The Journal of Immunology

Bordetella pertussis Adenylate Cyclase Blocks Induction of Bactericidal in through cAMP-Dependent Activation of the SHP-1 Phosphatase

Ondrej Cerny, Jana Kamanova,1 Jiri Masin, Ilona Bibova, Karolina Skopova, and Peter Sebo

The adenylate cyclase toxin– (CyaA) plays a key role in the virulence of pertussis. CyaA penetrates comple- ment receptor 3–expressing and catalyzes uncontrolled conversion of cytosolic ATP to the key second messenger molecule cAMP. This paralyzes the capacity of and macrophages to kill by complement-dependent oxidative burst and opsonophagocytic mechanisms. We show that cAMP signaling through the kinase A (PKA) pathway activates Src homology domain 2 containing protein tyrosine phosphatase (SHP) 1 and suppresses production of bactericidal NO in cells. Selective activation of PKA by the cell-permeable analog N6-benzoyladenosine-39,59-cyclic monophosphate interfered with LPS-induced inducible NO synthase (iNOS) expression in RAW264.7 macrophages, whereas inhibition of PKA by H-89 largely restored the production of iNOS in CyaA-treated murine macrophages. CyaA/cAMP signaling induced SHP phosphatase–dependent dephosphorylation of the c-Fos subunit of the transcription factor AP-1 and thereby inhibited TLR4- triggered induction of iNOS gene expression. Selective small interfering RNA knockdown of SHP-1, but not of the SHP-2 phosphatase, rescued production of TLR-inducible NO in toxin-treated cells. Finally, inhibition of SHP phosphatase activity by NSC87877 abrogated B. pertussis survival inside murine macrophages. These results reveal that an as yet unknown cAMP- activated signaling pathway controls SHP-1 phosphatase activity and may regulate numerous receptor signaling pathways in leukocytes. Hijacking of SHP-1 by CyaA action then enables B. pertussis to evade NO-mediated killing in sentinel cells of innate immunity. The Journal of Immunology, 2015, 194: 4901–4913.

he agent produces cells (7). The AC catalyzes uncontrolled conversion of cytosolic an RTX (Repeats in ToXin) adenylate cyclase (AC) toxin– ATP to the key second messenger molecule cAMP and interferes T hemolysin (CyaA, ACT, or AC-Hly) that paralyzes bac- with signaling of opsonins through receptors of complement and tericidal functions of myeloid phagocytes (1–6). The 1706- Igs on neutrophils, macrophages, or dendritic cells (7). As a result, residue-long bifunctional toxin binds the complement receptor 3 production of bactericidal (ROS) and (aMb2 integrin, CD11b/CD18, or Mac-1). The toxin forms small formation of extracellular traps is rapidly inhibited (8– cation-selective (hemolytic) pores in cellular membrane and its 10). In macrophage cells, the produced cAMP causes transient main cytotoxic activity consists in delivery of the N-terminal inhibition of RhoA GTPase activity and elicits massive actin -activated AC enzyme domain into the cytosol of rearrangements and unproductive membrane ruff- ling (11). As a result, and killing of complement- opsonized bacteria by neutrophils and macrophages are blocked Laboratory of Molecular Biology of Bacterial Pathogens, Institute of Microbiology of the ASCR, v.v.i., Czech Academy of Sciences, 142 20, Prague 4, Czech Republic near-instantaneously (8, 11, 12). Moreover, CyaA/cAMP signaling also skews the TLR-triggered maturation of dendritic cells and 1Current address: Yale University School of Medicine, New Haven, CT. impairs their capacity to present Ags to T cells for induction of Received for publication November 24, 2014. Accepted for publication March 10, 2015. adaptive cellular immune responses (13). This work was supported by Czech Science Foundation Grants P302/12/0460 (to J.M.), In the absence of opsonization, however, a fraction of internalized 13-14547S (to P.S.), and the Institutional Reasearch Project No. RVO61388971. O.C. B. pertussis bacteria was repeatedly found to escape killing and to is a doctoral student of the Faculty of Science of the Charles University in Prague, survive within primary human macrophages for several days (14, Czech Republic; I.B. and K.S. are doctoral students of the Institute of Chemical Technology in Prague, Czech Republic. 15). Because inducible NO production belongs to the major bacte- Address correspondence and reprint requests to Dr. Peter Sebo, Institute of Micro- ricidal mechanisms in macrophages, this indicated that B. pertussis biology of the ASCR, v.v.i., Videnska 1083, 142 20 Prague 4, Czech Republic. might possess an as yet unknown mechanism that enables it to es- E-mail address: [email protected] cape killing by NO produced by the inducible NO synthase (iNOS, The online version of this article contains supplemental material. NOS2). iNOS expression is, indeed, induced in macrophages upon Abbreviations used in this article: AC, adenylate cyclase; BGA, Bordet-Gengou agar; activation by bacterial LPS, or IFN-g or TNF-a (16). Production of 6 BMDM, bone marrow–derived macrophage; 6-Bnz-cAMP, N -benzoyladenosine- the highly active Ca2+-independent iNOS enzyme in macrophage 39,59-cyclic monophosphate; 8-CPT-cAMP, 8-(4-chlorophenylthio)-29-O-methylade- nosine-39,59-cyclic monophosphate; CyaA, AC toxin–hemolysin; DAF-FM, 4-amino- cells results, indeed, in production of toxic amounts of NO, and this 5-methylamino-29,79-difluorofluorescein diacetate; db-cAMP, N6,29-O-dibutyrylade- plays an important role in host defense against pathogens, including nosine-39,59- cyclic monophosphate; Epac, exchange protein directly activated by cAMP; IBMX, 3-isobutyl-1-methylxanthine; iNOS, NOS2, inducible NO synthase; B. pertussis (17, 18). Regulation of NO production in monocytic MOI, multiplicity of ; PKA, ; PT, pertussis toxin; ROS, cells then largely depends on regulation of iNOS gene expression reactive oxygen species; siRNA, small interfering RNA; STAT1, signal transducer level (19). This results from the interplay of transcription factors, and activator of transcription 1. such as NF-kB, the signal transducer and activator of transcription 1 Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 (STAT1), and the AP-1, respectively (20, 21). Involvement of the www.jimmunol.org/cgi/doi/10.4049/jimmunol.1402941 4902 iNOS EXPRESSION ABLATION BY cAMP SIGNALING CyaA TOXIN transcription factors NF-kB and AP-1 in iNOS expression further DMEM containing 10% (v/v) FCS in the presence of 10% (v/v) conditioned appears to be modulated by the activity of the SHP-1, which was media of L929 CMG cells and antibiotic–antimycotic solution (0.1 mg/ml also shown to modulate the activity of the transcription factor streptomycin, 100 U/ml penicillin, and 0.25 mg/ml amphotericin; Sigma- Aldrich) for 7 d at 37˚C in a humidified air–CO2 (5%) atmosphere. STAT1 and of the IRF-1 (22, 23). RAW264.7 mouse macrophage cells (ATCC cat. no. TIB 71) were grown in Production of the iNOS enzyme was previously observed to play RPMI 1640 medium supplemented with 10% (v/v) FCS. Prior to assays, the a role in host defense against B. pertussis infection, indicating that phosphate-buffered RPMI 1640 medium was replaced with HEPES-buffered 2+ the pathogen may be sensitive to NO-mediated killing. Indeed, DMEM medium with 10% (v/v) FCS, which contains 1.9 mM Ca , and the cells were allowed to rest in DMEM for 2 h before toxin addition. naive iNOS-deficient mice were found to be more susceptible to B. pertussis respiratory challenge than were wild-type mice, and Determination of ROS production upon by a whole-cell , the iNOS- ROS production was measured using a luminol-based assay as described deficient mice were less efficiently protected from infection than (28). Briefly, 5 3 105 RAW264.7 macrophages in HBSS supplemented their wild-type littermates (17). It has, however, not been explored with 1% glucose and 2 mM CaCl2 were incubated for 3 min at 37˚C with whether B. pertussis uses manipulation of iNOS expression in 150 mM luminol, before cells were transferred to wells containing the macrophages as part of its virulence strategy. We report that al- given activator (e.g., human complement–opsonized zymosan at 10 mg/ml, complement-opsonized B. pertussis at multiplicity of infection [MOI] ready low CyaA doses inhibit iNOS expression and NO produc- 10:1, or unopsonized B. pertussis at an MOI 10:1, respectively). Lumi- tion in mouse macrophages and that the AC enzyme activity of the nescence was recorded using a Safire2 microplate reader (Tecan), and data CyaA toxin is essential for extended survival of B. pertussis in output is given in relative light units integrated over 1 h of measurement. macrophage cells. We show that cAMP signaling leads to acti- Opsonization of bacteria with human complement vation of tyrosine phosphatase SHP-1 and that this blocks TLR- induced NO production. Randomized fresh human blood was purchased at the transfusion unit of the Thomayer Hospital in Prague, and complete human sera were obtained by centrifugation at 1200 rpm, 20 min, 17˚C. Prior to use, the sera were Materials and Methods controlled by ELISA for the absence of any detectable amounts of Abs Abs and reagents recognizing the PT, CyaA, and FHA Ags of B. pertussis (data not shown). For opsonization by human complement, 1 3 108 heat-killed B. pertussis 0111:B4 LPS, 3-isobutyl-1-methylxanthine (IBMX), cells (70˚C, 30 min), or 1 mg of zymosan, were incubated with 50% human and N6,29-O-dibutyryladenosine-39,59- cyclic monophosphate (db-cAMP) serum for 30 min at 37˚C under gentle shaking, and the suspensions were were obtained from Sigma-Aldrich. N6-benzoyladenosine-39,59-cyclic washed twice with serum-free HBSS. monophosphate (6-Bnz-cAMP) and 8-(4-chlorophenylthio)-29-O-methyl- adenosine-39,59-cyclic monophosphate (8-CPT-cAMP) were from NO assay BIOLOG Life Science Institute. NSC87877 and H-89 were purchased Production of NO was measured either as nitrite accumulation in the from Calbiochem, DAE-NONOate was from Cayman Chemical, and hy- medium, using the Griess reaction (29), or as accumulation of both ni- drogen peroxide was from Lach-Ner. Abs were from Santa Cruz Bio- trite and nitrate, using fluorescent 4-amino-5-methylamino-2’,7’-difluoro- technology (anti–SHP-1, anti–SHP-2), from Technology fluorescein diacetate ([DAF-FM]; Life Technologies) (30). RAW264.7 (anti–P–c-Fos, anti-iNOS, anti–b-actin, anti-p65, anti-IRF1, anti–P-IkB, cells were seeded into 96-well