OUP user menu

In vitro effects of synbiotic fermentation on the canine faecal microbiota

Eva Ogué-Bon, Christina Khoo, Anne L. McCartney, Glenn R. Gibson, Robert A. Rastall
DOI: http://dx.doi.org/10.1111/j.1574-6941.2010.00915.x 587-600 First published online: 1 September 2010


Stirred, pH-controlled anaerobic batch cultures were used to investigate the in vitro effects of galacto-oligosaccharides (GOS) alone or combined with the probiotic Bifidobacterium bifidum 02 450B on the canine faecal microbiota of three different donors. GOS supported the growth of B. bifidum 02 450B throughout the fermentation. Quantitative analysis of bacterial populations by FISH revealed significant increases in Bifidobacterium spp. counts (Bif164) and a concomitant decrease in Clostridium histolyticum counts (Chis150) in the synbiotic-containing vessels compared with the controls and GOS vessels. Vessels containing probiotic alone displayed a transient increase in Bifidobacterium spp. and a transient decrease in Bacteroides spp. Denaturing gradient gel electrophoresis analysis showed that GOS elicited similar alterations in the microbial profiles of the three in vitro runs. However, the synbiotic did not alter the microbial diversity of the three runs to the same extent as GOS alone. Nested PCR using universal primers, followed by bifidobacterial-specific primers illustrated low bifidobacterial diversity in dogs, which did not change drastically during the in vitro fermentation. This study illustrates that the canine faecal microbiota can be modulated in vitro by GOS supplementation and that GOS can sustain the growth of B. bifidum 02 450B in a synbiotic combination.

  • batch cultures
  • galacto-oligosaccharides
  • Bifidobacterium bifidum
  • synbiotic


The effects of supplementing food products with probiotics and prebiotics have been studied extensively in humans, and recently, have become an important area of research in companion animal nutrition. The majority of studies in dogs have concentrated on the prebiotic effects of inulin and/or fructo-oligosaccharides (FOS) (Swanson et al., 2002a, b; Vanhoutte et al., 2005; Verlinden et al., 2006). These carbohydrates have been shown to selectively stimulate the growth of indigenous bacteria such as Lactobacillus spp. and Bifidobacterium spp. and also to increase short-chain fatty acid (SCFA) concentrations, while decreasing the production of putrefactive components (Swanson et al., 2002b; Flickinger et al., 2003a; Propst et al., 2003). Lactobacilli and bifidobacteria are believed to be beneficial in humans, although their impact on the health status of dogs is less well established. The prebiotic effect of galacto-oligosaccharides (GOS) has also been studied in humans and pigs (Rycroft et al., 2001; Smiricky-Tjardes et al., 2003; Tzortzis et al., 2005), but the potential of GOS as a canine prebiotic has not been explored extensively. Tzortzis et al. (2004) evaluated, in batch culture systems, the in vitro fermentation properties of a synthesized mixture of GOS (α-linked, as opposed to the β-linked GOS developed commercially) alone or within synbiotic combinations on the canine faecal microbiota. The novel mixture induced an increase in the concentrations of bifidobacteria when combined with either Lactobacillus acidophilus or Lactobacillus reuteri; however, the increase was higher with the GOS+L. reuteri synbiotic. A decrease in the Clostridium clusters I and II was also observed with GOS and with GOS+L. reuteri after 24 h of fermentation.

Controlled feeding studies in dogs are the method of choice by which to determine the effects of potential canine prebiotics, probiotics or synbiotics. Because of the lack of information on the canine gut microbiota, a number of issues should be considered before feeding animals with prebiotics, probiotics or synbiotics. For example, the doses to be administered and specific animal characteristics, as well as economical considerations, should be taken into account before feeding trials are commenced (Flickinger et al., 2003b). Although the results obtained from in vitro studies do not always correspond to those obtained in feeding trials, in vitro studies offer the advantage of frequent sampling and help to develop a clearer picture of the canine intestinal microbiota and the potential effects of prebiotics, probiotics and synbiotics on its composition. By definition, a synbiotic should manage the gut microbiota towards a healthier composition by enhancing the growth and activity of the probiotic with the prebiotic (Gibson & Roberfroid, 1995). Only two synbiotic combinations (FOS+L. acidophilus and galactosyl melibiose mixture+L. reuteri) have been evaluated as synbiotics for dogs (Swanson et al., 2002a; Tzortzis et al., 2004). The aim of the current study was to evaluate one synbiotic combination for its ability to enhance the growth of the probiotic and to modulate the canine intestinal microbiota in in vitro batch culture systems.

Materials and methods

Unless otherwise stated, all chemicals were supplied by Sigma Aldrich (UK).

Bacterial strains

Four commercially available probiotic lactobacilli [Lactobacillus plantarum 115 400B, L. acidophilus 14 150B (Danisco Inc.; New Century), L. acidophilus and Lactobacillus rhamnosus (PrimaLac Microbials; Clarksdale, Kansas)] and two commercially available probiotic bifidobacteria [Bifidobacterium longum 05 and Bifidobacterium bifidum 02 450B (Danisco Inc.)] were investigated. Probiotics were chosen on the basis of established commercial application and availability for subsequent in vivo studies. All test strains were stored on Cryobank cryogenic beads (Prolab Diagnostics, UK) at −70 °C for long-term storage.


Three commercially available carbohydrates with purported prebiotic properties were investigated: short-chain FOS [scFOS Nutraflora®, 95% scFOS, degree of polymerization (DP) 3–5; GTC Nutrition Co., CO], inulin (Inulin IPS Raftifeed®, 90–95% inulin, average DP=23; Beneo-Orafti, Belgium) and GOS (Bi2muno, 57% of mono and disaccharides, 25% of DP=2 trans-GOS, 18% of DP≥3; Clasado, UK). Glucose (BDH, UK) was used as a nonselective control.

