Microbiota and metabolome responses in the cecum and serum of broiler chickens fed with plant essential oils or virginiamycin

This study investigated the cecal microbiota and serum metabolite profile of chickens fed with plant essential oils (PEO) or virginiamycin (VIRG) using high-throughput 16S rRNA gene sequencing and untargeted metabolomics approach. The main aim of this work was to explore the biochemical mechanisms involved in the improved growth performance of antibiotics and their alternatives in animal production. The results showed that both PEO and VIRG treatment significantly increased the relative abundance of phyla Bacteroidetes and decreased the abundance of phyla Firmicutes and genus of Lactobacillus in cecal microbiota of chickens. Compared to the control group (CT group), the relative abundance of genus of Alistipes, unclassified Rikenellaceae, Roseburia, and Anaeroplasma was enriched in the PEO group; that of genus Bacteroides, Lachnospiraceae, and unclassified Enterobacteriaceae was enriched in the cecal microbiota of the VIRG group. Untargeted metabolomics analyses revealed that the PEO treatment modified 102 metabolites and 3 KEGG pathways (primary bile acid biosynthesis and phenylalanine metabolism) in the cecal microbiota, and 81 metabolites and relevant KEGG pathways (fructose and mannose metabolism, biosynthesis of unsaturated fatty acids, and linoleic acid.) in the serum of the chicken. Compared to the CT group, VIRG treatment group differed 217 metabolites and 10 KEGG pathways in cecal contents and 142 metabolites and 7 KEGG pathways in serum of chickens. Pearson’s correlation analysis showed that phyla Bacteroidetes and genus of Bacteroides, Alistipes, and unclassified Rikenellaceae (in the VIRG and PE group) were positively correlated with many lipid metabolites. However, phyla Firmicutes and genera Lactobacillus (higher in the CT group) were negatively correlated with the lipid and thymine metabolism, and positively correlated with hydroxyisocaproic acid, cytosine, and taurine. This study shows that dietary supplementation with PEO and VIRG altered the composition and metabolism profile of the cecal microbiota, modified the serum metabolism profile.


Effects of dietary plan essential oils or virginiamycin on caecum and serum metabolites.
A non-targeted LC-MS-based metabolomics platform was used to analyze the cecum contents and the serum metabolite profiles of chicken fed supplemented with PEO or VIRG. According to the variable importance in the projection (VIP) value > 1, in 95% jack-knifed confidence intervals and P < 0.05, detailed information about the different biomarker metabolites has been shown in Supplementary Table 2. Compared to the CT group, 102 different metabolites with 53 LC-MS/MS(ESI+) and 49 LC-MS/MS(ESI-), were detected in cecum contents of the PEO group; 81 different metabolites  were detected in the serum of the PEO group. VIRG group have 217 different metabolites (96 ESI+ and 121 ESI-) in cecum contents, and 142 different metabolites (82 ESI+ and 60 ESI-) in the serum as compared to the CT group. There were 305 different metabolites (148 ESI+ and 157 ESI-) in cecum contents, and 119 different metabolites  in the serum between PEO and VIRG groups.
