FACIN, a Double-Edged Sword of The

Home , Coccus

FACIN, a Double-Edged Sword of the Emerging Periodontal Pathogen Filifactor alocis : A Metabolic Enzyme Moonlighting as a Complement Inhibitor This information is current as of September 29, 2021. Monika Jusko, Beata Miedziak, David Ermert, Michal Magda, Ben C. King, Ewa Bielecka, Kristian Riesbeck, Sigrun Eick, Jan Potempa and Anna M. Blom J Immunol 2016; 197:3245-3259; Prepublished online 16 September 2016; Downloaded from doi: 10.4049/jimmunol.1600739 http://www.jimmunol.org/content/197/8/3245 http://www.jimmunol.org/ References This article cites 55 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/197/8/3245.full#ref-list-1

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

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 © 2016 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

FACIN, a Double-Edged Sword of the Emerging Periodontal Pathogen Filifactor alocis: A Metabolic Enzyme Moonlighting as a Complement Inhibitor

Monika Jusko,* Beata Miedziak,* David Ermert,* Michal Magda,* Ben C. King,* Ewa Bielecka,*,† Kristian Riesbeck,‡ Sigrun Eick,x Jan Potempa,†,{ and Anna M. Blom*

Periodontal disease is one of the most common inflammatory infectious diseases worldwide and it is associated with other syn- dromes, such as cardiovascular disease or rheumatoid arthritis. Recent advances in sequencing allowed for identification of novel periodontopathogens such as Gram-positive Filifactor alocis, but its virulence mechanisms remain largely unknown. We confirmed that F. alocis is a prevalent species in periodontitis patients, and we also observed strong correlation of this bacterium with clinical

parameters, highlighting its role in the pathogenesis of the disease. Further, we found that preincubation of human serum with Downloaded from F. alocis resulted in abolished bactericidal activity and that F. alocis was surviving readily in full blood. We demonstrated that one of the key contributors to F. alocis complement resistance is a unique protein, FACIN (F. alocis complement inhibitor), which binds to C3, resulting in suppression of all complement pathways. Interestingly, FACIN is a nonclassical cell surface protein, a cytosolic enzyme acetylornithine transaminase, for which we now identiﬁed a moonlighting function. FACIN binds to C3 alone, but more importantly it also captures activated complement factor 3 within the complex with factor B, thereby locking in the convertase in an inactive state. Because of the indispensable role of alternative pathway convertase in amplifying complement http://www.jimmunol.org/ cascades, its inhibition by FACIN results in a very potent downregulation of activated complement factor 3 opsonization on the pathogen surface, accompanied by reduction of downstream C5 cleavage. The Journal of Immunology, 2016, 197: 3245–3259.

he oral microbiome consists of .600 prevalent bacterial based methods identified several bacterial clusters within a com- taxa (1), and unlike the other human microbiomes, it plex subgingival bacterial plaque (biofilm) and pointed out the T eventually causes a disease in the majority of individuals importance of three major consensus periodontopathogens: Tan- during their lifetime. The complex interactions of oral bacteria nerella forsythia, Porphyromonas gingivalis, and Treponema with human immune defenses result in chronic inflammation, denticola (8), collectively named as red complex. Subsequently, progressive destruction of periodontal tissues, pocket formation, those three Gram-negative species and their virulence factors have by guest on September 29, 2021 bone resorption, and in severe cases, tooth loss. Gingivitis (gin- been intensively studied, aiming at the identification of patho- gival inflammation without any bone loss and no pockets .3 mm) genesis mechanisms. affects .50% of the adult population, and periodontitis (three or More recent research resulted in a concept that periodontitis is a more teeth with pockets $4 mm) is present in 30% of adults, with result of a dysbiosis in the oral microbiota, leading to the formation ∼8% of severe cases suffering complete loss of dentition (2). of a pathogenic biofilm of an altered composition and increased Furthermore, periodontitis has been proven to be a significant risk bacterial counts, which, in turn, causes complement-dependent factor of systemic diseases, such as coronary heart disease, obe- inflammation of tooth-supporting tissues leading eventually to sity, atherosclerosis, and rheumatoid arthritis (3–7). Periodontitis alveolar bone loss (9, 10). A trigger for such alterations in the oral has traditionally been viewed as an infectious disease; however, biofilm can be provided by low-abundant keystone pathogens, as identification of a single bacterial species etiologic for the pro- shown for P. gingivalis in mouse models of periodontitis (9). gression of the disease has been unsuccessful. Instead, culture- Subsequent periodontal destruction can then be mediated by

*Section of Protein Chemistry, Department of Translational Medicine, Lund University, UMO-2011/01/B/NZ6/00268 and 2012/04/A/NZ1/00051, and the Polish Ministry 205 02 Malmo¨, Sweden; †Department of Microbiology, Faculty of Biochemistry, of Science and Higher Education Grant 2975/7.PR/13/2014/2. B.M. and M.M. are Biophysics and Biotechnology, Jagiellonian University, 30-387 Krakow, Poland; students at the University of Rzeszo´w, Poland. The Faculty of Biochemistry, Biophys- ‡Section of Clinical Microbiology, Department of Translational Medicine, Lund Univer- ics and Biotechnology of Jagiellonian University is a partner of the Leading National x sity, 202 13 Malmo¨, Sweden; Laboratory of Oral Microbiology, Department of Peri- Research Center supported by the Ministry of Science and Higher Education. odontology, University of Bern, 3010 Bern, Switzerland; and {Centre for Oral Health and Address correspondence and reprint requests to Prof. Anna M. Blom, Section of Systemic Diseases, University of Louisville School of Dentistry, Louisville, KY 40202 Protein Chemistry, Department of Translational Medicine, Lund University, Inga ORCIDs: 0000-0003-4600-9070 (D.E.); 0000-0001-6776-8741 (B.C.K.); 0000-0002- Maria Nilsson Street 53, S-205 02 Malmo¨, Sweden. E-mail address: anna. 2631-6711 (E.B.); 0000-0001-6274-6965 (K.R.); 0000-0002-4619-2461 (S.E.); 0000- [email protected] 0002-1348-1734 (A.M.B.). Abbreviations used in this article: AP, alternative pathway; C3b, activated com- Received for publication April 25, 2016. Accepted for publication August 20, 2016. plement factor 3; CP, classical pathway; DAF, decay-accelerating factor; FACIN, F. alocis complement inhibitor; FB, factor B; FD, factor D; GCF, gingival crevicular This work was supported by Swedish Research Council Grant K2012-66X-14928-09- fluid; GluD, glutamate dehydrogenase; GMFI, geometric mean fluorescence inten- 5, Swedish Government Funds for Clinical Research, the Torsten So¨derberg Founda- sity; GVB2+, barbiturate buffer with dextrose and gelatin; LB, Luria–Bertani; LP, tion, the O¨ sterlund Foundation, the Greta and Johan Kock Foundation, the Gustav V lectin pathway; MAC, membrane attack complex; NCBI, National Center for Bio- 80-Year Anniversary Foundation, the Knut and Alice Wallenberg Foundation, and the technology Information; NHS, normal human serum; pAb, polyclonal Ab; PVDF, Inga-Britt and Arne Lundberg Foundation. J.P. is supported by European Commission polyvinylidene difluoride. Grants P7-HEALTH-F3-2012-306029 TRIGGER and FP7-PEOPLE-2011-ITN-290246 RAPID National Institutes of Health, National Institute of Dental and Craniofacial Ó Research Grants DE 09761, DE 022597, and DE023207, Science Center, Poland Grants Copyright 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1600739 3246 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM pathobionts, commensals that begin to thrive under inflammatory lornithine transaminase involved in arginine catabolism. Further- conditions and evoke disease-associated symptoms (11, 12). Fur- more, we describe that in addition to its indispensable metabolic thermore, recent advances in sequencing techniques allowed for function for the bacterium, the protein provides compelling defense identification of novel species within the subgingival dental biofilm, against complement. To underline its novel role in immune evasion, which had previously been unrecognized because of culture diffi- we named it FACIN,thatis,F. a loc is complement inhibitor. culties (13–15). These facts facilitated complex comparative studies of bacterial communities between health and disease, and pointed Materials and Methods out shifts at all taxonomic levels, identifying species correlated with Determination of F. alocis and mRNA expression of disease states (16–18). NAD-specific glutamate dehydrogenase and acetylornithine Filifactor alocis is a Gram-positive bacterium, only recently aminotransferase (FACIN) in clinical samples recognized as a periopathogen. In comparison with the traditional In this analysis, obtained samples and data from 21 patients (15 with an periopathogens, F. alocis is abundant in diseased periodontal untreated severe chronic periodontitis, 6 with gingivitis) attending the pockets, whereas it is hardly detectable in healthy or periodontitis- Centre for Periodontology, Department of Cariology, Endodontology and resistant patients (19, 20). It has been found in patients suffering Periodontology, University Hospital of Leipzig, Leipzig, Germany, were from different forms of the disease, including chronic and gener- included. Samples were obtained from the deepest site per quadrant. The sampling of subgingival plaque was performed with cotton rolls. Each paper alized aggressive periodontitis, as well as endodontic infections (19, point was inserted to the gingival pocket until resistance level. After 30 s, the 21–23). Yet, little is known about pathogenicity of F. al oc is or its paper points were removed, pooled into one tube placed on ice, and ability to persist in the periodontal pocket. One report showed that immediately stored at 220˚C with RNAlater Storage solution (Sigma-