Determination of growth characteristics with different carbohydrates

Probiotic growth on each test carbohydrate was measured using cultivation techniques. Lactobacilli were grown in 10 mL of glucose-free De Man, Rogosa and Sharpe (MRS) medium (De Man et al., 1960), while bifidobacteria were grown in 10 mL of glucose-free WC medium (Wilkins & Chalgren, 1976) containing one of the carbohydrate substrates (1% w/v). A control with 1% glucose (w/v) was also included. The 10 mL carbohydrate medium was inoculated with 100 μL of bacterial suspension in the late exponential phase (OD∼0.8–0.9) and incubated for 24 h at 37 °C under anaerobic conditions (H2 10%; CO2 10%; N2 80%; MACS 1000 Anaerobic cabinet, Don Whitley Scientific, UK). Serial dilutions were performed in phosphate-buffered saline (PBS) solution (0.1 M, pH 7.4; Oxoid, UK) and 20 μL aliquots of each dilution were plated onto MRS agar or WC agar to determine live bacterial counts after 0, 4, 8, 12 and 24 h of incubation. All experiments were carried out in triplicate on three different days. Specific growth rates were calculated using the following equation: μ=((ln Xt2−ln Xt1)/t2t1), where ln Xt1 and ln Xt2 are the number of CFU at time 1 (t1) and time 2 (t2), respectively, during the exponential growth phase and μ is the specific growth rate (h−1). Bifidobacterium bifidum 02 450B was selected based on the high growth on the GOS material.

Probiotic strain and selection of a rifampicin-resistant variant

To track B. bifidum 02 450B in mixed culture, rifampicin-resistant variants were selected by successive overnight anaerobic incubation in WC broth with increasing rifampicin concentrations (from 0.0001 to 100 μg mL−1) as described by Saulnier et al. (2008). The variants were tested for fermentation of GOS at 1% w/v as described before.

Donor screening for rifampicin-resistant faecal bifidobacteria

Adult Golden Retriever dogs were screened for rifampicin resistance in their faecal bifidobacterial and lactobacilli populations, in order to avoid masking from indigenous species. Faecal samples from 10 healthy dogs were collected on site, kept in an anaerobic cabinet and processed within 10 min of collection. A 1/10 w/v dilution in PBS was prepared and the samples were homogenized in a stomacher (Seward, UK) for 2 min at normal speed. Aliquots (20 μL) of this preparation were plated onto Rogosa agar (Rogosa et al., 1951) and Beerens agar (Beerens, 1990) containing 100 μg mL−1 of rifampicin for lactobacilli and bifidobacteria, respectively. Plates were incubated in the anaerobic cabinet at 37 °C for 48 h. Dogs were excluded as donors when growth occurred. Three dogs were selected and fed with a standard commercial dog food (Table 1) for 20 days before faecal sample donation for the in vitro fermentation runs. Dogs were free of any known metabolic and gastrointestinal diseases, were not taking probiotic or prebiotic supplements and had not taken antibiotics 6 months before the faecal samples were collected.

View this table:

Standard diet composition

Hill's Science Plan, adult, large breed, chicken
Chicken (minimum chicken 25%; minimum chicken and turkey combined 33%)
Ground maize
Chicken and turkey meal
Soybean meal
Animal fat
Maize gluten meal
Vegetable oil
Dried whole egg
l-Carnitine supplement
Potassium chloride
Naturally preserved with mixed tocopherols, citric acid and rosemary extract

Stirred, pH-controlled, anaerobic faecal batch cultures

Probiotic B. bifidum 02 450B, prebiotic GOS and their synbiotic combination were evaluated in anaerobic 24-h stirred, pH-controlled, batch cultures inoculated with canine faecal material. A negative control (NC) containing neither the probiotic nor the prebiotic was also included. Four batch fermenters were therefore run in parallel. The experiment was performed in triplicate, using one faecal sample from a different donor for each run.

Sterile basal medium (135 mL) was aseptically added to each fermenter (300 mL working volume). Basal medium comprised (g L−1 unless stated otherwise) peptone water (Oxoid) 2, yeast extract (Oxoid) 2, NaCl (BDH) 0.1, K2HPO4 (BDH) 0.04, KH2PO4 (BDH) 0.04, MgSO4·7H2O (BDH) 0.01, CaCl2·2H2O (BDH) 0.01, NaHCO3 (BDH) 2, l-cysteine hydrochloride 2, bile salts (Oxoid) 0.5, vitamin K 10 μL, Tween 80 (BDH) 2 mL and haemin 0.0005. Resazurin (0.001 g L−1) was added as an indicator of anaerobiosis. The fermenters were sparged overnight with O2-free N2 at a rate of 15 mL min−1. Each vessel was then inoculated with 15 mL of a fresh 10% w/v faecal homogenate, prepared as described above. Culture pH was maintained at 6.8 (±0.1) using automated pH controllers (Electrolab Ltd, UK) and the addition of 0.5 M NaOH or HCl, and temperature was maintained at 39 °C using a circulating water bath. The contents of the vessels were agitated constantly using a magnetic flea and stirrer. GOS was added to the basal medium just before addition of the faecal homogenate to yield a final concentration of 1% w/v and was the sole carbon source available during the 24-h fermentation period. In vessels inoculated with probiotic, 1 mL (∼109 cells) of the rifampicin-variant strain was added. To prepare this culture, the B. bifidum 02 450B was grown anaerobically overnight, in 10 mL of WC or MRS broth containing 100 μg rifampicin mL−1. Before addition, 1 mL of the probiotic culture was centrifuged for 5 min at 3000 g. The cells were then washed twice in PBS, centrifuged for 5 min at 3000 g and resuspended in 1 mL of PBS. In addition, 1 mL of the probiotic culture was serially diluted and plated onto Beerens agar containing 100 μg rifampicin mL−1 to determine the initial inoculum.

