A total of 26 bacterial strains were used in this study ( Supplementary Table 1 ). They were acquired from the German Collection of Microorganisms and Cell Culture GmbH (DSMZ, Braunschweig, Germany) and the American Type Culture Collection (ATCC, Manassas, Virginia, USA) and obtained from our own collection. The strains were stored at −80°C in 25% (v/v) anaerobic glycerol stocks and routinely cultivated at 37°C, without shaking, in Hungate tubes containing anaerobic modified yeast extract casitone and fatty acid medium (mYCFA; O₂-free CO₂ head gas) consisting of (per liter): 10 g Amicase (Sigma-Aldrich, St. Louis, MO, USA), 2.5 g yeast extract (Sigma-Aldrich), 1.5 g meat extract (Sigma-Aldrich), 4 g NaHCO₃, 0.45 g K₂HPO₄, 0.45 g KH₂PO₄, 0.9 g NaCl, 0.9 g (NH₄)₂SO₄, 90 mg MgSO₄, 90 mg CaCl₂, 10 mg hemin, 1 g L-cysteine·HCl, 10 μg biotin, 10 μg cobalamin, 30 μg p-aminobenzoic acid, 50 μg folic acid, 150 μg pyridoxamine, and 5.75 mL of volatile fatty acid solution (Bircher et al., 2018). Except when otherwise stated, the medium was supplemented with glucose, cellobiose, and soluble starch (2 g/L each) as carbon sources, and the pH was adjusted to 6.5. When appropriate, mYCFA was also supplemented with L-tryptophan (Trp; from 5 to 24.5 mM). The procedures for anaerobic medium preparation in Hungate tubes were performed as previously described (Bircher et al., 2018).

A total of 13 strains ( Supplementary Table 1 ), closely related to known tryptamine producers or previously identified in silico as potential tryptamine producers (Williams et al., 2014), were reactivated from glycerol stocks by a 1.25% (v/v) inoculation of anaerobic mYCFA. Pre-cultures were incubated for 24 h at 37°C, and screening was initiated by inoculating 1.25% (v/v) of pre-cultures in 8 mL of fresh mYCFA containing 24.5 mM Trp. After 24 and 48 h of incubation at 37°C, 1 mL of bacterial cultures was centrifuged (14,000 × g, 10 min, 4°C), and the supernatants were directly processed for tryptamine (Try) quantification (described below). Each strain was screened using three independent replicates.

To identify growth phases associated with Try production, R. gnavus DSM 108212, B. hansenii DSM 20583, C. sporogenes DSM 795, and C. nexile DSM 1787 were reactivated in mYCFA (1.25%, v/v) for 24 h at 37°C. The pre-cultures were subsequently transferred (1.25%, v/v) into mYCFA, grown to an OD₆₀₀ of ~0.3, and re-inoculated (1.25%, v/v) into mYCFA containing 15 mM tryptophan to initiate the kinetic analysis. OD₆₀₀ and extracellular Try concentrations (supernatants: 14,000 × g, 10 min, 4°C) were monitored throughout 48 h of growth at 37°C. Each strain was evaluated using three independent replicates.

To investigate the impact of Trp concentration on Try production, R. gnavus DSM 108212, B. hansenii DSM 20583, and C. sporogenes DSM 795 were reactivated in mYCFA (1.25%, v/v) for 24 h at 37°C. The analysis was then initiated by transferring the pre-cultures (1.25%, v/v) into mYCFA containing 10 or 24 mM Trp at pH 6.5. OD₆₀₀ and extracellular Try concentrations (supernatants: 14,000 × g, 10 min, 4°C) were monitored after 6, 24, and 48 h of growth at 37°C. Each strain was evaluated using three independent replicates.

The effect of mild acid stress on the induction of Try production was first evaluated with actively growing cells of R. gnavus ATCC 29149. The cells were reactivated in mYCFA (1.25%, v/v) for 24 h at 37°C. The pre-cultures were then transferred (1.25%, v/v) into mYCFA containing 5 mM Trp and grown to an OD₆₀₀ of ~0.2, after which the pH was reduced to pH 5.5 and 5.0 using 2.5 M HCl anaerobic solution. The cells were incubated for a total of 24 h at 37°C, and the Try concentration in supernatants (14,000 × g, 10 min, 4°C) was evaluated 2 and 18 h after stress induction. Each condition was tested in three independent replicates.