plates (105 per well) and incubated with the anti-IkB, anti-STAT1, anti–P-Y701 STAT1, anti–P-S727 STAT1, anti–P- indicated inhibitors for 1 h prior to toxin addition. After 24 h of continued Y564 SHP-1), from ECM Biosciences (anti–P-Y563 SHP-1, anti–P-S591 incubation, 100-ml aliquots of culture supernatants were mixed with 100 SHP-1), or from BD Biosciences (anti–SHP-1). All small interfering RNAs ml Griess reagent [2% sulfanilamide, 0.2% N-(1-napthyl)-ethylenediamine (siRNAs) (ON TARGETplus SMARTpool, si-PTPN6, si-PTPN11, and dihydrochloride, 2.5% orthophosphoric acid], and OD was measured non-targeting siRNA control) as well as DharmaFECT 4 were obtained 540 using a Safire2 microplate reader (Tecan). For measurement with DAF- from Dharmacon. FM, the fluorescent probe (10 mM) was added to cell suspensions 22 h Bacterial strains after stimulation. At 2 h later, cells were washed five times with ice-cold PBS, and fluorescence (excitation, 485 nm; emission, 520 nm) was mea- Bacterial strains were derived from B. pertussis Tohama I (B.p. cyaA-wt), sured using a Safire2 microplate reader (Tecan). obtained as strain CIP 81.32 from the Collection of Institute Pasteur, Paris, France. The B.p. DcyaA and B.p. cyaA-AC2 strains, carrying an in-frame cAMP assay deletion of the cyaA open reading frame on the chromosome, or secreting 2 Accumulation of cAMP in RAW264.7 cells was measured as described earlier an enzymatically inactive CyaA-AC toxoid (24), were constructed using (11). Briefly, cells were incubated for the indicated times with CyaA at 10 the pSS4245 allelic exchange vector, generously provided by Dr. S. Stibitz ng/ml in DMEM in the presence or absence of 100 ng/ml LPS or were (25). Plasmid constructs were transformed into E. coli SM10 lpir and infected with live B. pertussis bacteria at the MOI 10:1, respectively. The transferred by conjugation into recipient B. pertussis cultured for 4 d under 2 reaction was stopped by the addition of 0.2% Tween 20 in 50 mM HCl. The modulating conditions (Bvg ) on BGA plates (Bordet-Gengou agar, samples were boiled for 15 min at 100˚C, neutralized with 150 mM un- Becton Dickinson) containing 15% defibrinated sheep blood and 50 mM buffered imidazole, and cAMP concentration was determined by ELISA. MgSO4. B. pertussis clones having the plasmid construct integrated into the chromosome were selected for 5 d at 37˚C on plates that contained 50 Determination of arginase activity mM MgSO4, 500 mg/ml streptomycin, 30 mg/ml ampicillin, and 40 mg/ml kanamycin and were restreaked on the same plates for an additional 5 d, Arginase activity was measured as previously described (31). RAW264.7 5 before plating on BGA lacking MgSO4, to select for allelic exchange. The cells (10 per well) were seeded into 96-well plates and incubated with obtained clones were characterized for phenotypic change, and the pres- indicated concentrations of CyaA for 24 h and lysed in 50 ml buffer ence of introduced as well as the absence of undesired unmarked mutations containing 50 mM Tris HCl (pH 7.5), 0.1% Triton X-100, 50 mM NaF, 10 was verified by PCR sequencing of relevant portions of the cyaA gene. mM Na4P2O7, 1 mM Na3VO4, 50 nM Calyculin A, and the Complete Mini Production of key virulence factors like PT, PRN, FHA, and CyaA was EDTA-free mixture inhibitor (Roche). After 30 min at room verified by immunoblotting of bacterial lysates, using specific Abs (26). temperature, 10 mM MnCl2 was added and the samples were heated for 10 The plasmid pBBRcyaGFP encoding GFP under control of the cyaA gene min at 55˚C. An equal volume of 0.5 M L-arginine was added, and the promoter was introduced into Bordetella strains by conjugation, using 50 mixture was incubated for 60 min at 37˚C. The reaction was stopped by the mg/ml kanamycin and 150 mg/ml cephalexin for selection. addition of 8 volumes of a mixture of H2SO4 and H3PO4 in water (1:3:7) with half a volume of 9% a-isonitrosopropiophenone dissolved in ethanol. Cell cultures and handling Samples were boiled for 45 min at 100˚C and were chilled for 10 min in the dark, and absorbance was read at 540 nm. The L929 CMG cells (a kind gift of T. Brdicka, Institute of Molecular Genetics of the ASCR, v.v.i., Prague, Czech Republic) were used for M-CSF Immunodetection of production (27). Bone marrow macrophage-like cells (bone marrow–derived macrophages [BMDMs]) were obtained from femoral and tibial bones of RAW264.7 or BMDM cells (106 per well) were cultured in 12-well plates 6-wk-old female C57BL/6 mice by cultivating bone marrow–derived cells in in RPMI and replaced into DMEM medium for 2 h, and inhibitors were The Journal of Immunology 4903 added 1 h before the addition of CyaA. Cells were washed twice with ice- jugated phalloidin (0.5 mg/ml; Sigma-Aldrich) and DNA with DAPI (10 cold PBS and lysed with 1% Nonidet P-40 in 20 mM Tris-HCl (pH 8.0) mg/ml; Sigma-Aldrich) in PBS supplemented with 3% BSA (v/v) for 30 buffer containing 100 mM NaCl, 10 mM EDTA, 10 mM Na4P2O7,1mM min. Samples were mounted on glass coverslips in Mowiol, and images Na3VO4, 50 mM NaF, 10 nM Calyculin A, and Complete Mini protease were captured using the CellR Imaging Station based on the Olympus inhibitors (Roche). SDS-PAGE–separated proteins were transferred onto IX81 fluorescence microscope using a 1003/1.35 oil objective. Images nitrocellulose membranes and probed by the indicated mAbs and sec- were captured in 0.1-mm Z-stack layers across the complete cell, and ondary Ab-peroxidase conjugates (1/5000; GE Healthcare); chemilumi- the three-dimensional image was reconstructed using three-dimensional nescence was revealed using the West Femto Maximum Sensitivity deconvolution. Internalized bacteria, emitting only green fluorescence Substrate (Pierce); and signals were quantified using the ImageQuant LAS (GFP+) and cell surface–associated yellow bacteria (emitting both red 4000 imaging system (Fuji) and analyzed using AIDA software (v. 3.28; [Cy5] and green [GFP] fluorescence), were counted for at least 50 cells per Raytest Isotopenmessgeraete). sample and experiment. RNA isolation and quantitative real-time PCR SHP-1 activity assay RAW264.7 cells (2 3 106 per well) were incubated in DMEM with 100 ng/ A total of 5 3 106 RAW264.7 cells were incubated in DMEM with 100 ng/ ml LPS and/or CyaA toxin at the indicated concentrations for 24 h, washed ml LPS and/or 10 ng/ml CyaA and washed twice with ice-cold PBS. Cells twice with ice-cold PBS, and lysed with TRIzol Reagent (Life Technol- were lysed and SHP-1 was immunoprecipitated from postnuclear fractions ogies). Total RNA was extracted using the RNeasy Mini Kit (QIAGEN), using the Pierce Co-Immunoprecipitation (Co-IP) Kit. The columns with including a treatment with DNase I (Ambion), and 1 mg total RNA was bound SHP-1 were washed six times with reaction buffer (25 mM imid- reverse transcribed into cDNA in a 25-ml reaction using M-MLV Reverse azole pH 7.2, 45 mM NaCl, 1 mM EDTA in phosphate-free water). Then Transcriptase (Promega) and 0.5 mg random hexamers and oligo(dT) 250 mM tyrosine phosphopeptide (RRLIEDAEpYAARG; Upstate Bio- mixture. Quantitative PCR was performed on a Bio-Rad CFX96 instru- technologies) was loaded on the columns. After 30 min at 37˚C, the free ment, using SYBR Green JumpStart Taq ReadyMix (Sigma-Aldrich) and phosphate was determined in column eluates using Malachite Green gene-specific primers (Supplemental Table I). A total of 200 nM of each Phosphate Assay (ScienCell). SHP-1 was next eluted from the columns primer with 40 ng reverse-transcribed RNA was used in a 20-ml qPCR and detected by immunoblotting. Specific SHP-1 activity was calculated as reaction, with an initial step at 95˚C for 2 min, followed by 40 cycles of the amount of free phosphate normalized to the total eluted SHP-1 protein 95˚C for 15 s, 60˚C for 30 s, and 72˚C for 30 s. The RPL37 gene was used amount. as reference, and relative gene expression was quantified using amplifi- cation efficiency values (32). NF-kB translocation assay siRNA silencing Translocation of NF-kB into the nucleus was analyzed by confocal mi- croscopy. A total of 105 RAW264.7 cells were seeded onto glass coverslips The siRNA transfection protocol (Dharmacon) was optimized for 1 d before the experiment and after incubation with or without LPS (100 RAW264.7 cells, using FITC-labeled siGlo RNA interference control. A ng/ml) and CyaA (10 ng/ml) for 3 h; the cells were fixed with 4% para- total of 2 3 104 cells per well were seeded into 96-well plates and grown formaldehyde and permeabilized with 0.1% Triton X-100. Upon blocking overnight in RPMI medium without antibiotics and transfected with 50 nM with PBS-5% BSA (v/v), NF-kB staining was performed at 4˚C overnight siRNA using 0.05% DharmaFECT 4 reagent. Transfection was repeated in PBS-2% BSA (v/v) buffer with anti-p65 Ab (dilution 1/100) and 48 h later to increase silencing efficacy. After an additional 48 h, the cells revealed using goat anti-rabbit secondary Ab conjugated with FITC were repeatedly washed with prewarmed DMEM medium and tested (Sigma-Aldrich). Glass coverslips were mounted in Mowiol, and images for NO production and/or target-protein expression. Unspecific effects were captured using Leica confocal microscope TCS SP2 (Wetzlar). were controlled in all experiments by including untreated, mock-treated (DharmaFECT 4 reagent only), and nontargeting siRNA-transfected cells, NF-kB reporter assay respectively. RAW264.7 cells were transfected using the FuGENE HD transfection 5 In vitro killing assay reagent according to the manufacturer’s instructions. Briefly, cells (10 per well) were seeded into 96-well plates in RPMI medium without antibiotics. B. pertussis suspensions were grown in liquid Stainer–Scholte medium On the next day, a mixture of pNF-kB d2EGFP with pEGFP/mCherry-N1 (33) to OD600 = 1. DAE-NONOate or hydrogen peroxide was added, and and FuGENE HD in the ratio 2:8 was added to the cell culture. After 18 h, after 2 h of incubation at 37˚C, the serial dilutions were plated on BGA for cells were washed with preheated DMEM medium and incubated in the determination of viable CFUs after 5 d of growth. presence or absence of CyaA (10 ng/ml) and LPS (100 ng/ml). After 6 h, the ratio of mCherry and GFP expression was analyzed using flow RAW264.7 cell infection by B. pertussis cytometry (LSR II; BD). Cell