(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



SssP1, a Fimbria-like component of Streptococcus suis, binds to the vimentin of host cells and contributes to bacterial meningitis [1]

['Zihao Pan', 'College Of Veterinary Medicine', 'Nanjing Agricultural University', 'Nanjing', 'Oie Reference Laboratory For Swine Streptococcosis', 'Moe Joint International Research Laboratory Of Animal Health', 'Food Safety', 'Nanjing China', 'Peijuan He', 'Shanxi Animal Health']

Date: 2022-10

Streptococcus suis (S. suis) is one of the important pathogens that cause bacterial meningitis in pigs and humans. Evading host immune defences and penetrating the blood-brain barrier (BBB) are the preconditions for S. suis to cause meningitis, while the underlying mechanisms during these pathogenic processes are not fully understood. By detecting the red blood and white blood cells counts, IL-8 expression, and the pathological injury of brain in a mouse infection model, a serine-rich repeat (SRR) glycoprotein, designated as SssP1, was identified as a critical facilitator in the process of causing meningitis in this study. SssP1 was exported to assemble a fimbria-like component, thus contributed to the bacterial adhesion to and invasion into human brain microvascular endothelial cells (HBMECs), and activates the host inflammatory response during meningitis but is not involved in the actin cytoskeleton rearrangement and the disruption of tight junctions. Furthermore, the deletion of sssP1 significantly attenuates the ability of S. suis to traverse the BBB in vivo and in vitro. A pull-down analysis identified vimentin as the potential receptors of SssP1 during meningitis and following Far-Western blot results confirmed this ligand-receptor binding mediated by the NR2 (the second nonrepeat region) region of SssP1. The co-localisation of vimentin and S. suis observed by laser scanning confocal microscopy with multiplex fluorescence indicated that vimentin significantly enhances the interaction between SssP1 and BBB. Further study identified that the NR 216-781 and NR 1711-2214 fragments of SssP1 play critical roles to bind to the BBB depending on the sialylation of vimentin, and this binding is significantly attenuated when the antiserum of NR 216-781 or NR 1711-2214 blocked the bacterial cells, or the vimentin antibody blocked the BBB. Similar binding attenuations are observed when the bacterial cells were preincubated with the vimentin, or the BBB was preincubated with the recombinant protein NR 216-781 , NR 1711-2214 or sialidase. In conclusion, these results reveal a novel receptor-ligand interaction that enhances adhesion to and penetration of the BBB to cause bacterial meningitis in the S. suis infection and highlight the importance of vimentin in host-pathogen interactions.

Streptococcus suis (S. suis) is considered an important zoonotic pathogen capable of causing meningitis in humans. Penetrating the blood-brain barrier (BBB) is one of the preconditions for S. suis to cause meningitis, while its underlying mechanism is incompletely understood. Here we identified a previously uncharacterised pathogenic mechanism associated with S. suis meningitis mediated by the interaction between bacterial SRR glycoproteins and a host cytoskeletal component. During the bacterial infection, SRR protein SssP1 is exported to assemble a fimbria-like component, which drives a strong binding effect with the BBB depending on the sialylation of vimentin. This interaction contributes to the bacterial adhesion to and penetration of the BBB and induces a robust inflammatory response during meningitis. This overall observation underscores the significance of host cell surface vimentin interactions in microbial pathogenesis and markedly improves our understanding of host barrier penetration during meningitis.

In this study, our group identified a previously uncharacterised pathogenic mechanism associated with meningitis mediated by the interaction between a bacterial SRR glycoprotein and a host cytoskeletal component. During the bacteremia infection by S. suis, SRR protein SssP1 is exported to assemble a fimbria-like component, which drives a strong binding effect with the vimentin of BBB. This interaction contributes to the bacterial adhesion to and penetration of the BBB and induces a robust inflammatory response during meningitis.

Vimentin is broadly expressed in mesenchymal cells of mesodermal origin, such as platelets, neutrophils, activated macrophages, vascular endothelial cells, and brain microvascular endothelial cells [ 21 – 26 ], and historically is viewed as a cytosolic intermediate filament protein that forms static cytoskeletal networks important for cell structural integrity [ 27 , 28 ]. Interestingly, vimentin also involves in many cellular processes such as cell adhesion, cellular signalling, autophagy, and bacterial infections [ 29 – 32 ]. Several studies have shown that vimentin can activate the inflammation pathways by regulating the activity of NOD2 and NLRP3 inflammasome [ 33 – 35 ]. The surface proteins of bacterial pathogens play critical roles in causing meningitis via the interaction with the host components (such as the vimentin) [ 36 ]. For example, the BspC of Group B Streptococcal interacts with vimentin to promote adherence to brain endothelium to cause meningitis [ 37 ]. The surface-localised vimentin can mediate the adhesion of Listeria monocytogenes and Escherichia coli to HBMECs via their virulence factor InlF and IbeA, respectively [ 27 , 38 ].

