TargeTron Inactivation of Chlamydia trachomatis gseA Results in a Lipopolysaccharide 3-Deoxy-d-Manno-Oct-2-Ulosonic Acid-Deficient Strain That Is Cytotoxic for Cells

ABSTRACT

All members of the family Chlamydiaceae have lipopolysaccharides (LPS) that possess a shared carbohydrate trisaccharide antigen, 3-deoxy-d-manno-oct-2-ulosonic acid (Kdo) that is functionally uncharacterized. A single gene, genus-specific epitope (gseA), is responsible for attaching the tri-Kdo to lipid IVA. To investigate the function of Kdo in chlamydial host cell interactions, we made a gseA-null strain (L2ΔgseA) by using TargeTron mutagenesis. Immunofluorescence microscopy and immunoblotting with a Kdo-specific monoclonal antibody demonstrated that L2ΔgseA lacked Kdo. L2ΔgseA reacted by immunoblotting with a monoclonal antibody specific for a conserved LPS glucosamine-PO₄ epitope, indicating that core lipid A was retained by the mutant. The mutant strain produced a similar number of inclusions as the parental strain but yielded lower numbers of infectious elementary bodies. Transmission electron microscopy of L2ΔgseA-infected cells showed atypical developmental forms and a reduction in the number of elementary bodies. Immunoblotting of dithiothreitol-treated L2ΔgseA-infected cells lysates revealed a marked reduction in outer membrane OmcB disulfide cross-linking, suggesting that the elementary body outer membrane structure was affected by the lack of Kdo. Notably, lactic acid dehydrogenase release by infected cells demonstrated that L2ΔgseA was significantly more cytotoxic to host cells than the wild type. The cytotoxic phenotype may result from an altered outer membrane biogenesis structure and/or function or, conversely, from a direct pathobiological effect of Kdo on an unknown host cell target. These findings implicate a previously unrecognized role for Kdo in host cell interactions that facilitates postinfection host cell survival.

INTRODUCTION

The obligate intracellular bacterium Chlamydia trachomatis is among the most prevalent human pathogens and is associated with disease of the genital tract and eye (1, 2). Other Chlamydia spp. cause diseases in humans and a wide variety of different animal species, and some of these infections are zoonotic (3, 4). Productive chlamydial development is initiated by host cell attachment by a metabolically limited infectious form, the elementary body (EB). Internalized EB then develops into a metabolically active and replicative form, the reticulate body (RB) (5). After several rounds of replication, there is an asynchronous secondary differentiation back to the EB, after which the bacteria are freed from the host cell and the cycle continues.

The EB:RB differentiation cycle is associated with structural changes in the outer membrane complex (OMC) (5). While some OMC components are predominately found in EBs (OmcA and OmcB), there are also OMC components that are common to both EBs and RBs. These include the serovariant major outer membrane protein (MOMP) and LPS that possesses a tri-Kdo antigenic structure common to all chlamydial species (6, 7). Outside the genus Chlamydia, the Kdo structure is shared by a limited number of strains within Acinetobacter spp. (8). Chlamydial tri-Kdo contains a unique glycosidic linkage (Kdo-(2->8)-Kdo) that constitutes the epitope recognized by chlamydial genus-specific monoclonal antibodies (MAbs) (9, 10). A single gene, genus-specific epitope (gseA), is responsible for attaching three Kdo sugars to the unique lipid IVA, generating the genus-common LPS structure (7, 11). The selective pressures leading to maintenance of this shared LPS Kdo structure are unclear and are in contrast to many other pathogenic Gram-negative species, in which structural variation is common and participates in pathogen survival and disease pathology (12). Also, in contrast to LPS produced by other bacteria, chlamydial LPS is a poor stimulator of classical pattern recognition receptor-associated innate immune functions (13), including endotoxic shock (13, 14). Antibody to the chlamydial Kdo is nonneutralizing in vitro. In fact, anti-LPS antibody enhances EB infectivity for cultured cells (15) and blocks the ability of neutralizing antibodies to bind to the EB surface (16). Despite extensive structural studies of chlamydial LPS, little is known about the role of the genus-common KDO structure in the pathogenesis of chlamydial infection. To investigate a role for KDO in host-cell interactions, we used a targeted mutagenesis approach (17) to generate a gseA-null strain (L2ΔgseA). Here, we conducted a basic biological characterization of the L2ΔgseA strain. We demonstrate that it results in significant cytotoxicity of host cells that is associated with a reduction in disulfide cross-linking of the EB cysteine-rich 6- kDa OMC protein.

