In this study, a total of 10 individuals with constipation were included. The inclusion criteria were as follows: those who met the Rome IV diagnostic criteria for constipation (15); age 20–50 years; Wexner scale requirements (score out of 30, with a positive correlation between score and severity of constipation (16), and defecating frequency of 1–3 times/week. Following were the exclusion criteria: use of any probiotic product in 2 weeks prior to the beginning of the study; suffering from gastrointestinal, neurological, cardiovascular, endocrine, renal, and/or other chronic diseases that may affect intestinal motility; suffering from chronic diarrhea; those who underwent gastric banding for weight loss or gastric bypass surgery and use of conventional laxatives, and pregnant or lactating women. The study design received the approval of the Ethics Committee of Hangzhou Normal University (approval number: 20190061). All participants in this study participated voluntarily and provided signed informed consent.

The fecal samples from constipated person were collected in sterile stool collection tubes, transported to the laboratory at low temperature, and stored at −80°C until subsequent use. All samples were divided into four groups for the in vitro fecal fermentation assay as described previously by Ignacio et al. with appropriate modifications (17). Specifically, 0.8 ± 0.02 g of the sample was placed in a HALO-F100 fecal processor (Suzhou Hailu Biotech, Jiangsu, China) to obtain a 10% fecal suspension. Subsequently, 0.5 mL of the clarified suspension was injected into the fermentation flasks containing YCFA medium (18). Samples were divided into the following 4 groups: blank control (no additional addition), BL-99 (5 × 10⁹ cfu/mL), FOS (concentration of 4 g/L, purchased from Baolingbao Biology, Shandong, China), and BL-99_FOS (BL-99 concentration of 5 × 10⁹ cfu/mL and FOS concentration of 4 g/L). All samples were tested after 24 h of fermentation at 37°C.

The FOS degradation rate was determined using thin-layer chromatography following the method described previously by Wang et al. with appropriate modifications (18). Briefly, 0.2 μL of the sample was placed at the marked position on the chromatography plate. Subsequently, the blow-dried chromatography plate was quickly immersed in the middle of the chromatography tank, covered with a lid, and left until the level of the developing solvent reached the top of the plate; next, it was removed and quickly blown dry. Further, the dried plate was quickly submerged in the color development agent. Next, the liquid on the plate was blown dry to ensure color development. The scanned images were processed using the scanner through the “TLC” software. Finally, the FOS degradation rate was calculated using the following equation.

The H₂, CO₂, CH₄, and H₂S concentrations and total gaseous release from in vitro fermentation were measured using a gas analyzer (HL-QT01, Hailu Biotech, Hunan, China) following the procedure described by Wang et al., with appropriate modifications (19). Specifically, the gas analyzer was calibrated by detecting the gasses in a control medium without any inoculum. The fermented gas was made to enter the gas analyzer through a sampling tube, wherein the gas content was determined by a gas sensor. Finally, the measurements were analyzed using the “Multi-Gas Analyzer.exe software.

The short-chain fatty acid (SCFAs) levels in various groups of fermentation broth were determined by gas-phase method according to the protocol described by Henrique-Bana et al. with appropriate modifications (20). Briefly, 0.5 mL of 10% fecal suspension was mixed with 0.1 mL of 25% metaphosphoric acid for 24 h at −20°C. Subsequently, this mixture was thawed, centrifuged for 20 min at 4°C, 3000 0067, and filtered using a 0.22 μm membrane for further use. A gas chromatograph (GC, Shimadzu, GC-2010 Plus, Kyoto, Japan) equipped with the DB-FFAP column (30m × 250 μm × 0.25 μm, Agilent Technologies, Inc.) was used for measuring SCFA content. Measurements included levels of total acid, propionic acid, acetic acid, and butyric acid (purity [standards of propionic acid, acetic acid, and butyric acid] >99%, chromatographic grade, purchased from Dr. Ehrensorfer Standards, Germany), with 2-ethylbutyric acid as the internal standard.

The gas chromatography conditions were as follows: Agilent DB-FFAP capillary column (30 m × 0.25 mm × 0.25 μm, column temperature of 75°C, Agilent Technologies); 20°C/min to 180°C for 1 min and 50°C/min to 220°C for 1 min; injection port temperature: 250°C; injection volume; 1.0 μL; split ratio: 5:1; carrier gas type: high purity nitrogen; detector type: FID; temperature: 250 °C; make-up gas concentration: 20 mL/min; air flow rate: 300 mL/min, and hydrogen flow rate: 30 mL/min.

