Wild-type zebrafish (Danio rerio; AB strain, Zebrafish International Research Center, Eugene, OR, United States) were raised in a flow-through system at 28 ± 0.5°C under a 14/10 h light/dark cycle at the Zebrafish Drug Screening Center, Mie University, Japan. Ten AB wild type 6 months-old zebrafish were transferred to 2 L of rearing water at 28°C. The fish were fed at 9:00 a.m. and 5:00 p.m. on weekdays. For 1 month, P. sabiae cells (1.0 × 10⁹ cell/L) were administered in the rearing water twice a day at feeding times. Paraburkholderia sabiae was purchased from the German Collection of Microorganisms and details for the culture is described in Supplementary online materials. The experimental procedures were performed in accordance with the Japanese Animal Welfare Regulatory Practice Act on Welfare and Management of Animals (Ministry of the Environment of Japan) and in compliance with ARRIVE guidelines¹. Ethical approval from the local Institutional Animal Care and Use Committee was not sought, because this law does not mandate the protection of zebrafish.

Anxiety in zebrafish was evaluated using the novel tank diving test (Haghani et al., 2019; Anwer et al., 2021) from 14:00 p.m. to 15:30 p.m. An observation tank (149 mm × 38 mm × 100 mm) was filled with rearing water at a height of 10 cm. The zebrafish were transferred from the rearing tank to a restraint tank (37 mm × 7 mm × 65 mm) with a net. The zebrafish were placed in a restraint tank for 5 min to cause anxiety and transferred to the observation tank. When the zebrafish were transferred to the observation tank, their movements were immediately captured as videos using an iPad Pro for 2 min. The average speed, acceleration, total travel distance, and exploration rate in the observation tank were analyzed using the ToxTrac software (Rodriguez et al., 2018). Supplementary Figure S1 shows a schematic of the evaluation of zebrafish anxiety.

DNA was extracted from the zebrafish intestine using a Quick-DNA Fecal/Soil Microbe Miniprep Kit (Zymo Research, Irvine, CA, United States). The 16S rRNA V3/V4 region in the DNA was amplified by PCR, and Library preparation and sequencing were performed using the MiSeq Reagent Kit v3 (Illumina, San Diego, CA, USA) at 2 × 300 bp. The microbiome data were analyzed by phylogenetic investigation of communities by reconstruction of unobserved states 2 (PICRUSt2) ver. 2.3.0 b 32 (Douglas et al., 2020). Linear discriminant analysis effect size (LEfSe; Segata et al., 2011). For details, please see the Supplementary online materials.

Zebrafish brains were thoroughly homogenized using a pestle in 100 μl PBS cooled on ice. The supernatant was collected by centrifugation at 14,500 rpm for 2 min. One microliter of the 10-fold diluted supernatant was used to measure taurine and serotonin concentrations using the Taurine Assay Kit (Cell Biolabs, San Diego, CA, USA) and Serotonin ELISA Kit (Immusmol, Bordeaux, France), respectively, according to the manufacturers’ protocols.

Six months-old zebrafish were exposed to P. sabiae bacterial cells for 1 month, and their anxiety-like behavior was evaluated using a novel tank diving test, which has become a standard test for determining anxiety-like behavior in zebrafish (Haghani et al., 2019; Supplementary Figure S1; Supplementary movies S1, S2). The average speed, average acceleration, and total distance tended to be higher in P. sabiae-exposed zebrafish. The average speeds were 45.1 mm/s and 60.8 mm/s, the average accelerations were 137.8 mm/s² and 190.0 mm/s², and the total distances were 5,566 mm and 7,501 mm in the control and the P. sabiae-exposed zebrafish, respectively (Figure 1A). Furthermore, P. sabiae-exposed zebrafish migrated significantly more extensively in the observation tank. The exploration rate in the observation tank was 32% for the control zebrafish, whereas that for the P. sabiae-exposed zebrafish was 49% (p < 0.05; Figures 1A,B). There was no significant difference in exploration rate between the genders (Supplementary Figure S2). The response of adult zebrafish in the novel tank environment, that is, to initially stay at the bottom of the tank and eventually acclimating to the rest of the tank, is interpreted as a precautionary anti-predator response and anxiety reduction, respectively (Haghani et al., 2019). Our results indicated that P. sabiae administration reduced anxiety in zebrafish.

