Chicken intestine microbiota following the administration of lupulone, a hop-based antimicrobial

Authors


  • Editor: Julian Marchesi

  • Present addresses: Mark G. Wise, BioMerieux, Centre Christophe Merieux, 5 rue des Berges, 38024 Grenoble Cedex 01, France. Gregory R. Siragusa, Danisco USA, W227 N752 Westmound Drive, Waukesha, WI 53186, USA.

Correspondence: Glenn E Tillman, USDA FSIS OPHS Eastern Laboratory, 950 College Station Road, Athens, GA 30605, USA. Tel.: +1 706 546 3694; fax: +1 706 546 3144; e-mail: glenn.tillman@fsis.usda.gov

Abstract

The use of antibiotic growth promotants in poultry rearing is a public health concern due to antibiotic resistance in bacteria and the harborage of resistance genes. Lupulone, a hop β-acid from Humulus lupulus, has been considered as a potential feed additive growth promotant. Here, the effect of lupulone was evaluated for its effect on the microbiota of the chicken intestine. The intestinal microbiota of broilers was quantified after the addition of 125 mg L−1 lupulone to water and challenge with Clostridium perfringens. Microbial DNA was extracted from the broiler midgut and cecal sections and bacterial groups were quantified using real-time PCR. The predominant cecal bacterial groups were Clostridium leptum subgroup 16S rRNA gene Cluster IV, Clostridium coccoides subgroup 16S rRNA gene Clusters XIVa and XIVb and Bacteroides, whereas Lactobacillus, the Enterobacteriaceae family and Enterococcus dominated the midgut. Lupulone at 125 mg L−1 significantly decreased the C. perfringens subgroup 16S rRNA gene Cluster I, which contains several pathogenic species, in both the midgut and the cecum and Lactobacillus in the midgut. No significant changes were noted in the overall microbiota for the cecum or the midgut. Lupulone warrants further evaluation as a botanical agent to mitigate C. perfringens overgrowth in antibiotic-free reared poultry.

Introduction

Subtherapeutic antibiotic use in livestock feed has been a common practice in the United States since the 1950s. The use of these agents, also referred to as antibiotic growth promotants (AGP), improves feed efficiency and conversion in food animals, lowers morbidity and mortality and improves the general health of food animals (Jukes & Williams, 1953). AGP are thought to exert their growth-promoting effect by altering the intestinal microbiota of the animal. The proposed mechanisms of AGP's effects on the intestinal microbiota include nutrient protection from microbial utilization, improved absorption of nutrients due to a thinner intestinal barrier, decrease of toxin production by intestinal bacteria and reduction of subclinical intestinal infections (Gaskins et al., 2002). More recently, Niewold (2007) proposed a direct mechanistic action of the AGP molecules on host intestinal cells and immune function.

Despite the benefits from AGP use in livestock feed, there are concerns that antibiotics in feed contribute to antibiotic resistance and positively select for resistance genes in pathogenic bacteria found in humans. Removing AGP from broiler diets can result in an increase of poultry necrotic enteritis (NE), which can subsequently lead to flock condemnation, production losses or the therapeutic use of antibiotics to treat clinical symptoms (Baba et al., 1997). Coccidiostat drugs, such as salinomycin, generally help control outbreaks of NE. These drugs also have an antibacterial effect on the pathogenic bacterium Clostridium perfringens, which is known to be the major causative agent of NE (Van Immerseel et al., 2004). Poultry exhibiting NE have one or several disposing factors, including diets high in wheat or barley, diets with protein derived from animal sources or infection with a coccoidia protozoan, Eimeria (Baba et al., 1997; Drew et al., 2004; Dahiya et al., 2005).

