The Coxiella burnetii type IV secretion system substrate CaeB inhibits intrinsic apoptosis at the mitochondrial level

Authors

  • Leonie Klingenbeck,

    1. Mikrobiologisches Institut – Klinische Mikrobiologie, Immunologie und Hygiene, Universitätsklinikum Erlangen, Friedrich-Alexander Universität Erlangen-Nürnberg, Erlangen, Germany
    Search for more papers by this author
  • Rita A. Eckart,

    1. Mikrobiologisches Institut – Klinische Mikrobiologie, Immunologie und Hygiene, Universitätsklinikum Erlangen, Friedrich-Alexander Universität Erlangen-Nürnberg, Erlangen, Germany
    Search for more papers by this author
  • Christian Berens,

    1. Lehrstuhl für Mikrobiologie, Department Biologie, Friedrich-Alexander Universität Erlangen-Nürnberg, Erlangen, Germany
    Search for more papers by this author
  • Anja Lührmann

    Corresponding author
    • Mikrobiologisches Institut – Klinische Mikrobiologie, Immunologie und Hygiene, Universitätsklinikum Erlangen, Friedrich-Alexander Universität Erlangen-Nürnberg, Erlangen, Germany
    Search for more papers by this author

For correspondence. E-mail anja.luehrmann@uk-erlangen.de; Tel. (+49) 9131 8522577; Fax (+49) 9131 85 1001.

Summary

Manipulation of host cell apoptosis is a virulence property shared by many intracellular pathogens to ensure productive replication. For the obligate intracellular pathogen Coxiella burnetii anti-apoptotic activity, which depends on a functional type IV secretion system (T4SS), has been demonstrated. Accordingly, the C. burnetii T4SS effector protein AnkG was identified to inhibit pathogen-induced apoptosis, possibly by binding to the host cell mitochondrial protein p32 (gC1qR). However, it was unknown whether AnkG alone is sufficient for apoptosis inhibition or if additional effector proteins are required. Here, we identified two T4SS effector proteins CaeA and CaeB (C. burnetii anti-apoptotic effector) that inhibit the intrinsic apoptotic pathway. CaeB blocks apoptosis very efficiently, while the anti-apoptotic activity of CaeA is weaker. Our data suggest that CaeB inhibits apoptosis at the mitochondrial level, but does not bind to p32. Taken together, our results demonstrate that C. burnetii harbours several anti-apoptotic effector proteins and suggest that these effector proteins use different mechanism(s) to inhibit apoptosis.

Introduction

Coxiella burnetii is a Gram-negative, obligate intracellular pathogen that causes Q-fever, a worldwide zoonotic disease (Maurin and Raoult, 1999). Infection in humans occurs by inhalation of infectious material transmitted from domestic livestock, and infection by as few as 10 bacteria can result in disease (Baca and Paretsky, 1983). Q-fever is often asymptomatic or a mild flu-like illness, but can develop into an atypical pneumonia or hepatitis. Furthermore, the disease can be associated with chronic or even fatal consequences, usually in form of a bacterial endocarditis (Maurin and Raoult, 1999). In natural infections, C. burnetii is mainly found in mononuclear phagocytes (Stein et al., 2005), whereas in vitro, this pathogen infects a variety of cell types, including epithelial, fibroblast and monocyte/macrophage-like cell lines (Voth and Heinzen, 2007). Importantly, establishing an intracellular niche for replication requires bacterial protein synthesis (Howe and Mallavia, 2000; Romano et al., 2007), suggesting direct involvement of bacterial proteins. Sequencing of the C. burnetii genome revealed the presence of genes encoding a type IV secretion system (T4SS) related to the Dot/Icm system of Legionella pneumophila (Seshadri et al., 2003). Recently, it was shown that C. burnetii requires this Dot/Icm system to ensure its survival inside host cells (Beare et al., 2011; Carey et al., 2011) and notably for apoptosis protection (Beare et al., 2011). C. burnetii has been shown to interfere with host cell apoptosis (Lührmann and Roy, 2007; Voth et al., 2007). How this occurs mechanistically is not understood, but it was hypothesized that effector proteins translocated into the host cell by the T4SS are required for apoptosis protection. This hypothesis is supported by the fact that T4SS effector proteins from other pathogenic bacteria have anti-apoptotic activities. The Bartonella protein BepA was found to interfere with apoptosis by a process that involved increasing cytosolic cAMP levels (Schmid et al., 2006). The L. pneumophila effector SdhA is required for inhibition of apoptosis, but the mechanism is not well understood (Laguna et al., 2006). Another L. pneumophila effector, SidF, has been shown to contribute to apoptosis resistance through interactions with the host BH3-only proteins BNIP3 and Bcl-rambo (Banga et al., 2007). Indeed, an anti-apoptotic C. burnetii effector protein, AnkG, has been identified. AnkG was shown to inhibit pathogen-induced apoptosis, probably by binding to the pro-apoptotic host cell protein p32 (Lührmann et al., 2010). However, whether AnkG is the only anti-apoptotic effector protein or whether it works in concert with other effector proteins is unknown. Functional redundancy has been demonstrated among L. pneumophila effector proteins (Ragaz et al., 2008), and this functional redundancy was suggested to be an important property enabling pathogens to enhance infection efficiency (Hubber and Roy, 2010). It is very likely that C. burnetii effector proteins are also functionally redundant. Therefore, we speculated that C. burnetii harbours additional anti-apoptotic effector proteins. So, we analysed 18 recently identified C. burnetii effector proteins for such an activity (Carey et al., 2011).

