Desulfovibrio fairfieldensis-Derived Outer Membrane Vesicles Damage Epithelial Barrier and Induce Inflammation and Pyroptosis in Macrophages

Sulfate-reducing bacteria Desulfovibrio fairfieldensis is an opportunistic pathogen that widely exists in the human intestine and can cause severe infectious diseases. However, the mechanisms contributing to its pathogenesis remain of great interest. In this study, we aim to investigate the outer membrane vesicles (OMVs) secreted by D. fairfieldensis and their pathogenic effect. The OMVs separated by ultracentrifugation were spherical and displayed a characteristic bilayer lipid structure observed by transmission electron microscopy, with an average hydrodynamic diameter of 75 nm measurement using the particle size analyzer. We identified 1496 and 916 proteins from D. fairfieldensis and its OMVs using label-free non-target quantitative proteomics, respectively. The 560 co-expressed proteins could participate in bacterial life activities by function prediction. The translocation protein TolB, which participates in OMVs biogenesis and transporting toxins was highly expressed in OMVs. The OMVs inhibited the expression of tight junction proteins OCCLUDIN and ZO-1 in human colonic epithelial cells (Caco-2). The OMVs decreased the cell viability of monocyte macrophages (THP-1-Mφ) and activated various inflammatory factors secretion, including interferon-γ (IFN-γ), tumor necrosis factor (TNF-α), and many interleukins. Further, we found the OMVs induced the expression of cleaved-gasdermin D, caspase-1, and c-IL-1β and caused pyroptosis in THP-1-Mφ cells. Taken together, these data reveal that the D. fairfieldensis OMVs can damage the intestinal epithelial barrier and activate intrinsic inflammation.


Introduction
The human gut contains a large number of microbes, which play an important role in human diseases and health. Sulfate-reducing bacteria (SRB) is a type of bacteria that exists widely in anaerobic environment, obtains energy by degrading organic matter, and reduces sulfate to sulfide. SRB is mainly composed of six categories in the human intestine, among which the most abundant species is Desulfovibrio. SRB is a major producer of hydrogen sulfide in the gut, and high concentrations of H 2 S in the gut can adversely affect the gut environment and gut microbiota through toxicity [1,2]. Studies have confirmed that SRB increase obviously in the intestine of patients with IBD and an increase in Desulfovibrio is particularly significant [3]. Among all Desulfovibrio species, Desulfovibrio fairfieldensis bic incubator. The DNA was extracted, and the 16S rRNA gene was amplified and sequenced. Then the sequence was blasted against NCBI Ref Seq for taxonomic identification.

Purification and Characterization of Bacterial OMVs
OMVs were enriched by the ultracentrifugation method [18]. Briefly, D. fairfieldensis was cultured under anaerobic conditions at 37 • C until optical density (600 nm) reached 1.5. The bacteria-free supernatant was collected by centrifugation at 12,000× g for 15 min at 4 • C and then filtered through a 0.22 µm filter. OMVs were pelleted by ultracentrifugation at 200,000× g for 2 h at 4 • C in an ultracentrifuge [19] (Hitachi CP100WX, Tokyo, Japan). After removing the supernatant, OMVs were re-suspended in sterile PBS. We selected a 100 kDa (Merck Millipore, Billerica, MA, USA) ultrafiltration membrane to remove impurities such as flagella in the crude extract of OMVs, and then used a 50 kDa (Merck Millipore, Billerica, MA, USA) ultrafiltration membrane to enrich OMVs, and resuspended the retentate with sterile PBS [20]. The protein content of OMVs was assessed by the bicinchoninic acid protein assay kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) according to the manufacturer's instructions. The particle size of OMVs was detected by a particle size analyzer (Malvern, Worcestershire, UK).

Transmission Electron Microscopy (TEM)
OMVs were added to carbon coated copper mesh grid to hold for 1 min and stained with 1% uranyl acetate for 1 min. After the mesh grid dried, OMVs were imaged with the 80 kV Transmission electron microscope (Hitachi H-7650, Tokyo, Japan) [21].