Compared to the CT group, dietary supplement with virginiamycin increased the levels of (S)-Equol, LysoPE (16:0/0:0), Caproic acid, D-Maltose, and ribitol in the cecum of the chicken. However, 9 other metabolites including nucleotides (Cytosine, Thymine, Ademine, and Uracil), Myristoleic acid, and Pyridoxal (Vitamin B6) decreased (Fig. 3B). About 10 lipid metabolites such as oleic acid, 9,10-DiHOME, 1-Stearoyl-2-oleoyl-sn-glycerol-3-phosphocholine (SOPC), Adynerin, betaine, DL-lactate, and dopamine were higher in the  Changes of microbial composition in the cecum of broiler chickens fed without or with PEO or VIRG. Microbial composition at the phylum level (A) and genus level (B) each bar represents the relative abundance of each bacterial taxa of chicken. Bacterial taxa significantly differentiated between CT, PEO group, and VIRG group (C) and was identified by linear discriminant analysis coupled with effect size (LEfSe) using the default parameters. CT: the basal diet; PEO: the basal diet supplemented with plant extracts; VIRG: the basal diet supplemented with virginiamycin. Significantly differential metabolites in the cecum and serum of broiler chickens fed without or with PEO or VIRG. Differential metabolites on PEO vs. CT in the cecum (A), VIRG vs. CT and VIRG in the cecum (B); differential metabolites on PEO vs. CT in the serum (C), VIRG vs. CT and VIRG in the serum (D). Metabolites accountable for class discrimination with VIP > 1 and P < 0.05 were listed. CT: the basal diet; PEO: the basal diet supplemented with plant extracts; VIRG: the basal diet supplemented with virginiamycin. acid, and Pyridoxal (Vitamin B6) in the cecum and D-Mannose in the serum. Additionally, the relative abundance of unclassified Rikenellaceae positively correlated with 1-Oleoyl-sn-glycero-3-phosphocholine, adenine, cytidine, and undecanoic acid in cecum (Fig. 4A) and negatively correlated with (S)-Equol, p-Hydroxyphenylacetic acid, and ribitol in the cecum and in the serum of chicken (Fig. 4B). The genus Alistipes positively correlated with 1-Palmitoylglycerol, myristoleic acid, and caprylic acid in the cecum and D-Mannose and 9,10-DiHOME in the serum of chicken, negatively correlated with β-Homoproline and p-Hydroxyphenylacetic acid in caecum and in citrate and urea in serum.
The high proportion of bacteria in the cecal microbiota of the VIRG group positively correlated with ribitol and (S)-Equol in the cecum and oleic acid, SOPC, and adynerin in the serum, and negatively correlated with cytidine, hydroxyisocaproic acid, and Pyridoxal (Vitamin B6) in the cecum. Additionally, the relative abundance of phylum Bacteroidetes and genera Bacteroides positively correlated with LysoPE(16:0/0:0), salicylic acid, and tamsulosin in the cecum and betaine, taurochenodeoxycholate, and dopamine in the serum while negatively correlating with hydroxyisocaproic acid, cytosine, taurine, and Pyridoxal (Vitamin B6) in the cecum and taurine in the serum. The relative abundance of phylum Firmicutes positively correlated with hydroxyisocaproic acid, cytosine, taurine, and Pyridoxal (Vitamin B6) in the cecum and taurine in the serum, while negatively

Effects of dietary plan essential oils or virginiamycin on KEGG pathway in caecum and serum.
Based on these metabolites (VIP > 0.1, P < 0.05), the software Metaboanalyst 3.0 was used to enrich the relevant KEGG pathways. The relevant KEGG pathways enriched by metabolites between PEO within the CT group are shown in Table 1. As compared to the CT group, three relevant KEGG pathways were significantly enriched for these metabolites in the cecum of the PEO group, including cecal primary bile acid biosynthesis (up-regulated by Chenodeoxycholate (1.86 FC (Fold change)) and Glycocholic acid (2.80 FC)); cecal Phenylalanine metabolism (up-regulated by Salicylic acid (2.02 FC) and Pyrimidine metabolism. There were six relevant KEGG pathways enriched in the serum of chickens of the PEO group (Fig. 5C), including fructose and mannose metabolism (up-regulated by D-mannose (1.46 FC) and D-allose (1.13 FC)), biosynthesis of unsaturated fatty acids, Linoleic acid metabolism, and serum fatty acid biosynthesis were enriched.
