CHAPITRE 5 : Changes of lactic and probiotic culture characteristics during continuous immobilized-cell fermentation with mixed strains

Table des matières

Y. Doleyres, I. Fliss et C. Lacroix

Centre de recherche en sciences et technologie du lait STELA

Université Laval

Québec, Qc, G1K 7P4, Canada

Les effets de l’immobilisation cellulaire et de la fermentation continue sur une période de 17 jours sur les caractéristiques probiotiques et technologiques de bactéries lactiques produites dans l’effluent ont été étudiés. Le procédé de fermentation comprenait deux réacteurs en série, le premier réacteur (R1) contenant des cellules de Bifidobacterium longum ATCC 15707 et de Lactococcus lactis subsp. lactis biovar. diacetylactis MD immobilisées séparément dans des billes de gel de polysaccharides, et le second réacteur (R2) opéré avec les cellules libres relarguées du premier réacteur. La fermentation fut conduite pendant 17 jours à différentes températures entre 32 et 37°C. La tolérance des cellules libres produites dans l’effluent des deux réacteurs à des stress variés, tels la lyophilisation, l’addition de peroxyde d’hydrogène, les conditions gastro-intestinales simulées, la présence de nisine et d’antibiotiques a augmenté avec le temps de fermentation et fut en général plus élevée après 6 jours de fermentation que celle de cellules produites de manière conventionnelle par fermentation en ‘batch’ avec des cellules libres. Des changements dans la paroi et la membrane cellulaire pourraient être responsable de cette adaptation non spécifique des cellules, puisque ce phénomène d’augmentation de tolérance cellulaire fut observé pour tous les stress et toutes les souches testées. La réversibilité de la tolérance de B. longum acquise aux antibiotiques, mais pas de L. diacetylactis , a été montrée lors de fermentations en ‘batch’ successives, dont la première avait été inoculée avec des cellules prélevées dans l’effluent de la fermentation continue avec cellules immobilisées. Notre étude a montré que la fermentation en continu avec cellules immobilisées pourrait être utilisée pour élaborer des produits industriels contenant des probiotiques et des bactéries lactiques avec des caractéristiques améliorées, potentiellement utiles pour la santé humaine.

The effect of immobilization and long-term continuous-flow culture was studied on probiotic and technological characteristics of lactic acid bacteria produced in the effluent medium. The fermentation process consisted of two reactors in series, with a first reactor (R1) containing cells of Bifidobacterium longum ATCC 15707 and Lactococcus lactis subsp. lactis biovar. diacetylactis MD separately immobilized in polysaccharide gel beads, and a second reactor (R2) operated with free cells released from the first reactor. The culture was carried out for 17 days at different temperatures ranging from 32 to 37°C. The tolerance of free cells produced in the effluent medium of both reactors to various stresses, including freeze-drying, hydrogen peroxide, simulated gastro-intestinal conditions, nisin, and antibiotics, markedly increased with culture time and was generally higher after 6 days than that of cells produced during conventional free-cell batch (FC) fermentations. Changes in cell wall and cell membrane could be partly responsible for this non-specific environmental adaptation of cells, as this phenomenon of increased cell tolerance was observed for all stresses and both strains tested. The reversibility of the acquired tolerance of B. longum , but not L. diacetylactis , to antibiotics was shown during successive FC batch cultures initially inoculated with cells taken from the effluent of the continuous immobilized-cell fermentation. Our study showed that continuous culture with immobilized cells could be used to manufacture industrial products containing probiotic and lactic acid bacteria with enhanced characteristics and potentially improved benefits on human health.

Key words : immobilization, continuous fermentation, cell physiology, lactic bacteria, probiotics.

