Screening of Lactiplantibacillus plantarum 67 with Strong Adhesion to Caco-2 Cells and the Effects of Protective Agents on Its Adhesion Ability during Vacuum Freeze Drying

Adhesion to the intestinal tract provides the foundation for Lactobacillus to exert its benefits. Vacuum freeze-drying (VFD) is currently one of the main processing methods for Lactobacillus products. Therefore, the effects of VFD on the adhesion and survival of Lactiplantibacillus plantarum 67 were investigated in this study. The results show that L. plantarum 67 exhibits remarkable tolerance following successive exposure to simulated saliva, gastric juice and intestinal juice, and also has a strong adhesion ability to Caco-2 cells. The adhesion and survival rates of L. plantarum 67 significantly decreased after VFD in phosphate-buffered saline (PBS), whereas they significantly increased in protective agents (PAs) (p < 0.05). Scanning electron microscope observations show that L. plantarum 67 aggregated more to Caco-2 cells in PAs than in PBS, and its shape and size were protected. Proteomics detection findings indicated that differentially expressed proteins (DEPs) related to adhesins and vitality and their pathways in L. plantarum 67 were significantly affected by VFD (p < 0.05). However, the expression of DEPs (such as cold shock protein, cell surface protein, adherence protein, chitin-binding domain and extracellular transglycosylase, membrane-bound protein) was improved by PAs. Compared with PBS, the PAs significantly adjusted the phosphotransferase system and amino sugar and nucleotide sugar metabolism pathways (p < 0.05). VFD decreased the adhesion and vitality of L. plantarum 67, while the PAs could exert protective effects by regulating proteins and pathways related to adhesion and vitality.


Introduction
Adhesion is a fundamental property for bacterial survival in the intestinal tract of the body. This property can help bacteria to colonize the intestinal tract, increase communication between bacteria and the gastrointestinal tract and promote probiotic functions. Therefore, a crucial aspect of the function of probiotics in vivo is the attachment of lactic acid bacteria (LAB) to intestinal epithelial cells (IECs) [1]. LAB can compete with intestinal pathogens for adhesion sites at the mucosal interface through competition, inhibition and repulsion, and the beneficial substances from its metabolism can also inhibit the invasion and colonization of pathogens [2]. It is also found that LAB can regulate immunity in vivo by activating macrophages and enhancing the activity of natural killer cells and immunoglobulin secretion [3].
The specific binding of proteins, polysaccharides, lipoteichoic acid and other adhesins on the surface of LAB to adhesion receptors in IECs or the intestinal mucus layer is essential for their adhesion [4]. Moonlighting proteins are important proteins with multiple functions, and individual polypeptide chains have two or more physiological functions, such as

Scanning Electron Microscopy
The coculture of LAB and Caco-2 cells was centrifuged at 8000× g for 10 min at 4 • C, and then the cells were resuspended in 2.5% (w/v) glutaraldehyde phosphate buffer (Solarbio Technology Co., Ltd., Beijing, China) to fix the cells. The fixed cells were dehydrated with different grades of ethanol [15]. After sputter-coating, the coculture cells were observed following the reference method [16].

Protein Extraction
The treatments were arranged as follows: C group (LAB without VFD; n = 3), N group (LAB with VFD in the PBS; n = 3) and P group (LAB with VFD in the PAs; n = 3). LAB with different treatments was centrifuged and then washed three times by PBS (pH 7.2) for protein extraction following Yang et al. [17]. The BCA Protein Assay Kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) was used to measure the concentration of the protein.
Sodium dodecyl sulfate polyacrylamide gel electrophoresis ( Figure S1) was used to detect the quality of extracted protein according to Fan et al. [18]. Briefly, the extracted proteins were mixed with 5× loading buffer and boiled for 5 min. The mixtures were loaded on 5% gel for 30 min, and then run on a 12% separating gel until they reached the bottom. The separated gel was dyed and then washed with water; finally, it was scanned by Image Scanner (GE HealthCare Technologies Inc., Chicago, IL, USA).

