Melatonin regulates the neurotransmitter secretion disorder induced by caffeine through the gut-brain axis in zebrafish ( Danio rerio )

Background : Melatonin has been widely used as a "probiotic agent" capable of producing strong neurotransmitter secretion regulatory effects. In the present study, a zebrafish neural hyperactivity model was established using caffeine induction, and the regulation and mechanism of melatonin and probiotic on zebrafish neurotransmitter secretion disorder were explored. Results : Disorders of brain neurotransmitter secretion caused by caffeine, including that of dopamine (DA), γ-aminobutyric acid (γ-GABA), and 5-hydroxytryptamine (5-HT), were improved after interference treatment with melatonin or the probiotic. Metagenomic sequencing demonstrated that the melatonin-treated zebrafish gradually restored their normal intestinal microbial structure, while probiotic supplementation may restructure a new microbiome. Additionally, supplementation with melatonin significantly regulated intestinal microbial functional features. Based on this activity, we supposed that gut microbes play an essential role in the regulation of neurotransmitter secretion disorder via a process that involves melatonin. Germ-free (GF) zebrafish were used to verify the hypotheses. Validation experiment results revealed that the effect on the zebrafish in the GF group could not achieve that on the zebrafish in the melatonin group after adding the same dose of melatonin, and subsequent real-time PCR and metabolic pathway analysis confirmed the conclusion. Meanwhile, the content of acetic acid and propionic acid in the zebrafish gut decreased after caffeine induction and increased significantly after melatonin treatment. Conclusions : In the present research, we identified the potential mechanism of melatonin regulation of neurotransmitter secretion disorder through the gut-brain axis, laying a foundation for exploring the prevention and treatment of some neuropsychiatric disorders by improving the intestinal microbiota. The results indicated that a significant determined. The results showed that the contents of acetic acid and propionic acid decreased in the intestine of tested fish after caffeine induction, while the content increased after melatonin treatment and was higher than that in the fish school without melatonin treatment, corresponding to our previous speculation that melatonin may the

Since melatonin mitigates the effects on multiple neurological diseases, gut-brain axis-related studies have also highlighted the role of the gut microbiota in neuromodulation. We hypothesized that the intestinal microbiota plays an important role in the improvement of neurotransmitter secretion disorders via melatonin. Additionally, we tested whether probiotics, which regulate the structure and metabolism of the intestinal microbiota, have the same effect as melatonin. To address these questions, we constructed a model of neural hyperactivity in zebrafish and modified this disorder by exogenous melatonin and probiotic administration. We utilized metagenomic approach to assess how exogenous melatonin and probiotics improve the neurotransmitter secretion disorder of zebrafish by adjusting the intestinal microbiota to better understand their roles and mechanisms in regulating neurotransmitter secretion disorders. Moreover, germ-free (GF) zebrafish were also used to verify the important role of the gut microbiota in the regulation of neurotransmitter secretion disorders by melatonin. In the present research, we identified the potential mechanism of melatonin regulation of neurotransmitter secretion disorder through the gut-brain axis, laying a foundation for exploring the prevention and treatment of some neuropsychiatric disorders by improving the intestinal microbiota.

Alteration of neurotransmitters in the zebrafish brain
To investigate the effects of different treatments on neurotransmitter secretion in the zebrafish head, the levels of neural hyperactivity-related neurotransmitters in each group were determined (Fig. 1B).
On day 1, the concentration of DA in the caffeine, melatonin and probiotic groups was significantly increased, while the concentrations of γ-GABA decreased, and the concentration of 5-HT did not change significantly compared with those in the control group. The imbalance indicated that caffeine had successfully induced neurotransmitter secretion disorders in the zebrafish brain. After 14 days of equal treatment in different experimental groups, the level of the neurotransmitters in each group changed. In the caffeine group, the contents of DA, γ-GABA, and 5-HT returned to almost normal levels. While the level of DA gradually returned to normal in the melatonin group. The γ-GABA levels in the melatonin and probiotic groups were significantly higher than those in the control group (145.98%, p=0.0456; 154.66%, p=0.0414), and the concentration of 5-HT in the melatonin group was each group after 14 days of treatment for comparative analysis (Fig. 2B), and further screened the differential genus at each time point to observe the variation of specific genus with time in each group ( Fig. 3A and Fig. S3). Based on determining the differences in microbial composition, we further refined the differential genera among the groups on day 1 and day 14 ( Table 1, Table S3 and Table S4). After two-day caffeine exposure, the intestinal microbial community structure of zebrafish was disturbed. Many of the differentially abundant bacterial genera showed a tendency to decrease compared with those in the control group (p<0.05). After 14 days of different treatments, the intestinal microbiota composition of zebrafish in each experimental group changed in different ways (Table S5). There was no significant difference compared with those in the control group, indicating that melatonin supplementation gradually reversed caffeine-induced intestinal microbiota dysbiosis. For the probiotic group, we observed that probiotics introduced differentially abundant genera, and there were 6 differentially abundant genera compared with those in the caffeine group.

