Inactivates Neutrophil Proteins S100A8/A9 Bacterial Surface

Bacterial Surface Protein L Binds and Inactivates Neutrophil Proteins S100A8/A9 Bo Åkerström and Lars Björck This information is current as J Immunol 2009; 183:4583-4592; Prepublished online 14 of September 28, 2021. September 2009; doi: 10.4049/jimmunol.0901487 http://www.jimmunol.org/content/183/7/4583 Downloaded from

References This article cites 57 articles, 23 of which you can access for free at: http://www.jimmunol.org/content/183/7/4583.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 28, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Bacterial Surface Protein L Binds and Inactivates Neutrophil Proteins S100A8/A91

Bo Åkerstro¨m2 and Lars Bjo¨rck2

Finegoldia magna is an anaerobic bacterial species that is part of the normal human flora on all nonsterile body surfaces, but it is also a significant opportunistic pathogen causing a wide range of infections. Some isolates of F. magna that are more frequently associated with clinical infection express protein L, a surface protein containing multiple homologous domains (B1-B5) that bind Igs through interactions with Ig L chains. The present study shows that the N-terminal A domain of protein L binds S100A8/A9, antibacterial proteins present in large amounts in the cytoplasm of neutrophils, but also extracellularly in tissues during inflammation. As a result, protein L-expressing F. magna are protected against killing by S100A8/A9. Igs and S100A8/A9 were found to interact independently with protein L, demonstrating that this bacterial surface protein is

capable of manipulating both adaptive and innate immune defense mechanisms. The Journal of Immunology, 2009, 183: Downloaded from 4583–4592.

any bacterial species express Ig-binding surface proin the release of proinﬂammatory mediators (14, 15). These teins, among which staphylococcal protein A (1) and different observations indicate that the presence of protein L at M streptococcal protein G (2, 3) are the most well the surface of F. magna enhances the potential pathogenicity of known. These proteins are widely used as biomedical tools to

the bacteria. http://www.jimmunol.org/ detect and bind Abs, and they interact with the H chains of IgG. Members of the human S100 protein family are Ca2ϩ-binding Apart from protein L, an Ig L chain-binding molecule (4), all proteins containing helix-loop-helix motifs, so-called EF hands bacterial proteins to date described that bind Ig in a nonimmune (16). The expression of the S100 proteins is usually regulated fashion interact with Ig H chains. Protein L is expressed by by environmental and developmental factors, and is cell and ϳ10% of clinical isolates of Finegoldia magna (5), an anaer- tissue specific. Two members, S100A8 and S100A9, are major obic Gram-positive bacterium that is part of the normal flora in cytosolic constituents (ϳ40% of the total protein content) in the skin, the mouth, the upper respiratory tract, the gastrointes- neutrophils and monocytes. The apparent molecular masses on tinal tract, and the female genito-urinary tract. F. magna is the SDS-PAGE are 8 and 14 kDa, respectively, and they are also most significant opportunistic pathogen of the anaerobic normal called myeloid-related proteins 8 and 14, migration inhibitory by guest on September 28, 2021 bacterial flora, causing soft tissue abscesses, bone/joint infec- factor-related proteins 8 and 14, or calgranulin a/b (17, 18). tions, wound infections, and vaginosis (6, 7). Protein L-express- Their expression is up-regulated in tissue macrophages, kera- ing F. magna strains are frequently isolated from patients with tinocytes, and epithelial cells in inflammation and cancer (19, vaginosis (5), and when expressed at the surface of Streptococ- 20). S100A8 and S100A9 are secreted from activated phago- cus gordonii, protein L promotes the adhesion of these bacteria cytes by a novel tubulin-dependent mechanism (21), and they to the vaginal mucosa of mice (8). Protein L preferentially binds are found in extracellular fluid and plasma during inflammation Ig L chains of the ␬ type through interactions with the variable (20). The two proteins form heteromeric complexes in vivo, and domain, but without interfering with the Ag binding site (9– at millimolar Ca levels the heterotetramer, (S100A8/A9) ,is 12). The molecule contains multiple homologous Ig binding 2 predominating (22, 23). The tetramer was also named calpro- domains, called B repeats (13), and it activates human basophils tectin due to its ability to inhibit growth of various bacterial and and mast cells by cross-linking surface-associated IgE, resulting fungal species (24). This bacteriostatic property has been as- cribed to the Zn-binding capacity of the tetramer, and it has been suggested that the effect is a result of depletion of bacterial Department of Clinical Sciences, Division of Infection Medicine, Lund University, nutrient Zn ions (25) or altered structure of the protein complex Lund, Sweden (26). The observation that the S100A8/A9 complex adheres to Received for publication May 11, 2009. Accepted for publication July 21, 2009. amastigotes in skin lesions of Leishmania major-infected mice The costs of publication of this article were defrayed in part by the payment of page (27) indicates that S100A8/A9 could also have a function in the charges. This article must therefore be hereby marked advertisement in accordance defense against this parasite. with 18 U.S.C. Section 1734 solely to indicate this fact. The interaction between the B repeats of protein L and Ig L 1 This work was supported by the Swedish Research Council (Projects 7144 and 7480), Swedish Government Funds for Clinical Research (ALF), Swedish Society for chains is well documented, but the properties of the N-terminal A Medical Research, Royal Physiographic Society (Lund), Foundations of Greta and domain (Fig. 3A shows a schematic depiction of protein L) are ¨ Johan Kock, Torsten and Ragnar So¨derberg, Alfred Osterlund, the Blood and Defence unknown. The starting point for the present investigation was the Network, Lund University, the Crafoord Foundation, the Thelma Zoe´ga Foundation, and Hansa Medical AB. finding that a protein L construct containing the A domain bound 2 Address correspondence and reprint requests to Dr. Bo Åkerstro¨m or Dr. Lars S100A8/A9 with high affinity and specificity, whereas the Ig-bind- Bjo¨rck, Lund University, BMC, B14, Solvegatan 19, SE-22184, Lund, Sweden. ing B repeats did not. A characterization of the interaction between E-mail addresses: [email protected] and [email protected] protein L and S100A8/A9 and its consequences for F. magna bi- Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 ology is the theme of this work. www.jimmunol.org/cgi/doi/10.4049/jimmunol.0901487 4584 PROTEIN L BINDS AND INACTIVATES S100 A8/A9

