FEMPs were acquired in our lab according to the previous method. DSS (Mw, 50000 Da) was purchased from Sigma-Aldrich Trading Co., Ltd. (Shanghai). Biodewax and clear solution, 4′,6-diamidino-2-phenylindole (DAPI), and ethylenediamine tetraacetic acid (EDTA) (pH 8.0) antigen retrieval solution were bought from Wuhan Servicebio Technology Co., Ltd. Ethanol was purchased from Sinopharm Chemical Reagent Co., Ltd, and BCA protein quantitative detection kit, β-actin, and TBS buffer solution were bought from Wuhan Servicebio Technology Co., Ltd.

A total of 60 Balb/c mice (male, 8 weeks old, SPF level) were obtained from Beijing Charles River Co., Ltd (Beijing, China) and housed in the lab in the Animal Model Laboratory Building at Jilin University. The housing condition was adjusted to 20–23°C of temperature, 40–70% of humidity, and a light-dark cycle of 12 h per day. Before the beginning of the experiment, the mice were made to acclimate to the environment and were given food and water freely. All the animal procedures in the whole experiment were implemented in accordance with the guidelines for laboratory animal care and use at Jilin University. The animal experiments were reviewed and confirmed by the Jilin University animal ethics committee (Approval No. 20200483). The entire experiment's anesthetic-required procedures were performed under the influence of isoflurane.

For the allocation of the 60 mice (22–24 g), the five groups were settled, which were named CK (the short name of control check), CK + FDP (control check and drink FEMPs freely), DSS + FDW (dextran sulfate sodium and drink water freely), DSS + GP (DSS and gavage FEMPs), and DSS + FDP (DSS and drink FEMPs freely). The integral experiment period was divided into the intestinal damage-making period (period 1, days 0–7) and the treatment period (period 2, days 7–21). During period 1, the DSS, the DSS + GP, and the DSS + FDP groups were made to drink 3% DSS solution instead of making them drink water to run the intestinal-damaged colitis model. Meanwhile, the CK and the CK + FDP groups were provided with free food and water. During period 2, the DSS + GP group was asked to obtain 200 mg/kg/day FEMPs by gavage. The CK + FDP and the DSS + FDP were required to drink FEMPs freely for 4 h and drink water for the other 20 h per day. The DSS + FDW group was asked to drink water. All the groups were given free mouse food, and the body status of all the mice was recorded during the experiment periods. The mice were euthanized according to animal procedures and guidelines. The colons were harvested and weighted, then cryopreserved at −80 °C.

Incubate sections were deparaffinized and rehydrated with xylene and absolute ethanol before being placed in an EDTA antigen retrieval buffer with a pH value of 8.0. Maintained at a sub-boiling temperature for 8 min, we washed the sections three times with PBS (pH 7.4). Then, 3% BSA was added here to cover the marked tissue and block non-specific binding in 30 min. After that, the primary antibody and the secondary antibody were provided. DAPI counterstain was carried out in the nucleus, threw away the liquid carefully, then the slips were covered with the anti-fade mounting medium. The incubated sections were detected and envisioned by fluorescent microscopy.

The western blot analysis was developed under the protocol of a previous study with some modifications (20). The key proteins (ZO-1 and MUC-2) in colon tissue lysates were removed by an SDS-polyacrylamide gel and transferred electrophoretically to a polyvinylidene difluoride (PVDF) membrane. We placed the transferred membrane into the TBST incubation tank for a quick wash before blocking the membrane for 30 min with a blocking buffer (5% milk) at room temperature. After that, the membrane was incubated with appropriate dilutions of primary antibodies. The membrane was incubated with a dilution of 1:5000 of conjugated secondary antibody in blocking buffer at a temperature condition in a room for 30 min. After that, the film was washed three times (5 min each time) with TBST. The acquired pictures were obtained with darkroom development techniques for chemiluminescence. The performance of ECL was in accordance with the manufacturer's description, adding ECL reagents for 1–2 min at room temperature, and the WB images were captured under the various times of exposure.