The construction of the calm index of microbiota (CIM)
Based on the above results, we found that both melatonin and probiotics regulate the intestinal microbiota. Then, we selected the differentially abundant genera to draw a ternary diagram (Fig. S4). Figure 3B-6D, most of the bacterial genera showed a downward trend compared with those in the control group at baseline after caffeine induction, while a few specific bacteria were enriched in the gut of the zebrafish induced by caffeine (Fig. 3B). On day 14, the microbial community structure of each group showed some recovery, with the different bacteria decreasing (Fig. 3C).

As shown in
To more directly observe the disturbance degree of caffeine interference, melatonin and probiotics supplement on intestinal microbiota, a "calm index of microbiota" was derived based on the distinguishing bacteria between control and treatment groups. 27 differentially abundant bacteria between the control and caffeine-induced zebrafish on day 1 were chosen as markers to construct the CIM day 1 , and the CIM day 1 of each sample was calculated. The results indicated that a significant difference existed in the CIM day 1 between the control and the other three groups (Fig. 3D), and after recalculation on day 14, the CIM day 1 was improved ( Fig. S5). Meanwhile, 11 differentially abundant genera among groups on day 14 were selected to build the CIM day 14 . Based on the CIM day 14 , we observed that the gaps between the melatonin group, probiotic group, and control group were smaller than that in the caffeine group (Fig. 3D). This indicated that melatonin played a further role in promoting the recovery of the intestinal microbiome relative to that in the caffeine group. The formulas were concluded as follow: [Due to technical limitations, the formulas could not be displayed here. Please see the supplementary files section to access the formulas.]

Functional features of the intestinal microbiota in different treatments
To investigate how caffeine interference and the supplementation of melatonin and probiotic affect metabolic pathways by regulating intestinal microbiota, we performed shotgun metagenomic sequencing for samples collected on day 14 (n=20). We annotated the assembled genes and compared with KEGG database to further annotate the microbial metabolic pathways. The Z score greater than 1.6 represent the significant differences. Analysis results after 14 days of equal treatment showed that different treatments significantly altered the metabolism of zebrafish intestinal microbiota ( Fig. 4 and Fig. S6). Compared with normal zebrafish, arginine and proline metabolism (ko00330), glyoxylate and dicarboxylate metabolism (ko00630), propanoate metabolism (ko00640), butanoate metabolism (ko00650) and biotin metabolism (ko00780) were enriched in the caffeine group. Glycolysis / gluconeogenesis (ko00010), citrate cycle (ko00020), fructose and mannose metabolism (ko00051), lipopolysaccharide biosynthesis (ko00540), pyruvate metabolism (ko00620), thiamine metabolism (ko00730) and phosphotransferase system (ko02060) were enriched in the control group. However, the relative abundance of these microbial metabolic pathways in the zebrafish intestines of the melatonin and probiotic groups approached the normal zebrafish in the control group. Collectively, the caffeine induction and melatonin and probiotic supplementation markedly altered the metabolism of the zebrafish gut microbiota.

