Gut Microbiota Mediates Oolong Tea Polyphenols Alleviation of Circadian Rhythm Disturbance

Tea polyphenols can be digested in the intestinal tract so as to promote the growth of helpful gut microbiota, and through the production of catechin, tryptophan, acetic acid and other active substances which involved in the modulation of circadian rhythms mechanism to improve the circadian rhythm disorders and neurological weakness, while its specic mechanism is still unclear. The interaction between host circadian rhythm and gut microbes through the gut-brain axis (GBA) provides new clues for tea polyphenols to improve host health. Our present research mainly investigated the metabolites of the gut microbiota and the heterogeneous expression of circadian rhythm genes in the liver and hypothalamic, and then revealed the modulatory role of oolong tea polyphenols (OTP) of circadian rhythm disorder via the GBA. We used 16S rDNA and untargeted metabolomics sequencing techniques to analyze the effects of OTP on intestinal ora diversity and abundance of specic ora in mice with diurnal disorders, and to screen out marker metabolites which may be involved in circadian rhythm regulation. Transcriptomics and 10X single-cell sequencing were used to evaluate the effects of OTP on circadian rhythm genes in liver and hypothalamus and hypothalamus cell types in mice with diurnal disorders. The Y maze and Morris water maze experiments were investigated the effects of OTP on long-term and short-term memory impairment.


Introduction
The circadian rhythm is an indispensable and important factor in the host's life cycle, which enables the host metabolism regularly within 24 h [1]. The mammalian circadian clock system consists of a master clock located in the suprachiasmatic nucleus of the hypothalamus and a peripheral clock composed of organs such as the intestine, liver, and heart, etc. Through them, metabolism of the host was synchronized with the light-dark cycle [2]. Interestingly, the biological clock can improve the learning ability which depended on the hippocampus and exert a powerful in uence on cognitive function [3]. Disturbance of circadian rhythm is usually accompanied by sleep disturbance, which was one of the main health problems nowadays [4]. People with circadian rhythm disorders, such as shift workers or people who frequently y over time zones, were more likely to suffer from insomnia, depression, and neurodegenerative diseases [2]. Study using animals with knock-out speci c clock genes in the system or tissue has shown that changes in the circadian rhythm can lead to abnormal metabolic phenotypes, including obesity, dyslipidemia, impaired insulin secretion, etc. [5].
In mammals, host cellular and behavioral processes showed 24 h circadian oscillations synchronized with the external light/dark cycle [6]. The gut microbiota also has circadian rhythm uctuations, and the host circadian rhythm disturbance could affect the composition and function of the intestinal microbiota [7]. The gut microbiota was a complex micro-ecological environment, with trillions of microbes inhabiting the intestines of adults [8]. From a genetic point of view, more than 99% of human genes were microbial genes [9]. Recently research found that the loss or variation of the gut microbiota may affect neurogenesis and blood-brain barrier functions, as well as the host's cognitive function, learning ability, and emotion [10], which were partly similar to the harmful effects caused by circadian rhythm disorders.
Most of the gut microbiota coexist with intestinal epithelial cells. It also acted a major role in food intake and metabolism, maintaining the intestinal barrier and the structure and function of the gastrointestinal tract [11].
Tea polyphenols can improve memory impairment caused by rhythm disorders through mechanisms related to the biological clock [12]. Epigallocatechin gallate (EGCG) is the most abundant active ingredient in tea polyphenols (TP). As the O-methylated form of EGCG, (-)-epigallocatechin 3-O-(3-Omethyl) gallate (EGCG3''Me), which was rich in oolong tea polyphenols (OTP) showed a potential activity similar to prebiotics [13]. OTP not only modulate the composition of the intestinal ora, but also have a positive regulatory effect on the key circadian rhythm genes in the liver [14]. Due to the low bioavailability, most OTP were metabolized by the intestinal ora into metabolic derivatives in the colon to exert biologically active functions [15]. Active metabolites can act as an intermediary between the intestine and the brain to participate in the modulation of host circadian rhythm disorders [16], but the relationship among intestinal metabolites, intestinal microbiota and rhythm genes is still unclear. Single-cell RNA sequencing (scRNA-seq) has recently been used to understand the development and interaction of heterogeneous cells, and to detect the expression and transcription subtypes of a large number of genes throughout the genome [17]. This allows us to fairly evaluate multiple different cells and genes in the hypothalamus at the single-cell level, to better explore the link between the gut and the hypothalamus.
In this study, a mouse model of circadian rhythm disturbance was established to investigate OTPmediated changes in intestinal microbiota, metabolites, and clock genes. Subsequently, the feces of OTP ingested mice were transplanted, and the effects of OTP intervention on mice with circadian rhythm disorders were characterized internally and externally through maze behavior experiment and hypothalamus scRNA-seq. Through fecal microbiota transplantation, a comprehensive analysis of the role of OTP in the internal and external circadian clock may provide valuable theoretical support for revealing that OTP can improve circadian rhythm disturbance via the GBA.