RESULTS

Generation of a C. trachomatis L2 gseA-null strain.

A gseA-specific TargeTron vector was constructed using the methods described by Weber and Faris (17). PCR analysis of genomic DNA from penicillin-resistant progeny demonstrated that 1.8 kb of additional DNA was present at the gseA locus, consistent with the insertion of the recombinant bla and flanking sequence into the gseA open reading frame (Fig. 1A). Growth of the mutant strain (L2ΔgseA) was assessed by immunofluorescence of methanol-fixed infected cells using MAbs to both LPS (EVI-HI) and the major outer membrane protein (L2I-45). These analyses showed that while a MAb directed against C. trachomatis MOMP reacted with both the parental strain (L2) and L2ΔgseA, a MAb against the genus-common Kdo epitope on chlamydial LPS did not react with L2ΔgseA (Fig. 1B). This result demonstrated that TargeTron-based interruption of gseA led to a C. trachomatis mutant that does not synthesize the Kdo epitope.

FIG 1.

A transposon insertion in gseA removes the genus-common Kdo antigen from C. trachomatis. (A) PCR-based analysis of the gseA-null strain (L2ΔgseA; lane 2) and the parent L2/434/Bu (lane 3), using oligonucleotide primers flanking gseA. DNA standards are in lane 1, and PCR product sizes are indicated. (B) Immunofluorescence microscopy comparing the WT and the mutant strain. Infected monolayers of both L2/434/Bu and L2ΔgseA were methanol-fixed 32 hpi and then stained with anti-MOMP (green) and anti-LPS (red). DNA was labeled with DAPI (blue). The scale bar in the bottom-left corner indicates 10 microns for each image. (C) Immunoblot of McCoy cell monolayer lysates either mock infected (lane 2), L2/434/Bu infected (lane 3), or L2ΔgseA infected (lane 4) and labeled with the anti-MOMP and anti-LPS MAbs. Molecular mass standards are in lane 1, with values indicated in kilodaltons.

Immunoblotting of uninfected and infected cell culture monolayers with a mixture of anti-MOMP MAb and anti-Kdo MAb was used to characterize the antigenic nature of the gseA mutant. The truncated chlamydial LPS migrates approximately 10 kDa following electrophoresis on polyacrylamide gels (6). Based on immunoblotting, the expected 40 kDa MOMP protein was present in both the parental and mutant strains, while LPS was only detected in the parental strain (Fig. 1C). These data demonstrate that L2ΔgseA does not express the tri-Kdo structure. (Fig. 2A, red labeling). Further immunoblotting using a MAb specific to chlamydial lipid A (S55-5) (18) was used to determine if a Kdo-deficient lipid A structure was retained in the L2ΔgseA mutant strain. This antibody is specific for a conserved phosphorylated glucosamine residue common to most lipid A molecules (Fig. 2A, green labeling). Immunoblotting of cell lysates from mock-infected, L2/434/Bu and L2ΔgseA-infected McCoy cells with MAb S55-5 demonstrated that the epitope was present in lysates of the L2ΔgseA strain. These findings support the conclusion that the L2ΔgseA strain, despite lacking tri-Kdo, maintained the core LPS fatty acid structure. In contrast, MAb S55-5 did not react with L2/434/Bu-infected cell lysates. We speculate that this lack of reactivity with wild-type (WT) LPS was the result of steric hindrance introduced by the presence of the terminal tri-Kdo structure that prevented the Ab from binding the intact LPS molecule (18).

