Homologues of the Chlamydia trachomatis and Chlamydia muridarum Inclusion Membrane Protein IncS Are Interchangeable for Early Development but Not for Inclusion Stability in the Late Developmental Cycle

ABSTRACT Chlamydia trachomatis is an obligate intracellular bacterium, which undergoes a biphasic developmental cycle inside a vacuole termed the inclusion. Chlamydia-specific effector proteins embedded into the inclusion membrane, the Inc proteins, facilitate inclusion interaction with cellular organelles. A subset of Inc proteins engages with specific host factors at the endoplasmic reticulum (ER)-inclusion membrane contact site (MCS), which is a discrete point of contact between the inclusion membrane and the endoplasmic reticulum (ER). Here, we report that the C. trachomatis Inc protein CTL0402/IncSCt is a novel component of the ER-inclusion MCS that specifically interacts with and recruits STIM1, a previously identified host component of the ER-inclusion MCS with an unassigned interacting partner at the inclusion membrane. In comparison, the Chlamydia muridarum IncS homologue (TC0424/IncSCm) does not interact with or recruit STIM1 to the inclusion, indicating species specificity. To further investigate IncS function and overcome the recently reported early developmental defect of the incS mutant, we achieved temporal complementation by expressing IncS exclusively during the early stages of the developmental cycle. Additionally, we used allelic exchange to replace the incSCt open reading frame with incSCm in the C. trachomatis chromosome. Inclusions harboring either of these strains progressed through the developmental cycle but were STIM1 negative and displayed increased inclusion lysis 48 h postinfection. Expression of incSCt in trans complemented these phenotypes. Altogether, our results indicate that IncS is necessary and sufficient to recruit STIM1 to C. trachomatis inclusion and that IncS plays an early developmental role conserved in C. trachomatis and C. muridarum and a late role in inclusion stability specific to C. trachomatis. IMPORTANCE Obligate intracellular pathogens strictly rely on the host for replication. Specialized pathogen-encoded effector proteins play a central role in sophisticated mechanisms of host cell manipulation. In Chlamydia, a subset of these effector proteins, the inclusion membrane proteins, are embedded in the membrane of the vacuole in which the bacteria replicate. Chlamydia encodes 50 to 100 putative Inc proteins. Many are conserved among species, including the human and mouse pathogens Chlamydia trachomatis and Chlamydia muridarum, respectively. However, whether the function(s) of Inc proteins is indeed conserved among species is poorly understood. Here, we characterized the function of the Inc protein IncS conserved in C. trachomatis and C. muridarum. Our work reveals that a single effector protein can play multiple functions at various stages of the developmental cycle. However, these functions are not necessarily conserved across species, suggesting a complex evolutionary path among Chlamydia species.

IncS homologues revealed that IncS performs an early developmental role conserved in C. trachomatis and C. muridarum and a late role in inclusion stability specific to C. trachomatis.

RESULTS
IncS Ct specifically interacts with and recruits STIM1 to the Chlamydia trachomatis inclusion membrane. To identify the STIM1-interacting partner at the inclusion membrane, we performed a pilot experiment with coimmunoprecipitation (co-IP) followed by mass spectrometry. Lysates from HEK293 cells expressing FLAG-tagged STIM1 and infected, or not, with wild-type C. trachomatis for 24 h were immunoprecipitated (IP) using anti-FLAG antibodies. Analysis of the samples by mass spectrometry revealed the presence of the C. trachomatis inclusion membrane protein CTL0402 in the coimmunoprecipitate of the infected samples. We have recently renamed CTL0402 IncS Ct (29); this nomenclature will be used here. To confirm the inclusion localization of endogenous IncS Ct , HeLa cells infected with wild-type C. trachomatis for 24 h were fixed and stained with anti-IncS antibodies. Confocal microscopy revealed the presence of IncS Ct -positive patches, reminiscent of MCS, associated with the inclusion (Fig. 1A). To validate the IncS Ct -STIM1 interaction, we first generated a wild-type C. trachomatis strain expressing mCherry constitutively and IncS Ct fused to a 3ÂFLAG tag under the control of an anhydrotetracycline (aTc)-inducible promoter and confirmed the aTc-dependent inclusion localization of IncS Ct (see Fig. S1A and B in the supplemental material). We next performed co-IP experiments. HEK293 cells expressing cyan fluorescent protein (CFP)-STIM1 were infected for 24 h with C. trachomatis encoding the inducible IncS Ct -3ÂFLAG, in the presence or absence of aTc. Untransfected and/or uninfected cells served as controls. Lysates were immunoprecipitated using anti-FLAG-conjugated beads. Total lysates and IP samples were analyzed by Western blotting (WB) using anti-green fluorescent protein (GFP) antibodies, to detect CFP-STIM1, and anti- These results show that IncS Ct interacts with STIM1. To address whether IncS Ct was sufficient to recruit STIM1 to the inclusion membrane, HeLa cells expressing CFP-STIM1 were infected with the C. trachomatis IncS Ct -3ÂFLAG-inducible strain, in the absence or presence of aTc. The samples were fixed 24 h postinfection (p.i.), stained with anti-FLAG antibodies, and analyzed by confocal microscopy (Fig. 1C, upper panels). In the absence of aTc, IncS Ct was not detected at the inclusion membrane and there were small STIM1 patches on the inclusion, as previously reported (26) (Fig. 1C, upper left panels, 2aTc). In the presence of aTc, IncS Ct localized to the inclusion membrane, leading to an apparent increase in STIM1 recruitment to the inclusion and to the colocalization of IncS Ct with STIM1 (Fig. 1C, upper right panels, 1aTc). Quantification of STIM1 association with the inclusion in the absence or presence of aTc confirmed that STIM1 recruitment to the inclusion was significantly increased upon IncS Ct expression (Fig. 1D, STIM1). These results indicate that IncS Ct expression leads to STIM1 recruitment to ER-inclusion MCS.
