Binding of Human Fibrinogen to MRP Enhances Streptococcus suis Survival in Host Blood in a αXβ2 Integrin-dependent Manner

The Gram-positive bacterium Streptococcus suis serotype 2 (S. suis 2), an important zoonotic pathogen, induces strong systemic infections in humans; sepsis and meningitis are the most common clinical manifestations and are often accompanied by bacteremia. However, the mechanisms of S. suis 2 survival in human blood are not well understood. In our previous study, we identified muramidase-released protein (MRP), a novel human fibrinogen (hFg)-binding protein (FBP) in S. suis 2 that is an important epidemic infection marker with an unknown mechanism in pathogenesis. The present study demonstrates that the N-terminus of MRP (a.a. 283–721) binds to both the Aα and Bβ chains of the D fragment of hFg. Strikingly, the hFg-MRP interaction improved the survival of S. suis 2 in human blood and led to the aggregation and exhaustion of polymorphonuclear neutrophils (PMNs) via an αXβ2 integrin-dependent mechanism. Other Fg-binding proteins, such as M1 (GAS) and FOG (GGS), also induced PMNs aggregation; however, the mechanisms of these FBP-hFg complexes in the evasion of PMN-mediated innate immunity remain unclear. MRP is conserved across highly virulent strains in Europe and Asia, and these data shed new light on the function of MRP in S. suis pathogenesis.

cell surface 16 . However, Enolase is a functional important enzyme in both prokaryotic and eukaryotic organisms 17 and its gene couldn't be knocked out in our previous study 15 . The mechanism of Enolase in immune evasion remained unclear until now. MRP is a 136-kDa protein that associates with the cell wall of S. suis via an LPxTG motif but is also released into culture supernatants 18 ; it is often reported as an epidemic infection marker for highly virulent strains in Europe and Asia and has low amino acid sequence similarity to other proteins with known functions [19][20][21] . In our previous study, rat antiserum to MRP had highly bactericidal activity against S. suis 05ZYH33 in human blood 22 , and the MRP-hFg interaction improved the antiphagocytic capacity of S. suis 2 against PMNs 15 , suggesting that MRP might be involved in the survival of the bacteria in the host blood. Therefore, the specific mechanism of MRP involvement in the evasion of innate immunity by S. suis in host blood needs to be clarified.
The results of this study indicate that MRP directly binds to the Aα and Bβ chains of the D fragment in hFg. Furthermore, the interaction of soluble (released) MRP with hFg enhanced the survival of S. suis 2 in human blood by disrupting the phagocytic activity of PMNs by inducing their aggregation in an α X β 2 integrin-dependent manner. These results provide useful insight into the molecular mechanism of S. suis pathogenesis and the biological basis of MRP as an important infection marker.

Results
Localization of the hFg-binding region in MRP. PMNs are the first line of defense against bacterial infections in the human blood. In our recent study, hFg enhanced the survival of S. suis 2 in a PMN killing assay. MRP was one of two major SsFBPs among the putative SsFBPs identified in that assay 15 . However, the specific binding mechanism remained unclear. The putative functional domain of MRP was not predicted in an analysis of the protein sequence with Interproscan (http://www.ebi.ac.uk/Tools), except for a cell wall attachment consensus motif, LPxTG ( Supplementary Fig. 1A). Like most surface proteins of Gram-positive bacteria 23,24 , MRP contains a proline-rich repeat sequence (Fig. 1A). To determine the Fg-binding region of MRP, four recombinant fragments of MRP (Fig. 1A) were engineered mainly according to its predicted secondary structure ( Supplementary Fig.  1B) by Psipred analysis (http://bioinf.cs.ucl.ac.uk/psipred) as follows: MRP-N (a.a. , the N-terminus of MRP), MRP-N1 (a.a. 48-282, the α -coil-rich region), MRP-N2 (a.a. 283-721, mainly consisting of β -sheets), and MRP-C (722-1257, a proline-rich region mainly formed by β -sheets). An interaction between recombinant MRP and immobilized hFg was identified for his MRP-N (a.a.  and his MRP-N2 (a.a. 283-721) (Fig. 1B), which increased as a function of the protein concentration. We detected no Fg-binding activity for his MRP-N1 (a.a.  or the C-terminal his MRP-C (a.a. 722-1257) (Fig. 1B). Similarly, when the system was reversed, immobilized his MRP-N and his MRP-N2, but not his MRP-N1 and his MRP-C, showed binding to Fg (Fig. 1C), indicating that a.a. 283-721 encompass the Fg-binding domain of MRP. The binding constants of his MRP-N were further evaluated using bio-layer interferometry (BLI). The K D value for the binding of his MRP-N to hFg was 340 nM (Fig. 1D), which was close to the value for the M1 protein in GAS 25 . Additionally, wild-type S. suis 05ZYH33 adhered to immobilized hFg were significantly higher than those observed with ∆ MRP bacteria or C∆ MRP-N (which express only the soluble binding region, MRP-N) (Fig. 1E), indicating that MRP is an important FBP of S. suis. These results provide important clues about the function of MRP in S. suis 2-induced pathogenesis.
