| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL NUTRITION |
* Departament de Ciència Animal i dels Aliments, Universitat Autònoma de Barcelona, 08193, Bellaterra, Barcelona, Spain;
and
CReSA, Centre de Recerca en Sanitat Animal, Campus Universitari de Bellaterra, Universitat Autònoma de Barcelona, 08193, Bellaterra, Barcelona, Spain
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
Key Words: antibiotic • microbiota • quantitative polymerase chain reaction • restriction fragment length polymorphism • weaning pig
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
). To prevent gastrointestinal disorders and to improve the postweaning growth rate, feed-grade antibiotics are used regularly. Despite their general use, their exact mode of action is not entirely clear, and different mechanisms have been proposed. Most are based on a reduction in bacterial numbers, with a resulting decrease in the production of growth depressing microbial metabolites and in the competition for nutrients with the host (Anderson et al., 1999; Hardy et al., 2000). However, other mechanisms related to the selection of a healthier microbial community could also be implicated.
Concerns about bacterial resistance to antibiotics and general food safety issues have encouraged intensive research on new feed additives to maintain the growth promotant effects of antibiotic growth promotants without their potential drawbacks. The addition of different organic acids to the feed is one of the most widely used alternatives to the use of antibiotic growth promotants, and their effects have been related to a reduction in the growth of some bacteria (Partanen, 2001). Herbs have been known since ancient times to have antimicrobial, antioxidant, and antifungal properties. Some of these compounds have been reported to improve animal performance because of their stimulating effect on salivation and pancreatic enzyme secretions or by having a direct bactericidal effect on gut microflora (Hardy, 2002). Carvacrol from oregano has demonstrated strong antimicrobial properties (Dorman and Deans, 2000); cinamaldehyde from cinnamon has shown antioxidant and antimicrobial effects (Mancini-Filho et al., 1998); and capsaicin from chili stimulates gastric secretions (Platel and Srinivsan, 2000).
The experiment reported here aimed to evaluate the effects of an antibiotic, an acidifier, and a plant extract mixture, on the load, metabolic activity, and community structure of the gastrointestinal microbiota of early-weaned pigs.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
Animals and Housing
A total of 32 early-weaned pigs [(Large White x Landrace) x Pietrain; mixed males and females] from a commercial herd were selected from 8 different litters. No creep feeding was provided during the lactation period. The pigs were weaned between 18 and 22 d of age at an average initial BW of 6.0 ± 0.10 kg and were housed in the Universitat Autònoma de Barcelona facilities according to their initial weight in 8 pens (4 animals per pen). The 32 pigs were housed in the same room and separated by solid walls of 60 cm in height with bars into the top up to 80 cm. Each pen had its own feeder and nipple drinker. The weaning room was equipped with automatic heating and forced ventilation and the temperature was gradually reduced from 29 to 25 ° C during the experiment.
Dietary Treatments and Feeding Regimen
Four dietary treatments were used. A control diet was formulated (CT; Table 1) to which 3 different additives were added: 0.04% avilamycin (AB; Maxus; Elanco Animal Health, Madrid, Spain), 0.3% sodium butyrate (AC; Nature S.A., Barcelona, Spain), or 0.03% plant extract mixture (XT). The plant extract mixture was standardized as 5% (wt/wt) carvacrol (Origanum spp.), 3% cinnamaldehyde (Cinnamonum spp.), and 2% capsicum oleoresin (Capsicum annum) in an inert fatty carrier that represented the remaining 90%. Chromic oxide (0.02%) was included as a digestibility marker in all diets. Pigs were fed the experimental diets ad libitum for 3 wk after weaning, and they had free access to water throughout the experiment.
|
For DNA analysis, samples of about 1 g each of digesta from the stomach, distal jejunum, cecum, and the distal colon were kept in weighed tubes with 3 mL of ethanol as a preservative. Samples were also taken from the mucosal layer of the jejunum: a segment of 4 cm was longitudinally cut and gently washed with a sterile saline solution. The mucosal layer was scraped with a spatula, and 250 to 500 mg were placed in weighed capped tubes and immediately snap-frozen in liquid N. Samples were kept at –80 ° C until analysis. Digesta samples (approximately 50 g) from the ileum, cecum, proximal colon, distal colon, and rectum were taken for purine base analysis. Samples were frozen and lyophilized until analysis. For the study of microbial enzymatic activities, samples of 5 g of digesta from the cecum and distal colon were snap-frozen in liquid N and stored at –80 ° C until analysis.
