Implications
Administration of antibiotics to animals has come under intense scrutiny because of their contribution to the dissemination of antibiotic-resistant bacteria into the food chain. Algal and yeast β-glucans are biologically active resources from nature. With respect to animal production, an important goal is to determine the optimal microbiota for the animal and then be able to manipulate the microbiota through the diet, supplements, etc. to obtain the desired microbiota. β-glucans derived from macroalgae and yeast aid in gut health during periods of intestinal imbalance, leading to the alteration in the microbial community composition and hence offering an easily accessible alternative to antibiotics. This paper focuses on the use of a molecular and a cultivation-based method to monitor fluctuations in the gastrointestinal microbiota caused by algal and yeast β-glucan supplementation of the pig diet. Molecular analysis of the Lactobacillus community in response to algal supplements in the porcine gastrointestinal tract is uncommon to date.
Introduction
Weaning is a critical period in a piglet's life and is accompanied by nutritional, social and environmental stress. In the immediate post-weaning period, the balance between the development of so-called healthy commensal microbiota and the establishment of a bacterial intestinal disease can be easily tipped towards disease expression (Hopwood and Hampson, Reference Hopwood and Hampson2003). There is a wide interest in developing management and feeding strategies to stimulate gut development and health in young pigs, with the aim of improving productivity of pigs around weaning, while minimising the use of antibiotics and expensive feed ingredients, such as milk products (de Lange et al., Reference de Lange, Pluske, Gong and Nyachoti2010).
Aerobic and facultative anaerobes including Escherichia coli, lactobacilli and streptococci colonise the porcine gastrointestinal tract (GIT) immediately after birth (Smith and Jones, Reference Smith and Jones1963; Mackie et al., Reference Mackie, Sghir and Gaskins1999). These species provide a reducing environment, allowing for the establishment of the obligate anaerobes in the small intestine. These genera include the Bacteroides, Bifidobacterium and Clostridium. Lactobacilli establish early in the piglet's lifetime and they remain a predominant member of the small intestine microbiota (Tannock et al., Reference Tannock, Fuller and Pedersen1990; Naito et al., Reference Naito, Hayashidani, Kaneko, Ogawa and Benno1995; Stewart, Reference Stewart1997).
Considering the high complexity of the gut microbiota, the Lactobacillus and Bifidobacterium populations served as an index of a health-promoting microbiota. These microbial populations are polyfunctional bacteria with industrially important implications (Giraffa et al., Reference Giraffa, Chanishvili and Widyastuti2010). Most lactobacilli present in the GIT of pigs are autochthonous, as they form stable populations throughout the life of the animal host. Unlike the human stomach, which is lined with a glandular mucosa, the stomachs of pigs are lined with non-glandular, squamous stratified epithelium and form a layer of bacterial cells (Walter, Reference Walter2008). These regions are densely colonised by lactobacilli, which adhere directly to the epithelium and form a layer of bacterial cells (Fuller and Brooker, Reference Fuller and Brooker1974).
The effects of a prebiotic are essentially indirect because it selectively feeds one or a limited number of microorganisms, thus causing a selective modification of the host's intestinal (especially colonic) microbiota. It is not the prebiotic by itself but rather the changes induced in microflora composition that is responsible for its effects (Wang, Reference Wang2009). Indeed, the most important bacterial genera targeted for selective stimulation are the indigenous bifidobacteria and lactobacilli (Teitelbaum and Walker, Reference Teitelbaum and Walker2002; Wang, Reference Wang2009). The consumption of feed additives may affect the microbial population dynamic within the GIT causing shifts in the specifically aforementioned populations. Research and development into biotic products has helped pave the way towards finding a successful replacement for antibiotic growth promoters during the weaning phase of pig production (Reilly et al., Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008; Gahan et al., Reference Gahan, Lynch, Callan, O'Sullivan and O'Doherty2009; Janczyk et al., Reference Janczyk, Pieper, Smidt and Souffrant2010).
For newly weaning piglets, the dietary supplementation of prebiotics is generally regarded as a comparatively straightforward approach to improve microbiota composition and functionality of both the small and large intestine (Bauer et al., Reference Bauer, Williams, Smidt, Mosenthin and Verstegen2006). Seaweeds are rich in polysaccharides that are resistant to digestion and adsorption in the small intestine but are completely or partially fermented in the large intestine by the microbiota present (Devillé et al., Reference Devillé, Damas, Forget, Dandrifosse and Peulen2004; Reilly et al., Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008). Seaweed extracts can potentially be used as feed additives (Gardiner et al., Reference Gardiner, Campbell and O'Doherty2008) for both performance improvement and pathogenic bacteria reduction (Allen et al., Reference Allen, Pond and Saker2001; Gahan et al., Reference Gahan, Lynch, Callan, O'Sullivan and O'Doherty2009).