Confocal Laser Imaging
THP-1-Mϕ cells were cultured on a chamber slide, and the diluted green dye DiO (Beyotime Biotechnology, Shanghai, China) which can label the lipid structure with a final concentration of 5 µM was incubated with OMVs for 30 min at 37 • C. Subsequently, the excess dye was removed through centrifugation using a 50 kDa molecular weight ultrafiltration tube (Merck Millipore, Billerica, MA, USA). The labeled OMVs were resuspended in PBS, added to THP-1-Mϕ cells, and incubated in the cell culture medium for 60 min. The diluted red fluorescent probe phalloidin (Beyotime Biotechnology, Shanghai, China) was used to label the cytoskeleton proteins in THP-1-Mϕ cell membranes. After fixation with neutral formaldehyde, the THP-1-Mϕ cell nuclei were labeled with DAPI (Beyotime Biotechnology, Shanghai, China).

Proteomics Analysis
D. fairfieldensis and their OMVs were subjected to label-free non-target quantitative proteomics analysis (BGI Genomics, Wuhan, China). The peptides separated by liquid phase chromatography were ionized by a nanoESI source and then passed to a tandem mass spectrometer Q-Exactive HF X for DDA (Data Dependent Acquisition) mode detec- tion. The off-machine data were identified using the Andromeda engine integrated by MaxQuant, filtered at the spectrum level with PSM-level FDR <=1%, and at the protein level with protein-level FDR <=1% is further filtered. The final identified protein sequences were all from Uniprot-Desulfovibrio species number Database. Welch's t-test was used for quantitative analysis of proteomics data. Fold change > 1.5 and p < 0.05 were used as the screening criteria for significantly different proteins.

Detection of Inflammatory Factors
After incubation with OMVs for 12 h, the cell culture supernatant of THP-1-Mϕ cells was used to detect the secreted inflammatory factors using Bio-Plex Pro Human Cytokine Grp I Panel 27-plex Luminex liquid suspension chip through the Luminex 200 platform according to the manufacturer's instruction (Luminex Corporation, Austin, TX, USA).

Statistical Analysis
The data are represented by mean ± SEM. Student's t test was used for analysis of differences between groups. p < 0.05 was considered significantly different. GraphPad Prism 8.2 (La Jolla, CA, USA) was used for statistical analysis and data visualization.

Isolation and Identification of OMVs Secreted by D. fairfieldensis
We isolated Desulfovibrio from human feces with the enrichment medium, followed by molecular biological identification of the 16S rRNA gene. We obtained strains belonging to D. intestinalis, D. simplex, D. legallii, D. Piger, and D. fairfieldensis. Then, we measured the growth pH and H 2 S production ability of these strains. We found that the growth pH of D. fairfieldensis changed most obviously (Supplementary Table S1), and its H 2 S production ability was also the highest (Supplementary Table S2). Due to the association with gastrointestinal tract infections, D. fairfieldensis has been suggested to have more pathogenic potential than other Desulfovibrio [6]. We focused on D. fairfieldensis in the following experiment and we successfully purified OMVs from D. fairfieldensis. TEM showed that these OMVs were spherical and displayed a characteristic bilayer lipid structure (Figure 1a,b). The average hydrodynamic diameter of OMVs is 74.68 nm, as determined by the particle size analyzer (Figure 1c). The expression of OmpF, which is a positive marker of OMVs, confirmed that OMVs originated from blebbing of the outer membrane of D. fairfieldensis (Figure 1d). by molecular biological identification of the 16S rRNA gene. We obtained strains belonging to D. intestinalis, D. simplex, D. legallii, D. Piger, and D. fairfieldensis. Then, we measured the growth pH and H2S production ability of these strains. We found that the growth pH of D. fairfieldensis changed most obviously (Supplementary Table S1), and its H2S production ability was also the highest (Supplementary Table S2). Due to the association with gastrointestinal tract infections, D. fairfieldensis has been suggested to have more pathogenic potential than other Desulfovibrio [6]. We focused on D. fairfieldensis in the following experiment and we successfully purified OMVs from D. fairfieldensis. TEM showed that these OMVs were spherical and displayed a characteristic bilayer lipid structure ( Figure  1a,b). The average hydrodynamic diameter of OMVs is 74.68 nm, as determined by the particle size analyzer (Figure 1c). The expression of OmpF, which is a positive marker of OMVs, confirmed that OMVs originated from blebbing of the outer membrane of D. fairfieldensis ( Figure 1d).