Protein Digestion and Tandem Mass Tag Labeling
The methods of protein digestion and tandem mass tag (TMT) labeling of peptides were performed according to the reference [19]. In detail, the extracted protein was mixed with triethylammonium bicarbonate (TEAB, Sigma, Roedermark, Germany; final concentration 100 mmol/L) and tris (2-carboxyethyl) phosphine (final concentration 10 mmol/L), which were incubated at 37 • C for 60 min. And then, iodoacetamide was added (final concentration 40 mmol/L), and reacted at 25 • C without light for 40 min. Next, precooled acetone (acetone: sample volume = 6:1) was added and centrifuged at 10,000× g for 20 min after being precipitated at −20 • C for 4 h. Precipitate was mixed with TEAB and trypsin for digestion overnight at 37 • C.
The TMT reagent (No. A44522, Thermo Fisher Scientific Inc., Waltham, MA, USA) was added to the peptide samples and incubated at 25 • C for 2 h. Then, hydroxylamine was added and was incubated at 25 • C for 30 min. At last, the products were combined in a tube and a vacuum concentrator was used to drain them.

Protein Identification and Quantification and Bioinformatics
The raw data were searched, identified and quantified following Wen et al. [20]. The database used in this study was uniprot-proteome_UP000000432_unique.fasta. Based on a false discovery rate (FDR) of less than 1%, a 95% confidence level for protein identification was applied to all data. To facilitate protein identification, a minimum of one distinct peptide identification was used. Subsequently, the differentially expressed proteins (DEPs) with thresholds of fold change (FC) > 1.5 or FC < 0.67 (p < 0.05) were identified as significantly regulated proteins. The function of each protein was identified by the Gene Ontology (GO) terms and classified by the GO enrichment analysis approach (http://beta.geneontology.org/ accessed on 22 March 2022). The functional subclasses and metabolic pathways of DEPs were identified using the Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.kegg.jp/kegg/pathway accessed on 28 March 2022) pathway enrichment analysis.

Statistical Analysis
SPSS Statistics 20 (International Business Machines Corporation, Armonk, NY, USA) was used to analyze the data, and the mean and standard deviation were used to express values. The differences were considered significant at p < 0.05. All graphs were plotted using Origin 2018 (OriginLab Corporation, Northampton, MA, USA) except for proteomicrelated graphs. The proteomic data were analyzed on the Majorbio Cloud Platform (www. majorbio.com accessed on 17 March 2022). Table 1 shows that OGT stress at different stages had different influences on the viability of LAB. The viability of LAB was slightly affected by simulated saliva alone, with which the survival rates were all more than 18%. In contrast, the survival rates were both significantly lower after sequential exposure to simulated saliva + gastric juice and simulated saliva + gastric juice + intestinal juice (p < 0.05). Only L. plantarum (67, L160, L167) and L. fermentum (L7 and L73) had a survival rate greater than 18% after simulated OGT stress in turn. We also found that S. thermophilus spp. in this study had a lower survival rate than other strains after exposure to simulated OGT stress in turn (p < 0.05). 20.80 ± 0.10 yA 13.06 ± 0.09 uB 0.04 ± 0.00 zC Significant differences in the survival rates of strains are indicated by different lowercase letters in the same column (p < 0.05). The survival rate of the same strain is denoted by different capital letters in the same row when there are significant differences (p < 0.05). The results are presented as the mean ± SD (n = 3).

Adhesion to Caco-2 Cells
The adhesion ability of LAB with strong resistance to OGT stress was tested with the Caco-2 cells. L. plantarum 67 had a substantially greater adhesion rate to Caco-2 cells (37.55%) than other strains (p < 0.05; Figure 1).   Figure 2 shows that VFD significantly affected the adhesion and viability of L. plantarum 67 to Caco-2 cells. The adhesion rate and survival rate of L. plantarum 67 significantly decreased after VFD in the PBS (p < 0.05) and significantly increased in the PAs (p < 0.05). Therefore, the adverse effects of VFD on L. plantarum 67 adhesion and viability could be mitigated by PAs. Figure 2 shows that VFD significantly affected the adhesion and viability of L. plantarum 67 to Caco-2 cells. The adhesion rate and survival rate of L. plantarum 67 significantly decreased after VFD in the PBS (p < 0.05) and significantly increased in the PAs (p < 0.05). Therefore, the adverse effects of VFD on L. plantarum 67 adhesion and viability could be mitigated by PAs.