FIG 2.

Western blotting with Mab S55-5, which labels an epitope centered on the lipid A phosphates, demonstrates that the core lipid A structure is present in L2ΔgseA developmental forms. (A) Chlamydial LPS structure, as determined by Rund et al. (29). Color coding shows the core lipid A backbone (black), the phosphates recognized by Mab S55-5 (green), and the tri-Kdo carbohydrates that are missing in L2ΔgseA (red). (B) Immunoblot of C. trachomatis lysates probed with Mab S55-5. Lanes: (1) Protein molecular weight standards; (2) mock-infected McCoy cell lysate; (3) lysate of L2/4340infected cells; (4) Lysate of L2ΔgseA-infected cells.

Genome sequence analysis of L2ΔgseA.

Strain L2ΔgseA was subjected to full genome sequence analysis and compared to the specific LGV/434/Bu isolate used as the parent. The TargeTron-sourced vector sequence was identified at position 996 in gseA. The transposition process led to a 1,782-bp insertion within the gseA coding sequence, which was consistent with the amplification product identified by PCR (Fig. 1A). Three additional positions in the genome were mutated in the L2ΔgseA genome, including nonsynonymous base changes in mutL and the hypothetical gene CTL_0309 and a single base deletion in recF. These additional mutations occurred in differing fractions of the total reads for each gene, indicating that they are present as subpopulations in the mutant strain.

Strain L2ΔgseA has a decreased growth rate relative to that of LGV/434/Bu.

A one-step growth curve was conducted with L2ΔgseA, and its growth was compared to the parental strain. These results demonstrated that L2ΔgseA produces infectious EBs but at a slower and less productive rate than the parent strain (Fig. 3A). Phase-contrast microscopy of infected cells showed that both the number of inclusions and general inclusion structure were similar in cells infected with either wild-type or mutant strains. However, inclusions formed by L2ΔgseA appeared more granular in nature and less densely populated with developmental forms (Fig. 3B).

FIG 3.

Growth characteristics of C. trachomatis strain L2ΔgseA. (A) Elementary body production in a one-step growth curve for the WT strain (triangles) and L2ΔgseA (squares). This figure is the average of data from two independent experiments. Each strain was grown at an MOI of 0.5 in McCoy cells and harvested at the indicated time points, and the titers of recoverable inclusion-forming units (IFUs) were determined in McCoy cells. (B) Phase-contrast microscopy of inclusion development in McCoy cells infected and cultured for different amounts of time. The box in each micrograph shows a higher magnification of the image for each time point and strain.

Two hypotheses were examined to explain the lower abundance and slower growth of L2ΔgseA. First, we examined whether the absence of wild-type LPS led to altered outer membrane structure, which might affect RB:EB differentiation and growth. This possibility was investigated by studying OmcB disulfide cross-linking in WT- and L2ΔgseA-infected cells. OmcB is an abundant cysteine-rich outer membrane protein found only in EB (19). Disulfide cross-linking of OmcB confers in part the rigidity and structural integrity of the chlamydial outer membrane (OM). Immunofluorescence with anti-OmcB antisera demonstrated that OmcB was expressed in both LGV/434/Bu and L2ΔgseA inclusions. OmcB fluorescent staining was less punctate and was more diffuse in L2ΔgseA (Fig. 4A). Differences in IFA staining were confirmed ultrastructurally by transmission electron microscopy (TEM) (Fig. 5A). TEM images showed that L2ΔgseA produced more atypical, larger RB forms, and EBs were less abundant than in WT-infected cells.

FIG 4.

The distribution and cross-linking of outer membrane complex protein OmcB is altered in strain L2ΔgseA. (A) Immunofluorescence of McCoy cells infected with either L2/434/Bu or L2ΔgseA and fixed 42 hpi and then labeled with antibodies to OmcB (green) and MOMP (red). (B) Immunoblotting of mock-infected, L2/434/Bu-infected (L2), and L2ΔgseA-infected McCoy cells with an antiserum directed at OmcB. Electrophoresed samples either did not include (−) or did include (+) 70 mM DTT during the cell lysis step. Molecular mass standards (kDa) are indicated at the left. The arrow points to monomeric OmcB (60 kDa).

