Bacterial Diversity at Different Sites of the Digestive Tract of Weaned Piglets Fed Liquid Diets

Bacterial diversity was studied using PCR-DGGE, cloning and sequencing. DNA was isolated from digesta samples from stomach, ileum and colon of 28 weaned piglets (Large White×Mong Cai) fed dry control feed, naturally fermented liquid feed (FE) and a liquid diet with inclusion of rice distiller’s residue feed. General bacterial diversity was described using DGGE analysis of the V3 region of 16S rDNA. The microbial populations in the stomach and the ileum were considerably influenced by the diet, while only marginal effects were observed in the colon. There was a large variation of the microbial flora in the stomach between individuals fed non-fermented diets. In contrast, animals fed diet FE had a more uniform microbial flora in the stomach and the ileum compared to the other diets. In total 47 bands from the DGGE profiles were cloned. In stomach, most frequently lactic acid bacteria were found. Feeding diet FE resulted in the occurrence of Pediococcus species in stomach and ileum. In pigs fed the other diets, Lactobacillus gallinarum, Lactobacillus johnsonii and Lactobacillus fermentum were found in stomach and ileum. Most of the sequences of bands isolated from colon samples and several from ileum matched to unknown bacteria, which often grouped within Prevotellaceae, Enterobacteriaceae, Bacteroidaceae and Erysipelotrichaceae. This study demonstrates that fermented liquid feed affects bacterial diversity and the specific microflora in stomach and ileum, which provides a potential to modulate the gut microflora with dietary means to increase the abundance of beneficial bacteria and improve piglets’ health. (


INTRODUCTION
Diarrhoea caused by enterotoxigenic Escherichia coli (ETEC) was reported as a disease associated with industrialized pig production in Asia (Ranald et al., 2000;Thuy et al., 2006).This disease leads to considerable economic losses for producers.Early weaned piglets are more susceptible to enteric disease if the housing, environment and nutrition are not optimal.In order to solve this problem, the pig producers can use antibiotics to treat piglets or use prophylactic vaccination of sows.In central Vietnam, smallholder farmers do not practice prophylactic vaccination of sows for the control of ETEC in piglets but frequently use antibiotics to treat piglets.This may have negative effects on the environment and will contribute to the development of bacterial strains that are resistance towards antibiotics (Levy, 1982;Bates et al., 1993).
Diet has an impact on the microbial flora in the pig intestine (Varel and Yen, 1997;Konstantinov et al., 2003;Konstantinov et al., 2004;Wang et al., 2007).Therefore, modulations of pig feed, such as using pre-biotic ingredients, supplementing with pro-biotics and organic acids, and fermenting liquid diets, have been applied in recent years to improve pig gut health status (Pluske et al., 2005).Fermented pig feed is generated by mixing solid carbohydrate-rich material with a liquid phase, water or a liquid co-product from food or ethanol production.During a period of incubation, microbial fermentation occurs resulting in a feed characterized by low pH, high concentration of lactic acid bacteria and organic acids.Fermented liquid feeds in particular have potential to improve pig health and performance by inhibiting the growth of pathogenic bacteria such as E. coli in the feed and in the gastrointestinal tract (Scholten et al., 1999;Hansen et al., 2000;van Winsen et al., 2001a;van Winsen et al., 2001b;Canibe and Jensen, 2003), and can easily be applied under smallholder farm conditions.Hong et al. (2009) have recently shown that weaned piglets fed a naturally fermented liquid diet (FE) and a diet with inclusion of rice distiller's residue (RDR) feed had lower pH and higher concentrations of organic acids in the stomach, ileum and mid-colon than piglets fed a dry control diet.Moreover, in weaned piglets fed diets FE and RDR, counts of lactic acid bacteria in stomach and ileum were higher and counts of E. coli and total coliforms were lower than in those fed the dry control diet.Digesta samples collected from different sites of digestive tract of weaned piglets in the study by Hong et al. (2009) were used get a more detailed picture of the dietary impact on the bacterial diversity by using molecular methods previously applied in both humans and animals (Suau et al., 1999;Simpson et al., 2000;Donskey et al., 2003).The specific aim was to investigate the bacterial diversity at different sites of the digestive tract in weaned piglets by using the culture independent DGGE method combined with cloning and sequencing for microbial identification.