Contents of tea catechins in OTP
In the present study, high purity (> 95%) of OTP was prepared (Supplementary Table S1). Results indicated that the EGCG content was the highest, making it the main tea catechin, while the content of EGCG3''Me was high.
In uence of OTP on body weight, food and water consumption of the mouse models In Fig. S1a-c, the body weight of the constant dark group (CD group) mice increased rapidly after the rst week and was higher than that of the 12 h normal light-dark cycle control group (CT group) and group under CD condition supplement with OTP (OTP group). After four weeks of OTP feeding, although the fractional body weight showed an increasing trend, it was slower than that of the CD group. The average body weight in OTP group was (23.88 ± 3.99) g, which was lower than the average body weight of the CD group mice (25.77 ± 5.09) g in the same period. Eight weeks later, the average weight of the veri cation group (VF group) was (27.86 ± 5.17) g. The growth rate of mice in the VF group was signi cantly lower than that in the CD group from the fth week, and the body weight gradually approached the CT group (P < 0.05). But there was no signi cant difference in food and water intake during the feeding process among all groups (P > 0.05).
In uences of OTP on cognitive ability of mice with circadian rhythm disturbance through Y-maze and Morris water maze In order to further observe the in uence of OTP on the memory impairment caused by circadian rhythm disturbance, behavioral veri cation was carried out through Morris water maze and Y maze experiments. As shown in Fig. S2a, the total number of times of entering the maze arm of the three groups of mice didn't show a signi cant difference (P > 0.05). In Fig. S2b, the accuracy of spontaneous alternation of the arm in the CD group was 45.3 ± 1.6%, which was signi cantly lower than that of 62.8 ± 2.3% in the CT group (P < 0.05). After OTP intervention, the spontaneous alternation accuracy rate of mice in the OTP group increased to 60.4 ± 3.2%. The desire to explore and short-term working memory of mice in the CD group were the weakest. The results showed that circadian rhythm disorder will damage the short-term spatial memory ability of mice, while OTP can alleviate this adverse effect, and has no obvious effect on its exercise ability.
During the water maze test, the escape latency and distance of different groups differed signi cantly. On the 6 th day of the experiment, the escape latency to the platform for the rst time was reduced to 26.0 ± 0.4 s in CD group which was signi cantly higher than that of the CT group (13.8 ± 0.16 s) and the OTP (16.3 ± 0.19 s) group (Fig. S3a, b). After the platform was removed, the swimming time percentages of the mice in the CT group, CD group, and OTP group in the quadrant of the original platform were 37 ± 2.0%, 24 ± 1.3%, and 30.5 ± 1.6%, respectively (Fig. S3c, d). It can be seen that the disorder of the circadian rhythm has signi cant negative effects on the activity state of the mice in the target quadrant, and OTP has improved this phenomenon.
The in uence of OTP on the diversity of intestinal microbiota in mouse models As shown in Table 1, the Chao1 index of all groups showed an upward trend after 4 weeks, while the CD group was signi cantly lower than that of the CT group and OTP group at the 4 th week (P ≤ 0.05). It showed that CD conditions signi cantly reduce the abundance of intestinal ora in the mouse model.
After supplementing OTP, the diversity of the total microbial community increased. Both Shannon and Simpson indexes showed that the fecal microbial diversity of the OTP group was signi cantly higher than that of the CD group (P ≤ 0.05). Moreover, principal coordinate analysis (PCA) showed that the fecal microbiota from the CT-4, CD-4 and OTP-4 could be separated clearly (Fig. S4).
As shown in Fig. 1a, the relative abundance of Firmicutes and Proteobacteria showed a signi cant downward trend after supplementing the mice with OTP for 4 weeks compared with the CD group, while the Bacteroidetes showed a signi cant increasing trend, and the CT group showed the same trend (P < 0.05). The ratio of Firmicutes/Bacteroidetes (F/B) in CD group was 158.72%, which was also signi cantly higher than that of 53.94% in OTP group, indicating that OTP supplementation can effectively inhibit the growth of Firmicutes and help maintain the balance of the intestinal microbiota. In addition, the number of Bacteroidetes in the CD group was signi cantly lower than that in the CT group (50.18 ± 20.87% vs 37.94 ± 10.69%).
The representative micro ora with relative abundance at the family level were shown in Fig. 1b.
Muribaculaceae belongs to the phylum Bacteroidetes, whose diversity provides new data for ecological research, and was named by Ilias Lagkouvardos et al [18]. After 4 weeks of feeding, the relative abundance of Muribaculaceae in the CD group was 0.06% which was signi cantly lower than that in the CT and OTP groups (P < 0.05). The relative abundance of Muribaculaceae in the CD group was signi cantly lower than that in the CT and OTP groups (P < 0.05). In particular, the relative abundance of Lactobacillaceae reached 33.33% in the CD group, which was signi cantly reduced after OTP treatment (P < 0.05).
In order to further characterize the differences between the intestinal microbiota of mice under different treatment conditions, we performed a statistical analysis on the relative abundance of the microbiota in the samples at the genus level (Fig. 1c). Among CD-4 and OTP-4, the most signi cant differences were also Bacteroides, Parabacteroides and Clostridium. On the other hand, the relative abundance of Muribaculum was signi cantly reduced under the dark treatment for 4 consecutive weeks, and it could be increased in the OTP group (P < 0.05). The relative abundance of Clostridiales, Lachnospiraceae, Eubacterium and Lactobacillus in the OTP group decreased. Results showed that the intervention of OTP was bene cial under the conditions of microbial imbalance.
Effects of OTP on the diurnal oscillation of microbiota in mice with circadian rhythm disturbance In the CT, CD and OTP groups, Firmicutes, Bacteroidetes, Proteobacteria, Tenericutes were the dominant ora at the phylum level. Different diurnal oscillation patterns of different microbiota were observed in the three groups as shown in Fig. 2a. For the CT group of mice, the main bacterial communities at the phylum level all showed obvious rhythm changes. However, the oscillation regularity of intestinal ora of mice in CD group was not strong. Interestingly, we found that OTP can signi cantly improve the suppression of periodic changes in phylum levels caused by continuous darkness. For example, we found that at zeitgeber time (ZT) 24, the relative abundance of Firmicutes in the CD group reached the highest, while both the CT and OTP groups were smaller than the CD group. The relative abundance of Bacteroides reached a high level at night (ZT24 and ZT48) which showed a trend higher than that of the CD group. Proteobacteria was commonly used as a microbial marker of intestinal dysregulation [19]. The relative abundance of the Proteobacteria in the OTP group was signi cantly lower than that in the CD group, revealing that supplementation with OTP can positively improve the diurnal oscillation of intestinal microbiota caused by circadian rhythm disturbance. The oscillation of Prevotella, Bacteroides, and Parabacteroides in the OTP group was closer to that in the CT group, and they both showed obvious rhythmic oscillation, while the oscillation of the CD group was disorder. Prevotella in the CT group and OTP group showed a trend of increasing during the day and decreasing at night. At ZT12, the relative abundance of Prevotella in the CD group was signi cantly lower than that in the CT group and OTP group. The relative abundance of Bacteroides in the CT group and OTP group reached the highest at ZT24, while the CD group reached the highest peak at ZT48.
