| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ANIMAL NUTRITION |
,2
* University of Aarhus, Faculty of Agricultural Sciences, Department of Animal Health, Welfare and Nutrition; and
Department of Genetics and Biotechnology, PO Box 50, 8830 Tjele, Denmark
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: digestive tract • fermented cereal • growth • liquid diet • microbiota • piglet
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Lindecrona et al., 2003; Boesen et al., 2004). However, the effects of fermented liquid feed on growth performance are more variable (Russell et al., 1996; Jensen and Mikkelsen, 1998; Lawlor et al., 2002). Possible contributing factors to the impaired growth performance of fermented liquid feed fed-pigs are microbial degradation of free amino acids in the mixture before feeding and low palatability (Brooks et al., 2001; Pedersen, 2001). Low pH combined with a high concentration of certain fermentation metabolites (i.e., acetic acid and biogenic amines) has been suggested to impair palatability of fermented liquid feed, particularly when fed to piglets (Brooks et al., 2001; Moran, 2001).
It can then be hypothesized that the palatability of fermented liquid feed can be improved by feeding fermented liquid feed with greater pH and decreased concentration of acetic acid and biogenic amines. These characteristics are found in feed prepared by fermenting the cereal grains alone and adding the remaining dietary ingredients immediately before feeding. Furthermore, a strategy like this minimizes the time available for the dietary microflora to decarboxylate free amino acids present in the mixture and has shown good results on growth performance traits in pigs (Scholten, 2001; Pedersen et al., 2002; Pedersen, 2006).
The current study was designed to investigate 1) the microbial and nutritional characteristics of liquid feed based on fermented cereal grains as compared with traditionally fermented liquid feed and dry feed and 2) whether the former fermentation strategy compared with the latter could improve growth performance of piglets while maintaining the beneficial effects on gastrointestinal ecology.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Animals and Housing
One hundred and twenty piglets (Danish Landrace x Yorkshire x Duroc) from 40 litters were used. Piglets were weaned at 28 ± 1 d of age and a BW of 8 kg (SD = 1.1) and moved to pens (184 x 82 cm, of which 82 x 82 cm was slatted) where they were housed during the whole study. The animals were allotted based on litter, sex, and initial BW to each of the dietary treatments. There were 24 pens with 5 pigs per pen. No physical contact between pigs housed in different pens was allowed. The temperature of the nursery was maintained at 28°C during the first 2 wk, 24°C during the following 2 wk, and 20°C during the last 2 wk of the experiment.
Diets and Feeding
Preexperimental Period.
Three dietary treatments were designed using a nonpelleted, nonheated weaner diet (Table 1) ground in a roller mill. The first diet consisted of the diet fed as a dry meal (DRY). The second diet, fermented, liquid cereal grain feed (FLG), consisted of the dietary cereal grains (barley and wheat) and water mixed in a ratio of 1:2.5 (wt/wt) in a closed tank. The mixture was agitated during 5 min every hour. Three times daily, at 0730, 1430, and 2100, 50% of the mixture stored in the tank was removed, discarded, and replaced with an equal amount of fresh cereals and water. For the third diet, termed fermented liquid feed (FLF), all dietary ingredients and water (at approximately 15°C) were mixed in the ratio 1:2.5 (wt/wt) in a closed tank and the same procedure described above for the FLG diet was followed.
|
The animals fed the DRY diet were allowed ad libitum access to feed during the whole study and those on the FLG and FLF diets were fed manually twice daily, at 0730 and 1430, so that the whole ration was consumed within approximately 1 h. The DRY group had ad libitum access to feed to simulate common practice in pig production, and the FLG and FLF groups were fed restrictively to avoid further fermentations in the trough. All pigs were fed in troughs with a length of 82 cm. Additional fresh water was available for all piglets from nipple drinkers.
Experimental Procedure
Eight samples of the fermented cereal grains, the FLG diet, and the FLF diet were taken weekly in the morning before the fresh cereal grains or compound feed and water were added to the tanks. The pH, microbial composition, and the concentration of organic acids and of ethanol were measured in all samples. Terminal RFLP (T-RFLP) profiles, DM content, nitrogen, lysine, threonine, methionine, cystine, sugars, and nonstarch polysaccharides (NSP) were measured in some of the samples.
