One of the major pathogenic bacteria associated with acute diarrhoea is enterotoxigenic Escherichia coli (ETEC)Reference Hampson and Gyles1, Reference Bhan2. Diarrhoea caused by ETEC is a persisting health problem in children and young animalsReference Bhan2–Reference Nagy and Fekete4. Treatment options for acute gastroenteritis mainly rely on oral rehydration solution, although the volume, frequency or duration of diarrhoea are not reduced using conventional oral rehydration solutionReference Bhan2, Reference Mahalanabis5.
Tempeh is a traditional fermented food made from soaked dehulled and cooked whole soyabeans inoculated with a mould, usually of the genus Rhizopus Reference Nout and Kiers6. An important function of the mould in the fermentation process is the synthesis of biomass and enzymes that hydrolyse soyabean constituents and contribute to the development of a desirable texture, flavour and aroma of the product. The process also inactivates or eliminates certain anti-nutritional factors and the enzymatic degradation improves the nutritional qualityReference Nout and Kiers6.
Tempeh has been reported to contain an antibacterial substanceReference Wang, Ruttle and Hesseltine7–Reference Kobayasi, Okazaki and Koseki9 and in vitro tempeh extracts are able to inhibit adhesion of ETEC to piglet small intestinal brush border membranesReference Kiers, Nout, Rombouts, Nabuurs and van der Meulen10. Rabbits infected with ETEC and fed tempeh showed reduced diarrhoea compared with rabbits fed diets without tempehReference Karmini, Affandi, Hermana, Sudarmadji, Suparmo and Raharjo11. Perfusion of small intestinal segments of piglets with pre-digested tempeh strongly reduced ETEC-induced intestinal fluid lossesReference Kiers, Nout, Rombouts, van Andel, Nabuur and van der Meulen12 and in ETEC-challenged weaned piglets the severity of diarrhoea was reduced by a diet containing tempeh compared with a control diet containing toasted soyabeansReference Kiers, Meijer, Nout, Rombouts, Nabuurs and van der Meulen13. In malnourished children tempeh was reported to be beneficial in terms of duration of diarrhoea episodes and rehabilitation period when supplemented to their dietReference Kalavi, Muroki, Omwega and Mwadime14–Reference Karyadi and Lukito16.
The present study was undertaken to investigate whether the reduction of ETEC-induced intestinal fluid loss in piglets by tempeh can be attributed to unsoluble or to water-soluble tempeh fractions of either low or high molecular mass.
Material and methods
Tempeh
Dehulled yellow-seeded soyabeans (Glycine max) were soaked overnight for about 16 h in tap-water while undergoing accelerated lactic acid fermentation using naturally acidified soaking water as an inoculumReference Nout, de Dreu, Zuurbier and Bonants van Laarhoven17. Subsequently, the beans were washed with tap-water and cooked (100°C) in fresh tap-water at a ratio of 1:3 for 20 min, and cooled to room temperature within 20 min by evaporation of adhering water by spreading them on perforated trays. Sporangiospore suspension was obtained by scraping off the sporangia from pure slant cultures of Rhizopus microsporus var. microsporus LU 573 grown on malt extract agar (Oxoid, CM 59) for 7 d at 30°C and suspending them in sterile distilled water with 0·85 % NaCl and 0·1 % peptone. After inoculation of the cooked soyabeans with the sporangiospore suspension (1 %, v/w), the beans (450 g) were packed in hard-PVC, perforated boxes (205 × 90 × 45 mm) and incubated at 30°C for 72 h. The resulting fresh tempeh was cut into 1 cm dice and dehydrated for 6 h at 60°C, ground using a 1·0 mm screen and stored at − 20°C until use.