Using the MOBIO Power Fecal DNA Isolation Kit (Qiagen, Germany), the total fecal DNA was extracted following the protocol described by Marsaux B and Nicolucci A C (21, 22), with appropriate modifications. In the V3-V4 region of the 16S rRNA gene, the sample-specific barcode sequences were amplified using the primer pair, 806R (5′-GTGGACTACHVGGGTWTCTAAT-3′) and 338F (5′-GTACTCCTACGGGAGGCAGCA-3′). The resulting amplicon libraries (Roche KAPA Library Quantification Kit) were equally combined and subjected to quantification and further sequenced on the Illumina MiSeq platform in the PE300 mode. Using the Quantitative Insights into Microbial Ecology version 2.0 (QIIME2) pipeline, we identified the sequences by barcoding. Sequences of acceptable quality (i.e., minimum mass fraction = 25, min/max length = 200/750, no mismatches in primer sequences, and no base ambiguity) were retained. Further, these were clustered into operational taxonomic units (OTUs) by a redundancy algorithm embedded in the DADA2 package and classified using the Silva database (version: SSU138.1). Subsequently, we constructed phylogenetic trees for the representative sequences and calculated the β-diversity by weighted principal coordinate analysis (PCoA), and α-diversity using the Shannon index. Species correlations with SCFAs and gasses were shown using correlation heatmap plots based on the Spearman rank correlation coefficients.

Data were expressed as mean ± standard deviation (M ± SD). All statistical analyses were performed on the SPSS 23.0 software (SPSS Inc. Chicago, IL, United States). The one-way analysis of variance was used to make comparisons between samples. P < 0.05 was the indicator of statistical significance. The graphs were plotted using (GraphPad Software Inc., San Diego, CA, United States) GraphPad Prism 8.

The ten adult female participants in the study had an average age of 28.9 ± 5.2 years, mean height of 162.8 ± 7.1 cm, and mean weight of 59 ± 10.6 kg. They complied with the Wexner scale (out of 30) after inclusion. The basic information of the participants is listed in Table 1.

The FOS degradation rate in the fermentation broth after in vitro fermentation is shown in Figure 1A. After 24 h of fermentation, the FOS degradation rates in FOS and BL-99_FOS groups were 65.25 ± 12.82% and 87.58 ± 6.30%, respectively. These markedly significant differences (P < 0.001) in the FOS degradation rates between the two groups indicated that BL-99 could utilize FOS.

Gas production was measured in each group after 24 h of in vitro fermentation (Figures 1B–F). Relative to the YCFA group, the CO₂ production in the FOS group was significantly lower (P < 0.01), whereas that in the BL-99_FOS group was highly significantly lower (P < 0.001). CO₂ production was also lower in the BL-99 group, however, this difference was not statistically significant. CH₄ content decreased in the BL-99, FOS, and BL-99_FOS groups, however, the comparisons were not significant. H₂ content increased in the BL-99, FOS, and BL-99_FOS groups but the differences were not statistically significant. H₂S production was lowered by a highly significant level in the FOS and BL-99_FOS groups (P < 0.001). H₂S production was reduced in the BL-99 group, however, the difference was not statistically significant. The total production of the four gasses was highly significantly lowered in the FOS and BL-99_FOS groups (P < 0.01).

After completion of the in vitro fermentation, the contents of SCFAs, including total acid(A), acetic acid(B), propionic acid(C), and butyric acid(D) were measured in the fermentation broth (Figures 2A–D). Relative to the YCFA group, the acetic acid content was markedly high in the FOS group (P < 0.05) and even higher in the BL-99_FOS group (P < 0.01). The propionic acid content reduced significantly in the FOS group (P < 0.05) and decreased further significantly in the BL-99_FOS group (P < 0.01). There were no significant differences in the contents of butyric acid and total acids among the BL-99, FOS, and BL-99_FOS groups (P > 0.05). However, the YCFA group showed the lowest total acid content (21.48 ± 7.11 μmol/mL), while the BL-99_FOS group exhibited the highest value of the total acid content (24.52 ± 3.66 μmol/mL).

16S rRNA high-throughput sequencing was performed for 40 samples from different groups on the Illumina Miseq double-end sequencing platform. A total of 3,315,045 raw sequences were obtained after filtering and splicing. The average number of reads per sample was 82,876 ± 275 reads. Subsequently, all chimeric sequences were excluded using the Usearch tool by the Uchime algorithm. The remaining sequences were clustered into OTUs at 97% similarity cut-off and unique representative OTU sequences were identified. After sequence annotation, OTUs with no results and those annotated as archaea were removed. The remaining OTUs were used for further analyses. Finally, in all samples, a total of 608 OTUs categorized into 9 phyla, 18 classes, 39 orders, 66 families, 146 genera, and 280 species were annotated.

