|Home | About | Journals | Submit | Contact Us | Français|
The intracellular bacterial pathogen Coxiella burnetii directs biogenesis of a phagolysosome-like parasitophorous vacuole (PV), in which it replicates. The organism encodes a Dot/Icm type IV secretion system (T4SS) predicted to deliver to the host cytosol effector proteins that mediate PV formation and other cellular events. All C. burnetii isolates carry a large, autonomously replicating plasmid or have chromosomally integrated plasmid-like sequences (IPS), suggesting that plasmid and IPS genes are critical for infection. Bioinformatic analyses revealed two candidate Dot/Icm substrates with eukaryotic-like motifs uniquely encoded by the QpH1 plasmid from the Nine Mile reference isolate. CpeC, containing an F-box domain, and CpeD, possessing kinesin-related and coiled-coil regions, were secreted by the closely related Legionella pneumophila Dot/Icm T4SS. An additional QpH1-specific gene, cpeE, situated in a predicted operon with cpeD, also encoded a secreted effector. Further screening revealed that three hypothetical proteins (CpeA, CpeB, and CpeF) encoded by all C. burnetii plasmids and IPS are Dot/Icm substrates. By use of new genetic tools, secretion of plasmid effectors by C. burnetii during host cell infection was confirmed using β-lactamase and adenylate cyclase translocation assays, and a C-terminal secretion signal was identified. When ectopically expressed in HeLa cells, plasmid effectors trafficked to different subcellular sites, including autophagosomes (CpeB), ubiquitin-rich compartments (CpeC), and the endoplasmic reticulum (CpeD). Collectively, these results suggest that C. burnetii plasmid-encoded T4SS substrates play important roles in subversion of host cell functions, providing a plausible explanation for the absolute maintenance of plasmid genes by this pathogen.
Coxiella burnetii is a highly infectious intracellular bacterium that causes Q fever, a zoonotic disease that typically presents as an acute, influenza-like illness. Rare but serious chronic disease can also occur and normally manifests as endocarditis. C. burnetii displays an extensive array of animal reservoirs, with humans exposed to the pathogen primarily via contact with infected domestic livestock. Inhalation of contaminated aerosols is the main route of C. burnetii transmission to humans, with alveolar mononuclear phagocytes considered the pathogen's initial target cell (reviewed in reference 38).
Multiple in vitro studies indicate that C. burnetii replicates in a parasitophorous vacuole (PV) with lysosomal characteristics (34, 65). The early PV interacts with autophagosomes, which may provide nutrients to activate pathogen metabolism (51). Following an initial phagosome stall, the PV fuses with lysosomes and continually engages the endosomal pathway, as indicated by delivery of fluid-phase markers (28). During early stages of infection, C. burnetii differentiates from a nonreplicating small-cell-variant morphological form into a replicating large-cell-variant form (27). To accommodate pathogen growth, the maturing PV expands by continual heterotypic fusion with endolysosomal compartments to ultimately occupy most of the host cell cytoplasm. PV expansion into a spacious structure visible by phase-contrast light microscopy occurs coincident with entry of C. burnetii into the log phase of its growth cycle (~1 to 3 days postinfection [dpi]) (17). The lysosomal character of the PV is shown by the presence of lysosomal membrane proteins, active acid hydrolases, and an acidic lumen (~pH 5) (34, 65).
C. burnetii's historical obligate intracellular nature has stymied attempts to identify pathogen factors required for successful infection. Lipopolysaccharide is the only confirmed virulence determinant of C. burnetii and is thought to shield the bacterial cell surface from innate immune recognition (42, 59). Predicted virulence factors include C. burnetii proteins that modulate host cell processes. For example, C. burnetii protein synthesis is required for PV maturation (33), autophagosome interactions (51), apoptosis subversion (68), and activation of the prosurvival kinases Akt and Erk1/2 (66). Maintenance of host cell viability by induction of prosurvival responses is considered a pathogenic strategy that accommodates C. burnetii's slow growth rate.