The most abundant polysaccharides in brown seaweeds are laminarin, fucoidan and alginic acid (Zvyagintseva et al., Reference Zvyagintseva, Shevchenko, Chizhov, Krupnova, Sundukova and Isakov2003; Gahan et al., Reference Gahan, Lynch, Callan, O'Sullivan and O'Doherty2009). The chemical structure of laminarin consists mainly of linear β-(1-3) linked glucan with some random β-(1-6) linked side chains depending on the variety of seaweed (Yun et al., Reference Yun, Estrada, van Kessel, Park and Laarveld2003; Brown and Gordon, Reference Brown and Gordon2005). Water-soluble β-glucans such as those derived from Laminaria digitata contain small levels of β-(1-6) linked side chains, whereas water-insoluble β-glucans such as those derived from Laminaria hyperborea only contain β-(1-3) linked residues (Read et al., Reference Read, Currie and Bacic1996; Reilly et al., Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008). β-(1-3) glucans have been shown to exhibit prebiotic properties increasing bifidobacteria and lactobacilli spp. in the large intestine (Mathers and Annison, Reference Mathers and Annison1993; Jaskari et al., Reference Jaskari, Kontula, Siitonen, Jousimies-Somer, Mattila-Sandholm and Poutanen1998), whereas β-(1-6) glucans stimulate the host immune system to protect against fungal, bacterial and viral infections (Michel et al., Reference Michel, Bernard and Lahaye1999; Devillé et al., Reference Devillé, Gharbi, Dandrifosse and Peulen2007). β-glucans isolated from Saccharomyces cerevisiae have been shown to have antibacterial properties (Kogan et al., Reference Kogan, Masler, Šandula, Navarová and Trnovec1989) as well as an ability to stimulate the release of cytokines, such as tumour necrosis factor-α from macrophages (Majtán et al., Reference Majtán, Kogan, Kováčová, Bíliková and Šimúth2005; Gahan et al., Reference Gahan, Lynch, Callan, O'Sullivan and O'Doherty2009). Fucoidans, the family of sulphated heteropolysaccharides, are composed mainly of α-(1-3) linked l-fucose resides (Percival and McDowell, Reference Percival and McDowell1967; Gahan et al., Reference Gahan, Lynch, Callan, O'Sullivan and O'Doherty2009). They are known to have antibacterial properties as they have the ability to inhibit the attachment of certain bacterial species to porcine gastric mucin in vitro (Shibata et al., Reference Shibata, Iimuro and Uchiya2003), as well as having antiviral and anti-inflammatory properties (Marais and Joseleau, Reference Marais and Joseleau2001; Reilly et al., Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008).
An important aspect of this gastrointestinal study is an understanding of the natural variability of Lactobacillus community, for example, in response to specific feed additives. Denaturing gradient gel electrophoresis (DGGE) was applied to circumvent the limitations of conventional cultivation for the analysis of Lactobacillus communities (Vaughan et al., Reference Vaughan, de Vries, Zoetendal, Ben-Amor, Akkermans and de Vos2002; Corsetti and Settanni, Reference Corsetti and Settanni2007). Specifically, DNA is extracted from the cells of all the species represented in the habitat and a hypervariable region of the 16S rRNA gene is amplified by PCR. The mixture of 16S rRNA gene fragments is subjected to DGGE to separate the fragments and obtain a profile of the community (Walter et al., Reference Walter, Hertel, Tannock, Lis, Munro and Hammes2001).
In this study, we investigated the impact of β-glucans isolated from different sources on the microbiota of the porcine GIT. β-glucans were isolated from the seaweeds L. digitata and L. hyperborea, as well as the yeast S. cerevisiae. Standard bacteriological methodology was used to get an overview of β-glucan-induced changes in coliforms, lactobacilli and bifidobacteria communities in different parts of the porcine GIT after feeding animals with the different β-glucans. In addition, molecular typing by PCR-DGGE was used to assess specific changes occurring in the Lactobacillus community.