Key Adhesin and Survival Rate of L. plantarum 67
The adhesion rate of L. plantarum 67 significantly decreased under LiCl treatment, compared with those under control (without treatment), and BSA and NaIO4 treatments before VFD (p < 0.05; Figure 2A). This indicates that surface proteins were the key adhesins of L. plantarum 67. The adhesion rate showed the same changing trend after L. plantarum 67 was vacuum freeze-dried in the PBS and the PAs, respectively (p < 0.05; Figure 2A). This indicates that surface proteins were also the key adhesin after VFD (p < 0.05; Figure 2A). Hence, it is suggested that VFD did not alter the key adhesin of L. plantarum 67.  The adhesion rate of L. plantarum 67 significantly decreased under LiCl treatment, compared with those under control (without treatment), and BSA and NaIO 4 treatments before VFD (p < 0.05; Figure 2A). This indicates that surface proteins were the key adhesins of L. plantarum 67. The adhesion rate showed the same changing trend after L. plantarum 67 was vacuum freeze-dried in the PBS and the PAs, respectively (p < 0.05; Figure 2A). This indicates that surface proteins were also the key adhesin after VFD (p < 0.05; Figure 2A). Hence, it is suggested that VFD did not alter the key adhesin of L. plantarum 67.

SEM Observation of L. plantarum 67 Adhering to Caco-2 Cells
L. plantarum 67 without VFD adhered closely to Caco-2 cells, and the L. plantarum 67 cells were rod-shaped with obtusely rounded ends and had a smooth surface, intact morphology and uniform cell size ( Figure 3A). After VFD in the PBS, L. plantarum 67 adhered loosely to the Caco-2 cells, and the cells were collapsed cell folds (red arrows) without uniform cell size ( Figure 3B). However, after VFD in the PAs, L. plantarum 67 showed higher aggregation and adhesion ability to Caco-2 cells than that in the PBS, and the cells' shapes and sizes showed little difference from those without VFD ( Figure 3C). Thus, VFD damaged the adhesion and shape of L. plantarum 67, which could be improved by PAs.
L. plantarum 67 without VFD adhered closely to Caco-2 cells, and the L. plantarum 67 cells were rod-shaped with obtusely rounded ends and had a smooth surface, intact morphology and uniform cell size ( Figure 3A). After VFD in the PBS, L. plantarum 67 adhered loosely to the Caco-2 cells, and the cells were collapsed cell folds (red arrows) without uniform cell size ( Figure 3B). However, after VFD in the PAs, L. plantarum 67 showed higher aggregation and adhesion ability to Caco-2 cells than that in the PBS, and the cells' shapes and sizes showed little difference from those without VFD ( Figure 3C). Thus, VFD damaged the adhesion and shape of L. plantarum 67, which could be improved by PAs.

Number of DEPs in L. plantarum 67 after Vacuum Freeze-Drying
A total of 416 proteins were significantly differentially expressed in the N/C group (p < 0.05; Figure 4A Figure 4C). Table 2 presents the DEPs, mainly including some chaperone proteins, moonlighting proteins and cell surface proteins, and so on.

Number of DEPs in L. plantarum 67 after Vacuum Freeze-Drying
A total of 416 proteins were significantly differentially expressed in the N/C group (p < 0.05; Figure 4A), with 233 upregulated DEPs and 183 downregulated DEPs (p < 0.05; Figure 4A). A total of 341 proteins were significantly differentially expressed in the P/C group (p < 0.05; Figure 4B), with 196 upregulated DEPs and 145 downregulated DEPs. When compared with the PAs, 53 proteins were upregulated, and 7 proteins were downregulated after L. plantarum 67 was vacuum freeze-dried in the PBS (p < 0.05; Figure 4C). Table 2 presents the DEPs, mainly including some chaperone proteins, moonlighting proteins and cell surface proteins, and so on. Table 2. Some DEPs related to adhesion and survival in L. plantarum 67 before and after VFD. The names of DEPs are shown in Table S1. "-" indicates that the proteins' FCs were not greater than 1.2 or not less than 0.83, and there was not a significant difference (p > 0.05) in this group.

Number of DEPs in L. plantarum 67 after Vacuum Freeze-Drying
A total of 416 proteins were significantly differentially expressed in the N/C group (p < 0.05; Figure 4A Figure 4C). Table 2 presents the DEPs, mainly including some chaperone proteins, moonlighting proteins and cell surface proteins, and so on.