Tempeh was pre-digested as described earlierReference Kiers, Nout and Rombouts18. It was suspended in distilled water (5 g/30 ml) and incubated while stirring with 2 ml α-amylase solution consisting of 125 000 units/l α-amylase (A-1031; Sigma Chemical Co., St Louis, MO, USA), 1·5 g/l NaCl, 1·5 g/l K2HPO4, 0·5 g/l Na2CO3 (pH 7·0) for 30 min at 37°C. Next, the pH was adjusted to 4·0 using 5 m-HCl and the suspensions were incubated with 8 ml stomach-medium (0·1 g/l lipase (Rhizopus F-AP15; Amano Pharmaceuticals, Nagoyy, Japan), 0·125 g/l pepsin (P-6887; Sigma), 3·1 g/l NaCl, 1·1 g/l KCl, 0·6 g/l Na2CO3, 0·11 g/l CaCl2, pH 4·0) for 1 h at 37°C. The pH was then adjusted to 6·0 using solid NaHCO3. Finally, 10 ml of a 2 % pancreatic solution (20·0 g/l pancreatin (P-1750; Sigma), 5·0 g/l bile (B-3883; Sigma), 5·0 g/l NaCl, 0·68 g/l KH2PO4, 0·3 g/l Na2HPO4, 0·84 g/l NaHCO3) was added and the suspensions were incubated for 30 min at 37°C. After pre-digestion the slurry was diluted using distilled water to 6·5 % DM.
Part of the suspension was centrifuged at 3000 g for 15 min at 4°C. The supernatant was filtered through a filter aid (Celite®545, 22 140; Fluka, Buchs, Switzerland) to remove residual coarse particles and finally filtered through a 0·22 μm filter (Steritop SCGPT05RE; Millipore, Billerica, MA, USA).
Supernatant of pre-digested tempeh was fractionated into permeate and retentate using ultra-filtration through a spiral wound membrane with molecular weight cut-off of 5 kDa (S2K328; Koch, Wilmington, MA, USA) against distilled water. In order to obtain fractions having the same volume as that of the original supernatant, permeate and retentate volumes were adjusted using rotary vacuum evaporation at 40°C.
Analysis
DM content of pre-digested supernatant, permeate and retentate was determined by drying aliquots until constant weight and nitrogen content was measured using a NA2100 nitrogen analyser (Interscience, Breda, The Netherlands). After centrifugation (10 min, 1300 g), sodium, potassium and chloride concentrations were determined using an Electrolyte 4+ analyser (Nova Biomedical, Waltham, MA, USA) and osmolality was determined using a cryoscopic osmometer (Osmomat; Gonotec, Berlin, Germany).
Gel permeation chromatography was performed on a LC-10Ai HPLC (Shimadzu Benelux, 's-Hertogenbosch, The Netherlands) equipped with a Superdex Peptide column (17-5003-01; Pharmacia Biotech, Piscataway, NJ, USA) and elution at 30°C with 0·1 % (v/v) trifluoroacetic acid and 30 % (v/v) acetonitrile at 0·5 ml/min. The eluate was monitored using a UV detector at 200 nm. Calibration was performed using proteins and peptides ranging from 200 to 7000 Da.
High-performance size-exclusion chromatography was performed on a SP8800 HPLC (Spectra Physics, Mountain View, CA, USA) equipped with three columns (each 300 × 75 mm) of Bio-Gel TSK in series (40XL, 30XL and 20XL; Bio-Rad Labs, Hercules, CA, USA) in combination with a TSK guard column (40 × 6 mm) and elution at 30°C with 0·2 m-NaNO3 at 0·8 ml/min. The eluate was monitored using a refractive index detector. Calibration was performed using dextrans ranging from 180 Da to 500 kDa.
Net intestinal absorption
All procedures involving animal handling and testing were reviewed and approved by the Animal Care and Ethics Committee of the Animal Sciences Group Lelystad, The Netherlands.
Piglets (crossbred Yorkshire × (Large White × Landrace)) were weaned at 3 weeks of age. About 2 weeks after weaning biopsies from the duodenal mucosa were taken using a fiberscope (Olympus GIF XP10; Olympus, Hamburg, Germany) and receptor status was determinedReference Sellwood, Gibbons, Jones and Rutter19. Sixteen piglets that expressed the receptor (K88/F4) involved in binding of the ETEC strain were used in the experiments 3 weeks after weaning.
The small intestinal segment perfusion test was carried out essentially as described beforeReference Nabuur, Hoogendoorn, van Zijderveld and van der Klis20. Up to ten segments, with a cranial inflow and a caudal outflow tube and 20 cm long, were situated between 35 and 65 % of the total length of the small intestine.
