C. trachomatis serovar L2 (434/Bu) was used as the parent strain in these studies. Previously described strains include C. trachomatis Δtarp (Ghosh et al., 2020), ΔtmeB (Mueller et al., 2016), Δtepp (Keb et al., 2021), ΔtmeA (Mueller et al., 2016) (referred to herein as ΔtmeA polar), and tmeA-lx (McKuen et al., 2017) (referred to herein as non-polar ΔtmeA). C. trachomatis ΔtmeAB and ΔtmeAB-tarp were generated via FRAEM (Mueller et al., 2016) and are described below. Mobilization of respective plasmids into C. trachomatis L2 was achieved using CaCl₂-mediated chemical transformation (Wang et al., 2012). Manipulations leveraging fluorescence reporting to generate engineered strains were accomplished according to established protocols (Mueller et al., 2017; Wolf et al., 2019). Chlamydiae were routinely maintained in either HeLa 229 epithelial cell monolayers (CCL-1.2; ATCC) or McCoy cell monolayers (CRL-1696; ATCC). Unless otherwise indicated, all cultures were grown in RPMI 1640 medium containing 2 mM L-glutamine (Life Technologies) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS; Sigma) at 37°C in an environment with 5% CO2 and 95% humidified air. Infections were performed using density gradient-purified EBs (Scidmore, 2005). Chlamydiae were cultivated in the presence of 600 ng/ml penicillin G (PenG; Sigma), 500 µg/ml spectinomycin (Spec; AlfaAesar), 1 µg/ml cycloheximide (Sigma), and 50 ng/ml anhydrotetracycline (aTc) where appropriate.

Ectopic expression of TmeB was accomplished by transformation of WT or ΔtmeB chlamydiae with pCOMP-TmeB. pCOMP-TmeB was generated by mobilization of tmeB with its endogenous promoter into pComA II (Mueller et al., 2016). Amplification was accomplished using ΔtmeA genomic DNA and primers 694pro@NmPgfpF (5’-CGGTTCCTGGCCTTTTGCTGG GTACGGAAATACTATCTCCAGCTCAAAGC-3’) and 695@gfpNmPR (5’- GCCCCGCCC TGCCACTCATCGTTAGATATTCCCAACCGAAGAAGGATCTTCC-3’). For ΔtmeAB, pSUmC expressing gfp and aadA flanked by 2.2 kb of chlamydial flanking DNA was generated. Sequences 5’ of tmeA were amplified from L2 genomic DNA with ctl0063-SalI-S (5’-GCAAAAGAGCTGATCCTCCGTCACTGCAGGTACCGCTCCAGCGTTGCGTATTGTTTGTGG-3’) and ctl0063-SalI-AS (5’-CGTATAATGTATGCTATACGAAGTAGG AATGCCTCCGCCGAAGCAATAACTTTTAATTCC-3’) whereas sequences downstream of tmeB were amplified with ctl0064-SbfI-S (5’-GTATAGCATACATTATACGAAGTTA TGCATTTTCTAATAGGGAAGAGGATAAATAGCGTG-3’) and ctl0064-SbfI-AS (5’-CTTTGATCTTTCTACGGGGTCTGACGCCCGACCATTTTACTTTGGAATAAG TGTGATATC-3’). Amplicons were sequentially cloned into pSUmC using Gibson Assembly of SalI-cut or SbfI-cut plasmid, respectively. The tmeA and tmeB double deletion strain was generated by transformation of C. trachomatis L2 with pSUmC-ΔtmeAB and allelic replacement using FRAEM as described (Mueller et al., 2017). A ΔtmeAB-tarp triple mutant was subsequently generated using lateral gene transfer. McCoy cells were co-infected at an MOI of 5 with a 1:1 ratio of ΔtmeAB and Δtarp. Cultures were maintained without antibiotic selection for two 48 hr passages followed by 4 serial passages in the presence of spectinomycin and penicillin G to select for isolates carrying both the tmeAB and tarp deletions. All Chlamydia strains were clonally isolated as described (Mueller et al., 2017) by 2 sequential limiting dilution passages in 384 plates. Chromosomal deletions were confirmed using specific primers and qPCR as described for tarp (Ghosh et al., 2020), or tmeA and tmeB (Mueller et al., 2016) and whole genome sequencing (SeqCenter) was employed to verify integrity of the remaining genome.