Verification experiment through a GF zebrafish model
To verify that the gut microbiota plays an essential role in the function of melatonin, a GF zebrafish model was used. Considering that the effects of neurotransmitter regulation were best in the melatonin group, we only retained four groups: the control, caffeine, GF, and melatonin groups. After caffeine interference, the zebrafish showed significant hyperactivity, and the rest time was significantly less than that in the control group (Fig. 5B). However, we did not observe any significant difference in the phenotype including body weight and body length among different groups (Fig. 5A, C). Moreover, after adding the same dose of melatonin, the rest time which could reflect brain activity was different between the GF and melatonin groups (Fig. 5B). Correspondingly, the neurotransmitter secretions γ-GABA, 5-HT, DA and their related synthetic genes were also found to be different To better explain the mechanism by which the intestine microbiota function in the process of melatonin regulation of neurotransmitter secretion disorders, we also determined the contents of metabolites in the gut of zebrafish. Since germ-free zebrafish had almost no intestinal microbiota, while the SCFAs were mainly produced by colonic anaerobic fermentation of undigested carbohydrates, SCFAs were not detected in the GF group here. Results showed that the contents of acetic acid and propionic acid in the intestine of zebrafish in the caffeine and melatonin groups decreased significantly after the treatment with caffeine, while the contents increased remarkably after melatonin treatment, resulting in levels that were obviously higher than those in the caffeine group ( Fig. 6D, E).

Melatonin regulated neurotransmitter secretion disorders through the gut-brain axis
Through the verification experiment, we highlighted the importance of intestinal microbiota in the melatonin-induced neurotransmitter regulation effects. Based on the Spearman′s rank correlation coefficient, we constructed a network of the correlations among melatonin, genera, metabolic pathways, SCFAs, and neurotransmitters. As shown in Fig. 7A, supplementation with melatonin inhibited the growth of Shewanella, Deefgea and Enterobacteriaceae in the zebrafish gut, which were negatively correlated with propanoate metabolism, butanoate metabolism and phosphotransferase system, whereas a negative correlation was observed between Aeromonadaceae and Rhizobiales and melatonin. Additionally, it was found that melatonin supplementation elevated the levels of acetic acid and propionic acid, representing SCFAs, in the intestine, and the generated SCFAs further stimulated the secretion of neurotransmitters in the brain of zebrafish, accompanied by decreased levels of DA and elevated levels of γ-GABA and 5-HT, which revealed the potential mechanism by which melatonin regulates neurotransmitter secretion disorders through the gut-brain axis (Fig. 7B).