FIG 5.

Evidence for toxicity of C. trachomatis L2ΔgseA in McCoy cell monolayers. (A) Appearance of late C. trachomatis inclusions with both phase-contrast (images 1 and 2) and TEM (images 3 and 4). Monolayers of L2/434/Bu (image 1) and L2ΔgseA (image 2) were examined by phase-contrast microscopy after 42 h of growth. Images 3 and 4 show TEM images of McCoy cells infected with L2/434/Bu (image 3) and L2ΔgseA (image 4). Cells were infected for 32 h prior to fixation for TEM. (B) Cellular toxicity as measured by LDH release in McCoy cells infected with the WT and mutant strains. Each value represents the LDH release compared to detergent-lysed monolayers. The multiplicity of infection was 0.5 in these assays.

We performed immunoblotting with anti-OmcB antiserum on infected cells to investigate where there were differences between strains in OmcB interactions with the OM (Fig. 4 B). Infected cell lysates were treated or not treated with dithiothreitol (DTT), a reducing agent that disassociates OmcB disulfide bonds, allowing the identification of both monomeric OmcB and its processed polypeptides contained in disulfide cross-linked EB OMC (19, 20). As described previously (21), we found by immunoblotting that LGV/434/Bu monomeric OmcB (circa 60 kDa) solubility was DTT dependent. In the presence of DTT, OmcB polypeptides of circa 30 and 20 kDa, which are processed from mature OmcB (19, 20) were detected. Notably, anti-OmcB immunoblotting of L2ΔgseA lysates showed that the OmcB monomer and the processed polypeptides were largely independent of DTT treatment. These data show that the removal of the LPS tri-Kdo in L2ΔgseA is associated with a reduction in OmcB disulfide cross-linking in chlamydial EBs.

L2ΔgseA exhibits increased toxicity to host cells.

Phase-contrast microscopy of infected cells at later time points postinfection exhibited an infected cell morphology characteristic of cell toxicity in L2ΔgseA-infected cells, while the L2/434/Bu-infected cells retained a normal morphological appearance (Fig. 5A). Toxicity in L2ΔgseA-infected cells was quantified using a lactate dehydrogenase (LDH) release assay. McCoy cells infected with the mutant or wild-type strains were collected at 12, 24, and 40 h postinfection and assayed for LDH release. Cells infected with both strains demonstrated minimal LDH release at 12 h postinfection (hpi). However, L2ΔgseA-infected cells exhibited significantly higher levels of LDH release than WT-infected cells at 24 and 40 hpi. These data support the conclusion that the absence of Kdo on LPS directly, or indirectly, results in host cell cytotoxicity.

Secretion of chlamydial proteins to the host cytosol is similar in the parental and mutant strains.

Chlamydiae have different secretion systems to deliver cargo to the host cytosol (5, 22–25). We used immunofluorescence microscopy to qualitatively assess the ability of L2ΔgseA to deliver secreted proteins to the host cytosol. Chlamydial inclusion membrane proteins are delivered by a type III secretion system to the inclusion membrane of infected cells (22). IncA (Fig. 6A and B) was localized to the inclusion membrane similarly in both WT- and mutant-infected cells. The C. trachomatis protease CPAF is secreted to the host cytoplasm via the type II machinery (23, 24). There was no difference in CPAF secretion to the cytosol in cells infected with either the WT or the L2ΔgseA-knockout strain (Fig. 6C and D). The plasmid-regulated protein GlgA is secreted to the host cytosol by a plasmid-dependent secretion system (25). Cells infected with either L2ΔgseA or the parental strain secreted GlgA to the cytosol (Fig. 6E and F). These data support the conclusion that proteins secreted to host cytosol by type II, type III, or the plasmid-dependent mechanism are not affected by LPS lacking Kdo.

FIG 6.