Identification of the hFg-binding site of MRP. Most FBPs of pathogens have specific Fg-binding mechanisms 7 ; thus, the region within fibrinogen that is responsible for MRP binding was investigated. Fg consists of subunits D and E, each containing three polypeptide chains (Aα , Bβ and γ ). When separated by SDS-PAGE under reducing conditions, fibrinogen appeared as three bands corresponding to the Aα (63.5 kDa), Bβ (56 kDa) and γ (47 kDa) chains ( Fig. 2A). Both the Aα and Bβ chains were readily detected when transferred to a polyvinylidene difluoride (PVDF) membrane and were probed with purified biotinylated his MRP-N and his MRP-N2, but not his MRP-N1 ( Fig. 2A). To identify the MRP-binding site(s) in Fg, isolated D and E fragments of Fg (Fg-D and Fg-E, respectively) were tested in an ELISA-type binding assay. The data indicate that his MRP-N2 (Fig. 2Bi) and his MRP-N ( Fig. 2Bii) bound well to Fg-D and intact Fg, but did not bind to Fg-E. An analysis of the MRP-Fg-D interaction by BLI further confirmed that Fg-D, but not Fg-E, bound to his MRP-N2 (Fig. 2Ci) and his MRP-N (Fig. 2Cii), providing insight into a possible mechanism by which MRP alters Fg to allow S. suis 2 to evade innate immunity.
The presence of MRP in culture supernatants and the bacterial cell surface. MRP is a substrate of sortase A in a National Collection of Type Cultures (NCTC) European strain 26 . However, the protein is also detectable in the culture supernatants from serotype 2 strains isolated from European diseased pigs 16 . To further explore the function of MRP in 05ZYH33, the location of MRP was detected in 05ZYH33 cells. Western blotting indicated that MRP was detected in the culture supernatant in the 05ZYH33 strain ( Fig. 3A) but was not detected in the mutant strain (Fig. 3A). To elucidate the function of soluble (released) MRP, a complementary strain (C∆ MRP-N) was constructed for the soluble expression of MRP-N (which is identified as the hFg-binding region in Fig. 1). Western blotting analysis indicated that MRP-N was present in the supernatant from the C∆ MRP-N strain (Fig. 3A). Additionally, the surface location of MRP was visualized using confocal laser scanning microscopy in the WT strain 05ZYH33 (Fig. 3Baiii), but not in the mutant (Fig. 3Bbiii) and complementary strains (Fig. 3Bciii). These data indicated that MRP was released into culture supernatant and that the complementary strain (C∆ MRP-N) expressing the secreted MRP-N was successfully constructed.
MRP facilitates the survival of S. suis 2 in human blood in an α X β 2 integrin-dependent manner.
