Effect of fermentable carbohydrates on piglet faecal bacterial communities as revealed by denaturing gradient gel electrophoresis analysis of 16S ribosomal DNA

Authors

  • Sergey R. Konstantinov,

    1. Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, Netherlands
    Search for more papers by this author
  • Wei-Yun Zhu,

    1. Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, Netherlands
    2. Wageningen Institute of Animal Sciences, Wageningen University, Wageningen, Netherlands
    3. College of Animal Science and Technology, Nanjing Agricultural University, 210095 Nanjing, PR China
    Search for more papers by this author
  • Barbara A. Williams,

    1. Wageningen Institute of Animal Sciences, Wageningen University, Wageningen, Netherlands
    Search for more papers by this author
  • Seerp Tamminga,

    1. Wageningen Institute of Animal Sciences, Wageningen University, Wageningen, Netherlands
    Search for more papers by this author
  • Willem M. de Vos,

    1. Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, Netherlands
    2. Wageningen Centre for Food Sciences, P.O. Box 557, 6700 AN Wageningen, Netherlands
    Search for more papers by this author
  • Antoon D.L. Akkermans

    Corresponding author
    1. Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, Netherlands
      *Corresponding author. Tel.: +31 (317) 48 34 86; Fax: +31 (317) 48 38 29. antoon.akkermans@algemeen.micr.wag-ur.nl
    Search for more papers by this author

*Corresponding author. Tel.: +31 (317) 48 34 86; Fax: +31 (317) 48 38 29. antoon.akkermans@algemeen.micr.wag-ur.nl

Abstract

The effect of fermentable carbohydrates (sugar beet pulp and fructooligosaccharides) on the faecal bacterial communities of weaning piglets was analysed using 16S rDNA-based approaches. Amplicons of the V6–V8 variable regions of bacterial 16S rDNA were analysed by denaturing gradient gel electrophoresis (DGGE), cloning and sequencing. Differences in piglet faecal bacterial community structure were determined based on the Dice coefficients for pairwise comparison of the DGGE fingerprints and revealed significant changes in the faecal microbiota immediately after weaning. Piglets fed with fermentable carbohydrates showed a higher bacterial diversity and a more rapid stabilisation of the bacterial community compared with that of the animals fed with the control diet. Thirteen dominant DGGE bands were matched with sequences that showed 91–97% similarity to those derived from the Clostridium coccoides group and the Clostridium leptum subgroup. Amplicons related to Ruminococcus-like species were found in all DGGE fingerprints derived from pigs on the diet containing sugar beet pulp and fructooligosaccharides, but not in pigs on the control diet. These results indicate that these bacteria may play a role in the utilisation of dietary fibres.

1Introduction

The large intestine of pigs and other production animals is densely colonised with bacteria, but little is understood about their activity, which affects animal performance, health and food safety. There is considerable interest in understanding and influencing the intestinal microbiota, notably because of the urgent need to replace antibiotics as growth promoters in animal production. Moreover, the pig's intestinal tract is considered to be an appropriate model system for the human intestinal tract [1]. Stimulating the fermentation of specific carbohydrates in the intestine has been suggested to influence the development of the bacterial community. Such fermentation may promote the production of volatile fatty acids that have been shown to inhibit the growth of certain pathogens, e.g. Salmonella[2,3].

Using traditional culture techniques, it has been shown that the majority of the faecal and colon microbiota isolated from adult swine were Gram-positive obligate anaerobes [4]. Most of the isolates were found to belong to the genera Streptococcus, Lactobacillus, Fusobacterium, Eubacterium, and Peptostreptococcus. The Gram-negative organisms comprise about 10% of the total culturable bacteria belonging to the Bacteroides and Prevotella groups. Such culture-dependent methods are highly sensitive and accurate to monitor those viable bacteria of which the growth requirements are known. However, one of the limitations of these methods is that plating in selective media relies on the assumption that all bacterial groups show equal plating efficiency or viability. These limitations of conventional culture-dependent detection techniques have led to the development of rRNA-based approaches by which the presence and identification of the bacteria is based on the sequence diversity of the 16S rRNA gene [5,6], although this approach may be biased as well [7,8]. Recent phylogenetic analysis of the small subunit (16S) rRNA from pig intestine has revealed that the intestinal microbial community is very complex and that the majority of the bacterial species colonising the intestine have not been characterised [9,10]. A combination of polymerase chain reaction (PCR) and DNA fingerprinting techniques, such as temperature gradient gel electrophoresis (TGGE) and denaturing gradient gel electrophoresis (DGGE), has been used successfully to describe the microbial diversity in complex ecosystems [11–14], including the mammalian gastrointestinal (GI) tract [15–17].