The alpha diversity in the gut microbial composition of the different groups was evaluated using the OTUs. As shown in Figures 3A–E, no statistically significant differences in the α-diversity indexes among the groups (P > 0.05) were found. As compared with to the YCFA group (0.23 ± 0.12), the Simpson index was reduced in the BL-99 group (0.19 ± 0.13), the FOS group (0.13 ± 0.09), and the BL-99_FOS group (0.15 ± 0.12), while the Ace, Shannon, Chao, and Sobs indices were all higher in FOS and BL-99_FOS groups. These results suggested that FOS and BL-99_FOS could significantly increase the microbial alpha diversity of the intestinal microbiota.

Principal coordinate analysis based on Bray–Curtis distances was performed to identify the beta-diversity in the intestinal microbiota at the OTU level to further examine the differences in the overall microbial composition among the four groups (Figure 3F). The P-value for PCoA cluster analysis was 0.022 (P < 0.05), indicating that there were statistically significant differences in the overall microbial composition among groups. As shown in Figure 3F, a large degree of dispersion within the YCFA group, and varying degrees of aggregation in BL-99, FOS, and BL-99_FOS groups were inferred. In addition, the microbial compositions in FOS and BL-99_FOS groups were significantly different from those in the YCFA group, suggesting that constipation led to changes in the structure of the intestinal microbiota. However, FOS and BL-99_FOS could induce new homeostasis in the previously disturbed intestinal microbiota.

The differences in intestinal microbiota compositions among YCFA, BL-99, FOS, and BL-99_FOS groups were compared at the genus and phylum levels using OTU taxonomy and composition analysis (Figure 4). At the phylum level (Figure 4A), the intestinal microbiota of the population with constipation mainly comprised of Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria. Relative to the YCFA group, the relative abundances of Firmicutes and Actinobacteria were enhanced, while that of Proteobacteria reduced in the BL-99, FOS, and BL-99_FOS groups. As compared to the BL-99 group, the relative abundances of Firmicutes and Actinobacteria were increased, while that of Proteobacteria reduced in the FOS group. Additionally, these effects were more pronounced in the BL-99_FOS group.

Relative abundances of the intestinal microbiota at the genus level were observed (Figure 4B). We focused on the microorganisms with relative abundances greater than 1%, which included Escherichia-Shigella, Bifidobacterium, Bacteroides, fecalibacterium, Klebsiella, ParaBacteroides, Lactobacillus Prevotella, Collinsella, Megamonas, Enterococcus, Streptococcus, Citrobacter, Subdoligranulum, Dialister, Blautia, Megasphaera, and Phascolarctobacterium. Subsequently, the relative abundances of dominant species at the genus level were compared among different groups (Figure 4C). Relative to the YCFA group, the relative abundances of Bifidobacterium and fecalibacterium showed a highly significant increase in the FOS and BL-99_FOS groups (both P < 0.001), whereas those of Blautia and Subdoligranulum were markedly enhanced in the FOS group (P < 0.05); Blautia and Subdoligranulum in the BL-99_FOS group were further significantly higher (P < 0.01); Escherichia-Shigella in FOS and BL-99_FOS groups showed a markedly significant decrease (P < 0.01); Bilophila in the FOS group increased significantly (P < 0.001), and Bilophila in the BL-99_FOS group showed a markedly significant increase (P < 0.01).

To investigate the correlation between the intestinal microbiota, gas composition, and SCFAs, the Spearman correlation coefficient was used to evaluate the association between microbial composition at the genus level and significantly variable acetic acid, CO₂, and H₂S levels (Figure 5). The Spearman correlation heatmap showed that Bifidobacterium was markedly significantly positively correlated with acetic acid content (P < 0.001) and highly significantly negatively correlated with H₂S content (P < 0.001). Fecalibacterium was significantly positively correlated with acetic acid content (P < 0.05) and highly significantly negatively associated with H₂S content (P < 0.01); Blautia was significantly positively related to acetic acid content (P < 0.05) and highly negatively correlated with CO₂ and H₂S contents (P < 0.01). Escherichia-Shigella was highly significantly negatively correlated with acetic acid content (P < 0.01) and highly positively correlated with H₂S content (P < 0.001). Bilophila was highly significantly negatively correlated with acetic acid content (P < 0.05) and highly positively correlated with H₂S content (P < 0.001), and Lactobacillus was significantly negatively correlated with H₂S content (P < 0.001).