The relevant KEGG pathways enriched by metabolites between VIRG within the CT group are shown in Table 2. Dietary supplement virginiamycin altered 10 cecal KEGG pathways and seven serum KEGG pathways of chicken. D-mannose and L-leucine up-regulated the KEGG pathway of Lysosome and the mTOR signaling pathway respectively. D-glucosamine 6-phosphate up-regulated and L-alanine down-regulated the KEGG pathway of

Discussion
Since the ban on antibiotic growth-promotion (AGP) in many countries, several effective alternatives to AGPs have been developed in recent years. One of them, PEO or phytogenic additives are considered as appropriate candidates due to their safety benefits 15 14,23,24 . PEO might promote growth by altering gut microflora and hence improving absorption of nutrients 6 , increasing the absorption of micronutrients in the small intestine 7 , stimulating the production and activity of digestive enzymes (such as trypsin and amylase) 25,26 , and stimulating digestive and physiology metabolism 27 . This study shows that both PEO and VIRG treatments increased the relative abundance of phyla Bacteroidetes and decreased the relative abundance of phyla Firmicutes and genera of Lactobacillus and streptococcus in the cecal microbiota of chickens. Meanwhile, the relative abundance of phyla Bacteroidetes and genera Alistipes, Roseburia, and unclassified Rikenellaceae increased in the cecal microbiota of the PEO group. These finding were consistent with previous studies that reported that PEO decreased the number of Lactobacilli 28,29 . Next, some researchers reported that there was a significant increase in the Lactobacillus in the intestines of chickens fed with PEO or Macleaya cordata extracts 18 . These differences may be due to the difference in the source and breed of PEOs, feeding patterns, and sampling parts of the chicken 14 .
The intestinal microbiota plays multiple roles in the intestinal morphology, immunity, nutrient absorption and metabolism, and host health 1,2 . Phyla Bacteroidetes are gram-negative bacteria that ferment polysaccharides and other indigestible carbohydrates and produce short-chain fatty acids (SCFAs) that are gut-friendly. Bacteroidetes are related to fat accumulation in chickens 30 . Alistipes and unclassified Rikenellaceae, which belong to phylum Bacteroidales, are generally considered beneficial to the host gut 31 . Alistipes can produce succinic acid and other long-chain fatty acids such as C15 32 . Rikenellaceae can produce propionic and succinic acids by fermentation of glucose, lactose, mannose, and melibiose, and formed the iso-methyl branched-chain fatty acid or long chain saturated acids 33 . Genus Roseburia belongs to the Lachnospiraceae family, a butyrate-producing organism, with a high capacity to form conjugated linoleic acid from linoleic acid 34   www.nature.com/scientificreports www.nature.com/scientificreports/ that phylum Bacteroidetes were positively correlated with LysoPE (16:0/0:0) and (s)-equol in the cecum and betaine, taurochenodeoxycholate, and dopamine in the serum, while negatively correlating with hydroxyisocaproic acid, cytosine, and taurine in the cecum and taurine in the serum. Moreover, unclassified Rikenellaceae and Alistipes were positively correlated with 1-Palmitoylglycerol, myristoleic acid, and Pyridoxal (Vitamin B6) in the cecum and D-Mannose in the serum. These metabolites participated in many biological functions. e.g. LysoPE (16:0/0:0), a lysophospholipid, that serves important signaling functions. Equol may enhance the actions of soy isoflavones, due to its greater affinity for estrogen receptors and have unique antiandrogenic and antioxidant activities 5 . Dopamine is a major transmitter in the extrapyramidal system of the brain and is important in regulating movement. A family of receptors (dopamine receptors) mediates action, which plays a major role in the reward-motivated behaviour 35 . The enrichment of the relative abundance of Phyla Bacteroidetes and genera Alistipes, unclassified Rikenellaceae might be related with biosynthesis of fatty acid and lipid metabolism. Recently, some researchers have reported that the increase in the proportion of Phyla Bacteroidetes is related to promoting animal growth performance. On the contrary, Although Lactobacillus is considered to be a beneficial probiotic to the host intestinal health and growth 36 , our results showed that the proportion of Lactobacillus in the cecum negatively correlated with lipid metabolites, such as LysoPE (16:0/0:0) and (S)-Equol in the cecum and alpha-linolenic acid, all cis-(6,9,12)-linolenic acid, linoleic acid, taurochenodeoxycholate, and dopamine in the serum; and positively correlated with some metabolites which accelerated lipid peroxidation (such as hydroxyisocaproic acid and cytosine), down-regulated KEGG pathway related to primary bile acid biosynthesis and citrate cycle (TCA cycle). These results were consistent with De Boever et al. 37 , who reported that Lactobacillus impaired lipid absorption and consequently resulted in the dietary energy losses 37 . Recently, some reports showed that some Lactobacillus strains had a negative influence on the growth performance 30,38 , some strains of Lactobacillus are retailed as weight loss probiotics while others are reported with the ability to reduce obesity for increasing appetite and feed consumption 39,40 . Therefore, alteration of intestinal flora may result to improve of animal performance.