Subsequently, the effect of acid stress on Try production and survival was tested in resting cells of R. gnavus DSM 108212 and C. sporogenes DSM 795, as previously described (Otaru et al., 2021). Briefly, after reactivation in mYCFA (1.25%, v/v) for 24 h at 37°C, the cells were transferred (1.25%, v/v) into mYCFA and grown for 15 h to reach the late exponential phase (OD₆₀₀ ~1.0). In an anaerobic chamber (10% CO₂, 5% H₂, and 85% N₂, Coy Laboratory Products, Inc., Ann Arbor, MI, USA), 1-mL bacterial cultures were harvested, centrifuged (6000 × g, 8 min, room temperature), and washed with sterile 0.9% NaCl solution. Acid stress was induced by resuspending the cell pellets in 1 mL carbohydrate-depleted mYCFA with or without 5 mM Trp at three different pH conditions each, i.e., 6.7, 5.4, and 4.3, followed by anaerobic incubation at 37°C. Cell viability, as well as Try and Trp concentrations (supernatants: 14,000 × g, 10 min, 4°C), was measured after 1 h of incubation. Each condition was tested in three independent replicates.

Cell viability was determined as described before (Otaru et al., 2021), by membrane integrity analysis using flow cytometry with two different dyes, i.e., the cell-impermeable molecule propidium iodide (PI, Life Technologies, Zug, Switzerland) and the cell-permeable SYBR Green I stain (SG, Life Technologies). Briefly, the bacterial cultures were diluted with phosphate-buffered saline (PBS) solution to achieve an approximate concentration of 10⁷ cells/mL. The resulting dilutions (30 μL) were then mixed with 267 μL PBS and 3 μL SG (1× concentration) to assess the total cell counts or with 3 μL SGPI solution (SG: 1× concentration; PI: 4 μM) to assess the intact cells. The mixtures were incubated for 30 min at room temperature in the dark. Samples were analyzed with a Cytomics FC 500 (Beckman Coulter GmbH, Krefeld, Germany) equipped with an air-cooled argon ion laser emitting 20 mW at 488 nm. Data analysis was performed with Kaluza Analysis 2.1 (Beckman Coulter GmbH). Total (SG-stained samples) or intact cells (SGPI-stained samples) were selected using the green and red detection channels. Cell viability (%) was calculated as the ratio of intact to total cells.

Fresh fecal samples from two healthy toddlers (toddler #1: female, 14-month-old; toddler #2: male, 17-month-old) and two healthy adults (adult #1: male, 34-year-old; adult #2: female, 23-year-old) were used to conduct fecal-derived ex vivo batch fermentations. Within an anaerobic chamber (10% CO₂, 5% H₂, and 85% N₂; Coy Laboratory Products, Inc.), fecal slurries were prepared by mixing ~2 g feces, 10 autoclaved glass beads (borosilicate, diameter: 3 mm) and 20 mL anaerobic peptone water (Oxoid™, Thermo Scientific, Waltham, USA), and the suspensions were vortexed for 2 min. After the sedimentation of fecal particles, the upper layers were transferred into Hungate tubes which were tightly closed for use outside the anaerobic chamber. Balch-type tubes containing 20 mL toddler-mYCFA (per liter: 3 g lactose, 0.3 g xylan, 0.3 g arabinogalactan, 0.3 g starch, and 2 g mucin as added carbon sources in mYCFA) or adult-mYCFA (per liter: 1 g cellobiose, 1 g xylan, 1 g arabinogalactan, 1 g starch, and 1 g mucin as added carbon sources in mYCFA) were anaerobically spiked with 2 mL of 0.1 M HCl solution (control) or 2 mL of tryptamine (stock at 5.9 mg/mL tryptamine in 0.1 M HCl; final concentration: 2.4 mM tryptamine). Finally, 0.2 mL of fecal slurry was inoculated in the corresponding media to initiate the fermentation, and the tubes were incubated at 37°C for 48 h. Samples (2 mL) were harvested anaerobically using sterile syringes after 0, 5, 24, and 48 h and centrifuged at 14,000 × g for 10 min at 4°C to separate cell pellets (for 16S rRNA community analysis) and supernatants (for metabolite quantification and pH), and the resulting samples were stored at -80°C until further analysis.