Apoptosis is a programmed cell death pathway crucial for immune system maintenance and removal of damaged or infected cells (Byrne and Ojcius, 2004). Two main pathways lead to apoptosis: the extrinsic and the intrinsic pathway.

The extrinsic cell death pathway is initiated in response to stimulation of death receptor proteins at the cell surface by extracellular stimuli, including tumour necrosis factor (TNF) and Fas ligand (FasL or CD95L), whose cognate receptors are TNF receptor I and Fas ligand receptor (Fas, APO-1, CD95) respectively (Suda et al., 1993). Binding to Fas causes assembly of the death-inducing signalling complex (DISC), containing pro-caspase 8 and the death-receptor adapters FADD/TRADD (Boldin et al., 1995; Kischkel et al., 1995; Jin and El-Deiry, 2005). Activation of caspase 8 mediates apoptosis either directly by activating the effector caspases 3 and 7 (Strasser et al., 2009) or through proteolysis of the BH3-only member BID, generating truncated BID (t-BID) which translocates to the mitochondria and activates mitochondrial (intrinsic) apoptosis (Li et al., 1998; Luo et al., 1998).

The intrinsic cell death pathway involves activation of Bax and Bak, which is regulated by the Bcl-2 protein family (Willis and Adams, 2005; van Delft and Huang, 2006), comprised of positive and negative regulators of apoptosis. Anti-apoptotic members include Bcl-2, Bcl-XL, Mcl-1, A1 and Bcl-w. The pro-apoptotic members, Bad, Bik, Hrk, Bim, Noxa and Puma belong to the BH3-only proteins. The ratio of positive and negative apoptosis regulators expressed in a cell plays a critical role in activating Bax and Bak. The finding that alternatively spliced variants of several members of the Bcl-2 family have different activities with respect to their pro- or anti-apoptotic potential adds to the complexity of apoptosis regulation (Boise et al., 1993; Marani et al., 2002; Renshaw et al., 2004). Once activated, Bax and Bak oligomerize and permeabilize the mitochondrial membrane, resulting in the release of cytochrome C and activation of caspase 9 through the apoptosome (Adams and Cory, 2002). Activated caspase 9 leads to caspase 3 and 7 activation. These key downstream effectors initiate robust proteolysis and free a dedicated DNase that fragments chromatin (Enari et al., 1998; Cory and Adams, 2002).

Here, we demonstrate that C. burnetii harbours additional anti-apoptotic T4SS effector proteins. Our study identified two effector proteins, named CaeA and CaeB, that inhibit the intrinsic apoptotic pathway. Ectopic expression of CaeB results in very robust inhibition of staurosporine- and UV-light-induced apoptosis. However, changes in the protein levels of Bcl-2 proteins were not observed. By activating defined steps in the apoptosis cascade using an inducible expression system, we narrowed down the point of action for CaeB to downstream of Bax activation and upstream of caspase 3 activation. As ectopically expressed CaeB localizes to mitochondria (Carey et al., 2011), we analysed the mitochondrial outer membrane permeabilization (MOMP) in cells stably expressing CaeB. We detected repressed MOMP as a result of CaeB expression. However, CaeB did not interfere with mitochondrial targeting of Bax, suggesting that CaeB inhibits apoptosis at the mitochondrial level.

Results

Several C. burnetii effector proteins interfere with apoptosis induction

So far, AnkG is the only C. burnetii effector protein known to inhibit host cell apoptosis (Lührmann et al., 2010). One open question regarding AnkG function was whether other effector proteins contribute to the C. burnetii-mediated anti-apoptotic activity (Broederdorf and Voth, 2011). We addressed this question by analysing recently identified T4SS effector proteins (Carey et al., 2011) for their interference with intrinsic apoptosis. Therefore, we ectopically produced these effector proteins transiently in CHO cells and measured apoptosis after stimulation with staurosporine, a broad protein kinase inhibitor and a potent inducer of intrinsic apoptosis. Nuclear fragmentation visualized by DAPI staining was counted to measure apoptosis. The expression of some of the GFP-tagged effector proteins displayed an even higher percentage of apoptotic cells after staurosporine treatment compared with cells expressing GFP. However, neither the expression of GFP, nor the expression of the GFP fusion proteins caused any change in nuclear morphology. This suggested that none of the expressed proteins promoted pro-apoptotic activity, but that the expression of some of these proteins might amplify an apoptotic signal. Furthermore, as shown in Fig. 1A and B, only the expression of GFP-CBU1524 and -CBU1532 protected the cells significantly from staurosporine-induced apoptosis compared with cells expressing GFP alone.

Figure 1.

Identification of two anti-apoptotic C. burnetii effector proteins.

A. CHO-FcR cells were transiently transfected with plasmids encoding GFP as control or the indicated GFP-tagged C. burnetii effector proteins followed by treatment with or without staurosporine (2 μg ml−1) for 4 h at 37°C in 5% CO2. The cells were fixed, permeabilized and the nuclei were stained with DAPI, before scoring the nuclear morphology of GFP-expressing cells. Data represent average values ± SD of 100 counted nuclei from GFP-expressing cells per sample from three independent experiments. *P < 0.05, **P < 0.01.

B. Representative immunofluorescence micrographs show CHO-FcR cells transiently transfected with GFP, GFP-CBU1524 and GFP-CBU1532. The cells were treated with or without 1 μM staurosporine for 6 h at 37°C in 5% CO2 followed by fixation, permeabilization and staining of the nuclei with DAPI (blue).

C. HEK293 cells stably expressing GFP as control, GFP-CaeA or GFP-CaeB were incubated with the indicated staurosporine concentrations for 4 h at 37°C in 5% CO2. Proteins were separated by SDS-PAGE, transferred to a PVDF membrane and probed with anti-cleaved PARP antibody to visualize cell death or anti-actin as protein loading control. Depicted is one representative experiment out of two independent experiments with similar results.

