Effect of Dietary Yeast Culture Supplementation on the Growth Performance and Cecal Microbiota Modulation of Geese

YC increases the relative abundance of Firmicutes that promote energy utilization and nutrition absorption, thereby improving the growth performance of geese. This dietary strategy based on feed additives is an effective method to maintain the health of the geese GIT and to improve growth eciency.


Introduction
Antibiotics combat disease in animals and improve production e ciency and quality, resulting ultimately in economic pro ts. Therefore, antibiotics have been widely used as feed additives for animal husbandry.
However, their overuse has caused many problems, such as presence of antibiotic residue, drug resistance and others [1] . Therefore, it is imperative to explore feed additives with similar effects to antibiotics, but which are safer and more effective. Yeast cultures (Ycs) are micro-ecological products formed by yeast after su cient anaerobic fermentation on speci c media and are mainly composed of yeast metabolites, fermented medium and a few yeast cells. YCs are rich in vitamins, saccharides, minerals, enzymes, growth-promoting factors and amino acids, which may bene t animal growth, metabolism, and health. YC is one of the most important feed additives [2] as it can improve yield and quality of animal by-products, such as milk [3] and eggs [4] . YC can also enhance animal growth performance, as measured by average daily gain, body weight and average daily intake [5][6][7] , and also can strengthen immunity and disease resistance [8][9][10] . The constituents of YC are live yeast and various yeast metabolites. These may act on the gastrointestinal tract (GIT) directly to maintain the balance of the microorganism and promote development of the GIT. For example, YC can signi cantly affect community structure and composition of gut microbiota of grass carp, increasing the abundance of bene cial bacteria [11] . Addition of YC in the diet resulted in an increase in bene cial bacteria, making greater compensatory weight gains in weanling pigs [12] . The effect of YC on hindgut microbial communities was evaluated in horse [13] where it increased the abundance of ber-degrading bacteria.
Geese are important poultry in China (90% of the world's production), and their cecum are more developed compared with other poultry, so it is the ideal model to study the interaction between GIT microbiota and diet [14] . Supplementation of geese diets with yeast fermanted feed regulate cecal micro ora that is bene cial of growth performance [15] . Addition of yeast supplement in the diets improve intestinal micro ora of goose and lead to better carcass hygiene [16] . However, few studies have investigated the effect of YC on the GIT microbiota of geese. Furthermore, the variety in GIT microbiota composition and structure can re ect the effect of dietary on the host. Thus, the present study was conducted to investigate the effect of YC supplementation on the GIT microbiota of geese. Methods Animals, diets, feeding and experimental design Three hundred Sichuan white geese with healthy and similar body weight (BW: 95.57 ± 2.42 g) were randomly divided into ve groups: i) basal diet (control), ii) basal diet + 0.5%YC (treat1), iii) basal diet + 1.0% YC (treat2), iv) basal diet + 2.0% YC (treat3), and v) basal diet + 4.0% YC (treat4). The commercial YC (Baihuibang, Beijing Enhalor Biotechnology Co., Ltd., Beijing, China) was a fermented product composed of Saccharomyces cerevisiae grown on a medium, which contains ~ 15.0% CP, ~ 3.5% crude fat, ~ 8.7% crude ber, ~ 14.2% amino acid, ~ 3.3% mannan, ~ 14.0% β-glucan and other micro-components. Each group was fed four times a day (at 7:30, 12:30, 17:00 and 21:00) for 10 weeks. Two corn-soybean based basal diets were formulated to be fed during starter (0-4 weeks) and grower (5-10 weeks; Table 1) periods. Each group consisted of 3 replicates pens with 20 geese per pen. Geese were allowed access to feed (granule form) and water ad libitum throughout the experimental period. The daily temperature inside the house throughout the experimental period was (21.32 ± 2.39) ºC, and the relative humidity was (84.03 ± 5.15)%.

Growth performance measurements
The BW was measured on days 1 and 70 with empty stomach, and feed consumption was recorded every morning before the feeding. Average daily gain (ADG), average daily feed intake (ADFI) and feed conversion ratio (FCR) were calculated for the total period.

Intestinal content collection
Three 70 day old geese were randomly selected from each pen and killed. Nine cecum contents samples from each group were extracted and pooled by 3 (n = 3). The cecum was dissected with sterile instruments and washed in 70% ethanol and sterile water to avoid transient bacteria. The content from cecum was squeezed out and mix thoroughly, then collected into sterilized tubes and immediately stored in liquid nitrogen.
16S rRNA PCR ampli cation and sequencing Bacterial genomic DNA was extracted from 100 mg of each sample using the QIAamp DNA Stool Mini Kit (Qiagen, Germany) following the manufacturer's protocol. The bacterial V3-V4 region of the 16S rRNA was ampli ed using the following primers: 319F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTC TAAT-3′), and a sequencing adapter was added to the end of the primers. Then the PCR products were puri ed, quanti ed, and homogenized to establish a sequencing library. The constructed library was subjected to library quality inspection, and the quali ed library was sequenced as described in Fadrosh, Ma, Gajer, Sengamalay, Ott, Brotman and Ravel [17] using Illumina HiSeq 2500 (Illumina, San Diego, CA, USA). All control and YC-treated samples were included in the same sequencing run.