DNA was isolated from the cell pellets obtained from 2 mL of batch fermentation samples using the FastDNA Spin Kit for Soil (MP Biomedicals, Illkirch-Graffenstaden, France), following the manufacturer’s instructions. The bacterial community was analyzed using 16S rRNA gene sequencing. Briefly, a two-step PCR approach was used to target the V4 region using primers 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) (Microsynth AG, Balgach, Switzerland). The samples were barcoded using Nextera XT v2 indexes (Illumina, San Diego, CA, USA) and pooled in equimolar concentration. Sequencing was performed with the Illumina MiSeq platform (Genetic Diversity Centre, ETH Zurich, Switzerland) using v2 chemistry supplemented with PhiX 20 pM (10%) and 250 × 2 read length. Raw data were processed using the DADA2 R package (v1.14.1) to obtain exact amplicon sequence variants (ASVs) using the metabaRpipe R package (Constancias and Mahé, 2022) as previously described (Kundra et al., 2022). Taxonomy was assigned to ASVs using DADA2 against SILVA database (v138.1). Determination of the phylogenetic relatedness of ASVs and alpha diversity (observed and Shannon indexes) and beta diversity analyses (weighted and unweighted UniFrac) were performed using the phyloseq (McMurdie and Holmes, 2013) and the DivComAnalyses packages (Constancias and Sneha-Sundar, 2022), and differential abundance was established using DESeq2 (Love et al., 2014).

The concentrations of Try and Trp in bacterial supernatants were determined by ultra-performance liquid chromatography equipped with a diode array detector (UPLC-DAD) and pre-column derivatization with diethyl ethoxymethylenemalonate (DEEMM), as described before (Otaru et al., 2021). Briefly, the derivatization was performed by mixing 200 μL of supernatant sample with 350 μL of borate buffer (1 M H₃BO₃ adjusted to pH 9 with NaOH), 150 μL methanol, 8 μL of internal standard (2 g/l L-2-aminoadipic acid, Sigma-Aldrich), and 7 μL DEEMM (VWR International GmbH). The reaction mix was incubated at room temperature in an ultrasound bath for 45 min and was subsequently heated at 70°C for 2 h to stop the derivatization. The samples were passed through a 0.2-μm nylon membrane filter and stored at 4°C until UPLC-DAD analysis. An ACQUITY UPLC H-Class system (Waters Corp., Milford, MA, USA) coupled to a diode array detector at 280 nm was used to detect the derivatized molecules. The separation (1-µL injection per sample) was performed at 40°C using an ACQUITY BEH C18 VanGuard pre-column (1.7-μm particle size, 2.1 mm × 5 mm; Waters Corp.) connected to an ACQUITY BEH C18 column (1.7-μm particle size; 2.1 × 100 mm; Waters Corp.), a flow rate of 0.46 mL/min, and a gradient consisting of (A) 25 mM acetate buffer (pH 6.6), (B) 100% methanol, and (C) 100% acetonitrile [gradient described in detail in Otaru et al. (2021)]. The detection limit for Try and Trp was 10 µM. The data were processed using Empower 2 software (Waters Corp.).

Carbohydrates (glucose, galactose, and lactose), intermediate metabolites (formate, succinate, and lactate), short-chain fatty acids (SCFAs; acetate, butyrate, and propionate), and branched-chain fatty acids (BCFAs; isobutyrate and isovalerate) were quantified by high-performance liquid chromatography equipped with a refractive index detector (HPLC-RI) as previously described (Poeker et al., 2018). Briefly, 200 μL of bacterial supernatant was passed through a 0.2-μm nylon membrane filter prior to HPLC-RI analysis. The separation was carried out with a LaChrom HPLC-System (Merck-Hitachi, Tokyo, Japan) using a SecurityGuard Cartridges Carbo-H (4 × 3.0 mm; Phenomenex Inc., Torrance, CA, USA) connected to a Rezex ROA-Organic Acid H+ (300 × 7.8 mm; Phenomenex Inc.) column. The samples (40-μL injection) were eluted at 40°C under isocratic conditions (10 mM H₂SO₄ and flow rate 0.4 mL/min), and the analytes were quantified using a refractive index detector L-2490 (Merck Hitachi). Data were processed using EZChrom software (Agilent, Santa Clara, CA, USA).