infection assays were performed as described by Lamberti et al. (15). Ethics statement Briefly, RAW264.7 cells (105 per well) were seeded overnight into 24-well plates in RPMI medium with 10% heat-inactivated FCS, but without All animal experiments were approved by the Animal Welfare Committee antibiotics. Prior to infection, RPMI 1640 medium was replaced by of the Institute of Microbiology of the Academy of Sciences of the Czech DMEM (containing 1.9 mM Ca2+ and 10% [v/v] HIFCS), and exponen- Republic. Animals were handled according to the Guidelines for the Care tially growing B. pertussis cells expressing GFP from the pBBRcyaGFP and Use of Laboratory Animals, the Act of the Czech National Assembly, plasmid were added at MOI 10:1. The bacteria were gently spun onto Collection of Laws no. 149/2004, inclusive of the amendments, on the Pro- macrophages at 640 3 g for 5 min to facilitate attachment. Coincubation tection of Animals against Cruelty, and Public Notice of the Ministry of was continued for 1 h at 37˚C, and unattached bacteria were removed by Agriculture of the Czech Republic, Collection of Laws no. 207/2004, on rinsing cells three times with prewarmed DMEM, before the cells were care and use of experimental animals. fixed for microscopic analysis, or before 100 mg/ml polymyxin B was added for 1 h to kill the attached extracellular bacteria. Polymyxin B Statistical analysis m concentration was next decreased to 20 g/ml in the culture medium to The significance of differences in values was assessed by the Student t test. avoid killing of internalized bacteria during continued cell culture (24, 48, and 72 h, respectively). Cells were washed three times with prewarmed medium without antibiotics and lysed with sterile water. Lysates were Results serially diluted in Stainer–Scholte medium and plated on BGA for deter- B. pertussis is sensitive to NO-mediated killing and blocks mination of B. pertussis CFUs after 5 d of growth. iNOS expression and NO production in macrophages through Fluorescence microscopy of cell-associated bacteria CyaA-catalyzed formation of cAMP After 1 h of incubation with B. pertussis/pBBRcyaGFP bacteria, the CyaA toxin production was previously found to be required for RAW264.7 cells were washed three times with prewarmed DMEM me- persistence of unopsonized B. pertussis in macrophages (14). It dium and fixed with 4% paraformaldehyde in PBS for 20 min. Non- remained, however, unclear whether the pore-forming hemolytic phagocytosed bacteria attached to the surface of macrophage cells were decorated with rabbit polyclonal antiserum against B. pertussis (a kind gift activity of CyaA contributed to survival of B. pertussis in mac- of Dr. B. Vecerek) and stained with goat anti-rabbit IgG conjugated with rophages, or whether this depended entirely on the capacity of the Cy5. F-actin was stained with tetramethylrhodamine isothiocyanate–con- cell-invasive AC enzyme of CyaA to elevate cAMP concentrations 4904 iNOS EXPRESSION ABLATION BY cAMP SIGNALING CyaA TOXIN in cells. Therefore, survival in murine RAW264.7 macrophage cells of a hemolytic B. pertussis strain producing a catalytically inactive CyaA-AC2 toxoid unable to convert cellular ATP to cAMP (B.p. cyaA-AC2) was compared at MOI 10:1 with persis- tence of the parental B. pertussis Tohama I strain (B.p. cyaA-wt) producing intact CyaA. As a negative control, a mutant not pro- ducing CyaA owing to the deletion of the entire cyaA open reading frame (B.p. DcyaA) was used. As shown in Fig. 1, asso- ciation of cyaA-AC2 or of DcyaA bacteria with the RAW264.7 macrophages after 1 h of coincubation was not impaired by the inability to produce enzymatically active CyaA toxin. When compared with the cyaA-wt strain, internalization of bacteria producing the AC2 toxoid (cyaA-AC2), or not producing CyaA at all (DcyaA), was even enhanced (∼120%). Although the differ- ences were not statistically significant, this observation would go well with the previously shown capacity of CyaA to provoke cAMP-triggered unproductive actin cytoskeleton rearrangements and macrophage ruffling (11). In contrast to enhanced initial up- take, the survival of the cyaA-AC2 and DcyaA bacteria within murine macrophages was significantly impaired in 24 h, as shown in Fig. 1B. After 48 h of incubation, about two orders of magni- tude lower viable counts of the mutants were recovered from RAW264.7 cells. Hence, the enzymatic AC activity of CyaA was crucial for extended survival of unopsonized B. pertussis inside murine macrophages. CyaA-catalyzed elevation of cAMP was previously shown to efficiently block oxidative burst in neutrophils (8, 10). However, as shown in Fig. 2A, despite responding by production of ROS to stimulation by complement-opsonized zymosan, the RAW264.7 cells produced little or no ROS upon incubation with heat-killed B. pertussis cells, irrespective of whether these were opsonized by complement or not. It hence appeared unlikely that the unopson- ized bacteria, eliciting even lower ROS production than opsonized bacteria, were killed in RAW264.7 cells by oxidative burst. Macrophage cells are, however, known to produce substantial amounts of bactericidal NO (34). Moreover, B. pertussis was shown to persist much better inside activated macrophages from iNOS-deficient mice (17). Therefore, we tested whether suppres- sion of bactericidal NO production by CyaA contributed to sur- vival of unopsonized B. pertussis inside macrophages. FIGURE 1. Survival of B. pertussis inside macrophages. RAW264.7 As shown in Fig. 2B, indeed, B. pertussis is rather susceptible to cells were infected at MOI 10:1 for 1 h (37˚C, 5% CO2) by centrifuging 3 g NO-mediated killing in vitro, because exposure to 400 mM DEA- (640 , 5 min) onto macrophages the GFP-expressing wild-type B. pertussis Tohama I (B.p. cyaA-wt), the bacteria producing enzymatically NONOate as a donor of NO radicals (35, 36) significantly de- 2 2 inactive CyaA-AC toxoid (B.p. cyaA-AC ), or B. pertussis lacking the creased viability of the bacteria, albeit less than upon oxidative cyaA gene (B.p. DcyaA), respectively. (A) Upon coincubation for 1 h at 37˚C, killing by 100 mMH2O2 (37). As further shown in Fig. 2C, sig- the cells were washed and fixed with 4% paraformaldehyde (20 min). nificantly decreased NO production was observed in RAW264.7 Extracellular bacteria were stained with rabbit anti-B. pertussis serum and macrophages infected with bacteria producing intact CyaA detected with the Cy5-labeled goat anti-rabbit IgG conjugate. Tetrame- (B.p. cyaA-wt), compared with RAW264.7 cells coincubated with thylrhodamine isothiocyanate––phalloidin (1 mg/ml) was used for actin heat-inactivated B. pertussis, or with macrophages infected with staining and DAPI (10 mg/ml) for visualization of nuclei. Representative B.p. cyaA-AC2 or B.p. DcyaA bacteria, respectively. This sug- images from one out of five independent experiments are shown. Scale bar, B gested that signaling of CyaA-produced cAMP interfered with 5 mm. ( ) The RAW264.7 cells were infected at MOI 10:1 as above, and TLR-triggered induction of NO production. As indeed shown in after 1 h of coincubation the cells were washed and replaced for 1 h into DMEM with 100 mg/ml of polymyxin B to kill extracellular bacteria. The Fig. 2D, a detectable increase of intracellular cAMP concentration infected cells were incubated for an additional 24, 48, or 72 h prior to lysis, in RAW264.7 cells was observed already in 5 min of exposure to and determination of CFUs of recovered bacteria was made. Values rep- CyaA concentrations as low as 10 ng/ml that correspond to toxin resent the means 6SD from $10 independent experiments (n = 10). **p , amounts detected in nasopharyngeal fluids from B. pertussis– 0.001 versus B.p. wild-type control. infected humans and primates (38). The CyaA-catalyzed cAMP production in cells then peaked within 1 h, and intracellular cAMP concentrations remained elevated for 24 h, irrespective of whether from infection of macrophages by wild-type B. pertussis for 24 h the cells were activated by the addition of E.coli LPS (100 ng/ml) at the MOI 10:1, as used in the intracellular survival assay (cf. or not. Hence, signaling of LPS used for induction of NO pro- Fig. 1). It should, however, be noted that some increase of cAMP duction did not interfere with cAMP elevation in macrophage levels was observed also upon prolonged macrophage infection by cells. The peak amounts of cAMP produced upon exposure to 10 the AC enzyme–deficient cyaA-AC2 mutant (Fig. 2D). This was ng/ml of CyaA were fully comparable to cAMP levels resulting most likely due to the action of the pertussis toxin (PT) produced The Journal of Immunology 4905

FIGURE 2. B. pertussis is sensitive to oxidative killing. (A) RAW264.7 macrophages were activated with complement-opsonized zymosan, with comple- ment-opsonized B.p. cyaA-wt, or with unopsonized B.p. cyaA-wt bacteria. Production of ROS was measured over a period of 1 h. Means 6SD from three independent experiments performed in triplicates (n = 9) are given. **p , 0.001 versus negative control. (B) Exponentially growing cultures of B. pertussis