C. burnetii protein effectors of host functions are likely delivered to the cytosol by a type IV secretion system (T4SS) with homology to the Dot/Icm machinery of Legionella pneumophila (57, 63). The refractory nature of C. burnetii to genetic manipulation has necessitated using L. pneumophila as a surrogate host to identify Dot/Icm substrates. These proteins often possess eukaryotic motifs/domains predicted to functionally mimic or antagonize the activity of host cell proteins (3, 14, 19, 46, 64). Indeed, multiple C. burnetii proteins with eukaryotic-like ankyrin repeat domains (Anks) are secreted by L. pneumophila in a Dot/Icm-dependent fashion (48, 67). An intriguing subset of secreted bacterial proteins contains eukaryotic F-box domains (2). F-box domain-containing proteins are components of the mammalian E3 ubiquitin ligase enzyme complex, which directs ubiquitination of target proteins, resulting in their proteasome-dependent degradation or functional alteration (2). C. burnetii isolates collectively contain three F-box-encoding open reading frames (ORFs): CBUA0014, CBU0355, and CBU0814 (2, 7).
CBUA0014 is present on the QpH1 plasmid, originally characterized from the C. burnetii Nine Mile reference isolate (55). All C. burnetii isolates examined to date maintain a related autonomously replicating plasmid or have chromosomally integrated plasmid-like sequences (IPS) (6, 29, 44, 55, 56, 62). Nucleotide sequences have been determined for QpH1, QpRS, QpDG, QpDV, and IPS of the G and S isolates (7). Early studies with a limited number of isolates demonstrated a correlation between plasmid content and disease presentation, i.e., isolates derived from ticks, infected animals, and patients with acute Q fever harbored QpH1, while isolates from patients with chronic disease carried QpRS or IPS (56). However, a recent genotyping study of 173 C. burnetii isolates revealed that QpH1 is also carried by some human chronic Q fever isolates, while showing correlations between QpDV and acute infection and QpRS and chronic infection (24). Whether plasmid type confers human disease potential is unresolved. However, genetically distinct C. burnetii isolates can clearly be grouped into pathotypes with different virulence in animal models of acute Q fever (52). Aside from genes involved in plasmid maintenance and segregation, plasmid genes primarily encode hypothetical proteins. Nonetheless, the absolute maintenance of plasmid sequences by all C. burnetii isolates suggests that they are critical for pathogen survival.
The presence of F-box-containing CBUA0014 on QpH1 prompted us to investigate whether this and other plasmid ORFs encode Dot/Icm substrates. Using L. pneumophila as a secretion model, we found that three QpH1-specific proteins (including the CBUA0014 protein) and three of five hypothetical proteins encoded by all C. burnetii plasmids and IPS are translocated into the host cell cytosol by the Dot/Icm T4SS. Importantly, plasmid effectors were also translocated into the cytosol by C. burnetii, demonstrating secretion in a native setting. Effector proteins contained a C-terminal region with similarity to the predicted L. pneumophila Dot/Icm translocation signal, and deletion analysis revealed that this region is necessary for secretion by C. burnetii. All plasmid effector genes were expressed during in vitro C. burnetii infection, and the encoded proteins trafficked to different subcellular sites when ectopically expressed. These results suggest that C. burnetii plasmids play an important role in host cell modifications.
Bacteria used in this study are described in Table Table1.1. C. burnetii Nine Mile phase II, clone 4 (RSA439), was propagated in Vero (African green monkey kidney) cells (CCL-81; ATCC, Manassas, VA) and purified as previously described (16, 58). L. pneumophila strains were cultured on charcoal yeast extract (CYE) agar plates. For plasmid selection, CYE plates contained 10 μg/ml chloramphenicol. For culture of DotA-deficient L. pneumophila LELA3118, plates also contained 25 μg/ml kanamycin. L. pneumophila transformations were conducted as previously described (67). THP-1 human monocytic cells (TIB-202; ATCC) and HeLa (human epithelioid carcinoma) cells (CCL-2; ATCC) were maintained in RPMI 1640 medium (Invitrogen, Carlsbad, CA) containing 10% fetal calf serum (Invitrogen) at 37°C and 5% CO2. THP-1 cells were differentiated into macrophage-like cells by use of phorbol 12-myristate 13-acetate (PMA; EMD Biosciences, San Diego, CA) as previously described (68). Escherichia coli TOP10 (Invitrogen) was used for recombinant DNA procedures. For host cell-free growth of C. burnetii, 6-well plates, T-75 flasks, or 0.2-μm-filter capped Erlenmeyer flasks containing ACCM (47) were inoculated with organisms and incubated at 37°C in a 2.5% O2/5% CO2 environment.
mRNA quantification was performed using the QuantiGene reagent system v.2.0 and custom-designed probes according to the manufacturer's directions (Panomics, Santa Clara, CA). THP-1 cells were incubated with C. burnetii at a multiplicity of infection (MOI) of 25 for 2 h, and then extracellular bacteria were washed from monolayers. Cells were lysed at various time points with QuantiGene lysis buffer supplemented with 150 ng/ml proteinase K (Invitrogen) and solubilized by incubation for 30 min at 55°C, followed by 3 freeze-thaw cycles. Lysates were diluted in QuantiGene lysis buffer and combined with blocking buffer and probes. Lysates were loaded into a 96-well capture plate and incubated for 16 to 20 h at 55°C. mRNA was detected by luminescence over 1,000 ms with a Safire2 microplate reader (Tecan, Mannedorg, Switzerland). mRNA from uninfected THP-1 cells was used to determine the background signal, and this value was subtracted from each infected cell sample. Transcriptional signals were normalized to C. burnetii genome equivalents established as previously described (17).