On d 14 and at a mean BW of 9.1 kg (SD = 1.60), 1 littermate from each pen was killed, 3 h after the morning meal, with an overdose of pentobarbital (Faculty of Life Sciences, Frederiksberg C, Denmark) followed by exsanguination. The gastrointestinal tract (GIT) was immediately removed and divided into 8 segments: stomach, 3 equal (length) segments of the small intestine, cecum, and 3 equal (length) segments of the colon including the rectum. The total contents of each segment were weighed and the pH was measured within 5 min. Microbial determinations were performed immediately in digesta from the stomach, the caudal segment of the small intestine, the cecum, and the midcolon. Samples from all 8 segments of the GIT were stored at –20°C to be analyzed for DM (5- to 10-g samples), and for short-chain fatty acid (SCFA), lactic acid, and succinic acid concentrations (approximately 10-g samples). Samples (2 g) from the stomach and midcolon were collected and stored at –20°C for T-RFLP analysis.
The remaining piglets were used to measure the effect of the diets on growth performance during the first 6 wk postweaning.
Feed intake by pen was recorded daily and individual BW was recorded weekly.
Analytical Methods
Dry matter content of digesta was determined by freeze-drying the samples. To express the results of chemical analyses of the diets in DM percentage, DM was determined by drying the samples at 103°C to constant weight (European Union, 1971). Ethanol was determined according to Beutler (1984). Total nitrogen was determined by the Dumas method using an elemental analyzer, model CNS-2000 (Leco, St. Joseph, MI), as described by Hansen (1989), and amino acid analyses were carried out according to Mason et al. (1980) and Commission Directive 2000/45/EC of the European Union (2000). Sugars (glucose, fructose, and sucrose) and fructans were analyzed by the enzymatic-colorimetric method of Larsson and Bengtsson (1983). Non-starch polysaccharides were determined by a modification of the Uppsala procedure and that of Englyst et al. (1982), as described by Bach Knudsen (1997).
The concentrations of SCFA, lactic acid, and succinic acid were measured as described by Jensen et al. (1995), with some modifications. Ten grams of intestinal sample was diluted 10-fold with a 0.028 M sodium hydroxide solution containing 11.1 mmol/L of internal standard (2-ethylbutyric acid, Sigma-Aldrich Denmark A/S, Vallensbæk Strand, Denmark) and homogenized for 2 min. One milliliter of the diluted sample was extracted by adding 0.5 mL of concentrated HCl and 2 mL of diethyl ether and vortex-mixing for 30 s. After centrifugation (3,000 x g for 10 min), 50 µL of the ether layer was transferred to a 100-µL vial and 10 µL of the derivatization reagent N-methyl–N-t-butyldimethylsilyl-trifluoroacetamide (Sigma-Aldrich Denmark A/S) was added. The reaction mixture was vortex-mixed and incubated at 80°C for 20 min, followed by a further incubation at room temperature for 48 h. The standard mixture of SCFA was the same as that described by Jensen et al. (1995). Quantification of SCFA, as well as lactic acid and succinic acid, was performed on a Hewlett Packard gas chromatograph (Model 6890, Hewlett Packard, Agilent Technologies, Naerum, Denmark) equipped with a flame-ionization detector and a 30-m ZB-5 column with an internal diameter of 0.32 mm and coated with 5%-phenyl 95%-dimethylpolysiloxane with a film thickness of 0.25 µm. The samples were injected with an autoinjector (Model G 1513A, Hewlett-Packard), and the chromatograms were integrated using HP GC ChemStation software (Agilent Technologies). Detector and injector temperatures were set to 250°C. The carrier gas was helium, with a pressure of 62.6 kPa. A sample volume of 2 µL was injected with a split ratio of 20. The compounds were eluted with a temperature gradient of the following shape: held at 70°C for 3 min, increased to 110°C at 10°C/min, further increased to 290°C at 20°C/min, and held for 5 min.
