Introduction

Chlamydia psittaci is a zoonotic pathogen that can be transmitted from birds, which serve as its primary host, to mammalian species, including humans. In humans, C. psittaci causes the respiratory disease psittacosis, which is characterized by influenza-like symptoms, including fever, chills, headache, coughing, and dyspnoea. The severity of psittacosis ranges from asymptomatic cases to severe disease and even death.

C. psittaci belongs to the genus of Chlamydia, a group of gram-negative, obligate intracellular pathogenic bacteria1,2. During the intracellular development of Chlamydia, the bacteria differentiate between two forms: The infectious elementary bodies (EBs) and the non-infectious, replicative reticulate bodies (RBs)3,4,5. After primary attachment of EBs at the plasma membrane of the host cell, EBs invade by using the host-cellular endocytosis or phagocytosis pathway3,6. Inside of the endocytic vacuole, Chlamydia built their replicative niche termed the chlamydial inclusion by type-3 secretion of effector proteins into the inclusion membrane (Inc-proteins) or into the cytosol7. After differentiation of EBs to RBs, rapid replication of Chlamydia occurs within the inclusion, and this growth is supported by the acquisition of host cellular metabolites and reorganization of cellular organelles8,9,10,11,12,13. For example, the host cellular ceramide transporter (CERT) is recruited to C. psittaci inclusions and is crucial for inclusion growth and infectious progeny formation13. After rapid replication, RBs perform secondary differentiation back to EBs and C. psittaci egresses from the host cell, either by lytic egress or by non-lytic Chlamydia-containing sphere (CCS) formation14. This sequential process is calcium dependent and involves loss of inclusion membrane integrity, proteolytic activity, an increase in the intracellular calcium concentration, plasma membrane blebbing, and the final detachment of the whole cell as a CCS filled with infectious EBs14. In contrast, other chlamydial species as C. trachomatis can perform extrusion formation, another unique non-lytic egress pathway, which is characterized as a cytoskeleton-dependent packaged release process, where both the inclusion membrane and the plasma membrane remain intact and infectious bacteria are released without host cell death15,16,17,18.

Besides its functional role as a sphingolipid transporter, CERT has also a mechanistic role during chlamydial infections in stabilizing ER-inclusion membrane contact sites (MCS)19,20. MCS between the ER and chlamydial inclusions were at first characterized for C. trachomatis and are supported by the interaction between the bacterial Inc-protein IncD, which is localized in the inclusion membrane and CERT, which localizes to the host ER membrane20. However, no IncD analog of C. psittaci is characterized yet, although the recruitment of CERT to C. psittaci inclusions was described3,13. For C. trachomatis, additional proteins are characterized that stabilize ER-inclusion MCS and thus also stabilize the chlamydial inclusion. These proteins include the ER-resident proteins VAPA/B, interacting with the chlamydial Inc-protein IncV and the ER calcium sensor STIM1, interacting with the chlamydial Inc-protein IncS21,22,23,24,25,26,27,28. Interestingly, STIM1 depletion inhibits egress of C. trachomatis by extrusion formation and deletion of IncS of C. trachomatis enforces egress by host cell lysis, both suggesting that the stability of ER-inclusion MCS regulate egress of Chlamydia spp.17,28. Interestingly, for C. psittaci no IncS homolog, but a IncV homolog was described based on genomic analysis29.

Here, we examine how CERT recruitment to C. psittaci inclusions influences the regulation of C. psittaci egress. Our findings demonstrate that the recruitment of CERT to the C. psittaci inclusion is diminished in late infections, coinciding with the formation of CCS. The absence of CERT at the inclusion membrane of C. psittaci either by knockout of CERT or ectopic expression of a CERT variant lacking the PH domain in CERT-KO cells, resulted in the formation of premature, non-infectious CCS at the mid-infection stage. The findings suggest that the presence of CERT at the inclusion membrane depends on the developmental progress of the bacteria, regulates the stability of chlamydial inclusions, and thereby influences the time point at which egress of C. psittaci occurs via CCS formation.

Results

Temporal profile of C. psittaci CCS formation correlates with a reduction in CERT recruitment

The recruitment of CERT to C. psittaci inclusions was exclusively studied for early and middle infections, and not for late infections13. Therefore, this study concentrated on the localization of CERT during late C. psittaci infections in the context of bacterial egress by CCS formation. At first, we analyzed CERT recruitment to C. psittaci inclusions by performing immunofluorescence staining using antibodies specific for CERT, and bacterial Hsp60. Hsp60-positive C. psittaci-inclusions strongly recruit CERT at 24 h post infection (h pi). However, this recruitment was reduced to weak or no recruitment at 48 h pi (Fig. 1A, right and left inclusion, respectively). Similar to this, ectopically expressed eGFP-CERT localized in close proximity to C. psittaci inclusions at 24 h pi, but not at 48 h pi (Fig. 1B).