The sequence information for FEMPs was based on the database in our lab and silicon digestion. The ExPASy PeptideCutter program, a database available at https://web.expasy.org/peptide_cutter/, was used here to collect the FEMPs' information. All the proteins and peptides in fermented egg milk and the sequences of LC-MS/MS results were collected based on our previous study (15) and uploaded into the PeptideCutter database. Pepsin and trypsin were selected as the digestive enzymes to perform the digestion process. All sequences of FEMPs were uploaded to the Pharmmapper platform at http://www.lilab-ecust.cn/pharmmapper/index.php. To screen the targets, FEMPs could be regulated (21). The targets were recorded according to the Fit score. The Uniprot platform (http://www.uniprot.org/) was utilized here to filter the Homo targets. Intestinal barrier-related targets were obtained from the GeneCard database (https://www.genecards.org). Besides, the targets obtained from references linked to intestinal health were used to supplement the targets set.

To clarify whether FEMPs could regulate the relationship between targets, we established and analyzed a PPI network of the targets. The STRING database (http://string-db.org/) (22), a powerful platform for analyzing current and possible protein-protein interactions, was used to establish functional protein association networks based on computational prediction, knowledge conversion between different organisms, and interactions aggregated with other related databases. The top 53 targets were input to the STRING database to analyze the potential co-interaction, and the STRING database depicted the figure of the PPI network.

To understand the underlying effect and mechanism of intestinal barrier functions of the core targets, we performed the GO and KEGG enrichment analysis based on the Metascape software (https://metascape.org) (23). In the custom analysis process, homo sapiens was set as the analysis species. GO biological processes (BP), GO cellular components (CC), and GO molecular functions (MF) were ticked in this part. The parameters of the analysis were shown: Min overlap, 3. P-value cutoff, 0.01, Min enrichment, 1.5. All the results were visualized by a free online bioinformatics database (http://www.bioinformatics.com.cn/).

FEMPs' regulation of the pathway was analyzed by Metascape software, and the related genes were marked by the website of the KEGG pathway (https://www.genome.jp/kegg/pathway.html). The targets FEMPs could influence in the pathway were marked in a different color, and the figure was visualized via Adobe Illustrator software.

Based on the computer calculation, molecular docking, a considerable tool widely used to reveal the relationship and interaction between receptors and ligands, was used here to analyze the action of FEMPs on the targets. The calculation has been extensively utilized and confirmed to predict the interaction energy between different molecules. In the current research, molecular docking was performed based on the Autodock Tools and Autodock Vina software. Based on the node degree analyzed by the PPI network (Supplementary Table S1), the AKT showed the highest degree. Herein, the AKT target was selected as the receptor to perform the molecular docking process, and the FEMPs were developed as ligands. All the files of receptors and ligands with the PDBQT format were set as the input files for molecular docking experiments. The crystal structure of AKT (1UNQ, PDB doi: 10.2210/pdb1UNQ/pdb) was downloaded from the database of the RCSB Protein Data Bank (http://www.rcsb.org/pdb), and the organism of the protein was selected as “Homo.” All the molecules of water were excluded from the crystal structure of AKT (1UNQ) to prepare for the molecular docking process. The binding site detection was according to the reference. In this study, a grid box of 25 × 25 × 25 Å was established at the surrounding binding site location, and the grid was generated suitably for peptide docking. The flexible style of the peptides was selected. The hydrogen bonds and cooperative interactions between the AKT residues (1UNQ) and ligands were visualized and analyzed.

Previous references have confirmed the significance and effect of TJs in the occurrence and preservation of UC (24, 25). ZO-1, a kind of vital peripheral membrane protein in the colon, maintained the junction network's relationship. It could enhance the colon barrier by linking claudins, occludin, and other proteins. MUC-2 was the vital component of mucus, with the effect of forming an indispensable barrier against the pathogen in the tissue of the intestine. Besides, MUC-2 was reported to preserve the mucus layers of colon tissue. The content of MUC-2 was also related to the number of intestinal goblet cells (26), which might benefit the immune function of the colon. Our study investigated the expression of these two vital proteins by immunohistochemistry staining and western blotting.

