with different letters are signi cantly different (P, 0.01). B B

8 PROTECTION OF NUTRACEUTICAL BACTERIA 2083 ACKNOWLEDGMENT This work was supported by the National Sciences and Engineering Research Council of Canada. REFERENCES FIGURE 10. Saponi cation of pectin under slightly alkaline conditions , 32). The lower survival rate for immobilized bi dobacterium cells in beads without membranes than for free cells led to complementary experiments. A diffusion test was carried out to evaluate the nature of bile diffusion inside the two types of beads (Table 1). The results obtained showed that incubation for 1 or 3 h in a 2 or 4% bile solution did not signi cantly affect the amounts of bile inside the beads (P, 0.12). As described earlier, the transacylation reaction involves the carboxylic group of the pectin to which a methyl group is attached. After the transacylation reaction has taken place, the methyl group is released into the medium and then separated from the membrane-coated beads by ltration. The membrane-free beads still have this methyl group, which contributes to the high mortality rate for immobilized bacteria. Indeed, in the bile solutions (at ph 7.5 and at 378C), the pectin undergoes a saponi cation reaction (52) (Fig. 10) that produces methanol, which has a lethal effect on the immobilized bacteria. Hence, this chemical reaction can explain the lower survival rates for cells in the membrane-free gel beads. This study has demonstrated that the immobilization of bi dobacteria in a mixed gel improves the survival of these bacteria under conditions simulating the human gastrointestinal tract and the resistance of these bacteria to these conditions. Indeed, the reduction in the count of immobilized cells in an acidic solution (ph 2.5) containing pepsin was,2.5 log after 2 h of incubation, while free cells did not survive. The double membrane coating enhanced cell resistance to acidic conditions. As was the case for the bile treatments, the cells in the membrane-coated beads performed better than the cells in the membrane-free beads, which were more sensitive than the free cells. These results indicate that cell encapsulation in double-membrane gel beads could be used to increase the survival of healthy probiotic bacteria during their transit through the gastrointestinal tract. In fact, this matrix could protect the bacteria both in the carrier food and in the human stomach. For instance, an initial immobilized-cell population of 1010 cells per g could reach the intestine at a level on the order of cells per g and hence provide a bene cial health effect (3, 33) Andresen, I. L., and O. Smidsrod Temperature dependence of the elastic properties of alginate gels. Carbohydr. Res. 58: Bondu, Proprie te s fonctionnelles des pectines, p In M. Colonna and M. Thibault (ed.), Proprie te s fonctionnelles des polysaccharides. APRIA: Association pour la Promotion Industrie Agriculture, Paris. Bouhnik, Y Survival and effects in humans of bacteria ingested in cultured milks. Lait 73: Boye, J. I., I. Alli, A. A. Ismail, B. F. Gibbs, and Y. Konishi Factors affecting molecular characteristics of whey protein gelation. Int. Dairy J. 5: Champagne, C. P., F. Mondou, Y. Raymond, and D. Roy Effect of polymers and storage temperature on the stability of freezedried lactic acid bacteria. Food Res. Int. 29: Champagne, C. P., Y. Raymond, F. Mondou, and J. P. Julien Studies on the encapsulation of Bi dobacterium longum cultures by spray-coating or co-crystallization. Bi dobacteria Micro ora 14(1): Clark, P. A., L. N. Cotton, and J. H. Martin Selection of bi dobacteria for use as dietary adjuncts in cultured dairy foods. II. Tolerance to stimulated ph of