To evaluate the effect of Try on the growth of 18 intestinal strains ( Supplementary Table 1 ), all strains were reactivated (1.25%, v/v) into anaerobic mYCFA until saturation (approximately ~20 h) at 37°C and were then diluted to OD₆₀₀ 0.3 with mYCFA. The assay was then initiated in the anaerobic chamber by inoculating (1%, v/v) the diluted bacterial suspensions into 96-well plates (SPL Life Sciences Co. Ltd., Gyeonggi-do, South Korea) containing 200 µL mYCFA with incremental concentrations of tryptamine (from 0.5 to 8 mM). All plates also contained mYCFA without Try (control wells; unperturbed growth). The preparation of the medium for growth in the multiwell setup in the anaerobic chamber was based on the method previously described (Zünd et al., 2024). The plates were incubated at 37°C in a Tecan Infinite M200 PRO plate reader (Tecan Group Ltd., Männedorf, Switzerland), and OD₆₀₀ was monitored for 24 h. Subsequently, all wells were re-inoculated (1%, v/v) into 96-well plates containing fresh mYCFA without tryptamine. These plates were incubated at 37°C, and OD₆₀₀ was monitored for an additional 24 h. To quantify growth, we calculated the area under the curve (AUC) from OD₆₀₀ measurements using the trapezoidal rule. To determine growth with Try as compared to the control (% growth), AUCs were normalized by dividing the mean AUC for specific Try concentrations by the mean AUC of unperturbed growth (control) of the same strain within the same plate. Each condition was tested in two independent replicates.

Data were analyzed and visualized using GraphPad Prism (v9.2.0) or R (v3.6.3) software. Statistical analysis was performed by multiple unpaired t-test, including Holm–Šídák test to correct for multiple comparison. The significance level was set to p < 0.05. The effects of Try treatment on microbiota alpha and beta diversity were tested using unpaired t-test (normal distribution assessed using the Shapiro–Wilk test) and PERMANOVA analysis, respectively, using the phyloseq R package (McMurdie and Holmes, 2013).

To identify tryptamine-producing strains in vitro, we first screened 13 intestinal strains closely related to known Try producers (i.e., Ruminococcus gnavus ATCC 29149 and Clostridium sporogenes ATCC 155797) or previously identified in silico as potential Try producers (i.e., Blautia hansenii, Clostridium nexile, and Enterocloster asparagiforme) (Williams et al., 2014). We evaluated extracellular Try production after 24 and 48 h of growth in mYCFA containing 24.5 mM Trp. This elevated concentration was chosen to enhance the detectability of tryptamine using our analytic method by amplifying the metabolic capabilities of the tested strains. Six strains produced tryptamine, namely, C. sporogenes DSM 795 (Clostridiaceae family), C. nexile DSM 1787, E. asparagiformis DSM 15981, B. hansenii DSM 20583, R. gnavus DSM 108212, and the positive control R. gnavus ATCC 29149 (Lachnospiraceae family) ( Figure 1A ). No tryptamine was detected in Lachnospiraceae bacterium DSM 24404, Blautia hydrogenotrophica DSM 10507, Blautia obeum DSM 25238, Blautia producta DSM 14466, Mediterraneibacter glycyrrhizinilyticus DSM 17593 (Lachnospiraceae family), Ruminococcus gauvreauii DSM 19829, and Ruminococcus bromii ATCC 27255 (Oscillospiraceae family) under the tested conditions (limit of detection: 10 µM; Figure 1A ). The lowest tryptamine concentration was detected in E. asparagiformis DSM 15981 (16 ± 1 µM after 48 h) and the highest in R. gnavus DSM 108212 (3883 ± 110 µM after 48 h; Figure 1A ). Next, we evaluated whether Try production was associated with a particular bacterial growth phase. Each newly identified tryptamine-producing strain was grown in mYCFA containing 15 mM Trp (E. asparagiformis DSM 15981 was excluded because the quantified Try was close to the limit of detection of the method), and OD₆₀₀ and extracellular Try levels were monitored during 48 h ( Figure 1B ). In all four tested strains, Try formation was mixed-growth-associated (Sakthiselvan et al., 2019), with initial detection at low levels during the exponential growth phase (before 5 h growth), and additional production during the stationary phase to reach the highest levels after 48 h (ranging from 134 ± 2 µM for C. nexile DSM 1787 and up to 2,746 ± 325 µM for R. gnavus DSM 108212; Figure 1B ). Finally, we assessed whether the concentration of available tryptophan could affect tryptamine production. Increasing Trp concentration in the media from 10 to 24 mM resulted in a significant increase of Try production in R. gnavus DSM 108212 (+ 2.7 mM max. after 24 h), B. hansenii DSM 20583 (+ 1.2 mM max. after 24 h), and C. sporogenes DSM 795 (+ 0.5 mM max. after 48 h), with no major impact on the growth ( Supplementary Figure 1 ).

