HeLa (human cervical epithelial carcinoma cells) cells used in this study were kindly provided by the Institute of Dermatology (PUMC, Nanjing, PRC). Chlamydia muridarum strains including the wild type C. muridarum Nigg strain (WT), the plasmid free (CMUT3), the intact plasmid transformant (Intact), the pgp3 deletion mutant (Δpgp3) [from Dr. Guangming Zhong's lab at the University of Texas Health Science Center at San Antonio, USA] were propagated, purified, aliquoted, and stored as described previously in the reference (Zhong et al., 2001). The new Pgp3 domain deletions were modified from the intact plasmid transformant as described below. For chlamydial infection, cells grown in 24-well plates with or without coverslips, 6-well plates or flasks containing DMEM (Gibco, New York, USA) with 10% fetal bovine serum (FBS, Institute of Hematology, CAMS &PUMC, Tianjin, China) at 37°C in an incubator supplied with 5% CO2 were inoculated with chlamydial organisms as described previously (Zhong et al., 2001).

For making pgp3 domain deleted mutants, primers listed in Table 1 were used to amplify DNA fragments lacking different pgp3 domain from the plasmid pGFP::CM by PCR using AccuPrime pfx SuperMix (Life technologies, Grand Island, NY). The desired PCR products were fused to produce the appropriate plasmids using the in-fusion HD cloning kit as described (Liu et al., 2014a). Plasmids were extracted from bacterial colonies with GFP and the extracted plasmids were partially digested by BamHI and XhoI. Plasmid with the desired fragments after digestion was fully sequenced and transformed into E. coli K12 ER2925 for amplification. The amplified plasmids designated as pGFP::CM pgp3Δn, pGFP::CM pgp3Δm or pGFP::CM pgp3Δc were used for transforming chlamydial organisms.

The plasmid of pGFP::CM pgp3Δn, GFP::CM pgp3Δm or pGFP::CM pgp3Δc was introduced into the plasmid-free C. muridarum strain CMUT3 in the form of a purified EB by following the protocol published previously (Liu et al., 2014a). The organisms of CMUT3-pGFP::CM pgp3Δn, CMUT3-pGFP::CM pgp3Δm or CMUT3-pGFP::CM pgp3Δc were plaque-purified as described previously (Zhong et al., 2001) for in vitro and in vivo experiments as described below.

An immunofluorescence assay was used to detect Pgp3 and C. muridarum by a triple staining technique as described previously (Gong et al., 2013). Briefly, infected HeLa cells monolayer grown on coverslips were fixed with 4% paraformaldehyde for 30 min at room temperature, followed by permeabilization with 2% (wt/vol) saponin (Sigma) for an additional 60 min and then blocking. The cell samples were incubated with antibody and chemical staining. Hoechst (blue; Sigma) was used to mark DNA. Rabbit antibodies against C. muridarum plus secondary antibody conjugated with Cy2 (green; Abnova) was used to mark chlamydial inclusions. Mouse primary antibodies against Pgp3 protein in combination with a secondary antibody conjugated with Cy3 (red; Abnova) were used to mark the Pgp3 protein. Immunofluorescence images were acquired by using a confocal laser scanning microscope (Leica, Germany) and processed using Leica confocal software.

Plaque size assay was carried out for evaluating in vitro growth properties of C. muridarum organisms as described previously (Huang et al., 2015). Briefly, Chlamydia muridarum organisms were inoculated onto McCoy cells monolayer in 12-well plates and centrifuged at 1,200 rpm for 1 h at room temperature (RT). Then infected cells were cultured with overlay medium (1 × Dulbecco's modified Eagle's medium, 10% fetal bovine serum, 1 μg/ml cycloheximide, and a final concentration of 0.55% of agarose). The cells were allowed to incubate at 37°C in an atmosphere of 5% CO2 for 5 days before stained with 0.03% Neutral Red for 1 h at RT. After taking pictures, the diameters of plaques were measured with the custom MATLAB program plaque Detector, which can be freely accessed at http://www.mathworks.com/matlabcentral/fileexchange/48860-plaque-detector.

