Sec-mediated secretion by Coxiella burnetii

Coxiella burnetiiproteins are present in growth medium supernatant

The Dot/Icm type IVB secretion system of C. burnetii has been extensively studied [9, 10, 39]. However, little is known about other secretion systems of C. burnetii that are presumably important for intracellular parasitism. To determine if C. burnetii secretes proteins during axenic growth, bacteria were cultivated in ACCM-2 without neopeptone to eliminate media proteins. Following 7 days of growth, supernatant was concentrated and analyzed by SDS-PAGE and silver staining (Figure 1). Many proteins were detected, with the majority having a molecular weight below 20 kDa. In a discovery experiment to generate a list of potentially secreted proteins to further investigate, SDS-PAGE was conducted again and proteins stained with Coomassie G-250 to allow analysis by microcapillary reverse-phase HPLC nano-electrospray tandem mass spectrometry (μLC/MS/MS). A list of 105 proteins was generated (Additional files 1 and 2) with functions assigned based on the annotated genome of the C. burnetii Nine Mile RSA493 reference strain [18]. Sixteen proteins were annotated as hypothetical exported proteins, which represents 36% of the total proteins with this annotation in the predicted C. burnetii proteome [18]. Twenty-nine proteins, such as translation initiation factor 1 (InfA) and ribosomal protein subunit L31P (RpmE), were predicted as cytoplasmic using the PSORTb v3.0.2 bacterial protein subcellular localization prediction program [40]. This result could be explained by a small amount of bacterial lysis releasing abundant cytoplasmic proteins that are then detected by highly sensitive mass spectrometry. The only Dot/Icm type IVB secretion system substrate detected was CBU0937 [39]. However, type IVB-dependent secretion of CBU0937 was demonstrated using L. pneumophila as a surrogate host, and the protein contains a predicted signal sequence, which are typically not associated with Dot/Icm type IVB effectors [41]. Thus, CBU0937 may represent a false positive type IVB effector. Nonetheless, the lack of identified C. burnetii Dot/Icm type IVB secretion system substrates in culture supernatants indicates secretion via this mechanism requires host cell-derived signals.

Expression of epitope-tagged proteins by C. burnetiitransformants confirms secretion

To confirm active secretion of proteins by C. burnetii into growth media, we generated 55 genetic transformants expressing individual proteins, under the control of an inducible TetA promoter, that contain a C-terminal 3xFLAG-tag (Additional file 2). Proteins identified by mass spectometry were selected for epitope-tagging based on predictions obtained using PSORTb, TMHMM [42], SignalP 3.0 [43], BLAST and PubMed bioinformatics tools. Each protein was first analyzed by a BLAST search to identify potential homologs. If a homolog was identified, PubMed searches were conducted to determine if the function and/or the cellular location of the homolog had been characterized. The predicted cellular location was also obtained using PSORTb, TMHMM and SignalP. Based on these analyses, proteins that were unlikely to be secreted, such as malate dehydrogenase, were eliminated from further study.

Expression of FLAG-tagged proteins by C. burnetii transformants was induced by addition of anhydrotetracycline (aTc) following 48 h of growth of individual transformants in ACCM-2. C. burnetii and culture supernatants were harvested 24 h later. Immunoblotting of culture supernatants with anti-FLAG antibody confirmed secretion of 27 of the 55 candidate proteins (Figure 2, Table 1 & Additional file 3). FLAG-tag positive bands were not due to cell lysis as bands were not observed following probing of individual supernatants with antibody directed against EF-Ts, an abundant cytoplasmic protein (Figure 2 & Additional file 3). To ensure negative secretion was not due to a lack of protein expression, bacterial pellets were also analyzed by immunoblotting using the anti-FLAG antibody. With the exception of CBU0089a, CBU1138, CBU1681, and CBU2027, expression of all tagged proteins was confirmed (Additional file 3).

