Hangzhou evergreen Co., Ltd provided the acetic acid. Trypsin, Yeast extract, L-cysteine, Hemoglobin solution, Sodium chloride, Calcium chloride solution, KH₂PO₄, K₂HPO₄, and Magnesium sulfate solution were purchased from the Shanghai Sangon Biotech Bioengineering Co., Ltd. The reagents used were all analytically pure. PLNC8 was mixed in equimolar proportions from the high purity oligopeptide PLNC8α and PLNC8β. And PLNC8α and PLNC8β were synthesized by the GL Biochem (Shanghai) Ltd.

Morning fresh feces were collected from eight healthy subjects from Key Laboratory for Food Microbial Technology of Zhejiang Province. These subjects had not taken antibiotics in the past 3 months. They ate traditional Chinese food, had healthy dietary habits, and did not smoke or abuse alcohol. This research was open to all subjects and our research was approved by the Ethics Committee of Zhejiang Academy of Agricultural Sciences (Ethics Committee Agreement Number: 20212203-04).

Fresh adults’ feces were fermented with synthetic PLNC8 in vitro (25). The 0.8 g of fresh fecal sample were mixed with pretreated PBS (autoclavation, 0.1 mM, pH 6.8) to prepare a 10% (W/V) fecal homogenate suspension. After filtration, the fecal supernatant was transferred to a centrifuge tube and placed in an anaerobic work station (Do Whitley Scientific) for the later use. The fecal supernatant (500 μl) was inoculated in the autoclaved YCFA medium (26; 5 ml) containing the testing samples (NC83, NC830) and fermented at 37°C for 24 h. The fermentation broth (2 ml) was centrifuged in a sterile centrifuge tube at 12,000 rpm for 5 min. The supernatant was absorbed and stored in a separate EP tube in –20°C refrigerator for the detection of SCFAs, while the precipitate was used for DNA extraction and subsequent analysis of bacterial structure. This work set a low concentration of PLNC8 (NC83, 3 μM) and a high concentration of PLNC8 (NC830, 30 μM) as the experimental group and a blank group (F) as the blank control.

Before the SCFAs were determined, feces samples were collected and processed. The processing was as described in section “in vitro Fermentation of PLNC8.” Gas chromatography (GC) was used to identify SCFAs in samples from the mimic intestinal system. Propionic acid, butyric acid, acetic acid, valeric acid, and isobutyric acid among the SCFAs were detected in the mimic intestinal system and crotonic acid was used as internal standard. After homogenization of 50 mg of colon tissue with 1 ml of 6% phosphoric acid solution, all homogenates were transferred to an injection vial and sealed with a further 1 ml of the internal standard solution. The GC conditions were as follows: temperature program started at 80°C, ramped up to 220°C and held for 0.5 min; high-purity helium was the carrier gas, with a flow rate of 1 ml/min and a split ratio of 10:1; the injection temperature was 200°C, and the flame ionization detector temperature was 250°C, followed by a static headspace injection with an incubation temperature of 85°C and an incubation time of 30 min; the injection needle temperature reached 95°C, and 1 ml of sample was injected. In the selected ion monitoring mode of mass spectrometry, the ion source temperature was 250°C. The areas of resolved peaks on extracted ion chromatogram were recorded to plot the standard curve. Then, the peak area ratios were used to compute the concentration of each component in each sample using the standard curve.

Bacterial genomic DNA was obtained from fermented fecal samples using the QIAamp^® PowerFecal^® DNA kit (cat 51604) following the manufacturer’s instructions (QIAGEN, Hilden, Germany). The integrity of DNA was confirmed by 1% agarose gel electrophoresis (27). Bacterial DNA was extracted and preserved on dry ice, and immediately sent to Biomarker Technologies for high-throughput sequencing and microbial community diversity analysis. The sequencing platform was miseqpe250. With a minimum single sample data size of 30,000, the 16S rDNA amplification primers were 338F/806R, with a forward primer sequence of ACTCCTACGGGAGGCAGCAG and a reverse primer sequence of GGACTACHVGGGTWTCTAAT. To obtain the requisite high-quality sequences, the collected dates were screened, optimized, and concatenated, followed by quality control and the removal of questionable sequences. These high-quality sequences were clustered under 97% similarity. OTUs were obtained for species classification after chimeric screening. Finally, the abundance data in each sample’s OTUs were counted for further analysis.