Probiotics are defined as microbial cells which transit the gastrointestinal tract and which, in doing so, benefit the health of the consumer (Tannock et al., 2000). Among these micro-organisms, bifidobacteria are already used in many probiotic dairy products including milk, yogurt, ice cream, and cheese. However, a thorough knowledge of the abilities of cells to survive manufacture and storage of the product is required to successfully develop foods containing high viable concentrations of these micro-organisms (Heller, 2001). Furthermore, probiotic cultures must reach the gastrointestinal tract in a viable state and in significant numbers to be beneficial to the host, which requires that they survive the acidic condition in the stomach and bile in the small intestine (Stanton et al., 2001). Presently, the only technique used to increase cell resistance to stresses occurring during production, storage or digestion relies on incubation of free cells under starving or other stressful conditions, such as heat, high concentrations of salt, bile salts, or hydrogen peroxide, or low pH (Gilliland and Rich, 1990; Reilly and Gilliland, 1999; Desmond et al. 2002). Cells can also be physically encapsulated after their production to create protection to different stresses such as oxygen, moisture, and acidity (Rao et al., 1989; Lee and Heo, 2000).

Immobilized cell technology has been shown to result in very high cell volumetric productivity of batch or continuous cultures compared with free-cell (FC) cultures, due to the high cell density maintained in the reactor (Doleyres et al., 2002a, b; chapters 2, 3, 4; Lamboley et al., 1997; 1999; 2001). The growth of immobilized cells results in high cell release into the bulk medium and efficient production of pure or mixed strain lactic acid bacteria (LAB) culture (Lamboley et al., 1997). Moreover, high stability of continuous immobilized cultures with mixed strains was obtained over long periods which allowed for constant biomass production, with a stable and controlled strain ratio (Lamboley et al., 1997; 1999; 2001). Other advantages of cell immobilization have been reported in previous studies, such as high resistance to bacterial and phage contaminations and high biological stability, including genetic stability (Champagne et al., 1994; Huang et al., 1996; Macedo et al., 1999). However, it has been reported that immobilization of living microbial whole cells may change the cell growth rate and morphology and sometimes alter the metabolic activity (Teixeira de Mattos et al., 1994; Krishnan et al., 2001). Several authors have observed an increased tolerance to product inhibition of immobilized cells (Teixeira de Mattos et al., 1994; Krisch and Szajani, 1997). Furthermore, increased resistance of cells to environmental stress factors, such as sanitizers or antibiotics, have been observed after prolonged cultures for specific systems (Jouenne et al., 1994; Trauth et al., 2001).

Recently, the production of a mixed model lactic culture containing a dominant LAB strain, Lactococcus lactis subsp. lactis biovar. diacetylactis MD, and a probiotic culture, Bifidobacterium longum ATCC 15707, was reported during a 17-day continuous immobilized-cell (IC) culture at different temperatures between 32 and 37°C (chapter 4). The two-stage fermentation system, composed of a first reactor containing cells of the two strains separately immobilized in κ-carrageenan/locust bean gum gel beads and a second reactor operated with free cells released from the first reactor, allowed continuous and stable production of a concentrated mixed culture at 35°C. In addition, composition and strain ratio of the culture produced in the effluent medium was influenced by temperature in the range studied.

In the present work, we report on the changes in physiological characteristics of cells produced during the 17-day continuous IC fermentation carried out at different temperatures (chapter 4) and compared cell characteristics to that of cells from conventional batch FC fermentations. B. longum and L . diacetylactis produced during FC batch and IC continuous fermentations were assessed for their tolerance to various stresses, by measuring important technological characteristics, such as survival to freeze-drying and hydrogen peroxide, sensitivity to nisin Z, and probiotic characteristics, such as cell survival to simulated gastro-intestinal conditions and tolerance to antibiotics.

The continuous IC culture was carried out as previously described (chapter 4). Briefly, a two-stage fermentation system was used for the continuous production of a mixed culture composed of L. diacetylactis and B. longum . The first flat-bottomed custom-built bioreactor (R1) contained 20 ml of the two types of pre-colonized gel beads immobilizing a pure culture, for a total culture volume of 120 ml. The second 600 ml useful-volume stirred tank reactor (R2, Bioflo model C30, New Brunswick Scientific Co., Edison, NJ, USA) was operated in series with free cells released from the first reactor R1. Carbon dioxide was injected in the headspace of the two reactors to maintain anaerobic conditions during the culture. The flow rate of supplemented MRS medium was set at 240 ml/h (dilution rate of 2 h-1 calculated for R1) during the 17-day culture and pH was controlled at 6.0 by addition of 6 N NH4OH. Temperature in both reactors was first set at 37°C, then decreased to 32°C at day 6, and finally increased to 35°C at day 13 until the end of the experiment. Broth samples were taken from the two reactors at 3-day intervals to test cell tolerance to different stress conditions.