Both quantitative and qualitative aspects of the faecal bacterial community may be sensitive to changes in environmental factors. Stress factors such as weaning, dietary change and transportation can cause a decrease in lactobacilli and an increase of coliforms in the GI tract [18]. The introduction of high levels of fibres in the diets of pigs has been found to stimulate growth of bacteria with cellulolytic and xylanolytic activities [19,20]. Moreover, changes in porcine colon and faecal bacterial population have been demonstrated when the animals were fed different diets and after introduction of an exogenous Lactobacillus strain [16,21].

Here we describe the changes in time of the predominant faecal bacterial community in weaning piglets that were fed diets containing fructooligosaccharides (FOS) and/or sugar beet pulp (SBP). The results indicated that addition of fermentable carbohydrates to the diet results in a rapid stabilisation of the bacterial community and the emergence of notably Ruminococcus obeum-like species.

2Materials and methods

2.1Experimental approach

To describe bacterial diversity in the piglet GI tract, total DNA was isolated from faecal samples and used as a template for PCR amplification. Amplicons of the V6–V8 regions of bacterial 16S rDNA were analysed using DGGE. The band position in different gels was compared using a marker which consisted of amplified V6–V8 regions from nine clones with different mobility. Additionally, a clone library of 16S rDNA amplicons (Escherichia coli positions 8–1510) was prepared from a faecal sample of a piglet which had been fed for 13 days with a diet containing SBP and FOS. Cloned amplicons that showed the same mobility upon DGGE as specific bands in the DNA-derived profile were further characterised by nucleotide sequence analysis.

2.2Animals, diets, and sampling

Three identical, but independent feeding trials with healthy conventionally raised piglets (crossbred Hypor×Pietrain) were started immediately at the time of weaning (25–28 days old). To prevent cross-contamination between litters, as well as the stress which mixing litters would cause, piglets from the same litter were kept together and therefore fed the same diet. However, it was for this reason that the experiment was repeated three times, involving nine litters in total (three litters per diet). At the beginning of each experiment, three litters containing four piglets each were offered one of the three diets with different amounts of non-digestible (fermentable) carbohydrates (control, SBP or SBP/FOS) (Table 1). The diets were composed in such a way that total energy and protein content were comparable. The setup of the trials and the number of the piglets and their codes are given in Table 2. All piglets from one litter were separated into two respiration chambers and maintained at an initial temperature of 24°C from day 1 to day 6. At 6 days, one half of the piglets were exposed to a temperature of 15°C for a week, while the other half were kept at 24°C for another week. Faecal samples were collected per rectum from days 1, 2, 5, 6, 7, 8, 9, 13 and stored at −20°C. The faecal samples from the first feeding trial were analysed within 6 months after the experiment was finished. The faecal samples from the other two feeding trials were kept at −20°C for 1 year.

Table 1.  Formulation of the diets used for the in vivo experiment
  1. aNo added copper or antibiotics.

  2. bDiamol, diatomaceous shell powder, inert material used as a marker for nutritional studies.

Ingredients (% (w/w))aControl dietSBPSBP/FOS
Maize starch51.541.8544.2
FOS002.5
SBP (10–15% sugar)0105
Fish meal (70.6% crude protein)202020
Dextrose151515
Soy oil0.51.40.9
Cellulose (Arbocel)555
Premix (maize)111
Soy isolate434
CaCO30.350.170.29
CaH2PO40.340.330.33
KHCO310.850.85
l-Lysine HCl00.040.02
dl-Methionine0.170.180.18
l-Threonine0.110.120.12
l-Tryptophan0.060.060.06
Diamolb111
Crude protein17.9718.0318
Total energy (kcal kg−1)252725262527
Table 2.  Details of piglets investigated, including diet, experiment number, environmental temperature (°C) from day 6, and litter number
  1. aExp., experiment number.