Multiple behaviors, including stress and anxiety, are modulated by neuropeptides (Rana et al., 2022). The altered expression of bdnf in distinct brain areas and BDNF signaling associated with social behavior have been frequently demonstrated in gut-brain axis models in mice (Bercik et al., 2010; Diaz Heijtz et al., 2011; Desbonnet et al., 2015). We also evaluated the expression levels of anxiety-related genes, neuropeptide Y npy (Shiozaki et al., 2020), isotocin (oxtI) (Godwin and Thompson, 2012), and bdnf (Bjorkholm and Monteggia, 2016). oxti expression tended to increase in the brains of P. sabiae-exposed fish, whereas npy expression decreased (no significance, Supplementary Figure S3). bdnf expression also tended to increase in the P. sabiae-exposed fish brain, similar to the results of a previous zebrafish study (Borrelli et al., 2016).

The gut microbiomes of the five P. sabiae-exposed zebrafish were analyzed using the 16S rRNA V3/V4 region of their gut microbial DNAs. The 30,000–50,000 pair-reads, 93.1–94.1 Q20 values, and 83.5–85.3 Q30 values were obtained for each sample. A total of 450 OTUs were generated from the analysis using Qiime2 (Supplementary Table S2). Despite the continuous feeding of the bacterial cells to the zebrafish for 1 month, the observed OTUs in the P. sabiae-exposed zebrafish gut microbiome tended to be higher than those in the control zebrafish. Chao1, the richness index of alpha diversity in the gut microbiome, was significantly higher in the P. sabiae-exposed zebrafish than in the control zebrafish (p < 0.05; Figure 2A). No significant difference in the evenness index of the Shannon diversity index and the indicator of phylogenetic diversity, Faith_pd, was observed between these zebrafish gut microbiomes. P. sabiae was not detected in the gut microbiome analysis (Supplementary Table S2); thus, P. sabiae functioned without colonizing the zebrafish gut.

LEfSe analysis was used to compare the bacterial populations between the P. sabiae-exposed and control zebrafish gut microbiomes (Supplementary Table S3; Segata et al., 2011). At the phylum level, Fusobacteria and Proteobacteria were dominant in the zebrafish gut microbiome, which is consistent with previous reports (Borrelli et al., 2016; Davis et al., 2016). Actinobacteria and Chloroflexi populations were smaller in the gut microbiome of P. sabiae-exposed zebrafish than in that of control zebrafish (Figure 2B). At the genus level, Cetobacterium was detected as the major bacterium in the gut microbiome of zebrafish. The populations of Actinomycetales, including Nocardiaceae (p < 0.01), Nocardia (p < 0.01), Gordoniaceae (p < 0.01), Gordonia (p < 0.01), Nakamurellaceae (p < 0.01), and Aeromonadaceae (p < 0.05) were significantly reduced, whereas those of Rhizobiales, including Xanthobacteraceae (p < 0.05) and Bradyrhizobiaceae (p < 0.05), Rhodospirillaceae (p < 0.01), and Pirellulaceae (p < 0.01) were significantly increased in the P. sabiae group (Figure 2C; Supplementary Table S3).

