C. trachomatis serovar L2 (LGV 434) was propagated in HeLa 229 cells (ATCC CCL-2.1) or McCoy B cells (ATCC CRL-1696) and purified by Renografin density gradient centrifugation (Scidmore, 2005).

In previous studies we had generated a number of in-frame Tarp deletions which were expressed as mutant GST-Tarp fusion proteins from pGEX-6p-1 (GE Health Sciences) plasmids (Jiwani et al., 2013). Tarp domain deletion mutants included: phosphorylation domain deletion (Δphos; deletion of D125 to Y424), proline rich domain deletion (ΔPRD; deletion of S625 to N650), actin binding domain deletion (ΔABD; A748 to K758), and F-actin binding domain 1 and 2 deletion (ΔFAB 1&2; deletion of L871 to G1005). These mutant Tarp alleles were subcloned into the chlamydial shuttle vector pCtSV.1 in a two-step process. First, wild type Tarp sequence was amplified from C. trachomatis (LGV 434) genomic DNA (Qiagen genomic purification kit, Valencia, CA). The forward (5′ACTCCGCGGTATTGCAT TTCTTCACAAACGTTACC-3′) and reverse (5′TATA TACAATTGTTACAGGTCCTCTTCAGATATTAGTTTTTGTTCTCCTACGGTATCAATCAGTG AGC-3′) DNA primers (Integrated DNA Technologies, Coralville, IA) were engineered to amplify 200 bases of putative Tarp promoter sequence and an in frame 3′ c-myc epitope tag by PCR with SacII and MfeI linkers. PCR products were purified (Qiagen), digested with restriction enzymes (New England Biolabs, Beverly, MA) and cloned into linearized pCtSV.1. This procedure resulted in the parent pCtSV.Tarp plasmid in which all other plasmids engineered to express Tarp mutants were generated. pCtSV.Tarp mutant derivatives were generated by exchanging the mutant DNA sequence from those pGEX-6p-1 clones described above. For example, pCtSV.TarpΔphos resulted from DNA exchange with digested Tarp DNA sequence flanking the phosphorylation domain with restriction sites BstAP1 and BmgB1 from pGEX-6p-1 TarpΔphos. Similarly, the other pCtSV.Tarp mutant clones were generated albeit with unique restriction enzymes which flanked the corresponding domain: The proline rich domain with BmgB1 and Bsm1, the actin binding domain with Bsm1 and Nco1, and the F-actin binding domains 1 and 2 with Nco1 and Mfe1. All engineered vectors were confirmed to be free of extraneous mutations by DNA sequence analysis and all in frame domain deletions were verified. All chlamydial shuttle vectors were purified from E. coli K12 ER2925 cells (New England Biolabs) and transformed into C. trachomatis (LGV 434) (Wang et al., 2011). All C. trachomatis transformants were density gradient purified and the number of infectious (inclusion) forming units (IFUs) was determined by indirect immunofluorescence of infected host cells with serially diluted purified elementary bodies (EBs).

Pyrene actin polymerization assays were performed as previously described (Jiwani et al., 2013). Briefly, monomeric pyrene-labeled actin was prepared by diluting 100 μg of lyophilized pyrene actin (cytoskeleton Inc. Denver, CO) in 2 mL of 5 mM Tris (pH 8.0), 0.2 mM CaCl₂, 0.2 mM ATP (G buffer) and incubated for 1 h at room temperature, followed by an additional 1 h incubation at 4°C. Monomeric pyrene actin was obtained by collecting the supernatant after a 2 h 100,000 × g 4°C spin in a Beckman Optima TLX Ultracentrifuge using a TLA 100.3 rotor (Beckman Coulter). Approximately 20 μg of pyrene-labeled actin was gently mixed with 5 μg of GST fusion proteins in a volume of 500 μL for 10 min before the addition of 1/20th volume of polymerization buffer (500 mM KCl, 20 mM MgCl₂, 10 mM ATP). The reaction was monitored over 1 h with an LS 55 Luminescence spectrophotometer directed by FL WinLab software version 4.0 (Perkin-Elmer, Beaconsfield, Bucks, United Kingdom) with 2.5-nm bandwidth at 365-nm excitation wavelength and 2.5-nm bandwidth at 407-nm emission wavelength.

