C. burnetii Nine mile phase II strain stocks were cultivated in African green monkey kidney Vero cells (CCL-81; ATCC, Manassas, VA) and purified as previously described (Morgan et al., 2010). The C. burnetii T4BSS ΔdotA mutant strain (Beare et al., 2012) and C. burnetii T4BSS ΔdotA mutant complemented strain were generously provided by Dr. Bob Heinzen (NIAID Rocky Mountain Laboratories, Montana) and grown as previously described (Beare et al., 2012). Non-adherent human monocytic leukemia derived THP-1 cells (TIB-202; ATCC) were grown in 75-cm² tissue culture flasks in RPMI 1640 medium (Gibco, Carlsbad, CA) supplemented with 1 mM sodium pyruvate, and 10% fetal bovine serum (FBS) at 37°C in 5% CO2 (Mahapatra et al., 2010). Hela 229 epithelial cells (CCL-1.2; ATCC) were grown in RPMI 1640 supplemented with 10% FBS and gentamicin (Invitrogen) at 37°C in 5% CO₂ and 95% humidified air. Infections of THP-1 cells with C. burnetii NMII were initiated in 24-well tissue culture plates at a multiplicity of infection (MOI) of 25. Bacteria were added to 2 × 10⁶ THP-1 cells per well and incubated at 37°C for 4 h to allow close host cell-bacteria contact. Fresh media was then added to each well for a final concentration of 10⁶ cells/ml. This time point represents T = 0 for infectious studies.

To assay C. burnetii modulation of NF-κB activation, uninfected or C. burnetii-infected THP-1 cells were incubated for 72 hpi. Cells were then incubated in media with (+) or without (−) bacteriostatic levels (10 μg/ml) of Cm for an additional 24 h (Mahapatra et al., 2010). Total protein lysates were collected from cell pellets using laemmli sample buffer (Bio-Rad, Hercules, CA) containing protease and phosphatase inhibitor cocktails (Sigma). An NF-κB activation control was generated by treating uninfected THP-1 cells with 50 ng/ml of recombinant human TNF-α (R&D systems) for 8 h prior to total cellular protein collection (Voth et al., 2007). For temporal analysis, protein lysates from paired infected and uninfected THP-1 cells were collected at 0, 24, 48, 72, 96, and 120 hpi, with one experimental set transiently treated with 10 μg/ml of Cm for the final 24 h while the other set was mock-treated. Cell culture media was exchanged daily using centrifugation to harvest the cells and removal of the spent media followed by suspension of the cells in fresh media ± Cm. Infected and uninfected cells were handled identically and a minimum of three experiments (N = 3) were carried out for each time point and condition. Temporal analysis of infected and uninfected differentiated THP-1 cells treated with Rifampin to inhibit mRNA synthesis was performed by treating cells with phorbol 12-myristate 13-acetate (PMA; 200 nM; EMD Biosciences, San Diego, CA) overnight. Cells were infected with C. burnetii, and Rifampin (10 μg/ml) was added (+Rif) or not (−Rif) at the time of infection. Total protein lysates were collected at 2, 24, 48, 72, and 96 hpi for immunoblot analysis as described below.

