Gut metagenomic analysis reveals prominent roles of Lactobacillus and cecal microbiota in chicken feed efficiency

Interactions between the host and gut microbiota can affect gut metabolism. In this study, the individual performances of 252 hens were recorded to evaluate feed efficiency. Hens with contrasting feed efficiencies (14 birds per group) were selected to investigate their duodenal, cecal and fecal microbial composition by sequencing the 16S rRNA gene V4 region. The results showed that the microbial community in the cecum was quite different from those in the duodenum and feces. The highest biodiversity and all differentially abundant taxa between the different efficiency groups were observed in the cecal microbial community with false discovery rate (FDR) <0.05. Of these differentially abundant cecal microbes, Lactobacillus accounted for a greater proportion than the others. The abundances of Lactobacillus and Akkermansia were significantly higher while that of Faecalibacterium was lower (FDR < 0.05) in the better feed efficiency (BFE) group. Phylogenetic investigation of communities by reconstruction of unobserved states (PICRUSt) analysis revealed that the functions relating to glycometabolism and amino acid metabolism were enriched in the cecal microbiota of the BFE group. These results indicated the prominent role of cecal microbiota in the feed efficiency of chickens and suggested plausible uses of Lactobacillus to improve the feed efficiency of host. Supplementary information The online version of this article (doi:10.1038/srep45308) contains supplementary material, which is available to authorized users.


Introduction 31
The gastrointestinal tract is the major site of food digestion and nutrient absorption. 32 Cecum is the chief functional section in the distal intestine, and its importance in birds' 33 metabolism has received increasing attention 1,2 . The cecum, which is full of microbial 34 fermentations, plays important roles in preventing pathogen colonization, detoxifying 35 harmful substances, recycling nitrogen and absorbing additional nutrients 3 . The 36 digestibility and the ability to metabolize crude fiber or other nutrients are lower in 37 birds with a cecectomy than in normal birds 4 . In addition, significant absorption of 38 glucose was observed in the cecum 5 , and a higher ability to actively absorb sugars at 39 low concentrations was found in the cecum compared with the jejunum 6 . Located at 40 the beginning of the intestine, the duodenum is crucial for feed digestion and 41 absorption; it has a lower pH than the hindgut and is the region that absorbs most 42 glucose 7 and other nutrients within the small intestine 8,9 . 43 Although the cecum and the duodenum themselves are important, interactions 44 between the gut and commensal microbes may exert a significant influence on the 45 function of the intestine. Previous studies showed that the digestion of uric acid, 46 cellulose, starch and other resistant carbohydrates in the cecum was associated with 47 the cecal microbial members 3, 10, 11 . In a recent study, numerous oligosaccharide-and 48 polysaccharide-degrading enzyme-encoding genes and several pathways involved in 49 the production of short-chain fatty acids (SCFAs) were observed in the cecal 50 metagenome of the chicken 12 . The SCFAs were produced mainly by microbial 51 fermentation in the hindgut and could be absorbed through the mucosa and 52 4 catabolized for energy by the host 13 ; the SCFAs also inhibited acid-sensitive 53 pathogens by lowering the pH 14 . Due to the rapid flow of the highly fluid, digested 54 material and a higher acidity, the number of microbes in the duodenum was lower 55 than that in the posterior intestine. Lactobacilli and Lactobacillaceae were observed 56 to be the predominant microbes in the duodenum of chickens 15 and mice 16 , 57 respectively. However, the relationship between the duodenal microbiota and the host 58 nutritional metabolism is poorly understood. Feces have been widely used for 59 metagenomic studies, because their easy to collection, allowing a continuous 60 observation of the changes during a period without complicated operations or 61 sacrifices, however, the microbial relationships between feces and intestinal segments 62 in layer chickens are still unclear and need to be explored [15][16][17] . 63 For farm animals, great attention is paid to feed efficiency which is a 64 comprehensive trait to evaluate the efficacy of nutrient and energy metabolism. 65 Improving feed efficiency can decrease the cost to producers, preserve additional 66 edible resources for humans, and reduce the excrement effluent and the emission of 67 greenhouse gases. The feed conversion ratio (FCR) and residual feed intake (RFI) are 68 the major indices for assessing the feed efficiency of animals. FCR has been used in 69 breeding for a long time because of its convenience and effect on improving growth. 70 However, in contrast to RFI, FCR does not include variability in the maintenance 71 requirement for feed intake 18 and does not distribute normally 19 . Koch et al. 20 72 proposed the concept of RFI, which accounts for both maintenance requirements and 73 growth. Because of its phenotypic independence from maintaining body weight and 74 5 body weight gain, RFI has been proposed for measuring feed efficiency in breeding, 75 with the heritability of RFI in chickens ranging from 0.2 to 0.8 21-24 . Feed efficiency is 76 a complex trait because it is influenced not only by the host genetics and 77 physiological state but also by the intestinal microbiota, which would affect the Lactobacillus salivarius to E. coli in the better efficiency group. Nevertheless, the 87 feed efficiency in both studies was represented by FCR, and the relationships between 88 RFI and the gut microbiota remain to be understood.