All 27 secreted proteins contained a predicted signal peptide, with 19 annotated as hypothetical proteins (Table 1). This is not surprising given the unique host-pathogen relationship of C. burnetii and the fact that 40.3% of the open reading frames of the Nine Mile reference strain encode hypothetical proteins [18]. Secretion of proteins annotated as enhanced entry proteins (EnhB1, EnhB2 and EnhA5) was confirmed by the FLAG-tag assay. These proteins are homologous to L. pneumophila proteins originally thought to facilitate pathogen entry into host cells (EnhA, B & C) [44]. However, a more recent study of L. pneumophila EnhC demonstrates a role for this protein in peptidoglycan remodeling [45]. Secretion of Com1 and FkpA (Mip) was confirmed, both of which also have homologs in L. pneumophila. Little is known about their roles in C. burnetii pathogenesis, although Com1 is known to be outer membrane associated [46] and FkpA has peptidyl-prolyl cis-trans isomerase (PPIase) activity [47]. The three remaining secreted proteins with predicted functions were ArtI (CBU0482), an arginine-binding protein, SodC (CBU1822), a Cu-Zn superoxide dismutase, and a M16 family peptidase (CBU1902).

C. burnetiisecretes FLAG-tagged proteins during growth in host cells

We next examined whether proteins secreted by C. burnetii during axenic growth were also secreted during growth in mammalian host cells. Vero cells were infected for 5 days with C. burnetii transformants expressing the FLAG-tagged secreted proteins CBU0110, CBU1135 or CBU1984. aTc was added to induce protein expression, then infected cells lysed 18 h later with 0.1% Triton X-100, which solubilizes host cell membranes, but not C. burnetii[13]. Cell lysates were centrifuged, then the pellets (containing C. burnetii and host cell debris) and supernatants were analyzed by immunoblotting using α-FLAG and α-EF-Ts antibodies (Figure 3). FLAG-tagged proteins were detected in the supernatant, indicating secretion occurs during host cell infection. FLAG-tagged proteins were also present in the bacterial pellet showing the rate of protein synthesis is greater than the rate of secretion. EF-Ts was only detected in the pellet, thereby eliminating bacterial lysis as a source of FLAG-tagged protein in supernatants.

Secretion of FLAG-tagged proteins requires an intact signal sequence

All verified secreted proteins contained a predicted N-terminal signal sequence. Signal sequences direct transport of proteins across the inner membrane via the Sec translocase [48]. To determine if transport to the periplasm was necessary for secretion, the secreted proteins CBU0110, CBU0915, CBU1135, CBU1173 and CBU1984 were expressed as before, but without their signal sequences. Immunoblotting for C-terminal FLAG-tags revealed that each of the five proteins was present in cell pellets, but not culture supernatants (Figure 4). Thus, a signal sequence, and therefore, a transient periplasmic location is necessary for secretion.

Potential secretion mechanisms

C. burnetii Sec-mediated secretion could occur by the mechanisms depicted in Figure 5. Type I-like secretion is predicted by the presence of a tolC homolog (CBU0056) in the C. burnetii genome. Genome analysis also makes T4P-mediated secretion conceivable as 13 T4P genes are present in the C. burnetii Nine Mile reference strain genome (Additional file 4). Eleven of these genes share homologs with the T4P genes of F. novicida, a bacterium that employs T4P-mediated secretion (Additional file 4). However, we did not detect pili on the surface of C. burnetii using a procedure that visualized pili on F. tularensis LVS [49] (Additional file 5). OMVs are produced by a large variety of microbes [50]. Figure 6 depicts what appear to be C. burnetii OMVs being produced by bacteria growing in media and within Vero cells, suggesting OMVs contribute to Sec-mediated secretion of proteins by C. burnetii.

C. burnetiiand mammalian cell lines

C. burnetii Nine Mile phase II (RSA439, clone 4) was used in these studies [62]. For general bacterial culture, organisms were propagated microaerobically in ACCM-2 + 1% fetal bovine serum (FBS, Invitrogen) at 37°C [37]. E. coli TOP10 (Invitrogen) or Stellar™ (BD Clontech) cells were used for recombinant DNA procedures and cultivated in Luria-Bertani (LB) broth. E. coli transformants were selected on LB agar plates containing 10 μg/ml of chloramphenicol. African green monkey kidney (Vero) cells (CCL-81; ATCC) were cultured using RPMI 1640 medium (Invitrogen) containing 10% FBS (Invitrogen).

