Modulation of innate immune signaling by a Coxiella burnetii eukaryotic-like effector protein

Significance

Coxiella burnetii is a stealth pathogen that evades innate immune recognition by inhibiting the NF-κB signaling pathway. This process is mediated by the bacterial Dot/Icm secretion system; however, the bacterial effector/s, as well as the molecular mechanism involved in this process have remained unknown to date. Here, by investigating C. burnetii proteins with eukaryotic-like features (EUGENs), we discovered a new effector protein, NopA (nucleolar protein A), which localizes at nucleoli of infected cells and perturbs nucleocytoplasmic transport by manipulating the intracellular gradients of the GTPase Ran. In doing so, NopA reduces the nuclear levels of transcription factors involved in the innate immune sensing of pathogens and single-handedly down-modulates the expression of a panel of cytokines.

Abstract

The Q fever agent Coxiella burnetii uses a defect in organelle trafficking/intracellular multiplication (Dot/Icm) type 4b secretion system (T4SS) to silence the host innate immune response during infection. By investigating C. burnetii effector proteins containing eukaryotic-like domains, here we identify NopA (nucleolar protein A), which displays four regulator of chromosome condensation (RCC) repeats, homologous to those found in the eukaryotic Ras-related nuclear protein (Ran) guanine nucleotide exchange factor (GEF) RCC1. Accordingly, NopA is found associated with the chromatin nuclear fraction of cells and uses the RCC-like domain to interact with Ran. Interestingly, NopA triggers an accumulation of Ran-GTP, which accumulates at nucleoli of transfected or infected cells, thus perturbing the nuclear import of transcription factors of the innate immune signaling pathway. Accordingly, qRT-PCR analysis on a panel of cytokines shows that cells exposed to the C. burnetii nopA::Tn or a Dot/Icm-defective dotA::Tn mutant strain present a functional innate immune response, as opposed to cells exposed to wild-type C. burnetii or the corresponding nopA complemented strain. Thus, NopA is an important regulator of the innate immune response allowing Coxiella to behave as a stealth pathogen.

The nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) family of transcription factors regulates the expression of genes associated with diverse cellular functions and plays a central role in regulating the innate and acquired host immune response to bacterial infections (1, 2). Under physiological conditions, the transcription factors of the NF-κB family are sequestered in the cytoplasm by specific interactions with nuclear factor kappa-light polypeptide gene enhancer in B cells inhibitor alpha (IκBα), which mask the nuclear localization signal (NLS) on transcription factors. Exogenous signals, including recognition of the tumor necrosis factor (TNF) by TNF receptor or the bacterial lipopolysaccharide (LPS) by toll-like receptor 4 (TLR4), activate the NF-κB signaling pathway by triggering the phosphorylation and proteasomal degradation of IκBα, thus unmasking the NLS on transcription factors. The signal is then recognized by importin-α and members of the importin-β family, which mediate the translocation of transcription factors to the nucleus through nuclear pore complexes (1). Energy for nuclear transport of NLS-containing proteins is provided by intracellular gradients of the small GTPase Ras-related nuclear protein (Ran), which interacts with the importin complexes upon nuclear import. GDP-bound Ran is largely cytoplasmic and nuclear translocation triggers the conversion to the GTP-bound form by means of the Ran guanine nucleotide exchange factor (GEF) RCC-1 (regulator of chromosome condensation-1). In its GTP-bound form, Ran triggers the dissociation of importins from the cargo and importin complexes recycle back to the cytoplasm. There, Ran GTPase activating protein (RanGAP) generates Ran-GDP, which dissociates from importin complexes (3).

Given its pivotal role in the antimicrobial response, it is not surprising to observe that a considerable number of bacterial pathogens deploy effector proteins that modulate the NF-κB signaling pathway (1, 2). These are mostly involved in phosphorylation, ubiquitination, and proteasomal degradation of components of the NF-κB complex, whereas others modulate NF-κB-mediated transcription (1, 2). Interestingly, it has recently been reported that Salmonella and Orientia tsutsugamushi effector proteins can interfere with nucleocytoplasmic transport, thereby inhibiting nuclear translocation of the p65/RelA transcription factor (4, 5).

The Q fever pathogen Coxiella burnetii is an obligate intracellular bacterium that relies on the translocation of effector proteins by a defect in organelle trafficking/intracellular multiplication (Dot/Icm) type 4b secretion system (T4SS) to replicate within large autolysosomal-like compartments inside infected cells (6, 7). Bioinformatics analysis identified over 140 C. burnetii genes encoding candidate effector proteins (7); however, the majority of these remain underinvestigated due to the technical constraints associated with the genetic manipulation of this organism. A subset of effector proteins is involved in the biogenesis of Coxiella-containing vacuoles (CCVs), by rerouting membrane traffic to the bacterial replicative niche, while other effectors manipulate the apoptotic and inflammatory pathways to ensure intracellular persistence (6). Importantly, C. burnetii behaves as a stealth pathogen, evading the host innate immune response by down-modulating the NF-κB and the inflammasome signaling pathways (8, 9). The C. burnetii effector protein IcaA (inhibition of caspase activation A) inhibits NOD-like receptor family pyrin domain containing 3 (NLRP3)-mediated inflammasome activation induced by caspase-11 (8), whereas the NF-κB signaling pathway is down-modulated in a Dot/Icm-dependent manner, by perturbing the nuclear translocation of the p65/RelA subunit, without affecting the overall cellular levels of p65 (9). However, the bacterial effector/s involved in this process remains uncharacterized (9). We have previously reported the large-scale phenotypic characterization of the C. burnetii transposon mutant library, which led to gaining important insights into the function of the Dot/Icm secretion system, and which highlight an important set of virulence determinants (10–12). Importantly, several genes involved in intracellular replication of C. burnetii encode proteins with predicted eukaryotic-like domains, which prompted us to investigate eukaryotic-like genes (EUGENs) on a genome-wide scale. Here, we identify and validate the Dot/Icm-mediated translocation of seven C. burnetii EUGENs. Among these, NopA (nucleolar protein A) displays four regulation of chromosome condensation (RCC) repeats, which are partially homologous to the seven repeats found in the bladed β-propeller structure of the Ran GEF RCC1 (13–15). Similarly to RCC1, NopA also localizes at the nucleus of infected or transfected cells; it is found associated with the chromatin nuclear fraction; and it uses the RCC-like domain to interact with Ran. Differently from RCC1, however, NopA accumulates at nucleoli and sequesters Ran, thus perturbing nucleocytoplasmic transport. Indeed, NopA perturbs nuclear translocation of p65 upon cell treatment with TNF-α or challenge with C. burnetii. Conversely, transposon insertions in the nopA gene restore nuclear translocation of p65 during infections, to levels that are similar to those observed with the Dot/Icm-deficient C. burnetii dotA mutant. Accordingly, myeloid cells challenged with the C. burnetii nopA or dotA mutant strains present a functional innate immune response, as opposed to myeloid cells exposed to wild-type (WT) C. burnetii or the nopA complemented strain.

Results

Identification of C. burnetii EUGENS.

