Effects of a dietary β-(1,3)(1,6)-D-glucan supplementation on intestinal microbial communities and intestinal ultrastructure of mirror carp (Cyprinus carpio L.)

Authors


Abstract

Aim

To assess the effects of dietary Saccharomyces cerevisiae β-(1,3)(1,6)-d-glucan supplementation (MacroGard®) on mirror carp (Cyprinus carpio L.) intestinal microbiota and ultrastructure of the enterocyte apical brush border.

Methods and Results

Carp were fed either a control diet or diets supplemented with 0·1, 1 or 2% w/w MacroGard®. Culture-dependent microbiology revealed that aerobic heterotrophic bacterial levels were unaffected by dietary MacroGard® after 2 and 4 weeks. No effects were observed on the allochthonous lactic acid bacteria (LAB) populations at either time point; however, reduced autochthonous LAB populations were observed at week 4. PCR-DGGE confirmed these findings through a reduction in the abundance of autochthonous Lactococcus sp. and Vagococcus sp. in MacroGard®-fed fish compared with the control-fed fish. Overall, sequence analysis detected microbiota belonging to the phyla Proteobacteria, Firmicutes, Fusobacteria and unidentified uncultured bacteria. DGGE analyses also revealed that dietary MacroGard® reduced the number of observed taxonomical units (OTUs) and the species richness of the allochthonous microbiota after 2 weeks, but not after 4 weeks. In contrast, dietary MacroGard® reduced the number of OTUs, the species richness and diversity of the autochthonous microbiota after 2 weeks, and those parameters remained reduced after 4 weeks. Transmission electron microscopy revealed that intestinal microvilli length and density were significantly increased after 4 weeks in fish fed diets supplemented with 1% MacroGard®.

Conclusions

This study indicates that dietary MacroGard® supplementation modulates intestinal microbial communities of mirror carp and influences the morphology of the apical brush border.

Significance and Impact of the Study

To the authors' knowledge, this is the first study to investigate the effects of β-(1,3)(1,6)-d-glucans on fish gut microbial communities, using culture-independent methods, and the ultrastructure of the apical brush border of the enterocytes in fish. This prebiotic-type effect may help to explain the mechanisms in which β-glucans provide benefits when fed to fish.

Introduction

The intestinal microbial communities and their metabolites play an integral role in the ontogeny of teleosts. They influence the host's metabolic and nutritional balance, the immunological status and physiological processes (Cerf-Bensussan and Gaboriau-Routhiau 2010; Merrifield et al. 2010; Sekirov et al. 2010). Intestinal microbial communities consist of allochthonous (digesta-associated, transient) and autochthonous (mucosa-associated, indigenous) microbiota (Ringø and Birkbeck 1999; Ringø et al. 2003). The piscine intestine itself is known to be one of the main portals for disease. Environmental changes, for example in the diet, can lead to imbalances in the intestinal microbial populations and disruption of the mucosal barrier, thus allowing for pathogenic invasion (Lødemel et al. 2001; Ringø et al. 2003, 2007; Vine et al. 2004; Birkbeck and Ringø 2005).

As a result, efforts have been directed towards strengthening the intestinal health of fish reared under intensive aquaculture conditions, especially since the EU ratified the ban on non-medical use of antibiotics in animal nutrition from 2006 (EU 2003) (EC No 1831/2003). Functional dietary supplements such as β-glucans are frequently employed as immunomodulators in animal husbandry. β-glucans have been shown to trigger innate immune responses, increase disease resistance and enhance growth performance of fish (Dalmo and Bøgwald 2008; Soltanian et al. 2009). Although a number of studies have examined the effects of prebiotics on the teleost intestinal microbial communities (Merrifield et al. 2010; Ringø et al. 2012), the effects of β-glucans on the intestinal microbial communities have received little attention. The chemical structure of β-glucans is very similar to prebiotics, consisting of β-glycosidic linked monomeric units. Prebiotics, defined as non-digestible food ingredients, beneficially affect the host by selectively stimulating the growth and/or activity of intestinal bacteria associated with health and well-being (Gibson et al. 2004). Recent studies have reported that prebiotics can modulate the gut microbiota of fish, for example, inulin in Atlantic salmon (Salmo salar L.) (Bakke-McKellep et al. 2007) and Arctic charr (Salvelinus alpinus L.) (Ringø et al. 2006a), and mannan oligosaccharides (MOS) in rainbow trout (Oncorhynchus mykiss) (Dimitroglou et al. 2009) and gilthead sea bream (Sparus aurata) (Dimitroglou et al. 2010). However, the effects of β-glucans, which are often utilized for their immunostimulatory properties, on the gut microbiota of fish have seldom been investigated. To the authors knowledge, only one study has reported the effect of dietary β-(1,3)(1,6)-d-glucans (chrysolaminarin and MacroGard®) on gut microbial populations of fish (Skjermo et al. 2006). This study revealed some evidence of the potential for β-glucans to modulate or disrupt the gut microbiota of fish; however, due to the limited number of replicates and the culture-dependent methods used, no comprehensive conclusions can be drawn.

Recent investigations using culture-independent techniques have provided quantitative information on the vast diversity of bacterial species that inhabit the gastrointestinal (GI) tract of teleosts (Merrifield et al. 2010; Nayak 2010). Using massively paralleled sequencing, van Kessel et al. (2011) reported that the allochthonous microbiota of common carp (Cyprinus carpio L.) was highly diverse and composed of Fusobacteria (46%), Bacteroidetes (21%), Planctomycetes (12%), Gammaproteobacteria (7%), Firmicutes (approximately 4%) and Verrucomicrobiae (1%). Several studies have reported that members of the Firmicutes and Bacteroidetes produce β-glucanases [e.g. Bacillus circulans (Rombouts and Phaff 1976a,b), Arthrobacter spp. (Kitamura et al. 1974; Vršanská et al. 1977; Doi and Doi 1986), Flavobacterium dormitator (Mori et al. 1977)].

Therefore, we hypothesize that the GI microbiota of fish may be influenced by inclusion of dietary β-glucans. As β-glucans are becoming more commonly utilized in aquafeeds, the potential impact on the gut microbiota warrants further investigation.