Materials and Methods some cases, binding experiments were also performed in TH broth con- Bacteria taining 0.5% Tween 80 (pH 5.5). F. magna (strains 312, 505, 564, and 644) and group G streptococcal Bacterial growth assay strains (G42, G43, G46, and G148) are clinical isolates from the De- F. magna strains 312 and 505 were grown under strict anaerobic conditions partment of Clinical Microbiology, Lund University Hospital. Strepto- in TH broth (pH 7.5) containing 0.5% Tween 80, to A620 0.5–0.7. A total coccus pneumoniae strains D39 and PR218 are from G. Pozzi, Univer- of 4 ␮l of these bacterial suspensions was added to 200 ␮l of TH (pH 5.5 sity of Siena, and the Streptococcus pyogenes strains (AP1, AP4, AP6, or 7.5) containing 0.5% Tween 80, in UV-transparent cuvettes. At the same AP12, and AP49) are from Institute of Hygiene and Epidemiology (Prague, time, various amounts of S100A8/A9 in 1–2 ␮l of PBS, or 1–2 ␮lofPBS Czech Republic). The two Staphylococcus aureus strains Cowan I and alone as a control, were added to the cuvettes. These manipulations were Wood 46 are from T. Foster, Trinity College (Dublin, Ireland). All strains 3 done in an anaerobic workstation (Electrotek). A620 was determined after were grown in Todd-Hewitt (TH) broth (Difco) at 37°C. In the case of F. 24 h of incubation. In some experiments, the strains were grown at pH 5.5 magna strains, TH contained 0.5% Tween 80, and these isolates were to A620 0.3–0.5 in the UV-transparent cuvettes, followed by the addition of grown under strict anaerobic conditions. 5 ␮g of S100A8/A9. Growth curves were obtained by cultivating the bacteria under strict anaerobic conditions and measuring A at different time Proteins, Abs, sequencing, and radiolabeling 620 points. Protein L (B1-B4) was expressed and purified, as described (13). Factor Xa (FXa) was purchased from ICN Pharmaceuticals. Human IgG was pur- Construction and cloning of a protein L (A-C2) vector chased from Sigma-Aldrich. Protein PAB was isolated from F. magna A DNA fragment coding for an N-terminal FXa-site (isoleucine-glutamic (strain ALB8) culture medium, as described (28). Protein A was purchased acid-glycine-arginine), followed by the domains A, B1, B2, B3, B4, B5, from Sigma-Aldrich, and protein H and M1 protein were purified, as de- C1, and C2 (A-C2) of protein L, and flanked by NarI- and SalI-cleavage scribed (29). The synthetic peptides DNF29, corresponding to aa positions sites, was constructed by PCR, using 10 ␮g of DNA from F. magna strain Downloaded from 66–94 of protein L (A-C2), and MLT24, corresponding to the N-terminal 312 as template and 250 ng of each of the two primers, 5Ј-GCTCAGGCG 24 aa of S100A8, were purchased from Innovagen. Mouse monoclonal GCGCCGATCGAGGGAAGGGCTGATGAACCTATTGATCTTG-3Ј and anti-S100A8/A9 (27E10) was purchased from Abcam. Polyclonal anti- 5Ј-AGGTCGACTTATTATTCAGCTTCTACTGGTGATAATAC-3Ј. The S100A8 Abs were raised in rabbits immunized with MLT24. N-terminal PCR product was purified by agarose gel electrophoresis, cleaved with amino acid sequence analysis was done by 12 cycles of Edman degradation NarI and SalI, and ligated with the expression vector pHD389 (34), cleaved (Protein Analysis Center, Karolinska Institute) of bands separated by SDS- by the same enzymes. Single-positive clones of the Escherichia coli strain PAGE and transfered to polyvinylidene difluoride (PVDF) membranes LB392, transformed with the ligated vector, were selected by PCR, using (Immobilon-P; Millipore), as described (30). Proteins were labeled with the same primer pair, and blotting with 125I-labeled human IgG after trans- http://www.jimmunol.org/ 125 I (Amersham Biosciences; IMS.30) using the chloramin-T method (31). fer of colonies to nitrocellulose membranes (13). Protein-bound iodine was separated from free iodide by gel chromatography on a Sephadex G-25 column (PD10; Amersham Biosciences). A sp. Expression and purification of protein L (A-C2) act. of ϳ0.1–0.2 MBq/␮g proteins was obtained. Expression of protein L (A-C2) was induced in a 50-ml culture of trans- Electrophoresis, blotting, slot binding, and chromatography formed E. coli bacteria (strain LB392) by raising the temperature to 42°C for 3 h. The bacterial pellet was lysed on ice for 5 min by addition of 3.2 SDS-PAGE was performed using 13.5% slab gels in the buffer system mg of lysozyme (Sigma-Aldrich) in 20 ml of 0.25 M Tris-HCl (pH 8), 0.5 described (32), including 2% v/v 2-ME and using Rainbow molecular mass mM EDTA added to 10 mM, and the lysate centrifuged at 10,000 ϫ g for standards purchased from GE Healthcare. The polyacrylamide gels were 20 min. Protein L (A-C2) was then purified by affinity chromatography on stained with Coomassie brilliant blue R-250. For immunoblotting, the gels a human IgG-Sepharose column; dialyzed against 20 mM Tris-HCl (pH by guest on September 28, 2021 were transferred to PVDF membranes (Immobilon-P; Millipore) and then 8.0), 0.1 M NaCl, and 2 mM CaCl ; and cleaved by incubation with FXa 125 2 incubated with antisera or I-labeled proteins, as described (33), using 1 (4 U/mg protein L) for8hatroom temperature. The cleaved protein L mM CaCl2 in the incubation and washing buffers. Images of the mem- (A-C2) was then finally purified by gel chromatography on Sephacryl branes were developed using Fuji FLA 3000 phosphor imaging system S-300 (GE Healthcare), dialyzed against 2 mM NH4HCO3, and freeze (Fujifilm Sweden). In slot-binding experiments, proteins were applied to dried. A smaller form of protein L, protein L (A -C2), lacking the first 69 125 70 Immobilon-P membranes, followed by incubation with I-labeled protein aa at the N terminus, was seen in the bacterial lysate, copurified on the 125 Lor I-labeled S100A8/A9, and developed by phosphor imaging. Pro- IgG-Sepharose column, and separated from full-length protein L on the gel teins were coupled to cyanogen bromide-activated Sepharose CL-4B (GE chromatography step. Healthcare) at a final density of 5–7 mg/ml gel and packed in 2-ml columns. Affinity chromatography was done by applying the sample, rinsing Neutrophil isolation and tissue extracts with PBS/Ca/Zn (10 mM phosphate buffer (pH 7.4), 120 mM NaCl, 3 mM Whole human blood from healthy donors was layered on Polymorph-prep KCl, 1 mM CaCl2, and 0.1 mM ZnCl2), and eluting bound samples with 0.1 ϫ M glycine-HCl (pH 2.3), immediately neutralizing with 1/10 vol 1 M Tris- (Medinor) and centrifuged at 400 g for 35 min at 18°C. The neutrophil layer was recovered and suspended in 50 ml of Ca- and Mg-containing HCl (pH 8.5). Gel chromatography was run manually on a 150-ml column ϫ packed with Sephacryl S-300 (GE Healthcare), equilibrated with 20 mM PBS/1.6 mM MgSO4/1.8 mM CaCl2. After centrifugation at 350 g for 10 min, erythrocytes were removed by hypotonic lysis in water (10:1) for 20 s Tris-HCl, 150 mM NaCl, and 0.02% NaN3 (pH 8.0) at 4°C. The column ϫ ϫ was eluted at a flow rate 3 ml/h, and the eluted fractions were analyzed by and reconstitution with 10 PBS. The cells were then pelleted at 250 g absorbance at 280 nm and SDS-PAGE. (5 min), counted using a hemocytometer, and resuspended in Na medium (5.6 mM glucose, 127 mM NaCl, 10.8 mM KCl, 2.4 mM KH2PO4, 1.6 mM Bacterial binding assay MgSO4, 10 mM HEPES, and 1.8 mM CaCl2; pH adjusted to 7.3 with NaOH) at a concentration of 107 cells/ml. After purification, the neutro- F. magna (four strains), human group C and G streptococci (four strains), phils were gently rotated end-over-end at room temperature. Tissue ex- S. pyogenes (five strains), S. aureus (two strains), and S. pneumoniae (two tracts from fresh placenta tissue were generated, as previously reported strains) were grown in TH broth at 37°C. Bacteria were washed and re- (35), and briefly described in the paragraph below. suspended in PBS containing 0.02% (w/v) of NaN3 and 0.05% of Tween 20 (PBSAT), and the bacterial concentration was adjusted to 2 ϫ 109 Purification of S100A8 and A9 bacteria/ml. A total of 200 ␮l of bacterial suspensions, undiluted or diluted, Frozen, pelleted neutrophil granulocytes were thawed in 5 ml of 0.1 M was incubated with 25 ␮lof125I-labeled protein in PBSAT (ϳ104 cpm) for Tris-HCl, 1 mM CaCl , 0.1 mM ZnCl , and 1 mM PMSF, and centrifuged 30 min at room temperature. A 2.0-ml volume of PBSAT was added, and 2 2 at 14,000 ϫ g for 20 min. S100A8/A9 was purified from the supernatant the suspensions were centrifuged for 15 min at 1800 ϫ g. The radioactivity by affinity chromatography on protein L (A-C2)-Sepharose, and then of the pellets was measured in a gamma counter, and binding was ex- removing Igs, remaining from plasma, by affinity on protein L (B1- pressed as a percentage of the total radioactivity added. Inhibition exper- B4)-Sepharose. Alternatively, the proteins were purified from human iments were performed by adding excess amounts of unlabeled protein. In placenta extracts (35) using sequential protein L (A-C2) and protein (B1-B4) affinity chromatography. Briefly, 200 g of a normal term pla- 3 Abbreviations used in this paper: TH, Todd-Hewitt; FXa, factor Xa; PBSAT, PBS centa, taken within 3 h after delivery, was homogenized in 200 ml of 50 containing 0.02% (w/v) of NaN3 and 0.05% of Tween 20; PVDF, polyvinylidene mM Tris-HCl (pH 8.0), 0.25 M sucrose, 2 mM EDTA, 1 mg/L pepsta- difluoride; SPRIA, solid-phase RIA. tin, 5 mg/L antipain, and 10 mg/L leupeptin, using a Potter-Elvehjem The Journal of Immunology 4585