The searching algorithm for type IV effector proteins (S4TE) 2.0 (16) was used to identify C. burnetii eukaryotic-like genes (EUGENs) encoding candidate effector proteins. This allowed the identification of 56 genes, which were validated using the Protein families database (PFAM), Simple Modular Architecture Research Tool (SMART), Conserved Domain Database (CDD), and Eukaryotic Linear Motif (ELM) databases (SI Appendix, Table S1). Of these, 20 candidate EUGENS were retained for further analysis (SI Appendix, Table S2), based on the S4TE score (16), the eukaryotic-like domain encoded, and the presence of corresponding transposon mutants in our library (10). cbu0072 (ankA), cbu0201 (ankC), cbu0447 (ankF), cbu0781 (ankG), and cbu1213 (ankI) encode ankyrin repeats (17, 18); cbu0295, cbu0547, and cbu1457 (cig43) encode tetratricopeptide repeats; cbu0175 and cbu1379a encode predicted Ser/Thr kinases; cbu0801 (rimI), cbu0505 (cig14), and cbu1799 encode acetyltransferases; cbu0096 encodes a predicted phospholipase D; cbu0519 (dedA) encodes a SNARE-like domain-containing protein; cbu1206 encodes a predicted sterol reductase; cbu1217 encodes a protein with four regulation of chromosome condensation (RCC) repeats; cbu1724 encodes a predicted F-box protein; cbu1366 (cig40) encodes a coiled-coil domain-containing protein; and cbu0542 (ligA) encodes a predicted DNA ligase (SI Appendix, Table S2). Of note, Dot/Icm-dependent translocation of proteins encoded by 8 of these genes has been previously validated using Legionella pneumophila as a surrogate system (17, 19, 20) (SI Appendix, Table S2). Selected genes were cloned into pXDC61K-blaM vector, thus generating N-terminal fusions with β-lactamase, and transformed into C. burnetii Nine Mile II (NMII) RSA439. The expression of 16 out of 20 chimeric proteins was validated by Western blot using an anti-β-lactamase antibody (SI Appendix, Fig. S1A). Candidate effector protein translocation was assessed at 6, 12, 24, 48, and 72 h postinfection using the β-lactamase assay. C. burnetii expressing β-lactamase alone or β-lactamase-tagged CvpB (CBU0021) (12) were used as negative and positive controls, respectively. CvpB, CBU0295, and CBU1217 were efficiently translocated from 12 h postinfection, whereas AnkA, F, and G were translocated from 24 h postinfection (Fig. 1 A and B). Finally, AnkC, CBU0175, and CBU1724 were also translocated, albeit less efficiently, at later time points of infection (Fig. 1 A and B). Plasmids encoding translocated effectors were then transformed into the C. burnetii dotA::Tn strain to validate their Dot/Icm-dependent secretion at 72 h postinfection. The expression of 6 chimeric proteins was validated by Western blot using an anti-β-lactamase antibody (SI Appendix, Fig. S1B). None of the effector proteins were secreted by the Dot/Icm-defective mutant as expected (Fig. 1A). Next, cbu0072 (ankA), cbu0295, cbu0447 (ankF), cbu0781 (ankG), and cbu1217 were cloned into a pLVX-mCherry vector to tag effector proteins at their N-terminal domain and their localization was investigated in noninfected and C. burnetii-infected U2OS cells (Fig. 1C). AnkA and CBU0295 were mostly diffuse in the cytoplasm and did not localize at CCVs in infected cells. AnkF displayed a punctate pattern in the cytoplasm, which partially colocalized with the lysosomal marker LAMP1 in noninfected and infected cells alike (Fig. 1C). Differently from previous reports, indicating a translocation of AnkG from mitochondria to the nucleus of transfected cells following staurosporine treatment (21), in our hands, this effector protein displayed nuclear localization even in the absence of staurosporine, in both infected and noninfected cells (Fig. 1C). Of note, CBU1217 was exclusively localized at subnuclear structures in over 90% of either infected or noninfected cells (Fig. 1C).

Fig. 1.

Identification of C. burnetii EUGENs. (A) U2OS cells were challenged with C. burnetii strains expressing BLAM-tagged versions of candidate EUGENs for 6, 12, 24, 48, and 72 h. The percentage of BLAM-positive, infected cells was automatically calculated using CellProfiler over the total number of infected cells per each condition. Empty, BLAM empty vector. The Dot/Icm-dependent translocation of the effectors that were efficiently secreted was validated in the dotA::Tn mutant strain at 72 h postchallenge. (B) Representative images of positive (blue) cells treated as in A. (C) Noninfected or GFP-expressing C. burnetii-infected U2OS cells were transfected with plasmids encoding N-terminally tagged mCherry versions of the effector proteins validated in A (red). At 24 h after transfection, cells were fixed and labeled with Hoechst (blue) and an anti-LAMP1 antibody (white). White arrows point at CBU1217 subnuclear localization. (Scale bars, 10 μm.)

The Effector Protein CBU1217 Localizes at Nucleoli in Infected and Transfected Cells.

The localization of CBU1217 was further investigated by cloning the gene into a pJA-LacO-4HA plasmid, to express the effector protein carrying an N-terminal 4×HA tag in C. burnetii, under the control of an isopropyl β-d-1-thiogalactopyranoside (IPTG) promoter, and monitor its localization during infection. GFP-expressing WT C. burnetii or the dot/icm mutant dotA::Tn (10) were transformed either with pJA-LacO-4HA or with pJA-LacO-4HA-cbu1217. Expression of 4HA-CBU1217 was validated by Western blot using an anti-HA antibody (SI Appendix, Fig. S1C). U2OS cells were challenged with the transformed C. burnetii strains and NopA localization was assessed, in the presence or absence of IPTG, using anti-HA and anti-fibrillarin antibodies, and Hoechst dye. Infections by C. burnetii transformed with pJA-LacO-4HA-cbu1217 in the absence of IPTG did not show specific HA labeling (Fig. 2A). Addition of 1 mM IPTG triggered 4HA-CBU1217 expression, which colocalized with fibrillarin in over 90% of HA-positive cells (Fig. 2B). The intracellular localization of 4HA-CBU1217 was lost when cells were infected with the dotA::Tn mutant transformed with pJA-LacO-4HA-cbu1217 in the presence of IPTG (Fig. 2D), confirming that CBU1217 is a Dot/Icm substrate. Induction of the expression of the HA tag alone did not show specific localization (Fig. 2C). We thus named the new C. burnetii EUGEN NopA, for nucleolar protein A.

Fig. 2.

Intracellular localization of CBU1217/NopA and role in C. burnetii replication during infection. U2OS cells were challenged either with WT GFP-tagged C. burnetii (white) transformed with plasmids encoding 4HA-tagged CBU1217/NopA (A and B, red) or the 4HA tag alone (C, red), or with the GFP-tagged dotA::Tn mutant (D, white), transformed with plasmids encoding 4HA-tagged CBU1217/NopA, all under the control of an IPTG-inducible promoter. At 72 h postinfection, cells were fixed and labeled with Hoechst (blue) and anti-fibrillarin antibodies (green). I, IPTG-induced; NI, noninduced. Arrow points at 4HA-CBU1217/NopA localization in infected cell. U2OS cells were transfected with plasmids encoding HA-tagged versions of either full-length (E) or the indicated deletion mutants (F and G) of HA-tagged NopA. At 24 h after transfection, cells were either fixed and labeled with Hoechst (blue), an anti-fibrillarin antibody (green) and an anti-HA antibody (red, Center), or lysed and processed for cell fractionation (Right). Cell fractions were analyzed by Western blotting using anti-fibrillarin and anti-GAPDH antibodies as nuclear and chromatin (Nu and Ch) and cytoplasmic (Cy) markers, respectively, and anti-HA antibodies to reveal NopA localization. (Scale bars, 10 μm.) (H) Genome equivalents (GEs) calculated using TaqMan real-time PCR with DNA purified from infected spleens of 5 SCID mice per group on day 14 after challenge with 1 × 10⁶ GEs of the strains shown. (I) Replication index calculated as the ratio between spleen GE at the time of necropsy and the input GE of the strains listed in the figure legend. (J) Spleen weight as a percentage of total body weight at the time of necropsy on day 14 after infection with 1 × 10⁶ GEs of the strains listed in the figure legend. Values are the mean of three independent infections, with error bars indicating SDs from the mean. ****P < 0.0001, ***P < 0.001, two-way ANOVA (H) and one-way ANOVA (J), Dunnett’s multiple comparisons test.