At 15 min before the perfusion started, the odd-numbered segments were injected with 5 ml ETEC (5 × 109 colony forming units, O149:K91:K88ac, producing LT and STb) and the even-numbered segments with 5 ml PBS. In each piglet, pairs of segments (a non-infected and an adjacent ETEC-infected) were perfused with either saline (supplemented with 0·1 % glucose and 0·1 % casamino acids and serving as an internal control to determine the maximum response to the infection) or any of the tempeh products, using a Latin-square design. Each segment was perfused with 64 ml product over 8 h, by injecting 2 ml product every 15 min.
Three consecutive experiments were carried out. In experiment 1 the total pre-digested tempeh, its supernatant, saline (and another not relevant soyabean product) were tested in four piglets in a 4 × 4 Latin-square design. In experiment 2 saline, the supernatant of the pre-digested tempeh, the permeate and the retentate obtained after ultra-filtration were tested in eight piglets in a replicated 4 × 4 Latin-square design. In experiment 3 with four piglets saline, the supernatant of pre-digested tempeh, the supernatant of undigested tempeh (and another not relevant soyabean product) were tested in a 4 × 4 Latin-square design.
At the end of the small intestinal segment perfusion test the product remaining in the segments was blown out into the drainage bottles. The piglets were killed by injection of sodium pentobarbital (200 mg/kg body weight) and the segments were cut from the mesenterium and their length was measured.
Net fluid, sodium, chloride and solute absorption were calculated from the difference between the volume and concentration of inflow and outflow divided by the surface area (length × circumference) of each segment. Reduction in net fluid absorption upon ETEC infection was determined by subtracting net fluid absorption in ETEC-infected segments from net fluid absorption in non-infected segments perfused with the same perfusion fluid.
Statistics
Results of non-infected and ETEC-infected segments perfused with the same product were compared using the Student's paired t test. The effect of the perfusion fluid on net fluid absorption upon ETEC infection was analysed with ANOVA for a Latin-square design with piglet, pair of segment and treatment as model effects. Results on net fluid absorption and osmolality in experiment 2 were analysed using linear regression analysis. All statistics were performed with GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, CA, USA). Net absorption of fluid, DM, sodium, chloride and solutes are presented as means and their standard errors.
Results
Pre-digested tempeh and its supernatant
In experiment 1 ETEC infection in saline-perfused segments resulted in a decrease in net fluid absorption of more than 500 μl/cm2 compared with non-infected segments (Fig. 1). The reduction in net fluid absorption upon ETEC infection was significantly less when segments were perfused either with pre-digested tempeh or its supernatant. The difference in reduction in net fluid absorption upon ETEC infection between pre-digested tempeh (75 μl/cm2; 14 % compared with saline) and its supernatant (282 μl/cm2; 53 % compared with saline) was substantial but not significantly different.
Fig. 1 Fluid loss upon enterotoxigenic Escherichia coli (ETEC) infection after perfusion with saline, pre-digested tempeh and its supernatant. Values are means with their standard errors depicted by vertical bars. a,b Mean values with unlike superscript letters were significantly different (P < 0·05).
Tempeh supernatant, permeate and retentate
Ultra-filtration of the supernatant of pre-digested tempeh resulted in a permeate fraction containing only low molecular weight compounds and a retentate fraction containing all compounds >5 kDa and some residual low molecular weight compounds (Fig. 2). Similar chromatograms were observed for gel permeation chromatography protein/peptide analysis and areas under the curve showed that around 50 % of the total DM in supernatant of pre-digested tempeh consisted of proteinaceous or protein-derived compounds (nitrogen × 6·25) (Table 1). After ultra-filtration, the permeate contained the majority (approximately 65 %) of the initial nitrogen. The osmolality of the retentate was the lowest and about a third of the permeate and a fourth of the supernatant.
Fig. 2 High-performance size-exclusion chromatography elution patterns of supernatant of pre-digested tempeh (●) and the permeate (□) and retentate (○) fraction obtained after ultra-filtration. ↓ , Molecular weights of dextran standards.
