Host immunity against bacteria typically involves antibodies that recognize the microbial surface and promote phagocytic killing. Methicillin-resistant Staphylococcus aureus (MRSA) is a frequent cause of lethal bloodstream infection; however, vaccines and antibody therapeutics targeting staphylococcal surface molecules have thus far failed to achieve clinical efficacy. S. aureus secretes coagulase (Coa), which activates host prothrombin and generates fibrin fibrils that protect the pathogen against phagocytosis by immune cells. Because of negative selection, the coding sequence for the prothrombin-binding D1-D2 domain is highly variable and does not elicit cross-protective immune responses. The R domain, tandem repeats of a 27-residue peptide that bind fibrinogen, is conserved at the C terminus of all Coa molecules, but its functional significance is not known. We show here that the R domain enables bloodstream infections by directing fibrinogen to the staphylococcal surface, generating a protective fibrin shield that inhibits phagocytosis. The fibrin shield can be marked with R-specific antibodies, which trigger phagocytic killing of staphylococci and protect mice against lethal bloodstream infections caused by a broad spectrum of MRSA isolates. These findings emphasize the critical role of coagulase in staphylococcal escape from opsonophagocytic killing and as a protective antigen for S. aureus vaccines.
Staphylococcus aureus colonizes the human nares and skin, and also causes soft tissue and bloodstream infections (David and Daum, 2010). Drug-resistant strains, designated MRSA (methicillin-resistant S. aureus), emerged with antibiotic use for the prevention or therapy of staphylococcal infections. MRSA infections are associated with increased failure of antibiotic therapy and increased mortality (David and Daum, 2010). To address this public health crisis, vaccines and antibody therapeutics have been developed, each targeting molecules on the staphylococcal surface, including capsule, polyglycerol phosphate lipoteichoic acid, iron-regulated surface determinant protein B (IsdB), and clumping factor A (ClfA; Spellberg and Daum, 2012). However, the corresponding clinical trials failed to reach their designated endpoints (Shinefield et al., 2002; Fowler et al., 2013).
A distinguishing feature of clinical S. aureus isolates is their ability to clot human plasma. This trait is based on the secretion of coagulase (Coa; Fig. 1 A; Tager, 1956), which associates with human prothrombin to form enzymatically active staphylothrombin, cleaving the A and B peptides of fibrinogen and generating fibrin fibrils (Friedrich et al., 2003). Staphylothrombin does not cut other endogenous substrates of thrombin, causing exuberant polymerization of fibrin while avoiding activation of other clotting and inflammatory factors (Panizzi et al., 2004; McAdow et al., 2012b). The resulting fibrin meshwork protects bacteria from phagocytes and is essential for the formation of S. aureus abscess lesions (Smith et al., 1947; Cheng et al., 2010). Activation of prothrombin is mediated by the N-terminal D1-D2 domain of Coa and blocked by specific antibodies that provide protection from S. aureus bloodstream infection in animal models (Rammelkamp et al., 1950; Cheng et al., 2010). Because of negative selection, coa is one of the most variable genes in the core genome of S. aureus. Up to 50% sequence variation occurs in the coding sequence for the D1-D2 domain, and the corresponding products can be categorized into serotypes without cross-protecting epitopes for the neutralization of staphylothrombin (Watanabe et al., 2009; McAdow et al., 2012a). S. aureus secretes a second staphylothrombin, designated von Willebrand factor binding protein (vWbp) with the conserved D1-D2 domain structure mediating association with prothrombin (Bjerketorp et al., 2004). This complex displays different catalytic activity than Coa-staphylothrombin, generating fibrin fibrils at a reduced rate and contributing to abscess formation without affecting staphylococcal escape from phagocytosis (Kroh et al., 2009; Guggenberger et al., 2012). The structural gene for vWbp, vwb, displays limited sequence variation, and is presumably not subject to negative selection (McAdow et al., 2012a).