Metabolomic analysis revealed that chenodeoxycholate and glycocholic acid up-regulated the KEGG pathway of primary bile acid biosynthesis in the cecal contents of PEO group, Primary bile acids are also critical to the digestion and absorption of fat 41 . In the blood of chickens of PEO group, six lipid metabolites (such as LysoPC (18:1(9Z)) and LysoPE (16:0/0:0)) enriched fatty acid biosynthesus, Lyso PCs perform essential functions in the lipid metabolism of organisms 9,42 ; 1-Palmitoylglycerol is the biosynthetic precursor of phosphatidic acid: the major component of lysophosphatidic acid, which is a pluripotent lipid mediator which controls growth, motility, and differentiation 43 . In addition, D-Mannose up-regulated KEGG pathway of fructose and mannose metabolism and ABC transporters in serum of the PEO group. ABC transporters utilize the energy of ATP binding and hydrolysis to transport various substrates across cellular membranes 44 . Above all, dietary supplementation with PEO alters the cecal and blood metabolic pathways, The biosynthesis pathways of fatty acids and unsaturated fatty acids that are closely related to host lipid metabolism might be an important mechanism for growth promoters of PEO 13 .
The mechanism of improving performance of AGPs is related to reduced incidences of subclinical infections, stability of the microbial ecology, thinning of the intestinal wall, suppression of inflammation, and the reduction of bioamines and toxins produced by the bacteria 45,46 . Recently, extensive research on the effects of antibiotics on intestinal biochemicals has been carried out using culture-independent methods and metabolomic analysis 9,10,47 . Many reports showed that the relative abundance of phyla Bacteroidetes increased and that of genus Lactobacillus decreased in the intestines of chickens after the administration of antibiotic 9,10,47-50 , Higher numbers of lactobacilli were previously implicated in broiler growth depression due to competition in nutrient uptake or impaired fat absorption 30,48 . Some other studies showed that proportion of phyla Firmicutes and genus of Lactobacillus increased in the gut of the chicken after administration of antibiotics [51][52][53] . Metabolomic analysis showed that 4-EOTC treatment improved the synthesis of lysophosphatidylcholine (LysoPC), as indicated by the lipid biomarkers LysoPC (16:0), LysoPC (18:3), LysoPC (20:3), and LysoPC (20:4) in the blood of Wistar rats. VIRG treatment altered 218 biochemicals (156 increased, 62 decreased) in the ileum of chickens, including many long chain saturated and polyunsaturated fatty acids; several lysophospholipids also increased in the process 10 . This observation was corroborated by our data that VIRG treatment increased the relative content of phyla Bacteroidetes and decreased that of genus Lactobacillus in the cecum, enriched lipid metabolites and biosynthesis of fatty acid and linoleic acid, and reduced some metabolites (such as urea) that may be toxic or harmful to growth of host.
Compared to the VIRG group, PEO group also increased the content of D-Mannose and arachidonic acid, docosapentaenoic acid, linoleic acid in the cecum of chickens. These metabolites are essential fatty acids with several biological functions: as the precursor that are metabolized by various enzymes to a wide range of bioactive components for growth and proper health 54 . Arachidonic acid is catalyzed by cyclooxygenase (COX) to generate prostaglandins (PGs), which regulate fever, inflammation, smooth muscle contraction, and translocation of water and salt in the kidney 55,56 . In addition, PEO treatment increases the biosystem of Pyridoxal (VB6) and Pantothenate in the cecal contents. Pyridoxal participates protein and sugar metabolism 57 , and the active form of VB6 is linked with adipogenesis 58 . These results indicate that PEO might alter different lipid metabolites and vitamins from AGP.
