Bile Salt Hydrolase Activity of Enterococci Isolated from Food: Screening and Quantitative Determination

Transcription

1 725 Journal of Food Protection, Vol. 64, No. 5, 2, Pages Copyright q, International Association for Food Protection Research Note Bile Salt Hydrolase Activity of Enterococci Isolated from Food: Screening and Quantitative Determination CHARLES M. A. P. FRANZ, INGRID SPECHT, PETRA HABERER, AND WILHELM H. HOLZAPFEL* Federal Research Center for Nutrition, Institute for Biotechnology and Molecular Biology, Haid-und-Neu Strasse 9, D-763 Karlsruhe, Germany MS -265: Received 3 August 2/Accepted 2 October 2 ABSTRACT One hundred seventeen enterococcal strains isolated from food (47 Enterococcus faecium, 48 Enterococcus faecalis, 6 Enterococcus durans, 2 Enterococcus gallinarum, 3 Enterococcus casseli avus, and Enterococcus malodoratus) were screened for bile salt hydrolase (BSH) activity on de Man, Rogosa, and Sharpe agar medium containing taurocholic acid and calcium chloride. The highest incidence of BSH-active strains was observed for E. faecalis (8%) followed by E. faecium (5%) and E. durans (44%). Isolates were grouped into four putative activity groups (no, low, medium, and high activity) based on the size of precipitation zones observed in the screening experiment. Our results showed that assumptions on BSH activity based on the size of bile precipitation zones in screening experiments did not correlate with actual activity as quanti ed by high-pressure liquid chromatography, but the screening assay is useful for assessing the presence or absence of BSH activity. Interest in the application of lactic acid bacteria and Bi dobacterium strains as probiotics has increased dramatically during the last few years. Important selection criteria for successful probiotic strains include physiological features and functional properties pertaining to health-promoting or health-sustaining activity (3, 23). Suggested functional effects include inhibition of pathogenic microorganisms, strengthening of the gut mucosal barrier, antimutagenic and anticarcinogenic activities, increase of the immune response, and lowering of blood cholesterol levels ( 3, 2 22). Cholesterol lowering effects have been linked to bacterial bile salt hydrolase (BSH) activity. BSH activity is mediated by various gram-positive intestinal bacteria, including members of the genera Enterococcus, Peptostreptococcus, Bi dobacterium, Clostridium, Bacteroides, and Lactobacillus (2, 4). The conjugated BSH enzyme (E.C ) deconjugates bile salts by liberating the glycine and/or taurine moiety from the side chain of the steroid core. It was hypothesized that deconjugation of bile salts may contribute to lower cholesterol levels because free bile acids may be excreted more likely from the gastrointestinal tract than conjugated bile salts (2). If enhanced fecal loss of bile acids occurs as a result of bacterial BSH activity, it may increase the demand for cholesterol as a precursor for de novo synthesis of bile salts, which in turn may lower cholesterol levels (5, 8). However, this BSH hypothesis is controversial because it contradicts current knowledge on the passive adsorption kinetics of free bile acids in the gastrointestinal tract (, 9). Observations based on pig and * Author for correspondence. Tel: ; Fax: ; wilhelm.holzapfel@bfe.uni-karlsruhe.de. minipig feeding experiments with BSH active lactobacilli strongly suggest a cholesterol-lowering effect (6, 9). Klaver and van der Meer (5) suggested that the in vitro cholesterol reduction by some Lactobacillus spp. results from the coprecipitation of cholesterol with deconjugated bile salts. Colonies of Bi dobacterium spp. grown on agar containing cholesterol and bile were shown later to form cholesterol-containing precipitates on the agar surface (6). It was assumed that the precipitate formed as a result of the bacterial extracellular BSH activity (6). In our previous studies, BSH-active enterococci were detected on de Man, Rogosa, and Sharpe (MRS) agar supplemented with the sodium salt of taurodeoxycholic acid and CaCl 2 (9). This study aimed to screen selected food enterococcal