Table 1 DM, nitrogen, sodium and chloride content and osmolality of the pre-digested supernatant and the permeate and retentate fractions obtained after ultra-filtration (Mean values with their standard errors)
In experiment 2 there was an inverse linear relationship between osmolality of the perfused fluids and net fluid absorption in non-infected segments (Fig. 3), with the net fluid absorption of retentate being the highest. Upon ETEC infection a significant decrease in net fluid absorption was observed for saline and the permeate (386 (sem 69) and 300 (sem 62) μl/cm2, respectively). The decrease in net fluid absorption upon ETEC infection and perfusion with either supernatant or retentate was not significant, being 125 (sem 37) and 140 (sem 42) μl/cm2, respectively. ETEC infection resulted in a decrease in net sodium and chloride absorption (Table 2). The reduction in sodium and chloride absorption was highest when perfused with saline and permeate and less when perfused with supernatant and retentate.
Fig. 3 Net fluid absorption in non-infected (□) and enterotoxigenic Escherichia coli-infected (■) segments perfused with saline and tempeh supernatant, permeate and retentate. Values are means with their standard errors depicted by bars. There was an inverse linear relationship between osmolality and net fluid absorption for non-infected segments (net fluid absorption = 1091 − 1·45 × osmolality; r 2 0·87).
Table 2 Average net absorption of sodium, chloride and solutes after perfusion with saline and tempeh supernatant, permeate and retentate (Mean values with their standard errors)
ETEC, enterotoxigenic Escherichia coli.
a–d Mean values within a column with unlike superscript letters were significantly different (P < 0·05).
Mean values were significantly different from those of the non-infected group: *P < 0·01; **P < 0·001.
Pre-digested and undigested tempeh supernatant
Pre-digestion of tempeh did not result in liberation of soluble carbohydrate polymers >8·5 kDa, whereas numerous low molecular weight compounds were formed (Fig. 4). Pre-digestion resulted in a marked increase in nitrogen levels (in low molecular weight compounds) and osmolality (Table 3).
Fig. 4 High-performance size-exclusion chromatography elution pattern of the supernatant of undigested (●) and pre-digested tempeh (○). ↓ , Molecular weight of dextran standards.
Table 3 Nitrogen content and osmolality of undigested and pre-digested supernatant of tempeh (Mean values with their standard errors)
The difference in net fluid absorption between non-infected and ETEC-infected segments in experiment 3 amounted to almost 400 μl/cm2 for perfusion with saline. Reduction in net fluid absorption upon ETEC infection was not different for undigested and for pre-digested tempeh supernatant (211 (sem 72) and 214 (sem 83) μl/cm2, respectively) and represented a reduction of 54 % compared with saline.
Discussion
Previously we demonstrated that pre-digested tempeh stimulates DM absorption and reduces ETEC-induced fluid and electrolyte losses in piglet small intestinal segments, whereas this was not the case for non-fermented soyabeansReference Kiers, Nout, Rombouts, van Andel, Nabuur and van der Meulen12. The protective effect of pre-digested tempeh against ETEC-induced fluid loss appears to be determined in part by the presence of the insoluble matrix of tempeh. The presence of insoluble material could probably modify fluid absorption as has been suggested for viscosity-enhancing agentsReference Go, Harper, Sia, Teichberg and Wapnir21, Reference Wapnir, Wingertzahn and Teichberg22. Whereas pre-digested tempeh as such reduced loss in net fluid absorption by about 85 % compared to saline, the supernatant of pre-digested tempeh reduced loss in net fluid absorption by about 50 %.
Although low osmolality promotes net fluid absorption in both non-infected and secreting intestineReference Kiers, Hoogendoorn, Nout, Rombouts, Nabuur and van der Meulen23–Reference Thillainayagam, Hunt and Farthing25, the difference between net fluid absorption in non-infected and ETEC-infected segments is fairly independent of osmolalityReference Kiers, Hoogendoorn, Nout, Rombouts, Nabuur and van der Meulen23. Therefore the lower decrease in net fluid absorption upon ETEC infection for perfusion with supernatant and retentate cannot be attributed to osmolality and we hypothesize that tempeh components interfere in the pathogenesis of the ETEC infection, resulting either in reduction of secretion or stimulation of absorption, or both.