F. al oc is induces secretion of proinflammatory cytokines from Aldrich) until further analysis. DNA and RNA were extracted by using Downloaded from innuPREP DNA/RNA Mini Kit (Analytik Jena) according to the manu- gingival epithelial cells, which may lead to their apoptosis (24). facturer’s instruction. For determination of F. alocis, primers (Fa-forward: Furthermore, in a coculture with P. gingivalis, F. al oc is exhibits 59-CAG GTG GTT TAA CAA GTT AGT GG-39, Fa-reverse: 59-CTA AGT improved ability of biofilm formation and increased adherence and TGT CCT TAG CTG TCT CG-39) were used (31). Real-time PCR using invasion to epithelial cells (25). A proteome analysis of F. al oc is GoTaq qPCR Master Mix (Promega) was performed according to the man- strains identified several potential virulence factors, including pro- ufacturer’s recommendation. The amplification specificity was checked using a mixture of 40 species associated with periodontitis. Furthermore, the teases, adhesion molecules, neutrophil-activating protein A, and specificity of the amplification was always verified with the use of melting http://www.jimmunol.org/ calcium-binding acid repeat protein (26). However, so far there have curves. For quantification, the results from unknown plaque specimens were been scarce reports revealing how F. a loc is resists major compo- projected on the counted pure culture standard curves of F. al oc is ATCC nents of host immunity such as the complement system. 35896. Real-time PCR detection of other periodontopathogens was performed as described previously (32). Complement plays a fundamental role in immunity, and its cDNA was generated from RNA by using GoScript Reverse Transcription subversion by periodontal bacteria is one of the hallmark features of System (Promega AG). Real-time PCR (as described earlier) was applied to periodontitis, leading to exacerbated inflammation and contributing assess expression of F. alocis glutamate dehydrogenase (GluD) and FACIN. The to the dysbiosis of oral plaque (27). In homeostasis, complement primers for GluD (59-GGTGGAGGTAAAGGTGGAGT-39;59-ACAGCTAC- 9 9 plays a fundamental role in immunity. Upon pathogen recognition, TTTAGCGCCTTTG-3 )andforFACIN (5 -CCGTTTCCTAACGCTTTCGC . 39;59-AGTGGAGCTGAAGCAAACGA-39) were designed based on ac- by guest on September 29, 2021 the complement cascade proceeds through sequential activation cession no. NC_016630.1, locus_tag = HMPREF0389_RS03275 and and proteolytic cleavage of a series of serum proteins. Depending locus_tag = HMPREF0389_RS03780. Correlation coefficients of F. a loc is on the molecular trigger, three pathways of complement activation with clinical parameters were determined by using Spearman test with have been distinguished, namely, classical pathway (CP), lectin statistical software (IBM SPSS Statistics 21.0 for Windows; SPSS). pathway (LP), and alternative pathway (AP). All pathways merge Ethics statement at the stage of C3 activation leading to opsonization of the path- The local ethical review board in Lund approved collection of blood and ogen with iC3b, which facilitates phagocytosis. Notably, C3 is the sera from healthy human volunteers. The ethical committee of Bern central complement protein capable of covalently binding to dif- University approved collection of periodontal plaques and GCF. Written ferent surfaces. Furthermore, anaphylatoxins released during ac- informed consent was obtained from patients and volunteers, and the in- tivation of the cascade, C3a and C5a, activate inflammatory cells vestigation was performed according to the principles of the World Medical and attract phagocytes to the site of infection. The end result of the Association Declaration of Helsinki (version 2008). complement cascade is formation of the membrane attack com- Sera and proteins plex (MAC), which lyses Gram-negative bacteria. Normal human serum (NHS) was obtained from eight healthy volunteers. Sera The expression of membrane-bound as well as the recruitment of deficient from various complement components as well as matching NHS soluble complement inhibitors provide the protection for the host were obtained from Quidel. Heat-inactivated sera (DNHS) were prepared by cells when this powerful system is activated. In gingival crevicular incubation at 56˚C for 35 min. Purified complement proteins C3, activated fluid (GCF), filling the pathological periodontal pockets, com- complement factor 3 (C3b), C3d, C3c, factor B (FB), and factor D (FD) were purchased from Complement Technology. C3met (C3b-like molecule) was plement proteins and their activation fragments can be found at 70– derived from C3 by incubation with 0.1 M methylamine hydrochloride pH 8 80% of their concentration in serum (28, 29). Therefore, bacteria for 1 h at 37˚C and subsequent dialysis against 50 mM of Tris-HCl, 150 mM of oral biofilms are in constant contact with this system and must of NaCl, pH 8. BSA was purchased from AppliChem. a-1-Antitrypsin in- use various complement evasion strategies to establish successful hibitor was a kind gift of late Prof. Laurell. infection. F. alocis is found in high abundance in diseased peri- Bacterial strains and their culture odontal sites, indicating increased potential of this bacterium to survive and thrive under proinflammatory conditions. As an F. a loc is type strain ATCC 35896 and a low passage clinical isolate D62-D (kind gift from Prof. Hansel M. Fletcher and Prof. Aruni Wilson), asaccharolytic species, F. alocis is well-equipped with enzymes P. g ing iv al is W50 (33), and Veillonella parvula CCUG 5122 (Culture Col- using specific amino acids, with arginine being a preferred sub- lection, University of Goteborg) were grown on fastidious anaerobic blood strate, supporting the growth of the bacterium (26, 30). agar (FAA) plates at 35˚C in an anaerobic chamber (Concept 400; Biotrace) In this study we identified a novel complement inhibitory protein with an atmosphere of 90% N2,5%CO2, and 5% H2. Moraxella catarrhalis RH4 and a mutant deficient in UspA1 and UspA2 (RH4 DuspA1, DuspA2) of F. al oc is, which binds C3, resulting in suppression of all com- (34) were grown on chocolate agar plates (supplemented with Zeocin plement pathways. Thus, we demonstrated a moonlighting function [phleomycin D1] 7 mg/ml and chloramphenicol 1.5 mg/ml for the double for a nonclassical cell surface protein, a cytosolic enzyme acety- mutant). Staphylococcus aureus 8325-4 was grown on tryptic soy broth at The Journal of Immunology 3247

37˚C with vigorous shaking. Overnight culture of S. aureus was diluted to was evaluated by the same procedure, except for using primary F(ab9)2 OD600 = 0.1 in fresh media and incubated at respective conditions until fragments anti-human C3c during detection, because of problems with they reached exponential growth at OD600 = 0.3–0.4. Escherichia coli unspecific binding of pAbs by these bacteria. The geometric mean fluo- DH5a was cultured in Luria-Bertani (LB) broth until exponential growth rescence intensity (GMFI) was calculated for all the samples using FlowJo phase (OD600 = 0.4–0.5). software (Tree Star). Bactericidal assays Biofilm formation: C3b deposition on biofilm versus planktonic F. alocis ATCC 35896 and P. gingivalis W50 were harvested from plates, cultures of bacteria washed once in PBS, adjusted to OD600 of 1.0, and incubated in anaerobic For biofilm formation, F. alocis strains ATCC 35896 and P. gingivalis W50 m conditions for 30 min with pronase (50 g/ml protease from Streptococcus were cultured in brain-heart infusion medium and incubated overnight at griseus; Sigma-Aldrich), 0.02% NaN3, or PBS alone, as indicated. Heat-killed 35˚C in an anaerobic chamber. The next day both strains were inoculated F. alocis and P. gingivalis were prepared by incubation for 10 min at 72˚C. to fresh medium to OD600 0.15. For planktonic bacteria these cultures were After treatment, bacteria were washed once in barbiturate buffer with dextrose further grown in sterile 50-ml tubes. For biofilm formation, cultures at and gelatin (GVB2+; 5 mM of veronal buffer pH 7.3, 140 mM of NaCl, 0.1% OD600 0.15 were placed in six-well cell culture plates (1 ml per well). Both gelatin, 1 mM of MgCl2,and5mMofCaCl2) and adjusted to the concen- types of cultures were incubated for 24 h at 35˚C in an anaerobic chamber. 3 9 D 2+ tration of 5 10 CFU/ml. NHS and NHS diluted in GVB were incu- After incubation, supernatants from biofilms were aspirated, bacteria were bated for 60 min at 37˚C in anaerobic conditions alone or with differently washed once with PBS, and the biofilm was gently disrupted to single-cell 3 8 treated F. alocis or P. gingivalis (5 10 CFU) at concentrations of 0.75 or suspension by scraping and pipetting, and resuspended in PBS to OD 0.5. m a 600 1.13% in a total volume of 200 l. Bacteria E. coli DH5 were cultured until Planktonic cultures were sedimented by centrifugation, washed once in PBS, exponential growth phase, harvested, washed once in GVB2+, and adjusted to 7 6 and resuspended to OD600 0.5. The protocol for CFSE staining and com- the concentration of 10 or 10 CFU/ml, respectively. After 60-min incubation plement deposition on F. al oc is by flow cytometry was followed as de- of NHS/DNHS with F. alocis or P. gingivalis,100mlofE. coli suspension

scribed earlier. Downloaded from was added to all the samples and further incubated for 30 min at 37˚C. For bactericidal assays with purified recombinant proteins, E. coli DH5a was Surface convertase formation on F. alocis prepared as described earlier and adjusted to the concentration of 106 CFU/ml in GVB2+.TheE. coli suspension (100 ml) was incubated for 30 min at 37˚C F. alocis ATCC 35896 was harvested from plate culture, washed, and m with 0.75% NHS pretreated (30 min) with increasing concentrations of tested resuspended in Binding Buffer to OD600 0.3. Bacteria (50 l) were used proteins in a total volume of 200 ml. After incubation, aliquots were removed, per sample and incubated with purified AP convertase components: C3 diluted serially, and spread onto LB agar plates. Plates were incubated for (6.3 ng), FB (31 ng), and FD (312 ng) added during steps 1 or 2. In step 1, bacteria were incubated with or without C3 for 1 h at 37˚C with gentle 12hat37˚CafterwhichE. coli CFU were counted and the numbers of http://www.jimmunol.org/ 3 surviving bacteria were calculated. No F. alocis CFUwereobservedonagar shaking, harvested by centrifugation (10 min, 15,000 g), and washed m plates under these conditions. twice. In step 2, bacteria were resuspended in Binding Buffer (50 l) containing either individual AP convertase factors or their combinations, Whole blood killing assay and incubated for another 1 h at 37˚C. As a control, convertase components were combined in a sample without bacteria and allowed to interact for 1 h F. alocis strains ATCC 35896 and D62-D, P. gingivalis W50, V. parvula at 37˚C. After incubation, bacteria were pelleted, washed as described CCUG5122, and S. aureus 8325-4 were washed in PBS buffer and adjusted to m 7 earlier, suspended in reducing sample buffer (50 l), boiled, and separated the approximate concentration of 5 3 10 CFU/ml. Bacteria (40 ml) were m m m on two identical 10% Laemmli gels (5- l sample per well). Proteins were mixed with 360 l of freshly collected human blood treated with 50 g/ml transferred onto polyvinylidene difluoride (PVDF) membranes. On one Refludan (lepirudin; Pharmion) as anticoagulant, and incubated anaerobically membrane FB were detected with pAbs (Quidel), whereas C3 was detected for 30 and 60 min. After incubation, aliquots were removed, diluted serially, on the other (pAbs; Sigma-Aldrich). by guest on September 29, 2021 and plated on appropriate agar media. The survival for each species was calculated as a percentage of recovered CFU compared with the inoculum. Isolation of cell wall proteins of F. alocis Complement deposition of F. alocis by flow cytometry To isolate cell wall proteins, we harvested F. a loc is ATCC 35896 from plate cultures, and washed and resuspended them in a buffer (50 of mM Tris-HCl, F. alocis ATCC 35896 from 7-d-old agar plate culture was harvested, 1 mM of EDTA, pH 8) to OD600 .1.5. The bacterial suspension was washed once in PBS, adjusted to OD600 of 0.5, and incubated for 45 min centrifuged at 7560 3 g for 20 min at 4˚C. The pelleted bacteria were in- with 20 mM of CFSE. After prestaining with CFSE, bacteria were washed 2+ cubated on ice for 5 min. Thereafter, bacteria were resuspended in a buffer once in GVB (CP/LP) or Mg-EGTA (2.5 mM veronal buffer [pH 7.3] (50 of mM Tris-HCl, 1 mM of EDTA, pH 8) with 20% sucrose containing containing 70 mM of NaCl, 140 mM of glucose, 0.1% gelatin, 7 mM of 1 mM of PMSF (Sigma-Aldrich) and centrifuged as described earlier. The MgCl2, and 10 mM of EGTA; AP) and adjusted to OD600 of 0.3. There- m D bacterial pellet was dissolved in an ice-cold solution of mutanolysin after, 140 l of bacteria was mixed with 1–30% NHS or NHS diluted in (5000 U/ml in 0.1 M K HPO , pH 6.2; Sigma-Aldrich) and incubated for 2 h GVB2+ or Mg-EGTA and incubated for 1 h at 37˚C in a total volume of 2 4 m at 37˚C with shaking. Subsequently, cell debris was spun down by centri- 200 l. Similar experiments were performed using commercial NHS and fugation at 14,000 3 g for 5 min, and the supernatant containing cell wall– deficient sera (Quidel). F. alocis ATCC 35896, treated as described earlier, associated proteins was aliquoted and frozen at 280˚C. Concentrations of was incubated with 15% C1q-, C2-, FB-deficient sera or NHS (all sera in proteins were determined using bicinchoninic acid assay (Pierce). active and heat-inactivated form) in GVB2+ or Mg-EGTA. Then bacteria were washed once in FACS buffer (50 mM of HEPES, 100 mM of NaCl, [125I]-C3b binding to the cell wall–associated cell proteins of pH 7.4, 1% BSA, 30 mM of NaN3) and incubated with specific polyclonal F. alocis Abs (pAbs) against C3c (Dako) for 30 min at room temperature, followed by 30-min incubation with secondary F(ab9)2 fragments conjugated with Samples of cell surface proteins of F. alocis ATCC 35896 were separated Alexa Fluor 647 (Life Technologies). Using the same protocol, we tested by 2D-PAGE on two identical gels. In brief, cell surface proteins were C3b deposition on FITC-labeled zymosan (Molecular Probes) from se- precleaned from 2D-PAGE interfering substances using 2D-Clean-Up Kit lected sera. To evaluate binding of different C3 protein fragments by (GE Healthcare, Little Chalfont, U.K.) and suspended in rehydration buffer F. alocis and M. catarrhalis, we used a similar protocol, whereby bacteria (7 M urea, 2 M thiourea, 4% CHAPS, 0.002% bromophenol blue with 1% were incubated with purified C3, C3b, C3met, C3d, and C3c proteins in DeStreak [GE Healthcare] and 0.5% IPG buffer [GE Healthcare]). The Binding Buffer (10 mM of Hepes, pH 7.2, 1 mM of MgCl2, 2 mM of first-dimension Immobiline DryStrips, pH 3–11 (GE Healthcare), were run CaCl2) supplemented with 0.5% BSA, and detection was performed with by adding 35 mg of protein in the rehydration buffer for 3 h at 300 V, 5 h at F(ab9)2 fragments of the rabbit anti-C3d Ab (Dako) followed by the goat linear gradient of 300–3500 V, and 11 h at 3500 V. After equilibration in anti-rabbit Alexa 647–labeled secondary Abs (Invitrogen). The sensitivity buffer containing 10 mg/ml DTT and 25 mg/ml iodoacetamide, the Dry- of the interaction to ionic strength was tested by the incubation of bacteria Strips were loaded on Criterion 10% Tris-HCl precast gels (BioRad) and with C3met in the presence of increasing concentrations of NaCl starting subjected to electrophoresis at 200 V for 4–5 h. Subsequently, proteins from from physiological amount (0.15–2.5 M NaCl). For testing the effect of one gel were transferred on PVDF membrane using Transblot Turbo system FACIN on deposition of C3b, NHS (5%) was supplemented with 3–100 (Bio-Rad). The membrane was blocked with Quenching Solution (50 mM of mg/ml purified recombinant protein, and the incubation with bacteria was Tris-HCl, 150 mM of NaCl, 0.1% Tween 20, pH 8 containing 3% fish gelatin carried out for 45 min. Finally, bacteria were washed twice, resuspended in [Nordic]) and incubated with [125I]-labeled C3b (labeled with Iodobeads 125 fixing buffer (BD) diluted 1:10 in H2O, and analyzed using flow cytometry [Pierce]) overnight. [ I]-C3b binding to cell wall–associated proteins of (CyFlow space; Partec, Germany). For M. catarrhalis, the C3b deposition F. al oc is was visualized using the PhosphorImager Typhoon FLA 9500 3248 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM

(GE Healthcare Life Sciences). The other gel was stained with GelCode Blue were washed once in PBS and incubated for 30 min at room temperature Stain Reagent (Thermo Scientific) and destained in distilled water. After with rabbit a-FACIN (prepared by Agrisera using recombinant FACIN identifying candidate proteins responsible for C3b binding on the membrane expressed in E. coli) and preimmune serum, respectively. After washing, using autoradiography, corresponding spots were cut out from GelCode bacteria were incubated with Alexa Fluor 647 F(ab9)2 fragments of goat Blue-stained gel and subjected to high performance liquid chromatography- anti-rabbit IgG (Invitrogen) for 30 min at room temperature. Before flow tandem mass spectrometry analysis at the Swegene Center for Integrative cytometry analysis, bacteria were washed again in PBS and amounts of Biology at Lund University core facility (Lund University). surface bound FACIN were acquired. Recombinant protein expression Hemolytic assays Genomic DNA was isolated from F. alocis strain ATCC 35896 using GenElute To assess activity of the CP, we washed sheep erythrocytes (Ha˚tunalab) Bacterial Genomic DNA Kit (Sigma-Aldrich). The two genes are available three times with DGVB2+ buffer (2.5 mM of veronal buffer pH 7.3, 70 mM online (http://www.ncbi.nlm.nih.gov/gene/) under National Center for Bio- of NaCl, 140 mM of glucose, 0.1% gelatin, 1 mM of MgCl2, and 5 mM of technology Information (NCBI) accession numbers HMPREF0389_01649 CaCl2). The cells were incubated with a complement-fixing Ab (Ambo- (encoding NAD-specific GluD) and HMPREF0389_01570 (encoding acety- ceptor; Boehringwerke; diluted 1:3000 in DGVB2+ buffer) at a concen- lornithine transaminase, FACIN) were amplified from genomic DNA by PCR. tration of 109 cells/ml for 20 min at 37˚C. After two washes with DGVB2+, PCR products were purified and cloned into the pET-28a expression vector 5 3 108 cells/ml was incubated for 1 h at 37˚C with 0.15% NHS diluted in (Novagen) using NcoI/XhoI (GluD) or XbaI/XhoI (FACIN) restriction sites. DGVB2+ buffer (total volume, 150 ml). Before incubation with erythro- The obtained recombinant constructs encoding F. alocis proteins with His-tags cytes, NHS was preincubated with various concentrations of recombinant on C-termini were transformed into E. coli strain BL21 (DE3) (New England proteins for 15 min at 37˚C. The samples were centrifuged and the amount Biolabs) under the control of the T7 promoter. Transformed E. coli were grown of lysed erythrocytes was determined by spectrophotometric measurement in LB broth at 37˚C to OD600 of 0.6–0.7, cooled to 30˚C, and expression of of released hemoglobin (405 nm; Varian 50 MPR Microplate Reader). To recombinant proteins was induced by the addition of 1 mM of isopropyl-1-thio- assess activity of the AP, we washed rabbit erythrocytes three times with 8 Downloaded from b-D-galactopyranoside. For expression of GluD the culture was continued for Mg-EGTA and then used them at a concentration of 5 3 10 cells/ml for 16 h at 30˚C. For expression of FACIN, bacteria were cultured for 4 h at 37˚C. 1-h incubation at 37˚C with 1.5% NHS diluted in Mg-EGTA buffer (total volume, 150 ml). The serum used was pretreated with recombinant proteins GluD purification for 15 min at 37˚C. The samples were centrifuged and the amount of lysed erythrocytes was determined spectrophotometrically. Cells were harvested by centrifugation (15 min, 6000 3 g, 4˚C), resuspended in PBS, and subsequently lysed by sonication (cycle of 60 3 1s Complement activation assays pulse, 2 s break at amplitude of 70%) using sonicator (Branson Sonifier Digital 450; Branson Ultrasonics, Danbury, CT). The cell lysate was Microtiter plates (MaxiSorp; Nunc) were incubated overnight at 4˚C with http://www.jimmunol.org/ clarified by centrifugation (30 min, 15,000 3 g, 4˚C), filtered through a 50 ml of a solution containing 2 mg/ml human aggregated IgG (Immuno), 0.45-mm syringe filter, and applied at room temperature onto a 1-ml col- 100 mg/ml mannan (M-7504; Sigma-Aldrich), 20 mg/ml zymosan (Z-4250; umn with Ni-NTA resin equilibrated with 50 mM of sodium phosphate Sigma-Aldrich) in 75 mM of sodium carbonate (pH 9.6). Between each step buffer, pH 7.5 with 0.5 M NaCl and 5 mM of imidazole. Recombinant of the procedure, the plates were washed four times with 50 mM of Tris-HCl, GluD-6xHis was eluted using increasing concentrations of imidazole 150 mM of NaCl, and 0.1% Tween 20 (pH 7.5). The wells were blocked with 10–500 mM in 50 mM of sodium phosphate buffer, pH 7.5 with 0.5 M Quenching Solution for 2 h at room temperature. NHS was diluted in GVB2+ NaCl. Fractions containing eluted protein were pooled and dialyzed to buffer (CP and LP) or Mg-EGTA (AP) and used at a concentration of 1% in 50 of mM Tris-Cl, pH 7.6, 150 mM of NaCl. CP (C3b), 2% in LP (C3b), 4% in AP (C3b), or 2% in CP (MAC). NHS was mixed with various concentrations of recombinant proteins, preincubated for FACIN purification