Discussion
Melatonin, also known as the "darkness hormone", is essential for calming nerves and sleep regulation [14] . It has been reported that exogenous melatonin has a positive sleep-promoting effect on diurnal vertebrates, including humans [6] . Mounting evidence has also indicated that intestinal microorganisms affect brain function and psychological state of the host through the gut-brain axis.
However, it is not clear whether and how melatonin regulates the neural hyperactivity state of the brain through the intestinal microbiota. Meanwhile, probiotics, as micro-ecological modulators of the intestine, can also have a major influence on the nervous system [15,16] , and whether they can regulate the neural excitability of the organism through the gut-brain axis is also worth further exploration. Herein, a zebrafish neural hyperactivity model was established to explore the regulatory effect and mechanism of melatonin and Lactobacillus plantarum HNU082 on neurotransmitter secretion disorder.
There is growing evidence suggesting that the gut microbiota produce and release many similar neurotransmitters to help regulate mood and promote sleep, while possibly affecting the CNS through the gut-brain axis [9,17] . In addition to the variation in the levels of neurotransmitters, the structure of the zebrafish intestinal microbiota is disordered after caffeine induction, potentially indicating that the imbalance of the intestinal microbiota may be related to the disorder of neurotransmitters. After 14 days of equal treatment, the intestinal microbiota in the various groups gradually recovered, among which the similarity of the melatonin group and control group microbiota was the highest, while the caffeine group microbiota still had a large difference. Moreover, we found that Cetobacterium, the dominant genus in fish intestine [18] which plays a crucial role in the synthesis of vitamin B12 and fermentation of carbohydrates and peptides [19,20] , was decreased significantly after caffeine induction and that self-healing and melatonin and probiotic supplementation failed to increase its content significantly. The bacterium Shewanella, which can restore only the insulin-secreting β cells to the average level in sterile zebrafish [21] , also was significantly decreased after neural hyperactivity, while the content of Shewanella in the melatonin group increased compared with that in the selfhealing caffeine group. Overall, melatonin supplementation may restore the original intestinal microbiota structure to a great extent, while the probiotic may tend to shape a new intestinal microbiota balance by introducing additional changes in the gut-specific genera. Previous studies have reported that melatonin supplementation can significantly affect the metabolism of the intestinal microbiota, including amino acid metabolism and drug metabolism, while altering the composition of the intestinal microbiota in weaned mice [22] . We also observed that amino acid metabolism, carbohydrate metabolism, cofactor and vitamin metabolism, and membrane transport changed significantly in different groups. Current research has indicated that SCFAs are implicated in a range of neuropsychiatric disorders, with the ability to cross the blood-brain barrier, and are critical substances that connect the intestine to the central nervous system [23] . The metabolites of SCFAs, such as propanoate and butanoate, were found to be stimulated during self-healing to restore the neurotransmitter secretion levels in the caffeine group. Previous studies have shown that SCFAs can promote the production of colonic 5-HT through their effects on enterochromaffin cells [24] , especially propanoate, which has been shown to alter the DA, 5-HT and GABA systems [25] . Furthermore, 5-HT activates the vagus nerve. Additionally, 5-HT is the precursor of melatonin, which may be a means to connect the brain to the intestinal microbes. Nevertheless, the SCFA metabolism in the melatonin and probiotic groups was significantly lower than that in the caffeine group. We hypothesized that supplementation might promote the production of SCFAs during neural hyperactivity regulation, which accelerates the recovery of neurotransmitter secretion in the brain and gradually regulates the hyperactive state. Therefore, the metabolism of SCFAs gradually tended towards normal in the melatonin and probiotic groups.
To verify whether gut microbes play a real role in the recovery of brain neurotransmitter secretion, germ-free animals were used, this is a valid and frequently used in vivo experimental model for studies on the host gut microbiota in health and various diseases, such as cancer [26] , diabetes [27] , and cardiovascular [28] and metabolic diseases [29] . Mice, zebrafish, dogs, chickens, pigs, and many types of other germ-free models have been appropriately used in various research studies to verify the link between the intestinal microbiota and some diseases [30−34] . As the importance of gut-brain axis in brain development has gradually been confirmed, germ-free animals have gradually become an accepted model for gut-brain axis studies, which has become an important research means and valuable experimental tool for the study of the intestinal microbiota in host behaviour and various brain diseases [35−38] . In our validation experiment, the recovery of neurotransmitters in the GF group did not reach the level as in the normal zebrafish after adding the same dose of melatonin, indicating the key role of the intestinal microbiota in the regulation process of neurotransmitter secretion by melatonin, which subsequent real-time PCR again confirmed. To better explain the mechanism by which the intestinal microbiota plays a role in the regulation of neurotransmitter secretion, the content of SCFAs in the gut of zebrafish was determined. The results showed that the contents of acetic acid and propionic acid decreased in the intestine of tested fish after caffeine induction, while the content increased after melatonin treatment and was higher than that in the fish school without melatonin treatment, corresponding to our previous speculation that melatonin may promote the secretion of neurotransmitters in the brain by affecting intestinal microbes and producing SCFAs.
In addition, we constructed the CIM for zebrafish based on a variety of specific genera in the zebrafish gut to quantify the degree of the effect of caffeine induction and subsequent melatonin and probiotic supplementation on the host gut microbiome which impacted the emotional state of zebrafish. Guo et al. [39] have examined the gut microbiome between gout patients and healthy people and then built a microbial gout index based on specific bacteria for the diagnosis of gout. Microbial indicators are currently widely used in intestinal research. Herein, a significant gap was found in CIM day 1 between normal zebrafish and caffeine-treated zebrafish on day 1, but the index showed a significant improvement after 14 days of equal treatment. Moreover, it can also be seen from CIM day 14 that melatonin and the probiotic played a specific role in promoting the recovery of the intestinal microbiota of zebrafish relative to that of the caffeine group. The CIM established in this study may provide a new means for analysing neural hyperactivity disorder.
Collectively, melatonin supplementation gradually restored the normal homeostasis and metabolism of the intestine and promoted the production of SCFAs in the intestine by regulating the gut microbiota to accelerate the recovery of the neurotransmitter secretion level through the gut-brain axis. The present research revealed the potential mechanism by which melatonin regulates neurotransmitter secretion disorder through the gut-brain axis, supporting future development of psychotropic drugs targeting the intestinal microbiota.