Animals and experimental design
The experiment was performed at the experimental farm of the faculty of Animal Science, Hue Agriculture and Forestry University.Samples from a total of 28 pigs (Large White×Mong Cai) breed from four litters, with a body weight (BW) of 9.5 kg (SD 1.3 kg), were used in this study.At the start of the experiment (Day 0) four piglets, one from each litter, were exsanguinated and then eviscerated for collection of digesta samples.Digesta samples were collected to estimate the initial bacterial composition and general bacterial diversity.The remaining piglets from each litter were distributed randomly by sex into 3 dietary treatments, with four replicates of 2 pigs per replicate housed in the same pen, and were fed for a total of 42 days.Five hours after feeding, on the last day of the experiment, four piglets per treatment, one from each replicate, were exsanguinated by a blow to the head and then eviscerated immediately for collection of digesta samples to estimate the final bacterial composition and general bacterial diversity.

Feeds and feeding
Locally available feed resources were used to formulate three iso-nitrogenous experimental diets.Rice and maize were grinded through a 2 mm screen before mixing with the other feed ingredients.Rice distiller's residue was produced from the same batch of rice by a local rice wine producer and was delivered on a daily basis.The experimental diets were prepared in mash form and are shown in Table 1.The control diet (CO) was fed in raw form.The natural fermented diet (FE) had the same ingredient composition as diet CO and was prepared every day in a 20-liter tank by mixing the feed ingredients with warm boiled water (35-40°C) in a ratio of 1:1.5 to allow fermentation to take place.After 72 h of fermentation at room temperature (about 30°C) the fermented diet was fed to the piglets.In the diet RDR, cassava root meal, maize meal and soybean meal was replaced with rice distiller's residue.Chromium oxide was used as a digesta flow marker and was added at 5 g per kg dry matter of diet.Piglets were fed three times per day (06.00 h, 12.00 h and 18.00 h) and had free access to drinking water via automatic drinking nipples.The daily feed allowance was 4% of body weight.

Digesta sampling
In order to estimate the general bacterial diversity at different sites of the digestive tract, digesta samples from the stomach, ileum and colon of each piglet were collected at day 0 and day 42.Immediately after collection, the samples were put on ice, froozen and kept at -20°C until further analysis.
gastrointestinal tract of piglets at day 0 and after 42 days feeding of the experimental diets collected into sterile tubes, put on ice and immediately transferred to the laboratory.All of samples were stored at -20°C until DNA isolation.The genomic DNA of the microflora was extracted using the QIAamp ® DNA Stool Mini Kit.DNA concentrations were measured using a NanoDrop ® ND-1000 spectrophotometer (NanoDrop technologies, Wilmington, USA).Genomic DNA was stored at -20°C until analysis.
PCR amplification: To investigate the microbial diversity of weaning piglets, the V3 variable region of bacteria 16S rDNA was amplified by PCR using the forward primer 341F: 5'-CCTACGGGAGGCAGCAG-3' with GC clamp at the 5' end: CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCA CGG GGG GCC TAC GGG AGG CAG CAG and reverse primer 534R: 5'-ATTACCGCGGCTGCTGG-3' targeting the V3 region of the 16S rRNA genes (Muyzer et al., 1993;Simpson et al., 1999).Working solution of DNA templates were diluted to a final concentration of 30 ng μl -1 .PureTaq Ready-To-Go PCR Beads were used for PCR.The PCR mix contained 15 μM of each primer, 45 ng of DNA template in a final volume of 25 μl.
PCR was performed in MJ Mini TM Personal Thermal Cycler (BioRad) using the following program: 94°C for 5 min; 10 cycles of 94°C for 45 s; 60°C for 45 s; 72°C for 1 min (the annealing temperature was decreased of 0.5°C every cycle) and 16 cycles of 94°C for 45 s; 55°C, for 45 s; 72°C, for 1 min.PCR products were stored at 4°C before checking by electrophoresis (Sambrook and Russell, 2001).PCR products were stored at -20°C until DGGE analysis.
DGGE: Denaturing gradient gel electrophoresis (DGGE) analysis of PCR-fragments with a GC-clamp was performed by using a DCode TM Universal Mutation Detection System (Bio-Rad), using a 7% polyacrylamid-gel in 1×TAE buffer and with 30-60% linear gradients of denaturant (Muyzer et al., 1993).Electrophoresis was performed at 60°C for 17 h at 130 V.After electrophoresis, DGGE gels were stained with SYBR ® Gold dye for 30 min at room temperature, checked under UV light and photographed using the Gel DOC TM 2000 Gel Documentation System (Bio-Rad).Selected bands (see Results) were cut out from the gel using a sterilized scalpel and frozen until further analysis.