SDS-PAGE and silver staining of C. burnetiiculture supernatants

Two 40 ml C. burnetii cultures in ACCM-2 lacking neopeptone were grown in 125 ml Erlenmyer flasks for 7 days with shaking at 75 rpm. The bacteria were combined and pelleted by centrifugation for 5 min at 20,000 × g, then the supernatant was passed through a 0.22 μm syringe filter before being concentrated ~400-fold using a 3000 MWCO centrifugal filter (Millipore). The concentrated supernatant was separated by SDS-PAGE using a 16.5% gel and visualized by staining with the Silver Quest kit (Invitrogen).

Microcapillary reverse-phase HPLC nano-electrospray tandem mass spectrometry (μLC/MS/MS)

Five 40 ml C. burnetii cultures in ACCM-2 lacking neopeptone were grown in 125 ml Erlenmyer flasks for 7 days with shaking at 75 rpm. The bacteria were combined and pelleted, then the supernatant passed through a 0.22 μm syringe filter before being concentrated ~500-fold using a 3000 MWCO centrifugal filter. The concentrated supernatant was separated by SDS-PAGE using a 16.5% gel and visualized by staining with Coomassie G-250-based SimplyBlue SafeStain (Invitrogen). The protein containing lane was cut into 10 equal sections that were washed twice with 50% acetonitrile, then stored at -20°C prior to shipping to the Harvard Mass Spectrometry and Proteomics Resource Laboratory, FAS Center for Systems Biology, Northwest Bldg Room B247, 52 Oxford St, Cambridge MA. Gel sections were subjected to tryptic digestion and the resulting peptides sequenced by tandem mass spectrometry. Peptides were analyzed by microcapillary reverse-phase HPLC, directly coupled to the nano-electrospray ionization source of an LTQ-Orbitrap XL mass spectrometer (LCMSMS). Using a custom version of Proteomics Browser Suite (PBS; ThermoFisher Scientific), MS/MS spectra were searched against the C. burnetii subset of the NCBInr protein database concatenated to sequences of common laboratory contaminants. Methionine was allowed a variable modification for methionine sulfoxide and cysteine a fixed modification of carboxyamidomethyl cysteine. Peptide-spectrum matches were accepted with PBS filter sets to attain an estimated false discovery rate of <1% using a decoy database strategy. Searches were performed with 2 missed cleavages, semi-tryptic, at 30 ppm mass tolerance, accepting only +/- 2.5 ppm. A minimum of 2 unique peptides were required to identify a protein.

Construction of pJB-CAT-TetRA-3xFLAG

The TetRA promoter/operator fragment was PCR amplified from pMiniTn7T-CAT::TetRA-icmDJB [9] using Accuprime Pfx (Invitrogen) and the primers TetRA-pJB-F and TetRA-3xFLAG-R obtained from Integrated DNA Technologies (Additional file 6). pJB-CAT-P1169-3xFLAG [63] was digested with EcoRI and PstI (New England Biolabs) to remove the P1169 promoter that was replaced with the TetRA fragment using the In-Fusion PCR cloning system (BD Clontech).

Construction of plasmids encoding C-terminal FLAG-tagged proteins and transformation of C. burnetii

Genes were PCR amplified with Accuprime Pfx and the primer sets listed in Additional file 6. SignalP 3.0 [43] was used to determine the location of signal sequences for the cloning of genes lacking this sequence. pJB-CAT-TetRA-3xFLAG was digested with PstI (New England Biolabs) followed by insertion of gene-encoding PCR products using the In-Fusion PCR cloning system (BD Clontech). C. burnetii was transformed with plasmid constructs as previously described [37].

Immunoblotting of C. burnetiitransformant culture supernatants

Transformed C. burnetii expressing C-terminal 3xFLAG-tagged proteins were cultivated in ACCM-2 + 1% FBS for 48 h, then expression of tagged proteins induced by addition of anhydrotetracycline (aTc, final concentration = 50 ng/ml). Cell pellets and growth medium were collected 24 h after induction. One milliliter of supernatant from each sample was concentrated by trichloroacetic acid (TCA) precipitation (17% final TCA concentration) prior to analysis by immunoblotting. Detection of proteins present in ACCM and/or the bacterial pellet was conducted by immunoblotting following separation of proteins by SDS-PAGE using a 4-20% gradient gel (Pierce). Nitrocellulose membranes were incubated with monoclonal antibodies directed against FLAG (Sigma) or elongation factor Ts (EF-Ts; a generous gift of James Samuel, Texas A&M University) [64]. Reacting proteins were detected using anti-mouse IgG secondary antibodies conjugated to horseradish peroxidase (Pierce) and chemiluminescence using ECL Pico or Femto reagent (Pierce).