89
Next-generation sequencing techniques have been used to study microbiota 90 composition and extend the understanding of the interactions between the host and 91 commensal microbes in feed efficiency studies. However, the microbial communities 92 have not been compared among the foregut (duodenum), hindgut (cecum) and feces in 93 hens, and the interactions between the feed efficiency evaluated using RFI and gut 94 microbiota need to be explored.

96
Phenotypic and sequencing data. The daily feed intake (FI), daily egg mass (EM), 97 average body weight (BW) and residual feed intake (RFI) at 32-44 (T1) and 57-60 98 (T2) weeks of age are listed in Table 1. The RFI value of the better feed efficiency 99 (BFE) group was found to be significantly lower (P<0.01) than that of the poor feed 100 efficiency (PFE) group. The FIs of the BFE group were 17.0 and 24.3 percent lower 101 than those of the PFE group in T1 and T2, respectively. No significant difference was 102 found in the EM and BW between the two groups.

103
The 16S rRNA gene-based sequencing produced millions of raw reads. After feces were similar at the phylum level, in which Firmicutes and Bacteroidetes were 111 the major microbes. However, the relative abundances of these two phyla were 112 quantitatively different among the three sites ( Table 2). Firmicutes accounted for 113 more than 50% of the duodenal and fecal community, while only approximately 26% 114 was observed in the cecal community. In contrast, greater than 50% Bacteroidetes was 115 observed in the cecal microbial community, while less than 20% was found in both 116 duodenum and feces. In the BFE group, the Firmicutes in the duodenum and 117 7 Verrucomicrobia in the cecum were more abundant (P<0.01), while the Fusobacteria 118 in duodenum was less abundant (P<0.01) than in the PFE group ( Table 2). 119 At the genus level, the top five abundant genera are shown in Figure 1.    Table 3).

152
Abundance differences in the microbiota between the BFE and PFE groups. To 153 investigate the differences in microbial abundance between the contrasting feed 154 efficiency groups, Mann-Whitney tests between the two groups were performed, and 155 the negative logarithms of the false discovery rate (FDR) values are shown in Figure   156 5. It is clear that all significantly different (FDR<0.05) taxa were present in cecum.

157
The significantly different taxa in the cecum (FDR<0.05) and suggestively different 158 taxa in the feces (FDR<0.1) are listed in Supplementary Table. Of these taxa, 12 159 genera were more abundant in the BFE group than in the PFE group, while 7 genera 160 were more abundantly in the PFE group ( Figure 6). Notably, there were significantly 161 9 higher proportions of Lactobacillus and Akkermansia (FDR<0.05) in the BFE group.

162
The comparison also revealed the difference at the species level. The relative 163 abundance of 5 species, Bacteroides coprophilus, Lactobacillus delbrueckii, 164 Veillonella dispar, Lactobacillus reuteri and Prochlorococcus marinus, were 165 significantly higher in the BFE group (FDR<0.05), whereas 3 species, 166 Faecalibacterium prausnitzii, Parabacteroides distasonis and Thermobispora bispora, 167 were found to be significantly higher in the PFE group (FDR<0.05).   (Figure 9a, 9c). Notably, a potential harmful function relating to     Akkermansia spp., a widely studied microorganism that is inversely associated with 249 obesity 46, 47 , was found to be more abundant in the cecum of the BFE group.

250
Akkermansia has been reported to be a mucin degradation-specialized bacterium that 251 utilizes mucus as a sole carbon and nitrogen source 48 . An increase in Akkermansia has 252 been shown to protect the niche from IBDs 49 , obesity 46, 50 , and type I and type II 253 diabetes mellitus 51, 52 .

254
Interestingly, a potentially beneficial microbe, Faecalibacterium prausnitzii, was 255 more abundant in the PFE group. This species has been found to be strongly reduced in  feed efficiency by cecal-oriented and differential microbiota-oriented alterations, and 281 indicated that some segments (e.g., the cecum, which has not been well considered to 282 date), should receive more attention for strategies to improve the health and nutrition 283 of the host. "EMDc" and "BWG" represent the expected feed intake, mean body weight, metabolic 300 body weight, corrected egg mass production(adjusted abnormal egg) and body weight 301 gain, respectively. RFI was then calculated from the actual FI by subtracting the 302 expected FI. The RFI value, which was used to assess the feed efficiency, was 303 negatively correlated with the feed efficiency.                                      between the different feed efficiency groups. "R" is the index of ANOSIM that indicates 535 the similarity of comparison group pairs. "R" ranges from -1 to 1: the pairs are more similar when 536 the R index is closer to 0 and the pairs are different from each other when the R index is close to 1.