  2. bT, environmental temperature from day 6 till day 13.

Exp.aDiets
 SBP/FOSSBPControl
 PigletT (°C)bLitterPigletT (°C)bLitterPigletT (°C)bLitter
1A1151B1152C1153
1A2151B2152C2153
1A3241B3242C3243
1A4241B4242C4243
2A5154B5155C5156
2A6154B6155C6156
2A7244B7245C7246
2A8244B8245C8246
3A9157B9158C9159
3A10157B10158C10159
3A11247B11248C11249
3A12247B12248C12249

2.3DNA isolation

Three grams (wet weight) of thawed faecal samples were homogenised in 50 ml of ice-cold 0.05 M potassium phosphate buffer (pH 7.0), and aliquots of 1 ml stored at −20°C. Total genomic DNA was extracted from pig faeces by a bead-beating method as previously described [17]. Agarose gel 1.2% (w/v) containing ethidium bromide was used to check the amounts of DNA visually.

2.4PCR amplification

Primers U0968f-GC (5′-CGC CCG GGG CGC GCC CCG GGC GGG GCG GGG GCA CGG GGG GAA CGC GAA GAA CCT TAC-3′) and L1401r (5′-CGG TGT GTA CAA GAC CC-3′) [22] were used to amplify V6–V8 regions of 16S rDNA. This primer pair is specific for bacterial 16S rDNA and yields amplicons of 470-bp length. PCR was performed using the Taq DNA polymerase kit from Life Technologies (Gaithersburg, MD, USA). PCR mixtures (50 μl) contained 0.5 μl of Taq polymerase (1.25 U), 20 mM Tris–HCl (pH 8.5), 50 mM KCl, 3.0 mM MgCl2, 50 mM each dNTP, 5 pmol of the primers U968-GC and L1401, 1 μl of DNA diluted to approximately 1 ng and UV sterile water. The samples were amplified in a thermocycler T1 Whatman Biometra (Göttingen, Germany) using the following program: 94°C for 5 min, and 35 cycles of 94°C for 30 s, 56°C for 20 s, 68°C for 40 s, and 68°C for 7 min last extension. Aliquots of 5μl were analysed by electrophoresis on 1.2% agarose gel (w/v) containing ethidium bromide to check the sizes and amounts of the amplicons.

2.5DGGE

The amplicons obtained from the faecal-extracted DNA were separated by DGGE according to the specifications of Muyzer et al. [11], using a Dcode TM system (Bio-Rad Laboratories, Hercules, CA, USA). Electrophoresis was performed in 8% polyacrylamide gel 37.5:1 acrylamide–bisacrylamide (dimensions 200×200×1 mm) using a 38–48% denaturing gradient [13]. The gels were electrophoresed for 16 h at 85 V in 0.5×TAE buffer [23] at a constant temperature of 60°C. The gels were stained with AgNO3[24].

2.6Analysis of the DGGE gels

DGGE analysis of all samples was repeated twice. All gels were scanned at 400 dpi and analysed using the software of Molecular Analyst/PC (version 1.12, Bio-Rad, Hercules, CA, USA). Firstly, a number of bands per lane was assessed using bands searching algorithm within the program. A manual check was done and the DGGE fragments constituting less than 1% of the total area of all bands were omitted. When one and the same DGGE band had a value lower than 1% of the total area of all bands in the first replicate and higher than 1% in the second one we calculated an average value based on the two runs. Bands above 1% of the total area of all bands in two replicates of one sample were considered as dominant DGGE bands and included in the further analysis. The similarity between the DGGE profiles was determined by calculating a band similarity coefficient (SD) (Dice: SD=2nAB/(nA+nB), where nA is the number of DGGE bands in line 1, nB represents the number of DGGE bands in lane 2, and nAB is the number of common DGGE bands [15,16,25,26]. Secondly, as a parameter for the structural diversity of the microbial community, the Shannon index of general diversity, H[27–29], was calculated using the following function: H′=−ΣPi log PI, where Pi is the importance probability of the bands in a lane. H′ was calculated on the basis of the bands on the gel tracks that were applied for the generation of the dendrograms by using the intensities of the bands as judged by peak height in the densitometric curves. The importance probability, Pi, was calculated as: Pi=ni/H′, where ni is the height of a peak and H′ is the sum of all peak heights in the densitometric curve.