The aim of the present study was to investigate the effects of a dietary administration of a Saccharomyces cerevisiae β-(1,3)(1,6)-d-glucan (MacroGard®) on the allochthonous and autochthonous microbiota in the distal intestine of mirror carp (Cyprinus carpio L.) after 2 and 4 weeks of feeding. Because possible changes in the composition of the intestinal microbiota may have implications for physiological processes in the mucosa, a further aim was to examine the ultrastructure of the apical brush border of the enterocytes of the same intestinal region.

Material and methods

Experimental diets

A nutritionally balanced basal diet meeting the known requirements of carp (NRC 2011) was prepared by TETRA GmbH (Melle, Germany). The main ingredients of the dietary formulation were fish protein concentrate (Sopropeche CPSP 90, Sopropeche, Wimille, France) containing at least 83% crude protein and 10% lipid and wheat starch (not less than 84% starch) (Table 1). This highly purified diet minimized the presence of other immunostimulatory and prebiotic ingredients. Four isonitrogenous and isolipidic extruded diets were produced using the basal formulation as the control diet and by supplementing three other diets with 0·1, 1 and 2% w/w MacroGard® (Biorigin, Sao Paulo, Brazil), a β-(1,3)(1,6)-d-glucan derived from the yeast Saccharomyces cerevisiae, at the expense of wheat starch (Table 1). The blended ingredients were extruded to form pellets of 2·5 mm diameter. Diets were stored in airtight bags at 4°C throughout the experiment.

Table 1. Formulations (g per 1000 g) and chemical composition (%; = 3) of the experimental diets
IngredientsControl0·1% M1% M2% M
  1. a

    Mineral/trace element premix per kg feed meal (prior to extrusion): 20·0 g monocalcium phosphate, 10·7 g manganese (II) sulfate, 5·7 g zinc sulfate, 4·1 g ferreous (II) sulfate, 0·1 g cobalt (II) acetate.

  2. b

    Vitamin premix per kg feed meal (prior to extrusion): 64 800 IU Vitamin A, 2700 IU Vitamin D3, 144 mg Vitamin E, 75·6 mg Vitamin B1, 57·6 mg Vitamin B2, 216 mg Calcium-D-pantothenate, 720 mg Niacin, 29 mg Vitamin B6, 126 μg Vitamin B12, 900 mg Inositol, 1134 mg Ascorbic acid.

  3. c

    Min. 35% Vitamin C activity.

  4. d

    = 3.

  5. e

    = 5.

  6. f

    = 2.

Fish protein concentrate450·0450·0450·0450·0
Wheat starch410·0409·0400·0390·0
Fish oil45·045·045·045·0
Soybean oil45·045·045·045·0
Cellulose25·6525·6525·6525·65
Min/trace element premixa20·6020·6020·6020·60
Vitamin premixb2·502·502·502·50
Stabilized vitamin Cc1·101·101·101·10
Ethoxyquin0·150·150·150·15
MacroGard®011020
Chemical composition (%DM)
Dry matterd92·9 ± 0·095·1 ± 0·094·8 ± 0·095·0 ± 0·1
Crude proteine41·4 ± 0·142·4 ± 0·343·7 ± 0·342·8 ± 0·2
Crude lipidd14·2 ± 0·014·5 ± 0·115·0 ± 0·415·1 ± 0·3
Ashd4·6 ± 0·04·5 ± 0·04·6 ± 0·04·6 ± 0·0
Gross energy (MJ kg−1)f22·3 ± 0·122·4 ± 0·022·5 ± 0·222·2 ± 0·1

Chemical analysis

Diets were analysed according to AOAC protocols (AOAC 2003). Briefly, dry matter (DM) was determined by drying (Gallenkamp Oven BS, OV-160, Manchester, UK) at 105°C until a constant weight was achieved. Crude protein (N x 6·25; Vapodest 40; Gerhardt Laboratory Instruments, Königswinter, Germany) was determined by automated Kjeldahl method after acid digestion. Crude lipid was determined using the Soxhlet method in a Soxtherm apparatus (model 41x; Gerhardt Laboratory Instruments). Ash content was determined by incineration at 550°C in a muffle furnace (Carbolite ESF3, Hope, UK) for 8 h. Gross energy was determined with a Parr Adiabatic Bomb Calorimeter 1356 (Parr Instrument Company, Moline, IL, USA).

Fish culture and feeding

The experiment was conducted at the aquarium facilities of the Aquatic Animal Nutrition and Health Research Group, Plymouth University, UK. One hundred and fifty mirror carp (Cyprinus carpio L.) were obtained from Hampshire Carp Hatcheries, Hampshire, UK. The fish, weighing approximately 7 g, were carefully acclimatized over 6 weeks whilst being fed the control diet at a rate of 2% of the body weight per day.

Thereafter, 15 carp were allocated to each of the eight 71-l experimental tanks (= 2 tanks per treatment). The initial average body weight at the start of the experiment was 11·26 ± 0·09 g. The carp were fed 4% of their body weight per day by hand at regular intervals (8:30, 11:00, 14:00 & 17:00) over a period of 4 weeks. The body weight was determined after 2 weeks by weighing whole tank body mass, and the feed quantities were adjusted accordingly.

Water quality was monitored daily by measuring oxygen saturation (94·1 ± 0·7%) and pH (6·57 ± 0·50). Ammonium, nitrite and nitrate were determined once a week and the levels ranged between 0 and 0·039 mg l−1, 0 and 0·004 mg l−1 and 16·11 and 31·02 mg l−1, respectively. The average water temperature was maintained at 23·7 ± 0·4°C with the aid of a heating unit. Mechanical filters were cleaned daily, and a partial water exchange (about 10%) was conducted weekly. The photoperiod was maintained at a 12h/12h light/dark cycle.