The plasmid pJB2581 was used for expression of CyaA fusion proteins in L. pneumophila (4). C. burnetii genes were amplified from genomic DNA by PCR using Accuprime Taq polymerase (Invitrogen) and gene-specific primers (Integrated DNA Technologies, Coralville, IA) where the 5′ primer incorporates a BamHI site and the 3′ primer incorporates a SalI or PstI site (see Table S1 in the supplemental material). Products were cloned into pCR2.1-TOPO (Invitrogen), plasmids were digested with either BamHI/SalI or BamHI/PstI (New England BioLabs, Ipswich, MA), and gene-containing fragments were ligated to similarly digested pJB2581 by use of a Ligate-IT system (U.S. Biologicals, Cleveland, OH).
Modified versions of pJB2581, designated pJB-CAT-BlaM and pJB-CAT-CyaA, were constructed for expression of BlaM and CyaA fusion proteins, respectively, in C. burnetii. The chloramphenicol acetyltransferase (CAT) gene was amplified by PCR from pJB2581 using primers CAT-P1169F and CAT-pJB2581-HindIIIrecR. The CBU1169 promoter (P1169) was amplified from C. burnetii genomic DNA using primers P1169-pJB2581-Ab-HindIIIrecF and P1169-R. The CAT gene was placed downstream from P1169 to create P1169-CAT using overlapping PCR and the primers P1169-pJB2581-Ab-HindIIIrecF and CAT-pJB2581-HindIIIrecR. P1169-CAT was cloned into HindIII-digested pJB2581 by use of an In-fusion kit (BD Clontech, Mountain View, CA) to create pJB-CAT. P1169 was then amplified by PCR from C. burnetii genomic DNA using the primers P1169-pJB2581-F and P1169-R and fused by overlapping PCR to either blaM, amplified by PCR from pXDC61 using the primers BlaM-p1169-F and BlaM-pJB-CAT-R, or cyaA, amplified by PCR from pJB2581 using the primers Cya-P1169-F and Cya-pJB-CAT-R. P1169-blaM and P1169-cyaA fragments were cloned into EcoRI/PstI-digested pJB-CAT by use of the In-fusion kit to create pJB-CAT-BlaM and pJB-CAT-CyaA, respectively. pJB-CAT-BlaM and pJB-CAT-CyaA were then used to generate plasmids encoding BlaM or CyaA fused to the N terminus of C. burnetii plasmid effector genes. C. burnetii genes were amplified by PCR using gene-specific primers and cloned into a unique SalI site in pJB-CAT-BlaM or pJB-CAT-CyaA by use of the In-fusion kit.
For green fluorescent protein (GFP) fusion constructs, C. burnetii genes were amplified by PCR with gene-specific primers where the forward primer contains CACC at the 5′ end for directional cloning and a 5′ Kozac sequence (ATGGGC) for mammalian expression. Products were cloned into pENTR-D/TOPO (Invitrogen) and then subcloned into pcDNA6.2/N-GFP using LR Clonase II (Invitrogen). Plasmid constructions were confirmed by sequencing. All plasmids and primers used in the study are listed in Table Table11 and Table S1 in the supplemental material.
ACCM-cultured C. burnetii was collected, washed with 10% glycerol, and then resuspended in 10% glycerol. Ten micrograms of plasmid DNA was mixed with 50 μl of C. burnetii in a 0.1-cm cuvette, and organisms were electroporated as previously described (5). Following electroporation, bacteria were cultured in ACCM for 24 h, and then chloramphenicol (3 μg/ml) was added to cultures for selection of transformants. The use of genes conferring resistance to chloramphenicol and ampicillin for C. burnetii genetic transformation studies at the Rocky Mountain Laboratories (RML) has been approved by the RML Institutional Biosafety Committee and the Centers for Disease Control and Prevention, Division of Select Agents and Toxins.