The concentrations of tyramine, putrescine, cadaverine, and histamine in the diets were determined as follows: 90 mL of trichloroacetic acid was added to the samples (10 g) and the mixture was homogenized for 30 s using an Ultra Turrax homogenizer (Buch and Holm A/S, Herley, Denmark). The samples were transferred to glass funnels with 50 mL of water and filtered on a Whatman GF/C filter paper under vacuum. The filtrate was transferred to a 200-mL volumetric flask and diluted to volume with water. Five milliliters of the solution was filtered through a 0.45-µm filter. The biogenic amines were separated by gradient elution using reverse phase HPLC chromatography with a Waters NOVA-PAK C18, 5 µm, 3.9 x 150 mm column (Waters A/S, Hedehusene, Denmark), postcolumn-derivatized with o-phthaldehyde, and detected with a fluorescence detector (Model 420, Waters). The mobile phases consisted of A) 0.1 N sodium acetate (adjusted with 100% acetic acid to pH 4.5) containing 1.172 g of L-octanesulfonic acid sodium salt (Sigma Aldrich, Brøndby, Denmark)/L, and B) 0.2 N sodium acetate (pH 4.5) containing 300 mL of methanol and 0.351 g of L-octanesulfonic acid sodium salt/L. The gradient used had the following profile: 0 to 30 min, 80% A and 20% B; 30 to 58 min, 0% A and 100% B; 58 to 71 min, 80% A and 20% B.
Microbiological Determinations.
Feed samples (10 g) were transferred into flasks containing 90 mL of peptone water containing 10 g/L of Bacto peptone (Merck, Darmstadt, Germany) and 1 mL/L of Tween 80 (Merck). The suspension was transferred to a plastic bag and homogenized in a stomacher blender (Interscience, St. Nom, France) for 2 min. Digesta samples (8 g, SD = 2.6 g) were transferred rapidly after collection, under a flow of CO2, into flasks containing 90 mL of a prereduced salt medium (Holdeman et al., 1977). The suspension was then transferred to a CO2-flushed plastic bag and homogenized as described for the feed samples. Then, 10-fold dilutions were prepared in peptone water for the feed samples and in prereduced salt medium for the digesta samples by the technique of Miller and Wolin (1974). The samples (0.1 mL) were plated on selective media. Lactic acid bacteria were enumerated on de Man, Rogosa, and Sharpe agar (Merck) after anaerobic incubation at 37 ± 1°C for 2 d and at 20°C for 3 d. Enterobacteriaceae were enumerated on McConkey agar (Merck) after aerobic incubation at 37 ± 1°C for 1 d. Yeasts and molds were enumerated on malt chloramphenicol agar [10 g/L of glucose (Merck); 3 g/L of malt extract (Merck); 3 g/L of yeast extract (Merck); 5 g/L of Bacto peptone (Merck); 50 mg/L of chloramphenicol (Sigma-Aldrich Chemie GmbH, Steinheim, Germany); and 15 g/L of agar (Merck)] following aerobic incubation at 37 ± 1°C for 2 d and at 20°C for 3 d. Incubation of samples at 37°C and 20°C was carried out to obtain 2 groups of microorganisms, the former growing best at the temperature of the GIT of piglets, and the latter at the temperature of the liquid feed.
Terminal-RFLP analysis was performed for finger-printing microbial communities, following the procedure described by Højberg et al. (2005). The peaks of the T-RFLP profiles were subjected to tentative identification by comparing them with the lengths of terminal restriction fragments (T-RF) obtained by in silico HhaI digestion of 16S rRNA gene sequences of a bacteria culture collection comprising a) approximately 100 commensal bacterial isolates, b) a clone library comprising approximately 350 16S rRNA gene clones (operational taxonomic units or otus), and c) selected GenBank sequences of bacterial type strains. Most of these bacteria and clones have been isolated or derived from the GIT of pigs. However, strains of lactic acid bacteria isolated from fermented feed and food products were also included as references for the T-RFLP analysis. According to the analysis, some of the peaks in the electrophoretograms seemed to represent unique strains, but often several strains rendered identical T-RF. Clearly, the reference collection of 16S rRNA gene sequences did not cover all bacterial species present in the GIT of pigs as well as in the feedstuff, and therefore some sample T-RF did not have a matching counterpart.
Calculations and Statistical Methods
When feed or digesta samples had counts below detectable levels, the minimum detectable level was applied. When the minimum detectable level is applied to one treatment, statistical differences between treatments only indicate minimal differences, and when it is applied to several treatments for a given microbial group, the P-value is only approximate. Before carrying out statistical analysis of microbial counts, logarithmic conversion of the data was performed.