(OD600 = 1) were divided into aliquots to which corresponding volumes of solvent control (ethanol), or of 400 mM DAE-NONOate solution, or of 100 mMH2O2 were added for 2 h. Following plating on BGA, the numbers of viable CFUs were counted upon 5 d of growth. The values are means 6SD from at least five independent experiments performed in duplicates (n = 10). **p , 0.001 versus control. (C) RAW264.7 macrophages were incubated with heat-inactivated (HI) B.p. cyaA-wt,orliveB.p. cyaA-wt, B.p. cyaA-AC2,orB.p. DcyaA cells for 24 h before NO production was measured in culture supernatants. Values represent the means 6SD from three experiments performed in triplicates (n =9).**p , 0.001 versus HI control. (D) RAW264.7 macrophages were treated with 10 ng/ml CyaA, in the presence or absence of 100 ng/ml of LPS for the indicated times, or were infected for 24 h with the indicated B. pertussis strains, as described in Fig. 1. Concentration of cAMP was measured in cell lysates by ELISA. One representative result out of three independent experiments performed in triplicates is shown. (E and F) RAW264.7 macrophages or BMDM, respectively, were treated in the presence or absence of 100 ng/ml of LPS from E. coli with the CyaA variants (CyaA concentrations are given in ng/ml) or with 1 mM db-cAMP for 24 h. 10 mM IBMX was added 1 h prior to db-cAMP addition to cells, and the produced NO was measured by the Griess reagent or by the DAF-FM probe, as indicated in Materials and Methods. Values represent means 6SD from three independent experiments performed in triplicates (n =9).**p , 0.001 versus LPS-treated control. (G) RAW264.7 cells, or BMDM, were treated as in (E). After 24 h, iNOS protein levels were detected by immunoblotting. For experiments with RAW264.7 cells, one representative blot out of three independent experiments is shown (n = 3). For experiments with BMDMs, cells were isolated from three animals, and one blot representative out of three experiments is shown (n =3). Mean 6SD iNOS amounts were determined by densitometric analysis.*p , 0.01, **p , 0.001 versus LPS-treated control. (H) (Figure legend continues) 4906 iNOS EXPRESSION ABLATION BY cAMP SIGNALING CyaA TOXIN by the mutant strain, as PT can deregulate the activity of the versive cAMP-dependent signaling elicited by CyaA does not cellular adenylyl cyclase enzymes through inhibitory ADP- yield any clear . ribosylation of the Gia subunits of the trimeric G proteins (39). The PT-triggered cAMP elevation occurs, however, several hours CyaA blocks iNOS expression by cAMP-dependent activation after PT penetration into cells and requires external stimuli acti- of protein kinase A vating the endogenous AC enzyme, while yielding substantially cAMP activates signaling of protein kinase A (PKA) and of the lower levels of cAMP than what CyaA does produce within exchange protein directly activated by cAMP (Epac) (44). To minutes of contact with phagocytes (40). Hence, likely because of determine whether PKA or Epac activities were involved in sup- the delay in the onset of PT-mediated cAMP elevation, which pression of LPS-induced iNOS expression by cAMP signaling, takes 8–12 h to manifest (40), the PT-mediated elevation of cAMP RAW264.7 cells were preincubated with 10 mM of the phospho- did not prevent the rapid killing of the cyaA-AC2 bacteria inside diesterase inhibitor IBMX and stimulated by LPS in the presence RAW264.7 cells (cf. Fig. 1). of 1 mM cell-permeable cAMP analogs that either activate both As next shown in Fig. 2E, exposure to as little as 1 ng/ml of PKA and Epac, such as db-cAMP, or activate selectively only the CyaA did significantly decrease LPS-induced NO production in PKA (6-Bnz-cAMP) or only Epac (8-CPT-cAMP), respectively. RAW264.7 macrophages, as detected by the nitrite-determining As shown in Fig. 4A, signaling of 1 mM db-cAMP or of the PKA- Griess reaction and corroborated by the use of the fluorescence specific activator 6-Bnz-cAMP provoked as significant an inhibi- probe DAF-FM, which would also detect the nitrite eventually tion of LPS-triggered NO production as the exposure of cells to converted to nitrate. The inhibition of NO production was clearly CyaA toxin (10 ng/ml). In contrast, treatment with the Epac- and specifically due to the elevation of cAMP concentration by the selective activator 8-CPT-cAMP had little effect on LPS- AC enzyme activity of the CyaA toxin, as the LPS-inducible NO triggered NO production. The PKA-activating 6-Bnz-cAMP and production was also inhibited upon RAW264.7 cell exposure to the db-cAMP treatments, but not signaling of 8-CPT-cAMP (Epac membrane-permeable cAMP analog db-cAMP (1 mM). In contrast, activation), also blocked iNOS protein production in RAW264.7 cell exposure to the nonenzymatic CyaA-AC2 toxoid, unable to macrophages, as shown in Fig. 4B. An insignificant effect of the raise cAMP levels even at concentrations as high as 100 ng/ml, had Epac activator was, however, also observed. It remains to be no effect on induction of NO production by LPS. As shown in clarified if it was due to a low level of nonspecific activation of Fig. 2F, CyaA also inhibited NO production in BMDM cells, PKA by 8-CPT-cAMP or whether Epac signaling also contributed confirming that primary murine macrophages were highly suscep- to some extent to suppression of iNOS gene expression. Never- tible to CyaA-provoked inhibition of bactericidal NO induction. theless, as shown in Fig. 4C, activation of PKA by toxin-produced As further revealed by immunodetection with specific Abs cAMP appeared to be the dominant pathway through which the in Fig. 2G, exposure to 1 ng/ml of CyaA already resulted in a CyaA-produced cAMP suppressed iNOS-mediated NO production significant decrease of iNOS production in LPS-stimulated in macrophages. The inhibition could be reversed, albeit not fully, RAW264.7 cells, or primary BMDMs. Moreover, no iNOS pro- by preincubation of cells with the PKA inhibitor H-89 (10 mM). tein was detected in cells exposed to higher toxin concentrations Besides possibly incomplete inhibition of PKA by H-89 at high (10 or 100 ng/ml) or to 1 mM db-cAMP. Indeed, a strong decrease cAMP levels produced by CyaA, the incomplete restoration of of iNOS mRNA level was detected by quantitative PCR in so iNOS expression in H-89–treated cells was likely due to some treated RAW264.7 cells, as shown in Fig. 2H. Hence, treatment nonspecific off-target effects of H-89. As also shown in Fig. 4C, with the cell-permeable cAMP analog fully reproduced the impact the H-89 inhibitor (10 mM) itself caused some observable re- of CyaA toxin action, whereas cell exposure to the CyaA-AC2 duction of LPS-stimulated iNOS expression. H-89 is, indeed, toxoid had no effect. known to be not entirely specific for PKA and to influence ac- Previously, Cheung and coworkers (41) suggested that CyaA tivities of other AGC kinases at the concentrations used in this might interfere with bactericidal NO production through induction study (45). Moreover, the more specific inhibitors of PKA, such as of arginase expression. A cAMP-dependent increase of arginase mPKI or cAMPS, were less potent in reversal of the over- activity was, indeed, observed in CyaA-treated RAW264.7 cells, whelming cAMP-mediated impact of CyaA action on iNOS ex- as shown in Fig. 3A. However, as shown in Fig. 3B, the production pression (data not shown). It remains, therefore, possible that H-89 of NO was not restored in CyaA-treated macrophages either upon may also interfere with induction of iNOS expression at some step selective inhibition of arginases by 100 mM nor-NOHA or at downstream of PKA signaling. Collectively, these results show saturating concentrations of the arginase substrate L-arginine that the CyaA toxin inhibits iNOS expression predominantly (12.8 mM), as shown in Fig. 3C. It can thus be concluded that through cAMP-mediated hijacking of the PKA signaling pathway. rather than by enhancement of arginase activity and degradation cAMP signaling through PKA causes dephosphorylation of AP-1 of the L-arginine substrate, the CyaA-produced cAMP signaling suppressed NO production in LPS-stimulated RAW264.7 cells iNOS expression depends to large extent on the activity of the through inhibition of iNOS gene expression. transcription factor AP-1 (20, 21). Therefore, we examined Intriguingly, induction of arginase would be indicative of whether cAMP signaling of CyaA affected the regulatory phos- macrophage polarization toward the M2 phenotype; however, in- phorylation of the c-Fos subunit of AP-1. As shown in Fig. 5, no duction of the M1 phenotype–associated cyclooxygenase-2 en- activating phosphorylation of c-Fos was detected in untreated zyme activity in murine macrophages by CyaA was previously cells, whereas activation of RAW264.7 macrophages by LPS reported (42) and we did not observe any alteration in the pro- resulted in hyperphosphorylation of c-Fos, detected as two bands tein levels of the Mox macrophage markers NRF2 and HO-1 in P–c-Fos immunoblots. Exposure of LPS-activated RAW264.7 (Supplemental Fig. 1) (43). It appears, therefore, that the sub- cells to 10 ng/ml of CyaA then provoked a near-complete disap-

RAW264.7 cells were treated as in (E). After 24 h, iNOS mRNA was quantified by quantitative RT-PCR. Relative iNOS mRNA levels normalized to iNOS mRNA extracted from LPS-treated control cells are expressed as fold change. Means 6SD of three independent experiments performed in duplicates (n =6) are given. **p , 0.001 versus LPS-treated control. N.C., negative control. The Journal of Immunology 4907

FIGURE 4. PKA activation by CyaA inhibits NO production (A) RAW264.7 cells were preincubated for 1 h in the presence or absence of FIGURE 3. CyaA induces increase in arginase activity. (A) RAW264.7 10 mM IBMX before LPS (100 ng/ml) and CyaA (10 ng/ml), 8-CPT-cAMP, cells were incubated with CyaA (ng/ml) and 100 ng/ml LPS for 24 h before 6-Bnz-cAMP, or db-cAMP (1 mM) were added, respectively. NO production total arginase activity was measured in cell lysates. The values are means 6 was detected after 24 h. Values represent the means 6SD from three inde- SD from three independent experiments performed in triplicates (n =9). pendent experiments performed in triplicates (n = 9). (B) RAW264.7 cells (B) RAW264.7 cells were preincubated for 1 h in the presence or absence were treated as in (A). CyaA was added to the cells concomitantly with LPS, of 100 mM nor-NOHA prior to activation with 100 ng/ml of LPS and/or and iNOS was detected by immunoblotting. One representative blot out of addition of 10 ng/ml CyaA for 24 h. Values represent means 6SD from three three independent experiments is shown. (C) RAW264.7 cells were pre- independent experiments performed in triplicates (n = 9). (C) RAW264.7 incubated for 1 h, with or without the PKA inhibitor H-89 (10 mM), prior to cells were incubated with 100 ng/ml of LPS and 10 ng/ml of CyaA in the stimulation with 100 ng/ml LPS and 10 ng/ml of CyaA. After 24 h, the levels presence of the indicated concentration of L-arginine for 24 h. Values rep- of iNOS protein were detected by immunoblotting. One representative blot resent the means 6SD from three independent experiments performed in out of three independent experiments is shown. The means 6SD of densi- triplicates (n =9).**p , 0.001 versus LPS-treated control. tometric analysis from three experiments are shown (n =3).**p , 0.001 versus LPS-treated control, †p , 0.01 versus each other. pearance of the upper hyperphosphorylated c-Fos isoform, whereas the hypophosphorylated lower band of P–c-Fos was tion of c-Fos, whereas the action of CyaA was almost fully detected at comparable levels as in LPS-activated and mock- reproduced by the PKA-activating analogs db-cAMP (1 mM) and treated cells. In contrast, specific activation of Epac with 1 mM 6-Bnz-cAMP (1 mM), respectively. Compared with CyaA action, 8-CPT-cAMP did not affect TLR-(LPS)–triggered phosphoryla- the activation of PKA by the cAMP analogs yielded a less com- 4908 iNOS EXPRESSION ABLATION BY cAMP SIGNALING CyaA TOXIN