Functional composition tables based on EC number, Kyoto Encyclopedia of Genes and Genomes (KEGG) ortholog, cluster of orthologous groups, and METACYC were generated using PICRUSt2 (Supplementary Tables S4–S7; Douglas et al., 2020). Functional analysis based on METACYC revealed that the mevalonate pathway (PWY-6174), superpathway of polyamine biosynthesis (PWY-6565), and chitin derivative degradation (PWY-6906) were detected only in the P. sabiae-exposed zebrafish gut microbiome. In contrast, the functions of the superpathways of taurine degradation (PWY-1541) (p < 0.01), creatinine degradation (PWY-4722) (p < 0.01), and aerobic toluene degradation via catechol (PWY-5178) (p < 0.001) were significantly reduced in the P. sabiae-exposed zebrafish gut microbiome (Supplementary Table S7). Regarding the neurotransmitter taurine, glutamate decarboxylase [EC:4.1.1.15] (p < 0.01; Supplementary Table S4) and the functions of tauY: taurine dehydrogenase large subunit [EC:1.4.2.-] (K07256) (p < 0.05), tauA: taurine transport system substrate-binding protein (K07256) (p < 0.05; Supplementary Table S5), ABC-type taurine transport system, periplasmic component (COG4521; p < 0.05), and ATPase component (COG4525) (p < 0.01; Supplementary Table S6) were also reduced in the P. sabiae-exposed zebrafish gut microbiome. The reduced functions in the metabolic pathways of taurine and hypotaurine were represented by the KEGG Mapper (Kanehisa et al., 2022; Figure 2D).

Because PICRUSt2 predicted downregulation of the superpathway of taurine degradation (PWY-1541) in the zebrafish gut microbiome, we next evaluated taurine levels in zebrafish. The concentration of taurine (2-aminoethanesulfonic acid) in the P. sabiae group was significantly (p < 0.01) higher than that in the control zebrafish brain:32.2 and 10.4 nmol/mg-brain, respectively (Figure 3A). We also measured the taurine level of intestinal contents (microbiota) and blood, and found that P. sabiae administration shows a tendency (p < 0.2) to increase taurine in the intestinal contents (Supplementary Figure S4). The expression level of the taurine biosynthesis gene encoding cysteine dioxygenase cdo, was 2.0 times higher than that in the P. sabiae-exposed zebrafish brains, compared with that in the control zebrafish brains (p < 0.1), while no difference was observed in those of the other taurine biosynthesis genes encoding cysteamine dioxygenase ado, cysteine sulfinic acid decarboxylase csad, and taurine transporter tauT (Figures 3B,C).

In addition to taurine, serotonin (5-hydroxytryptamine) and GABA are also well-known key players in anxiety in human and zebrafish (Davis et al., 2016; Marcinkiewcz et al., 2016). Several probiotic treatments affect serotonin-related gene expressions in zebrafish (Borrelli et al., 2016; Davis et al., 2016). The expression level of the serotonergic gene tph2 was significantly increased in the P. sabiae group, however that of the other serotonergic genes slc4a6a, slc4a6b, htr1aa, mao, tph1a, and tph1b were not significantly different between these zebrafish brains (Supplementary Figure S5). No significant difference of the serotonin concentrations was observed in these zebrafish brains: (6.0 and 6.2 pmol/mg-brain), according not to be predicted by PICRUSt2. For GABAergic genes, the expression levels of gabra1 and gad1 were not affected by P. sabiae administration (Supplementary Figure S6).

Paraburkholderia is a new genus in gram-negative β-proteobacteria, which can be distinguished from Burkholderia (Sawana et al., 2014). Paraburkholderia is often isolated from the rhizosphere and plant organs, such as root nodules, and it functions in nitrogen fixation and plant growth promotion (Esmaeel et al., 2018). Furthermore, Paraburkholderia has been detected in guts, ranging from those of insects to humans (Takeshita and Kikuchi, 2020; Chen et al., 2022; Tang et al., 2022). The presence of Burkholderia-Caballeronia-Paraburkholderia in human gut microbiomes remarkably correlated with body mass index, osteoarthritis, and cholangiocarcinoma (Zhang et al., 2021; Ning et al., 2022). Opisthorchis viverrini infection is a key risk factor for cholangiocarcinoma and increases Paraburkholderia in the gut microbiomes of hamsters (Haonon et al., 2021). The abundance of Paraburkholderia has also been discussed with regards to ulcerative colitis (Liu et al., 2021) and breast cancer-related fatigue (Lv et al., 2022).