As mentioned above, NopA encodes four RCC repeats in its C-terminal domain (Fig. 2E). In the eukaryotic protein RCC1, seven repeats are arranged in a seven-bladed propeller, which associates with nuclear chromatin and acts as a GEF for Ran, thus regulating nucleocytoplasmic protein transport (3). To determine the role of the RCC-like domain in NopA localization and function, the effector protein was cloned into a pRK5-HA plasmid to generate HA-tagged NopA. U2OS cells transfected with pRK5-HA-NopA were processed for immunofluorescence using Hoechst dye and anti-HA and anti-fibrillarin antibodies. In parallel, HA-NopA localization was investigated by Western blot using U2OS cells transfected as above and lysed and separated into cytoplasmic, nuclear, and chromatin fractions. Full-length NopA (NopA_FL) localized at nucleoli in over 90% of transfected cells, confirming our observations in the context of C. burnetii infections (Fig. 2E). Western blot analysis confirmed that NopA is excluded from the cytoplasmic fraction and localized at the soluble and chromatin nuclear fractions (Fig. 2E). Next, we generated HA-tagged NopA deletions to exclude (NopA_N-ter; amino acids [aa] 1 to 195, Fig. 2F) or include (NopA_C-ter; aa 196 to 497, Fig. 2G) the RCC repeats. Ectopically expressed HA-NopA_N-ter was excluded from nuclei and remained diffuse in the cytoplasm (Fig. 2F), whereas HA-NopA_C-ter retained the nucleolar localization (Fig. 2G). Cell fractionation confirmed the cytoplasmic localization of HA-NopA_N-ter and the nuclear localization of HA-NopA_C-ter, as well as the association with the chromatin fraction (Fig. 2 F and G). Thus, despite the lack of typical nuclear or nucleolar localization signals, the C-terminal domain of NopA encoding the RCC-like domain, is necessary and sufficient for the nucleolar targeting of the effector protein. The role of the RCC repeats in the intracellular localization of NopA was further dissected by generating increasing deletions of single RCC repeats (numbered from 1 to 4 from the N-terminal) from either the N-terminal or C-terminal ends of HA-NopA_C-ter (SI Appendix, Fig. S2A). The intracellular localization of each construct was tested by immunofluorescence and cell fractionation following ectopic expression in U2OS cells. Interestingly, this revealed that the first RCC repeat is critical for targeting NopA to the nucleus as removal of this repeat from NopA_C-ter displaces the protein to the cytoplasm (SI Appendix, Fig. S2 B and C). The first two RCC repeats (RCC12; aa 196 to 310) alone localize within the nucleus but are excluded from nucleoli (SI Appendix, Fig. S2E) and instead localize at promyelocytic leukemia (PML) bodies (SI Appendix, Fig. S2G). This localization remains unchanged with the addition of the third RCC repeat (SI Appendix, Fig. S2 D and F), and it is only with the addition of the complete NopA_C-ter that the protein localizes at nucleoli (Fig. 2G), suggesting the presence of a nucleolar-targeting motif in the fourth RCC repeat. Unfortunately, we were unable to express detectable amounts of single RCC repeats (RCC1 and RCC4, SI Appendix, Fig. S2A).

NopA Is Not Involved in C. burnetii Intracellular Replication.

Given the early translocation of NopA observed using the β-lactamase assay, we determined the time course of NopA production during infection. To this aim, we have complemented the nopA mutation, using a mini Tn7 transposon to integrate a WT copy of HA-tagged nopA, under the regulation of its predicted endogenous promoter, in the chromosome of the C. burnetii Tn227 strain, which carries the transposon insertion closest to the nopA start codon (10). Protein expression was then monitored by Western blot, using an anti-HA antibody, from cells challenged with the complemented nopA::Tn strain for 12, 24, 48, and 72 h. By this approach, detectable amounts of NopA were observed from 12 h postinfection (SI Appendix, Fig. S1D).

We previously reported that transposon insertions in nopA do not affect bacterial replication in Vero cells (10). To further investigate the role of NopA in C. burnetii infections, bacterial replication and virulence of the wild-type, dotA::Tn, nopA::Tn, and the nopA::Tn complemented strain (nopA::Tn Comp.) described above, were tested using either bone marrow-derived macrophages (BMDMs) or a severe combined immunodeficiency (SCID) mouse model of infection. Confirming our initial observations, transposon insertions in nopA do not affect C. burnetii replication (Fig. 2 H and I) or virulence (as determined here by splenomegaly measurements, Fig. 2J).

NopA Interacts with the Small GTPase Ran.

Given that NopA localizes at nucleoli and presents four out of the seven RCC repeats present in the eukaryotic Ran-GEF RCC1, we investigated whether NopA can interact with Ran. To this aim, U2OS cells incubated either with the nopA::Tn mutant or the complemented strain expressing 4HA-tagged NopA under the control of the predicted endogenous promoter (nopA::Tn Comp.). Twenty-four hours postinfection, cells were lysed, separated into cytoplasmic, nuclear, and chromatin fractions, and NopA was immunocaptured from cell fractions using an anti-HA antibody. As expected, NopA was not detected in cells infected with the nopA::Tn mutant strain, whereas it was efficiently isolated from the nuclear and chromatin fractions of cells challenged with the complemented strain (Fig. 3A). Of note, whole cell lysates of cells incubated with the complemented strain also presented an accumulation of Ran in the chromatin fraction, which was not observed in cells challenged with the nopA::Tn mutant strain (Fig. 3A). Importantly, Ran was efficiently detected, together with NopA, in immunoprecipitates from the nuclear and chromatin fractions of cells challenged with the nopA::Tn complemented strain, indicating indeed an interaction with the C. burnetii effector protein (Fig. 3A). The NopA/Ran interaction was further investigated in U2OS cells, following the ectopic expression of either HA-tagged NopA_N-ter, NopA_C-ter, or the C. burnetii effector protein CvpF as negative control (22). Unfortunately, under these conditions, we were unable to immunoprecipitate full-length NopA from transfected cells. Similarly to infected cells, 24 h posttransfection, cells were lysed, separated into cytoplasmic, nuclear, and chromatin fractions, and NopA truncations and CvpF were immunocaptured from cell fractions using an anti-HA antibody. As expected, NopA_N-ter and CvpF were efficiently isolated from the cytoplasmic fractions, whereas NopA_C-ter was isolated from the nuclear and cytoplasmic fractions (Fig. 3B). In agreement with what we observed in infected cells, the ectopic expression of NopA_C-ter triggered an accumulation of Ran to the chromatin fractions (Fig. 3B). Moreover, Ran was specifically detected in the nuclear and chromatin fractions upon immunocapturing of NopA_C-ter, confirming an interaction between the two proteins (Fig. 3B). Of note, no interaction was detected between Ran and NopA_N-ter, despite their shared cytoplasmic localization (Fig. 3B). Furthermore, NopA_C-ter did not interact with other small GTPases such as DRP1 or RAB26, nor with the nucleolar marker fibrillarin (Fig. 3B). Conversely, the C. burnetii effector protein CvpF (22), was readily immunocaptured from the cytoplasm of transfected cells and interacted with RAB26 as reported (22) (Fig. 3B).

Fig. 3.

NopA interacts with the small GTPase Ran. (A) U2OS cells challenged for 24 h with either the C. burnetii nopA transposon mutant (nopA::Tn) or the corresponding complemented strain (nopA::Tn Comp.) were lysed and processed for cell fractionation. Whole cell lysates (WCLs) were probed with the indicated antibodies, as well as anti-GAPDH and anti-fibrillarin antibodies as cytoplasmic (Cy), and nuclear/chromatin (Nu/Ch) markers, respectively. Following immunoprecipitation with anti-HA-coated magnetic beads, the presence of Ran and that of fibrillarin (as a negative control) was assessed using specific antibodies (IP HA). (B) U2OS cells transfected with HA-tagged versions of either the N-terminal domain (NopA_N-ter), the C-terminal domain (NopA_C-ter) of NopA, or CvpF as negative control were lysed and processed for cell fractionation. WCLs were probed with the indicated antibodies, as well as anti-GAPDH and anti-fibrillarin antibodies as cytoplasmic (Cy), nuclear/chromatin (Nu/Ch) markers, respectively. Following immunoprecipitation with anti-HA-coated magnetic beads, the presence of candidate interacting proteins was assessed using specific antibodies (IP HA). (C) U2OS cells were transfected with plasmids encoding GFP1-9 in combination with plasmids encoding either the GFP10 and GFP11 tags alone as negative control (Top row), GFP10 and GFP11 linked by a leucine zipper motif (GFP10-zip-GFP11) as positive control (Middle row), or GFP10-Ran and GFP11-NopA (Bottom row). At 24 h after transfection, cells were fixed and labeled with Hoechst (blue) and anti-GFP antibodies (red) to reveal nuclei and the expression of GFP1-9, respectively. Protein/protein interaction was assessed following the reconstitution of GFP (Reconst. GFP, green). Arrows point at nuclei in which Ran and NopA interact. (D) The percentage of cells presenting GFP reconstitution over the total number of GFP1-9-positive cells was calculated. Values are means ± SD from two independent experiments. Asterisks indicate statistically significant variations (n.s., nonsignificant, ****P < 0.0001, one-way ANOVA, Dunnett’s multiple comparison test). (Scale bars, 20 μm.)