apparatus with a tight-fitting teflon pestle. The homogenate was centrifuged at 10,000 ϫ g for 10 min. The supernatant was centrifuged at 100,000 ϫ g for 90 min and applied to the columns. Recombinant S100A8 and S100A9 were expressed separately using pET3a vectors with S100A8- and S100A9-encoding inserts (gifts of N. Hogg, Leuko- cyte Adhesion Laboratory, Imperial Cancer Research Fund, London, U.K.) and purified, as described above. The recombinant proteins were only used for affinity measurements with BIAcore. Fluorescence microscopy For immunofluorescence on bacteria exposed to neutrophil lysate, purified neutrophils were disrupted by nitrogen cavitation. The cells were added in a pressurized cell disruption bomb (4639; Parr Instrument) and equilibrated to 600 psi for 5 min (Na medium with Complete Mini protease inhibitors; Roche Diagnostics; 1 tablet per 10 ml solution). The bacteria were incubated with the dispupted neutrophils (12 ϫ 106 per ml) or only Na medium for 5 min at 37°C. Bacteria or bacteria/neutrophil samples were washed three times and fixed in 2% PFA for 45 min at room temperature. This was followed by incubation with blocking FIGURE 1. SDS-PAGE and Western blotting of protein L-binding mol- buffer (Na medium, 5% donkey serum, 50 mM glycine) for 10 min and ecules in human plasma, neutrophils, and placenta. Human plasma, neu- two washes. Primary mouse Ab against S100A8/A9 was added at 1:100 trophil lysate, and placenta extract were separately subjected to affinity overnight. Following two washes, secondary goat Fab against mouse Downloaded from Igs (Alexa 594 conjugated; Invitrogen) was added and incubated for chromatography on protein L (B1-B4)-Sepharose (lanes 1) and protein L 1 h. Finally, after two washes in Na medium, the samples were resus- (A-C2)-Sepharose (lanes 2) columns. Bound proteins were eluted, sepa- pended in 50 ␮l of Na medium, adhered to poly(L-lysine) (m.w. rated by SDS-PAGE (13.7%), and either stained by Coomassie brilliant 150,000; Sigma-Aldrich)-coated glass coverslips for 60 min, and then blue (left panel, stain) or transferred to PVDF membranes. The membranes mounted using ProLong Gold antifade reagent (Invitrogen). were either stained and protein bands cut out from the gel and sequenced by Edman degradation, or blotted against 125I-labeled protein L (B1-B4) or Electron microscopy 125I-labeled protein L (A-C2) (right panel, blot). The two bands identified http://www.jimmunol.org/ F. magna strains 312 and 505 were separately grown in TH broth contain- as S100A8 and S100A9 by their amino acid sequences are denoted. ing 0.5% Tween 80 (pH 5.5) under strict anaerobic conditions in UV-

transparent cuvettes to an OD620 of 0.5–0.7. S100A8/A9, labeled with colloidal thiocyanate gold, as previously described (36), was added to the described in the legends to the figures. The pH dependence was studied ␮ cultures (5 g/ml growth medium). After another 22 h of incubation, sam- using 0.1 M Na-acetate (pH 4–6.5) or 0.1 M Tris (pH 7–8.5) plus 1 mM ples from the cultures were adsorbed onto carbon-coated copper grids for CaCl2 and 0.1 mM ZnCl2. After washing, the wells were cut, and bound 2 min, washed briefly on two drops of water, and negatively stained on two 125I-labeled proteins were estimated by counting in a Wallac Wizard 1470 drops of 0.75% uranyl formate. The grids were rendered hydrophilic by gamma counter (PerkinElmer Life Sciences). glow discharge at low pressure in air. Specimens were observed in a Jeol JEM 1230 electron microscope operated at 60 kV accelerating voltage. Molecular modeling Images were recorded with a Gatan MultiScan 791 charge-coupled device by guest on September 28, 2021 camera. To construct a representative figure of protein L comprising part of the A domain, the B1 domain, and the B2 domain (aa 66–227), aa 70–77 were BIAcore surface plasmon resonance interaction analysis added to the N terminus of the nuclear magnetic resonance structure of the ϳ ␮ B1 domain (aa 78–155; Protein Data Bank code 2PTL) (38). The amino Protein L was diluted in 10 mM sodium acetate (pH 4; 10 g/ml) and acids were added in an extended conformation and minimized. The B2 immobilized via amine coupling to flow cells of a CM5 sensorchip (BIA- ϳ domain, which has an almost identical amino acid sequence as the B1 core). Immobilization levels were optimized to 600 response units. A domain, was represented by the B1 domain structure truncated with the first flow cell subjected to the immobilization protocol, but without any addition six residues and was arranged with the N terminus overlapping the COOH of protein, was used as control. For affinity measurements, the binding and terminus of the B1 domain. The binding site for the Ig L chain was derived dissociation phases were monitored in a BIAcore 2000 instrument. In con- from the crystal structure of the protein L-Ab complex (Protein Data Bank trol experiments for possible mass transfer limitations, S100A8/A9 was code 1HEZ) (12). injected over the protein L surface at different flow rates. No differences in initial binding were observed at 5 ␮l/min or above, indicating no limitations to any combinations. S100 proteins were then injected at different Results concentrations (typically 6.3–100 nM) at 35 ␮l/min and 25°C over the flow Purification of protein L (A-C2) cell in running buffer: 10 mM HEPES (pH 7.5), 150 mM NaCl, 0.005% Protein L (A-C2), comprising the domains A, B1-B5, and C1-C2 surfactant P20, 1 mM CaCl2, and 0.1 mM ZnCl2. In some experiments, the calcium and zinc-chloride were replaced by 3.4 mM EDTA. Between ex- (see Fig. 3A), was expressed in E. coli and isolated from the periments, the surfaces were strictly regenerated with pulses of 0.1 M periplasma by affinity chromatography on IgG-Sepharose, as de- NaHCO3 (pH 12) and with running buffer containing 2 M NaCl, followed scribed in Materials and Methods. To obtain a homogenous N by an extensive wash procedure after reaching baseline. After X and Y terminus, the protein was cleaved by FXa at a site introduced im- normalization of data, the blank curves from the control flow cell of each mediately N-terminal of the A domain, and purified by gel chro- injected concentration were subtracted. The association (ka) and dissociation (kd) rate constants were determined simultaneously using the equation matography. The final yield was 11 mg/L E. coli culture, the N- for 1:1 Langmuir binding in the BIA Evaluation 4.1 software (BIAcore). terminal sequence was determined to ADEPIDLEKL, and the The binding curves were fitted locally, and the equilibration dissociation apparent Mr on SDS-PAGE was 65 kDa. constants (KD) were calculated from mean values of the obtained rate constants. Protein L (A-C2) binds S100A8/A9 Solid-phase RIA To investigate whether protein L (A-C2) has affinity for human Binding between protein L and S100 proteins was studied by solid-phase proteins other than Ig, human plasma, neutrophil lysate, and ex- RIA (SPRIA), essentially as described (37). Briefly, the proteins were tract from human placenta were analyzed by Western blotting us- coated to the plastic by incubation at 1–4 ␮g/ml in PBS overnight at 4°C. ing protein L (A-C2) as the probe. Apart from a band of 25 kDa After washing with 0.15 M NaCl, 0.05% Tween 20, and 0.1% BSA, the ϳ wells were incubated 1–3 h at room temperature with 125I-labeled protein corresponding to Ig L chains, two additional bands 10 kDa were Lor125I-labeled S100A8/A9 (10–50 ng/ml) in 20 mM Tris-HCl (pH 8.0), clearly detected in two of the materials: the placenta extract and