Proteins were separated on SDS 5–15% polyacrylamide gels (BIORAD, Hercules, CA) and either stained with Imperial protein stain (Pierce, Rockford, IL) or transferred to 0.45 μm pure nitrocellulose transfer and immobilization membrane (Schleicher and Schuell, Keene, NH). Immunoblotting employed peroxidase conjugated secondary antibodies (Chemicon International, Temecula, CA) and Supersignal West Pico chemiluminescent substrate (Pierce). The anti-actin C4 monoclonal antibody was purchased from Chemicon International. The anti-actin polyclonal antibody was purchased from Cytoskeleton, Inc. The anti-phosphotyrosine 4G10 monoclonal antibody was purchased from Upstate (Millipore). The anti-chlamydial EB polyclonal antibody, the Momp monoclonal antibody and the GAPDH monoclonal antibody were all purchased from Pierce. The anti-c-myc monoclonal antibody was purchased from Genscript (Piscataway, NJ). The anti-chlamydial Hsp60 A57-B9 monoclonal antibody was purchased from Thermo Fisher Scientific (Waltham, MA). Polyclonal rabbit antibodies directed toward C. trachomatis L2 LGV 434 Tarp (CT456) were developed at Rocky Mountain Laboratories as previously described (Clifton et al., 2004).

Intrinsically fluorescent EBs from C. trachomatis transformants were purified from cell cultures supplemented with CellTracker™ Red CMTPX Dye as previously described (Carabeo et al., 2007). Briefly, CMPTX-labeled C. trachomatis EBs (MOI ~50) were permitted to attach to HeLa 229 host cells for 30 min at 4°C. HeLa 229 cells were prepared in 24 well plates with cover slips and grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) and 1% L-glutamine for 24 h prior to infection. The cultures were rinsed with cold HBSS and the temperature shifted to 37°C by the addition of pre-warmed DMEM plus 10% FBS. The cultures were then incubated at 37°C for 1 h. The cultures were fixed with 4% paraformaldehyde at room temperature for 15 min and rinsed with PBS. The cells were not permeabilized. Extracellular EBs were labeled for 1 h with a monoclonal antibody specific for chlamydial major outer membrane protein (MOMP). After four washes in PBS, secondary antibody conjugated to Alexa 488 was added for 1 h. Coverslips were rinsed and mounted in ProLong Gold antifade reagent (Invitrogen, Carlsbad, CA). Cells were examined with a Zeiss Axio Observer A1 microscope equipped with a phase-contrast and epifluorescence optics. Images were obtained using an AxioCam MRm camera controlled by Axio Vision 4.8.2 and further processed using Adobe Photoshop CS2. The number of green (external) and red (total) EBs was determined for each host cell. These data were then used to determine the percentage of internalized EBs. Twenty fields of view were taken from each cover slip and these percentages were then averaged together to give a final invasion rate.

C. trachomatis infected cells underwent subcellular fractionation as previously described (Cox and Emili, 2006). Briefly, C. trachomatis infected McCoy or HeLa 229 cells maintained at 37 or 4°C or host cells alone incubated at 37°C were removed from flasks and suspended in 100 mM KCl, 10 mM HEPES (pH 7.7), 2 mM MgCl₂, and 2 mM ATP (Buffer A) and disrupted by sonication delivered in three consecutive 30 s intervals (~ 2000 joules) using an ultrasonic sonicator processor XL equipped with a microtip (Misonix Incorporated, Farmingdale, NY). All cell lysates underwent subcellular fractionation by sequential centrifugation in which supernatants and pellets were separated. Lysates were initially subject to an 800 × g spin for 15 min at 4°C. The 800 × g supernatants were then subjected to a 10,000 × g spin for 30 min at 4°C. The remaining 10,000 × g supernatant underwent a 100,000 × g spin for 1 h at 4°C. Protein sample buffer was added to all pellets and supernatants and proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes for immunoblotting with antibodies specific for c-myc, Tarp, actin, GAPDH, Momp, and chlamydial EBs. Subcellular fractionation experiments were conducted in both McCoy (Figure 6) and HeLa 229 (Supplemental Figure 1) cells.