Since we previously reported that STIM1 colocalized with VAP and CERT at ER-inclusion MCS (26), we next investigated whether IncS Ct also affected VAP and CERT inclusion association. HeLa cells expressing CFP-VAPA or CFP-CERT were infected with the C. trachomatis IncS Ct -3ÂFLAG-inducible strain. The samples were fixed 24 h p.i., stained with anti-FLAG antibodies, and analyzed by confocal microscopy (Fig. 1C, middle and lower panels). In the absence of aTc, small VAPA and CERT patches were observed on the inclusion (Fig. 1C, middle and lower left panels, respectively), which is in line with previous reports (17,24). Addition of aTc did not increase the apparent recruitment of VAPA or CERT to the inclusion (Fig. 1C, middle and lower right panels), which was confirmed by quantification of VAPA and CERT association with the inclusion (Fig. 1D, VAPA and CERT).
Altogether, these results indicate that IncS Ct is a novel component of ER-inclusion MCS that specifically recruits STIM1 to the inclusion membrane.
A STIM1 region between aa 234 and 535 is necessary and sufficient for the IncS Ctdependent recruitment of STIM1 to the inclusion membrane. We next sought to identify the minimal domain of STIM1 required for the STIM1-IncS Ct interaction. STIM1 is a multidomain protein with the N terminus residing in the ER lumen and the C terminus facing the cytosol (28). A series of mCherry-tagged truncated STIM1 constructs, previously described (26) and presented in Fig. 2A, were assessed for their ability to interact with IncS Ct in co-IP experiments. HEK293 cells expressing mCherry-STIM1 constructs, full length (1-685) or truncated in their C-terminal region (1-535 or 1-246), were infected with a wild-type C. trachomatis strain expressing GFP constitutively and IncS Ct -3ÂFLAG under the control of the aTc-inducible promoter ( Fig. S1C and D), in the presence or absence of aTc. Lysates were immunoprecipitated, using anti-FLAG-conjugated beads. Total lysates and IP samples were analyzed by Western blotting using anti-mCherry antibodies, to detect mCherry-STIM1, and anti-FLAG antibodies, to detect IncS Ct (Fig. 2B). Full-length mCherry-STIM1 (1-685) and the construct with the deletion of the Ser/Pro-rich (S/P) region and poly(Lys)-rich (K) region (1-535) successfully interacted with IncS Ct -3ÂFLAG, while the mCherry-STIM1 construct truncated for the cytosolic domains (1-246) did not. This result prompted us to test an internal deletion of the region located between amino acids (aa) 253 and 535 (D253-535). The corresponding mCh-STIM1 construct failed to immunoprecipitate IncS Ct , indicating that the region between aa 253 and 535 is necessary for STIM1-IncS Ct interaction.
FIG 2 A STIM1 region between amino acids 234 and 535 is necessary and sufficient for the IncS Ct -dependent recruitment of STIM1 to the inclusion membrane. (A) Schematic representation of the STIM1 constructs employed in this study, indicating the major domains of the STIM1 protein, their respective cellular localization (ER lumen or cytosol), and the amino acid residue in the truncated protein. EF, EF-hand; SAM, sterile alpha motif; TM, transmembrane domain; CC1, coiled-coil 1; CAD, CRAC activation domain; S/P, serine-proline-rich region; K, lysine-rich region. (B) Coimmunoprecipitation (IP) of IncS Ct -3ÂFLAG from lysates of HEK293 cells expressing the indicated mCherry-STIM1 constructs and infected for 24 h with C. trachomatis expressing IncS Ct -3ÂFLAG under the control of an aTc-inducible promoter in the absence (2aTc) or the presence (1aTc) of aTc. Numbers at left are molecular masses in kilodaltons. (C) Three-dimensional confocal micrographs of STIM1/2 DKO MEF cells expressing the indicated mCherry-STIM1 construct (red) and infected for 24 h with C. trachomatis expressing GFP constitutively (green) and IncS Ct -3ÂFLAG under the control of the aTc-inducible promoter for 24 h, in the absence (2aTc) or presence of aTc (1aTc). Scale bar, 5 mm. (D) Quantification of inclusion association (AU) of the indicated mCh-STIM1 constructs when IncS Ct expression is induced (1aTc, black dots) or not (2aTc, gray dots), as in panel C. Each dot represents one inclusion. n = 31 inclusions. Unpaired t test, ***, P , 0.0001; nd, none detected. (E) Coimmunoprecipitation (IP) of IncS Ct -3ÂFLAG from lysates of HEK293 cells expressing GFP or GFP-STIM1 (234-535) and infected for 24 h with C. trachomatis expressing IncS Ct -3ÂFLAG under the control of an aTc-inducible promoter in the absence (2aTc) or the presence (1aTc) of aTc. (F) Three-dimensional confocal micrographs of STIM1/2 DKO MEF cells expressing GFP-STIM1 (234-535) (green) infected for 24 h with a C. trachomatis strain expressing mCherry constitutively (red) and IncS Ct -3ÂFLAG under the control of an aTc-inducible promoter, in the absence (2aTc, upper panels) or the presence (1aTc, lower panels) of aTc. The merge is shown on the right. Scale bar, 5 mm. (G) Quantification of GFP-STIM1 (234-535) association with the inclusion when IncS Ct expression is induced (1aTc, black dots) or not (2aTc, gray dots) in arbitrary units (AU), as in panel F. Each dot represents one inclusion. n = 44 inclusions. Unpaired t test, ***, P , 0.0001.