D. HEK293 cells stably expressing GFP, GFP-CaeA or GFP-CaeB were exposed to UV-light (200 J m−2 or 800 J m−2) and incubated for 6 h at 37°C in 5% CO2. Proteins were separated by SDS-PAGE, transferred to a PVDF membrane and probed with anti-cleaved PARP antibody to visualize cell death or anti-actin as protein loading control. The result of one representative experiment out of three independent experiments with similar results is shown.

Structural features and host cell localization phenotypes of the anti-apoptotic effector proteins CBU1524 and CBU1532

CBU1524 (C. burnetii anti-apoptotic effector A, CaeA) displayed nuclear localization when ectopically expressed (Carey et al., 2011 and Fig. 1B) and contains a predicted coiled-coil region and several predicted nuclear localization sequences (NLS). CBU1532 (CaeB) displayed mitochondrial localization when ectopically expressed (Carey et al., 2011). A comparison of the C. burnetii Nine Mile genome to the genome of the C. burnetii Dugway strain isolated from a rodent (Stoenner and Lackman, 1960), which does not appear to cause clinical disease, revealed nucleotide deletions in the C. burnetii Nine Mile CaeA und CaeB genes leading to shorter proteins compared with the homologues found in the C. burnetii Dugway strain (Carey et al., 2011). Thus, in C. burnetii Nine Mile, which was used in this study, CaeA is 25.1 kDa and CaeB 16.5 kDa while the homologues in C. burnetii Dugway are 36.9 kDa and 49.7 kDa respectively. Importantly, it was shown previously that the genes encoding CaeA and CaeB in C. burnetii Nine Mile are transcribed during infection. In addition, both proteins are translocated in a T4SS-dependent manner, demonstrating that these proteins are bona fide effector proteins (Carey et al., 2011).

CaeA and CaeB have different anti-apoptotic properties

Next, the anti-apoptotic activity of these two effector proteins in another cell line with two different apoptotic inducers was evaluated. Thus, HEK293 cells stably expressing GFP, GFP-CaeA or GFP-CaeB were analysed for their ability to interfere with staurosporine- or UV-light-induced apoptosis. UV-light induces DNA damage that results in the induction of intrinsic apoptosis (Kulms and Schwarz, 2002). Apoptosis was measured by assaying the presence of cleaved PARP (poly ADP-ribose polymerase) by immunoblotting. Proteolytic cleavage of nuclear PARP inactivates DNA repair activity and is a marker for the terminal stages of apoptosis. As shown in Fig. 1C and D, GFP-expressing cells contained cleaved PARP, which accumulated with increasing staurosporine concentration or increased UV-light strength. In contrast, the expression of CaeB completely blocked PARP cleavage under both conditions, while the expression of CaeA substantially reduced PARP cleavage after UV-light induction, but did not inhibit staurosporine-induced PARP cleavage. These results indicate that both effector proteins target the intrinsic apoptosis pathway. However, as the expression level of GFP-CaeA is higher than the expression level of GFP-CaeB (Fig. S1A and B), the difference in activity cannot be explained by the expression level, but rather by a difference in the anti-apoptotic mechanism(s).

CaeB is a potent apoptosis inhibitor

To determine how potent CaeB inhibits apoptosis, we intensified the apoptotic stimulus by increasing the staurosporine concentrations and the incubation times. Apoptosis was again measured by the presence of cleaved PARP assayed by immunoblotting. Increasing staurosporine concentrations resulted in stronger signals of cleaved PARP in GFP-expressing cells. In clear contrast, we did not observe PARP cleavage even with a concentration of 4 μM staurosporine when GFP-CaeB was expressed (Fig. 2A). However, GFP-CaeB-expressing cells are not completely unsusceptible to apoptosis induction, as PARP cleavage was detected after a 16 h treatment with staurosporine (Fig. 2B). Thus, CaeB is a very potent anti-apoptotic protein. However, it does not completely prevent a cellular response to an apoptotic stimulus. In order to quantify the anti-apoptotic activity of CaeB we determined the percentage of apoptotic cells after staurosporine treatment by using a TUNEL assay. As shown in Fig. 2C, 20.4% of cells expressing GFP were TUNEL positive after treatment, while only 2.7% of cells expressing GFP-CaeB were TUNEL positive, demonstrating apoptosis inhibition of greater than 85% caused by the expression of CaeB.

Figure 2.

CaeB inhibits intrinsic apoptosis efficiently.

A. HEK293 cells with stable expression of GFP as control and GFP-CaeB were treated with different concentrations of staurosporine ranging from values of 0.25 μM to 4 μM as indicated. After incubation for 4 h at 37°C in 5% CO2, proteins were separated using SDS-PAGE, transferred to a PVDF membrane and probed with antibodies directed against cleaved PARP as well as actin representing a loading control. Shown is one representative experiment out of two independent experiments.

B. HEK293 cells stably expressing GFP or GFP-CaeB were incubated with the indicated staurosporine concentrations for 16 h at 37°C in 5% CO2. Protein samples were loaded on a SDS-PAGE and immunoblots were probed with anti-cleaved PARP antibody to visualize cell death or anti-actin as loading control. The data represent one representative experiment out of two independent experiments with similar results.

C. HEK293 cells stably expressing GFP and GFP-CaeB were treated with or without 1 μM staurosporine for 16 h at 37°C in 5% CO2. After fixation and permeabilization the cells were stained with a TUNEL reaction mixture to detect DNA strand breaks of apoptotic cells. The percentage of TUNEL-positive cells was determined by flow cytometry. Shown are the average values ± SD of three independent experiments. **P < 0.01.