25 min at 37˚C (CP and LP) or 15 min on ice (AP), and incubated in the by guest on September 29, 2021 Cells were harvested by centrifugation (15 min, 3000 3 g, 4˚C) and resus- wells of microtiter plates for 35 (CP, LP) and 45 min (AP) at 37˚C for C3b pended in washing buffer (50 mM of Tris-HCl pH 8, 300 mM of NaCl) and and MAC. Complement activation was assessed by detecting deposited subsequently disrupted by sonication (cycle of 4 3 50 s pulse, 30 s break at complement factors using specific pAbs against C3d (C3b detection; Dako) or C5 (MAC detection on aggregated IgG-coated plate; Quidel pAbs) diluted amplitude of 70%) using sonication. The insoluble fraction (inclusion bodies) in Quenching Solution. Bound pAbs were detected with HRP-conjugated anti- was collected by centrifugation (40 min, 27,000 3 g,4˚C)andsuspendedin washing buffer with 1 M urea. The sonication and centrifugation step was rabbit, anti-goat, or anti-mouse secondary pAbs. Bound HRP-conjugated repeated three times. Finally, the protein pellet was suspended in 10 ml of pAbs were detected with 1,2-phenylenediamine dihydrochloride tablets with washing buffer with 1 M urea. Thereafter, 8 M urea in washing buffer was absorbance measured at OD490. added drop-wise to the protein suspension with gentle orbital shaking until a FB activation in NHS activated with zymosan transparent and homogenous solution was obtained (final volume, 100 ml). Proteins were dialyzed against washing buffer with 10% glycerol and 0.1 M NHS (5%) in Mg-EGTA was preincubated with 100 mg/ml recombinant EDTA at 4˚C in two steps: at 1:10 ratio for 4 h followed by 16 h at 1:100 protein for 15 min on ice (10-ml sample). Zymosan (12 mg; 10 ml) was ratio. Protein suspension was cleared from insoluble debris by centrifugation added and incubated with NHS at 37˚C for different time points to allow (40 min, 27000 3 g, 4˚C). The supernatant was divided in 50-ml portions and complement activation. The reaction was stopped by adding reducing sample incubated with 4 ml of Ni-NTA superflow resin (per 50 ml; Qiagen) for 2 h at buffer (10 ml). Samples were boiled, centrifuged, and the supernatants were 4˚C. Ni-NTA beads were sedimented by centrifugation (5 min, 5000 3 g, separated on 10% Laemmli gel (25 ml sample per well). Proteins were 4˚C), and bound protein was eluted with 50 mM of imidazole in washing transferred onto PVDF membranes and FB was detected with pAbs (Quidel). buffer in 5-ml fractions. Fractions containing eluted protein were pooled and dialyzed in 50 mM of Tris-Cl, pH 7.6, 150 mM of NaCl. FB detection on F. alocis surface after incubation in NHS C3-met binding F. alocis ATCC 35896 was harvested from 5-d-old FAA plate culture, washed, and resuspended in Mg-EGTA at a OD600 0.3. Recombinant proteins Proteins diluted in PBS (10 mg/ml) were coated onto microtiter plates (200 mg/ml) were preincubated with bacteria (50 ml) for 30 min at 37˚C in a (Maxisorp; Nunc) overnight at 4˚C. The plates were washed four times total volume of 75 ml. NHS was added to the samples and incubated for 25 with Immunowash (50 mM of Tris-HCl, 150 mM of NaCl, and 0.1% min at 37˚C, after which samples were centrifuged for 5 min at 14,000 3 g. Tween 20 [pH 7.5]). The wells were blocked with Quenching Solution for Bacterial pellets were washed twice in PBS with 0.05% Tween 20, resus- 2 h at room temperature. C3-met dilutions were prepared in Binding pended in reducing sample buffer (50 ml/sample), boiled, centrifuged to Buffer, added to the plate (50 ml per well), and incubated for 1 h at 37˚C. remove insoluble particles, followed by separation of the supernatant on a The plates were washed as described earlier and C3-met binding was 10% Laemmli gel (20-ml sample per well). Proteins were transferred onto detected using specific Abs against C3c (pAbs; Dako, Glostrup, Denmark). PVDF membranes, and FB was detected with pAbs (Quidel). Bound Abs were detected with HRP-conjugated anti-rabbit pAbs followed by detection with 1,2-phenylenediamine dihydrochloride tablets (Dako) Phagocytosis of zymosan by HL-60 cells and measurement at OD490. HL-60 cells (DSMZ, Braunschweig, Germany) were precultured in RPMI Flow cytometry of surface staining for FACIN on F. alocis 1640 with 10% FCS, 25 mM of HEPES for 1 d, then supplemented with 1.25% DMSO to induce differentiation, and cultured at these conditions for Harvested and CFSE-stained F. alocis were adjusted to OD600 = 0.2 in PBS 4 d. At day 4, cells were harvested and upregulation of a marker of myeloid and incubated with 50 mg of recombinant FACIN for 1 h at 37˚C. Bacteria differentiation, CD11b, was confirmed by FACS. Cells were thereafter The Journal of Immunology 3249 seeded to a 48-well plate at 1 3 105 cells per well. Serial dilutions of were killed significantly faster. The commensal V. parvula was also 2+ recombinant FACIN in GVB were preincubated with NHS (50%) for eradicated after that time. 15 min at 37˚C and subsequently added to samples containing zymosan A(S. cerevisiae) beads conjugated with fluorescein (4 3 105 beads in 2+ F. alocis is recognized by all pathways of the complement GVB per sample). After 30 min of serum opsonization (final NHS system and appears to proactively bind C3 concentration, 8%) at 37˚C, zymosan beads were harvested, washed once and resuspended in RPMI 1640, and added to the HL-60 cells (∼2 beads F. al oc is was incubated with NHS, a source of human complement per cell). Cells were allowed to take up beads for 1 h at 37˚C, after which diluted in two different buffers: GVB2+ and Mg-EGTA. GVB2+ time they were harvested, washed in FACS buffer, and subjected to flow cytometry analysis. The percentage of FITC+ cells (indicating uptake of allows activation of CP and LP (already at low NHS concentrations) zymosan beads) was calculated for all the samples using FlowJo software or all complement pathways (at higher NHS concentrations). (Tree Star). Mg-EGTA buffer allows only AP activation. Then bacteria were 2+ Surface plasmon resonance analysis of interaction between washed and deposited C3b was detected. In NHS-GVB , C3b was FACIN and C3b deposited on F. al oc is cells in a dose-dependent manner, indicating activation of the CP/LP and AP (Fig. 2A). There was also a weak The affinity between FACIN and C3b was analyzed by surface plasmon fluorescence signal in DNHS-GVB2+, exceeding that of Ab control resonance using a Biacore 2000 (GE Healthcare). C3b (CompTech) was (sample in which serum was omitted), but this was not statistically immobilized on CM5 sensor chip (GE Healthcare) using standard amino coupling reaction to reach 2500 reference units. FACIN diluted to 30– significant (Fig. 2A). Surprisingly, in Mg-EGTA, we observed a 1000 nM concentrations in running buffer (50 mM of HEPES, pH 7.8 dose-dependent increase in fluorescence associated with bacteria in containing 100 mM of NaCl and 0.005% Tween 20) was injected at flow of both NHS and DNHS (Fig. 2B). The differences in GMFI between m 30 l/s for 100 s, and the dissociation was then followed for 300 s. The NHS and DNHS were not statistically significant, except for the Downloaded from signal from the control surface was subtracted. In all experiments, two consecutive injections of 2 M of NaCl, 100 mM of HCl were used to remove highest concentration (30%). We concluded that F. al oc is readily bound ligands during a regeneration step. The BIAevaluation 3.0 software absorbs C3 from human serum in which complement is inactivated, (Biacore) was used to determine affinity constants using the Langmuir 1:1 whereas the real level of opsonization via the AP is reflected mainly interaction model. The experiment was performed two independent times. by the difference between NHS and DNHS at 30%. Statistical analysis To confirm these results and to determine the role of particular

complement pathways in the activation of complement by F. alocis, http://www.jimmunol.org/ One- or two-way ANOVA (GraphPad Prism) was used to calculate the human sera deficient in C1q, C2, or FB were used to determine p values to estimate whether the observed differences between experi- 2+ mental results were statistically significant. deposition of C3b. We found that deposition of C3b in GVB was diminished in serum deficient in C1q (initiator of CP) in comparison with NHS (Fig. 2C) and decreased in serum deficient in C2 Results (a common component of CP and LP), indicating that CP and LP are F. alocis avoids complement, as well as phagocytosis, and activated by F. alocis (Fig. 2C). Depletion of FB also resulted in produces several virulence factors that affect bactericidal largely reduced C3b opsonization, indicating that a significant activity of human serum amount of deposited C3b was generated by the amplification via the To date, no reports have investigated the ability of F. a loc is to resist AP. Furthermore, depletion of FB from serum in Mg-EGTA reduced by guest on September 29, 2021 complement attack. To assess the effect of F. al oc is on bactericidal deposition of C3b compared with NHS, confirming the role of AP in activity of human serum, we preincubated it with NHS and sub- opsonization of F. alocis, whereas C1q depletion as expected had no sequently added serum-sensitive Gram-negative E. coli DH5a to effect (Fig. 2C). As a control, we analyzed C3b deposition on zy- test its survival. Incubation of NHS with F. a loc is resulted in mosan particles (Fig. 2D). As expected, we found significant abolished bactericidal activity of serum toward E. coli DH5a amounts of C3b deposited from NHS and C1q-depleted NHS, but not (Fig. 1A). Only intact, alive F. a loc is was able to counteract the from FB-depleted sera, because opsonization of zymosan, in contrast bactericidal activity of NHS, whereas whole F. al oc is killed by heat with bacteria, occurs predominantly via the AP. Once again, in all or azide (Fig. 1A) did not have any effect, implicating involvement heat-inactivated deficient sera in both buffers, there was a significant of active processes, such as protein export systems. P. gi ng iv al is is level of fluorescence signal compared with Ab controls, in some known to disarm complement using surface-bound as well as se- cases even exceeding that of corresponding active serum, implicating creted proteases such as gingipains (35). The effects of P. gingivalis direct absorption of C3 to the bacterial surface, but not so much to and F. al oc is on E. coli survival in NHS were similar (Fig. 1B). In zymosan, in a process unrelated to traditional complement activation. contrast with P. g ing iv al is, the effect of F. alocis on the survival of Because periodontal bacteria normally do not grow in fluid phase E. coli could be partially blocked through pronase treatment or heat but rather as complex sessile communities, we sought to determine inactivation. Our data proved that the diminished bactericidal ac- how biofilm formation may affect complement opsonization of tivity of NHS was not due to simple depletion of complement F. alocis. To exclude a possible bias because of differences in caused by its binding to the surface of F. al oc is . Furthermore, we complement activation by larger bacterial aggregates in compar- demonstrated that bacterial surface proteins contribute to the effect ison with single cells, we disrupted bacterial biofilms before because their digestion with pronase significantly reduced the effect opsonization with human serum. Strikingly, C3b opsonization of of F. alocis on NHS (Fig. 1A). F. alocis grown as a biofilm was markedly reduced compared with Because F. alocis appears to exploit multiple strategies of com- the planktonic form of the culture (Fig. 2E). These results indi- plement inhibition, we tested how this translates into its survival in cated that biofilm formation upregulates mechanisms of comple- whole human blood, in which complement-opsonized bacteria are ment resistance in F. alocis. phagocytized. To this end, we compared survival of F. alocis with two As we observed direct C3 binding by F. alocis, we decided to other oral bacteria, that is, P. gingivalis, a very potent periopathogen, compare it with M. catarrhalis, which has been shown to bind and and V. parvula, a commensal species being part of a normal oral flora. neutralize C3 on its surface via its surface proteins UspA1 and In addition, we included the Gram-positive bacterium S. aureus,one UspA2, and efficiently limit complement activation (36, 37). of leading pathogens causing bacteremia. We found that both strains When we incubated M. catarrhalis strain RH4 and its mutant of F. alocis survived well in the blood, showing .50% survival rate devoid of UspA1 and UspA2 in different concentrations of NHS after 1-h incubation (Fig. 1C). In contrast, P. gingivalis and S. aureus and DNHS in Mg-EGTA (in the same conditions as for F. alocis), 3250 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM

we observed dose-dependent, complement-mediated C3b deposition in NHS for both strains, and signiﬁcant C3 binding from DNHS only in case of wild type RH4 strain (Fig. 2F), conﬁrming previously published data. The pattern of C3 binding to M. catarrhalis RH4 was similar to those observed for F. alocis.

F. alocis binds directly C3 and blocks AP convertase formation Next, we sought to investigate which part of the C3 molecule mediates the interaction with the F. alocis surface. C3 is composed of two chains designated a and b, and during complement activation it is sequentially cleaved into various fragments (Fig. 3A). Puriﬁed C3, C3b, C3c, C3d, as well as C3-met (molecule conformationally re- sembling C3b but retaining C3a) were incubated with F. alocis for 1 h. We observed pronounced binding to the bacterium of intact C3, as well as all its derived fragments (Fig. 3B). M. catarrhalis,usedas a positive control, bound as expected large amounts of C3, C3met, and C3b, despite using only a fourth of the recombinant proteins compared with F. alocis. These results indicate that there is more

than one binding site for F. alocis on the C3 molecule, and that Downloaded from strong binding occurs via both C3d and C3c parts of the molecule. Furthermore, we found that C3 interaction with F. a l oci s was mostly mediated by ionic interactions, as the binding was gradually decreased at increasing ionic strength (Fig. 3C). Next, we investigated the inﬂuence of C3 binding by F. alocis on