In this study, we focused on the function of Paraburkholderia and demonstrated that zebrafish exposed to the Paraburkholderia type strain, P. sabiae DSM 23623 (Chen et al., 2008), moved more extensively in the novel tank diving test (Figure 1), suggesting the possibility that the P. sabiae-exposed treatment reduced anxiety in zebrafish. This is the first report to describe the function of Paraburkholderia in animal behavior.

A loss in the diversity of the gut microbiome and dysbiosis caused by diet, infection, stress, etc., leads to abnormalities in brain function and mental health (Rogers et al., 2016). Antibiotic dysbiosis is also associated with an elevated risk of anxiety (Lurie et al., 2015). Borrelli et al. (2016) and Davis et al. (2016) reported that bacterial administration modulates gut microbial communities in zebrafish. Here, we show that P. sabiae administration increased the richness of zebrafish gut microbiomes, as shown by the high values of the diversity index, Chao1 (Figure 2A). In addition to the higher richness, we observed a reduction of Actinobacteria in the gut microbiome in the P. sabiae-exposed zebrafish, which represented reduced anxiety-like behavior (Figures 1, 2), which was similar to the case of L. rhamnosus administration (Borrelli et al., 2016). Paraburkholderia exposure treatment unexpectedly worked as a paraprobiotic (Tsilingiri and Rescigno, 2013), because P. sabiae did not colonize zebrafish guts (Supplementary Table S2).

Metagenomic profiling using PICRUSt revealed functional alterations in the KEGG pathway involved in taurine metabolism (Figure 2B; Supplementary Table S7), and we demonstrated that taurine concentration was significantly (p < 0.05) increased in the P. sabiae-exposed zebrafish brain (Figure 3A). The presence of taurine in the colon depends on the gut microbiome, and it is transported from the colon into the systemic circulation (Matsumoto et al., 2017; Sharon et al., 2019; Li et al., 2020), including the central nervous system, through the blood–brain-barrier via the taurine transporter TauT (Benrabh et al., 1995). In addition, taurine is synthesized in the brain neurons and astrocytes (Vitvitsky et al., 2011). In this study, P. sabiae exposure upregulated the taurine synthesis gene, cdo, in zebrafish brains (Figure 3C), suggesting the existence of indirect mechanisms induced by P. sabiae ingredients, which activate taurine synthesis in the brain.

Taurine, a small organic compound, is a major constituent of bile that exists mainly in the large intestine. It possesses many biological functions, including osmoregulation, membrane stability, intracellular calcium metabolism, anti-oxidation, and anti-inflammation in various tissues (Marcinkiewicz and Kontny, 2014; Menzie et al., 2014). In vertebrates, taurine is thought to have anxiolytic properties in vertebrates (Zhang and Kim, 2007). For example, mouse administered 0.5 mmol/kg taurine for 7 days showed anxiolytic-like properties in the plus-maze test (Chen et al., 2004). The taurine-deficient mice without CSAD exhibited increased anxiety-like behavior in the plus-maze test, which was compensated by the oral treatment with 0.2% taurine (Park et al., 2019). In our results (Figure 3), taurine concentration was significantly (p < 0.05) higher in the P. sabiae-exposed zebrafish brains, suggesting that the increased level of taurine is the cause of reduced anxiety in the P. sabiae-exposed zebrafish. Consistent with our results, Mezzomo et al. reported that treatment with taurine (3.2 mmol/L) decreased anxiety-like behaviors induced by acute exposure to conspecific alarm substances in zebrafish (Mezzomo et al., 2016, 2019). Interestingly, Mezzomo et al. detected anxiolytic effects only in the light and dark test, not in the novel tank test as we did in the present study, in the case of acute taurine administration. In other words, the light and dark test may also provide greater improvement in the P. sabiae group.

Thus, additional research into the molecular mechanisms underlying the increase in the zebrafish gut microbiome diversity, along with the increase in taurine concentration in the zebrafish brain, following P. sabiae administration is required, and future studies will aim to explore the same.