L. pneumophila transformant cultures were incubated with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) (ICN Biomedicals, Costa Mesa, CA) for 2 h to induce protein expression. Cultures were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting using a mouse monoclonal antibody directed against CyaA (clone 3D1; Santa Cruz Biotechnology, Santa Cruz, CA). Reacting proteins were detected using an anti-mouse IgG secondary antibody conjugated to horseradish peroxidase (Pierce, Rockford, IL) and chemiluminescence using ECL Pico reagent (Pierce). To confirm fusion protein expression by C. burnetii transformants, THP-1 cells (1 × 106 cells/well) were infected for 2 to 3 days and lysed with SDS-PAGE sample buffer, and lysates were immunoblotted. CyaA expression was detected as described above. BlaM expression was assessed using a mouse monoclonal antibody directed against β-lactamase (QED; Bioscience, San Diego, CA).
L. pneumophila CyaA assays were performed as previously described, using a cyclic AMP (cAMP) enzyme immunoassay (GE Healthcare, Piscataway, NJ) (67). For C. burnetii CyaA assays, THP-1 cells (1 × 106 cells/well) were infected with C. burnetii transformants (MOI of 100) for 3 days, cells were lysed, and then lysates were processed as described above for CyaA assays. Positive secretion of CyaA-effector fusion proteins was scored as ≥2.5-fold more cytosolic cAMP then that for cells infected with organisms expressing CyaA alone. In L. pneumophila assays, CyaA fused to C. burnetii AnkG was used as a positive control (48, 67), and confirmation of Dot/Icm-dependent secretion was conducted by repeating the assay with the L. pneumophila DotA− mutant LELA3118.
For BlaM translocation assays, THP-1 cells were cultured in black, clear-bottomed, 96-well plates and infected with C. burnetii transformants (MOI of 100) for 2 days. Monolayers were loaded with the fluorescent substrate CCF4/AM (LiveBLAzer-FRET B/G loading kit; Invitrogen) in a solution containing 15 mM probenecid (Sigma). Cells were incubated in the dark for 1 h at room temperature and then fluorescence quantified on a Safire2 microplate reader, with excitation detected at 405 nm and emission detected at 450 nm. Average fluorescence from 8 uninfected wells was subtracted from results from experimental wells. Results are presented as fold increase in fluorescence above that for cells infected with C. burnetii expressing BlaM alone (negative control). As with the CyaA assay, a fluorescence value ≥2.5-fold above that for the negative control was considered positive for effector translocation. Infected cells processed for β-lactamase activity were also visualized by epifluorescence microscopy using an inverted Nikon TE2000 microscope. Representative images (×20 magnification) were captured with a β-lactamase ratiometric filter set (Chroma Technology, Rockingham, VT) and a Coolsnap HQ2 digital camera (Photometrics, Tucson, AZ) controlled by Metamorph software (Molecular Devices, Inc., Sunnyvale, CA). Images were processed using Adobe Photoshop (Adobe Systems, San Jose, CA).
Uninfected or infected HeLa cells cultured on 12-mm glass coverslips were transfected with individual GFP fusion constructs by use of Effectene reagent (Qiagen). At 18 h posttransfection, cells were processed for fluorescence microscopy. Cells were incubated with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) to stain host and bacterial DNA. A rabbit polyclonal anti-LC3B antibody (Cell Signaling Technology, Danvers, MA) was used to label autophagosomes, mouse monoclonal antiubiquitin (clone FK2; Sigma) was used to identify ubiquitinated proteins, and rabbit polyclonal anticalnexin (Stressgen, Ann Arbor, MI) was used to detect the endoplasmic reticulum (ER). Antibodies were detected with anti-rabbit or anti-mouse secondary antibodies conjugated to Alexa Fluor 488 (Invitrogen). Fluorescence microscopy was performed using a Nikon Ti-U microscope, and images were acquired with a 60× oil immersion objective and a DS-Qi1Mc camera (Nikon, Melville, NY). Images were processed using NIS-Elements software from Nikon.