2.7Statistical analysis

For statistical analysis calculations were made to determine band similarity coefficient (SD), number of DGGE bands, and the Shannon index of general diversity. Differences between diets for these parameters were tested for significance using Tukey's Studentised range test of multiple comparisons [30] according to the following: Y=μ+Di+εij, where Y is the result, μ the mean, D the effect of the diet, and εij the error term. All statistical analyses were performed using the SAS GLM procedure [31].

2.8Cloning of the PCR amplified products

PCR was performed with a Taq DNA polymerase kit from Life Technologies using primers 8f and 1510r [32], which amplify the bacterial 16S rDNA. Amplification was carried out as described previously [17]. The PCR product was purified with the QIAquick PCR purification kit (Westburg, Leusden, The Netherlands) according to the manufacturer's instructions. Purified PCR product was cloned into a pGEM-T (Promega, Madison, WI, USA). Ligation was done at 4°C overnight followed by transformation into competent E. coli JM109. The colonies of ampicillin-resistant transforms were transferred with a sterile toothpick to 15 μl TE buffer and boiled for 15 min at 95 °C. Immediately, PCR was performed with pGEM-T-specific primers T7 (5′-AAT ACG ACT CAC TAT AGG-3′) and SP6 (5′-ATT TAG GTG ACA CTA TAG-3′) [17] to check the size of the inserts using the cell lysate as a template. Plasmids containing an insert of approximately 1.6 kb were used to amplify V6–V8 regions of 16S rDNA. The amplicons were compared with the bands of DGGE profiles that comprised more than 1% of the total area of all bands. Plasmids containing an insert corresponding to a dominant band were grown in Luria broth liquid media (5 ml) with ampicillin (100 μg ml−1). Plasmid DNA was isolated using the Wizard Plus purification system (Promega, Madison, WI, USA) and used for sequence analysis.

2.9Preparation of the clone-specific probe

Probes for clones A9 (AF349429), A17 (AF349417), and A22 (AF349420) were generated by PCR amplification of the V6 region from position 971 to 1057 of the 16S rDNA as described by Felske et al. [33]. PCR products were purified with the QIAquick PCR purification kit. The DNA was resolved in 50 μl UV sterile water and then labelled using the Prime-a-Gene Labeling System Kit (Promega, Madison, WI, USA) and [α-32P] ATP as specified by the manufacturer and used as a probe.

2.10Southern blot hybridisation

Electroblotting of denaturing gradient gels was carried out according to Muyzer et al. [13]. PCR products were transferred from the polyacrylamide gels to Hybond-N+ nucleic acid transfer membranes (Amersham International, Bucks, UK) using a Bio-Rad model SD semi-dry electrophoretic transfer cell. Prehybridisation and hybridisation were done as previously described [9]. A detection screen (Molecular Dynamics, Hercules, CA, USA) was incubated with the hybridised membrane and the probe signals were detected with a Phosphor Imager SF (Molecular Dynamics, Hercules, CA, USA).

2.11Sequence analysis

Purified plasmid DNA (approximately 1 μg ml−1) was used for sequence analysis of the cloned 16S rDNA by using a Sequenase (T7) sequencing kit (Amersham Life Sciences, Slough, UK) according to the manufacturer's specifications using the T7 and SP6 primers or the 1100r primer (5′-GGG TTG CGC TCG TTG-3′) 5′-end labelled with IRD-800. Sequences were automatically analysed on a LI-COR DNA Sequencer 4000L (LiCor, Lincoln, NE, USA) and corrected manually. These sequences were also compared to those available in public databases by using BLAST analysis of sequences from the Ribosomal Database Project [34]. The partial and complete 16S rDNA sequences were checked for chimerical constructs by the RDP CHECK_CHIMERA program [35]. None of the sequences were found to be PCR-generated chimeras.