As observed in HeLa cells, induction of IncS Ct expression led to an apparent and significant increase of full-length STIM1 recruitment to the inclusion membrane ( Fig. 2C and D, 1-685). A similar result was observed with STIM1 lacking the S/P and poly(K) regions ( Fig. 2C and D, 1-535). However, and in agreement with the co-IP results (Fig. 2B), IncS Ct failed to recruit STIM1 lacking the cytosolic domain, or internally truncated for the 253-535 region, to the inclusion membrane ( Fig. 2C and D, 1-246 and D253-535, respectively). Altogether, these results indicate that the region located between aa 253 and 535 of STIM1 is necessary for IncS Ct -STIM1 interaction at the inclusion membrane.
We next investigated the sufficiency of the 253-535 region in the IncS Ct -STIM1 interaction. We first generated a translational fusion between GFP and the 234-535 region of STIM1 (GFP-STIM1 234-535 ) and tested whether this construct was sufficient to interact with IncS Ct in co-IP experiments. HEK293 cells expressing GFP or GFP-STIM1 234-535 were infected, or not, with C. trachomatis encoding the inducible IncS Ct -3ÂFLAG construct, in the presence or absence of aTc. Lysates were immunoprecipitated using anti-FLAG-conjugated beads. Total lysates and IP samples were analyzed by Western blotting using anti-GFP antibodies, to detect GFP or GFP-STIM1 234-535 , and anti-FLAG antibodies, to detect IncS Ct (Fig. 2E). In contrast to GFP alone, GFP-STIM1 234-535 was successfully immunoprecipitated upon induction of IncS Ct expression (Fig. 2E, last lane, IP-FLAG blot).
We next evaluated the ability of IncS Ct to recruit STIM1 234-535 to the inclusion membrane. STIM1/2 DKO MEFs cells expressing GFP-STIM1 234-535 were infected with C. trachomatis expressing mCherry constitutively and IncS Ct -3ÂFLAG under the control of the aTc-inducible promoter, in the absence or in the presence of aTc. The samples were fixed 24 h p.i. and analyzed by confocal microscopy (Fig. 2F). In the absence of IncS Ct induction, STIM1 234-535 was not detected at the inclusion (Fig. 2F, top panels). In contrast, in the presence of aTc, STIM1 234-535 localized to inclusion (Fig. 2F, bottom panels). Quantification confirmed the IncS Ct -dependent recruitment of STIM1 234-535 to the inclusion (Fig. 2G).
Altogether, these results indicate that the region encompassing aa 234 to 535 of STIM1 is necessary and sufficient for the IncS Ct -dependent recruitment of STIM1 to the inclusion.
The Chlamydia muridarum IncS homologue does not interact with STIM1. To determine if the IncS Ct -dependent recruitment of STIM1 to the inclusion membrane is conserved in other Chlamydia species encoding IncS homologues, we investigated STIM1 association with C. muridarum inclusions. Capitalizing on the fact that human and murine STIM1 share 96.79% identity and that the 253-535 region important for IncS interaction is 100% identical between the two species, human STIM1 was used for these studies. HeLa cells expressing CFP-STIM1 were infected with C. muridarum expressing mCherry constitutively and analyzed by confocal microscopy at 24 h p.i. To our surprise, STIM1 was not recruited to C. muridarum inclusions (Fig. 3A). This result prompted us to evaluate whether the C. muridarum IncS homologue (TC0424, referred to as IncS Cm ) was capable of interacting with STIM1. We generated a C. trachomatis strain expressing GFP constitutively and IncS Cm -3ÂFLAG under the control of the aTc-inducible promoter. Lysates from HeLa cells expressing mCherry-STIM1 and infected with C. trachomatis strains expressing inducible IncS Ct -or IncS Cm -3ÂFLAG, in the absence or presence of aTc, were immunoprecipitated using anti-FLAG-conjugated beads. Total lysates and IP samples were analyzed by Western blotting using anti-mCherry antibodies, to detect mCherry-STIM1, and anti-FLAG antibodies, to detect IncS Ct or IncS Cm (Fig. 3B). As opposed to IncS Ct , IncS Cm failed to immunoprecipitate with STIM1 (Fig. 3B, compare lanes 2 and 4 of IP:FLAG blot). To further validate the lack of IncS Cm -STIM1 interaction, STIM1/2 DKO MEF cells expressing mCherry-STIM1 were infected with C. trachomatis strains expressing inducible IncS Ct -or IncS Cm -3ÂFLAG, in the absence or presence of aTc. The samples were fixed 24 h p.i. and analyzed by confocal microscopy (Fig. 3C). As shown in Fig. 1C and D and Fig. 2C and D, IncS Ct expression led to a significant increase in STIM1 association with the inclusion (Fig. 3C, upper two panels, and Fig. 3D, IncS Ct ). In contrast, IncS Cm expression did not result in a significant increase in STIM1 recruitment to the inclusion membrane (Fig. 3C, lower two panels, and Fig. 3D, IncS Cm ). Altogether, these results indicate that STIM1 does not associate with C. muridarum inclusions and that IncS Cm does not interact with STIM1.