Control samples used in the tests for strain sensitivity to antibiotics, freeze-drying, simulated gastro-intestinal conditions, hydrogen peroxide and nisin, were produced using the same procedure as for the inoculum used for immobilization. Control and effluent samples were centrifugated (10000 g at 4°C for 10 min) and suspended to their initial volume in a solution of 100 g/l skim milk and 100 g/l glycerol, prior to freezing at –80°C until testing.

The resistance of cells produced during continuous IC culture in the effluents from R1 and R2 to different environmental stresses was tested as a function of fermentation time and temperature, and compared with a control treatment. Control cells were produced during conventional batch FC fermentation in MRS medium inoculated at 2% and incubated for 18 h, anaerobically at 37°C (for B. longum ) or aerobically at 30°C (for L. diacetylactis ), with no pH control.

B. longum survival to simulated gastric and duodenal juices, studied separately, increased progressively with fermentation time, with no apparent effect of temperature, during continuous IC culture (Figure 5.3a). Samples at day 3 exhibited lower survival rates than control cells (0.02±0.01% and 0.08±0.03% in R1 compared with 2±1% and 21±9% for control culture, for gastric and duodenal juices, respectively). However, cell survivals reached very high values after 15 days continuous IC culture (69±4% in gastric juice and 70±7% in duodenal juice for R1). Survivals of cells from R2 were higher than that for cells from R1 for all fermentation times, with maxima of 86±3% and 90±13% after 15 days for gastric and duodenal juices, respectively (Figure 5.3a).

A different behaviour was observed for L. diacetylactis (Figure 5.3b). Cell survival also increased with culture time but was apparently much more dependent on fermentation temperature. During the first fermentation period of 6 days at 37°C, survival of cells from R1 to gastric and duodenal juices increased from 30±4% and 42±6% at day 3 to 60±9% and 78±12% at day 6, respectively. During this period, cell survival for samples from R2 to gastric and duodenal juices were higher than for R1, and also increased from day 3 to day 6 to reach very high values of 79±12% and 82±12%, respectively. Decreasing fermentation temperature to 32°C resulted in a progressive decrease of L. diacetylactis cell survival, to 37±5% and 40±6% for gastric and duodenal juices in R1, respectively, and to 66±10% and 72±11% for gastric and duodenal juices in R2 respectively, at day 12. Then, increasing temperature to an intermediate value of 35°C increased cell survival to gastric and duodenal juices, to 67±10% and 73±11% in R1, and 74±11% and 77±12% in R2 at day 15, respectively. L. diacetylactis control cell survival to gastric and duodenal juices was 54±8% and 66±10%, respectively.

B. longum (Figure 5.4a) and L. diacetylactis (Figure 5.4b) cells produced during continuous IC fermentation in reactors R1 and R2 for different culture times showed increased tolerance to chloramphenicol compared with control cells from FC batch culture. The tolerance to chloramphenicol increased with time for the two strains, with samples from day 3 exhibiting lower tolerance (29.7±2.5 and 31.3±1.6 mm inhibition for B. longum and L. diacetylactis in R1, respectively) than control cells (25.8±2.0 and 26.0±1.1 mm inhibition for B. longum and L. diacetylactis , respectively). However, cells produced after day 6 showed higher tolerance (19.2±1.7 and 22.2±1.3 mm inhibition for B. longum and L. diacetylactis in R1 after 15 days culture, respectively) than the control (Figure 5.4). Cell tolerance to chloramphenicol of samples from R2 was not significantly (p>0.05) different from that of R1 for each fermentation time.

