In a 28-day trial. A total of 150-day-old unsexed of newly hatched chicks from Jing Hong laying strain were randomly allocated into two treatment groups (n = 75/group) as follows: (1) birds fed on the basal diet (CON), and (2) birds fed on basal diet and received fecal microbiota transplantation (FMT). Each dietary treatment consisted of 5 replicates (15 birds/ replicate). The chicks were housed in growth metal cages in an enclose poultry farm of Huazhong Agricultural University throughout the study period, each cage was equipped with one manual feeder and drinker to ensure the animals’ nutritional requirements. A tray with a parchment paper was placed under each cage for fecal sample collection. The diet was formulated following the guidelines of Chinese Feeding Standards using the Feed database in China (Beijing, 2004), diet ingredients and nutrient composition is provided in Table 1. All birds were fed on the same iso-nitrogenous and iso-caloric diet and water was provided ad-libitum.

Seventy-five newly hatched Cobb-500 broiler chicks were used daily as a stool donor. The FMT was prepared following the methods (Matsuoka et al., 2014). Briefly, 10 g of fresh fecal samples were collected daily in the morning into the sterile tube (50 ml). The white part of the excreta was removed immediately, because it mainly comprises uric acid, and the fecal part was mixed with 0.75% saline in 1:6 ratios (6 ml of 0.75% saline for each gram of feces) and span at low speed (800 × g) for 3 min at 4°C to separate undigested feed and particulate material from the microbial fraction. Keeping the mixture on ice until precipitates were fully settled down, the supernatant was collected and filtered with the sterile gauze to get fecal suspension.

A total of 150 females of the recipients Jing Hong chicks similar age with the donor Cobb-500 (one-day-old) were randomly assigned into two groups (FMT and CON), each group consisted from 75 chicks into 5 replicates (n = 15 bird/ replicate). Each laying chick in FMT group was provided with an oral fresh dosage of 1 ml of FMT solution daily from day 1 to day 28 of age, at early morning and before having access to feed and water. The transplant was orally administered using 1-ml oral dosage via oral-syringes with soft flexible tips that were positioned at the back of the chicks’ tongues. The chicks were then supervised to ensure that they had swallowed the material and feed was with-held for 60 min after administration.

Live body weight was recorded weekly and used to determine the weekly body weight gain and feed intake was recorded weekly, FCR was then calculated. At 28 days of age, 11 healthy chicks of group average BW were selected and euthanized. The chicks’ breasts (pectoralis major) and thigh (drumstick) muscles were skinned, de-boned and weighted; their livers weight, and the lengths (cm) of their duodenum, jejunum, ileum, and ceca were measured similar to Abdelatty et al. (2021).

To obtain samples for microbial genes analysis, fresh samples from the cecum (n = 11/ group) were aseptically collected 5–10 min after slaughter under anaerobic conditions and placed into sterile 2-ml cryo-tubes (Sarstedt, Nümbrecht, Germany), then, samples were immediately stored at −80°C until total microbial DNA extraction. Then, all the sample tubes were instantaneously snap-frozen in liquid nitrogen, and subsequently stored at −80°C for extraction of total microbial DNA later.

The microbial DNA was extracted from 300-μl cecal samples using a DNA stool mini kit (QIAmp DNA Stool mini Kit; QIAGEN, Hilden, Germany) following the manufacturer’s instructions. A Qubit 2.0 Fluorometer (Thermo Fisher Scientific Co., Ltd., Shanghai, China) with a Qubit dsDNA HS assay kit (Life Technologies) was used to quantify DNA concentration.

An aliquot of each of the extracted DNA samples was used as a template for PCR amplification. Additionally, 16S ribosomal RNA, library preparation, and DNA sequencing were performed by a commercial provider (Personalbio Co., Ltd., Shanghai, China). To generate an approximate amplicon size of 570 bp, primers (forward 5′ CCTAYGGGRBGCASCAG GNG 3′, reverse 5′ GGACTACNNGGGTATCTAAT 3′) targeting V3-V4 region amplicons were used for amplification and PCR products were purified (Busato et al., 2022; Tawfik et al., 2022). The PCR conditions were as follows: initial denaturation, annealing, and extension were carried out and repeated at 94°C for 4 min, 94°C for 30 s, 50°C for 45 s, and 72°C for 30 s for 25 cycles.

