Expression and Localization of Predicted Inclusion Membrane Proteins in Chlamydia trachomatis

Abstract

Chlamydia trachomatis is an obligate intracellular pathogen that replicates in a membrane-bound vacuole termed the inclusion. Early in the infection cycle, the pathogen extensively modifies the inclusion membrane through incorporation of numerous type III secreted effector proteins, called inclusion membrane proteins (Incs). These proteins are characterized by a bilobed hydrophobic domain of 40 amino acids. The presence of this domain has been used to predict up to 59 putative Incs for C. trachomatis; however, localization to the inclusion membrane with specific antibodies has been demonstrated for only about half of them. Here, we employed recently developed genetic tools to verify the localization of predicted Incs that had not been previously localized to the inclusion membrane. Expression of epitope-tagged putative Incs identified 10 that were previously unverified as inclusion membrane localized and thus authentic Incs. One novel Inc and 3 previously described Incs were localized to inclusion membrane microdomains, as evidenced by colocalization with phosphorylated Src (p-Src). Several predicted Incs did not localize to the inclusion membrane but instead remained associated with the bacteria. Using Yersinia as a surrogate host, we demonstrated that many of these are not secreted via type III secretion, further suggesting they may not be true Incs. Collectively, our results highlight the utility of genetic tools for demonstrating secretion from chlamydia. Further mechanistic studies aimed at elucidating effector function will advance our understanding of how the pathogen maintains its unique intracellular niche and mediates interactions with the host.

INTRODUCTION

Chlamydiae are Gram-negative obligate intracellular pathogens that are of human and veterinary importance. Chlamydia trachomatis is a major cause of bacterial sexually transmitted infections and is the causative agent of trachoma, a leading cause of infectious blindness. The species includes at least 15 serovars that differ in disease state and severity, including trachoma (serovars A to C), a sexually transmitted infection (serovars D to K), and a systemic disease referred to as lymphogranuloma venereum (LGV) (serovars L1 to L3) (1). C. pneumoniae is a respiratory pathogen that is a common cause of pneumonia and bronchitis and is associated with chronic diseases, such as atherosclerosis (2). C. psittaci is a zoonotic pathogen associated with a rare but serious pneumonia (3). Other Chlamydia species, including C. muridarum, C. caviae, and C. felis, are associated with other animals, i.e., mice, guinea pigs, and cats, respectively (4–6).

All chlamydiae exhibit a biphasic developmental cycle in which the bacteria alternate between an infectious, metabolically dormant elementary body (EB) and the replicative form, referred to as the reticulate body (RB) (7). Upon infection, the EB is internalized by endocytosis into a membrane-bound compartment that is rapidly modified by the pathogen to establish a replicative niche termed the inclusion (8). The inclusion avoids fusion with endocytic compartments/lysosomes but traffics along microtubules to the peri-Golgi region (9). The inclusion undergoes extensive interactions with the host to acquire nutrients, including lipids, amino acids, and iron, all while avoiding activation of the innate immune response (9, 10). Additionally, the inclusion engages many host organelles, including the endoplasmic reticulum, Golgi apparatus, mitochondria, and cytoskeleton (9, 11, 12). At the conclusion of the developmental cycle, EBs are released from the host cell by either cell lysis or extrusion (13) for subsequent rounds of infection.

At each stage of the developmental cycle, the bacterium translocates bacterial virulence proteins, termed effector proteins, that manipulate host pathways and facilitate formation of its unique replicative niche. Among these are a subset of type III secreted proteins that are inserted into the inclusion membrane. These inclusion membrane proteins (Incs) are incorporated into the membrane so that they are exposed to the host cytosol and are poised to mediate crucial host-pathogen interactions (14). While Incs lack significant sequence similarity to other proteins or each other, one characteristic feature is a bilobed hydrophobic domain of approximately 40 amino acids (15). To date, the presence of this domain has been used to predict 59 putative Incs for C. trachomatis (16–19). Of these, localization to the inclusion membrane has been demonstrated for only about half of the proteins (16, 17, 20–22).