IncS Ct is necessary for STIM1 association with the C. trachomatis inclusion. We next investigated whether IncS Ct was necessary for STIM1 association with C. trachomatis inclusion. However, we recently showed that a C. trachomatis incS Ct mutant does not progress through the early stages of the C. trachomatis developmental cycle (29). To bypass the early developmental defect of the incS Ct mutant, we capitalized on the fact that the severe growth defect of the C. trachomatis incS Ct mutant can be complemented by inducing incS Ct transcription during the first 6 to 8 h p.i. of the developmental cycle (29). Under this condition, we found that IncS Ct is present at the inclusion 8 h p.i. (Fig. 4A). To determine if stopping incS Ct transcription, shortly after the early stage of the developmental cycle, would result in loss of IncS Ct at the inclusion at later time points, HeLa cells were infected with the C. trachomatis incS Ct mutant in the presence of aTc for the entire duration of the experiment or for only the first 9 h ( Fig. 4B and C). The samples were fixed at 24 h, 48 h, and 60 h p.i., stained with anti-FLAG antibodies, and analyzed by confocal microscopy. Full complementation of the C. trachomatis incS Ct mutant resulted in IncS Ct -positive inclusions that increased in size over time (Fig. 4C, left panels). Addition of aTc in the early stages of the development cycle also resulted in inclusion growth over time, indicating successful complementation of the early developmental defect. However, while inclusions were IncS Ct positive prior to aTc removal (Fig. 4A), IncS Ct was no longer detectable at later time points, indicating that temporal complementation of IncS expression was successfully achieved (Fig. 4C, right panels). To next determine whether IncS Ct was essential for STIM1 recruitment to C. trachomatis inclusions, HeLa cells expressing mCherry-STIM1 were infected with the C. trachomatis incS Ct mutant under the conditions described above to achieve temporal complementation. The samples were fixed at 24 h p.i. and analyzed by confocal microscopy. Under complementation during the entire developmental cycle, IncS Ct expression led to STIM1 association with the inclusion (Fig. 4D, left panels). However, preventing IncS Ct inclusion localization beyond the early stages of the developmental cycle resulted in STIM1-negative inclusion (Fig. 4D, right panels). Altogether, these results indicate that STIM1 is not recruited to C. trachomatis inclusions in the absence of IncS Ct .
Characterization of the allelic replacement of incS Ct with incS Cm at the C. trachomatis incS Ct locus. To further explore the respective properties of incS Ct and incS Cm , we set out to generate a C. trachomatis strain expressing IncS Cm instead of IncS Ct . To this end, we designed a floxed cassette allelic exchange mutagenesis (FLAEM) (32)-based  (Fig. S2). The corresponding C. trachomatis strain, which also expresses mCherry constitutively from a plasmid, is referred to as the C. trachomatis SWAP strain.
To evaluate the timing of expression of IncS Cm from the endogenous incS Ct promoter, HeLa cells were infected with the C. trachomatis SWAP strain, fixed at the indicated time points, stained with anti-FLAG antibodies, and analyzed by confocal microscopy (Fig. 5B). IncS Cm -3ÂFLAG was detected in association with the mCherry-positive inclusions as early as 4 h p.i., and throughout the developmental cycle until 48 h p.i. (the latest time point  Stage and Species Specificity of Effector Functions mSphere SWAP strain. The samples were fixed 24 h p.i., stained with anti-FLAG antibodies, and analyzed by confocal microscopy (Fig. 5C). As previously observed, wild-type C. trachomatis inclusions displayed small STIM1 patches (Fig. 5C, upper left panels). In contrast, despite the presence of IncS Cm at the inclusion, C. trachomatis SWAP inclusions were negative for STIM1 (Fig. 5C, upper right panels). Quantification of the percentage of STIM1 inclusions confirmed these observations (Fig. 5D, STIM1). To determine if C. trachomatis SWAP inclusions were also defective for recruitment of other known components of the ERinclusion MCS, we evaluated the recruitment of VAPA and CERT. In contrast to the lack of STIM1 association, C. trachomatis SWAP inclusions displayed small VAPA and CERT patches, similar to wild-type inclusions (Fig. 5C, middle and upper panels, respectively; Fig. 5D, quantification), indicating the specific lack of STIM1 association with C. trachomatis SWAP inclusions. Altogether, these results indicate that IncS Ct is necessary for STIM1 recruitment to C. trachomatis inclusions and that expression of IncS Cm cannot complement this phenotype. This is in contrast with the interchangeability of IncS Ct and IncS Cm for progression through the early stages of the developmental cycle, suggesting shared and distinct functions for the C. trachomatis and C. muridarum IncS homologues.
Lack of IncS Ct results in inclusion lysis at late stages of the developmental cycle. STIM1 depletion does not affect the production of C. trachomatis infectious progeny at 48 h p.i. (26,27); however, exit via extrusion is reduced (27). To investigate the potential role of IncS Ct in C. trachomatis exit via extrusion, we first attempted to measure extrusion of the C. trachomatis SWAP strain. To this end, HeLa cells were infected with wild-type C. trachomatis or the SWAP strain and the production of infectious progeny or extrusion was analyzed at 48 h p.i. The C. trachomatis SWAP strain produced similar amounts of infectious particles as wild-type C. trachomatis (Fig. S3A). However, cells infected with the C. trachomatis SWAP strain displayed a high proportion of lysed inclusions, usually not observed at 48 h p.i., that prevented us from quantifying extrusion defects. To further characterize the timing of inclusion lysis of the C. trachomatis SWAP strain, HeLa cells infected with wild-type C. trachomatis or the C. trachomatis SWAP strain expressing mCherry were monitored live at different time points p.i. (30 h, 48 h, 60 h, 72 h, and 96 h p.i.). At each time point the percentage of inclusions that had undergone lysis was quantified (Fig. 6A). Compared to the wild-type strain, a significant increase in the percentage of lysed inclusions was observed starting at 48 h p.i. with the C. trachomatis SWAP strain.