Isolation and amplification of DNA bands from the DGGE-gels
Fourty-seven DNA bands cut out from DGGE gel were cloned for sequencing.To elute the DNA from the gel, the frozen gel pieces (see above) were thawed at room temperature for 1 h, placed at -70°C for 1 h and finally thawed over night at 4°C.One μl of the eluate (about 10 ng of DNA) was used as a template in a PCR with the primers 341F and 534R (7.5 μM each) in a final volume of 12.5 μl.
PCR and detection of the PCR products was performed as described above.All PCR products were purified by using QIAquick Gel Extraction Kit of QIAGEN.PCR products were stored at -20°C until cloning.

Cloning of amplified DNA-molecules
The amplified molecules were cloned into the vector pCR ® 4-TOPO ® using the TOPO TA Cloning ® Kit for Sequencing (Invitrogen, Stockholm, Sweden) and the resulting plasmids were transformed into chemically competent E. coli according to the supplier's manual.Bluewhite selection was applied by adding 40 mg ml -1 X-gal to the LB medium.

Analyzing positive colonies
For each cloned band, five white colonies were randomly selected from the transformation plates.Cells from each selected colony were transferred to a PCR tube using a sterile toothpick.PCR was performed with PureTaq Ready-To-Go PCR Beads and the plasmid specific primers M13F (5'-GTA AAA CGA CGG CCA G-3') and M13R (5'-CAG GAA ACA GCT-ATG AC-3').For each tested colony also a frozen stock culture was prepared from an overnight culture in LBA medium (trypton 10 g L -1 , NaCl 5 g L -1 , yeast extract 5 g L -1 , ampicillin 100 mg L -1 ) at 37°C, by mixing one culture volume with one volume of concentrated glycerol.Stock cultures were stored at -70°C.
PCR was performed in MJ Mini TM Personal Thermal Cycler (Bio-Rad) using the following program: 94°C for 5 minutes followed by 35 cycles at 94°C for 30 seconds, 55°C for 30 seconds, 72°C for 30 seconds.The program was finished by a final 7 minutes extension at 72°C.The final PCR products were checked in 2% agarose gels.
All PCR products were purified by using the QIAquick Gel Extraction Kit (QIAGEN, Hilden, Germany).

Sequencing
Sequencing was performed by MCLAB (Molecular cloning laboratories; South San Francisco, USA) using the purified PCR-product as template and the primer M13F or M13R.

Analyses of sequence data
All determined sequences were compared with known strains using NCBI-BLAST2 Nucleotide search (www.ebi.ac.uk/; http://blast.ncbi.nlm.nih.gov/).In cases where only matches to uncultured bacteria were obtained, the sequences were investigated in the database of the ribosomal database project (rdp, http://rdp.cme.msu.edu/).This at least allowed a determination to the family level, in some cases even to the genus or species level.If sequences of known species were found in the rdp-search, they were aligned to the identified sequences to determine the degree of identity.Multiple alignments were performed using the ClustalW-programme (www.ebi.ac.uk/).

General microbial diversity in different parts of the digestive tract
Amplification of the V3 region yielded clear products without by-products, for all samples with a size of approximately 200 bp as determined by electrophoresis in 2% agarose gels.
DGGE-profiles of the bacterial populations were successfully generated and in Figure 1 the profile obtained from the stomach samples is shown.As can be seen, some bands were dominant in the profile, while others were faint and thus difficult to see.We checked the gels at different exposure times to visualise also weaker bands.The results of all samples from stomach, ileum and colon are schematically summarised in Figure 2.
In the stomach (st) samples, a clear impact of the diet on the microbial population was seen (Figure 1 and Figure 2A).One striking feature of the general microbial flora in animals fed diets with non-fermented material (samples prior the experiment started (day 0), CO and RDR) was the strong variation between the different individuals.In contrast, the four individuals fed FE showed a uniform banding pattern.The figure also shows that the different diets influenced the general microbial diversity in the stomach.At least the upper and the lower band of the four clearly visible bands in the DGGE-profile of the animals fed FE-diet seem to represent species that were not or only to a low amount present in individuals fed the other diets.This impression was confirmed by sequencing the according bands (see below).
In ileum (il) samples also an influence of the diets was found.Although there were more individual differences than in the colon, animals fed the FE diet had similar banding patterns, three patterns were almost identical by strong DNA bands on DGGE (e.g.il11, il12, il13) (Figure 2B).Cloning and sequencing showed that in some cases bands with similar migration pattern represented the same species, but there were a number of exceptions (see below).
The colon (co) samples looked less diverse than those from the other intestinal regions.Although there were some differences between the individuals, most of the bands that were obtained in one treatment were also found in individuals fed other diets.Considerable differences were found to the animals before starting the diets (day 0).These individuals also showed a substantial difference to each other (Figure 2C).