To investigate the role of MRP in immune evasion in human blood, the survival of mrp-deficient mutants (∆ MRP) was evaluated after the addition of his MRP truncations. Interestingly, the viability of strain ∆ MRP in human blood was increased by the exogenous addition of his MRP-N or his MRP-N2, but not his MRP-N1 or his MRP-C (Fig. 4A). . Bound proteins were detected with a monoclonal anti-his antibody (1:5000) and an HRP-conjugated anti-mouse antibody (1:10,000). The background values obtained for the control (incubated with PBST containing 1% BSA, but without truncated MRP) were subtracted. (C) Immobilized his MRP-N 48-721 , his MRP-N1 100-282 , his MRP-N2 283-721 and his MRP-C 722-1257 (100 nM) were incubated with the indicated concentrations of biotinylated hFg, and bound proteins were detected with HRP-conjugated streptavidin (1:8000). The background values obtained for the control (incubated with PBST containing 1% BSA, but without biotinylated hFg) were subtracted. The assays were performed in triplicate, and the mean ± SD is shown. (D) The binding constants of his MRP-N were evaluated by BLI. Label-free binding of hFg to immobilized his MRP-N was conducted in the Octet, which uses streptavidin-coated biosensors to detect protein:protein interactions via BLI. Baseline-corrected BLI response curves (shown in black, with the lower curve corresponding to a lower concentration of injected analyte) were globally fitted to the 1:1 binding model. Kinetic parameters obtained from the fitting are shown. (E) Binding of S. suis to immobilized hFg. The S. suis strains WT, ∆ MRP and C∆ MRP-N (~1 × 10 4 CFU per well) were incubated with wells pretreated with hFg or BSA (1, 5, 10 μ g/ml per well). The values represent triplicate wells and the mean ± SD of the CFU of binding bacteria per well.   Therefore, it is possible that the biological effects of MRP are exerted upon its release into the extracellular milieu. As expected, C∆ MRP-N rescued the survival defect of the mutant ∆ MRP bacteria in the blood, bringing the survival to levels similar to WT (Fig. 4B). Additionally, in a specific antibody-blocking assay, antiserum against MRP-N decreased the WT and C∆ MRP-N bacterial survival, but not to ∆ MRP in human blood (Fig. 4C), suggesting that MRP could be a promising candidate for antibody-based therapies for S. suis induced-sepsis. These results indicate that soluble (released) MRP is responsible for immune evasion by S. suis in human blood.
The mechanistic basis of MRP in immune evasion by S. suis 2 was further investigated in a bactericidal assay using a peptide derived from the Aα chain of Fg (Gly-Pro-Arg-Pro, GPRP), which acts as a β 2 integrin antagonist and has been shown to impair the M1-Fg complex-induced aggregation of PMNs 27 . A second peptide, Gly-His-Arg-Pro (GHRP), which does not have this effect 27 , was also tested as a control. GPRP decreased the growth of WT and C∆ MRP-N bacteria in human blood (Fig. 4D), while no effects were observed for the ∆ MRP bacteria. The control peptide GHRP or H 2 O had no effect on the survival of WT, ∆ MRP and C∆ MRP-N in human blood, suggesting that the β 2 integrins on PMNs were involved in this process. α X β 2 and α M β 2 are general integrins expressed by phagocytes; neutrophils interact with hFg through the β 2 family of integrins, with the N-terminal portion of the Aα chain of Fg binding to α X β 2 28 while the C-terminal portion of the γ chain binds to α M β 2 integrins 29 . To determine which type of β 2 integrin mediates the formation of MRP-hFg complexes for immune evasion by S. suis 2, monoclonal antibodies against the α M (clone CBRM1/5) and α x (clone BU15) integrin subunits and an isotype control mouse IgG1 were used in blocking assays. Like GPRP, the antibody against the α x integrin subunit decreased the survival of WT and C∆ MRP-N, but not ∆ MRP bacteria in human blood, while there was no interference detected using the antibody against α M or the control IgG1 (Fig. 4E).
These results suggest that the N-terminus of MRP (MRP-N) might interact with hFg and exert some effects on PMNs via α X β 2 integrin crosslinking to improve the survival of S. suis 2 in human blood.
The soluble (released) MRP was involved in the immune evasion of S. suis 2 by interacting with hFg. PMN-mediated innate immunity is the first line of defense against bacterial pathogens in human blood.
To verify the hypothesis that MRP improves the immune evasion of S. suis 2 by interacting with hFg, a simplified model of whole blood, a PMN killing assay, was performed in the presence of fresh serum, fresh plasma or fresh serum supplemented with hFg (the serum and plasma were isolated from the same donor). Interestingly, as in the whole blood, the mutant ∆ MRP bacteria had reduced viability compared to the WT bacteria in serum containing hFg at physiological concentrations (2-4 mg/ml) (Fig. 5A), and C∆ MRP-N rescued the viability of ∆ MRP to the level of WT in serum with hFg, but not in serum alone (Fig. 5A). Similarly, ∆ MRP bacteria had reduced viability compared to WT and C∆ MRP-N bacteria in plasma (Fig. 5B), but not in serum alone; additionally, supplementation of ∆ MRP bacteria with MRP-N, but not MRP-C, rescued their viability in plasma but had no effect in serum alone (Fig. 5B). These results indicate that the N-terminus of MRP (MRP-N) is responsible for the immune evasion of S. suis by interacting with hFg.