A primary challenge facing the food animal industry and the public health community is to effectively limit or remove the use of AGP in feed while still producing food animals humanely and efficiently. A number of AGP alternatives with little relationship with traditional human antimicrobial therapy have been evaluated. Among these are bacteriocins, antimicrobial peptides, bacteriophages, probiotics, competitive exclusion cultures and plant products (Nurmi & Rantala, 1973; Teuber & Schmalreck, 1973; Chen & Stern, 2001; Joerger, 2003). Extracts from the hop plant, Humulus lupulus, used in beer brewing for centuries, exert antimicrobial effects on Gram-positive bacteria (Teuber & Schmalreck, 1973; Srinivasan et al., 2004). The efficacy of whole hop flowers as growth promotants in chicken feed has been demonstrated (Cornelison et al., 2006), but as with other AGP in feed, little is known about the effective mechanisms of hop extracts on growth promotion. In one of the few studies on this topic, it has been shown that xanthohumol, another constituent of hop, had no effect on the diversity of the microbiota of the rat gastrointestinal tract (Hanske et al., 2005)

Notwithstanding the genetics of the chicken itself, the sheer diversity of microorganisms found in the chicken gastrointestinal tract has made understanding its ecology challenging. Cultural and molecular techniques, such as 16S rRNA gene analysis and G+C% profiling, have helped to elucidate a qualitative representation of the chicken gastrointestinal tract microbiota. The midgut intestine has been shown to be dominated by Lactobacillus species, whereas the cecum has been shown to contain mostly low G+C content, Gram-positive bacteria, such as Clostridium species, Fusobacterium species, Bacteroides and Lactobacillus (Barnes et al., 1972; Salanitro et al., 1974; Lan et al., 2002; Lu et al., 2003; Dumonceaux et al., 2006). To understand how AGP or the alternatives positively affect growth and performance, it is important to determine both the normal intestinal microbiota of the chicken gastrointestinal tract and the microbiota of the gastrointestinal tract after AGP use (Dumonceaux et al., 2006).

In this study, we quantified the gastrointestinal microbiota of broilers fed lupulone, a β-acid derived from hops, and challenged with NE-associated C. perfringens strains (Wise & Siragusa, 2006). The aim of this study was to investigate the effect of lupulone on the microbiota of the broiler midgut and cecum.

Materials and methods

Experimental group design and sample collection

All animal experiments were conducted in compliance with Agricultural Research Service – USDA Laboratory Animal Care and Use Committee standards for care, feeding, euthanasia and disposal. The present study was run in duplicate and samples from both trials were combined to obtain the data. Newly hatched broiler chickens (Gallus gallus domesticus) were assigned to one of five groups (n=4) as shown in Table 1. Broilers were fed diets consisting of nonmedicated broiler starter diet feed (corn–soy based 23% crude protein, 6% fat, 2.5% fiber, 1.0% calcium and 0.48 available phosphorous; 3100 kcal kg−1 estimated metabolic energy) obtained from the University of Georgia, Department of Poultry Science feed mill. On Days 13–22, broilers in Groups 3 and 5 (lupulone treatment groups) were given 125 mg L−1 (307 μM) lupulone through cage watering systems. Lupulone was administered as described previously (Siragusa et al., 2008). The midrange level of 125 mg L−1 lupulone was chosen based on prior laboratory studies (data not shown) showing no significant difference in broiler weight of chickens for administered levels of lupulone ranging from 62.5 to 250 mg L−1. Beginning on Day 14, broilers in the challenge Groups 4 and 5 were administered 0.1 mL of a three-strain C. perfringens cocktail (per bird total dosage of ∼107 cells) per os for three consecutive days (Days 14–16 of growth). The strains were propagated in prereduced brain–heart infusion broth under anaerobic conditions for 16 h at 37 °C in screw cap tubes before challenge dosage preparation. The C. perfringens cocktail consisted of three C. perfringens Type A strains obtained from commercially reared birds from an NE outbreak as described previously (Siragusa et al., 2008). Challenge with C. perfringens was meant as a colonization model rather than an NE model. Birds from Group 1 were sacrificed by cervical dislocation on Day 14, and birds from Groups 2–5 were sacrificed on Day 22. Cecal and midgut sections were removed from the bird gastrointestinal tracts and placed in individual, sterile Whirl-Pack bags.

Table 1.   Experimental group design with days of treatment, challenge and sacrifice
GroupCodeChallengeTreatment with 125 mg L−1 lupuloneSacrifice
  • *

    Day 14 baseline no challenge, (0 mg L−1) lupulone (n=4).

  • Day 22 no challenge, (0 mg L−1) lupulone (n=4).

  • Day 22 no challenge, (125 mg L−1) lupulone (n=4).

  • §

    Day 22 challenge with Clostridium perfringens, (0 mg L−1) lupulone (n=4).

  • Day 22 challenge with Clostridium perfringens, (125 mg L−1) lupulone (n=4).

