Modulation of anti-pathogenic activity in canine-derived Lactobacillus species by carbohydrate growth substrate
R.A. Rastall, School of Food Biosciences, The University of Reading, PO Box 226, Whiteknights, Reading RG6 6AP, UK (e-mail: firstname.lastname@example.org).
Aims: To investigate the effect of various carbon sources on the production of extracellular antagonistic compounds against two Escherichia coli strains and Salmonella enterica serotype Typhimurium by three canine-derived lactobacilli strains.
Methods and Materials: Cell-free preparations, pH neutralized, were used in antibiotic disc experiments as an initial screening. The bacteria/carbohydrate combinations that showed inhibition of the growth of those pathogens, were further investigated in batch co-culture experiments. The cell-free supernatants of the cultures, that decreased the population number of the pathogens in the co-culture experiments to log CFU ml−1 ≤ 4, were tested for inhibition of the pathogens in pure cultures at neutral and acidic pH.
Conclusions: The results showed that the substrate seems to affect the production of antimicrobial compounds and this effect could not just be ascribed to the ability of the bacteria to grow in the various carbon sources. L. mucosae, L. acidophilus and L. reuteri, when grown in sugar mixtures consisting of α-glucosides (Degree of Polymerization (DP) 1–4) could produce antimicrobial compounds active against all three pathogens in vitro. This effect could not be attributed to a single ingredient of those sugar mixtures and was synergistic. This inhibition had a dose-response characteristic and was more active at acidic pH.
Significance and Impact of the Study: Knowledge of the effect that the carbon source has on the production of antimicrobial compounds by gut-associated lactobacilli allows the rational design of prebiotic/probiotic combinations to combat gastrointestinal pathogens.
Numerous investigators have reported the ability of lactic acid bacteria (LAB) to produce antimicrobial substances active against certain pathogenic and spoilage organisms (Mehta et al. 1983) in various ecosystems resulting in a change of the bacterial population in their microenvironment (Olsen et al. 1995).
The LAB, especially the lactobacilli and bifidobacteria, have received great attention as means to maintain a healthy balance of the microflora in the large intestine (Abdel-Bar et al. 1987).
The defence mechanisms of the gastrointestinal tract provide an effective barrier against infection by pathogenic micro-organisms, but allow the establishment of a normal bacterial flora in the gut. The dynamic process that results in the occurrence or prevention of an infection has previously been described as the balance of pathogenic microbes, their specific virulence characteristics and the status of the host-defence mechanism (Duncan et al. 1999). The antagonistic effects are attributed to the decrease of pH and specific effects of the production of primary metabolites such as lactic acid and hydrogen peroxide and the secretion of specific bacteriocins produced by LAB during fermentation (Axelsson 1998).
Lactobacilli of human intestinal origin such as L. reuteri and L. acidophilus, have been shown to exhibit antagonistic activity against both Gram-positive and Gram-negative bacteria (Silva et al. 1987; Drago et al. 1997). Strains belonging to the L. reuteri group produce reuterin (Axelsson 1998), the production of which is stimulated when L. reuteri is co-cultured with other bacteria, e.g. E. coli (Talarico and Dobrogosz 1989).
Many strains belonging to the L. acidophilus group have been reported to produce antimicrobial compounds, which show a great variety regarding their inhibition spectrum (De Vuyst et al. 1996; Contreras et al. 1997; Zamfir et al. 1999).
There is, however, growing interest in the use of probiotic LAB in companion animals, particularly dogs, as their diet can be controlled. Very little is known of the gut microflora of dogs and, similarly, very little is known about the interactions of the dog commensal flora with invading gastrointestinal pathogens. Dogs have been suggested as being reservoirs of pathogenic bacteria such as salmonellae and Campylobacter spp. without expressing clinical symptoms (Baker et al. 1999). Syngre et al. (1993) suspected companion animals being reservoirs of infection for VTEC O157 and, because of the close associations that people have with those animals, suggested that they pose a risk by direct contact.