Proteomic Characterization of D. fairfieldensis and Its OMVs
We identified 1496 and 916 proteins from D. fairfieldensis and its OMVs by HPLC-MS/MS analysis, respectively. Venn diagram displayed that 560 proteins were co-expressed in bacteria and its OMVs (Figure 2a). The proteins isolated from the parent bacteria and OMVs were predicted to be derived from cytoplasm, inner-membrane, periplasm, outer-membrane, and extracellular matrix. According to the recent hypothesis of extracellular microvesicle formation mechanism, it can be predicted that D. fairfieldensis can form OMVs by explosive lysis. A total of 50 differentially expressed proteins were identified (Figure 2b, Supplementary Table S3). Among them, five proteins were upregulated in the

Proteomic Characterization of D. fairfieldensis and Its OMVs
We identified 1496 and 916 proteins from D. fairfieldensis and its OMVs by HPLC-MS/MS analysis, respectively. Venn diagram displayed that 560 proteins were co-expressed in bacteria and its OMVs (Figure 2a). The proteins isolated from the parent bacteria and OMVs were predicted to be derived from cytoplasm, inner-membrane, periplasm, outermembrane, and extracellular matrix. According to the recent hypothesis of extracellular microvesicle formation mechanism, it can be predicted that D. fairfieldensis can form OMVs by explosive lysis. A total of 50 differentially expressed proteins were identified (Figure 2b, Supplementary Table S3). Among them, five proteins were upregulated in the OMVs, including translocation protein TolB, which participates in OMVs biogenesis and transporting toxins (Figure 2b).
We then performed metabolic pathway analysis on the differential proteins between OMVs and D. fairfieldensis. The metabolic pathways enriched by these differential proteins have significant differences, including microbial metabolism involved in environmental changes (microbial metabolism in diverse environments), amino acid biosynthesis pathway (biosynthesis of amino acids), and antibiotic biosynthesis pathway (biosynthesis of antibiotics), metabolic pathways (metabolic pathways), and RNA-polymerase-related pathways (RNA polymerase) (Figure 2c).  OMVs. Red dots represent proteins significantly upregulated in OMVs, blues dots represent proteins significantly down-regulated in OMVs, and gray dots represent proteins without significant changes when compared with D. fairfieldensis. (c) Metabolic pathways enriched by differential proteins between OMVs and D. fairfieldensis. The X-axis enrichment factor (Rich Factor) is the number of differential proteins annotated to the pathway divided by all the proteins identified in the pathway. The larger the value, the greater the proportion of differential proteins to the pathway annotation. The dot size in the figure represents the number of differential proteins annotated to the pathway.
We then performed metabolic pathway analysis on the differential proteins between OMVs and D. fairfieldensis. The metabolic pathways enriched by these differential proteins have significant differences, including microbial metabolism involved in environmental changes (microbial metabolism in diverse environments), amino acid biosynthesis pathway (biosynthesis of amino acids), and antibiotic biosynthesis pathway (biosynthesis of antibi- Proteins involved in translation, ribosomal structure, biogenesis, as well as energy production and conversion were present in the D. fairfieldensis and OMVs through protein function prediction (Figure 3a,b). Proteins involved in cytoskeleton category only exist in D. fairfieldensis (Figure 3a). In addition, proteins involved in the transport and metabolism of carbohydrates and coenzymes in the OMVs reflected that the OMVs could participate in and interfere with the life-related activities of the host (Figure 3b). Proteins involved in translation, ribosomal structure, biogenesis, as well as energy production and conversion were present in the D. fairfieldensis and OMVs through protein function prediction (Figure 3a,b). Proteins involved in cytoskeleton category only exist in D. fairfieldensis (Figure 3a). In addition, proteins involved in the transport and metabolism of carbohydrates and coenzymes in the OMVs reflected that the OMVs could participate in and interfere with the life-related activities of the host (Figure 3b).