The effect of diet on digesta over a range of intestinal segments was analyzed according to the following normal mixed model:
 |
where Y = the observed response; µ = the overall mean; 
= the effect of diet, i = 1, 2, 3; 
= the effect of segment, j = 1, ..., 5; () = the effect of the interaction between diet and segment; 
= the effect of block, defined as 3 pens with 5 piglets in each and each pen assigned to each of the dietary treatments (k = 1,..., 8); U = the variance component that accounts for the correlation between measurements made on the same animal (p = pig), ~N (0, 
ij[
l(i 
j) + l (i = j)]), a heterogeneous compound symmetry variance structure was assumed; and 
= the residual error, ~N (0,2) represents the unexplained random error.
The effect of diet on ADG and ADFI, with pen as the experimental unit, was estimated using a simple ANOVA based on the following GLM:
 |
where Y = the observed response, µ = the overall mean; = the effect of diet; = the effect of block; ß = the effect of initial BW IWij; and = the residual error, ~N (0, 2),
The effect of diet on G:F, with pen as the experimental unit, was estimated using a simple ANOVA based on the following GLM:
 |
where Y = the observed response, µ = the overall mean; = the effect of diet; = the effect of block; and = the residual error, N (0, 2).
The analyses were performed with SAS software (SAS Inst. Inc., Cary, NC). When there was an overall effect of diet, at an of P = 0.05, differences between means were compared pairwise using an F-test.
The T-RFLP profiles of samples from the stomach and midcolon were compared pairwise by log-linear models. The P-values were computed by Monte Carlo simulation. Fisher’s test was applied to compare pairwise the frequency of a fragment between treatments.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
No statistical comparisons among groups of samples were performed because there was only one independent replicate of each sample type sampled at various time points.
The pH of the fermented cereals was 3.85 (Table 2). Mixing the fermented cereals with the remaining ingredients of the diet to obtain the FLG treatment resulted in a greater pH, with an average value of 5.00. The pH of FLF was 4.45. Fermenting only the cereals promoted the growth of yeasts able to grow at 20 and 37°C to a greater extent than fermenting the whole diet (FLF), whereas the density of lactic acid bacteria was similar in both treatments. Fermenting the cereals alone (< 3 log cfu/g) or fermenting the whole diet (< 3.5 log cfu/g) reduced the numbers of Enterobacteriaceae compared with the DRY diet (5.4 log cfu/g). The numbers of Enterobacteriaceae in the FLG diet were below detectable levels (3 log cfu/g) in 7 of 8 samples, whereas 4 out of 8 FLF samples contained detectable numbers. Molds were not detected in any of the samples (detection level = 3 log cfu/g). The FLF diet had the greatest concentration of acetic acid and lactic acid, whereas the fermented grain alone and the FLG diet contained the greatest concentrations of ethanol (Table 2).
|
). Terminal restriction fragments with sizes 263 to 264, 404 to 405, 586 to 588, and 589 to 591 bp were present at high frequency (
3 out of 7 samples per group) in the fermented cereal grains and the FLG diet, and at low frequency (frequency 
2) in the FLF samples, whereas fragments with sizes 406 to 407, 624, and 649 to 650 bp were present at high frequency (frequency 4) in the FLF samples and at low frequency (frequency 2) in the other 2 groups. At the same time, several fragments were common for all 3 diets; that is, they were present at a high frequency in all 3 sample types. The most frequent fragments present in almost all samples analyzed were a fragment of 598 bp, representing (for example) Lactobacillus pentosus, Lactobacillus plantarum, Lactobacillus paraplantarum, Lactobacillus brevis, and potentially Lactobacillus acidophilus (597 bp), or Lactobacillus coryniformis (599 bp) and a fragment of 604 to 605 bp representing (for example) Lactobacillus sakei and Lactobacillus curvatus ssp. melibiosus. The data also revealed that most of the identified T-RF belonged to the lactic acid bacteria group and that a large proportion of the fragments present at high frequency in all or some of the diets could not be identified by the approximately 500 reference sequences included in the analysis.
|
). The concentration of sugars decreased after fermentation of the whole diet and the concentration of total NSP was slightly decreased in the FLG and FLF diets compared with the DRY diet. The total concentrations of the biogenic amines measured were approximately 71 mg/kg of DM in the DRY diet, approximately 334 mg/kg of DM in the FLG diet, and 1,186 mg/kg of DM in the FLF diet, with cadaverine being the amine measured at the greatest concentration in all diets.
|
The amount of gastric fresh digesta collected from the piglets fed the 3 diets was similar (P 0.34) ranging between 327 and 379 g (data not shown). All pigs had digesta along the GIT, which indicates that the pigs had consumed feed on the day they were killed and the previous days. No differences in pH along the GIT among diets were detected (data not shown).