All C. muridarum organisms were inoculated onto HeLa monolayers in 6-well plates at an MOI of 0.8. The infected cells were harvested at the 24th h after infection. The cells before and after infection were lysed with 0.1% sodium dodecyl sulfate (SDS) and used as PCR templates for titrating the changes of genomic copy number using primers designed from16s rRNA gene. The sequences of primers were listed as follows: P1 5′ cgcctgaggagtacactcgc 3′, P2 5′ ccaacacctcacggcacgag 3′. The experiment was repeated three times with duplicate in each.

The wild type C. muridarum Nigg strain or plasmid-free C. muridarum CMUT3 with or without transformation with the full plasmid or plasmids with pgp3 different domain deletion was used to infect female C3H/HeJ mice or CBA/J intravaginally with 2 × 10⁵ inclusion-forming units (IFUs). To increase mouse susceptibility to the infections, each mouse was injected with 2.5 mg medroxyprogesterone (Depo-Provera; Pharmacia Upjohn, Kalamazoo, MI) subcutaneously 5 days prior to the infection.

To monitor live organism shedding from lower genital tract, vaginal swabs were taken on different days after infection. To quantitate live chlamydial organisms, each swab was dissolved in 500 μl of ice-cold SPG and vortexed with five glass beads, and the chlamydial organisms released into the supernatants were titrated on HeLa cell monolayers in duplicate as described previously (Liu et al., 2014b). The total number of IFUs per swab was calculated based on the number of IFUs per view, the number of views per coverslip, dilution folds, inoculation doses and total sample volumes. The calculated total number of IFUs/swab was converted into log₁₀, and the log₁₀ IFUs were used to calculate the mean and standard deviation at each time point. These animal experiments were carried out according to the principles of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Ethics Committee of Tianjin Medical University General Hospital.

On the 60th day after infection, mice were sacrificed to harvest their entire genital tracts including the vagina, the uterus, the oviduct, and the ovary for evaluating tissue pathology. Before removing the genital tract tissues from the mice, an in situ gross examination was performed for evidence of oviduct hydrosalpinx or any other related abnormalities. The genital tract tissues were isolated entirely and laid on a blue background to acquire images. The oviduct hydrosalpinx were visually scored based on their dilation sizes using a scoring system as described previously (Chen et al., 2014). Mice with hydrosalpinx on either side of the oviducts were determined to be hydrosalpinx positive and the severity of hydrosalpinx was scored based on the following criteria: No oviduct dilation found with a stereoscope inspection is defined as no hydrosalpinx and a score of zero (0); Hydrosalpinx is visible only under stereoscope but not naked eyes (1); Hydrosalpinx is visible with naked eye but the size is smaller than the ovary (2); The size of hydrosalpinx is equal to the ovary (3) or larger than the ovary (4). Scores from both sides of the oviducts from the same mouse were combined as the total gross pathology score for that mouse. Both the incidence and severity scores of oviduct hydrosalpinx were statistically analyzed between mice infected with different C. muridarum organisms. The researchers who scored the pathology were blinded to the experimental groups.

For the observation of histological pathology, the isolated mouse genital tract tissues were fixed in 10% neutral formalin and embedded in paraffin and serially sectioned longitudinally. The sections were stained with hematoxylin and eosin (H&E) as described before (Liu et al., 2014b). The H&E stained sections were observed under microscope for severity of inflammation and pathologies based on the modified schemes established previously (Murthy et al., 2011). Scoring for inflammatory cell infiltration was as follows: 0, no significant infiltration; 1, infiltration at one single focus; 2, infiltration at two to four foci; 3, infiltration at more than four foci; 4, confluent infiltration. For observing oviduct dilation and inflammatory infiltration, image from each mouse was taken under a 40X and 100X objective lens.