RESULTS AND DISCUSSION
R domain of coagulase supports
S. aureus bloodstream infection
The C-terminal domain of Coa is conserved and comprised of tandem repeats of a 27-residue peptide which binds fibrinogen (Fig. 1 A; Watanabe et al., 2009; Panizzi et al., 2011). The number of tandem repeats varies between Coa molecules from different isolates of S. aureus (Watanabe et al., 2009). To characterize the contribution of the R domain to the pathogenesis of staphylococcal disease, we generated isogenic S. aureus variants with a truncated coa, lacking the R domain (coaΔR), in either wild-type or Δvwb backgrounds. When probed by immunoblotting with Coa- and vWbp-specific antibodies and compared with Coa from wild-type staphylococci, S. aureus coaΔR and coaΔR/Δvwb strains secreted a truncated protein into the extracellular medium (Fig. 1 B). mAb 5D5, which recognizes the D1 domain of Coa, bound to both Coa and CoaΔR, whereas mAb 3B3, specific for the R domain, only bound Coa, but not CoaΔR (Fig. 1, A and B). When inoculated into calcium-chelated mouse blood and incubated for 24 h, wild-type S. aureus produced a firm clot, whereas mock-infected blood did not (Fig. 1 C). Staphylococci rely on secretion of both coagulases for clotting, as only Δcoa/Δvwb but not Δcoa or Δvwb variant strains displayed a defect in this assay (Fig. 1 C). Compared with their respective parent strains, the coaΔR and coaΔR/Δvwb mutants were not defective for clotting (Fig. 1 C).
When inoculated intravenously into mice, wild-type S. aureus Newman causes a lethal bloodstream infection within 2–3 d, whereas Δcoa or Δvwb mutations each cause a delay in time-to-death that is additive for the Δcoa/Δvwb mutant (median survival time 60 h [wild-type], 108 h [Δcoa or Δvwb], and 180 h [Δcoa/Δvwb]; Fig. 1, D and E). Surprisingly, the coaΔR mutation also caused a delay in time-to-death (median survival time 72 [coaΔR] and 126 h [coaΔR/Δvwb]), which could be quantified in strains with (wild-type vs. coaΔR, P = 0.0308; Δcoa vs. coaΔR, P = 0.0229) or without vwb expression (Δvwb vs. coaΔR/Δvwb, P = 0.043; Δcoa/Δvwb vs. coaΔR/Δvwb, P = 0.0084). Thus, the R domain, although dispensable for staphylothrombin-mediated clotting, contributes to the pathogenesis of S. aureus infection in mice.
R domain enables assembly of the staphylococcal fibrin shield
Full-length Strep-tagged Coa (CoaST), Coa truncated for the R domain (CoaΔR\ST), and R domain alone (RST) were purified and used for affinity chromatography experiments with citrate plasma (Fig. 2 A). CoaST and RST retained molar excess of fibrinogen, whereas CoaΔR/ST retained only equimolar amounts of fibrinogen (Fig. 2 A). This can be explained by the equimolar association between fibrinogen and the exosite of staphylothrombin within CoaST or CoaΔR/ST, whereas the R domain of CoaST and RST associates with 3–4 mol of fibrinogen (Fig. 2 A). As expected, CoaST and CoaΔR/ST bound prothrombin via their D1-D2 domain, whereas RST did not (Fig. 2 A). Staphylococci display surface proteins, for example ClfA, that promote association of bacteria with fibrinogen (McDevitt et al., 1994; McAdow et al., 2012a). Mixed with dilute plasma, mid-log staphylococcal cultures formed fibrin clots that, when centrifuged, sedimented with the bacteria and could be solubilized with urea (Fig. 2 B). When analyzed by Coomassie-stained SDS-PAGE, fibrin was found to be associated with the bacterial sediment, whereas albumin remained in the supernatant of agglutinated staphylococci (Fig. 2 B). Immunoblotting revealed that full-length Coa sedimented with the bacterial clot, whereas CoaΔR did not (Fig. 2 B). Association of Coa with staphylococci occurred in the presence of the fibrin clot and was not observed for staphylococcal cultures centrifuged without human plasma (Fig. 2 B). To visualize the contribution of the R domain toward staphylococcal fibrin formation, mCherry-expressing bacteria were added to plasma samples with Alexa Fluor 88–conjugated fibrinogen and clot formation was viewed by fluorescence microscopy. Unlike wild-type staphylococci, which generated large fibrin deposits in the vicinity of bacteria, the coaΔR mutant produced long fibrin strands that were only loosely associated with the pathogen (Fig. 2 C). Thus, by augmenting the recruitment of soluble fibrinogen, the C-terminal repeats favor Coa-induced fibrin clots and limit diffusion of Coa away from staphylococci, thereby localizing the staphylothrombin-generated fibrin shield in the immediate vicinity of the bacteria.