Finally, the direct interaction between NopA and Ran was further investigated using the tripartite split-GFP interaction sensor (23). Briefly, the assay is based on a tripartite association between two GFP β-strands (GFP10 and GFP11), fused to proteins of interest, and the complementary GFP1-9 detector. If proteins interact, GFP10 and GFP11 self-associate with GFP1-9 to reconstitute a functional GFP. pCDNA3-zipper-GFP10 and pCDNA3-zipper-GFP11 were used as negative control, whereas a plasmid encoding GFP10 and GFP11 linked by a zipper motif (GFP10-zip-GFP11) was used as positive control (23). ran cDNA was cloned into the pCDNA3-GFP10-zipper plasmid to generate the GFP10-Ran, whereas nopA, rcc1, and fbl (the gene encoding fibrillarin) were cloned into the pCDNA3-zipper-GFP11 plasmid to generate the corresponding GFP11 fusion proteins. Combinations of the above-mentioned constructs with a pCMV plasmid encoding GFP1-9 were used for triple transfections in U2OS cells. After fixation, an anti-GFP antibody was used to identify cells expressing GFP1-9 (which is not fluorescent) and protein interactions were analyzed by monitoring GFP reconstitution. As expected, coexpression of GFP1-9 with GFP10 and GFP11 did not result in the reconstitution of GFP (Fig. 3 C, Top row, and 3D). The coexpression of GFP1-9 with GFP10-zip-GFP11 led to the reconstitution of GFP fluorescence in over 93% of transfected cells, demonstrating the functionality of the assay (Fig. 3 C, Center row, and 3D). Importantly, over 60% of cells expressing GFP1-9 in combination with GFP10-Ran and GFP11-NopA showed reconstitution of GFP, with a fluorescent signal detected at nuclei, with a strong accumulation at nucleoli (Fig. 3 C, Bottom row, and 3D). On the contrary, the expression of GFP1-9 in combination with GFP10-Ran and GFP11-RCC1 allowed reconstitution of GFP fluorescence, which was homogeneously detected in the nucleus in over 73% of transfected cells (SI Appendix, Fig. S3 A and D). Lack of GFP reconstitution upon expression of GFP1-9 in combination with GFP10-Ran and GFP11-fibrillarin indicated that the shared nucleolar localization was not sufficient for GFP reconstitution (SI Appendix, Fig. S3 A and D).

Ectopic expression of either mCherry-NopA_FL or mCherry-NopA_C-ter in combination with GFP-Ran in U2OS cells also confirmed the colocalization of both proteins at nucleolar structures labeled with the anti-fibrillarin antibody (SI Appendix, Fig. S3B), Conversely, Ran-GFP accumulation at nucleoli was lost when the small GTPase was ectopically expressed in U2OS cells in combination either with mCherry alone or mCherry-NopA_N-ter (SI Appendix, Fig. S3B). Collectively, these observations indicate that NopA specifically interacts with Ran and may sequester it at nucleoli.

NopA Preferentially Interacts with GDP-Bound Ran and Triggers an Increase in Ran-GTP.

To determine whether NopA displays preferential binding to Ran in its GDP- versus GTP-bound form, a GFP-trap assay was carried out on U2OS cells cotransfected with plasmids encoding HA-NopA_C-ter in combination with either GFP alone, GFP-Ran, GFP-Ran_T24N (GDP-locked), GFP-Ran_Q69L (GTP-locked), or GFP-Ran_N122I (nucleotide-free form). Similar to RCC1, NopA displayed preferential binding to either GDP-locked Ran_T24N or the nucleotide-free form Ran_N122I (Fig. 4A).

Fig. 4.

NopA increases the intracellular levels of Ran-GTP. (A) The GFP-trap assay was carried out in U2OS cells expressing HA-tagged NopA_C-ter in combination with either GFP alone, GFP-Ran, GFP-Ran_T24N (GDP-locked), GFP-Ran_Q69L (GTP-locked), or GFP-Ran_N122I (guanosine-free). WCLs (Upper) were probed with anti-GFP and anti-HA antibodies to assess the expression of the GFP-tagged proteins and HA-tagged NopA_C-ter, and anti-tubulin antibodies as loading control. Protein/protein interactions were assessed using anti-GFP and anti-HA antibodies following GFP capture (GFP-trap, Lower). (B) GTP-bound Ran was pulled down using RanBP1-coated beads from cell lysates of U2OS cells challenged for 24 h with either WT C. burnetii (WT), a nopA transposon mutant (nopA::Tn), the corresponding complemented strain (nopA::Tn Comp.), or the Dot/Icm-defective mutant (dotA::Tn). Noninfected cells (NI) were used as control. WCLs were probed with anti-C. burnetii (NMII), anti-Ran, and anti-β-tubulin antibodies. GTP-bound Ran was revealed using an anti-Ran antibody (IP RanBP1). (C) GTP-bound Ran was pulled down using RanBP1-coated beads from cell lysates of U2OS cells expressing either the HA tag alone, HA-tagged versions of either full-length (NopA_FL), the N-terminal domain (NopA_N-ter), the C-terminal domain (NopA_C-ter) of NopA, or CvpB. WCLs were probed with anti-HA antibodies to assess the expression of the HA-tagged versions of NopA and anti-Ran and anti-tubulin antibodies as loading controls. GTP-bound Ran was revealed using an anti-Ran antibody (IP RanBP1). The signal ratio of GTP-bound Ran over the total amount of Ran is indicated for experiments illustrated in B and C. Values are mean ± SD from three independent experiments. n.s., nonsignificant, ****P < 0.0001, **P < 0.007, *P < 0.02, one-way ANOVA, Dunnett’s multiple comparison test.

Next, we investigated whether NopA binding to Ran can affect the Ran GDP/GTP ratio that is required to fuel nucleocytoplasmic transport. U2OS cells were challenged either with WT C. burnetii, the dotA::Tn, nopA::Tn, or the nopA::Tn complemented strains. Noninfected cells were used as control. Twenty-four hours postinfection, cells were lysed and incubated with agarose beads coated with the Ran effector Ran-binding protein 1 (RanBP1), to specifically pull down the GTP-bound form of Ran. Indeed, infection with WT C. burnetii triggered a 40-fold increase in the intracellular levels of Ran-GTP, as compared to noninfected cells (Fig. 4B). This phenotype was lost in cells challenged with the dotA::Tn mutant strain and only an 8.5-fold increase was observed in cells challenged with the nopA::Tn mutant strain. Increased levels of Ran-GTP were largely restored (35-fold increase) in cells exposed to the complemented strain (nopA::Tn Comp., Fig. 4B). The effects of NopA on the intracellular levels of Ran-GTP were further investigated in U2OS cells transfected with plasmids encoding either HA alone, HA-NopA, HA-NopA_N-ter, HA-NopA_C-ter, or the C. burnetii effector protein CvpB (12) used here as negative control. A threefold increase in the intracellular levels of Ran-GTP was observed in cells expressing either HA-NopA or HA-NopA_C-ter, as compared to cells transfected with HA alone or HA-NopA_N-ter (Fig. 4C). As expected, ectopic expression of CvpB had negligible impact on the intracellular levels of Ran-GTP (Fig. 4C). These observations suggest that NopA sequestration of Ran at nucleoli leads to an increase in the intracellular levels of Ran-GTP, which may negatively regulate nuclear import (24).

NopA Perturbs Protein Translocation to the Nucleus.