HeLa 229 cells were seeded into 6 well plates (2 × 10⁵ cells/well) and grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) and 1% L-glutamine for 24 h. Individual wells were infected with wild type C. trachomatis L2 (LGV 434) or C. trachomatis transformants. All host cells and bacteria were collected from select wells (cells scraped off the bottom of each well and collected in 15 mL tubes, and sonicated for 30 s using a microtip equipped Misonix sonicator) at 0, 12, 24, 36, and 48 h. Cell lysates were then frozen at −80°C until all time points had been collected. Cell lysates were thawed on ice and diluted and then placed onto HeLa cells grown on 16 mm circular cover slips contained within 24 well plates and grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% L-glutamine. After a 40 h incubation, infected cells were then immunostained and observed under a fluorescent microscope for inclusion formation. Twenty fields of view were taken from each cover slip (the experiment was performed in triplicate) and cover slip counts were averaged. Averages were plotted using GraphPad Prism software.

The Tarp actin binding domain (ABD, amino acids 748–758) and the proline rich polymerization domain (PRD, amino acids 625–650) have previously been found to be required for Tarp mediated polymerization of pyrene actin in vitro (a summary of Tarp protein domains is provided in Figure 1; Jewett et al., 2006). One model for Tarp mediated actin nucleation advocates that Tarp oligomerizes via the proline rich domain. Tarp effectors, which come together following secretion into the host cell cytosol, associate with monomeric actin to form an actin nucleus, affectedly reducing the critical concentration of actin required for actin filament formation (Jewett et al., 2006). Since Tarp is hypothesized to function in a homo-oligomeric complex, we tested whether TarpΔPRD and TarpΔABD were able to disrupt wild type Tarp mediated actin polymerization in a dominant negative fashion. To investigate whether mutant Tarps could inhibit actin polymerization induced by wild type Tarp, purified recombinant wild type Tarp proteins and TarpΔABD or TarpΔPRD were combined at equal molar ratios and analyzed for actin nucleation in pyrene actin polymerization assays (Figure 2). Similar to our previous findings, when TarpΔPRD or TarpΔABD alone was introduced to pyrene actin, an increase in the rate of actin polymerization was not observed compared to actin alone controls (Figures 2B,C). The amount of TarpΔPRD employed in these experiments was lower compared to previous studies to ensure that TarpΔPRD would not appreciably sequester monomeric actin (Jiwani et al., 2013). At higher concentrations, TarpΔPRD sequesters monomeric actin and may disrupt native Tarp function by lowering the concentration of available actin in the reaction. As predicted, when (TarpΔPRD or TarpΔABD) were mixed with an equal amount of wild type Tarp, the rate of actin polymerization was reduced (Figures 2B,C). Although the rate of pyrene actin polymerization was reduced in the presence TarpΔPRD or TarpΔABD, some Tarp mediated actin nucleation was still observed. To confirm that the reduction of actin polymerization was not the consequence of steric hindrance caused by excess Tarp in the reaction, an additional control (TarpΔphos) which is not predicted to interfere with actin polymerization was tested. Tarp which lacks the tyrosine rich phosphorylation domain (TarpΔphos, harbors the deletion in amino acids 125–424) has previously been shown to nucleate actin to equivalent rates as compared to wild type Tarp (Jiwani et al., 2013). To test whether Tarp mediated actin polymerization might be affected by TarpΔphos (Figure 2D), purified Tarp and TarpΔphos were mixed at equal molar ratios and tested in pyrene actin polymerization assays. TarpΔphos combined with wild type Tarp enhanced the rate of actin polymerization due to the two-fold molar increase of functional actin nucleators (Figure 2E). These data support the hypothesis that biochemically wild type Tarp may experience reduced actin polymerization kinetics when associated with or competing with defective TarpΔPRD or TarpΔABD. Conversely, wild type Tarp mediated actin polymerization is not inhibited by TarpΔphos.