The counts of lactic acid bacteria able to grow at 37°C were greater in the caudal small intestine of the DRY-fed piglets than in those fed the FLF diet (P < 0.05; Table 5). In the midcolon, the numbers were greatest in the DRY-fed pigs (P < 0.01). The counts of lactic acid bacteria able to grow at 20°C in the stomach and caudal small intestine were greater in the FLF and FLG groups than in the DRY group (P < 0.05), whereas no differences were detected in the hindgut. The number of yeasts able to grow at 37°C in the stomach and caudal small intestine was greatest in the FLG group (P < 0.01). In the cecum and colon, the differences were only significant (P < 0.05) between the pigs fed the FLG and the FLF diets. The piglets fed the DRY diet had the lowest numbers of yeasts able to grow at 20°C along the whole GIT, and the animals fed the FLG diet had the greatest. Although the Enterobacteriaceae counts in the small intestine and hindgut were numerically lowest in the FLG group, the values of the 3 diets were not different (P > 0.05).
|
. The T-RFLP profiles of stomach samples from the 3 dietary groups were computed pairwise by log-linear models. The P-values computed by Monte Carlo simulation showed a difference (P < 0.05) among the 3 contrasts. Fisher’s test showed that the DRY-fed piglets had the greatest frequency of a fragment of 254 to 255 bp, and the lowest frequency of a fragment of 596 to 598 bp; the FLG group had the greatest frequency of a fragment of 581 to 583 bp; and the FLF had the greatest frequency of fragments with sizes 406 to 407 and 606 to 608 bp (for tentative identification of the T-RF, see Table 6). The T-RFLP profiles of digesta samples from the midcolon showed a difference (P < 0.001) between the DRY and FLG groups and a tendency (P < 0.10) to a difference between the DRY and FLF groups, whereas the FLG and FLF groups showed similar profiles. The DRY group had lower frequency of fragments of 233 to 234, 236 to 237, 574 to 578, and 591 to 593 bp than the FLG group, and greater frequency of a fragment of 365 to 366 bp and a decreased frequency of a fragment of 816 to 819 bp than the FLF group. The number of T-RF was 45, 35, and 43 in the stomach and 41, 49, and 49 in the midcolon for the DRY, FLG, and FLF diets, respectively (data not shown).
|
0.01; Table 7). The concentration in the cecum in the FLF group was greater than in the FLG group (P < 0.05), with the same tendency (P < 0.10) in the caudal small intestine. The gastric concentration of ethanol was lowest in the piglets fed the DRY diet (P 0.05). In the stomach and midsmall intestine, ethanol concentration was greatest in the FLG-fed piglets, and, in the caudal small intestine, greater than in the FLF group (P < 0.01). The concentration of the sum of acetic, propionic, and butyric acids in the caudal small intestine was greater in the FLG group than in the animals fed the DRY diet (P < 0.05). In the cecum, the concentration was greater in the pigs fed the DRY diet than in those fed the FLF (P = 0.05); in the midcolon, it was greatest in the DRY group (P < 0.05).
|
Three piglets from the DRY group and 2 from the FLF group died during the experiment. The effect of diet on growth performance showed the greatest ADG for the DRY-fed piglets in the period 14 to 42 d and for the entire experimental period (Table 8). The FLG fed-piglets had a greater ADG than those fed FLF during 14 to 42 d postweaning, and showed the same tendency (P < 0.10) for the 6-wk postweaning period. Feeding the DRY diet resulted in the greatest ADFI in all the periods tested. Although the ADFI values in all periods measured were numerically greater after feeding the FLG diet than the FLF diet, no differences were detected (P > 0.05). The G:F during the 6-wk postweaning period tended to be greater (P < 0.10) for the DRY group compared with the FLG and FLF groups.
|
|
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The chemical and microbial characteristics of fermented liquid feed and fermented, liquid cereal grains have previously been reported to differ (Scholten et al., 2001a,b), which is confirmed in the current study. The observed pH value of the fermented, liquid cereal grains below 4 is in line with values obtained by Scholten (2001) and Moran et al. (2006) when fermenting liquid milled wheat. Scholten (2001) measured a pH value of 3.9 after 24 h of incubation and a pH value of 3.7 after 48 h of incubation at 24°C. Cereal grains have a lower buffering capacity than compound diets (Scholten et al., 2001b; Lawlor et al., 2005), which allows a faster pH decrease with incubation time in the former.