The MRP-hFg interaction results in the integrin-triggered aggregation of PMNs and PMN exhaustion.
The effects of the MRP-hFg interaction on PMNs in the PMN killing assay were further investigated by confocal laser scanning microscopy. The WT (Fig. 6A) and C∆ MRP-N (Fig. 6C) strains induced PMN aggregation in serum supplemented with hFg; however, infection by either the ∆ MRP strain (Fig. 6B) or the WT strain in serum alone had no effect (data not shown), suggesting that the aggregation was caused by the MRP-hFg interaction. Moreover, dead bacteria and PMNs (red staining by propidium iodide [PI]) were the principal constituents of these aggregates induced by the WT or C∆ MRP-N strains (Fig. 6A,C and their magnified images in Supplementary Fig. 2), and greater numbers of green extracellular viable bacteria (green staining by SYTO9) were observed in these groups (white arrows in Fig. 6A,C and in Supplementary Fig. 2) than with the ∆ MRP strain (Fig. 6B). Because soluble MRP enhanced the survival of S. suis 2 in PMN killing assays by interacting with hFg, the supernatants were assessed to determine whether they could induce PMN aggregation. As expected, the WT and C∆ MRP-N-supernatants had a similar effect on PMNs in the presence of serum supplemented with hFg (Fig. 6J,L), but the ∆ MRP supernatant did not (Fig. 6K). These results suggest that soluble (released) MRP induces PMNs aggregation and death by interacting with hFg.
Furthermore, GPRP, but not GHRP, inhibited the aggregation induced by the WT (Fig. 6D,E) and C∆ MRP-N (Fig. 6G,H) strains, as did a monoclonal antibody against α X integrin subunits (Fig. 6F,I). These results were (C) Antibodies against the MRP-N-binding region decrease S. suis viability in human blood. Bacteria were pretreated with preimmune (control) or anti-MRP-N serum from rabbits immunized with MRP-N (1:5 dilution). Significant differences were found between preimmune and anti-MRP-N sera in the WT (P < 0.001) and C∆ MRP-N (P < 0.001) groups, but not in the ∆ MRP group (P > 0.05). (D) Viability of S. suis 2 in human blood in the presence or absence of GPRP (50 μ g/ml) and GHRP (50 μ g/ml). Significant differences were found between GPRP treatment and other treatments in the WT (P < 0.05) or C∆ MRP-N (P < 0.001) groups but not in the ∆ MRP group (P > 0.05). (E) Viability of S. suis 2 in human blood as determined by a bactericidal assay after treatment with the specific anti-α M β 2 antibody (10 μ g/ml), the specific anti-α x β 2 antibody (10 μ g/ml) or their isotype control mouse IgG1 (10 μ g/ml). Significant differences were found between anti-α x β 2 treatment and other treatments in the WT (P < 0.001) and C∆ MRP-N (P < 0.001) groups but not in the ∆ MRP group (P > 0.05). In Figs. A-E, each symbol represents an individual human blood sample. In Figs. A, C, D and E, the horizontal lines indicate the mean for each group. In Figs. A, D and E, post hoc tests were used for statistical analysis. In Fig. C, an unpaired Student's t test was used. In  Fig. B, the horizontal lines indicate the median for each group, and a Mann-Whitney test was used.
consistent with the bactericidal assay in Fig. 5A,B, suggesting that the MRP-hFg interaction results in the α X β 2 integrin-triggered aggregate formation of PMNs and reduces the capacity of PMNs to clear S. suis in the blood by phagocytosis.
To further analyze PMN activation by S. suis, dihydrorhodamine (DHR)123, a fluorogenic substrate of active oxygen, was used to detect oxidative activity. The mean fluorescence intensity (MFI) as determined by flow cytometry in PMNs treated with the WT strain was twice that of PMNs treated with ∆ MRP. The MFI of PMNs was 2-fold lower in the presence of GPRP relative to GHRP upon treatment with the WT strain opsonized in hFg-supplemented serum (Fig. 7), but not in serum alone (data not shown). These results suggest that the MRP-hFg interaction results in the excessive activation and exhaustion of PMNs via a β 2 integrin-dependent mechanism.   Fig. 6. GPRP, Gly-Pro-Arg-Pro; GHRP, Gly-His-Arg-Pro; anti-α x β 2 , monoclonal antibody against α x integrin subunits; Sup, supernatant.
