The Chlamydia pneumoniae Inclusion Membrane Protein Cpn1027 Interacts with Host Cell Wnt Signaling Pathway Regulator Cytoplasmic Activation/Proliferation-Associated Protein 2 (Caprin2)

We previously identified hypothetical protein Cpn1027 as a novel inclusion membrane protein that is unique to Chlamydia pneumoniae. In the current study, using a yeast-two hybrid screen assay, we identified host cell cytoplasmic activation/proliferation-associated protein 2 (Caprin2) as an interacting partner of Cpn1027. The interaction was confirmed and mapped to the C-termini of both Cpn1027 and Caprin2 using co-immunoprecipitation and GST pull-down assays. A RFP-Caprin2 fusion protein was recruited to the chlamydial inclusion and so was the endogenous GSK3β, a critical component of the β-catenin destruction complex in the Wnt signaling pathway. Cpn1027 also co-precipitated GSK3β. Caprin2 is a key regulator of the Wnt signaling pathway by promoting the recruitment of the β-catenin destruction complex to the cytoplasmic membrane in the presence of Wnt signaling while GSK3β is required for priming β-catenin for degradation in the absence of Wnt signaling. The Cpn1027 interactions with Caprin2 and GSK3β may allow C. pneumoniae to actively sequester the β-catenin destruction complex so that β-catenin is maintained even in the absence of extracellular Wnt activation signals. The maintained β-catenin can trans-activate Wnt target genes including Bcl-2, which may contribute to the chlamydial antiapoptotic activity. We found that the C. pneumoniae-infected cells were more resistant to apoptosis induction and the anti-apoptotic activity was dependent on β-catenin. Thus, the current study suggests that the chlamydial inclusion protein Cpn1027 may be able to manipulate host Wnt signaling pathway for enhancing the chlamydial anti-apoptotic activity.

In the current study, we used the C. pneumoniae inclusion membrane protein Cpn1027 Cterminal region covering residues 210-527, designated as Cpn1027(210-527), as a bait to screen a HeLa cDNA library in a yeast two-hybrid assay and found that Cpn1027 interacted with host Caprin2, a regulator of the β-catenin destruction complex. Both a red fluorescence protein (RFP)-Caprin2 fusion protein and the endogenous GSK3β, a critical component of the β-catenin destruction complex, were recruited to the C. pneumoniae inclusion. The recruitment of the β-catenin destruction complex to the C. pneumoniae inclusion correlated with the C. pneumoniae anti-apoptotic activity and knocking down of β-catenin attenuated the C. pneumoniae anti-apoptotic activity.
To detect the interaction of Cpn1027 with GSK3β, HEK 293T cells were transfected with pDsRed-Cpn1027 or pDsRed alone. 24h after transfection, a mouse anti-GSK3α/β monoclonal antibody (BD Transduction Laboratories) was used to precipitate the cell lysates as described above. The precipitates were used for Western blot detection of RFP.

Western blot assay
The Western blot assay was carried out as described elsewhere [36,37] for detecting components in the pellets brought down by agarose bead-immobilized protein G-antibody complexes (co-immunoprecipitation) or bead-immobilized GST fusion proteins (GST pull-down) or cultured cells. The pellets were boiled for 5 min in 2% sodium dodecyl sulfate (SDS) sample buffer while the cultured cells after various treatments including C. pneumoniae infection, siRNA transfection and apoptosis induction as described in sections below were lysed 2% SDS sample buffer and then boiled for 5 min. The boiled samples were resolved in SDS polyacrylamide gel and then transferred to nitrocellulose membranes for labeling with various primary and secondary antibodies. The primary antibodies used include the following: mouse polyclonal antibody against Cpn1027, rabbit polyclonal antibody against RFP/GST, goat polyclonal antibody against GST (GE Healthcare), mouse monoclonal antibody against β-actin (Sigma), mouse monoclonal antibody against GSK3α/β (BD laboratories), rabbit polyclonal antibody against Axin (Cell Signaling, Danvers, MA 01923), mouse monoclonal antibody against β-catenin (BD Laboratories), mouse monoclonal antibody against Bcl-2 (BD laboratories), mouse monoclonal antibody against chlamydial HSP60 (BC7.1), and rabbit polyclonal antibody against CK1α (Cell Signaling). The primary antibody binding was probed with a HRP (horseradish peroxidase)-conjugated secondary antibody and visualized with an enhanced chemiluminescence (ECL) kit (Santa Cruz Biotechnology). Semi-quantitative densitometry was performed using Bio-Rad Quantity One software.