Ex vivosecretion assay

The assay was performed essentially as described by Pan et al.[13]. Briefly, Vero cells cultured in 6-well tissue culture plates were infected for 5 days with C. burnetii expressing 3xFLAG-tagged proteins under the control of a TetA promoter. Protein expression was then induced with aTc (final concentration = 400 ng/ml) for 18 h. Cells were lysed with 0.1% Triton X-100 plus protease inhibitor cocktail (Sigma) in 1× phosphate buffered saline (1.5 mM KH₂PO₄, 2.7 mM Na₂HPO₄-7H₂O, 155 mM NaCl, [pH 7.2]). Lysates were centrifuged for 10 min at 16,000 × g and the supernatant passed through a 0.22 μM syringe filter before TCA precipitation. Pellet and supernatant samples were separated by SDS-PAGE, transferred to nitrocellulose and probed with anti-FLAG and anti-EF-Ts antibodies.

Transmission electron microscopy (EM) of C. burnetiigrown in ACCM-2

C. burnetii was grown in ACCM-2 for 2 or 6 days, then the cells were pelleted and fixed in 2.5% (vol/vol) glutaraldehyde with 0.05 M sucrose in 0.1 M sodium cacodylate buffer for 2 h. Cells were post fixed in 0.5% reduced osmium using a Pelco Biowave microwave (Ted Pella) at 250 W under a 15-in Hg vacuum (all other chemical steps retained these settings) for 2 min on/2 min off/2 min on. Next, tannic acid (1%) was added and samples microwaved, followed by addition of 1% uranyl acetate and microwaving. Samples were dehydrated in a graded ethanol series for 1 min under vacuum and infiltrated with 1:3, 1:1, and 3:1 (Epon/Araldite resin/ethanol), microwaved for 5 min on/5 min off/5 min on, then finally embedded in Epon/Araldite resin. Thin sections (80 nm) were cut using a Leica UC6 (Leica Microsystems) and sections stained with 1% uranyl acetate. Samples were viewed on a Hitachi H-7500 transmission electron microscope (Hitachi) at 80 kV, and digital images were acquired with a Hamamatsu XR-100 digital camera system (AMT).

Scanning EM of C. burnetiiinfected Vero cells

Vero cells infected with C. burnetii for 48 h were fixed, postfixed, and dehydrated as described for transmission EM except that 1% reduced osmium was used for postfixation. Samples were then dried to the critical point in a Bal-Tec cpd 030 drier (Balzer). Cells were dry-fractured by very lightly applying a small piece of adhesive tape to the apical surface that was subsequently gently removed. Cells were coated with 75 Å of iridium in an IBS ion beam sputter (South Bay Technology). Samples were imaged on a Hitachi S-4500 scanning electron microscope (Hitachi).

Transmission EM of negative stained C. burnetii and F. tularensisLVS

A fixation and staining protocol optimized for preservation and visualization of pili was employed. F. tularensis subsp. holarctica Live Vaccine Strain (LVS) from a frozen stock was streaked onto a modified Mueller-Hinton plate that was incubated for 48 h at 37°C, 7% CO₂. Two milliliters of Chamberlain’s defined medium was inoculated with F. tularensis LVS at 0.1 OD/ml and grown ~16 h at 37°C, 200 rpm. The cells were pelleted, washed 2× with 1× PBS, then fixed with 4% paraformaldehyde (PFA). C. burnetii was cultured for 4 days in ACCM-2 + 1% FBS, the cells pelleted, washed 2× with 1× PBS, then fixed with 4% PFA. A 5 μl aliquot of fixed bacteria was allowed to settle on a formvar/carbon-coated grid for 5 min. Liquid was removed with filter paper and the samples washed with dH₂O. Samples were stained with 2% ammonium molybdate for 2 min. Remaining stain was removed with filter paper. Samples were viewed on a Hitachi H-7500 transmission electron microscope (Hitachi) at 80 kV, and digital images were acquired with a Hamamatsu XR-100 digital camera system (AMT).

Availability of supporting data

All supporting data are included as additional files.