In previous studies, we identified the S. suis strain CZ130302 that could cause acute meningitis with the representative symptoms of central nervous system infection (by video records) in the mouse infection model [ 15 ]. We then confirmed this phenotype that significantly correlated with a 4,647 aa Serine-rich repeat (SRR) glycoprotein, which transported by the SecY2/A2 (also known as accessory Sec) system [ 16 ]. Numerous studies have demonstrated that the SRR family proteins played critical roles in bacterial pathogenicity by mediating the attachment to host cell surfaces using the nonrepeat (NR) domains [ 17 , 18 ]. In Streptococcus pneumoniae, the NR domain of PsrP binds to the lung epithelial cell surface factor keratin 10, thus significantly promoting the pneumococcal infection [ 19 ]. While the NR domains of Hsa and GspB were confirmed to enable Streptococcus gordonii to effectively colonise the damaged endocardium and cause infective endocarditis via the attachment to the sialylated ligands on the platelet membrane [ 20 ].

Numerous studies focus on meningitis caused by S. suis and try to clarify the underlying mechanisms tentatively. Different biological techniques were employed to screen the virulence factors required for S. suis meningitis infection, including the selective capture of transcribed sequences (SCOTS) approach on a porcine brain microvascular endothelial cell infection model [ 8 ], the bacterial transcriptomic analysis after co-incubation with the swine cerebrospinal fluid (CFS) in vitro [ 9 ], and the TnYLB-1 transposon mutagenesis strategy using a BBB infection model in vitro [ 10 ]. Although many vital genes, small RNAs (sRNA) and extracellular components were successfully identified using these techniques, only a few ones were verified to facilitate meningitis during S. suis infection experimentally, such as the sRNA rss04 regulating CPS synthesis [ 11 ], the Serine/Threonine protein kinase destroying the tight junctions (TJs) protein claudin-5 of human brain microvascular endothelial cells (HBMECs) [ 12 ], the MRP inhibiting the adherens junctions (AJs) protein p120-catenin of BBB [ 13 ], and the suilysin inducing actin rearrangement of BBB [ 14 ]. These findings revealed the processes of transcellular and paracellular traversal during S. suis crossing the BBB models in vitro, while the underlying pathogenesis of meningitis caused by S. suis in vivo is not yet understood fully since the lack of high-virulent strains and suitable animal infection models to reproduce the typical meningitis symptoms perfectly.

Streptococcus suis (S. suis) is considered an emerging zoonotic pathogen, causing severe infections in pigs and humans, in which meningitis is one of the most life-threatening manifestations [ 1 ]. To cause meningitis, S. suis must contact and traverse the blood-brain barrier (BBB), separating the circulating blood from the brain extracellular fluid in the central nervous system [ 2 ]. As an essential defensive structure, the BBB mainly comprises brain microvascular endothelial cells, neuroglial cells, and peripheral cells [ 3 , 4 ]. The processes of bacterial pathogens penetrating the BBB are multifactorial and require complicated interactions with host cells [ 5 , 6 ]. Different pathogens rely on varied mechanisms to penetrate this barrier via transcellular or paracellular routes or through infected Trojan horse mechanisms [ 7 ].

All data are represented as mean ± SEM of triplicate samples (*** P < 0.001, ** P < 0.01, *P < 0.05). (A) The adhesion levels of S. suis strain CZ130302 were detected using the HBMECs pretreated with the NR 216-781 , NR 1711-2214 and NR 1225-1781 fragments. (B) The adhesion levels of S. suis strain CZ130302 to HBMECs were detected when anti-NR 216-781 or anti-NR 1711-2214 serums were added during co-incubation at MOI 100:1. The values were normalised against cells incubated with the pre-immune serum. (C) The adhesion levels of S. suis strain CZ130302 were detected using the HBMECs pretreated with the vimentin or Recombinant-keratin-2 (KRT2). KRT2 protein was use as the negative control. The cells without any treatments were used as the blank control. (D) The adhesion levels of S. suis strain CZ130302 to HBMECs were detected when the anti-vimentin or negative serums were added during the co-incubation at the MOI 100:1. (E) Confocal microscopy observation of vimentin skeleton and bacterial tracer. The images showed that the nuclei were stained with DAPI in blue, the vimentin was stained by IIF in red, and the S. suis cells with green fluorescent plasmid pKSM410-rss29 were observed in green.