Until recently, a lack of genetic tools significantly hampered the identification of chlamydial virulence factors and instead necessitated reliance on heterologous expression systems for identification of chlamydia type III secretion system (T3SS) substrates. To date, several secreted proteins have been identified using Yersinia pseudotuberculois, Shigella flexneri, or Salmonella enterica serovar Typhimurium as a surrogate host (23–27). Recent advances, including plasmid transformation (28) and generation of a shuttle vector system with multiple reporters, have enabled experiments that confirm secretion of C. trachomatis proteins into the host cell cytosol and localization to the inclusion membrane during infection (29). In the current study, we employed these genetic tools to assess the localization of the predicted Inc proteins. Using this approach, we demonstrated that 10 previously unverified Incs were localized to the inclusion membrane. Many predicted Incs remained associated with the bacteria and did not localize to the inclusion membrane. Importantly, we determined the localization of several additional Incs that had been predicted but not yet characterized.

MATERIALS AND METHODS

Bacterial and cell culture.

C. trachomatis serovar L2 (LGV/434/Bu) was propagated in HeLa 229 cells (American Type Culture Collection), and infectious EBs were purified using a Renografin density gradient, as previously described (30). Wild-type and Y. pseudotuberculosis ΔyscS were maintained on heart infusion broth (HIB). HeLa and Vero CCL-81 cells (ATCC) were grown in RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) at 37°C and 5% CO₂.

Comparison of Inc serovars.

To evaluate the conservation of Incs among C. trachomatis serovars, the DNA sequences from the following strains were used: D/UW/-3/CX (NC_000117.1), L2/434/Bu (NC_010287.1), and A/HAR-13 (NC_007429.1). The corresponding DNA sequences were aligned using Clustal Omega to identify N-terminal truncations (NT), C-terminal truncations (CT), alternative start sites (AS), or frameshifts (FS) compared to D/UW/-3/CX.

Plasmid construction.

To assess the localization of Incs, Chlamydia genes were PCR amplified from L2/434/Bu genomic DNA using the primers listed in Table S1 in the supplemental material and were cloned into the NotI/SalI or the NotI/KpnI sites of pBomb-tet-mcherry (29). Each open reading frame (ORF) was expressed as a C-terminal fusion to a Flag tag by addition to the 3′ primer. To address secretion of predicted Incs by Yersinia, the first 40 amino acids of each ORF were fused to neomycin phosphotransferase (NPT) and cloned into the pFlag-CTC vector (Sigma, St. Louis, MO) using GeneArt seamless cloning (Invitrogen). The integrity of all constructs was verified by DNA sequencing.

Chlamydia transformation.

C. trachomatis serovar L2 was transformed with each expression construct as previously described (29). Briefly, plasmid DNA, prepared from methylation-deficient Escherichia coli K-12 ER2925 (New England BioLabs, Ipswich, MA), was transformed into C. trachomatis L2 (LGV/434/Bu) density gradient-purified EBs using CaCl₂ buffer (10 mM Tris [pH 7.5], 50 mM CaCl₂). Following 4 passages under selection with 0.1 U of penicillin G/ml, the expression of the fusion protein was induced using 10 ng/ml anhydrotetracycline hydrochloride (aTc), and expression of the Flag-tagged fusion protein was verified using immunofluorescence microscopy. Transformed bacteria, expressing the fusion protein, were plaque cloned in Vero cells, and individual plaques were picked and expanded.

Immunofluorescence.

HeLa cells were seeded at 10⁵/ml onto glass coverslips. After 24 h, the cells were infected at a multiplicity of infection (MOI) of 1, and expression of the fusion protein was induced with 10 ng/ml of tetracycline. Eighteen hours postinfection, the cells were fixed with 100% methanol and blocked using 1% bovine serum albumin (BSA), and nuclei were visualized using DAPI (4′,6-diamidino-2-phenylindole) (Invitrogen). EBs were stained with anti-EB antiserum, and the Flag-tagged fusion protein was detected using an anti-Flag antibody (Sigma). Primary antibodies were visualized by staining with DyLight 594-goat anti-mouse (Jackson Laboratories, Las Vegas, NV) and DyLight 488-goat anti-rabbit (Jackson Laboratories) secondary antibodies. For CT134, cells were fixed at 24 h postinfection with 4% paraformaldehyde, and the Flag-tagged fusion protein was detected using an anti-hemagglutinin (HA) antibody. To determine if newly validated Incs localize to microdomains, infected cells were stained with phosphorylated-Src (p-Src) antibodies (Millipore, Darmstadt, Germany). Images were captured on a Nikon Eclipse 80i fluorescence microscope, and the images were analyzed using Nikon Elements software. The data are representative of at least 3 independent experiments with at least 100 infected cells per experiment.