2.12Nucleotide sequence accession numbers

The sequences of the pig faecal 16S rDNA clones were deposited in the GenBank database. The new sequences, with their accession numbers in parentheses, are: A2 (AF349418), A4 (AF349425), A4–1 (AF349426), A8 (AF349428), A9 (AF349429), A11 (AF349430), A12 (AF349415), A13 (AF349416), A17 (AF349417), A20 (AF349419), A22 (AF349420), A29 (AF349421), A29–1 (AF349422), A36 (AF349423), A39 (AF349424), A47 (AF349427).

3Results

3.1Outline of the feeding trials and analysis of the pig faecal bacterial community before weaning

A feeding trial was performed in triplicate in which a total of 36 piglets were fed 13 days one of three diets: a basic diet without fermentable carbohydrates (control), one containing 10% SBP, and one containing 5% SBP and 2.5% FOS (Table 2). On day 6 of each experiment, the ambient temperature was reduced from 24 to 15°C for half of the piglets, but no significant effect of this was found for any aspect of the microbial community structure, so it is not discussed further.

In order to assess the variability between the bacterial populations in the piglets before weaning, faecal samples were taken immediately prior to the feeding trial (day 1). A representative analysis of the PCR fragments generated with primers 968GC and 1401r and analysed by DGGE is shown in Fig. 1. This gel consists of faecal samples from 12 piglets on day 1. The number of bands in the DGGE profiles varied from 20 to 28. Several bands were common in all samples before weaning (Fig. 1). A comparative analysis of DGGE profiles of faecal amplicons from day 1 from all piglets of nine litters was performed. Dice coefficients for pairwise comparison between DGGE profiles of piglets within the same litter were calculated and found to vary between 47 and 83.7%.

Figure 1.

DGGE of PCR products of V6–V8 regions of 16S rDNA of piglet faecal samples from day 1 of experiment 2. A5, A6, A7, A8, piglets on diet SBP/FOS; B5, B6, B7, B8, piglets on diet SBP; C5, C6, C7, C8, piglets on a control diet. M, marker. I, II, III indicate bands common to most piglet faecal samples.

The Shannon index of diversity, H′, from the DGGE banding pattern of a sample was calculated on the basis of the number and relative intensity of bands on a gel lane. Thus, H′ was used as a parameter that reflects the diversity of the whole microbial community. The H′ index calculated for all piglets from all three replicate experiments at day 1 ranged from 1.06 to 1.46 (mean 1.38±0.086). These values indicated differences in the DGGE band intensity and number of bands between the individual piglets at the beginning of each experiment.

3.2Effect of dietary fermentable carbohydrates on the development of the faecal bacterial community of weaned piglets

Faecal samples from 36 individual pigs that participated in the trial (Table 2) were analysed by DGGE. The samples were collected immediately post weaning and were used to monitor the effect of non-digestible carbohydrates on the microbial community. The faecal samples collected from the piglets of the first feeding trial were used to validate the reproducibility of the DGGE fingerprints. The DNA extraction from each sample was repeated twice. After PCR amplification and separation on DGGE, two aliquots of one faecal sample had a band similarity coefficient (SD) higher than 95% (data not shown), therefore they were considered as identical. Further, amplicons of 16S rDNA DGGE from all faecal samples obtained at days 1, 2, 5, 6, 7, 8, and 13 were separated by DGGE on 18 gels. The DGGE profiles of faecal samples from pigs on one of three different diets were compared. The similarity indices of all possible pairs of gel tracks were calculated. In parallel, analyses were performed based on the Shannon index of diversity, based on the mean of the DGGE bands of 12 piglets on the same diet in time and the average number of common bands nAB within the individual piglet between days 5 and 13.

The mean values of the Dice coefficients for pairwise comparison of DGGE profiles for the piglets on the same diet examined during the three experiments are shown in Fig. 2. During the first 2 days after weaning no major changes in the SD values were observed for the 12 piglets fed with the control diet and that containing SBP and FOS. However, there were marked differences in banding patterns between day 1 and day 5 after weaning as well as between day 2 and day 5 for all faecal samples from pigs on the three diets. Differences were found both in the position of specific bands and in the number of bands (data not shown). Changes in the prominent bands were reflected in the lowest Dice coefficients for pairwise comparison between samples from days 1 and 5 and between days 2 and 5. These results suggest that during the week following introduction of the solid diet, dynamic changes were occurring in the bacterial community.

Figure 2.

