The Chlamydia effector CpoS modulates the inclusion microenvironment and restricts the interferon response by acting on Rab35

ABSTRACT The obligate intracellular bacterium Chlamydia trachomatis inserts a family of inclusion membrane (Inc) proteins into the membrane of its vacuole (the inclusion). The Inc CpoS is a critical suppressor of host cellular immune surveillance, but the underlying mechanism remained elusive. By complementing a cpoS mutant with various natural orthologs and variants of CpoS, we linked distinct molecular interactions of CpoS to distinct functions. Unexpectedly, we found CpoS to be essential for the formation of inclusion membrane microdomains that control the spatial organization of multiple Incs involved in signaling and modulation of the host cellular cytoskeleton. While the function of CpoS in microdomains was uncoupled from its role in the suppression of host cellular defenses, we found the ability of CpoS to interact with Rab GTPases to be required not only for the manipulation of membrane trafficking, such as to mediate transport of ceramide-derived lipids (sphingolipids) to the inclusion, but also for the inhibition of Stimulator of interferon genes (STING)-dependent type I interferon responses. Indeed, depletion of Rab35 phenocopied the exacerbated interferon responses observed during infection with CpoS-deficient mutants. Overall, our findings highlight the role of Inc–Inc interactions in shaping the inclusion microenvironment and the modulation of membrane trafficking as a pathogenic immune evasion strategy. IMPORTANCE Chlamydia trachomatis is a prevalent bacterial pathogen that causes blinding ocular scarring and urogenital infections that can lead to infertility and pregnancy complications. Because Chlamydia can only grow within its host cell, boosting the intrinsic defenses of human cells may represent a novel strategy to fight pathogen replication and survival. Hence, CpoS, a Chlamydia protein known to block host cellular defenses, or processes regulated by CpoS, could provide new opportunities for therapeutic intervention. By revealing CpoS as a multifunctional virulence factor and by linking its ability to block host cellular immune signaling to the modulation of membrane trafficking, the present work may provide a foundation for such rationale targeting and advances our understanding of how intracellular bacteria can shape and protect their growth niche.


Chlamydia strains and infection
Experiments were carried out with C. trachomatis strain L2/434/Bu (CTL2, ATCC VR-902B) and the derivatives generated in this study (Table S2).Two different procedures were used to prepare infection inocula.First, to prepare crude bacterial preparations, bacteria were propagated in Vero cells and harvested at about 36-48 hpi by H2O-mediated lysis of host cells.This was followed by addition of 1/4 volume of 5X SPG (sucrose-phosphate-glutamate) buffer (375 g/l sucrose, 2.5 g/l KH2PO4, 6 g/l Na2HPO4, 3.6 g/l glutamic acid, pH 7.5), centrifugation at low speed (500 x g, 10 min, 4°C) to remove host cell debris, brief sonication, and centrifugation at higher speed (17,000 x g, 10 min, 4°C) to pellet bacteria.Bacteria were then resuspended in 1X SPG buffer, briefly sonicated, and stored at -80°C.Second, for selected experiments, EBs present in crude bacterial preparations were further purified by density gradient centrifugation.In brief, we first generated O95NaCl, i.e. a 95% dilution of omnipaque 350 (780 mOsmol/kg, GE Healthcare) supplemented with additional 211.25 mM NaCl for osmolality adjustment.Bacteria in SPG buffer were then sonicated and layered on top of 8 ml of a 30% dilution of O95NaCl (diluted in SPG), followed by ultracentrifugation (40,000 x g, 40 min, 4°C).Subsequently, the bacteria found in the pellet were resuspended in 1X SPG buffer, sonicated, and layered on top of a gradient consisting of 3 ml of 54% O95NaCl, 5 ml of 44% O95NaCl, and 2 ml of 40% O95NaCl (O95NaCl dilutions in SPG), followed by ultracentrifugation (40,000 x g, 1 h, 4°C).Purified EBs were then collected from the 44%/54% interface, diluted in 1X SPG buffer, and pelleted by centrifugation (30,000 x g, 30 min, 4°C).Subsequently, the bacteria were resuspended in 1X SPG buffer, again briefly sonicated, and stored at -80°C.Bacterial preparations prepared by either of the two approaches were shown to be free of Mycoplasma contamination by PCR, as described above for cell lines.To determine the number of infectious bacteria (i.e. the inclusionforming units (IFUs)) contained in the preparations, monolayers of Vero cells in 96-well plates were infected with serial dilutions of the bacteria.At 28 hpi, cells were fixed with 4% formaldehyde and inclusions were stained with antibodies targeting the chlamydial protein Slc1 (procedure described below).Inclusions were counted using a Cellomics ArrayScan VTI HCS automated imaging system (ThermoFisher).To conduct infections, cells were seeded in multiwell plates, followed by addition of bacteria (number of IFUs/cell as specified), centrifugation (1500 x g, 30 min, 23°C), and incubation (37°C, 5%CO2) for the indicated periods of time.