R domain antibody protects mice against bloodstream infection
Mouse mAbs were raised by immunizing mice with full-length Coa of S. aureus Newman. 13 antibodies reactive to Coa, but not to vWbp or IsdA controls, were characterized for their affinity and specificity to D1, D2, D1-D2, D1 lacking the first 18 residues (D1Δ1-18), L (linker), and R domains (Fig. 1 A). Two mAbs targeting the variable or conserved domains of Coa, 5D5 and 3B3, were used for further study. mAb 5D5, which bound to the D1 domain within the first 18 residues of D1 that insert into the prothrombin-active site to generate active staphylothrombin (Table S1), prevented CoaST binding to prothrombin but not to fibrinogen (Fig. 3, A and B). mAb 3B3, on the other hand, bound to the R domain (Table S1) and blocked CoaST association with fibrinogen but not with prothrombin (Fig. 3, A and B). Further, mAb 5D5, but not mAb 3B3, inhibited S. aureus Newman–mediated clotting of mouse blood in vitro; however, neither 5D5 nor 3B3 inhibited S. aureus agglutination of EDTA-rabbit plasma (Fig. 3, C and D). Purified mAbs 5D5 or 3B3 were injected at a concentration of 5 mg antibody/kg body weight into the peritoneal cavity of BALB/c mice and compared with IgG1 isotype control mAb (Fig. 4). Both 5D5 and 3B3 provided protection against lethal bloodstream infection with S. aureus Newman (IgG1 vs. 5D5, P < 0.0001; IgG1 vs. 3B3, P < 0.0001; Fig. 4 A). Similar results were obtained when the S. aureus Δvwb variant was used as a challenge strain (IgG1 vs. 5D5, P = 0.0011; IgG1 vs. 3B3, P = 0.0004; Fig. 4 B). In ELISA assays, mAb 3B3 was observed to bind coagulase from different serotypes, including type II (CoaN315), type III (CoaUSA300), type IV (CoaMRSA252 and Coa85/2082), and type VII (CoaWIS; Table S2). In contrast, mAb 5D5 recognized only CoaUSA300 and to a lesser degree CoaWIS (Table S2). When analyzed for the prevention of lethal bloodstream infections, both 3B3 and 5D5 provided protection against MRSA strain USA300, with a type III coagulase similar to S. aureus Newman (IgG1 vs. 5D5, P = 0.0007; IgG1 vs. 3B3, P < 0.0001; Fig. 4 C). However, only mAb 3B3 protected mice against lethal bloodstream challenge with S. aureus N315 (IgG1 vs. 5D5, P = 0.1186; IgG1 vs. 3B3, P < 0.0001), MRSA252 (IgG1 vs. 5D5, P = 0.5993; IgG1 vs. 3B3, P < 0.0001), and MRSA isolate WIS (IgG1 vs. 5D5, P = 0.4243; IgG1 vs. 3B3, P < 0.0001; Fig. 4, D–F).
S. aureus agglutination in human blood
Blood from human volunteers was anticoagulated with desirudin to inhibit endogenous thrombin without affecting staphylothrombin (McAdow et al., 2011). Blood cells were removed by centrifugation, and 0.5 ml human plasma was inoculated with S. aureus Newman (5 × 106 CFU). Staphylococcal CFU were enumerated at timed intervals (0 and 60 min incubation at 37°C). Within 60 min, CFU for wild-type S. aureus dropped from 5 × 106 (100%) to 0.15 × 106 (3%), whereas CFU for the isogenic Δcoa/Δvwb variant were not reduced (Fig. 5 A). Treatment of plasma samples with streptokinase (SK), the plasminogen activator of fibrinolysis, did not affect bacterial CFU in the 0 min samples, and liberated wild-type S. aureus agglutinated over 60 min (Fig. 5 A). USA300 LAC agglutinated in human plasma and replicated quickly to generate a large bacterial load. USA300 LAC agglutination did not occur in defibrinated human serum (Fig. 5 A).