Fig. 1
figure 1

In late infections, CERT recruitment to C. psittaci inclusions is reduced and CERT-KO induce early egress of C. psittaci by CCS formation at 24 h pi. (A) Representative immunofluorescence images of C. psittaci-infected HeLa cells (MOI 2) at 24 and 48 h pi. PFA-fixed cells were stained for C. psittaci and endogenous CERT using a mouse-anti-Hsp60 (Cy3) and a rabbit-anti-CERT (AF488) antibody, respectively. DNA was counterstained using DAPI. n = 3. (B) Representative fluorescence images of HeLa cells transiently expressing eGFP-CERT infected with C. psittaci (MOI 2) at 24 and 48 h pi. Cells were PFA-fixed and immunostained for C. psittaci using a mouse-anti-Hsp60 (Cy3) antibody and DNA was counterstained using DAPI. n = 3. (C) Quantification of CERT recruitment to C. psittaci inclusions at 24 and 48 h pi. Based on immunofluorescence images of C. psittaci-infected HeLa cells (MOI 2) at 24 and 48 h pi, the fluorescence intensity (in artificially units) of CERT at the inclusions was quantified. At least 59 inclusions per condition were analyzed. Data show mean ± SEM; n = 3; ***p < 0.001 (Student’s t‐test). (D) CCS formation in HeLa CERT-KO and KO control cells. Cells were infected with C. psittaci (MOI 2), medium was replaced at 20 h pi and CCS in the supernatant were visually quantified at 24 h pi. Data show mean ± SEM; n = 3; **p < 0.01 (Student’s t‐test).

Full size image

Subsequently, the recruitment of CERT to C. psittaci inclusions was quantified at 24 and 48 h pi. The mean fluorescence intensity of CERT antibody labelling at C. psittaci inclusions was significantly reduced at 48 h pi in comparison to 24 h pi, from 30.7 to 14.6 relative units (Fig. 1C).

We therefore sought to determine whether the loss of CERT at the inclusion membrane is associated with egress by CCS formation. To this end, HeLa CERT-KO cells were infected with C. psittaci. We hypothesized that the absence of CERT at the inclusion membrane during intracellular development could lead to premature formation of CCS. We therefore quantified CCS in the supernatant of C. psittaci-infected cells at 24 h pi and compared CERT-KO with KO control cells. An almost tenfold increase in formed CCS was observed in the absence of CERT, from 2.0 × 10⁴ CCS/cm² in KO control cells to 19.6 × 10⁴ CCS/cm² in CERT-KO cells at 24 h pi (Fig. 1D). Interestingly, 84% of the inclusions detected in CERT-KO cells at 16 h pi formed CCS between 16 and 24 h pi (Supplemental Figure S1 A). In line with this observation, the number of CCS formed between 68 and 72 h pi is reduced under CERT-KO conditions compared to control cells (Supplemental Figure S1 B).

In conclusion, the data demonstrate that the absence of CERT at the inclusion membrane of C. psittaci correlates temporally with egress by CCS formation, indicating that alterations in CERT recruitment to C. psittaci inclusions regulate CCS formation.

Localization of CERT regulates CCS formation

CERT is composed of three functional domains: The N-terminal PH domain, mediating Golgi targeting in uninfected cells, the central FFAT motif, mediating ER targeting, and the C-terminal domain, which is a ceramide-specific lipid-binding START domain (Fig. 2A). During C. trachomatis, C. muridarum, and C. psittaci infections CERT is recruited to the inclusion by its PH-domain13,20,30.

Fig. 2
figure 2

In contrast to full CERT and CERT variants lacking the START- or FFAT-domain, a CERT variant lacking the PH domain is not recruited to C. psittaci inclusions and cannot prevent premature CCS formation. (A) CERT is composed of the N-terminal PH domain, mediating Golgi targeting in uninfected cells, the central FFAT motif, mediating ER targeting, and the C-terminal START-domain, that is a ceramide-specific lipid binding domain. (B) Representative fluorescence images of HeLa CERT-KO cells transiently expressing eGFP-CERT or eGFP-CERT variants lacking the START, FFAT or PH domain and controls infected with C. psittaci (MOI 2) at 24 h pi. Cells were PFA-fixed and immunostained for C. psittaci using a mouse-anti-Hsp60 (Cy3) antibody; DNA was counterstained using DAPI; n = 4. (C) CCS formation in HeLa CERT-KO cells transiently expressing eGFP‐CERT or eGFP-CERT variants lacking the START, FFAT or PH domain and controls. Cells were infected with C. psittaci (MOI 2), medium was replaced at 20 h pi and CCS in the supernatant were visually quantified at 24 h pi. CCS number was normalized on the transfection efficiency. Data show mean ± SEM; n = 4; *p < 0.05, **p < 0.01 (Student’s t‐test).