Sampling and enumeration of the probiotic rifampicin variant

Samples were obtained after 0, 5, 10 and 24 h of batch culture fermentation. Portions (5 mL) of culture fluid were removed from each vessel and prepared for FISH analysis and DNA extraction. Samples were also serially diluted and plated onto Beerens agar containing 100 μg rifampicin mL−1, in triplicate, to follow probiotic counts.


Sample fixing

The FISH technique was used to enumerate selected groups of bacteria belonging to the canine faecal microbiota as described by Martín-Peláez et al. (2008), but with minor modifications. Briefly, 375 μL of culture fluid was diluted in 1125 μL of ice-cold 4% v/v paraformaldehyde. The samples were mixed thoroughly and then stored at 4 °C for 4 h, to fix the bacterial cells. Samples were then centrifuged at 13 000 g for 5 min, the supernatants were discarded and the cells were washed twice in PBS. The cell pellet was resuspended in 150 μL of filter-sterilized PBS, to which 150 μL of ice-cold, filter-sterilized 96% ethanol was added. The samples were mixed thoroughly and stored at −20 °C.

Pretreatment of cells before the use of probe Lab158

For each sample, one tube of cells fixed in PBS/ethanol was centrifuged for 5 min at 13 000 g. The supernatant was removed carefully. The cell pellet was washed twice in 1 mL of filter-sterilized PBS (5 min, 13 000 g). The cell pellet was then resuspended in 145 μL of lysozyme/lipase buffer [25 mM Tris/HCl, 585 mM sucrose, 5 mM CaCl2, 0.3 mg taurocholic acid mL−1, 10 mM EDTA, 2 mg lysozyme mL−1, 1 mg lipase mL−1 (porcine pancreas type II); pH 7.6] and incubated at 37 °C for 2 h. After this, the sample was centrifuged for 5 min at 13 000 g, and again washed twice in 1-mL filter-sterilized PBS. The supernatant was removed carefully, and the pellet was resuspended in 150 μL of ice-cold filter-sterilized PBS, to which 150 μL of ice-cold 96% ethanol was added. Samples were processed for FISH analysis within 2 days of this procedure being carried out.

Enumeration of bacteria

Details of the oligonucleotide probes used in this study are shown in Table 2. 4′,6-Diamidino-2-phenylindole dihydrochloride (DAPI), a stain for double-stranded DNA, was used to enumerate the total bacteria. Cy3-labelled oligonucleotide probes were obtained from Sigma Genosys (Sigma Aldrich). Slides were viewed under a Nikon E400 Eclipse microscope. DAPI slides were visualized with the aid of a DM400 filter; probe slides were visualized with the aid of a DM575 filter. Fifteen random fields of view were counted per sample, and the numbers of specific bacteria and total bacteria in each sample were determined using the following equation: Embedded Image

View this table:

FISH probes used enumeration of bacterial populations in samples from stirred, pH-controlled anaerobic batch cultures

ProbeTemperature (°C)Sequence (5′–3′)Target molecule and organisms detected by probeReferences
Bac3034648CCAATGTGGGGGACCTT16S rRNA gene; most Bacteroides sensu stricto and Prevotella spp., all Parabacteroides spp., Barnesiella spp. and Odoribacter splanchnicusManz et al. (1996)
Bif1645050CATCCGGCATTACCACCC16S rRNA gene; most Bifidobacterium spp. and Parascardovia denticolensLangendijk et al. (1995)
Chis1505050TTATGCGGTATTAATCTYCCTTT16S rRNA gene; members of Clostridium clusters I (most) and II (all)Franks et al. (1998)
EC15313737CACCGTAGTGCCTCGTCATCA23S rRNA gene; Escherichia coliPoulsen et al. (1994)
Erec4825050GCTTCTTAGTCARGTACCG16S rRNA gene; most members of Clostridium cluster XIVa, Syntrophococcus sucromutans, (Bacteroides) galacturonicus and (Bacteroides) xylanolyticus, Lachnospira pectinschiza and Clostridium saccharolyticumFranks et al. (1998)
Lab1585050GGTATTAGCAYCTGTTTCCA16S rRNA gene; most Lactobacillus, Leuconostoc and Weissella spp., Lactococcus lactis, all Vagococcus, Enterococcus, Melisococcus, Tetragenococcus, Catellicoccus, Pediococcus and Paralactobacillus spp.Harmsen et al. (1999)

where DF is the dilution factor (300/375=0.8); ACC the average cell count of 15 fields of view; DFsample the dilution of the sample used with a particular probe or stain (e.g. 10 × for Bif164 counts); 6732.42 the area of the well divided by the area of the field of view; and 50 takes the cell count back to per millilitre of sample.

SCFA analysis

Samples taken from the batch culture vessels were centrifuged at 13 000 g for 5 min to remove all particulate material. Supernatants were then filtered using 0.2 μm polycarbonate syringe filters (Whatman, UK) and 20 μL was injected into an HPLC system (Merck, NJ) equipped with refractive index detection. The column used was an ion-exclusion REZEX-ROA organic acid column (Phenomenex Inc., UK) maintained at 84.2 °C. Sulphuric acid in HPLC-grade water (0.0025 mmol L−1) was used as the eluent and the flow rate was maintained at 0.5 mL min−1. Quantification of the samples was obtained by comparing with calibration curves of lactic, acetic, propionic and butyric acids in concentrations ranging between 6.25 and 100 mM.