1Baseline*NoneNoneDay 14
2NCL0NoneNoneDay 22
3NCL125NoneDays 13–22Day 22
4CL0§Days 14–16NoneDay 22
5CL125Days 14–16Days 13–22Day 22

DNA extraction from intestinal samples

Midgut and cecal samples for DNA extraction were removed by squeezing the luminal contents (approximately 1.0 g) into 15 mL conical tubes. The samples were diluted 1 : 10 with phosphate-buffered saline (PBS). Samples were held at 2–8 °C until the DNA extraction step. Samples were vortexed vigorously for 1 min before microbial DNA extraction following the manufacturer's instructions for the MoBio UltraClean Fecal DNA Kit (Solana Beach, CA). Briefly, 1 mL of each mixture was transferred to the Bead Tubes provided in the kit. Samples were centrifuged at 10 000 g for 5 min and the supernatant was discarded. Five hundred microliters of Bead Solution (guanidine thiocyanate), 60 μL of S1 solution (sodium dodecylsulfate) and 200 μL of inhibitor removal solution (IRS) solution (proprietary IRS) were added to each tube and bacterial cell walls were mechanically lysed at the maximum speed on the MoBio Vortex Adaptor for 10 min. The tubes were centrifuged for 3 min at 10 000 g and the supernatant was transferred to labeled, sterile microcentrifuge tubes. The supernatant was mixed with 250 μL of Solution 2 (proprietary acetate solution), vortexed and placed on ice for 5 min. The tubes were centrifuged at 12 000g for 3 min, and the supernatants were removed and mixed with 1 mL of Solution 3 (guanidine HCl and isopropanol). The mixture was added to spin filter columns, washed with 300 μL of Solution 4 (ethyl alcohol) and purified DNA was eluted with 50 μL of Solution 5 (Tris aminomethane/hydrochloride). Purified DNA was stored at −20 °C until use as a template in PCR.

Quantitative real-time PCR

All PCR reactions consisted of 20 μL volumes in a 96-well format for the Applied Biosystems 7300 instrument. Reaction volumes and primer sequences, as described previously (Wise & Siragusa, 2007), included 10 μL 2 × SYBR Green MasterMix (Applied Biosystems, Foster City, CA), 1 μL of each group-specific forward primer at 10 μM (final concentration of 0.5 μM), 1 μL of group-specific reverse primer at 10 μM (final concentration of 0.5 μM), 1 μL of bovine serum albumin at 2.5 mg mL−1 (final concentration 125 μg mL−1), 5 μL of nuclease-free water and 2 μL of purified DNA template. Because of the presence of inhibitors in intestinal samples, all templates were diluted 1 : 10 in MoBio Elution Buffer before adding to PCR reactions (Wise & Siragusa, 2005). Cycling conditions were 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, primer-specific annealing temperatures (Table 2) for 1 min and 78 °C for 30 s. A dissociation step was included after amplification to analyze the melting curves of the amplified product. Data were obtained at the 78 °C step to ensure that primer-dimers were not measured as amplification of the target. For this assay, the temperature of 78 °C was empirically determined to be above the melting temperature of primer dimers, but below the melting temperature of the amplified products (Wise & Siragusa, 2007).

Table 2.   Bacterial groups assayed and their primer annealing temperatures
Target groupRepresentative isolatePrimer annealing temperature (°C)Product melting temperature (°C)*
  • *

    Approximate melting temperature of the PCR products was determined empirically from standards assayed with each PCR run. The melting temperatures for PCR products of unknowns were compared with that of the standards assayed with each PCR run.

Domain BacteriaEscherichia coli6083.5
C. leptum subgroup (Cluster IV)C. leptum ATTC 290656084.4
C. coccoides subgroup (Cluster XIVa and XIVb)C. coccoides ATTC 292365083.8
Bacteroides groupB. fragilis ATTC 252855982.2
Bifidobacterium spp.Bifidobacterium sp. wild type6087.8
Enterobacteriaceae familyE. coli TOP106383.8
Lactobacillus groupLactobacillus sp. wild type5882.7
C. perfringens subgroup (Cluster I)C. perfringens ATTC 131246081.0
Enterococcus spp.Enterococcus faecalis ATTC 194336182.4
Veillonella spp.Veillonella parvula ATTC 107906285.5
Atopobium spp.Atopobium minutum ATTC 332676184.3
Campylobacter spp.Campylobacter jejuni NCTC 111686183.9

Quantification of bacterial groups in intestinal samples

Major and minor groups of the midgut and ceca were quantified using absolute quantification as described previously (Wise & Siragusa, 2007). Standard curves were constructed using 16S rRNA gene products from the representative isolates from each group shown in Table 2. Amplified gene products were purified and cloned into pCR4-TOPO using the TOPO-TA Cloning Kit (Invitrogen, Carlsbad, CA). Plasmids containing the insert for each group were purified and quantified using a spectrophotometer. The number of target gene copies was calculated from the mass of DNA and the number of base pairs in the insert and plasmid. The gene copies were serially diluted from 2 × 109 to 2 × 101. Standards for each bacterial group were amplified along with the unknowns. The cycle threshold, the point at which the fluorescence passed the threshold, was determined for the standards, and unknowns were extrapolated from the curve. Gene copy numbers for the unknowns were converted to log10 for a normal distribution and reported per gram intestinal material.

Statistical analyses

The average log10 gene copy numbers and SDs were determined for the bacterial groups in all the experimental groups. Bacterial groups not detected in at least 50% of the birds were reported as not detected (ND). For bacterial groups with at least 50% detection, count averages were determined from nonzero counts. Microbiota comparisons were made between lupulone treatment groups and control groups, challenge groups with nonchallenge groups and interactions of challenge and lupulone. Presence–absence analyses were performed on all groups using Fisher's exact test. For nonzero counts in each group, a profile analysis using anova with SAS® 9.1 was performed at the Biostatistics Consulting Service College of Public Health, The University of Georgia. Further, similarity of treatments was assessed by unconstrained correspondence analysis performed in r (R development Core Team, 2008) using the CCA routine of the vegan package based on mean copies  g-1.

Results

The mean bacterial group counts and SEs of the means for the cecum and midgut are shown in Tables 3 and 4, respectively. There was no apparent significant effect of challenge (P=0.28) on the overall bacterial groups; thus, data were aggregated for the consideration of the effect of lupulone on the individual bacterial groups of the gastrointestinal tract. Data for Groups 2 and 4 (0 lupulone) were combined and considered as the control group. Individual bacterial group data for Groups 3 and 5 (125 mg L−1 lupulone) were combined and considered the lupulone treatment group. The Enterobacteriaceae family, Clostridium coccoides subgroup, Atopobium, Bacteroides, Enterococcus, Clostridium leptum subgroup, Lactobacillus and Bifidobacterium were detected in 100% of the control group samples taken in the cecum, and C. perfringens subgroup was detected in seven out of eight samples. Campylobacter and Veillonella were undetected in the cecum and the midgut for all of the experimental group samples. All other bacterial groups were detected in 100% of the lupulone treatment group samples, except the C. perfringens subgroup and the C. leptum subgroup, which were both detected in seven out of eight samples.

Table 3.   Mean counts of bacterial groups in the cecum expressed as log10 16S rRNA gene-specific gene copies per gram of intestinal material
 Total bacteriaEnterobacteriaceae familyC. coccoides subgroup Cluster XIVa, XIVbC. leptum subgroup Cluster IVAtopobium genusBacteroides groupEnterococcus genusLactobacillus groupC. perfringens subgroup Cluster IBifidobacterium group
  • *

    Day 14 baseline no challenge, (0 mg L−1) lupulone (n=4).

  • Day 22 no challenge, (0 mg L−1) lupulone (n=4).

  • Day 22 no challenge, (125 mg L−1) lupulone (n=4).

  • §

    Day 22 challenge with Clostridium perfringens, (0 mg L−1) lupulone (n=4).

  • Day 22 challenge with Clostridium perfringens, (125 mg L−1) lupulone (n=4).