human stomachs. Cult. Dairy Prod. J. 28(4): Clark, P. A., and J. H. Martin Selection of bi dobacteria for use as dietary adjuncts in cultured dairy foods. III. Tolerance to simulated bile concentrations of human small intestines. Cult. Dairy Prod. J. 29(3):18, Dore, D Biochimie clinique. Griffon d argile, Sainte-Foy, Quebec, Canada. Dumay, E De naturation thermique de la beta-lactoglobuline et proprie te s ge li antes des concentre s prote iques de lactose rum, p In D. Lorient, B. Colas, and M. Le meste (ed.), Proprie te s fonctionnelles des macromole cules alimentaires. Cahiers de l ENSBANA, Paris. Dunne, C., L. O Mahony, L. Murphy, et al In vitro selection criteria for probiotic bacteria of human origin: correlation with in vivo ndings. Am. J. Clin. Nutr. 73(Suppl. 2):386S 392S. Edwards-Levy, F. M., and M. C. Levy Serum albumin-alginate coated beads: mechanical properties and stability. Biomaterials 20: Elofsson, C., P. Dejmek, M. Paulsson, and H. Burling Characterization of a cold-gelling whey protein concentrate. Int. Dairy J. 7: Font de Valdez, G., G. S. De Giori, A. Pesce de Ruiz Holgado, and G. Oliver Rehydration conditions and viability of freeze-dried lactic acid bacteria. Cryobiology 22: Fraser, J. E., and G. F. Bickerstaff Entrapment in calcium alginate, p In G. F. Bickerstaff (ed.), Methods in biotechnology. Humana Press, Totowa, N.J. Grill, J. P., S. Perrin, and F. Schneider Bile salt toxicity to some bi dobacteria strains: role of conjugated bile salt hydrolase and ph. Can. J. Microbiol. 46: Groboillot, A. F., D. K. Boadi, D. Poncelet, and R. J. Neufeld Immobilization of cells for application in the food industry. Crit. Rev. Biotechnol. 14: Gue rin, D Unpublished data. Hayat, M. A Principles and techniques of scanning electron microscopy. Biological applications, vol. 3. Van Nostrand Reinhold, New York. Hughes, D. B., and D. G. Hoover Bi dobacteria: their potential for use in American dairy products. Food Technol. 45:74, 76, 78 80, 83. Ju, Z. Y., and A. Kilara Properties of gels induced by heat, protease, calcium salt, and acidulant from calcium-ion aggregated whey protein isolate. J. Dairy Sci. 81:

9 2084 GUE RIN ET AL. 22. Kailasapathy, K., K. Sultana, G. Godward, et al Preliminary studies on enhanced viability and improved delivery of probiotic bacteria by encapsulation. BRG Workshop Trondhein 8:O Khalil, A. H., and E. H. Mansour Alginate encapsulated bi dobacteria survival in mayonnaise. J. Food Sci. 63: Kos, B., J. Suskovic, J. Goreta, and S. Matosic Effect of protectors on the viability of Lactobacillus acidophilus M92 in simulated gastrointestinal conditions. Food Technol. Biotechnol. 38: Laliberte, A Techniques instrumentales en biologie me dicale, vol. 2. Odile Germain, Sainte-Foy, Quebec, Canada. 26. Lankaputhra, W. E. V., and N. P. Shah Survival of Lactobacillus acidophilus and Bi dobacterium spp. in the presence of acid and bile salts. Cult. Dairy Prod. J. 30(3): Laroia, S., and J. H. Martin Bi dobacteria as possible dietary adjuncts in cultured dairy products a review. Cult. Dairy Prod. J. 25(4):18, Lee, K.-Y., and T.-R. Heo Survival of Bi dobacterium longum immobilized in calcium alginate beads in simulated gastric juices and bile salt solution. Appl. Environ. Microbiol. 66: Levy, M. C., and F. Edwards-Levy Coating alginate beads with cross-linked biopolymers: a novel method based on a transacylation reaction. J. Microencapsul. 13: Liu, P., and T. R. Krishnan Alginate-pectin-poly-L-lysine particulate as a potential controlled release formulation. J. Pharm. Pharmacol. 