In order to determine whether mutant Tarp effectors might disrupt endogenous Tarp function in vivo, we engineered a chlamydial shuttle vector to express epitope tagged wild type and mutant Tarp effectors (Figure 3A). C. trachomatis genes are temporally regulated so as to match their function with the correct window in the chlamydial developmental cycle. To promote coordinated expression of the mutant Tarp alleles with that of the endogenous Tarp gene, mutant Tarp effectors expressed from the chlamydial shuttle vector were engineered under the control of ~200 nucleotides of DNA upstream of the annotated Tarp gene, which we have termed the Tarp promoter (tarPp) (Figure 3). Mutant Tarp constructs included deletions of the phosphorylation domain, the proline rich domain, the actin binding domain and a Tarp truncation which resulted in the deletion of the F-actin binding domains 1 and 2 (Figure 3A). C. trachomatis L2 transformed with each shuttle vector were expanded under antibiotic selection for several passages, density gradient purified and tested for the presence of the c-myc epitope tag by western blot analysis. All of the transformants produced c-myc tagged proteins of the expected size (Figure 3B). Interestingly, endogenous Tarp in C. trachomatis expressing TarpΔphos appears to be lower compared to the relative abundance of Tarp in the other transformants. This may hold true for the other transformants (ABD, PRD, and FAB1 and 2) but cannot be visualized in the western analysis since mutant and endogenous Tarp migrate to the same position on the protein gels.

We hypothesized that production of a dominant negative Tarp complex in vivo would have a significant effect on the ability of elementary bodies to invade host cells. To test the invasion potential of the five different C. trachomatis clones expressing mutant Tarp effectors from the C. trachomatis shuttle vector, we performed invasion assays to quantitate the number of elementary bodies which entered a host cell in a 1 h time period (Figure 4A). As predicted, C. trachomatis expressing Tarp lacking the actin binding domain (TarpΔABD) demonstrated a significant reduction in host cell invasion compared to wild type C. trachomatis L2 and C. trachomatis harboring pCtSV.Tarp. Surprisingly and in contrast to the in vitro actin polymerization studies, C. trachomatis expressing TarpΔphos displayed a significant reduction in invasion relative to all of the clones analyzed, including C. trachomatis expressing TarpΔABD (Figure 4A). No altered invasion phenotype was observed for C. trachomatis expressing TarpΔPRD or TarpΔFAB1 and 2. In order to examine if the presence of a dominant-negative Tarp might disrupt chlamydial development, growth curves were performed for C. trachomatis expressing TarpΔphos and wild type controls (Figure 4B). No significant changes to chlamydial growth was observed between wild type and C. trachomatis expressing TarpΔphos as measured by the number of EBs harvested from infected cells at 0, 12, 24, 36, and 48 h post infection (Figure 4B and data not shown).

We hypothesized that C. trachomatis expressing TarpΔphos might show a reduction in endogenous Tarp phosphorylation following attachment and entry into host cells as TarpΔphos is likely to compete with endogenous Tarp for translocation into the host cell. TarpΔphos is missing the tyrosine residues known to be phosphorylated by Src family kinases (SFKs) therefore TarpΔphos is not expected to be phosphorylated by host cell tyrosine kinases after translocation into the host cell. In order to test this, wild type C. trachomatis or C. trachomatis harboring the plasmids presented in Figure 3 were added to host cells and protein lysates were generated at 0 and 1 h post infection. Tarp has previously been identified as the predominant phosphotyrosine modified protein of ~150 kDa which is observed in the wild type C. trachomatis infected cells. Interestingly, phosphorylated Tarp was reduced in those bacteria expressing TarpΔphos, but not in the other transformants (Figure 5).