strains for BSH activity, determine the incidence of BSH activity among strains, and semiquantify BSH activity on the basis of the size of the precipitation zone. Furthermore, it aimed to correlate the results from semiquantitative screening experiments with quantitative determination of BSH activity of selected enterococcal strains by high-pressure liquid chromatography (HPLC). MATERIALS AND METHODS Source and maintenance of cultures. A total of 45 isolates of enterococci were collected from six laboratories and were identi ed to species level by physiologicaltests, determination of sugar fermentation patterns using the API 5CH test (biomérieux Germany, Nürtingen, Germany), and sodium dodecyl sulfate-polyacrylamide gel electrophoresis of total soluble cell proteins, whereas randomly ampli ed polymorphic DNA polymerase chain reaction, pulsed- eld gel electrophoresis, and ampli ed fragment length polymorphism studies provided information on biodiversity (unpublished results). The 45 enterococcal strains were isolated

2 726 FRANZ ET AL. J. Food Prot., Vol. 64, No. 5 from food, clinical, and veterinary sources (ratio of 6::), and of these, 7 strains of food origin were selected to study functional properties relating to cheese fermentation and use as probiotics. These strains consisted of Enterococcus faecium (48 strains), Enterococcus faecalis (47 strains), Enterococcus durans (6 strains), Enterococcus gallinarum (2 strains), Enterococcus casseli avus (3 strains), and Enterococcus malodoratus ( strain). Strain information is available in the catalogue of enterococci of the FAIR- E collection (24). Catalogue and strains are available at the Belgian Coordinated Collection of Microorganisms/Laboratory of Microbiology Gent (BCCM/LMG) Bacteria Collection, Laboratory of Microbiology, University of Gent, Ledegankstraat 35, B- 9 Gent, Belgium. Stock cultures in MRS broth (Merck, Darmstadt, Germany) with 5% (vol/vol) glycerol added were kept at 288C. Cultures were subcultured at least twice before experimental use in MRS broth, using a % (vol/vol) inoculum of an overnight (8-h) culture and subsequent incubation at 378C for a further 8 h. Screening for BSH activity. Cultures were screened for BSH activity by spotting ml of culture grown in MRS broth onto BSH screening medium, which consisted of MRS agar plates supplemented with.5% (wt/vol) sodium salt of taurodeoxycholic acid (Sigma, Deisenhofen, Germany) and.37 g/liter of CaCl 2 (3). Plates were incubated anaerobically (atmosphere of 8% N 2, % CO 2, and % H 2 ) in an anaerobic work chamber (MK3 anaerobic work station, dw Scienti c, Shipley, England) at 378C for 72 h, after which the BSH activity was semiquanti ed by measuring the diameter of precipitation zones. Screening experiments were performed in duplicate. Isolates were grouped into one of six activity classes based on the diameter of the precipitation zone on BSH screening medium:, no precipitation;, precipitation on colony surface but not extending beyond the edge of the colony (up to 8 mm); 2, precipitation zone of up to mm; 3, precipitation zone of up to 2 mm; 4, precipitation zone of up to 5 mm; 5, precipitation zone of up to 7 mm; and 6, precipitation zone of up to 2 mm. Isolates belonging to activity classes and 2 were hypothesized to have low BSH activity, those belonging to activity classes 3 and 4 to have medium BSH activity, and those belonging to activity classes 5 and 6 to have high BSH activity. Quantitative determination of BSH activity by HPLC. One strain each of E. durans, E. faecium, and E. faecalis belonging to activity class was chosen as a control for quantitative determination of BSH activity by HPLC. In addition, for each of the low- and medium-activity groups, one strain each of E. faecium and E. faecalis was chosen, whereas for the high-activity group, one strain each of E. faecalis and E. durans was chosen. The E. durans strain served as a representativeof the high-activity group because it was the only one that was grouped in activity class 6, a level that was not attained by either E. faecium or E. faecalis strains. For BSH activity the deconjugation of glycocholic