Full size image

Initially, we infected CERT-KO cells ectopically expressing distinct eGFP-CERT variants with C. psittaci and examined the localization of the eGFP-CERT variants through immunofluorescence staining at 24 h pi. In cells that had not been transfected with the CERT-GFP construct, no GFP signal was observed (Fig. 2B). In CERT-KO cells transfected with full-length eGFP-CERT, the protein was observed to localize in close proximity to C. psittaci inclusions (Fig. 2B), a finding that is comparable to those observed in wildtype HeLa cells (Fig. 1B). A comparable staining pattern was observed following transfection with eGFP-CERT lacking the START or FFAT domain (Fig. 2B). As published, eGFP-CERT lacking the PH domain was not recruited to C. psittaci inclusions13.

Fig. 3
figure 3

CERT-KO-induced early CCS are formed by an analogous sequence of events as mature CCS. (A) During early CCS formation, plasma membrane blebbing occurs and the inclusion membrane destabilizes. HeLa CERT-KO cells transiently expressing eGFP were infected with C. psittaci (MOI 2) and monitored live starting at 20 h pi using a CLSM. Panels show representative images of a section plane of a CCS-forming CERT-KO cell; The inclusion is marked with an asterisk; n = 3. (B) EqtSM is recruited to C. psittaci inclusions during early CCS formation. CERT-KO cells transiently expressing EqtSM-HaloTag were infected with C. psittaci (MOI 2). At 20 h pi, cells were labeled with 200 nM Janelia Fluor 585 HaloTag Ligand (Promega) and monitored live. Panel shows representative images of an EqtSM-HaloTag-recruiting C. psittaci inclusion during early CCS formation; n = 3. (C) A DEVD-cleaving protease is activated during early CCS formation. C. psittaci-infected CERT-KO and KO control cells (MOI 2) were stained with Incucyte Caspase-3/7 Dye for Apoptosis at 16 h pi and monitored using a CLSM for 8 h. Mean fluorescence intensity of infected cells at 16, 20, and 24 h pi was normalized to mean fluorescence intensity of uninfected cells at respective time points. Data show mean ± SEM; n = 3; *p < 0.05 (Student’s t-test). (D) Early CCS formation depends on the extracellular calcium concentration. CERT-KO cells were infected with C. psittaci (MOI 2). At 20 h pi, culture medium was replaced by serum-free medium supplemented with the indicated concentrations of calcium chloride. CCS in the supernatant were visually quantified at 24 h pi. Data show mean ± SEM; n = 3; *p < 0.05 (Student’s t‐test). (E, F) CERT-KO cells were infected with C. psittaci (MOI 2). At 20 h pi, medium was replaced by serum-free medium supplemented with 0 mM calcium chloride (E) or 1.8 mM calcium chloride (F). At 22 h pi, cells were labeled with the calcium sensor Rhod-3 for 1 h. At 24 h pi, DNA was counterstained using Hoechst and images were acquired. Representative images of intact cells (0.0 mM CaCl2) and CCS formation (1.8 mM CaCl2) are shown; n = 2.

Full size image

Next, we analyzed CCS formation after infecting CERT-KO cells ectopically expressing different eGFP-CERT variants with C. psittaci. In untransfected CERT-KO cells, 36.0 × 104 CCS/cm2 were formed compared to 4.9 × 104 CCS/cm2 in untransfected KO control cells (Fig. 2C). In the conditions in which the CERT variants were introduced by transient transfection, the CCS numbers were normalized to the transfection efficiencies. Upon transfection of CERT-KO cells with full length eGFP-CERT, the normalized CCS formation (4.5 × 104 CCS/cm2) resembles that detected in the control cells, indicating that full length CERT can complement the premature egress phenotype seen in the CERT-KO cells. Similarly, transfection of CERT-KO cells with eGFP-CERT lacking the START or FFAT domain significantly decreased CCS formation to 7.9 × 104 and 6.1 × 104 CCS/cm2, respectively. In contrast, transfection of CERT-KO cells with eGFP-CERT lacking the PH domain did not decrease CCS formation (40.7 × 104 CCS/cm2) (Fig. 2C).