51: Mangino, M. E Physicochemical aspects of whey protein functionality. J. Dairy Sci. 67: Marteau, P., M. Minekus, R. Havenaar, and J. H. J. Huis In t Veld Survival of lactic acid bacteria in a dynamic model of the stomach and small intestine: validation and the effects of bile. J. Dairy Sci. 80: Marteau, P., P. Pochart, Y. Bouhnik, and J. C. Rambaud Survival and effects of Lactobacillus acidophilus and bi dobacteria from cultured milks in the human digestive tract. Cah. Nutr. Diet. 29: Martinsen, A., G. Skjak-Braek, and O. Smidsrod Alginate as immobilization material. I. Correlation between chemical and physical properties of alginate gel beads. Biotechnol. Bioeng. 33: Martinsen, A., I. Storro, and G. Skjak-Braek Alginate as immobilization material. III. Diffusional properties. Biotechnol. Bioeng. 39: McDowell, R. H New reactions of propylene glycol alginate. J. Soc. Cosmet. Chem. 21: McKay, J. E., G. Stainsby, and E. L. Wilson A comparison of the reactivity of alginate and pectate esters with gelatin. Carbohydr. Polym. 5: Mitsuoka, T Taxonomy and ecology of bi dobacteria. Bi dobacteria Micro ora 3(1): Moe, S. T., K. I. Draget, G. Skjak-Braek, and O. Smidsrod Alginates, p In A. M. Stephen (ed.), Food polysaccharides and their applications. Marcel Dekker, New York. Mohamed, S. B., and G. Stainsby Ability of various proteins to form thermostable gels with propylene glycol alginate. Food Chem. 13: Mohamed, S. B., and G. Stainsby Lysine availability in protein-alginate ester gels. Food Chem. 14:1 10. Mohamed, S. B., and G. Stainsby The digestibility of gelatin complexed with propylene glycol alginate. Food Chem. 18: Morr, C. V., and E. Y. W. Ha Whey protein concentrates and isolates; processing and functional properties. Crit. Rev. Food Sci. Nutr. 33: Ouwerx, C., N. Velings, M. M. Mestdagh, and M. A. V. Axelos Physico-chemical properties and rheology of alginate gel beads formed with various divalent cations. Polym. Gels Netw. 6: Rao, A. V., N. Shiwnarain, and I. Maharaj Survival of microencapsulated Bi dobacterium pseudolongum in simulated gastric and intestinal juices. Can. Inst. Food Sci. Technol. J. 22: Silva Cunha, A., J. L. Grossiord, F. Puisieux, and M. Seiller Insulin in w/o/w multiple emulsions: preparation, characterization and determination of stability towards proteases in vitro. J. Microencapsul. 14: Sultana, K., G. Godward, N. Reynolds, R. Arumugaswamy, P. Peiris, and K. Kailasapathy Encapsulation of probiotic bacteria with alginate-starch and evaluation of survival in simulated gastrointestinal conditions and in yoghurt. Int. J. Food Microbiol. 62: Thibault, J. F., and P. Colonna Proprie te s fonctionnelles: pectine et amidon, p In D. Lorient, B. Colas, and M. Le meste (ed.), Proprie te s fonctionnelles des macromole cules alimentaires. Cahiers de l ENSBANA, Paris. Trindade, C. S. F., and C. R. F. Grosso The effect of the immobilisation of Lactobacillus acidophilus and Bi dobacterium lactis in alginate on their tolerance to gastrointestinal secretions. Milchwissenschaft 55: Velings, N. M., and M. M. Mestdagh Physico-chemical properties of alginate gel beads. Polym. Gels Netw. 3: Verheul, M., and S. P. F. M. Roefs Structure of particulate whey protein gels: effect of NaCl concentration, ph, heating temperature, and protein composition. J. Agric. Food Chem. 46: Voragen, A. G. J., and W. Pilnik Pectins, p In A. M. Stephen (ed.), Food polysaccharides and their applications. Marcel Dekker, New York. Zavaglia, A. G., G. Kociubinski, P. Perez, and G. De Antoni Isolation and characterization of Bi dobacterium strains for probiotic formulation. J. Food Prot. 61: Zhou, Y., E. Martins, A. Groboillot, C. P. Champagne, and R. J. Neufeld Spectrophotometric quanti cation of lactic bacteria in alginate and control of cell release with chitosan coating. J. Appl. Microbiol. 84: