Pasture flock chicken cecal microbiome responses to prebiotics and plum fiber feed amendments

Pasture flock chicken cecal microbiome responses to prebiotics and plum fiber feed amendments S. H. Park,∗,† A. Perrotta,‡ I. Hanning,§,# S. Diaz-Sanchez,¶,∗∗ S. Pendleton,¶ E. Alm,‡ and S. C. Ricke∗,†,1 ∗

Cell and Molecular Biology Program, University of Arkansas, Fayetteville; † Center for Food Safety and Department of Food Science, University of Arkansas, Fayetteville; ‡ Department of Civil and Environmental Engineering and Biological Engineering, and the Center for Microbiome Informatics and Therapeutics, MIT, Cambridge, MA; § Department of Science, Lincoln International Academy, Managua, Nicaragua; # The Graduate School of Genome Sciences and Technology, University of Tennessee, Knoxville; ¶ Department of Food Science and Technology, University of Tennessee, Knoxville; and ∗∗ SaBio IREC (CSIC-UCLM-JCCM), Ciudad Real, Spain tives enrich for specific bacterial community members and modulate the diversity of the microbiome. We applied synthetic learning in microbial ecology (SLiME) analysis to interpret 16S rRNA microbial community data and identify specific bacterial operational taxonomic units (OTU) that are predictive of the particular feed additives used in these experiments. The results suggest that feed can influence microbiome composition in a predictable way, and thus diet may have indirect effects on weight gain and feed conversion through the microbiome.

ABSTRACT When prebiotics and other fermentation substrates are delivered to animals as feed supplements, the typical goal is to improve weight gain and feed conversion. In this work, we examined pasture flock chicken cecal contents using next generation sequencing (NGS) to identify and understand the composition of the microbiome when prebiotics and fermentation substrates were supplemented. We generated 16S rRNA sequencing data for 120 separate cecal samples from groups of chickens receiving one of 3 prebiotics or fiber feed additives. The data indicated that respective feed addi-

Key words: microbiome, prebiotics, pasture flock chicken 2017 Poultry Science 00:1–11 http://dx.doi.org/10.3382/ps/pew441

INTRODUCTION

The ceca are present at the junction of the small and large intestines and some fermentation does occur in the small intestine (Clench and Mathias, 1995), but is dependent on the type of substrate. Relative to the small intestines, the ceca are not a significant source of nutrients and cecectomy does not impair growth (Son and Karasawa, 2000). However, these pouches are heavily colonized with bacteria, which do provide the bird with small amounts of short-chain fatty acids (SCFA) and vitamins (Baurhoo et al., 2007). In chickens, a majority of the fermentation of prebiotics occurs in the ceca. However, the ceca are also the primary habitat for the zoonotic pathogens Salmonella and Campylobacter (Bailey et al., 1991; Patterson & Burkholder, 2003; Ricke, 2003, 2015; Hofacre et al., 2005; Hume, 2011). Typically, cecal samples are collected to culture target organisms of food safety concern or gastrointestinal indigenous bacteria that are assumed to have an impact on bird intestinal health or growth rates as the primary focus (Huyghebaert et al., 2011). However, there is a prevalence of bacteria that are uncultivable with conventional culturing methods (Ricke and Pillai, 1999; Kim and Mundt, 2011) and strong evidence that the microbial community is a factor dictating weight gain and

In agriculture, the addition of prebiotics to feed usually has one or 2 goals: 1) replacing sub-therapeutic dosages of antibiotics to obtain similar growth promoting results and/or 2) increasing the populations of beneficial bacteria by providing fermentation substrates that may subsequently reduce or eliminate zoonotic pathogens (Jones and Ricke, 2003; Patterson and Burkholder 2003; Hofacre et al., 2005; Callaway et al., 2008; Gaggia et al., 2010; Ricke, 2015). Broiler chickens raised on pasture flock have different management requirements compared to conventional systems in both rearing conditions as well as nutritional strategies. These chickens are grown in outdoor environments with feeds that do not include antimicrobial agents or growth promoters except natural feed supplements and dietary amendments (AMS/USDA, 2008; Park et al., 2013).  C 2017 Poultry Science Association Inc. Received March 5, 2016. Accepted November 16, 2016. 1 Corresponding author: [email protected]

1

2

PARK ET AL.

feed conversion (Hume, 2011; Huyghebaert et al., 2011; Stanley et al., 2012). Thus, understanding the structure of the entire chicken gut microbiome and its response to different dietary amendments may be an important step in increasing their ability to harvest energy from the diet efficiently. Prebiotic treatments can reduce zoonotic pathogens by supporting beneficial bacterial populations that potentially outcompete pathogens by lowering lumen pH, and increasing the concentrations of antimicrobial substances (Hume, 2011; Ricke, 2015). Fructooligosaccharides (FOS) and galactooligosaccharides (GOS) have both been used as food ingredients and as prebiotics for human consumption (Sako et al., 1999; Patterson and Burkholder, 2003; MacFarlane et al., 2006; Torres et al., 2010; Ricke, 2015). These oligosaccharides are composed of chains of fructose or galactose possessing glycosidic bonds that resist hydrolysis by endogenous enzymes and are available for fermentation by bacteria in the intestinal tract (Hidaka and Mirayama 1991; Alles et al., 1996; Donalson et al., 2007, 2008a, b; Park and Oh, 2010). Fermentation of prebiotics by intestinal bacteria produces SCFA, lactate, CO2 , and hydrogen and reduces the pH favoring the growth of beneficial bacteria and promoting calcium absorption (Gibson and Roberfroid, 1995; Ricke, 2003, 2015). Fruit pomaces, specifically plum, also can be considered as a prebiotic-like compound because they possess components that are indigestible by the host and may promote the growth of beneficial bacteria (Milala et al., 2013; Jarvis et al., 2015; Roto et al., 2015; Hutkins et al., 2016). In addition to providing a fermentation substrate, fruit pomaces also can supply additional health promoting factors such as polyphenols, anthocyanins, and antioxidants (Milala et al., 2013). Given the differences in the composition of these three feed additives and the potential differences in fermentation location and products, we hypothesized that the cecal microbial communities of the chickens would delineate among treatments in response to each particular feed amendment. In this study, our objective was to evaluate the effects of 2 prebiotics (FOS and GOS) and a plum fiber-based feed amendment on pasture raised chicken cecal content microbiome.

MATERIALS AND METHODS Bird housing A total of 340 day-of-hatch Cornish Cross White Plymouth Rock commercial broiler chicks was acquired from a local hatchery (Cobb 500; Cobb-Vantress, Fayetteville, AR) and raised on a pasture flock operation as part of the study described previously by Hanning et al. (2012). The chickens were divided into 4 groups with 85 chickens in each pen. The first 2 wk after hatch, the chickens were located in conventional housing pens with approximately 50 ft2 (4.65 m2 ). After 2 wk of age, the chickens were moved into outside pens on pas-

ture flock measuring approximately 80 ft2 (7.23 m2 ) for an additional 4 weeks. Every 2 wk, 10 birds from each group were randomly selected up to 6 wk for ceca extraction. A total of 120 birds (30 birds [2, 4, and 6 wk]) per group) was utilized for microbiome analysis. Further specific growth and environmental conditions are discussed in detail by Hanning et al. (2012). The University of Arkansas Institute Animal Care and Use Committee (IACUC) approval was exempted for this research because all birds were raised in an off-campus commercial facility and limited to microbiological evaluation.

Feed additives and feed formulation The feed additives were added to the starter, grower, and finisher feeds in each group and feed and water were made available ad libitum. Each group consisted of 85 chickens and treatments were as follows: 1) control (no feed additive); 2) plum fibers (California Dried Plum Board, Sacramento, CA) added at one kg per ton of feed; 3) FOS (GTC Nutrition, Golden, CO) added at one kg per ton of feed; and 4) GOS (GTC Nutrition) added at 2 kg per ton of feed, respectively. Those dosage levels were recommended by the manufacturers. Feed was supplemented with the prebiotics throughout the experimental period. Every 2 wk, a total of 10 chickens from each group was selected randomly, euthanized, and transported to the laboratory for necropsy. Both ceca were extracted and stored at −80◦ C until DNA extraction could be done.

Campylobacter isolation A sub-sample of cecal contents was collected for culture and quantification of Campylobacter. From the cecal contents, a sample weighing approximately one g was weighed and suspended in phosphate buffered saline (PBS) at a ratio of 1:9. Dilution series were conducted from this suspension and plated onto Campylobacter Line Agar (CLA). All plates were incubated in a microaerophilic atmosphere at 42◦ C for 48 hours. With this method, the limit of detection was 100 CFU per gram of cecal contents.

DNA isolation Total DNA from cecal contents was isolated using a Qiagen Stool Mini Kit (Qiagen, Valencia, CA) with modifications to increase DNA yields. Briefly, garnet beads (0.7 mm, Mo Bio Laboratories Inc., Carlsbad, CA) were utilized to lyse cells. Samples were centrifuged to pellet unhomogenized materials and the suspension was transferred into a microcentrifuge tube containing glass beads (0.1 mm, Mo Bio Laboratories Inc.). Vortexing was vigorously performed for 10 min and the samples were incubated at 95◦ C for 6 minutes. The remainder of the extraction protocol followed the

CECAL MICROBIOME OF PASTURE RAISED CHICKEN

manufacturer’s instructions. DNA concentration and purity were measured using a NanoDrop ND-1000 (Thermo Scientific, Wilmington, DE) and subsequently stored at −20◦ C until used.

16s rRNA amplicon library preparation and sequencing Ten nanograms of DNA from each cecal content were used in a paired-end Illumina HiSeq sequencer (Illumina, San Diego, CA) using a 2-step 16S rRNA PCR amplicon approach targeting the V4 region of the 16S rRNA. Primers U515F and E786R were used as described in Preheim et al. (2013). Samples were processed in sets no greater than 93 in a 96-well plate format with at least 3 negatives per plate. Each time point and treatment group was represented equally on each plate. All libraries were sequenced on a single lane for paired end 108 nt HiSeq sequencing runs at the Biomicro Center (MIT, Cambridge, MA) along with 284 other samples from independent studies.

sequencing data analysis and taxonomic assignment A total of 72,649,144 Illumina HiSeq reads was generated, 20,048,322 of which represented the prepared sequencing libraries for this study. Quality filtering and de-multiplexing resulted in an average of 169,901 reads per sample. The FASTQ files were filtered using the split library fastq.py protocol as part of the QIIME default pathway (Caporaso et al., 2010) with the following variables: The minimum length required was 100 bp (-min per read length 100); the last bad character used was Q (phred = 17, calculated error rate of 0.02) (–last bad character Q), no bad characters were allowed (– max bad run length 0), no barcode errors were allowed (– max barcode errors 0), no new cluster formation was allowed (-C), and a 7 base barcode (–barcode type 7). All other parameters were default parameters. The diversity and primer region were removed using the trim.seqs program in Mothur (Schloss et al., 2009). The resulting 77 bp reads were then aligned to the greengenes database (gg otus 4feb2011) and clustered into operational taxonomic units (OTU) using the QIIME pick otus.py command (-m uclust ref; all other parameters were default parameters). OTU tables subsequently were generated using the QIIME make otu table.py protocol. Samples were filtered for read counts, excluding samples with less than 1,000 reads, using the QIIME command filter samples from otu table.py –n 1000. Five samples (A2.4wk, A3.2wk, A6.6wk, C5.6wk, and D2.2wk) were excluded from all further analysis due to a low number of sequence reads associated with those samples. After initial quality processing, a total of 930 OTU was identified.

3

Microbiome analyses The diversity observed within a sample, alpha diversity, was measured using Shannon’s diversity index with sample diversity indices = [‘Shannon’] and dist mat (metric = ‘JS’) commands, respectfully, in pysurvey (https://bitbucket.org/yonatanf/pysurvey). The Shannon diversity metric used was Shannon entropy and is defined as H = −Sumi (pi ∗ log pi ), with pi being the fraction of OTUi . The effective number of species of order 1 represented in each sample was calculated by taking the inverse of the natural log of the calculated Shannon entropy. Diversity observed among samples, beta diversity, was measured using a weighted UniFrac (QIIME package, evenly sampled with 5,716 sequences per sample for all UniFrac calculations) and Spearman rank distances. Discriminating differences in presence, absence, and relative abundance at the OTU, genera, family, and phyla level were calculated using a Mann Whitney Utest. All P-values calculated were corrected by estimating the False Discovery Rate. All boxplots shown were generated using iPython (verion 2.7.3) and R (version 3.2.3). All assigned OTU were included for diversity calculations; all other analyses, unless noted otherwise, included only OTU with a relative abundance greater than 1e-03 considered for each treatment group within a time point. Correlations with metadata such as age, treatment, and pathogen count were conducted using synthetic learning in microbial ecology (SLiME) analysis (Papa et al., 2012). All OTU abundance data were analyzed with a log transformation (–x log). All meta data categories were analyzed without transformation except for pathogen count and weight data, which were run with a box cox transformation (–y boxcox). Analyses were run using 3 forests (-n 3) and the data were shuffled 100 times for P-value calculations (-s 100). All other parameters were default parameters. Correlations between OTU distributions and collected metadata were calculated using Kendall tau correlation coefficients for continuous data and receiver operator characteristic (ROC) curves with area under the curve (AUC) calculations for discrete data. Contributions of specific OTU to these correlations was calculated using percent increase mean squared error (IncMSE) and mean decrease accuracy (MDA) for continuous and discrete metadata, respectively.

RESULTS AND DISCUSSION Potential prebiotic feed additives can impact the gut microecology by changing the microbiota profile, mucin secretion, and histology by selectively enriching those taxa that can ferment the feed additives (Kaplan and Hutkins, 2003; Rebol´e et al., 2010; Roto et al., 2015; Hutkins et al., 2016). The end products of fermentation are dictated by the fermentation substrates provided (Baurhoo et al., 2007; Rebol´e et al., 2010; Roto

4

PARK ET AL.

et al., 2015). Oligosaccharides such as FOS and GOS can be converted to lactate by lactic acid bacteria or other metabolites (Shoaf et al., 2006; Ricke, 2015). Xu et al. (2003) demonstrated that broilers raised with FOS supplemented feeds increased beneficial bacteria such as Lactobacillus and Bifidobacterium while reducing Escherichia coli in the gut ecosystem. Shoaf et al. (2006) reported that GOS has an ability to inhibit the adherence of enteropathogenic E. coli on epithelial tissue cell. Aside from FOS and GOS, De Maesschalck et al. (2015) evaluated the beneficial effects of prebiotic xylo-oligosaccharides (XOS) on broiler growth performance and gut microbiota. They reported that XOS stimulated butyrate-producing bacteria and increased lactobacilli strains as well as potentially having positive effects on gastrointestinal function such as increase of villus length. For the birds used in the current study, Hanning et al. (2012) reported chicken growth performance results on villi and crypt measurement and body weights. At 2 wk of age, there were no significant differences in body weight but the GOS supplemented group exhibited lower body weights than other groups at 6 wk of age. For villi and crypt measurements at the end of the study (8 wk), FOS supplemented birds had the longest villi, while there

was no significant difference in crypt depth among groups.

Feed additives and bacterial taxonomic comparison From each treatment group described in Hanning et al. (2012), 9 to 10 chickens were collected and each microbial community in the cecal contents was sequenced. When viewing the samples individually within a treatment group, there was variability among the individual samples and among time points (Fig. S1). Bacteria belonging to the classes Bacteroidia and Clostridia were highly abundant across all time points (Fig. S1). The average abundance of the phyla, families, and genera was calculated for each treatment (Table 1). At the level of phylum, the Firmicutes were dominant in all groups at 2 wk of age (Table 1). The relative abundance of Firmicutes increased while the Bacteroidetes decreased at the 4-week time point for all treatments, but not the control; this trend continued into the 6-week time point for the FOS group only. Firmicutes appeared to be more prevalent than the Bacteroidetes in all cecal populations among the treatments, but this ratio

Table 1. Relative abundance of phyla, family, and genus for each treatment at each time point. Only those phylum, family, and genus that were present at a minimum of 4% relative abundance at any time point in any treatment are shown. Control

FOS

GOS

Plum

Taxonomic level

2 wk

4 wk

6 wk

2 wk

4 wk

6 wk

2 wk

4 wk

6 wk

2 wk

4 wk

6 wk

Phylum

Firmicutes Bacteroidetes Proteobacteria Tenericutes

59 31 9 2

52 43 3 1

47 41 6 3

51 40 5 5

64 27 5 2

66 22 4 3

51 35 11 2

69 23 4 2

53 37 3 2

55 36 6 2

67 25 4 2

49 30 13 4

Family

Ruminococcaceae Lachnospiraceae Bacteroidaceae Enterobacteriaceae Clostridiales Lactobacillaceae Catabacteriaceae Bacteroidales Rikenellaceae Porphyromonadaceae Rikenellaceae Halomonadaceae Alcaligenaceae

30 23 30 6 2 2 1 1 0 0 – – –

20 14 7 0 10 2 4 18 15 4 – – –

16 13 5 0 7 5 2 4 29 2 – – –

26 21 40 – 1 3 – 0 0 – 0 – –

31 21 6 – 7 2 – 7 12 – 12 – –

27 18 7 – 10 5 – 2 11 – 11 – –

19 23 23 – 1 6 – 10 1 – – 5 4

36 16 10 – 8 4 – 3 9 – – 0 2

18 21 15 – 5 3 – 2 17 – – 0 1

20 26 35 4 3 5 – 1 – – 0 – –

27 24 5 2 6 8 – 4 – – 13 – –

21 12 6 0 7 3 – 2 – – 19 – –

Genus

Bacteroides Ruminococaceae Lachnospiraceae Enterobacteriaceae Clostridiales Lactobacillus Catabacteriaceae Bacteroidales Rikenellaceae Alistipes Blautia Clostridium Faecalibacterium Eubacterium Alcaligenaceae Coprococcus

30 20 16 6 2 2 1 1 0 0 – – – – – –

7 15 6 0 10 2 4 18 3 12 – – – – – –

5 12 7 0 7 5 2 4 26 3 – – – – – –

40 21 11 – 1 3 – 0 – 0 3 3 1 – – –

6 21 11 – 7 2 – 7 – 12 2 4 5 – – –

7 19 8 – 10 5 – 2 – 10 4 4 2 – – –

23 14 9 – 1 6 – 10 – 1 – 4 2 5 4 –

10 23 8 – 8 4 – 3 – 7 – 3 6 1 2 –

15 13 14 – 5 3 – 2 – 15 – 4 2 0 0 –

35 14 15 4 3 5 – 1 0 0 – 3 2 – – 5

5 21 13 2 6 8 – 4 1 12 – 4 4 – – 2

6 16 6 0 7 3 – 2 7 12 – 2 1 – – 1

The numbers stand for relative abundance (%)

CECAL MICROBIOME OF PASTURE RAISED CHICKEN

varied by age and treatment. Although a higher ratio of Firmicutes to Bacteriodetes is correlated with weight gain in humans and mice, the ratio of these phyla in the chicken ceca did not correlate with weight (Ley et al., 2006). At the family taxonomic level, differences among the treatment groups were more obvious (Table 1). Ruminococcaceae, Lachnospiraceae, and Bacteroidaceae were the most abundant families in all groups, but differences in relative abundances of those taxa among the groups were observed over time. It appeared that each treatment had a unique impact on the microbial diversity. For example, the Halomonadaceae and Alcaligenaceae were present only in the GOS treatment group. Similarly, the Porphyromonadaceae and Catabacteriaceae were present only in the control group. Interestingly, the beneficial bacteria of the family Lactobacillaceae were underrepresented in all the samples. Similar to the family level, some genera were unique to specific treatment groups or the control (Table 1). Blautia was present only in the FOS group, while Eubacterium and Alcaligenaceae were present only in the GOS group. Coprococcus was unique to the plum group and Catabacteriaceae was present only in the control group, respectively. While diversity increased in the FOS and plum fiber supplemented groups over time, the control group did not gain a large number of additional OTU (Fig. 1). Interestingly, Fusobacterium was present at 6 wk in nearly all the cecal samples from chickens receiving plum fibers. Fusobacterium is a known butyrate producer, and fermentation of pectin produces butyrate (Duncan et al., 2002). Butyrate has numerous beneficial impacts on health, but for the chicken specifically, Salmonella growth is inhibited by butyrate (Donalson et al., 2008a). Thus, plum fibers may be a favorable choice as a feed additive for potential control of Salmonella in the chicken cecum. Variations in these microbial communities could produce differences in metabolite profiles, vitamins, and impacts on host-cellular expression. However, Qu et al.

5

(2008) found that even though diversity and evenness of bacterial taxa may differ among cecal samples, the functional gene content remained the same. Therefore, microbiome functionality may remain stable regardless of variations in community OTU abundance. However, this question of functional stability could not be addressed in this study, as metagenomic sequencing was not conducted. We observed shifts in abundance, presence, and absence of OTU in all groups over the experimental period. Some OTU were present in specific groups and absent from others. However, we observed that weighted UniFrac clustering distances between samples were highly dependent on age, not feed additive (Fig. 2). It has been shown previously that diversity is affected by factors other than age, including diet (Stanley et al., 2012). In this study, the diet formulation was changed at 2 wk of age (Hanning et al., 2012) and we have shown previously in other studies that adjustments to the diet may create microbial shifts (Hanning and Diaz-Sanchez, 2015). In these trials, feed was formulated to contain greater protein content for the first 2 wk to promote skeletal growth. This could partially explain the differences in clustering between the 2 and 4 wk age groups, but not between the 4 and 6 wk groups as the diet was the same for wk 4 and 6. However, 4 and 6 wk chickens digest nutrients differently as they age, which results in physiological variations, including enzymatic secretion, nutrient digestion, fat, and nutrient deposition. For example, while chickens develop the ability to digest fats with age, initially, fat digestion is very poor (Krogdahl, 1985). The ability to digest fats increases gradually until 6 wk of age (Katongole and March, 1980). Similarly, the digestion of carbohydrates improves with age because older birds are more resistant to amylase inhibitors present in the maize (Cowieson, 2005). These physiological differences may have impacted the intestinal ecology and substrates resulting from digestives that would be available for microbial succession downstream in the ceca.

Figure 1. Shannon diversity for A) control, B) FOS, C) GOS, and D) plum fiber treated groups. ∗ represents a q-value equal to or less than 0.05 compared with control group at same time point.

6

PARK ET AL.

Figure 2. Principal Coordinate Analysis (PCoA) plots of beta diversity exhibit the clustering to understand microbial population distance of each bird based on age (panel A) and treatment (panel B).

Microbial diversity trends of the cecal microbiome associated with feed additive type

Table 2. Correlative analysis of meta data and microbial community distribution. Category

Kendall tau coefficient

AUC

P-value

In accordance with previous observations on the strong temporal effects of host skeletal and weight gain growth on microbial gut communities (Lu et al., 2003), clustering samples by beta diversity showed that host age, and therefore developmental stage, represented the primary clustering distance (Fig. 2). This indicates that although microbial populations can be shifted with prebiotics, there remains a very strong host effect upon the microbiome in the chicken ceca. Alpha-diversity (Shannon index) increased over time for all treatment groups, but samples from chickens fed FOS (Fig. 1B) and plum fiber (Fig. 1D) exhibited significantly greater diversity than the control group at the 6-week time point (Fig. 1A). The increase in diversity over time was not surprising as it is well known that diversity increases with age in many animals as well as humans (Rodriguez et al., 2015). Specific to the chicken, the intestinal tract is considered sterile at hatch but quickly becomes colonized with aerobic bacteria. The intestines are relatively flat with minimal mucosa (Hanning and Diaz-Sanchez, 2015). These aerobic bacteria stimulate the histological development and mucosa secretion, which leads to a suitable environment for anaerobes. The anaerobes easily outcompete the aerobic bacteria and are usually dominant by 2 wk of age (Hanning and Diaz-Sanchez, 2015). Correlation analysis of metadata and microbial community distribution was conducted using SLiME (Papa et al., 2012), which identified 4 OTU that were significantly predictive of important host factors including age, weight, pathogen count, and treatment. Kendall tau correlation coefficients, AUC calculations for ROC

Age Weight Pathogen count Control vs. all FOS vs. control GOS vs. control Plum vs. control

0.79 0.65 0.41 – – – –

– – – 0.89 0.95 0.93 0.89

< 10e-12 < 10e-12 8.53e-12 2e-4 8.41e-11 1.03e-10 2.36e-1

curves, and corresponding P-values for all correlations are listed in Table 2. An OTU identified as belonging to the genus Alistipes increased with age in all treatment groups and was indicated as a predictive feature for age (>0.6% IncMSE), weight (>150% IncMSE), and Campylobacter populations (>10% IncMSE). As seen in Fig. 3, this OTU had significantly greater relative abundance in treatment groups compared to the control group at the 6-week time point. A Lactobacillus intestinalis OTU was indicated in our analysis as being a predictive feature for Campylobacter populations (>25% IncMSE) and treatments (FOS vs. control > 0.0025 MDA; GOS vs. control 0.0025 MDA, plum fibers vs. control 0.003 MDA) (Fig. 4). This OTU was particularly predictive for the FOS treatment compared to the control at 2 wk of age, which supports other reports that FOS selectively enriches for Lactobacillus (Swanson et al., 2002). Lactobacillus species are known to be capable of hydrolysing bile salts, which have been associated with host gut health (Feighner and Dashkevicz, 1988; Tannock et al., 1989; Swennen et al., 2006) and the inhibition of pathogenic bacterial colonization (Dec et al., 2014). Although Lactobacillaceae were present in all groups at all time points (Table 1), they were relatively underrepresented in the

CECAL MICROBIOME OF PASTURE RAISED CHICKEN

7

Figure 3. Relative percent abundance of Alistipes (genus) OTU in A) control, B) FOS, C) GOS, and D) plum fibers treated groups. This OTU was highly predictive of age (0.65% IncMSE), weight (150% IncMSE), and pathogen count (> 10% IncMSE). ∗ represents a q-value equal to or less than 0.05 and ∗∗ represents a q-value equal to or less than 0.005 compared with control group at same time point.

Figure 4. Relative percent abundance of Lactobacillus intestinalis OTU in A) control, B) FOS, C) GOS, and D) plum fibers treated groups. This OTU was predictive of treatment particularly for FOS compared to control at the 2-week time point, mean decrease accuracy of > 0.015, and pathogen count, > 20% increase in mean squared error. ∗ represents a q-value equal to or less than 0.05 and ∗∗ represents a q-value equal or less than 0.005 compared with control group at same time point.

ceca ranging from 2 to 8% relative abundance of the population. In contrast to the ceca, this genus dominates the small intestines, which is a region that favors facultative anaerobes and bile salt resistant species, while the cecal environment appears to be more suitable for strict anaerobes (Knarreborg et al., 2002; AmitRomach et al., 2004; Saengkerdsub et al., 2007; Saengkerdsub and Ricke, 2014). Prebiotics can selectively enrich bacterial taxa, and our analyses indicated that the FOS and GOS treatments selected for Faecalibacterium prausnitzii (Fig. 5B and C). This OTU had a relative abundance just below our analysis threshold but was included, as it is well covered in the published literature. F. prausnitzii is highly abundant in the human colon where it is known to produce butyrate (Lopez-Siles et al., 2012). Butyrate has been shown to reduce Salmonella infec-

tion, inhibit inflammation, and can promote cellular mucosal turnover rates (Cerisuelo et al., 2014; Parson et al., 2014). F. prausnitzii has been identified in several publications as being involved in butyrate production and bile acid hydrolysis and is enriched by prebiotic treatments in humans and mice (Feighner and Dashkevicz, 1988; Ramirez-Farias et al., 2009; Lopez-Siles et al., 2012). The F. prausnitzii OTU was almost absent from all control samples but was significantly enriched in the FOS and GOS treatment groups in the 6-week time point (Fig. 5). Thus selective enrichment of this species by FOS and GOS could be beneficial in terms of host health. It also should be noted that F. prausnitzii is one of the few colonic community members that can ferment pectin (Lopez-Siles et al., 2012). However, this OTU was not enriched in the plum pomace group.

8

PARK ET AL.

Figure 5. Relative percent abundance of a Faecalibacterium prausnitzii OTU in A) control, B) FOS, C) GOS, and D) plum fibers treated groups. This OTU was predictive of treatment particularly FOS and GOS compared to control at the 6-week time point, mean decrease accuracy of > 0.005. ∗∗ represents a q-value equal to or less than 0.005 compared with control group at same time point.

Numerous additional OTU were indicated as predictive of treatment in this analysis but due to pen effect in treatment groups and low relative abundance, only OTU indicated in previously published literature and a relative abundance greater than 1e-03 were considered. It should be noted that many of the low abundance OTU were identified as belonging to the family Ruminococcacea whose members have been reported previously to exhibit a butyrogenic effect and to be enriched by prebiotic supplementation (Walker et al., 2011).

Microbiome characterization and Campylobacter populations in the ceca Due to the relevance of Campylobacter in the poultry industry (Horrocks et al., 2009), we conducted separate culturing counts for Campylobacter and correlated these counts with the cecal community members. In all groups, Campylobacter counts were not significantly different and concentrations ranged from 6 to 8 log cfu per gram of intestinal content (data not shown). Control of Campylobacter has been difficult and control methods have had little success (Willis and Reid, 2008; Horrocks et al., 2009; Gaggia et al., 2010). One reason may be the symbiotic relationship Campylobacter has within the bacterial community (Indikova et al., 2015). Campylobacters do not code for phosphofructokinase, which is necessary for sugar fermentation (Velayudhan and Kelly, 2002). Thus, it relies on other bacteria for secondary metabolites. Second, Campylobacter also can act as a hydrogen sink favoring the fermentation process (Laanbroek et al., 1978). Since excess hydrogen inhibits acetate production, Campylobacter can promote a consistent concentration of acetate benefiting the rest of the microbial community. In fact, it is for this reason that some suggest Campylobacter is actually beneficial to the chicken intestinal health (Sergeant et al., 2014). We did observe that Alistipes and Lactobacillus intestinalis OTU appeared to be predictive of Campy-

lobacter population levels. Campylobacter species can utilize a wide range of organic acids including succinic acid, the major metabolic product of Alistipes species (Kaakoush et al., 2014). Similarly, lactate is the major fermented by-product of Lactobacillus, and Campylobacter possesses 2 enzyme pathways for the utilization of L-lactate. Although these 2 bacterial OTU were predictive of Campylobacter, it is unclear if there was a true mutualistic relationship. Since L. intestinalis was present at a very low abundance numerically, any metabolic by-products produced by L. intestinalis that Campylobacter would also be low in abundance. Alistipes was present in greater abundance than L. intestinalis and this OTU was also predictive of age. Independently, Campylobacter is also predictive of age and it is well known that this zoonotic pathogen colonizes older chickens that have a mature gut histology and microbiota (Dhillon et al., 2006). Given that Campylobacter typically becomes detectable when chickens are 2 to 3 wk of age, the connection between Alistipes and Campylobacter may be a factor of age rather than metabolite sharing. In the future, in vitro studies focused on co-culturing of Campylobacter and these organisms’ cecal contents are needed to delineate any potential interdependent metabolic relationship.

CONCLUSIONS In conclusion, the data collected here provide genomic foundations for further refinement of prebiotic administration. The data gathered here can be used to complement general assessments with more targeted approaches of particular groups of cecal bacteria. In addition, targeted assays that quantify specific populations could potentially be designed for use in other studies as well as establish interdependent microbial metabolic groups occurring in the chicken ceca.

CECAL MICROBIOME OF PASTURE RAISED CHICKEN

SUPPLEMENTARY DATA Figure S1. Relative abundance of taxonomic class level in cecal contents of each bird by age and treatment. Each bar represents one bird. Supplementary data are available at PSCIEN online.

ACKNOWLEDGMENTS Research funding was awarded to author SCR by the USDA/CSREES (Cooperative State Research, Education, and Extension Service, Washington DC) National Integrated Food Safety Initiative (NIFSI, Washington DC) grant # 406 2008 51110, USDA CREES Food Safety Consortium grant, the Plum Board (Sacramento, CA), and a donation of FOS and GOS prebiotics from GTC Nutrition Co. (Golden, CO).

REFERENCES Alles, M. S., J. G. A. J. Hautvast, F. M. Nagengast, R. Hartemink, K. M. J. Van Laere, and J. B. M. J. Jansen. 1996. Fate of fructooligosaccharides in the human intestine. Br. J. Nutr. 76:211–221. Amit-Romach, E., D. Sklan, and Z. Uni. 2004 Microflora ecology of the chicken intestine using 16S ribosomal DNA primers. Poult. Sci. 83:109–1098. AMS/USDA. 2008. The national organic program. Agricultural Marketing Service-USDA, Washington, DC. Available: http://www.ams.usda.gov/about-ams/programs-offices/ national-organic-program. Accessed 16 February 2016. Bailey, J. S., L. C. Blankenship, and N. A. Cox. 1991. Effect of fructooligosaccharide on Salmonella colonization of the chicken intestine. Poult. Sci. 70:2433–2438. Baurhoo, B., L. Phillip, and C. A. Ruiz-Feria. 2007. Effects of purified lignin and mannan oligosaccharides on intestinal integrity and microbial populations in the ceca and letter of broiler chickens. Poult. Sci. 86:1070–1078. Callaway, T. R., T. S. Edrington, R. C. Anderson, R. B. Harvey, K. J. Genovese, C. N. Kennedy, and D. J. Nisbet. 2008. Probiotics, prebiotics and competitive exclusion for prophylaxis against bacterial disease. Anim. Health Res. Rev. 9:217–225. Caporaso, J. G., J. Kuczynski, J. Stombaugh, K. Bittinger, F. D. Bushman, E. K. Costello, N. Fierer, A. G. Pena, J. K. Goodrich, J. I. Gordon, G. A. Huttley, S. T. Kelly, D. Knights, J. E. Koenig, R. E. Ley, C. A. Lozupone, D. McDonald, B. D. Muegge, M. Pirrung, J. Reeder, J. R. Sevinsky, P. J. Turnbaugh, W. A. Walters, J. Widmann, T. Yatsunenko, J. Zaneveld, and R. Knight. 2010. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods. 7:335–336. Cerisuelo, A., C. Mar´ın, F. S´ anchez-Vizca´ıno, E. A. G´ omez, J. M. de la Fuente, R. Dur´ an, and C. Fern´ andez. 2014. The impact of a specific blend of essential oil components and sodium butyrate in feed on growth performance and Salmonella counts in experimentally challenged broilers. Poult. Sci. 93: 599–606 Clench, M. H., and J. R. Mathias. 1995. The avian cecum: A review. Wilson Bull. 107:93–121. Cowieson, A. J. 2005. Factors that affect the nutritional value of maize for broilers. Anim. Feed Sci. Tech. 119:293–305. De Maesschalck, C., V. Eeckhaut, L. Maertens, L. De Lange, L. Marchal, C. Nezer, S. De Baere, S. Croubels, G. Daube, J. Dewulf, F. Haesebrouck, R. Ducatelle, B. Taminau, and F. Van Immerseel. 2015. The effects of xylo-oligosaccharides on performance and microbiota in broiler chickens. Appl. Environ. Microbiol. 81:5880– 5888. Dec, M., A. Puchalski, R. Urban-Chmiel, and A. Wernicki. 2014. Screening of Lactobacillus strains of domestic goose origin against

9

bacterial poultry pathogens for use as probiotics. Poult. Sci. 93: 1–9. Dhillon, A. S., H. L. Shivaprasad, D. Schaberg, F. Wier, S. Weber, and D. Bandli. 2006. Campylobacter jejuni infection in broiler chickens. Avian Dis. 50:55–58. Donalson, L. M., W. K. Kim, V. I. Chalova, P. Herrera, J. L. McReynolds, V. G. Gotcheva, D. Vidanovi´c, C. L. Woodward, L. F. Kubena, D. J. Nisbet, and S. C. Ricke. 2008a. In vitro fermentation response of laying hen cecal bacteria to combinations of fructooligosaccharide prebiotics with alfalfa or a layer ration. Poult. Sci. 87:1263–1275. Donalson, L. M., J. L. McReynolds, W. K. Kim, V. I. Chalova, C. L. Woodward, L. F. Kubena, D. J. Nisbet, and S. C. Ricke. 2008b. The influence of a fructooligosaccharide prebiotic combined with alfalfa molt diets on the gastrointestinal tract fermentation, Salmonella Enteritidis infection, and intestinal shedding in laying hens. Poult. Sci. 87:1253–1262. Donalson, L. M., W. K. Kim, V. I. Chalova, P. Herrera, C. L. Woodward, J. L. McReynolds, L. F. Kubena, D. J. Nisbet, and S. C. Ricke. 2007. In vitro anaerobic incubation of Salmonella enterica serotype Typhimurium and laying hen cecal bacteria in poultry feed substrates and a fructooligosaccharide prebiotic. Anaerobe. 13:208–214. Duncan, S. H., G. L. Hold, H. J. M. Harmsen, C. S. Stewart, and H. J. Flint. 2002. Growth requirements and fermentation products of Fusobacterium prausnitzii, and a proposal to reclassify it as Faecalibacterium prausnitzii gen. nov., comb. nov. Int. J. Sys. Evol. Microbiol. 52:2141–2146. Feighner, S. D., and M. P. Dashkevicz. 1988. Effect of dietary carbohydrates on bacterial cholytaurine hydrolase in poultry intestinal homogenates. Appl. Environ. Microbiol. 54: 337–342. Gagg`ıa, F., P. Mattarelli, and B. Biavati. 2010. Probiotics and prebiotics in animal feeding for safe food production. Int. J. Food Microbiol. 141:S15–S28. Gibson, G. R., and M. B. Roberfroid. 1995. Dietary modulation of the human colonic microbiota: Introducing the concept of prebiotics. J. Nutr. 125:1401–1412. Hanning, I., A. Clement, C. Owens, S. H. Park, S. Pendleton, E. E. Scott, G. Almeida, F. Gonzalez Gil, and S. C. Ricke. 2012. Assessment of production performance in 2 breeds of broilers fed prebiotics as feed additives. Poult. Sci. 91: 3295–3299. Hanning, I., and S. Diaz-Sanchez. 2015. The functionality of the gastrointestinal microbiome in non-human animals. Microbiome. 3:51. Hidaka, H., and M. Mirayama. 1991. Useful characteristics and commercial applications of fructo-oligosaccharides. Biochem. Soc. Trans. 19:561–565. Hofacre, C. L., G. F. Mathis, and M. A. Quiroz. 2005. Natural alternatives to prevent necrotic enteritis. Int. Poult. Prod. 13: 7–9. Horrocks, S. M., R. C. Anderson, D. J. Nisbet, and S. C. Ricke. 2009. Incidence and ecology of Campylobacter jejuni and coli in animals. Anaerobe. 15:18–25. Hume, M. E. 2011. Historic perspective: Prebiotics, probiotics, and other alternatives to antibiotics. Poult. Sci. 90: 2663–2669. Hutkins, R. W., J. A. Krumbeck, L. B. Bindels, P. D. Cani, G. Fahey, Jr., Y. J. Goh, B. Hamaker, E. C. Martens, D. A. Mills, R. A. Rastal, E. Vaughan, and M. E. Sanders. 2016. Prebiotics: Why definitions matter. Curr Opin Biotech. 37:1–7. Huyghebaert, G., R. Ducatelle, and F. van Immerseel. 2011. An update on alternatives to antimicrobial growth promoters for broilers. Vet. J. 187:182–188. Indikova, I., T. J. Humphrey, and F. Hilbert. 2015. Survival with a helping hand: Campylobacter and microbiota. Front. Microbiol. 6:1266. Jarvis, N., C. A. O’Bryan, S. C. Ricke, and P. G. Crandall. 2015. The functionality of plum ingredients in meat products: A review. Meat Sci. 102:41–48. Jones, F. T., and S. C. Ricke. 2003. Observation on the history of the development of antimicrobials and their use in poultry feeds. Poult. Sci. 82:613–617.