Disruption of the Tight Junction Structure of Intestinal Epithelium by D. fairfieldensis OMVs
The tight junction proteins, including OCCLUDIN, ZO-1, and ZO-2, play an essential role in the gut barrier and mucosal repair [22,23]. We found that a higher amount of OMVs (0.5-4 µg/mL) significantly inhibited the proliferation of Caco-2 cells, but the cell viability rate was still above 80% (Figure 4a). We also found that the blank culture medium D. fairfieldensis (Med) did not impact the gene expression of tight junction proteins in Caco-2 cells (Figure 4b). Although the corresponding amount of the D. fairfieldensis culture supernatant (Df_Med) downregulated the relative gene expressions of ZO-1, ZO-2, and OCCLUDIN, its OMVs (1 µg/mL) exhibit a more substantial gene expression inhibition effect (Figure 4b). We further confirmed that OMVs markedly reduced the protein level of ZO-1 and OCCLUDIN in Caco-2 cells compared with other groups by Western blotting (Figure 4c). These data indicated that D. fairfieldensis OMVs impair the expression of tight junction proteins of the intestinal epithelium, which may facilitate their transfer to the host. The tight junction proteins, including OCCLUDIN, ZO-1, and ZO-2, play an essenti role in the gut barrier and mucosal repair [22,23]. We found that a higher amount of OMV (0.5-4 μg/mL) significantly inhibited the proliferation of Caco-2 cells, but the cell viabili rate was still above 80% (Figure 4a). We also found that the blank culture medium D fairfieldensis (Med) did not impact the gene expression of tight junction proteins in Caco cells (Figure 4b). Although the corresponding amount of the D. fairfieldensis culture supe natant (Df_Med) downregulated the relative gene expressions of ZO-1, ZO-2, and OC CLUDIN, its OMVs (1 μg/mL) exhibit a more substantial gene expression inhibition effe (Figure 4b). We further confirmed that OMVs markedly reduced the protein level of ZO 1 and OCCLUDIN in Caco-2 cells compared with other groups by Western blotting (Fi ure 4c). These data indicated that D. fairfieldensis OMVs impair the expression of tig junction proteins of the intestinal epithelium, which may facilitate their transfer to th host.

Pyroptosis of THP-1 Macrophages Caused by D. fairfieldensis OMVs
Pyroptosis is a regulated death pathway of cells [24,25]. Its main feature is that Gasdermin D protein is activated by caspase-1 cleavage to form peptides containing Gasdermin D N-terminal active domain, resulting in cell membrane perforation, cell rupture, and the release of contents, resulting in inflammation [26][27][28]. The TEM images of THP-1-Mφ cells stimulated by OMVs showed that the nuclear chromatin diffused and distributed around the nuclear membrane. At the same time, inflammasomes wrapped in multilayer membranes appeared. The microvilli of THP-1-Mφ cells were sparse, and vesicles were exuding in the cells, which was consistent with the characteristics of pyroptosis (Figure 6a). The OMVs induced the expression of the major inflammasome components,