The adhesion and adhesion inhibition assays were performed to verify the pathogenic roles of SssP1 amplified by the interaction with the vimentin of HBMECs. As shown in Fig 7A and 7B , the adhesion ability significantly attenuated when the CZ130302 cells were blocked by the antiserum of NR 216-781 or NR 1711-2214 (P < 0.01, P < 0.001), or the HBMECs were preincubated with the recombinant protein NR 216-781 or NR 1711-2214 (P < 0.05, P < 0.01), while this defect was not observed in the NR 1225-1781 or anti-NR 1225-1781 serum treated sample. Following assays were performed by blocking the vimentin of HBMECs using antiserum or blocking the SssP1 of S. suis by preincubating with vimentin. The results confirmed that the adhesion ability of the CZ130302 strain to HBMECs was significantly attenuated when the bacterial cells were preincubated with vimentin ( Fig 7C ) (P < 0.001). Similar attenuations were observed when HBMECs were blocked by the vimentin antibody ( Fig 7D ) (P < 0.001). Expectedly, the above attenuations were restored in the KRT2 treated strain CZ130302 or anti-KRT2 serum treated HBMECs to a similar level with the control sample. In addition, the indicated bacterial strains with the green fluorescent plasmid pKSM410-rss29 were observed using the fluorescence microscope to visualise the bacterial adhesion of S. suis. As shown in Fig 7E , the adhesion ability was significantly attenuated to a similar level of ΔsssP1 when the vimentin antibody blocked the CZ130302. All the above results confirmed that the full virulence of S. suis to meningitis is significantly associated with the interaction between SssP1 and vimentin of HBMECs.

Many SRR proteins from diverse pathogens have been reported to utilise their siglec-like (Ig-like) domains to bind platelet sialoglycans preferentially, such as SrpA, GspB, and Hsa [ 20 ]. We then used the Web service of CD search and SWISS MODEL to analyze the potential domains within SssP1-NR region, and found that NR 1711-2214 contains an Ig-like domain and a herpes virus major outer envelope glycoprotein (BLLF1) domain, while NR 216-781 does not contain any known domains ( Fig 6A ). Otherwise, the NR 1225-1781 contains another Ig-like domain, while it could not bind to the vimentin ( Fig 6B and 6C ). Therefore, whether the cell surface vimentin contains sialic acid modifications and whether the sialic acids are necessary for the SssP1-vimentin interaction need to be addressed. The sialidase (Sigma) was used to pretreat the vimentin protein and HBMECs. The Far-Western blot assay showed that sialidase treatment abolished the NR 216-781 and NR 1711-2214 binding to vimentin ( Fig 6D ), while the hyaluronidase (Sigma) and bioled sialidase treatment had no effect on this phenotype. Significantly different from the original cells, the HBMECs pretreated by sialidase significantly reduced the NR 216-781 and NR 1711-2214 bindings ( Fig 6C ), further suggesting that both the above fragments interacted with vimentin in a sialylation-dependent manner. Furthermore, the sialidase treatment caused the bacterial adhesion of strain CZ130302 to HBMECs reduced to the similar level of ΔsssP1 ( S4 Fig ), suggesting that the sialylation of cell surface was required for the SssP1 mediated bacterial adhesion.

(A) Schematic diagram of the NR domain divided into 5 fragments according to the protein structure. (B) The Far-Western blot was performed to identify the interaction between vimentin and the indicated fragments of NR domain. (C) Contributions of the NR 216-781 and NR 1711-2214 fragments in bacterial adhesion to HBMECs. Several indicated cell samples were pretreated for 0.5 h at 37°C using the sialidase (Sigma). The images show the DAPI-stained nuclei in blue, and the adhered proteins detected by goat anti-mouse IgG-FITC in green. The NR 1225-1781 was used as control. (D) The Far-Western blot analyses of NR domains using the sialidase (Sigma) pretreated vimentin (for 0.5 h at 37°C). The hyaluronidase (Sigma) pretreated vimentin was used as control.

To determine the more accurate/smaller fragment that is required to mediate SssP1 attachment to vimentin, the NR2 domain was re-divided into five overlapping recombination fragments (NR 1-282 , NR 216-781 , NR 720-1298 , NR 1225-1781 and NR 1711–2214 ) according to its structure and function ( Fig 6A ), and immunising rabbits prepared the polyclonal antibodies against them, respectively. Far-Western blot results verified that two of the five smaller recombinant proteins, NR 216-781 and NR 1711-2214 , could directly interact with vimentin ( Fig 6B ), while the remaining three samples and the blank control could not detect the specific bands. In addition, both the proteins of NR 216-781 and NR 1711-2214 showed an adhesive ability to HBMECs detected by indirect immunofluorescence ( Fig 6C ), and the NR 1225-1781 and blank control showed the negative result correctly. These results demonstrate that the NR domains NR 216-781 and NR 1711-2214 of SssP1 can promote the interaction with HBMECs via vimentin to activate its following pathogenic roles during S. suis infection.