Material and methods
All procedures described in this study were conducted under experimental license from the Irish Department of Health in accordance with the Cruelty to Animals Act, 1876 and the European Communities (Amendments of the Cruelty to Animals Act, 1876) Regulations, 1994.
Dietary treatments, experimental design and sample collection
Forty-eight 49-day-old pigs with an initial live weight of 17.9 ± 2.2 kg were blocked on the basis of live weight and randomly assigned to one of the four dietary treatments. The pigs had been previously weaned at 26 days of age. The pigs were housed in fully slated pens (1.7 × 1.2 m) and had ad libitum access to feed and water. The baseline diet was composed of wheat and soya (68.7% wheat, 26.5% soya bean meal, 2.3% minerals and vitamins and 2.5% soya oil, University College Dublin, Dublin, Ireland). The seaweed β-glucans were supplied by Bioatlantis Ltd (Tralee, Co. Kerry, Ireland), whereas yeast β-glucan (alkyl polyglycoside 3-6 gluco polysaccharide) was purchased from Biothera (Eagan, MN, USA).
The experiment was set up as a complete randomised study. Twelve pigs were fed one of four treatments: T1, baseline diet (control); T2, baseline diet supplemented with β-glucan from L. digitata (250 g/tonne), referred to as LD β-glucan; T3, baseline diet supplemented with β-glucan from L. hyperborea (150 g/tonne), referred to as LH β-glucan; andT4, baseline diet supplemented with β-glucan from S. cerevisiae (150 g/tonne), referred to as Y β-glucan. Pigs were housed in pens, with two animals confined to each. Only pigs being fed the same diet were penned together. To initially examine the effect of each diet on the porcine GIT microbiota, faecal samples were collected from 40 pigs before and after the feeding period. On day 29 of the trial, 32 of the 48 pigs were humanely slaughtered by Euthanol injection (Pentobarbitone Sodium BP), at a rate of one injection of 0.7 ml per 1.4 kg BW. Luminal samples from the stomach, ileum, caecum, proximal colon and distal colon were collected, diluted onefold in cryoprotective medium (Crowther, Reference Crowther1971) and stored at −80°C for later analysis. The cryoprotective medium was prepared by dissolving 1 g of Lab-Lemco (Oxoid, Basingstoke, Hampshire, UK) and 10 ml of glycerol in 90 ml of water, adjusting the pH to 7.3 and autoclaving at 110°C for 20 min.
Bacterial strains and growth conditions
The strains Lactobacillus acidophilus DSM20079T, Lactobacillus agilis DSM20509T, Lactobacillus amylovorus DSM20531T, Lactobacillus brevis DSM20054T, Lactobacillus delbruecki DSM20076T, Lactobacillus johnsonii DSM10533T, Lactobacillus intestinalis DSM6629T, Lactobacillus mucosae DSM13345T, Lactobacillus plantarum DSM20174T, Lactobacillus reuteri DSM20016T, Lactobacillus ruminis DSM20403T and Lactobacillus salivarius DSM20555T were used as reference strains. Lactobacilli were routinely cultured at 30°C or 37°C on MRS medium (Microbilogy Voigt global distribution ltd, Lawrence, USA).
Microbial cell counts of bifidobacteria, lactobacilli and coliforms
Samples from the ileum, caecum, proximal and distal colon from 32 animals were subjected to microbial cell count analysis. Briefly, digesta samples (∼10 ± 1 g) were aseptically removed from the ileum, caecum and colon of each animal immediately after slaughter, stored in sterile containers (Sarstedt, Nümbrecht, Oberbergischer Kreis, Germany) on ice and transported to the laboratory. A 1.0 g sample was removed from each digesta sample, was serially diluted in maximum recovery diluent (MRD, Oxoid) and spread plated onto selective agars as follows.
Lactobacillus spp. were isolated on de Man, Rogosa, Sharp agar (MRS, Oxoid) with overnight (18 to 24 h) incubation at 37°C in a 5% CO2 environment, as recommended by the manufacturers (Oxoid). The API 50 CHL (BioMérieux SA, Marcy l'Etoile, France) kit was used to confirm lactobacilli colonies.
Bifidobacterium spp. was isolated on Wilkins Chalgren agar (Oxoid), containing 5% (v/v) defribrinated blood (Biological Labs Europe, Heidelberg, Germany) with anaerobic incubation for 24 to 48 h in a Wise Anaerobic Chamber (Don Whitley Scientific, Shipley, West Yorkshire UK) at 37°C, with an atmosphere of 10% hydrogen, 10% CO2 and 80% nitrogen and confirmed using API 50 CHL (BioMerieux).