0.15 M NaCl, 1 mM CaCl2, and 0.1 mM ZnCl2, with various additions, as the neutrophil lysate (data not shown). These materials and plasma 4586 PROTEIN L BINDS AND INACTIVATES S100 A8/A9 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 2. Characterization of the binding between S100A8/A9 and protein L. A–C, SPRIA, coating microtiter plate wells with S100A8/A9 (1 ␮g/ml), followed by incubation with 125I-labeled protein L (A-C2) (50 ng/ml) in the presence of a dilution series of nonlabeled IgG (E) or protein L (A-C2) (F) (A), EDTA (B), or heparin (C). D, Heparin-Sepharose affinity chromatography of 125I-labeled protein L (A-C2) (0.5 ␮g/ml) in the presence of 0.5 mg/ml S100A8/A9 (F) or buffer only (E). Arrows indicate sample application and elution of the bound proteins with 0.1 M glycine-HCl (pH 2.3). E, SPRIA, coating microtiter plate wells with protein L (A-C2) or protein L (B1-B4) and incubating with 125I-labeled S100A8/A9 in buffers with various pH, as described in Materials and Methods. F, BIAcore surface plasmon resonance analysis of immobilized protein L (A-C2) and S100A8/A9 in the fluid phase at 100, 50, 25, 12.5, and 6.75 ␮g/ml (graphs shown from top to bottom). The S100A8/A9 proteins used were purified from neutrophil lysates. A, B, C, and E, Mean values Ϯ SD of three determinations are shown. were separately subjected to affinity chromatography on either a eluates, but the bands corresponding to S100A8/A9 only reacted protein L (B1-B4)- or a protein L (A-C2)-Sepharose column. with radiolabeled protein L (A-C2), again demonstrating that the Eluted proteins were separated by SDS-PAGE, and the bands (Fig. A-C2 fragment, but not the B1-B4 fragment, binds S100A8/A9. It 1, stain) were identified by N-terminal amino acid sequence anal- should be noted that the Western blot experiments show that pro- ysis after transfer to PVDF membranes. All three sources con- tein L (A-C2) binds each monomer separately. Similar results were tained different Ig H chains (50–70 kDa) and Ig L chains (25 kDa), obtained with the neutrophil eluates, whereas blotting of plasma and in the neutrophil and placenta material two additional bands at eluates with 125I-labeled protein L (A-C2) did not show any bands 8 and 14 kDa were eluted from the protein L (A-C2) column, but in the position of S100A8/A9 (data not shown). not from the protein L (B1-B4) column. These bands, identified as The binding between purified S100A8/A9 and protein L (A-C2) S100A8 and S100A9, respectively, were not present in any of the was then further characterized. As shown in Fig. 2A, IgG did not eluates from plasma (Fig. 1, left panel, stain). 125I-labeled protein inhibit the binding of S100A8/A9 to protein L (A-C2), suggesting L (A-C2) or protein L (B1-B4) were used as probes in Western separate binding sites. The binding was inhibited by EDTA, which blot analysis of the placenta proteins eluted from the two columns is consistent with a conformational dependence of S100A8/A9 on (Fig. 1, right panel, blot). The probes detected Ig L chains in both Ca2ϩ ions (Fig. 2B). Heparin inhibited the binding to ϳ50% (Fig. The Journal of Immunology 4587

2C), and protein L (A-C2) bound strongly to a heparin-Sepharose column in the presence, but not the absence, of S100A8/A9 (Fig. 2D). A binding maximum was seen at pH 5.5–7.5 (Fig. 2E), and

the equilibrium dissociation constant (KD) of the binding between protein L (A-C2) and neutrophil S100A8/A9 was determined at pH 7.5 by surface plasmon resonance analysis to 2.4 nM. Similar dissociation constants were obtained for S100A8 and S100A9 separately, using recombinant proteins (Fig. 2F). Addition of EDTA to the buffers abolished the binding, and protein L (B1-B4)-coupled chips did not display any binding to S100A8/A9 (data not shown).

Deﬁnition of the binding site for S100A8/A9 on protein L The interaction of S100A8/A9 with different fragments of protein L was examined (Fig. 3, A and B). Binding was seen to protein L

(A-C2) and the N-terminally truncated protein L (A70-C2) at approximately the same strength, as measured by SPRIA and West- ern blotting. A binding was also seen to the synthetic peptide DNF29 (corresponding to aa residues 66–94 of intact protein L). Downloaded from This interaction was weaker and could only be detected by inhibition of the 125I-labeled protein L binding to microtiter plates coated with S100A8/A9. No binding was seen to protein L (B1- B4) or to protein PAB from F. magna strain ALB8. The latter protein contains a C domain with 69% amino acid sequence iden-

tity to the C2 domain of protein L (28), suggesting that C domains http://www.jimmunol.org/ are not involved in the interaction with S100A8/A9. The strong

binding seen with the A70-C2 construct and the lack of binding

with the B1-B4 construct indicate that amino acids D70-K80,lo- cated where the A domain is connected to the B1 domain (K80 is the N terminus of B1; see Fig. 3, A and C), are essential for the binding to S100A8/A9. The weak interaction with the DNF29 pep-

tide, which contains the amino acids D70-K80, is probably due to the absence of tertiary structure of the peptide, and the results described above cannot exclude that amino acid residues in B1 are by guest on September 28, 2021 also part of the binding site. However, it is clear from the com- petition experiments (Fig. 2A) that the binding sites on protein L for Ig L chains and S100A8/A9 are different.

Binding of S100A8/A9 to whole bacteria and isolated bacterial proteins A strong binding of S100A8/A9 was seen to two protein L-carrying F. magna strains (312 and 564), whereas only background FIGURE 3. Mapping of the binding site of protein L on S100A8/A9. A, binding was detected to two strains (505 and 644) lacking protein Construction of protein L fragments. The domains of protein L are desig- L (Fig. 4, A and B). All (four of four) tested group G streptococcal nated A, B, C, W, and M (13). The protein is bound to the bacterial surface strains displayed a weak binding, and the five tested strains of S. through the COOH-terminal membrane-spanning M domain and the cell pyogenes bound S100A8/A9 at or just above background level. wall-associated W domain. The function of the C domains is unknown, None of the S. aureus or the S. pneumoniae strains showed affinity whereas the B repeats are responsible for Ig L chain binding. The N- terminal A domain ends with aa residue 79. Protein L (A-C2) and an for the protein (Fig. 4A). As expected from the experiments described above, the binding of S100A8/A9 to F. magna 312 could N-terminally truncated fragment of protein L (A70-C2), lacking the first 69 aa, were prepared as described in this work. The construct L (B1-B4) has not be inhibited by IgG (Fig. 4C). We also tested the binding of been described (13), and the peptide DNF29 was synthesized commer- radiolabeled S100A8/A9 to a number of purified bacterial Ig-bind- cially. The binding to S100A8/A9, purified from neutrophil lysates, was ing surface proteins applied in various amounts to Immobilon fil- analyzed by Western blotting, slot binding, and SPRIA. Protein L (A-C2) ters (Fig. 4D). Binding was seen to protein L (A-C2) and protein and protein L (A70-C2) displayed positive binding with all methods (ϩϩ); G from group G and C streptococci, whereas protein A from S. protein L (B1-B4) did not bind with any method (Ϫ), and DNF29 displayed aureus and proteins H and M1 protein from S. pyogenes did not ϩ 125 weak binding in SPRIA ( ). B, Inhibition of the binding of I-labeled bind the radiolabeled probe. All of these initial binding experi- protein L (A-C2) to microtiter plate-coated S100A8/A9 by various protein ments were performed at pH 7.5. L constructs. Microtiter plate wells were coated with S100A8/A9 (1 ␮g/ ml), followed by incubation with 125I-labeled protein L (A-C2) (50 ng/ml) in the presence of 0.2 mg/ml protein L fragments. Control wells were coated with buffer only. Mean values of four experiments Ϯ SD are shown. Materials and Methods and is based on the nuclear magnetic resonance The inhibition by DNF29 was statistically significant (p ϭ 7.7 ϫ 10Ϫ5). structure of the B1 domain and its Ig L chain-interacting site (38, 56), and