S. aureus phagocytosis and opsonophagocytic killing (OPK) were measured in blood samples from 20 healthy human volunteers infected with 5 × 106 CFU USA300 LAC for 60 min. Bacterial CFU were quantified with or without SK treatment (Table S3). Control blood samples were pretreated with cytochalasin D (CD), thereby preventing S. aureus phagocytosis (Mimura and Asano, 1976). At a challenge dose of 10 bacteria per leukocyte, the assay quantifies OPK of 5 × 106 CFU USA300 LAC as the percent CFU reduction from 0 to 60 min in SK-treated blood. Phagocytes in blood samples of volunteer A killed 2.552 × 106 CFU (51.04%) within 60 min (Fig. 5 B). A fraction (64.62%) of the total staphylococcal load could be enumerated in blood without SK treatment (Table S3). When pretreated with CD, 97.92% of staphylococcal CFU were agglutinated in blood from volunteer A. Agglutination was calculated as the percent S. aureus CFU requiring SK treatment for enumeration after 60 min incubation. For volunteer A, 35.38% of the staphylococcal load had agglutinated within 60 min, whereas 64.62% had been phagocytosed (Table S3). Phagocytes in blood samples from volunteer G were unable to kill S. aureus: 99.68% of the inoculum was recovered in SK blood (Fig. 5 B). Here, 21.93% of the bacterial load had been phagocytosed, whereas 78.07% were agglutinated (Table S3). USA300 LAC expanded in blood samples from volunteer I to 204.42% of the initial inoculum; 85.75% of the load were agglutinated (Fig. 5 B). On the basis of these phenotypes, we categorized human blood samples as staphylococcal killer, controller, or prey (Table S3). This classification applies only to S. aureus, as both killer and prey blood samples were active in phagocytosis and OPK of Staphylococcus epidermidis, a commensal that does not express coagulases (Fig. 5 C). Antibody titers against the D1-D2 or the C-terminal R domain were not correlated with OPK of USA300 LAC in human blood (Table S3).
R domain antibody promotes phagocytosis of fibrin-coated staphylococci
When added to blood samples of volunteer B (prey), mAb 3B3 reduced the bacterial load to 63%, whereas USA300 LAC expanded to 128% in blood without antibody (3B3 vs. mock, P < 0.05; Fig. 5 D). Pretreatment of blood with CD abolished phagocytosis and OPK of USA300 LAC in the presence of mAb 3B3 (Fig. 5 D). S. aureus Newman–expressing GFP was inoculated into mouse blood, and neutrophils were isolated by GR1-staining and flow cytometry (Fig. 5 E). Although phagocytosis of staphylococci occurred in the absence of antibody, association of staphylococci with neutrophils was increased in the presence of mAb 3B3 (Fig. 5 E). Further, GFP fluorescence did not increase after 30 min, indicating that bacterial replication had been arrested (Thammavongsa et al., 2013). Antibody-mediated uptake of staphylococci was not observed in neutrophils from S. aureus coaΔR samples (Fig. 5 E). Neutrophil uptake of wild-type S. aureus was accompanied by uptake of fibrin, detected by adding Alexa Fluor 488–conjugated human fibrinogen to blood samples and measuring neutrophil fluorescence (Fig. 5 F). Mouse blood infected with S. aureus was Giemsa stained, which revealed large clumps of fibrin-agglutinated staphylococci outside of neutrophils (Fig. 5 G). When treated with mAb 3B3, staphylococci appeared to be internalized by mouse neutrophils (Fig. 5 G). Mouse blood was infected with S. aureus and analyzed for CFU after 30 and 60 min incubation. Compared with mock control, mAb 3B3 promoted phagocytic killing of staphylococci. As expected, OPK was blocked by pretreatment with CD (Fig. 5 H). OPK of S. aureus was quantified in vivo in mice with intravenous challenge of S. aureus followed by CFU enumeration in cardiac blood 30 min post infection. mAb 3B3 reduced the bacterial load in mice infected with wild-type S. aureus but not in mice infected with the coaΔR variant (Fig. 5 I).