Scientific RepoRts | 6:26966 | DOI: 10.1038/srep26966 Discussion S. suis is an emerging pathogen and a leading cause of skin wounds and bloodstream, brain, and soft tissue infections 30,31 . The most common clinical manifestations of infection are sepsis and meningitis 30 , which are often accompanied by bacteremia, as is the case for Streptococcus pneumoniae-and Neisseria meningitidis-associated meningitis 32 . Outbreaks of S. suis 2 infection in China have highlighted the importance and relatively high frequency of a severe sepsis syndrome that shares features with toxic shock syndrome and is associated with high mortality 1,3 . However, little is known of the mechanism of S. suis survival in human blood, and only a few virulence factors, such as Fhb 24 and SLY 33 , are known to enable pathogens to evade the host immune response by inhibiting complement activation.
MRP is often reported as an epidemic infection marker in Europe and Asia [19][20][21] . Although this protein was reported not to be a critical virulence factor in an intranasal infection model in Anthony piglets 34 , the real contribution of MRP to bacterial virulence needs to be identified in a rational infection model because different infection models might lead to different conclusions 35 . For example, the Canadian virulent strain 89-1591 (MRP -) was reported as avirulent by Rasmussen et al. 36 but was shown to be virulent in experimental models applied by other researchers 37 . Several North American virulent strains were reported to be MRP − genotype strains 38 . However, most North America virulent strains (MRP -) belong to ST25 strains, which were different from ST7 strains based on multilocus sequence typing 39 . The highly virulent Chinese strain 05ZYH33 used in this study belongs to ST7 strains which exhibit strong invasiveness and high pathogenicity, has resulted in the description of a new disease form of streptococcal toxic shock syndrome (STSS) and a putative pathogenicity island (termed 89 K) 31 . We believe that the real contribution of MRP to different strains need to consider the background of strains.
PMNs are the first line of defense against bacterial infections in human blood; in our recent study, MRP enhanced the survival of S. suis 2 in PMN killing assays by interacting with hFg 15 . However, the specific mechanism of this enhancement needed to be clarified. In this study, MRP was identified as a novel FBP of S. suis 2 that showed high binding affinity to hFg, suggesting that the MRP-hFg interaction may be critical for MRP function. The present study revealed that binding occurs via the interaction of the N-terminus of MRP (a.a. 283-721) and The MFI of DHR123 was measured. In the PMN killing assay, the PMNs were stimulated with the WT or Δ MRP strains in the absence or presence of GPRP or GHRP. PMNs without S. suis stimulation served as the control. DHR123 (1 μ M) was used to detect oxidative activity. The data are expressed as the means ± SD of three replicates. * P < 0.05; n.s., no significance. the D fragment of Fg, similar to that observed in the adhesins of other gram-positive bacteria 7,11 . Nonetheless, there are some notable differences from other FBPs: first, the binding region of MRP has no sequence homology to other known FBPs, which share a common domain architecture that enables fibrinogen binding through a "dock, lock, and latch" (DLL) mechanism [40][41][42][43] . Second, in contrast to other FBPs that bind the Aα , Bβ or γ chain singly, MRP displayed specific binding to both the Aα and Bβ chains in a manner similar to Lysin, an FBP of Streptococcus mitis 44 , but different from that of the M1 protein, which binds to coiled coils in the β and γ chains of hFg 45 . Studies have been initiated to determine the crystal structure of the MRP-hFg complex, which can reveal the details of this interaction.