Microbiology sampling procedures

For the evaluation of the intestinal microbial communities, mucosa and digesta samples from the distal intestine were taken after week 2 (= 4) and week 4 (= 3). Fish were euthanized with an overdose of tricaine methanesulfonate (MS-222; Pharmaq Ltd., Fordingbridge, UK), buffered with sodium bicarbonate (math formula), followed by destruction of the brain (PIL No. 30/9104 under PPL No. 30/2644). The body surface was cleaned with 70% ethanol, the body cavity was opened under aseptic conditions, the entire intestine excised, and the digesta was removed by gentle squeezing (Merrifield et al. 2009a). The mucosal tissue samples were taken from the distal end of the intestine and washed thoroughly three times with sterile phosphate-buffered saline (PBS) in order to remove digesta. At this point, mucosa and digesta subsamples were used for culture-dependent analysis or stored at −20°C in sterile molecular biology-grade 1·5-ml microcentrifuge tubes for culture-independent analysis.

Culture-dependent analysis of the intestinal microbiota

One hundred milligram of each sample (mucosa and digesta) was 10-fold diluted with sterile PBS and homogenized in a macerator (MSE, London, UK). Samples were serially diluted 10-fold with sterile PBS. One hundred microlitres was spread onto TSA (tryptone soy agar, Oxoid, Basingstoke, UK) plates to determine aerobic heterotrophic bacterial levels and MRS (de Man, Rogosa, Sharpe agar, Oxoid) plates to determine lactic acid bacteria (LAB) levels. Samples were plated in duplicate and incubated for 7 days at 25°C before enumeration of colony-forming units (CFUs) from statistically viable plates (i.e. plates containing 30–300 colonies).

Culture-independent analysis of the intestinal microbiota

DNA extraction and PCR amplification of the 16S rRNA V3 region were performed as described by Merrifield et al. (2009a) with slight modifications. Briefly, DNA was extracted from 200 mg of sample (digesta or mucosa) with an additional lysozyme pretreatment (50 mg ml−1 in TE buffer for 30 min at 37°C). PCR amplification used primers P2 and P3, after Muyzer et al. (1993), and 1 μl of DNA template. The PCR was performed with the following parameters: touchdown thermal cycling at 94°C for 5 min and then 20 cycles of 94°C for 1 min, 65°C for 1 min (decreasing 1°C every 2nd cycle) and 72°C for 1 min; this was followed by 15 cycles of 94°C for 1 min, 55°C for 1 min and 72°C for 1 min. A final extension for 5 min at 72°C was performed.

A 40–60% DGGE was conducted on the V3 PCR amplicons using a DCode Universal Mutation Detection System (Bio-Rad laboratories, Segrate, Italy) as described by Merrifield et al. (2009a). Twenty microlitres of each PCR product was used, and the gel was run for 16 h at 60°C and 65 V in 1× TAE buffer (66 mmol l−1 Tris, 5 mmol l−1 Na acetate and 1 mmol l−1 EDTA). DGGE bands were visualized by staining the gel with SYBR® Green (Molecular Probes, Eugene, OR, USA). The gel was scanned using Bio-Rad Gel Doc XR (Bio-Rad laboratories), and negative images were taken with the software Quantity One, version 4.6.3 (Bio-Rad laboratories), and optimized for analysis by enhancing contrast and greyscale. Bands of interest were selected and excised for sequence analysis. The DNA was eluted in molecular-grade water at 4°C for at least 12 h.

The DNA template was then subjected to a re-PCR using P1 (forward primer lacking the GC clamp) and P2 (reverse primer as used for initial amplification). Subsequently, 25 μl of the PCR product was purified using Diffinity Rapid Tips (Diffinity Genomics Inc., West Henrietta, NY, USA) and then sequenced by GATC laboratories (GATC Biotech, Konstanz, Germany). The nucleotide sequences were submitted to a BLAST search in GenBank (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to obtain the closest known alignment identities. The abundance (i.e. intensity units) of the OTUs identified was quantified and presented as a percentage relative to the abundance of the OTU in the control group.

Transmission electron microscopy (TEM)

Samples from the distal intestine of five fish per treatment (= 5) were obtained from fish fed the control, the 0·1% and the 1% MacroGard® supplemented diets for TEM analysis. Rectangular samples of circa 2 mm3 were taken both at week 2 and at week 4. Samples were fixed immediately after excision in 2·5% glutaraldehyde with 0·1 mol l−1 cacodylate buffer, pH 7·2, and refrigerated at 4°C for later analysis (Merrifield et al. 2009b). Samples were then rinsed twice with 0·1 mol l−1 sodium cacodylate buffer for 15 min each and postfixed in OsO4 for 1 h. Afterwards, samples were rinsed again twice with 0·1 mol l−1 sodium cacodylate buffer and dehydrated with graded alcohol solutions of 30, 50, 70, 90 and 100% (twice) for at least 15 min each. Alcohol was removed by gradual replacement with low-viscosity resin at 30, 50, 70 and 100% for at least 12 h in each solution. A second step in 100% resin was followed for 24 h. Samples were placed in Beem® embedding capsules (Beem® Inc., West Chester, PA, USA), and the resin was polymerized at 60°C (overnight). Resin blocks were trimmed using a glass knife and then sectioned using a diamond knife. Ultrathin sections (approximately 80 nm) from each sample were placed on copper grids and stained with saturated uranyl acetate for 15 min, rinsed with distilled water and poststained with Reynolds lead citrate for 15 min (Reynolds 1963). Sections were screened with a JSM 1200EX transmission electron microscope at 120 kV (Jeol; Tokyo, Japan). TEM images were analysed using Image J version 1.42 (National Institutes of Health, Bethesda, MD, USA) to measure the microvilli length as described by Hu et al. (2007) with slight modifications. Briefly, the length of 10 well-orientated longitudinal microvilli was evaluated in five ultrathin sections per sample (50 measurements per sample; in total, 250 measurements per treatment). Furthermore, the microvilli density was assessed after Daniels et al. (2010) with a slight modification. Briefly, the number of microvilli present on the enterocyte surface was evaluated and then standardized to 1 μm in five ultrathin sections per sample (five measurements per sample; in total, 25 measurements per treatment).