amounts of the SHP-1 protein were produced in cells repeatedly transfected with SHP-1–specific siRNA, in which exposure to CyaA provoked only a mild decrease of the LPS-induced NO pro- duction. Moreover, in the absence of CyaA, the cells with SHP-1 knocked down were reproducibly responding to activation by LPS with a ∼50% higher production of NO than control cells or than cells transfected by nontargeting or SHP-2–specific siRNA. It can thus be concluded that it was specifically the SHP-1 tyrosine phosphatase that translated the PKA-mediated signaling of toxin-produced cAMP into inhibition of iNOS expression and loss of NO production. To corroborate that cAMP/PKA signaling activated the SHP-1 phosphatase, its activity was measured in the material pulled down by SHP-1–specific Ab from lysates of CyaA-treated cells. As documented in Fig. 6E, ∼2-fold higher SHP-1 phosphatase activity was, indeed, detected in immunoprecipitates from LPS- activated cells exposed to 10 ng/ml of CyaA for 10 or 30 min, compared with lysates from cells exposed to LPS only. Intrigu- ingly, this activation of SHP-1 was only transient, and at 60 min of cell exposure to CyaA the levels of SHP-1 activity in cells de- creased back to the steady-state levels. To gain more insight into FIGURE 5. CyaA provokes dephosphorylation of the P–c-Fos subunit the possible mechanism of SHP-1 activity regulation by signaling of AP-1 via PKA-mediated signaling. RAW264.7 macrophages were of cAMP, we examined the phosphorylation of SHP-1 on residues preincubated with or without 10 mM IBMX for 1 h before addition of known to be involved in upregulation of the specific phosphatase LPS (100 ng/ml) and/or CyaA (10 ng/ml), or db-cAMP, 6-Bnz-cAMP, activity. As shown in Fig. 6F, at 30 min of cell exposure to toxin, or 8-CPT-cAMP (1 mM), respectively. After 24 h, the active hyper- no alteration of the activatory phosphorylation of Y564 and Y536 phosphorylated P–c-Fos form (slower migrating upper band) and the in- active hypophosphorylated form (lower band) were detected by immu- residues was detected, whereas decrease of the inhibitory phos- noblotting and are indicated by arrows, respectively. **p , 0.001 versus phorylation on S591 in 30 min was reproducibly observed. The LPS treated control, †p , 0.01, ‡p , 0.001 versus IBMX control. exact mechanism of cAMP-dependent regulation of SHP-1 ac- tivity will, however, require further analysis, as at 60 min of CyaA action, at which the SHP-1 activity was already back to the initial plete dephosphorylation of the hyperphosphorylated P–c-Fos form level, no increase of the inhibitory phosphorylation of the S591 (upper band), whereas it increased the relative detected amounts residue was observed (data not shown). of the hypophosphorylated P–c-Fos isoform. Altogether, these Inhibition of SHP phosphatase activity abrogates B. pertussis data suggest that activation of PKA signaling by CyaA-produced survival in macrophages cAMP provokes dephosphorylation of the hyperphosphorylated P–c-Fos subunit of the AP-1 transcription factor that plays a cru- It was next important to examine whether SHP-1 activation by CyaA cial role in iNOS gene expression in LPS-activated macrophages. accounted for the extended survival of unopsonized B. pertussis inside mouse macrophages. As shown in Fig. 7A, treatment with the CyaA/cAMP blocks iNOS expression through activation of the SHP-1/2 inhibitor NSC87877 (500 nM) enhanced by about 2- to 3- SHP-1 phosphatase fold the total number of B. pertussis cells associated with and in- No protein phosphatases directly activated by cAMP signaling have ternalized by the RAW264.7 macrophages. This finding matches as yet been identified, but the tyrosine phosphatase SHP-1 is known with an earlier observation that inhibition of the SHP-1 phosphatase to be involved in the regulation of numerous receptor signaling enhances the phagocytic activity of macrophages (47, 48). In sharp pathways in leukocytes and it was previously implicated also in the contrast, the presence of the SHP inhibitor completely abrogated the regulation of iNOS expression (46). Therefore, we examined capacity of B. pertussis to survive within macrophages, as docu- whether SHP-1 activity played a role in cAMP/PKA-regulated mented in Fig. 7B. Essentially no viable bacteria could be recovered dephosphorylation of c-Fos. As shown in Fig. 6A, the CyaA- from macrophages incubated with NSC87877 already at 24 h post provoked dephosphorylation of the hyperphosphorylated (upper infection, whereas in the absence of the SHP inhibitor, the bacterial band) form of P–c-Fos was partly inhibited upon pretreatment of viability inside cells was decreasing gradually over 72 h. Collec- LPS-activated cells with NSC87877 (500 nM), an inhibitor of the tively, these data show that activation of the SHP-1 phosphatase by SHP-1 and SHP-2 phosphatases. Moreover, increased amounts of CyaA-elicited signaling of cAMP enables unopsonized B. pertussis the hypophosphorylated (faster migrating) isoform of c-Fos were bacteria to extend their survival inside macrophage cells (Fig. 8). observed in NSC87877-treated cells exposed to CyaA, indicating that SHP-1/2 activity may be regulated by cAMP signaling. Discussion Moreover, pretreatment with NSC87877 also restored in part the We show that cAMP-activated signaling through PKA yields ac- expression of iNOS and production of NO in toxin-treated tivation of the tyrosine phosphatase SHP-1. Hijacking of this RAW264.7 cells (Fig. 6B, 6C). Therefore, SHP-1 and SHP-2 novel cAMP-regulated signaling pathway by the adenylate cyclase expression was selectively knocked down by siRNA, and toxin then enables B. pertussis to evade rapid NO-mediated killing cAMP-induced inhibition of iNOS expression was examined. As inside macrophage cells. As summarized in the model shown in shown in Fig. 6D, CyaA fully suppressed production of NO in Fig. 8, the CyaA-cAMP–triggered activation of SHP-1 leads to RAW264.7 macrophages transfected with nontargeting siRNA or dephosphorylation of the P–c-Fos subunit of the transcription with SHP-2–specific siRNA. Such cells, indeed, expressed normal factor AP-1 and results in loss of TLR-induced expression of the levels of the SHP-1 protein, as detected in immunoblots of iNOS enzyme and loss of production of bactericidal levels of NO. cell lysates with SHP-1–specific Abs. In contrast, no detectable This novel mechanism adds to the broad spectrum of immuno- The Journal of Immunology 4909

FIGURE 6. SHP-1 mediates CyaA-provoked inhibition of NO production. (A) RAW264.7 macrophages were pretreated, with or without the 500 nM NSC87877 inhibitor of SHP-1/2 phosphatases, for 1 h before 100 ng/ml LPS and 10 ng/ml of CyaA were added for 24 h and the phosphorylation status of P–c-Fos was analyzed by immunoblotting of cell lysates. One representative blot out of three independent experiments is shown. The mean values 6SD from densitometric analysis are shown (n = 3). (B) RAW264.7 macrophages were treated as above, and after 24 h the production of NO was detected. Values represent the means 6SD from three independent experiments performed in triplicates (n = 9). (C) RAW264.7 cells were treated as in (A), and after 24 h, the levels of iNOS protein were detected by immunoblotting. One representative blot out of three independent experiments is shown. The means 6SD from densitometric analysis of three experiments is shown (n = 3). (D) RAW264.7 cells were transfected with nontargeting, SHP-1–, or SHP-2–specific siRNA, respectively. LPS (100 ng/ml) or CyaA (10 ng/ml) or both were added, and incubation was continued for 24 h. The production of SHP proteins was analyzed by immunoblotting. Values represent the means 6SD from three independent experiments performed in triplicates (n = 9). (E) RAW264.7 or BMDM cells were treated with 10 ng/ml of CyaA for the indicated times in the presence of 100 ng/ml LPS. SHP-1 was immunoprecipitated, and its phosphatase activity was measured and expressed as relative phosphatase activity. Mean values 6SD from three independent experiments are given. (F) RAW264.7 cells were treated with the indicated CyaA variants for 30 min, and the phosphorylation status of SHP-1 was analyzed by immunoblotting. One representative blot out of three independent experiments is shown. Mean values 6SD from densitometric analysis are given (n = 3). **p , 0.001 versus LPS-treated control, †p , 0.01 versus each other, ‡p , 0.001 versus NSC87877 control. subversive outcomes of CyaA-catalyzed synthesis of cAMP in of iNOS gene expression in rat vascular smooth muscle and host phagocytes. mesangial cells, elevation of cAMP in other rodent cells yielded Somewhat controversial results were previously reported on the reduction of iNOS expression (49–51). It has, indeed, been pre- role of cAMP in the regulation of iNOS expression in different cell viously observed that modulation of cAMP levels in RAW264.7 types and tissues. Although cAMP was found to be a strong inducer macrophages may impact on LPS-induced expression of iNOS and 4910 iNOS EXPRESSION ABLATION BY cAMP SIGNALING CyaA TOXIN

FIGURE 8. Scheme of CyaA-provoked signalization that blocks iNOS expression and NO production in macrophage cells. After binding of CyaA and delivery of the AC domain into the cytosol of macrophages, the toxin is activated by binding of calmodulin, and its enzymatic AC domain cat- alyzes uncontrolled conversion of cytosolic ATP into cAMP, thus acti- vating PKA signaling. By an as yet uncharacterized mechanism, this leads to the enhancement of activity of the SHP-1 phosphatase, which yields dephosphorylation and loss of activity of the c-Fos subunit of the tran- scription factor AP-1. Loss of P–c-Fos function then prevents iNOS ex- pression and NO production in TLR-activated macrophages.