As the production of antimicrobial compounds by LAB is a result of their fermentation and is also highly affected by the growth medium composition, the aim of this study was to determine the effect of the type of carbon source on the production of antimicrobial substances by three canine-derived lactobacilli against a human pathogen Escherichia coli 0157:H7 (VT−) and two canine pathogens E. coli HE320 and Salmonella enterica serotype Typhimurium DT104.
Bacterial strains and culture conditions
All candidate probiotic Lactobacillus strains were obtained from the culture collection of the Waltham Center for Pet Nutrition (Waltham-on-the Wolds, Leicestershire, UK), having originated from canine large intestine biopsy samples taken from sedated healthy adult dogs using a previously described technique (Rolfe et al. 2002). The strains L. mucosae NCIMB 41149, L. acidophilus NCIMB 41085, L. reuteri NCIMB 41152, E. coli O157:H7 (VT−), E. coli HE 320 and S. enterica serotype Typhimurium DT 104 were maintained at −70°C. The working cultures of lactobacilli were maintained in MRS broth (Oxoid Basingstoke, Hampshire, UK) and MRS agar (MRS broth plus 1·5% agar) slabs. All pathogens were maintained in Mueller–Hinton broth (Oxoid) and Mueller–Hinton agar (Oxoid) stabs. All working cultures were stored at 4°C.
Actilight [fructo-oligosaccharides (FOS) of low degree of polymerization] was supplied by Eridania Beghin-Say (Vilvoorde, Belgium). Xylan, maltose, glucose, melezitose, sucrose, palatinose, lactose, melibiose, cellobiose, raffinose, laevan, tagatose, stachyose and gentiobiose were obtained from Sigma (Poole, Dorset, UK). Isomalto-oligosaccharides (IMO) were supplied by Showa Sangyo (Tokyo, Japan) and xylo-oligosaccharides (XOS) by Suntory (Tokyo, Japan). Panorich® and Biotose® were supplied by Nihon Shokuhin Kako Co., Ltd (Tokyo, Japan). Panorich®, a high panose syrup, is made from corn starch by hydrolysis and transglucosylation reactions with β-amylase and transglucosidase. The standard composition of Panorich® is (w/w): 30% panose, 23% glucose, 17% maltose, 16% branched oligosaccharides (DP ≥ 4), 9% isomaltose, 3% maltotriose and 2% isomaltotriose. Biotose® is a starch sweetener consisting of 41% glucose, 20% isomaltose, 9% isomaltotriose, 9% panose, 7% maltose and 14% others.
In order to increase the oligosaccharide fraction of Panorich®, a GyrosepTM 300 stirred cell (Techmate Ltd, Milton Keynes, UK) was modified to permit the use of pressures up to 50 bar. Flat sheet membranes with an effective membrane area of 40 cm2 were employed. A polytetrafluoroethylene-coated magnetic stirrer bar was centrally positioned gripped on a stainless steel bar supported on the top plate, and the stirring rate was adjusted so that the depth of the vortex was no more than one-third of the stirred solution level. Reverse stirring was also applied to avoid the creation of lamina flow. The pressure source was a nitrogen gas pressure cylinder.
The membranes used were supplied by Intersep Ltd (Wokingham, UK). One nanofiltration membrane (NF-TFC-50 thin film composite) was used for three runs, and one ultrafiltration membrane with low nominal molecular weight cut off (UF-CA-1 cellulose acetate, MW cut off 1000) was used for one. All these membranes had an integral porous support, but a further support was used when placing the membranes into the cell in order to maximize filtration efficiency. The final permeate (Panorich*) had the same sugar composition to Panorich® but higher concentration of the oligosaccharide fraction (2·96% glucose, 5·43% of disaccharides and 91·6% of oligosaccharides with DP ≥ 3).
Bacterial growth was measured with an automatic turbidometer, the Bioscreen C system (Labsystems, Helsinki, Finland), which records kinetic changes in the absorbance of liquid samples in a multiwell plate. Each well of the plate was filled with 200 μl glucose-free MRS medium containing one of the carbohydrate substrates (1%, w/w) and inoculated with 50 μl of bacterial cell suspension (105 CFU ml−1). As it is not possible to generate anaerobic conditions in the Bioscreen system, the strains were transferred to the multiwell plate inside an anaerobic cabinet (10 : 10 : 80; H2 : CO2 : N2). Access of oxygen to the culture was minimized by closing the gap between the cover and the bottom part of the plate with adhesive tape. All strains were incubated for 24 h at 37°C in triplicates in two sets of experiments.