To determine if this phenotype was due to the lack of IncS Ct , we generated a C. trachomatis SWAP strain harboring a complementation plasmid expressing mCherry constitutively and IncS Ct fused to a hemagglutinin (HA) tag under the aTc-inducible promoter. The strain is referred to as C. trachomatis SWAP1pTet-IncS Ct . The inducible expression and inclusion localization of IncS Ct -HA was confirmed by immunofluorescence (IF) (Fig. S3B); the strain produced similar numbers of infectious progeny in the presence or absence of aTc as the wild-type C. trachomatis and original C. trachomatis SWAP strains (Fig. S3A). We Stage and Species Specificity of Effector Functions mSphere next infected HeLa cells with the C. trachomatis SWAP1pTet-IncS Ct strain in the absence or presence of aTc and monitored inclusion lysis over time (Fig. 6B). In the absence of aTc, the C. trachomatis SWAP1pTetIncS Ct strain displayed a significant increase of inclusion lysis in the late stage of the developmental cycle, as observed for the original SWAP strain (Fig. 6B, gray bars). In comparison, aTc induction of the incS Ct complementation allele resulted in a significant decrease in lysis (Fig. 6B, black bars). These results indicate that the increase in inclusion lysis of the C. trachomatis SWAP strain is due to the absence of IncS Ct , rather than the expression of IncS Cm . To further confirm our conclusion, we analyzed the timing of inclusion lysis of the C. trachomatis incS Ct mutant upon temporal complementation, as described in Fig. 4. HeLa cells were infected with the C. trachomatis incS Ct conditional mutant, in the presence of aTc during the entire developmental cycle or only during the first 9 h. Inclusion lysis was monitored live and quantified at the indicated time points (Fig. 6C). As observed with the C. trachomatis SWAP strain, in the absence of IncS Ct at the inclusion membrane past the early stages in the developmental cycle, a significant increase in inclusion lysis was observed 48 h p.i. and onward (Fig. 6C, gray bars). This phenotype was complemented by expression of IncS Ct throughout the developmental cycle (Fig. 6C, black bars).
Altogether, these results indicate that the absence of IncS Ct on the C. trachomatis inclusion results in increased inclusion lysis in the late stages of the developmental cycle.

DISCUSSION
Advances in Chlamydia genetic manipulation to characterize effector function and function conservation across species. The identification and characterization of Chlamydia effector proteins have been significantly impaired by the genetic intractability of the organism. In the past years, the Fluorescence-Reported Allelic Exchange Mutagenesis (FRAEM)/FLAEM methodology has been instrumental in achieving targeted gene inactivation and cis complementation (33)(34)(35). Additionally, we recently reported a blueprint for the FRAEM-based generation of conditional mutants in essential genes, such as genes critical for progression through the early stages of the developmental cycle, like incS (29). We showed that induction of a complementation allele exclusively during the first 8 h of the developmental cycle was sufficient to complement the early developmental defect of the C. trachomatis incS Ct mutant (29). Here, we show that temporal complementation, exclusively in the early stages of the developmental cycle, rescued the early developmental defect of the C. trachomatis incS Ct mutant, while revealing and allowing us to investigate an additional role of IncS Ct during the late stages of the developmental cycle. While the success of this powerful approach highly depends on the stability of the effector protein of interest, it will most certainly benefit the field when dissecting the role of effector proteins at a specific stage of the developmental cycle. Similar to IncS Ct , which appears to play multiple roles both early and late, we anticipate that, given the reduced size of the Chlamydia genome (36), effector proteins displaying multiple functions may be common in Chlamydia.
Another significant advance in Chlamydia genetic manipulation presented here is the use of the FLAEM methodology to specifically and precisely engineer the C. trachomatis chromosome to achieve the replacement of a specific ORF with a different ORF. Replacing the incS Ct ORF with its C. muridarum homologue, in the C. trachomatis chromosome, was instrumental in investigating potential conservation of IncS function between C. trachomatis and C. muridarum. While IncS Ct and IncS Cm shared a conserved function in the early stages of the developmental cycle, IncS Cm failed to rescue the late lysis defect, pointing toward different functions later in the developmental cycle, which could be attributed to the fact that IncS Ct and IncS Cm share only 60% and 70% identity and similarity, respectively (see Fig. S5 in the supplemental material). Overall, the differences in primary amino acid sequence between IncS Ct and IncS Cm are restricted to individual amino acids evenly distributed throughout the IncS proteins. However, a few motifs consisting of 5 to 10 aa are poorly conserved and will be the focus of future studies. Overall, our results highlight that effector function may not be fully conserved among Chlamydia species, which is an important consideration as the field moves forward, especially when C. muridarum should be the default species for in vivo studies, because of the inability of C. trachomatis to establish long-lasting infection due to failure to counteract the host interferon gamma response (37).
Our FLAEM-based chromosome engineering strategy was also designed for expression of a tagged version of the protein of interest, allowing us to track expression of effector proteins from their endogenous promoter at the locus, thus eliminating any artifact of overexpression and/or improper timing of expression. The methodology described here could be easily applied for further manipulation of the Chlamydia genome, such as introducing a point mutation(s) in promoters or coding regions of interest. Furthermore, while the original FRAEM/FLAEM tools were restricted to C. trachomatis (33,34), one can envision that the recent development of a minimal replicon for allelic replacement in other Chlamydia species, such as C. muridarum (38), will be instrumental in applying the method described here to investigate effector function in C. muridarum.
IncS Ct -STIM1 interaction at ER-inclusion MCS. Through gain-and loss-of-function studies, we showed that the C. trachomatis inclusion membrane protein IncS Ct interacts with, recruits, and colocalizes with STIM1 at the inclusion, thus identifying IncS Ct as a novel component of ER-inclusion MCS. Identification of the IncS Ct -STIM1 complex is in line with the current model that ER-inclusion MCS localized protein complexes are composed of specific inclusion membrane proteins that each interact with a defined set of host factors (17,24,25).