Gene Ontology (GO) Enrichment Analysis
The GO functional analysis shows that the Biological Process (BP) categories (such as DNA metabolic processes, malonyl-CoA biosynthetic process and malonyl-CoA metabolic process) were extremely significantly enriched after L. plantarum 67 was vacuum freezedried in the PBS, compared with those without VFD (p < 0.01; Figure 5A), while aminoglycan catabolic process, "de novo" UMP biosynthetic process, glycosaminoglycan catabolic process, and peptidoglycan catabolic process were extremely significantly enriched in the PAs, compared with those without VFD (p < 0.01; Figure 5B). (C) P/N group. Column color gradients indicate the significance of enrichment, where P or FDR less than 0.001 is marked as ***, P or FDR less than 0.01 is marked as **. (C) P/N group. Column color gradients indicate the significance of enrichment, where P or FDR less than 0.001 is marked as ***, P or FDR less than 0.01 is marked as **.
The catalytic complex, external encapsulating structure and extracellular region in the Cellular Component (CC) categories were extremely significantly enriched after L. plantarum 67 was vacuum freeze-dried in the PBS and the PAs, compared with those without VFD (p < 0.01; Figure 5A,B).
The acetyl-CoA carboxylase activity, catalytic activity, acting on DNA, CoA carboxylase activity, hydrolase activity, acting on glycosyl bonds, hydrolase activity, hydrolyzing Oglycosyl compounds, ligase activity, forming carbon-carbon bonds, lysozyme activity and peptidoglycan muralytic activity in the Molecular Function (MF) categories were extremely significantly enriched after L. plantarum 67 was vacuum freeze-dried in the PBS and the PAs, compared with those without VFD (p < 0.01; Figure 5A,B). These categories contained a large number of DEPs. These results may suggest that VFD had a great influence on the molecular function of L. plantarum 67 cells.
In addition, the lysozyme activity, peptidoglycan muralytic activity, hydrolase activity, acting on glycosyl bonds and hydrolase activity, hydrolyzing O-glycosyl compounds in the MF categories were extremely significantly enriched after L. plantarum 67 was vacuum freeze-dried in the PAs when compared with those in the PBS (p < 0.01; Figure 5C), and they were also extremely significantly enriched in the N/C and P/C group. At the same time, it was found that the BP categories accounted for 15 among the top 20 most significant categories (p < 0.01; Figure 5C), and contained a large number of DEPs, which may indicate that PAs had a significant influence on the function of the biological process occurring when L. plantarum 67 was vacuum freeze-dried when compared with PBS.
A total of 3 DEPs were associated with the phosphotransferase system (PTS) (map02060, p = 0.0387) and 3 DEPs were associated with amino sugar and nucleotide sugar metabolism (map00520, p = 0.0462) after L. plantarum 67 was vacuum freeze-dried in the PAs, compared with those in the PBS ( Figure 6C).
A total of 3 DEPs were associated with the phosphotransferase system (PTS) (map02060, p = 0.0387) and 3 DEPs were associated with amino sugar and nucleotide sugar metabolism (map00520, p = 0.0462) after L. plantarum 67 was vacuum freeze-dried in the PAs, compared with those in the PBS ( Figure 6C).  The p value represents whether the enriched results are statistically significant, and the smaller the p value, the more statistically significant it is; generally, a p value less than 0.05 is considered a significant enrichment item for this function.

Discussion
After simulating saliva conditions, the survival rate of most LAB was greater than 40% ( Table 2), indicating that the pH 7.0 saliva is a suitable environment for the survival of LAB [21]; however, the fact that the expression of the hsp1 gene of LAB was upregulated may be one of the reasons that the survival rate of some strains was higher than 100% (Table 2) [22]. After successive exposure to simulated saliva and gastric juice conditions, the survival rate of LAB was significantly lower than that when exposed to simulated saliva ( Table 2; p < 0.05), which may suggest that a low pH environment had a significant impact on the growth of strains; but the activity of the proton transfer membrane ATPase and glutamic acid decarboxylase of LAB were aroused when subjected to gastric acid stress to keep the balance of intracellular and extracellular pH, thereby contributing to LAB maintaining a high survival rate ( Table 2) [23]. After successive exposure to simulated saliva, gastric juice and intestinal juice conditions, the genes related to bile stress of LAB were encoded, which may also be conducive to some strains having a higher survival rate (Table 2) [24,25].
During VFD, the intracellular enzyme outflow caused by cell dehydration easily leads to the decline of cell vitality [26], and the cells of L. plantarum 67 may collapse and rupture ( Figure 3B) with the increasing degree of dehydration. However, the collapsing and rupturing of L. plantarum 67 cells were improved ( Figure 3C) by the added PAs due to the reduction in the permeability of the cell membrane and the formation rate of ice crystals, and improvements regarding changes in ionic strength caused by electrolyte concentration inside and outside the cells [27,28].
The calcium ions in skim milk can activate the mediated signaling pathway, promote specific binding between LAB and surface receptors of Caco-2 cells, and enhance the adhesion ability of LAB to Caco-2 cells [29]. Furthermore, sodium glutamate can maintain the adhesion ability of LAB by improving the activity and expression of surface proteins during VFD [30]. It is suggested that PAs can affect the adhesion of LAB by affecting their associated proteins and pathways. Therefore, the effects of PAs (16% skim milk, 10% inulin, 4% sucrose, 7% sodium glutamate, and 12% trehalose) on the adhesion ability of L. plantarum 67 during VFD according to proteins and pathways were further investigated by proteomics in this study.