CBUA0014 is specific to QpH1 (Fig. (Fig.11 A and Table Table2)2) and encodes a protein with a ubiquitination-related F-box domain (2, 7), which comprises 47 amino acids (aa) of this small, 77-aa protein. Immediately downstream of CBUA0014 is QpH1-specific CBUA0015, which encodes a protein with a predicted 55-aa coiled-coil domain (CCD). Common in eukaryotic proteins and increasingly found in secreted bacterial effectors, CCDs are α-helical conformations that mediate protein-protein interactions (8, 20). The CBUA0015 protein also contains a 25-aa region with similarity to kinesin-like protein 1 (KLP1) of Giardia spp. Kinesin is a microtubule-associated motor protein involved in intracellular transport of vesicular cargo (30). Because bacterial effector proteins often mimic host protein activities, the eukaryotic domains in the CBUA0014 and CBUA0015 proteins suggested that these proteins might be Dot/Icm substrates. To test this hypothesis, Legionella pneumophila was used as a surrogate model to assay secretion (48, 67). THP-1 cells were infected with L. pneumophila transformants harboring plasmids encoding the CBUA0014 or CBUA0015 protein N-terminally fused to the C terminus of Bordetella pertussis adenylate cyclase (CyaA) (60). The CyaA assay relies on effector-mediated delivery of the fusion protein to the cytosol, where the CyaA moiety is activated by binding cytosolic calmodulin, resulting in supraphysiological levels of cAMP (60). Elevated cAMP levels were observed in lysates from cells infected with wild-type L. pneumophila, but not a DotA− strain (53), expressing either CyaA fusion protein (Fig. (Fig.1B).1B). Thus, the CBUA0014 and CBUA0015 proteins, now denoted CpeC (Coxiella plasmid effector protein C) and CpeD, respectively, are Dot/Icm substrates. CBUA0016 is QpH1 specific, encodes a hypothetical, hydrophilic protein, and resides in a predicted operon with cpeD (39). This genetic arrangement suggested that the CBUA0016 protein might also be a T4SS substrate. Indeed, the CBUA0016 protein, now designated CpeE, was also translocated into host cells in a Dot/Icm-dependent manner (Fig. (Fig.1B1B).
Genus-specific hypothetical proteins lacking eukaryotic sequence similarity are also common T4SS substrates (23). This is not unexpected considering the broad repertoire of unique host-pathogen interactions. Therefore, we investigated whether five plasmid ORFs (CBUA0006, CBUA0007a, CBUA0012, CBUA0013, and CBUA0023) encoding hypothetical proteins that are conserved among all C. burnetii plasmids (Fig. (Fig.1A)1A) and IPS (6, 7) are Dot/Icm substrates. As shown in Fig. Fig.22 A, infection of THP-1 cells with wild-type, but not DotA−, L. pneumophila expressing the CBUA0006, CBUA0013, and CBUA0023 proteins fused to CyaA resulted in significantly increased cAMP levels, indicating Dot/Icm-dependent translocation into the host cytosol (Fig. (Fig.2A).2A). These proteins are now designated CpeA, CpeB, and CpeF, respectively (Fig. (Fig.2A2A and Table Table2).2). Conversely, the CBUA0007a and CBUA0012 proteins were not translocated (Fig. (Fig.2A).2A). By immunoblotting, all CyaA fusion proteins were equally expressed (data not shown); thus, negative results for CBUA0007a and CBUA0012 were not due to lack of synthesis.
Homologs of C. burnetii T4SS effectors in other bacterial species are unusual (67). While not a confirmed homolog, the N-terminal 250 aa of CpeA show ~50% similarity to the N termini of LPP1878 and LPL0189 from the L. pneumophila Paris and Lens strains, respectively. Therefore, we tested whether these proteins are secreted. As shown in Fig. Fig.2B,2B, both LPP1878 and LPL0189 caused increased levels of cAMP in host cells when expressed in wild-type L. pneumophila, suggesting a potential shared effector protein between these related pathogens.
L. pneumophila Dot/Icm substrate expression is transcriptionally regulated to influence distinct infection events (36). To confirm that the six plasmid effectors are expressed by C. burnetii during infection and to resolve the kinetics of expression, we determined the transcriptional profile of each effector gene over a 7-day time course of infection. This time course represents the pathogen's growth cycle in human macrophages from initial infection through stationary phase (34). As shown in Fig. Fig.3,3, maximal expression of QpH1-specific and conserved plasmid effector genes occurred at 2 and 3 days, respectively, after infection of THP-1 cells. Thus, expression of plasmid effector genes is temporally regulated, with highest expression observed concomitant with early-log-phase growth and rapid PV expansion.