LPS-activated macrophage cells (cf. Fig. 2B). Indeed, PT action takes hours to translate into cAMP increase in cells, and it depends on additional receptor signaling–mediated activation of the en- dogenous AC enzyme. It is, therefore, plausible to propose that the steep increase of cAMP concentration, accomplished early upon bacterial contact with phagocytes by the CyaA toxin, was required for prevention of TLR-activated iNOS expression. This activity is likely made possible by two unique features of CyaA. First, CyaA exhibits an extraordinarily rapid mechanism of target cell pene- tration. Translocation of the AC domain into cells occurs by an -independent mechanism directly across the cytoplas- mic membrane of cells, and it proceeds with a half-time of only ∼30 s (55–58). Second, the calmodulin-activated catalytic domain FIGURE 7. SHP-1 inhibition abrogates survival of B. pertussis inside of CyaA possesses an extremely high specific AC enzyme activity, macrophages. Prior to infection with GFP-expressing B.p. cyaA-wt, the with a turnover number of ∼2000 s21 (58). As a result, CyaA RAW264.7 macrophages were preincubated for 1 h, with or without SHP- concentrations as low as those used in this study (10 ng/ml), and A 1/2 inhibitor NSC87877 (500 nM). ( ) After 1 h, the cells were washed, detected in mucosal fluids of infected infants and experimentally fixed, and analyzed as described in the legend to Fig. 1A. Representative challenged baboons (38), are not only sufficient for ablation of images from at least five independent experiments are shown. Scale bar, 5 mm. (B) The RAW264.7 cells were treated as described in the legend to superoxide production and inhibition of neutrophil extracellular Fig. 1B. Cell lysates were plated on BGA, and bacterial CFUs were trap release by host neutrophils (8–10) but also provoke inhibition counted after 5 d. Values represent the means 6SD from $10 independent of bactericidal NO production in macrophages. This finding experiments (n = 10). **p , 0.01 versus control. explains why CyaA expression is crucial for the extended survival of unopsonized B. pertussis bacteria in macrophages. These results further highlight the unique role played by CyaA in sub- NO production (52, 53). We show that as little of the CyaA toxin version of innate immunity mechanisms, explaining the impor- as 1 ng/ml does produce sufficient amounts of cAMP to cause tance of the role played by CyaA in the early phases of bacterial a strong reduction of iNOS expression in LPS-activated macro- colonization of host airways (2–4, 59). phage cells (cf. Fig. 2G). In agreement with a previous report Although major differences exist in iNOS expression regulation in (54) and in contrast to CyaA, the PT produced by the mutant humans and mice, the expression of bactericidal iNOS in human cyaA-AC2 bacteria was unable to increase cAMP concentrations phagocytes and the role of NO in innate immunity in humans are now through deregulation of the endogenous AC to a level that would well established (60). iNOS expression in human macrophages, provoke a block of induction of bactericidal NO production in however, requires, besides activation by LPS, a simultaneous in- The Journal of Immunology 4911 volvement of several cytokine signals (e.g., IFN-g, TNF-a, and IL- Fig. 6E) remains unknown but correlates in time with the cAMP- 1b). This requirement makes the deciphering of signaling path- provoked transient inhibition of the small GTPase RhoA activity ways leading to iNOS expression in human phagocytes difficult. in CyaA-treated murine macrophages, which was maximal at 30 In contrast, LPS alone is sufficient for strong iNOS expression min from toxin addition (11). It will therefore be of interest to in murine macrophages (21). Therefore, mouse cells like the determine if the same regulatory circuit is involved in modulation RAW264.7 macrophages are preferentially used for analysis of of both RhoA and SHP-1 activities by toxin-produced cAMP signaling that leads to bactericidal iNOS expression in phagocytes signaling. PKA, indeed, likely exerts its effect on SHP-1 activity (52, 53, 61). Using these cells, we reveal in this study the prominent by an indirect mechanism, as no phosphorylation consensus se- role played by cAMP signaling in the regulation of activity of the quence for PKA (-RRXS/T- where X stays for any ) is transcription factors involved in iNOS expression in phagocytes. found in the SHP-1 sequence, in contrast to SHP-2. We then This remains scarcely documented in the literature. Activation of conclude that it is selectively the cAMP-activated PKA signaling the NF-kB transcription factor in human keratinocytes by db-cAMP that leads to activation of SHP-1. This is based on the observation or forskolin has previously been reported (62), which accords with that the Epac-specific cell-permeable activator 8-CPT-cAMP the observed triggering of NF-kB translocation into cell nuclei by exerted little, if any, effect on c-Fos phosphorylation, iNOS ex- the cAMP-elevating activity of CyaA (cf. Supplemental Fig. 2). pression, and NO production. In contrast, the PKA-specific acti- However, NF-kB activation alone does not lead to iNOS expression, vator 6-Bnz-cAMP mimicked to a large extent the impact of CyaA and Hickey and coworkers (63) observed previously that cAMP action on cells. Hence, cAMP-dependent activation of PKA production by CyaA caused a decrease of levels of both IRF1 and clearly played a dominant role in CyaA-provoked suppression of IRF8 transcription factors. Similarly, cAMP elevation by cholera iNOS expression in macrophages. However, it cannot be defini- toxin was found to cause inhibition of IRF1–IRF8 interaction (64). tively concluded at present that only PKA signaling was involved Furthermore, IRF1 and STAT1 activities were also found to be and Epac did not play any role in the control of TLR-induced inhibited upon VPAC1-mediated PKA activation by a mechanism iNOS expression. PKA inhibition with H-89 did not fully re- independent of suppressor of cytokine signaling 1/3 (65). Hence, the verse the impact of CyaA action on iNOS expression. Moreover, previous observations that IRF1 and STAT1 activities are regulated H-89 is known to exert some off-target effects (45), and H-89 on by cAMP agree with our results showing that CyaA activity con- its own was found to partially impair iNOS expression (cf. comitantly decreased STAT1 phosphorylation and IRF1 levels (cf. Fig. 4C). It is thus possible that owing to high levels of cAMP Supplemental Fig. 2D). Although STAT1 activation was previously produced in cells by CyaA, the H-89 inhibitor at the usual con- found to be inhibited via cAMP/PKA signaling–triggered activation centration of 10 mM was just unable to completely inhibit the fully of SHP-2 (66), we observed in this study a highly SHP-1–specific activated PKA signaling. effect of cAMP increase on LPS-triggered iNOS expression in In summary, modulation of iNOS expression by a cAMP/PKA- murine macrophage cells. Moreover, inhibition of STAT1 by CyaA regulated SHP-1 activity, as reported in this article, indicates that occurred later than inhibition of iNOS expression (cf. Supplemental SHP-1 is playing a rather central role in the control of iNOS Fig. 2E), showing that STAT1 inhibition did not account for the expression and bacterial survival in phagocytes. Along this line, inhibition of iNOS expression. In contrast, the effects of CyaA- selective silencing of SHP-1 expression led to restoration of provoked cAMP signaling on phosphorylation of the c-Fos sub- NO production in CyaA toxin–treated RAW264.7 cells, and in- unit of AP-1 correlated in time with the observed inhibition of iNOS hibition of SHP-1/2 phosphatases by NSC87877 provoked a expression. steep decrease of B. pertussis survival inside phagocytes. These The c-Fos subunit of the transcription factor AP-1 was, indeed, observations reveal an as yet undescribed cAMP/PKA-regulated shown to undergo hyperphosphorylation at multiple sites upon signaling mechanism manipulated by the CyaA toxin that the macrophage activation by LPS. This increases its activity (67) and whooping cough agent B. pertussis uses for hijacking of the accounts for the appearance of several P–c-Fos bands, or a smear tyrosine phosphatase SHP-1. These findings open the way for of hyperphosphorylated P–c-Fos, in immunoblots with P–c-Fos– development and testing of novel SHP-1–specific inhibitors as specific Abs. The analyzed CyaA-produced cAMP signaling then potential drugs for treating pertussis in the early phases of caused dephosphorylation of the hyperphosphorylated P–c-Fos B. pertussis . into the hypophosphorylated c-Fos. This finding suggests that P–c-Fos was, indeed, the primary target of PKA activation–de- Acknowledgments pendent signaling in the course of CyaA action and that de- We thank Scott Stibitz for the generous gift of the pSS4245 allelic ex- phosphorylation of P–c-Fos accounted for inhibition of iNOS change vector; Tomas Brdicka for L929 CMG cells; Branislav Vecerek for expression. the plasmid pBBRcyaGFP and the anti-B.p. Ab; and Sona Kozubova, Intriguingly, manipulation of iNOS expression through a very Hana Lukeova, and Iva Marsikova for excellent technical help. different mechanism of protease-dependent activation of SHP-1 and a subsequent modulation of c-Fos activity has previously Disclosures been observed in mouse macrophages infected with Leishmania The authors have no financial conflicts of interest. donovani (46, 68, 69). Moreover, LPS-triggered iNOS expression in murine macrophages could previously be blocked by activation of SHP-1 (70). To the best of our knowledge, however, we de- References 1. Glaser, P., A. Danchin, D. Ladant, O. Barzu, and A. Ullmann. 