acid (GCA) added to a bacterial culture was determined by measuring the amounts of this compound during incubation. Cultures were grown in 25 ml of MRS broth (ph 6.5) for 8 h at 378C in the MK3 anaerobic work station to reach a cell density of 3 9 CFU/ml. After growth, the culture was diluted : using MRS broth (ph 6.5) containing mm (4.656 mg/ml) GCA (Sigma; G-2878) to reach an initial GCA concentration of approximately 2.3 mg/ml, which is the average concentration of this acid at the beginning of digestion in the duodenum. The cultures were incubated for a further 6 h under anaerobic conditions at 378C. Duplicate samples (.5 ml) were removed immediately after addition of MRS broth containing GCA and after.5, 3, 4.5, and 6 h. Samples were prepared for HPLC analysis by adjusting the ph to 7.5 with 5 N KOH to stop BSH activity. These samples were centrifuged at 7,937 3 g in a microcentrifuge for min, and the supernatant was diluted : using phosphate-buffered saline (ph 7.5). GCA was extracted using Sep Pak C8 columns (Waters, Eschborn, Germany), which were pre-equilibrated with methanol and distilled water according to the manufacturer s instructions. The column was washed with ml of distilled water followed by 3 ml of % acetone and ml of distilled water, and GCA was eluted with 3 ml of HPLC grade methanol. The eluate was evaporated at 378C in a block heater for 3 to 5 h, after which the precipitate was dissolved in ml of HPLC grade methanol. For HPLC quanti cation of GCA, the sample was rst diluted :2 in HPLC grade methanol and 2 ml was injected to a C 8 column (25 cm by 4.6 cm by 5 mm; Supelcosil ABZ, Sulpelco, Inc., Bellefonte, Pa.) tted with an appropriate precolumn, both of which were kept at 258C. The running buffer used was 7% methanol and 3%.7 M sodium acetate (ph 3.), and the column was run isocratically at a ow rate of ml/min. GCA was detected using a SHIMADZU SPD-A UV-Vis detector at 25 nm. A standard stock solution of GCA was made by dissolving mg in ml of methanol. The stock solution was diluted in methanol to obtain a concentration of mg/ml. For quanti - cation of GCA the standard was used as a reference, and the concentration (mg/ml) of both the reference and samples was calculated by multiplication of the area of the plotted curve with the response factor and the dilution factor. The means from duplicate samples in two replicate experiments were determined and used to plot bile salt deconjugation curves. RESULTS Incidence and semiquantitative determination of BSH activity among enterococcal strains. Among the 48 E. faecium strains, 24 (5%) were BSH positive, whereas of the 47 E. faecalis strains, 38 (8%) exhibited BSH activity (Table ). Seven (44%) of the 6 E. durans strains were BSH positive, whereas both the E. gallinarum strains and one of the E. casseli avus strains tested were BSH positive. The only E. malodoratus strain tested was BSH negative (Table ). Among the BSH-positive E. faecium strains, 22 were hypothesized to have low BSH activity, whereas the remainder exhibited medium BSH activity (Table ) on the basis of size of precipitation zones. Among the positive E. faecalis strains, 6 were hypothesized to have low BSH activity, 2 isolates to have medium activity, and the remainder to have high activity. For the E. durans strains about equal numbers belonged to hypothetical low-, medium-, and high-activity classes, whereas the E. gallinarum and E. casseli avus strains showed medium to high BSH activity (Table ). An example for each of the activity classes on BSH screening medium is shown in Figure. Quantitative determination of BSH activity by HPLC. By determining the concentration of GCA in HPLC quanti cation experiments, it could be shown that the strains E. faecium FAIR-E 9, E. faecalis FAIR-E 37, and E. durans FAIR-E 2, which were negative in the screening experiments, did not hydrolyze bile salt. The concentration of GCA added to cultures of these strains remained at the same level (approximately 2 mg/ml) throughout the 6- h incubation period (data not shown). The strains E. fae-