Statistical analysis

All presented data are means ± standard deviation (SD). Statistical analysis was performed using SPSS statistics, version 18 (SPSS Inc., Chicago, IL, USA). The data were checked for normality and homogeneity of variance and then subjected to a one-way analysis of variance (anova). Significant differences between control and treatment groups were determined by a post hoc LSD test and were accepted at the < 0·05 level.

The DGGE banding patterns were transformed into intensity matrices in order to evaluate similarities between treatments using the software Quantity one, version 4.6.3 (Bio-Rad Laboratories), after Schauer et al. (2000). Band intensities were measured and analysed using Primer V6 software (Clarke and Gorley 2006), and similarity percentages (SIMPER) were determined. A one-way analysis of similarity (anosim) was used for pairwise comparison to determine differences between DGGE banding profiles (Abell and Bowman 2005).

The total number of operational taxonomical units OTUs (S), the Margalef species richness (d = (S − 1)/log(N)), the Shannon diversity index (H′ = × Σ (pi (ln pi)) and the Pielou evenness (J' = H′/log(S)) were calculated, where N is the total number of individuals (total intensity units) and pi represents the proportion of the total number of individuals in the ith species. These parameters were subjected to a one-way anova.

Results

Culture-dependent analysis of the intestinal microbiota

The effects of a dietary MacroGard® supplementation on viable counts of aerobic heterotrophic bacteria and LAB in the carp intestine were determined by a culture-based approach using TSA and MRS agar plates, respectively. Figure 1 displays the allochthonous and autochthonous viable cell counts after week 2 and week 4. After 2 weeks of feeding, the levels of both allochthonous and autochthonous aerobic heterotrophic bacteria fluctuated at around log 6·5 CFU g−1 across treatments (Fig. 1a). Figure 1b shows that the allochthonous LAB reached levels of approximately log 5 CFU g−1 (present in all replicates), whereas the respective autochthonous levels were lower at levels of log 3 CFU g−1 (detectable in three of the four replicates of the 0·1% MacroGard® group). After 4 weeks of feeding, the allochthonous aerobic heterotrophic bacteria mirrored the results from week 2 with levels of around log 6·5 CFU g−1 (Fig. 1c); the levels of the respective aerobic heterotrophic autochthonous bacteria, however, decreased from log 6 CFU g−1 in the control group to log 3 CFU g−1 in the 2% MacroGard® group (detectable in two of the replicates). A clear reduction in the abundance of autochthonous LAB was observed in fish fed β-glucans. Indeed, levels were too low to be statistically viable (i.e. below 30 CFU on the lowest dilution plate) in two fish from the 1% MacroGard® group and one fish from the 2% MacroGard® group. The levels decreased from log 3·9 CFU g−1 in the control group to log 1 CFU g−1 in the 1% MacroGard® group and log 1·4 CFU g−1 in the 2% MacroGard® group.

Figure 1.

(a–d) Viable counts (CFU g−1) of allochthonous (black bars) and autochthonous (light grey bars) aerobic heterotrophic bacteria and lactic acid bacteria in the carp intestine after 2 weeks (= 4) and 4 weeks (= 3) of feeding control (0%) or MacroGard® (0·1, 1, 2% M)-supplemented diets. Statistically viable levels of bacteria were recovered from all replicates unless indicated by the number present in parenthesis.

Culture-independent analysis of the intestinal microbiota

Week 2

The influence of dietary MacroGard® supplementation on the intestinal microbial diversity of carp was investigated using culture-independent techniques. Figures 2-5 display the 16S rRNA V3 PCR-DGGE fingerprints of the allochthonous and autochthonous microbiota after 2 and 4 weeks of feeding together with the respective dendrograms and the nonmetric multidimensional scaling analysis plots. Tables 2 and 3 show the microbial ecological parameters derived from those PCR-DGGE fingerprints at week 2 and 4, respectively.

Figure 2.

(a) Denaturing gradient gel electrophoresis (DGGE) profiles of PCR-amplified products of the V3 region of the 16S rRNA gene from week 2 allochthonous samples (numbers indicate OTUs sequenced). (b) Bray–Curtis dendrogram demonstrating the similarity. (c) nonmetric multidimensional scaling analysis plots showing clusters at different similarity levels (%). (image_n/jam12313-gra-0001.png) Control; (image_n/jam12313-gra-0001.png) 0·1% Macrogard; (image_n/jam12313-gra-0003.png) 1% Macrogard; (image_n/jam12313-gra-0004.png) 2% Macrogard; (image_n/jam12313-gra-0005.png) 50; (image_n/jam12313-gra-0006.png) 60; (image_n/jam12313-gra-0007.png) 70. Carp were fed control (0%) or MacroGard® (01, 1, 2% M)- supplemented diets for 4 weeks.

Figure 3.

(a) Denaturing gradient gel electrophoresis (DGGE) profiles of PCR-amplified products of the V3 region of the 16S rRNA gene from week 2 autochthonous samples (numbers indicate OTUs sequenced). (b) Bray–Curtis dendrogram demonstrating the similarity. (c) nonmetric multidimensional scaling analysis plots showing clusters at different similarity levels (%). (image_n/jam12313-gra-0008.png) Control; (image_n/jam12313-gra-0009.png) 0·1% Macrogard; (image_n/jam12313-gra-0010.png) 1% Macrogard; (image_n/jam12313-gra-0011.png) 2% Macrogard; (image_n/jam12313-gra-0012.png) 50; (image_n/jam12313-gra-0013.png) 65; (image_n/jam12313-gra-0014.png) 80. Carp were fed control (0%) or MacroGard®(01, 1, 2% M)- supplemented diets for 4 weeks..

Figure 4.

(a) Denaturing gradient gel electrophoresis (DGGE) profiles of PCR-amplified products of the V3 region of the 16S rRNA gene from week 4 allochthonous samples. (b) Bray–Curtis dendrogram demonstrating the similarity. (c) nonmetric multidimensional scaling analysis plots showing clusters at different similarity levels (%). (image_n/jam12313-gra-0015.png) Control; (image_n/jam12313-gra-0016.png) 0·1% M; (image_n/jam12313-gra-0017.png) 1% M; (image_n/jam12313-gra-0018.png) 2% M; (image_n/jam12313-gra-0019.png) 80; (image_n/jam12313-gra-0020.png) 85; (image_n/jam12313-gra-0021.png) 90; (image_n/jam12313-gra-0022.png) 95. Carp were fed control (0%) or MacroGard®(0-1, 1, 2% M)-supplemented diets for 4 weeks.