DNA Extraction.
Digesta samples (400 mg) preserved in ethanol were precipitated by centrifugation (13,000 x g for 5 min), and DNA from the precipitate was extracted and purified using the commercial QIAamp DNA Stool Mini Kit (Qiagen, West Sussex, UK). The recommended lysis temperature was increased to 90 ° C, and a prior incubation step with lysozyme was added (10 mg/mL, 37 ° C, 30 min) to improve the bacterial cell rupture. The DNA was eluted in 200 mL of Qiagen Buffer AE (Qiagen, West Sussex, UK) and was stored at –80 ° C. The DNA from the mucosal layer scrapings was harvested using the same commercial kit. The DNA from pure cultures of Lactobacillus acidophilus (CECT 903NT) and Escherichia coli (CECT 515NT) was harvested using the same Qiagen Kit. Pig genomic DNA was obtained from blood samples using the Mammalian Genomic DNA extraction kit (Camgen, Cambridge Molecular Technologies Ltd., Cambridge, UK).
Quantitative PCR.
The primers used to quantify the different bacterial groups are listed in Table 2. The oligonucleotides were based on regions of identity within the 16S rDNA and were adapted from published specific primers or probes using the Primer Express Software (Applied Biosystems, Foster City, CA). This software was used to check for primer-dimer, internal hairpin configurations, the melting temperature, and percentage guanine and cytosine values within possible primer/probe sets. The different primers were also checked for their specificity using the database similarity search program BLAST (Altschul et al., 1990), and the absence of amplification of porcine DNA was tested empirically by PCR using DNA extracted from pig blood.
|
For absolute quantification, PCR products obtained from the amplification of the whole 16S rDNA of Escherichia coli (CECT 515NT) and Lactobacillus acidophilus (CECT 903NT) were used to construct the standard curves, and the PCR conditions corresponded to those published by Leser et al. (2002). An amplified gene from E. coli was used for absolute quantification of the total bacteria and enterobacteria and an amplified gene from L. acidophilus for quantification of the lactobacilli. The functions describing the relationship between Ct (threshold cycle) and x (log copy number) for the different assays were: Ct = –3.19x + 53.66 (R2 = 0.99) for total bacteria; Ct = –2.60x + 46.82 (R2 = 0.99) for lactobacilli; and Ct = –2.32x + 43.88 (R2 = 0.99) for enterobacteria.
PCR-RFLP Analysis.
To analyze the total bacteria, a fragment of the 16S rDNA gene was amplified from DNA extracts by PCR using primers specific to conserved sequences flanking variable regions V3, V4, and V5: 5'-CTACGGGAGGCAGCAGT-3' (forward) and 5'-CCGTCWATTCMTTTGAGTTT-3' (reverse). Primer and PCR reaction conditions were those described by Lane (1991). The reaction was performed using a Gen-eAmp PCR System 9700 (PE, Biosystems, Warrington, UK) thermocycler. The DNA amplification conditions were 94 ° C (4 min); 35 cycles of denaturation at 94 ° C (1 min), annealing at 45 ° C (1 min) with an increment of 1 ° C per cycle, extension at 72 ° C (1 min 15 s); and a final extension at 72 ° C (15 min). After visual confirmation of the PCR products with agarose gel electrophoresis, 4 independent enzymatic restrictions were carried out (AluI, RsaI, HpaII, CfoI; F. Hoffmann-LaRoche Ltd. Group, Basel, Switzerland). The digestions were carried out for 3 h, as recommended by the manufacturer, with appropriate restriction buffers at the recommended temperature. Different fragments were separated using a 2% high resolution agarose gel.