Yersinia T3SS assay.

To assess secretion of predicted Incs, the first 40 amino acids of each Chlamydia ORF were fused to NPT and cloned into pFlag-CTC. The resulting constructs were transformed into wild-type (pIB102) and ΔyscS Y. pseudotuberculosis YPIII as previously described (24). Transformed bacteria were cultured in HIB supplemented with 5 mM calcium or in medium lacking calcium (20 mM MgCl₂ and 5 mM EGTA) to inhibit or promote secretion, respectively. After incubation for 2 h at 26°C, protein expression was induced with 0.1 mM IPTG (isopropyl-β-d-thiogalactopyranoside), and the temperature was shifted to 37°C to induce secretion. After 4 h, the cultures were separated into supernatant and pellet fractions, and the supernatants were trichloroacetic acid (TCA) precipitated. Each fraction was normalized to equivalent volumes based on the optical density (OD) of the original culture. The fractions were analyzed by immunoblotting, and the blots were probed with anti-Flag antibody.

RESULTS

Genetic conservation of inclusion membrane proteins among C. trachomatis serovars.

Inclusion membrane proteins possess a hydrophobic domain of 40 amino acids (15). The presence of the domain has been used to predict 59 potential Incs for C. trachomatis (15–21). Whereas many Incs are poorly conserved between species (18, 19), most are conserved among different C. trachomatis serovars (Table 1). To evaluate the conservation of Incs between C. trachomatis serovars, the corresponding Incs from strains L2/434/Bu and A/HAR-13 were compared to those from D/UW/-3/CX. At least seven (CT036, CT115, CT179, CT192, CT226, CT300, and CT383, and/or their orthologs) display N-terminal truncations or are annotated as having alternative start sites (Fig. 1; see Fig. S1 in the supplemental material). Additionally, CT135, CT227, CT244, and CT358 and/or their orthologs harbor frameshift mutations or alternative stop sites within the last 30 bp. Because they still express the bilobed hydrophobic domain, these proteins may still be Incs, although possible effector functions may be disrupted.

TABLE 1.