Taken together, expression of CERT variants revealed that the ability of CERT to be recruited to the inclusion correlates with its capacity to prevent premature egress and suggest that CERT recruitment to the inclusion controls CCS formation.

CERT-KO induces early CCS formation by a sequence of events analogous to mature CCS formation

Next, we asked if the early egress seen in infected CERT-KO cells has the same mechanistic characteristics as the (mature) CCS formation that occurs in wildtype cells at the culmination of C. psittaci development. Mature CCS are formed by a process with distinct characteristics14. An increase in intracellular calcium precedes protease activity that can be detected by cleavage of a DEVD tetrapeptide-containing substrate. Loss of inclusion membrane integrity can be detected by recruitment of EqtSM caused by exposure of sphingomyelin to the cytoplasm. Plasma membrane blebbing then leads to eventual detachment of the CCS14. Thus, we examined the sequence of events described for mature CCS during the early CCS formation induced in CERT-KO cells.

At first, we monitored early CCS formation using HeLa CERT-KO cells transiently expressing eGFP starting at 20 h pi. Before CCS formation, the bacterial inclusion excluded the cytosolic eGFP (Fig. 3A). Then in all analyzed cells, CCS formation initiated with blebbing of the cellular plasma membrane and influx of eGFP into the inclusion lumen. This was followed by enlargement of the plasma membrane blebs and subsequent detachment of the entire host cell, which completed the formation of the CCS in the supernatant of the cell culture.

Fig. 4
figure 4

Premature, CERT-KO-induced CCS contain mainly non-infectious reticulate bodies. (A) CCS in the supernatant of C. psittaci-infected CERT-KO and KO control cells (MOI 2) were separated from free bacteria by centrifugation (300 x g, 5 min, RT) at 24 h pi. Infectious progeny was titrated after glass bead lysis and numbers were normalized to genome copy numbers determined by qRT-PCR. Data show mean ± SEM; n = 3. (B) Transmission electron microscopy (TEM) of a thin section through the chemically fixed supernatant of C. psittaci-infected HeLa CERT-KO cells (MOI 2, 24 h pi). Numerous Chlamydia, mostly in RB stage, were found; n = 3. The image shows a group of RBs (*) which are associated with cellular debris, which is the main component of the supernatant. (C) Representative fluorescence images of an early live CCS isolated of the supernatant of C. psittaci-infected HeLa CERT-KO cells (MOI 2, 24 h pi). The surrounding membrane was visualized using the membrane marker FM 4–64 and DNA was counterstained by Hoechst; n = 3. (D) Representative fluorescence images of a PFA-fixed early CCS isolated from C. psittaci-infected HeLa CERT-KO cells (MOI 2, 24 h pi). Bacteria inside the CCS were detected using a chlamydial Hsp60 (Cy3) antibody and the DNA was counterstained using DAPI; n > 3.

Full size image

Next, we asked if EqtSM is recruited to the C. psittaci-inclusion when early CCS formation starts. To test this, we transiently expressed EqtSM-HaloTag in CERT-KO cells, labeled them with HaloTag Ligand and performed live cell imaging starting at 20 h pi. We observed many cells recruiting EqtSM to the inclusion coincident with the start of plasma membrane blebbing (Fig. 3B) and followed by CCS formation.

In addition, we tested if proteolytic cleavage of a DEVD-containing substrate is increased in adherent CERT-KO cells during early CCS formation. Indeed, we detected a significant increase in fluorescence intensity of a cleaved DEVD-containing substrate compared to uninfected cells from 1.0-fold at 16 h pi to 1.8-fold and 1.9-fold at 20 and 24 h pi, respectively (Fig. 3C). In comparison, in control cells the intensity was only slightly increased from 1.1-fold at 16 h pi to 1.2-fold and 1.3-fold at 20 and 24 h pi, respectively.

Furthermore, the role of calcium signaling for early CCS formation was investigated. During mature CCS formation, CCS numbers increased with extracellular calcium concentrations14. Similarly, during early CCS formation in CERT-KO cells, we observed a dose-dependent increase in CCS numbers from 3.4 × 104 CCS/cm2 to 26.1 × 104 CCS/cm2 when increasing the extracellular calcium from 0.0 to 1.8 mM CaCl2 (Fig. 3D).

Subsequently, we analyzed the intracellular calcium concentration during early CCS formation using the Rhod-3 membrane-permeant cytosolic calcium sensor under conditions containing 0.0 and 1.8 mM calcium chloride. At 0.0 mM calcium chloride, we did not detect an increase in cytosolic calcium concentration in C. psittaci-infected CERT-KO cells at 24 h pi (Fig. 3E). However, at 1.8 mM calcium chloride, we observed C. psittaci-infected CERT-KO cells with blebbing membranes and an increased cytosolic calcium concentration (Fig. 3F).