Figure 5.

(a) Denaturing gradient gel electrophoresis (DGGE) profiles of PCR-amplified products of the V3 region of the 16S rRNA gene from week 4 autochthonous samples. (b) Bray–Curtis dendrogram demonstrating the similarity. (c) nonmetric multidimensional scaling analysis plots showing clusters at different similarity levels (%). (image_n/jam12313-gra-0023.png) Control; (image_n/jam12313-gra-0024.png) 0·1% M; (image_n/jam12313-gra-0025.png) 1% M; (image_n/jam12313-gra-0026.png) 2% M; (image_n/jam12313-gra-0027.png) 60; (image_n/jam12313-gra-0028.png) 70; (image_n/jam12313-gra-0029.png) 80. Carp were fed control (0%) or MacroGard® (0-1, 1, 2% M)-supplemented diets for 4 weeks.

Table 2. Microbial community analysis of the allochthonous and autochthonous microbial communities of carp from DGGE fingerprints after 2 weeks of feeding control (0%) or MacroGard® (0·1, 1, 2% M)-supplemented diets
 Microbial ecological parametersSimilarity
N RichnessEvennessDiversitySIMPER (%)Control0·1% M1% M2% M
  1. a,bDifferent superscripts indicate a significant difference (< 0·05).

  2. N = number of operational taxonomical units; Richness = Margalef species richness, Diversity = Shannon's diversity index, SIMPER = similarity percentage within group replicates. Similarity = pairwise comparison of weighted Bray–Curtis similarity.

Week 2 allochthonous
Control28·00 ± 3·27a2·63 ± 0·28a0·95 ± 0·013·16 ± 0·1257·5 ± 17·910050·16 ± 12·6256·11 ± 6·2546·99 ± 11·18
0·1% M23·00 ± 3·37b2·17 ± 0·28b0·96 ± 0·023·00 ± 0·1254·7 ± 18·3 10060·66 ± 10·8250·56 ± 15·03
1% M24·00 ± 2·71a,b2·23 ± 0·24b0·95 ± 0·013·02 ± 0·1167·5 ± 3·0  10054·81 ± 8·06
2% M21·25 ± 2·22b1·99 ± 0·16b0·96 ± 0·022·92 ± 0·1142·0 ± 18·0   100
Week 2 autochthonous
Control34·00 ± 4·55a3·00 ± 0·36a0·99 ± 0·003·47 ± 0·13a75·5 ± 8·110073·10 ± 4·7567·48 ± 7·3453·81 ± 8·78
0·1% M29·33 ± 2·08a,b2·63 ± 0·17a,b0·98 ± 0·003·31 ± 0·07a,b80·6 ± 1·5 10077·72 ± 8·8962·07 ± 9·64
1% M28·75 ± 2·06b2·60 ± 0·18a,b0·98 ± 0·003·30 ± 0·06b73·2 ± 9·2  10064·62 ± 8·53
2% M25·00 ± 3·37b2·31 ± 0·27b0·98 ± 0·003·15 ± 0·13b67·8 ± 7·9   100
Table 3. Microbial community analysis of the allochthonous and autochthonous microbial communities of carp from DGGE fingerprints after 4 weeks of feeding control (0%) or MacroGard® (0·1, 1, 2% M)-supplemented diets
 Microbial ecological parametersSimilarity
N RichnessEvennessDiversitySIMPER (%)Control0·1% M1% M2% M
  1. a,bDifferent superscripts indicate a significant difference (< 0·05).

  2. N = number of operational taxonomical units, Richness = Margalef species richness, Diversity = Shannon's diversity index, SIMPER = similarity percentage within group replicates. Similarity = pairwise comparison of weighted Bray–Curtis similarity.

Week 4 allochthonous
Control30·67 ± 1·152·76 ± 0·090·98 ± 0·003·36 ± 0·0492·75 ± 3·3410089·62 ± 2·0886·37 ± 1·8781·91 ± 1·28
0·1% M32·00 ± 0·002·87 ± 0·020·98 ± 0·003·40 ± 0·0195·17 ± 2·47 100·0087·67 ± 3·9284·51 ± 4·96
1% M28·67 ± 1·152·60 ± 0·080·98 ± 0·003·28 ± 0·0490·53 ± 3·02  100·0083·12 ± 3·50
2% M31·67 ± 2·312·83 ± 0·170·98 ± 0·003·39 ± 0·0885·81 ± 9·42   100
Week 4 autochthonous
Control28·00 ± 3·27a2·31 ± 0·32a0·97 ± 0·013·14 ± 0·16a70·81 ± 1·85a10070·00 ± 7·6771·22 ± 5·8255·77 ± 5·32
0·1% M23·00 ± 3·37b1·83 ± 0·31a,b0·97 ± 0·012·87 ± 0·19a,b82·48 ± 6·08b 10083·67 ± 3·8069·58 ± 7·06
1% M24·00 ± 2·71a,b1·99 ± 0·22a,b0·98 ± 0·012·98 ± 0·14a87·36 ± 2·48b  10071·94 ± 6·46
2% M21·25 ± 2·22b1·53 ± 0·09b0·97 ± 0·022·69 ± 0·01b84·61 ± 3·78b   100

Figure 2 illustrates low similarities between treatments of the allochthonous samples taken after 2 weeks. This is confirmed by the microbial ecological parameters displayed in Table 2, which show approximately 47–61% similarities between treatments. MacroGard® reduced the number of OTUs from 28·00 ± 3·27 in the control group to 23·00 ± 3·37 (= 0·033) and 21·25 ± 2·22 (= 0·007) in the 0·1 and 2% MacroGard® group, respectively; the reduced numbers of OTUs in the 1% MacroGard® group (24·00 ± 2·71) were approaching significance (= 0·077). Similarly, the species richness was significantly reduced in all MacroGard® groups to 2·17 ± 0·28 (= 0·020), 2·23 ± 0·24 (= 0·038) and 1·99 ± 0·16 (= 0·003) from 2·63 ± 0·28 in the control treatment. Species diversity, evenness and SIMPER remained unaffected.