the ability of the molecule to promote complement activation. http://www.jimmunol.org/ When C3 was preincubated with the bacterium for 1 h, it lost the capacity to interact with FB and FD, the other two components of the AP C3 convertase, as indicated by the lack of FB conversion to Bb (Fig. 3D, lane 7) and inhibition of C3 a-chain cleavage (Fig. 3E, lane 7). In contrast, when all components of the convertase were added to bacteria at the same time, C3 interacted with FB and FD, as shown by appearance of Bb (Fig. 3D, lane 11) and C3 a-chain cleavage (Fig. 3E, lane 11). by guest on September 29, 2021 Cell surface attached proteins of F. alocis are responsible for C3 binding We then attempted to identify a protein of F. alocis responsible for interaction with C3. Surface proteins of Gram-positive bacteria are attached to the cell wall envelope, so we isolated this fraction from F. alocis and separated all proteins using two-dimensional electrophoresis. The proteins were transferred to a PVDF membrane and incubated with [125I]-C3b. We observed most pronounced binding of radiolabeled C3b to five distinct bands of approximate sizes 40–45 kDa (Fig. 4A, 4B). We identified these proteins using mass spec- FIGURE 1. F. alocis produces several virulence factors that affect trometry. The high performance liquid chromatography-tandem mass bactericidal activity of human serum and resists phagocytosis in blood. (A) spectrometry analysis resulted in the identification of several peptides B F. alocis ATCC 35896 or ( ) P. gingivalis were incubated for 30 min in in each protein, which after BLAST search were matched with anaerobic conditions with PBS (untreated), pronase, or 0.02% NaN3. Heat- F. alocis proteins with high scores (Table I). We preselected the two killed F. alocis and P. gingivalis were prepared by incubation for 10 min at most abundant proteins with highest matches: NAD-specific GluD 72˚C. After treatment, F. alocis or P. gingivalis (5 3 108 CFU) were washed once in GVB2+ and incubated with 0.5% (A)or1%(B) NHS or and acetylornithine aminotransferase (FACIN). To characterize GluD DNHS for 60 min at 37˚C in anaerobic conditions. Next, 100 mlofE. coli and FACIN in more detail, we recombinantly expressed them with [in 106 (A)or105 CFU (B)] was added to all the samples and incubated His-tag in E. coli. The purity of proteins was evaluated by SDS- further for 30 min at 37˚C. Finally, aliquots were removed, diluted serially, PAGE and Coomassie staining (Fig. 4C). Resulting proteins were spread onto LB agar plates, and incubated at 37˚C. E. coli colonies were coated on plates, and their interaction with soluble C3-met was an- counted and the numbers of surviving bacteria were calculated. No colo- alyzed. We found that both GluD and FACIN showed significant C3- nies of F. alocis were formed on LB under these aerobic conditions. The met binding, comparable with that of UspA2, a known C3-interacting C dashed line divides samples that were tested in separate experiments. ( ) protein from M. catarrhalis (Fig. 4D). Further analysis of this in- Bacteria were incubated for 30 and 60 min at 37˚C with freshly collected teraction using plasmon surface resonance techniques revealed that human blood. After incubation, aliquots were removed, serially diluted, 2 FACIN strongly interacts with C3b (K =8.93 10 8 M; Fig. 4E). and plated on appropriate agar plates. Survival was calculated as a per- D centage of survival compared with the inoculum. (A and B) Statistical We sought to verify the previously suggested surface localization significance of observed differences was estimated compared with un- of FACIN on F. alocis; therefore, we analyzed protein surface treated F. alocis or P. gingivalis, respectively, and (C) between commensal exposition using flow cytometry (Fig. 4F). Flow cytometry anal- V. parvula and other species. Significance was estimated using two-way ysis using specific Ab revealed that FACIN is present on F. alocis. ANOVA and Bonferroni posttest: ***p , 0.001, **p , 0.01, *p , 0.05. Addition of recombinant FACIN to the bacteria led to a trend The Journal of Immunology 3251 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 2. F. alocis is recognized by all complement pathways leading to C3b deposition and also binds directly C3. (A–C) F. alocis was incubated for 1 h with increasing concentrations of NHS (%) or heat-inactivated NHS (ΔNHS, %) diluted in GVB2+ (A)orMg-EGTA(B). Similarly, (C) F. alocis or (D) zymosan particles were incubated with NHS, C1q-, C2-, FB-deficient serum and ΔNHS diluted in GVB2+,orwithNHS,ΔNHS and FB-, and C1q-deficient serum diluted in Mg-EGTA. (E) F. a loc i s grown in a planktonic or a biofilm culture was incubated for 1 h with NHS and ΔNHS in GVB2+.(F) M. catarrhalis RH4 and its isogenic mutant lacking UspA1/A2 were incubated with increasing concentrations of NHS (%) and ΔNHS (%) in Mg-EGTA. The dashed line separates results obtained for the two strains. (A–F) Deposited C3b/bound C3 was detected on bacterial surface with specific pAbs using flow cytometry. Deposition/binding of C3b/C3 is shown as GMFI. (E) Data were normalized to the highest GMFI obtained for the sample containing planktonic bacteria in 15% NHS. Means of three independent experiments are presented with bars indicating SD. Either standard (C–E) or matched (A, B,andF) two-way ANOVA (yielding in all cases p , 0.0001) and Tukey or Bonferroni posttest were used to calculate significance: ****p , 0.0001, ***p , 0.001, **p , 0.01, *p , 0.05. (A, B,andE) The statistical significance of differences is shown compared with Ctrl (control: bacteria without serum, incubated with detection Abs), (C and D) compared with NHS, and (E) between two culture forms. We observed deposition of C3b via both CP and AP with higher degree of deposition on the planktonic form compared with bacteria grown in biofilm. Furthermore, C3 was also absorbed from ΔNHS in a complement-independent manner compared with what was previously observed for M. catarrhalis.

toward an increased signal (adjusted p = 0.06), indicating that gingivitis patients (n = 6). The majority of the analyzed peri- F. alocis may also bind soluble FACIN to its surface. odontitis samples contained high bacterial counts ($105 per sampling site). Counts of F. alocis correlated strongly with those F. alocis is prevalent in subgingival plaque of chronic of P. gingivalis (r = 0.672, p = 0.001) and of T. forsythia (r = 0.631, periodontitis and gingivitis patients, and expresses both GluD p = 0.002). Importantly, F. al oc is showed strong correlation with and FACIN in vivo bleeding on probing, indicative of inﬂammation: r = 0.470, p = We found that F. alocis was present at very high prevalence rates 0.0032. There was no correlation of F. a loc is presence with plaque of 100% in both untreated chronic periodontitis (n =15)and index (indicative of oral hygiene). 3252 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 3. F. alocis binds directly C3 via ionic interactions and prevents C3 convertase formation. (A) Processing of C3 during complement activation. C3 convertases cleave C3 (composed of a-andb-chain) to C3b and the anaphylatoxin C3a. The thioester of C3b interacts covalently with molecules in its vicinity such as microbial surfaces. C3b is then inactivated by factor I in combination with various cofactors in several steps. First, the small fragment C3f is released from the middle part of a-chain, resulting in the formation of inactive iC3b, the major opsonin (not shown). Further cleavages liberate the large C3c fragment and leave the smaller C3dg fragment attached to microbial surface. Subsequently, C3dg fragment is truncated to yield stable C3d. C3 treated with methylamine (C3met) corresponds conformationally and functionally to C3b but still has C3a attached. (B) F. alocis and M. catarrhalis were incubated for 1 h with intact C3, C3-met, or various C3-fragments (100 or 25 mg/ml, respectively). (C) F. alocis was incubated for 1 h with C3-met in the presence of increasing concentrations of NaCl. (B and C) Bound proteins were detected on bacterial surface with specific Abs anti-C3c or anti-C3d using flow cytometry. We observed that F. alocis binds intact C3, as well as its fragments C3b/C3met, C3c, and C3d (B), and that this interaction is sensitive to ionic strength (C). Means of three independent experiments are presented with bars indicating SD. Statistical significance of observed differences compared with (B) corresponding Ctrl (control 2 bacteria without any protein, incubated with detection Abs) or (C) physiological NaCl (0.15 M) was estimated using one-way ANOVA (yielding p , 0.0001) and a Dunnett posttest. ****p , 0.0001, ***p , 0.001, **p , 0.01, *p , 0.05. (D and E) F. alocis was incubated for 1 h with C3 to allow binding of the protein to the bacterium (step 1), washed, and subsequently FB and FD were added to the samples, which were incubated further for another hour (step 2). As a control, all components of the APs (C3, FB, and FD) were added to the bacteria at the same time (step 2). Next, bacteria were washed and suspended in reducing sample buffer. Proteins were separatedby SDS-PAGE and blotted onto PVDF membrane. (D)ConversionofFBtoBb(D) and cleavage of a-chain of C3 (E), both indicative of active AP convertase formation on bacterial surface, were detected with specific pAbs. When C3 was added to the bacteria simultaneously with FB and FD, we observed its conversion to C3b, as well as activation of FB. However, C3 preincubated first with bacteria lost this ability.

To conﬁrm the expression of these two proteins in vivo, contrast, GCF samples rich in F. alocis demonstrated very we tested GCF samples with low and high F. alocis counts. The high amounts of DNA and mRNA for both of the analyzed samples with lower bacterial count showed little detect- proteins (Table II), suggesting expression of these two genes able DNA and no mRNA for GluD and FACIN (Table II). In in vivo. The Journal of Immunology 3253 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 4. Several of cell surface–attached proteins of F. al oc is bind C3. Surface proteins of F. alocis were isolated and separated using two-dimensional (2D) electrophoresis on two identical gels. Subsequently, the proteins were transferred on PVDF membrane and incubated with [125I]-C3b (A). Several distinct protein spots showing strongest binding of [125I]-C3b were cut out from corresponding GelCode-stained gel (B) and subjected to mass spectrometry analysis (Table I). Two of the identified proteins, FACIN (acetylornithine aminotransferase) and GluD, were expressed recombinantly with His-Tags and purified by affinity chromatography to high purity as shown after separation of proteins by electrophoresis followed by Coomassie staining (C). (D) FACIN, GluD, as well as positive (UspA2) and negative (a-1-antitrypsin inhibitor) controls were immobilized in microtiter plates and incubated with 3–50 mg/ml C3met. The binding was then detected with specific Abs. (E) Surface plasmon resonance analysis of interaction between FACIN and C3b. C3b was immobilized on a CM5 sensor chip and FACIN at increasing concentrations ranging from 30 to 1000 nM was injected. Affinity constants were fitted using 1:1 Langmuir interaction model, and shown are means of two independent measurements. The analysis of C3met binding to FACIN revealed a strong 28 interaction with KD = 8.9 3 10 M. (F) FACIN is located on the surface of F. alocis. In addition, recombinant FACIN added to the bacteria showed a trend toward binding to the surface. FACIN was detected using specific Abs using flow cytometry analysis. Means of three independent experiments are presented with bars indicating SD. Statistical significance of observed differences in binding (compared with negative control) was estimated using either a repeated- measures (D) or a standard (E and F) two-way ANOVA (yielding p = 0.0153) and a Bonferroni posttest (to indicate exact differences at a given concentration). ****p , 0.0001, **p , 0.01, *p , 0.05.

FACIN abolishes hemolytic and bactericidal activity of human pacity to FACIN, GluD did not appear to signiﬁcantly affect serum complement activity in serum (Fig. 5A, 5B). To further conﬁrm Because C3 lies at the center of activation of all complement the effect of FACIN on the bactericidal activity of human serum, a pathways, we used well-established assays evaluating comple- we used an E. coli DH5 model system whereby cells were in- ment activity to test the effect of both C3-binding proteins from cubated with NHS pretreated with various concentrations of F. alocis. The proteins were incubated at various concentrations FACIN and DAF. Surviving bacterial cells were determined by with human serum, and the hemolytic activity of the CP and AP colony counting. GluD again did not have any effect in this was analyzed in the pretreated sera. FACIN showed very strong experimental system (data not shown). In contrast, FACIN, inhibitionofbothCP(Fig.5A)andAP(Fig.5B),inanextent similarly to DAF, destroyed the bactericidal activity of human similar to that of human decay-accelerating factor (DAF) used as serum in a dose-dependent manner and rescued E. coli that are a positive control. In contrast, despite similar C3 binding ca- otherwise very sensitive to killing by NHS (Fig. 5C). 3254 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM

Table I. Identiﬁcation of surface proteins of F. alocis responsible for C3 binding by mass spectrometry

PMFb Score Total Peptides Matched Predicted m.w. of Spot NCBI Accession No.a Protein Descriptiona (emPAIc) (Unique Peptides) the Protein 1 HMPREF0389_01649 NAD-specific GluD 568/1.25 27 (24) 47231 1 HMPREF0389_00745 Acetyl-CoA acetyltransferase 418/0.86 9 (8) 41067 2 HMPREF0389_01649 NAD-specific GluD 1544/1.94 58 (51) 47231 2 HMPREF0389_00745 Acetyl-CoA acetyltransferase 115/0.47 8 (6) 41067 3 HMPREF0389_01649 NAD-specific GluD 794/1.94 40 (33) 47231 3 HMPREF0389_00745 Acetyl-CoA acetyltransferase 196/0.47 9 (5) 41067 4 HMPREF0389_01570 Acetylornithine aminotransferase 849/1.72 36 (32) 44653 4 HMPREF0389_01649 NAD-specific GluD 145/0.22 4 (4) 47231 5 HMPREF0389_00744 Butyryl-CoA dehydrogenase 665/1.15 18 (17) 41459 5 HMPREF0389_00567 Glyceraldehyde-3-phosphate dehydrogenase 102/0.18 2 (2) 37599 aAccession numbers and protein descriptions are from the NCBI: F. alocis genome project (http://www.ncbi.nlm.nih.gov/genomeprj/30485). bPMF score: peptide mass fingerprinting score from Mascot. cemPAI: Exponentially Modified Protein Abundance Index.