The rate of bacterial growth on a single carbohydrate source was determined by calculating the slope of the growth curve (h−1). The readings from the Bioscreen C system were converted to CFU ml−1 by means of calibration curves (R2 = 0·97).
Antibiotic disc assay for antimicrobial activity
For the initial screening for the production of any antimicrobial substances, the test organisms were grown in MRS broth containing one of the growth substrates (1%, w/w). After 24 h of incubation at 37°C in an anaerobic cabinet (10 : 10 : 80; H2 : CO2 : N2), cells were removed by centrifugation (30 000 g for 30 min) and the cell-free supernatants were neutralized with 1 m NaOH to pH 6·8.
Agar plates were carefully overlaid with 5 ml of Mueller–Hinton soft agar (0·7%, w/v) seeded with 0·5 ml of a 24 h culture of the respective pathogen. Filter paper discs (0·6 cm diameter) soaked in the cell-free supernatant of the test organisms, were then added to the soft agar surface. A maximum of four filter paper discs, spaced ca 3 cm apart were placed per plate. Blank media were used as controls. Plates were then incubated for 24 h at 37°C. After incubation, the degree of inhibition was measured as the diameter of the clear zones around the paper discs.
Batch culture assay for antimicrobial activity
The culture medium at its final concentration contained (g l−1): peptone water, 2·0; yeast extract, 2·0; NaCl, 0·1; K2HPO4, 0·04; KH2PO4, 0·04; MgSO4·7H2O, 0·01; CaCl2·2H2O, 0·01; NaHCO3, 2·0; cysteine HCl, 0·5; bile salts, 0·5. The following liquid additions were made (ml l−1): Tween 80, 2; vitamin K1, 0·2; haemin solution, 1. The pH of the media was adjusted to 7 and autoclaved at 121°C for 15 min. The carbohydrate substrate was filter sterilized (0·02 μm pore size) and added to the medium after autoclaving to give a final concentration of 10 g l−1. Batch culture fermentation systems (100 ml working volume) were inoculated with 1 ml of 24 h cultures of the test organism and 1 ml of 24 h culture of E. coli HE320 or E. coli O157:H7 (VT−) or S. enterica serotype Typhimurium DT104. Experiments were carried out in triplicate for 24 h at 37°C in an anaerobic cabinet (10 : 10 : 80; H2 : CO2 : N2). Samples were removed after 0, 4, 8, 12 and 24 h from each fermentation vessel for the enumeration of viable bacteria and measurement of the pH. Prior to the co-culture experiments, growth of each of the bacteria used was separately tested in the growth medium given above.
Liquid samples (1 ml) were serially diluted in peptone water (50% w/v) supplemented with 0·5 g l−1 cysteine HCl (pH 7) in an anaerobic cabinet. Plates were then inoculated from each dilution (up to 10−8) in triplicate. Growth medium (Oxoid Ltd, Basingstoke, Hampshire, UK) used for the enumeration of lactobacilli was Rogosa, for E. coli was MacConkey No. 3 and of S. enterica serotype Typhimurium was SS agar. Rogosa plates were incubated for 64 h at 37°C under anaerobic conditions, while MacConkey No. 3 and SS plates were incubated for 24 h at 37°C.
Serum tube assay for antimicrobial activity
Lactobacillus mucosae was grown in the culture medium, described before, containing Panorich® (1% w/v) and L. acidophilus and L. reuteri were grown in the same medium supplemented with Biotose® (1% w/v). The cultures were incubated for 24 h under anaerobic conditions. Bacterial cells were then separated by centrifugation (30 000 g for 30 min). Half of the resultant supernatant was adjusted to pH 6·8 with 1 m NaOH while the other half was kept at its original pH and sterilized by filtration (0·2 μm pore size). Supernatant fluid (100–250 μl) was then added to a range of serum tubes, which contained 5 ml of the same growth medium. In control experiments sterile media replaced the supernatant. Samples (0·05 ml) of growing cultures (in Tryptone Soya Broth medium) of E. coli 0157 : H7 (VT−), E. coli HE320 and S. enterica serotype Typhimurium were then inoculated in triplicate to separate tubes. Bacterial growth was determined by monitoring A550 changes during incubation at 37°C.