Identification of microorganisms represented by bands in the DGGE-gels
To obtain more detailed information about the microflora in the intestine, bands were excised from the gels, cloned and sequenced.In total 47 dominant DNA bands were investigated.For each band, five independent clones were sequenced.In ten of fifteen cases of DNA bands of stomach samples (e.g.st1, st2, st5, st7, st8, st10, st11, st12, st13, st14, Figure 2, Table 2) only one sequence per band was found.However, a substantial number of bands contained more than one sequence (Tables 2, 3 and 4).
In the stomach, most of the bands were formed from lactic acid bacteria.Several bands at similar positions in the gel but obtained from different animals had identical sequences like st1, st3, st5 and st7 in stomach, which all matched to Lactobacillus acidophilus/L.gasseri.These closely related species cannot be distinguished according to their V3-sequence (Berger et al., 2007).However, band st9 that was found in an individual fed FE matched to Pediococcus pentosaceus.Band st4 from an animal at t 0 was formed by two organisms, Staphylococcus gallinarum and L. jensenii.It has earlier been shown that sequences of different strains can show a similar running pattern in DGGE (e.g.Ercolini et al., 2001).We also found that bands that were isolated from different positions in the gel contained the same sequence.This includes the bands st8, st9 and st10, st11, st12 and st13.Bands st8 and st9 both included molecules that matched most closely to P. pentosaceus, although band st9 also included L. gallinarum and band st8 matched only to 96% to P. pentosaceus.Band st10, st11, st12 and st13 matched to P. acidilactici in the Blast searches.The results from the Blast searches with the sequences of the isolated bands in stomach are summarised in Table 2.
Among the bands from the ileum some extreme cases of different migration patterns of identical sequences were observed (Table 3 and Figure 2B identical sequences, similar to the sequence of uncultured Turicibacter, family Erysipelotrichaceae as well as il11 and il13 that represented P. acidilactici sequences.Many of these bands were obviously formed by more than one sequence, including bands in ileum and colon.Sequences matching to E. coli as those of bands il1, il2 and il3 cannot be distinguished from related enterobacteria, including Shigella flexneri or even a group of plant pathogens because sequences of the amplified V3-region of these organisms are identical (Naum et al., 2008).
In the colon samples, bands co4, co7, and co10 running to similar positions in the gel had identical sequences matching to Prevotellaceae (Figure 2C and Table 4).On the other hand, bands co11 and co12 contained identical sequences, matching to L. johnsonii.However, both co11 and co12 bands also contained sequences belonging to Prevotellaceae (Table 4).
The number of samples that matched to sequences of un-cultured and not further classified bacteria increased from stomach to ileum to colon.Band il8 in ileum (91% identity), and bands co3 and co6 in colon (85 and 94% identity, respectively) contained sequences that had sequence identities to the closest Blast hits of clearly less than 98%, i.e.where the sequences were isolated from strains that do not belong to any known bacterial species (Leser et al., 2002).