For immunoblot analyses, proteins were concentrated from purified EBs by trichloroacetic acid precipitation. Material was separated on 4 to 15% SDS-PAGE gels (Bio-Rad) and transferred to 0.45 µm PVDF membranes (Millipore). Primary antibodies used include: MOMP (Hower et al., 2009), TmeA (Hower et al., 2009), TmeB (Mueller and Fields, 2015), Tarp (Clifton et al., 2004), Tepp (kindly provided by Dr. Raphael Valdivia, Duke University), Hsp60 (A57-B9; Santa Cruz), N-WASP (30D10; Cell Signaling), and ARPC1 (13C9; Millipore). Where appropriate, immunoblots were probed with anti-phospho-tyrosine (4G10, Millipore) to detect phosphorylated Tarp and Tepp or with anti-Flag (Sigma) to detect flag-tagged proteins. We used peroxidase-conjugated secondary antibodies (Sigma) and Amersham ECL Plus (GE Healthcare UK Limited) detection reagent to visualize proteins.

Co-purification of host proteins with TmeA-FT, TmeB-FT, or Cpn0678-FT was performed essentially as described (Keb et al., 2021). Briefly, HeLa cells were transfected with ectopic expression vectors and seeded in one 6-well plate each. Additional controls included cells expressing eGFP or Cpn0677-FT. Cultures were harvested in NP40 buffer (50 mM Tris pH7.5, 150 mM NaCl, 1 mM EDTA, 0.5% NP-40) as described (Mirrashidi Kathleen et al., 2015) after 24 hrs, and incubated on ice for 1 hr. Soluble fractions were applied to equilibrated Anti-Flag M2 Affinity gel (Sigma) and incubated overnight at 4°C. The resin was washed 3x with lysis buffer and FT proteins were eluted with 3X Flag peptide in PBS. 6x Laemmli buffer was added to eluates prior to separation on SDS-PAGE gels for immunoblotting.

HeLa229 cells were prepared in 24-well plates with 12 mm coverslips, and invasion assays were performed essentially as described (Carabeo et al., 2002). Density gradient purified EBs were used at a multiplicity of infection (MOI) of 20. Where appropriate, HeLa cells were pretreated with 200 µM CK666 (all purchased from Sigma-Aldrich) for 15 min prior to infection and maintained with 200 µM CK666 during infection and subsequent incubation. All infections were carried out on ice with rocking for 1 hr then shifted to 37°C for 45 min. Cultures were then treated with 4% paraformaldehyde, and extracellular or intracellular EBs were differentially labeled with murine LPS-specific or rabbit MOMP (Hower et al., 2009) antibodies, respectively. Secondary antibody conjugated to Alexa-594 (anti-mouse) or Alexa-488 (anti-rabbit) were used for detection. Internal and external chlamydiae were enumerated in 10 fields of view. Percent EB internalization was calculated via the formula [(total EBs –external EBs)/total red EBs] x 100 = percent (%) invasion.

Pyrene actin polymerization assays were performed as previously described (Jiwani et al., 2013). Full-length TmeB was expressed as a GST fusion in pGEX-6p-1, purified to homogeneity over GST affinity column, and eluted by cleavage of the GST-tag with PreScission protease (Sigma). Remaining components were from the commercially available pyrene actin polymerization kit (Cytoskeleton). For polymerization assays, monomeric pyrene-labeled actin was prepared by diluting lyophilized pyrene actin (Cytoskeleton) in 5mM Tris (pH 8.0) 0.2mM CaCl₂ 0.2mM ATP (G buffer) and collection of the supernatant of a 90 minute, 100,000 xg, 4°C spin in a Beckman Optima MAX TL Ultracentrifuge using a TLA 55 rotor (Beckman Coulter). Approximately 30 µg of pyrene-labeled actin was mixed with 1-2µg of indicated proteins (TmeB, VCA, Arp2/3, GST) in a volume of 500 µl for 5 min before the addition of 1/20^th volume of polymerization buffer (500mM KCl, 20mM MgCl₂, 10mM ATP). The reaction (contained in a semi-microcuvette and holder assembly) was monitored for 30 min with a LS 55 Luminescence spectrophotometer equipped with the biokinetic accessory and 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.

Unless otherwise noted, presented data are representative of triplicate experiments where quantitative data were generated from experiments containing triplicate biological replicates. Calculation of standard deviation of the mean and assessment via Student’s t-test with Welch’s correction. Statistical analyses were performed using GraphPad Prism 6 version 6.04 (Graphpad Software Inc).