Sequences analysis
According to the overlap relationship between paired-end (PE) reads, the double-end sequences obtained from Hi-seq sequencing were merged into a single sequence tag, to control the quality of reads and lter the effect of merge. This included the following three main steps: 1) PE reads merging: used FLASH v1.2.7 to merge the reads of each sample through the overlap, obtaining the original tags data (raw tags); 2) tags ltering: used Trimmomatic v0.33 to lter the raw tags obtained by merging to obtain high quality tags data (clean tags); 3) chimera removal: used UCHIME v4.2 to identify and remove the chimera sequence to obtain the nal effective data (effective tags).

Bioinformatics analysis
Usearch [18] was used to cluster tags at a similarity level of 97% to identify operational taxonomic units (OTUs). OTUs were annotated based on the Silva (bacterial) and UNITE (fungi) taxonomy databases.
Taxonomic assignment was achieved using RDP classi er, clustering the sequences at 97% similarity with a con dence threshold of 0.8.
The α-and β-diversity of each sample was evaluated based on the OTU data (Supplementary Material S1). Mothur (version v.1.30) and QIIME were used to calculate α-and β-diversity, respectively. PICRUSt [19] was used to detect the composition of functional genetic of each group by comparing species information and analyzing differences between groups in KEGG Orthologs (KO) composition including function. This involved metabolism, genetic information processing, human diseases, environmental information processing, cellular processes and organismal systems. Bugbase was used to detect the potential pathogenic bacteria abundance at the phylum and genus level.

Statistical analysis
The data analyses were performed with SPSS 22.0 software (IBM Corporation, 2014) using one-way ANOVA with LSD multiple comparison tests. All data are shown as means ± SEM unless otherwise indicated. Differences were considered statistically signi cant when P < 0.05 throughout.  (Fig. 1b) suggests that bacterial abundance in the control was different from the YC groups. A total of 462 common OTUs were shared among the ve groups, and the most abundant taxa of the shared OTUs (Fig. 1c) indicate that Firmicutes, Bacteroidetes, Proteobacteria, Acidobacteria, Verrucomicrobia and Cyanobacteria form the six most dominant phyla. Seven unique OTUs were found in the YC-group compared with the control (Fig. 1d). Among them, four OTUs at the genus level, Butyrivibrio, Ruminococcus_1, DTU014, and Bacteroides were found in the YC-group, Caproiciproducens and Clostridiales_vadinBB60_group were found in the YC-group except treat1, and [Eubacterium]_ coprostanoligenes_group was found in the treat3 and trea4.

Variation in microbiota diversity
The addition of YC in the diet produced no signi cant differences (P > 0.05) in α-diversity, as measured using Shannon, Simpson, Ace and Chao1 indices ( Table 2). Using the Simpson's diversity index, the sequence is the same but in opposite order. These results indicate that the diet supplemented with 0.5% and 4% of YC caused the highest and the lowest species diversity, respectively. The Chao1 and Ace indices indicating that the diets supplemented with 2% and 4% YC had the highest and lowest microbial richness, respectively. The β-diversity, based on the overall community composition, indicated that the similarity between clustered GIT microbiota in the ve groups was due to the YC supplement (Fig. 2). This gure shows that group treat4 clearly separated from the other groups, using both weighted and unweighted PCoA analysis, whereas treat1 and treat3 formed a cluster. In addition, the replicates in treat 2 show a large variation. These results are signi cant since the sum of PC1 and PC2 was more than 50%, therefore these results likely represent the differences in GIT microbial community structure among these ve groups.  (Fig. 3b).

Predicted functions of microbiota
Based on functionality prediction, we investigated the KO composition in the ve groups. We found no signi cant differences between control, treat2, treat3 and treat4. However, neurodegenerative diseases, substance dependence, cancers: speci c types, carbohydrate metabolism, amino acid metabolism, global and overview maps, aging and sensory system achieve a statistical signi cant difference (P < 0.05) between control and treat1 (Fig. 5a). To determine whether the YC supplementation has a positive or negative effect on the GIT, and what levels are optimal, we predicted what bacteria were pathogenic and bene cial in the ve groups (Fig. 5b). At the genus level, the proportion of potential bene cial bacteria in treat2 was higher than in other groups, whereas the proportion of potential pathogenic bacteria in treat4 was higher than in other groups. Among these bacteria, the dominant potential pathogenic genera were Rikenellaceae, Bacteroides and Ruminococcus, whereas the prominent potential bene cial genera were Parabacteroides, Enterococcus, Streptococcus and Pseudomonas.