In uences of OTP on metabolites in mice with circadian rhythm disturbance The close grouping of QC samples in the PCA score chart further con rms the high repeatability and stability of the entire test process (Fig. S5). And the metabolites in the CT group, CD group and OTP group can be completely separated, which re ected the better quality of the data. The results showed that the circadian rhythm disorder altered the metabolism of normal mice. The OTP group and the CT group tended to be similar, proving that OTP has a certain effect on improving the circadian rhythm disorder. In the human metabolome database (HMDB) superclass analysis, whether in positive ion mode (PIM) or negative ion mode (NIM), the most abundant metabolites belong to lipids and lipid molecules (Fig. 3a).
Metabolomics analysis was performed to determine the changes of metabolites in mice with circadian rhythm disturbance and mice after OTP intervention. The results of mass spectroscopy showed that 104 signi cantly different secondary metabolites were obtained under PIM and NIM conditions (53 downregulated metabolites, 51 up-regulated metabolites) indicating signi cant differences in metabolites produced by OTP intervention in mice with circadian rhythm disturbance (Fig. 3b-c). Compared with the CD group, the most important characteristic after OTP intervention was a signi cant increase in the number of organic acids and derivatives (20 up-regulated metabolites) as well as lipids and lipid-like molecules (9 up-regulated metabolites). Especially isoleucyl-isoleucine, (R)-2-Benzylsuccinate, 4isopropylbenzoic acid, 2-phenoxyethanol, (-)-catechin which were the top 5 signi cantly different metabolites by sorting in descending order of the variable in uence on projection (VIP). Additionally, other metabolites associated with circadian rhythm and neuroprotection such as acetic acid, 3,4methylenesebacic acid, hydroxyphenyllactic acid, Trp-Asn and some other metabolites were also signi cantly up-regulated after OTP intervention. Inversely, the levels of carcinogens o-toluidine and acylcarnitine, which can cause neurological disorders, were found to be signi cantly reduced after OTP supplementation indicating that OTP can reverse the metabolic disorders caused by constant dark conditions.
After the accurate analysis of the differential metabolites by the KEGG database, the pathways with signi cant differences in the enrichment of differential metabolites whose metabolite levels decreased signi cantly after treatment under constant dark conditions mainly included "glycerophospholipid metabolism", "tryptophan metabolism", "arginine biosynthesis", etc. (Fig. 3d). After OTP intervened in mice living in constantly dark conditions, the differential metabolite enrichment pathway, which signi cantly increased the level of metabolites, was changed. As shown in Fig. 3e, the identi ed differential metabolite enrichment pathways were "pyruvate metabolism", "glycolysis/gluconeogenesis", "glyoxylate and dicarboxylate metabolism" and "tryptophan metabolism". Pyruvate metabolism was enriched to the highest degree, and the pathway mechanism involved was shown in the Fig. 4, from which it could be seen that acetic acid was signi cantly up-regulated after OTP intervention.
The correlation between intestinal microbiota and intestinal differential metabolites As shown in Fig. 5a, the microbiota most relevant to the signi cantly different intestinal metabolites between the CD group and OTP group were Bacteroidetes, Acidobacteria, Patescibacteria and Firmicutes. The metabolites that were signi cantly positively correlated with Bacteroidetes were (4-aminobutyl) guanidine, 2-phenoxyethanol, acetic acid, orcinol, isoleucyl-isoleucine, Pro-Leu, polyethylene oxidized, 3,4methylenesebacic acid, Gln-Ile, O-acetyl-l-carnitine, Trp-Asn, etc (P < 0.05). Among them, the correlation between acetic acid, polyethylene oxidized, Gln-Ile, O-acetyl-l-carnitine and Bacteroidetes was extremely signi cant (P < 0.01). Interestingly, metabolites mentioned above were negatively correlated with Firmicutes, and their signi cance remained unchanged. The correlation network diagram of intestinal differential metabolites and intestinal ora in the CD and OTP groups was shown in Fig. 5b, and the relationship between different phylum and metabolites can be clearly seen. The metabolites that were signi cantly positively correlated with Firmicutes were acylcarnitine 16:1, 2'-deoxyadenosine and agmatine (P < 0.05). Our data showed that intestinal ora and intestinal metabolites have a good correlation, suggesting that OTP helped the intestinal ora to repair the intestinal environment and promote the production of metabolites.
Impacts of OTP on the expression of hepatic clock genes in mouse models of circadian rhythm disorder In order to further study the molecular mechanism of OTP affecting circadian rhythm in mice, we used transcriptome sequencing technology to perform high-throughput analysis of the peripheral clock tissueliver. As shown in the Fig. S6, the heat map was used to re ect the expression transcription pro le of liver clock genes in circadian rhythm mice within 48 h. Transcriptome data showed that 214 and 208 genes in the CT and OTP group respectively showed circadian rhythm, while only 156 genes in the CD group showed circadian rhythm. This re ected that although continuous darkness will reduce the number of clock genes in the liver of mice, the supplementation of OTP can improve this situation.
To fully grasp the expression of circadian rhythm genes, we analyzed the mRNA expression of Bmal1, Clock, Cry1, Cry2, Per1, Per2, Per3, N l3 and Dbp in the liver within 48 h. As shown in the Fig. 6, most of the liver clock genes have unique circadian rhythmic expression, especially the rhythmic genes in the CT group showed a relatively regular oscillation pattern. While this regular oscillation trend did not appear well in the CD group after constant dark treatment. In particular, genes Cry1, Cry2, Dbp, Per1, Per2 showed a distinctly different oscillation disorder from the CT group. Interestingly, even in the same environment as the constant dark treatment, the oscillation amplitude and phase of rhythmic genes (Cry1, Cry2, Dbp, Per1, Per2) expression began to approach the CT group after supplementing OTP. In the CT group and OTP group, the expression of genes Bmal1 and N l3 showed a downward trend during the light period (ZT12, ZT36) and an upward trend during the dark period (ZT24, ZT48). However, the genes Cry1, Dbp, Per1, Per2 and Per3 showed the opposite expression trend. The Clock and Cry2 genes that lost their rhythm after dark treatment showed a regular oscillation pattern in the OTP group, although the trend was different from the CT group. At the same time, the intervention of OTP also signi cantly increased the oscillation amplitude of genes Clock and Cry2. All these indicated that supplementing OTP can improve the amplitude and phase shift of liver rhythm genes caused by circadian rhythm disorder to a certain extent, and make it closer to the normal level.
scRNA-seq identi es different cell clustering in the hypothalamus of mice A total of 29,635 effective cells were detected in this experiment, of which 14,114 cells belonged to the VF group, 12,035 cells belonged to the CD group, and 3486 cells belonged to the CT group. The t-SNE clustering analysis of the above-mentioned cells in the CT, CD and VF group was performed by an automatic clustering algorithm, and the results of different samples and subgroups of cells were shown in the Fig. 7a-b, respectively. The different colors in the picture represented different clusters and the legend on the right showed the names of each cluster. The proportion of cells in each cluster in the sample and the percentage of different sample cells in each cluster were shown in Fig. S7.
The results showed that there were 22 clusters in the CD group and 25 clusters in the VF group. After OTP intervention, the number of cells in the cluster2, cluster3, cluster4, cluster11, cluster15, cluster18, cluster19, cluster20, cluster21, cluster22, cluster23, cluster24 and cluster25 in the VF group was signi cantly higher than that in the CD group. Among them, the number of cells in cluster15, cluster19, cluster23, cluster24, and cluster25 in the CD group was 0, which can be used as clusters unique to OTP for cell type heterogeneity identi cation research.