The S100A8/A9 proteins used were puriﬁed from neutrophil lysates. C, the structure of a protein L-Ab complex (12, 57). Amino acids (D70-K80) Three-dimensional model of protein L (A70-B2) indicating binding sites for of the A domain involved in the binding to the S100A8/A9 complex are S100A8/A9 and Ig L chains. The model was constructed as described in marked in red color. 4588 PROTEIN L BINDS AND INACTIVATES S100 A8/A9 Downloaded from http://www.jimmunol.org/

FIGURE 4. Binding of S100A8/A9 to bacteria and bacterial proteins. A, 125I-labeled S100A8/A9 was incubated with various bacterial species and strains, as described in Materials and Methods. The binding is shown as percentage of added radiolabeled protein (mean Ϯ SEM). B, Binding of 125I-labeled S100A8/A9 to F. magna strains 312 and 505 at various concentrations starting at 2 ϫ 1010 bacteria/ml. Each point is the mean Ϯ SEM of three determinations. C, The binding of 125I-labeled S100A8/A9 (ϳ0.1 ␮g/ml) to F. magna strain 312 (2 ϫ 1010 bacteria/ml) in the presence of various

concentrations of nonlabeled S100A8/A9 or human IgG. Binding is shown as percentage of inhibition (mean Ϯ SEM) after subtracting the background by guest on September 28, 2021 binding of tubes without cells. D, Slot binding of 125I-labeled S100A8/A9 to dilutions series of puriﬁed bacterial proteins immobilized on PVDF membranes. S100A8/A9, puriﬁed from neutrophil lysates, was used for all experiments.

S100A8/A9 released from disrupted neutophils bind to protein L different concentrations of S100A8/A9 (added to the cultures at

expressing F. magna bacteria OD620: 0.4–0.5). At this pH, no growth inhibition was recorded High concentrations of S100A8/A9 are found in the cytosol of for the two strains, even at the highest concentration tested (8 ␮ neutrophils and monocytes, and the proteins are released from ac- g/ml). Tween 80 at this concentration did not affect the binding tivated neutrophils by unknown tubulin-dependent mechanisms to the bacteria at pH 7.5 (data not shown). In the skin, at mucosal (21). As mentioned above, they are also found at high concentra- surfaces in the vagina, and in mouth and tissue abscesses where tions extracellularly in inflamed tissues. To investigate whether protein L-expressing strains of F. magna are commonly isolated ϳ S100A8/A9 released from neutrophils interact with F. magna, neu- (5, 39), the pH is acidic at 5–6 (40–43). The effect of trophils were mechanically disrupted by nitrogen cavitation. Bac- S100A8/A9 was therefore tested in TH broth also at pH 5.5. In teria of strains 312 and 505 were incubated with this material, and these experiments, increasing concentrations of S100A8/A9, analysis of the bacteria with immunofluorescence (Fig. 5) revealed present from the time of inoculation, inhibited the bacterial growth that S100A8/A9 bound strongly to protein L-expressing F. magna in a dose-dependent manner, but 312 bacteria were significantly 312 bacteria (Ͼ99% of the bacteria were stained), but only weakly more resistant. Thus, S100A8/A9 at 8 ␮g/ml resulted in an 80% to bacteria of strain 505 (ϳ10%). These experiments were also growth inhibition of strain 505, but only 35% inhibition of strain performed at pH 7.5. 312 (Fig. 6A). The difference between the strains was further pro- nounced when S100A8/A9 was added to bacterial cultures at ex- Protein L confers resistance to S100A8/A9-induced bacterial ponential growth phase (A620: 0.4–0.5). Under these conditions, killing the growth of 505 bacteria was completely inhibited, whereas no S100A8/A9 has been reported to have antibacterial activity (24), significant effect was seen on bacteria of strain 312 (Fig. 6B). To but a possible effect on anaerobic bacteria has not been investi- investigate whether the growth inhibition of strain 505 at pH 5.5 is gated previously. We hypothesized that S100A8/A9 is antibacte- due to a physical interaction between S100A8/A9 and the bacterial rial also for F. magna, but that strains carrying protein L are pro- surface, the binding of 125I-labeled S100A8/A9 was tested to both tected against killing by S100A8/A9. This hypothesis was tested in 312 and 505 bacteria in TH broth (pH 5.5). The background bind- microcultures under strict anaerobic conditions, in which the ing to the test tubes with no bacteria present was 1.7% Ϯ 0.4 growth of F. magna strains 312 and 505 was measured in TH (mean Ϯ SEM), compared with 8.5% Ϯ 0.9 for 312 and 6.8% Ϯ medium containing 0.5% Tween 80 (pH 7.5) in the presence of 0.6 for 505 bacteria. Thus, although the two strains bind The Journal of Immunology 4589 Downloaded from http://www.jimmunol.org/ FIGURE 6. Protein L-expressing F. magna bacteria are resistant to S100A8/A9 killing. S100A8/A9 isolated from neutrophils was used for all experiments. Bacteria were cultivated under strict anaerobic conditions at pH 5.5, and growth was measured by reading the absorbance at 620 nm. Results are shown as mean value Ϯ SD of triplicate experiments. The growth of strains 312 and 505 was statistically compared at different time .p Ͻ 0.001, according to Student’s t test. A, F ,ءءء ;p Ͻ 0.05 ,ء ;points magna strain 312 (F) expressing protein L or strain 505 (Ⅺ) devoid of protein L was grown for 24 h at 37°C in the presence of various concen-

trations of S100A8/A9. The absorbances at 24 h are relative to the mean by guest on September 28, 2021 value of respective strain with no addition of S100A8/A9. B, S100A8/A9 FIGURE 5. S100A8/A9 from disrupted neutrophils bind to protein L- was added to the cultures to a ﬁnal concentration of 23 ␮g/ml 22 h after expressing F. magna bacteria. F. magna 312 or F. magna 505 bacteria inoculation. F. magna strain 312 with (f) or without addition (F)of were incubated with mechanically disrupted neutrophils or with buffer S100A8/A9, and strain 505 with (Ⅺ) or without addition (E)of alone. After washing, the bacteria were stained with anti-S100A8/A9 and S100A8/A9. Alexa 594-conjugated secondary Abs. Differential interference contrast (DIC; left column) and immunoﬂuorescence (right column) show localiza- Ͼ tion of S100A8/A9 to a high degree with F. magna 312 ( 99%), and to a (Fig. 7, lower panel). In summary, the results described in this much lesser extent with F. magna 505 (Ͻ10%). None of the strains show paragraph demonstrate that protein L protects F. magna from kill- staining following incubation with buffer alone. Scale bar: 10 ␮m. ing by S100A8/A9.