All analyses were carried out using paired t-test, assuming equal variances and considering both sides of the distribution (two-tailed distribution). Differences were considered significant at P ≤ 0·05.
Growth of bacteria on various carbohydrates
The growth of L. mucosae, L. acidophilus or L. reuteri with various carbohydrates as carbon source was studied for 24 h under anaerobic conditions as an initial screening for the substrates to be used in studies on induction of antimicrobial activity. Although the results of the Bioscreen C system do not give the actual growth rates of those cultures (Table 1), because of the small volume of the culture, they were used as an indication of the preferences of the strains of interest regarding the type of carbon source. All four strains showed ability to grow in α-glucosides (maltose, Biotose®, Panorich®, IMO) and galacto-oligosaccharides (lactose, raffinose, melibiose). Lactobacillus reuteri was the only one to grow well in FOS (Actilight).
Table 1. Growth rate and survival fraction of the lactobacilli on various carbohydrates
|Actilight||2·76 ± 0·27f||5·58 ± 0·24b||9·90 ± 0·28e|
|Biotose®||8·10 ± 0·21c||11·70 ± 0·67d||6·54 ± 0·30c|
|Cellobiose||7·62 ± 0·15c||11·46 ± 0·63d||1·26 ± 0·19k|
|Gentiobiose||5·88 ± 0·16b||4·74 ± 0·45a||10·26 ± 0·25e|
|Glucose||8·40 ± 0·24c||12·06 ± 0·49d||6·90 ± 0·29c|
|IMO||6·12 ± 0·19b||4·20 ± 0·69a||5·04 ± 0·21a|
|Lactose||8·28 ± 0·30c||10·26 ± 0·55e||7·26 ± 0·29c|
|Laevan||0·06 ± 0·02||0·06 ± 0·03||1·68 ± 0·17g|
|Maltose||6·96 ± 0·12b||10·20 ± 0·37e||5·88 ± 0·26a|
|Melezitose||3·24 ± 0·17°||0·78 ± 0·17h||1·86 ± 0·20g|
|Melibiose||7·98 ± 0·24c||11·34 ± 0·42d||4·74 ± 0·25a|
|Palatinose||9·36 ± 0·17e||9·48 ± 0·50e||9·06 ± 0·20e|
|Panorich®||8·28 ± 0·19c||8·34 ± 0·34c||5·82 ± 0·11b|
|Raffinose||7·68 ± 0·20c||10·86 ± 0·46d||7·02 ± 0·27c|
|Stachyose||6·66 ± 0·19b||7·44 ± 0·32c||5·76 ± 0·33b|
|Sucrose||6·84 ± 0·23b||10·32 ± 0·48e||5·64 ± 0·27b|
|Tagatose||1·92 ± 0·27g||0·06 ± 0·03||3·84 ± 0·23a|
|XOS||3·84 ± 0·18a||4·14 ± 0·19a||7·50 ± 0·33c|
|Xylan||0·06 ± 0·03||0·96 ± 0·17k||4·44 ± 0·25a|
Antibiotic disc assay for production of antimicrobial activity
Antibiotic discs soaked in cell-free extract of the Lactobacillus cultures indicated that the supernatant fluid exerted an inhibitory effect toward E. coli HE320, E. coli O157 : H7 VT− and S. enterica serotype Typhimurium DT 104 (Table 2). The inhibitory effect could neither be attributed to competition for growth substrates (as the supernatant fluid was cell free) nor an acidic environment (the extract pH was adjusted to 6·8). Interestingly the inhibitory activity did not correlate with the carbohydrates that supported the highest growth rates. Lactobacillus mucosae showed inhibitory effects when grown in gentiobiose and Panorich® (against all pathogens), IMO and maltose (against E. coli) and raffinose (against E. coli O157 : H7 VT− and S. enterica serotype Typhimurium). Lactobacillus acidophilus and L. reuteri inhibited all three pathogens when grown in Biotose®.