Given the role of Ran in nucleocytoplasmic traffic, and the previously reported observation that C. burnetii infections modulate nuclear translocation of p65 by a Dot/Icm-dependent mechanism (9), we investigated whether NopA affects the nuclear localization of p65, which follows the activation of the NF-κB signaling pathway. U2OS cells transfected with plasmids encoding either HA- or mCherry-tagged versions of NopA were either left untreated or challenged with 10 ng/mL TNF-α for 30 min, and the nuclear translocation of p65 was monitored using an anti-p65 antibody either by fluorescence microscopy or Western blot following cell fractionation. Cells expressing either HA- or mCherry-tagged CvpB or the tags alone were used as controls. TNF-α treatment efficiently activated the NF-κB pathway, as indicated by the significant degradation of IκBα (Fig. 5A). Accordingly, p65 was readily relocalized to the nucleus of cells expressing either the HA or mCherry tags alone or tagged versions of the C. burnetii effector CvpB (Fig. 5 B–D). However, p65 translocation was largely inhibited in cells expressing either HA-NopA or mCherry-NopA (Fig. 5 B–D). In all cases, the intracellular levels of p65 remained largely unaltered. To determine whether NopA modulates the intracellular levels of p65 by perturbing its nuclear import or by accelerating its nuclear export, U2OS cells expressing either mCherry-NopA, mCherry-CvpB, or mCherry alone as controls, were incubated for 4 h with 5 nM leptomycin B (LMB), a fungal metabolite that blocks nuclear export by covalently binding to exportin 1. As p65 shuttles continuously between the nucleus and the cytoplasm, treatment with LMB in mCherry or mCherry-CvpB expressing cells led to an accumulation of the transcription factor in the nucleus (Fig. 5D and SI Appendix, Fig. S4). Interestingly however, ectopic expression of mCherry-NopA significantly prevented p65 nuclear accumulation in response to LMB treatment (Fig. 5D and SI Appendix, Fig. S4). A similar phenotype was observed in cells treated with LMB for 4 h, followed by a 30-min incubation with TNF-α (Fig. 5D and SI Appendix, Fig. S4). These data indicate that indeed, NopA perturbs nuclear import.

Fig. 5.

Overexpression of NopA interferes with the nuclear translocation of p65. Representative Western blot of U2OS cells expressing either the HA tag alone, HA-NopA, or HA-CvpB left untreated or incubated with 10 ng/mL TNF-α for 30 min, lysed, and processed for cell fractionation. Whole cell lysates (A, WCLs) were used to assess the overall levels of p65 and IκBα and nuclear fractions (B) to monitor p65 translocation to the nucleus (nuclear fraction). The signal ratio of p65 over tubulin or fibrillarin and of IκBα over tubulin is indicated for experiments illustrated in A and B. Values are mean ± SD from three independent experiments. (C) Representative images of U2OS cells expressing mCherry-NopA or mCherry-CvpB and treated as in A. The localization of p65 was monitored using an anti-p65 antibody and Hoechst staining of nuclei. Asterisks indicate transfected cells. (D) CellProfiler was used to identify mCherry-expressing U2OS cells and to measure the median of the ratios of p65 fluorescence intensity at nuclei versus cytoplasm. Values are means ± SEM from two independent experiments where a minimum of 200 nuclei were measured per condition. Asterisks indicate statistically significant variations [n.s., nonsignificant, ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.1, one-way ANOVA, Dunnett’s (A and B) and Bonferroni (D) multiple comparison test]. (Scale bars, 10 μm.)

Next, we tested whether the perturbation of nuclear import triggered by NopA was specific to p65 and C. burnetii infections. The nuclear translocation of the transcription factor IRF3 was monitored in U2OS cells cotransfected with 3FLAG-tagged IRF3 in combination with either mCherry alone, mCherry-NopA, or mCherry-CvpB, and infected with the Sendai virus for 18 h. Noninfected cells were used as control (SI Appendix, Fig. S5A). Similar to what we reported for p65, IRF3 was readily translocated to the nuclei of cells expressing either mCherry alone or mCherry-CvpB but remained largely cytoplasmic in cells expressing mCherry-NopA (SI Appendix, Fig. S5 A and B).

NopA Is Involved in the Silencing of the Innate Immune Response during C. burnetii Infections.

p65 nuclear translocation was further monitored in U2OS cells noninfected or challenged either with 10 ng/mL TNF-α for 30 min, WT C. burnetii, the Dot/Icm-defective mutant dotA::Tn, the nopA::Tn, or the complemented strain (nopA::Tn Comp.), for 24, 48, and 72 h by fluorescence microscopy and, for the 72 h time point, by Western blot following cell fractionation. IκBα was significantly degraded in all conditions as compared to noninfected cells, indicating an efficient activation of the NF-κB pathway (Fig. 6A). Translocation of p65 to the nucleus was readily detected in cells treated with TNF-α, either by Western blotting (Fig. 6B) or by immunofluorescence (Fig. 6 C and D). Cells challenged with WT C. burnetii or the nopA::Tn complemented strain showed a small but significant increase in nuclear p65 fluorescence as compared to noninfected, untreated cells (Fig. 6 B–D). However, incubation with either the dotA::Tn or the nopA::Tn mutants triggered an accumulation of p65 to the nucleus which was comparable to the TNF-α treatment (Fig. 6 B–D). Measurement of p65 nuclear translocation by immunofluorescence, which was specifically measured in infected cells, resulted in a stronger phenotype as compared to Western blot analysis, which was carried out on the total cell population.

Fig. 6.

NopA interferes with the nuclear translocation of p65 during C. burnetii infections. Representative Western blot of U2OS cells challenged for 72 h with GFP-tagged strains of WT C. burnetii (WT), the Dot/Icm-defective dotA transposon mutant (dotA::Tn), the nopA transposon mutant (nopA::Tn), or the corresponding complemented strain (nopA::Tn Comp.). Noninfected cells (NI) and cells treated with 10 ng/mL TNF-α (TNF-α) for 30 min were used as negative and positive controls, respectively. Cells were lysed and fractionated to isolate nuclear fractions. Whole cell lysates (A, WCLs) were used to assess the overall levels of p65 and IκBα and nuclear fractions (B) to monitor p65 translocation to the nucleus (nuclear fraction). Normalized densitometry of indicated protein ratios was calculated. Values are means ± SD from two independent experiments. (C) Representative images of U2OS cells treated as in A. The localization of p65 (red) was monitored using an anti-p65 antibody and Hoechst staining of nuclei (blue). White arrows indicate nuclei of infected cells. (D) U2OS cells were treated as in A for 24, 48, and 72 h. CellProfiler was used to identify infected and total U2OS cells and to measure the median of the ratios of p65 fluorescence intensity at nuclei versus cytoplasm. Values are means ± SEM from two independent experiments where a minimum of 400 nuclei were measured per condition. In all cases, n.s., nonsignificant, ***P < 0.0001, one-way ANOVA, Dunnett’s multiple comparison test. (Scale bars, 10 μm.)

To investigate the downstream effects of perturbing the nuclear translocation of transcription factors involved in the immune response to C. burnetii infections, differentiated THP-1 macrophages were exposed to either WT C. burnetii, the Dot/Icm-deficient dotA::Tn mutant, the nopA::Tn mutant, or the corresponding complemented strain (nopA::Tn Comp.) for 24, 48, and 72 h. Total RNA was extracted from cell lysates and qRT-PCR analysis was used to monitor the expression of a panel of cytokines (Fig. 7A and SI Appendix, Fig. S6A). A slight increase in the mRNA expression levels of all tested cytokines was observed in cells exposed to WT C. burnetii or the nopA::Tn complemented strain, as compared to noninfected cells. Interestingly however, cells exposed to either the dotA::Tn mutant or the nopA::Tn mutant displayed a comparable, significant increase in the production of the majority of the cytokines tested, ranging from a 2-fold increase to a 100-fold increase for IL8 (Fig. 7A and SI Appendix, Figs. S6A and S7). Down-modulation of the innate immune response was further confirmed by monitoring TNF-α and IFN-α production in THP-1 macrophages infected as above for 72 and 96 h. As C. burnetii effectors are known to perturb the secretory pathway of infected cells (20, 25), THP-1 cells were treated with brefeldin A (BFA) 24 h prior to fixation and the intracellular levels of TNF-α and IFN-α were assessed by flow cytometry (Fig. 7 B and C), following the application of a specific gating to isolate the population of infected cells (SI Appendix, Fig. S6B). A significant increase in the intracellular levels of both cytokines was observed in cells infected either with the nopA::Tn or the dotA::Tn strains as compared to cells infected with WT C. burnetii or the nopA::Tn complemented strain (Fig. 7 B and C). Overall, our data indicate that C. burnetii uses the Dot/Icm secretion system to down-modulate the NF-κB signaling pathway as previously reported (9), and that NopA is a key effector for this process.