Further analysis showed that seven candidate proteins have the potential to interact with the NR 1-1298 and NR 1225-2214 jointly ( Fig 5B ), and the protein named vimentin showed the highest score, which represents the ability to interact with SssP1 based on the protein loads measured in the pull-down assays. Far-Western blot assay confirmed that the NR 1-1298 and NR 1225-2214 could directly interact with vimentin ( Fig 5C ), while the other two selected proteins, keratin 2 and 10, showed negative results. The green fluorescent tracer plasmid pKSM410-rss29 was constructed and introduced into a wild-type strain and ΔsssP1 for the following fluorescence microscope, respectively ( S3 Fig ), to visualise the localisation of S. suis and vimentin in brain endothelial cells. HBMECs infected with either CZ130302 or ΔsssP1 were fixed but not permeabilised to permit labelling of only extracellular bacteria and surface-expressed vimentin by incubating with antibodies to vimentin. We observed that the surface vimentin co-localised with the bacterial cells of the wild-type strain, while this was not observed for HBMECs infected with the ΔsssP1 mutant ( Fig 5D ). These results suggested that SssP1 is required for the interaction with host extracellular vimentin during S. suis infection.

(A) Schematic diagram of the domain architecture of SssP1 and the location of NR 1-1298 and NR 1225-2214 . (B) The analysis of the screened host proteins by pull-down assays of SssP1-NR 1-1298 and SssP1-NR 1225-2214 is based on the data shown in Tables 1 and S2 . The binding score represents the ability to interact with SssP1 based on the protein loads measured in the pull-down assays. (C) The Far-Western blot was performed to identify the interaction between vimentin and the indicated NR fragments of SssP1. (D) Visual observation of the localisation of vimentin and S. suis co-incubated with the HBMECs. The vimentin skeleton was stained by indirect immunofluorescence (IIF) in red, the nuclei were counterstained with DAPI in blue, and the S. suis cells with green fluorescent plasmid pKSM410-rss29 were observed in green.

Normally, most SRRPs contain two nonrepeat regions (i.e., NR1 and NR2) and two serine-rich repeat regions (i.e., SRR1 and SRR2) [ 47 ]. Like other members of the SRRPs family [ 48 ], the basic structure of SssP1 is conserved. However, the ligand-binding region (NR2 domain) between SRR1 and SRR2 is distinctive among SRRPs from different species and strains [ 47 , 49 ]. It is why different SRRPs have different interacting receptors in the pathogenic process. Therefore, to find the target receptors on the brain microvascular endothelial cells which amplify the pathogenic effects of SssP1 in adhesion, invasion, and inflammatory response, the NR2 domain of SssP1 was divided into two segments (NR 1-1298 and NR 1225-2214 ) to express by prokaryotes since the large and complication of its structure ( Fig 5A ). Then, the two segments were purified and interacted with HBMECs for the pull-down screening, respectively. The results showed that 15 and 69 candidate proteins targeted by NR 1-1298 and NR 1225-2214 were identified, respectively ( S2 Table ). Since the blank control group also detected eight proteins, although they showed lower interaction scores, we corrected the data according to the blank control and listed the proteins showing top 7 scores as the potential candidates from NR 1-1298 and NR 1225-2214 binding groups in Table 1 , respectively.

Many meningitis-causing bacterial pathogens cross the BBB by modifying or rearranging the actin cytoskeleton and disrupting the epithelial tight junction proteins (TJs) [ 41 , 42 ], including S. suis [ 43 ]. To determine whether SssP1 affects actin filaments, HBMECs were infected with CZ130302 or ΔsssP1 at the MOI 10:1 and then stained with Alexa Fluor 660 phalloidin. Cells were serum-starved to inactivate Rho family proteins and enhance the response to activators [ 44 ]. In the absence of S. suis, there is a notable paucity of actin filaments visible in HBMECs at the three time points ( Fig 4E ). In the presence of S. suis, both ΔsssP1 and wild-type strains caused the production of stress fibres, filopodia, lamellipodia, and a long protrusion of actin filament stained in HBMECs after 1 h post-infection ( Fig 4E ). However, it was clear that the distribution of actin filaments showed no apparent differences in the presence of CZ130302 and ΔsssP1 at the same time. Otherwise, ZO-1, occludin, and claudin-5 are significant components of TJs [ 45 , 46 ], and these proteins are localised at TJs of the BBB endothelial cells. Western blotting showed that S. suis caused the significant decreases of these three proteins on the infected monolayer of HBMECs at 8 h post-infection ( Fig 4F ), while there were no significant differences observed between the ΔsssP1 and wild-type strain treated cells. These data suggested that both the rearrangement of the actin cytoskeleton and disruption of TJs caused by the S. suis infection is independent of SssP1.