Day 14 mean9.947.429.289.205.288.366.575.443.403.87
(SE)*0.020.220.140.090.200.040.220.470.270.04
NC0 mean9.917.249.199.515.268.136.285.634.953.82
(SE)0.060.090.120.030.280.150.270.330.850.14
NC125 mean9.687.179.018.995.437.856.395.933.213.76
(SE)0.100.180.260.070.100.130.220.590.160.09
C0 mean10.006.829.479.655.628.146.096.164.464.07
(SE)§0.070.220.230.170.230.100.270.390.610.25
C125 mean10.016.689.489.755.748.205.865.723.674.10
(SE)0.030.410.170.090.100.220.200.630.070.17
Table 4.   Mean counts of bacterial groups in the midgut expressed as log10 16S rRNA gene-specific gene copies per gram of intestinal material
 Total BacteriaEnterobacteriaceae familyC. coccoides subgroup Cluster XIVa, XIVbC. leptum subgroup Cluster IVBacteroides groupEnterococcus GenusLactobacillus groupC. perfringens subgroup Cluster IBfidobacterium group
  • Bacterial groups not present in at least 50% of the birds are expressed as not detected (ND).

  • *

    Day 14 baseline no challenge, (0 mg L−1) lupulone (n=4).

  • Day 22 no challenge, (0 mg L−1) lupulone (n=4).

  • Day 22 no challenge, (125 mg L−1) lupulone (n=4).

  • §

    Day 22 challenge with Clostridium perfringens, (0 mg L−1) lupulone (n=4).

  • Day 22 challenge with Clostridium perfringens, (125 mg L−1) lupulone (n=4).

Day 14 Mean7.356.523.05NDND5.393.513.20ND
(SE)*0.430.400.160.370.640.24
NC0 Mean7.306.843.274.04ND5.976.193.76ND
(SE)0.300.380.390.210.690.440.35
NC125 Mean7.407.153.734.193.016.234.782.68ND
(SE)0.500.310.150.130.410.510.440.37
C0 Mean7.336.263.874.543.685.366.383.43ND
(SE)§0.250.730.240.120.350.990.820.27
C125 Mean6.917.134.314.723.106.274.932.78ND
(SE)0.670.320.170.240.130.590.610.33

In the midgut, Enterobacteriaceae, C. coccoides subgroup and Enterococcus were detected in 100% of the samples taken for all groups. Lactobacillus was undetected in one sample from the lupulone treatment group and one sample from the control group. The C. leptum subgroup was undetected in two samples for the lupulone treatment group and twice for the control group. The presence–absence analysis for C. leptum subgroup showed a significant difference (P=0.0163) in presence for the challenge groups vs. the baseline Day 14 group. In the midgut, the C. perfringens subgroup was detected in seven out of eight for the control group and six out of eight for the lupulone treatment group. Atopobium was detected in two out of eight for the midgut samples for the lupulone treatment group, but was undetected in the control group. Bacteroides was detected in four out of eight for both the lupulone treatment group and the control group. Bifidobacteria were not detected in the midgut for the control group, but was detected in one out of eight for the lupulone treatment group. Because there was no statistical difference (P=0.46) between ingesta from the two different gut sections (i.e. cecum or midgut), we were able to combine data for all Gram-positive bacteria, regardless of the region, and found that lupulone had no significant effect (P=0.15) on the Gram-positive bacterial groups analyzed. There was no statistically significant effect of lupulone on the aggregated total Gram-negative bacteria assayed as well (P=0.51). However, two bacterial groups showed a significant effect when considered individually. The counts for C. perfringens subgroup in both the cecum (P=0.0253) and the midgut (P=0.0179) were significantly lower in the lupulone treatment group than in the control group. Lupulone treatment groups had significantly lower (P=0.0275) Lactobacillus counts in the midgut than did the control groups. Figures 1 and 2 show the mean bacterial group counts for the lupulone-treated group and the control untreated group of the cecum and midgut, respectively.

Figure 1.

 Bacterial groups for the cecum by treatment group. The experimental groups are identified as follows: Day 14, baseline; Day 14 HL0, 0 lupulone; HL125, treatment with lupulone (*P=0.0253).

Figure 2.

 Bacterial groups for the midgut by treatment group. The experimental groups are identified as follows: Day 14, baseline Day 14; HL0, 0 lupulone; HL125, treatment with lupulone (*P=0.0179, **P=0.0275).