C. trachomatis expressing both endogenous Tarp and TarpΔphos demonstrated a reduction in host cell invasion and Tarp phosphorylation. These results raised the possibility that type three secretion is altered and/or inhibited in C. trachomatis expressing TarpΔphos. To determine whether C. trachomatis expressing TarpΔphos was capable of type three secretion we examined whether endogenous Tarp and plasmid encoded epitope tagged TarpΔphos could be recovered from the soluble fraction following subcellular fractionation of host cells (Figure 6A). The fractionation profile of wild type C. trachomatis infected cells is presented as Supplemental Figure 1 for reference. Endogenous Tarp and epitope tagged TarpΔphos can be distinguished by their unique molecular weights and the presence or absence of the c-myc epitope tag (Figure 3). Protein supernatant samples sequentially obtained from 800, 10,000, and 100,000 × g centrifugal spins indicated that the endogenous Tarp and TarpΔphos effectors co-fractionated and were detectable in fractions that were distinct from intact EBs, which pellet at 10,000 × g (Figure 6B). The endogenous Tarp and TarpΔphos proteins were observed in the 100,000 × g pellet and the 100,000 × g soluble fraction, the latter represents the host cell cytosolic fraction as defined by the presence of the soluble eukaryotic protein GAPDH (Figure 6B). We hypothesize that most of the secreted tarp is restricted to the 100,000 × g pellet due to interactions with actin filaments (short actin filaments pellet at 100,000 × g). The portion of Tarp found in the 100,000 × g supernatant (soluble fraction) may represent those proteins which have not yet associated with host cell molecules of the cytoskeleton. Non-secreted chlamydial antigens identified by the anti-EB and anti-MOMP antibodies were not detected in fractions beyond the 10,000 × g pellet, indicating that these later protein fractions did not contain lysed EBs. It has been shown previously that type three secretion by C. trachomatis is temperature dependent and can be inhibited at low temperatures (Jamison and Hackstadt, 2008). Consistent with this finding, endogenous and mutant Tarp were not detected in the host cell cytosolic fraction of host cells maintained at 4°C during the course of the C. trachomatis infection. Together these data indicate that C. trachomatis expressing TarpΔphos is capable of type three secretion mediated delivery of endogenous and mutant Tarp into the host cell cytosol. Although the data presented in Figure 6B do not provide quantitative measures of WT and TarpΔphos secretion, the experiment indicates that both endogenous and TarpΔphos are actively secreted into the host cell.

Previous studies with purified recombinant Tarp effectors revealed that Tarp multimerization is mediated by a region of ~25 amino acids rich with prolines subsequently called the proline rich domain (PRD) (Jewett et al., 2006). This model suggests that plasmid encoded TarpΔphos and endogenous Tarp may associate following translocation into the cell perhaps resulting in the observed dominant-negative invasion phenotype. To further test this hypothesis, we sought to disrupt the ability of TarpΔphos to form heteromeric complexes with wild type Tarp through genetic deletion of the proline rich domain from the pCtSV TarpΔphos construct. According to our model, a TarpΔphosΔPRD double mutant lacking both the phosphorylation domain and the proline rich oligomerization domain would be unable to associate with endogenous Tarp, rescuing the dominate negative phenotype caused by the TarpΔphos protein. As predicted, C. trachomatis transformants harboring pCtSVTarpΔphosΔPRD demonstrated wild type levels of chlamydial entry after 1 h (80% internalized EBs) and displayed no defect in chlamydial growth (Figure 7 and data not shown).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.