Conservation and localization of predicted Incs

Inc in C. trachomatis straina:			Previously reported localization	Reference(s)	Localization in current studyb	Proposed localization
D/UW-3/CX	L2/434/Bu	A/HAR-13
CT005	CTL0260	CTA0006	Inclusion membrane	17	Inclusion membrane	Inclusion membrane
CT006	CTL0261	CTA0007	Undefined		Inclusion membrane	Inclusion membrane
CT036	CTL0291 (NT)	CTA0038	Undefined		Bacteria	Undefined
CT058	CTL0314	CTA0062 (FS)	Intrainclusion	16	Bacteria	Intrainclusion
CT079	CTL0335	CTA0084	Undefined		Bacteria	Undefined
CT089	CTL0344	CTA0094	Inclusion membrane	16, 55	Not tested	Inclusion membrane
CT101	CTL0356	CTA0107	Inclusion membrane	17, 20	Not tested	Inclusion membrane
CT115 (incD)	CTL0370 (AS, FS)	CTA0122 (AS)	Inclusion membrane	16, 17, 21	Not tested	Inclusion membrane
CT116 (incE)	CTL0371	CTA0123	Inclusion membrane	16, 17, 21	Inclusion membrane	Inclusion membrane
CT117 (incF)	CTL0372	CTA0124	Inclusion membrane	16, 17, 21	Inclusion membrane	Inclusion membrane
CT118 (incG)	CTL0373	CTA0125	Inclusion membrane	16, 17, 21	Not tested	Inclusion membrane
CT119 (incA)	CTL0374	CTA0126	Inclusion membrane	16, 17, 21, 22	Not tested	Inclusion membrane
CT134	CTL0389	CTA0141	Undefined		Inclusion membrane	Inclusion membrane
CT135	CTL0390	CTA0142 (CT)	Undefined		Inclusion membrane	Inclusion membrane
CT147	CTL0402	CTA0156	Inclusion membrane	16, 56	Inclusion membrane	Inclusion membrane
CT164	CTL0419A	CTA0174	Undefined		Not expressed	Undefined
CT179	CTL0431 (AS)	CTA0195	Undefined		Inclusion membrane	Inclusion membrane
CT192	CTL0444 (NT)	CTA0210 (NT)	Intrainclusion	16	Inclusion membrane	Inclusion membrane
CT195	CTL0447	CTA0213	Intrainclusion	16	Bacteria	Intrainclusion
CT196	CTL0448	CTA0214	Undefined		Not expressed	Undefined
CT214	CTL0466	CTA0234	Undefined		Bacteria	Undefined
CT222	CTL0475 (FS)	CTA0244	Inclusion membrane	16, 17, 20	Inclusion membrane	Inclusion membrane
CT223	CTL0476	CTA0245	Inclusion membrane	15–17	Inclusion membrane	Inclusion membrane
CT224	CTL0477	CTA0246	Undefined	17	Inclusion membrane	Inclusion membrane
CT225	CTL0477A	CTA0247	Inclusion membrane	16, 17	Not expressed	Inclusion membrane
CT226	CTL0478 (NT)	CTA0248 (NT)	Inclusion membrane	16, 17, 57	Inclusion membrane	Inclusion membrane
CT227	CTL0479	CTA0249 (FS, CT)	Undefined	17	Inclusion membrane	Inclusion membrane
CT228	CTL0480	CTA0250	Inclusion membrane	16, 17, 33	Not tested	Inclusion membrane
CT229	CTL0481	CTA0251	Inclusion membrane	15–17, 42	Inclusion membrane	Inclusion membrane
CT232 (incB)	CTL0484	CTA0254	Inclusion membrane	16	Inclusion membrane	Inclusion membrane
CT233 (incC)	CTL0485	CTA0255	Inclusion membrane	15, 16	Inclusion membrane	Inclusion membrane
CT244	CTL0496	CTA0266 (FS, CT)	Undefined		Bacteria	Bacteria
CT249	CTL0500A	CTA0271	Inclusion membrane	16, 17, 58	Not expressed	Inclusion membrane
CT288	CTL0540	CTA0310	Inclusion membrane	15, 16	Inclusion membrane	Inclusion membrane
CT300	CTL0552 (NT, FS)	CTA0322 (FS)	Undefined		Not expressed	Undefined
CT324	CTL0576	CTA0348	Undefined		Not expressed	Undefined
CT345	CTL0599	CTA0374	Undefined		Inclusion membrane	Inclusion membrane
CT357	CTL0611A	CTA0387	Undefined		Not expressed	Undefined
CT358	CTL0612 (FS, CT)	CTA0389	Inclusion membrane	16	Not expressed	Inclusion membrane
CT365	CTL0619	CTA0397	Undefined		Bacteria	Undefined
CT383	CTL0639	CTA0418 (AS)	Intrainclusion	16	Inclusion membrane	Inclusion membrane
CT440	CTL0699	CTA0480	Inclusion membrane	16	Not expressed	Inclusion membrane
CT442	CTL0701	CTA0482	Inclusion membrane	15, 16, 59	Inclusion membrane	Inclusion membrane
CT449	CTL0709	CTA0491	Undefined		Inclusion membrane	Inclusion membrane
CT483	CTL0744	CTA0530	Inclusion membrane	17	Bacteria	Inclusion membrane
CT484	CTL0745	CTA0531	Intrainclusion	15, 16	Bacteria	Intrainclusion
CT529	CTL0791	CTA0578	Inclusion membrane	16, 47	Bacteria	Inclusion membrane
CT565	CTL0828	CTA0615	Intrainclusion	16, 17	Bacteria	Bacteria
CT616	CTL0880	CTA0669	Undefined		Bacteria	Bacteria
CT618	CTL0882	CTA0671	Inclusion membrane	16, 60	Not expressed	Inclusion membrane
CT642	CTL0010	CTA0697	Undefined		Bacteria	Undefined
CT728	CTL0097	CTA0790	Undefined		Bacteria	Undefined
CT788	CTL0156	CTA0858	Undefined		Bacteria	Undefined
CT789	CTL0157	CTA0859	Undefined		Bacteria	Undefined
CT813 (lnaC)	CTL0184	CTA0885	Inclusion membrane	16, 17	Not tested	Inclusion membrane
CT819	CTL0191	CTA0893	Undefined		Bacteria	Undefined
CT846	CTL0218	CTA0922	Undefined		Bacteria	Undefined
CT850	CTL0223	CTA0927	Inclusion membrane	16, 17, 20	Not tested	Inclusion membrane