Because the decarboxylation of amino acids into biogenic amines often constitutes a mechanism of acid stress tolerance in bacteria (Guan and Liu, 2020), we next investigated if tryptamine production was modulated by pH and could potentially be induced at low pH to protect cells against acid stress-mediated damage.

First, we investigated if higher Try levels were produced when applying acid stress to an equivalent amount of actively growing cells. R. gnavus ATCC 29149 was grown at 37°C until OD₆₀₀ ~0.2 in mYCFA (with 5 mM Trp), after which the pH was reduced to 5.0 and 5.5 (from 6.2; control), and Try concentration was quantified 2 and 18 h after the acid stress. Compared to unperturbed control cells, the Try concentration normalized per OD₆₀₀ did not significantly increase post-stress, yet slightly higher but non-significant normalized Try levels were observed at pH 5.0 compared to the control (6,293 ± 203 μM/OD₆₀₀ vs. 5,333 ± 421 μM/OD₆₀₀; p = 0.07) at 18 h post-stress ( Supplementary Figure 2 ).

To further examine Try production in response to acid stress and assess if Try production was associated with increased cell viability under acidic conditions, R. gnavus DSM 108212 and C. sporogenes DSM 795 were grown in mYCFA until the late exponential growth phase (OD₆₀₀ ~1.0). Subsequently, they were exposed for 1 h to carbohydrate-depleted mYCFA at three different pH values (i.e., 6.7, 5.4, and 4.3), with and without the precursor Trp (5 mM). In the absence of Trp, tryptamine was not produced under any of the tested pH conditions ( Figure 2 ). When Trp was supplemented, maximum Try production was detected at the highest pH of 6.7 (386 ± 36 µM for R. gnavus DSM 108212 and 282 ± 5 µM for C. sporogenes DSM 795), decreased at pH 5.4 (289 ± 14 µM for R. gnavus DSM 108212 and 75 ± 2 µM for C. sporogenes DSM 795), and was not detected at pH 4.3 in both strains ( Figure 2 ). When normalized per number of intact cells, the Try produced remains highest at pH 6.7 versus pH 5.4 (3.2 nM/cell vs. 1.7 nM/cell for R. gnavus and 1.9 nM/cell vs. 0.5 nM/cell for C. sporogenes). On the other hand, Trp was consumed by both strains at all pH levels, including at pH 4.3 where no tryptamine was detected, with 2,164 ± 568 µM and 3,807 ± 522 µM Trp consumed by R. gnavus DSM 108212 and C. sporogenes DSM 795, respectively ( Figure 2 ). No difference of cell viability was observed in response to Try production in R. gnavus DSM 108212 at pH 6.7 and 5.4 ( Figure 2A ). In C. sporogenes DSM 795, no difference of cell viability was observed at pH 6.7, but a slight increase of viable cells was observed at pH 5.4 in the presence of Trp (89.1% vs. 83.5%, p = 0.002; Figure 2B ). For both strains, no viable cells were detected after 1 h of exposure to pH 4.3 in the presence or absence of Trp ( Figure 2 ).

Our data altogether suggest that tryptamine production may not be induced by acid stress in R. gnavus and C. sporogenes under the tested conditions and therefore may not constitute an effective mechanism of protection against acid stress.