An inhibitory peptide (IP)—DRQIKIWFQNRRMKWKKNGLLSGDEDFSS (Novus Biologicals)—that competitively inhibits phosphorylation of RelA (S529/S536) and a custom synthesized control peptide (CP)—RMDRKWKQIFQNKIWRKSSDELLNDFGGD (Thermofisher scientific)—were used to determine if inhibition of NF-κB activation alters C. burnetii survival/growth in host cells. Initially, the IP's ability to suppress NF-κB activation in uninfected THP-1 cells was determined. Briefly, 150 μm IP or CP was added to 10⁶ THP-1 cells and incubated at 37°C, 5% CO₂ for 4 h. Recombinant human TNF-α (200 ng/ml; R&D systems) was then added to induce NF-κB activation. Controls included untreated THP-1 cells, TNF-α only treated cells, IP only treated cells, and CP only treated cells. After an additional 12 h incubation, cells were pelleted and total protein lysates collected. After confirmation via immunoblotting that the IP does inhibit NF-κB activation, it was used to determine the requirement of NF-κB activation for C. burnetii survival/growth. Briefly, 10⁶ THP-1 cells treated with either 10 μg/ml of Cm (Sigma), 200 ng/ml of TNF-α (R&D systems), or IP (150 μm) were inoculated with purified C. burnetii NMII at an MOI of 25. Prior to inoculation, cells were treated with IP for 4 h or TNF-α for 1 h. To ensure consistent cellular response, fresh TNF-α was added every 12 h and IP every 24 h, respectively. At 72 hpi, cells were harvested for (i) total protein lysates, (ii) IFA after cytospin as previously described (Mahapatra et al., 2010) to detect PV formation, or (iii) C. burnetii infectious unit enumeration by sonication lysis of infected cells, serial dilution, and infection of HeLa cell monolayers followed by Fluorescent Forming Unit (FFU) calculation as described (Coleman et al., 2004; Omsland et al., 2009). Briefly, IFA was used to visualize infected cells using a Nikon Eclipse TE 2000-S fluorescence microscope. Fifteen randomly chosen fields of view (FOV) visualized at 20x magnification were counted per well to determine the average number of PVs per FOV. For FFU/ml determination, the FOV average times the area FOV/well and inoculum volume (100 ul) were multiplied by the dilution factor to determine infectious particles from each sample treatment.

Cell lysates were separated by 12% SDS-PAGE and transferred to a nitrocellulose membrane (Pierce, Rockford, IL). The membranes were blocked for 1 h at room temperature with 5% nonfat milk in Tris-buffered saline (150 mM NaCl, 100 mM Tris-HCl, pH 7.6) containing 0.1% Tween-20 (TBST) (Voth et al., 2007). Following blocking, membranes were incubated overnight at 4°C with 5% nonfat milk in TBST having primary antibodies for the various target proteins (see following). Detection of NF-κB was performed using a rabbit monoclonal anti-human primary antibody specific to the phosphorylated Serine 536 form of RelA (Cell Signaling Technology, Danvers, MA). Activation of the non-canonical NF-κB pathway was assayed using rabbit anti-human polyclonal antibody against p100 (the precursor), and p52 (active form of NF-kappaB2, Cell Signaling Technology, Danvers, MA). Determination of C. burnetii density within infected THP-1 cell lysates was performed using rabbit polyclonal antibody against the C. burnetii 27-kDa outer membrane protein Com1 as previously described (Morgan et al., 2010). Mouse monoclonal antibodies directed against human β-actin (Sigma, Saint Louis, MO) were employed as a loading control. After primary antibody incubation, the membranes were washed 3X with TBST then incubated with either anti-rabbit or anti-mouse IgG secondary antibody conjugated to horseradish peroxidase (KPL, Gaithersburg, MD) for 1 h at room temperature, washed 3X in TBST and detected by chemiluminescense following the manufacturer's directions (ECL SuperSignal West Pico Chemiluminescent Substrate, Pierce, Rockford, IL). Visualization and digital imaging of immunoblots was performed on a FluorChem HD2 Imaging System (Alpha Innotech Corporation, Leandro, CA).

The signal density of the detected bands in experimental samples were analyzed by ImageJ (version 1.46 h) as described previously (Schneider et al., 2012). Briefly, relative phosphorylated RelA band intensity was normalized to β-actin and quantified with respect to uninfected THP-1 cells (Figure 1A) and C. burnetii infected THP-1 cells at 24 hpi (Figure 2A). To examine the role of non-canonical NF-κB signaling pathway, relative NF-κB p100 and NF-κB p52 band intensity was normalized to β-actin and quantified with respect to C. burnetii infected THP-1 cells at 24 hpi (Figure 3). In Figure 4D, relative C. burnetii Com1 band intensity was normalized to β-actin and quantified with respect to C. burnetii infected THP-1 cells at 72 hpi. For Figure 5A, relative phosphorylated RelA band intensity was normalized to β-actin and quantified with respect to uninfected THP-1 cells.