Hypothalamus cell types and marker genes identi cation
The result with the highest corresponding cell type of each cluster was selected as the result of cell type identi cation of this cluster. The results of cell identi cation in the CT group, CD group and VF group were shown in Fig. 7c. The speci c results of the cell types identi ed by each cluster and their percentages were shown in Table S2. As shown in Fig. 7d, the cell types of the two groups of samples were annotated respectively, and the heterogeneity of the expression of each cell type in the two groups of samples could be clearly and intuitively seen. The main cell types identi ed by the CD group and VF group of samples included astrocytes, oligodendrocytes, microglia, macrophages, neurons, epithelial cells, endothelial cell, and broblast. Fibroblasts only uniquely existed in the VF group, but not in the CD group.
We focused on the analysis of cluster15, cluster19, cluster23, cluster24, and cluster25 to determine the cell heterogeneity between the CD group and the VF group. Cluster15 and cluster25 were identi ed as oligodendrocytes, which acted an indispensable role in neuron protection and development [20], and the number of cells in the VF group was signi cantly higher than that in the CD group. In cluster19, 79.11% were identi ed as oligodendrocytes, 16% were neurons, and 4.89% were astrocytes which were important parts of the central nervous system cells. Cluster23 and cluster24 were identi ed as macrophages and broblasts, respectively. After fecal bacteria transplantation, the number of these cells in the hypothalamus was signi cantly increased.
Circadian rhythm genes expression of each cell types in and around the hypothalamus The expression of marker genes in each cluster was shown in Fig. S8, the high expression of different marker genes re ected the unique genes in each cluster. The average gene expression pro les of 30,691 gene samples that were common in 8 cells was shown in Table S3.
We selected a total of 16 co-expressed rhythmic genes (Clock, Bmal1, Cry1, Cry2, Per1, Per2, Per3, Bhlhe41, Nr1d1, Nr1d2, Dbp, Rorb, Tim, Dbt, Sirt, N l3) from the top 8 major cell types, which were mainly core clock genes and genes involved in coding clock gene proteins. The expression levels of circadian genes in broblasts, epithelial cells, endothelial cell, neurons and astrocytes were relatively high (Fig.   S9a). Moreover, the expression level in VF group was higher than that in CD group, especially the percent expressed ratio of rhythm genes Cry2, Per3, Bhlhe41, Nr1d1, Nr1d2, Dbp, Rorb in each cell in the VF group (Fig. S9b).
The speci c expression of circadian rhythm genes in each cell and different groups was shown in Fig. 8a. The percent expressed ratio and average expression of the rhythm gene Clock in the neurons of the VF group were the highest. Although the percent expressed ratio of Clock in oligodendrocytes in the VF group was similar to that of neurons, the average expression was much lower than neurons. Meanwhile, it was found that the average expression of other rhythm genes Clock, Bmal1, Cry1, Per2, Per3, Nr1d2, Dbp and Rorb in neurons of the VF group was also relatively high. The t-SNE map was used to cluster and identify the expression patterns of circadian rhythm genes in neurons of the CD and VF group, as shown in Fig.  8b. From the t-SNE graph of gene expression, it can be re ected that Clock, Cry2, Per3, Bhlhe41, Nr1d1, Nr1d2, Dbp and Rorb have a relatively high expression ratio in neurons in the VF group. From this we speculate that neurons were one of the main cells expressing circadian rhythm genes, which was consistent with the ndings of Branchaccio et al [20]. And OTP can enhance the expression of circadian rhythm genes in them.
The violin plot can clearly show the expression of circadian rhythm genes in each cell in the hypothalamus (Fig. 8c). Compared with the CD group, the Clock, Cry2, Per3, Bhlhe41, Nr1d1, Nr1d2, Dbp and other genes related to the circadian rhythm in the VF group were better expressed in each cell. In the CD group, the top 3 cells with the expression levels of the core clock genes Clock and Bmal1 were astrocytes, oligodendrocytes and epithelial cells. While in the VF group, the top 3 cells with the Clock expression levels were astrocytes, neurons and epithelial cells. The top 3 cells with Bmal1 expression level in the VF group were consistent with the CD group.