CaeB does not alter the steady-state level of Bcl-2 proteins

The ratio of positive and negative apoptosis regulators expressed in a cell plays a critical role in activating Bax and Bak and, thus, determines whether the cell will survive or undergo apoptosis. We therefore examined if GFP-CaeB expression alters the steady-state levels of Bcl-2 proteins. HEK293 cells stably expressing GFP or GFP-CaeB were either left untreated or treated with the staurosporine concentrations indicated. Immunoblot analysis revealed no obvious difference in the steady-state levels of the anti-apoptotic Bcl-2-like proteins Bcl-2, Bcl-XL and Mcl-1 (Fig. 3A). Furthermore, there were no obvious difference in the protein levels of the pro-apoptotic, BH3-only proteins Bim and Puma, while there was an upregulation of Bid in GFP-CaeB-expressing cells. Furthermore, no differences were detected in the protein levels of Bak and Bax (Fig. 3B). These data suggest that CaeB anti-apoptotic activity is not mediated by modulating the steady-state levels of Bcl-2 proteins and, thus, most likely acts downstream.

Figure 3.

CaeB does not alter the steady-state level of Bcl-2 proteins. HEK293 cells stably expressing GFP as control and GFP-CaeB were treated with the indicated staurosporine concentrations for 4 h at 37°C in 5% CO2. Protein extracts were separated by SDS-PAGE and immunoblots were probed with antibodies directed against (A) Bcl-2, Bcl-XL, Mcl-1 and actin as loading control, (B) Bid, Bim, Puma, Bak, Bax and actin. Shown is one representative immunoblot out of at least three independent experiments with similar results.

Narrowing down the point of CaeB action

The results obtained suggest that CaeB interferes with the apoptotic cascade downstream of Bcl-2 proteins. To analyse where CaeB interferes with the apoptosis pathway, we used an inducible expression system that allows induction at defined steps of the apoptosis pathway. Thus, HEK293 cells stably expressing GFP or GFP-CaeB were transfected with a Tet-On containing regulator plasmid (Krueger et al., 2006) and a TRE-containing response plasmid (Danke et al., 2010) encoding full-length human Bax. Expression of Bax was induced with doxycycline and apoptosis induction was measured by the presence of cleaved PARP assayed by immunoblotting. Cells transfected with both regulator and response plasmid did not show cleaved PARP under non-inducing conditions (data not shown). Under inducing conditions, cells expressing GFP displayed PARP cleavage (Fig. 4A). In contrast, cells expressing GFP-CaeB showed a 70% reduction in PARP cleavage compared with cells expressing GFP. These data suggest that CaeB inhibits apoptosis downstream of Bax activation. Next, we analysed whether CaeB could interfere with apoptosis induction at a very late step within the cascade. Thus, HEK293 cells stably expressing GFP or GFP-CaeB were transfected with the same Tet-On regulator plasmid and a TRE-containing response plasmid (Knott et al., 2005) encoding a constitutively active form of human caspase 3 (Srinivasula et al., 1998). Under inducing conditions, cells expressing GFP, as well as cells expressing GFP-CaeB, displayed the same degree of PARP cleavage (Figs 4B and S2), while no PARP cleavage was detected under non-inducing conditions (Fig. S2). Thus, CaeB expression cannot prevent apoptosis once caspase 3 is cleaved and thereby activated, indicating that CaeB acts downstream of Bax, but upstream of caspase 3 activation. Next, we analysed whether CaeB inhibits activation of effector and initiator caspases after apoptosis-induction by staurosporine. Thus, cleavage of the initiator caspase 9 and the effector caspase 7 was assayed by immunoblotting. As shown in Fig. 4C, cells expressing GFP and treated with staurosporine displayed caspase 9 and caspase 7 cleavage, as well as PARP cleavage. In contrast, cells expressing GFP-CaeB were completely protected from caspase 9 and caspase 7 cleavage, indicating that CaeB interferes with the apoptotic cascade at the level of caspase 9 activation or upstream. The results obtained suggest that CaeB acts between Bax activation and caspase 9 activation.

Figure 4.

Narrowing down the point of CaeB action.

A and B. HEK293 cells stably expressing GFP as control and GFP-CaeB were transfected with a pTet-On regulator plasmid and a pTRE response plasmid encoding in (A) Bax and (B) activated caspase 3. Five hours post transfection, expression was activated by addition of doxycycline (1 μg ml−1) for 18 h at 37°C in 5% CO2. Protein extracts were separated on SDS-PAGE, transferred to a PVDF membrane and probed with antibodies against cleaved PARP or actin to control equal protein loading. One representative immunoblot from three independent experiments with similar results is shown. Densitometric analysis of the cleaved PARP/actin ratio was performed using ImageJ. The data shown represent average values ± SD from at least three independent experiments. **P < 0.01.

C. HEK293 cells stably expressing GFP and GFP-CaeB were treated with the indicated staurosporine concentrations for 4 h at 37°C in 5% CO2. Protein extracts were separated by SDS-PAGE and immunoblots were probed with antibodies directed against cleaved PARP, cleaved caspase 7, cleaved caspase 9 and actin. Shown are representative immunoblots out of at least three independent experiments with similar results.

