A total of 400 one-day-old male lion-head geese with similar birth weight (82.6 g ± 1.4 g) was randomly divided into five treatments: the control treatment (CON) and the HF supplementation treatments, HF was supplementally arithmetically increased, which includes 0.25, 0.50, 0.75, and 1.0%. Each treatment contains four replications, with 20 birds in each replicate. All birds received the floor-rearing system, each replicate was allotted an 8-m long and 4-m wide compartment. Each compartment was 1 m high from the ground. All birds were provided with a two-phase feeding procedure (Day 0–21 as the starting phase, while Day 22–70 was the finishing phase), feed and water were provided ad libitum throughout the experiment. The room temperature was maintained at 37°C for the first week and then reduced by 3°C each week until reaching 24°C. The lighting schedule was 23 h light and 1 h dark during all experiment periods.

Raw honeycomb used in this study was provided by the Institute of Bee Research, Jiangxi Agricultural University, Jiangxi, China. Honeycomb was frozen at −18°C for 24 h, and then quickly pulverized and sieved through a 20-mesh screen for further investigation. HF in the present study was extracted in the laboratory using a 30-min long ultrasonic extracting procedure combined with a centrifugation at 3,000 rpm for 10 min. The liquid level of ultrasonic cleaner (Elmasonic X-tra Flexl, Elma, Konstanz, Germany) was kept between 3 and 5 cm. The post-extracting liquor was filtered after cooling and collected into an Erlenmeyer flask. The flavonoids content was measured for about 45% active ingredient.

For the experimental diets in each phase, a master-batch of the basal diet (negative control) was prepared in mash, and the additives were added afterward. The composition of the experimental diets and the nutrients are shown in Table 1.

On Day 70, all parts throughout the gastrointestinal tract were sampled which includes muscular and glandular stomachs, duodenum, ileum, jejunum, cecum, and colorectum. The duodenum is connected to the muscular stomach, following with the jejunum and the ileum, which are in the second segment of the whole intestine. The cecum is located at the junction of the small intestine and the large intestine, which possessed two branched parallel bowel segments. Whereafter, the weight and length of each part of the intestine were measured to investigate the gastrointestinal development. Furthermore, tight junctions related genes ZO1 and ZO2 of intestinal epithelial cells which play important roles in maintaining the intestine barrier function also determined the expression quantity in both ileum and jejunum through reverse transcription-PCR (RT-PCR).

Cecal samples were collected from all slaughtered birds. All samples were rapidly frozen with liquid nitrogen and then stored at −80°C for further analysis. Based on the above-mentioned parameters, optimum supplement treatment was selected for the further microbiome analysis compared with the CON treatment. DNA from each sample was extracted using CTAB/SDS method followed by the measurement of DNA concentration and purity. The V4 region of the 16S rRNA gene was amplified using the universal primers 520F and 802R (F: GTGCCAGCMGCCGCGGTAA and R: GGACTACHVGGGTWTCTAAT). All PCR reactions were carried out with Phusion High-Fidelity PCR Master Mix (New England Biolabs). Samples with a bright main strip between 400 and 450 bp were chosen for further analysis. The mixture of PCR products was purified with Qiagen Gel Extraction Kit (Qiagen, Hilden, Germany) and, subsequently, sequencing libraries were generated using TruSeq^® DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, CA). The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. At last, the library was sequenced on Illumina HiSeq 4000 platform (Illumina Inc., San Diego, CA).

Quality filtering of raw tags was performed under specific filtering conditions to obtain the high-quality clean tags according to the Quantitative Insights Into Microbial Ecology (QIIME, V1.7.0) quality controlling process. Sequences within similarity >97% were assigned into the same operational taxonomic unit (OTU). For each representative sequence, the GreenGene Database was used based on the SILVA classifier algorithm to annotate taxonomic information.

Differential analysis of growth performances, carcass performances, and gastrointestinal development parameters was verified through a normal distribution test using SAS procedure “proc univariate data = test normal.” Subsequently, one-way ANOVA S-N-K test was applied to investigate the differences among the 5 treatments. Results were presented as mean ± SEM. OTU abundances of cecal bacteria first conducted a transformation of normal distribution using log₂, and then Student's T-test of SAS 9.2 was applied for the differential analysis. Alpha diversity and beta diversity in our samples were calculated with QIIME (Version 1.7.0) and displayed with R software (Version 3.15.3). PCoA analysis was displayed by WGCNA package, stat packages, and ggplot2 package in R software (Version 3.15.3). The OTU abundance of ruminal bacteria was first transformed into normal distribution using the log₂ transformation, and then the Student's T-test of SAS 9.2 was applied to analyze the differences of bacteria. P < 0.05 was significant. Spearman correlations between bacteria communities and production performances, carcass performances, and intestinal development parameters were assessed using the PROC CORR procedure of SAS 9.2 and then the correlation matrix was created and visualized in a heatmap format using R software (Version 3.15.3).

Mortality and culling rate of each treatment was recorded daily, all treatments showed 1–2 deaths of birds throughout the trials, which indicated no significant changes were found among all treatments. Differential analysis of gradient HF supplement on growth performances was first evaluated including FI, BWG, and FCR in the starting phase, the finishing phase, and the whole phase. Results are shown in Table 2. No significant enhancive effects were acquired after HF supplement on both FI and BWG for either the starting or finishing phases. It is noteworthy that FCR showed a gradually decreasing trend with the increase of HF supplement in the finishing and the whole phase.

Subsequently, carcass characteristics were measured, and the results are presented in Table 3. Totally, HF supplement significantly promoted the dressed weight and semi-eviscerated weight while it tended to promote the eviscerated weight. Besides, supplemented with 0.5 and 0.75% of HF significantly increased the tight weight compared with other treatments. No significant changes were found for other parameters after HF supplement.