Main experimental design
In experiment 1, fish were equally distributed into 24 glass tanks (15 fish/tank) after 14 days of acclimation, each containing 5 L of water. Then, tanks were randomly assigned to four groups to conduct the experiment (Fig. 1A): (1) control group: normal feed, regular filtered water; (2) caffeine group: 200 μM caffeine was added to the water for two consecutive days to alter zebrafish activity.
This concentration of caffeine was enough to cause neural hyperactivity disorders in zebrafish.
Thereafter the zebrafish were maintained conventionally with standard feed and regular filter water; (3) melatonin group: 200 μM caffeine was used for two consecutive days to change zebrafish activity, and then the fish were fed regularly, but the water was replaced with filtered water containing 1 μM melatonin (melatonin powder was dissolved in anhydrous ethanol to 100 mM and eventually diluted to a final concentration of 1 μM); and (4) probiotic group: zebrafish were treated with 200 μM caffeine for two consecutive days to change the fish activity. Then they were maintained in regular filtered water, and fed with normal fish feed containing 108 CFU of Lactobacillus plantarum HNU082/g of feed every day. Unconsumed food, faeces, and dead fish were removed in a timely manner, and the corresponding fresh filtered water was replaced in each tank every day; the water for the melatonin group was replaced with filtered water containing 1 μM melatonin to maintain the constant melatonin concentration, while the water for the other three groups was replaced with regular filtered water.

Validation experimental design
To verify the critical role of the intestinal microbiota in the secretion of related neurotransmitters, germ-free (GF) zebrafish were also studied (Fig. 1A). The zebrafish were purchased from the China zebrafish resource centre, and the GF zebrafish model was established as previously reported [40].
Considering that the effects of neurotransmitter regulation were best in the melatonin treated fish, we used only 4 groups, the control, caffeine, GF and melatonin groups. The rearing and treatment conditions of the control, caffeine, and melatonin groups were the same as those in the experiment above. Two-day caffeine induction was conducted, and melatonin supplementation was performed for 14 days, as in the other treatment groups. Water and feed were sterilized by irradiation with ultraviolet radiation in advance. On day 1 and day 14 after two-day caffeine induction in experiment 2, the mean static time of the fish school in five minutes under night-time dark conditions in the control, caffeine, melatonin, and GF groups were observed and recorded.

Zebrafish maintenance and feeding
Adult zebrafish (Danio rerio) obtained from the China zebrafish resource centre were maintained in the aquatic experimental animal facility of the College of Food Science and Engineering, Hainan University. The experimental and animal care procedures were conducted following the recommendations of the Ethics Committee of Hainan University. All fish were housed for 14 days in 50 L tanks supplied with aerated and dechlorinated tap water to acclimate to the laboratory conditions before experiments. The acclimation period and experiments were performed at a temperature of 28±0.5°C and with a 14 h light/10 h dark cycle according to the standards of zebrafish care [41].
Commercial fodder was provided once daily at a fixed time, which was administered by 3% of the total weight of the fish in each tank. Dead zebrafish were removed promptly, and the cumulative mortality did not exceed 2% throughout the acclimation period. For the GF group, whole tanks were kept in the clean bench ventilated with filtered sterile air to maintain germ-free conditions until the end of the experiment. Sterile food was performed through ultraviolet sterilization. Water was treated by filter-sterilized system and irradiated overnight with ultraviolet in the clean bench to keep sterilize, and ensured that the temperature was maintained at room temperature during morning transfer (Pham et al., 2008).

Sample collection for 16S rRNA and shotgun metagenomic sequencing and SCFAs determination
Several zebrafish were randomly selected from each group on day 1, 3, 7 and day 14 following the two-day caffeine induction. Fish were euthanized promptly by an ice water bath, and dissection was conducted vice the anus with an aseptic scalpel after the body was sterilized with 75% alcohol. The entire intestine from the esophagus to the anus was collected, and three fish were pooled as a sample, taking a total of 6 replicates in each group for DNA extraction and a total of 3 replicates in each group for short-chain fatty acid (SCFA) determination. After centrifugation at 12000 rpm for 20 min at 4°C, the head supernatants were subsequently collected and stored at -80°C for ELISA. In the verification experiment, another 3 heads were collected as a sample, with a total of 3 replicates in each group, to conduct gene expression analysis by real-time PCR.