CaeB does not interfere with the recruitment of Bax to the mitochondria

Next, we analysed whether CaeB interferes with Bax activation and mitochondrial targeting. Upon activation, Bax undergoes extensive conformational changes, leading to mitochondrial targeting and oligomerization. Activation and oligomerization of Bax and Bak are most likely required for the formation of a pore within the mitochondrial membrane, resulting in MOMP and release of cytochrome C. To measure mitochondrial targeting of Bax, cells stably expressing GFP and GFP-CaeB were separated into cytosolic and mitochondrial fractions. Successful separation and equal protein loading was determined by immunoblot analysis using antibodies directed against a cytosolic marker protein (GAPDH) and two mitochondrial marker proteins [the matrix marker pyruvate dehydrogenase subunit E1-alpha (PDH-E1-α) and the inner membrane marker ATP synthase subunit alpha (C-V-α)]. The level of Bax in the cytosolic and mitochondrial fractions was determined by immunoblot analysis using an anti-Bax antibody (6A7) directed against activated Bax and an anti-Bax antibody (D2E11) directed against total Bax. Cells expressing GFP and GFP-CaeB were either mock-treated or treated with UV-light to induce apoptosis. The treatment with UV-light led to a reduction of total and activated Bax in the cytosolic fraction and to an increase of total and activated Bax in the mitochondrial fraction (Fig. 5A), demonstrating mitochondrial targeting of Bax after apoptosis induction. HEK293 cells stably expressing GFP and GFP-CaeB displayed a similar ratio of mitochondrial targeting of Bax. However, only cells expressing GFP showed cleavage of PARP and, thus, the onset of apoptosis. These data suggest that CaeB interferes with the apoptotic cascade downstream of the activation and mitochondrial targeting of Bax.

Figure 5.

CaeB inhibits intrinsic apoptosis at the mitochondrial level.

A. HEK293 cells stably expressing GFP and GFP-CaeB were exposed to UV-light (800 J m−2). After incubation for 6 h at 37°C in 5% CO2 cells were lysed, homogenized and mitochondria were isolated as described in Experimental procedures. Protein extracts were separated by SDS-PAGE, transferred to a PVDF membrane and probed with antibodies directed against Bax (clone 6A7 and D2E11) and cleaved PARP. Successful isolation of mitochondria was analysed by using a monoclonal antibody cocktail including PDH-E1-α and C-V-α (to visualize mitochondrial fraction) and GAPDH (to visualize cytoplasmic fraction). Shown are representative immunoblots out of at least three independent experiments with similar results.

B and C. HEK293 cells stably expressing GFP or GFP-CaeB were treated with the indicated staurosporine concentrations. After 4 h at 37°C in 5% CO2, the cells were stained with 1,1′,3,3,3′,3′-hexamethylindodicarbo-cyanine iodide [DilC1(5)] to visualize mitochondrial outer membrane permeabilization (MOMP). Cells were analysed by flow cytometry and the loss of DilC1(5) staining after staurosporine treatment was quantified. (B) Shown is one representative experiment out of at least three independent experiments with similar results. (C) The percentage of induced MOMP was calculated by subtracting MOMP-positive cells [DilC1(5)-negative cells] of uninduced from staurosporine-induced samples. Shown are the average values ± SD of three independent experiments. *P < 0.05, **P < 0.01.

D. HEK293 cells were transfected with GFP or the indicated GFP-tagged plasmids and precipitated from cell lysates with an anti-GFP antibody. Immunoblot analysis was used to detect either p32 in the anti-p32 blot or GFP (27 kDa), GFP-AnkG (65.6 kDa), GFP-CaeB (43.5 kDa) in the anti-GFP blot within immunoprecipitates (IP) and lysates. Shown are representative immunoblots out of three independent experiments with similar results.

CaeB interferes with mitochondrial outer membrane permeabilization

As ectopically expressed CaeB localizes to mitochondria (Fig. S3; Carey et al., 2011), we analysed the mitochondrial outer membrane permeabilization (MOMP) of staurosporine treated HEK293 cells stably expressing GFP or GFP-CaeB by staining with 1,1′,3,3,3′,3′-hexamethylindodicarbo-cyanine iodide [DilC1(5)]. DilC1(5) accumulates primarily in mitochondria with active mitochondrial membrane potential. During apoptosis, MOMP occurs and as a result the number of cells with reduced DilC1(5) staining increases. As shown in Fig. 5B and C, with increasing staurosporine concentrations more GFP-expressing cells displayed MOMP. In contrast, in cells expressing GFP-CaeB, the same number of cells showed MOMP, regardless of the staurosporine concentration applied. These data demonstrate a roughly 80% reduction in MOMP caused by the expression of GFP-CaeB, suggesting that CaeB inhibits apoptosis at the mitochondrial level.

CaeB does not bind to p32

AnkG was shown to interfere with apoptosis, possibly by binding to the pro-apoptotic host cell mitochondrial protein p32. As CaeB localizes to mitochondria and seems to interfere with the apoptotic cascade at the mitochondrial level, we analysed whether the anti-apoptotic activity of CaeB is mediated by binding to p32. Thus, we expressed GFP, GFP-CaeB and GFP-AnkG in HEK293 cells, precipitated proteins from the cell lysates with an anti-GFP antibody and evaluated the co-immunoprecipitation of endogenous p32 by immunoblot analysis. As shown in Fig. 5D, we detected co-immunoprecipitation of p32 with GFP-AnkG, but not with GFP and GFP-CaeB, demonstrating that CaeB does not interfere with apoptosis by binding to p32.