Enterobacterial colonies were isolated in MacConkey agar (Oxoid), following aerobic incubation at 37°C. API 20E (BioMerieux) were used to confirm enterobacterium colonies. Bacterial counts were expressed per gram of digesta.
DNA extraction
Total DNA was extracted from the Lactobacillus reference strains listed above. Briefly, colonies grown overnight on agar plates were harvested using sterile deionised water to generate a 1 ml cell suspension. Bacterial cells were washed twice using 1 ml of phosphate-buffered saline (PBS, pH 8.3). DNA was extracted from 200 μl of cell suspension. For cells lysis, 5 μl of lysozyme (10 mg/ml in Tris-HCl, pH 8) and 10 μl of mutanolysin (20 mg/ml in bdH20) were added and incubated at 37°C for 1 h. DNA was extracted and purified using the Roche ‘High Pure PCR template preparation kit’ (Roche Diagnostics Ltd, West Sussex, UK) according to the manufacturer's instructions. The eluted DNA was stored at −20°C for further analysis.
Microbial DNA was extracted from the different GIT samples as previously described (Walter et al., Reference Walter, Hertel, Tannock, Lis, Munro and Hammes2001) with some modifications. Total cells from gastrointestinal samples were thawed on ice, centrifuged at low speed for debris removal and washed three times with PBS solution and once with sterile distilled water for samples containing large amounts of contaminants, that is, lower GIT samples. The pellet was then incubated with a lysis buffer containing 1 ml P1 buffer (Qiagen), RNase (Sigma Aldrich, Gillingham, Dorset, UK) (100 μg/ml), lysozyme (20 mg) and mutanolysin (1000 U/ml) – at 37°C for 1 h. After phenol–chloroform extraction and ethanol precipitation, the extracted DNA was dissolved in 100 μl Tris-HCl (10 mM, pH 8).
PCR amplification of microbial DNA for DGGE analysis
PCR was performed using a T3000 Thermocycler (Bio-Sciences Ltd., Dublin, Ireland). The reaction mixture (50 μl) contained 25 pmol of each primer, 0.2 mM of each deoxyribonucleotide triphosphate, 1/10 volume of 10× reaction buffer, 20 mM tetramethylammonium chloride, 25 μg of bovine serum albumin, 2.5 U of Taq DNA polymerase (Promega, Sunnyvale, CA, USA) and 50 μg of DNA.
For PCR-DGGE analysis of DNA extracted from pure Lactobacillus colonies or lactic acid bacteria (LAB) cells extracted from gastrointestinal samples, the primers Lac1 and Lac2GC (Walter et al., Reference Walter, Hertel, Tannock, Lis, Munro and Hammes2001) and the primers L1GC and HDA2 (Meroth et al., Reference Meroth, Walter, Hertel, Brandt and Hammes2003) were used. These primers are specific for the genera Lactobacillus, Weissella, Leuconostoc and Pediococcus. Reference ladders were constructed by PCR amplification of the mix of DNA extracted from the reference strains.
Co-migration of PCR-DGGE bands through the polyacrylamide matrix may arise, whereby the unambiguous identification of some species is not possible. The majority of Lactobacillus species could be differentiated according to migration distances of their respective 16S rRNA fragments, with the exception of L. acidophilus, which co-migrated with Lactobacillus crispatus and Lactobacillus sanfranciscensis (Meroth et al., Reference Meroth, Walter, Hertel, Brandt and Hammes2003). For rapid identification, the primers Lac1 and Lac2-GC were used as they revealed migration distances that differed from those observed with the primers L1-GC and HDA2. Molecular ladders provide an alternative to carrying out excessive amounts of sequencing, which may be unsuccessful.
DGGE and excision of DNA fragments
DGGE analysis of PCR amplicons was performed using a DCode™ Universal Mutation Detection System (Bio-rad, Hercules, CA, USA) as previously described (Walter et al., Reference Walter, Hertel, Tannock, Lis, Munro and Hammes2001), using a denaturing gradient of 32.5% to 50% urea and formamide. A 100% denaturing solution was defined as 7 M urea and 40% formamide. Excision and purification of DNA fragments from DGGE gels were performed as previously described by Ben Omar and Ampe (Reference ben Omar and Ampe2000).