The higher bacterial loads of wild-type strain in brain tissue might induce higher IL-8 levels (Figs 1B and 2A ). Thus, the defect of ΔsssP1 inducing inflammatory response needs confirmation whether it was caused by indirectly reducing bacterial load or by the interaction of SssP1 with the host directly. We then measured the related data in HBMECs after interacting with the equal bacterial cells of ΔsssP1 or wild-type strains at the MOI 10:1. The results showed that the transcriptional level of IL-8 encoding gene in HBMECs interacted with ΔsssP1 after 0.5 h were significantly lower than the wild-type strain ( Fig 4C ) (P < 0.01). Furthermore, the releasing levels of IL-8 at 4 h post-infection also showed a significant decrease infected by the ΔsssP1 compared with the wild-type strain ( Fig 4D ) (P < 0.001). These observations suggested that SssP1 can directly stimulate HBMECs to produce a more robust inflammatory response.

All data are represented as mean ± SEM of triplicate samples (**** P < 0.0001, *** P < 0.001, ** P < 0.01). (A, B) The adhesion and invasion levels of the indicated S. suis strains to HBMECs were detected at the MOI 100:1. (C) The transcriptional level of IL-8 encoding gene in the HBMECs interacted with different S. suis strains were detected using an RT-qPCR assay at the indicated times post-infection. (D) The releasing levels of IL-8 in the cultural supernatant of the HBMECs infected with different S. suis strains were measured at 4 h post-infection. (E) Cytoskeletal changes of HBMECs in response to infection by S. suis CZ130302 and ΔsssP1. Cells were stained with phalloidin to show actin filaments and were imaged using confocal microscopy. Control HBMECs were incubated without bacteria. Otherwise, HBMECs were incubated with S. suis for 0.5, 1.0, and 1.5 h before phalloidin staining. (F) Levels of the three major components of TJs, ZO-1, occludin and claudin-5, were detected using a Western blot assay to the total protein samples from the HBMECs infected with the indicated strains. β-actin was used as a loading control.

The following assays were performed to explore the underlying mechanisms of SssP1 in the bacterial penetration of BBB during S. suis infection. As shown in Fig 4A , the adhesion rates of ΔsssP1 and SssP1 T182A to HBMECs were decreased by nearly 90% compared with the wild-type strain (P < 0.0001), indicating that SssP1 has a strongly adhesive ability to HBMECs, which may enhance the virulence to cause meningitis. Likewise, the abilities of ΔsssP1 and SssP1 T182A invading the HBMECs were significantly attenuated compared with the wild-type strain ( Fig 4B ) (P < 0.0001). The invasion/ahersion ratios were then measured, and the values stayed around 0.18 both in the CZ130302 and ΔsssP1 infection groups, suggesting that the attenuation of bacterial invasion in ΔsssP1 might be caused by the adhesion defect indirectly.

All data are represented as mean ± SEM of triplicate samples (**** P < 0.0001, ** P < 0.01, *P < 0.05). (A) The curves showed the TEER values of three types of models within 4 days. The HBMECs and astrocyte cell HA1800 were respectively or jointly inoculated into a transwell chamber to establish three types of BBB models in vitro ( S1 Fig ). (B) The curves showed the TEER values of the HBMEC&HA18000 joint BBB model infected with the indicated S. suis strains. (C) The penetration rates of three types of models infected with the indicated strains were measured, and the data showed the calculated value of the bacterial CFU in the bottom chamber/ upper + bottom chambers. (D) The fluorescence values of the bottom chambers at OD428 were measured to show the permeability of Fluorescein (FLU) in three type of BBB models infected with the indicated bacterial strains. The boiled CZ130302 was used as control.