Discussion

Apoptosis has been identified as an important innate immune response that enables cells damaged by infectious agents to be cleared from the body. Thus, pathogens have evolved mechanisms to interfere with host cell apoptosis. While some bacteria activate apoptosis to escape from the ‘wrong’ host cell, others inhibit apoptosis to keep their safe harbour alive in order to survive and multiply. Especially for obligate intracellular pathogens, modulation of host cell apoptosis is an important and common virulence mechanism. Interestingly, most of the pathogens that inhibit apoptosis can also induce apoptosis under certain conditions. How these pathogens regulate host cell survival/death at the molecular level is still largely unknown. This is also the case for the obligate intracellular pathogen C. burnetii, for which inhibition of host cell apoptosis was first described in 2007 (Lührmann and Roy, 2007; Voth et al., 2007). So far, anti-apoptotic activity has been identified for several T4SS effector proteins from other pathogenic bacteria (Laguna et al., 2006; Schmid et al., 2006; Banga et al., 2007). This information, together with the reported anti-apoptotic function of the C. burnetii effector protein AnkG (Lührmann et al., 2010), suggests that substrates of the T4SS might be important molecules regulating host cell survival.

Information about C. burnetii in general and its virulence factors specifically is limited, as this pathogen has been rather difficult to study, due to a historic lack of genetic tractability. In 2009, a medium that supports host cell-free (axenic) growth of C. burnetii was defined (Omsland et al., 2009), opening the door for the development of genetic manipulation. Since then several groups have successfully transformed C. burnetii (Chen et al., 2010; Carey et al., 2011; Voth et al., 2011). Furthermore, by using transposon mutagenesis, it was possible to isolate T4SS mutants that were unable to translocate effector proteins into the host cell (Beare et al., 2011; Carey et al., 2011). These T4SS mutants are not able to multiply intracellularly and cannot inhibit host cell apoptosis, clearly demonstrating that C. burnetii-induced inhibition of host cell apoptosis depends on a functional T4SS and, thus, on the activity of effector proteins. However, whether apoptosis inhibition induced by C. burnetii is exclusively mediated by the anti-apoptotic effector protein AnkG was unclear. For the related pathogen L. pneumophila, functional redundancy among effector proteins has been demonstrated, suggesting that C. burnetii might also harbour multiple effector proteins that interfere with such an essential cellular activity. Therefore, we screened 18 newly identified (Carey et al., 2011) effector proteins for their anti-apoptotic activity. Expression of two of the effector proteins tested displayed inhibition of staurosporine-induced apoptosis in CHO cells (Fig. 1A). However, while CaeB blocked staurosporine- and UV-induced apoptosis also in HEK293 cells, CaeA only interfered with UV-induced apoptosis (Fig. 1C and D). This discrepancy might be due to differences in the genetic composition of the two cell lines used (Lee et al., 1997) and/or differences in the activation patterns of the two inducers (Lei and Davis, 2003; Willis and Adams, 2005) and, therefore, different signalling events after apoptosis induction. Consequently, these differences in anti-apoptotic activity between the two effector proteins suggest that they might interfere with different steps of the apoptosis cascade. CaeB was investigated further as it displayed more robust anti-apoptotic activity (Fig. 1C and D). Apoptosis is regulated by the ratio of the pro-apoptotic BH3-only proteins and the anti-apoptotic Bcl-2-like family members. Not surprisingly, these proteins are targeted by several pathogens. For instance, Chlamydia trachomatis inhibits host cell apoptosis by a process that involves degrading pro-apoptotic BH3-only proteins, possibly by the secreted protease CPAF (chlamydial protease/proteasome-like activity factor) (Fischer et al., 2004; Dong et al., 2005). However, C. burnetii infection does not alter the steady-state level of these apoptotic regulators (Lührmann and Roy, 2007; Voth et al., 2007). In agreement with these findings, we also did not detect a CaeB-induced modulation of Bcl-2 proteins, with the exception of Bid (Fig. 3A and B). Full-length Bid is upregulated in HEK293 cells stably expressing GFP-CaeB under all conditions tested (Fig. 3B). However, although this is an interesting observation, it might not be of relevance for the anti-apoptotic activity observed. In order to narrow down where CaeB interfered with the apoptotic cascade, we used a doxycycline-inducible expression system (Krueger et al., 2006; Danke et al., 2010), which allowed us to activate the apoptotic pathway at defined steps. Thus, CaeB activity was narrowed down to upstream of caspase 3 and downstream of Bax activation (Fig. 4A and B). As CaeB expression also prevents cleavage of the initiator caspase 9, the anti-apoptotic activity could be further narrowed down to downstream of Bax and upstream of caspase 9 activation. However, it is also possible that CaeB directly interferes with Bax activation and/or mitochondrial targeting. By separating the cytosolic from the mitochondrial fraction, we were able to demonstrate that after staurosporine treatment Bax traffics from the cytosol to the mitochondria in cells stably expressing GFP and GFP-CaeB (Fig. 5A). This clearly indicates that CaeB does not prevent activation, conformational change and mitochondrial targeting of Bax. As CaeB was shown to localize to mitochondria (Fig. S3 and Carey et al., 2011), and as C. burnetii-induced apoptosis inhibition is mediated through a process that involves preventing cytochrome C release from mitochondria (Lührmann and Roy, 2007), we investigated whether CaeB interferes with mitochondrial function. The mitochondrion is the central organelle in intrinsic apoptosis. Permeabilization of its outer-membrane (MOMP) leads to the release of pro-apoptotic proteins from the mitochondrial inner-membrane and is a crucial event driving intrinsic apoptosis (Tait and Green, 2010). Our data demonstrate that CaeB expression reduced MOMP by around 80% (Fig. 5B and C). Thus, CaeB does not prevent trafficking of activated Bax to the mitochondria, but prevents MOMP, suggesting that CaeB blocks the apoptotic signal within the mitochondria. Only a few bacterial proteins have been identified that target the mitochondria to modulate host cell apoptosis. Some of these effectors induce apoptosis while the Anaplasma phagocytophilum T4SS effector Ats1, the Neisseria meningitidis porin PorB and the T3SS effector HopG1 from the plant pathogen Pseudomonas syringae display anti-apoptotic activity (Rudel et al., 2010). Ats1 contains an N-terminal sequence signal required for mitochondrial localization. Mitochondria-localized Ats1 was shown to prevent mitochondrial targeting of Bax and cytochrome C release from the mitochondria (Niu et al., 2010). In contrast to Ats1, CaeB does not interfere with mitochondrial targeting of Bax. Thus, the anti-apoptotic mechanism applied by these two T4SS effector proteins seems to be different.