Gene Ontology (GO) and KEGG analysis of different expression genes (DEGs) in liver and hypothalamus
The results of liver transcriptome sequencing showed that after 4 weeks of OTP intervention, GO terms enriched in the biological process mainly included "negative regulation of transcription by RNA polymerase II", "regulation of transcription, DNA-templated" and "negative regulation of transcription, DNA-templated". It showed that OTP could improve the effects of continuous darkness by affecting the biological process of the host. The cellular components were mainly enriched in "nucleus", "cytoplasm", and "membrane" etc. Among the molecular functions were "protein binding", "metal ion binding", and "DNA binding" etc (Fig. 9a). In organisms, different expression genes (DEGs) coordinate with each other to perform their biological functions. On the basis of GO annotation classi cation, KEGG analysis was used to enrich the results for biological pathway analysis [21]. Comparing the two groups of CD and OTP, 396 unigenes were annotated in the KEGG database and mapped to 191 metabolic pathways. Among them, "circadian rhythm", "ubiquinone and other terpenoid-quinone biosynthesis" and "fatty acid elongation" were highly enriched in KEGG pathways which were closely related to the perfection of circadian rhythms and protein phosphorylation in vivo (Fig. 9b).
As shown in Fig. 9c, the top 3 in the biological process eld in hypothalamus were mainly "positive regulation of transcription by RNA polymerase II", "multicellular organism development", and "signal transduction". In terms of cellular component, "membrane", "cytoplasm", and "nucleus" were mainly involved. The main processes in molecular function were "protein binding", "metal ion binding", and "nuclearide binding" which were highly similar to the GO analysis results of liver sequencing. After fecal microbiota transplantation, the DEGs involved in the KEGG pathways between the CD group and VF group have become richer. Mainly re ected in "human diseases", "metabolism" and "organismal systems" as shown in Fig. 9d. According to the degree of gene enrichment in each category, the top 3 pathways associated with "human diseases" were "neurodegenerative disease", "infectious disease: parasitic" and "endocrine and metabolic disease". There were "energy metabolism" (25 DEGs), "lipid metabolism" (28 DEGs) and "carbohydrate metabolism" (26 DEGs) that have a higher percentage of gene expression in "metabolism". In the "organismal systems" category, there were "nervous system", "endocrine system", and "immune system" where the number of DEGs expression distributions was more than 60, the proportion of the number of DEGs expressions in the total gene expression of the them were rose to 14.01%, 8.81% and 7.66%, respectively.