^aC. trachomatis serovars were compared to identify N-terminal (NT) or C-terminal (CT) truncations, frameshift mutations (FS), and alternative start sites (AS).

^bLocalization is in boldface for newly validated Incs.

FIG 1.

Variations in Inc proteins from C. trachomatis serovars. Several Incs from C. trachomatis strain D/UW/-3/CX were compared to those of L2/434/Bu and A/HAR-1. Shown are N-terminal or C-terminal truncations (△), frameshift mutations due to insertions (+) or deletions (−), and alternative start sites (†).

Localization of predicted Incs to the inclusion membrane.

To determine whether the predicted Incs that had not yet been localized to the inclusion membrane were actually Incs, we expressed each predicted Inc as a C-terminal Flag fusion. C. trachomatis L2 was transformed with each construct, and localization was assessed at 18 h postinfection. Of the 50 confirmed or predicted Incs tested in the current study, we demonstrated localization to the inclusion membrane for several predicted Incs (Fig. 2; see Fig. S2 in the supplemental material), including many previously defined Incs, highlighting the utility of the system (see Fig. S2 in the supplemental material). Importantly, we demonstrated localization to the inclusion membrane for 10 previously unverified Incs (CT006, CT134, CT135, CT179, CT192, CT224, CT227, CT345, CT383, and CT449). Whereas the majority of Incs exhibited uniform circumferential staining around the inclusion, some were restricted to specific areas around the inclusion (see below).

FIG 2.

Localization of predicted Incs to the inclusion membrane. Predicted inclusion membrane proteins were expressed as C-terminal fusions to a Flag tag in C. trachomatis L2. HeLa cells were infected at an MOI of 1, and after 18 h, the cells were fixed with methanol and probed with anti-L2 and anti-Flag M2 primary antibodies and Alexa Fluor 488 (green) and 594 (red) secondary antibodies, respectively. Nuclei were visualized using DAPI (blue). The scale bar represents 10 μm. Ten previously undefined Incs localized to the inclusion membrane. The data are representative of the results of 3 independent experiments, with at least 100 infected cells observed per experiment.

Surprisingly, we found that numerous predicted Incs (CT195, CT244, CT365, CT483, CT484, CT529, CT565, CT616, CT788, CT789, and CT846) did not localize to the inclusion membrane and instead remained associated with the bacteria (Fig. 3). A few predicted Incs (CT036, CT058, CT079, CT214, CT642, CT728, and CT819) had weak expression levels but appeared to be associated with the bacteria. Overexpression of several of these proteins resulted in the presence of enlarged or aberrant RBs (CT483, CT484, CT616, CT788, and CT789). Similar results were obtained with various infection and induction times, indicating that these phenotypes and lack of localization to the inclusion membrane were not artifacts of being expressed out of the normal infection phase (data not shown). Of these, all but CT058, CT195, CT484, and CT565 were previously undefined, and these 4 had been previously reported as intrainclusion. Collectively, these results suggest these proteins may not be Incs and instead may represent bacterial membrane proteins.