Increased cell tolerance as a function of fermentation time was observed for various antibiotics. Indeed, L. diacetylactis cells produced in the effluent of R2 after 15 days continuous IC culture were significantly (p<0.05) more tolerant to ampicillin, bacitracin, chloramphenicol, gentamycin, kanamycin, neomycin, penicillin, polymyxin B, streptomycin, tetracyclin, and vancomycin than control cells, whereas B. longum showed increased tolerance to ampicillin, bacitracin, chloramphenicol, erythromycin, neomycin, penicillin, and vancomycin (Table 5.1). However, no increase in tolerance to erythromycin, novobiocin and rifampin was observed for L. diacetylactis and to gentamycin, novobiocin, rifampin, and tetracycline for B. longum as a result of cell immobilization.

Control cells of B. longum were resistant to kanamycin, polymyxin B, and streptomycin (6 mm of inhibition diameter corresponds to the diameter of the antibiotic disk) (Table 5.1), and the same result was obtained for cells produced during continuous IC fermentation for all fermentation times (data not shown).

The sensitivity of B. longum and L. diacetylactis to nisin Z as a function of continuous IC fermentation time and reactors (R1 or R2) is reported in Figure 5.5. The ordinate intercept of the regression equations of the inhibition diameter as a function of log nisin Z concentration was used to measure nisin sensitivity of a cell sample. The intercept progressively decreased with fermentation time for B. longum (4.2, -2.3, -6.3, -10.4, and -15.3 mm of inhibition after 3, 6, 9, 12, and 15 days in R1), and for L. diacetylactis (0.4, -5.8, -5.0, -9.8, and -12.8 mm of inhibition after 3, 6, 9, 12, and 15 days in R1), indicating that cells became progressively less sensitive to nisin with time during continuous IC culture (Figure 5.5).

For a given strain and fermentation time, the ordinate intercept of samples from the effluent of R2 (1.3, -6.4, -9.2, -12.6, -16.5, and –2.6, -4.2, -7.6, -13.7,-14.7 mm of inhibition after 3, 6, 9, 12, and 15 days in R2 for B. longum and L. diacetylactis , respectively) was lower than that from cells in the effluent of R1 (Figures 5.5c and 5.5d). Control cells (ordinate intercept of –1.6 and –1.5 mm for B. longum and L. diacetylactis ) exhibited lower sensitivity to nisin than samples from IC culture after day 3, but higher sensitivity than samples from day 6 and the following days.

For a given strain, the slope of the regression progressively increased with fermentation time from 5.0 to 10.2 and from 5.4 to 9.9 after 3 and 15 days for B. longum in R1 and R2, respectively, and from 5.7 to 10.2 and from 6.6 to 10.0 after 3 and 15 days for L. diacetylactis in R1 and R2, respectively). The slope of the regression for control cells were 5.4 and 5.5 for B. longum and L. diacetylactis , respectively.

Tolerance to ampicillin, bacitracin, penicillin and chloramphenicol of B. longum cells sampled from the effluent of R2 after 15 days continuous IC fermentation that were subcultured decreased with the number of batch cultures (Figure 5.6). The inhibition diameter of 19.0±1.5 mm for ampicillin and cells from R2 after 15 days progressively increased to 24.5±0.8 mm after three batch cultures, and remained stable during subsequent batch cultures (25.0±0.6 mm) (Figure 5.6a). After the third batch culture, the inhibition diameter with ampicillin was not significantly different from control cells (25.2±1.1 mm). A similar decrease in tolerance to penicillin, bacitracin and chloramphenicol was observed for B. longum during repeated batch cultures, and tolerance to these antibiotics was similar to that for control cells, after 3, 5, and 7 batch cultures, respectively (Figure 5.6a).