The establishment of the BBB model in vitro provides a stable qualitative and quantitative evaluation tool for further studying the mechanism of bacterial meningitis [ 39 , 40 ]. In this study, the human brain microvascular endothelial cell (HBMEC) and astrocyte cell HA1800 were respectively or jointly inoculated into a Transwell chamber to establish three types of BBB models in vitro ( S1 Fig ). All three models showed significantly higher trans-endothelial electrical resistance (TEER) values than the blank control ( Fig 3A ), indicating that the barrier had already formed successfully. Significantly, the HBMEC& HA1800 model showed the highest TEER value. Thus, it was used to assess the potential roles of SssP1 in the destruction of BBB by monitoring the changing curve of TEER values. Expectedly, the deletion of sssP1 caused a significantly less decrease of TEER value compared to the wild type in the inoculated BBB model ( Fig 3B ). We then measured the bacterial load within the bottom chambers, which showed the ability of S. suis crossing the BBB in vitro. In 2 h post-infection, wild-type strain showed significantly higher penetrability than ΔsssP1 in all Transwell chamber BBB models ( Fig 3C ) (P < 0.05; P < 0.01; P < 0.0001), suggesting that SssP1 affects the ability of bacteria to penetrate the BBB in vitro. In order to detect the BBB integrity more sensitively, the sodium Fluorescein (FLU) was added to this infection model and measured the fluorescence value of the bottom chambers at OD428 to show the destructiveness of the indicated bacterial strains to the BBB. As shown in Fig 3D , the solutions showed significantly lower FLU values treated by ΔsssP1 than wild-type strain at 2 h post-infection. Similar results were observed in the human umbilical vein endothelial cell (HUVEC) model after bacterial infection ( S2 Fig ). This result indicated that the penetration of ΔsssP1 to BBB was significantly attenuated compared to the wild-type strain.

(A) The bacterial load of the wild-type and ΔsssP1 strains in the brain of the mice infected with the indicated strains. Data are represented as mean ± SEM of triplicate samples (**** P < 0.0001). (B) The curves showed the bacterial loads in CFS of the mice challenged with different S. suis strains until 20 h post-infection. Mean values and SEMs of triplicate samples are indicated. (C, D) Evans Blue (EB) permeability in brains of the mice infected with the indicated strains. EB concentration of the infected brains was measured. Data are represented as mean ± SEM of triplicate samples (** P < 0.01).

The bacterial loads of CZ130302 and ΔsssP1 in CFS and brain tissue weremeasured after tail intravenous injection in mouse infection model to evaluate the effect of the sssP1 gene on the capacity of invasion and colonisation during bacterial meningitis. As shown in Fig 2A , the bacterial load of ΔsssP1 strain in brain tissue was also notably reduced in comparison to wild-type strain at 20 h post-infection (P < 0.0001). The bacterial load of CZ130302 in CFS increased gradually to approximately 100 CFU/μL at 20 h post-infection, while CFS from the mice challenged with ΔsssP1 strain, there were no bacterial cells detected ( Fig 2B ). These results suggested that SssP1 might promote S. suis to penetrate the BBB in vivo. Thus, the Evans Blue (EB) dye permeability assay was performed to assess the integrity of the BBB in mice challenged with the indicated S. suis strains. Accordingly, the brains of mice infected with ΔsssP1 showed lower EB than the wild-type strain ( Fig 2C and 2D ) (P < 0.01). These observations suggest that the deletion of sssP1 significantly attenuates the ability of S. suis to penetrate the BBB in the mouse infection model.

Since the inoperability of constructing the complementary strain by reintroducing a 13,944 bp sssP1 gene, a mutant (SssP1 T182A ) with a stop codon at position 182 bp was constructed by artificially changing T to A, which was used as an alternative method to confirm that ΔsssP1 is a non-polar deletion mutant. (A) The red blood and white blood cell counts of cerebrospinal fluid (CFS) from the mice infected with the indicated strains. Data are represented as mean ± SEM of three independent repeats (**** P < 0.0001, *** P < 0.001). (B) Transcriptional level of IL-8 encoding gene in the brains of the mice infected with the indicated strains. Data are represented as mean ± SEM of three independent repeats (**** P < 0.0001). (C) Histopathological analysis of the brains from the mice challenged with CZ130302, ΔsssP1 or SssP1 T182A . Scale, 50 μm.

A mouse infection model challenged by tail intravenous injection was used to assess the potential roles of SssP1 in meningitis, and the cerebrospinal fluids (CFS) were collected at 12 h post-infection. As shown in Fig 1A , both the red blood cell count and white blood cell count from the ΔsssP1 and SssP1 T182A infection group were significantly reduced to a comparable level of blank control compared with the wild-type infection group. The activation of inflammatory response is a common feature in bacterial meningitis, and interleukin-8 (IL-8) is a general marker for inflammation level assessment. In this study, the brain tissues of mice infected with ΔsssP1 and SssP1 T182A showed a significantly lower transcriptional level of IL-8 than wild-type at 12 h post-infection ( Fig 1B ) (P < 0.0001), suggesting that the gene sssP1 is significantly associated with inflammation. Furthermore, there were marked lesions in mice brains challenged with CZ130302 strain, like bleeding point, cell swelling, brain vascular dilatation and congestion, and inflammatory cell infiltration, as shown under the microscope ( Fig 1C ). However, the group infected with ΔsssP1, SssP1 T182A or PBS has no obviously histopathological changes. Those results suggested that the gene sssP1 facilitates the process of S. suis causing meningitis.