So far, only one C. burnetii effector protein with anti-apoptotic activity has been identified (Lührmann et al., 2010). AnkG might interfere with apoptosis by binding to the mitochondrial host cell protein p32. Here, we describe two additional anti-apoptotic effector proteins, CaeA and CaeB. For CaeB, we provided evidence that this effector inhibits apoptosis at the mitochondrial level. The fact that there are at least three different C. burnetii effector proteins with anti-apoptotic activities suggests that there is functional redundancy among the effector proteins. However, CaeB did not co-immunoprecipitate with p32, indicating that CaeB, in contrast to AnkG, most likely does not interfere with the apoptotic cascade by interacting with the pro-apoptotic mitochondrial protein p32. Thus, these data indicate that although the general activities (inhibition of apoptosis) of AnkG and CaeB are redundant, the molecular mechanism(s) associated with each effector protein is probably distinct. Hence, C. burnetii contains several anti-apoptotic effector proteins with distinct molecular activities, which provides a much broader protection from host immune intervention. This enables the bacteria to survive in different host cells and is therefore beneficial for intracellular survival.

Experimental procedures

Reagents, cell lines and bacterial strains

Unless otherwise noted, chemicals were purchased from Sigma Aldrich. Complete Protease inhibitor cocktail mixture and Fugene 6 Transfection Reagent were from Roche. Staurosporine was from Cell Signaling. Cell lines were cultured at 37°C in 5% CO2 in media containing 10% heat-inactivated fetal bovine serum (Biochrom) and 1% pencillin-streptomycin (Invitrogen). CHO (Chinese hamster ovary) fibroblasts were grown in minimal essential medium alpha medium (Invitrogen), HeLa (human epithelial) and HEK293 (human embryonic kidney) cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen). To construct HEK293 cells stably expressing GFP, GFP-CaeA and GFP-CaeB the corresponding plasmids were transfected into HEK293 cells and selected by culturing in media supplemented with 1.5 mg ml−1 geneticin (Roth).

Plasmids and primers

Plasmid and primers used in this study are listed in Tables S1 and S2.

Plasmid construction

Cloning CaeA and CaeB into the expression vector pEGFP-C2. The genes of CaeA and CaeB were amplified from C. burnetii Nine Mile phase II clone 4 genomic DNA by standard PCR using the primers CaeA-fwd and CaeA-rev as well as CaeB-fwd and CaeB-rev. The resulting PCR product was restricted with BglII and PstI, followed by ligation with likewise restricted pEGFP-C2.

Cloning the expression vector for the human pro-apoptotic protein Bax. The gene of the human Bax isoform 1 was amplified by standard PCR from the vector pMSCV-Venus-Bax (Spencer et al., 2009) using the primers TREtight-hBax-fwd and TREtight-hBax-rev. The resulting PCR product was restricted with EcoRI and EcoRV which was followed by ligation with likewise restricted pWHE655 (Danke et al., 2010) to give pWHE655-hBax. The cDNA encoding the activated reverse Caspase 3 was amplified from the plasmid pWHE556 with the primer pair 5′Seq-pBi and pUHD1. The PCR product was restricted with NheI and XbaI and ligated with pWHE169 (Knott et al., 2005), which was restricted with NheI and partially restricted with XbaI. The resulting plasmid, containing a bidirectional Tet-controlled promoter expressing EGFP and revCasp3, was designated pWHE544.

Apoptosis assays

Nuclear fragmentation assay

CHO cells were plated on coverslips in 24-well dishes at a density of 2.5 × 104 cells per well. After an overnight incubation, cells were transfected with the indicated plasmids. Eighteen hours post transfection, the cells were incubated with staurosporine (2 μg ml−1) for 4 h at 37°C in 5% CO2. The cells were fixed with 4% paraformaldehyde (Alfa Aeser) in PBS (Biochrom) for 20 min at room temperature, permeabilized with ice-cold methanol for 30 s, quenched with 50 mM NH4Cl (Roth) in PBS for 15 min at room temperature. The cells were mounted using ProLong Gold with DAPI (Invitrogen) to visualize the nucleus.

Immunoblotting

HEK293 cells stably expressing GFP, GFP-CaeA or GFP-CaeB were seeded in a 12-well plate at a density of 1 × 105 cells per well. The cells were treated with the indicated staurosporine concentrations for the indicated times at 37°C in 5% CO2. Proteins were separated by SDS-PAGE and transferred to a PVDF membrane (Millipore). The membranes were probed with antibodies directed against cleaved PARP (9541), Bcl-2 (2870), Bcl-XL (2764), Mcl-1 (4572), Bid (2002), Bim (2933), Puma (4976), Bak (6947), Bax (5023), cleaved caspase 7 (9491), cleaved caspase 9 (9501) from Cell Signaling, cleaved PARP (611038) from BD Biosciences and actin (A 2066) from Sigma-Aldrich. The proteins were visualized by using horseradish peroxidase-conjugated secondary antibodies (Dianova) and a chemiluminescence detection system (Thermo Scientific or Millipore).