Table 2. Inhibition of pathogenic micro-organisms by lactobacilli grown on various carbon sources
|Actilight||LG||LG||–||LG||LG||6·5 ± 0·5§||LG||LG||–|
|Biotose®||LG||7·0 ± 0·3§||4·2 ± 0·5*||LG||5·3 ± 0·4*||3·0 ± 0·2°||LG||6·1 ± 0·5§||7·1 ± 0·4§|
|Cellobiose||–||–||LG||–||–||LG||3·8 ± 0·5°||–||NA|
|Gentiobiose||6·1 ± 0·6§||–||–||5·0 ± 0·4*||–||–||6·0 ± 0·3§||3·2 ± 0·3°||3·2 ± 0·4|
|IMO||1·4 ± 0·5||4·2 ± 0·5*||4·2 ± 0·6*||4·6 ± 0·4*||–||–||–||1·7 ± 0·4||4·5 ± 0·9*|
|Lactose||–||–||–||–||–||6·2 ± 0·5§||–||4·3 ± 0·5*||5·9 ± 0·9§|
|Maltose||3·8 ± 0·7°||–||6·8 ± 0·8§||3·4 ± 0·5°||–||1·3 ± 0·4||–||–||6·5 ± 0·7§|
|Melibiose||–||–||–||–||–||5·6 ± 0·5*||–||–||–|
|Palatinose||–||–||–||–||–||–||1·3 ± 0·4||1·0 ± 0·5||–|
|Panorich®||6·2 ± 0·4§||–||–||5·0 ± 0·3*||–||–||9·0 ± 0·4α||–||–|
|Raffinose||6·8 ± 0·5§||–||–||–||–||5·8 ± 0·7§||3·4 ± 0·5°||1·9 ± 0·5||–|
|Sucrose||–||1·3 ± 0·5||–||–||–||–||–||3·5 ± 0·4°||–|
|XOS||LG||LG||–||LG||LG||3·7 ± 0·5||LG||LG||–|
Antimicrobial activity in batch co-cultures
Each batch culture fermenter was inoculated with pure overnight cultures of one of the Lactobacillus strains and one of the pathogens. After 24 h of incubation the substrates that induced an inhibitory effect against the pathogens were: Panorich® and maltose for L. mucosae (Table 3), Biotose® for L. acidophilus (Table 4) and L. reuteri and glucose, maltose for L. reuteri (Table 5). In the cases of L. mucosae + Panorich®, L. acidophilus + Biotose® and L. reuteri + Biotose® the number of pathogens had been reduced to log CFU ml−1 < 4, while the number of lactobacilli were at least equal to the number inoculated. In all three cases the decrease in the number of pathogens could be observed after 12 h of incubation, when the potential probiotic strains had reached their stationary phase and cannot be attributed to the acidic environment as the pH had stabilized after 8 h of incubation (Fig. 1). In all other experiments the number of lactobacilli increased during 24 h of fermentation, except for L. mucosae and L. reuteri when grown on maltose, in which cases there were similar, although less severe reduction in the number of pathogens (Tables 3, 4 and 5).
Table 3. Changes in the bacterial population (log CFU ml−1) in co-culture experiments after 24 h of incubation under anaerobic conditions with Lactobacillus mucosae as the test organism (results are mean values of three replicates ± s.d.)