While the respective role of the IncD-CERT-VAP and IncV-VAP complexes in lipid acquisition and ER-inclusion tethering is fairly well established (17,(23)(24)(25), the role of the IncS Ct -STIM1 complex remains elusive. The concomitant expression and inclusion localization of IncS Ct with the recruitment of STIM1 to the inclusion, starting early and lasting throughout the C. trachomatis developmental cycle (26), are in line with a shared function for IncS Ct and STIM1. However, a functional connection has yet to be established. Depletion of STIM1 did not result in, nor did it rescue, the early developmental defect or the late inclusion lysis phenotype observed with the incS Ct mutant, suggesting that STIM1 is not involved in these processes (Fig. S4). Moreover, because the increased inclusion lysis observed in the absence of IncS Ct prevented assaying for a potential extrusion defect, it remains to be determined whether the extrusion defect observed upon STIM1 depletion (27) is due to a lack of IncS Ct -STIM1 interaction at ER-inclusion MCS.
A recent study by Chamberlain et al. may help shed light on the function of IncS Ct -STIM1 in midcycle (39). More specifically, the authors reported impaired SOCE-dependent NFAT (nuclear factor of activated T cells) nuclear translocation in C. trachomatis-infected cells and proposed that STIM1 retention at ER-inclusion MCS, instead of proper relocalization to ER-PM MCS, may explain this phenotype and consequently lead to the downregulation of chemokine and cytokines downstream of NFAT signaling. Based on these observations, future studies focused on NFAT-dependent signaling pathways in cells infected with the incS Ct mutant may help shed light on the biological significance of the IncS Ct -STIM1 interaction at ER-inclusion MCS.
Inclusion membrane proteins and inclusion membrane stability. The inclusion membrane cloaks the bacteria from host surveillance pathways and allows for a safe replicative niche, until completion of the developmental cycle and the release of the infectious progeny. It was proposed that some inclusion membrane proteins play a structural role in maintaining inclusion membrane integrity (15). Moreover, a role in the specific inhibition of innate immune pathways and/or inclusion targeting by components of these pathways is supported by the fact that premature inclusion membrane rupture in the middle stages of the developmental cycle triggers host cell death pathways such as apoptosis, upon infection with C. trachomatis strains lacking CT229/CpoS, CT233/IncC, and CT383 (40)(41)(42). IncS Ct -negative inclusions were lysed earlier than wild-type inclusions; however, in comparison to strains lacking CT229/CpoS, CT233/IncC, or CT383, lysis occurred much later in the developmental cycle, suggesting that if IncS is also involved in counteracting innate immune defenses, it does so in the late development cycle.
The Inc protein CT0390 has been implicated in the end of the developmental cycle bacterial exit via lysis through the activation of STING (43). The proposed role of STIM1 in retaining STING in the ER, and thereby preventing STING activation, (44), led us to investigate if the IncS Ct -STIM1 complex could interfere with activation of the STING pathway to maintain integrity of the inclusion until completion of the developmental cycle. However, genetic and pharmacological inhibition of STING did not rescue the lysis phenotype of IncS Ct -negative inclusions (data not shown).
The results presented here are partially in agreement with a study by Dimond et al. proposing that IncS stabilizes the inclusion membrane and prevents host cell death by apoptosis during midcycle (45). However, some discrepancies remain in the timing of inclusion membrane destabilization: midcycle (45) versus late (this study). The study by Dimond et al. (45) relied on C. trachomatis/C. muridarum chimera strains generated by lateral gene transfer, in which large regions of the C. trachomatis chromosome were replaced with regions of the C. muridarum chromosome. A chimera strain (RC826), in which a segment of the C. trachomatis chromosome, including the incS Ct ORF and some of the immediate downstream sequences, was replaced with the homologous C. muridarum sequence, presented premature inclusion rupture during midcycle, followed by cell death with signs of apoptosis, resulting in a strain that was challenging to amplify. Through backcross experiments, the phenotype of RC826 was linked to the exchange of incS Ct for incS Cm and by extension the lack of IncS Ct expression. However, complementation of RC826 by expression of IncS Ct in trans was not performed; therefore, the phenotype of the chimera could also be due to IncS Cm expression in the C. trachomatis/C. muridarum chimeric background. The C. trachomatis SWAP strain presented here potentially allows us to settle between these two possibilities. Compared to RC826, the precise replacement of the incS Ct ORF with the incS Cm homologue in an intact C. trachomatis genetic background did not result in any inclusion instability until late in the developmental cycle, and this phenotype was complemented by expression of IncS Ct in trans. These results suggest that the midcycle inclusion rupture of RC826 may be due to a combinatory effect of the IncS Ct/Cm exchange and of the C. trachomatis/C. muridarum genetic background. Regardless, both strains point toward a role for IncS Ct in inclusion stability.
In future studies, it will be important to consider the proper timing of IncS-dependent inclusion rupture, as it may have significant biological implications relevant to the characterization of the exact molecular mechanism resulting in the lysis of IncS Ct -negative inclusions. Based on our results, we favor a role for IncS Ct in maintaining inclusion stability late in the developmental cycle. A non-mutually exclusive hypothesis, potentially in line with a role of IncS Ct -STIM1 in exit via extrusion, would be that an increase in inclusion lysis may be reflective of a shift in the balance of bacterial exit toward lysis rather than extrusion (9).
Conclusion. Through the use of innovative genetic approaches, we have dissected the species specificity and temporal function of the Chlamydia inclusion protein IncS. Our study revealed that in addition to the previously reported conserved role in the early stages of developmental (29), IncS performs a species-specific role in inclusion stability during the late developmental cycle. Altogether, our results highlight that Inc protein function may display multiple functions depending on the stage of the developmental cycle and may not be fully conserved among species.