Pyroptosis of THP-1 Macrophages Caused by D. fairfieldensis OMVs
Pyroptosis is a regulated death pathway of cells [24,25]. Its main feature is that Gasdermin D protein is activated by caspase-1 cleavage to form peptides containing Gasdermin D N-terminal active domain, resulting in cell membrane perforation, cell rupture, and the release of contents, resulting in inflammation [26][27][28]. The TEM images of THP-1-Mϕ cells stimulated by OMVs showed that the nuclear chromatin diffused and distributed around the nuclear membrane. At the same time, inflammasomes wrapped in multilayer membranes appeared. The microvilli of THP-1-Mϕ cells were sparse, and vesicles were exuding in the cells, which was consistent with the characteristics of pyroptosis (Figure 6a). The OMVs induced the expression of the major inflammasome components, including cleaved-Gasdermin D, caspase-1, and c-IL-1β in THP-1-Mϕ cells (Figure 6b). The addition of the caspase inhibitor Z-VAD-FMK can inhibit the cleavage of Gasdermin D and pro-IL-1β ( Figure 6b). These results indicated that D. fairfieldensis OMVs could induce inflammatory cell death by activating the pyroptosis pathway.
R PEER REVIEW 10 of 13 including cleaved-Gasdermin D, caspase-1, and c-IL-1β in THP-1-Mφ cells (Figure 6b). The addition of the caspase inhibitor Z-VAD-FMK can inhibit the cleavage of Gasdermin D and pro-IL-1β (Figure 6b). These results indicated that D. fairfieldensis OMVs could induce inflammatory cell death by activating the pyroptosis pathway.

Discussion
OMVs can participate in regulating the life activities of bacteria and can be used as a messenger of bacteria to communicate with the host [29,30]. In this study, D. fairfieldensis OMVs were identified and characterized. The functional analysis of the protein composition of OMVs revealed that its protein composition was roughly similar to that of the parent bacteria. D. fairfieldensis OMVs destroyed the tight junction barrier of intestinal epithelial cells. D. fairfieldensis OMVs also significantly induced the production of inflammatory factors and caused cell death of macrophages. These results provide a theoretical basis for further research on the physiological and pathological functions of the D. fairfieldensis OMVs.
The different protein profiles of D. fairfieldensis and its secreted OMVs indicate that they may have different biological functions. To confirm the secretion mode of OMVs, we located the proteins present in OMVs and found that they contained proteins in the outer membrane, inner membrane, periplasm, and cytoplasm. Therefore, we speculated that D. fairfieldensis produced OMVs through explosive lysis. In addition, we found that the proteins related to the carbohydrate and coenzyme transport and metabolism reflect to a certain extent that it participates in the function of bacteria and interferes with the life-related activities of the host. For example, the translocation protein TolB, which was upregulated in OMVs, plays an important role in the toxic effects of OMVs. TolB belongs to the Tol-Pal system, a highly conservative membrane system in Gram-negative bacteria, which is also an essential component of OMVs. The Tol-Pal system plays an important role in maintaining the stability and integrity of the cell membrane [31,32]. It is an essential device for pathogens to transport toxins and can also cause the host immune response [33]. Previous studies have shown that OMVs can promote bacterial colonization and regulate immune response [34,35]. The specific function of these OMVs may be related to the differential proteins in the executive function of OMVs. We found higher doses of OMVs lead to more cell death in THP-1-Mφ macrophages, either due to increased TolB proteins, or due to increased pyroptois. Whether TolB was the essential pathogenic agent in D. fairfieldensis OMVs needs further exploration.
As an important pathogenic agent secreted by Gram-negative bacteria, OMVs can easily penetrate the intestinal epithelium and even transport to a distant location in the host, thereby activating downstream cell signals and promoting the development of the