3 J. Food Prot., Vol. 64, No. 5 BILE SALT HYDROLASE ACTIVITY OF ENTEROCOCCI 727 TABLE. BSH activity of food enterococci as determined on BSH screening medium a Activity class Species (no.) BSH positive No activity () Low activity 2 Medium activity 3 4 High activity 5 6 E. faecium (48) E. faecalis (47) E. durans (6) E. casseli avus (2) E. gallinarum (2) E. malodoratus () 24 (5) 38 (8) 7 (44) 2 24 (5) 9 (9.4) 9 (56.25) (4.7) 3 (7.8) 2 (28.5) 2 (5) 3 (34.2) (4.3) 2 (8.3) 5 (39.5) (4.3) () 5 (3.2) (4.3) () 2 (5.2) (4.3) () () (4.3) a The diameter of the precipitation zone around each colony was taken as an indication of presumptive BSH activity. Numbers in parentheses are the percentages of BSH-positive strains. Percentage values were not determined for E. casseli avus, E. gallinarum, and E. malodoratus due to low number of strains investigated. calis FAIR-E 339 and E. faecium FAIR-E 54 belong to the hypothetical low-activity class but differed from each other markedly in their quantitative BSH activity as determined by HPLC. E. faecalis FAIR-E 339 indeed showed a low BSH activity, decreasing the concentration of GCA from 2, mg/ml to approximately,5 mg/ml during the 6-h incubation period. In contrast, E. faecium FAIR-E 54 showed a high BSH activity, reducing the GCA concentration to 7 mg/ml during 6-h incubation, even though this strain belonged to the low-bsh-af nity class (Fig. 2). The strains E. faecalis FAIR-E 44 and E. durans FAIR-E 23 belong to the hypothetical high-bsh-activity class. These strains indeed showed a high BSH activity, because they decreased the level of GCA to 29 mg/ml (FAIR-E 23) and 36 mg/ml (FAIR-E 44) after 6 h of incubation (Fig. 2). The strains E. faecium FAIR-E 84 and E. faecalis FAIR-E 88 belong to the hypothetical medium-bsh-activity class. However, E. faecium FAIR-E 84 showed high BSH activity, decreasing the GCA concentration to mg/ ml, whereas E. faecalis FAIR-E 88 decreased the GCA concentration to only approximately,27 mg/ml during 6 h of incubation (Fig. 2). This was noticeably less than the level of reduction observed for E. faecium FAIR-E 54, which belongs to the low-bsh-activity class according to screening results (Fig. 2). DISCUSSION FIGURE. Screening for BSH activity of enterococcal strains on BSH screening medium and examples for grouping into BSH activity classes based on diameter of bile precipitation zone: (a) no activity (FAIR-E 9), (b) low activity (FAIR-E 339), (c) medium activity (FAIR-E 84), and (d) high activity (FAIR-E 23). Strains of each of the enterococcal species (E. faecium, E. faecalis, E. durans, E. casseli avus, and E. gallinarum) in this study showed BSH activity. This is not surprising since enterococci are well known to be commensals of the gastrointestinal tract of humans and animals, and in this ecological niche, these bacteria come into contact and interact with bile salts. Deconjugation of bile salts was proposed to serve as a detoxi cation mechanism by reducing the toxicity of the protonated form of the bile salts, which can cause intracellular acidi cation similar to organic acids, to the less toxic and relatively weakly acidic deconjugated form (4, 6). The food enterococcal strains in this study were mainly isolated from cheese. The fact that many of these strains exhibit BSH activity may be explained by the fact that cheese enterococci isolates are known to originate as fecal contaminants in the dairy environment as a result of poor hygienic practices (7, 8, 9). The incidence of BSH activity appeared to be higher for E. faecalis than for E. faecium (8 versus 5%, respectively), but the reason for this is not known. Although the incidence of BSH activity among E. durans isolates was comparable to that of E. faecium isolates (44% versus 5%, respectively), the incidence

