Sulfated dietary fiber protects gut microbiota from antibiotics

Abstract

Background: Antibiotics, while essential for combating pathogens, also disrupt commensal bacteria, leading to gut microbiota imbalance and associated diseases. However, strategies to mitigate such collateral damage remain largely underexplored.

Result: In this study, we found that fucoidan, a marine polysaccharide derived from brown seaweed, provides broad-spectrum growth protection against multiple classes of antibiotics for human gut microbial isolates in vitro and for fecal communities ex vivo. This protective effect is dependent on the structural integrity, molecular weight, and sulfur content of the polysaccharide. Transcriptomic analysis showed that while fucoidan had minimal impact on baseline gene expression, it counteracted about 60% of the genes induced by kanamycin, suggesting a potential inhibition of kanamycin. Mass spectrometry results further showed that this inhibition may be due to the non-specific binding of fucoidan to kanamycin in solution. Finally, animal model experiments revealed that fucoidan facilitated the recovery of gut microbes following antibiotic treatment in vivo.

Conclusion: These findings suggest fucoidan could serve as a potential intervention to help protect gut microbiota during antibiotic therapy. Further studies are needed to evaluate its clinical potential and ensure it does not compromise antimicrobial efficacy. Video Abstract.

Keywords: Antibiotics; Dietary fiber; Fucoidan; Gut microbiome; Microbial community; Sulfated polysaccharide.

Fig. 1

Fucoidan and sulfated polysaccharides protect E. coli isolates from ampicillin and kanamycin. A E. coli DH5α cell growth kinetics under ampicillin (MIC: 3.75 μg/ml), with 22 dietary supplements (1 mg/ml). H₂O is added in the control (black line); #P5: ulvan polysaccharides from Ulva spp. #P16: fucoidan from Undaria pinnatifida. Error bars represent the standard error (n = 3). The y-axis is cells’ optical density measurement at 600 nm (OD600) with arbitrary unit (a.u.). B Dose response of ampicillin-treated cells to ulvan polysaccharides (#P5). #P5 concentration ranges from 0.1–5 mg/ml. Error bars represent the standard error (n = 3). C Human fecal isolate E. coli_P09 growth kinetics under kanamycin (MIC: 40 μg/ml), with 38 substrates (1 mg/ml). The colored lines are samples with fucoidans from different macroalgae or manufacturers. Error bars represent the standard error (n = 3). D Inhibition zone tests for E. coli_P09 cells grown on agar plates with or without fucoidan (#P35, 2 mg/ml). Kanamycin (2 µl, 40 µg/ml) was pipetted in the plate center. Plates became yellowish with fucoidan addition. Images were taken after 16 h of incubation at 37 °C. E The diameter of the inhibition zones was measured by a vernier scale. Error bars represent the standard error of three plates. **p < 0.01 (Student’s t-test). F Dose response of E. coli_P09 cells to kanamycin in the presence of 3 mg/ml fucoidan #P35 in the medium. Control is 40 μg/ml of kanamycin without fucoidan supplement. Error bars represent the standard error (n = 3)

Fig. 2

Screening of dietary supplements that protect cells from antibiotics. A total of 38 dietary supplements (1 mg/ml) and 8 antibiotics were screened for 3 E. coli strains, including the DH5α strain and two isolates from human fecal samples (E. coli_P09 and E. coli_P58). Numbers boxed in purple represent monosaccharides, in orange represent oligosaccharides, and in blue represent polysaccharides. * represents fucoidans from different species and sources (Table S1). Cell growth was monitored by a plate reader for at least 12 h at 30 min intervals. The fold changes in growth were calculated by dividing the OD of samples with each supplement (1 mg/ml) by the OD of control at corresponding time points. The maximum fold change for each antibiotic-supplement combination is plotted. Fold change > 3.0 is colored in red, and fold change < 3.0 is not included (in grey)

Fig. 3

Molecular weight, structural integrity, and degree of sulfation determine polysaccharide protection against antibiotics. A Large molecular size is required for cell protection by fucoidan. Fucoidan #P16 (3 mg/ml) was separated into two fractions using a 10-kDa cutoff centrifugal filter. Each fraction was added to E. coli_P09 cell cultures treated with ampicillin at their MIC. Error bars represent the standard error (n = 3). B Hydrolyzed fucoidan was unable to protect cells from ampicillin. Fucoidan #P16 was hydrolyzed by acid and heating. Error bars represent the standard error (n = 3). C Strong (#P35-121 °C for 3 h) and mild (#P35-80 °C for 12 min) hydrolyzation undermined fucoidan’s (#P35, 1 mg/ml) protection against kanamycin. Error bars represent the standard error (n = 3). D Highly sulfated large dextran confers cell protection against kanamycin. E. coli_P09 cells were grown at kanamycin at their MIC with dextrans or dextran sulfates (1 mg/ml) added into the medium. Error bars represent the standard error (n = 3)