L. pneumophila has been invaluable in identifying C. burnetii T4SS substrates (35, 48, 67). However, direct demonstration of effector secretion by C. burnetii would be ideal. Recent advances in C. burnetii genetic transformation (5) and the discovery that RSF1010 ori-containing plasmids autonomously replicate in the organism (15) allowed us to develop both CyaA and β-lactamase (BlaM) secretion assays for use in C. burnetii. Like the CyaA assay, the BlaM assay is an enzymatic reporter assay that relies on delivery of a BlaM-effector chimera to the cytosol. Here, the BlaM moiety cleaves the β-lactam ring of a cell-loaded fluorescent compound, resulting in blue cytosolic fluorescence (10, 18). The RSF1010 plasmid backbone used to create the C. burnetii expression vectors pJB-CAT-CyaA and pJB-CAT-BlaM (see Fig. S1 in the supplemental material) was derived from pJB2581 (4), the same plasmid used in our L. pneumophila CyaA assays. Fusion protein expression is driven by the C. burnetii CBU1169 promoter, which encodes Hsp20 and is constitutively expressed by C. burnetii (P. A. Beare, unpublished results).
Plasmid effector ORFs were cloned downstream and in frame with cyaA or blaM. THP-1 cells were infected with C. burnetii transformants cultured axenically under antibiotic selection, and then CyaA and BlaM assays were conducted at 3 and 2 dpi, respectively. By immunoblotting, CyaA and BlaM effector fusion proteins were all expressed in THP-1 cells (Fig. (Fig.44 A and data not shown). THP-1 cells infected with C. burnetii expressing CpeA, CpeB, CpeD, CpeE, and CpeF fused to CyaA had ≥2.5-fold-higher levels of cAMP than cells infected with C. burnetii expressing CyaA alone, indicating translocation into the cytosol (67) (Fig. (Fig.4A).4A). CyaA-CpeC induced significantly more cAMP accumulation than CyaA alone; however, the level was slightly below the 2.5-fold cutoff. Based on the same elevation in blue fluorescence relative to the level for THP-1 cells infected with C. burnetii expressing BlaM alone, CpeA, CpeB, CpeC, CpeD, and CpeE were positive for translocation in the BlaM translocation assay (Fig. (Fig.4B).4B). Levels of fluorescence with BlaM-CpeF were significantly higher than levels with BlaM alone but fell below the 2.5-fold cutoff. Furthermore, the CBU0007a and CBU0012 proteins were not secreted using the BlaM assay (data not shown), which corresponds with results from the L. pneumophila CyaA assay. Collectively, as assessed by one or both secretion assays, all plasmid effectors were secreted by C. burnetii during intracellular growth.
Deletion analysis indicates that an intact C terminus is required for translocation of effectors by L. pneumophila's Dot/Icm T4SS (9, 43). Salient features of the putative C-terminal secretion signal include depletion of negative amino acids in positions −1 to −6, enrichment of serine and threonine in positions −3 to −11, and enrichment of hydrophobic amino acids in positions −1 to −3 (9, 43). While a lineup of the 20 C-terminal residues of each C. burnetii plasmid effector protein showed little sequence identity, C termini had the same general features of the putative L. pneumophila translocation signal, including the lack of negatively charged residues and enrichment of hydrophobic residues in positions −1 to −6 (Fig. (Fig.55 A). To test the importance of the C terminus for C. burnetii secretion, BlaM translocation assays were conducted with C. burnetii transformants expressing BlaM-CpeD fusion proteins with C-terminal deletions of 2, 5, 7, or 10 aa. All fusion proteins were equally expressed by C. burnetii following infection of THP-1 cells (data not shown). Deletion of only 2 aa impaired secretion, while deletion of 7 aa or more diminished secretion to near negative-control levels (Fig. (Fig.5B).5B). Thus, C. burnetii Dot/Icm substrates contain a C-terminal region important for translocation.
To probe effector function, we ectopically expressed plasmid T4SS substrates in HeLa cells as fusions to green fluorescent protein (GFP) to visualize the subcellular distribution. This commonly used approach provides clues about bacterial virulence factor function, as effector proteins typically modulate a host factor/process associated with the targeted subcellular site (11, 18, 26, 45, 67). Specific localization of GFP-CpeE and GFP-CpeF was not observed, while GFP-CpeA and GFP-CpeC localized to small cytoplasmic puncta, GFP-CpeB to vesicular structures, including the PV membrane, and GFP-CpeD to filamentous structures (Fig. (Fig.66 and data not shown). Further analysis demonstrated that CpeB trafficked to autophagosomes, as evidenced by colocalization with LC3B (Fig. (Fig.6).6). CpeD partially localized to the endoplasmic reticulum (ER), as indicated by colocalization with calnexin, and caused accumulation and eventual breakdown of the ER network (Fig. (Fig.6).6). Finally, CpeC colocalized with ubiquitin-rich structures (Fig. (Fig.6),6), consistent with the presence of an F-box domain in this protein.