Quantitative data including the diameters of plaque sizes was analyzed using Student's t-test while the overall IFU shedding in form of Log10 was analyzed using Kruskal-Wallis test. All semi-quantitative data including the pathology scores was analyzed using Mann–Whitney U rank-sum test. The qualitative data including incidence rates was analyzed using Fisher's exact test.

We generated mutants harboring deletion constructs of either the pgp3 N-terminus, middle domain, or C- terminus by modifying pGFP::CM, resulting in pGFP::CM pgp3Δn, pGFP::CM pgp3Δm, or pGFP::CM pgp3Δc, respectively (Figure 1A). These plasmids were transformed into the C. muridarum plasmid-free clone CMUT3, and the resulting strains were cultured with L929 cell monolayers for 12 h without ampicillin and subsequently with ampicillin for 20 h. After four to five passages, GFP-positive inclusions were enriched, and a single clone of each transformant was selected by plaque assay and purified. The CMUT3 transformants obtained were named CMUT3-pGFP::CM pgp3Δn, CMUT3-pGFP::CM pgp3Δm, and CMUT3-pGFP::CM pgp3Δc (Figure 1B), referred to hereafter as pgp3Δn, pgp3Δm, and pgp3Δc, respectively.

We assessed Pgp3 expression in the three stable transformants. C. muridarum Nigg (wild type [WT]), CMUT3, and CMUT3 transformed with pGFP::CM (CMUT3-pGFP::CM, intact) or with deletion of full-length pgp3 (CMUT3-pGFP::CMΔpgp3 [Δpgp3]) were used as the controls. The plasmid-free CMUT3, Δpgp3, pgp3Δm, and pgp3Δc lacked the corresponding signals of Pgp3 (Figure 2, Figure S1A in the Supplemental Material). The failure to detect mutated Pgp3 in pgp3Δm and pgp3Δc organisms may be due to the absence of their epitopes, for the mutated Pgp3 expressions were detectable at transcriptional level (Figure S2). Besides, the His-tagged mutated Pgp3 expressions were verified in E. coli expression system (data not shown). The Pgp3 protein in pgp3Δn was detected only in chlamydial inclusion bodies, but not in host cell cytoplasm (Figure 2, Figure S1). We further performed a Western blot of native gel to compare the protein levels of Pgp3 in the three Pgp3 mutants, using intact organisms as positive control and Δpgp3 as a negative control. The result was consistent with that from immunofluorescence staining (Figure S1). Therefore, Pgp3 with deletion of its N-terminus was not secreted into the host cell cytoplasm, whereas the middle and C-terminal domains of Pgp3 are essential for polyclonal antibody recognition.

We evaluated the effect of Pgp3 domain deficiency on induction of hydrosalpinx by C. muridarum. C3H/HeJ mice were infected with WT, CMUT3, plasmid-competent (intact), Δpgp3, pgp3Δn, pgp3Δm, or pgp3Δc C. muridarum intravaginally. Sixty days after infection, the mice were euthanized and their genital tracts were harvested for assessment of oviduct pathology (Figure 3A). CMUT3 and Δpgp3 did not induce hydrosalpinx; in contrast, 80% of the C3H/HeJ mice infected with WT or intact C. muridarum developed severe hydrosalpinx (severity score 4.17 ± 2.27), which is consistent with previous reports (Lei et al., 2014; Liu et al., 2014b). However, pgp3Δn and pgp3Δm failed to induce this pathology in the mouse oviduct, and pgp3Δc induced hydrosalpinx in only 20% of C3H/HeJ mice (severity score 0.2 ± 0.6). Therefore, we then infected CBA/J mice, which are more susceptible to C. muridarum infection (Chen et al., 2014), with the above-mentioned C. muridarum strains. CMUT3 induced hydrosalpinx in 20% of the mice. Δpgp3, pgp3Δn, and pgp3Δm did not induce hydrosalpinx in CBA/J mice, whereas 40% of the mice infected with pgp3Δc showed mild pathology in the oviduct (severity score 0.8 ± 1.3). Therefore, the Pgp3 C-terminus is less important for C. muridarum pathogenicity than are the N-terminal and middle domains. However, deletion of any of the three Pgp3 domains significantly reduced the pathogenicity of C. muridarum in the mouse genital tract.