In summary, this study investigated the microbial profile and metabolites in the cecum and serum of chickens fed with PEO or VIRG by combined microbiome and metabolomic analysis. Our results demonstrated that PEO treatment increased the relative abundance of phyla Bacteroidetes and genera Alistipes, unclassified Rikenellaceae and decreased that of phyla Firmicutes and genus of Lactobacillus in the cecum. Many Lipid metabolites and KEGG pathway of fatty acid biosynthesis were enriched in cecum and serum of chickens in the PEO and VIRG group. These finding provide better understanding of the mechanism of promoting performance of PEO or VIRG, which could further provide useful information for developing an effective and safe alternative to AGP in poultry industry. Animals, treatment, management, and ethics statement. One-day-old male Cobb 500 chickens (234) were randomly divided into three groups, and each group was replicated six times with 12 chickens in each replicate. Each group was fed with: (1) a basal diet (CT), without any additives; (2) a basal diet with 400 mg/kg (PEO); (3) a basal diet with 30 mg/kg virginiamycin (VIRG). Virginiamycin purchased from Phibro Animal Health Corporation (USA). The corn-soybean-based basal diet was used, and the nutritional level of the diet was prepared following the NRC (1994) feeding standards. All experiments were performed in accordance with the approved guidelines and regulations. The diet composition and nutrient levels are shown in Table 3.
Chickens were reared in a coop (100 cm × 80 cm) with two water nipples per pen. Chickens were raised routinely and immunized according to the normal immunization procedures. The diet was available ad libitum throughout the trial period. The temperature of the house was maintained at the optimum temperature of broiler chickens at 33-34 °C for the first week, and decreased from the second week by 3 °C per week until it reached 22 °C. All experimental protocols were approved by the China Agricultural University Animal Care Committee (permit number SYXK20171208).
Sampling collection. At 28 days of age, six chickens per treatment group (one chicken per repeat) were randomly selected to collected blood from the left brachial vein. After blood was collected, the chicken was injected intravenously pentobarbital sodium (30 mg/kg body weight) and cervical dislocation was executed about 2 g of cecal digestive fluid was collected and placed in two sterilizing tubes, immediately frozen in liquid nitrogen and stored at −80 °C. One sample was used for DNA extraction and Pyrosequencing, and second one for global metabolomic analysis. The blood samples were centrifuged for 15 min (1,500 g, 4 °C). serum sample was stored at −80 °C for UPLC-Q-TOF/MS analysis. DNA Extraction and 16S rDNA Amplicon Pyrosequencing. Total bacterial genomic DNA was extracted from each digested sample using the Fast DNA SPIN extraction kits (MP Biomedicals, Santa Ana, CA, USA). DNA extracts were stored at −20 °C prior to further analysis. Extracted DNAs were measured using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and agarose gel electrophoresis.  Sequence analysis. The sequencing data were processed using the Quantitative Insights Into Microbial Ecology (QIIME, v1.8.0) pipeline, following Caporaso et al. (2010) 60 . Briefly, raw sequencing reads with complete barcode matches were assigned to the appropriate sample and identified as valid. Sequences < 150 bp long, with average Phred scores < 20, containing ambiguous bases, or with mononucleotide repeats longer than 8 bp were considered low-quality and were excluded from further analysis 61 . Paired-end reads were assembled using FLASH 62 . After chimera detection, UCLUST 63 was used to group the remaining high-quality sequences into OTUs based on a minimum sequence identity of 97%. A representative sequence was identified for each OTU using default parameters. The OTUs were taxonomically classified using BLAST; each representative sequence was searched against the Greengenes Database 64 and the best hit was selected 65 13 . LEfSe (Linear discriminant analysis effect size) was performed to detect differential abundant taxa across groups using the default parameters 66 . Operational taxonomic units were clustered with a 97% similarity threshold. Alpha diversity analysis included Shannon diversity index, Chao1 richness estimate, and observed species richness. Principal component analysis (PCA) was performed at the genus level 13 . LEfSe (Linear discriminant analysis effect size) was performed to detect differential abundant taxa across groups using the default parameters 66 .