Here, we show that C. burnetii plasmids and IPS encode Dot/Icm T4SS substrates. The functional relevance of C. burnetii plasmids has been elusive, despite description of the molecules over 25 years ago and subsequent nucleotide sequencing (55, 56). An essential role in host cell modification provides the first plausible explanation for absolute maintenance of plasmid sequences by all C. burnetii isolates.
C. burnetii plasmid effectors were initially identified using L. pneumophila as a surrogate model (48, 67). This screen indicated that CpeA, CpeB, CpeC, CpeD, CpeE, and CpeF are secreted in a Dot/Icm-dependent manner. Previous confirmation of C. burnetii secretion of an effector originally identified using L. pneumophila was achieved for AnkF by immunoblotting the soluble cytosolic fraction of infected cells with specific antibody (48). A major accomplishment of the current work is direct demonstration of secretion by C. burnetii using genetic methods. This screen was made possible using an RSF1010 ori-containing plasmid that autonomously replicates in C. burnetii, thereby providing a system for heterologous gene expression (15). When expressed in C. burnetii, five plasmid effectors were positive in the CyaA assay and five were positive in the BlaM assay. We are unsure why CpeC and CpeF were negative in the CyaA and BlaM assays, respectively. Fusion proteins are expressed at similar levels in C. burnetii; thus, negative results are not associated with translational levels. However, native CpeC is ~9 kDa in size, and attachment of CyaA (43 kDa) to this small protein may alter folding and subsequent secretion. In agreement with this hypothesis, CyaA-CpeC was also translocated at low levels in L. pneumophila compared to the other plasmid effectors. The RSF1010 ori plasmid used in this study has a copy number of 3 to 6 in C. burnetii (P. A. Beare and R. A. Heinzen, unpublished data), and effectors are constitutively expressed. Thus, overexpression of effector fusions might inhibit the secretion apparatus, resulting in a negative readout. However, all C. burnetii effector transformants productively infect THP-1 cells to form a typical PV, suggesting that translocation of the effector repertoire required for intracellular growth is unaltered. In the L. pneumophila CyaA assay, isopropyl-β-d-thiogalactopyranoside (IPTG) is added to cultures before infection to induce fusion protein expression. A similar system of temporally induced expression may be optimal for C. burnetii translocation assays. Finally, fusion partners may affect chaperone recognition of some effectors. Little is known about chaperone function in C. burnetii type IV secretion, although some, but not all, C. burnetii CyaA-Ank fusions require the chaperone IcmS for translocation by L. pneumophila. Nonetheless, combined results from both assays indicate that all six plasmid effectors are secreted by C. burnetii.
The C termini of C. burnetii plasmid effectors show amino acid enrichments similar to the proposed C-terminal secretion signal of L. pneumophila Dot/Icm T4SS substrates (9, 43). We previously demonstrated that deletion of 10 amino acids from the C terminus of C. burnetii AnkI eliminates Dot/Icm-mediated secretion by L. pneumophila (67). Here, we show that deletion of only 2 residues from the C terminus of CpeD adversely affects translocation by C. burnetii, with deletion of 7 or 10 aa lowering secretion to near negative-control levels. Collectively, these data demonstrate a critical C-terminal secretion signal in C. burnetii T4SS substrates.
Because Dot/Icm-translocated CpeC, CpeD, and CpeE are specific to QpH1, their functions are strictly associated with isolates that maintain this plasmid (24). Functional analysis of specific and conserved plasmid effectors would be aided by phenotyping respective gene knockouts. Unfortunately, methods of allelic exchange in C. burnetii are currently unavailable. Therefore, to provide clues about effector protein activity, we ectopically expressed plasmid effectors as GFP fusion proteins in mammalian cells and monitored their subcellular localization.
Consistent with the presence of an F-box domain (31), GFP-CpeC traffics to ubiquitin-positive structures throughout the cytoplasm. Ubiquitination mediated by bacterial F-box-containing proteins can modify protein function or target proteins for degradation by the proteasome (2). This process modulates diverse host functions, including innate immune signaling, inflammation, and apoptosis (2). L. pneumophila produces five F-box-containing proteins that are Dot/Icm substrates (22), the most thoroughly characterized being AnkB (LegAU13). AnkB forms a functional Skp-Cullin-F-box E3 ubiquitin ligase complex (37) and is targeted to the cytosolic face of the L. pneumophila-containing vacuole by a process requiring host cell farnesylation (50). The substrate(s) (bacterial or host) targeted by AnkB for ubiquitination is unknown, as is its role in L. pneumophila parasitism, with some (1, 37, 49), but not all (22), strains with inactivated AnkB showing strong infection defects. Further characterization of CpeC-interacting proteins will provide insight into the specific function of this protein.