Discussion
As a basic component of the living body, the circadian rhythm is an endogenous and guidable biological process [22]. Each cell of the host contained an internal biological clock, which performed normal turnover via a series of rhythmic genes or molecular oscillations [23]. Irregular lifestyle and diet or circadian rhythm disorders caused by long-term shift work may cause the host's central-peripheral circadian clock disorder and increase the risk of neurodegenerative diseases, cognitive impairment and other metabolic diseases [24]. The study by Thaiss et al. [6] found that transplanting the gut microbiota with jet lag into mice resulted in a signi cant increase in host fat. In our research, it was found that after transplanting the intestinal ora of mice supplemented with OTP into mice with circadian rhythm disorders, the weight of the mice increased extremely slowly and approached mice living in a normal environment. This showed that the intake of OTP can improve the negative response of circadian rhythm disorders to weight.
The host's circadian rhythm pattern can be measured by biological and behavioral indicators. The activity-rest cycle is the characteristic that we spend 24 h a day inducing physiological and cellular adaptation to various processes [25]. In addition to having the extensive effects on physiology, the circadian rhythm also affects the host's learning and memory abilities. Similar to the cognitive impairment of mice with circadian rhythm disorders observed in this experiment, study has shown that mice with disrupted circadian rhythm genes also showed poor performance in various learning behavior tasks [26]. Similarly, the disturbance of humans' circadian rhythm can also lead to the decline of cognitive function. For example, ying across time zones for several years would increase the cognitive impairment of the host [27]. Four weeks after supplementing the OTP, the long-term cognitive learning impairment assessed in the Morris water maze and the short-term memory impairment in the Y-maze were reversed. This phenomenon was similar to the ndings of Biasibetti et al. [28] that tea catechins can reduce oxidative stress in peripheral and brain tissues and inhibit behavioral changes related to cognitive de cits.
Previous study by others has indicated that disruptions in circadian rhythm led to variations in gut microbiota characterized by an increase in pro-in ammatory bacteria and a decrease in bacteria which produce anti-in ammatory short-chain fatty acids (SCFAs) [29]. In this study, we observed that the value of F/B was higher than that of normal mice in mice with circadian rhythm disorders. After OTP intervention, the ratio of F/B in the gut microbiota of mice with circadian rhythm disorders decreased, and Bacteroidetes increased signi cantly. Additionally, supplementation of OTP for 4 weeks also signi cantly increased the relative abundance of Muribaculaceae which was belong to Bacteroidetes. The relatively high relative abundance of Muribaculaceae (previously named S24-7) was related to the extension of lifespan [30]. Muribaculaceae has been also reported to contribute to the production of SCFAs [31]. SCFAs were the main metabolites of the gut microbiota. As signal molecules, SCFAs can activate multiple pathways, such as activating AMP-activated protein kinase in peripheral tissues (such as liver and muscle) and cholesterol or lipids metabolism etc. [32]. Additionally, studies have shown that SCFAs may affect cognitive function and mood through a variety of mechanisms, included regulation of histone acetylation and methylation, promotion of secretion of various hormones and neurochemicals (such as serotonin), and induction of vagal nerve signals [33,34]. It was observed that the relative abundance of acetate and butyrate-producing bacteria in mice supplemented with OTP showed an increasing trend in our research.
Metabolites can drive the relationship between the host and microbiota. At the same time, metabolites were considered to be cellular transmitters of circadian rhythms in various tissues [35]. Microbial-derived SCFAs or lactate were found to regulate the phase of the host's peripheral clock in vivo [36]. Trp is a dietary essential amino acid, which can be directly metabolized in the gastrointestinal tract through the microbiota or serotonin pathway and transmit signals from the intestine to endogenous or exogenous neurons [37]. The metabolism of tryptophan by the microbiota included the conversion of tryptophan to indole and its derivatives, of which 5-Hydroxyindole-3-acetic acid (5-HIAA) is main metabolite [37]. The enrichment pathways that signi cantly down-regulated differential metabolites in the CD group included tryptophan, glycerophospholipid metabolism, etc. Study has shown that phospholipid compounds involved in glycerophospholipid metabolism participate in diseases such as atherosclerosis, vascular dementia, and spinal cord injury by activating the PPARγ pathway [38]. Choline is a strong organic base that exists in sphingomyelin and can be used as a precursor of the neurotransmitter acetylcholine [39].
The content of choline and 5-HIAA in mice that have lived in constantly dark conditions for a long period has a downward trend which indicated that the disorder or destruction of the circadian rhythm may affect the synthesis and transport of neurotransmitters, etc.
The activity of TP depended largely on its conversion in the intestine [40]. Due to the limited bioavailability of TP in the body, most of TP exerted functionally activities by their metabolic derivatives [13]. In this study, OTP and its metabolites have shown various positive effects on metabolic disorders caused by circadian rhythm disorders. After OTP intervention in mice with circadian rhythm disorders, the content of multiple metabolites which originally were have a down-regulated trend in CD group was now signi cantly increased, and tended to the level of the CT group. Some representative metabolites were screened out, such as acetic acid, (-)-catechin, O-acetyl-L-carnitine, 2-indolecarboxylic acid, 3-hydroxy-3-methylglutaric acid, etc. The protein kinase C-related properties of catechins make it have a positive impact on the anti-neurodegenerative disease Alzheimer's disease [41]. The metabolic levels of various amino acids such as tryptophan, glutamine (Gln), alanine, and tyrosine (Tyr) were signi cantly increased after OTP supplementation. Metabolic syndrome and neurological features (signi cantly anxiety, depression, and autism), were mostly affected by the nal product of tryptophan metabolism [37]. 2-indolecarboxylic acid was one of the tryptophan pathways and has been found to be involved in immune regulation, in ammation and affect intestinal function [42]. Gln can effectively enhance the function, proliferation and life cycle of small intestinal epithelial cells, and may indirectly support the optimal neuropsychological environment by protecting the intestinal wall from damage and chronic in ammation [43]. In another study, after long-term supplementation of Gln, the ratio of F/B, a common indicator of obesity, was signi cantly reduced. An optimized intestinal microecological environment was more likely to perform neurotransmitter synthesis and other functions more effectively [44]. Tyr was also a precursor of neurotransmitters formation, and the increase of Tyr can increase the level of neurotransmitter in the body [45]. The supplementation of OTP can increase the level of Gln and Tyr. An optimized intestinal microecological environment was more likely to perform neurotransmitter synthesis and other functions more effectively. Abnormal "pyruvate metabolism" often leads to cancer, heart failure and neurodegenerative diseases. KEGG enrichment analysis of differential metabolites signi cantly up-regulated after OTP intervention showed that the degree of enrichment of "pyruvate metabolism" was the highest which can be explained that the intervention of OTP is bene cial to the reversal of these abnormal phenomena [46].
The liver and intestine were both typical peripheral clock models in the biological clock system. They were synchronized with the circadian rhythm of the central clock in the suprachiasmatic nucleus (SCN) and affected by meal time and dietary composition [47]. The supplementation of OTP can increase the number of rhythmic genes in the liver of mice and positively improved the disorder of the expression patterns of liver clock genes. The microbiota with circadian rhythm oscillations accounts for about 60% of the gut microbiota [6]. And in this study, compared with the microbial population that normally undergoes the light-dark cycle, constant darkness changed the oscillation of some microbes, and this situation can be signi cantly alleviated by supplementing OTP. KEGG analysis showed that DEGs obviously have the highest levels in "protein processing in endoplasmic reticulum", "endocytosis", and "MAPK signaling pathway" after OTP supplementation. It can be seen that after OTP interferes with mice with circadian rhythm disorders, cell transcription and translation activities become active, and can regulate circadian rhythm disorders by participating in the formation of proteins.
Circadian rhythm can be found at the intracellular level and is coordinated by the SCN master clock located in the hypothalamus [48]. Patients with different degrees of Alzheimer's disease were considered to have circadian rhythm disorders, which were speci cally manifested as delayed attenuation of amplitude and phase [49]. The lack or disorder of clock gene expression in astrocytes can lead to neuroin ammation which was the main manifestation factor of neurodegeneration 3 . For instance, the absence of Bmal1 in the mouse brain disrupted the function of the biological clock and also caused the degeneration of the synapses of astrocytes [50]. Astrocytes can selectively remove synapses, ne-tune and reshape the brain's neural circuits during host learning, memory and exercise [48]. It has been con rmed that transplantation of human astrocytes into newborn mice can enhance the memory and learning ability of mice [51]. It has been con rmed that transplantation of human astrocytes into newborn mice can enhance the memory and learning ability of mice [52]. Oligodendrocytes provided metabolic support and nerve repair functions for neurons and axons, which were thought to be related to some mental diseases [53]. Neurons have always been considered a crucial part of SCN. The latest research shows that neurons can not only participate in the regulation of clock gene expression and maintain circadian rhythmic oscillations to coordinate the biological clocks of tissues and organs throughout the body, but also have the ability to autonomously encode circadian rhythms, so as able to initiate and maintain biological rhythmic activities and behaviors [20]. Among the 8 hypothalamic cell types identi ed in this study, we revealed that broblasts only exist in the VF group, and the number of astrocytes and oligodendrocytes was up-regulated. Rhythmic genes such as Clock, Cry2, Per3, Bhlhe41, Nr1d1, Nr1d2, Dbp, etc. showed better expression after OTP intervention. Notably, the expression percentage and average expression level of the rhythm gene Clock in the VF group neurons were both the highest. All these supports the perspective that OTP can improve the negative effects of circadian rhythm disorders by increasing the number of cells and gene expression associated with the circadian rhythms.
The functional enrichment analysis of signi cantly differential gene expression in hypothalamic cells can further support the above viewpoints. GO classi cation analysis suggested that OTP supplementation may ameliorate the negative effects of circadian disruption by altering cell composition [54]. The mechanism by which TP participate in the intestinal microbiota and its metabolism to regulate circadian gene expression includes the interaction of the basic development and metabolism of the microbiota, as well as interference with cell membrane function and energy metabolism [55]. KEGG pathway analysis showed that after fecal microbiota transplantation, DEGs were mainly enriched in "signal transduction", "cancer: overview", "nervous system" and "endocrine system". These pathways were closely associated with the health of the host.