The experimental data are manifested as mean ± standard deviation (SD) and GraphPad Prism 7.0 (GraphPad Software Inc., San Diego, CA, United States). Any differences were analyzed by one-way analysis of variance and Tukey’s multi range test and were considered significant when p < 0.05. Sequences were clustered using USEARCH (version 10.0) with 97% (default) similarity and we used 0.005% of the sequenced sequence count to filter OTUs. The DADA2 method in QIIME2 (version 2020.6) was used to denoise the data after quality control and 0.005% of the sequence count was used as a threshold to filter ASVS.

According to the European Molecular Biology Laboratory’s definition of enterotypes (28), there are three genus-level enterotypes. Enterotype 1 (ET B) is dominated by Bacteroides in the gut microbiota, enterotype 2 (ET P) is dominated by Prevotella in the gut microbiota, and enterotype 3 (ET F) is dominated by Ruminococcus in Firmicutes. ET B is engaged in the synthesis of biotin, riboflavin, pantothenic acid, and ascorbic acid, whereas ET P is involved in thiamine and folic acid synthesis. ET F is characterized by the predominance of Ruminococcus with mucin degradation activity and membrane transport of sugars (29). At the genus level, four human subjects’ gut microbiota was ET B (Figure 1B), whereas the other four were ET P (Figure 1D).

High-throughput sequencing of the number of gut bacteria in fecal samples was used to study the effect of PLNC8 on the gut microbiota. The microbial communities of three groups of fecal samples (F, NC83, and NC830) were examined using a high-throughput Illumina sequencing platform for analysis of 16S rDNA (V3 – V4 hypervariable region). Figures 1A–D shows the community composition of the phylum (A and C) and genus (B and D) levels of ET B and ET P, respectively. The dominant microbiota of ET B was Bacteroides, and the dominant microbiota of ET P was Prevotella. ET B and ET P had similar gut microbiota species at the phylum level. Both contained Firmicutes, Bacteroidetes, Actinobacteria, Proteobacteria, Fusobacteria, Verrucomicrobia, and Cyanobacteria, differing only in the abundance of their microbiota (Figures 1A,C). However, at the genus level, ET B contained Parabacteroides, Fusicatenibacter, Akkermansia, Parasutterella, Anacrostipes, and Streptococcus, genera not found in ET P. Meanwhile, ET P contained the genera Dialister, Mitsuokella, Ruminococcaceae_UCG-002, Ruminococcaceae_UCG-014, Ruminococcaceae_UCG-005, Catenibacterium, and Barnesiella, none of which were found in ET B (Figures 1B,D). In addition, at the phylum level, compared to group F, Bacteroidetes, Actinobacteria, and Proteobacteria had increased, and Verrucous and Firmicutes had decreased in the bacteriocin groups in both ET B and ET P. At the genus level, the bacteriocin groups had more Bacteroides, Bifidobacterium, Megamonas, Escherichia-Shigella, Parabacteroides, and Lactobacillus and less Streptococcus than group F in ET B. Furthermore, bacteriocins increased the abundance of Prevotella_9, Bifidobacterium, Escherichia-Shigella, Mitsuokella, and Collinsella in ET P.

Short-chain fatty acids are major fermentation products metabolized by intestinal anerobic bacteria; they have beneficial effects on the human body (30–32). Figure 2 shows that after 24 h of fermentation, the level of SCFAs did not increase. SCFAs for ET B (Figures 2A–E) and ET P (Figures 2F–J) were not significantly different. The NC83 group exhibited a greater level of isovaleric acid than the F and NC830 groups in ET B. Furthermore, the NC830 group had a higher acetic acid level than the F and NC83 groups. However, no significant changes in SCFAs levels were seen for ET P in any of the three groups.

Alpha diversity evaluates the variation of microbes in individual samples, including the Shannon, Chao, Ace, and Simpson indices (Table 1 and Figure 3; 33). Alpha diversity of ET B (Figures 3A,D) shows that the Shannon index of group F was significantly higher than that of group NC83, while the Chao, Ace, and Simpson indices were not significant between groups F, NC83, and NC830. However, ET P differed from ET B in that the Simpson index was higher in group F than in NC83 or NC830. Furthermore, the other indices did not change significantly.