An intriguing observation was that the soluble MRP-hFg interaction enhanced the survival of S. suis 2 in human blood and induced the aggregation and hyperactivation of PMNs via the α X β 2 integrin receptor, which was independent of complement inhibition (data not shown). However, M proteins are important virulence factors for GAS that inhibit complement deposition on the bacterial surface via the binding of hFg to the surface M proteins 8,46 . Although the soluble (released) M1 induced PMN aggregation and HBP release and was blocked by the β 2 integrin antagonist GPRP 25 , it was not determined which member of the β 2 integrin subfamily was involved in the crosslinking of the M1-hFg complex and the immune evasion of GAS. Despite hFg being a physiologically relevant ligand for α M β 2 , the integrin engagement of hFg is critical for leukocyte function and innate immunity in vivo 47 . α X β 2 , but not α M β 2 , was involved in the excessive activation of PMNs induced by the MRP-hFg complex in the present study. Although other FBPs, including M1 25 and FOG 10 , induce PMN aggregation, the specific receptor on PMNs that mediates crosslinking to the FBP/hFg complexes and the relationship of this process to immune evasion by pathogens remained unclear.
In summary, MRP is a known infection marker of S. suis 2 in most Chinese and European clinical isolates, but the role of MRP in the evasion of PMN-mediated innate immunity by S. suis remained unclear and controversial.
Here, we showed that the N-terminus of MRP (a.a. 283-721) binds to both the Aα and Bβ chains of the D fragment in hFg. Moreover, the MRP-hFg interaction aggregates PMNs via the α X β 2 integrin receptor and promotes pathogen survival in host blood. These data shed new light on the function of MRP in the S. suis-host interaction. More objective consideration of strains and infection models is needed to reevaluate the contribution of MRP to the virulence of ST-7 strains in China and ST-1 strains in Europe.

Methods
Human blood, serum and plasma. The healthy donors who provided the serum and plasma in this study provided written informed consent in accordance with the Declaration of Helsinki. Approval was obtained from the Institutional Medical Ethics Committee of the Academy of Military Medical Sciences. The human blood, serum and plasma used in this study were obtained in accordance with the approved guidelines.
Bacterial strains and growth conditions. The highly virulent S. suis 2 strain 05ZYH33, originally isolated from a deceased STSS patient following an outbreak in Sichuan, China in 2005, was used in this study. S. suis 2 strains were cultured as previously described 24 . The bacterial strains and plasmids used in this study are listed in Supplementary Table 1

ELISA-type binding assay.
To detect the binding of MRP to immobilized hFg, 96-well plates were coated overnight at 4 °C with 100 μ l of 2.5 μ g/ml Fg/Fg-D/Fg-E (Merck). After blocking with 3% BSA (bovine serum albumin) in PBST (phosphate-buffered saline with 0.05% Tween-20), recombinant MRP proteins were added and incubated for 1 h at room temperature. Bound MRP was detected by incubation with an anti-his monoclonal antibody (1:5000, Sigma) and HRP (horseradish peroxidase)-conjugated anti-mouse antibodies (1:10,000; Santa Cruz). The binding was quantified by measuring the absorbance at 450 nm. Background values obtained for the control (incubated with PBST containing 1% BSA, without truncated MRP) were subtracted. To detect the binding of hFg to immobilized MRP, 96-well plates were coated overnight at 4 °C with 100 μ l of 100 nM truncated MRP. After blocking with 3% BSA in PBST, biotinylated hFg was added, and the plates were incubated for 1 h at room temperature. Bound hFg was detected by incubation with HRP-conjugated streptavidin (1:8000; GE Health). Background values obtained for the control (incubated with PBST containing 1% BSA, without biotinylated hFg) were subtracted. The assays were performed in triplicate, and the mean ± SD is shown.
Scientific RepoRts | 6:26966 | DOI: 10.1038/srep26966 BLI analysis of the MRP-hFg interactions. All BLI experiments were performed in PBST on a BLItz system (ForteBio, Inc.) as previously described 48 . Streptavidin-coated biosensors with immobilized biotinylated MRP-N proteins were exposed to different concentrations of hFg, Fg-D, or Fg-E. The association and dissociation rate constants (k a and k d , respectively) for the interactions were derived from curve fitting and were used to calculate the dissociation constant, K D (K D = k d /k a ).