Dice coefficients for pairwise comparison between piglets on the same diet in time. Diet Control, mean values of SD from the 12 piglets fed with control diet, experiments 1, 2 and 3. Diet SBP/FOS, mean values of SD from the 12 piglets, experiments 1, 2 and 3. Diet SBP, mean values of SD from the 12 piglets, experiments 1, 2 and 3. x, comparison between the DGGE profiles from different days. y, Dice coefficients for pairwise comparison in %. Bars indicate the standard deviation between the DGGE profiles from piglets on the same diet.

To examine the impact of fermentable carbohydrates in the diet on the stability of the dominant DGGE bands, Dice coefficients for pairwise comparisons were calculated between DGGE amplicons from days 5–6, 6–7, 7–8, and 8–13 (Fig. 2). A comparison of the results from piglets fed either of the three diets showed that those piglets fed the diet containing SBP and that containing SBP and FOS had a higher similarity index during a period of almost 10 days compared to piglets fed with the control diet. This suggests that the stimulation of fermentation by the presence of fermentable carbohydrates (SBP or SBP/FOS) in the diet resulted in faster stabilisation of the bacterial community at this critical period after weaning. However, the mean values of the Dice coefficients for pairwise comparison between the groups were not significantly different (P>0.05), due to the large individual variation between piglets on the same diet.

In order to compare the diversity of the predominant bacterial population in piglets fed one of the three diets, the numbers of predominant fragments in the DGGE profiles and the Shannon index of general diversity were calculated at days 1, 2, 5, 6, 7, 8, and 13 for the three experiments (Fig. 3). Statistical analysis showed that at the beginning of each experiment up to day 8, neither the number of DGGE bands nor the Shannon index was statistically different among the treatment groups. However, at the end of the experiments (day 13) the bacterial community diversity of faecal populations from piglets fed with the SBP/FOS diet (H′=1.43±0.06) and SBP (H′=1.39±0.08) was greater (P<0.05) than in piglets fed the control diet (H′=1.18±0.11) (Fig. 3A). The total number of bands at day 13 was also significantly higher (P<0.05) in the groups fed with diets containing fermentable carbohydrates (for SBP/FOS diet 32±4, for SBP diet 29±5) compared to the control diet (19±2) (Fig. 3B.) In addition, the relative mobility of each band within all 18 DGGE gels was calculated in order to find the number of common bands (nAB) between days 5 and 13. The mean of the number of DGGE bands per track from day 13 of all piglets fed the three diets was compared with the mean of the number of common DGGE bands. The mean values of nAB were 14±3 or 42% of all bands per track in the samples from pigs fed with diet containing SBP and FOS, 12±3 or 41% of all bands in samples from piglets fed with a diet containing only added SBP, and higher (P<0.05) than for pigs fed with the control diet 6±2 or 30%. The mean of the common bands for the group fed with the diet containing SBP and FOS compared with that of the group fed SBP was not significantly different (P>0.05).

Figure 3.

Analysis of DGGE banding patterns in time: days 1, 5, 6, 7, 8, 13. A: Diversity index. ♦, piglets fed with SBP; ▪, piglets fed with SBP/FOS; ▴, control. B: Number of DGGE amplicons in time according to the three diets.

These differences in the Shannon index of general diversity, total number of DGGE amplicons and also the number of common bands indicate a shift in the bacterial populations present in the GI tract that occurred due to the presence of fermentable carbohydrates in the diet.

3.3Identification of cloned 16S rDNA sequences in DGGE patterns

To study the effect of fermentable carbohydrates (SBP and FOS) on the phylogenetic diversity of the predominant bacteria, the 16S rDNA from a faecal sample of a single pig fed a diet SBP/FOS was amplified, cloned and sequenced. V6 and V8 regions of the 16S rDNA were amplified of the cell lysates of 73 transformants. The mobility of these amplicons after DGGE were compared to those obtained from rDNA of the piglets fed 13 days with the SBP/FOS diet. Thirty-six clones were assigned to one of the 16 dominant bands in the DGGE profiles, while 37 did not match any of the detectable bands. The majority (15 out of 16) of the dominant bands showed less than 97% similarity with known sequences in the database (Fig. 5). This indicates that most of these sequences were derived from new, as yet undescribed bacterial species. The phylogenetic analysis based on 16S rRNA gene showed that, except the three clones A8 (AF349428), A4 (AF349425) and A4–1 (AF349426) (Fig. 5), the remaining 13 sequences were related to species of the Clostridium coccoides rRNA group of species (cluster XIVa) and the Clostridium leptum subgroup [14].