DNA extraction and PCR amplification

DNA was extracted from samples taken from each vessel using the phenol–chloroform method. Cells stored in 50% glycerol at −70 °C for 1 month approximately were centrifuged for 5 min (13 000 g) and washed with 1 mL of sterile PBS. Cells were then resuspended in 500 μL of TES buffer (4 mM Tris-HCl, 1 mM Trizma-base®, 5 mM NaCl, 0.5 mM EDTA), followed by the addition of 8 μL of lysozyme (10 mg mL−1, 50 000 U mg−1) and 2 μL of mutanolysin (1 mg mL−1, 4000 U mg−1). The suspension was mixed by vortexing and incubated at 37 °C for 30 min. Samples were removed and placed on ice, and 10 μL of proteinase K (20 mg mL−1, 30 U mg−1) and 10 μL of RNAse (10 mg mL−1, 70 U mg−1) were added to the cell suspension. Samples were mixed by vortexing and incubated at 65 °C for 1 h. After incubation, 100 μL of 10% sodium dodecyl sulfate was added to the cell suspension and mixed gently and incubated for a further 15 min at 65 °C. Samples were then removed and cooled on ice for 30 min and 620 μL of phenol/chloroform/water (25 : 24 : 1, v : v : v) was added. Samples were carefully mixed and centrifuged for 10 min (5000 g). The aqueous layer was removed and placed into a sterile centrifuge tube. Ice-cold ethanol (1 mL) was added and the samples were mixed by inversion. Samples were placed on ice for 30 min and stored overnight at −20 °C. Samples were centrifuged for 5 min (13 000 g) and the supernatant was discarded. Samples were left to dry at room temperature for ∼3 h. The DNA pellets were resuspended in 50 μL sterile-distilled water and stored at −20 °C. PCRs were performed using the GoTaq® Polymerase kit (Promega, UK). PCR amplification of ∼200 bp of the variable V3 region of the 16S rRNA gene corresponding to positions 341–534 was performed as described by Waldram et al. (2009).

Nested PCR was used to generate Bifidobacterium-specific profiles for the batch culture samples. Almost complete 16S rRNA genes were amplified using the universal 16S rRNA gene primers pA (5′-AGAGTTTGATCCTGGCTCAG-3′; Escherichia coli numbering 9–28) and pH (5′-AAGGAGGTGATCCAGCCGCA-3′; E. coli numbering 1521–1540) (Hoyles et al., 2004). The programme used for amplification was an initial denaturation at 94 °C for 5 min, followed by 25 cycles at 95 °C for 30 s, 55 °C for 30 s and 72 °C for 1 min, with a final extension period at 72 °C for 5 min. A QIA-quick PCR purification kit (QIAgen Ltd, UK) was used to clean up the PCR products (according to the manufacturer's instructions). For Bifidobacterium genus-specific PCR with primers Bif164-GC-r (5′-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGCCACCGTTACACCGGGAA-3′) and Bif662-f (5′-GGGTGGTAATGCCGGATG-3′, Satokari et al., 2001), which produce amplicons of ∼520 bp, 1 μL of the purified PCR products was used. The PCR thermocycling programme was as follows: 95 °C for 5 min; 35 cycles of 94 °C for 30 s, 62 °C for 20 s and 68 °C for 40 s; 62 °C for 20 s; and 68 °C for 7 min. All PCR products were run on a 1.5% w/v agarose gel in 1 × TAE buffer (National Diagnostic, UK) containing ethidium bromide (5 μL of a 0.4 mg mL−1 stock solution) to check for amplicons within the predicted size range.

Denaturing gradient gel electrophoresis (DGGE)

DGGE was performed as described by Waldram et al. (2009), except that electrophoresis was run at a constant voltage of 100 V on the V20-HCDC DGGE system (BDH) with a gradient of 35–60% for P2/P3 and 40–50% for Bif164-GC-r/Bif662-f. PCR samples (5 μL) were applied directly to polyacrylamide gels. After electrophoresis, gels were silver-stained according to the method of Sanguinetti et al. (1994), with the minor modifications described by Martín-Peláez et al. (2010). The gel images were scanned using Canon scanning software (Canon Ltd, UK).

Scanned DGGE images were analysed using quantity one® software (Bio-Rad, UK) by scoring for the presence or absence of bands at different positions in each lane. DGGE profiles of samples from the gel were compared using a similarity matrix. Dendrograms were constructed using unweighted pair-group method with arithmetic mean (UPGMA) analyses.

Sequencing of the DNA in bands extracted from DGGE gels

Bands from the Bifidobacterium-specific DGGE gel were excised, processed and analysed as described by Waldram et al. (2009). All newly generated sequences have been deposited in the EBI database (accession numbers: FN831729FN831761).

Statistical analyses

All data were analysed by anova using the GLM procedure of sas (v.9.1; SAS Institute Incorporated, Cary, NC).

Growth rates, batch culture data, probiotic growth and SCFA were analysed according to the following model: Embedded Image

where Yi is the dependent variable, μ is the overall mean, αi is the effect of the carbohydrate (or probiotic or time) and ɛi represents the unexplained random error.

Differences between means were assessed using the PDIFF option of sas adjusted by Tukey–Kramer. Statistical significance was accepted at P<0.05.


Influence of carbohydrates on the probiotic growth in pure culture

Probiotic growth varied on the test substrates (Table 3), although all probiotics showed high growth rates with glucose (the control). Bifidobacterium bifidum 02 450B grew best on GOS, with the highest growth rate and the highest cell density after 12 h. Bifidobacterium longum 05 also showed good growth on GOS, but significantly lower than that of B. bifidum 02 450B. Lactobacillus acidophilus PrimaLac displayed the highest growth rate on FOS (significantly higher than all the other probiotics); however, after 12 h of fermentation, cell densities were lower compared with the rest of the probiotics. With the exception of L. rhamnosus PrimaLac, all probiotic strains presented the lowest cell densities after 12 h on inulin; however, probiotic growth rates on inulin were similar to the rest of the substrates.