Numerous chemical and physical factors, such as temperature, acid concentration, pH, and buffering capacity influence the rate and extent of growth of various microorganisms in liquid feed, and consequently their sequence of appearance during fermentation (Gänzle et al., 1998; Canibe et al., 2001; Brandt et al., 2004). The lower concentration of lactic acid and acetic acid in the FLG diet compared with the FLF is probably a result of differences in the mentioned characteristics of the matrix (i.e., pH, buffer capacity). Brandt et al. (2004) observed that acidification of dough resulted in a major shift of the microbial population from lactobacilli to yeasts. The low pH reached in our fermented cereal grains most likely resulted in growth inhibition of lactic acid bacteria, thereby further reducing acid production (Fleet and Mian, 1987; Gänzle et al., 1998; Brandt et al., 2004). Yeasts are acid tolerant (Gänzle et al., 1998; Brandt et al., 2004) and continue to grow if fermentative carbohydrates remain after lactic acid bacteria are inhibited by low pH values. The greater number of yeasts and the greater concentration of ethanol in the FLG compared with the FLF observed in the current study could be explained by the mentioned dynamics during fermentation. Identification of the yeasts present in the mixtures (at least to the species level) would be needed in order to speculate on the effects of feeding liquid feed containing these microorganisms to pigs.
In the current study, a T-RF of 598 bp tentatively identified as L. plantarum was present in all dietary samples. However, other results from our laboratory (Shlimon et al., 2006) clearly indicated Pediococcus pentosaceus as the most dominant bacterium in fermented liquid feed and fermented, liquid cereal grains (barley and wheat). The previous result indicates that the dominating bacterial population in FLF/fermented, liquid cereal grains is not the same every time a new fermentation is initiated, even when the composition of the fermented ingredients is similar. A T-RF of 604 to 605 bp tentatively identified as L. sakei or L. curvatus ssp. melibiosus was also detected in almost all samples. Lactobacillus sakei, although isolated in sourdough (Valcheva et al., 2005), is a food-associated organism found naturally in fresh meat and fish and considered a transient inhabitant of the human gut (Walter et al., 2001; Chaillou et al., 2005).
The concentrations of biogenic amines, which are microbial metabolites of amino acid decarboxylation, were, as expected, greatest in the FLF and lowest in the DRY diet.
Gastrointestinal Tract
In the current study, no effect of feeding FLF on pH along the GIT of piglets was detected compared with feeding dry feed. The gastric pH of the DRY-fed piglets was relatively low (pH 3.8, SEM = 0.26), which makes further reductions difficult to achieve. In line with our results, Moran (2001) measured an average gastric pH of 3.9 in piglets fed dry feed, and did not detect differences (P > 0.05) between the animals fed dry feed and those fed FLF.