Immunofluorescence of parent and mutant strains with antibodies to secreted chlamydial proteins. (A to F) McCoy cell monolayers were infected with either L2/434/Bu (panels A, C, and E) or L2ΔgseA (panels B, D, and F) for 32 to 40 h prior to fixation with paraformaldehyde, permeabilization with Triton X-100, and labeling. DNA is labeled blue in each image with DAPI. Antibody targets are indicated to the left of each pair of labeled images. The antibodies used in the labeling are MAb 3H7 (IgG1), which labels the type III-secreted inclusion membrane protein IncA (22), Anti-LPS MAb EVI-HI (IgG2a), anti-CPAF MAb 100a (IgG1), (23), anti-GlgA MAb F443G (IgG1), (25), and L2I-45, an antibody to MOMP (25). The arrows point to labeling for IncA in panels A and B, CPAF in panels C and D, and GlgA in panels E and F. The scale bar in panel C represents 10 microns for panels A to D.

DISCUSSION

The genus-common antigen possessed by all Chlamydia spp. was first identified as a carbohydrate by Dhir and colleagues (26). Subsequent research demonstrated that the antigen is represented by the tri-Kdo oligosaccharide in the chlamydial LPS molecule (9, 10). The highly conserved LPS is a poor inducer of classic canonical and noncanonical LPS activation of innate immunity (13, 14), but its role in the pathogenesis of chlamydial host cell interactions important to infection or disease remains unclear.

The terminal enzyme in chlamydial LPS synthesis is gseA, which encodes a glycosyltransferase that attaches the tri-Kdo genus-common antigen to lipid A (7, 11). The TargeTron system was used to generate a mutant strain, L2ΔgseA, by insertional inactivation of the gseA coding sequence. Immunofluorescence microscopy and immunoblotting with anti-Kdo MAb demonstrated that the mutant did not produce the genus-common tri-Kdo LPS antigen. While this provided evidence that the tri-Kdo was missing in L2ΔgseA, it remained unknown if the remaining core LPS fatty acid structures were made following gseA inactivation. This was addressed with a MAb that recognizes a lipid A structure that lacks Kdo sugars (18). The MAb reacted in immunoblots with LPS of L2ΔgseA-infected cells, leading to the conclusion that the mutant strain retains the ability to make the chlamydial core lipid A structure in the absence of a terminal tri-Kdo.

In a previous study, Nguyen et al. characterized the effects of small-molecule inhibitors of LpxC on chlamydial growth and development (27). LpxC participates in lipid A synthesis in Gram-negative bacteria, leading to a block in their ability to synthesize LPS. Treatment of Chlamydia spp. with these inhibitors leads to typical early and midcycle growth of the bacteria, but development terminated with inclusions containing only RB forms. We explored intracellular growth of L2ΔgseA and determined that, while EB production was slower than that of the wild-type parent strain, EBs were produced, and the developmental cycle was completed. Therefore, while Nguyen et al. showed that inhibiting an early step in LPS synthesis led to chlamydial development that was interrupted during the RB-to-EB transition, we found that a core lipid A structure was sufficient to allow completion of the developmental cycle and the production of infectious EBs.

Three abundant proteins in the outer membrane of EBs are heavily disulfide cross-linked and form a complex, the OMC, that resists solubilization in the absence of a reducing agent (5, 19). We hypothesized that the slower growth of L2ΔgseA may result from disruption in OMC protein:protein interactions in the absence of the Kdo structure at the EB surface. This was studied by immunoblotting, with anti-OmcB antibodies, of infected monolayers lysed in detergent with and without DTT. These experiments demonstrated that the cross-linking of OmcB was markedly reduced in cell lysates infected with L2ΔgseA, indicating a disruption in normal EB OM biogenesis. Future studies will explore in detail the nature of protein:protein interactions in the outer membrane of the Kdo-deficient mutant and how OMC structural differences associated with the lack of Kdo might affect chlamydial growth and host cell toxicity.