Ectopically expressed GFP-CpeD partially colocalizes with the ER and disrupts the organelle's dispersed architecture. CpeD contains a region of homology to a kinesin-related Giardia protein. Kinesins are well-characterized motor proteins that direct plus-end vesicular cargo trafficking along microtubules (30) and mediate processes such as Golgi complex-directed secretory transport (69). CpeD also contains a CCD that may mediate protein-protein interactions. When ectopically expressed, CCD-containing Dot/Icm effectors of L. pneumophila can alter Saccharomyces cerevisiae secretory transport and/or cause accumulation of ER-derived structures in CHO cells (9, 18). Recent evidence suggests that C. burnetii manipulates host secretory processes, with Rab1 recruitment to the PV involved in proper vacuole formation (13). Additionally, the ER protein Bip associates with a fractionated PV (32), and the PV membrane decorates with calnexin (D. E. Voth, unpublished results). Thus, CpeD and potentially other Dot/Icm substrates may influence ER function. It should be noted that CpeD was also identified in a proteomic screen of C. burnetii secreted proteins (54).
The third QpH1-specific Dot/Icm substrate, CpeE, showed no specific localization when ectopically expressed. CpeE is a hypothetical protein lacking any obvious homology to known eukaryotic or prokaryotic proteins. CpeE was originally identified as a hydrophilic protein specific to acute Q fever isolates and was termed CbhE′ in reference to unique restriction mapping on QpH1 (40, 41). cpeE and cpeD reside in a putative operon (39), a prediction supported by similar expression profiles. Like that for cpeC, maximal expression of both genes occurs at 2 days postinfection, coincident with the onset of rapid PV enlargement. Thus, CpeE and CpeD may act together to modulate PV formation or other host cell functions.
Of the five hypothetical proteins encoded by all C. burnetii plasmids and IPS, three (CpeA, CpeB, and CpeF) are Dot/Icm substrates. This effector group is maximally expressed at 3 days postinfection, 1 day later than QpH1-specific effectors. CpeA is similar to two hypothetical L. pneumophila proteins, LPL0189 and LPP1878, that we demonstrate are also Dot/Icm substrates. The relatedness of these three secreted proteins may correlate with a common biological function during intracellular growth. Only GFP-CpeB showed specific subcellular trafficking in localizing to LC3B-positive autophagosome-derived vesicles, including the PV membrane. Autophagosomes normally remove unwanted material, such as damaged organelles, from the host cell cytoplasm and mature into autophagolysosomes, where this material is degraded (61). Numerous intracellular pathogens subvert autophagic signaling to provide a suitable growth environment (12). For example, Mycobacterium tuberculosis inhibits autophagosome/lysosome fusion to avoid lysosomal destruction (21). Conversely, exogenous induction of autophagy benefits C. burnetii replication (25, 51), and the organism actively engages autophagosomes a few minutes after infection (25, 51). It is intriguing to speculate that CpeB benefits C. burnetii by acting alone, or in concert with other effectors, to induce autophagy.
In summary, C. burnetii plasmids and IPS are enriched in genes encoding Dot/Icm substrates. Functional characterization of these proteins will provide needed insight into C. burnetii host cell parasitism. Host cell-free cultivation of C. burnetii (47) may provide a means to cure the pathogen of its plasmid. Generation of an isogenic strain lacking plasmid sequences will help define the role of these molecules in host cell infection and pathotype-specific virulence.
We thank Anita Mora for graphic illustrations, Katja Mertens and James Samuel for unpublished information regarding the autonomous replication of RSF1010 ori-containing plasmids in C. burnetii, and Seth Winfree for BlaM microscopy. We thank Raphael Valdivia and Sergey Konstantinov for helpful discussions and Xavier Charpentier for the generous gift of L. pneumophila Paris and Lens genomic DNA.
This work was supported by funding from NIH NIAID grants K22AI081753 (D.E.V.) and R01AI087669 (D.E.V.), the Arkansas Biosciences Institute (D.E.V.), and the Intramural Research Program of the National Institutes of Health, National Institute of Allergy and Infectious Diseases (R.A.H.).
Published ahead of print on 7 January 2011.
†Supplemental material for this article may be found at http://jb.asm.org/.