Immunofluorescence assay
HeLa cells grown on coverslips were fixed with 4% paraformaldehyde (Sigma) dissolved in PBS for 20 min at room temperature followed by permeabilization with 0.1% Triton X-100 (Sigma) for an additional 10 min. After washing and blocking, the cell samples were subjected to antibody and chemical staining. DAPI (blue; Sigma) was used to visualize nuclear DNA. A rabbit anti-chlamydial organism antibody (R12AR39, raised with C. pneumoniae AR39 organisms; unpublished data) or, in some cases, a rabbit polyclonal antibody against IncA (kindly provided by Dr. Ted Hackstadt, ref: [8], plus a goat anti-rabbit IgG secondary antibody conjugated with Cy2 (Jackson ImmunoResearch Laboratories) was used to visualize chlamydial inclusions. Mouse monoclonal antibodies raised against GSK3α/β (BD laboratories) or β-catenin (BD laboratories) plus a goat anti-mouse IgG conjugated with Cy3 (red, Jackson ImmunoResearch) were used to visualize the corresponding antigens. For the transfected cell samples, the RFP-fusion proteins were visualized via the fusion tag RFP (red). The images were acquired with an Olympus AX-70 fluorescence microscope equipped with multiple filter sets (Olympus) as described previously [36,37]. The single color images were superimposed with the software Sim-plePCI. All microscopic images were processed using Adobe Photoshop (Adobe Systems).

siRNA Knockdown and apoptosis induction
HeLa 229 cells were infected with Chlamydia pneumoniae AR39 at a multiplicity of infection of 0.5 in the presence of 2 μg/ml cycloheximide for different time points as indicated in individual experiments following an infection procedure as described elsewhere [15][16][17][18]. Twenty-4 hours after infection, the culture media were replaced with fresh media without cycloheximide. For knocking down β-catenin, HeLa cells were transfected with 100nM SignalSilence β-cateninspecific siRNA I (Cell Signaling) or with 100nM SignalSilence Control siRNA (Cell Signaling #6568) using the Lipofectamine 2000 following the protocol recommended by the manufacturer (Invitrogen). 24h after transfection, the cultures were either harvested for Western blot detection of the protein levels of β-catenin and Bcl-2 or induced to undergo apoptosis. For apoptosis induction, HeLa cells with or without C. pneumoniae or siRNA transfection were treated with 2μM staurosporine (STS) (Sigma) for 4 h. The mock treatment cells were incubated in culture medium alone. The cell samples were processed for immunofluorescence assay as described below. The cells fixed in 4% paraformaldehyde (Sigma) were processed for labeling both Chlamydia (green) and DNA (blue) as described above. Cells were evaluated for chromatin condensation [23,24,37] and the presence of either C. pneumoniae infection. For each sample, cells from ten random fields were counted under an Olympus AX-70 fluorescence microscope and the percentage of apoptotic cells was calculated as the number of apoptotic cells divided by the total number of cells counted,~100 cells for each coverslip.