The size and the intensity of the bands within each lane of a gel were analyzed by the Gene Tools software (Syngene, Cambridge, UK), and the degree of microbial biodiversity was measured as the total number of different bands obtained from the 4 independent restriction digestions.
For pairwise comparisons of the banding patterns and the construction of dendrograms, similarity matrices were generated based on the Manhattan distance (Kaufmann and Rousseeuw, 1990) that takes into account the size and the intensity of the bands generated.
Purine Bases Analysis.
Purine bases (adenine and guanine) in lyophilized digesta samples (40 mg) were determined by HPLC (Makkar and Becker, 1999). For this analysis, purine bases were hydrolyzed from the nucleic acid chain by incubation with 2 mL of 2 M HClO4 at 100 ° C for 1 h, including 0.5 mL of 1 mM allopurinol as an internal standard.
Microbial Enzymatic Activities.
The microbial enzymes were extracted from the digesta contents by hydrolysis of bacterial cells with lysozyme (5 mg/mL, 37 ° C, 3 h) following the method described by Silva et al. (1987). After the incubation period, samples were centrifuged (23,000 x g for 15 min) and the enzymes from the supernatant were kept frozen (–80 ° C) until analysis. Polysaccharidase activities of the enzymatic extract were determined by assay of reducing sugars released from purified substrates according to the Nelson-Somogyi method (Ashwell, 1957). The substrates were suspended in 0.1 N sodium phosphate buffer (pH 6.7). The samples (0.05 mL) were incubated (30 min, 40 ° C) with 0.45 mL of each substrate solution containing carboxymethylcellulose (Sigma-Aldrich Química S. A., Madrid, Spain), xylan from oat spelts (Sigma-Aldrich Química), soluble starch from potato (Panreac, Barcelona, Spain), and waxy starch from corn (Sigma-Aldrich Química). Activities against these 4 substrates were referred to as CMCase, xylanase, amylase, and amylopectinase, respectively. After the incubation period, the reaction was stopped by denaturing the enzyme proteins (100 ° C for 10 min), and the amount of reduced sugars was quantified spectrophotometrically at 600 nm. Dilutions of glucose (0, 25, 50, and 100 µg/mL) were used as the standard curve. The activity of the enzymatic extract was expressed as micromoles of neutral sugars released per milliliter of extract per minute, and is referred to the purine bases concentration (bacterial enzymatic activity).
Statistical Analysis
The effect of diet on microbial counts, biodiversity, purine base concentration, and enzymatic activities in a given intestinal segment was tested with an ANOVA using the GLM procedures of SAS (version 8.1, SAS Inst. Inc., Cary, NC). The individual pig was used as the experimental unit. When treatment effects were established (P < 0.05), treatment least squares means were separated using the probability of differences (PDIFF) function adjusted by Tukey-Kramer (SAS). Purine bases concentrations along different intestinal segments in each animal were analyzed as repeated measures using the PROC MIXED procedure of SAS. Statistical significance was accepted at P < 0.05.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
Changes in the Total Microbial Counts
The total microbial population was quantified along the whole gastrointestinal tract using quantitative PCR (qPCR; Figure 1). In the foregut, the counts, expressed as log 16S rDNA copies/g of fresh matter (FM), increased from 8.0 ± 1.16 in the stomach to 11.1 ± 0.88 in the jejunum, showing a considerable increase of more than 3 log units. The microbial population intimately attached to the jejunum mucous membrane was also quantified and although mean values were lower than counts in the lumen (10.2 ± 0.94 log 16S rDNA copy number/g of FM) differences did not reach statistical significance. The cecum and colon digesta showed mean values of 12.4 ± 0.13 and 12.3 ± 0.93 log 16S rDNA copy number/g of FM, respectively. These values represent an increase of more than 1 log unit compared with the total counts in the jejunum. No significant differences in total bacterial loads related to experimental diets were found in any part of the analyzed gastrointestinal tract.