A second possible reason for the less productive growth of L2ΔgseA is the finding that L2ΔgseA was toxic to host cells. Purified chlamydial LPS is significantly less toxic to mammalian cells than is Escherichia coli LPS (28). The reduction in endotoxic activity is associated with the inability of chlamydial LPS to activate innate immune processes in mammalian cells and has been correlated with the unique structure of chlamydial lipid A (10, 29). Microscopic evidence of host cell toxicity late in development was evident in cells infected with L2ΔgseA, and this was supported by LDH release experiments comparing WT and mutant strains. We hypothesize that several factors may be associated with the increased toxicity. First, as our data demonstrate, the OMC is altered in L2ΔgseA, as measured by high levels of non-cross-linked, monomeric OmcB and its breakdown products, within infected cells. A C-terminal fragment of OmcB is released into the host cytoplasm during infection (20), and it is possible that this may have unidentified effector functions that are of higher abundance and, therefore, dysregulated in L2ΔgseA. A second possibility is that the secretion of known and unknown effector proteins is affected by the lack of Kdo sugars at the bacterial cell surface, leading to a more active cellular response to intracellular infection. Chlamydiae are known to actively block apoptotic processes, and secreted effector proteins provide protection against cell death pathways (30–32). We did not examine this question directly in our study, but we do provide evidence that each of the known secretion pathways in Chlamydia can function in L2ΔgseA (Fig. 6). A final possibility is that the KDO sugars function by directly preventing host cell death in C. trachomatis-infected cells, and the removal of this structure from LPS triggers cell death pathways that might be used against intracellular infections. LPS on the surface of Gram-negative bacteria is constantly shed, either directly or in the form of extracellular membrane vesicles, and such shedding is known to have a variety of effects on the host. There is evidence that C. trachomatis sheds LPS during infections of monolayers (33, 34), although the scope and significance of this during infection remain unclear (35). We expect that the gseA knockout strain will be a unique tool for examining host stress responses following chlamydial infection. Future experiments will explore the role of the unique genus-common Kdo structure during infection of hosts by Chlamydia spp.

MATERIALS AND METHODS

TargeTron mutagenesis.

Chlamydia trachomatis LGV-434-Bu (L2) was the parental strain used for TargeTron mutagenesis of gseA (CTL_0460). This parental strain was used as a control in each subsequent experiment. Site-specific inactivation of gseA was accomplished using the methods of Weber and Faris (17). The pACT vector was modified with gseA-targeting oligonucleotides IBS (AAAAAAGCTTATAATTATCCTTATTTGTCGGAGGAGTGCGCCCAGATAGGGTG), EBS1 (CAGATTGTACAAATGTGGTGATAACAGATAAGTCGGAGGAACTAACTTACCTTTCTTTGT), and EBS2 (TGAACGCAAGTTTCTAATTTCGATTACAAATCGATAGAGGAAAGTGTCT). Targeting sequences were then ligated into digested TargeTron vector and transformed into dam^–/dcm^– E. coli (New England Biolabs). Kanamycin-resistant colonies were selected and expanded, and plasmids were extracted with the QIAPrep Spin Miniprep kit (Qiagen). A sequence-confirmed construct was expanded in E. coli and purified for use in transformation of C. trachomatis.