We report that S. aureus evolved coagulase-mediated assembly of a fibrin shield to protect the pathogen against uptake by phagocytes. The R domain drives the formation of the bacterial fibrin shield that protects bacteria but also exposes Coa for antibody deposition. To avoid neutralizing antibody responses against Coa, coding sequence for the D1-D2 domain is subject to negative selection, generating S. aureus variants that are not be neutralized by antibodies against another serotype (Watanabe et al., 2009; McAdow et al., 2012a). We also show that antibody against the R domain can target staphylococci for OPK. Successful vaccines generally rely on antibodies against bacterial surface structures to implement pathogen destruction (Robbins et al., 1996). S. aureus escapes antibody-mediated destruction by a number of different immune evasion mechanisms, blocking neutrophil chemotaxis, phagocytosis, complement activation, and antibody deposition (Spaan et al., 2013). To address this, vaccine developers rely on a standardized assay, measuring OPK in cultured HL60 phagocytes supplemented with complement and antibody but not with hemostasis factors (Nanra et al., 2013). This assay does not, however, assess the immune evasive attributes of coagulase and may overestimate the role of antibodies in promoting OPK.
MATERIALS AND METHODS
Bacterial growth, strains, and plasmids
S. aureus and Escherichia coli were grown in tryptic soy and Luria broth or agar, with ampicillin (100 µg ml–1) or chloramphenicol (10 µg ml–1) when necessary. Earlier work reported S. aureus Newman and its variants Δcoa, Δvwb, and Δcoa/Δvwb with or without plasmid expressing GFP or mCherry (Cheng et al., 2010). pKOR1 was used to introduce the coaΔR allele (deletions of codons 470–605) into wild-type or Δvwb Newman (Bae and Schneewind, 2006). Earlier work generated E. coli plasmids for purification of full-length mature Coa (S. aureus Newman, USA300, N315, MRSA252, 85/2082, or WIS; McAdow et al., 2012a; Thomer et al., 2013) or Coa Newman domains (D1, D1-D2, D1Δ1-18, D2, and L; McAdow et al., 2012a). Plasmid pET15b-rST harbors coding sequence for the R domain (codons 470–605) and a C-terminal Strep tag.
Identification of coagulases in cultures and clots
To examine the secretion of coagulases, cultures of staphylococci were grown to an optical density A600 0.4 (∼108 CFU ml–1). Proteins in the supernatant, i.e., 1 ml of centrifuged culture, were precipitated with 75 µl of trichloroacetic acid 100% (wt/vol), washed with acetone, dried, and solubilized in 50 µl sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, and 0.01% bromophenol blue). To examine the fate of coagulase in fibrin clots, 950 µl of bacterial culture (∼108 CFU ml–1) or broth were mixed with 50 µl of PBS or human citrate plasma for 10 min at 37°C and centrifuged at 13,000 g for 10 min to separate soluble and clotted materials. 4 M urea was used to solubilize fibrin clots before separation of extracts by SDS-PAGE. Proteins were visualized with Coomassie staining or transferred to polyvinylidene difluoride (PVDF) membranes for immunoblotting using rabbit affinity-purified antibodies against Coa (α-Coa) or vWbp (α-vWbp; Thomer et al., 2013) and mouse affinity-purified monoclonal antibodies 3B3 or 5D5.
Pull down experiments
CoaST, CoaΔR/ST, and RST were purified over Strep-Tactin-Sepharose (IBA) following methods described earlier for Coa subdomains and Coa strain variants (McAdow et al., 2012a; Thomer et al., 2013). All purified proteins were stored in PBS. For pull-down experiments, citrate plasma from healthy human volunteers (500 µl) diluted 1:1 in PBS was applied by gravity flow over Strep-Tactin-Sepharose beads precharged or not with 100 nmol of purified CoaST, CoaΔRST, or RST. Bound proteins were recovered by boiling the resin in sample buffer and analyzed by SDS-PAGE separation followed by Coomassie staining or immunoblot.