|Panorich||0·02 ± 0·01|| || ||−2·70 ± 0·15*||4·47|
|2·42 ± 0·13α||−4·70 ± 0·27§|| || ||4·00|
|2·39 ± 0·14α|| ||−4·60 ± 0·23§|| ||4·36|
|Glucose||1·07 ± 0·09β|| || ||1·60 ± 0·10α||3·60|
|1·25 ± 0·10β|| ||1·51 ± 0·14α|| ||3·54|
|1·00 ± 0·09β||1·04 ± 0·06β|| || ||3·68|
|IMO||1·73 ± 0·12c|| || ||1·80 ± 0·13c||4·10|
|1·74 ± 0·11c||1·27 ± 0·08β|| || ||4·06|
|1·93 ± 0·12α|| ||1·79 ± 0·13c|| ||3·95|
|Maltose||1·17 ± 0·09β|| || ||−1·89 ± 0·12f||3·60|
|0·56 ± 0·05w||−2·67 ± 0·16*|| || ||3·40|
|0·56 ± 0·04w|| ||−1·23 ± 0·09w|| ||3·70|
|Panorich*||1·20 ± 0·10 β|| || ||2·10 ± 0·16α||4·30|
|1·50 ± 0·13c||1·57 ± 0·10c|| || ||4·35|
|1·03 ± 0·09β|| ||−0·04 ± 0·03|| ||4·50|
|Gentioiose||1·90 ± 0·15α|| || ||2·72 ± 0·16*||4·60|
|2·11 ± 0·17α||2·69 ± 0·19*|| || ||4·60|
|Cellobiose||0·85 ± 0·06f|| || ||1·02 ± 0·08β||5·50|
Table 4. Changes in the bacterial population (log CFU ml−1) in co-culture experiments after 24 h of incubation under anaerobic conditions with L. acidophilus as the test organism (results are mean values of three replicates ±s.d.)
|Glucose||2·30 ± 0·18α|| || ||0·38 ± 0·02*||3·10|
|2·43 ± 0·18α||0·90 ± 0·06w|| || ||3·18|
|1·88 ± 0·13β|| ||0·81 ± 0·04w|| ||3·10|
|Maltose||3·37 ± 0·21c|| || ||0·38 ± 0·03*||3·90|
|3·37 ± 0·20c||0·90 ± 0·05w|| || ||3·95|
|2·80 ± 0·18f|| ||0·81 ± 0·06w|| ||3·90|
|Panorich*||0·30 ± 0·07*|| || ||0·27 ± 001*||4·80|
|0·21 ± 0·02*||0·74 ± 0·03w|| || ||4·67|
|0·42 ± 0·06*|| ||−0·01 ± 0·01|| ||3·98|
|Biotose||1·50 ± 0·10β|| || ||−3·30 ± 0·19§||3·90|
|0·30 ± 0·02*||−3·10 ± 0·16§|| || ||3·90|
|0·70 ± 0·05w|| ||−4·00 ± 0·18°|| ||4·00|
Table 5. Changes in the bacterial population (log CFU ml−1) in co-culture experiments after 24 h of incubation under anaerobic conditions with L. reuteri as the test organism (results are mean values of three replicates ± s.d.)
|Glucose||1·56 ± 0·02α|| || ||−0·38 ± 0·03§||3·60|
|1·45 ± 0·10α||−1·31 ± 0·08w|| || ||3·80|
|2·05 ± 0·12β|| ||−0·32 ± 0·04§|| ||3·90|
|IMO||2·48 ± 0·15β|| || ||1·59 ± 0·08α||4·93|
|2·78 ± 0·13β||1·48 ± 0·05α|| || ||3·80|
|3·73 ± 0·13c|| ||1·63 ± 0·10α|| ||4·50|
|Maltose||3·86 ± 0·18c|| || ||−0·70 ± 0·05#||3·70|
|3·51 ± 0·09c||−1·66 ± 0·03w|| || ||3·60|
|3·91 ± 0·12c|| ||−0·54 ± 0·07#|| ||3·60|
|Panorich*||0·30 ± 0·06d|| || ||3·20 ± 0·13*||4·50|
|0·22 ± 0·03d||2·10 ± 0·08β|| || ||4·30|
|0·66 ± 0·05°|| ||0·73 ± 0·07°|| ||4·20|
|Gentioiose||4·01 ± 0·21c|| ||1·63 ± 0·09α|| ||4·50|
|Biotose||0·60 ± 0·03°|| || ||−3·30 ± 0·16*||4·10|
|0·90 ± 0·04f||−3·40 ± 0·11*|| || ||4·10|
|1·10 ± 0·04f|| ||−3·20 ± 0·15*|| ||4·10|
Because Panorich® and Biotose® are mixtures of carbohydrates, further investigations were carried out to identify the active components. Three different patterns could be observed in the data depending on the lactobacillus strain used. When L. mucosae (Table 4) was grown on glucose, IMO and Panorich* all of the pathogens increased in population number to greater extent than did the L. mucosae. Only with maltose was a decrease in the pathogen population number obvious while with Panorich* S. enterica serotype Typhimurium did not show any change in the population number. Lactobacillus acidophilus (Table 5) increased in population to a greater extent than did any of the pathogens, but at the same time no pathogen inhibition could be observed. Lactobacillus reuteri (Table 5) was the only one of the strains tested that could inhibit the growth of the pathogens in glucose and maltose, although to a lesser extent than when grown on Biotose®.