1988. Bordetella scribe in this article for the first time the activation of SHP-1 by pertussis adenylate cyclase: the gene and the protein. Tokai J. Exp. Clin. Med. 13 a mechanism dependent on cAMP elevation and PKA activity. (Suppl): 239–252. Intriguingly, although the cAMP levels increased with time of 2. Goodwin, M. S., and A. A. Weiss. 1990. Adenylate cyclase toxin is critical for colonization and pertussis toxin is critical for lethal infection by Bordetella toxin action (cf. Fig. 2D), and presumably the PKA activity pertussis in infant mice. Infect. Immun. 58: 3445–3447. remained high over the duration of cell exposure to CyaA, the 3. Khelef, N., H. Sakamoto, and N. Guiso. 1992. Both adenylate cyclase and he- activation of SHP-1 by toxin-produced cAMP signaling was only molytic activities are required by Bordetella pertussis to initiate infection. Microb. Pathog. 12: 227–235. transient. The mechanism underlying downregulation of SHP-1 4. Vojtova, J., J. Kamanova, and P. Sebo. 2006. Bordetella adenylate cyclase toxin: activity after prolonged cell exposure to CyaA (60 min; cf. a swift saboteur of host defense. Curr. Opin. Microbiol. 9: 69–75. 4912 iNOS EXPRESSION ABLATION BY cAMP SIGNALING CyaA TOXIN

5. Weiss, A. A., E. L. Hewlett, G. A. Myers, and S. Falkow. 1984. Pertussis toxin 34. Noda, T., and F. Amano. 1997. Differences in nitric oxide synthase activity in and extracytoplasmic adenylate cyclase as virulence factors of Bordetella per- a macrophage-like cell line, RAW264.7 cells, treated with lipopolysaccharide tussis. J. Infect. Dis. 150: 219–222. (LPS) in the presence or absence of interferon-gamma (IFN-gamma): possible 6. Wolff, J., G. H. Cook, A. R. Goldhammer, and S. A. Berkowitz. 1980. Cal- heterogeneity of iNOS activity. J. Biochem. 121: 38–46. modulin activates prokaryotic adenylate cyclase. Proc. Natl. Acad. Sci. USA 77: 35. Dyet, K., and J. Moir. 2006. Effect of combined oxidative and nitrosative stress 3841–3844. on Neisseria meningitidis. Biochem. Soc. Trans. 34: 197–199. 7. Guermonprez, P., N. Khelef, E. Blouin, P. Rieu, P. Ricciardi-Castagnoli, 36. Potter, A. J., S. P. Kidd, J. L. Edwards, M. L. Falsetta, M. A. Apicella, N. Guiso, D. Ladant, and C. Leclerc. 2001. The adenylate cyclase toxin of M. P. Jennings, and A. G. McEwan. 2009. Thioredoxin reductase is essential for Bordetella pertussis binds to target cells via the alpha(M)beta(2) integrin protection of Neisseria gonorrhoeae against killing by nitric oxide and for (CD11b/CD18). J. Exp. Med. 193: 1035–1044. bacterial growth during interaction with cervical epithelial cells. J. Infect. Dis. 8. Confer, D. L., and J. W. Eaton. 1982. impotence caused by an invasive 199: 227–235. bacterial adenylate cyclase. Science 217: 948–950. 37. Mandell, G. L., and E. W. Hook. 1969. Leukocyte bactericidal activity in 9. Eby, J. C., M. C. Gray, and E. L. Hewlett. 2014. Cyclic AMP-mediated sup- chronic granulomatous disease: correlation of bacterial hydrogen peroxide pression of neutrophil extracellular trap formation and by the Borde- production and susceptibility to intracellular killing. J. Bacteriol. 100: 531– tella pertussis adenylate cyclase toxin. Infect. Immun. 82: 5256–5269. 532. 10. Pearson, R. D., P. Symes, M. Conboy, A. A. Weiss, and E. L. Hewlett. 1987. 38. Eby, J. C., M. C. Gray, J. M. Warfel, C. D. Paddock, T. F. Jones, S. R. Day, Inhibition of oxidative responses by Bordetella pertussis adenylate J. Bowden, M. D. Poulter, G. M. Donato, T. J. Merkel, and E. L. Hewlett. 2013. cyclase toxin. J. Immunol. 139: 2749–2754. Quantification of the adenylate cyclase toxin of Bordetella pertussis in vitro and 11. Kamanova, J., O. Kofronova, J. Masin, H. Genth, J. Vojtova, I. Linhartova, during respiratory infection. Infect. Immun. 81: 1390–1398. O. Benada, I. Just, and P. Sebo. 2008. Adenylate cyclase toxin subverts 39. Katada, T., and M. Ui. 1982. Direct modification of the membrane adenylate phagocyte function by RhoA inhibition and unproductive ruffling. J. Immunol. cyclase system by islet-activating protein due to ADP-ribosylation of a mem- 181: 5587–5597. brane protein. Proc. Natl. Acad. Sci. USA 79: 3129–3133. 12. Weingart, C. L., and A. A. Weiss. 2000. Bordetella pertussis virulence factors 40. Bokoch, G. M., T. Katada, J. K. Northup, E. L. Hewlett, and A. G. Gilman. 1983. affect phagocytosis by human neutrophils. Infect. Immun. 68: 1735–1739. Identification of the predominant substrate for ADP-ribosylation by islet acti- 13. Adkins, I., J. Kamanova, A. Kocourkova, M. Svedova, J. Tomala, H. Janova, vating protein. J. Biol. Chem. 258: 2072–2075. J. Masin, B. Chladkova, L. Bumba, M. Kovar, et al. 2014. Bordetella adenylate 41. Cheung, G. Y., P. Dickinson, G. Sing, M. Craigon, P. Ghazal, R. Parton, and cyclase toxin differentially modulates toll-like receptor-stimulated activation, mi- J. G. Coote. 2008. Transcriptional responses of murine macrophages to the gration and stimulatory capacity of dendritic cells. PLoS ONE 9: e104064. adenylate cyclase toxin of Bordetella pertussis. Microb. Pathog. 44: 61–70. 14. Friedman, R. L., K. Nordensson, L. Wilson, E. T. Akporiaye, and D. E. Yocum. 42. Perkins, D. J., M. C. Gray, E. L. Hewlett, and S. N. Vogel. 2007. Bordetella 1992. Uptake and intracellular survival of Bordetella pertussis in human mac- pertussis adenylate cyclase toxin (ACT) induces cyclooxygenase-2 (COX-2) in rophages. Infect. Immun. 60: 4578–4585. murine macrophages and is facilitated by ACT interaction with CD11b/CD18 15. Lamberti, Y. A., J. A. Hayes, M. L. Perez Vidakovics, E. T. Harvill, and (Mac-1). Mol. Microbiol. 66: 1003–1015. M. E. Rodriguez. 2010. Intracellular trafficking of Bordetella pertussis in human 43. Kadl, A., A. K. Meher, P. R. Sharma, M. Y. Lee, A. C. Doran, S. R. Johnstone, macrophages. Infect. Immun. 78: 907–913. M. R. Elliott, F. Gruber, J. Han, W. Chen, et al. 2010. Identification of a novel 16. Muhl,€ H., M. Bachmann, and J. Pfeilschifter. 2011. Inducible NO synthase and macrophage phenotype that develops in response to atherogenic phospholipids antibacterial host defence in times of Th17/Th22/T22 immunity. Cell. Microbiol. via Nrf2. Circ. Res. 107: 737–746. 13: 340–348. 44. Peters-Golden, M. 2009. Putting on the brakes: cyclic AMP as a multipronged 17. Canthaboo, C., D. Xing, X. Q. Wei, and M. J. Corbel. 2002. Investigation of role controller of macrophage function. Sci. Signal. 2: pe37. of nitric oxide in protection from Bordetella pertussis respiratory challenge. 45. Bain, J., L. Plater, M. Elliott, N. Shpiro, C. J. Hastie, H. McLauchlan, Infect. Immun. 70: 679–684. I. Klevernic, J. S. Arthur, D. R. Alessi, and P. Cohen. 2007. The selectivity of 18. MacMicking, J., Q. W. Xie, and C. Nathan. 1997. Nitric oxide and macrophage protein kinase inhibitors: a further update. Biochem. J. 408: 297–315. function. Annu. Rev. Immunol. 15: 323–350. 46. Blanchette, J., I. Abu-Dayyeh, K. Hassani, L. Whitcombe, and M. Olivier. 2009. 19. Bian, K., F. Ghassemi, A. Sotolongo, A. Siu, L. Shauger, A. Kots, and F. Murad. Regulation of macrophage nitric oxide production by the protein tyrosine 2012. NOS-2 signaling and cancer therapy. IUBMB Life 64: 676–683. phosphatase Src homology 2 domain phosphotyrosine phosphatase 1 (SHP-1). 20. Bonaparte, K. L., C. A. Hudson, C. Wu, and P. T. Massa. 2006. Inverse regu- Immunology 127: 123–133. lation of inducible nitric oxide synthase (iNOS) and arginase I by the protein 47. Kant, A. M., P. De, X. Peng, T. Yi, D. J. Rawlings, J. S. Kim, and D. L. Durden. tyrosine phosphatase SHP-1 in CNS glia. Glia 53: 827–835. 