View this table:

Growth characteristics of probiotic strains with different carbohydrates in pure culture after 12 h of incubation

StrainNo. of bacteria [log10(CFU mL−1)] with a substrateGrowth rate (μ) of exponential-phase growth
B. bifidum 02 450B10.45aw9.69bu9.52cw9.93bx0.18<0.0010.34av0.24bx0.23bx0.28abx0.060.02
B. longum 0510.19ax9.37bw9.08cx10.28ay0.12<0.0010.27abw0.22bx0.25bx0.32ay0.050.04
L. acidophilus 14 150B9.26cy10.47ax8.47cy9.96bx0.10<0.0010.16bx0.25ax0.15byz0.25axz0.03<0.001
L. acidophilus PrimaLac8.55bz8.76by8.16cz9.40az0.17<0.0010.19cxz0.36ay0.10dz0.28bx0.06<0.001
L. plantarum 115 400B9.52cy10.56ax9.50cw10.30by0.13<0.0010.21yz0.22x0.24x0.23z0.020.14
L. rhamnosus PrimaLac9.42aby8.31cz9.18bx9.56az0.24<0.0010.25bxy0.15cz0.22bxy0.35ay0.03<0.001
P-value (n=3)<0.001<0.001<0.001<0.001<0.001<0.001<0.001<0.001
  • Values are given as least squares means. Superscripts (a, b, c, …) within a row for the same parameter (either no. of bacteria or μ) indicate significant differences among carbohydrates (P<0.05). Subscripts (w, x, y, z, …) within a column for the same parameter (either no. of bacteria or μ) indicate significant differences among probiotics (P<0.05).

  • RSD, residual SD.

Modulation of bacterial populations in stirred, pH-controlled, anaerobic batch cultures

FISH analysis was used to determine how B. bifidum 02 450B, GOS and their synbiotic combination affected the dominant members of the canine faecal microbiota. Total bacteria increased after 24 h in all systems. In general, most bacterial groups remained at similar levels throughout the fermentation for NC and B. bifidum 02 450B systems, with the exception of Bif164 and EC1531 (Table 4). Bif164 levels were statistically higher after 10 h in the B. bifidum 02 450B system (P=0.01), while EC1531 levels increased significantly in the NC system after 10 h of fermentation. However, after 24 h, the EC1531 population in the NC system was similar to the initial levels. Systems supplemented with GOS led to notable changes in the faecal microbiota. Bif164 levels increased throughout both the GOS and the synbiotic fermentations; however, the synbiotic system had significantly higher Bif164 counts than GOS alone at 10 and 24 h. Similarly, Lab158 counts increased significantly in the prebiotic and synbiotic fermentations compared with systems inoculated with probiotic and NC systems, which showed a decrease in Lab158 counts. The presence of GOS also had an effect on EC1531 levels, with significantly higher counts compared with B. bifidum 02 450B and NC fermentation. In contrast, Chis150 counts were similar in all systems after 10 h; however, significantly higher counts were observed after 24 h in the NC fermentation, whereas the lowest levels were observed in the synbiotic vessel. Bac303 counts were significantly higher in the prebiotic and synbiotic systems after 10 h, but after 24 h of fermentation, there were no significant differences between treatments. Erec482 counts increased in the probiotic, prebiotic and synbiotic vessels throughout the fermentation period, but remained stable in the NC system. After 24 h of fermentation, Erec482 levels were significantly higher in the probiotic, prebiotic and synbiotic systems compared with the NC vessel.

View this table:

In vitro investigation of the effects of GOS, Bifidobacterium bifidum 02 450B and their synbiotic combination on the canine faecal microbiota

TreatmentMean no. of bacteria [log10(cells mL−1)] detected with probe x at time (h)
GOS+B. bifidum 02 450B9.309.6510.068.198.33a8.236.61b8.07b7.78b7.137.266.61b
B. bifidum 02 450B9.319.429.948.268.05b8.136.23a6.81c6.56c7.187.007.25ac
P treatment (n=9)0.340.100.120.590.040.010.04<0.001<0.0010.470.10<0.001
P donor (n=3)0.410.610.820.780.0010.100.910.600.56<0.0010.750.78
GOS+B. bifidum 02 450B7.428.51b8.27b7.788.14ab8.32ab7.097.83a7.64a
B. bifidum 02 450B7.377.63c7.18c7.857.96b8.18bc7.247.28b6.96b
P treatment (n=9)0.14<0.001<0.0010.060.030.0080.10<0.0010.003
P donor (n=3)0.500.870.960.55<0.0010.0070.170.080.96
  • Values are given as least squares means. Superscripts (a, b, c, …) within a column indicate significant differences among treatments (P<0.05).

  • RSD, residual SD.

Monitoring of B. bifidum 02 450B growth

The growth of B. bifidum 02 450B was enhanced by the presence of GOS in the batch culture fermentations (Fig. 1). Bifidobacterium bifidum 02 450B counts were similar in both probiotic and synbiotic systems after 5 h of fermentation. Thereafter, B. bifidum 02 450B levels reverted to the initial levels in the probiotic system. In the synbiotic system, B. bifidum 02 450B levels increased significantly between T5 and T10 and were maintained at the elevated level thereafter.


Growth of rifampicin-resistant Bifidobacterium bifidum 02 450B in stirred, pH-controlled, anaerobic batch cultures over a 24-h fermentation. Rifampicin-resistant B. bifidum 02 450B was grown either with GOS (black bars) or without GOS (grey bars). Values are given as mean±SD of three runs (different canine donor for each). Different letters indicate significant differences between time points within the same treatment (P<0.05).

