Coxiella burnetii Employs the Dot/Icm Type IV Secretion System to Modulate Host NF-κB/RelA Activation

Abstract

Coxiella burnetii is the causative agent of Q fever and an obligate intracellular pathogen in nature that survives and grows in a parasitophorous vacuole (PV) within eukaryotic host cells. C. burnetii promotes intracellular survival by subverting apoptotic and pro-inflammatory signaling pathways that are typically regulated by nuclear transcription factor-κB (NF-κB). We and others have demonstrated that C. burnetii NMII proteins inhibit expression of pro-inflammatory cytokines and induce expression of anti-apoptotic genes during infection. Here, we demonstrate that C. burnetii promotes intracellular survival by modulating NF-κB subunit p65 (RelA) phosphorylation, and thus activation, in a Type Four B Secretion System (T4BSS)-dependent manner. Immunoblot analysis of RelA phosphorylated at serine-536 demonstrated that C. burnetii increases NF-κB activation via the canonical pathway. However, RelA phosphorylation levels were even higher in infected cells where bacterial protein or mRNA synthesis was inhibited. Importantly, we demonstrate that inhibition of RelA phosphorylation impairs PV formation and C. burnetii growth. We found that a T4BSS-defective mutant (CbΔdotA) elicited phosphorylated RelA levels similar to those of wild type C. burnetii infection treated with Chloramphenicol. Moreover, cells infected with CbΔdotA or wild type C. burnetii treated with Chloramphenicol showed similar levels of GFP-RelA nuclear localization, and significantly increased localization compared to wild type C. burnetii infection. These data indicate that without de novo protein synthesis and a functional T4BSS, C. burnetii is unable to modulate NF-κB activation, which is crucial for optimal intracellular growth.

Keywords: Coxiella burnetii; NF-κB; Q fever; obligate intracellular; type four secretion system.

Figure 1

Immunoblot analysis of C. burnetii modulation of NF-κB activation. (A) Top panel—immunoblot detection of phosphorylated RelA. Bottom panel—β-actin loading control. Uninfected THP-1 cells without Cm (U−Cm). Uninfected THP-1 cells with Cm (U+Cm). Infected THP-1 cells without Cm (I−Cm). Infected THP-1 cells with Cm (I+Cm). All samples were collected at 72 hpi. TNF-α-treated cells were used as positive controls for RelA phosphorylation. (B) Graph—difference in RelA protein phosphorylation levels relative to normalized β-actin. The Y-axis represents fold changes in phosphorylated RelA and the X axis indicates samples. Results represent the mean of three independent experiments. Error bars represent ± SD. Statistically significant differences between U−Cm and I−Cm or I+Cm are represented as ^*P ≤ 0.001—(Student's t-test).

Figure 2

Temporal analysis of NF-κB activation in C. burnetii-infected THP-1 cells. (A) Representative immunoblots showing NF-κB activation and controls over a time course of C burnetii infection. Top panel—RelA probed with a monoclonal antibody. Middle panel—phosphorylated RelA probed with a monoclonal antibody against S536. Bottom panel—β-actin loading control. Time (hpi) at which untreated (I−Cm) and Cm-treated (I+Cm) C. burnetii-infected THP-1 cells were harvested is indicated above lanes. Arrow indicates lane break. (B) Fold change of phosphorylated RelA vs. time in the presence and absence of Cm. Results of densitometric analysis are presented as the mean of three experiments. Error bars show ±S.E.M. Statistical differences were calculated using a t-test for paired samples. ^*Signifies P < 0.05 of I-Cm samples compared to 24 hpi. ^**Signifies P < 0.05 between paired (I−Cm to I+Cm) samples at each time point. (C) A representative immunoblot showing phosphorylated RelA levels in PMA-differentiated THP-1 cells treated or mock-treated with Rifampin 2—96 hpi. Top panel—RelA control probed with a monoclonal antibody. Middle panel—phosphorylated RelA from infected cells mock-treated with Rifampin (−Rifampin). Bottom panel—phosphorylated RelA from infected cells treated with Rifampin (+Rifampin). Time (hpi) at which cells were harvested is indicated above each lane. Un = Uninfected cells.