|
). Both groups increase in number from the jejunum to the cecum; however, whereas the enterobacteria showed an increase of around 4 log units, lactobacilli increased only by around 2 logs. Expressing the difference between both bacteria groups as a ratio of logarithms (lactobacilli:enterobacteria ratio) the cecum showed lower ratio values than the jejunum. Between dietary treatments, XT promoted an increase in the lactobacilli:enterobacteria ratio in the cecum when compared with the CT (P = 0.02), which can be explained by an increase in lactobacilli numbers (P = 0.02). The AC diet also showed greater lactobacilli:enterobacteria ratio mean values; however, differences with CT diet were not significant (P = 0.49).
|
). The effect of different additives on biodiversity, measured as number of bands, was evident when considering the decrease in number of bands for the CT diet (28.9, 38.5, 38.8, and 32.0 for CT, AB, AC, and XT respectively; P = 0.03). The effects of the experimental diets on microbial composition were more clearly distinguished by the cluster analysis. A dendrogram comparing different banding patterns is shown in Figure 2. Distinct clusters according to the different diets were observed. The acidifier diet promoted the biggest structural changes (63.5% similarity) followed by the XT diet (66.9%), and then the AB diet (73.3%).
|
. All the treatments show an increase in PB concentration from the ileum to the cecum that is quantitatively larger for the CT and AB diet than for the AC and XT diet. In the hindgut, the evolution of the PB concentration also shows differences between diets (diet x intestinal section, P = 0.01). Whereas the CT diet reached maximum values at the end of the colon, with an abrupt decline from the colon to rectum, the rest of the treatments reached maximum values in previous intestinal sections. Finally, PB concentration in the rectum was similar for all the treatments.
|
. The data showed high variability, precluding the ability to find any differences between treatments.
|
|
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
using a similar PCR methodology described a decrease in the total bacteria in the ileum of growing pigs after 2 wk of treatment with 40 ppm of tylosin. This decrease was observed until d 21; however, the effect disappeared after 28 d, which reflects the microbial community’s adaptive response, replacing susceptible strains with resistant organisms. In our case, this new equilibrium could have been reached sooner (2 wk), which would explain the absence of an antibiotic effect on the total bacteria counts.
The values for the total bacteria in the stomach, jejunum, cecum, and colon estimated by qPCR were similar to those described by other authors for culturable bacteria in pigs of a similar age (Jensen and Jorgensen, 1994; Krause et al., 1995; McFarland, 1998). It is fair to remark that values were always close to the highest levels, probably because of the trend of qPCR to overestimate microbial populations. Some authors, when comparing qPCR with culture methods, have also described discrepancies of 1 or even 2 log units (Huijsdens et al., 2002; Nadkarni et al., 2002). These discrepancies could be explained mainly by the presence of a high number of viable but not culturable bacterial cells in the digesta samples (Rigottier-Gois et al., 2003), the amplification and later quantification of free DNA from dead bacteria, and the multiplicity of 16S rDNA genes per genome in prokariotic organisms (Fogel et al., 1999).
In the light of the absence of significant effects on the total microbial counts, it seems feasible that antibiotics and other alternatives such as organic acids or plant extracts could act not by reducing the total size of the microbial population but by promoting the selection of particular bacteria. In this respect, the different spectra that antibiotics have and also the specific susceptibility of bacteria to different organic acids are well known (Cherrington et al., 1991). Possemiers et al. (2004) using qPCR could not detect changes in the total microbial population after adding an antibiotic to an in vitro simulator of the human microbial ecosystem; however, using group specific primers they could detect a decrease in the number of bifidobacteria.