Purified L2/434/Bu EBs were mixed with 6 μg of the purified plasmid in sterile transformation buffer (10 mM Tris, pH 7.5, 50 mM CaCl₂). This mixture was incubated at room temperature for 30 min and then distributed into each well of a 6-well plate containing Hanks balanced salt solution (HBSS)-washed monolayers of L929 cells. This tray was then centrifuged (545 × g, 1 h, 20°C), followed by removal of inocula and addition of Dulbecco’s modified Eagle medium plus 10% fetal bovine serum (FBS; DMEM-10; Corning, Inc., Corning, NY, USA). Plates were incubated at 37°C in 5% CO₂ for 5 h, followed by replacement of the antibiotic-free DMEM-10 with DMEM-10 plus 1 mM cycloheximide and 10 IU/mL of penicillin G. Cultures were then incubated for 48 h. For each subsequent passage, the medium was removed, and infected cells were harvested via scraping. EBs were released from the infected cells by vortexing with borosilicate beads. These suspensions were centrifuged at 545 × g for 5 m at 4°C, and the supernatants were inoculated onto to 6-well plates of McCoy cells and centrifuged. The inoculum was then removed, DMEM-10 with cycloheximide and penicillin G was added, and the plates were incubated for 48 h. These culture cycles were continued until penicillin-resistant EBs were culturable from the monolayers. Initial screens of candidate gseA-interrupted strains were checked with PCR using oligonucleotide primers pET26b-CT208 forward (GGAATTCCATATGATAAGACGTTGGTTAACATC) and pET26b-CT208 reverse (CCGCTCGAGGATTTTCATGCAAGTAATTTGG).

Genome sequencing and analysis.

Sequencing libraries were prepared from both the parent and mutant strains using the TruSeq DNA PCR-free library preparation kit, in accordance with the manufacturer’s instructions (Illumina, San Diego, CA). Briefly, 500 ng genomic DNA was sheared on the M220 instrument (Covaris, LLC., Woburn, MA) to generate an average insert size of 350 bp. The final TruSeq PCR-free library was quality assessed using the Agilent Bioanalyzer and quantified on the CFX96 real-time PCR system (Bio-Rad, Hercules, CA) using the Kapa library quantification kit (Roche, Indianapolis, IN). Libraries were sequenced as 2 × 150-cycle paired-end reads on the MiSeq instrument using the Nano kit v2 (Illumina). Raw fastq files were trimmed of adapter sequences using cutadapt v1.12 (https://doi.org/10.14806/ej.17.1.200), with the -m 20 parameter setting. Adapter-trimmed reads were quality trimmed and filtered using the FastX toolkit v0.0.14 (http://hannonlab.cshl.edu/fastx_toolkit), removing reads with a phred score below 20 and read length below 40. Processed reads were mapped to the genome sequence of C. trachomatis strain L2/434/Bu (NC_010287.1) using Bowtie2 v2.2.9 with -no-mixed and -no-unal parameter settings (36). Quality-filtered reads were also mapped to bla from the TargeTron vector (17) nucleotide sequence. Sequence alignment/map files were converted to BAM format, sorted, and indexed using SAMtools v1.10 (37). Duplicate reads were removed using Picard Mark Duplicates (http://broadinstitute.github.io/picard/), and variants were detected using the GATK package v4.2.5.0 HaplotypeCaller (38). The genome sequence data can be accessed through BioProject no. PRJNA934942.

Immunofluorescence.

McCoy cells (seeded at 2 × 10⁵ cells/mL) on coverslips were infected with parent or mutant strains at a multiplicity of infection (MOI) of 0.5. Infected cells were then fixed either with ice-cold methanol for 10 min or with 4% paraformaldehyde for 30 min. Paraformaldehyde-fixed cells were then permeabilized with 0.1% Triton X-100 for 2 min. Fixed cells were then washed three times with phosphate-buffered saline (PBS). Wells were then blocked for 30 m with 2% bovine serum albumin (BSA) in PBS at 37°C on a rocker platform. Antibodies were incubated on fixed cells for 1 h at 37°C and then washed with 2% BSA in PBS. Appropriate secondary antibodies and DAPI (4′,6-diamidino-2-phenylindole) were then added, and the cells were rocked again for 1 h at 37°C on the rocker and then removed. The wells were washed three times with 1× PBS and then kept in PBS until they were mounted. Coverslips were then inverted onto microscope slides with ProLong Gold antifade mountant (Thermo Fisher Scientific) and were imaged with a Nikon Eclipse 80i microscope; images processed with ImageJ and Nikon NIS elements software.

Quantitative one-step growth curves.