10 µl of bacterial suspension (∼108 CFU ml–1) was added to 90 µl of freshly collected mouse blood anticoagulated with sodium citrate (10 mM final concentration) in a sterile plastic test tube (BD). Samples were incubated at room temperature, and blood coagulation was verified by tipping the tubes to 45° angles at timed intervals. Where indicated, antibodies were added at a final concentration of 3 µM. Statistical analysis was performed by two-tailed Student’s t test using Prism (GraphPad Software).
For visualization of bacteria in clots, 5 µl of staphylococci expressing mCherry (∼108 CFU ml–1) were mixed for 5 min with 5 µl of human citrate plasma supplemented with 5% Alexa Fluor 488–conjugated human fibrinogen (Life Technologies). Images of samples placed on glass slides were captured on a SP5 tandem scanner spectral 2-photon confocal microscope (Leica) using a 100× objective. For assessment of agglutination, 1 ml staphylococci (∼108 CFU ml–1) was incubated with 1:500 SYTO9 (Invitrogen) for 15 min, washed twice, and suspended in 1 ml PBS. Bacteria were incubated 1:1 for 15 min with human citrate plasma on glass microscope slides. Where indicated, antibodies were added at a final concentration of 3 µM. Images were captured on a live cell total internal reflection fluorescence microscope (IX81; Olympus) using a 20× objective. The threshold function in ImageJ software (National Institutes of Health) was used to convert the image into a dichromatic format in which staphylococci are black and the background is white. Statistical significance was determined by two-way analysis of variance using Prism.
Production of monoclonal antibodies against coagulase
Three 8-wk-old BALB/c female mice (The Jackson Laboratory) were immunized by intraperitoneal injection with 100 µg of purified recombinant CoaNM emulsified 1:1 in Complete Freund’s Adjuvant (DIFCO) for the first immunization. On days 21 and 42, animals were boosted with 100 µg CoaNM emulsified 1:1 in Incomplete Freund’s Adjuvant (DIFCO). On days 31 and 52, animals were bled and screened by ELISA on MaxiSorp (Nunc) 96-well flat bottom plates coated with Coa. 79 d after the initial immunization, mice that showed strong immunoreactivity to antigen were boosted with 25 µg Coa in PBS. 3 d later, splenocytes were harvested and fused with the mouse myeloma cell line SP2/mIL-6, an IL-6–secreting derivative of SP2/0 myeloma cell line. Hybridomas were screened by ELISA and antigen-specific clones subcloned by limiting dilution to produce monoclonal antibody-secreting hybridomas arising from single cells. Hybridoma cell lines were grown until a density of 106 cells ml−1 in DMEM-10 medium with 10% FBS and left spending for 6 wk. Antibodies were purified from filtered culture supernatants by affinity chromatography as previously described (McAdow et al., 2012a; Thomer et al., 2013).
To determine the binding affinity and specificity of mAbs, Nunc MaxiSorp 96-well plates were coated with the various Coa variant serotypes and subdomains prepared at a concentration of 20 nM in 0.1 M sodium bicarbonate and affinities were measured as described earlier (McAdow et al., 2012a). ELISA plates coated with vWbp and IsdA served as negative controls. The ability of mAbs to interfere with the binding of prothrombin or fibrinogen was measured as described previously (McAdow et al., 2012a), and statistical analyses were performed using one-way ANOVA with Bonferroni post-test. Half-maximal IgG titers in serum from human volunteers for binding to purified Hla, D1-D2ST, or RN12D were determined by ELISA as described previously (McAdow et al., 2012a). RN12D is a translational hybrid between SpAKKAA, a variant of SpA that does not bind immunoglobulin, and two 27-residue repeats of the R domain from CoaNewman, with Asn12Asp at position 12 of each repeat, followed by a C-terminal Strep tag; purified RN12D for is defective fibrinogen binding.