Fig. 7.

NopA inhibits cytokines production. (A) Differentiated THP-1 cells were challenged either with GFP-expressing WT C. burnetii (WT), the Dot/Icm-defective dotA transposon mutant (dotA::Tn), the nopA transposon mutant (nopA::Tn), or the corresponding complemented strain (nopA::Tn Comp.) for 24, 48, 72, and 96 h. The expression of TNF-α and IFN-α4 cytokines was assessed by RT-qPCR for the indicated time points. Dot plots from a representative experiment showing intracellular staining of TNF-α (B) and IFN-α (C) in cells infected for 72 and 96 h and treated with BFA for the last 24 h. Infected cells were first gated on the GFP⁺ population and the percentage of cells expressing TNF-α and IFN-α was assessed. Flow cytometry data are presented on graphs as fold relative to WT. Values are means ± SD from three independent experiments. n.s., nonsignificant, ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.1. Full statistical analysis for the 72-h time point illustrated in A is available in SI Appendix, Fig. S7.

Discussion

Intracellular bacterial pathogens and symbionts establish intimate interactions with their eukaryotic hosts, which have evolved by coevolution over time. Part of their adaptation to their intracellular niches has been mediated by transkingdom acquisition and functional integration of eukaryotic genes in bacterial genomes (26). Indeed, EUGENs represent a hallmark of intracellular bacteria, and are rarely observed in free-living bacteria. Importantly, many EUGENs from intracellular bacteria produce candidate or validated effector proteins that are translocated into host cells through dedicated type III or type IV secretion systems (27). Thus, EUGENs are predicted to play an important role in the establishment of parasitic or symbiotic bacterial lifestyles.

In this study, bioinformatics analysis combined with translocation assays led to the identification of seven C. burnetii effector proteins encoding eukaryotic-like domains involved in protein/protein interactions, protein/chromatin interactions, and posttranslational modifications. CBU0447 and CBU0175 are conserved among C. burnetii strains, whereas the remaining five EUGENS present some degree of polymorphism (7). Upon ectopic expression in epithelial cells of translocated ankyrin repeat-containing proteins, AnkA (CBU0072) was largely cytoplasmic, whereas AnkF (CBU0447) seemed to associate with membranes that partially colocalized with the lysosomal marker LAMP1. AnkG (CBU0781), which was previously reported to localize at mitochondria and translocate to the nucleus upon staurosporine treatment of transfected cells (21), partially localized to the nucleus even in the absence of staurosporine in our hands. It is thus possible that other Ank proteins modify their intracellular localization at different stages of infection.

Here, we have focused our study on CBU1217, which encodes four RCC repeats in its C-terminal domain (aa 196 to 497). RCC repeats are found in the regulation of chromosome condensation 1 (RCC1) eukaryotic protein (28). In eukaryotes, the RCC domain consists of seven homologous repeats of 51 to 68 amino acid residues, arranged in a β-propeller fold (15). A single RCC domain constitutes the majority of the protein in the case of the RCC1 subgroup of the RCC1 superfamily, whereas multiple RCC domains can be found, either alone or in combination with other functional domains of the other subgroups of the superfamily (13). As such, RCCs are versatile domains that can be involved in protein/protein or protein/chromatin interactions, guanine nucleotide exchange factor (GEF), and posttranslational modifications including ubiquitination and phosphorylation (13). RCC1 is primarily found in association with histones H2A and H2B on chromatin (29) and acts as a GEF for the small GTPase Ran, a master regulator of nucleocytoplasmic transport during interphase and mitotic spindle assembly during mitosis (30).

Among vacuolar bacterial pathogens, the L. pneumophila effector protein LegG1 encodes an RCC-like domain (RLD) consisting of three out of the seven RCC repeats typically found in eukaryotes (31). Of note, LegG1 localizes at Legionella-containing vacuoles (LCVs) where it recruits and activates Ran to promote microtubule polymerization and LCV mobility (32). Differently from LegG1, C. burnetii NopA encodes an additional RCC repeat and, despite the lack of typical nuclear or nucleolar localization signals, exclusively localizes at nuclei with a strong enrichment in the chromatin fraction, which is consistent with RCC1 localization. NopA RCC repeats are necessary and sufficient to target the protein to nucleoli and exert its functions. Moreover, the first RCC repeat seems to be critical for targeting NopA to the nucleus, as removal of this repeat from NopA_C-term displaces the protein to the cytoplasm. Interestingly, the first two RCC repeats (aa 196–310) alone localize at PML bodies. This localization remains unchanged with the addition of the third RCC repeat and it is only the expression of the complete NopA_C-term that triggers protein localization at nucleoli, suggesting the presence of a nucleolar-targeting motif in the fourth RCC repeat.

Similarly to the eukaryotic protein RCC1, NopA interacts with Ran, with preferential affinity for the GDP-bound form and promotes the activation of Ran. Differently from RCC1, however, NopA also triggers a nucleolar accumulation of Ran. Thus, the observed increase in the intracellular levels of GTP-bound Ran may result from either GEF activity of NopA (which has been reported for RCC1), or via the observed sequestration of Ran at nucleoli, which would prevent GTP-bound Ran to recycle back to the cytoplasm, where Ran GTPase-activating proteins (GAPs) stimulate GTP to GDP conversion. As we were unable to purify sufficient amounts of either full-length NopA or NopA_C-ter, we could not assess for the moment whether NopA has intrinsic GEF activity. Interestingly, a residual increase in the intracellular levels of Ran-GTP was still observed in cells challenged with the nopA::Tn mutant strain, as compared to infections with the Dot/Icm-defective dotA::Tn strain. This may suggest that other C. burnetii effectors may have a role in the modulation of Ran activity.

Of note, mutations in nopA do not affect C. burnetii intracellular replication (10). However, increasing the intracellular levels of Ran-GTP results in a global alteration in the nucleocytoplasmic transport of proteins (24). It has been reported that during infections, C. burnetii requires Dot/Icm activity to down-modulate the NF-κB pathway by perturbing the nuclear translocation of the p65 transcription factor (9). Here we demonstrate that NopA is one of the effector proteins involved in this process, as indicated by the strong inhibition of nuclear translocation of p65 upon treatment of cells with TNF-α or infection. The modulation of the NF-κB signaling pathway has been reported for a number of bacterial pathogens and viruses (1, 2). In most cases, bacteria interfere with the degradation of IκBα and the release of p65 or by triggering the proteasomal degradation of p65 itself. Other bacteria, including L. pneumophila and Shigella flexneri may also inhibit the innate immune response downstream of p65 nuclear translocation, at the level of transcription and mRNA processing, respectively (2). Finally, an emerging number of bacterial effectors inhibit NF-κB activation by modulating the nuclear translocation and/or accumulation of p65, by interfering with nucleocytoplasmic protein transport. The Salmonella SPI-2 T3SS effector protein SpvD accumulates importin-α in the nucleus by binding exportin Xpo2, thereby preventing p65 nuclear import (4). O. tsutsugamushi uses ankyrin repeat-containing effector proteins Ank1 and 6 by coopting the function of both importin-β and exportin 1, thus accelerating p65 nuclear export (5). Here we show that the NF-κB pathway is readily activated upon C. burnetii infections as shown by efficient IκBα degradation. However, NopA perturbs nuclear accumulation of p65 by triggering the nuclear accumulation of GTP-bound Ran, resulting in an imbalanced Ran gradient across cells. In turn, this leads to a defective nuclear import of proteins, as also demonstrated by challenging cells ectopically expressing NopA with leptomycin B to block nuclear export.

Considering that these bacterial effectors manipulate common adaptors and GTPases involved in nucleocytoplasmic transport, it would be of interest to monitor their effect on a broader panel of proteins and investigate how infected cells respond to these perturbations. For example, other C. burnetii effector proteins have been described to localize at the nucleus of infected cells (21, 33, 34). In this perspective, it is important to note that nuclear translocation of p65 is not completely ablated during C. burnetii infections, and that the strongest phenotypes are observed at 48 and 72 h postinfection, which is compatible with a reduced, but still detectable translocation of protein to the nucleus at earlier time points. Here we show that indeed the perturbation of nuclear import by NopA affects a broader class of proteins, also outside the context of C. burnetii infections, as indicated by the perturbation of nuclear translocation of IRF3 in response to Sendai virus infection, in cells ectopically expressing NopA.