Protein expression levels of HEK293 cells stably expressing GFP, GFP-CaeA and GFP-CaeB were analysed with anti-GFP rabbit serum from Invitrogen (A6455).

Mitochondrial outer membrane permeabilization (MOMP)

HEK293 cells stably expressing GFP or GFP-CaeB were seeded in a 12-well plate at a density of 3 × 105 cells per well. The cells were treated with different staurosporine concentrations ranging from 0.25 μM to 2 μM for 4 h, followed by staining with 50 nM 1,1′,3,3,3′,3′-hexamethylindodicarbo-cyanine iodide [DilC1(5)] from Molecular Probes (34151) for 20 min at 37°C in 5% CO2 according to instructions of the manufacturer. The percentage of DilC1(5)-negative cells was analysed by flow cytometry.

TUNEL assay

HEK293 cells stably expressing GFP or GFP-CaeB were seeded in a six-well plate at a density of 6 × 105 cells per well. After treatment with 1 μM staurosporine for 16 h at 37°C in 5% CO2 cells were fixed with 2% paraformaldehyde in PBS and permeabilized using 0.1% Triton X-100 in 0.1% sodium citrate. DNA strand breaks were detected by incubating cells with TUNEL reaction mixture containing the terminal deoxynucleotidyl transferase and fluorescently labelled nucleotides using the manufacturer's protocol (Roche, 12156792910). The amount of TUNEL-positive cells was analysed by flow cytometry.

Apoptosis induction with UV-light

HEK293 cells with stable expression of GFP, GFP-CaeA and GFP-CaeB were seeded at a concentration of 3 × 105 cells per well. Cells were washed with PBS before exposure with 200 J m−2 and 800 J m−2 UV-light (Stratagene). After 6 h incubation at 37°C in 5% CO2 samples were prepared for Western blot analysis.

Tet-On expression system

HEK293 cells stably expressing GFP or GFP-CaeB were seeded in a 12-well plate at a density of 1 × 105 cells per well. The cells were transfected using polyethylenimine (Polysciences) with 100 ng of pWHE125-P regulator plasmid and 100 ng of the response plasmids pWHE655-hBax or pWHE544. For titration of DNA amounts, concentrations of response plasmid pWHE544 ranging from 1 ng to 200 ng were used and the total of 300 ng was filled up with salmon sperm DNA (Invitrogen). Five hours post transfection, protein expression was induced by adding 1 μg ml−1 doxycycline to the medium. After incubation for 18 h at 37°C in 5% CO2, samples were prepared for Western blot analysis.

Isolation of mitochondria

HEK293 cells stably expressing GFP or GFP-CaeB were seeded in a 100 × 20 mm cell culture dish at a density of 3.5 × 106 cells per dish. Cells were washed with PBS before exposure with 800 J m−2 UV-light (Stratagene). After 6 h incubation at 37°C in 5% CO2 a total of 3.5 × 107 cells per sample were washed with PBS and incubated with 800 μl of lysis buffer [250 mM Sucrose, 20 mM Hepes (pH 7.5), 1.5 mM MgCl2, 10 mM KCl, 1 mM EDTA (pH 8.5), 1 mM EGTA (pH 8.0), 1 mM DTT, 1× protease inhibitor] for 10 min at 4°C. The cells were homogenized using a dounce tissue grinder from VWR (432-5005) and centrifuged for 10 min at 1000 g at 4°C. The supernatant was used for a second centrifugation for 25 min at 10 000 g at 4°C to isolate mitochondria (pellet) from the cytosolic fraction (supernatant). Different fractions were analysed by immunoblot analysis using antibodies directed against Bax clone D2E11 (Cell Signaling, 5023), Bax clone 6A7 (BD Biosciences, 556467), cleaved PARP (BD Biosciences, 611038), GFP (Invitrogen, A6455) and ApoTrac Apoptosis Fraction Analysis mAb Cocktail (Invitrogen, 459160).

Co-immunoprecipitation

HEK293 cells were seeded in a six-well plate at a density of 2 × 105 cells per well and transiently transfected with the indicated plasmids. On the following day, cells were washed with PBS and incubated with 300 μl of lysis buffer [20 mM Hepes (pH 7.5), 200 mM NaCl, 1 mM EDTA, 0.1% (v/v) Nonidet P-40, 10% (v/v) Glycerol, 1× protease inhibitor, 1 mM DTT] for 30 min on ice. After centrifugation for 10 min and 14000 r.p.m. at 4°C, supernatants were incubated with anti-GFP rabbit serum from Invitrogen (A6455) for 2 h at 4°C. Complexes were precipitated by adding 40 μl of protein A/G PLUS-Agarose (Santa Cruz) and incubated for 45 min at 4°C. Beads were washed three times with washing buffer [20 mM Hepes (pH 7.5), 100 mM NaCl, 1 mM EDTA, 0.1% (v/v) Nonidet P-40] and samples were analysed by immunoblot using mouse anti-GFP (Roche, 11814460001) and mouse anti-p32 (Covance, MMS-604R) antibodies.

Statistical analysis

An unpaired Student's t-test was used for statistical analysis.

Acknowledgements

This work was supported by the Deutsche Forschungsgemeinschaft (DFG) to A.L. (LU1357/2-1), by the SFB796 to C.B. and by the ERA-NET PathoGenoMics 3rd call (0315903) to A.L. We thank Dr Sabrina Spencer for the plasmid pMSCV-Venus-Bax, Eric Cambronne for critically reading the manuscript and the Lührmann laboratory for helpful discussions.

Ancillary