Previous studies showed a significant modulatory effect of bacteria-derived tryptamine on intestinal homeostasis (Williams et al., 2014; Bhattarai et al., 2018), with indications of a potential direct impact of Try on the gut microbiota function and composition (Dopkins et al., 2021). To evaluate this interaction, we performed ex vivo fecal batch fermentation and exposed four healthy gut microbiota (two adults and two toddlers) to 2.4 mM tryptamine. We then assessed how Try could influence the metabolism and composition of gut microbial communities throughout the fermentation period (i.e., 0, 5, 24, and 48 h). The concentration of 2.4 mM was selected based on the levels produced by high Try producers in vitro ( Figure 1 ) and on previous studies evaluating colonic tissues response to Try in Ussing chamber assays (Williams et al., 2014; Bhattarai et al., 2018).

Overall, Try supplementation induced mild effects on the carbohydrate metabolism of all tested intestinal microbial communities. A minor increase of monosaccharide and disaccharide concentrations was observed in the presence of Try as compared to the control early during the fermentation (5 h) in toddler #1 (7.3 mM vs. 2.9 mM lactose; p < 0.05), toddler #2 (0.27 mM vs. 0.11 mM glucose and 0.27 mM vs. 0.04 mM galactose; p < 0.05), adult #1 (0.56 mM vs. 0.17 mM glucose and 0.77 mM vs. 0.20 mM galactose; p < 0.05), and adult #2 (2.2 mM vs. 1.4 mM; p < 0.05) ( Supplementary Figure 3 ). Since polysaccharides (i.e., fibers) were supplemented in both toddler-mYCFA and adult-mYCFA, this could suggest an increased release of mono/di-saccharides or, more likely, a reduction in carbohydrate consumption by gut microbes. After 5 h, the pH of the milieu was indeed higher in the presence of Try as compared to the control (except for adult #2; Supplementary Figure 4 ), and there was a significant reduction in the concentration of the intermediate metabolite succinate [1.7 mM vs. 3.1 mM in toddler #1, 0.4 mM vs. 1.3 mM in adult #1, and 0.4 mM vs. 1.0 mM in adult #2 (p < 0.05); no difference in toddler #2], but not lactate or formate ( Figure 3 ).

After 48 h, the end products from carbohydrate (SCFAs) and protein fermentation (BCFAs) were affected by Try supplementation in a donor-specific manner. In adult #2, the concentrations of both SCFAs and BCFAs were significantly altered, with a decrease of acetate (60.8 mM vs. 76.7 mM, p = 0.001) and an increase of butyrate (31.4 mM vs. 26.5 mM, p = 0.0002), propionate (27.8 mM vs. 15.1 mM, p < 0.0001), isobutyrate (6.7 mM vs. 1.2 mM, p < 0.0001), and isovalerate (8.7 mM vs. 1.4 mM, p < 0.0001; Figure 3 ). In adult #1, Try supplementation did not alter SCFAs but increased isobutyrate (1.9 mM vs. 0.9 mM, p = 0.01) and isovalerate levels (2.5 mM vs. 1.2 mM, p = 0.02; Figure 3 ). In toddlers’ microbiota, no significant changes of SCFAs and BCFAs were detected in the presence of Try, except for a small but significant decrease of propionate in toddler #1 (30.4 mM vs. 33.1 mM, p = 0.003; Figure 3 ). With all four tested gut microbial communities, no significant reduction of Try concentration was observed throughout the fermentation ( Supplementary Figure 5 ), indicating that tryptamine was not metabolized.

We suspected that the observed metabolic changes could be associated with an alteration of the microbial community. Therefore, we performed 16S rRNA metabarcoding for 5- and 48-h fermentation samples from all donors. At the phylum level, Try mediated a significant decrease of the mean relative abundance of the phylum Bacteroidota after 5 h in all donors (toddler #1: 25% vs. 32%, p = 0.02; toddler #2: 10% vs. 16%, p = 0.03; and adult #2: 25% vs. 28%, p = 0.02), except for adult #1 where a non-significant reduction was observed (16% vs. 29%, p = 0.09; Supplementary Figure S6A ). A similar pattern was observed at the class level for Bacteroidia ( Supplementary Figure 6B ; Supplementary Table 2 ). The reduction of Bacteroidota remained significant after 48 h only in toddler #1 (15% vs. 30%, p = 0.0003), along with an increase of Actinobacteriota (34% vs. 29%, p = 0.005) and Firmicutes (50% vs. 40%, p = 0.0001). A bloom of Proteobacteria was observed in adult #2 in the presence of Try (48 h: 5% vs. 2%, p = 0.0002; Supplementary Figure 6A ).