An indirect immunofluorescence assay (IFA) was used to enumerate PV and infectious C. burnetii. For immuno-staining, methanol-fixed C. burnetii-infected cells were incubated 1 h with polyclonal Guinea pig anti-C. burnetii primary sera diluted in PBS, 2% BSA. After 3 PBS washes, samples were incubated 1 h with Alexa488-conjugated goat anti-Guinea pig secondary IgG (Molecular probes, Thermo Fisher) also diluted in PBS, 2% BSA. DAPI (4',6-diamidino-2-phenylindole) was used in the secondary incubation to stain total DNA. Labeled cells were visualized using a Nikon Eclipse TE 2000-S fluorescence microscope with a Nikon DS FI1 camera and NIS-ELEMENTS F 3.00 software (Mahapatra et al., 2010). A minimum of 15 fields per sample were counted using the 20x objective. All experiments were performed with 3 biological samples and statistical analyses were performed using a paired Student's t-test.

Hela cells were transiently transfected with the GFP-RelA plasmid (Addgene) using X-tremeGENE 9 DNA Transfection Reagent (Roche Life Sciences), as described by the manufacturer. Transfected cells were grown overnight and then inoculated with either C. burnetii NMII, C. burnetii ΔdotA (Beare et al., 2012), or a C. burnetii dotA-complemented strain (Beare et al., 2012). Separate cells were also infected with C. burnetii NMII concurrent with the addition of Cm (Mahapatra et al., 2010). C. burnetii ΔdotA strain and dotA-complemented strains were generously provided by Dr. Robert A. Heinzen (Rocky Mountain Laboratories, NIAID, NIH, Hamilton, MT) (Beare et al., 2012). All strains were used at an MOI of 25. At 24 and 48 hpi, cells were methanol-fixed and IFA performed as above using Guinea pig anti-C. burnetii primary antibody and Alexa555-conjugated goat anti-Guinea pig secondary IgG (Molecular probes, Thermo Fisher). DAPI was used to stain total DNA. Subcellular co-localization of GFP-RelA and DAPI in transfected cells was visualized as above using a 60x oil objective. A minimum of 100 transfected cells were scored per sample. Data shown are the means of three independent experiments and statistical analyses were performed using a paired Student's t-test.

RelA is one of the most extensively studied NF-κB complex subunits (Sakurai et al., 2003; Bonizzi and Karin, 2004; Perkins, 2007). Similar to cRel and RelB, it contains a 300-amino acid region with homology to the Rel proto-oncogene (RH domain) and a transactivation domain (Sakurai et al., 2003; Bonizzi and Karin, 2004; Perkins, 2007). The RH domain harbors motifs for nuclear localization and binding to specific DNA sequences, while the transactivation domain contains phosphorylation sites that remain bound to the inhibitor IκB while in the cytoplasm (Sakurai et al., 2003; Bonizzi and Karin, 2004; Perkins, 2007). Phosphorylation of serine 536 (S536) in the transactivation domain is required for optimal activation (Sakurai et al., 2003; Bonizzi and Karin, 2004; Viatour et al., 2005; Perkins, 2007). To determine if NF-κB is modulated by C. burnetii, NF-κB activation was assayed in THP-1 cells via detection of RelA phosphorylation during infection and treatment with Cm (Sakurai et al., 2003). Figures 1A,B reveal that NF-κB is activated during infection as evidenced by increased levels of phosphorylated RelA, and that C. burnetii protein synthesis is required to modulate the level of activation. Compared to uninfected cells, RelA phosphorylated protein levels increased ~10-fold in C. burnetii-infected cells. However, phosphorylated RelA levels were ~20-fold higher in infected cells treated with Cm. These results indicate that C. burnetii proteins modulate NF-κB signaling during infection.