Similarly, Fig. 3 illustrates low similarities between treatments of the autochthonous microbiota after 2 weeks. This is confirmed by the microbial ecological parameters displayed in Table 2, which show approximately 54–78% similarities between treatments. Dietary MacroGard® reduced the number of OTUs from 34·00 ± 4·55 in the control group to 29·33 ± 2·08 (= 0·088), 28·75 ± 2·06 (= 0·044) and 25·00 ± 3·37 (= 0·002) in fish fed the 0·1, 1 and 2% MacroGard®-supplemented diets, respectively. Additionally, the species richness decreased from 3·00 ± 0·36 in the control group to 2·63 ± 0·17 (= 0·098), 2·60 ± 0·18 (= 0·055) and 2·31 ± 0·27 (= 0·003) and Shannon's diversity index decreased from 3·47 ± 0·13 in the control group to 3·31 ± 0·07 (= 0·073), 3·30 ± 0·06 (= 0·044) and 3·15 ± 0·17 (= 0·001) in the 0·1, 1 and 2% MacroGard® groups, respectively. Species evenness and SIMPER remained unaffected by dietary MacroGard®.

Week 4

At week 4, the allochthonous microbial profiles from each treatment group were highly similar (of approximately 82–90% similarity) (Table 3 and Fig. 4). The microbial ecological parameters revealed the total number of OTUs (28·67–32·00), the species richness (2·60–2·87), the species evenness (0·98), the species diversity (3·28–3·40) and the SIMPER similarity (85·81–95·17) remained unaffected by dietary MacroGard® (Table 3).

However, Fig. 5 illustrates that the similarity of the autochthonous microbiota remained low between treatments at week 4. Indeed, similarity was approximately 56–84% between the treatments (Table 3). Dietary MacroGard® supplementation reduced the number of OTUs from 28·00 ± 3·27 in the control group to 23·00 ± 3·37 (= 0·045), 24·00 ± 2·71 (= 0·131) and 21·25 ± 2·22 (= 0·004) in fish fed the 0·1, 1 and 2% MacroGard®-supplemented diets (Table 3). Similarly, the species richness was reduced from 2·31 ± 0·32 in the control group to 1·83 ± 0·31 (= 0·052), 1·99 ± 0·22 (= 0·163) and 1·53 ± 0·09 (= 0·006) and the Shannon diversity index from 3·14 ± 0·16 in the control group to 2·87 ± 0·19 (= 0·052), 2·98 ± 0·14 (= 0·211) and 2·69 ± 0·01 (= 0·005) in the 0·1, 1 and 2% MacroGard® groups, respectively. Species evenness was unaffected. A significant increase in SIMPER similarity in all MacroGard® groups was observed from 70·81 ± 1·85% in the control group to 82·48 ± 6·08% (= 0·006), 87·36 ± 2·48% (= 0·001) and 84·61 ± 3·78% (= 0·003) in the 0·1, 1 and 2% MacroGard® groups, respectively.

Sequence analysis

A number of OTUs were excised for sequence analysis from the PCR-DGGE gels (Tables 4 and 5). These OTUs belonged to the phyla Proteobacteria, Firmicutes and Fusobacteria or were unidentified uncultured bacteria. Potentially pathogenic (e.g. Aeromonas spp., Lactococcus garvieae), probiotic (e.g. Vagococcus spp., Weissella spp., Lactococcus spp.) and beneficial (e.g. Cetobacterium) microbiota were detected.

Table 4. Closest relatives with similarity (%) to the respective relatives and mean relative abundance for the sequences obtained from the PCR-DGGE of the allochthonous and autochthonous microbial communities from carp after 2 weeks of feeding control (0%) or MacroGard® (0·1%, 1%, 2% M)-supplemented diets
Band No.Closest relativeIdentity (%)Mean relative abundance (%)
Control0·1% M1% M2% M
  1. ‘−′ = species absent; ‘+′ = species present at least in one replicate.

Week 2 allochthonous
1 Lactococcus garvieae 10010089·8130·5149·0
2Aeromonas sp.99+
3Uncultured Propionigenium sp.1001000·075·936·5
4Vagococcus sp.10010079·9136·260·4
5Uncultured bacterium99100120·1120·358·0
6Uncultured bacterium10010057·983·957·4
7Uncultured Propionigenium sp.10010043·3133·565·4
8 Aeromonas hydrophila 99+++
9 Aeromonas hydrophila 98100116·8116·8111·3
10 Aeromonas hydrophila 10010098·3123·799·0
11Uncultured Aeromonas sp.100++
Week 2 autochthonous
12 Weissella cibaria 10010083·093·373·7
13Uncultured Lactococcus sp.100100108·281·561·7
14Vagococcus sp.10010062·145·050·3
15Aeromonas sp.10010096·081·567·5
16 Aeromonas hydrophila 10010089·080·865·7
Table 5. Closest relatives with similarity (%) to the respective relatives and mean relative abundance for the sequences obtained from the PCR-DGGE of the allochthonous and autochthonous microbial communities from carp after 4 weeks of feeding control (0%) or MacroGard® (0·1, 1, 2% M)-supplemented diets
Band No.Closest relativeIdentity (%)Mean relative abundance (%)
Control0·1% M1% M2% M
Week 4 allochthonous
17 Lactobacillus sakei 94100105·386·587·8
18Vagococcus sp.100100104·993·193·9
19Uncultured Streptococcus sp.99100100·092·587·0
20 Chlostridium tetani 93100104·974·695·8
21Uncultured bacterium100100103·493·2104·2
Week 4 autochthonous
22Uncultured bacterium971000·00·00·0
23 Lactococcus garvieae 10010038·9145·4112·7
24Uncultured Lactococcus sp.10010083·085·269·9
25Vagococcus sp.9910086·993·378·5
26Uncultured Cetobacterium sp.9910079·980·040·1
27 Aeromonas hydrophila 1001000·00·00·0

Week 2

In the allochthonous microbiota after 2 weeks of feeding, the relative abundance of L. garvieae (OTU 1) increased in fish fed the 1 and 2% MacroGard®-supplemented diet (130·5 and 149·0% relative to the control abundance, respectively). Reductions in relative abundance were seen in an uncultured Propionigenium sp. (OTU 3) and an uncultured bacterium (OTU 6) with all MacroGard® concentrations. Minor changes in relative abundance were observed for Aeromonas hydrophila strains (OTU 8–10) and uncultured Aeromonas sp. (OTU 11) and Aeromonas sp. (OTU 2).