For each library, amplified libraries of V3-V4 region amplicons were pooled and sequenced using the Illumina Miseq 2000 platform sequencer, including 250-bp paired-end reads generated with a 7-cycle index read (sequences with an overlap longer than 10 bp without any mismatch were assembled). The resultant overlapping paired-end reads were stitched and quality filtered using Microsynth after removing barcode and primer sequences. To perform sequence quality control, the pre-filtered and stitched reads were processed using the QIIME software (Quantitative Insights into Microbial Ecology, v1.8.0).¹ Briefly, the raw tags were merged based on the overlap of two reads. Then, clean tags were created by pretreating the raw tags and removing the chimeric sequences to generate effective tags. Finally, quantitative analysis of the sequences was performed using the QIIME software, and the chimeric sequences were eliminated using the USEARCH software (v5.2.236).²

Quality sequences were counted for each sample after the removal of the chimeric sequences (shorter than 160 bp), and statistical estimations were created for the distribution of sequence length using the R software to characterize the length distribution of the sequences contained in each sample. Then, sequences were clustered into OTUs with 97% similarity using open-reference OTU picking and UCLUST (Edgar, 2010). In addition, Specaccum analysis was applied to check if all sample sizes and the OTU abundance matrix were sufficient to estimate community richness (Supplementary Figure S1). Finally, the taxonomy of OTUs was determined using the Greengenes default database in QIIME (DeSantis et al., 2006). The most abundant OTUs between groups (CON and FMT) were then classified against the NCBI nucleotide database using BLASTN for taxonomic classification, targeting the 16S rRNA marker. For the analysis of bacteria and archaea, the databases for the 16S rRNA gene, Greengenes (Release 13.8)³ and RDP (Ribosomal Database Project, Release 11.1),⁴ were used by default to identify the OTU diversity among the samples and between the groups. A Venn diagram was constructed to calculate the total number of OTUs per sample (i.e., group) using the R software.

A rarefaction depth of 12,000 sequences was used to perform α- and β-diversity analyses among samples; the α-diversity was determined using the unweighted and weighted UniFrac distance (Lozupone et al., 2007). The analysis of α-diversity (Shannon, Simpson, Chao1, and ACE indices) was applied to find the drivers of variation in the microbial community structure in the samples (Shannon, 1948; Simpson, 1949; Chao and Shen, 2004). In addition, rarefaction curves of β-diversity (PCA, PCoA, NMDS, and UPGMA cluster analyses) in the samples were calculated using a maximum rarefaction depth of 12,000 sequences and the observed OTU index (Ramett, 2007). The orders of phylogenetic tree construction (MEGAN and KRONA) were used for microbial interactive visualization and to visually display the results of species annotation between groups based on OTU tags. Then, a heat map was constructed according to the distribution of the top 50 abundances and the degree of similarity between the samples. The order analyses of PLS-DA, LDA, PERMANOVA, and ANOSIM were performed to determine the variation in community structure between groups for screening key species.

To predict the metabolic functioning of bacteria based on total genome sequences from the 16S rRNA gene, Kyoto Encyclopedia of Genes and Genomes (KEGG) function annotation of the sequence was carried out based on Tax4Fun, then visualized using the statistical analysis of metagenomic profiles (STAMP) software package (Langille et al., 2013). KEGG orthologies (KOs) were categorized into KEGG level 2 pathways with the removal of the non-microbial pathways. The classified genes were divided into six categories, namely, Metabolism, Genetic Information Processing, Environmental Information Processing, Cellular Processes, Organismal Systems, and Human Diseases, each of which was further divided into multiple levels. The nearest sequenced taxon index (NSTI) was used to provide a measure of the availability of sequenced reference genomes for each OTU in a sample. It calculates the average branch length for each sample between an OTU and the nearest sequenced reference genome based on the Greengenes reference phylogeny, accounting for OUT abundance. Finally, the R software was used for the cluster analysis of the top 50 most abundant functional predictions in each sample, presented by the heat map (Langille et al., 2013).

The analysis of β-diversity included PCA and PCoA, which were calculated using weighted UniFrac and unweighted UniFrac. One-way ANOVA and the LM procedure of the R software version 3.2.2, R Core were applied to estimate the different bacterial taxa in CON and FMT groups (R Development Core Team, 2015). Each chick individual was considered an experimental unit, and FMT was included as a fixed effect in the statistical model. All differences were considered significant at p < 0.05 and were considered trends when p < 0.10. Pairwise comparison was performed using T tests.