Antimicrobial activity of Lactobacillus culture supernatants
In all the experiments inhibition on the growth of the pathogens can be observed with supernatants from each of the Lactobacillus strains. This inhibition displayed a clear dose-response relationship with the amount of supernatant added and was particularly marked when the supernatants were not adjusted to pH 6·8.
The most active culture supernatant was obtained when Biotose® was used as the substrate, displaying maximum inhibition of all of the pathogens when 200 μl of supernatant were added, regardless the pH value (Figs 2 and 3), and especially when L. reuteri was used as the producer organism (Fig. 2). Culture supernatant from L. mucosae grown on Panorich® displayed the lowest inhibitory activity (Fig. 4) compared with the other supernatants tested.
The method by which LAB are thought to be inhibitory to pathogens is, in part, related to the production of organic acids as part of their normal metabolic processes (Yazawa and Tamura 1982). This mechanism assumes that the antimicrobial agent is the undissociated acid, which reaches increasingly higher proportions as the pH decreases (Eklund 1983). The results of this study indicate that acidity may not be the sole inhibitory agent. First, an inhibitory effect was not observed in every one of the batch cultures even though the pH value of the cultures were similar (Tables 3, 4 and 5). Furthermore, experiments in which pH neutralized supernatant fluids were taken from overnight cultures of the three Lactobacillus strains (L. mucosae + Panorich®, L. acidophilus + Biotose®, L. reuteri + Biotose®) and added to pure cultures of the three pathogenic bacteria confirmed this hypothesis (Figs 2, 3 and 4). It was also noticed that in the acidic environment the three pathogenic bacteria were much more sensitive to the antimicrobial agent produced by the Lactobacillus strains (Figs 2, 3 and 4), which agrees with the mode of action of most LAB bacteriocins.
For all three lactobacilli the inhibitory effect was induced by the presence of malto-oligosaccharide mixtures (Panorich®, Biotose®) in the growth medium. Lactobacillus mucosae showed similar inhibitory activity in the presence of maltose, but not in the presence of glucose, IMO, gentiobiose and cellobiose, suggesting that the α-1,4 linkage of glucose in malto-oligosaccharides may be important in the production of the antimicrobial agents (Table 3). Lactobacillus reuteri was the only one showing antimicrobial activity in the presence of glucose (Table 5), which is necessary for the production of reuterin (Rasch et al. 2002). Otherwise L. mucosae followed a similar pattern of inhibitory activity as L. reuteri.
Although no antimicrobial compound from lactobacilli in the intestinal environment has yet been characterised, these bacteria are part of the normal intestinal microflora that exerts a strong effect on the health of their hosts (Huis in't Veld and Havenaar 1997) by enhancing host resistance to bacterial and viral infections.
We have obtained consistent data indicating the ability of L. mucosae, L. acidophilus and L. reuteri, isolated from the canine intestine, to inhibit the growth of three pathogens, one of human origin (E. coli 0157 : H7 VT−) and two of canine origin (E. coli HE320 and S. enterica serotype Typhimurium) in vitro. This effect could not be attributed solely to the acidic environment of the culture and seemed to be regulated by the carbon source used, with some degree of synergy with respect to the sugar ingredients of the substrate mixtures. Further work needs to be carried out in order to characterize the active components produced by these Lactobacillus strains, but the fact that they are gut associated, allows us to speculate that they may be useful to combat gastrointestinal pathogens.