10

PARK ET AL.

Kaakoush, N. O., N. Sodhi, J. W. Chenu, J. M. Cox, S. M. Riordan, and H. M. Mitchell. 2014. The interplay between Campylobacter and Helicobacter species and other gastrointestinal microbiota of commercial broiler chickens. Gut Pathog. 6:18. Kaplan, H., and R. W. Hutkins. 2003. Metabolism of fructooligosaccharides by Lactobacillus paracasei 1195. Appl. Environ. Microbiol. 69:2217–2222. Katongole, J. B. D., and B. E. March. 1980. Fat utilization in relation to intestinal fatty acid binding protein and bile salts in chicks of different ages and different genetic sources. Poult. Sci. 59:819– 827. Kim, T., and E. Mundt. 2011. Metagenomic analysis of intestinal microbiomes in chickens. Pages 185–194 in In High-Throughput Next Generation Sequencing. S. C. Ricke, and Y. M. Kwon, eds. Human Press, United Kingdom. Knarreborg, A., M. A. Simon, R. M. Engberg, B. B. Jensen, and G. W. Tannock. 2002. Effects of dietary fat source and subtherapeutic levels of antibiotic on the bacterial community in the ileum of broiler chickens at various age. Appl. Environ. Microbiol. 68:5918–5924. Krogdahl, A. 1985. Digestion and absorption of lipids in poultry. J. Nutr. 115:675–685. Laanbroek, H. J., L. J. Stal, and H. Veldkamp. 1978. Utilization of hydrogen and formate by Campylobacter spec. under aerobic and anaerobic conditions. Arch Microbiol. 119:99–102. Ley, R. E., P. J. Turnbaugh, S. Klein, and J. L. Gordon. 2006. Microbial ecology: Human gut microbes associated with obesity. Nature. 444:1022–1023. Lopez-Siles, M., T. M. Khan, S. H. Duncan, J. M. Harmsen, L. J. Garcia-Gil, and H. J. Flint. 2012. Cultured representatives of two major phylogroups of human colonic Faecalibacterium prausnitzii can utilize pectin, uronic acids, and host-derived substrates for growth. Appl. Environ. Microbiol. 78:420–428. Lu, J., U. Idris, B. Harmon, C. Hofacre, J. J. Maurer, and M. D. Lee. 2003. Diversity and succession of the intestinal bacterial community of the maturing broiler chicken. Appl. Environ. Microbiol. 69:6816–6824. MacFarlane, S., G. T. Macfarlane, and J. T. Cummings. 2006. Review article: Prebiotics in the gastrointestinal tract. Aliment Pharmacol. Ther. 24:701–714. Milala, J., M. Kosmala, M. S´ ojka, K. Kolodziejczyk, M. Zbrze´zniak, and J. Markowski. 2013. Plum pomaces as a potential source of dietary fibre: Composition and antioxidant properties. J. Food Sci. Technol. 50:1012–1017. Papa, E., M. Docktor, C. Smillie, S. Weber, S. P. Preheim, D. Gevers, G. Giannoukos, D. Ciulla, D. Tabbaa, J. Ingram, D. B. Schauer, D. V. Ward, J. R. Korzenik, R. J. Xavier, A. Bousvaros, and E. J. Alm. 2012. Non-invasive mapping of the gastrointestinal microbiota identifies children with inflammatory bowel disease. PloS One. 7:e39242. Park, A. R., and D. K. Oh. 2010. Galacto-oligosaccharide production using microbial β -galactosidase: Current state and perspectives. Appl. Microbiol. Biotechnol. 85:1279–1286. Park, S. H., I. Hanning, A. Perrota, B. J. Bench, E. Alm, and S. C. Ricke. 2013. Modifying the gastrointestinal ecology in alternatively raised poultry and the potential for molecular and metabolomics assessment. Poult. Sci. 92:546–561. Parson, B. N., P. Wigley, H. L. Simpson, J. M. Williams, S. Humphrey, A.-M. Salisbury, A. J. M. Watson, S. C. Fry, D. O’Brien, C. L. Roberts, N. O’Kennedy, A. V. Keita, J. D. S¨ oderholm, J. M. Rhodes, and B. J. Campbell. 2014. Dietary supplementation with soluble plantain non-starch polysaccharides inhibits intestinal invasion of Salmonella Typhimurium in the chicken. PLoS One. 9:e87658. Patterson, J. A., and K. M. Burkholder. 2003. Application of prebiotics and probiotics in poultry production. Poult. Sci. 82: 627–631. Preheim, S. P., A. R. Perrotta, A. M. Martin-Platero, A. Gupta, and E. J. Alm. 2013. Distribution-based clustering: Using ecology to refine the operational taxonomic unit. Appl. Environ. Microbiol. 79:6593–6603. Qu, A., J. M. Brulc, M. K. Wilson, B. F. Law, J. R. Theoret, L. A. Joens, M. E. Konkel, F. Angly, E. A. Dinsdale, R. A. Edwards, K. E. Nelson, and B. A. White. 2008. Comparative metagenomics