Effect of Moonlight Proteins on L. plantarum 67 Adherence and Survival
After VFD in the PBS and the PAs, the expression of 50S ribosomal proteins (L7/L12, L27; Accession: Q88YW7, Q88WN3) of L. plantarum 67 was significantly downregulated, while the expressions of EF-G (Accession: Q88XY8, F9USR0) was significantly upregulated, compared with those without VFD (p < 0.05; Table 2). These changes can reduce cellular energy consumption so as to favor the survival of L. plantarum 67 in the adverse environment of VFD [31]. They can also promote the secretion of cytoplasmic proteins into the extracellular system, thus helping to maintain the adhesion of L. plantarum 67 (Figure 2A) [7].
Stress proteins and folding proteins are the main companion proteins in LAB, including cold shock protein (Csp; Accession: F9UMI2, P96349) and small heat shock proteins (sHsps; Accession: F9USV1, F9UTM5). SHsps can prevent damage caused by cold stress [32]. The expression of Csp 2 and sHsps in L. plantarum 67 was markedly reduced in the N/C and P/C groups (p < 0.05; Table 2), while the expression of Csp 2 was significantly upregulated in the P/N group (p < 0.05; Table 2). However, Csp could improve the transcription and translation efficiency of cellular RNA and maintain the normal physiological function of LAB when the activity of other proteins was inhibited in the low temperature environment [33]. These results suggest that the adhesion and vitality of L. plantarum 67 could be improved by PAs accordingly to regulate the expression of moonlight proteins (Figure 2).

Effect of Cell Surface Proteins and Other Proteins on the Adhesion and Survival of L. plantarum 67
Cell surface protein has important physiological functions, such as maintaining cell structure and shape, regulating LAB's adhesion to IECs and participating in intercellular recognition [34]. MUB is an important adhesion protein on the surface of Lactobacillus [35], which has a characteristic secretory signal sequence and an LPXTG-anchoring motif at its N-terminus and C-terminus, respectively. MUB can realize its adhesion function by recognizing sialic acid residues of mucin chains [36,37]. In this study, the expression of cell surface proteins (CscA/DUF916 family, CscB family, LPXTG-motif cell wall anchor; Accession: F9ULM1, F9USA9, F9UU92, F9UNI8, F9UU91, F9UU93, F9UUC6, F9ULS6, F9UM21, F9US12, F9US24) and adherence protein, chitin-binding domain (Accession: F9UP60) and mucus-binding protein (MUB; Accession: F9UME2, F9UP14, F9UR18, F9USM7) in L. plantarum 67 were significantly downregulated in PBS after VFD (p < 0.05; Table 2), while being significantly upregulated in the PAs (p < 0.05; Table 2). Hence, the PAs could facilitate the synthesis of MUB and improve the expression of adhesion protein, which was conducive to maintaining the adhesion of L. plantarum 67 (Figure 2A).
The hydroxyl groups contained in trehalose and sucrose in the PAs can replace the lost water, form hydrogen bonds with the phosphate and protein polar groups in the LAB cell membrane, and wrap on the protein surface in the form of a "hydration membrane", thereby protecting the integrity of the LAB cell membrane and protein structure and function during the VFD [38,39]; and at the same time, a glassy layer is formed on the surface of the LAB cell to wrap the biological structure, limiting the intramolecular mobility of the protein and protecting the function of the surface proteins of LAB [40]. In addition, sodium glutamate in the PAs can also form hydrogen bonds with the polar groups of LAB cell surface proteins so as to maintain the structural and functional integrity of LAB surface proteins [30]. Therefore, PAs may also maintain the adhesion ability of L. plantarum 67 by protecting the activity and function of surface proteins, adhesion proteins and mucus-binding proteins.