Gene disruption in Chlamydia
CTL0476 (ipaM) was disrupted in CTL2 via the TargeTron approach (10).The primers IBS-CTL0476, EBS1d-CTL0476, and EBS2-CTL0476 (Table S3), as well as EBS universal (Merck), were used to retarget vector pDFTT3 (10) for this purpose.CTL0481 (cpoS) was disrupted using a derivative of pDFTT3, which was retargeted towards CTL0481 (3) and modified (as described previously (11)) to contain a cat (chloramphenicol resistance) gene instead of a bla (β-lactamase resistance) gene in the intron.Bacteria were transformed using the CaCl2 approach (12) and selected in presence of 1 U/ml penicillin G (Merck) or 0.5 µg/ml chloramphenicol (Merck), first added at 12 hpi.Bacteria were then plaque-purified (3) in presence of 5 U/ml penicillin G or 1 μg/ml chloramphenicol.Intron insertion at correct target sites was verified by PCR (using primers shown in Table S3) and sequencing of the resulting PCR products (Eton Bioscience).

Gene expression in Chlamydia
To enable expression of CpoS orthologs and variants with a C-terminal FLAG tag, a DNA fragment encoding a FLAG tag and a stop codon was inserted between the KpnI and SalI restriction sites of vector p2TK2-SW2 (9).DNA fragments coding for CpoS proteins (without stop codon) and their promoters (210 bp sequence upstream of the start codon) were then PCRamplified from genomic DNA (using primers shown in Table S3), or obtained as synthetic gene blocks (IDT), and inserted between the AgeI and KpnI restriction sites of the vector.To enable expression of MYC-tagged Incs, the genes encoding the Incs (CTL0476 (ipaM), CTL0374 (incA), CTL0370 (incD)) and their promoters (210 bp sequence upstream of the start codon) were PCRamplified (using primers shown in Table S3), digested with AgeI (or SgrAI in case of CTL0476) and NheI, and inserted into AgeI/NheI digested p2TK2-SW2 (9).Vector p2TK2-SW2-mCherry (9), enabling constitutive expression of mCherry, was a kind gift from Isabelle Derré (University of Virginia).Vector pTL2-tetO-IncB-GFP11x7-flag (8), enabling inducible expression of IncB (CT232) tagged at its C-terminus with seven copies of GFP11 (RDHMVLHEYVNAAGIT) and one copy of FLAG (DYKDDDDK), was a kind gift from Kevin Hybiske (University of Washington).Bacteria were transformed as described above.Expression of IncB-GFP11x7-FLAG was induced by addition of 4 ng/ml anhydrotetracycline (Clontech) at 0 hpi (i.e. after the centrifugation step that followed addition of the bacteria).

Quantification of cell death and IFN responses
LDH activity in culture supernatants, an indicator for host cell lysis, was quantified using the photometric in vitro cytotoxicity kit (Merck), according to the manufacturer's instructions.Activity detected in cell-free medium was subtracted and values were normalized to activity detected in a total cell lysate to calculate the percentage of dead cells.The induction of type I IFN-dependent genes was assessed using an A2EN-ISRE-luciferase reporter cell line as previously described (3).Cells were lysed by addition of a 1:1 mixture of Hanks' Balanced Salt Solution (HBSS; Gibco) and Britelite plus reagent (PerkinElmer), before measurement of luminescence.
Luminescence values were normalized to the mean luminescence detected in uninfected (non-siRNA-treated) wells.Absorbance and luminescence measurements were made using an EnSpire 2300 (PerkinElmer), SpectraMax i3 (Molecular Devices) or Infinite 200 (Tecan) plate reader.The same instrument was used for all replicates of the same experiment.

Determination of infectious progeny
Infectious progeny was determined as previously described (3).In brief, confluent monolayers of HeLa cells in 96-well plates were infected using a low infection dose (< 50% infected cells).At 36 hpi, cell lysates were prepared by H 2O-based cell lysis.IFUs in the initial inoculum (input) and collected cell lysates (output) were quantified by infecting confluent monolayers of Vero cells with serial dilutions, followed by the fluorescence microscopic determination of inclusion numbers at 28 hpi (as described above for the titering of bacterial stocks).From the IFUs detected in the input and the output, the number of IFUs formed per infected cell was determined and then normalized to the IFU production observed for CTL2.

Ceramide (sphingolipid) acquisition
To measure the intra-inclusion accumulation of ceramide (and/or ceramide-derived lipids), cells were seeded in 96-well plates and infected (for the experiment shown in Fig 5D , mCherryexpressing derivatives of the indicated strains were used).At 14 hpi, cells were incubated for 15 min at 4°C, rinsed with cold HBSS (Gibco), and then incubated for 30 min at 4°C in HBSS containing 5 µM NBD C6-ceramide complexed to BSA (ThermoFisher).Subsequently, the cells were rinsed twice with growth medium and incubated for another 6 hours in medium (37°C, 5% CO 2).BFA (3 µg/ml) was added to some wells at 12 hpi and was also present during the incubation with ceramide, but not thereafter.After the 6-hour incubation period, cells were