FACIN inhibits complement-dependent activation and bound to the surface of F. alocis when FACIN was present during deposition of C3b on the surfaces incubation (Fig. 7B). C3bBb convertase is inherently labile and its Downloaded from Each complement pathway is a cascade of events proceeding in a proteolytically active component Bb quickly dissociates from the consecutive manner and leading to opsonization of a target with complex. FACIN appeared to stabilize Bb on the bacterial surface, activated forms of complement proteins, C4b (CP/LP) and C3b (all yet based on the largely reduced ability of AP convertase (Fig. 6C) pathways), and finally resulting in the assembly of terminal MAC. to generate C3b molecules, such stabilized convertase was no Using a microtiter plate-based assay whereby the plate surface is longer able to perform C3 cleavage. coated with pathway-specific ligands, the deposition of successive FACIN prevents C3b opsonization on F. alocis and blocks http://www.jimmunol.org/ complement factors in the presence of a potential inhibitor can be C3b-mediated phagocytosis detected with specific Abs. Complement activation was initiated by Το analyze the overall effect of FACIN on opsonization of F. al oc is aggregated human Igs or mannan for the CP and LP, respectively. in serum, we preincubated NHS with purified recombinant FACIN We found that deposition of C4b on such surfaces was not affected and incubated the bacteria in such sera followed by surface detec- by FACIN (data not shown), whereas there was a significant re- tion of C3b by flow cytometry. FACIN decreased opsonization of duction of deposited C3b molecules in both cases (Fig. 6A, 6B). F. a loc is with C3b from human serum in a dose-dependent manner Furthermore, using Mg-EGTA buffer and plates coated with zy- (Fig. 8A). At the highest concentration of FACIN, the remaining

mosan, we analyzed AP. C3b deposition by AP was significantly by guest on September 29, 2021 signal from detected C3b was comparable with that in DNHS, limited by FACIN (Fig. 6C). Inhibition of the central point of presumably indicating the amount of C3 passively bound by the complement activation by FACIN translated to an inhibition of bacteria (Fig. 8A). Next, we sought to investigate how such reduced downstream MAC deposition, as expected (Fig. 6D). C3b opsonization would affect phagocytosis of the bacteria. To our FACIN inhibits AP convertase surprise, in the preliminary experiments with purified human neu- When screening for complement inhibitory mechanisms of trophils as well as a model cell line HL60, we observed a strong F. alocis, we observed inhibition of the AP convertase when C3 inhibition of uptake and phagocytosis of F. al oc is, even in the ab- was allowed to interact with the bacterial surface before addition sence of purified FACIN (data not shown). We concluded that of FB and FD (Fig. 3D, 3E). To test whether FACIN was re- F. a loc is has a unique mechanism of direct phagocyte evasion. To sponsible for this effect, we preincubated it with serum proteins be able to selectively look on the uptake of C3b-opsonized targets and then activated complement by addition of zymosan. The re- in the presence of FACIN, without influence of other bacterial action was stopped after different time points by adding sample factors, we replaced the bacteria with zymosan particles and sub- buffer. After 20 min we observed pronounced AP convertase jected these to phagocytosis by HL60 cells. Our differentiation formation, as shown by efficient FB conversion to Bb and Ba protocol with DMSO-induced specific expression of CD11b on the (Fig. 7A). The convertase activity was reduced in the presence of surface of HL60 (data not shown), which is a component of the FACIN and almost completely inhibited in the presence of the complement receptor CR3 found on neutrophils mediating uptake strong complement inhibitor DAF (Fig. 7A). Furthermore, we of iC3b-coated particles (38). The cells then took up zymosan in a investigated the amount of FB and its activated derivative Bb manner dependent on complement opsonization. We found that in accumulating on the surface of F. alocis incubated in NHS, in the the presence of FACIN, phagocytosis of zymosan was significantly presence of recombinant FACIN. Surprisingly, we found more Bb reduced (Fig. 8B), demonstrating the functional outcome of FACIN-mediated inhibition of C3b opsonization.

Discussion Table II. Detection of gluD and FACIN genes and mRNA in Subgingival dental plaque bacteria evolved various mechanisms of subgingival plaque interacting with complement found in GCF (39–42). These in-

Samples with Low F. alocis Samples with High F. alocis teractions include both inhibitory and stimulatory effects, which Count ,105 per Site (n = 11) Count $105 per Site (n = 10) may seem contradictory at first, but our current perception is changing toward appreciating the fine-tuned adaptations of mi- DNA, n (%) mRNA, n (%) DNA, n (%) mRNA, n (%) crobes coexisting with the host in the dynamic inflammatory en- GluD 2 (18.2) 0 (0) 9 (90) 8 (80) vironment. Essentially, oral biofilm bacteria attempt to establish a FACIN 5 (45.5) 0 (0) 10 (100) 5 (50) balance between complement inhibition, leading to a decrease in The Journal of Immunology 3255

their immune clearance, and activation, providing a flow of nutrients in inflammatory exudates (35, 40, 43). Various factors produced by oral bacteria facilitate their sophisticated manipula- tion of host defense mechanisms. In this regard, the role of proteolytic enzymes of periodontal species in the corruption of innate immunity with focus on complement-directed actions is well- documented (35, 43–45). As shown in these studies, proteases of periodontal bacteria, on one hand, inhibit complement by degradation of various key components, such as mannose-binding lectin, ficolins, C3, C4, but on the other hand, they activate it by increasing C1q deposition or releasing C5a by C5 cleavage, and the final outcome may differ depending on the milieu dynamics (35, 43–45). Pathogens acting in a proinflammatory way may seem counterintuitive. Yet, C5 convertase-like activity of P. gingivalis protease gingipain was shown to be a crucial feature in the immune subversion and dysbiosis induction by this pathogen (12). Eventually, in progressing periodontitis the bacteria manage to establish an advanced plaque with complex composition and ar-

chitecture, and composed of numerous species interacting together Downloaded from and acting as one organized community. Analysis of in vivo grown subgingival plaques from periodontitis patients demonstrated that Gram-positive F. alocis is implicated in the pathogenic structure of oral bioﬁlm, and it predominantly colonizes apical parts of the pocket in proximity to the soft tissues, hence staying in close

contact with host defenses (23). Complement is the major defense http://www.jimmunol.org/ against all invading pathogens, presumably even more important in the confined space of the gingival pocket, with limited access of blood leukocytes. C3 is the central complement protein that directly, by proceeding to generation of MAC, or indirectly, by fa- cilitating phagocytosis, enables clearance of all invaders. When searching for immune evasion mechanisms of Gram-positive periodontal bacterium F. alocis, we quickly discovered that this pathogen aims to efficiently limit a key step of innate immunity, C3b opsonization. Intriguingly, F. alocis does so very efficiently by guest on September 29, 2021 by using its cytoplasmic enzyme, FACIN, which can be secreted and is also presented on the bacterial surface where it moonlights as a potent complement inhibitor. To our knowledge, this study is one of the very first reports regarding F. alocis, in which its immune evasion strategy is characterized in detail. Importantly, very few molecules, either natural or synthetic, have been shown to act at the C3 level, which is what classifies FACIN as a rare complement inhibitor. Several studies have already demonstrated a high prevalence and abundance of F. alocis in diseased periodontal sites with little detection in healthy or periodontitis-resistant individuals (21–23), and we confirmed these findings in our patient samples. In addi- FIGURE 5. F. aloc is complement inhibitor FACIN destroys bactericidal tion, we found further evidence that the bacterium contributes to activity of human serum. (A) CP. NHS (0.15%) was supplemented with various destructive inflammation and tooth attachment loss, by demon- concentrations of F. alocis proteins FACIN and GluD, BSA (negative control), strating its strong correlation with clinical parameters such as or DAF-Fc (positive control) and preincubated for 15 min at 37˚C, after which probing depth and bleeding on probing. Interestingly, counts of 2+ B sheep erythrocytes sensitized with Abs and diluted in DGVB were added. ( ) F. alocis strongly correlated with those of two major perio- AP. NHS (1.5%) was preincubated with increasing concentrations of proteins dontopathogens, P. gingivalis and T. forsythia, placing this novel for 15 min at 37˚C. Serum was then added to rabbit -erythrocytes diluted in species in the same rank of importance. These data certainly in- Mg2+-EGTA. (A and B) After 1-h incubation, the degree of lysis was estimated by measurement of released hemoglobin (absorbance at 405 nm). Lysis ob- dicate that the old concept of red complex bacteria as the most tained in the absence of any protein was set as 100%. (C) E. coli DH5a were pathogenic periodontal species (8) certainly needs to be revisited incubated with 0.75% NHS pretreated with increasing concentrations of pro- and put in context of most recently discovered oral bacteria. teins, and the surviving bacteria were enumerated after overnight culture on LB Because F. alocis reaches high numbers in the disease, it must agar plates. The survival was expressed as percentage of initial number bacteria produce factors to avoid immune recognition. We demonstrate that in samples (% of inoculum). (A–C) An average of three independent experi- all complement pathways are important for the recognition of ments is presented with bars indicating SD. Statistical significance of observed F. alocis and contribute to C3b opsonization (Fig. 2). Furthermore, differences (compared with no protein added) was estimated using two-way the vast majority of the C3b molecules opsonizing F. alocis are , repeated-measures ANOVA (with p 0.0001 for all tests) and a Bonferroni generated by the AP amplification loop, because depletion of FB posttest (to indicate exact differences at a given concentration). ****p , had the largest impact on complement deposition on the pathogen 0.0001, ***p , 0.001, **p , 0.01, *p , 0.05. surface. Strikingly, the bacterium grown as a biofilm showed 3256 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM

FIGURE 6. FACIN inhibits all pathways of complement. (A–D) Proteins were preincubated with 1% NHS (A), 2% NHS (B and D) diluted in GVB2+, or 4% NHS in Mg-EGTA (C) and added to microtiter plates coated with IgGs (A and D), mannan (B), or zymosan (C). After 35 (A and B)or 45 min (C and D) of incubation, the plates were washed and deposited C3b (A–C), and MAC (D) were detected with speciﬁc Abs. Absorbance obtained in the absence of any protein was set as 100%. An average of three independent experiments is presented with bars indicating SD. Sta- tistical signiﬁcance of observed differences compared with a condition in which no protein was added was estimated using two-way repeated- Downloaded from measures ANOVA (with p , 0.0001 for all tests) and a Bonferroni posttest (to indicate exact differences at a given concentration). ****p , 0.0001, ***p , 0.001, **p , 0.01, *p , 0.05. http://www.jimmunol.org/

significant reduction in C3b deposition compared with the plank- complement inhibitor (Fig. 5). We further demonstrated that inde- tonic form (Fig. 2E), suggesting increased expression of comple- pendently of the initial recognition pathway, the amount of C3b ment inhibitors under more physiological-like growth conditions. molecules produced as a result of complement amplification was Importantly, biofilm was thoroughly dispersed before the opsoni- reduced by 50–75% in the presence of FACIN, and a downstream by guest on September 29, 2021 zation process to avoid possible bias by morphological differences step of MAC formation was also strongly diminished (Fig. 6). By between the two growing stages. Similar effect of biofilm on limiting C3b opsonization, the bacterium can inhibit phagocytosis. complement evasion was previously reported for another human Importantly, the dominant cellular population in the gingival sulcus pathogen Streptococcus pneumoniae (46). In keeping with a key is neutrophils, which tend to form a specific structure called a role of C3b in the bacterial clearance, we found that F. al oc is targets leukocyte wall along the margins of the periodontal plaque (48). this crucial step in complement activation. We observed that C3 is Therefore, we initially attempted to perform phagocytosis assays not only deposited on bacterial surface via complement activation with use of primary human neutrophils. To our surprise, F. al oc is but also passively bound, and that both C3c and C3d are equally was phagocytosed very poorly by these cells, independently of important for the latter binding (Fig. 3B). These results suggest that serum opsonization (data not shown), suggesting the existence of a there is either a large binding site on C3 interacting with certain direct mechanism for inhibition of phagocytosis, which should be bacterial protein, or that C3 is captured by more than one bacterial studied in the future. Hence we used a simpler model to analyze the molecule. Our analysis of surface proteins of F. alocis interacting effect of FACIN on phagocytosis. We confirmed that in contrast with C3b indicated several C3 binding proteins on F. al oc is (Fig. 4). with undifferentiated cells, DMSO-treated HL60 cells selectively We expressed two of these, GluD and FACIN, and demonstrated took up zymosan particles when these were opsonized with serum, that both have the ability to bind C3. Both of these proteins are indicating an iC3b-CR3–mediated interaction. Such uptake of zy- intracellular enzymes, normally involved in the bacterial meta- mosan was inhibited when opsonization was performed in the bolism. Yet, we also identified them in the fraction of proteins presence of FACIN (Fig. 8B). C3b opsonization mediates not only anchored to the cell wall of the bacterium. This finding was not uptake of bacteria but also other key processes, such as Ab gen- surprising because release of conserved cytoplasmic proteins is eration, and C3 conversion precedes C5a generation. We can widely spread among Gram-positive and Gram-negative bacteria. speculate that by shutting down C3b opsonization, FACIN will One example of these is an enolase of S. pneumoniae,whichbinds disturb interactions of F. al oc is with various immune cells, although plasminogen and human complement inhibitor C4b-binding protein because of the aforementioned reasons we do not provide more (47). We believe that we identified two moonlighting proteins of experimental evidence in this regard. F. a loc is, which next to their enzymatic intracellular function ac- Interestingly, FACIN was previously identified in both the quired additional nonenzymatic roles in immune evasion. Whereas membrane fraction and secreted to the growth medium, when a GluD appeared to bind C3 without influencing its function, FACIN detailed analysis of the F. alocis proteome was performed (26). showed a strong inhibitory activity on complement, and we focused This allows us to speculate that FACIN may play a role both on our study on characterizing its molecular mechanism. Recombinant the surface of the bacterium and in a secreted form. Indeed, when FACIN showed a strong inhibitory effect on serum hemolytic and serum was preincubated with soluble FACIN and then added to the bactericidal activity, comparable with that of DAF, a potent human activating surface (IgG/mannan/zymosan-coated plate [Fig. 6] or The Journal of Immunology 3257 Downloaded from http://www.jimmunol.org/ FIGURE 7. FACIN inhibits AP convertase formation. (A) NHS (5%) was preincubated with FACIN or DAF for 15 min on ice, and subsequently complement was activated in the samples by adding zymosan for 0, 10, 20, or 30 min. Reactions were stopped by adding reducing sample buffer and boiling for 5 min. Samples were then centrifuged and supernatant from the samples were separated by SDS-PAGE followed by blotting the proteins onto the PVDF membrane. FB conversion to Ba and Bb was detected with specific pAbs. As a control, convertase was formed from purified compo- FIGURE 8. FACIN inhibits deposition of C3b on F. alocis and inhibits nents. (B) F. al oc is was incubated with increasing NHS concentrations in the phagocytosis. (A) F. alocis was incubated for 45 min with 5% NHS or by guest on September 29, 2021 presence of BSA or FACIN for 20 min at 37˚C. Bacteria were then washed, ΔNHS diluted in GVB2+. NHS was also supplemented with 3–100 mg/ml suspended in reducing sample buffer, and boiled. The amount of surface- FACIN. Deposited C3b was detected using specific Abs and flow cytom- attached FB/Bb was evaluated by Western blotting with specific pAbs. etry. Deposition/binding of C3b/C3 is shown as GMFI. (B) Zymosan beads conjugated with fluorescein were opsonized with NHS in the absence or presence of FACIN, and subsequently added to differentiated HL-60 cells bacteria [Fig. 8A]) there was limited C3b deposition on such at 2:1 ratio for 1 h at 37˚C. The uptake of fluorescent zymosan particles surfaces. When serum was preincubated with FACIN and then was evaluated by flow cytometry, and the frequency of positive cells was supplemented with zymosan to activate complement, we observed calculated. Data were normalized to the signal obtained in NHS in each A B abolished FB conversion to Ba and Bb (Fig. 7A). From these data experiment. ( and ) An average of three independent experiments is presented with bars indicating SD. Statistical significance of observed we can conclude that there is an overall decrease in the AP con- differences compared with NHS was estimated using one-way ANOVA vertase formation as well as its activity in the presence of F. alocis [yielding p , 0.0001 in both (A)and(B)] and a Dunnett’s posttest; protein. Furthermore, we detected the amount of Bb accumulating ****p , 0.0001, **p , 0.01, *p , 0.05. on the surface of F. alocis from NHS pretreated with FACIN (importantly here we washed away the supernatant after incubation); surprisingly, we found more Bb on F. alocis in such con- locally high deposition of C3b (amplified by AP) is a prerequisite ditions. Normally, because of the inherent instability of C3bBb for C3 convertases switch to C5 convertases. We speculate that convertase, Bb dissociates quickly and irreversibly from the FACIN first binds to C3/C3b without preventing its initial inter- complex and by doing that loses its activity, which is a part of action with FB. Presumably, once the complex is formed, FACIN convertase regulatory mechanism (49). Yet, in the presence of then locks the whole AP convertase in a more stable yet inactive FACIN, Bb appeared to be accumulated on the bacterial surface. state, and thereby limits C3 conversion. In fact such sophisticated When we coated C3-met on the plate and added FB in the pres- mechanism has already been exemplified in nature, by the family ence of FACIN, we did not observe any change in FB binding of staphylococcal complement inhibitors SCIN (50, 51). The (data not shown), even though FACIN also binds to C3-met SCIN proteins are potent inhibitors of all complement pathways (Fig. 4D). These data indicate that AP convertases are in fact and they act by stabilizing C3 convertases and hindering move- formed on the surface of the bacterium. We can also conclude that ments of essential catalytic domains toward C3, resulting in the the binding site of FACIN on C3 does not overlap with the one for cleavage inhibition (42, 43). As a result, SCIN proteins potently FB. Yet, the activity of AP convertase appears to be diminished by decrease phagocytosis and killing of S. aureus by neutrophils (50). FACIN, based on the decrease in C3b generation, which then feeds Molecular details of the interaction of SCIN with C3 convertases back as an overall lower number of newly formed convertases. were revealed after crystalizing the C3bBb complex stabilized by Furthermore, the formation of the cytolytic MAC was diminished the staphylococcal inhibitor (52), as well as in a report presenting in the presence of FACIN (Fig. 6D), consistent with the fact that complex binding studies with use of surface plasmon resonance 3258 FILIFACTOR ALOCIS INHIBITS THE COMPLEMENT SYSTEM