We developed floxed allelic exchange mutagenesis (FLAEM) to enable marker-less gene deletion in C. trachomatis. This approach ( Figure 1A ) was previously used (Keb and Fields, 2020) to excise a blaM-gfp cassette replacing the tmeA locus in our original ΔtmeA null (Mueller et al., 2016) created by fluorescence-reported allelic exchange mutagenesis (FRAEM). The marker-induced polar effect on downstream tmeB was relieved, resulting in increased TmeB abundance (Keb and Fields, 2020). During the course of our TmeA functional studies, we noticed the new non-polar ΔtmeA (referred to previously as ΔtmeA-lx) strain appeared to exhibit a more robust invasion defect than the parent polar strain. We directly tested for a potential difference by performing side-by-side invasion studies ( Figure 1B ). Compared to WT, polar ΔtmeA reduced invasion efficiency by ca 16% whereas the non-polar strain was reduced by ca 50%. Corresponding protein abundance was examined in material concentrated from gradient-purified EBs ( Figure 1C ). As expected, immunoblot data confirmed loss of TmeA and lower abundance of TmeB in the polar ΔtmeA strain. Interestingly, a pronounced overabundance of TmeB was detected in non-polar ΔtmeA compared to WT EBs. TmeB overabundance was specific to the non-polar ΔtmeA strain since WT levels of TmeB were detected in strains lacking tepp or tarp ( Figure S1 ). These data indicate an inverse correlation between invasion efficiency and TmeB abundance when TmeA is absent.

We next generated a strain containing a complete deletion of both tmeA and tmeB to address whether the invasion defect manifested by ΔtmeA strains was due solely to the absence of TmeA. C. trachomatis L2 was transformed with pSUmC-ΔtmeAB, and FRAEM was employed to replace the entire coding region for tmeA and tmeB with a gfp-aadA cassette. Gene deletion was first confirmed in a clonal isolate via qPCR ( Figure S2A ) and then via immunoblot using specific antibodies to probe EB lysates ( Figure 2A ). Neither TmeA nor TmeB were detected whereas co-secreted effectors Tarp and Tepp remained unchanged compared to WT. Tarp is required for efficient invasion, and a triple mutant was generated using lateral gene transfer. Co-infections containing Spec^r ΔtmeAB and PenG^r Δtarp (Ghosh et al., 2020) were cultivated in the presence of both antibiotics to select a recombinant strain carrying both mutations. Respective deletions of a clonal isolate were confirmed via qPCR ( Figure S2B ). Subsequent immunoblot ( Figure 2B ) analysis confirmed the loss of TmeA, TmeB, and Tarp in a single strain. Invasion efficiency of each strain was compared to WT ( Figure 2C ). While levels of ΔtmeAB invasion did not differ from WT, the additional loss of Tarp manifested a significant decrease in internalization efficiency. These data raise the possibility that TmeA, but not Tarp, is required for efficient invasion only in the presence of TmeB.

Non-physiological levels of TmeB in non-polar ΔtmeA also raised the possibility that TmeB could be directly detrimental in the absence of co-expressed TmeA. C. trachomatis WT and ΔtmeB were transformed with pCOMP constitutively expressing TmeB to address this possibility. A 5-10 fold increase in TmeB abundance was evident via immunoblot analysis of EB lysates ( Figure 3A ). These strains were then assayed for invasion efficiency after allowing 30 min for internalization ( Figure 3B ). As shown previously (Keb et al., 2021), loss of tmeB did not interfere with invasion efficiency compared to WT. The presence of excess TmeB in WT and ΔtmeB, however, did correlate with decreased invasion efficiency. These data implicate non-physiological levels of TmeB as a contributing factor to the observed impairment of chlamydial entry into epithelial cells.