The 16S rRNA gene amplicon sequences from cecal contents of both CON and HF treatment samples were conducted to investigate the effects of HF supplementation treatments on gastrointestinal microbiota. Taxonomy results of all bacteria are shown in Supplementary Table 1. A total of 10 phyla and more than 200 genera were identified in the present study, the average length of sequence reads was about 410 nt. All the results were subsequently analyzed for α-diversity and β-diversity.

Interactive analysis between the most abundant bacteria and production performance, carcass performance, and intestinal development parameters were conducted and the result is shown in Figure 3.

Integrally, phenotype parameters could be separated into two big clusters. The first, which mainly consists of ileum, cecum, FI, and breast, was positively correlated with Butyricicoccus, Prevotella, Clostridiales, Streptococcus, Lachnoclostridium, and Bifidobacterium, while negatively correlated with Faecalibacterium, Ruminococcaceae, Escherichia Shigella, and Sellimonas. The other cluster showed the reverse correlations with bacterial genera compared with the first one which included jejunum, abdominal fat, and FCR. Especially, Butyricicoccus and Eisenbergiella remarkably negatively correlated with FCR, which may provide the target for reducing FCR and promoting feed efficiency.

The intestinal functions typically regulated by the epithelial absorptivity and gut microbiota digestibility. The nutrients' absorption mainly occurs at the upper parts of the gastrointestinal tract, which is regulated by high levels of acids, abundant digestive enzymes, and antimicrobials (21). Besides, the intestinal epithelial cells (IECs) exist as a layer of columned cells that generated a functional barrier to protect the intestinal mucosa from pathogenic microorganisms, and an interchange channel to absorb nutrients into circulation (22). The epithelium also contains plentiful functional proteins, which conduct the substances interchange or achieve informative communication with other metabolites as the target spot (23).

Flavonoids in the intestines were considered as efficient antimicrobial metabolites, which showed a broad-spectrum antimicrobial capacity and inhibited most pathogenic micro-organisms and thus improved epithelium development. Moreover, flavonoids in the intestinal tract interacted with the functional proteins including inhibition of inflammatory signaling such as nuclear factor-kappa B (NK-kB) and up-regulation of tight-junction proteins such as ZO1 to fortify the intestinal tight junction barrier and structure (18, 24). In the present study, the expression of tight-junction proteins ZO1 and ZO2 showed a significant increase with HF supplementation. This might partly enhance intestinal barrier functions and further intestinal absorptivity.

Apart from the enzymatic digestion in the upper parts of the gastrointestinal tract, colonic and cecal conditions support a diverse community of bacteria that are capable of fermenting complex substances undigested in the small intestine (21, 25). In our research, the microbiota community is dominated by bacteria, with more than 90% of the species belonging to Firmicutes and Bacteroidetes, which is in line with Tremaroli and Backhed (26). What is noteworthy is that the ratio of Firmicutes to Bacteroidetes was positively correlated with the energy metabolism, the higher Firmicutes boosts the concentrate's digestibility, and thus more energy was provided. In our research, Firmicutes significantly increased after HF supplement. This may partly explain the increased body weight after HF supplement.

Particularly, Lactobacillus and Bifidobacterium, which played important roles in maintaining intestinal homeostasis and fortifying the intestinal mucus layer (27, 28) in the present study were significantly proliferated in HF treatment. Higher abundance of cecal Lactobacillus positive regulated a better feed efficiency (2), while increased Bifidobacterium correlated with reduced adiposity and levels of microbe-derived inflammatory molecules (29). These changes protected intestinal structure, enhanced intestinal barrier functions, and further promoted the intestinal absorptivity.

To date, host and bacteria interactions turned into a great causal factor that shaped the gastrointestinal tract, affected host fitness, growth rates, and carrying capacity through providing more energy and scavenging inflammatory molecules (7, 30). Thus, the host-associated bacteria interacting with the epithelial barrier is thought to be the main underlying mechanism that impacts the geese growth performance. Just as we here focused on the nutrient's metabolism, the uppermost energy supplier such as carbohydrates, and the main nutritional anti-inflammatory metabolite—bile acid—are considered as the main substances mediating the connection between host and bacteria to the HF supplement (31, 32).

The carbohydrates, including cellulose, xylans, and starch, yield the most energy by colon and cecum microbiota. The further fermented products such as short-chain fatty acids (SCFAs) have profound effects on gut health and energy absorption and utilization (26). Of the SCFAs produced from microbial fermentation, butyrate is particularly important as an energy substrate for epithelial development and cellular metabolism in the colonic and cecal epithelium (33). In the present research, the main butyrate generating bacteria such as Clostridium and Butyricicoccus remarkably increased after HF supplement (34, 35). This might indicate that HF stimulated the proliferation of butyrate generating bacteria and further provided more energy for gut epithelial development.

Bile acid (BA) metabolism might be another causative factor that regulated intestinal absorptivity. In recent research, the BA signaling pathway was proved to be critical in both lipid and carbohydrate metabolism, also in maintaining esoteric glucose and cholesterol homeostasis as well as immune states (36, 37). In addition, bile acids can regulate gut microbial composition both directly and indirectly by activation of innate immune response genes in the small intestine (37). BA is also a crucial secondary metabolite that is utilized by gut microbiota and communicated with epithelial cells. Research further confirmed that over 80% of microbiota-host interactions were with secondary Bas (8). Intriguingly, flavonoids could be responsible for increased BA production (38). These findings provided us a strong promotive effect of HF on gastrointestinal development and growth performances.