Nucleotide sequencing and sequence analysis
DNA sequences of 16S rRNA PCR fragments obtained from purified DGGE bands were provided by MWG-biotech (Ebersberg, Germany). The closest relatives of the partial 16S sequences were obtained by searches in GeneBank DNA database (Altschul et al., Reference Altschul, Gish, Miller, Myers and Lipman1990). An identity of the partial sequences to the type strains sequences >98% was used as a criterion for species identification.
Statistical analysis
A parametric single factor ANOVA test was used to analyse the microbial counts for three distinct bacterial communities, namely lactobacilli, bifidobacteria and coliforms (Supplementary Material S1). A parametric F-test: two-sample for variances (Supplementary Material S2) and t-test: two-sample assuming unequal variances (Supplementary Material S3) were also conducted on the coliform community. A non-parametric Wilcoxon test and Friedman test were used to analyse the data generated from DGGE banding patterns of the Lactobacillus species community from animals fed one of the four diets (Supplementary Material S4).
Results
Influence of the β-glucan diets on counts of Lactobacillus, Bifidobacterium and Coliforms
Microbial cell counts were used to determine the effects of the different β-glucans on three distinct communities of bacteria, namely, lactobacilli, bifidobacteria and coliforms. API analysis was used to confirm bacterial community classification, that is, 50 colonies were randomly selected per population and tested using API strips, all of which were positive.
Microbiological data demonstrated that all three β-glucans had the ability to reduce coliform counts to varying degrees in the ileum, caecum, proximal and distal colon in comparison with the control diet (Figure 1a–d). However, the LH β-glucan resulted in a statistically significant decrease in coliform counts in the proximal colon of pigs, from 1.56 × 104 CFU/g in the control to 1.62 × 102 CFU/g (Figure 1c). Independent from the diet used, counts of Lactobacillus and Bifidobacterium were not significantly different in any of the GIT samples investigated.
Microbiological counts of total coliforms, total bifidobacteria and total lactobacilli from the ileum, caecum, proximal colon and distal colon samples after 28 days of feeding control diet, LD β-glucan diet, LH β-glucan diet and Y β-glucan diet. (Error bars signify the standard error, i.e., the standard deviation of the sampling distribution.) LD = Laminaria digitata; LH = Laminaria hyperborea.
Statistical analysis was performed on the microbial counts to determine whether at least two treatment group means were significantly different. Therefore, the null hypothesis for the ANOVA was that all treatments had an equal or identical impact of the gastrointestinal microbiota, that is, Ho: control diet = LD β-glucan diet = LH β-glucan diet = Y β-glucan diet. To reject this hypothesis at a 95% confidence interval, the P-value is required to be ⩽0.05. The null hypothesis could not be in rejected in the ileum, caecum, proximal colon and distal colon for any of the three microbial communities (Supplementary Material S1a to d). P-values were lowest for the coliform counts, which prompted further statistical analysis.
The F-Test: Two-Sample for Variances was performed comparing each of the supplemented diets with the control diet (Supplementary Material S2a and b). This test revealed that the variances of each of the treatment groups were different to the control group, which led to the use of the t-test for Two-Sample Assuming Unequal Variances (Supplementary Material S3a and b). The null hypothesis could be rejected for the LH β-glucan diet in the proximal colon where the t statistic was greater than the critical t-value and the P-value was <0.05.
Lactobacillus community shifts in faecal samples of pigs before and after β-glucan consumption
The fluctuations in the Lactobacillus community were monitored by PCR-DGGE, using primer pairs L1GC-HDA2 (Meroth et al., Reference Meroth, Walter, Hertel, Brandt and Hammes2003). The primer pair Lac1/Lac2GC was used to discriminate between co-migrating species (Walter et al., Reference Walter, Hertel, Tannock, Lis, Munro and Hammes2001). Overall, three Lactobacillus species are commonly found in the GIT of pigs, that is, L. amylovorus, L. salivarius and L. reuteri (Hammes and Hertel, Reference Hammes and Hertel2006). Of these three, two were identified in the animals fed the control diet, namely, L. amylovorus and L. reuteri (Figure 2). Feeding the control diet caused a shift in the faecal species composition of the animals, and the incidence of L. amylovorus detection increased (Figure 3).