SCFA levels

The SCFA levels throughout the batch fermentations were influenced by the presence of GOS (Table 5), with significantly higher levels compared with the probiotic and NC batch cultures at T10 and T24. Lactic acid levels increased significantly during the first 10 h for the probiotic, prebiotic and synbiotic systems, with markedly (P=0.06) higher levels in the synbiotic fermentations compared with GOS alone. However, lactate levels decreased in all three systems thereafter. Acetate and propionate levels increased in all four systems after 10 h of fermentation and they continued to increase after 24 h, with the exception of the GOS system, which presented statistically lower levels of acetate compared with 10 h of fermentation (P=0.01). Butyrate was not detected in any of the systems at 0 h, and was significantly higher in the prebiotic and synbiotic systems after 10 h compared with B. bifidum 02 450B and NC systems. Interestingly, butyrate levels continued to increase in the synbiotic system between T10 and T24, with significantly higher levels after 24 h of fermentation than all the other systems.

View this table:

SCFA levels in stirred, pH-controlled, anaerobic batch cultures over 24 h

TreatmentAmounts (mM) of SCFA present at time (h)
Lactic acidAcetic acidPropionic acidButyric acid
GOS+B. bifidum 02 450B0.4233.37a15.11a2.4930.35a31.02a1.0619.79a29.04aND8.01a11.26b
B. bifidum 02 450B0.351.94b0.71b2.6212.25b15.28c1.116.34b8.15bND1.64b1.56c
P treatment (n=9)0.91<0.001<0.00010.98<0.0010.0040.66<0.001<0.0010.001<0.001
P donor (n=3)<0.0010.350.430.320.020.460.01<0.001<0.0010.0030.56
  • Values are given as least squares means. Superscripts (a, b, c, …) within a column indicate significant differences among treatments (P<0.05).

  • RSD, residual SD; ND, none detected.


UPGMA analyses of the universal DGGE profiles demonstrated one main cluster corresponding to GOS treatment T10 and T24 samples (Fig. 2). Interestingly, the GOS profiles from each run formed subclusters of this prebiotic cluster (i.e. were most similar within each run rather than at each time point). A second nondescript cluster was also seen with the same similarity cut-off (i.e. >55% similarity). This cluster contained probiotic and NC treatment samples from runs 1 and 2, together with synbiotic samples for run 2 (T10 and T24) and for T10 run 1.


Comparison of the DGGE profiles (universal primers; P2 and P3) of batch culture fermentation samples from GOS, GOS+Bifidobacterium bifidum 02 450B (GOS+BB), B. bifidum 02 450B (BB) and NC vessels, using the Dice coefficient correlation and the UPGMA method. 10 and 24 h refer to the fermentation time; 1st, 2nd and 3rd indicate the different runs.

Bifidobacterial DGGE of batch culture fermentation samples demonstrated reasonably simple profiles (with 2–4 bands per sample) (Fig. 3). A common band was seen in all the profiles (indicated by the arrow on Fig. 3). At approximately the same position as this common band, another band was seen for profiles from T24 synbiotic samples (just above the common band), which was difficult to see on the scanned gel image. A band corresponding to that of the profile for B. bifidum 02 450B DNA was seen in some samples (most notably those from the probiotic and NC fermentation samples from runs 1 and 3), but not from all samples taken from systems inoculated with B. bifidum 02 450B.


Investigation of the bifidobacterial diversity during stirred, pH-controlled, anaerobic batch culture fermentations (three runs, different canine donor for each). DGGE profiles from samples taken at 10 h (A–D) and 24 h (E–H): A and E, GOS; B and F, GOS+Bifidobacterium bifidum 02 450B; C and G, B. bifidum 02 450B; D and H, NC; I, pure culture of B. bifidum 02 450B. Numbered bands were excised and sequenced (Table 5).

Of the 36 randomly selected clones (four clones for each excised band), 33 contained inserts that could be used for sequence analysis (Table 6). One clone sequence was identified as Halomonas spp. (band 2) and seven others showed the ‘best hit’ for uncultured clone sequences. The remaining 25 clone sequences all had the ‘best hit’ for sequences from 16S rRNA genes of bifidobacterial isolates.

View this table:

Details of Bifidobacterium spp. identified via sequencing of the bands numbered in Fig. 3

BandSelection criterion‘Best hit’ (GenBank accession no./sequence similarity)Number of clone sequences obtainedEMBL accession for clone sequence
1Present only in some run 1 samplesB. longum biovar infantis (AY699578, 99%)1FN178329
B. longum biovar infantis (AB186299, 100%)2FN178330, FN178331
Bifidobacterium spp. (AM117580, 98%)1FN178332
2Present in all run 2 and 3 samples and in run 1 synbiotic samplesUncultured (EF703282, 100%)1FN178333
Halomonas spp. (DQ356998, 100%)1FN178334
B. breve (AB116290, 100%)1FN178335
3Present in all run 1 samplesB. longum (AM990172, 100%)3FN178336, FN178337, FN178339
Uncultured (AB094951, 100%)1FN178338
4Present in run 1 synbiotic sampleB. bifidum (AB116284, 99%, 100%)2FN178340, FN178341
Uncultured Bifidobacterium spp. (AY826697, 97%)1FN178342
B. bifidum (DQ338568, 100%)1FN178343
5Present in run 2 synbiotic samplesUncultured actinobacterium clone (DQ1003, 98%)1FN178344
B. bifidum (AB116284, 100%, 98%)2FN178345, FN178346
B. bifidum (AB116282, 98%)1FN178347
6Present in all run 2 samplesB. bifidum (AB116282, 99%, 100%, 100%)3FN178348, FN178349, FN178350
7Present in run 3 synbiotic samplesB. bifidum (AB437356, 98%, 99%)2FN178351, FN178352
Uncultured actinobacterium clone (EF703282, 100%)1FN178353
8Present in all run 3 samplesB. longum (AM990172, 100%)3FN178354, FN178355, FN178357
Uncultured (FJ369633, 100%)1FN178356
9At the same level as that for probioticUncultured actinobacterium clone (EF703282, 98%)1FN178358
B. longum biovar infantis (AB116310, 100%)1FN178359
Bifidobacterium spp. (AM117580, 100%)2FN178360, FN178361
  • * ‘Best hit’ defined according to Abecia et al. (2007).