Despite the greater number of lactic acid bacteria able to grow at 37°C in the FLF and FLG diets compared with the DRY diet observed in the current study, there was no difference in the number of these bacteria in the stomachs of the animals fed the 3 diets. However, greater counts of lactic acid bacteria able to grow at 20°C were measured in the stomach and cranial small intestine of piglets fed the FLF and FLG diets compared with the DRY group, which is in line with previous data with growing pigs (Canibe and Jensen, 2003). The numbers of these 2 lactic acid bacteria groups in the DRY-fed animals is likely the reason for the results observed. The number of lactic acid bacteria able to grow at 37°C was already high in the stomachs of the DRY group (9.0 log cfu/g of digesta), which makes further increments difficult. However, the counts of lactic acid bacteria able to grow at 20°C was much lower in these animals (< 6.9 log cfu/g of digesta) and therefore the high number of lactic acid bacteria able to grow at 20°C loaded into the stomach of the animals by feeding with FLG and FLF could then be measured at this site. The lower counts of lactic acid bacteria able to grow at 37°C in the colon of the FLF and FLG groups compared with the DRY group is in line with previous results obtained with growing pigs (Hansen et al., 2000; Canibe and Jensen, 2003) and piglets (Mikkelsen and Jensen, 1997, 1998). As suggested by Canibe and Jensen (2003), a reduction of available substrates for microbial fermentation in the caudal segments of the GIT, due to disappearance of easily fermentable substrates in the feed, could partially explain the reduced number of lactic acid bacteria at this site. Lower counts of total anaerobes and lower fermentation capacity have been measured in the caudal segments of the GIT of FLF-fed pigs compared with those fed dry feed (Moran, 2001; Canibe and Jensen, 2003; Højberg et al., 2003). Moreover, the lower concentration of SCFA in the midcolon of piglets fed the FLG and FLF diets compared with those fed the DRY diet further supports the hypothesis of a lower microbial activity in the caudal GIT of animals fed the FLG and FLF diets. Another contributing factor to the lower microbial activity in the caudal GIT of FLF and FLG fed-animals could be a smaller load of substrate to the microflora inhabiting this site due to an enhanced digestibility in the small intestine, as suggested by Urlings et al. (1993).
The relative numbers of yeasts along the GIT of the piglets fed the 3 dietary treatments closely reflected those measured in the diets, and at the same time, were supported by the concentration of ethanol both in the diets and digesta from the GIT. The greater concentration of ethanol in the caudal small intestine compared with the midsmall intestine suggests that ethanol concentration did not only reflect the ethanol entering the GIT with the diet, but that production of ethanol had occurred at this site. Although heterofermentative lactic acid bacteria can produce ethanol, yeasts are the main producers (Damiani et al., 1996). The present data on ethanol suggest a greater density and activity of yeasts in the GIT of piglets fed the FLG diet compared with those fed the DRY or FLF diets. Some yeast species are used as probiotics (Bontempo et al., 2006), whereas others can cause disease (Schaller et al., 2005). Therefore, knowledge of the taxonomy and metabolism of the yeasts present in fermented diets and along the GIT of pigs is necessary before the consequences for the animals of these results can be discussed.
In the current study, no effect of diet on Enterobacteriaceae counts was detected at any site of the GIT. van Winsen et al. (2001) suggested that the most important factors contributing to the reduction of undesirable microorganisms in the GIT after feeding fermented liquid feed are pH and organic acid concentration. In the current study, the gastric pH in the DRY group was low (3.8) and the gastric concentration of lactic acid was high (74 mmol/kg of digesta), which can explain the relatively low number of Enterobacteriaceae at this site (and along the GIT) in this group, making further reductions by feeding FLG or FLF difficult to achieve. Also, the gastric pH was similar among all groups, which could explain the lack of effect of diet on Enterobacteria-ceae counts.
The T-RFLP data of samples from the stomach showed that a T-RF tentatively identified as Lactobacillus delbrueckii ssp. bulgaricus (255 bp) was present at a greater frequency in the DRY-fed piglets than in the other piglets. Lactobacillus delbrueckii, although typically inhabiting fermented dairy products such as yogurt and cheese (Zourari et al., 1992; Randazzo et al., 2002), has also been isolated in the pig GIT (Leser et al., 2002; Højberg et al., 2005). A T-RF tentatively identified as Lactobacillus reuteri/Lactobacillus fermentum (406 bp) or Lactobacillus durianis/Lactobacillus suebicus/Lactobacillus vaccinostercus (407 bp) was detected in more piglets fed FLF than in those fed the other 2 diets. Because this fragment was also present at the greatest frequency in the FLF, the results could indicate that the origin of this bacterium was the FLF. Although L. fermentum has been isolated in the GIT of pigs (Tannock et al., 1990), it is not represented