DGGE profiles obtained with primer pair L1GC/HDA2 and bacterial DNA isolated from porcine gastrointestinal samples after the feeding period. Gel (a) depicts a portion of the data from faecal samples collected before (B) and after (A) each diet treatment, whereas gel (b) and (c) correspond to data from stomach, ileum, caecum, proximal and distal colon of pigs fed the control, LD β-glucan, LH β-glucan or Y β-glucan diets. 3. Lane L, reference strains; 1, Lactobacillus plantarum; 2, Lactobacillus johnsonii; 3, Lactobacillus brevis; 4, Lactobacillus agilis; 5, Lactobacillus acidophilus/Lactobacillus amylovorus; 6, Lactobacillus salivarius; 7, Lactobacillus intestinalis; 8, Lactobacillus ruminis; 9, Lactobacillus mucosae; 10, Lactobacillus reuteri. Band 5 co-migrating species were discriminated using primer pair Lac1-Lac2GC. DGGE = denaturing gradient gel electrophoresis; LD = Laminaria digitata; LH = Laminaria hyperborea.
Lactobacillus species detected by group-specific PCR-DGGE in pigs faecal samples collected before and after each diet treatment. Animals were fed a control diet, control diet plus β-glucan from Laminaria digitata, control diet plus β-glucan from Laminaria hyperborea or control diet plus β-glucan from Saccharomyces cerevisiae.
The inclusion of algal or yeast β-glucan in the pig diets affected the Lactobacillus species distribution in the animals. L. amylovorus was present in approximately two-thirds of all animals before the feeding and was detectable in all animals after the trial, suggesting that this species is a stable member of the population. The detection of L. reuteri remained unchanged after feeding the control diet. Animals fed LD β-glucan and Y β-glucan had a reduced incidence of L. reuteri, whereas this decrease was not observed in animals fed LH β-glucan.
In pigs fed the control and LD β-glucan, L. johnsonii did not persist after the feeding period. On the contrary, L. johnsonii was detected in the faecal samples collected from pig one of 10 pigs fed LH β-glucan and four of nine pigs fed Y β-glucan.
L. mucosae was not detected in pig faecal samples fed the LD β-glucan. On the contrary, the species prevalence increased from zero to four out of 10 after feeding with LH β-glucan and from zero to four out of nine after feeding Y β-glucan.
L. ruminis was a minor member of the population in animals fed the control, LH β-glucan and Y β-glucan diets. Conversely, the incidence of L. ruminis increased after the feeding with LD β-glucan.
After feeding the control diet, the detection of L. brevis in faecal samples was unchanged. In the pig group fed LD β-glucan, L. brevis was no longer detected, whereas LH β-glucan enhanced the detection of this species, as all animals in this group were devoid of L. brevis before feeding, but this species was now detected in three out of 10 animals afterwards. L. agilis was absent from all animals after the feeding period.
The Wilcoxon test was used to determine whether there was a significant shift in the Lactobacillus community in faecal samples of animal before and after the consumption of each diet (Supplementary Material S4a). The null hypothesis was that the diets had no effect or that there was no difference between the community before and after the feeding period. The smaller the P-value, the greater the impact of the dietary treatment. The P-value for the Y β-glucan diet was the smallest. However, a P-value ⩽0.05 would be preferable in order to reject the null hypothesis.
Lactobacillus species detected throughout the GIT using PCR-DGGE after β-glucan consumption
Two animals from each of the four diet groups were slaughtered after the feeding trial. PCR-DGGE analysis of DNA extracted from the contents of the different sites in the GIT gave a comprehensive overview of the specific locations, where initial detection of the different Lactobacillus species occurred (Figure 4). The most common Lactobacillus species detected in the stomach and ileum was L. amylovorus. The species L. mucosae, L. reuteri and L. ruminis were also detected in these sites, albeit in fewer animals. L. johnsonii was only detected in the stomach of a single animal, but it was not found in other samples from the GIT of that animal.
Lactobacillus species detected in stomach, ileum, caecum, proximal and distal colon of pigs fed the control, LD β-glucan, LH β-glucan or Y β-glucan diets. LD = Laminaria digitata; LH = Laminaria hyperborea.
In the caecum, L. amylovorus was present in all animals, independent from the fed diet. L. johnsonii was present in all the caecum samples regardless of diet. L. mucosae was detected in eight of the eight caecum samples. L reuteri was detected in five of the eight samples. L ruminis was detected in six of the eight samples.
In the colon, L. amylovorous was present in all animals, independent of the diets fed to the pigs. L. mucosae was widely detected in the proximal colon regardless of diet. L. reuteri was detected in six out of eight samples from the proximal colon, with samples from Y β-glucan being devoid of this species. L. ruminis was detected in six out of eight proximal colon samples.