Cpn1027 interacts with host Caprin2
We used a C-terminal fragment (covering residues 210 to 527) of Cpn1027, designated as Cpn1027(210-527), as a bait to search for host cell binding partners from a HeLa cell cDNA library in a yeast two-hybrid screen assay. Our screen yielded several positive cDNA clones (data not shown) but only the clone encoding residues 245-1128 of Caprin2 was found to interact with Cpn1027 reproducibly in various in vitro assays (Fig 1). For example, when yeasts co-transformed with pEXP-AD502-Caprin2(245-1128) as prey plus pDBLeu-Cpn1027(210-527) as bait or the positive control prey Gal4AD-TRAF3 plus bait Gal4DB-LMP1 [34] or the negative control prey Gal4AD plus bait Gal4DB-LMP1 were grown on selective medium, the bait-prey interactions only occurred between Caprin2(245-1128) and Cpn1027(210-527) or TRAF3 and LMP1 but not Gal4AD alone and LMP1 (Fig 1A). In a co-immunoprecipitation assay, the RFP-Caprin2(245-1128) fusion protein but not the control RFP-APC4 fusion protein or RFP alone precipitated the FLAG-tagged Cpn1027(210-527) from the 293T cells cotransfected with plasmids coding for the RFP alone or RFP-and FLAG-tagged proteins ( Fig  1B). The Cpn1027 interaction with Caprin2 was further confirmed in a GST pull down assay, in which the GST-Caprin2 (245-1128) but not GST alone pulled the RFP-Cpn1027 fusion protein but not RFP alone down into the pellets (Fig 1C). We also attempted to use Cpn1027 to immunoprecipitate endogenous Caprin2 from cell lysates of mock or C. pneumoniae-infected cells but without success. This might be primarily due to the fact that the endogenous Caprin2 was not detectable in cell lysates (data not shown). We also failed to precipitate the endogenous Cpn1027 from C. pneumoniae-infected cell lysates (data not shown), which has been a common challenge for probing the interactions of endogenous Incs with host cell proteins. Despite these technical difficulties, the 3 independent assays above have validated the interactions of C. pneumoniae Cpn1027 with host protein Caprin2. In addition, a RFP-Caprin2 (245-1128) fusion protein but not RFP alone was detected around C. pneumoniae inclusions (Fig 1D), suggesting that Caprin2 can be recruited by C. pneumoniae inclusion membrane proteins.
3. Both Caprin2 and GSK3β, components of the β-catenin destruction complex, are recruited around the chlamydial inclusion membrane As Cpn1027 is an inclusion membrane-localized protein, interaction with Caprin2 should lead to the accumulation of Caprin2 in the vicinity of the chlamydial inclusion. We found that the ectopically expressed RFP-Caprin2 but not RFP alone was detected around the C. pneumoniae inclusion (Fig 1D), suggesting that Caprin2 is actively recruited to the C. pneumoniae inclusion membrane. In the presence of Wnt signaling, Caprin2 enhances the interaction of proteins in the β-catenin destruction complex with LRP5/6 to form a signalsome leading to the stabilization of cytosolic β-catenin [27]. In particular, Caprin2 promotes the GSK3β-mediated LRP5/6 interaction. Since Cpn1027 was found to interact with Caprin2, we next tested whether GSK3β was also recruited to the inclusion in C. pneumoniae infected cells. The endogenous GSK3β was detected at the periphery of the C. pneumoniae inclusion (Fig 3A). Furthermore, an anti-RFP antibody precipitated the endogenous GSK3β from HEK 293T cells expressing RFP-Cpn1027 but not RFP alone (Fig 3B). Vice versa, an anti-GSK3α/β antibody precipitated RFP-Cpn1027 but not RFP alone (Fig 3C). These observations together suggest that host cell GSK3β can be recruited around the C. pneumoniae inclusion by interacting with Cpn1027.