Other statistical observations of the effects of hops-by-challenge interaction and challenge effect were made. There was a significant increase (P=0.0257) in the C. coccoides subgroup from Day 14 baseline to all Day 22 groups including lupulone and control groups. Counts for C. coccoides subgroup in the cecum were significantly higher (P=0.0030) in birds that were challenged. The challenge group had significantly (P=0.0008) higher counts for the C. leptum subgroup in the cecum, but not in the midgut. The lupulone treatment and challenge interaction also had significantly higher (P=0.0110) C. leptum subgroup counts in the cecum. The baseline Day 14 group had significantly lower (P=0.0289) counts of the C. leptum subgroup than any of the Day 22 groups. A correspondence analysis was performed to show the similarity of treatment on the mean copies per taxa for each group. In the midgut, all treatments clearly diverged from the baseline for both Gram-positive and Gram-negative taxa (Fig. 3). In the cecum, the major effect was observed for the 125 mg L−1 lupulone challenge treatment for the Gram-positive taxa (Fig. 3). When the data were pooled based on lupulone treatment alone without consideration of challenge with C. perfringens, an effect of both temporal succession and lupulone treatment could be observed in divergence from the baseline sample (Fig. 4).

Figure 3.

 Similarity of treatments considered by intestinal tract location as determined by correspondence analysis. (a) The response for the midgut and (b) for the cecum. Experimental treatment designations are as follows: Base, Day 14; NC0, no challenge with C. perfringens (0 lupulone); NC125, no challenge with C. perfringens (treatment with lupulone); C0, challenge with C. perfringens (0 lupulone); C125, challenge with C. perfringens (treatment with lupulone).

Figure 4.

 Similarity of control treatments and lupulone-treated groups considered by intestinal tract location as determined by correspondence analysis. (a) The response for the midgut and (b) for the cecum. Experimental treatment designations are as follows: Base, Day 14; H0 (control) – 0 lupulone; H125, treatment with lupulone.

Discussion

In this study, we used quantitative real-time PCR to quantify the major bacterial groups of the broiler cecum and midgut with or without the addition of 125 mg L−1 lupulone (hop β-acids) to the broiler diet. Our study design included quantifying bacterial groups at Days 14 and 22. We quantified the microbiota at Day 14 to assess any temporal succession in the bacterial community by Day 22, which represents the approximate point at which NE is generally observed in the field. The purpose of the study was to determine whether lupulone would have an effect on the microbiota of the gastrointestinal tract. We found that the addition of lupulone to the broiler diet did not significantly affect the overall microbiota of the cecum or the midgut, but did have a significant effect on several individual bacterial groups. Both Lactobacillus and C. perfringens subgroup counts were lower in the midgut for the lupulone treatment groups than in the zero lupulone treatment groups. We also found that C. perfringens subgroup Cluster I counts were significantly decreased in the cecum for the lupulone treatment. We have previously found that the treatment of the birds with lupulone led to decreased plate counts of C. pefringens in our colonization model (Siragusa et al., 2008). Lactobacillus and C. perfringens are Gram-positive bacteria, and we had expected that hop extract would have an antimicrobial effect as found previously (Teuber & Schmalreck, 1973; Srinivasan et al., 2004). Interestingly, Lactobacillus counts did not significantly decrease with hop exposure in the cecum. We found that the C. coccoides subgroup, C. leptum subgroup, Bacteroides and Enterobacteriaceae dominated the cecal environment as shown in both culture and molecular studies (Barnes et al., 1972; Salanitro et al., 1974; Lan et al., 2002; Lu et al., 2003; Dumonceaux et al., 2006). Enterobacteriaceae, Enterococcus and Lactobacillus dominated the midgut in all the experimental groups. There was no significant shift in the microbiota with lupulone treatment, consistent with a previous study that found no significant changes in the microbial diversity of the rat gastrointestinal tract exposed to xanthohumol, a hop extract (Hanske et al., 2005). Although not significant, Enterobacteriaceae appeared to increase in the midgut with lupulone exposure as the Lactobacillus counts decreased. Lactobacillus is considered to be part of a normal gut and possibly occupies a niche that prohibits pathogenic organisms from colonizing the gastrointestinal tract. In our study, lupulone had an antimicrobial effect on Lactobacillus that could potentially allow the proliferation of bacteria from the Enterobacteriaceae family including Salmonella and Escherichia coli. This interaction was not statistically significant possibly due to the high variation inherent in the small data set.