During seven repeated FC batch cultures initially inoculated with L. diacetylactis cells taken from the effluent of R2 after 15 days continuous IC fermentation, cell tolerance to ampicillin, bacitracin, penicillin and kanamycin remained stable, with inhibition diameters averaging 18.7±1.2, 18.9±1.0, 15.0±1.1, and 7.1±0.2 mm, respectively (Figure 5.6b). These values were significantly lower than that for control cells (33.5±0.9, 27.7±1.3, 25.7±0.3, and 15.7±1.1 mm inhibition, for ampicillin, bacitracin, penicillin and kanamycin, respectively). The high tolerance to vancomycin of B. longum and L. diacetylactis cells acquired after 15 days continuous culture in R2 was stable during the seven successive batch FC cultures (6.0±0.0 mm corresponding to the diameter of the antibiotic disk) and significantly different from control cells (20.2±1.1 and 21.3±0.9 mm inhibition for B. longum and L. diacetylactis , respectively).

This work was carried out to study the effects of the cell production technology on tolerance to different stresses of two strains, B. longum and L. diacetylactis . Cells produced during conventional batch FC fermentation and long-term continuous IC fermentations were compared for their survival to different technological and environmental stresses, including freeze-drying, hydrogen peroxide, simulated gastric and intestinal conditions, antibiotics, and nisin Z. As a general trend for the different stresses applied, survivals of cells produced with immobilized cell technology significantly increased with fermentation time, and were higher after 3-6 day culture than that for cells produced during batch FC culture.

Commercial dairy starters are increasingly supplied in the freeze-dried form because such starters can eliminate the subculturing steps without the risks of thawing during transportation or storage of the frozen culture. Even though many compounds have been successfully used to improve the survival of LAB during freeze-drying, other means of increasing cell survival such as fermentation process are of interest to culture suppliers. For B. longum cells produced during continuous IC culture, cell survival after freeze-drying increased for fermentation times up to day 9, and then remained constant until the end of the experiment. Fermentation time and temperature apparently had an effect on B. longum survival, with a stable high value (9.3±3.1%) measured between day 9 and 12 for a low temperature of 32°C. Even though the effects of fermentation time and temperature on cell survival were superposed, our data suggest that a temperature lower than the optimal value for B. longum growth could increase cell resistance to freeze-drying. For L. diacetylactis , a slight non-significant increase in cell survival to freeze-drying was also observed as a function of fermentation time, but this characteristic was not different than for control cells. Moreover, no significant difference was observed for both strains between samples from R1 and R2 for a given fermentation time. The limited accuracy of the pour plate method used to determine cell survival could partly explain the absence of effects for L. diacetylactis .

Cell tolerance to hydrogen peroxide was tested because bifidobacteria are classified as anaerobic organisms, although some species can tolerate low concentrations of oxygen. Precautions are thus required to prevent toxic effects of oxygen when bifidobacteria are grown for industrial applications or used in products (Shah, 1997). Indeed, oxygen toxicity is generally considered to result from the effects of different compounds including superoxide, hydrogen peroxide, and hydroxyl radicals (Shoesmith and Worsley, 1984). Hydrogen peroxide produced by yogurt organisms, especially Lactobacillus delbrueckii subsp. bulgaricus , may also affect the viability of bifidobacteria, and the presence of hydrogen peroxide in acidic conditions, such as in yogurt, may cause synergistic inhibition of bifidobacteria (Lankaputhra et al., 1996). Our study showed that both B. longum and L. diacetylactis cells with enhanced tolerance to H2O2 concentrations of 20000 ppm or higher were produced during continuous IC fermentation, compared with conventional batch FC culture. This increase in tolerance was very high (40- and 3–fold higher survival rate at 30000 ppm H2O2 for B. longum and L. diacetylactis from IC culture compared with the control, respectively) and developed rapidly in the course of the culture (less than 3 days) compared with the slower adaptation of cells to the other stresses tested. Cell tolerance to H2O2 did not change after 3 days, with no effect of reactors.