S100A8/A9 to a similar degree at pH 5.5 in TH broth (in contrast Discussion to the much higher binding to protein L-expressing 312 bacteria at Protein L is an elongated, fibrous surface molecule of the human pH 7.5 shown in Fig. 4B), 312 is more resistant to the antibacterial anerobic commensal and potential pathogen F. magna. The mol- activity of the protein complex. In an attempt to visualize this ecule consists of multiple exposed domains, and five of these, activity, gold-labeled S100A8/A9 (23 ␮g/ml) was added to cul- the B1-B5 repeats, bind to human Ig L chains. In this work, we tures of 312 and 505 grown exponentially in TH broth (pH 5.5) for show that the outermost domain, the A domain, has high affinity 22 h. After another 18 h of incubation, when 505 in contrast to 312 for the neutrophil cytosol proteins S100A8 and S100A9. The bacteria had stopped to grow (see Fig. 6B), the two strains were S100A8/A9 binding site on protein L was localized to a short analyzed by electron microscopy. As shown in Fig. 7, the picture segment (Ͻ10 aa) most proximal to the neighboring B1 domain. is strikingly different. The cell wall of 312 bacteria is intact, and The binding was dependent on an intact B1 domain, suggesting the gold-labeled S100A8/A9 is associated with fibrous projections that the side groups forming the binding site are coordinated protruding from the bacterial surface, and no gold-labeled material by the folded B1 domain. Despite the closeness to the B1 do- is detected within the cell wall close to the membrane (upper main, the binding of Ig and S100A8/A9 did not inhibit each panel). Physicochemical analysis of protein L has revealed a other, indicating full sterical separation of the two binding sites. Stokes radius of 4.74 nm and a frictional ratio of 1.70, suggesting The N-terminal 70 aa of the A domain did not appear to par- an elongated fibrous structure (9), which fits nicely with the ticipate in the binding to S100A8/A9. This part of the molecule S100A8/A9-binding surface protrusions. Contrary to the intact 312 may therefore have other functions, or bind to other, hitherto bacteria, the cell walls and membranes of 505 bacteria are disin- unknown ligands of the human host. tegrated and gold-labeled S100A8/A9 is found at the membranes. The binding to S100A8/A9 was dependent on the presence of Frequent blebs with ejected cytoplasmic material are also seen divalent cations (Fig. 2B). The conformation of both these EF hand 4590 PROTEIN L BINDS AND INACTIVATES S100 A8/A9

mechanism. The results show that S100A8/A9 kills F. magna bacteria of strain 505 and that the bactericidal effect most likely is a consequence of an interaction of S100A8/A9 with the cell membrane, causing lysis of the bacteria. To our knowledge, no previous report has described a bactericidal effect of the proteins, although a decreased number of cells of the bacterial strain Capnocytophaga sputigena was noted after several days of growth in the presence of S100A8/A9 (53). The F. magna strain 312 was not killed by the addition of S100A8/A9, and the results strongly indicate that the resistance to killing is conferred by the surface protein L. The electron microscopy experiments suggest that protein L forms a barrier that binds the neutrophil proteins at a distance from the cell surface, thereby preventing bacterial lysis. The bactericidal effect of S100A8/A9 on the F. magna strain 505 was recorded at pH 5.5, but not at pH 7.5. The lower pH value is relevant in the skin, at mucosal surfaces in the vagina, and in mouth and tissue abscesses where protein L-expressing strains of F. magna are commonly isolated (5, 40–44). It is noteworthy that the acidic pH in tissues and wound ﬂuids dur- Downloaded from ing infections and inﬂammation coincides with elevated levels of S100A8/A9 (54), and that a similar pH dependence of antimicrobial activity vs Gram-positive and Gram-negative bacte- FIGURE 7. Electron microscopy analysis of F. magna exposed to gold- ria, as well as the fungus Candida albicans, was reported for labeled S100A8/A9. Strains 312 and 505 were grown in TH broth con- peptides with sequences derived from high m.w. kininogen and

taining 0.5% Tween 80 under strict anaerobic conditions at pH 5.5. During http://www.jimmunol.org/ histidine-rich glycoprotein (55). In that study, it was speculated exponential growth phase (OD620: 0.5–0.7), gold-labeled S100A8/A9 was added to the cultures (5 ␮g/ml growth medium). After another 22 h of that a positive charge of the histidines is a prerequisite for an incubation, the bacteria were subjected to negative staining and electron interaction with the bacterial membrane. Protonation of histi- microscopy. In 312 bacteria (upper panel), the cell wall is intact and the dine requires a pH below 6, the pKa value of the histidine side gold label is found in the periphery associated with ﬁbrous, hair-like pro- group. Both S100A8 and A9 have histidine-rich regions (45, jections. In 505 bacteria (lower panel), the membrane integrity is severely 47), and a possible explanation for the pH dependence of the disturbed, causing blebbing and ejection of cytoplasm (arrows). Scale bar: bactericidal effect may be that it is dependent on a positive 100 nm. charge of the histidine-rich regions of one or both of the sub-