The S. suis strain CZ130302, isolated from piglet meningitis case, was confirmed to cause typical meningitis symptoms in mouse infection model [ 15 ], thus was used to explore the potential mechanism of bacterial meningitis in this study. The SssP1, a 4,647 aa SRR glycoprotein (AWD32147.1) of CZ130302, was identified to assemble a fimbria-like component [ 16 ], and significantly associated with the occurrence of bacterial meningitis. Therefore, a sssP1 deletion mutant was constructed. The transcriptional levels of more than 50 upstream and downstream genes were assessed without significant changes in vivo and in vitro, which partially verified that the deletion operation did not cause significant polarity effects. It is regrettable for the inoperability of constructing the complementary strain by reintroducing a 13,944 bp sssP1 gene. Therefore, a mutant (SssP1 T182A ) with a stop codon at position 182 bp was constructed by artificially changing T to A, and identified to display similar phenotypes with the ΔsssP1 strain in growth, pathogenicity and deficiency of fimbria-like structure [ 16 ], which indicated that ΔsssP1 is a non-polar deletion mutant.

Discussion

Bacterial invasion and translocation across the BBB are critical steps during the development of meningitis, while the identities of specific factors necessary to facilitate S. suis infection have remained unclear. Here we report the discovery that both host vimentin and S. suis SssP1 are necessary for efficient penetration of the BBB and brain colonisation by S. suis. Furthermore, SssP1 drives a strong binding effect with the host vimentin, promoting the bacterial adhesion to and penetration of the BBB, and inducing a robust inflammatory response. This overall picture (Fig 8) of bacteria-host interaction markedly improves our understanding of host barrier penetration, and S. suis virulence during meningitis.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 8. Summary of the role of the SssP1-vimentin interaction in promoting meningitis. SssP1 interacts with vimentin on the surface of HBMECs to promote S. suis attachment and induce robust inflammatory responses, which contributes to destroying the BBB to cause meningitis indirectly. The transmission electron microscope (TEM) image of S. suis strain CZ130302 showing the fimbria-like construction of SssP1 proteins was observed by this study, which was is consistent with a previous study of our lab [16]. https://doi.org/10.1371/journal.ppat.1010710.g008

SssP1 is a 4,647 aa SRR glycoprotein exported by S. suis to assemble a fimbria-like component on the cell surface. Numerous studies have demonstrated that the SRR family proteins mediate the interaction with host cells using the nonrepeat (NR) domains, thus playing critical roles in bacterial pathogenicity [17,18]. Furthermore, the SRR proteins from different bacterial pathogens showed significant genetic variability, especially in the NR regions. Therefore, there is a great variety of substrates bind by the NR regions on host cells, such as keratin 4, keratin 10, and conjugated sialic acid [19,50–52]. In different S. suis strains, the SRR NR regions also showed significant variability, and the homologous NR region of SssP1 only encoded by some high-virulent S. suis isolates. In this study, vimentin was identified as the specific substrate bound by SssP1 on the surface of BBB model in vitro. Indeed, vimentin locates on the surface of numerous cells such as T cells, platelets, neutrophils, activated macrophages, vascular endothelial cells, and brain microvascular endothelial cells [21–25]. Thereinto, the T cells, platelets, neutrophils, and activated macrophages targeted by the SssP1-vimentin interaction may serve as inadvertent carriers of bacterial cells for the optimal proliferation in the bloodstream [20]. Relatively high pathogen concentrations in blood are considered a prerequisite for microorganisms to traverse the BBB [5], and the subsequent SssP1-vimentin interaction on vascular endothelial cells and brain microvascular endothelial cells may further facilitate pathogens penetrating this barrier. Vimentin on the surface has been reported to function as the primary receptor for bacterial adhesion or to activate the potential signalling pathways [53,54]. However, the vimentin also can be modified and secreted by macrophages [22] and endothelial cells [55], and its roles during microbial infections should warrant further studies.