2002. SHP-1 regulates Fcgamma receptor-mediated phagocytosis and the acti- 21. Pautz, A., J. Art, S. Hahn, S. Nowag, C. Voss, and H. Kleinert. 2010. Regulation vation of RAC. Blood 100: 1852–1859. of the expression of inducible nitric oxide synthase. Nitric Oxide 23: 75–93. 48. Lee, S. M., E. J. Kim, K. Suk, and W. H. Lee. 2011. Synthetic peptides con- 22. Forget, G., D. J. Gregory, L. A. Whitcombe, and M. Olivier. 2006. Role of host taining ITIM-like sequences of IREM-1 inhibit BAFF-mediated regulation of protein tyrosine phosphatase SHP-1 in Leishmania donovani-induced inhibition interleukin-8 expression and phagocytosis through SHP-1 and/or PI3K. Immu- of nitric oxide production. Infect. Immun. 74: 6272–6279. nology 134: 224–233. 23. Massa, P. T., and C. Wu. 1996. The role of protein tyrosine phosphatase SHP-1 in the 49. Koide, M., Y. Kawahara, I. Nakayama, T. Tsuda, and M. Yokoyama. 1993. regulation of IFN-gamma signaling in neural cells. J. Immunol. 157: 5139–5144. Cyclic AMP-elevating agents induce an inducible type of nitric oxide synthase in 24. Osickova, A., J. Masin, C. Fayolle, J. Krusek, M. Basler, E. Pospisilova, C. Leclerc, cultured vascular smooth muscle cells. Synergism with the induction elicited by R. Osicka, and P. Sebo. 2010. Adenylate cyclase toxin translocates across target inflammatory cytokines. J. Biol. Chem. 268: 24959–24966. without forming a pore. Mol. Microbiol. 75: 1550–1562. 50. Messmer, U. K., and B. Brune.€ 1994. Modulation of inducible nitric oxide 25. Buasri, W., A. Impoolsup, C. Boonchird, A. Luengchaichawange, synthase in RINm5F cells. Cell. Signal. 6: 17–24. P. Prompiboon, J. Petre, and W. Panbangred. 2012. Construction of Bordetella 51. Pang, L., and J. R. Hoult. 1997. Repression of inducible nitric oxide synthase and pertussis strains with enhanced production of genetically-inactivated Pertussis cyclooxygenase-2 by prostaglandin E2 and other cyclic AMP stimulants in J774 Toxin and Pertactin by unmarked allelic exchange. BMC Microbiol. 12: 61. macrophages. Biochem. Pharmacol. 53: 493–500. 26. Weber, C., C. Boursaux-Eude, G. Coralie, V. Caro, and N. Guiso. 2001. Poly- 52. Chang, S. Y., D. B. Kim, G. R. Ryu, S. H. Ko, I. K. Jeong, Y. B. Ahn, Y. H. Jo, morphism of Bordetella pertussis isolates circulating for the last 10 years in and M. J. Kim. 2013. Exendin-4 inhibits iNOS expression at the protein level in France, where a single effective whole-cell vaccine has been used for more than LPS-stimulated Raw264.7 macrophage by the activation of cAMP/PKA path- 30 years. J. Clin. Microbiol. 39: 4396–4403. way. J. Cell. Biochem. 114: 844–853. 27. Weischenfeldt, J., and B. Porse. 2008. Bone marrow-derived macrophages 53. Koide, N., T. Sugiyama, I. Mori, M. M. Mu, T. Yoshida, and T. Yokochi. 2003. (BMM): isolation and applications. CSH Protoc. 2008:pdb.prot5080. C2-ceramide inhibits LPS-induced nitric oxide production in RAW 264.7 mac- 28. Anderson, K. E., K. B. Boyle, K. Davidson, T. A. Chessa, S. Kulkarni, rophage cells through down-regulating the activation of Akt. J. Endotoxin Res. 9: G. E. Jarvis, A. Sindrilaru, K. Scharffetter-Kochanek, O. Rausch, L. R. Stephens, 85–90. and P. T. Hawkins. 2008. CD18-dependent activation of the neutrophil NADPH 54. Kuo, P. C., R. A. Schroeder, and S. T. Bartlett. 1997. Endotoxin-mediated oxidase during phagocytosis of Escherichia coli or Staphylococcus aureus is synthesis of nitric oxide is dependent on Gq protein signal transduction. Sur- regulated by class III but not class I or II PI3Ks. Blood 112: 5202–5211. gery 122: 394–402; discussion 402–393. 29. Marletta, M. A., P. S. Yoon, R. Iyengar, C. D. Leaf, and J. S. Wishnok. 1988. 55. Bumba, L., J. Masin, R. Fiser, and P. Sebo. 2010. Bordetella adenylate cyclase Macrophage oxidation of L-arginine to nitrite and nitrate: nitric oxide is an in- toxin mobilizes its beta2 integrin receptor into lipid rafts to accomplish trans- termediate. Biochemistry 27: 8706–8711. location across target cell membrane in two steps. PLoS Pathog. 6: e1000901. 30. Cortese-Krott, M. M., A. Rodriguez-Mateos, G. G. Kuhnle, G. Brown, 56. Gordon, V. M., S. H. Leppla, and E. L. Hewlett. 1988. Inhibitors of receptor- M. Feelisch, and M. Kelm. 2012. A multilevel analytical approach for detection mediated endocytosis block the entry of Bacillus anthracis adenylate cyclase and visualization of intracellular NO production and nitrosation events using toxin but not that of Bordetella pertussis adenylate cyclase toxin. Infect. Immun. diaminofluoresceins. Free Radic. Biol. Med. 53: 2146–2158. 56: 1066–1069. 31. Corraliza, I. M., M. L. Campo, G. Soler, and M. Modolell. 1994. Determination of 57. Gordon, V. M., W. W. Young, Jr., S. M. Lechler, M. C. Gray, S. H. Leppla, and arginase activity in macrophages: a micromethod. J. Immunol. Methods 174: 231–235. E. L. Hewlett. 1989. Adenylate cyclase from Bacillus anthracis and 32. Kubista, M., J. M. Andrade, M. Bengtsson, A. Forootan, J. Jona´k, K. Lind, Bordetella pertussis. Different processes for interaction with and entry into target R. Sindelka, R. Sjo¨back, B. Sjo¨green, L. Stro¨mbom, et al. 2006. The real-time cells. J. Biol. Chem. 264: 14792–14796. polymerase chain reaction. Mol. Aspects Med. 27: 95–125. 58. Rogel, A., and E. Hanski. 1992. Distinct steps in the penetration of adenylate 33. Stainer, D. W., and M. J. Scholte. 1970. A simple chemically defined medium for cyclase toxin of Bordetella pertussis into sheep erythrocytes. Translocation of the production of phase I Bordetella pertussis. J. Gen. Microbiol. 63: 211–220. the toxin across the membrane. J. Biol. Chem. 267: 22599–22605. The Journal of Immunology 4913

59. Sebo, P., R. Osicka, and J. Masin. 2014. Adenylate cyclase toxin-hemolysin 65. Delgado, M., and D. Ganea. 2000. Inhibition of IFN-gamma-induced janus relevance for pertussis vaccines. Expert Rev. Vaccines 13: 1215–1227. kinase-1-STAT1 activation in macrophages by vasoactive intestinal peptide and 60. Fang, F. C., and A. Vazquez-Torres. 2002. Nitric oxide production by human pituitary adenylate cyclase-activating polypeptide. J. Immunol. 165: 3051–3057. macrophages: there’s NO doubt about it. Am. J. Physiol. Cell. Mol. 66. Tai, Z., Y. Lin, Y. He, J. Huang, J. Guo, L. Yang, G. Zhang, and F. Wang. 2014. Physiol. 282: L941–L943. Luteolin sensitizes the antiproliferative effect of interferon a/b by activation of 61. Tsukamoto, K., K. Hazeki, M. Hoshi, K. Nigorikawa, N. Inoue, T. Sasaki, and Janus kinase/signal transducer and activator of transcription pathway signaling O. Hazeki. 2008. Critical roles of the p110 beta subtype of phosphoinositide 3- through protein kinase A-mediated inhibition of protein tyrosine phosphatase kinase in lipopolysaccharide-induced Akt activation and negative regulation of SHP-2 in cancer cells. Cell. Signal. 26: 619–628. nitrite production in RAW 264.7 cells. J. Immunol. 180: 2054–2061. 67. Pellegrino, M. J., and P. J. Stork. 2006. Sustained activation of extracel- 62. Qi, X. F., D. H. Kim, Y. S. Yoon, J. H. Li, S. B. Song, D. Jin, X. Z. Huang, lular signal-regulated kinase by nerve growth factor regulates c-fos protein stabilization and transactivation in PC12 cells. J. Neurochem. 99: 1480– Y. C. Teng, and K. J. Lee. 2009. The adenylyl cyclase-cAMP system suppresses 1493. TARC/CCL17 and MDC/CCL22 production through p38 MAPK and NF- 68. Nandan, D., R. Lo, and N. E. Reiner. 1999. Activation of phosphotyrosine kappaB in HaCaT keratinocytes. Mol. Immunol. 46: 1925–1934. phosphatase activity attenuates mitogen-activated protein kinase signaling and 63. Hickey, F. B., C. F. Brereton, and K. H. Mills. 2008. Adenylate cycalse toxin of inhibits c-FOS and nitric oxide synthase expression in macrophages infected Bordetella pertussis inhibits TLR-induced IRF-1 and IRF-8 activation and IL-12 with Leishmania donovani. Infect. Immun. 67: 4055–4063. production and enhances IL-10 through MAPK activation in dendritic cells. J. 69. Gomez, M. A., I. Contreras, M. Halle´, M. L. Tremblay, R. W. McMaster, and Leukoc. Biol. 84: 234–243. M. Olivier. 2009. Leishmania GP63 alters host signaling through cleavage- 64. la Sala, A., J. He, L. Laricchia-Robbio, S. Gorini, A. Iwasaki, M. Braun, activated protein tyrosine phosphatases. Sci. Signal. 2: ra58. G. S. Yap, A. Sher, K. Ozato, and B. Kelsall. 2009. Cholera toxin inhibits IL-12 70. Hardin, A. O., E. A. Meals, T. Yi, K. M. Knapp, and B. K. English. 2006. SHP-1 production and CD8alpha+ differentiation by cAMP-mediated inhibits LPS-mediated TNF and iNOS production in murine macrophages. inhibition of IRF8 function. J. Exp. Med. 206: 1227–1235. Biochem. Biophys. Res. Commun. 342: 547–555.