Discussion
OMVs can participate in regulating the life activities of bacteria and can be used as a messenger of bacteria to communicate with the host [29,30]. In this study, D. fairfieldensis OMVs were identified and characterized. The functional analysis of the protein composition of OMVs revealed that its protein composition was roughly similar to that of the parent bacteria. D. fairfieldensis OMVs destroyed the tight junction barrier of intestinal epithelial cells. D. fairfieldensis OMVs also significantly induced the production of inflammatory factors and caused cell death of macrophages. These results provide a theoretical basis for further research on the physiological and pathological functions of the D. fairfieldensis OMVs.
The different protein profiles of D. fairfieldensis and its secreted OMVs indicate that they may have different biological functions. To confirm the secretion mode of OMVs, we located the proteins present in OMVs and found that they contained proteins in the outer membrane, inner membrane, periplasm, and cytoplasm. Therefore, we speculated that D. fairfieldensis produced OMVs through explosive lysis. In addition, we found that the proteins related to the carbohydrate and coenzyme transport and metabolism reflect to a certain extent that it participates in the function of bacteria and interferes with the life-related activities of the host. For example, the translocation protein TolB, which was upregulated in OMVs, plays an important role in the toxic effects of OMVs. TolB belongs to the Tol-Pal system, a highly conservative membrane system in Gram-negative bacteria, which is also an essential component of OMVs. The Tol-Pal system plays an important role in maintaining the stability and integrity of the cell membrane [31,32]. It is an essential device for pathogens to transport toxins and can also cause the host immune response [33]. Previous studies have shown that OMVs can promote bacterial colonization and regulate immune response [34,35]. The specific function of these OMVs may be related to the differential proteins in the executive function of OMVs. We found higher doses of OMVs lead to more cell death in THP-1-Mϕ macrophages, either due to increased TolB proteins, or due to increased pyroptois. Whether TolB was the essential pathogenic agent in D. fairfieldensis OMVs needs further exploration.
As an important pathogenic agent secreted by Gram-negative bacteria, OMVs can easily penetrate the intestinal epithelium and even transport to a distant location in the host, thereby activating downstream cell signals and promoting the development of the disease. Enterohemorrhagic Escherichia coli can use outer membrane vesicles to target mitochondria, resulting in decreased mitochondrial transmembrane potential, cytochrome c transfer to cytosol, and apoptosis of intestinal epithelial cells [36]. OMVs secreted by intestinal microbes from colitis rats can down-regulate UGT1A1 expression in Caco-2 cells through a macrophage-mediated mechanism, thus causing intestinal ecological imbalance [37]. OMVs from E. coli BL21 significantly reduced the expression of the tight junction protein Ecadherin in Caco-2 and HT-29 cells, resulting in increased intestinal barrier permeability [38]. After co-culture of D. fairfieldensis OMVs with Caco-2 cells, we found that OMVs inhibited the proliferation of Caco-2 cells and decreased the expression of tight junction proteins in intestinal epithelial cells. Damage to the integrity of the intestinal barrier may result in the infiltration of D. fairfieldensis OMVs into the systemic circulation.
E.coli BL21-derived OMVs induced scorch death of a variety of cells, including bone marrow-derived dendritic cells, THP-1 macrophages, and HeLa cells [39]. Our research demonstrated that THP-1-Mϕ macrophages could phagocytose D. fairfieldensis OMVs, which induced pyroptosis. OMVs can carry a variety of proteins, LPS, DNA fragments, and other pathogen-related molecules, which may cause inflammation by activating different signaling pathways [40,41]. Studies have shown that OMVs play an important role in bacterial survival and the spread of virulence to the host. However, we still need more research to reveal the pathogenic mechanisms of D. fairfieldensis and its secreted OMVs. Our study lacks the in vivo experimental part and does not explore the role of OMVs in the complex environment in vivo. Although the toxicity of OMVs has been demonstrated by in vitro cell experiments, it still needs to perform in vivo experiments to investigate their toxic dose and mechanism of action.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells12010089/s1. Table S1: Changes of pH of the culture system of Desulfovibrio species over time. Table S2: The content of H 2 S in the culture medium of Desulfovibrio species after 5 days of culture. Table S3: Differential proteins of D. fairfieldensis and its OMVs.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patients to publish this paper.

Data Availability Statement:
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.