The microbial ecology of the canine large intestine has not been well explored, particularly the effects of probiotics and prebiotics on the canine intestinal microbiota. In this study, we evaluated the potential of a synbiotic to modulate the canine faecal microbiota in vitro. Initial screening was conducted to select the preferred synbiotic combination, which was then used for in vitro batch culture fermentations inoculated with canine faecal slurry. Changes in the microbial populations were determined using molecular-based techniques. We showed that GOS+B. bifidum 02 450B enhanced the growth of the probiotic and also has the ability to modulate the faecal microbiota by increasing bacterial groups that are considered to be health promoting, and reciprocally, by preventing the growth of bacterial groups that comprise of pathogenic species.

The growth of probiotics in pure culture with the prebiotics showed strain specificity. GOS induced the greatest growth of B. bifidum 02 450B and B. longum 05. The ability of bifidobacteria to utilize GOS has been demonstrated extensively (Gopal et al., 2001; Rada et al., 2008; Zanoni et al., 2008). This may be explained by the ability of bifidobacteria to produce β-galactosidase activity, an enzyme that hydrolyses GOS (Moller et al., 2001; Hinz et al., 2004; Goulas et al., 2007). FOS was well utilized by all the probiotic strains tested, with the exception of L. rhamnosus. This result was consistent with the study by Kaplan & Hutkins (2000), which showed that 12 out of 15 Lactobacillus strains were able to ferment FOS and L. rhamnosus GG could not ferment FOS. FOS is hydrolysed by extracellular bacterial β-fructosidases, but not all Lactobacillus spp. have the ability to produce these enzymes. In a recent study, Goh et al. (2007) demonstrated that the expression of the fosE gene (a β-fructosidase precursor) in the non-FOS fermenting L. rhamnosus GG enabled the recombinant strain to metabolize FOS. In the current study, both B. bifidum 02 450B and B. longum 05 were able to ferment FOS. This is in agreement with previous studies that reported the ability of bifidobacteria to produce β-fructosidases and to utilize FOS as a substrate (Warchol et al., 2002; Janer et al., 2004). Inulin was the prebiotic that supported the lowest cell densities with most of the probiotic strains in this study, similar to previous studies (Makras et al., 2005; Rossi et al., 2005).

The combination of GOS+B. bifidum 02 450B was deemed the most promising combination for further study and was subsequently selected for an additional study using 24 h pH-controlled, anaerobic batch cultures.

In order to determine whether the addition of a prebiotic could stimulate the growth of the probiotic in mixed microbial batch culture experiments, a rifampicin-resistant variant of the probiotic was induced to track the growth of the probiotic during the 24-h fermentation. This method has been applied in human and piglet studies (Simpson et al., 2000; Saulnier et al., 2008); however, in dogs (or with canine faecal inoculum), only two synbiotic studies exist (Swanson et al., 2002a; Tzortzis et al., 2004). In both studies, the fermentation properties of a synbiotic were assessed, but neither monitored the growth of the probiotic itself. In the present study, we demonstrated that addition of the prebiotic enhanced the growth of B. bifidum 02 450B throughout the fermentation. This increase was not as significant as in the pure culture fermentation, which is due to the complex ecological interactions, such as competition for the substrate, which take place in a mixed culture.

The addition of GOS stimulated the growth of bifidobacteria, which were further increased by the concomitant addition of the probiotic B. bifidum 02 450B (i.e. synbiotic administration). This is in agreement with the in vitro study of Tzortzis et al. (2004), which evaluated the effects of a galactosyl–melibiose mixture combined with L. reuteri. They observed the highest increases in bifidobacterial levels after 10 h in the synbiotic system. A further study by the same group (Tzortzis et al., 2005) used the same GOS as the current study and demonstrated a bifidogenic effect in pigs fed the higher dose of GOS (4%).

Clostridium cluster XIV also increased at the later time of the fermentation with all treatments. This bacterial group comprises both saccharolytic and proteolytic species, and so it is not surprising to observe an increase in this group even in the absence of carbohydrate (due to the fermentation of proteins present in the basal media) (Cummings et al., 1989; Gibson et al., 2004).

Changes in Lab158 numbers after 24 h were less pronounced with the addition of GOS, while a decrease was observed in the NC treatment. Using human faecal inocula, Rycroft et al. (2001) showed that GOS (Oligomate 55) did not alter the levels of this bacterial group after 5 h of fermentation and actually decreased after 24 h. An interesting finding was that the numbers of Clostridium clusters I and II, comprising of potentially pathogenic species, did not increase during the first 10 h in the synbiotic systems (with a significant decrease after 24 h). Decreases in Chis150 numbers have also been reported in previous in vitro studies with GOS (Rycroft et al., 2001; Tzortzis et al., 2004). In contrast, in an in vivo porcine study, faecal clostridia population numbers did not differ among the groups of pigs fed with either GOS or the control diet (Tzortzis et al., 2005).