In sum, this data revealed that CERT-KO induced early CCS are formed by an analogous sequence of events as mature CCS including plasma membrane blebbing, inclusion membrane destabilization, EqtSM recruitment, activation of a DEVD cleaving protease, dependency on the extracellular calcium concentration and intracellular calcium increase. This further supports that the presence and absence of CERT at the inclusion regulates CCS formation.

CERT-KO induces the formation of RB-containing, non-infectious CCS

At 24 h pi, C. psittaci inclusions contain mostly RBs, with only few bacteria starting to perform secondary differentiation to EBs31. Thus, we asked if bacteria in early-formed CCS are infectious. To test this, we separated CCS from free bacteria present in the supernatant of C. psittaci-infected CERT-KO and control cells at 24 h pi using differential centrifugation. We then quantified the infectious progeny (IFU) relative to the bacterial genome copy numbers associated with the harvested CCS. Although more CCS are released in the supernatants from CERT-KO cells than from control cells 24 h pi, the percentage of the progeny that are infectious are comparable (0.003 ± 0.0016% and 0.006 ± 0.0038%, respectively) (Fig. 4A). These values 24 h pi are extremely reduced compared to the percentage of infectious progeny in mature CCS released from HeLa cells at 48 h pi (3.2 ± 0.9%)14.

To assess the presence and morphology of the C. psittaci released from infected CERT-KO cells, we analyzed the culture supernatants using thin-section transmission electron microscopy (TEM). Among cellular debris, which is the main component of the supernatant, we found numerous Chlamydia, which were, with few exceptions, in RB stage (Fig. 4B). We further analyzed the morphology of early CCS by staining with the membrane dye FM 4–64 and with Hoechst. Fluorescence live cell confocal microscopy revealed that early CCS were surrounded by a membrane and contained concentrated DNA, which are both features of mature CCS (Fig. 4C)14. Bacteria could be detected in fixed early CCS with antibodies specific for bacterial Hsp60 (Fig. 4D). Bacteria were also labeled with the DNA marker DAPI, which also labeled DNA we identified as the host cell nucleus, which is also found within mature CCS14.

These data show that CCS released early in the absence of CERT rarely contain any infectious EBs although the CCS are morphologically similar to mature CCS. In wildtype infections, CERT recruitment is reduced after RB perform secondary differentiation to the EB stage, indicating that stage-controlled changes of the localization of CERT allow release of infectious progeny containing CCS.

Discussion

The stability of pathogen-containing vacuoles is crucial for intracellular vacuolar pathogens. On the one hand, stabilization of the vacuole facilitates pathogen growth in a protected niche, on the other hand, the vacuole must allow pathogen egress from host cells and transmission to other hosts32,33,34. For the zoonotic bacterial pathogen C. psittaci, the predominant egress strategy is via CCS formation, a process where bacteria are released from a destabilized inclusion to the host cytosol and the whole host cell subsequently detaches from the remaining cellular monolayer as a CCS14. This process depends on the extracellular calcium concentration, bacterial protease activity and a rapid increase in the cytosolic calcium concentration14. However, it remains unclear how the stage-dependent stabilization and destabilization of the inclusion membrane is regulated to ensure completion of intracellular development before egress by CCS formation.

In this study, we showed that CERT recruitment to the inclusion membrane is time-dependent: During RB replication at 24 h pi, CERT is recruited to the inclusion membrane. Here it can stabilize ER-inclusion MCS19,20. In contrast, during CCS formation at 48 h pi, CERT recruitment is reduced and the inclusion loses stability (Fig. 5). Premature loss of CERT recruitment, which can be induced by using CERT-KO cells, induces the formation of premature, non-infectious RB-containing CCS by a sequence of events analogous to mature CCS formation (Fig. 5)14. We propose that the time-dependent recruitment of CERT stabilizes the inclusion by ER-inclusion MCS during RB replication and that reduction of CERT at the inclusion after EB maturation allows destabilization of the inclusion by the loss of ER-inclusion MCS prior to CCS formation.

Fig. 5
figure 5

Graphical model of the role of CERT in regulating C. psittaci CCS formation. At 24 h pi, the recruitment of CERT to the C. psittaci inclusion promotes inclusion stability by CERT-stabilized ER-inclusion membrane contact sites. At 48 h pi, CERT recruitment to the C. psittaci inclusion is reduced and the consequential loss of the inclusion stability facilitates CCS formation. When CERT is not present at the inclusion membrane at 24 h pi (CERT-KO or CERT∆PH), this promotes early egress and facilitates the formation of premature, non-infectious CCS.