4 728 FRANZ ET AL. J. Food Prot., Vol. 64, No. 5 FIGURE 2. Deconjugation of GCA by food enterococcal strains as quantitatively determined by HPLC. Quantitative determination was done for enterococcal strains grouped into low-, medium-, and high-activity classes as determined in screening experiments. of BSH activity among the other enterococcal species could not be accurately determined because too few isolates were examined. Devriese et al. (7) showed that both E. faecalis and E. faecium could be isolated from feces of preruminant calves but only E. faecalis from ruminating young cattle and dairy cows. Speculatively, this may be a consequence of a higher incidence of BSH activity in E. faecalis strains. The screening experiments on medium using the taurine conjugated bile salt could be correlated with HPLC determinations using the glycine conjugated bile salt, because strains that showed a negative result in screening experiments with the taurocholic acid were shown to be also unable to deconjugate GCA in HPLC quanti cation experiments. Conversely, all strains that were positive in the screening experiments with taurocholic acid also deconjugated GCA in HPLC quanti cation experiments. Although it was initially assumed that size of bile precipitation zones observed in the screening for BSH activity may correlate to the level of BSH enzyme activity, our quantitative determinations by HPLC showed that this was not the case. The reason for lack of correlation of BSH screening experiments and HPLC quanti cation of BSH activity is not clear. It is possible that CaCl 2 ions present in the BSH screening medium, which are vital for bile precipitation, may be bound to the surface of the cell wall to different extents by the different enterococcal strains. Different ion concentrations and ph values also may contribute to differences in sizes of precipitation zones. Alternatively, this lack of correlation may be a result of the different types of bile salt used in the two methods for determining BSH activity. Previously, we chose enterococci isolates for studies on the cholesterol-lowering effect in minipig feeding trials on the basis of size of precipitation zones in screening experiments, assuming that large precipitation zones may be equated with high BSH activity (9). The results of this study show that this may not be the case, especially when considering hydrolysis of the glycine conjugate. GCA is the predominant form of conjugated bile salt in bile constituents of humans. Therefore, for screening of laboratory strains for the ability to deconjugate bile salts, the BSH screening medium may be adequate for distinguishing BSH-positive from BSH-negative strains. However, this method does not give information on the level of BSH activity of strains, especially with respect to the level of GCA deconjugation activity. ACKNOWLEDGMENT This study was carried out with nancial support from the Commission of the European Communities, Agriculture and Fisheries (FAIR) speci c RTD programme, CT Enterococci in Food Fermentations: Functional and Safety Aspects. It does not necessarily re ect its views and in no way anticipates the Commission s future policy in this area. REFERENCES. Aldini, R., M. Montagnani, A. Roda, S. Hrelia, P. L. Piagi, and E. Roda Intestinal absorption of bile acids in the rabbit: different transport rates in jejenum and ileum. Gastroenterology : Chikai, T., H. Nakao, and K. Uchida Deconjugation of bile acids by human intestinal bacteria implanted in germ-free rats. Lipids 22: Dashkevicz, M. P., and S. D. Feighner Development of a differential medium for bile-salt hydrolase-activity Lactobacillus spp. Appl. Environ. Microbiol. 55: De Boever, P., and W. Verstraete Bile salt deconjugation by Lactobacillus plantarum 8 and its implication for bacterial toxicity. J. Appl. Bacteriol. 87: De Rodas, B. Z., S. E. Gilliland, and C. V. Maxwell Hypocholestererolemic action of Lactobacillus acidophilus ATCC 432 and calcium in swine with hypercholesterolemia induced by diet. J. Dairy Sci. 79: De Smet, I., L. Van Hoorde, M. Vande Woestyne, H. Christiaens, and W. Verstraete Signi cance of bile salt hydrolytic activities of lactobacilli. J. Appl. Bacteriol. 79: Devriese, L. A., L. Laurier, P. De Herdt, and P. Haesebrouck Enterococcal and streptococcal species isolated from faeces of calves, young cattle and dairy cows. J. Appl. Bacteriol. 72: Driessen, F. M., and R. de Boer Fermented milks with selected intestinal bacteria: a healthy trend in new products. Neth. Milk Dairy J. 43: Du Toit, M., C. M. A. P. Franz, L. M. T. Dicks, U. Schillinger, P.