The results of immunohistochemistry staining revealed the two vital proteins locally. As Figure 1 depicts, the nucleus in intestine epithelial cells was blue, ZO-1 was red, and MUC-2 was green. The intensity and area of immunofluorescence staining represented the expression levels of the two proteins. Figure 1 shows that in the CK and CK+FDP groups, ZO-1 and MUC-2 were distributed around the nucleus, which meant the normal expression of the two proteins. The intestinal structure was complete and sound, and the intestinal barrier function could be normal and powerful. By contrast, DSS treatment induced an obvious loss in the content of ZO-1 and a considerable decrease in the expression of MUC-2, which showed that the gut structure was damaged and weak. Without the protection of the intestinal barrier, the function of the gut might be influenced. FEMP treatment could repair the colonic damage induced by DSS and enhance the expression of two vital proteins. In addition, the expression of ZO-1 and MUC-2 was upregulated. As the picture depicts, the fluorescence intensity of ZO-1 and MUC-2 was enhanced, which indicated that the location and expression of these two significant proteins were restored.

To further determine the relationship between the observed proteins and FEMPs treatment in experimental colitis, an experiment with the western blot was performed in our study. The proteins were extracted from the colons of the mice. As shown in Figure 2, both ZO-1 and MUC-2 displayed a decreasing trend in the DSS group, which meant these two proteins weakened in the colitis body. The proteins in the DSS+GP group and the DSS+FDP group were restored, which indicated that FEMPs might benefit from the increased levels of ZO-1 and MUC-2. Both the immunohistochemistry staining and western blotting revealed the positive effect of FEMPs on the damaged intestinal barrier.

To acquire information on targets that FEMPs could regulate, we developed target matching based on the Pharmmapper website. According to the network pharmacology method, a total of 52 targets (the top 52) were contained in the FEMPs targets database. The STRING website was applied to perform the PPI network. After that, a network containing 52 nodes and 223 edges was received, and the enrichment p-value of this network was 1.0e−16 (Figure 3). In the structure of the PPI network, the nodes represented the targets, and the edges represented the interactions. The more edges a node emerged, the more important role it would play in the network. Notably, the targets AKT1, CASP3, SRC, HPGDS, and MMP9 were connected with other targets, which revealed that these targets could generate more interaction and play a linkage role. Meanwhile, the node degree illustrated the same results that the above-mentioned targets have the potential to affect other targets (Supplementary Table S1).

To further demonstrate the information on interactions and functions of the PPI, the interaction analysis of the edges was calculated by the STRING database. The edges with a combined score larger than 0.90 are listed in Table 1. The results indicated that the interaction between CASP3 and XIAP, GLO1 and HAGH, INSR and PTPN1, PIK3R1 and SRC showed a strong relationship, with a combined score of 0.999. Besides, the co-interactions between PTPN1 and SRC, AKT1 and GSK3B, AKT1 and PIK3R1, and INSR and PIK3R1 revealed an intensive influence in the PPI network. All of these results demonstrated the potential interaction among the targets.

The GO and KEGG enrichment analyses of the above targets were carried out by Metascape software to further find out their potential biological functions. The parts of BP, CC, and MF were performed. The top 35 results (14 highly enriched in BP, 10 highly enriched in CC, and 11 highly enriched in MF) are shown in Figure 4. For BPs, the results indicated the core targets participated in the endopeptidase activity involved in the apoptotic process with an enrichment score of 52.8. Meanwhile, the glutathione metabolic process was enriched with an enrichment score of 50.7. The regulation of cysteine-type endopeptidase activity (with an enrichment score of 47.8) was shown here. Besides, the results also illustrated the process of response to peptides and the peptide metabolic process, which corresponds to the FEMPs. For the parts of CC, tertiary granule lumen, ficolin-1-rich granule lumen, and ruffle membrane were contained with enrichment scores of 31.7, 23.4, and 18.0, respectively. These results indicated that the targets could participate in the regulation of cell structures. We could observe from the MF enrichment results that the core targets could affect the regular activity involved in the apoptotic process with an enrichment score of 43.5, metallopeptidase activity with an enrichment score of 22.0, and protein serine/threonine/tyrosine kinase activity with an enrichment score of 10.4.