The differences in body weight, daily weight gain, internal organs weight, and length of the digestive tract sections between CON and FMT groups of laying chickens are shown in Table 2. FMT increased final body weight (p < 0.01) by 2.5% and average daily gain increased (p < 0.01) by 2.3%. Feed conversion ratio was lower in FMT group compared with the CON group (p = 0.01), breast muscle weight increased by 9.7% and thigh muscle weight increased by 17.63% in FMT chickens compared with the control counterpart (p = 0.01). Liver weight increased by 8.5% (p = 0.01). On the other side, duodenum, ilium, and cecum length was shorter in MFT group compared with the CON one (p < 0.05).

High-throughput sequencing generated from 22 chickens (n = 11/ dietary treatment) yielded a total of 775,945 reads (average of 35,270 and range of 32,836–38,767 reads per sample; Supplementary Figure S2). Average read length was 160 bp, and the distributions of sequence lengths shown in OTUs were generated based on the greengene database using quantitative insights into microbial ecology (QIIME) and characterized for different taxonomic levels including domain, phylum, class, order, family, and genus. The taxa considered as common in the samples were used in further analysis. Statistical number of OTUs at each classification level among groups of control and antibiotics-exposed chicks are presented in Supplementary Table S1. A total of 12 phyla, 23 classes, 35 orders, 60 families, 77 genera, and 31 species were identified in these samples Supplementary Table S2.

Total observed OTU count, α indexes (ACE and Chao1), and diversity indicators (Simpson and Shannon) between groups are summarized in Table 3. The total observed OTUs was significantly increased (p < 0.01) in the FMT group than in the CON group. In contrast, the means of ACE and Chao1 were significantly decreased (p < 0.01) in the FMT group than in the control group. Similarly, the Simpson and Shannon indices were significantly decreased in the FMT group than in the CON group. In addition, β-diversity indicators, principal component analysis (PCA), non-metric multidimensional scaling (NMDS), and principal coordinate analysis (PCoA) were obtained to measure the between-group and inter-group distance of the CON and FMT groups.

The PCA based on unweighted UniFrac distance is presented in Figure 1. The PCA plot revealed that the distance between samples of the CON group were smaller than that of the FMT group, indicating that FMT had a more diverse gut microbiota. The R value of the analysis of similarities (ANOSIM; 0.38) was correlated among samples within the CON and FMT groups (p < 0.01). Similarly, the weighted and unweighted values of NMDS indicated that samples of the CON group formed a tighter cluster than those of the FMT group. Regarding the R value obtained from the mutational multivariate analysis of variance (PERMANOVA), there were significant correlations in weighted (R = 0.25; p < 0.05) and unweighted (R = 0.19; p < 0.01) UniFrac distances among samples and between groups of the CON and FMT groups. In addition, the lowest group distance was significantly estimated (p < 0.05) within groups in the comparative analysis boxplot.

Figure 2 illustrates the effect of FMT on gut microbiota structure of Jing Hong chickens. The number of OTUs was significantly increased (p < 0.01) in the FMT group, with a high abundance of unique OTUs when compared with the CON group. Additionally, a heat map of the top 50 most abundant genera in the microbiome community combined with a cluster analysis revealed similar microbiome composition among samples of the CON and FMT groups. The differences between the CON and FMT groups regarding the relative abundance (percent of total sequences) of each phylum and family of taxa from the microbiome are presented in Figure 2 and Table 4.

At the phylum level, the relative abundances of Firmicutes increased and that of Bacteroidetes decreased (p < 0.01) in the FMT group, and the Firmicutes/Bacteroidetes ratio (F/B ratio) increased in the FMT group than in the CON group (p < 0.01; Table 4). The differences between the CON and FMT groups regarding the relative abundance (percent of total sequences) of each phylum and family of taxa from the microbiome are presented in Figure 2 and Table 4. At the phylum level, the relative abundances of Firmicutes was increased (FMT vs. CON, 89.05 vs. 75.39%) and that of Bacteroidetes was decreased (FMT vs. CON, 2.51 vs. 5.80%) in the FMT group (p < 0.05). Consequently the Firmicutes/Bacteroidetes ratio (F/B ratio) increased (FMT vs. CON, 35.45 vs. 12.99%) in the FMT group than in the CON group (p < 0.01; Table 4). On the other hand, the relative abundance of Proteobacteria (FMT vs. CON, 3.52 vs. 7.98%) and Campilobacterota (FMT vs. CON, 0.164 vs. 0.811%) were decreased in the FMT group chickens (p < 0.01; Table 4). Similarly, there was decreased (p < 0.05) in the relative abundance of phyla of Verrucomicrobia (FMT vs. CON, 0.740 vs. 1.36%) and TM7 (FMT vs. CON, 0.005 vs. 0.042%) in the FMT group than in the CON group as presented in Figure 2 and Table 4.