reveals host specific metavirulomes and horizontal gene transfer elements in the chicken cecum microbiome. PLoS One. 8:e2945. Ramirez-Farias, C., K. Slezak, Z. Fuller, A. Duncan, G. Holtrop, and P. Louis. 2009. Effect of inulin on the human gut microbiota: Stimulation of Bifidobacterium adolescentis and Faecalibacterium prausnitzii. Br. J. Nutr. 101:541–550. Rebol´e, A., L. T. Ortiz, M. L. Rodr´ıguez, C. Alzueta, J. Trevino, and S. Velasco. 2010. Effects of inulin and enzyme complex, individually or in combination, on growth performance, intestinal microflora, cecal fermentation characteristics, and jejunal histomorphology in broiler chickens fed a wheat- and barley-based diet. Poult. Sci. 89:276–286. Ricke, S. C. 2003. Perspectives on the use of organic acids and short chain fatty acids as antimicrobials. Poult. Sci. 82: 632–639. Ricke, S. C. 2015. Potential of fructooligosaccharide prebiotics in alternative and nonconventional poultry production systems. Poult. Sci. 94:1411–1418. Ricke, S. C., and S. D. Pillai. 1999. Conventional and molecular methods for understanding probiotic bacteria functionality in gastrointestinal tracts. Crit. Rev. Microbiol. 25:19–38. Rodriguez, J. M., K. Murphy, C. Stanton, R. P. Ross, O. I. Kober, and N. Juge. 2015. The composition of the gut microbiota throughout life, with an emphasis on early life. Microb. Ecol. Health Dis. 26:10. Roto, S. M., P. M. Rubinelli, and S. C. Ricke. 2015. An introduction to the avian gut microbiota and the effects of yeast-based prebiotic-type compounds as potential feed additives. Front. Vet. Sci. 2:28. Saengkerdsub, S., and S. C. Ricke. 2014. Ecology and characteristics of methanogenic archaea in animals and humans. Crit. Rev. Microbiol. 40:97–116. Saengkerdsub, S., R. C. Anderson, H. H. Wilkinson, W. K. Kim, D. J. Nisbet, and S. C. Ricke. 2007. Identification and quantification of methanogenic archaea in adult chicken ceca. Appl. Environ. Microbiol. 73:353–356. Sako, T., K. Matsumoto, and R. Tanaka. 1999. Recent progress on research and application of non-digestible galato-oligosaccharides. Int. Dairy J. 9:69–80. Schloss, P. D., S. L. Westcott, T. Ryabin, J. R. Hall, M. Hartmann, and E. B. Hollister. 2009. Introducing mothur: Opensource, platform-independent, community-supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75:7537–7541. Sergeant, M. J., C. Constantinidou, T. A. Cogan, M. R. Bedford, C. W. Penn, and M. J. Pallen. 2014. Extensive microbial and functional diversity within the chicken cecal microbiome. PLoS One. 9:e91941. Shoaf, K., G. L. Mulvey, G. D. Armstrong, and R. W. Hutkins. 2006. Prebiotic galactooligosaccharides reduce adherence of enteropathogenic Escherichia coli to tissue culture cells. Infect. Immun. 74:6920–6928. Son, J. H., and Y. Karasawa. 2000. Effect of removal of caecal contents on nitrogen utilization and nitrogen excretion in caecally ligated chickens fed on a low protein diet supplemented with urea. Br. Poult. Sci. 41:69–71. Stanley, D., S. E. Denman, R. J. Hughes, M. S. Geier, T. M. Crowley, H. Chen, V. R. Haring, and R. J. Moore. 2012. Intestinal microbiota associated with differential feed conversion efficiency in chickens. Appl. Microbiol. Biotechnol. 96:1361–1369. Swanson, K. S., C. M. Grieshop, E. A. Flickinger, L. L. Bauer, J. Chow, B. W. Wolf, K. A. Garleb, and G. C. Fahey Jr. 2002. Fructooligosaccharides and Lactobacillus acidophilus modify gut microbial populations, total tract nutrient digestibilities and fecal protein catabolite concentrations in healthy adult dogs. J. Nutr. 132:3721–3731. Swennen, K., C. M. Courtin, and J. A. Delcour. 2006. Non-digestible oligosaccharides with prebiotic properties. Crit. Rev. Food. Sci. Nutr. 46:459–471. Tannock, G. W., M. P. Dashkevicz, and S. D. Feighner. 1989. Lactobacilli and bile salt hydrolase in the murine intestinal tract. Appl. Environ. Microbiol. 55:1848–1851. Torres, D. P., M. D. P. F. Gon¸calves, J. A. Teixeira, and L. R. Rodrigues. 2010. Galacto-oligosaccharides: Production, properties,

CECAL MICROBIOME OF PASTURE RAISED CHICKEN applications, and significance as prebiotics. Compr. Rev. Food. Sci. F. 9:438–454. Velayudhan, J., and D. J. Kelly. 2002. Analysis of gluconeogenic and anaplerotic enzymes in Campylobacter jejuni: An essential role for phosphoenolpyruvate carboxykinase. Microbiology 148:685–694. Walker, A. W., J. Ince, S. H. Duncan, L. M. Webster, G. Holtrop, X. Ze, D. Brown, M. D. Stares, P. Scott, A. Bergerat, P. Louis, F. McIntosh, A. M. Johnstone, G. E. Lobley, J. Parkhill, and H. J. Flint. 2011. Dominant and diet-responsive groups

11

of bacteria within the human colonic microbiota. ISME J. 5: 220–230. Willis, W. L., and L. Reid. 2008. Investigating the effects of dietary probiotic feeding regimens on broiler chicken production and Campylobacter jejuni presence. Poult. Sci. 87:606–611. Xu, Z. R., C. H. Hu, M. S. Xia, X. A. Zhan, and M. Q. Wang. 2003. Effects of dietary fructooligosaccharide on digestive enzyme activities, intestinal microflora and morphology of male broilers. Poult. Sci. 82:1030–1036.