Growth performance
There was no signi cant difference in the initial BW of geese with different YC dietary groups. Dietary YC supplementation signi cantly affected the nal BW, ADG, ADFI, and F/G (Table 3). Compared with the control, the YC dietary signi cantly increased the nal BW and ADG, while signi cantly decreased the F/G, with treat3 being the most effective. Moreover, ADFI in treat3 was signi cantly reduced, and no signi cant differences were found among other groups.

Discussion
The GIT microbiota is a community of microorganisms inhabiting the animal gastrointestinal tract, which can be bene cial or harmful. In general, there is a balance between bene cial and harmful microbiota, and this has relevance for the immune system, metabolism [20] and diseases [21] . The composition of the GIT microbiota evolves throughout an individual's lifetime, and the balance may be altered by both endogenous and exogenous factors [22] , such as the diet [23] . Thus, the GIT microbial composition depends on the interaction between the microorganisms and their host, and with the diet of the host. Therefore, studying the GIT microbiota composition and its structure can be used to better understand how dietary YC affects growth performance of goose at the microbial level. In the present study, we chose ve commonly used YC diet supplementation levels (0, 0.5, 1.0, 2.0, and 4.0%) to study how YC modulates GIT microbiota. We used high-throughput sequencing of the V3 + V4 region of the 16S rRNA gene to explore the diversity and composition of GIT microbiota in geese.
The α-diversity analysis showed that there is no signi cant differences in species diversity or richness between these treatments, whereas β-diversity analysis based on PCoA showed a clear separate microbiota structure among the groups. Further, the α-diversity analysis suggested that some microorganisms are essential for the growth of the host. In addition, the presence or absence of microorganisms is not easily affected by the rearing environment or changes in diet, although the relative abundance may be affected. This is consistent with a report indicating that there are no signi cant differences in α-diversity when gilthead sea bream is supplemented with different mannanoligosaccharide levels [24] . However, laying hens with the higher fermentation yeast addition into diets had the lower total bacterial count [25] . The discrepancies between these and the present study may be caused by differences in yeast product used, host species and rearing conditions.
In the present study, there were clear differences in microbial composition between the different YC treatments. Firmicutes and Bacteroidetes were the two most dominant phyla in the bacterial communities of the goose. Similar results have been observed in chicken [26] , buffalo [27] and goose [28] , where the predominant genera were Alistipes and Desulfovibrionaceae. Phylum Firmicutes and Bacteroidetes are bene cial to the host. Firmicutes are tightly related to nutrition and to energy absorption [29] . In particular, the genera of this phylum degrade ber and cannot stand extremely acidic environments. In contrast, phylum Bacteroidetes plays a role in maintaining a healthy GIT, adapting to changing and extreme environments [30] . Furthermore, we found that as the amount of YC increased, the relative abundance of the phyla Firmicutes and Bacteroidetes decreased and increased, respectively, and a similar trend can be found in their respective genera (Fig. 3). These results are consistent with similar studies performed on gut microbiota in humans [31] . These results may be explained by considering that YC diet supplementation can enhance microbiota fermentation, resulting in an increase in the concentration of volatile fatty acids and a reduced pH of the GIT [32] . In contrast to Firmicutes, Bacteroidetes can adapt to this acidic environment.
The dominant bacteria in the GIT may play many important functions essential to the goose [33] . Our study indicates that 0.5% YC can signi cantly enhance carbohydrate and amino acid metabolism and global and overview maps (Fig. 5a). These biochemical processes might promote utilization of energy, absorption of nutrients and growth [34] . Remarkably, the proportion of potential bene cial bacteria after 1.0 and 2.0% YC treatment was higher than in the control (Fig. 5b). These results are caused by the presence of oligosaccharides in YC, which can minimize GIT pathogen concentration and stimulate the growth of bene cial bacteria [35] , thus alleviating the competition for diet between the host and the pathogenic bacteria. However, the proportion of potential bene cial bacteria in the sample treated with the highest amount of YC (treat4) was the lowest (Fig. 5b). High concentration of YC may induce the goose to develop immune tolerance, leading to wastage of energy and nutrients and suppressing the growth of both host and microbiota [36] . A similar result was obtained for growth performance, where low YC concentration (0.5%, 1% and 2%)improved average daily gain and feed conversion rate in goose, whereas high YC concentration (4%) did not.
In conclusion, dietary YC modulates the composition and relative abundance of GIT microbiota in goose and lead to better growth performance. In general, dietary YC is bene cial to GIT microbiota and its effect depends on concentration. Speci cally, 0.5% YC increased species diversity, 1% YC increased the proportion of bene cial bacteria and 2% YC improved species richness. Competing interests