LC-MS/MS Analysis for serum and cecal contents.
Determination of cecal contents and serum by using a UHPLC (1290 Infinity LC, Agilent Technologies) coupled to a quadrupole time-of-flight (AB Sciex TripleTOF 6600) in Shanghai Applied Protein Technology Co., Ltd. According to Wenqiang Fan's method, the HILIC separation was carried out samples were analyzed by using a 2.1 mm × 100 mm ACQUIY UPLC BEH 1.7 µm column (waters, Ireland) 69 . In the ESI positive and negative mode, the mobile phase contained A = 25 mM ammonium acetate and 25 mM aqueous ammonium hydroxide, B = acetonitrile. The gradient was 85% B for 1 minute, followed by a linear drop to 65% in 11 minutes, then a drop to 40% in 0.1 minutes for 4 minutes, and ultimately increased to 85% in 0.1 minutes for 5 minutes with a rebalancing period. ESI conditions were set as follows: the ion source Gas1 (Gas1) was 60, the ion source Gas2 (Gas2) was 60, the curtain gas (CUR) was 30, the source temperature was 600 °C, and the ion spray voltage fluctuation (ISVF) was ±5500 V. In the MS only acquisition, the instrument was set to collect data in the range of 60-1000 Da, and the TOF MS scanning accumulation time was set to 0.20 s/spectrum. In the automatic MS/MS acquisition, the instrument was set at the range of m/z and above 25-1000 Da, and the production scanning accumulation time was set at 0.05 s/spectrum 70 . The production scan is acquired using information dependent acquisition (IDA) with the selection of high sensitivity mode. The parameters were set as follows 71 : the collision energy (CE) was fixed at 35 V with ±15 eV; declustering potential (DP), 60 V(+) and −60V(−); exclude isotopes within 4 Da; and 10 candidate ions to monitor per cycle.
The raw MS data (wiff. scan files) were converted to MzXML files using ProteoWizard MSConvert before importing it into freely available XCMS software 67 . For peak selection, the following parameters were used: cent-Wave m/z= 25 ppm, peak width = c (10, 60), prefilter = c (10, 100). For peak grouping, bw = 5, mzwid = 0.025, minfrac = 0.5 were used 71 . CAMERA (Collection of Algorithms of MEtabolite pRofile Annotation) was used for annotation of isotopes and admixtures. In the extracted ion features, only more than 50% of the variables of at Scientific RepoRtS | (2020) 10:5382 | https://doi.org/10.1038/s41598-020-60135-x www.nature.com/scientificreports www.nature.com/scientificreports/ least one set of non-zero measurements were retained. The variation coefficient (CV) of metabolites in the QC samples was set at a threshold value of 30%, which was used as the standard for evaluating the repeatability of the metabolomics data set. Compound identification of metabolites was performed by comparing the accuracy of m/z value (<25 ppm), and MS/MS spectra with an in-house database established using reliable standards 72 .

Statistical analysis.
After normalizing the processed data to the total peak intensity, it was uploaded before importing into SIMCA-P (version 14.1, Umetrics, Umea, Sweden), for multivariate data analysis, including pareto scale principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) 71 . The importance of each variable in the OPLS-DA model in the projection (VIP) value was calculated to indicate its contribution to the classification. Metabolites with the VIP score > 1 were further applied to Student's t-test at univariate level to measure the significance of each metabolite, the P values less than 0.05 were considered as statistically significant. The correlative analysis between different cecal microbial species and microbial or serum metabolites was used for calculating by Spearman's correlations coefficient 13 .