Plates (24 wells) were seeded with McCoy cells at 4 × 10⁵. After an overnight incubation, cells were inoculated with the wild-type or mutant C. trachomatis strains at an MOI of 0.5 and centrifuged at 20°C for 1 h at 545 × g. Inocula were replaced with DMEM-10 with cycloheximide and penicillin G (in L2ΔgseA only). Directly after centrifugation, the 0 hpi time point cell were harvested, and the remaining wells were placed in the incubator at 37°C. Triplicate samples were harvested at 6, 18, 24, 32, and 42 hpi. Cells collected after each time interval were harvested by removing the medium, adding 500 μL of cold sucrose-phosphate-glutamic acid (SPG) to the well, and scraping. The lysate was vortexed with glass beads and centrifuged for 5 min at 4°C at 545 × g. Harvested bacteria were stored at −80°C. For quantification of EBs at each time point, 10-fold serial dilutions of the samples were prepared and inoculated onto McCoy cells in 48-well trays in triplicate. These inocula were centrifuged at 20°C and 545 × g for 1 h and then incubated in appropriate medium for 30 hpi before methanol fixation. Fixed cells were labeled with L2-I45 as described above. The average numbers of inclusions viewed on 10 fields of each well were calculated. Each sample was processed in triplicate for final average counts per well.

Significance was calculated using a two-sided unpaired t test, and the resulting data were imaged in GraphPad Prism v9. The growth curve was repeated twice.

Immunoblotting.

McCoy monolayers grown in 6-well plates were infected with WT or mutant strains and harvested at either 32 hpi or 44 hpi, as indicated in the results. Lysates were subjected to electrophoresis through 12% (for OmcB) or 15% (for LPS) acrylamide gels using a Tris-glycine-based buffer system. Samples were standardized for equivalent amounts of MOMP. Electrophoresed products were transferred onto polyvinylidene difluoride (PVDF) membranes, which were then blocked for 1 h in Tris-buffered saline (TBS) plus 5% powdered skim milk (TBSM). The indicated antibodies were diluted in 5% milk in TBS to the appropriate concentrations and then rocked at room temperature overnight. The membrane was washed three times in TBSM, followed by incubation in horseradish peroxidase antibody-conjugated secondary antibodies diluted in TBSM. Blots were washed 3 times in TBSM and imaged via the Bio-Rad GelDoc Go gel imaging system.

LDH assay.

The CytoTox 96 nonradioactive cytotoxicity assay (Promega) was used for the LDH assay according to the manufacturer’s instructions. Briefly, McCoy cells were infected with each strain at an MOI of 0.5, or mock-infected, in wells of a 96-well plate. The supernatant from cells from each time point was collected for the assay. Samples were centrifuged to remove cell debris, and the supernatant was assayed for LDH activity. Assay buffer was added to each well, and the reactions were incubated for 30 min at room temperature in the dark. Stop solution was added to each sample, and the absorbance was read at 490 nm. The percent lysis was calculated by comparing the LDH release from each test well to values obtained following detergent-mediated lysis of uninfected cells. Each data point was collected in triplicate, and the assay was conducted twice.

Transmission electron microscopy.

Thermanox plastic coverslips were placed in a 24-well plate, sterilized with ethanol, and then washed with sterile HBSS. Each well was seeded with McCoy cells at 4 × 10⁵ cells per well. Cells infected with each strain were cultured for 32 h and then fixed at 4°C in Karnovsky’s fixative, consisting of 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M Sorensen’s buffer. The samples were postfixed for 1 h with 0.5% osmium tetroxide and 0.8% potassium ferricyanide in 0.1 M sodium cacodylate, stained for 1 h with 1% tannic acid, and then stained with 1% aqueous uranyl acetate for 1 h. Samples were rinsed with distilled water (dH₂O) and dehydrated in a graded ethanol series through 100% ethanol. The samples were then infiltrated and embedded with Spurr’s resin. Thin sections were cut using an EM UC6 ultramicrotome (Leica Microsystems, Vienna, Austria) and viewed using an HT7800 transmission electron microscope at 80 kV (Hitachi High-Tech, Tokyo, Japan). Images were captured with a digital camera system (AMT, Woburn, MA).