Looking for ecological changes, the lactobacilli and enterobacteria populations were quantified in the jejunum using qPCR. The relationship between these bacterial groups has traditionally been considered as an index of desirable or undesirable bacteria in pigs, relating a high index with a greater resistance to intestinal disorders (Ewing and Cole, 1994). From the additives tested, XT showed the clearest effect, increasing the lactobacilli:enterobacteria ratio compared with the CT. Increases were observed in the cecum (P = 0.006) mainly due to an increase in the number of lactobacilli. Previous results with the same plant extract mixture also showed increases in the lactobacilli:enterobacteria ratio in the jejunum of weaned pigs due to an increase in lactobacilli numbers (Manzanilla et al., 2004). It is difficult to find an explanation for this promoting effect, taking into account that most of the in vitro studies with plant extracts have shown an unspecific antimicrobial effect (Hammer et al., 1999). However these consistent results seem to point to some kind of prebiotic effect on the lactobacilli population, either by a direct or indirect effect through an ecological change in the intestinal microbiota. Adding butyrate to diets also promoted greater mean values in the lactobacilli:enterobacteria ratio, although in this case differences compared with the CT were not significant (P = 0.17). There are few publications studying the inclusion of n-butyrate in diets for weaned piglets and its effects on microbial populations. Galfi and Bokori (1990) using 0.17% sodium n-butyrate in the diets of weaned piglets observed changes in the ileal microbiota with a decrease in the proportion of coliform bacteria with a simultaneous increase in lactobacilli, however these authors also found an increase in the ileal concentration of butyrate that we did not observe (data not shown). Van Immerseel et al. (2004) using microencapsulated butyric acid in young chickens could also demonstrate a decrease in Salmonella in the cecum colonization after an experimental infection. It is interesting to note that the same authors using other organic acids like formic and acetic acid observed the opposite effect with an increase in Salmonella in the cecum colonization. Probably many other factors, such as the activation or inhibition of different metabolic routes with different organic acids, are involved in the changes observed with the acidifiers and not only a simple effect caused by a lower pH. This complexity could explain the diverse and sometimes contradictory effects of different acidifiers on microbial populations described in the literature (Hebeler et al., 2000; Canibe et al., 2001; Février et al., 2001).
Avilamycin is an antibiotic mainly active against gram-positive bacteria; therefore, we could expect a decrease in lactobacilli numbers. However, we did not find this effect. Other authors using avilamycin (50 ppm) also did not find differences in bacterial numbers (Decuypere et al., 2002). Similarly, Collier et al. (2003) using tylosin (another macrolide active against gram positives) could not detect any decrease in lactobacilli but rather an increase, which is particularly intriguing and could reflect complex interactions between different species in the bacterial ecosystem. When using RFLP to analyze variations in the bacterial community, we evidenced changes in band patterns related to dietary treatments. It is interesting to point out how profiles for each treatment were clustered separately; the cluster for the animals that received the AC diet showed the most difference. Differences in the RFLP patterns were due to an increase in the biodiversity in the microbial ecosystem with the use of additives (number of bands) and also to a change in the species composition of the community (type of bands). From our results, it could be suggested that a more complex microbial community would have a greater robustness in response to changes in the intestinal environment promoted by different dietary ingredients or stress and that the beneficial effects of antimicrobial additives could be related to an improvement of the adaptive capacity of commensal microbiota as a natural barrier defense against the overgrowth of pathogens more than to a reduction in bacteria numbers. Other authors using similar fingerprinting techniques (denaturant gradient gel electrophoresis; McCracken et al., 2001) and comparing fecal microbial populations from rats receiving diets supplemented or not with antibiotics did not detect changes in the biodiversity, but they did detect how bacterial species that form each microbial community were significantly altered by the antibiotic. Similarly, Collier et al. (2003), working with pigs receiving 40 ppm of tylosin for more than 21 d, found that the number of denaturant gradient gel electrophoresis bands in ileal samples was similar to the number in the CT diet but that the banding patterns were treatment-dependent.