Beta diversity analysis is used to analyze changes in species composition at temporal and spatial scales. Unweighted principal coordinate analysis (PCoA) was performed to assess the beta diversity of fermentation products in groups F, NC83, and NC830 at the genus level (34). Figures 4A,B shows the PCoA scores for ET B (A) and ET P (B). PCoA plots provided a visual representation of the similarity in bacterial community composition for all samples, with PCoA1 and PCoA2 being the principal coordinate components responsible for the differences between samples (PCoA1 and PCoA2 explanatory values were 31.12 and 20.75% for ET B and 37.7 and 18.93% for ET P, respectively). In Figure 4, the distance between the individual samples in the F group was short, indicating that the individual samples of the F group have similar bacterial compositions. Meanwhile, the distances between individual samples in the bacteriocin groups were relatively long, indicating greater variability in bacterial composition among their individual samples. However, there was some overlap between the F group and the bacteriocin groups, which suggested that under the influence of PCoA1 and PCoA2, the composition of gut microbiota in these groups was partially similar and partially different. Bacteriocins as anti-microbial peptides influence the microbial consortium (35).

The LDA Effect Size (LEfSe) can determine the abundance of taxa between different groups. Figures 5A–D shows the LEfSe plots of the gut microbiota of ET B (A) and ET P (B), with the taxonomic classes shown to their right, in addition to the LDA plots of the gut microbiota of ET B (C) and ET P (D). LEfSe analysis (LDA threshold of 2.0) was used to compare the relative abundances of bacteria in groups. The LEfSe plots show that the microbiota was significantly higher in the two PLNC8 groups than in group F. For ET B, at the levels of phylum and class, Bacteroidetes in NC830 group and Actinobacteria in NC83 group were significantly different from those in the F group; at the order level, Bacteroidales and Desulfovibrionales were significantly enriched in the NC830 group, Bifidobacteriales and Propionibacteriales were significantly enriched in the NC83 group, Clostridiales, Erysipelotrichia, Xanthomonadales, and Coriobacteriales in the F group were significantly different from those in the other groups; at the family level, Bacteroidaceae and Desulfovibrionales in the NC830 group, and Bifidobacteriaceae and Propionibacteriales in the NC83 group were significantly enriched; at the genus level, Bacteroides in the NC830 group, Bifidobacterium, Cutibacterium, and Bilophila in the NC83 group, and Agathobacter, Streptococcus, Stenotrophomonas, Anaerostipes, Tyzzerella_4, and Eubacterium_hallii_group in the F group were significantly enriched. For ET P, at the class level, Actinobacteria and Gammaproteobacteria were significantly enriched in the NC830 group, and Clostridia and Alphaproteobacteria were significantly enriched in the F group; at the phylum level, Actinobacteria and Proteobacteria in the NC830 group were significantly different from those in the F group; at the order level, Bifidobacteriales and Enterobacteriales were enriched in the NC830 group, Clostridiales and Actinomycetales were significantly enriched in the F group; at the family level, Bifidobacteriaceae, Enterobacteriaceae, and Tannerellaceae in the NC830 group, Actinomycetaceae, Rhizobiaceae, and Lachnospiraceae in the F group were significantly enriched; at the genus level, Bifidobacterium and Parabacteroides in the NC830 group, Escherichia Shigella in the NC83 group, Subdoligranulum, Fournierella, and Anaerostipes in the F group were significantly enriched.

The correlations between strains and SCFAs were compared and assessed in accordance with the classification level. Figure 6 shows that there were significant differences in the correlation coefficients between the content of most SCFAs and species compositions. The results showed that Ruminiclostridium_5, Hemophilus, and Eggerthella had significantly different levels of correlation with isovaleric acid for ET B (Figure 6A). This indicates that under the exposure to different concentrations of PLNC8, these taxa in the gut microbiota can influence the production of acetic acid, valeric acid, propionic acid, butyric acid, and isobutyric acid, but they have no significant effect on isovaleric acid.

At the genus level, according to the results, Weissella, Ruminiclostridium_5, Rothia, Prevotella_9, Odoribacter, Megamonas, Lactobacillus, Lachnoclostridium, Holdemania, Fusicatenibacter, Flavonifractor, Escherichia-Shigella, Erysipelatoclostridium, Eggerthella, Dialister, Akkermansia, Actinomyces, [Ruminococcus]_torques_group, and [Ruminococcus]_gnavus_group affected the productions of SCFAs. The correlation between SCFAs and the gut microbiota of ET P was slightly different from that of ET B (Figure 6B). Propionic acid was negatively correlated with species composition, and isobutyric acid, isovaleric acid, valeric acid, acetic acid, and butyric acid were positively correlated with the gut microbiota, especially Slackia, Ruminococcaceae_UGG-002, Ruminococcaceae_UGG-014, Ruminococcaceae_2, Prevotella_9, Peptococcus, Fournierella, Desulfovibrio, Barnesiella, and Alistipes. Overall, the positive correlation between gut microbiota and SCFAs was more visible in ET P than in the other enterotypes.