To assess the dynamics of NF-κB activation throughout infection, we examined infected THP-1 cells from 24–144 hpi. Cells were either mock-treated or transiently treated with Cm. We hypothesized that NF-κB activation would respond directly to de novo bacterial protein synthesis depending on the stage of infection (early [24–48 hpi], mid [48–96 hpi], or late [96–144 hpi]). To test this hypothesis, we first investigated the total levels of RelA. As shown in Figure 2A, RelA levels at each time post-infection were consistent throughout infection, indicating C. burnetii does not modulate the relative expression of RelA. We next assessed RelA phosphorylation in the presence and absence of transient Cm treatment. Figures 2A,B show that in the absence of transient Cm treatment, phosphorylated RelA levels remain low at 24 hpi (early infection). However, when compared to 24 hpi, RelA phosphorylation levels significantly increase at 48 hpi (early to mid-infection) and remain elevated through 96 hpi (mid infection). Notably, transient application of Cm at 0, 24, and 48 hpi resulted in even higher levels of phosphorylated RelA at 24, 48, and 72 hpi, respectively, suggesting that C. burnetii proteins dampen NF-κB activation during infection. Interestingly, application of Cm at 72 hpi did not alter RelA phosphorylation levels at 96 hpi relative to infected cells in the absence of Cm. During late infection times (96–144 h), RelA phosphorylation was reduced and de novo C. burnetii protein synthesis did not alter NF-κB activation. Combined, these results indicate that infection of THP-1 cells by C. burnetii involves temporal modulation of NF-κB activation via RelA phosphorylation. While C. burnetii infection triggers NF-κB activation, de novo C. burnetii proteins significantly suppress NF-κB activation during early and middle stages of intracellular growth. To confirm these findings, we examined the effect of treatment with the mRNA synthesis inhibitor rifampin on temporal NF-κB activation during infection. Additionally, to determine if these observations were infection model-specific, we assessed infection of PMA-differentiated THP-1 macrophage-like cells. Cells were treated with rifampin at the time of infection, and RelA phosphorylation assessed at 2, 24, 48, 72, and 96 hpi. Figure 2C shows that RelA phosphorylation levels in differentiated THP-1 cells infected with C. burnetii increased relative to uninfected cells. Furthermore, inhibition of bacterial mRNA synthesis resulted in even higher levels of RelA phosphorylation that were evident through 96 hpi similar to Figure 2A. These observations support our findings that indicate C. burnetii proteins temporally modulate NF-κB activation during infection.

The results above demonstrate the involvement of RelA in NF-κB activation. NF-κB transcription factors are typically activated by either the canonical or non-canonical signaling pathway (Beinke and Ley, 2004; Bonizzi and Karin, 2004; Park et al., 2005; Viatour et al., 2005; Perkins, 2007). The canonical pathway transmits signals via RelA activation, while the non-canonical pathway functions through NF-κB p52 activation (Beinke and Ley, 2004; Bonizzi and Karin, 2004; Park et al., 2005; Viatour et al., 2005; Perkins, 2007). Formation of active p52 occurs via proteolytic processing of the p100 precursor (Beinke and Ley, 2004; Bonizzi and Karin, 2004; Park et al., 2005; Viatour et al., 2005; Perkins, 2007). Therefore, to determine if the non-canonical pathway is activated during C. burnetii infection of THP-1 cells, we used immunoblot analysis to detect NF-κB p100/p52. Figure 3 clearly demonstrates that C. burnetii does not modulate host cell NF-κB p100 and p52 levels over the course of infection (24—144 hpi) in the presence or absence of bacterial protein synthesis. When compared to C. burnetii-infected cells at 24 hpi, NF-κB p100 levels remain constant throughout infection, and addition of Cm does not significantly impact p100 expression (Figure 3B). In contrast, p52 levels are barely detectable and do not significantly change in the presence or absence of Cm (Figure 3C). Together, these data reveal that NF-κB activation in infected THP-1 cells does not involve non-canonical NF-κB signaling.