Since the chlamydial infectivity seems to be correlated with its ability to induce oviduct pathology, we evaluated the sizes of plaques formed by the three transformants, which can reflect chlamydial infectivity, growth rate, and exit from the cell in vitro. The plaques formed by CMUT3 were significantly smaller than those formed by WT or intact C. muridarum (Figures 4A,B). The plaques formed by each of the three Pgp3-domain-deficient transformants were smaller than those formed by WT or intact C. muridarum (Figure 4). Thus, deletion of any of the three Pgp3 domains reduced the in vitro infectivity, growth rate, and exit from the cell of C. muridarum. Additionally, qPCR was used to measure the change of genomic copy after a single step infection. It displayed that the growth rates of Pgp3-domain-deficient transformants were much lower than those of WT or intact C. muridarum (Figure 4C).

In addition, live organisms shedding of C. muridarum from the LGT of infected mice was evaluated (Figure 5) as an indirect measure of survival in vivo. As expected, C3H/HeJ mice infected with CMUT3 C. muridarum showed significantly reduced shedding from the LGT at days 3, 7, 14, and 21 compared with mice infected with intact C. muridarum. The Δpgp3 C. muridarum resulted in significantly reduced shedding from the LGT at 3, 7, and 14 days postinfection. Importantly, pgp3Δm and pgp3Δc exhibited reduced C. muridarum shedding from the LGT at 3, 7, and 14 days postinfection, whereas pgp3Δn showed reduced shedding at 3 and 7 days post-infection.

The duration of vaginal shedding was slightly longer in CBA/J than in C3H/HeJ mice, possibly because CBA/J mice are more susceptible to C. muridarum infection. Shedding of live organisms by mice infected with CMUT3 or Δpgp3 C. muridarum was decreased at 3, 7, 14, and 21 days postinfection. Moreover, CBA/J mice infected with the three Pgp3-domain-deficient transformants showed significantly reduced shedding of live organisms. Shedding of pgp3Δn was reduced at 3, 7, 14, 21, and 28 days postinfection, and shedding of pgp3Δm and pgp3Δc was reduced at 3, 7 and 14 days postinfection. Therefore, each of the three Pgp3 domains is required for infection of the mouse LGT by C. muridarum.

The chlamydial infectivity and host inflammatory infiltrations in oviduct are two factors contributing to oviduct hydrosalpinx. On this basis, we supposed that the three Pgp3-domain-deficient transformants had impaired infectivity in mouse genital tract. However, the host immune response stimulated by chlamydia in oviduct still remains unknown. To settle this issue, we assessed inflammatory cell infiltration in the oviduct tissue of C3H/HeJ and CBA/J mice at 60 days post-intravaginal infection, as this is essential for hydrosalpinx induction by C. muridarum (Chen et al., 2014; Lei et al., 2014; Liu et al., 2014b; Huang et al., 2015). As expected, intact and WT C. muridarum induced markedly greater inflammatory cell infiltration in oviduct tissue compared with CMUT3 (Figure 6), which is consistent with our previous report. Interestingly, the transformants deficient in either of the three Pgp3 domains or full-length Pgp3 induced a lower level of inflammatory cell infiltration in the oviduct. In contrast, pgp3Δc induced slightly greater inflammatory cell infiltration (severity score 3.05 ± 1.49 for C3H/HeJ and 3.53 ± 2.24 for CBA/J mice) compared with CMUT3. These results were consistent with the gross pathology, indicating that pgp3Δc-induced hydrosalpinx may be due, in part, to relatively robust inflammatory cell infiltration in oviduct tissue.

WT and intact C. muridarum induced considerably greater oviduct dilation compared with CMUT3, Δpgp3, pgp3Δn, pgp3Δm, or pgp3Δc C. muridarum. These findings are consistent with the gross pathology (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.