Knocking down β-catenin attenuate C. pneumoniae anti-apoptotic activity
Since β-catenin plays a critical role in cell survival [26,28] and C. pneumoniae is known to possess a profound anti-apoptotic activity, we next tested whether knocking down β-catenin in C. pneumoniae-infected cells can inhibit the C. pneumoniae anti-apoptotic activity (Fig 4). HeLa cells with C. pneumoniae infection for 24h were transfected with either control siRNA or siRNA targeting β-catenin. 24h after transfection, a portion of the HeLa cell samples were harvested for monitoring the protein levels of β-catenin and Bcl-2. We found that the siRNA targeting β-catenin but not the control siRNA significantly reduced the levels of both β-catenin and Bcl-2 but not β-actin (Fig 4A). The parallel cultures including the uninfected normal HeLa cells were also fixed for immunofluorescence assay for monitoring the effect of the β-catenin SiRNA treatment (Fig 4B). The β-catenin protein level in both the AR-39-infected and normal uninfected cells was significantly reduced by the β-catenin SiRNA treatment. More importantly, some of the remaining cell culture samples were stimulated with staurosporine for inducing apoptosis and the percent of apoptotic cells was compared between the β-catenin-specific siRNA and control siRNA groups (Fig 4B). We found that in the presence of the control siRNA, C. pneumoniae maintained a robust anti-apoptotic activity. However, the siRNA that targets β-catenin blocked the C. pneumoniae anti-apoptotic activity.

Discussion
In the current study, we have presented evidence that the Chlamydia pneumoniae inclusion membrane protein Cpn1027 can interact with the host Caprin2. First, in a yeast two-hybrid screening, the Caprin2-expressing clone was identified to be positive for interacting with Cpn1027, which was reproduced in a yeast co-transformation experiment. Second, the Cpn1027-Caprin2 interaction was validated in a co-immunoprecipitation assay, in which the RFP-Caprin2 fusion protein but not RFP fusion tag alone or RFP-APC4 fusion protein precipitated FLAG-Cpn1027. Third, a GST-Caprin2 fusion protein but not GST alone pulled down RFP-Cpn1027. Fourth, the interaction between Cpn1027 and Caprin2 was mapped to the C-termini of both molecules using a GST pull-down or ELISA assay. Finally, an RFP-Caprin2 fusion protein was recruited around C. pneumoniae inclusion membrane where Cpn1027 is localized. Since Caprin2 is a regulator in canonical Wnt signaling by facilitating the phosphorylation of the seven-transmembrane-spanning receptor Frizzled/LRP5/6 receptor complex and promoting the sequestration of the β-catenin destruction complex to the cytoplasmic membrane [28], the above observations suggest that C. pneumoniae may be able to recruit and sequestrate the βcatenin destruction complex around the inclusion membrane via the Cpn1027-Caprin2 interaction. This hypothesis is further supported by the finding that GSK3β, a key component of βcatenin destruction complex, was also recruited around the C. pneumoniae inclusion.
The observation that the C-terminal regions in both Cpn1027 and Caprin2 are responsible for their interactions indicates that their C-termini are accessible to each other. All chlamydial inclusion proteins possess a bi-lobed hydrophobic trans-membrane region in either their N-or C-terminal regions [19,40,41]. The Cpn1027's bi-lobed region is localized in its N-terminus pneumoniae AR39 organisms were processed for co-staining with a mouse anti-GSK3β plus a Cy3-conjugated goat anti-mouse IgG (red) and a rabbit anti-IncA antiserum plus a Cy2-conjugated goat anti-rabbit IgG (green). The DNA was labeled with DAPI (blue). Note that the endogenous GSK3β was detected around the chlamydial inclusions. The region marked with a white square in panel a was magnified and the magnified area was shown in panels b (3 color-overlay), c (GSK3β alone, red), d (IncA alone, green) and e (DNA alone, blue) respectively. (B) Lysates from HEK 293T cells transfected with pDsRed vector alone (expressing RFP alone, left lanes) or pDsRed-Cpn1027 plasmid [expressing RFP-Cpn1027, right lanes] were immunoprecipitated without (panel a) or with an anti-RFP antibody (b & c). The lysates or precipitates were analyzed by Western blot using an anti-RFP (panels a & b) or an anti-GSK3α/β (c) antibodies. Note that RFP-Cpn1027but not RFP alone coprecipitated the endogenous GSK3β. (C) The parallel lysates were precipitated without (panel d) or with an anti-GSK3α/β antibody (e & f). The lysate or precipitates were analyzed by using αGSK3α/β (panels d & e) or αRFP antibodies. Note that the endogenous GSK3β co-precipitated RFP-Cpn1027, but not RFP alone (f).