Figure 5.

Clones, with percentage of similarity to known sequences in GenBank and the RDP database and sequence length, that were retrieved from a piglet (A5) fed a diet with FOS and SBP for 13 days (see Table 2 and Fig. 4A). Fragments that are indicated by numbers were used for V6 Southern hybridisation.

Our investigations were intended to reveal whether the different diets led to the appearance of common PCR amplicons for all piglets fed that diet by 2 weeks after weaning. Sixteen clones assigned to the 16 prominent DGGE bands from a single pig were compared with the bands in the DGGE profiles from piglets fed all three diets. Based on identical mobility within the gel, three common DGGE bands were present only in pigs that were fed with the SBP/FOS diet. These bands, indicated with 1, 2 and 3 in Figs. 4 and 5, matched with clones A22, A9 and A17 and were related to Ruminococcus sp. (97%), R. obeum (95%), and Ruminococcus sp. (str. BIE 41) (93%). In addition, an amplicon with a sequence resembling that of Ruminococcus sp. (str. BIE 41) was also common in all DGGE profiles from the pigs on the SBP diet. None of the 16 clones could be matched to the patterns derived from the piglets fed with the control diet. The high resolution of DGGE does not exclude the possibility that two different 16S rDNA sequences might migrate to exactly the same position. Visual matches for the three Ruminococcus-like species A9 (AF349429), A17 (AF349417), and A22 (AF349420) (Fig. 4) were confirmed with clone-specific V6 probe Southern blot hybridisation. A highly specific hybridisation was obtained with clone A9 (AF349429) related to R. obeum (Fig. 4B). Similar results were obtained with probes A17 and A22 (data not shown).

Figure 4.

A: DGGE of PCR products of V6–V8 regions of 16S rDNA of faecal samples at 13 days after weaning, experiment 2. A5, A6, A7, A8, piglets on diet SBP/FOS; B5, B6, B7, B8, piglets on diet SBP; C5, C6, C7, C8, piglets on a control diet (Table 2). M, marker. Fragments that are indicated by numbers were identified by the clone library and V6 Southern hybridisation. The origin of the fragments and the corresponding clones are: 1, A22 Ruminococcus sp. (AF349420); 2, A9 R. obeum (AF349429); 3, A17 Ruminococcus sp. (str. BIE 41) (AF349417). B: Southern hybridisation of an electroblotted DGGE (A) with probe derived from clone A9 R. obeum (AF349429) indicated with number 2. The PCR products were loaded as described for panel A. 2, place of hybridisation of clone A9 with itself; A5, A6, A7, A8, place of specific hybridisation with PCR amplicons derived from piglets that were fed for 13 days with a diet containing SBP and FOS; B5, B6, B7, B8, piglets on diet SBP; C5, C6, C7, C8, piglets on a control diet (Table 2).

4Discussion

This study describes the application of DGGE, cloning, sequencing and specific V6 hybridisation of 16S rDNA to monitor the effect of diet containing non-digestible fermentable carbohydrates (SBP and FOS) on the development of the faecal bacterial community of weaning piglets. The results show that diet containing SBP with or without FOS leads to a higher bacterial diversity. A proposed mechanism is that non-digestible carbohydrates reach the large intestine and provide a source of energy for those members of the bacterial community which can utilise the dietary fibres comprising predominantly pectin and cell walls.

Recently, it has been shown that both DGGE and TGGE are sensitive tools that can be used to demonstrate the differences between bacterial communities from different individuals [15–17]. Zoetendal et al. [17] reported that TGGE profiles of 16S rDNA amplicons from faeces obtained from two unrelated individual persons were host-specific and stable over a period of at least 6 months. Similar results have been described for pig faecal microbiota using an approach based on DGGE and unique amplicons specific for each pig were observed even when the piglets were fed with the same diet [33]. This study also demonstrated a significant stability of the individual faecal bacterial population in pigs in the age period between 28 and 49 days and weaned at approximately 3 weeks. In the present study, we found a significant change in the pig faecal bacterial community after weaning at approximately 28 days.