To determine if C. burnetii-mediated NF-κB activation is essential for intracellular survival and growth, we treated THP-1 cells with an inhibitory peptide (IP) that competitively inhibits RelA (S529/S536) phosphorylation to suppress NF-κB activation. Inhibition of RelA phosphorylation was confirmed by immunoblotting using TNF-α to trigger RelA phosphorylation (Figure 4A). Figure 4A demonstrates the effectiveness of TNFα to induce RelA phosphorylation and the IP's ability to inhibit RelA phosphorylation. To measure C. burnetii development following induction (TNFα) or inhibition (IP) of NF-κB activation, we analyzed (i), the number of treated THP-1 cells with PV relative to infected, untreated cells, (ii) the number of infectious progeny produced in Hela cells by treated THP-1 cell lysates relative to infected, untreated THP-1 cell lysates, and (iii) levels of the major outer membrane protein Com1 in treated and untreated THP-1 cells as a measure of C. burnetii total protein. Figures 4B–D show that PV formation, infectious progeny, and C. burnetii Com1 levels were reduced by approximately 20–30% in THP-1 cells pretreated with TNF-α. Of particular interest, Figure 4B demonstrates that, compared to C. burnetii infected cells, PV are reduced by approximately 60% in IP-treated cells. Figures 4C–E confirm and support these observations, demonstrating that the number of infectious C. burnetii produced (Figure 4C) within IP-treated cells as well as the amount of Com1 detected in these cells (Figures 4D,E) was reduced by approximately 60%. Collectively, these results demonstrate that C. burnetii development is significantly impaired when NF-κB activation is inhibited. Additionally, unlike Cm treatment, C. burnetii is able to overcome high levels of activated NF-κB induced by TNF-α application. However, their growth is marginally inhibited.

Pathogenic microorganisms employ unique strategies to interfere with NF-κB signaling (Rahman and McFadden, 2011). Bacteria modulate NF-κB signaling (activation or suppression) depending on requirements of their intracellular lifestyle and niche (Rahman and McFadden, 2011). C. burnetii's close relative, L. pneumophila, utilizes over 10 Dot/Icm effectors to induce a biphasic pattern of NF-κB activation (Rahman and McFadden, 2011; Shin, 2012). To determine if the C. burnetii T4BSS plays a role in modulating NF-κB, we analyzed RelA phosphorylation in THP-1 cells infected with either wild type C. burnetii (± Cm), a C. burnetii T4BSS ΔdotA mutant, or a dotA-complemented strain. Figures 5A,B reveal that, compared to cells infected with wild type C. burnetii (I−Cm), levels of phosphorylated RelA were higher in cells infected with C. burnetii (I+Cm) or the ΔdotA mutant. However, levels of phosphorylated RelA were similar in cells infected with either wild type C. burnetii (−Cm) or the ΔdotA complemented strain. Together, these results suggest that C. burnetii protein synthesis and a functional T4BSS are required for NF-κB modulation. To confirm this observation and determine if this event extends to RelA nuclear translocation, we transiently transfected Hela cells with GFP-RelA and examined nuclear translocation when infecting cells with either wild type C. burnetii (± Cm), the ΔdotA mutant, or the ΔdotA-complemented strain. Migration of GFP-RelA into the nucleus indicates NF-κB activation. Using immunofluorescence microscopy we visualized RelA colocalization with DAPI (DNA stain). Figure S1 and Figure 5C show that at 24 and 48 hpi, GFP-RelA localized to the nucleus regardless of the C. burnetii strain used to infect the cells. However, quantification of the amount of GFP-RelA in the nucleus of infected cells revealed that wild type C. burnetii (+Cm) and the ΔdotA mutant had approximately 50% more cells with nuclear-localized GFP-RelA (Figure 5D). These data clearly demonstrate that NF-κB activation levels are significantly higher in infected cells treated with Cm and cells infected with the ΔdotA mutant, suggesting that T4BSS effectors play a crucial role in modulating NF-κB signaling.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.