FIG 3.

Several predicted Incs do not localize to the inclusion membrane. HeLa cells were infected at an MOI of 1 with C. trachomatis L2 expressing Flag-tagged predicted inclusion membrane proteins. Fusion protein expression was induced with 10 ng/ml aTc, and 18 h postinfection, the cells were fixed with methanol and probed with anti-L2 and anti-Flag M2 primary antibodies and Alexa Fluor 488 (green) and 594 (red) secondary antibodies, respectively. DNA was counterstained with DAPI (blue). The scale bar represents 10 μm. Several predicted Incs colocalized with the bacteria and did not localize to the inclusion membrane. The data are representative of the results of 3 independent experiments, with at least 100 infected cells observed per experiment.

Bacterium-associated proteins are not secreted by Y. pseudotuberculosis.

Y. pseudotuberculosis has been successfully used to demonstrate translocation of chlamydial Incs (24, 29, 31). To substantiate our findings that these predicted Incs, which do not localize to the inclusion membrane, are not secretion substrates, we employed a heterologous secretion assay. Fifteen of the bacterium-associated Incs were included in the assay. The first 40 amino acids of each protein were fused to NPT and expressed in wild-type Y. pseudotuberculosis. Bacteria were cultured in medium containing or lacking calcium to inhibit or promote secretion, respectively. Expression of the fusion protein was induced with 0.1 mM IPTG, and bacterial secretion was promoted by shifting the temperature from 26°C to 37°C. We included IncC, which has previously been shown to be secreted by Y. pseudotuberculosis in a T3SS-dependent manner (24, 29), as a positive control. As shown in Fig. 4, CT036, CT058, CT079, CT244, CT565, CT616, CT728, and CT819 were not detected in the supernatant, indicating they are not secreted. Two substrates, CT195 and CT365, were located in the supernatant fraction only under conditions supporting bacterial secretion. These data suggest both CT195 and CT365 may be type III secreted; however, we did not observe localization to the inclusion membrane for either (Fig. 3). Previous data indicated that CT195 is localized within the inclusion, and thus, it may be secreted into the inclusion lumen but not localized to the membrane and therefore not an Inc. CT483, CT484, CT788, CT789, and CT846 were not expressed in the Y. pseudotuberculosis system and therefore could not be validated for secretion. Of these, localization was previously undefined for CT788, CT789, and CT846, whereas CT484 had been demonstrated to be intrainclusion (16). Collectively, our results indicate several of the putative Incs that are not detected in the inclusion membrane are also not secreted via a T3SS and therefore are unlikely to be inclusion membrane proteins.

FIG 4.

Bacterium-associated predicted Incs are not secreted by Y. pseudotuberculosis. Transformed Y. pseudotuberculosis YPIII pIB102 (wild type [WT]) was grown in heart infusion broth with (+) or without (−) Ca²⁺ to inhibit or promote secretion, respectively. Cultures were grown at 26°C for 2 h, expression of the fusion protein was induced with 0.1 mM IPTG, and the cultures were shifted to 37°C for an additional 4 h. The cultures were separated into supernatant (S) and pellet (P) fractions and normalized to equivalent volumes based on the original OD. Samples were resolved using SDS-PAGE, and the immunoblots were probed with anti-Flag M2 antibodies. The data are representative of the results of at least 2 independent experiments.

Inclusion membrane proteins localize to microdomains.

The chlamydial inclusion membrane possesses localized structures, termed microdomains, that are enriched in host active Src family kinases and cholesterol, along with 4 inclusion membrane proteins, CT101, CT850, CT222, and IncB (20). Recent evidence suggests these microdomains serve as a platform to promote interactions between the inclusion and the cytoskeleton. These interactions include dynein, promoting trafficking along microtubules to position the inclusion at the microtubule-organizing center (MTOC) (20, 32), and the myosin motor complex, which is involved in egress of the inclusion (33). To determine if any other Incs localize to microdomains, we tested 23 Incs (Fig. 2; see Fig. S2 in the supplemental material) for colocalization with active Src. As shown in Fig. 5, four Incs (CT223, CT224, IncC, and CT288) were enriched in microdomains on the inclusion that correlated with active Src family kinases.