Conclusions
In this study, the intervention of OTP reduced the F/B ratio, the daily oscillation of the intestinal microbiota tended to be regular, and the intestinal microecological status was also better than that in the CD environment. After OTP mediated, the content of important metabolic targets closely associated with host metabolism such as Trp, Gln, Thr, Pro, alanine, 2-indolecarboxylic acid and some others has been signi cantly increased. The poor expression of liver clock genes has been improved, KEGG analysis showed that the signi cance differences between the OTP group and CD group were in "signal transduction", "endocrine system", "cancer: overview" and "immune system", etc. After fecal microbiota transplantation in the OTP group, the expression percentages of most rhythm genes such as Cry2, Per3, Bhlhe41, Nr1d1, Nr1d2, Dbp and Rorb were higher in hypothalamus cells than those in the constant dark environment. Behavioral experiments con rmed that the short-term and long-term cognitive learning abilities of mice after fecal microbiota transplantation were better than those in the CD group. These results indicated that OTP can signi cantly improve the metabolic imbalance caused by the circadian rhythm disorder, maintain the host's homeostasis, and improve cognitive ability.

Materials Chemicals and Reagents
Oolong tea was produced in Baifeng Tea Garden, Beilun District, Ningbo City, Zhejiang Province. Polyamide resin was purchased from Ocean Chemical Co., Ltd. (Qingdao, China). Sterile male mice C57BL/6J were purchased from Hunan Slack Experimental Animal Co., Ltd. Ordinary animal feed is provided by the Experimental Animal Center of Ningbo University.

Preparation of OTP
The dried sample was ground into a powder with a grinder, and the material passing through a 40-mesh sieve was stored in a sealed polyethylene bag at a temperature of 20 °C until use. Shortly, take the tea powder and add distilled water, and place it at 96 °C for 40 min. The extract was centrifuged at 4500 g for 15 min. Repeat the above process. According to the method we previously reported, the residue is dissolved, sorted and puri ed through a polyamide column. The eluate was analyzed using high performance liquid chromatography [56]. And collect, concentrate, load on the polyamide column and process the required fractions as described above. The results showed that the components containing OTP were concentrated and freeze-dried through a freeze-drying system to obtain nished products.

Animals and experimental design
Animal studies were carried out in the Centre for Laboratory Animals, Ningbo University (Permission No. SYXK [Zhejiang] 2013-0191). All the animal-related operations were strictly in compliance with the relevant laws and regulations on the use and care of laboratory animals in China. All mice were individually placed in a sterile barrier environment with a room temperature of 22.3 °C and a humidity of 55.5%. Adapt to the environment for 7 days under a 12h light-dark cycle period [ZT0 = 6 am, lights on; ZT12 = 6 pm, lights off]. Seven days later, they were randomly divided into 4 groups (18 animals in each group), namely CT group, CD group, VF group and OTP group. The CT group continued to grow in a 12 h light-dark cycle environment, the CD and VF group was placed in a completely dark environment, and the OTP group was placed in a completely dark environment and fed with 0.1% (w/w) OTP once a day [54].
At the same time, the CT, CD and VF group were given the same amount of sterile water per day. The body weight, water and food consumption of CT, CD, OTP and VF group of animals were recorded every week. After 4 weeks of continuous administration, laparotomy was conducted on 3 mice of CT, CD and OTP group under pentobarbital anesthesia every 12 h over 48 h, which can be divided into 5 groups according to time points and fresh feces were collected respectively for intestinal microbiota analysis. The liver and contents of the cecum were collected immediately after euthanasia for analysis of clock genes and differential metabolites.
After sampling, the intestinal ora of the remaining 3 mice in the OTP group were transplanted into the mice in the VF group for subsequent veri cation tests. Brie y, fresh fecal samples from the OTP group collected in the 4 th week as abovementioned were mixed with sterile saline and centrifuged at 3000 g for 5 min. The supernatant was collected and homogenized into 400 mg/mL faecal liquid, and gavage once a day 0.2 mL last 1 week [57]. The remaining mice in the CT and CD groups group were given the same amount of sterile water per day continue to be cultured for 4 weeks until the eighth week under the original culture conditions. After 8 weeks, conduct animal behavior experiments and repeat the above anesthesia operation to euthanize the mice, and then remove the hypothalamus.