doi:10.1371/journal.pone.0127909.g003 [15], suggesting that the C-terminus of Cpn1027 is exposed to the host cell cytoplasm, thus allowing the Cpn1027 C-terminus to access to the host cytoplasmic signaling protein Caprin2. It has been previously shown that the Chlamydia pneumoniae inclusion protein Cpn0585 may use its C-terminus to interact with multiple host cell Rab GTPases [9] while Cp0236 with the host NFkappaB activator 1 (Act1; ref: [22]. The C-terminus of Caprin2 is a C1q-related domain that is a homotrimer required for Caprin2 to interact with the Wnt signaling receptor complex LRP5/6 [28,42]. By interacting with the C1q-related domain of Caprin2, Cpn1027 anchored in the Chlamydia pneumoniae inclusion membrane may mimick the function of the Wnt The parallel cultures including the uninfected normal HeLa cells were also fixed for immunofluorescence assay for labeling C. pneumoniae AR39 (green), β-catenin (red) and DNA (blue) using the corresponding antibodies and dyes as listed in the materials and methods section. Note that the β-catenin SiRNA treatment significantly reduced the β-catenin protein level in both the AR-39-infected and normal uninfected cells. (C) The parallel cultures were further induced to undergo apoptosis using staurosporine (STS) stimulation for 4h. The apoptotic cells were identified using DNA staining. 10 random views were counted from each coverslip. Note that the C. pneumoniae AR39-infected cells were significantly resistant to apoptosis induction (column 1 v.s. 2) and the treatment with βcatenin-targeted siRNA but not control siRNA significantly neutralized the C. pneumoniae anti-apoptotic activity (column 2 v.s. column 4). Asterisks indicates significant differences of P<0.05; paired Student's t-test. The data came from 3 independent experiments each with duplicates. signaling receptor for stabilizing the Caprin2 trimer and recruiting the β-catenin destruction complex around the chlamydial inclusion (Fig 5).
GSK3β, a key component of β-catenin destruction complex, was also localized around the C. pneumoniae inclusion. Furthermore, in cells expressing RFP-Cpn1027 (Fig 3) or GST-Cpn1027 (data not shown), the Cpn1027 fusion co-precipitated the endogenous GSK3β and the endogenous GSK3β also co-precipitated the Cpn1027 fusion protein. These observations further In the absence of Wnt signals, cytosolic β-catenin is degraded by a β-catenin destruction complex composed of Axin, adenomatosis polyposis coli (APC), glycogen synthase kinase 3β (GSK3β) and casein kinase 1α (CK1α). Both GSK3β and CK1α can phosphorylate β-catenin, which triggers ubiquitination and subsequent degradation of β-catenin by proteasomes. B. The Wnt signaling pathway is activated by binding of Wnt ligands, to Frizzled and LRP5/6 at the cell surface. LRP5/6 is activated by phosphorylation, which recruits the β-catenin destruction complex to the plasma membrane. The cytoplasmic activation/proliferation-associated protein 2 (Caprin2) facilitates LRP5/6 phosphorylation by GSK3β and enhances the interaction between Axin and the cytoplasmic tail of LRP5/6, which promotes the sequestration of the β-catenin destruction complex. The cytoplasmic pool of β-catenin rises and translocates into the nucleus, where it binds to the TCF/LEF family of transcription factors, displacing co-repressors Groucho and HDAC and acts as a co-activator to stimulate the transcription of anti-apoptotic, Bcl-2. C. In the C. pneumoniae AR39-infected cells, signalosomes are formed around the inclusion possibly through the chlamydial Inc protein Cpn1027 interacting with the β-catenin destruction complex, composed of Caprin2, Axin, GSK3β and CKIα resulting in the activation of the Wnt signaling transduction pathway in the absence of extracellular stimulation of Wnt. This leads to an increase in the cytoplasmic pool of β-catenin, nuclear β-catenin and membrane-associated β-catenin. The β-catenin-transactivated anti-apoptosis genes may promote the survival of the C. pneumoniae AR39-infected cells, hence the survival of AR39.