Animal infection and immunization studies
6-wk-old female BALB/c mice (cohorts of 10; Charles River) anesthetized with 100 mg ml−1 ketamine and 20 mg ml−1 xylazine per kilogram of body weight were inoculated into the peri-orbital venous plexus with 100 µl of bacterial suspension in PBS at a concentration of 2 × 108 CFU ml−1 (USA300), 8 × 108 CFU ml−1 (Newman, N315, WIS), or 2 × 109 CFU ml−1 (MRSA252). mAbs were injected at a concentration of 5 mg kg−1 into the peritoneal cavity 10 h before challenges. Statistical analyses were performed by two-tailed log-rank test using Prism. To assess the fate of staphylococci in blood (in vivo blood survival assay), animals were euthanized by CO2 inhalation 30 min after infection and cardiac puncture was performed. Blood samples were treated with 0.5% saponin to lyse eukaryotic cells, serially diluted in PBS, and plated on agar for enumeration of CFU. Statistical analysis was performed using two-tailed Student’s t test. Animal experiments were performed in accordance with the institutional guidelines following experimental protocol review and approval by the Institutional Biosafety Committee and the Institutional Animal Care and Use Committee at the University of Chicago.
Bacterial survival in blood, opsonophagocytosis assay, and flow cytometry analysis
To measure bacterial replication and survival ex vivo, 0.5 ml of freshly drawn mouse or human blood anticoagulated with 0.005 mg desirudin per ml was incubated with 50 µl of a bacterial suspension containing 5 × 105 CFU (mouse) or 5 × 106 CFU (human). Where indicated, human blood was processed to generate desirudin-plasma or serum. Where indicated, 5% Alexa Fluor 488–conjugated human fibrinogen (Life Technologies), CD (0.04 mM), or purified mouse monoclonal antibodies (∼10 µg ml−1 final concentration) were added to the samples. After incubation at 37°C for 0, 30, or 60 min, 0.5 ml of PBS with 0.5% saponin or 0.5 ml agglutination lysis buffer (0.5% saponin, 200 U SK K, 100 µg trypsin, 2 µg DNase, 10 µg RNase per ml PBS) were added to each sample for 10 min at 37°C before plating on agar for enumeration of CFU. Treatment with agglutination lysis buffer is annotated as +SK in the figures. Statistical analysis was performed by two-tailed Student’s t test. For flow cytometry analysis, samples were incubated first with lysostaphin (10 µg ml−1) for 5 min to lyse extracellular bacteria and next with erythrocyte lysis buffer (QIAGEN) for 30 min on ice. Blood leukocytes were recovered after centrifugation at 400 g, washed three times, and suspended in PBS containing 1% FBS. Cells were stained with allophycocyanin-conjugated α-GR1 and analyzed using a FACSCanto (BD). The data were analyzed with the two-tailed Student’s t test. Human volunteers were enrolled under a protocol that was reviewed and approved by the University of Chicago's Institutional Review Board.
Online supplemental material
Table S1 shows the binding sites and affinities of mAbs 5D5 and 3B3 for coagulase from S. aureus Newman. Table S2 shows the affinity of mAbs 5D5 and 3B3 for coagulases from different clinical isolates of S. aureus. Table S3 analyzes agglutination, phagocytosis, and OPK of S. aureus in blood samples from 20 human volunteers.
This work was supported by grants from the National Institute of Allergy and Infectious Diseases (AI52747 to O.S. and AI110937 to D. Missiakas), the National Institutes of Health (HD009007 to M.E. McAdow) and the American Heart Association (14PRE19910021 to L. Thomer and PST4590023 to V. Thammavongsa).
The authors declare no competing financial interests.
Author contributions: L. Thomer, C. Emolo, V. Schneewind, H.K. Kim, M.E. McAdow, M. Kieffer, and W. Yu performed experiments; L. Thomer, V. Thammavongsa, H.K. Kim, M.E. McAdow, W. Yu, O. Schneewind, and D. Missiakas interpreted data; L. Thomer, V. Thammavongsa, O. Schneewind, and D. Missiakas designed experiments; and L. Thomer, O. Schneewind, and D. Missiakas wrote the paper.