Animal Behavioral Tests
The Y-maze spontaneous alternation experiment was mainly used to evaluate the short-term working memory ability of mice. Five mice in each group were randomly selected for maze test. After 1 h of dark adaptation, all mice were put into the center of the Y-maze device and explored freely in the Y-maze for 8 min. Record the arm-in sequence and the total number of arm-in-arms to calculate the correct rate of spontaneous alternation. Entering three different arms in a row was the correct alternate response. The correct rate of spontaneous alternation = [the number of correct spontaneous alternation / (total number of arm advance -2)] × 100% [58].
The MWM test was a classic experiment to test the long-term spatial learning and memory ability of mice, and it was also one of the most widely used tests in behavioral neuroscience [59]. The MWM experiment was completed in a total of 6 days. During the experiment, the water temperature was maintained at 20-25 °C. On the rst day, the water was colorless, the platform was clearly visible and each group of mice swam freely in the maze for 90 s to adapt to the environment. Non-toxic food-grade titanium dioxide was added to the water on day 2 to 6 until the water became opaque and white in order to clearly record the movement of the mice. The escape platform (set in quadrant II) is hidden 0.5-1.0 cm below the water surface, and put the mice into the maze from quadrants , and IV each time so that they can freely search for the platform within 60 s. On the 6 th day, the time and trajectory of each group of mice reaching the platform (set in the II quadrant) were recorded. Then remove the platform, record the swimming time and trajectory of the mouse within 60 s in the quadrant where the original escape platform was located (quadrant II), and calculate the percentage of the mouse's swimming time in the target quadrant. These data were measured by photoelectric sensors and analyzed using a video tracking system connected to a computer (SuperMaze software, Shenzhen Rayward Life Technology Co. Ltd, China).

DNA extractions and Intestinal Microbiota Analysis
DNA extraction and high-throughput sequencing were performed based on the methods we reported earlier, and some of the procedures were modi ed appropriately [14]. DNA from different samples was The combined pair sequences were analyzed using the QIME (Quantitative Insights Into Microot Ecology) (version 1.2.8) software package [60], and high quality reads were collected into the operational classi cation unit for subsequent analyses [61].

Metabolite extraction and Metabolomics data processing
Based on the research method of Qiu et al., [62] metabolites were extracted from the cecal content of CT, CD, and OTP groups for metabolomics analysis. The frozen cecal contents were crushed with liquid N 2 and thawed on ice. The precooled 50% methanol buffer was mixed with the thawed samples at a ratio of 6:1, vortexed for 1min and incubated at room temperature for 10min, followed by overnight storage at -20°C . After centrifugation, the supernatant was transferred to 96-well plates and stored at -80 °C for LC-MS analysis.
The ultra-performance liquid chromatography (UPLC) system (SCIEX, UK) was used for chromatographic separation. The reversed-phase separation adopts ACQUITY UPLC T3 column (100mm*2.1mm, 1.8mm, Waters, UK), the column temperature is 35℃, and the mobile phase and ow rate refer to the recommendations of Want et al. [63]. The eluted metabolites were detected using the TripleTOF 5600 Plus system (SCIEX, Warrington, UK) [64]. XCMS software was used to preprocess the obtained mass spectrometry data. The open databases KEGG and HMDB were used to compare and annotate metabolites. The selected differential metabolites were imported into the pathway analysis module in the

Statistical analysis
The data obtained in this study were analyzed by SPSS (SPSS Inc., Chicago, IL, USA, v 17.0.0). Comparison between groups was performed using one-way analysis of variance (ANOVA) with the post hoc Tukey test for multiple comparisons. P < 0.05 was considered to be statistically signi cant.

Declarations Ethics approval and consent to participate
Animal studies were carried out in the Centre for Laboratory Animals, Ningbo University (Permission No. SYXK [Zhejiang] 2013-0191). All the animal-related operations were strictly in compliance with the relevant laws and regulations on the use and care of laboratory animals in China.
Authors' contributions RY and XZ drafted the manuscript. RY and XZ performed the animal's test and supervised the work. RY, CH, YL, and RZ analyzed the metabolomics data. RY, ZW, and PW performed the 16S rRNA sequencing and analysis. RY, CH, YL, and RZ analyzed the RNA sequencing data. RY and XZ conducted the statistics study. All authors critically revised the manuscript. All authors read and approved the nal manuscript.  The relative abundance of intestinal ora at the phylum (a), families (b) and genus (c) from samples (the relative abundance of intestinal ora of each input sample was analyzed using STAMP). Effects of OTP on metabolomic changes in circadian rhythm disorder mice model. a. HMDB super class identi cation classi cation and annotation diagram of rst-level metabolites; b. volcano plots of the different metabolites in the OTP and CD groups; c. heat map of signi cantly differential metabolites between the CD group and OTP group; d. KEGG (KEGG: Kyoto Encyclopedia of Genes and Genomes) enrichment pathways of signi cantly down-regulated differential metabolites between the CD group and CT group; e. KEGG enrichment pathways of signi cantly up-regulated differential metabolites between the CD group and OTP group. * p < 0.05, **p < 0.01.  The correlation between intestinal microbiota and intestinal differential metabolites. a. correlation heat map between intestinal differential metabolites and microbiota in CD and OTP groups in NIM; b. network diagram of intestinal differential metabolites and microbiota. * p < 0.05, **p < 0.01.

Figure 6
The mRNA levels of Clock, Bmal1, Cry1, Cry2, Per1 and other liver clock genes at different time points of the day.   Effects of OTP intervention on gene enrichment pathway in liver and hypothalamus. (a) GO classi cation analysis in liver between the CD group and OTP group; (b) KEGG pathway analysis in liver between the CD group and OTP group; (c) GO classi cation analysis in hypothalamus between the CD group and VF group; (d) KEGG pathway analysis in liver between the CD group and VF group.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.