units. No killing of the bacteria by S100A8/A9 was seen at pH by guest on September 28, 2021 ϩ 7.5. The strong binding of S100A8/A9 by protein L also at this proteins is dependent on binding of Ca2 ions, and chelation of pH may therefore reflect another function of this interaction, calcium with EDTA therefore leads to a conformational change, and is currently under investigation. which may explain the loss of binding to protein L. The proteins Anaerobic species dominate the normal bacterial flora in hu- appear as a (S100A8/A9)2 tetramer when calcium loaded (22, 23), suggesting that protein L actually binds to the tetramer in vivo. mans, and the significance of anaerobic bacteria, both in the nor- However, the results from the Western blotting experiments, mal flora and as pathogens, is probably much underestimated due which were done in the presence of calcium, show that protein L to difficulties obtaining good quality specimens and experimental also interacts with each of the monomers when separated from problems related to the requirement for anaerobic growth condi- each other (Fig. 1). This was confirmed by BIAcore experiments in tions. F. magna is one of the most ubiquitous anaerobes, and which protein L bound to recombinant S100A8 and S100A9 alone among the anaerobic bacteria of the indigenous flora, it is also the (data not shown), indicating that protein L can bind to all four species most frequently isolated in pure culture from patients with subunits of the tetramer. The S100A8/A9 tetramer was reported to clinical infections (6). The majority of protein L-expressing F. bind directly to heparin, heparan-sulfate, and chondroitin-sulfate magna strains have been isolated from patients with wound infec- via a binding site on S100A9 (44). Approximately 50% of the tions, abscesses, and vaginosis, suggesting that the protein adds protein L binding was inibited by heparin (Fig. 2C), and it may be selective advantages to F. magna in an inflammatory environment speculated that the binding to one of the S100 subunits is inhibited where the pH is acidic and the concentration of S100A8/A9 is by heparin. high. For instance, the protein L-expressing 312 strain used in the A high-affinity Zn2ϩ binding site has been proposed in each of present work was isolated in pure culture from a patient with vagi- the two S100 chains (45). Several reports suggest that S100A8/A9 nosis, hospitalized after a spontaneous abortion. Although most of inhibits bacterial and fungal growth by zinc chelation and deple- the protein L is bound to the bacterial surface, some is also re- tion of zinc, a necessary component of microbial growth (20, 24, leased into the medium during growth (4, 5). As mentioned, pro- 25, 46–48). Furthermore, the proteins bind arachidonic acid, and tein L, by cross-linking cell-bound IgE, can induce a powerful depletion of arachidonic acid has also been suggested as a mech- inflammatory response (14, 15). The rationale for a member of the anism of bacterial growth inhibition (49, 50). Observations sug- normal flora to induce inflammation is not obvious, but there could gesting that S100A8/A9 contributes to mucosal innate immunity be nutritional-ecological reasons. Increased vascular permeability are that the complex inhibits invasion of epithelial cells by bacte- will cause an influx of nutrient-rich plasma, but the inflammation rial pathogens (51), and a recent study showing that S100A9 is will also activate a number of host defense mechanisms, produc- crucial for keratinocyte resistance to invasion by Listeria mono- tion of chemokines and antimicrobial peptides, recruitment of neu- cytogenes and Salmonella typhimurium (52). In the present study, trophils, etc. However, provided that a bacterium with proinflam- we provide evidence for another and more direct antibacterial matory capacity also can counteract these mechanisms more The Journal of Immunology 4591 efficiently than other species in the same ecological niche, this will 19. Wilkinson, M. M., A. Busuttil, C. Hayward, D. J. Brock, J. R. Dorin, and generate a selective advantage. V. van Heyningen. 1987. Expression pattern of two related cystic fibrosis-associated calcium-binding proteins in normal and abnormal tissues. J. Cell Sci. 91: The major finding of the present work is that a bacterial surface 221–230. protein specifically and independently interacts with both Igs and 20. Stritz, I., and I. Trebichavsky. 2004. Calprotectin: a pleiotropic molecule in acute antibacterial proteins. To our knowledge, this is a unique property and chronic inflammation. Physiol. Res. 53: 245–253. 21. Rammes, A., J. Roth, M. Goebeler, M. Klempt, M. Hartmann, and C. Sorg. 1997. of protein L that emphasizes the complexity of host-microbe re- Myeloid-related protein (MRP) 8 and MRP14, calcium-binding proteins of the lationships and demonstrates how diffuse the borderline is between S100 family, are secreted by activated monocytes via a novel, tubulin-dependent bacterial commensalism and pathogenicity. pathway. J. Biol. Chem. 272: 9496–9502. 22. Strupat, K., H. Rogniaux, A. van Dorsselaer, J. Roth, and T. Vogl. 2000. Calci- um-induced non-covalently linked tetramers of MRP8 and MRP14 are confirmed Acknowledgments by electrospray ionization-mass analysis. J. Am. Soc. Mass Spectrom. 11: 780–788. We are grateful to the following colleagues at the Division of Infection 23. Leukert, N., T. Vogl, K. Strupat, R. Reichelt, C. Sorg, and J. Roth. 2006. Cal- Medicine: Dr. Maria Allhorn for performing initial construction and ex- cium-dependent tetramer formation of S100A8 and S100A9 is essential for bio- pression of protein L (A-C2); Monika Heidenholm, Pontus Nordenfeldt, logical activity. J. Mol. Biol. 359: 961–972. and Dr. Hans Tapper for the fluorescence microscopy; Dr. Anders Olin for 24. Steinbakk, M., C. F. Naess-Andresen, E. Lingaas, I. Dale, P. Brandtzaeg, and BIAcore data; Maria Baumgarten and Dr. Mattias Mo¨rgelin for electron M. K. Fagerhol. 1990. Antimicrobial actions of calcium-binding leukocyte L1 microscopy; Dr. Arne Egersten, Dr. Artur Schmidtchen, and Dr. Ole Sø- protein, calprotectin. Lancet 336: 763–765. 25. Sohnle, P. G., M. J. Hunter, B. Hahn, and W. J. Chazin. 2000. Zinc-reversible rensen for stimulating discussions; and Ingbritt Gustafsson for performing antimicrobial activity of recombinant calprotectin (migration inhibitory factor- most of the experiments. Dr. Bjo¨rn Walse (Saromics) is also gratefully related proteins 8 and 14). J. Infect. Dis. 182: 1272–1275. acknowledged for the molecular modeling. 26. Murthy, A. R. K., R. I. Lehrer, S. S. L. Harwig, and K. T. Miyasaki. 1993. In vitro candidastatic properties of the human neutrophil calprotectin complex. J. Immu- Downloaded from nol. 151: 6291–6301. Disclosures 27. Goto, Y., C. Sanjoba, M. Asada, K. Saeki, T. Onodera, and Y. Matsumoto. 2008. The authors have no financial conflict of interest. Adhesion of MRP8/14 to amastigotes in skin lesions of Leishmania major-infected mice. Exp. Parasitol. 119: 80–86. 28. De Chateau, M., and L. Bjo¨rck. 1994. Protein PAB, a mosaic albumin-binding References bacterial protein representing the first contemporary example of module shuffling. 1. Forsgren, A., and J. Sjo¨quist. 1966. “Protein A” from S. aureus. I. Pseudo-im- J. Biol. Chem. 269: 12147–12151. mune reaction with human ␥-globulin. J. Immunol. 97: 822–827. 29. Åkesson, P., K. H. Schmidt, J. Cooney, and L. Bjo¨rck. 1994. M1 protein and http://www.jimmunol.org/ 2. Bjo¨rck, L., and G. Kronvall. 1984. Purification and some properties of strepto- protein H: IgGFc- and albumin-binding streptococcal surface proteins encoded by coccal protein G, a novel IgG-binding reagent. J. Immunol. 133: 969–974. adjacent genes. Biochem. J. 300: 877–886. 3. Reis, K. J., E. M. Ayoub, and M. D. P. Boyle. 1984. Streptococcal Fc-receptor. 30. Matsudaira, P. 1987. Sequence from picomole quantities of proteins electroblot- I. Isolation and partial characterization of the receptor from a group C strepto- ted onto polyvinylidene difluoride membranes. J. Biol. Chem. 262: 10035–10038. coccus. J. Immunol. 132: 3091–3097. 31. Greenwood, F. C., W. M. Hunter, and J. S. Glover. 1963. The preparation of 4. Bjo¨rck, L. 1988. Protein L: a novel bacterial cell wall protein with affinity for Ig 131I-labelled human growth hormone of high specific radioactivity. Biochem. J. L chains. J. Immunol. 140: 1194–1197. 89: 114–123. 5. Kastern, W., E. Holst, E. Nielsen, U. Sjo¨bring, and L. Bjo¨rck. 1990. Protein L, a bacterial immunoglobulin-binding protein and possible virulence determinant. 32. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the Infect. Immun. 58: 1217–1222. head of bacteriophage T4. Nature 227: 680–685. 6. Murdoch, D. A. 1998. Gram-positive anaerobic cocci. Clin. Microbiol. Rev. 11: 33. Wester, L., M. U. Johansson, and B. Åkerstro¨m. 1997. Physicochemical and ␣ by guest on September 28, 2021 81–120. biochemical characterization of human 1-microglobulin expressed in baculov- 7. Stephens, P., I. B. Wall, M. J. Wilson, K. E. Hill, C. E. Davies, C. M. Hill, irus-infected insect cells. Protein Expression Purif. 11: 95–103. K. G. Harding, and D. W. Thomas. 2003. Anaerobic cocci populating the deep 34. Dalbøge, H., E. B. Jensen, H. Tøttrup, A. Grubb, M. Abrahamsson, I. Olafsson, tissues of chronic wounds impair cellular wound healing processes in vitro. and S. Carlsen. 1989. High-level expression of active human cystatin C in Esch- Br. J. Dermatol. 148: 456–466. erichia coli. Gene 79: 325–332. 8. Ricci, S., D. Medaglini, H. Marcotte, A. Olseń, G. Pozzi, and L. Bjo¨rck. 2001. 35. Berggård, T., J. J. Enghild, S. Badve, C. M. Salafia, L. Lo¨gdberg, and ␣ Immunoglobulin-binding domains of peptostreptococcal protein L enhance vag- B. Åkerstro¨m. 1999. Histological distribution and biochemical properties of 1- inal colonization of mice by Streptococcus gordonii. Microb. Pathog. 30: microglobulin in human placenta. Am. J. Reprod. Immunol. 41: 52–60. 229–235. 36. Baschong, W. J., M. Lucocq, and J. Roth. 1985. “Thiocyanate gold”: small (2–3 9. Åkerstro¨m, B., and L. Bjo¨rck. 1989. Protein L: an immunoglobulin light chain- nm) colloidal gold for affinity cytochemical labeling in electron microscopy. His- binding protein: characterization of binding and physicochemical properties. tochemistry 83: 409–411. J. Biol. Chem. 262: 19740–19746. 37. Elbashir, M. I., B. H. K. Nilson, P. Åkesson, L. Bjo¨rck, and B. Åkerstro¨m. 1990. 10. Nilson, B. H. K., A. Solomon, L. Bjo¨rck, and B. Åkerstro¨m. 1992. Protein L from Antibody response in immunized rabbits measured with bacterial immunoglob- Peptostreptococcus magnus binds to the ␬ light chain variable domain. J. Biol. ulin-binding proteins. J. Immunol. Methods 135: 171–179. Chem. 267: 2234–2239. 11. Beckingham, J. A., S. P. Bottomley, R. Hinton, B. J. Sutton, and M. G. Gore. 38. Wikstro¨m, M., T. Drakenberg, S. Forseń,U.Sjo¨bring, and L. Bjo¨rck. 1994. 1999. Interactions between a single immunoglobulin-binding domain of protein L Three-dimensional solution structure of an immunoglobulin light chain-binding from Peptostreptococcus magnus and a human ␬ light chain. Biochem. J. 240: domain of protein L: comparison with the IgG-binding domains of protein G. 193–199. Biochemistry 33: 14011–14017. 12. Graille, M., E. A. Stura, N. G. Housden, J. A. Beckingham, S. P. Bottomley, 39. Murphy, J. P., C. J. Duggleby, M. A. Atkinson, A. R. Trowern, T. Atkinson, and D. Beale, M. J. Taussig, B. J. Sutton, M. G. Gore, and J. B. Charbonnier. 2001. C. R. Goward. 1994. The functional units of peptostreptococcal protein L. Mol. Complex between Peptostreptococcus magnus protein L and a human antibody Microbiol. 12: 911–920. reveals structural convergence in the interaction modes of Fab binding proteins. 40. Rentzsch, G., and J. Wilke. 1970. Measurements of pH values in vitro and in vivo Structure 9: 679–687. in chronic tonsillitis. Z. Laryngol. Rhinol. Otol. 49: 391–397. 13. Kastern, W., U. Sjo¨bring, and L. Bjo¨rck. 1992. Structure of peptostreptococcal 41. Bryant, O. D., A. L. Rashad, J. A. Mazza, and D. Hammond. 1980. ␤-Lactamase protein L and identification of a repeated immunoglobulin light chain-binding activity in human pus. J. Infect. Dis. 142: 594–601. domain. J. Biol. Chem. 267: 12820–12825. 42. Grinstein, S., C. J. Swallow, and O. D. Rotstein. 1991. Regulation of cytoplasmic 14. Patella, V., V. Casolaro, L. Bjo¨rck, and G. Marone. 1990. Protein L: a bacterial pH in phagocytic cell function and dysfunction. Clin. Biochem. 24: 241–247. Ig-binding protein that activates human basophils and mast cells. J. Immunol. 43. O¨ hman, H., and A. Vahlquist. 1994. In vivo studies concerning a pH gradient in 145: 3054–3061. human stratum corneum and upper epidermis. Acta Derm. Venereol. 74: 15. Genovese, A., G. Borgia, L. Bjo¨rck, A. Petraroli, A. de Paulis, M. Piazza, and 375–379. G. Marone. 2003. Immunoglobulin superantigen protein L induces IL-4 and ϩ 44. Robinson, M. J., P. Tessier, R. Poulsom, and N. Hogg. 2002. The S100 family IL-13 secretion from human Fc⑀RI cells through interaction with the ␬ light heterodimer, MRP-8/14, binds with high affinity to heparin and heparan sulfate chains of IgE. J. Immunol. 170: 1854–1861. glycosaminoglycans on endothelial cells. J. Biol. Chem. 277: 3658–3665. 16. Heizmann, C. W., G. Fritz, and B. W. Scha¨fer. 2002. S100 proteins: structure, functions and pathology. Front. Biosci. 7: 1356–1368. 45. Clohessy, P. A., and B. E. Golden. 1996. His-X-X-X-His motif in S100 protein, 17. Odink, K., N. Cerletti, J. Bru¨ggen, R. G. Clerc, L. Tarcsay, G. Zwadio, calprotectin: relation to microbistatic activity. J. Leukocyte Biol. 60: 674. G. Gerhards, R. Schlegel, and C. Sorg. 1987. Two calcium-binding proteins in 46. Sohnle, P. G., C. Collins-Lech, and J. H. Wiessner. 1991. The zinc-reversible infiltrate macrophages of rheumatoid arthritis. Nature 330: 80–82. antimicrobial activity of neutrophil lysates and abscess fluid supernatants. J. In- 18. Roth, J., T. Vogl, C. Sorg, and C. Sunderkotter. 2003. Phagocyte-specific S100 fect. Dis. 164: 137–142. proteins: a novel group of proinflammatory molecules. Trends Immunol. 24: 47. Clohessy, P. A., and B. E. Golden. 1995. Calprotectin-mediated zinc chelation as 155–158. a biostatic mechanism in host defense. Scand. J. Immunol. 42: 551–556. 4592 PROTEIN L BINDS AND INACTIVATES S100 A8/A9