To monitor the downstream effects of inhibiting the nuclear accumulation of transcription factors involved in immune sensing, we challenged differentiated THP-1 cells with WT C. burnetii or strains carrying mutations either in the Dot/Icm secretion system or in nopA. As expected, infections by the WT strain elicited a minor response in the expression of a panel of cytokines, including TNF-α, interleukins, and interferons, in agreement with the observation that C. burnetii is a stealth pathogen. Evasion of the innate immune response was unmasked by infections with the Dot/Icm-defective strain dotA::Tn, which triggered a significant cytokine response. Interestingly, infections by the nopA::Tn mutant strain largely phenocopied the dotA::Tn mutation, suggesting that NopA is critical for the down-modulation of the innate immune response.

Together, this work highlighted a number of C. burnetii eukaryotic-like effector proteins and showed that one of them, NopA, is responsible for evading the host innate immune response by interfering with nucleocytoplasmic transport.

Materials and Methods

Antibodies, reagents, bacterial strains, cell lines, and growth conditions used in this study are listed in SI Appendix.

Plasmids.

Plasmids used in this study are listed in SI Appendix, Table S4. DNA sequences were amplified by PCR using Phusion polymerase (New England Biolabs) and gene-specific primers (Sigma).

Plasmid Design for Secretion Assay in C. burnetii.

Selected genes from SI Appendix, Table S2 were amplified from C. burnetii RSA439 NMII genomic DNA using primer pairs indicated in SI Appendix, Table S5. PCR products were cloned into the pXDC61-BLAM plasmid to generate N-terminal-tagged fusion version of all candidate effector proteins.

Plasmid Design for Mammalian Cells Transfection.

Effector-coding genes were amplified from C. burnetii RSA439 NMII genomic DNA using primer pairs indicated in SI Appendix, Table S5. PCR products were cloned either into pLVX-mCherry-N2 or pRK5-HA plasmids to generate N-terminal-tagged mCherry or HA fusion versions of all effector proteins, respectively. For cloning of Ran in pcDNA3-GFP10-zipper, Ran was amplified using forward primers Ran-BspEI and reverse primers Ran-XbaI-rev. For cloning of NopA, RCC1, and fibrillarin in pcDNA3-zipper-GFP11, genes were amplified using forward primers NopA-NotI, RCC1-NotI, or FBL-NotI and reverse primers NopA-ClaI-rev, RCC1-ClaI-rev, or FBL-ClaI-rev. pGBKT7-containing eukaryotic sequence of Ran WT, RanT24N/Q69L/N122I mutants were kindly provided by Aymelt Itzen, Zentrum für Experimentelle Medizin Institut für Biochemie und Signaltransduktion, Hamburg, Germany. Ran WT and mutants were amplified from pGBKT7-Ran-WT, pGBKT7-Ran-T24N, pGBKT7-Ran-Q69L, and pGBKT7-Ran-N122I using primers pairs XhoI-Ran-F and Ran-XmaI-rev, and the PCR products were cloned into pLVX-GFP-N2.

Plasmid Design for nopA Complementation in C. burnetii.

For nopA complementation, the nopA putative promoter and nopA sequences were amplified using the primer pairs NheI-prom1217-fw/PstI-prom1217-Rev and CBU1217-BamHI/CBU1217-EcoRI-rev, respectively (SI Appendix, Table S5). The PCR products were cloned into pUCR6K-miniTn7-Kan-tetRA-4HA. Plasmids were electroporated in the nopA::Tn mutant strain Tn227 (10).

Beta-Lactamase Translocation Assay.

For C. burnetii effector translocation assays, cells were cultured in black, clear-bottomed, 96-well plates and infected with the appropriate C. burnetii strain (multiplicity of infection [MOI] of 100) for 24 and 48 h. C. burnetii-expressing BLAM alone or BLAM-tagged CBU0021 were used as negative and positive controls, respectively. Cell monolayers were loaded with the fluorescent substrate CCF4/AM (LiveBLAzer-FRET B/G loading kit; Invitrogen) in a solution containing 20 mM Hepes, 15 mM probenecid (Sigma) pH 7.3, in Hank’s balanced salt solution (HBSS). Cells were incubated in the dark for 1 h at room temperature and imaged using an EVOS inverted fluorescence microscope. Images were acquired using DAPI and GFP filter cubes. The image analysis software CellProfiler was used to segment and count total cells and positive cells in the sample using the 520-nm and 450-nm emission channels, respectively, and to calculate the intensity of fluorescence in each channel. Following background fluorescence subtraction using negative control samples, the percentage of positive cells was then calculated and used to evaluate effector translocation. A threshold of 20% of positive cells was applied to determine efficient translocation of bacterial effector proteins.

Immunofluorescence Staining and Microscopy.

Cells were fixed in 4% (wt/vol) paraformaldehyde in phosphate-buffered saline (PBS) solution at room temperature for 20 min. Samples were then rinsed in PBS solution and incubated in blocking solution (0.5% bovine serum albumine [BSA], 50 mM NH₄Cl in PBS solution, pH 7.4). Cells were then incubated with the primary antibodies diluted in blocking solution for 1 h at room temperature, rinsed five times in PBS solution, and further incubated for 1 h with the secondary antibodies diluted in the blocking solution. To visualize HA-tagged NopA or nuclear/nucleolar proteins, cells were fixed as previously described in 4% (wt/vol) paraformaldehyde in PBS solution. Then, cells were permeabilized with 0.5% Triton X-100 in PBS solution for 3 min at room temperature. Sample were then rinsed in PBS solution and incubated with blocking solution (0.1% Triton X-100, 5% [wt/vol] milk in PBS solution) for 1 h at room temperature. Cells were then incubated with the primary antibodies diluted in blocking solution for 1 h at 37 °C, rinsed five times in PBS solution, and incubated with the secondary antibodies for 1 h at room temperature. For all conditions, coverslips were mounted by using Fluoromount mounting medium (Sigma) supplemented with Hoechst 33258 for DNA staining. Samples were imaged with a Zeiss Axio Imager Z1 epifluorescence microscope (Carl Zeiss) connected to a CoolSNAP HQ² charge-coupled device (CCD) camera (Photometrics). Images were acquired alternatively with 100×, 63×, or 40× oil immersion objectives and processed with MetaMorph (Universal Imaging). ImageJ and CellProfiler software were used for image analysis and quantifications.

Immunoprecipitations and Pull-Down Assays.

For coimmunoprecipitation experiments, pLVX-GFP-N2-tagged WT Ran, RanT24N/Q69L/N122I mutants, or vector control were cotransfected with pRK5-HA-NopA_C-ter in U2OS cells. At 24 h posttransfection, cells were lysed in lysis buffer (10 mM Tris HCl, pH 7.5, 150 mM NaCl, 0.5 mM ethylenediaminetetraacetic acid [EDTA] 1% Nonidet P-40) supplemented with a protease inhibitor tablet (Complete; Roche) and incubated with 25 μL of GFP-Trap magnetic beads (Chromotek) for 2 h at 4 °C with rotation. The beads were then washed three times with wash buffer (10 mM Tris HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA), resuspended in Laemmli buffer 4×, and analyzed by Western blot.

Tripartite Split-GFP Assay.

U2OS were grown in Dulbecco’s modified Eagle medium (DMEM) supplemented in 10% (vol/vol) fetal calf serum (FCS) at 37 °C and 5% CO₂. For the interaction assay, U2OS cells were cotransfected with Lipofectamine 2000 (Gibco, Invitrogen) with plasmids encoding for GFP1-9, GFP10, and GFP11 fusions. At 24 h posttransfection, cells were fixed in 4% paraformaldehyde in PBS solution and processed for immunofluorescence. Protein/protein interactions were scored by calculating the percentage of GFP-positive cells over the total number of cells positive for the anti-GFP antibody.

Cell Fractionation.