Enzyme-linked immunosorbent assay (ELISA) and analysis of intestinal fatty acid metabolites by GC-MS
Genomic DNA of zebrafish intestine was extracted using a QIAamp® DNA Stool Mini Kit (Qiagen, Hilden, Germany) following the manufacturer′s instructions. The quality of all isolated DNA was assessed by A260/A280 using a NanoDrop instrument. Qualified DNA was preserved at -20°C for subsequent analysis. The PCR amplification procedure was conducted as previously described [46]. A set of 6-nucleotide barcodes was added to the specific primers 338F and 806R to amplify the 16S rRNA V3-V4 region of the genomic DNA. An Agilent DNA 1000 kit combined with an Agilent 2100 Biologic Analyzer (Agilent Technologies, USA) was used to quantify the PCR products. Sequencing was performed using the Illumina MiSeq high-throughput sequencing platform by Shanghai Personal-Bio Corporation.

Shotgun metagenomic sequencing and quality control
The zebrafish samples on day 14 (n=20, 5 samples each group) were subjected to shotgun metagenomic sequencing by using an Illumina HiSeq 2500 instrument. Libraries were prepared with a fragment length of approximately 300 bp. Paired-end reads were generated using 100 bp in the forward and reverse directions. The reads were trimmed using Sickle and were subsequently aligned to the human genome to remove the host DNA fragments.

Functional annotation and metabolic pathway analysis
The annotated amino acid sequences were aligned against the Kyoto Encyclopaedia of Genes and Genomes (KEGG) databases using BLASTp (e-value≤1e-5 with a bit-score higher than 60). The annotated sequences were assigned to the KEGG orthologue group (KO) according to the highest score. Reporter Z-scores were calculated to reveal the differences in enriched metabolic pathways between the control and PCOS groups, as previously described. Accordingly, a reporter score of >2.3 (90% confidence according to the normal distribution) was used as a detection threshold to significantly differentiate between pathways [47].

Bioinformatic analyses and statistical analyses for figure construction
A quality-control procedure was performed for the raw pair-end reads; bioinformatics analysis of highquality sequence data was conducted after removal of primer and barcode sequences by using the QIIME platform [48]. PyNAST [49] was used to align the trimmed sequences, and those under 100% sequence identity were clustered using UCLUST [50] to obtain a unique V3-V4 sequence set. On this basis, operational taxonomic units (OTUs) were clustered at a threshold level of 97% sequence identity, and the sequences with the highest frequency were selected as the representative sequences of OTUs. After ChimeraSlayer [51] was employed to remove the potentially chimeric sequences in the representative set of OTUs, a representative sequence of each OTU was assigned to a taxonomic level by Ribosomal Database Project (RDP) [52]. A phylogenetic tree of OTUs was constructed by using Fast Tree [53], and the alpha and beta diversity of bacterial communities were analysed on this basis. The sequence data reported in this paper have been deposited in the NCBI database (metagenomic data: PRJNA543612).

Statistical analyses and figure construction
Relatedness analysis was conducted with R Studio. Boxplot based on weighted and unweighted UniFrac distance were established using the ggplot2 package. Principal component analysis (PCA) was carried out by the ggord package, and the pheatmap package was used for cluster analysis and heatmap construction. Correlation analysis was calculated by using Spearman′s rank correlation coefficient and visualized as a network by Cytoscape. Statistical significance was analysed with Mann-Whitney test or T-test. The false discovery rate (FDR) based on the Benjamini-Yekutieli method [54] was used to control the multiple testing error rate. Data are presented as the means ± SEMs; a twosided test was conducted, and p < 0.05 was considered statistically significant. All statistical analyses were performed by GraphPad Prism, version 8.      which were compared with the KEGG database to further explain the microbial metabolic pathways. Z scores greater than 1.6 were selected as significant differences.   represents melatonin, blue nodes represent genera, orange nodes represent metabolic pathways, green nodes represent SCFAs, and yellow nodes represent neurotransmitters.
The node size is proportional to the mean abundance in the respective population. (B) Summary diagram of the mechanism of melatonin regulation of neurotransmitter secretion disorders through the gut-brain axis.

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