Some KEGG Pathways Enrichment Analysis
DNA damage repair pathways include (e.g.,) mismatch repair, homologous recombination, base excision repair, and tyrosine metabolism [43]. These pathways were subjected to significant differential changes in the N/C and P/C groups (p < 0.005; Figure 6). The expression of UvrABC system protein C (Accession: Q88VE9; Nucleotide excision repair pathway), DNA mismatch repair proteins MutL and MutS (Accession: Q88UZ8, Q88UZ7; Mismatch repair pathway), DNA polymerase III subunit alpha and DNA-directed DNA polymerase III delta chain (Accession: F9UQ72; Homologous recombination pathway) were significantly upregulated by VFD (p < 0.05; Table 2). In comparison to PAs, the expression of UvrABC system protein C in PBS was noticeably higher (p < 0.05; Table 2). The results show that PBS may cause more DNA damage than PAs, and the DNA damage repair pathways were employed by L. plantarum 67 to repair DNA lesions by VFD [44,45].
The DEPs of the PTS system, N-acetylglucosamine/galactosamine-specific EIIA component (Accession: F9URE0) and Putative PTS system EIIA component (Accession: F9UUF2) in L. plantarum 67 were all enriched in the phosphotransferase system (PTS) and amino sugar and nucleotide sugar metabolism pathways, which underwent significant differential changes in the P/N group (p < 0.05; Figure 6). However, the PTS system can not only regulate the carbon metabolism, but also improve their response to cold shock [46], and the Positive PTS system EIIA component may also be related to the transport of raffinose, which can promote the adhesion of Lactobacillus to intestine [47]. At the same time, we also found that galactokinase (Accession: Q88SE8) was enriched in the amino sugar and nucleoside sugar metabolism pathway, which can synthesize the UDP-galactose through galactose-1-phase, a precursor of EPS [48]. And the three proteins above were all significantly upregulated in P/N group (p < 0.05; Table 2), indicating that PAs may regulate the related proteins and pathways to protect the vitality and adhesion of L. plantarum 67 during VFD when compared with PBS.
In the cationic antimicrobial peptide resistance pathway (map01503), the DEPs of D-alanyl carrier protein 1 (Accession: Q88VM8) was enriched, which was critical carrier proteins for the synthesis of teichoic acid and lipoteichoic acid [49]. However, the expression of D-alanyl carrier protein 1 was significantly downregulated by VFD, while being significantly upregulated in PAs (p < 0.05; Table 2). The change indicates that the expression of protein and its pathway related to teichoic acid synthesis in L. plantarum 67 was improved, thus maintaining its adhesion by PAs.
In the peptidoglycan biosynthesis pathway (map00550), peptidoglycan was finally synthetized under the action of penicillin-binding protein (ClassA PBP), serine-type D-Ala-D-Ala carboxypeptidase (LMW PBP) and zinc D-Ala-D-Ala carboxypeptidase (VanY). It was found that the expression of LMW PBP (Accession: F9UN64) in L. plantarum 67 was significantly downregulated in the N/C group (p < 0.05; Table 2) while being significantly upregulated in the P/N group (p < 0.05; Table 2). Therefore, upregulating the expression of related proteins in the peptidoglycan pathway by the PAs can improve the adhesion of L. plantarum 67 [50].
The oligopeptide ABC transporters lipoprotein-binding protein (Accession: F9USS1) was significantly upregulated in the Quorum sensing pathway (map02024) in the P/N group (p < 0.05; Table 2). However, oligopeptide ABC transporters lipoprotein-binding protein is involved in biofilm formation [51,52], and could also promote Lactobacillus adhesion to host cells [53]. Therefore, the PAs can help to maintain the adhesion of L. plantarum 67 to host cells and enhance the vitality of L. plantarum 67 during VFD by promoting biofilm formation.

Conclusions
L. plantarum 67 exhibited strong resistance to OGT stress and strong adhesion to intestinal epithelial Caco-2 cells. The adhesion and vitality of L. plantarum 67 significantly decreased due to VFD, while adding PAs had a protective effect. Furthermore, the expressions of proteins related to adhesins (such as moonlight proteins, teichoic acid, exopolysaccharides and cell surface protein, and so on) and proteins related to survival (such as cold shock protein, sHsps, and so on) in L. plantarum 67 were regulated by the PAs after VFD. The pathways related to adhesion and survival were also adjusted by the PAs to resist the damaging effects on the adhesion and vitality of L. plantarum 67 caused by VFD.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/foods12193604/s1, Figure S1: SDS-PAGE electrophoresis map of L. plantarum 67 before and after vacuum freeze-drying. Table S1: DEPs in L. plantarum 67 before and after vacuum freeze-drying. Data Availability Statement: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.