DISCUSSION
This study aimed to investigate the impact of fermented feeds on the gastrointestinal flora in piglets.We analysed the bacterial flora in groups of animals fed different diets by culture independent DGGE-analysis.There is yet little known about the presence of bacterial strains in different parts of the digestive tract of pigs fed different diets (Leser et al., 2002;Konstantinov et al., 2003;Pedersen et al., 2005;Wang et al., 2007).Most studies have focussed on enumeration of harmful and beneficial bacteria in the digestive tract of piglets but not on their identification (van Winsen et al., 2001b;Scholten et al., 2002;Canibe and Jensen, 2003;Højberg et al., 2003).However, Leser et al. (2002) found that Lactobacillus amylovorus, L. johnsonii, and L. reuteri were the most common in pigs while L. sharpeae were common in the ileum and cecum of pigs fed fermented feed.Pedersen et al. (2005) showed that feeding wet wheat distiller's grain to piglets affected the composition of the lactic acid bacteria flora in their intestine.Canibe and Jensen (2003) reported that feeding fermented liquid cereal grain feed changed the bacterial population of the stomach.
The strongest impact of the diet in the current study was seen in the stomach, where all investigated groups showed different banding patterns.Sequencing also confirmed that a Lactobacillus acidophilus and L. gasseri could not be distinguished (Berger et al., 2007).b Determination according to the rdp search.
there were different microorganisms in the intestine of animals fed different diets.Interestingly, the diet consisting of fermented feed had the highest impact on the microbial flora in the animals tested in our study.In the stomach of piglets fed fermented feed the microbial flora was more uniform than in animals fed the other diets.It has earlier been shown that the supply of nutrients in the feed, especially the content of fermentable carbohydrates and of fibre, can influence the bacterial flora in the intestine (Durmic et al., 1998;Konstantinov et al., 2003).However, diet CO and FE in the current study had the same ingredient and nutrient composition.Thus, the intestinal flora was obviously mainly influenced by the microbial flora that developed on the feed.Probably, it was only in the FE diet that a number of microorganisms was reached that was high enough to substantially influence the microbial flora in the intestine.Organisms belonging to the genus Pediococcus were only found in animals fed the diet with fermented feed (FE).We have earlier shown that bacteria belonging to this genus can dominate feed fermentations (Olstorpe et al., 2008).Thus, it is likely that these organisms arose from the naturally fermented feed.In the ileum, some influence of a Determination according to the rdp search.b E. coli and Shigella could not be distinguished on the basic of the V3 sequence (Naum et al., 2008).c Lactobacillus acidophilus and L. gasseri could not be distinguished (Berger et al., 2007).
the feed was observed.Again, diet FE seemed to have the highest impact on the bacterial flora.Although there were greater individual differences and there were several bands in the DGGE pattern in individuals fed the FE-diet that were also observed in individuals fed other diets, some bands were unique for the FE-diet, and some identified organisms were only found in the FE-group.This again included bacteria belonging to the genus Pediococcus.The colon samples showed the smallest diet dependent differences.Differences seemed rather to be due to the individual than to diet.This is in accordance with earlier studies of the faecal flora of pigs, where it has been shown that individual differences in the microflora are often bigger than those due to the nutrition (Simpson et al., 2000).Lactic acid bacteria have frequently been found in the stomach of pigs (Fuller et al., 1978;Henriksson et al., 1995;Leser et al., 2002), which is confirmed by our results.The occurrence of Pediococcus species in the pig intestine has rarely been reported and no sequences arising from this genus have been found in a clone library made of 16S rDNA sequences isolated and amplified from pig intestine (Leser et al., 2002).This again indicates that these organisms were introduced by the diets into the intestine of the pigs.Some clones matched to species that have been earlier described to be abundant in the pig intestine, including L. acidophilus, E. coli/Shigella and bacteria belonging to the Prevotella group (Leser et al., 2000;Leser et al., 2002).We also found a considerable number of organisms that did not match to any known bacterial species.This has been described in other studies using molecular detection methods and supports the contention that great parts of the microbial flora in the pig intestine are still unknown (Leser et al., 2002).
DGGE is a method that has frequently been applied to analyze the gastrointestinal microflora in pigs (Simpson et al., 1999;Simpson et al., 2000;Collier et al., 2003;Konstantinov et al., 2003).The method is relatively easy to perform and provides a good general survey about the microbial flora in the intestine.However, there are also limitations.Due to the short length of the amplified sequences several organisms cannot clearly be identified.This became obvious in the case of bands that matched to the E. coli/Shigella group.In this case it is impossible to distinguish between E. coli, which is a common bacterium in the intestine of mammals, and Shigella flexneri, which is supposed to be pathogenic (Naum et al., 2008).The occurrence of chimeric bands, due to recombination processes has also been described (von Wintzingerode et al., 1997;Leser et al., 2002).However, we did not find such bands, which might also be due to the low molecular weight of the amplified molecules.Co-migration of bands with different sequences was found for 28 of the 47 cloned bands in the current study, a phenomenon that has been observed earlier (Ercolini et al., 2001;Gafan et al., 2005).Multiple bands have been observed when molecules with attached GC-clamps were (Nübel et al., 1996), which might also account for multiple bands of the same sequence in our data (Figure 2).Anyway, our results show that the method was appropriate to detect population changes as a result of feeding fermented diets.Especially in the stomach, a great impact of feeding fermented diet on the banding pattern was observed.Sequencing revealed the presence of strains belonging to P. acidilactici and P. pentosaceus, which were not found in piglets fed the other diets.P. acidilactici has been used as a probiotic in weaning piglets resulting in improved weight gain and increased resistance against infections (Giancamillo et al., 2008).Pediococcus strains have been shown to produce bacteriocines, which may have a potential against pathogenic bacteria.Thus, our study strongly suggests that the use of a fermented diet with a microbial flora dominated by Pediococcus strains may be preferable to traditional diets.

Table 2 .
16S rRNA gene sequences of the strong DNA bands from stomach detected by DGGE and cloning technique

Table 3 .
16S rRNA gene sequences of the strong DNA bands from ileum detected by DGGE and clonning technique

Table 4 .
16S rRNA gene sequences of the strong DNA bands from colon detected by DGGE and clonning technique a Determination according to the rdp search.