in our porcine culture collection comprising 130 different phylotypes, or in the porcine clone library of Leser et al. (2002). Lactobacillus fermentum proliferates in products like fermented cassava (Kostinek et al., 2005), sourdough (Kitahara et al., 2005), and cheese (Poznanski et al., 2004). Lactobacillus durianis and L. suebicus are both isolated from fermented plant products as well (Kleynmans et al., 1989; Leisner et al., 2002). In contrast, L. reuteri has been reported as one of the most abundant lactobacilli in pigs (Leser et al., 2002; Højberg et al., 2005), but in the current study, a T-RF of 406 bp was detected at very low frequency in the DRY-fed group. This finding would suggest that the detected T-RF of 406 bp was L. fermentum or another lactobacilli of dietary origin, and not L. reuteri. In addition, a T-RF of 404 to 405 bp was present at high frequency in the stomach and midcolon irrespective of diet (data not shown). Acknowledging the potential for methodological inaccuracies, this T-RF could represent L. reuteri. Similarly, the absence of a T-RF identified as L. pentosus, L. plantarum, L. paraplantarum, or L. brevis (598 bp) in the DRY-fed animals and its presence in most of the animals fed the other 2 diets as well as in the fermented grain and FLF suggests a dietary origin of this bacterium. Furthermore, the observation that the 2 T-RF (406 and 598 bp) were not detected at a high frequency in the midcolon of the same animals could indicate that these bacteria were not able to proliferate to a high extent in the GIT of piglets, which would support their dietary origin. Data from the midcolon indicated that Clostridium perfringens (233 bp) and Clostridium lituseburense (814 bp) were more frequent in the FLG- and FLF-fed animals than in those fed the DRY diet. Unfortunately, there are no published data on the effect of feeding FLF or fermented liquid grain on the number of clostridia in the GIT of pigs. Because the pathogenicity of these bacteria varies with the strain, it is not possible to predict the effect of their presence in the GIT on the health of the piglets.
Growth Performance
Growth performance of piglets fed the DRY diet presented in the current study is not directly comparable with the results of the FLG and FLF groups. Although the former were fed ad libitum to simulate common practice in pig production, the latter were restrictively fed. The FLG and FLF were not fed ad libitum to avoid further fermentation of the feed in the trough, which would result in less control of what the animals really ingested (i.e., the nutritional and microbial characteristics of the FLG and FLF treatments would change with time in the trough).
Feeding fermented liquid feed to piglets has given varying results on growth performance (Russell et al., 1996; Pedersen, 2001; Lawlor et al., 2002). Disappearance of free amino acids, mainly lysine, by microbial fermentation in fermented liquid feed was suggested by Pedersen (2001) as the main reason for the negative effect of feeding it on growth performance compared with nonfermented liquid feed.
The microbial degradation of free amino acids, and thereby the decrease in its dietary content, was minimized with the procedure followed to prepare the FLG diet. Moreover, because eventual feed residues were removed from the trough 1 h after feeding, this was the maximum period available for the microorganisms in the FLG diet to degrade the free amino acids. Nonetheless, no improvement of G:F was detected when feeding the FLG diet compared with the FLF diet. This is in contrast with the improved G:F measured by Scholten (2001) when feeding liquid feed containing 45% fermented wheat compared with feeding nonfermented liquid feed, and a study by Pedersen (2006) that fed piglets liquid diets containing liquid fermented grain (66% of the mixture) compared with nonfermented liquid feed.
It is noteworthy that a diet prepared as the FLG diet would be further fermented if left in the trough or pipelines for a longer period, which would partially reduce the advantages of such feeding strategy. On the other hand, a feeding schedule for the FLG and FLF groups with more frequent meals may have increased the growth performance of these animals (Moran, 2001).
Low pH combined with high concentration of some fermentation metabolites (e.g., acetic acid, biogenic amines) in fermented liquid feed has been suggested to impair its palatability (Brooks et al., 2001; Moran, 2001) and consequently, decrease feed intake. Because the FLG had greater pH and contained less acetic acid and biogenic amines than the FLF, liquid feed containing fermented, liquid cereal grains may improve feed intake while maintaining the health-promoting effects associated with feeding FLF. This latter observation was only numerically supported in the current study.
|
Footnotes |
|---|
1 Corresponding author: nuria.canibe{at}agrsci.dk
Received for publication November 10, 2006. Accepted for publication June 19, 2007.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. P. Li, X. J. Zhao, and J. Y. Wang Synergy of Astragalus polysaccharides and probiotics (Lactobacillus and Bacillus cereus) on immunity and intestinal microbiota in chicks Poult. Sci., March 1, 2009; 88(3): 519 - 525. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