Figure 3

Immunoblot analysis of NF-κB p100 and p52 in C. burnetii-infected cells. (A) Representative immunoblots showing NF-κB p100 and p52 in C. burnetii-infected THP-1 cells either treated with Cm (I+Cm) or left untreated (I−Cm). Blots were probed with a polyclonal rabbit antibody against NF-κB (p100/p52) of human origin (Top/Middle panels, respectively). Bottom panel—β-actin loading control. CD40-treated THP-1 cells served as a p52 positive control. Sample conditions and time point are indicated above each lane. Arrows indicate lane breaks. (B) Results of densitometric analysis showing fold changes of NF-κB p100 levels over time of infection with and without Cm after β-actin normalization. Error bars show ±S.E.M of three biological experiments. (C) Fold changes of NFκB p52 levels over time of infection with and without Cm after β-actin normalization. Error bars show ±S.E.M of three biological experiments.

Figure 4

C. burnetii development following NF-κB activation or inhibition. (A) Top panel—immunoblot detection of phosphorylated RelA in THP-1 cells. Bottom panel—β-actin loading control. Sample treatments (±) are indicated above immunoblot panels and correlate to the lanes below. (B) Inhibition of RelA phosphorylation in THP-1 cells impairs C. burnetii PV formation. 15 fields of view at 20x magnification (>100 cells/field of view) from three independent samples were analyzed for PV enumeration. Error bars show ±S.E.M. Statistical differences relative to I−Cm were calculated using a t-test for paired samples (^*signifies P < 0.05). (C) Infectious progeny enumeration in Hela cells from three independent samples as measured by IFA and Fluorescent Forming Units (FFUs). Error bars show ± SEM. Statistical differences relative to I−Cm were calculated using a t-test for paired samples (^*signifies P < 0.05). (D) Representative immunoblot showing Com1 levels in C. burnetii-infected THP-1 cells either untreated or treated with Cm, TNF-α, or RelA IP and harvested at 72 hpi. Top panel—β-actin control. Bottom panel—Com1 detected with a rabbit polyclonal antibody. (E) Differences in C. burnetii Com1 levels relative to normalized β-actin from three independent samples were analyzed. All samples were analyzed at 72 hpi. C. burnetii growth was examined in THP-1 cells either untreated (I−Cm), or treated with Cm (I+Cm), TNF-α, or RelA IP. Error bars represent ± SEM. Statistically significant differences relative to I–Cm are represented as ^*P ≤ 0.05, (Student's t-test).

Figure 5

C. burnetii employs the Dot/Icm T4BSS to modulate NF-κB signaling. (A) Immunoblot analysis of NF-κB activation in THP-1 cells (48 hpi) using wild type C. burnetii (± Cm), a DotA mutant (ΔdotA), and a dotA-complemented strain (ΔdotA) Comp. Top panel—detection of RelA. Middle panel—detection of phosphorylated RelA. Bottom panel—β-actin loading control. TNF-α-treated cells were used as positive RelA phosphorylation controls. (B) Differences in phosphorylated RelA levels relative to normalized β-actin. The Y-axis represents fold changes in phoshorylated RelA by densitometry and the X axis indicates samples. Results represent the mean of three independent experiments. Error bars represent ± SEM. For statistically significant differences, ^*signifies P < 0.05 when samples are compared to U–Cm, ^**signifies P < 0.05 when samples are compared between I–Cm to I+Cm or ΔdotA. ^***Signifies P < 0.05 when samples are compared between ΔdotA and ΔdotA Comp—(Student's t-test). (C) Immunofluorescent image showing localization of GFP-RelA (green). C. burnetii was visualized by Alexa-555 (red), and DNA/Nuclei with Dapi (blue). HeLa cells transiently transfected with GFP-RelA (green) vector were infected with the indicated C. burnetii strains for 48 h. (A–C) Uninfected cells. (D–G,I–L,N–Q,S–V) Cells infected with the indicated C. burnetii strain in which GFP-RelA and DAPI co-localize. Arrows indicate co-localization. (H,M,R,W) Merged micrographs representing cells infected with the indicated C. burnetii strain where GFP-RelA did not co-localize with DAPI. Bar, 10 μm. (D) Quantification of GFP-RelA/DAPI co-localization. A minimum of 100 transiently transfected Hela cells from each of three separate experiments was counted to determine GFP-RelA/DAPI co-localization. Error bars show ± SD. Statistically significant differences (^*P < 0.05, Student's t-test) are shown when samples are compared to U–Cm at 48 hpi.