MATERIALS AND METHODS
Ethics statement. All genetic manipulations and containment work were approved by the University of Virginia Biosafety Committee and are in compliance with section III-D-1-a of the NIH guidelines for research involving recombinant DNA molecules.
Cloning. Restriction enzymes and T4 DNA ligase were obtained from New England BioLabs (Ipswich, MA). PCR was performed using Herculase DNA polymerase (Stratagene). PCR primers were obtained Stage and Species Specificity of Effector Functions mSphere from Integrated DNA Technologies. Primers and cloning strategies are described in Table S1 in the supplemental material and detailed below. Construction of p2TK2 Spec -SW2 mCh(Gro L2 ) TetR-tetA P IncS Ct -33FLAG incDEFG terminator. DNA fragments corresponding to the tet repressor (TetR) and tetA promoter (tetA P ) (PCR A) and to the 3ÂFLAG and the incDEFG operon terminator (PCR C) were amplified by PCR from p2TK2-SW2 mCh(Gro) Tet-IncV-3F plasmid (24) using primers TetRSTOP5Kpn and TetAP-IncS Rv and primers 0402FLAG Fw and IncDTerm3Not, respectively. A DNA fragment corresponding to the ctl0402 (incS Ct ) ORF (PCR B) was amplified from C. trachomatis L2 genomic DNA by PCR using primers TetAP-IncS Fw and 0402 FLAG Rv. A DNA fragment corresponding to TetR-tetA P IncS Ct 3ÂFLAG incDEFG terminator (PCR D) was amplified by overlapping PCR using PCR A, B, and C as the template and primers TetRSTOP5Kpn and IncDTerm3Not. PCR D was cloned into the KpnI/NotI sites of p2TK2Spec-SW2 mCh(Gro L2 ) (46).
Construction of pSUmC IncS Ct -IncS Cm -33FLAG SWAP. Three-kilobase DNA fragments downstream and upstream of incS Ct (PCR A, right arm, and PCR B, half left arm, respectively) were amplified from C. trachomatis L2 genomic DNA via PCR using primers pSUmC3Dwn0402 5 2.1 and 3Dwn0402pSumC 3 2.1 and primers pSUmC3Up0402 5 and 3Up0402TC0424 3, respectively. A DNA fragment corresponding to IncS Cm -3ÂFLAG (PCR C) was amplified by PCR from p2TK2 Spec -SW2 GFP(nmP) TetR-tetA P IncS Cm -3ÂFLAG incDEFG terminator using primers 3 Up0402TC0424 5 and 3xFLAG-pSUmC 3. The DNA fragment corresponding to 3-kb upstream incS Ct -incS Cm -3ÂFLAG (PCR D, complete left arm) was amplified by overlapping PCR using primers pSUmC3Up0402 5 and 3ÂFLAG-pSUmC 3. Gibson assembly reaction using HiFi DNA assembly master mix (New England BioLabs) and following manufacturer instructions was used to sequentially clone PCR A (right arm) into the SbfI site and PCR D (left arm) into the SalI site of pSUmC-4.0 (32) so that each arm immediately flanked the aadA-gfp cassette.
Generation of the C. trachomatis SWAP strains. C. trachomatis L2 was transformed via CaCl 2 and isolated by plaque purifications as previously described (46). Floxed cassette allelic exchange mutagenesis via FLAEM was accomplished as described previously (34) with some modifications. Briefly, 1 mg of pSUmC IncS Ct -IncS Cm -3ÂFLAG SWAP plasmid was used to transform wild-type C. trachomatis L2, and transformants were selected with 500 mg/mL of spectinomycin and 50 ng/mL of aTc. Generation of the C. trachomatis SWAP strain was accomplished by first cultivating pSUmC IncS Ct -IncS Cm -3ÂFLAG SWAP-aadA-gfp transformants in the absence of aTc for multiple passages at a low multiplicity of infection (MOI; ;0.1), followed by plaque purification of Spec-resistant GFP-expressing chlamydiae. The allelic replacement was confirmed by PCR with the IncS 3-kb Ups Fwd and IncS 3-kb Dwn Rv primers (Table S1). The resulting strain was transformed with the plasmid pSU-Cre carrying ampicillin resistance (32). After 3 passages in the presence of 0.5 U penicillin G (PenG) and 50 ng/mL of aTc, the green fluorescence was lost. The excision of the aadA-gfp selection cassette was verified by PCR with the primers IncS 3 kb Ups Fwd and IncS 3 kb Dwn Rv (Fig. S2B) and TC0424 (3469-3489) and IncS Dwn Rv ( Fig. S2B and C; Table S1) and DNA sequencing (Fig. S2D). The strain was cured of the pSU-Cre plasmid after performing 4 passages in the absence of aTc, denoted by the loss of mCherry expression (colorless inclusions). The colorless bacteria were plaque purified and amplified. The resulting strain was transformed with the p2TK2 Spec -SW2 mCh(Gro L2 ) plasmid (46) or the complementation plasmid p2TK2 Spec -SW2 mCh(Gro) Tet-IncS Ct -HA. The resulting strains were named C. trachomatis SWAP and C. trachomatis SWAP1pTet-IncS Ct , respectively.
DNA transfection. DNA transfection was performed using X-tremeGENE 9 DNA transfection reagent (Roche), according to manufacturer's recommendations.