The mechanisms whereby high molecular weight soluble tempeh compounds exert their beneficial effects in the case of the observed reduced fluid and electrolyte losses remain to be determined. These may involve aspects of enzyme activity, interference with pathogenicity factors and stimulation of fluid absorption.
Tempeh contains a diversity of microbial enzyme activities. α-Galactosidase and protease activity were found in the supernatant of pre-digested tempeh and in its retentate but not in its permeate (JL Kiers, unpublished results). Enzymatic activity in the gut could degrade intestinal receptors for ETEC as was shown before for bromelainReference Chandler and Mynott26–Reference Mynott, Guandalini, Raimondi and Fasano28. Components in the tempeh extract could also interfere with the (binding of) enteroxin, as was shown elsewhere for certain toxin receptor analoguesReference Takeda, Yoshino, Adachi, Sato and Yamagata29 or interfere in the cascade of events leading to an increased chloride secretion and inhibited sodium chloride absorption, as was shown for polyphenolic compounds in plant extracts and boiled riceReference Greenwood-van Meerveld, Tyler, Kuge and Ogata30–Reference Mathews, MacLeod, Zheng, Hanrahan, Bennett and Hamilton32.
Tempeh components could also exert a pro-absorptive activity. This could be mediated through stimulated sodium–solute co-transport which is the basis for traditional oral rehydration therapyReference Farthing24. In the present study there was a net absorption of solutes both in the tempeh supernatant as well as in the permeate and retentate. Probably this reflected the absorption of low molecular weight components such as monosaccharides and amino acids present in supernatant and permeate and in low concentrations in retentate. Since the reduction of fluid loss was found only in the retentate but not in the permeate, which contained higher levels of compounds supposed to enhance fluid absorption through sodium–solute co-transport, coupled sodium–solute absorption is unlikely to be responsible for the observed protective effect.
Vegetable polysaccharides, like starchReference Wapnir, Wingertzahn, Moyse and Teichberg33, Reference Wingertzahn, Teichberg and Wapnir34 and gum ArabicReference Turvill, Wapnir, Wingertzahn, Teichberg and Farthing35–Reference Wapnir, Teichberg, Go, Wingertzahn and Harper37, have been shown to enhance intestinal electrolyte and/or fluid absorption in normal and secreting rat small intestine. Gum Arabic has been shown to enhance net sodium absorption without altering net fluid absorption in normal rat jejunumReference Wapnir, Teichberg, Go, Wingertzahn and Harper37, whereas fluid absorption was increased in the case of diarrhoeaReference Turvill, Wapnir, Wingertzahn, Teichberg and Farthing35, Reference Wapnir, Wingertzahn, Moyse and Teichberg36. Enhanced absorption could be the result of increased accessibility of electrolytes and associated fluid to the microvillus membrane through the emulsifying properties of gum ArabicReference Wapnir, Wingertzahn, Moyse and Teichberg36, but also other physicochemical properties could play a roleReference Wapnir, Wingertzahn, Moyse and Teichberg33. Evidence for the possible involvement of tempeh polysaccharides in the protective effect described in the present study could be the presence of high molecular weight polysaccharides in the tempeh supernatant and its retentate. Additionally, the high-performance size-exclusion chromatography analysis of the supernatant of pre-digested tempeh was identical to the supernatant of undigested tempeh, especially with regard to high molecular weight compounds, and supernatant of undigested tempeh showed equal protection against ETEC-induced fluid loss as supernatant of pre-digested tempeh.
The present results warrant the protective role that tempeh could play in ETEC-associated diarrhoea and show the role of its high molecular weight soluble fraction. Further research is required to identify the component(s) in the high molecular weight soluble fraction which is responsible for the protective effect and to evaluate the specific mechanism(s) underlying the improved net fluid balance.
Acknowledgements
The authors thank Mirjana Eimermann, Gerrit de Vrij, Wendy Verwillegen (Numico Research BV, Wageningen), Gert-Jan de Graaf, Esther van Andel, Arie Hoogendoorn and Ank van Zijderveld-van Bemmel (Animal Sciences Group, Lelystad) for their excellent technical assistance. Financial support by the Dutch Ministry of Agriculture, Nature and Food Quality, as well as by Numico Research BV, Wageningen, The Netherlands is gratefully acknowledged.