Fig. 4

Transcriptomics and mass spectrometry for mechanisms of fucoidan protection. A Volcano plots of differentially expressed genes in each treatment group compared to control. From left to right: kanamycin over control (Kan/Ctl); fucoidan over control (Fcdn/Ctl); and kanamycin and fucoidan over kanamycin (Kan_Fcdn/Kan). The horizontal dash line corresponds to a Bonferroni-corrected significance value of 0.1 (qval). Dots above the dashed line (significant) are colored red, and dots below the dashed line (non-significant) are colored blue. B Heatmap of the top 100 differentially expressed transcripts between treatment and control. Transcripts were filtered with q-value < 0.05 and organized by hierarchal clustering. Each row is a transcript and each column is a sample, with the color bar at the top indicating which treatment/control group the sample is from. The unit for the color coding of the heatmap is log-transformed transcript counts per million. C Principle component analysis of the transcript counts in the transcriptomic dataset. Samples are colored by treatment groups. D Interaction analysis for differentially expressed genes induced by kanamycin and kanamycin_Fucoidan. Blue dots represent the log2 fold change data for a total of 904 shared DEGs between Kan/Ctl and Kan_Fcdn/Kan groups. The red line is y = -x. E MALDI-TOF results for kanamycin detection in solution. Top row: pure 40 µg/ml kanamycin solution (MW = 485 g/mol); 2nd row: solution with kanamycin and 1 mg/ml fucoidan #P16; 3rd row: solution with kanamycin and 10 mg/ml fucoidan #P16; 4.^th row: kanamycin with 10 mg/ml fucose #P3

Fig. 5

Fucoidan protects human fecal communities from antibiotics ex vivo. A Growth of fecal microbial community cultures treated with kanamycin or ampicillin, with or without fucoidan #P35 at 24 h. Left: donor H01; kanamycin: 20 µg/ml; fucoidan #P35: 1 or 3 mg/ml. Right: donor H02; ampicillin: 5 µg/ml. Two sample t-tests were used to compare the significance with p-value shown above. Significance was determined using two-sample t-tests, with 3 replicates in each group and p-values indicated above. B Community composition changes of the fecal microbial cultures in control, fucoidan and/or kanamycin groups for subject H01 (n = 3 for each group). Only the top 30 ASVs were presented at the genus level. C Clustering of the fecal community cultures in (B) by NMDS of Bray Curtis dissimilarity (left, PERMANOVA: R² = 0.874, p < 0.001) and weighted Unifrac distances (right, PERMANOVA: R² = 0.867, p < 0.001). D Community composition changes of the fecal microbial cultures in control, fucoidan and/or antibiotics (kanamycin or ampicillin) groups for subject H02 (n = 3 for each group). Only the top 30 ASVs were presented at the genus level. E Clustering of the fecal community cultures in (D) by NMDS of Bray Curtis dissimilarity (left, PERMANOVA: R² = 0.797, p < 0.001) and weighted Unifrac distances (right, PERMANOVA: R² = 0.786, p < 0.001)

Fig. 6

Fucoidan facilitated the recovery of gut microbes in mice after antibiotic treatment. A Schematic of the mouse study illustrating the feeding strategies, timeline, and fecal sampling. Groups of six-week C57BL/6 J female mice were randomly assigned into four groups: control (Ctl, N = 6 mice), kanamycin (K, N = 6), fucoidan (F, N = 6), and fucoidan_kanamycin (FK, N = 12). The ‘C’ represents the control diet, ‘F’ represents the fucoidan diet, and ‘K’ represents kanamycin water. The study consisted of three phases: pre-treatment (days −1 ~ 1, all mice on control diet/water), treatment (days 2 ~ 17, K ± F), and recovery (days 18 ~ 33, kanamycin was discontinued). During the treatment phase, mice in the K and FK groups received kanamycin-supplemented water for 6 h daily, followed by regular water for approximately 17 h, allowing for about 1 h per day for animal husbandry, weighing, and sample collection in between. Fucoidan diet was given continuously to mice in the F and FK groups. During recovery, the fucoidan diet was maintained for four additional days (days 18 ~ 21) in the F and FK groups to counter potential residual antibiotics, followed by a regular diet and water for the remaining period. B PCoA of weighted UniFrac distances showing microbiome community shifts across the experimental phases: pre-treatment (day 1), end of the treatment (day 17), and post-treatment recovery (day 33). Temporal changes in community structure were tested by PERMANOVA (R² = 0.081, p < 0.001). C PCoA of weighted UniFrac distances for beta-diversity among different treatment groups in the treatment period (days 13 and 17). Community structure significantly differed between treatments (PERMANOVA: R² = 0.203, p < 0.001). D Temporal changes of microbial abundance for a cluster of bacteria (ASVs) that recovered after kanamycin treatment in the presence of fucoidan. The fold changes in relative abundance for each taxon in the early (d3 + d5) and late (d13 + d17) treatment phase, as well as in the recovery phase (d27 + d33) were compared to the pre-treatment control group (on day 1). ∗ ∗ p < 0.01; and ∗ ∗ ∗ p < 0.001 (Paired t-tests). E Temporal changes of fold difference (compared to control) of five bacterial genera with the most significant differential abundance in the kanamycin (K) group and kanamycin_fucoidan (FK) group, respectively