Full size image

The role stability of ER-inclusion MCS plays in egress of Chlamydia spp. was previously discussed in terms of the IncS-STIM1 interaction at C. trachomatis inclusions25,27,28. C. trachomatis can egress from infected cells via extrusions, and unlike during CCS formation, the inclusion membrane remains intact when extrusions form. Destabilization of ER-inclusion MCS either by a lack of STIM1 induced by siRNA-mediated knock-down or by a lack of C. trachomatis IncS shifted C. trachomatis egress to reduce extrusion formation or increase inclusion lysis, respectively17,28. Notably, C. psittaci does not recruit STIM1 at 24 and 48 h pi (Supplemental Figure S2) and extrusion formation takes not place in C. psittaci infections14. Perhaps the absence of STIM1-containing MCS result in reduced stability of the C. psittaci inclusions compared to C. trachomatis inclusions, and this feature could preclude C. psittaci extrusion formation. Following this model, CERT recruitment would be more important for the stabilization of C. psittaci ER-inclusion MCS compared to C. trachomatis ER-inclusion MCS, which may explain the influence CERT recruitment has in regulating CCS formation in C. psittaci in a time-dependent manner. Interestingly, STIM1 is also not recruited to C. muridarum inclusions28. It is therefore tempting to speculate that also in C. muridarum egress could be regulated by the recruitment of CERT.

CERT does not only stabilize ER-inclusion MCS in chlamydial infections, but it also plays a crucial role for sphingolipid acquisition of Chlamydia spp.13,20,30,35,36,37. In C. trachomatis, C. muridarum and C. psittaci infections, chemical inhibition of CERT with the ceramide analogue HPA-12 reduces inclusion size and progeny formation13,20,30. Interestingly, CERT-KO does not prevent sphingolipid acquisition at the inclusion in C. psittaci infections, but HPA-12 treatment does, indicating that a CERT-independent, HPA-12-sensitive route of ceramide acquisition is used by C. psittaci in CERT-KO cells13. Accordingly, EqtSM recruitment to sphingomyelin exposed to the inclusion membrane prior to CCS formation can be observed during CCS formation in CERT-KO cells. In addition, expression of a CERT variant lacking the START domain, which is not able to transport ceramides, but localizes at the inclusion membrane, can compensate for the CERT-KO-induced premature CCS formation. These results further indicate that premature CCS formation in the absence of CERT is due to the loss of its mechanistic role in stabilizing ER-inclusion MCS and is not based on its role in sphingolipid acquisition during C. psittaci infection.

To further understand how the time-dependent recruitment of CERT is regulated, it will be crucial to identify the bacterial interaction partner of CERT within the C. psittaci inclusion membrane. For several chlamydial species, including C. trachomatis, C. muridarum, C. suis, C. caviae, and C. felis, the bacterial Inc-protein IncD interacts with the PH domain of CERT at the inclusion membrane13,21,30,35. However, no IncD analog has yet to be identified in C. psittaci, even though the PH domain of CERT is involved in its recruitment to the C. psittaci inclusion and in its interaction with IncD3,13.

Taken together, we showed that the recruitment of CERT stabilizes C. psittaci inclusions during RB replication and prevents premature CCS formation. Our data suggest that loss of CERT recruitment in late infections facilitates the destabilization of ER-inclusion MCS and the chlamydial inclusion membrane and consequently promotes CCS formation and bacterial egress.

Materials and methods

Cell culture, transient transfection and infection assays

Cell culture, transient transfection and infection assays were performed as described previously13,14. Briefly, HeLa cells (ATCC CCL-2) were cultivated in RPMI 1640 medium (Gibco) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich), 1 mM sodium pyruvate (Gibco), and 2 mM L-glutamine (Gibco) at 37 °C and 5% CO2 and passaged every 2 to 3 days. Transient transfection was performed with Lipofectamine 2000 reagent (Thermo Fisher) 4 h prior to infection according to manufacturer’s instructions. For infection assays, subconfluent HeLa cells were washed with infection medium (Dulbecco’s modified Eagle’s medium (DMEM, Gibco) with glucose (4.5 g L− 1) supplemented with 5% FBS, 1 mM sodium pyruvate, and 2 mM l‐glutamine) and incubated at 35 °C and 5% CO2 using the indicated multiplicity of infection (MOI) with C. psittaci strain 02DC15 38. C. psittaci-infected cell cultures were centrifuged (30 min, 600 x g, room temperature (RT)) at 30 min pi and washed with infection medium at 2 h pi.