Cell tolerance to simulated gastric and intestinal juices was tested because probiotic cultures must reach the gastrointestinal tract in a viable state and in significant numbers to be beneficial to the host, which requires that they survive the acidic conditions in the stomach and bile in the small intestine (Stanton et al., 2001). A very large increase in cell survival to both simulated gastric and duodenal juices was observed for B. longum as a function of continuous IC fermentation time, with a higher survival rate than for control cells after 9 days continuous culture. For L. diacetylactis , cell survival seemed to be more dependent on fermentation temperature. L. diacetylactis cells were less tolerant to simulated gastro-intestinal conditions when they were grown close to their optimal growth temperature (32°C) compared with higher temperatures (up to 37°C), indicating that a temperature stress in addition to the effect of immobilization generated cells that could better survive during digestion. In a previous study, we showed that a change in growth temperature between 32 and 37°C had more effect on growth kinetics and concentration in the effluent of continuous IC fermentation of L. diacetylactis than for B. longum (chapter 4).

Tolerance of probiotic cells to antibiotics is of great interest due to their possible use to reconstitute the intestinal microflora of patients suffering from antibiotic-associated colitis. Cell tolerance to different antibiotics greatly increased with culture time during continuous IC fermentation (Table 5.1). Cells generally exhibited lower tolerance after 3 days continuous culture than control cells but higher tolerance after day 6 and the following days. Escherichia coli cells immobilized in agar gel layers also exhibited increased resistance to an antibiotic (latamoxef) compared with free (planktonic) cells, with tolerance increasing with ageing of the immobilized culture (Jouenne et al., 1994). The authors explained this difference in sensitivity by a limitation of diffusion of latamoxef in the biofilm-like agar structure. However, in our study, an increased tolerance was observed for both released cells in R1 and cells that subsequently grew in a suspended state in R2. Moreover, tolerance to antibiotics of cells from R2 was generally higher than for cells from R1 for a given fermentation time, as shown for chloramphenicol. In a previous study, we reported growth inhibiting conditions in R2 due to high lactic and acetic acid production by the two strains, which can therefore have added a stress that contributed to further increase cell tolerance to antibiotics (chapter 4).

Bacteriocins from lactic acid bacteria are protein compounds showing antimicrobial activity against several gram-positive bacteria (Klaenhammer, 1993). Nisin is the best-known bacteriocin and the only one recognized as GRAS (generally recognized as safe) for several specific applications in more than 50 countries as a food preservative (Turtell and Delves-Broughton, 1998). However, one important limit for the use of nisin in dairy products or of nisin-producing lactic acid bacteria in mixed starters is its inhibitory effect on other suitable lactic strains (Bouksaim et al., 2000; Benech et al., 2002). Our study showed that cells produced during long-term continuous IC fermentation exhibited increased tolerance to nisin Z with fermentation time. Cells were more tolerant to nisin Z concentrations of ≤200 UI/ml than control cells after 6-9 days continuous culture. Moreover, the slopes of the regression equations increased with fermentation time, indicating that the increased tolerance to nisin Z was more pronounced for low nisin concentrations.

The low tolerance at day 3 observed in all tests might be explained by the fact that cells produced during the continuous IC fermentation at high dilution rate (2 h-1 calculated for the total volume of the first IC reactor R1, and 0.4 h-1 for the second FC reactor R2) are in exponential growth phase. Indeed, it is well known that cells in exponential phase are much more sensitive to the environment than cells in stationary growth phase such as control cells produced after 18 h batch FC cultures (Kolter, 1993; Hartke et al., 1994; Rallu et al., 1996). Moreover, control cells that are propagated without pH control are subjected to acid stress at the end of culture which is already commonly used to increase cell resistance to stresses occurring during production and storage (Hartke et al., 1996). However, compared with cells of both strains in exponential growth phase in R1, cells in R2 were in late exponential growth phase for L. diacetylactis and in stationary or decline/death growth phase for B. longum due to the the higher residence time in R2 compared with R1, resulting in an increase tolerance to environmental stresses (chapter 4).

Cell growth in gel beads is limited by diffusional limitations of both substrates and inhibitory products, in this case lactic and acetic acids (Arnaud et al., 1991; 1992; Lamboley et al., 1997). This leads to the development of steep gradients of inhibitory products, pH and biomass in colonized beads (Masson et al., 1994; Doleyres et al., 2002a; chapter 3). Recently, enhanced cell-to-cell signalling and cell-matrix interactions leading to coordinated behaviour of immobilized microorganisms were reported (Shapiro and Dworkin, 1997). Such studies must be carried out to better understand the effects of the local microenvironment of immobilized cells combined to long-term culture leading to a progressive increase of resistance characteristics for the two strains.