FIG 5.

Inclusion membrane proteins localize to microdomains. HeLa cells were infected at an MOI of 1 with C. trachomatis L2 expressing Flag-tagged Incs. After 18 h, the cells were fixed with methanol and probed with anti-phospho-Src and anti-Flag M2 primary antibodies and Alexa Fluor 594 and 488 secondary antibodies, respectively. Nuclei were visualized using DAPI. The scale bar represents 10 μm. Four Incs colocalized with active Src and localized to microdomains. The data are representative of the results of 2 independent experiments, with at least 100 infected cells observed per experiment.

DISCUSSION

Until recently, the genetic intractability of chlamydiae significantly hampered our understanding of this important pathogen and generated reliance on heterologous systems for identifying and characterizing chlamydial virulence factors. Recent advances, including the ability to transform Chlamydia (28) and to express epitope-tagged proteins (29), have enabled experiments that confirm localization and secretion of type II and type III secretion substrates from C. trachomatis. Chlamydial inclusion membrane proteins are characterized by a bilobed hydrophobic domain of 40 or more amino acids (15, 21, 34). While this signature has been a good predictor of inclusion membrane localization, it is not a confirmation of protein localization. The localization of several Incs has previously been confirmed using specific antibodies generated against predicted chlamydial Incs (15–18, 20, 21). With the prediction of new Incs and the availability of genentic tools, further studies were possible to define the localization of additional Inc proteins. Here, we used recently adapted genetic tools to demonstrate inclusion membrane localization for an additional 10 predicted Incs. We also found that several predicted Incs did not localize to the inclusion membrane. Furthermore, several of these putative Incs that did not localize to the inclusion membrane were not secreted in a heterologous type III secretion assay. While the negative results here and previously (16–18) do not eliminate the possibility that any of these are bona fide Incs, more extensive analyses will be required to define their true localization. Collectively, our results highlight the value of validating secretion in Chlamydia.

Many obligate intracellular pathogens have undergone significant genomic reduction in favor of their obligate lifestyle, requiring them to scavenge many essential nutrients from the host. Despite having a small genome of 1.04 Mbp (35), C. trachomatis is predicted to translocate numerous type III secretion substrates (25, 27, 36), representing over 13% of the proteome. Of these, up to 59 are predicted to localize to the inclusion membrane based on the presence of a bilobed hydrophobic domain. These proteins are incorporated into the inclusion membrane so that they are exposed to the host cytosol and thus are poised to mediate crucial host-pathogen interactions. While Incs are not well conserved between species, they are well conserved between C. trachomatis serovars, indicating they may play an integral role in the infection process. However, Incs share little sequence similarity with other proteins or with each other, complicating effector characterization. In spite of this, the functions of several Incs have been elucidated. IncA encodes SNARE-like motifs and has been implicated in mediating homotypic fusions of inclusions (14, 37). IncD recruits CERT, a ceramide endoplasmic reticulum transferase (38, 39). IncG is presumed to promote host cell viability by recruiting the adaptor protein 14-3-3β to the inclusion, sequestering proapoptotic BAD (40, 41). CT228 mediates chlamydial extrusion through an interaction with myosin phosphatase target subunit 1 (MYPT1) (33). CT229 interacts with Rab4 (42), CT813 (LnaC) binds host Arfs (12), and CT850 interacts with host dynein (32). However, the functions of the majority of Incs remain largely unknown. A recent comprehensive screen of potential binding partners to specific Incs may help shed light on additional interactions (43). Surprisingly, though, of the C. trachomatis Incs with confirmed binding partners, none of the interactions are conserved among all species of chlamydia. It is possible that not all Incs have specific cellular interacting partners, as structural roles for Incs have also been proposed (44).

Here, we demonstrate that numerous predicted Incs localize to the inclusion, 10 of which were previously unverified. While CT224 and CT227 belong to a gene cluster, encoding CT222 to CT229, which is widely believed to encode several inclusion membrane proteins, localization for these 2 proteins had not been shown prior to the current study. Additionally, CT192 and CT383 were previously reported to localize within the inclusion (16); however, in our study, we observed localization to the inclusion membrane, indicating they are Incs.