Quantification of CERT recruitment

To quantify the recruitment of CERT to C. psittaci inclusions, immunofluorescence assays were performed using a CERT antibody and a chlamydial Hsp60 antibody as primary antibodies (see Immunofluorescence assay). The fluorescence intensity of CERT of at least 59 individual inclusions per condition being positive for Hsp60 was analyzed under comparable imaging conditions.

Quantification of chlamydial genome copy number (GCN) by quantitative real-time PCR

Determination of bacterial genome copy numbers by quantitative real-time PCR (qRT-PCR) was performed as already described39.

Quantification of CCS

The number of CCS in C. psittaci infected cell cultures was determined as described previously14. In brief, cell cultures were washed with PBS at indicated time points and medium was changed to conditional medium and cultivation was continued for additional 4 h. The number of CCS in the supernatant of the infected cell cultures was quantified in 12 well plates. The average number of CCS of at least five visual fields was determined and normalized to the infected cell area.

Live cell imaging of premature C. psittaci egress in CERT-KO cells

To monitor early CERT-KO-induced egress of C. psittaci, live cell imaging of Chlamydia-infected CERT-KO cells transiently expressing eGFP was performed as described previously14. In brief, live cell microscopy of infected cells cultured in 8 well chambered coverslips (µ-Slide 8 Well, Ibidi) was performed at a Stellaris 8 Confocal Microscope (Leica Microsystems) equipped with a live cell chamber at 35 °C. Z-stacks of 12 slices with each 3 μm distance covering both adherent cells and the supernatant containing CCS were acquired every 2–2.5 min from 20 to 24 h pi.

Imaging of cytosolic sphingomyelin (SM) exposure

To image the cytosolic SM exposure at the inclusion membrane, CERT-KO cells transiently expressing a non-toxic, HaloTagged version of equinatoxin II (EqtSM-HaloTag) as cytosolic SM reporter were used40,41. Cell culture, transient transfection and infection with C. psittaci (MOI 2) were performed in 8 well chambered coverslips (µ-Slide 8 Well, Ibidi). At 19.5 h pi, cells were labeled with 200 nM Janelia Fluor 585 HaloTag Ligand (Promega) in infection medium for 30 min. Subsequently, cells were washed with infection medium and live cell imaging was performed at a Stellaris 8 Confocal Microscope (Leica Microsystems) equipped with a live cell chamber at 35 °C. Z-stacks of 12 slices with each 3 μm distance were acquired every 2:10 min for a total of 4 h.

Determination of proteolytic DEVD cleaving activity

Determination of proteolytic DEVD cleaving activity was performed as described previously14. For this, the Incucyte Caspase-3/7 Red Dye for Apoptosis (Sartorius) was used as described in the manufacturer’s protocol. Cells were cultured and infected with C. psittaci (MOI 2) in 8 well chambered coverslips (µ-Slide 8 Well, Ibidi). Incucyte Caspase-3/7 Red Dye was diluted in infection medium to a final assay concentration of 833 nM and added to C. psittaci-infected cell cultures at 16 h pi. Live imaging was performed using an LSM 780 CLSM (Carl Zeiss) equipped with a live cell chamber at 35 °C. Z-stacks of 12 slices with each 3 μm distance covering both adherent cells and CCS were acquired every 5 min for a total of 8 h.

Calcium imaging

Imaging of intracellular calcium levels and distribution was performed by Rhod 3 staining using the Rhod-3 Calcium Imaging Kit (Thermo Fisher Scientific) following the manufacturer’s protocol as described previously14. In brief, CERT-KO cells cultured in 8 well chambered coverslips (µ-Slide 8 Well, Ibidi) were infected with C. psittaci (MOI 2) and at 20 h pi, infected cells were washed with PBS and medium was changed to calcium- and FBS-free infection medium (Cat. No. 21068028, Gibco) supplemented with the indicated concentrations of calcium chloride (Roth). At 22 h pi, cells were washed with phosphate buffered saline (PBS) twice and 150 µL per well of freshly prepared Rhod-3 loading buffer (1x PowerLoad concentrate, 10 µM Rhod-3 AM, 250 µM Probenecid in infection medium or calcium- and FBS-free infection medium) were added. After incubation (60 min, 35 °C, 5% CO2), cells were washed three times with PBS and 150 µL per well of incubation buffer (250 µM Probenecid in infection medium or calcium- and FBS-free infection medium) were added. After incubation (60 min, 35 °C, 5% CO2), cells were washed with PBS and 150 µL per well of 25 mg mL− 1 Hoechst 33,342 (Sigma-Aldrich) diluted 1:2000 in infection medium were added for live cell microscopy. Live cell microscopy was performed at a Stellaris 8 Confocal Microscope (Leica Microsystems).