Coimmunoprecipitation. For mass spectrometry samples, 6 Â 10 6 HEK293 cells expressing STIM1-3ÂFLAG seeded in 10-cm 2 dishes were infected for 24 h with C. trachomatis (MOI of 2) or not (control). At 23 h p.i. cells were incubated with 300 mM lactacystin (Sigma) for 1 h, washed once with phosphatebuffered saline (PBS), and lysed in 2 mL of lysis buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, protease inhibitor cocktail EDTA-free [Roche]) for 20 min at 4°C, rotating. All subsequent steps were performed at 4°C. The lysates were centrifuged at 13,000 rpm for 10 min, and the supernatants were incubated with 20 mL of anti-FLAG M2 affinity beads (Sigma) for 2 h, rotating. The protein-bound beads were washed three times with wash buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 2 mM EDTA, 1% Triton X-100), and proteins were eluted in 40 mL of elution buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 2 mM EDTA, 100 mg/mL 3ÂFLAG peptide [Sigma]) and separated via 10% SDS-PAGE. Samples were sliced for gel digestion and analyzed at the W.M. Keck Biomedical Mass Spectrometry Laboratory of the University of Virginia. We note that the experiment described above was designed as a pilot experiment that did not include biological or technical replicates. C. trachomatis proteins were not detected in the uninfected sample. Fewer than 30 C. trachomatis proteins were detected in the infected samples. More than one peptide were detected for only four proteins, namely, DnaK (21 peptides), IncS (16 peptides), FliN (10 peptides), and RpoC (5 peptides).
For coimmunoprecipitation experiments, the protocol was slightly modified. HEK293 or HeLa cells (5 Â 10 5 ) were seeded in 6-well tissue culture plates and transfected (1 mg DNA/well) for 24 h prior to C. trachomatis infection (MOI of 5). Infected cells were incubated in the presence of 20 ng/mL of aTc for 4 h prior to being lysed in 250 mL of lysis buffer. A 20-mL aliquot of the precleared lysates was collected (lysate), and samples were eluted in 30 mL (IP).
Immunoblotting. Protein samples were separated by SDS-PAGE and transferred to nitrocellulose membranes. The membranes were stained with a Ponceau S solution, to ensure even transfer, and rinsed with distilled water (dH 2 O) before blocking in blocking buffer (5% nonfat milk in 1Â PBS, 0.05% Tween) for 1 h at room temperature. Primary and horseradish peroxidase (HRP)-conjugated secondary antibodies were diluted in blocking buffer and incubated with the membranes overnight (ON) at 4°C and for 1 h at room temperature, respectively. HRP-conjugated secondary antibodies were detected with the Amersham ECL Western blotting detection reagent according to the manufacturer's recommendations and a Bio-Rad ChemiDoc imaging system.
Immunofluorescence, confocal microscopy, and quantification of protein associated with the inclusion. For immunofluorescence experiments, cells seeded on coverslips were transfected for 24 h prior to C. trachomatis infection (MOI of 1) and incubated in the presence of 2 ng/mL of aTc starting 4 h postinfection. C. trachomatis incS Ct conditional mutant infection was performed in the presence of 0.5 ng/mL aTc at all times otherwise indicated. At the indicated times, samples were fixed with 4% paraformaldehyde in PBS for 30 min. All steps were performed at room temperature. Coverslips were sequentially incubated with primary and secondary antibodies diluted in 0.1% Triton X-100 in 1Â PBS for 1 h. For CT147 staining cells were fixed with cold methanol (MeOH) for 10 min and then incubated ON at 4°C with anti-CT147. Coverslips were washed with 1Â PBS and mounted with DABCO antifade-containing mounting medium. Imaging was performed using the Leica DMi8 microscope equipped with the Andor iXon Ultra 888BV electron-multiplying charge-coupled device (EMCCD) camera and the confocal scanner unit CSU-W1 and driven by the IQ software. Images were processed using the Imaris software (Bitplane, Belfast, United Kingdom) as follows. For quantification, 3-dimensional reconstructions of the raw signal corresponding to each marker were generated using the Imaris imaging software as previously described (24). For each marker, the volume corresponding to the sum of the pixels, above the threshold corresponding to the signal in the cytoplasm, was determined. The host protein volumes were normalized to the corresponding inclusion volume to determine the inclusion association of the respective markers in arbitrary units. At least 30 inclusions were quantified per condition. Each experiment was performed in triplicate. One representative experiment is shown. The graphs were generated using GraphPad Prism. Statistical analysis was performed using the appropriate test as indicated in the figure legends.
Infectious progeny production. To determine the infectious progeny production, infected HeLa cells in 96-well plates were lysed 48 h postinfection in 100 mL sterile water, and 5-fold dilutions of the lysates were used to infect fresh HeLa cell monolayers. Infected cells were stained with Hoechst stain. The number of infected cells was determined by automated imaging using an ImageXpress automated system. Quantification of inclusion-forming units (IFUs) per milliliter was determined using the MetaXpress software. The fold growth was determined as the ratio between the IFUs per milliliter at 48 h and those at 0 h p.i.
Lysis assay. Confluent monolayers of HeLa cells or wild-type or STIM1/2 DKO MEFs plated in 96 wells were infected with the indicated C. trachomatis strains at an MOI of 0.005. To limit overgrowth of the MEF cells over the 5-day period of the assay, lysis assay in MEF cells was carried out in the presence of 1 mg/mL cycloheximide. Cycloheximide was not used to conduct the lysis assay in HeLa cells. At the indicated time points, the number of intact and lysed inclusions was manually quantified live with a fluorescence microscope, as described in the work of Bishop and Derré (43). The percentage of lysed inclusions was determined as the number of lysed inclusions/total of inclusions. One hundred to 150 inclusions were quantified per well, with 4 biological replicates per experiment.
Statistics. Each experiment was performed in triplicate. The sample size is indicated in the figure legends. The average 6 standard error of the mean (SEM) from one representative experiment is shown. The graphs were generated using GraphPad Prism. The appropriate statistical tests were used and are indicated in the figure legends.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.