doi:10.1371/journal.pone.0127909.g005 support our proposal that Cpn1027 can recruit the β-catenin destruction complex around chlamydial inclusions. However, at this moment, it remains unclear whether the interaction of Cpn1027 with GSK3β requires Caprin2. Since the endogenous Caprin2 is below our detection limit, we have not been able to determine whether the Cpn1027-GSK3β interaction is direct or indirect. Nevertheless, some chlamydial inclusion membrane proteins have been shown to interact with multiple host factors [9] and to simultaneously recruit multiple function-related host proteins around the inclusion [43]. Thus, regardless of the precise mechanisms on how the βcatenin destruction complex proteins are recruited around the C. pneumoniae inclusions, their close proximity to chlamydial inclusions may be biologically significant.
Caprin2 refers to cytoplasmic activation/proliferation-associated protein 2, which is normally distributed in the host cell cytosol. In response to the activation of Wnt signaling pathway, Caprin2 is recruited to the plasma membrane where it participates in the LRP5/6 signalosome formation by enhancing the interactions between Axin, GSK3β, and LRP5/6, resulting in the sequestration of the β-catenin destruction complex, inactivation of GSK3β, accumulation of β-catenin and translocation of β-catenin to the nucleus for trans-activating Wnt target genes [26][27][28]. On the contrary, in the absence of Wnt ligands, GSK3β in concert with other β-catenin destruction complex components Axin, APC, and CK1α, phosphorylates βcatenin, resulting in destruction of β-catenin. Cytoplasmic β-catenin levels are normally kept low through continuous proteasome-mediated degradation regulated by the β-catenin destruction complex. We propose that in Chlamydia pneumoniae-infected cells, Cpn1027 anchored in the inclusion membrane protein may mimick the function of the Wnt signaling receptor LRP5/6 by interacting with Caprin2 and recruiting β-catenin destruction complex around chlamydial inclusion, resulting in β-catenin accumulation and activation of Wnt signaling pathway-targeted genes as proposed in Fig 5. Thus, even in the absence of extracellular Wnt signaling, Cpn1027 may be able to deliver the Wnt-like signaling from inside cells.
The Chlamydia-triggered Wnt-like signaling may have multiple effects on the infected cells. In the current study, we found that the C. pneumoniae-infected cells were more resistant to apoptosis induction by staurosporine. β-catenin seemed to play an important role in the C. pneumoniae anti-apoptotic activity since knockdown of β-catenin significantly attenuated the C. pneumoniae anti-apoptotic activity. Thus, we proposed that preserving of β-catenin by sequestrating the β-catenin destruction complex around the chlamydial inclusions might represent an active mechanism for maintaining the C. pneumoniae anti-apoptotic activity. We are aware that inhibition of host cell apoptosis may involve multiple molecular mechanisms since maintaining host cell viability is fundamental for the slow growing C. pneumoniae organisms to complete their growth cycles. Since Cpn1027 is unique to C. pneumoniae, it should be cautious when applying the current finding to C. trachomatis. In C. trachomatis-infected cells, the βcatenin was found to recruit to the chlamydial inclusion [29], which may suppress the trans-activation activity of β-catenin. Nevertheless, the current study has provided the first experimental demonstration for the involvement of an inclusion membrane protein in manipulating the Wnt signaling pathway and promoting the C. pneumoniae anti-apoptotic activity.