The in vitro fermentation system used in the present study is a very simplistic representation of the microbiological conditions in the large bowel compared with faeces, not only because faeces have been used to represent the microbial ecosystem in the colon but also because SCFA in the colon are rapidly absorbed, while in the in vitro system, they remain in solution. The higher levels of SCFA in vitro could explain the differences in the cell densities of some microbial groups between in vitro and in vivo studies.

There was a considerable increase in the levels of SCFA with the addition of GOS. A significant increase in lactic acid was observed in the synbiotic fermentation at 10 h, which correlated with higher Lab158 and Bif164 levels. Thereafter, a decline in the concentrations of lactate was seen. Recent studies suggest that lactate is rapidly converted by bacteria in the human intestinal tract (Duncan et al., 2004; Bourriaud et al., 2005). Accordingly, an increase in butyrate concentrations was observed throughout the fermentation, particularly with GOS. Another explanation for elevated butyrate levels was an increase in the Clostridium cluster XIV. Barcenilla et al. (2000), who investigated butyrate-producing bacteria in human faecal samples, found that most of the species were related to Eubacterium spp. However, the results of the latter study were due to the fermentation of substrates other than GOS. Probiotic and NC fermentations mainly induced an increase in acetic and propionic acids. In the absence of the prebiotic, Lab158 did not increase; therefore, it is likely that dominant bacterial groups (BacteroidesPrevotella or Clostridium cluster XIV) were responsible for this SCFA production.

DGGE analysis of the microbial community demonstrated that inter-individual variation played a larger role in the microbial diversity than treatment, with a low similarity between the profiles of the different runs than was seen between different vessels within each run. This corroborated the findings of Simpson et al. (2002), who evaluated the diversity in faecal samples from dogs fed diets containing two different fibres. DGGE analysis showed few differences as a result of diet, but distinct differences between individual dogs (Simpson et al., 2002). Although we attempted to minimize factors that may influence the canine gut microbiota such as the breed and the diet of the dogs (Benno et al., 1992; Simpson et al., 2002), it appears that each individual harbours a unique faecal population as has been shown for humans and pigs (Simpson et al., 2000; Vanhoutte et al., 2006).

When amplification with primers Bif164-GC-r/Bif662-f was performed after DNA extraction, some of the samples did not generate a PCR product. In agreement with Satokari et al. (2001), the samples that did not yield a PCR product corresponded to those with low bifidobacterial counts. As such, a nested-PCR approach was used. Bifidobacteria are one of the main target groups for probiotic and/or prebiotic supplementation in mammals, but little is known about the bifidobacterial populations in dogs. Several studies have reported the presence of bifidobacteria in the canine intestinal contents (Beynen et al., 2002; Flickinger et al., 2003a; Mentula et al., 2005); however, these studies characterized the microbial groups using selective agars that lack selectivity (Greetham et al., 2002). In a study conducted by Vanhoutte et al. (2005), faecal bifidobacteria were not detected after PCR amplification with specific primers even after dogs were fed FOS. The faecal bacterial community is commonly used as a representative of the canine intestinal luminal contents; however, bacteria adhered to the intestinal mucus are neglected (O'Mahony et al., 2009). In the present study, the bifidobacterial DGGE profiles showed a low diversity across all three runs (i.e. for all three dogs). The observation of a common band in the Bif-DGGE profiles for all samples suggests that there was some consistency regardless of the treatment or the fermentation time. From the sequences obtained from the bands excised from the DGGE gel, it appeared that this common band corresponded to B. longum, although that from the run 2 sample was identified as B. bifidum. Even though a very narrow gradient was used for Bif-DGGE, it cannot be assumed that one band equals to one sequence (Jackson et al., 2000), particularly in the case of bifidobacteria whose 16S rRNA gene is extremely conserved among members of this genus (Leblond-Bourget et al., 1996). Moreover, identification of species could be subject to error due to the short lengths of the sequences obtained.

The band observed only in synbiotic samples was identified as B. bifidum, although the mentioned band did not denature at the same position as that corresponding to a pure culture of the probiotic. It therefore seems likely that the synbiotic increased the numbers of other indigenous B. bifidum strains from the canine faecal microbiota. The lack of a band at the same position as that corresponding to the pure culture of B. bifidum 02 450B in the synbiotic systems was unexpected. The bifidobacteria-specific primers used in the present study, however, include many species of bifidobacteria. It may be the case that the levels of B. bifidum 02 450B in the synbiotic system were lower compared with other bifidobacteria species and therefore could not be amplified and hence detected on the gel. Few studies have characterized canine intestinal bifidobacteria using genetic analysis. Kim & Adachi (2007) and Lamendella et al. (2008) found a low incidence of bifidobacteria in dogs, with Bifidobacterium animalis and Bifidobacterium pseudolongum being the species found most frequently. However, these two studies did not report the age, breed or health status of dogs, all factors that may influence bifidobacterial populations. Together with the data from this current study, the results demonstrate low bifidobacterial diversity in canine intestinal contents.

The results obtained from this study indicate that the synbiotic (GOS+B. bifidum 02 450B) induced greater modulation of the canine faecal microbiota in vitro compared with GOS alone. Bifidobacterium bifidum 02 450B growth was stimulated by the prebiotic throughout the fermentation; however, it was not detected by DGGE analysis. The synbiotic also increased the numbers of putative beneficial bacteria and decreased those considered detrimental in the gut microbiota of dogs. Additional research is needed in order to have a clearer picture of the canine gut microbiota and to understand the role played by prebiotics, probiotics and synbiotics in the autochthonous bacterial species of canines.


This work was funded by Hill's Pet Nutrition. Lesley Hoyles is thanked for help with the FISH methodology and preparation of the gel image for publication.


  • Editor: Julian Marchesi


View Abstract