Previous studies have provided the basis for a qualitative analysis of the resident microbiota, but few studies have quantified the bacterial groups. In this study, we used gene sequences from highly conserved regions of the 16S rRNA gene to develop standard curve assays for real-time PCR. However, quantitative molecular methods for detecting microorganisms in a complex community also have limitations such as DNA extraction efficiency, PCR inhibitors, PCR bias and detection of dead cells. Definitive determination of the specificity and sensitivity of group-specific primer sets would also require unfeasibly large-scale cloning and sequencing of reverse transcriptase-PCR amplicons.

The bacterial groups assayed were found in the literature to be the predominant groups in the broiler gastrointestinal tract microbiota based on cultural methods and clonal sequence libraries. The C. perfringens subgroup (Cluster I) consists of C. perfringens, Clostridium butyricum, Clostridium botulinum and several Eubacterium spp. In this study, we observed a reduction in the C. perfringens subgroup. The C. leptum subgroup (Cluster IV) contains C. leptum, Ruminococcus, Eubacterium and Fusobacterium prausnutzii. Many of the species found within Cluster IV are mesophilic and cellulolytic and can confer the benefit of cellulose degradation for the host. The C. coccoides subgroup consists of C. coccoides, Eubacterium rectale and other butyrate-producing bacteria. Butyrate promotes gut health through cell proliferation and colonic cell turnover (Gong et al., 2002).

The Lactobacillus group, consisting of Lactobacillus spp., Leuconostoc, Pediococcus, Aerococcus and Weissella, is generally considered beneficial for gut health and as a group is often used as probiotics (Hofacre et al., 2003; La Ragione et al., 2004; Smirnov et al., 2005). The reduction of the Lactobacillus subgroup in the midgut of the lupulone-treated birds could have a deleterious effect, allowing the colonization of pathogenic species, and needs to be further investigated. The Enterococcus group, including Enterococcus faecalis, did not appear to show a significant change, which could be an important point for future studies. Antimicrobial use has been shown to correlate with an increase in antibiotic resistance in enterococci (reviewed by Sullivan et al., 2001). Campylobacter and Veillonella had no observable counts, but were expected to be present in the cecum. The absence of these two groups could be attributable to using a model system not reared on litter or PCR-inhibiting substances rather than the true absence of the organisms. The midgut also possessed several bacterial groups that were not detected in all samples. Again, the absence of these groups was likely attributable to relying on a model rearing system rather than actual field conditions. An explanation for the high variability seen, especially in the midgut, was the inconsistent rate at which groups of birds defecate and feed. The unpredictability of defecation and gut peristalsis could make midgut profile assays highly variable.

A better understanding of the quantitative levels of gastrointestinal bacterial groups and their interactions with the host organism will facilitate the development of intervention methods that continue the current feed efficiency and conversion in conventionally raised poultry and yet do not create or select antibiotic-resistant bacteria, limiting their prevalence in the environment. In the present study, we administered a purified hop component, lupulone, a hop β-acid, to the drinking water of broilers to assess the effect on the microbiota of the midgut and cecum. As lupulone is a Gram-positive antimicrobial (Teuber & Schmalreck, 1973), we have not lost sight of its small, but measurable reduction of not only the C. perfringens subgroup, but the potentially beneficial Lactobacillus group as well. We would propose that much like using traditional antibiotics such as penicillin for treating gut clostridial disease, lupulone reduces levels of other bacteria besides the targeted group; yet with judicious and timed use of antibiotics, a normal pretreatment state of eubacteriosis is once again established following withdrawal of the antibiotic. It leaves to consideration that lupulone acts in a similar manner, although this hypothesis was not experimentally tested here. The fate and uptake of lupulone in a complex biological system like the chicken intestine would be both important and relevant to the effect on the microbiota of the intestines, thereby warranting additional study of lupulone in a pharmocologic mode. The study of various lupulone delivery methods addressing the host's metabolic effect on lupulone and vice versa would be important in understanding its mode of action in the gut. Finally, lupulone has been reported to act in a synergistic fashion with other antibiotics (Haas, 2010) such as polymyxin B, both reducing the minimum inhibitory concentration as well as expanding the active range of inhibition. This synergy is possibly worth exploring in the avian system from the standpoint of both reducing antibiotic resistance, use levels and potentially expanding the range of antimicrobial activity.

Acknowledgement

Glenn Tillman completed this work as part of the requirements for a M.Sc. degree for the Interdisciplinary Toxicology Program at The University of Georgia. The authors would like to thank Johnna Garrish for her help throughout the study.

Ancillary