At the genus level, there were significant increased the relative abundances of some symbiotic taxa such as Lactobacillus (FMT vs. CON, 17.87 vs. 9.12%), Lactococcus (FMT vs. CON, 2.51 vs. 5.80%) and Bifidobacterium (FMT vs. CON, 0.064 vs. 0.051%) in the FMT group compared with the CON group (p < 0.01; Figure 2; Table 4). On the other hand, there were significant decreased the relative abundances of some pathogenic taxa such as Enterococcus (FMT vs. CON, 0.55 vs. 0.86%), Helicobacter (FMT vs. CON, 1.027 vs. 1.589%) and Bacteroides (FMT vs. CON, 0.054 vs. 1.30%) in the FMT group in compared with the CON group (p < 0.01; Figure 2; Table 4).

The results of the LEfSe analysis based on the linear discriminant analysis (LDA) of the specific bacterial taxa that are associated with FMT are presented in Supplementary Figure S3. The LEfSe cladogram representative of the structure of the host–microbiota axis shows a significant shift of microbiota between groups, including a total of 64 bacterial taxa that were significantly different between the CON and FMT groups, with only seven taxa (Lactobacillallus, Lactobacillaceae, Lactobacillales, Bacilli, Dietzia, Dietziaceae, and Aeriscardovia) in the FMT group. The LDA score plot presented group-enriched taxa that were significant between the two groups (p < 0.05).

Figure 3 illustrates the comparison of abundance among groups at the levels of phylum and genus were performed using a Metastats analysis and indicated an increase of the phyla Firmicutes and Acidobacteria and of the genera Dietzia and Ruminococcus in the FMT group. The similarity between samples was evaluated by unweighted pair group method with arithmetic mean (UPGMA) analysis, which showed that the samples of the CON and FMT groups were mostly similar and distributed into a different cluster of hierarchical trees (Supplementary Figure S4).

The relative abundance comparison between the CON and FMT groups based on the second-level of KEGG pathways analysis is shown in Table 5 and Figure 4. A total of six categories of cellular processes, environmental information processing, genetic information processing, immune information processing, metabolism processing, and organismal systems processing were revealed between the CON and FMT groups.

In the category of cellular processes, the levels of cell growth and death, and transport and catabolism had a higher (p < 0.01) relative abundance in the FMT group than in the CON group. The level of membrane transport pathway in the category of environmental information processing was significantly increased (p < 0.01) in the FMT group than in the CON group (Table 5). Prediction of functional microbiome in the category of genetic information processing showed that three pathway levels of DNA replication and repair, transcription, and translation were significantly increased (p < 0.01) in FMT group. One level of infectious diseases pathway under the category of immune information processing was increased (p < 0.05) in the FMT group when compared with the CON group. Regarding the category of metabolism processing, the differences between groups (FMT and CON) in the pathway levels regarding amino acid metabolism (p < 0.01), secondary metabolites biosynthesis (p < 0.01), carbohydrate metabolism (p < 0.01), energy metabolism (p < 0.05), enzyme families (p < 0.05), and cofactors and vitamins (p < 0.05) were higher in the FMT groups than in the CON group. The pathway level of endocrine system was significantly increased (p < 0.01) in the FMT group when compared with the CON group.

The top 50 most abundant KEGG orthologous genes were clustered in the heat map combined with the analysis between the CON and FMT groups; samples of the control group formed clusters based on their similar microbial compositions. The Venn diagram of the common and unique functional groups revealed that the FMT group had more functional groups than the CON group. PLS discriminant analysis (PLS-DA) plot showed that samples belonging to the same group were more similar to each other than those from different groups.