5 J. Food Prot., Vol. 64, No. 5 BILE SALT HYDROLASE ACTIVITY OF ENTEROCOCCI 729 Haberer, B. Warlies, F. Ahrens, and W. H. Holzapfel Characterisation and selection of probiotic lactobacilli for a preliminary minipig feeding trial and their effect on serum cholesterol levels, faeces ph and faeces moisture content. Int. J. Food Microbiol. 4: Fernandes, C. F., and K. M. Shahani. 99. Anticarcinogenic and immunological properties of dietary lactobacilli. J. Food Prot. 53: Fuller, R Probiotics in man and animals. J. Appl. Bacteriol. 66: Havenaar, R., and J. H. J. Huis in t Veld Probiotics: a general view, p In B. J. B. Wood (ed.), The lactic acid bacteria, vol.. The lactic acid bacteria in health and disease. Elsevier Applied Science, London. 3. Holzapfel, W. H., P. Haberer, J. Snel, U. Schillinger, and J. H. J. Huis in t Veld Overview of gut ora and probiotics. Int. J. Food Microbiol. 4: Hylemon, P. B Metabolism of bile acids in intestinal micro- ora, p In H. Danielsson and J. Sjövall (ed.), Sterols and bile acids. Elsevier, New York. 5. Klaver, F. A. M., and R. van der Meer The assumed assimilation of cholesterol by lactobacilli and Bi dobacterium bi dum is due to their bile salt-deconjugation activity. Appl. Environ. Microbiol. 59: Kociubinski, G., P. Pérez, and G. de Antoni Screening of bile resistance and bile precipitation in lactic acid bacteria and bi dobacteria. J. Food Prot. 62: Litopoulou-Tzanetaki, E. 99. Changes in numbers and kinds of lactic acid bacteria during ripening of Kefalotyri cheese. J. Food Sci. 55: López-Dõ az, T. M., J. A. Santos, T. J. González, B. Moreno, and M. L. Garcõ a Bacteriological quality of traditional Spanish blue cheese. Milchwissenschaft 5: Marteau, P., M. F. Gerhardt, A. Myara, E. Bouvier, F. Trivin, and J. C. Rambaud Metabolism of bile salts by alimentary bacteria during transit in the human small intestine. Microbial Ecol. Health Dis. 8: MikesÏ, Z., M. Ferencõ k, E. Janová, L. Erbringer, and I. Ciznár Hypocholesterolemic and immunostimulatory effects of orally applied Enterococcus faecium M-74 in man. Folia Microbiol. 4: Pool-Zobel, B. L., B. Bertram, M. Knoll, R. Lambertz, C. Neudecker, U. Schillinger, P. Schmezer, and W. H. Holzapfel Antigenotoxic properties of lactic acid bacteria in vivo in the gastrointestinal tract of rats. Nutr. Cancer 2: Salminen, S., M. Deighton, and S. Gorbach Lactic acid bacteria in health and disease, p In Lactic acid bacteria. Marcell Dekker Inc., New York. 23. Salminen, S., E. Isolauri, and E. Salminen Clinical uses of probiotics for stabilizing the gut mucosal barrier: successful strains for future challenges. Antonie van Leeuwenhoek 7: Vancanneyt, M., K. Kersters, and J. Swings Catalogue of enterococci of the FAIR-E collection. Published for the EU-project FAIR-CT-378 by the BCCM/LMG Bacteria Collection, Gent, Belgium.