In the distal colon, L. johnsonii was detected in a two out of eight samples. L. brevis was detected in six out of eight animals. L. reuteri appeared to be most common in the distal colon samples where it was detected in all samples regardless of the diet. L. ruminis diminished in the distal colon, where it was now detected in four out of eight samples as compared with the caecum and proximal colon.
The Friedman test was also used to determine whether the Lactobacillus species colonising the GIT differed significantly from the stomach, ileum, caecum, proximal colon and distal colon. The null hypothesis was that the diets had no effect or that there was no difference between the communities in each of the GIT sites. The null hypothesis could be rejected with a P-values <0.05 (Supplementary Material S4b).
Discussion
The objective of this study was to determine whether seaweed and yeast β-glucans additives were capable of promoting the growth of beneficial bacteria, while suppressing the growth of potentially pathogenic microbes when used as feed additive in the diet of the weaning pig. To this end, the target communities were monitored using molecular- and culture-based techniques, as fluctuations in these microbial communities relay the effectiveness of the prebiotics under investigation. Microbiological data in this study showed that all three β-glucans had the ability to reduce coliform counts to varying degrees in comparison with the control diet. LH and Y β-glucan caused a one to two log reduction in coliforms present in the proximal and distal colon. Although Y β-glucan had the most consistent reducing effect on coliforms present in all four GIT segments examined, the only statistically significant reduction was caused by the LH β-glucan in the proximal colon. All three β-glucans did not affect the growth of lactobacilli or bifidobacteria communities. Although a reduction in growth of both these bacterial genera was not detected, it is important to note that an increase in microbial counts was also not apparent.
Reilly et al. (Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008) reported a reduction in enterobacteria but also bifidobacteria and lactobacilli in the caecum and/or colon of weaned pigs fed L. digitata containing both laminarin and fucoidan. A similar effect was demonstrated on the beneficial microbiota in pigs fed extracts from L. hyperborea, however, there was no reduction in enterobacteria counts. A combination of both Laminariaspp. extracts resulted in a reduction of enterobacteria and lactobacilli in the caecum and colon but no effects on bifidobacteria were observed. The results from our study showed that the β-glucan exerted a similar reducing effect on coliform counts.
Janczyk et al. (Reference Janczyk, Pieper, Smidt and Souffrant2010) demonstrated that dietary supplementation with alginate resulted in higher enterococci counts in the distal small intestine, caecum and proximal colon of pigs, in comparison with counts obtained from animals fed inulin or a control diet. Counts of lactobacilli decreased in all GIT segments in the experimental group after the feeding period (Janczyk et al., Reference Janczyk, Pieper, Smidt and Souffrant2010). Although alginate is a polysaccharide derived from Phaeophyta, its chemical structure, molecular and chain conformation are different from the laminarin component, thus possessing not only different biological properties but also behaving as a substrate for different LAB phylotypes.
Although the β-glucans did not cause an increase in the growth of lactobacilli, a shift in the microbial community composition was evident. L. amylovorus was a stable member of the Lactobacillus community and it was not affected by any of the diets investigated. The prevalence of this species in pig faeces coincides with its use as a microbial marker of pig manure (Marti et al., Reference Marti, Dabert, Ziebal and Pourcher2010). Konstantinov et al. (Reference Konstantinov, Awati, Smidt, Williams, Akkermans and de Vos2004) demonstrated that the combination of dietary fibre and oligosaccharides present in the diet may specifically stimulate the L. amylovorus-like population along the gut of weaning pigs.
The LD β-glucan caused a shift in the community whereby L. ruminis emerged after the feeding period. L. ruminis has been identified as the dominant lactic acid producing bacterium in the caecum and rectum of the pig (Jassim, Reference Jassim2003). The LH β-glucan caused the persistence of L. brevis and L. mucosae. L. mucosae was originally identified in the small intestine of pigs in an attempt to isolate L. reuteri strains carrying the putative colonisation factor gene for mucus binding (Roos et al., Reference Roos, Karner, Axelsson and Jonsson2000), potentially allowing adherence to mucus loosely associated with gastrointestinal niches. L. johnsonii was negatively affected by the presence of the LH β-glucan within the weaner diet. The reduction in this species may indicate that a shift in the bacterial community towards species that utilise the substrates present in the diet has occurred. Although lactobacilli utilise a large variety of carbohydrates, the spectrum of which can vary significantly between species and strains (Pieper et al., Reference Pieper, Janzcyk, Zeyner, Smidt, Guiard and Souffrant2008), the fluctuations in Lactobacillus species may not be solely based on substrate availability.