(51). A similar approach would be required to gain a full insight 13. Dymock, D., A. J. Weightman, C. Scully, and W. G. Wade. 1996. Molecular analysis of microflora associated with dentoalveolar abscesses. J. Clin. Micro- into FACIN mechanism, and our study provides a solid basis for biol. 34: 537–542. such further investigations. When it comes to limitations of further 14. Aas, J. A., B. J. Paster, L. N. Stokes, I. Olsen, and F. E. Dewhirst. 2005. Defining research, because of an indispensable role of FACIN in the main the normal bacterial flora of the oral cavity. J. Clin. Microbiol. 43: 5721–5732. 15. Kumar, P. S., A. L. Griffen, M. L. Moeschberger, and E. J. Leys. 2005. Identi- metabolic pathways of F. alocis, we can speculate that generating fication of candidate periodontal pathogens and beneficial species by quantitative a deficient strain would not be possible. 16S clonal analysis. J. Clin. Microbiol. 43: 3944–3955. We also present indirect evidence for the role of other virulence 16. Griffen, A. L., C. J. Beall, J. H. Campbell, N. D. Firestone, P. S. Kumar, Z. K. Yang, M. Podar, and E. J. Leys. 2012. Distinct and complex bacterial factors in F. alocis complement defense. When F. alocis was profiles in human periodontitis and health revealed by 16S pyrosequencing. treated with sodium azide, it lost its inhibitory activity toward ISME J. 6: 1176–1185. complement (Fig. 1A). This simple experiment gives further ev- 17. Li, Y., J. He, Z. He, Y. Zhou, M. Yuan, X. Xu, F. Sun, C. Liu, J. Li, W. Xie, et al. 2014. Phylogenetic and functional gene structure shifts of the oral microbiomes idence for the importance of secreted or cell wall–anchored pro- in periodontitis patients. ISME J. 9: 1879–1891. teins, in F. alocis complement resistance, as sodium azide inhibits 18. Kumar, P. S., E. J. Leys, J. M. Bryk, F. J. Martinez, M. L. Moeschberger, and secretory mechanism of Gram-positive bacteria (53). When A. L. Griffen. 2006. Changes in periodontal health status are associated with bacterial community shifts as assessed by quantitative 16S cloning and se- F. alocis survival was analyzed in whole human blood, we found quencing. J. Clin. Microbiol. 44: 3665–3673. that it can persist under these conditions, better than S. aureus or 19. Kumar, P. S., A. L. Griffen, J. A. Barton, B. J. Paster, M. L. Moeschberger, and E. J. Leys. 2003. New bacterial species associated with chronic periodontitis. J. the major periopathogen P. gingivalis (Fig. 6A). Considering that Dent. Res. 82: 338–344. S. aureus is responsible for fatal bloodstream infections (54) and 20. Wade, W. G. 2011. Has the use of molecular methods for the characterization of P. gingivalis and other oral bacteria are associated with systemic the human oral microbiome changed our understanding of the role of bacteria in the pathogenesis of periodontal disease? J. Clin. Periodontol. 38(Suppl. 11): 7–16. disease (55, 56), one can hypothesize that such comparable ability 21. Hutter, G., U. Schlagenhauf, G. Valenza, M. Horn, S. Burgemeister, H. Claus, and Downloaded from of F. alocis to resist phagocytosis may facilitate its dissemination U. Vogel. 2003. Molecular analysis of bacteria in periodontitis: evaluation of clone from the oral cavity to other sites in the organism, via the car- libraries, novel phylotypes and putative pathogens. Microbiology 149: 67–75. 22. Siqueira Jr., J. F., and I. N. Roˆças. 2003. Detection of Filifactor alocis in end- diovascular system. odontic infections associated with different forms of periradicular diseases. Oral Collectively, in this study we expand the knowledge about a Microbiol. Immunol. 18: 263–265. € recently discovered periodontal pathogen. Most importantly, we 23. Schlafer, S., B. Riep, A. L. Griffen, A. Petrich, J. Hubner, M. Berning, A. Friedmann, U. B. Go¨bel, and A. Moter. 2010. Filifactor alocis–involvement in characterize a novel function for FACIN in inhibition of C3 periodontal biofilms. BMC Microbiol. 10: 66. http://www.jimmunol.org/ convertases. We demonstrate various strategies of resisting com- 24. Moffatt, C. E., S. E. Whitmore, A. L. Griffen, E. J. Leys, and R. J. Lamont. 2011. Filifactor alocis interactions with gingival epithelial cells. Mol. Oral Microbiol. plement and phagocytes, which certainly allow F. alocis to thrive 26: 365–373. in the disrupted periodontal environment, while taking advantage 25. Aruni, A. W., F. Roy, and H. M. Fletcher. 2011. Filifactor alocis has virulence of the nutrients flowing in GCF supporting the growth of this attributes that can enhance its persistence under oxidative stress conditions and mediate invasion of epithelial cells by porphyromonas gingivalis. Infect. Immun. asaccharolytic pathobiont. 79: 3872–3886. 26. Aruni, A. W., F. Roy, L. Sandberg, and H. M. Fletcher. 2012. Proteome variation among Filifactor alocis strains. Proteomics 12: 3343–3364. Acknowledgments 27. Hajishengallis, G., and J. D. Lambris. 2012. Complement and dysbiosis in We thank Prof. Hansel M. Fletcher and Prof. Aruni Wilson (Loma Linda periodontal disease. Immunobiology 217: 1111–1116. University) for providing F. alocis strains. 28. Schenkein, H. A., and R. J. Genco. 1977. Gingival fluid and serum in periodontal by guest on September 29, 2021 diseases. II. Evidence for cleavage of complement components C3, C3 proac- tivator (factor B) and C4 in gingival fluid. J. Periodontol. 48: 778–784. Disclosures 29. Lally, E. T., W. P. McArthur, and P. C. Baehni. 1982. Biosynthesis of complement components in chronically inflamed gingiva. J. Periodontal Res. 17: 257–262. The authors have no financial conflicts of interest. 30. Uematsu, H., N. Sato, M. Z. Hossain, T. Ikeda, and E. Hoshino. 2003. Degradation of arginine and other amino acids by butyrate-producing asaccharolytic anaerobic Gram-positive rods in periodontal pockets. Arch. Oral Biol. 48: 423–429. References 31. Siqueira Jr., J. F., and I. N. Roˆças. 2004. Simultaneous detection of Dialister 1. Dewhirst, F. E., T. Chen, J. Izard, B. J. Paster, A. C. Tanner, W. H. Yu, pneumosintes and Filifactor alocis in endodontic infections by 16S rDNA- A. Lakshmanan, and W. G. Wade. 2010. The human oral microbiome. J. Bac- directed multiplex PCR. J. Endod. 30: 851–854. teriol. 192: 5002–5017. 32. Eick, S., A. Straube, A. Guentsch, W. Pfister, and H. Jentsch. 2011. Comparison of 2. Loesche, W. J., and N. S. Grossman. 2001. Periodontal disease as a specific, real-time polymerase chain reaction and DNA-strip technology in microbiological albeit chronic, infection: diagnosis and treatment. Clin. Microbiol. Rev. 14: 727– evaluation of periodontitis treatment. Diagn. Microbiol. Infect. Dis. 69: 12–20. 752 (table of contents). 33. Aduse-Opoku, J., N. N. Davies, A. Gallagher, A. Hashim, H. E. Evans, 3. Dietrich, T., M. Jimenez, E. A. Krall Kaye, P. S. Vokonas, and R. I. Garcia. 2008. M. Rangarajan, J. M. Slaney, and M. A. Curtis. 2000. Generation of lys- Age-dependent associations between chronic periodontitis/edentulism and risk gingipain protease activity in Porphyromonas gingivalis W50 is independent of coronary heart disease. Circulation 117: 1668–1674. of Arg-gingipain protease activities. Microbiology 146: 1933–1940. 4. Suvan, J., F. D’Aiuto, D. R. Moles, A. Petrie, and N. Donos. 2011. Association 34. Nordstro¨m, T., A. M. Blom, A. Forsgren, and K. Riesbeck. 2004. The emerging between overweight/obesity and periodontitis in adults. A systematic review. pathogen Moraxella catarrhalis interacts with complement inhibitor C4b bind- Obes. Rev. 12: e381–e404. ing protein through ubiquitous surface proteins A1 and A2. J. Immunol. 173: 5. Tonetti, M. S. 2009. Periodontitis and risk for atherosclerosis: an update on 4598–4606. intervention trials. J. Clin. Periodontol. 36(Suppl. 10): 15–19. 35. Popadiak, K., J. Potempa, K. Riesbeck, and A. M. Blom. 2007. Biphasic effect of 6. de Pablo, P., I. L. Chapple, C. D. Buckley, and T. Dietrich. 2009. Periodontitis in gingipains from Porphyromonas gingivalis on the human complement system. J. systemic rheumatic diseases. Nat. Rev. Rheumatol. 5: 218–224. Immunol. 178: 7242–7250. 7. Maddi, A., and F. A. Scannapieco. 2013. Oral biofilms, oral and periodontal 36. Hallstro¨m, T., T. Nordstro¨m, T. T. Tan, T. Manolov, J. D. Lambris, infections, and systemic disease. Am. J. Dent. 26: 249–254. D. E. Isenman, P. F. Zipfel, A. M. Blom, and K. Riesbeck. 2011. Immune evasion 8. Holt, S. C., and J. L. Ebersole. 2005. Porphyromonas gingivalis, Treponema of Moraxella catarrhalis involves ubiquitous surface protein A-dependent C3d denticola, and Tannerella forsythia: the “red complex”, a prototype polybacterial binding. J. Immunol. 186: 3120–3129. pathogenic consortium in periodontitis. Periodontol. 2000 38: 72–122. 37. Nordstro¨m, T., A. M. Blom, T. T. Tan, A. Forsgren, and K. Riesbeck. 2005. Ionic 9. Hajishengallis, G., and R. J. Lamont. 2012. Beyond the red complex and into binding of C3 to the human pathogen Moraxella catarrhalis is a unique mech- more complexity: the polymicrobial synergy and dysbiosis (PSD) model of anism for combating innate immunity. J. Immunol. 175: 3628–3636. periodontal disease etiology. Mol. Oral Microbiol. 27: 409–419. 38. Le Cabec, V., S. Carreńo, A. Moisand, C. Bordier, and I. Maridonneau-Parini. 10. Hajishengallis, G., T. Abe, T. Maekawa, E. Hajishengallis, and J. D. Lambris. 2002. Complement receptor 3 (CD11b/CD18) mediates type I and type II 2013. Role of complement in host-microbe homeostasis of the periodontium. phagocytosis during nonopsonic and opsonic phagocytosis, respectively. J. Semin. Immunol. 25: 65–72. Immunol. 169: 2003–2009. 11. Hajishengallis, G. 2014. Immunomicrobial pathogenesis of periodontitis: key- 39. Potempa, J., and R. N. Pike. 2009. Corruption of innate immunity by bacterial stones, pathobionts, and host response. Trends Immunol. 35: 3–11. proteases. J. Innate Immun. 1: 70–87. 12. Hajishengallis, G., S. Liang, M. A. Payne, A. Hashim, R. Jotwani, M. A. Eskan, 40. Krauss, J. L., J. Potempa, J. D. Lambris, and G. Hajishengallis. 2000. Com- M. L. McIntosh, A. Alsam, K. L. Kirkwood, J. D. Lambris, et al. 2011. Low- plementary Tolls in the periodontium: how periodontal bacteria modify com- abundance biofilm species orchestrates inflammatory periodontal disease through plement and Toll-like receptor responses to prevail in the host. Periodontol. 2000 the commensal microbiota and complement. Cell Host Microbe 10: 497–506. 52: 141–162. The Journal of Immunology 3259

41. Slaney, J. M., and M. A. Curtis. 2008. Mechanisms of evasion of complement by 49. Ponnuraj, K., Y. Xu, K. Macon, D. Moore, J. E. Volanakis, and S. V. Narayana. Porphyromonas gingivalis. Front. Biosci. 13: 188–196. 2004. Structural analysis of engineered Bb fragment of complement factor B: 42. Blom, A. M., T. Hallstro¨m, and K. Riesbeck. 2009. Complement evasion strat- insights into the activation mechanism of the alternative pathway C3-convertase. egies of pathogens-acquisition of inhibitors and beyond. Mol. Immunol. 46: Mol. Cell 14: 17–28. 2808–2817. 50. Rooijakkers, S. H., M. Ruyken, A. Roos, M. R. Daha, J. S. Presanis, R. B. Sim, 43. Jusko, M., J. Potempa, A. Y. Karim, M. Ksiazek, K. Riesbeck, P. Garred, S. Eick, W. J. van Wamel, K. P. van Kessel, and J. A. van Strijp. 2005. Immune evasion and A. M. Blom. 2012. A metalloproteinase karilysin present in the majority of by a staphylococcal complement inhibitor that acts on C3 convertases. Nat. Tannerella forsythia isolates inhibits all pathways of the complement system. J. Immunol. 6: 920–927. Immunol. 188: 2338–2349. 51. Ricklin, D., A. Tzekou, B. L. Garcia, M. Hammel, W. J. McWhorter, 44. Potempa, M., J. Potempa, T. Kantyka, K. A. Nguyen, K. Wawrzonek, G. Sfyroera, Y. Q. Wu, V. M. Holers, A. P. Herbert, P. N. Barlow, et al. 2009. A S. P. Manandhar, K. Popadiak, K. Riesbeck, S. Eick, and A. M. Blom. 2009. molecular insight into complement evasion by the staphylococcal complement Interpain A, a cysteine proteinase from Prevotella intermedia, inhibits comple- inhibitor protein family. J. Immunol. 183: 2565–2574. ment by degrading complement factor C3. [Published erratum appears in 2009 52. Rooijakkers, S. H., J. Wu, M. Ruyken, R. van Domselaar, K. L. Planken, PLoS Pathog. 5(3).] PLoS Pathog. 5: e1000316. A. Tzekou, D. Ricklin, J. D. Lambris, B. J. Janssen, J. A. van Strijp, and P. Gros. 45. McDowell, J. V., B. Huang, J. C. Fenno, and R. T. Marconi. 2009. Analysis of a 2009. Structural and functional implications of the alternative complement unique interaction between the complement regulatory protein factor H and the pathway C3 convertase stabilized by a staphylococcal inhibitor. Nat. Immunol. periodontal pathogen Treponema denticola. Infect. Immun. 77: 1417–1425. 10: 721–727. 46. Domenech, M., E. Ramos-Sevillano, E. Garcıá, M. Moscoso, and J. Yuste. 2013. 53. Navarre, W. W., and O. Schneewind. 1999. Surface proteins of gram-positive Biofilm formation avoids complement immunity and phagocytosis of Strepto- bacteria and mechanisms of their targeting to the cell wall envelope. Microbiol. coccus pneumoniae. Infect. Immun. 81: 2606–2615. Mol. Biol. Rev. 63: 174–229. 47. Agarwal, V., S. Hammerschmidt, S. Malm, S. Bergmann, K. Riesbeck, and 54. Corey, G. R. 2009. Staphylococcus aureus bloodstream infections: definitions A. M. Blom. 2012. Enolase of Streptococcus pneumoniae binds human com- and treatment. Clin. Infect. Dis. 48(Suppl. 4): S254–S259. plement inhibitor C4b-binding protein and contributes to complement evasion. J. 55. Bokarewa, M. I., T. Jin, and A. Tarkowski. 2006. Staphylococcus aureus: Immunol. 189: 3575–3584. staphylokinase. Int. J. Biochem. Cell Biol. 38: 504–509. 48. Delima, A. J., and T. E. Van Dyke. 2003. Origin and function of the cellular 56. Scannapieco, F. A. 2004. Periodontal inflammation: from gingivitis to systemic components in gingival crevice fluid. Periodontol. 2000 31: 55–76. disease? Compend. Contin. Educ. Dent. 25(7Suppl. 1): 16–25. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021