siRNA-mediated gene silencing
Cells were transfected with siRNAs (Dharmacon siGENOME Human SMART pools, i.e. a mix of four gene-specific siRNAs, Table S4) at a concentration of 25 nM using DharmaFECT-1 transfection reagent (Dharmacon).To reduce toxicity, the transfection medium containing siRNAs and reagent was removed after a 6-hour incubation and replaced with fresh growth medium.In control transfections, RNA-free siRNA buffer (Dharmacon) was added instead of siRNAs.To determine knockdown efficiencies by western blotting, protein samples were prepared (as described below) at 46 hours post transfection.To determine effects of knockdowns on the IFN response during infection, cells were infected at 32 hours (reporter cell assay) or at 36 hours (detection of STING/IRF3 phosphorylation) post transfection with siRNAs and then further processed at 14 hpi.To determine effects of knockdowns on the IFN response after DNA stimulation (detection of STING/IRF3 phosphorylation), cells were transfected with DNA at 46 hours post transfection with siRNAs and then further processed four hours later.To test the effect of Rab35 depletion on Rab35 effector recruitment to inclusions, cells were transfected with Rab35 effector expression plasmids at 48 hours post transfection with siRNAs, infected five hours later, and analyzed at 24 hpi.

Co-immunoprecipitation
HeLa cells were infected with the indicated Chlamydia strains and when indicated transfected with EGFP-Rab expression plasmids at 1 hpi (see above).At 26 hpi, cells were lysed in Pierce IP lysis buffer containing Pierce Protease and Phosphatase Inhibitor (ThermoFisher).Lysates were homogenized by passing through a needle (0.33 x 12 mm) and cleared by centrifugation (17000-20000 x g, 4°C, 20 min).For immunoprecipitation of FLAG-tagged proteins, ANTI-FLAG M2 magnetic beads (Merck) were incubated with the lysate (500-800 μl, 0.6-1.0 mg/ml protein) for 6 hours at 4°C and then washed four times with Pierce IP lysis buffer.Proteins were eluted by incubation in 0.1 M glycine/HCl (pH 3.0) followed by neutralization by addition of 1/5 volume of 0.5 M Tris/HCl, 1.5 M NaCl (pH 7.4).For the co-immunoprecipitation of IPAM, proteins were instead eluted by boiling (3 min, 100°C) of the beads in Laemmli buffer (final concentration: 50 mM Tris (pH 6.8), 10% glycerol, 2% SDS, 50 mM DTT, traces of bromophenol blue).For immunoprecipitation of EGFP-fusion proteins, GFP-Trap_MA magnetic beads (Chromotek) were incubated with lysate (1350 μl, 0.6 mg/ml protein; diluted in TBS/EDTA buffer (10 mM Tris/HCl, 150 mM NaCl, 0.5 mM EDTA pH 7.5)) for 5 hours at 4°C and then washed four times with TBS/EDTA buffer.Proteins were eluted by boiling, as described above.Aliquots from the initial lysate (lysate fraction, also known as input), lysate after incubation with beads (unbound fraction), and the buffer used for the last wash step (wash fraction) were also collected.Samples were stored at -20°C until analysis by western blotting.

Western blot analysis
Protein extracts were prepared by cell lysis in boiling 1% SDS buffer (50 mM Tris/HCl (pH 7.5), 1% SDS, 0.1 M NaCl), as previously described (3), quantified using a BCA protein assay (Pierce), and adjusted to equal protein content by dilution with 1% SDS buffer.Protein extracts (and Co-IP samples) were then mixed with loading buffer (Laemmli buffer (see above) or stained with Hoechst 33342 (2 µg/ml, 10 min), washed twice with growth medium, and imaged live with an automated fluorescence microscope (ImageXpress Micro XL, Molecular Devices) at about 21 hpi.Inclusions were detected in MetaXpress (Molecular Devices) based on mCherry fluorescence (in experiments using mCherry-expressing strains) or NBD fluorescence (in all other experiments).The average intra-inclusion NBD fluorescence intensity was determined and normalized to the average intra-inclusion NBD fluorescence observed during infection with CTL2.
Images were taken with various microscopy systems, including epifluorescence microscopes (Zeiss Axio Observer.Z1, Zeiss Axio Imager.Z2), confocal microscopes (Zeiss LSM 780, Leica SP8), and high-content imaging platforms (ImageXpress Micro XL system (Molecular Devices), Cellomics ArrayScan VTI HCS (ThermoFisher)).The percentage of cells displaying STING translocation (accumulation of STING-HA in post-ER vesicles; as previously shown (3)) recruitment of EGFP-Rab fusion proteins to the inclusion (enrichment at inclusion membrane; examples shown in Fig 5A), or recruitment of HA-tagged Rab effector proteins to the inclusion, was determined by manual inspection of images taken with the ImageXpress Micro XL or Zeiss Axio Imager.Z2 systems.Except for the assessment of Rab effector recruitment, this analysis was conducted blinded, i.e. the individual conducting the counting was unaware of the sample identity at the time of counting.