Although we observed localization to the inclusion for the L2 ortholog of CT192, genomic comparison of CT192 orthologs revealed a large N-terminal truncation for both 434/Bu and A/Har-13 relative to serovar D. Surprisingly, all orthologs of CT192 were annotated with alternative (TTG or GTG) start sites. The actual translational start site has not been determined for any Inc, and thus, some caution must be exercised in the selection of coding regions for expression studies. For example, despite identical nucleotide sequences over the relevant region, the A/Har-13 ortholog of CT383 is predicted to start at an ATG start codon 24 nucleotides upstream from an in-frame ATG codon predicted as the start codon of the D/UW3/Cx and LGV/434/Bu genes. Given the importance of the N-terminal type III secretion signal sequence (29), divergence in the region predicted to encode the secretion signal could explain failure to localize for some Inc orthologs.

Previous phylogenetic analyses of C. trachomatis Incs from 14 published genomes (19) revealed a high degree of genetic conservation among Incs, although four (CT115, CT116, CT223, and CT229) were identified as the most divergent. This divergence was comparable to that seen with the pmps (45), encoding a class of surface autotransporters that have been linked to clinical phenotypes or tissue tropism in C. trachomatis (46). Of these four Incs, CT115 appeared most diverse at the amino acid level, although it is unknown whether any of these mutations affect function.

We did not observe localization to the inclusion membrane for several predicted Incs and subsequently verified that a few of them are not secreted from Y. pseudotuberculosis. Of these, most were previously undefined, but 4 were observed to localize within the inclusion (16). The L2 homolog of CT036 possesses a 60-bp N-terminal truncation compared to D/UW/3/CX. Additionally, the L2 homolog of CT300 possesses a single-base-pair insertion within the first 70 nucleotides. The location of this frameshift may affect the secretion signal and the function of the Inc. It is possible that while these L2 proteins are not Incs, their corresponding homologs in other C. trachomatis serovars might be. Two proteins, CT728 and CT819, exhibit significant homology to other bacterial outer membrane proteins from Bacteroides and Porphyromonas, respectively, further substantiating our findings that they are not Incs. In the current study, only CT244, CT565, and CT616 had good expression and were not secreted by Yersinia, indicating they are likely not Incs. Unfortunately, the remaining predicted Incs were unable to be validated in both assays due to weak or no expression. We propose leaving these proteins undefined and list only those validated in both assays as not inclusion membrane proteins. Additionally, CT483 and CT529 were previously reported to localize to the inclusion membrane (16, 47); however, we did not observe inclusion membrane localization for either protein. The reason for this discrepancy is unclear, but we propose keeping the two proteins as Incs until secretion can be validated in a secondary assay.

An additional consideration is the passage histories of the reference strains used in this study. LGV/434/Bu, D/UW-3/Cx, and A/Har-13 are common laboratory strains. Recently, one of the putative Incs confirmed here, CT135, was shown to be highly polymorphic in vitro with laboratory passage (48, 49). Disruption of CT135 resulted in loss of virulence of C. trachomatis D/UW-3/Cx in vivo in a mouse genital tract model of infection (50). Interestingly, loss of CT135 is more common in urogenital and ocular strains of C. trachomatis and does not seem to occur in the LGV strains (48, 49).

The results presented in this study demonstrate localization to the inclusion membrane for 10 novel predicted Incs. Collectively, our results refine and expand the list of predicted Incs. Since the ability to transform chlamydia was first described, multiple groups have made great strides to further our ability to genetically manipulate chlamydia. Chemical mutagenesis has been described, and a library of defined mutants has been generated (12, 51, 52). Specific chlamydial genes have been knocked down using dendrimer-delivered antisense RNA (53), and more recently, a group II intron-based approach was adapted, allowing selectable site-specific gene inactivation (54). The plethora of genetic tools now available to characterize chlamydial secreted virulence factors will allow us to better understand how this unique pathogen generates its intracellular niche.