The evolution of purine bases concentration along the hindgut showed that the main differences were between diets CT and AB. When in the CT diet, the concentration of purine bases reached its maximum value at the distal colon, decreasing afterwards; the AB diet purine concentration reached its maximum values at the cecum. Previous results from our group described similar patterns when comparing animals receiving diets differing in the amount of resistant starch (Martinez-Puig et al., 2003). In that case, the concentration of purine bases decreased earlier in animals receiving the diet with a lower amount of fermentable starch. In the present work, experimental diets have the same ingredient composition except for the added additives; therefore, changes in fermentation should not be attributed to differences in dietary carbohydrates. However, changes in the extent of digestion and absorption of nutrients at the foregut level could potentially have promoted the arrival of different amounts of fermentable material to the hindgut and therefore changes in microbial carbohydrases activity, but we were unable to observe that occurrence. If additives did promote differences, they were not big enough to be detected by this methodology. It is interesting to point out that although the amylase and amylopectinase activities were comparable to those described by other authors in growing pigs (Morales et al., 2002), cellulase or xylanase activities were not detected. This lack of enzymatic bacteria activity could be related to an insufficient adaptation of microbiota to digesting complex carbohydrates like cellulose or hemicellulose in young animals.
|
IMPLICATIONS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
|
Footnotes |
|---|
2 Acknowledgments: The authors are grateful to Ana Pérez de Rozas for her collaboration in the DNA analysis.
3 Corresponding author: Susana.martin{at}uab.es
Received for publication October 18, 2004. Accepted for publication November 9, 2005.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONIMPLICATIONSLITERATURE CITED |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. G. Manzanilla, J. F. Perez, M. Martin, J. C. Blandon, F. Baucells, C. Kamel, and J. Gasa Dietary protein modifies effect of plant extracts in the intestinal ecosystem of the pig at weaning J Anim Sci, June 1, 2009; 87(6): 2029 - 2037. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mazzoni, M. Le Gall, S. De Filippi, L. Minieri, P. Trevisi, J. Wolinski, G. Lalatta-Costerbosa, J.-P. Lalles, P. Guilloteau, and P. Bosi Supplemental Sodium Butyrate Stimulates Different Gastric Cells in Weaned Pigs J. Nutr., August 1, 2008; 138(8): 1426 - 1431. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Windisch, K. Schedle, C. Plitzner, and A. Kroismayr Use of phytogenic products as feed additives for swine and poultry J Anim Sci, April 1, 2008; 86(14_suppl): E140 - E148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. A. Olukosi, J. S. Sands, and O. Adeola Supplementation of carbohydrases or phytase individually or in combination to diets for weanling and growing-finishing pigs J Anim Sci, July 1, 2007; 85(7): 1702 - 1711. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Nofrarias, E. G. Manzanilla, J. Pujols, X. Gibert, N. Majo, J. Segales, and J. Gasa Effects of spray-dried porcine plasma and plant extracts on intestinal morphology and on leukocyte cell subsets of weaned pigs J Anim Sci, October 1, 2006; 84(10): 2735 - 2742. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. G. Manzanilla, M. Nofrarias, M. Anguita, M. Castillo, J. F. Perez, S. M. Martin-Orue, C. Kamel, and J. Gasa Effects of butyrate, avilamycin, and a plant extract combination on the intestinal equilibrium of early-weaned pigs J Anim Sci, October 1, 2006; 84(10): 2743 - 2751. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Rn (magnitude of the signal generated by the PCR conditions). (B) DNA concentrations (16S rDNA gene copies/g of FM) were plotted vs. the threshold cycle value to construct the standard calibration curve. (C) Bacterial loads in the stomach, jejunum, cecum, distal colon digesta, and in the jejunum mucosal layer, measured by quantitative PCR (log 16S rDNA gene copies/g of FM) in early-weaned pigs receiving a control diet (CT) or the same diet with 0.04% avilamycin (AB), 0.3% butyric acid (AC), or 0.03% plant extract mixture (XT). Bars represent means and SEM.