For the KEGG pathway enrichment analysis, various signaling pathways related to cell proliferation, differentiation, morphogenesis, and apoptosis were analyzed. The results demonstrated that there were 23 pathways with a number < 5, as listed in Figure 5. The typical signaling pathway might influence the intestinal epithelial cells and intestinal barrier functions mentioned in this study. The PI3K-Akt signaling pathway, with a p-value of 4.41e−10, was enriched with the largest count of 10. These results showed that the targets could participate more in the PI3K-Akt signaling pathway. VEGF signaling pathway and EGFR tyrosine kinase inhibitor resistance were listed here with an enrichment score of 59.06 and 51.46, respectively. Meanwhile, many pathways related to intestinal health, such as the TLR signaling pathway (27), MAPK signaling pathway (28), and Rap1 signaling pathway, were enriched with a high enrichment score.

As a central and vital signal transduction pathway in many biological and physiological processes, the PI3K-Akt signaling pathway has a relationship with cell proliferation, morphology, apoptosis, migration, and synthesis (29). Besides, the PI3K-Akt signaling pathway has been confirmed to produce a relationship between intestinal health and colonic mucosa by previous references (30, 31). In our research, the connection between intestinal barrier function and the PI3K-Akt signaling pathway was confirmed again. The core targets of FEMPs could regulate and participate more in the PI3K-Akt signaling pathway. As Figure 6 depicts, cytokines, GF, RTK, FAK, PI3K, AKT, RAF1, GSK3, 4EBPs, and EIF4E were included in the PI3K-Akt signaling pathway. The result illustrated that the targets that FEMPs could regulate were distributed both upstream and downstream in the pathway, which could influence focal adhesion, protein synthesis, and cell cycle progression.

According to the result of the PPI network analysis, the target AKT revealed the highest degree, which indicated that this target might play a significant role in the PPI network and could affect other targets easily. Thus, the AKT target was chosen as the receptor in the molecular docking process. A widely known and powerful method, molecular docking, was developed here to demonstrate the interaction between the target and ligands. There have been many studies that have utilized molecular docking to define the interactions between different targets and peptides (19, 32, 33). The crystal structure of AKT (1UNQ, PDB doi: 10.2210/pdb1UNQ/pdb) was downloaded from the RCSB Protein Data Bank (http://www.rcsb.org/pdb) of the “Homo” organism (Figure 7A). The water molecules were removed during the molecular docking process. After the consummation of the calculation, the binding site was constructed. The ligands were set to bind at the site (Figure 7B). As the figure shows, the peptide was embedded in the active cavity at the docking site of the protein receptor and was fixed at the site by chemical force, forming a stable composite structure. The docking energy can numerically reflect how tightly the ligand binds to the receptor. Herein, the docking energy between peptides and the target was recorded. The result showed that FEMPs could bind to the receptor with a stable status and the top 15 FEMPs with their interaction information, as listed in Table 2. The docking energy ranged from −8.576 to −6.048 kcal mol⁻¹. The peptide with the sequence ESQNK showed a docking energy of −8.576 kcal mol⁻¹, which formed nine hydrogen bonds and two salt bridges in the combined interaction structure. For the interaction analysis of the combined structure, many residues contributed to the chemical bonds. The interactions between the peptide ESQNK and the target were visualized in Figure 7C. The interaction result revealed that GLU 114 and NH3+ ion, LYS 111 and O⁻ ion formed one salt bridge, respectively. Six residues (GLU 114, LEU 111, ALA 58, GLN 59, CYS 60, and ARG 76) could produce hydrogen bonds in conformation with other atoms.