In the autochthonous microbiota after 2 weeks of feeding, the abundance of LAB OTUs (Weissella cibaria, uncultured Lactococcus sp., Vagococcus sp.; OTUs 12–14) and aeromonads (Aeromonas sp., Aeromonas hydrophila; OTUs 15 and 16) decreased with increasing dietary MacroGard® supplementation. The presence of Vagococcus sp. (OTU 14) was clearly reduced in all MacroGard® groups (0·1%M = 62·1%, 1%M = 45·0% and 2%M = 50·3%). The uncultured Lactococcus sp. (OTU 13) was less prominent only at the 1 and 2% MacroGard® concentrations (81·5 and 61·7%, respectively). The relative abundance of Weissella cibaria (OTU 12) was marginally decreased with all dietary MacroGard® concentrations (0·1%M = 83·0%, 1%M = 93·3% and 2%M = 73·7%). Aeromonas sp. (OTU 15) decreased to 81·5 and 67·5% and Aeromonas hydrophila (OTU 16) decreased to 80·8 and 65·7% in the 1 and 2% MacroGard® groups.

Week 4

In the allochthonous microbiota after 4 weeks of feeding, the sequences retrieved from the excised bands were present in all samples and identified as Lactobacillus sakei (OTU 17), Vagococcus sp. (OTU 18), Streptococcus sp. (OTU 19), Clostridium tetani (OTU 20) and an uncultured bacterium (OTU 21). 0·1% MacroGard® supplementation had no effect on abundances, but marginal reductions were observed at 1 and 2% MacroGard® supplementation.

In the autochthonous microbiota after 4 weeks of feeding, OTUs identified as uncultured Lactococcus sp. (OTU 24) and Vagococcus sp. (OTU 25) were present in all samples, but the abundances were reduced in fish fed the MacroGard®-supplemented diets. The uncultured Lactococcus sp. decreased to 83·0 and 85·2% with the 0·1 and 1% MacroGard®-supplemented diets and further decreased to 69·9% with the 2% MacroGard®-supplemented diet. Similarly, the Vagococcus sp. was reduced to 86·9 and 93·3% in fish fed the 0·1 and 1% MacroGard®-supplemented diets and further reduced to 78·5% in fish fed the 2% MacroGard®-supplemented diet. An uncultured Cetobacterium sp. (OTU 26) was marginally reduced in the 0·1 and 1% MacroGard® groups (79·9 and 80·0%, respectively) and clearly reduced in the 2% MacroGard® group (40·1%; absent in one replicate). Lactococcus garvieae (OTU 23) was absent in one control replicate and in two replicates of fish fed the 0·1% MacroGard®-supplemented diet (38·9% abundance); the abundance was increased to 145·4 and 112·7% in the 1 and 2% MacroGard® groups, respectively. OTU 27, identified as Aeromonas hydrophila, was present in two of the three control replicates, but was not detectable in any of the replicates of the MacroGard® treatments.

Epithelial ultrastructure

The apical brush border of the enterocytes in the distal intestinal region from fish fed either the control, the 0·1% or the 1% MacroGard®-supplemented diets were examined using transmission electron microscopy. Overall, the epithelial architecture appeared healthy with integrated, columnar-shaped enterocytes. Numerous, well-developed microvilli with the presence of associated terminal webs and no damage to tight junctional complexes, desmosomes or intercellular spaces were observed. The microvilli length and density of intestinal enterocytes were analysed (Table 6). Dietary MacroGard® supplementation did not affect microvilli length or density after 2 weeks of feeding. After 4 weeks of feeding, however, the administration of 1% MacroGard® led to a significant increase in the microvilli length (< 0·025) and density (< 0·038) compared with fish fed the control and the 0·1% MacroGard®-supplemented diet.

Table 6. Transmission electron microscopy analysis of the microvilli length (in μm) and the microvilli density (microvilli μm−1) of distal intestinal enterocytes of carp after 2 and 4 weeks of feeding control (0%) or MacroGard® (0·1%, 1% M)-supplemented diets (= 5)
 Control0·1% M1% M
  1. a,bDifferent superscripts indicate a significant difference (< 0·05).

Microvilli length
Week 21·38 ± 0·201·22 ± 0·141·42 ± 0·19
Week 41·32 ± 0·15a1·46 ± 0·15a1·68 ± 0·12b
Microvilli density
Week 27·06 ± 0·257·28 ± 0·217·37 ± 0·40
Week 47·21 ± 0·42a7·36 ± 0·22a7·85 ± 0·33b

Discussion

The present study indicates for the first time that dietary β-(1,3)(1,6)-d-glucans may affect the intestinal microbial communities in teleosts. Carp intestinal microbial communities have been assessed using culture-dependent methods in the past and have been reported to consist mainly of aerobes and facultative anaerobes (Sugita et al. 1985), with Aeromonas species being highly abundant (Sugita et al. 1990; Namba et al. 2007). Enterobacteriaceae, Pseudomonas, Moraxella, Acinetobacter and Micrococcus (Sugita et al. 1990; Namba et al. 2007), Bacteriodetes (Sugita et al. 1990; Tsuchiya et al. 2008), Plesiomonas, Flavobacterium, Staphylococcus, Streptococcus, Bacillus and Clostridium (Sugita et al. 1990), Vibrio (Namba et al. 2007) and Cetobacterium (Tsuchiya et al. 2008) have also been detected. It is, however, generally accepted that evaluation of intestinal microbial communities by using culture-based techniques alone is inaccurate and restricted to culturable bacteria (Cahill 1990; Ringø et al. 2006b; He et al. 2011). To the authors' knowledge, there is no published information on the effects of β-glucans on the intestinal microbiota of carp, or indeed any teleost, using culture-independent methods. The present study therefore applied both culture-dependent and culture-independent methods to assess the effects of dietary β-(1,3)(1,6)-d-glucans (MacroGard®) on allochthonous and autochthonous microbial communities of mirror carp (Cyprinus carpio L.).