To further explore the pathogenic roles of SssP1, we identified two fragments of the NR2 regions, NR 216-781 and NR 1711-2214 , that are necessary for the interaction with vimentin during BBB penetration. SRR proteins have been reported to utilise their siglec-like (Ig-like) domains to bind platelet sialoglycans preferentially in diverse pathogens [20]. Our previous study has confirmed that the Ig-like domains of SssP1 showed binding capacities for human sialic acids and possessed apparent adherence to HBMECs [16]. In this study, two fragments, NR 1225-1781 and NR 1711-2214 contain an Ig-like domain, respectively, while only NR 1711-2214 could interact with host vimentin and HBMECs, suggesting that the other domain of NR 1711-2214 , BLLF1 (herpes virus major outer envelope glycoprotein domain), might play vital roles for binding to vimentin. Otherwise, NR 216-781 does not contain any known domains, and its interaction with vimentin or HBMECs could be abolished by the sialidase (Sigma) treatment, suggesting that NR 216-781 bound to vimentin in a sialylation-dependent manner, while which was not mediated by the Ig-like domain. The homologous of SssP1 or other SRR proteins are widely encoded in the bacterial pathogens of Streptococcal species, and most of them could bind the surface sialoglycans of host cells. Suppose all these SRR proteins from diverse bacterial species have similar capacities to interact with the vimentin of BBB, which will significantly promote the exploration of the meningitis pathogenesis in a broader range of bacterial pathogens. Furthermore, a greater understanding of SssP1-vimentin interactions may prove applicable to other pathogens and provide significant insight and possible targets for developing novel therapeutics for meningitic infections. Our study has demonstrated that when the surface vimentin of HBMECs was blocked, the bacterial adhesion was significantly impaired, which suggests a new therapeutic approach to protect the CNS against bacterial invasion. Several recent studies have reported the strategy to block host-pathogen interaction to protect the brain from pathogen invasion during pneumococcal meningitis [56, 57]. Significantly, an increasing number of reports have demonstrated that vimentin is a central meningitic factor utilised by multiple bacterial pathogens to facilitate the crossing of the BBB and colonisation of the brain [35,38,53]. Therefore, early antibody blockade to vimentin may effectively resist the brain invasion of diverse bacterial pathogens, which provides a longer time window for proper choice of antibiotic, and thereby increase survival [57].

Except for the bacterial penetration of the BBB, paracellular migration of leukocyte and inflammatory activation occur at the BBB, which are the triad hallmark features of bacterial meningitis [54,58]. Paracellular migration of polymorphonuclear leukocytes across the BBB is not only crucial for host defence against pathogens, but it may also cause significant damage to the central nervous system tissues with a massive haemorrhage [59]. Furthermore, NF-κB signalling activation can significantly promote the central nervous system’s inflammation, damaging the brain from inflammatory injury during bacterial meningitis [60, 61]. Our study confirmed the roles of SssP1 in these three hallmark features in vivo by detecting the red blood cell count, white blood cell count, mRNA level of IL-8 encoding gene, bacterial load in CFS and brain, and pathological injury of the brain during meningitis caused by S. suis (Fig 8). Several studies have reported that vimentin plays a detrimental role in the host defence against meningitic infection by modulating the NF-κB signalling pathway to increase polymorphonuclear leukocytes recruitment and neuronal inflammation [35,38,53]. Therefore, the SssP1-vimentin interaction may facilitate brain damage from two perspectives, (i) traversing the BBB indirectly via enhancing bacterial adhesion and other unidentified pathways (ii) activating the inflammatory response within the central nervous system. Speculatively, these two perspectives may represent an important mechanism responsible for a coordinated modulation of both bacterial and host factors that contribute to the BBB penetration and bacterial diffusion within the central nervous system. Their cooperative contributions may represent a new paradigm in bacterial meningitis and lead to the development of novel strategies for preventing and treating S. suis infection.

Many meningitis-causing bacterial pathogens cross the BBB via the transcellular route by modifying or rearranging the actin cytoskeleton or via the paracellular route by disrupting the epithelial tight junction proteins (TJs) [41,42]. However, our study found that both the rearrangement of the actin cytoskeleton and disruption of TJs caused by the S. suis infection is independent of SssP1, which may be caused by: (i) SssP1 indeed does not perform the above biological functions, (ii) other virulence factors significantly facilitating the above phenotypes have obscured the actual effect of SssP1 in the above assays, and (iii) the roles of SssP1 are interrupted by several unknown reasons in these assays. These results suggest that the SssP1-vimentin interaction may contribute to traverse the BBB via indirect or unidentified pathways. Several previous studies have demonstrated that the specific interaction between IbeA and its primary receptor vimentin is necessary for the caveolae/lipid raft-dependent entry of E. coli K1 into HBMECs [38,53]. Otherwise, the IbeA-vimentin interaction mediates the polymorphonuclear leukocytes transmigration across the BBB [54], which may help bacterial pathogens hiding within the polymorphonuclear leukocytes to traverse the BBB via the infected Trojan horse mechanism. Whether similar mechanisms exist in Sssp1-vimentin interaction warrants further studies.

In summary, all the findings in this study indicate that the pathogenic features of S. suis meningitis are driven by specific interactions between the meningitic virulence factor SssP1 and its primary receptor vimentin (Fig 8). This ligand-receptor relationship provides a new angle to elucidate the pathogenesis of S. suis.

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010710

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.