We found that the DGGE profiles of 16S PCR amplicons were not affected by reducing the environmental temperature from 24°C to 15°C. To our knowledge this is the first attempt to monitor changes in the faecal bacterial community structure prior to environmental stress, and further experiments with different temperatures or other forms of stress are needed to confirm these results.

The pig's intestinal microflora undergoes a rapid ecological succession during the period from birth to weaning [36]. During and after birth, the young animal becomes contaminated with a variety of microbes from the birth canal and the immediate environment [4]. The microflora remains fairly stable in terms of species after this initial colonisation, and for as long as the piglets receive sow milk [37]. However, the introduction of solid food causes major qualitative and quantitative alterations in the microflora. For example, strict anaerobes such as Bacteroides become established in the large intestine, and this corresponds with a decline in the number of facultative anaerobe organisms [38].

It has also been shown that the type of diet can influence the bacterial community structure in the colon of pigs [19–21]. The present results indicate that faecal microbiota from the piglets fed with a diet containing the fermentable carbohydrates SBP and FOS had more stable DGGE profiles from about day 5 of weaning. The DGGE profiles on day 5 and day 13 from piglets fed with a diet containing either SBP and FOS or SBP contained approximately twice as many common 16S rDNA amplicon bands compared with that from pigs fed the control diet. The higher number of common DGGE bands in piglets fed with the SBP/FOS diet may reflect an increasing diversity of the pig microbiota. However, an alternative explanation may be the appearance of species with multiple copy number of slightly different 16S rDNA operons [22,39].

Nearly half of all clones in our 16S rDNA library did not match any visible bands in the DGGE profile. DGGE visualised only the dominant bacterial fraction in pig faeces, while a cloning approach randomly selected 16S rDNA amplicons. It has been shown previously that all ribotypes matching clearly visible fingerprint bands only represent approximately one-half of all of the rRNA extracted from the soil [40]. This suggests that a large number of less dominant bacteria do not form detectable DGGE bands, although some of them can be selected using a cloning approach [17].

The 13 sequences corresponding to dominant bands in the DGGE profile of pigs fed with the SBP/FOS diet were related to two phylogenetic groups: the C. coccoides group 11 sequences and the C. leptum subgroup two sequences. The sequences fell into cluster XIVa [41], which included many Clostridium spp., Butyrivibrio spp., and Ruminococcus spp. Three clones related to R. obeum-like amplicons were found on three different positions in DGGE profiles. Similar results were observed for Ruminococcus-like sequences isolated from human faeces [17] which were found in different positions on the TGGE fingerprint. The result suggests an abundance of species related to Ruminococcus in pig faeces. The genus Ruminococcus consists of anaerobic cocci and the members of this group have been isolated previously from both human [42] and pig faeces [20].

Four of the clones, which corresponded with four dominant bands, showed the highest similarity with Clostridium and Eubacterium spp. Since their partial sequences showed a similarity lower than 97%, it is likely that they represented a new, as yet uncultured Clostridium sp.

Beyond the individual variations between DGGE profiles from pigs on the same diet there were also some bands in common. V6 Southern hybridisation confirmed the presence of three Ruminococcus-like species in the predominant DGGE amplicons derived from pigs fed with diet SBP/FOS.

It can be concluded that the diet containing fermentable carbohydrates affects the bacterial diversity in piglet faeces approximately 2 weeks after its administration. It has been shown that the pig microbiota changes quite dramatically after weaning, and that the addition of non-digestible but fermentable carbohydrates can lead to a higher bacterial diversity and more rapid stabilisation of the microbial community.

Acknowledgements

This research was partly supported by the Product Board Animal Feed, The Netherlands. S.R.K. is grateful for a fellowship from the Wageningen Institute of Animal Sciences, and W.-Y.Z. thanks the National Natural Sciences Foundation in China for its support (grant Nos. 39870588 and 30025034). We also thank Wilma Akkermans-van Vliet and Hans Heilig for technical assistance and advice for all the analyses and Dr. John A. Patterson for critically reading the manuscript.

Ancillary