Binding of S. suis to immobilized fibrinogen. The binding of S. suis to immobilized fibrinogen was detected as previously described 43 . The 96-well plates were coated overnight at 4 °C with 100 μ l of purified fibrinogen (1, 5, 10 μ g/ml) or BSA (1, 5, 10 μ g/ml) and were then washed four times with PBS. Overnight cultures of the WT, CΔ MRP-N and Δ MRP strains were harvested by centrifugation and then adjusted to almost equal the original inocula (~10 5 CFU [colony-forming units]/ml) in PBS. The wells, which were pretreated with hFg/BSA, were then incubated with 100 μ l of S. suis suspension for 30 min at 37 °C. The wells were washed four times with PBS to remove unbound bacteria and were then treated with 100 μ l of trypsin (2.5 mg/ml) for 10 min at 37 °C to release the attached bacteria. The number of bound bacteria was determined by plating serial dilutions of the recovered bacteria onto THB agar plates.
Generation of the ∆MRP mutant and the complementary CΔMRP-N strains. The mrp mutant was obtained from the 05ZYH33 WT strain (accession number YP_001198119) by in-frame deletion of the SSU05_0753 gene, as previously described 15,49 . In the present study, for complementation assays, we selected MRP-N, but not the entire MRP sequence for cloning because MRP exists as a soluble protein in the supernatant, and MRP-N (a.a. 48-721) was the hFg-binding region. Furthermore, it was difficult to construct a complementary strain expressing the entire MRP protein. The upstream promoter and a DNA fragment encoding MRP-N without the LPxTG motif were amplified using the primers CΔ MRP-N-F/CΔ MRP-N-R (Supplementary Table 1). The amplicon was subsequently cloned into the E. coli-S. suis shuttle vector pAT18, the recombinant plasmid pAT18::mrp-N was transformed into the Δ MRP strain, and the CΔ MRP-N strain was screened on THB agar with selective pressure from Em. Western blotting and confocal laser scanning microscopy analysis were used to confirm the soluble expression of MRP-N in CΔ MRP-N bacteria.
Bactericidal and PMN killing assays. The bactericidal assay was used to compare the growth of the WT, CΔ MRP-N and Δ MRP strains in human blood as described previously 24 . Briefly, diluted strain cultures (50 μ l at 2 × 10 4 CFU/ml) were combined with fresh human blood (450 μ l), and the mixtures were rotated at 37 °C; the samples were treated with 1% saponin, diluted, and plated onto blood agar. The percentage of live bacteria was subsequently calculated as (CFU on plate/CFU in original inoculum) × 100%. In the integrin-blocking assay, GPRP (50 μ g/ml), GHRP (50 μ g/ml), monoclonal antibodies against α M (clone CBRM1/5, eBioscience, 10 μ g/ml), α x integrin subunits (clone BU15, Acris Antibodies, 10 μ g/ml) and the isotype control mouse IgG1 antibody (eBioscience, 10 μ g/ml) were preincubated with the blood for 15 min at room temperature. In the PMN killing assay, PMNs were infected with S. suis 2 at a multiplicity of infection (MOI) of 1:15 (PMN:bacterium) in 50% non-immune human serum and were centrifuged at 380 × g for 5 min at 4 °C. The plates were then incubated at 37 °C in 5% CO 2 . Samples were taken at 60 min and analyzed immediately as described for the human blood. The colonies were counted, and the percentage of surviving S. suis 2 bacteria was calculated as follows: (CFU PMN+ / CFU PMN− ) × 100%. GPRP and GHRP were dissolved in H 2 O.
Confocal laser scanning microscopy analysis. The LIVE/DEAD viability kit (Invitrogen) was used to determine the viability of PMNs and S. suis. In the PMN killing assay, a mixture of PMNs and bacteria was stained with PI and SYTO 9 and then spotted onto Petri dishes (World Precision Instruments Inc.) using cover slips placed over the liquid to stabilize the cells. The dead (PI, red staining) and viable (SYTO 9, green staining) PMNs/ bacteria were then visualized by XYZ scanning using an FV1000 confocal laser scanning microscope (Olympus).
Visualization of MRP proteins on the bacterial surface was performed as previously described 50 . Briefly, an aliquot of streptococci (6 × 10 6 CFU/mL) was added to rabbit anti-MRP-N IgG (10 μ g/ml), incubated for 45 min on ice and washed. Three hundred microliters of the resuspension was then incubated with rhodamine (TRITC)-goat anti-rabbit IgG (Santa Cruz) (0.3 mg/ml) for 45 min on ice. The bacteria were then washed with PBS and spotted onto CITOGLAS slides (SHITAI). The slides were then analyzed with an OLYMPUS FV1000 confocal scanning microscope.