To explain the different effects of the β-glucans in the microbial community of the pigs, we need to consider their differences in structure. The molecular weight of each β-glucan was previously determined (Friedlaender et al., Reference Friedlaender, Cook and Martin1954). Both algal β-glucans used in this study were <5000 Da, whereas the yeast β-glucan used was between 5000 and 80 000 Da. Assuming that a higher molecular weight β-glucan, having a higher polymerisation, should be degraded further along the digestive tract, this would indicate that the prebiotic effects on animals fed yeast β-glucan should be more evident. With a higher molecular weight, this glucan potentially remained intact for a longer time period than the algal β-glucans, therefore exerting a prebiotic suppression of deleterious bacteria to a greater extent than both algal β-glucan examined. Johansen et al. (Reference Johansen, Bach Knudsen, Wood and Fulcher1997) examined the digestibility of carbohydrates in the GIT contents of pigs fed diets based on oat bran. Microscopic examination of digesta showed that β-glucan retained intact endospermic cell wall structure until the distal small intestine but was completely disrupted in the caecum. A high percentage of oat β-glucans have a molecular weight of about 80 000 Da. Owing to the size range of the LH β-glucan and Y β-glucan, it is thus possible that the higher molecular weight fractions remained intact until the lower digestive tract of the pigs.
The solubility of the β-glucan supplements within the porcine GIT is also important. Microbiological data revealed that LH β-glucan had a statistically significant suppressive effect on coliform counts in the proximal colon. L. hyperborea is mainly insoluble in water as it contains only linear β-(1-3) linked residues on its structure (Annan et al., Reference Annan, Hirst and Manners1965), hence the insoluble fraction may be passed into the colon (Reilly et al., Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008). However, certain algal fibres present in the seaweed are not a suitable substrate for fermentation by the microbial communities in the large intestine (Reilly et al., Reference Reilly, O'Doherty, Pierce, Callan, O'Sullivan and Sweeney2008). Goñi et al. (Reference Goñi, Gudiel-Urbano, Bravo and Saura-Calixto2001) suggested that certain algal fibres such as fucoidans were resistant to bacterial metabolism. Fucoidans from different species of brown algae inhibit the attachment of certain bacterial species in the porcine GIT, as well as preventing the binding of enterococci and streptococci species to the extracellular matrix proteins of animal cells (Shibata et al., Reference Shibata, Iimuro and Uchiya2003). This indicates that a component of the LH β-glucan behaves as a greater antimicrobial agent than LD β-glucan.
A non-selective reduction in the microbial community, as seen in the distal colon of animals, fed the Y β-glucan supplemented diet may be because of the agglutination properties of the extract (Lynch et al., Reference Lynch, Sweeney, Callan, O’ Sullivan and O’ Doherty2010). Yeast-derived β-glucans, which have a similar structure to algal β-glucans (laminarin component of L. hyperborea), have the ability to agglutinate certain bacterial species, which in turn would inhibit their attachment to epithelial cells, thus preventing the colonisation of mucosal surfaces (Mirelman et al., Reference Mirelman, Altmann and Eshdat1980; Kogan and Kocher, Reference Kogan and Kocher2007). The consistent reduction in the coliform count and also the reduction in Bifidobacterium counts in the distal colon evident in animals fed diets supplemented with Y β-glucan maybe partly because of the agglutination properties of the extract.
In conclusion, DGGE provided an excellent method by which to analyse the interspecies fluctuations caused by the β-glucan-based diets within the porcine GIT, with regard to the Lactobacillus community. The LH β-glucan had the most profound effect on reducing coliform counts ileum and proximal colon in pigs. The suppressive effect of the β-glucans is undoubtedly linked to their molecular weight and solubility. Remarkably, all β-glucan diets provoked the emergence of a more varied community of Lactobacillus. Algal and yeast β-glucans may provide a cheaper alternative to other prebiotic β-glucans, such as oat, which require a more complex extraction procedure to obtain a pure compound.
Acknowledgement
This work was supported by the Food Institutional Research Measure; project references 06/RDC462 and 08/RDCIT600.
Supplementary materials
For supplementary materials referred to in this article, please visit http://dx.doi.org/10.1017/S1751731113000165