U2OS cells were grown to 60% confluence in 100-mm Petri dishes before being transfected with 10 μg of pRK5-HA-NopA_N-ter or pRK5-HA-NopA_C-ter in JetPEI reagent (Polyplus-Transfection) according to the manufacturer’s recommendations. At 24 h after transfection, cells were washed in PBS and pelleted at 4 °C. U2OS cells cultured in 100-mm dishes were infected with the nopA::Tn mutant or the corresponding complemented strain (nopA::Tn Comp.) expressing a 4HA-tagged version of NopA. After 24 h of infection, cells were washed in PBS and pelleted at 4 °C. Transfected or infected cell pellets were subjected to cell fractionation as previously described (35). Where appropriate, cytoplasmic, nuclear, and chromatin fractions were subjected to immunoprecipitation using 40 μL of anti-HA magnetic beads (Sigma) for 2 h at 4 °C with rotation. Bound proteins were eluted using 80 μL of 100 μg/mL⁻¹ HA-peptide (Sigma) and then resuspended in Laemmli buffer 4× and analyzed by Western blot.

Ran Activation Assay.

For the analysis of enzymatic activity of NopA, U2OS cells were either infected or transfected and lysed with lysis buffer (25 mM Hepes, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 10 mM MgCl₂, 1 mM EDTA, 2% glycerol) containing a protease inhibitor tablet (Complete; Roche). Cell lysates were then centrifuged for 10 min at 14,000 × g at 4 °C. For Ran-GTP immunoprecipitation, 40 μL of RanBP1 beads (Cell Biolabs, Inc.) were incubated with cell lysates for 1 h at 4 °C, and then washed three times with lysis buffer, subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS/PAGE), and visualized by Western blotting using an anti-Ran antibody (1:4,000, Sigma). GTP-bound Ran levels were determined by calculating the signal ratio of GTP-bound Ran over the total amount of Ran.

NF-κB/IRF3 Translocation Assays.

To analyze NF-κB translocation, U2OS cells were grown to 60% confluence before being transfected as previously described. At 24 h posttransfection, cells were incubated with media containing 10 ng/mL TNF-α for 30 min at 37 °C. Alternatively, cells were preincubated with media containing 5 nM LMB for 4 h at 37 °C, followed by a TNF-α treatment as indicated above where needed. For C. burnetii infection assays, cells were infected with C. burnetii and incubated at 37 °C for 1 to 3 d. Cells were then fixed in 4% paraformaldehyde in PBS solution and processed for NF-κB immunostaining. The image analysis software CellProfiler was used to segment all nuclei using the Hoechst staining and cell contours using nuclei as seeds and p65 labeling. Cytoplasm was segmented by subtracting nuclei from cell objects. Next, mCherry signal was used to identify and isolate the subpopulation of transfected cells, and single cell measurements of the ratio of the mean p65 fluorescence in the nucleus versus cytoplasm were calculated for each condition. For infection assays, CellProfiler was used to identify and isolate the population of infected cells based on the GFP fluorescence associated with the strains of C. burnetii used in this study and nuclear p65 fluorescence was specifically measured as described above in the subpopulation of infected cells. To analyze IRF3 translocation, pLVX-mCherry-N2-tagged NopA, CvpB, or empty vector were cotransfected with pcDNA3-3×FLAG-tagged IRF-3 in U2OS cells. At 24 h posttransfection, cells were infected with a defective-interfering H4 Sendai virus (36) provided by D. Garcin, Department of Microbiology and Molecular Medicine, University of Geneva, Geneva Switzerland, and used at 50 hemagglutination units (HAU)/mL for 18 h at 37 °C. Cells were then fixed in 4% paraformaldehyde in PBS solution and processed for FLAG immunostaining. IRF3 nuclear translocation was measured as described above for p65.

Densitometry.

Regions of Interest (ROIs) were obtained from each band of interest and the intensity was measured using ImageJ. For each band, the same ROI was used for background calculation and removal from areas adjacent to each band. For the experiments illustrated in Fig. 5, the intensity of bands from samples treated with TNF-α were normalized for the intensity of the corresponding untreated sample. For the experiments illustrated in Fig. 6, the intensity of bands from samples challenged with C. burnetii or treated with TNF-α were normalized for the intensity of the noninfected (NI) sample.

qRT-PCR Analysis of Cytokine mRNA.

Total RNA was extracted from THP-1 cells using the RNeasy Micro kit and submitted to DNase treatment (Qiagen), following manufacturer’s instructions. RNA concentration and purity were evaluated by spectrophotometry (NanoDrop 2000c, Thermo Fisher Scientific). A total of 500 ng of RNA was reverse transcribed with both oligo-dT and random primers, using PrimeScript RT Reagent Kit (Perfect Real Time, Takara) in a 10-mL reaction. Real-time PCR reactions were performed in duplicates using Takyon ROX SYBR MasterMix blue dTTP (Eurogentec) on an Applied Biosystems QuantStudio 5, using the following program: 3 min at 95 °C followed by 40 cycles of 15 s at 95 °C, 20 s at 60 °C, and 20 s at 72 °C. Cycle threshold (Ct) values for each transcript were normalized to the geometric mean of the expression of RPL13A, B2M, and ACTB (i.e., reference genes) and the fold changes were determined by using the 2^-ΔΔCt method. Primers used for quantification of transcripts by real-time quantitative PCR are indicated in SI Appendix, Table S5.

SCID Mouse Infections.

SCID (C.B-17/LcrHsd-Prkdcscid) mice were purchased from Envigo and housed in the Texas A&M Health Science Center (TAMHSC) animal facility. All animal procedures were done in compliance with Texas A&M University Institutional Animal Care and Use Committee (Animal Use Protocol, AUP 2016–0370). Infections were performed as described previously (37). Briefly, 6- to 8-wk-old female mice (SCID or C57BL/6) were infected with 1 × 10⁶ viable C. burnetii phase II strain via intraperitoneal (i.p.) injection. Inoculum concentrations were confirmed by serial dilution spot plating on acidified citrate cysteine medium-2-defined (ACCM-D) agarose as described previously (38).

Mouse Tissue Collection, Processing, and DNA Purification.

At 10 d (competitive infections) or 14 d postinfection (single infections), the mouse spleens were removed and weighed at necropsy to determine splenomegaly (spleen weight/body weight). Each spleen was added to 1 mL PBS and homogenized using an Omni tissue homogenizer equipped with plastic tips. A total of 100 μL of homogenate was added to 400 μL of TriZol LS (Invitrogen) for RNA extraction. For DNA extraction, 100 μL of homogenate was added to 900 μL tissue lysis buffer (Roche) plus 100 μL of proteinase K and incubated at 55 °C overnight. The following day 100 μL of 10% SDS (wt/vol) was added and incubated at room temperature for 1 h. Lysed tissue samples were then processed using the Roche High Pure PCR template preparation kit according to the manufacturer’s instructions.

Enumeration of Coxiella in Mouse Spleens.

DNA purified from infected organs was used as template for TaqMan real-time PCR using primers and probe for com1 or primers and probe of IS1111 as described previously (37). Quantitative PCR was performed in 20-μL reactions with ABI TaqMan universal PCR mastermix run on an ABI StepOne Plus machine. The replication index reported for each mouse was calculated by dividing the number of genome copies recovered from spleens by the number of genome copies in the original inoculum.

Flow Cytometry.

For intracellular human TNF-α/IFN-α4 staining, 5 × 10⁴ THP-1 cells differentiated in phorbol 12-myristate 13-acetate (PMA) (200 ng/mL) for 2 d seeded in 24-well plates were infected with the indicated C. burnetii strain for 72 and 96 h. Cells were then treated with 1 μg/mL of BFA for the last 24 h. The following day, cells were fixed using 2% paraformaldehyde in PBS solution for 20 min at 4 °C. After washing with fluorescence-activated cell sorting (FACS) buffer (1% BSA in PBS solution), cells were permeabilized in FACS buffer supplemented with 0.1% saponin for 30 min at 4 °C and then stained with anti-TNF-α-PE and IFN-α-PE antibodies for 1 h at 4 °C. Infected cells were analyzed based on the GFP fluorescence associated with the strains of C. burnetii. Flow cytometry analyses were performed on a BD FACSCalibur flow cytometer using flow cytometry (CellQuest software, BD Biosciences). FlowJo software (Tree Star) was used to analyze data.

Data Availability.

All data discussed in the paper are available to readers either in the manuscript or in SI Appendix.