In the present study, cultivable allochthonous LAB levels were not affected by dietary MacroGard® at either sampling time point. Similarly, Smith et al. (2011) reported that laminarin, a low-molecular-weight (7·7 kDa), soluble β-(1,3)(1,6)-d-glucan, had no effect on allochthonous lactobacilli and bifidobacteria levels in the pig colon. Contradictory results have been reported in rats, where elevated allochthonous intestinal lactobacilli levels have been observed after dietary β-glucan supplementation (Dongowski et al. 2002; Snart et al. 2006).

In the present study, the autochthonous LAB levels were not affected after 2 weeks by dietary MacroGard® supplementation. After 4 weeks, however, autochthonous LAB levels decreased with increasing MacroGard® supplementation. Sequence analyses from DGGE profiles identified a number of LAB, and confirming the culture-based results, the abundance of the autochthonous LAB was reduced in the MacroGard®-fed fish. After 2 weeks of feeding, the abundance of Weissella cibaria, uncultured Lactococcus sp. and Vagococcus sp. (OTUs 12–14) decreased with increasing dietary MacroGard® supplementation. Similarly, after 4 weeks of feeding, the abundance of uncultured Lactococcus sp. (OTU 24) and Vagococcus sp. (OTU 25) was reduced in fish fed the MacroGard®-supplemented diets. To the authors' knowledge, there is no information available on the effects of dietary β-glucan supplementation on autochthonous LAB in fish or mammals. In the present study, the total number of cultivable aerobic heterotrophic bacteria levels were not affected by MacroGard® supplementation, which is in accordance with the preliminary results observed in Atlantic cod (Gadus morhua) (Skjermo et al. 2006). Culture-independent analyses revealed that dietary MacroGard® reduced the number of phylotypes and the species richness of the allochthonous microbiota in the carp intestine after 2 weeks, but not after 4 weeks of feeding. For the autochthonous microbiota, dietary MacroGard® reduced the number of phylotypes, the species richness and diversity after 2 weeks, and those parameters remained reduced after 4 weeks of feeding. Sequence analysis identified that the abundance of autochthonous Aeromonas hydrophila was reduced with dietary MacroGard®, whereas the effect on the abundance of allochthonous aeromonad OTUs was more variable. These results demonstrate that the effect of MacroGard® on the bacterial communities in the intestine of carp is different, with a more pronounced effect on the autochthonous microbiota, which may be reflective of a different community composition on the mucosal surface, than in the lumen, or potentially the observed changes may be a result of structural or immunological changes in the host's epithelial brush border or mucus biochemistry. A recent carp study may support the hypothesis that immunological changes in the epithelial brush border are related to microbial changes, either as a cause or as a resultant effect. Kühlwein et al. (2013) observed that dietary MacroGard® administration at 1 and 2% led to higher numbers of intraepithelial leucocytes and a trend towards increased numbers of goblet cells in the proximal intestine of mirror carp. In addition, the present study reveals structural changes in epithelial brush border.

After 4 weeks of feeding, the distal intestinal microvilli length and density were significantly increased in fish fed diets supplemented with 1% MacroGard®, whereas after 2 weeks, no differences were observed. To the authors' knowledge, no report on the effects of β-glucans on the apical brush border morphology of teleost enterocytes has been published to date. However, a number of studies have reported similar effects with dietary mannan oligosaccharides (MOS) isolated from Saccharomyces cerevisiae yeast cell walls. Increased microvilli length was observed in red drum (Sciaenops ocellatus) when MOS and other prebiotic oligosaccharides were included in the diet at 1% (Zhou et al. 2010) and in Pacific white shrimp (Litopenaeus vannamei) with MOS supplementations from 0·2 to 0·8% but not with 0·1% (Zhang et al. 2012). In addition, increased microvilli length and density were reported in European lobsters (Homarus gammarus L.) (Daniels et al. 2010) and rainbow trout (Oncorhynchus mykiss) (Dimitroglou et al. 2009) fed MOS-supplemented diets. The authors suggested that a well-developed, healthy mucosal epithelium may be advantageous in the defence of opportunistic indigenous bacterial infections in addition to a likely improved feed utilization.

In conclusion, the present study was the first to comprehensively investigate the time-dependent effects of a dietary immunostimulating β-(1,3)(1,6)-d-glucan (MacroGard®) on both allochthonous and autochthonous microbial communities in the teleost intestine. The results indicate that dietary MacroGard® may affect the composition of the carp intestinal microbial communities. Furthermore, positive effects on intestinal microvilli length and density were also observed. Indeed, these changes at 1 and 2% MacroGard® supplementation might be contributory factors to the improved growth performance recently observed in carp fed 1 and 2% dietary MacroGard® (Kühlwein et al. 2013). Future research is required to validate this hypothesis and to determine whether the present observations occur under commercial husbandry or in different fish species. If these effects are common across different rearing conditions, different fish species and different β-glucan sources, the mechanisms involved and the extent to which these changes contribute to the host benefits reported with the application of dietary β-glucans should be determined.

Acknowledgements

The authors wish to thank Ben Eynon and Peter Bond for their technical support. In addition, the authors are thankful to Biorigin (Brasil) and Tetra GmbH (Germany) for the provision of MacroGard® and the extruded diets, respectively. This research was supported by funding from the European Community's Seventh Framework Programme [FP7/2007-2013] under grant agreement no. PITN-GA-2008-214505.

Conflict of Interest

No conflict of interest declared.

Ancillary