For industrial benefits of these cell modifications, particular attention should be paid to the possible reversibility of the cell characteristics when they are further grown or incorporated in dairy products. Consequently, cells from the effluent of R2 after 15-day continuous IC fermentation were propagated for seven successive FC batch cultures to study the reversibility of the strain-characteristic modifications observed during long-term continuous IC culture, with tolerance to antibiotics as a model. Tolerance of B. longum cells to four antibiotics (ampicillin, bacitracin, penicillin, and chloramphenicol), which were selected for the test, progressively decreased with numbers of batch cultures and reached that of control cells after seven batch cultures or less. However, the high tolerance to vancomycin of B. longum cells measured after the 15-day continuous culture in R2 was stable during the seven successive FC batch cultures. On the other hand, the behaviour of L. diacetylactis cells was very different. For this culture, the tolerance to antibiotics was stable during seven successive FC batch cultures, and remained significantly higher than for control cells.

Several studies have shown that cells produced with immobilized cell technology exhibited a change in growth, morphology and physiology characteristics compared with cells produced during conventional FC cultures. Cachon et al. (1998) reported differences in cell physiology during batch cultures with and without pH control and continuous cultures with free and immobilized Lactococcus lactis , depending on the culture mode. The redox states, enzymatic pool and intracellular pH differed for immobilized and free-cell cultures. A shift in the metabolic pathway from homofermentative to heterofermentative has also been observed during continuous cultures with immobilized cells of Lactobacillus (Krishnan et al., 2001). Moreover, many authors observed an increased tolerance of immobilized cells to product inhibition (Krisch and Szajani, 1997), alcohols (Holcberg and Margalith, 1981; Curtain, 1986), phenols (Keweloh et al., 1989; Heipieper et al., 1991; Diefenbach et al., 1992), or quaternary ammonium sanitizers (Trauth et al., 2001). Immobilization as well as prolonged culture of immobilized cells led to increased cell tolerance to sanitizers (Trauth et al., 2001). In the last study, the authors tentatively explained this phenomenon of increased tolerance by a possible modification of the cell membrane and physiology, and protection of the cell membrane by the proximity of cells in a saturated matrix. Our study showed that the increased tolerance to various environmental stresses when cells were immobilized was not associated with cell- or strain-specific mechanisms or physical protection by cell contact and high density in the gel matrix, since tolerance was measured with released cells and for different stresses and strains. Our results suggest that the non-specific increase in cell resistance during long-term continuous IC fermentations could be partly due to a reversible increase in the thickness of the cell wall or due to modifications of the cell membrane, as proposed by Trauth et al (2001) for the increased tolerance to quaternary ammonium sanitizers of immobilized cells.

For industrial applications of immobilized cells for culture production with increased characteristics, the colonized beads should be stored under adequate conditions between two productions to ensure good viability and to avoid the approximately 6-day adaptation phase of newly immobilized cells. Moreover, particular attention should be paid to the possible reversibility of the characteristics of cells produced during continuous IC culture when they are further grown or incorporated in dairy products, which depends both on strains and characteristics. However, our study showed that immobilized cell technology combined with long-term continuous culture can be used to efficiently produce, in a one step process, cells with enhanced tolerance to environmental stresses, without the need for preconditioning treatments which are sometimes used for better survival of probiotics during production and use in functional foods, but eventually result in reduced cell activity and yield (Desmond et al., 2002). In addition, cells produced with continuous IC cultures are in exponential or early stationary growth phase, and they exhibited both a high viability and metabolic activity compared with starving cells produced with conventional batch cultures. These results together with the high productivities of immobilized cell technology for cell and metabolic product production (chapter 4), show that there is a crucial advantage of this process for producing biomass with controlled properties and improved stress resistance compared with conventional batch FC cultures, particularly in the area of probiotic culture applications.