Separation of CCS and free bacteria for reinfection assay

HeLa CERT-KO and KO control cells were initially infected with C. psittaci and incubated for 24 h. Then, supernatants were collected and centrifuged at low speed (5 min, 300 x g, RT) to pellet CCS and separate them from free bacteria in the remaining supernatant. Subsequently, CCS and free bacteria were subjected to glass bead lysis followed by infection of freshly seeded HeLa cells. Cells were fixed at 24 h post infection (pi), stained for Hsp60 (see immunofluorescence assay for more details), and numbers of infection-forming units (IFU) per mL were calculated from average inclusion counts in 10 fields of view per condition. IFU were normalized on GCN (see quantification of chlamydial genome copy number (GCN) by quantitative real-time PCR for more details).

Transmission electron microscopy

Transmission electron microscopy was performed as described previously14. In brief, at 24 h pi the culture supernatant of C. psittaci-infected (MOI 2) HeLa CERT-KO cells (sedimented by centrifugation for 5 min at 300 x g) were fixed with a mixture of glutaraldehyde and formaldehyde (2.5% and 1%, respectively; in 50 mM HEPES buffer). After incubation for 2 h at RT, suspensions were centrifuged (6000 x g, 10 min) and sediments were embedded in low-melting point agarose. Small sample blocks were extracted from the solidified agarose, post fixed in osmium tetroxide (1% in distilled water) followed by block contrasting with tannic acid (0.1% in 50 mM HEPES buffer) and uranyl acetate (2% in distilled water). Subsequently, samples were dehydrated by incubation in a graded ethanol series and embedded in epon resin. After polymerization for 48 h at 60 °C, ultrathin sections (approx. 70 nm) were prepared using an ultramicrotome (UC7; Leica) and stained with uranyl acetate (2% in distilled water, 20 min) followed by lead citrate (2 min) to increase contrast. Sections were examined with a transmission electron microscope (Tecnai 12, FEI Corp.) at 120 kV. Images were recorded using a CMOS-camera (Phurona, Emsis).

Isolation and staining of CCS

CCS in the supernatant of C. psittaci-infected cell cultures were separated by centrifugation (5 min, 300 x g, RT) at 24 h pi. For live staining, pellet was mixed with 10 mM FM4-64 in DMSO (Sigma-Aldrich) diluted 1:1000 in infection medium and 25 mg mL− 1 Hoechst 33,342 (Sigma-Aldrich) diluted 1:2000 in infection medium. CCS were transferred into a 15 well chambered coverslip (µ-Slide 15 Well 3D, Ibidi) and examined using an LSM 780 CLSM (Carl Zeiss). For immunofluorescence staining, pellet was mixed with 4% paraformaldehyde in PBS and transferred into a poly-l-lysine coated 8 well chambered coverslip (µ-Slide 8 Well, Ibidi). After 30 min of incubation at RT, staining was continued as described for immunofluorescence assays.

Immunofluorescence assay

For immunofluorescence assays, C. psittaci-infected (MOI 2) or uninfected HeLa cells, HeLa CERT-KO cells or HeLa KO control cells were fixed with 4% paraformaldehyde in PBS (30 min, RT) at the indicated time points. After washing with PBS, samples were blocked and permeabilized using 0.2% Triton X-100 and 0.2% bovine serum albumin (BSA) in PBS (20 min, RT). Incubation with primary antibodies (1 h, RT) was followed by washing and incubation with secondary antibodies and DAPI (25 µg ml− 1, Sigma-Aldrich) (1 h, RT). All antibodies were diluted using 0.2% BSA in PBS. Antibodies were used at the following concentrations: mouse anti‐Hsp60 (chlamydial, cHsp60; 1:600, Cat. No. ALX‐804‐072, Enzo Life Sciences), rabbit anti-CERT (1:200, Cat. No. PA5-115035, Invitrogen), rabbit anti-STIM1 (1:50, Cat. No. 11565-1-AP, Proteintech), Alexa Fluor 488‐coupled goat anti-rabbit IgG (1:100, Cat. No. 111‐545‐144, Dianova), and Cy3‐coupled goat anti-mouse IgG (1:200, Cat. No. 115‐165‐146, Dianova). All samples were analyzed at a Stellaris 8 Confocal Microscope (Leica Microsystems).