Possibilities for TmeB-specific interference with invasion include indirect effects such as competition for the secretion chaperone Slc1 within chlamydiae or more directly by manifesting a dysregulated activity within host cells. To test the former, HeLa cultures were infected for 1 hr with WT or ΔtmeB strains alone or ectopically expressing TmeB. Material from whole-cell lysates was probed with phospho-tyrosine-specific antibodies to detect translocated Tarp and Tepp ( Figure S3 ). Both were readily detected in all strains, indicating that negative impact of excess TmeB likely does not manifest by interfering with type III secretion. Given that operon-encoded genes are often functionally related, we next reasoned that TmeB could have an activity related to TmeA. HeLa cells were transiently transfected with plasmids encoding Flag-tagged TmeA or TmeB. HeLa cells were also nucleofected with plasmids expressing eGFP as a mock control. Plasmids encoding Flag-tagged C. pneumoniae Cpn0677 or Cpn0678, which are encoded by tmeA-syntenic genes found in C. pneumoniae, were used as specificity controls. Flag-tagged proteins were immunoprecipitated under mild lysis conditions using Flag-beads, and material was probed in immunoblots for N-WASP, which is a molecular target of TmeA ( Figure 4A ). We also probed for the Arp2/3 complex using antibodies specific for the ARPC1 subunit. As expected, all Flag-tagged proteins were recovered and neither N-WASP nor ARPC1 co-precipitated with Cpn0678. A cross-reactive band was detected in the GFP control with Flag antibodies, but neither N-WASP nor ARPC1 were detected. N-WASP was detected in the presence of TmeA and C. pneumonia Cpn0677. We also detected ARPC1 which presumably co-precipitates with the activated N-WASP complex. N-WASP did not co-precipitate with TmeB, yet ARPC1 was detected, raising the possibility that TmeB targets the Arp2/3 complex. We tested this hypothesis by assessing comparative invasion in the presence or absence of the pharmacologic Arp2/3 inhibitor CK666 ( Figure 4B ). We chose a CK666 dose sufficient to inhibit WT invasion ca. 50%. HeLa cells were infected with C. trachomatis WT, ΔtmeB or ΔtmeB+TmeB with or without 200 µM CK666 and processed for internalization counts at 30 min. Consistent with previous studies, cultivation with CK666 significantly decreased WT invasion. This was also detected for ΔtmeB. Interestingly, CK666 failed to additionally inhibit ΔtmeB+TmeB, indicating that the negative impact was not synergistic. Instead, the negative effect of over-expressed TmeB appears to be manifested through Arp2/3. In aggregate, these data suggest that TmeB targets the host Arp2/3 complex during invasion in an inhibitory fashion.

The pyrene-based actin fluorescence assay is a well-established approach to functionally investigate factors impacting the state of actin polymerization. Indeed, we have leveraged this approach to demonstrate Tarp-dependent actin polymerization (Jewett et al., 2006) and activation of N-WASP-dependent polymerization by TmeA (Keb et al., 2021). The verprolin homology, central, acidic (VCA) domain of N-WASP was used to activate Arp2/3 and test the impact of TmeB on actin polymerization. Purified preparations ( Figure 5A ) of protein components were incubated alone or in combination with pyrene-conjugated actin. As expected, the combination of Arp2/3 and VCA increased polymerization rates, as indicated by fluorescence intensity, compared to actin alone ( Figure 5B ). Addition of control GST did not interfere with this polymerization. In the presence of pure TmeB, the fluorescence intensity profile was similar to that of actin alone. TmeB-mediated inhibition of actin polymerization was dose-dependent as increasing levels of TmeB resulted in progressively dampened actin polymerization rates ( Figure 5C ). This decrease in actin polymerization depended on Arp2/3 since, in the absence of Arp2/3 and VCA, TmeB alone had no impact on polymerization rates ( Figure S4 ). In aggregate, these data indicate that one function of TmeB is to antagonize branched actin polymerization by interfering with Arp2/3 activity.

The VCA domain of N-WASP exhibits constitutive activity, and we previously used a version of N-WASP (residues 151-501) to show that TmeA activates N-WASP to promote Arp2/3-dependent actin polymerization (Keb et al., 2021). We therefore examined whether TmeB could interfere with actin polymerization stimulated by the TmeA-N-WASP complex by combining these proteins in the pyrene assay ( Figure 6A ). As expected, addition of TmeA-N-WASP with Arp2/3 enhanced fluorescence kinetics and intensity. Inclusion of TmeB interfered with this activation and slowed the overall polymerization rate. We next tested whether TmeB acts synergistically with CK666 ( Figure 6B ). Addition of low levels (17 µM) of CK666 to VCA + Arp2/3 resulted in inhibition of actin polymerization as expected. Inclusion of TmeB shifted the fluorescence profile further downward indicating a synergistic effect between CK666 and TmeB. This additive effect was not detected at a saturating concentration (100 µM) of CK666 ( Figure S4B ). Collectively, these data suggest that CK666 and TmeB-mediated inhibition of Arp2/3 may manifest via a similar mechanism.