Probiotic cultures produced with continuous IC fermentations could survive to a greater extent as they pass through the gastro-intestinal tract and could be used to manufacture products with new or enhanced health claims, as well as new food or pharmaceutical bio-ingredients. Probiotics with enhanced tolerance to antibiotics could also be used to reconstitute the intestinal microflora of patients suffering from antibiotic-associated colitis or to be given to farm animals which are eating growth-promoting concentrations of antibiotics in their food. One important limitation for using nisin in dairy products or nisin-producing lactic acid bacteria in mixed starters is the inhibitory effect of the peptide on suitable acidifying or aroma-producing lactic strains (Bouksaim et al., 2000; Benech et al., 2002). Indeed, immobilized cell technology could also be used to enhance lactic starter or probiotic tolerance to nisin prior to incorporation into nisin-containing products. Finally, a better cell survival to freeze-drying conditions and resulting increase in process cell yield would be of great economical interest to culture suppliers.

Table 5. 1 : Inhibition diameters (mm) measured by the disc assay method 1 of different

antibiotics on L. diacetylactis MD and B. longum ATCC 15707 produced during FC batch culture (control) or after 15 days continuous IC fermentation in the effluent of reactor R2.

1 A diameter of 6 mm corresponds to the diameter of the disk and indicates the absence of inhibition of the culture by the antibiotic.

Figure 5. 1 : B. longum ( ) and L. diacetylactis ( ) cell survival during freeze-drying of control (batch FC culture) and experimental (continuous two-stage IC fermentation) cultures for different culture times. For the IC culture, mean survival rates for the first (R1) and second (R2) reactors are reported because there was no significant effect of reactors (p>0.05). Temperature during continuous IC culture was set at 37°C from day 1 to day 6, 32°C until day 13, and 35°C thereafter.

Figure 5. 2 : B. longum ( ; ) and L. diacetylactis ( ; ) cell survival to hydrogen peroxide (H2O2) of control (batch FC culture, dark symbols), and experimental (continuous two-stage IC fermentation, open symbols) cultures. For the IC culture, mean survival rates of samples from the first (R1) and second (R2) reactors for different culture times are reported because there was no significant effect of reactors and culture time (p>0.05). Temperature during continuous IC culture was set at 37°C from day 1 to day 6, 32°C until day 13, and 35°C thereafter.

Figure 5. 3 : Cell survival to simulated gastric and duodenal juices for B. longum (a) and L. diacetylactis (b) from control (batch FC culture, ) and experimental (continuous two-stage IC fermentation) cultures, for reactors R1 ( = gastric juice; = duodenal juice) and R2 ( = gastric juice; = duodenal juice) and different culture times. Temperature during continuous IC culture was set at 37°C from day 1 to day 6, 32°C until day 13, and 35°C thereafter.

Figure 5. 4 : Tolerance to chloramphenicol of B. longum (a) and L. diacetylactis (b) cells produced during batch FC culture (control, ) or during continuous IC fermentation in reactors R1 ( ) and R2 ( ) at different culture times. The antibiotic sensitivity was tested by the disc assay method. Temperature during continuous IC culture was set at 37°C from day 1 to day 6, 32°C until day 13, and 35°C thereafter.

Figure 5. 5 : Tolerance to different nisin Z concentrations of cells produced during batch FC culture (control, l) or after 3 ( ), 6 ( ), 9 ( ), 12 ( ), or 15 (¡) days continuous IC fermentation: a) B. longum in reactor R1 ; b) L. diacetylactis in reactor R1 ; c) B. longum in reactor R2 ; d) L. diacetylactis in reactor R2. Temperature during continuous IC culture was set at 37°C from day 1 to day 6, 32°C until day 13, and 35°C thereafter.

Figure 5. 6 : Inhibition diameter (mm) measured by the disc assay method for different antibiotics ( = ampicillin; = bacitracin; = penicillin; = chloramphenicol) on B. longum (a) and L. diacetylactis (b) cells produced after successive batch FC cultures, inoculated (2%) with a culture sampled from reactor R2 after 15 days continuous IC fermentation. Control samples were produced during batch FC culture.