For all four tested microbiota, the observed ASVs (richness; Supplementary Figure 7A ) and the Shannon index (both richness and evenness; Figure 4A ) remained stable whether Try was present or absent. However, a small yet significant increase in the Shannon index was noted at 5 h in adult #1 ( Figure 4A ). The beta-diversity analyses showed a distinct clustering of samples treated with Try compared to the control samples (regardless of the time point) when considering the taxa abundance, as indicated by the weighted UniFrac distance matrix (p < 0.05 for all donors except adult #1; Figure 4B ). However, this pattern was not observed when considering the presence or absence of taxa only, as shown by the unweighted UniFrac matrix (all donors non-significant; Supplementary Figure 7B ). Taken together, our data indicated that Try exerts a mild effect on the overall gut microbiota community structure and on carbohydrate metabolism.

Although limited, the donor-specific alterations observed in all four fecal-derived microbiota samples ( Figures 3 , 4 ) suggest that certain microbes, rather than the whole community, may be particularly sensitive to the presence of Try. In turn, such specific growth inhibition could alter the microbial community structure and the cross-feeding interactions between microbes. To identify the specific taxa affected by Try, we performed differential abundance analysis at the ASV level (DESeq2 analysis). Considering the two timepoints (5 and 48 h) and the four donors, a total of 145 ASVs were altered by the presence of Try (log2fold change > 1; p < 0.05; Supplementary Figure 8 ). One ASV, specifically Blautia sp. (ASV005), decreased across all tested conditions. In all adult microbiota samples, Faecalibacterium sp. (ASV012), Blautia massiliensis (ASV010), and Blautia faecis (ASV077) were reduced, and Coprococcus comes (ASV024) was increased. Notably, Bacteroides vulgatus (ASV001) was decreased in all toddler #1 and adult #2 samples.

To determine whether Try possesses inhibitory properties against specific intestinal strains, we cultivated a range of gut bacteria, including some of those identified as downregulated ASVs in the DESeq2 analysis, and exposed them to Try, ranging from 0.5 to 8 mM. This range was chosen to ensure comparability with the 2.4-mM concentration used in ex vivo fecal cultivation and included a condition within the micromolar range. Percentage growth as compared to the control without Try was determined using area under the curve (AUC). Overall, the lowest concentration of Try tested (0.5 mM) had little to no effect on the growth of most tested strains, except for Bacteroides uniformis DSM 6597, which only grew to 68% of the control ( Figure 5 ). At 2 mM Try, all three Bacteroides strains tested were affected, with 61% (Phocaelicola vulgatus DSM 1447, previously B. vulgatus), 55% (B. thetaiotaomicron DSM 2079), and 32% (B. uniformis DSM 6597) growth compared to the control ( Figure 5 ). Exposure to 2 mM Try also significantly inhibited the growth of the three Faecalibacterium strains tested, with 26% (F. prausnitzii DSM 107838), 18% (F. duncaniae DSM 17677), and 11% (F. prausniitzii DSM 107840) growth compared to the control ( Figure 5 ). Blautia obeum DSM 25238 was not strongly affected by Try, with 69% growth observed at 4 mM ( Figure 5 ). Interestingly, the growth of the Try producers R. gnavus ATCC 29149 and DSM 108212 was reduced at 4 mM, with 32% and 38% growth compared to the control, respectively, which was not the case for C. sporogenes DSM 795 and B. hansenii DSM 20583 ( Figure 5 ). Several other strains, including Coprococcus eutactus DSM 107541, Anaerostipes caccae DSM 14662, and Anaerobutyricum hallii DSM 3353, were not inhibited by Try at any concentration ( Figure 5 ). Interestingly, upon reinoculation (1%, v/v) into Try-free fresh medium, most strains recovered from exposure to Try concentrations of 4 mM and below, except for the Faecalibacterium strains ( Supplementary Figure 9 ). Collectively, our data suggest that Try exhibits inhibitory properties, diminishing the growth of specific intestinal bacteria while others remain unaffected.