48. Lusitani, D., S. E. Malawista, and R. R. Montgomery. 2003. Calprotectin, an 53. Miyasaki, K. T., A. L. Bodeau, A. R. Murthy, and R. I. Lehrer. 1993. In vitro abundant cytosolic protein from human polymorphonuclear leukocytes, inhibits antimicrobial activity of the human neutrophil cytosolic S-100 protein complex, the growth of Borrelia burgdorferi. Infect. Immun. 71: 4711–4716. calprotectin, against Capnocytophaga sputigena. J. Dent. Res. 72: 517–523. 49. Klempt, M., H. Melkonyan, W. Nacken, D. Wiesmann, U. Holtkemper, and 54. Johne, B., M. K. Fagerhol, T. Lyberg, H. Prydz, P. Brantzaeg, 2ϩ C. Sorg. 1997. The heterodimer of the Ca -binding proteins MRP8 and MRP14 C. F. Naess-Andresen, and I. Dale. 1997. Functional and clinical aspects of the binds to arachidonic acid. FEBS Lett. 408: 81–84. myelomonocyte protein calprotectin. Mol. Pathol. 50: 113–123. 50. Siegenthaler, G., K. Roulin, D. Chatellard-Gruaz, R. Hotz, J. H. Saurat, 55. Kacprzyk, L., V. Rydengård, M. Mo¨rgelin, M. Davoudi, M. Pasupuleti, U. Hellman, and G. Hagens. 1997. A heterocomplex formed by the calcium- M. Malmsten, and A. Schmidtchen. 2007. Antimicrobial activity of histidine-rich binding proteins MRP8 (S100A8) and MRP14 (S100A9) binds unsaturated fatty peptides is dependent on acidic conditions. Biochim. Biophys. Acta 1768: acids with high affinity. J. Biol. Chem. 272: 9371–9377. 2667–2680. 51. Nisapakultorn, K., K. F. Ross, and M. C. Herzberg. 2001. Calprotectin expression inhibits bacterial binding to mucosal epithelial cells. Infect. Immun. 69: 56. Wikstro¨m, M., U. Sjo¨bring, T. Drakenberg, S. Forseń, and L. Bjo¨rck. 1995. 3692–3696. Mapping of the immunoglobulin light chain-binding site of protein L. J. Mol. 52. Champaiboon, C., K. J. Sappington, B. D. Guenther, K. F. Ross, and Biol. 250: 128–133. M. C. Herzberg. 2009. Calprotectin S100A9 calcium-binding loops I and II are 57. Enokizono, J., M. Wikstro¨m,U.Sjo¨bring, L. Bjo¨rck, S. Forseń, Y. Arata, essential for keratinocyte resistance to bacterial invasion. J. Biol. Chem. 284: K. Kato, and I. Shimada. 2000. NMR analysis of the interaction between protein 7078–7090. L and Ig light chains. J. Bacteriol. 182: 1419–1422. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021