Loss of the conserved “cryptic” plasmid from C. trachomatis and C. muridarum is pleiotropic, resulting in reduced innate inflammatory activation via TLR2, glycogen accumulation and infectivity. The more genetically distant C. caviae GPIC is a natural pathogen of guinea pigs and induces upper genital tract pathology when inoculated intravaginally, modeling human disease. To examine the contribution of pCpGP1 to C. caviae pathogenesis, a cured derivative of GPIC, strain CC13, was derived and evaluated in vitro and in vivo. Transcriptional profiling of CC13 revealed only partial conservation of previously identified plasmid-responsive chromosomal loci (PRCL) in C. caviae. However, 2-deoxyglucose (2DG) treatment of GPIC and CC13 resulted in reduced transcription of all identified PRCL, including glgA, indicating the presence of a plasmid-independent glucose response in this species. In contrast to plasmid-cured C. muridarum and C. trachomatis, plasmid-cured C. caviae strain CC13 signaled via TLR2 in vitro and elicited cytokine production in vivo similar to wild-type C. caviae. Furthermore, inflammatory pathology induced by infection of guinea pigs with CC13 was similar to that induced by GPIC, although we observed more rapid resolution of CC13 infection in estrogen-treated guinea pigs. These data indicate that either the plasmid is not involved in expression or regulation of virulence in C. caviae or that redundant effectors prevent these phenotypic changes from being observed in C. caviae plasmid-cured strains.
Citation: Frazer LC, Darville T, Chandra-Kuntal K, Andrews CW Jr, Zurenski M, et al. (2012) Plasmid-Cured Chlamydia caviae Activates TLR2-Dependent Signaling and Retains Virulence in the Guinea Pig Model of Genital Tract Infection. PLoS ONE 7(1): e30747. doi:10.1371/journal.pone.0030747
Editor: Kathleen A. Kelly, University of California Los Angeles, United States of America
Received: July 22, 2011; Accepted: December 28, 2011; Published: January 24, 2012
Copyright: © 2012 Frazer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The Hartwell Foundation (http://www.thehartwellfoundation.org/) to TD, National Institutes of Health via grants R21 AI083657 to COC, R01 AI070693 to RB, and R01 AI064749 to RI. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Chlamydiaceae are gram-negative obligate intracellular pathogens that infect ocular, genital and respiratory tissues in both humans and animals. The genomes of chlamydial species are highly conserved, which likely reflects the specific requirements of intracellular pathogens for survival and the limited opportunity for genetic exchange with other bacterial pathogens within this intracellular niche (reviewed by Stephens et al. ). This high degree of genetic-relatedness extends to carriage of a 7.5 kb “cryptic” plasmid by Chlamydia trachomatis (human), C. muridarum (mice), C. psittaci (birds), C. felis (cats), strains of C. pneumoniae that do not infect humans and C. caviae (guinea pigs). Plasmid-deficient C. trachomatis isolates are extremely rare, leading to speculation on the importance of the plasmid for chlamydial pathogenesis .
We developed a protocol for the derivation of plasmid-deficient chlamydiae and demonstrated its efficacy using C. muridarum . The plasmid-cured derivative of C. muridarum Nigg, strain CM972, displays distinctive phenotypic changes in vitro and in vivo when compared with its parent. CM972 is less infectious in vitro  and this is associated with significantly reduced chlamydial load in the oviducts of intravaginally inoculated mice , . In addition, CM972 lacks the ability to accumulate glycogen within the inclusion, a property shared by naturally occurring plasmid-deficient isolates , . CM972 also did not cause oviduct pathology or signal via Toll-like receptor 2 (TLR2) , which is important for the development of oviduct pathology in this model . Importantly, primary infection with CM972 prevented the development of pathology upon secondary challenge with wild-type Nigg. Recently, we demonstrated that these phenotypic changes are conserved in plasmid-cured C. trachomatis and identified a number of PRCL that, in addition to plasmid-encoded gene products, are candidate effectors of these virulence properties .
C. caviae is a natural pathogen of guinea pigs that causes inclusion conjunctivitis and respiratory infection in newborns. The guinea pig has been used to model sexual transmission of chlamydial infection from males to females  and to study genital tract infection and disease pathology in females , , . Nevertheless, C. caviae is more distant genetically from C. trachomatis and C. muridarum and differs phenotypically from them in several respects including intrinsic resistance to sulphonamides , non-fusing inclusions  and an inability to accumulate glycogen . We sought to examine the role of the conserved cryptic plasmid more broadly within the Chlamydiacae by examining the in vitro and in vivo consequences of curing the plasmid from C. caviae. We hypothesized that plasmid-deficient C. caviae would not activate TLR2 and would fail to cause oviduct disease in the guinea pig model. However, in contrast to plasmid-cured C. muridarum and C. trachomatis, plasmid-cured C. caviae strain CC13 signaled via TLR2 in vitro and elicited cytokine production in vivo similar to wild-type C. caviae. Furthermore, pathology induced by guinea pig genital tract infection with this strain was not reduced although we observed more rapid resolution of infection with CC13 in estrogen-treated guinea pigs. These data indicate that the association of the chlamydial plasmid with virulence is not universally conserved among chlamydial species.
Female 20-week-old outbred Hartley strain (Hilltop Labs, Scottdale, PA) guinea pigs were used for experiments. Guinea pigs were given food and water ad libitum in an environmentally controlled room with a cycle of 12 h of light and 12 h of darkness. All animals were determined to be seronegative for anti- C. caviae antibodies prior to infection. All animal experiments were pre-approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh Medical Center under protocol # 0807981.
Strains, cell lines and culture conditions
C. caviae, guinea pig inclusion conjunctivitis (GPIC) strain, was provided by Dr. Roger Rank and plaque-purified before use. Chlamydiae were cultured in L929 fibroblasts. Cells were infected at an approximate MOI of 0.5–1 before being centrifuged for 1 hour at 37°C. The cell culture medium was then removed and replaced with 1 X Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated FBS, gentamicin (20 µg ml−1) and 0.1 µg ml−1 cycloheximide. Infected cells were harvested into sucrose phosphate glutamate buffer at 40 hours post infection, sonicated, and maintained at −80°C. Bacteria were subsequently titrated by either the plaque assay  or as inclusion forming units (IFU)  using a genus-specific fluorescently tagged anti-chlamydial LPS monoclonal antibody (Biorad, Hercules, CA). Evaluation of plaquing efficiency under different culture conditions (e.g. with or without centrifugation) was achieved by the titration of individual strains by plaque-forming assay, with indicated modifications, in parallel with titration to estimate IFU. The efficiency of plaquing (EOP) was calculated as PFU ml−1 /IFU ml−1.
Total RNA was isolated from chlamydia-infected L929 cells 24 hours post-infection (MOI~1) using an RNeasy RNA isolation kit (Qiagen, Valencia, CA). For quantitative PCR analysis, the RNA (2 µg) was pretreated with DNase I (Ambion) and cDNA prepared using an iScript cDNA synthesis kit (BioRad, Hercules, CA) according to the manufacturer's instructions. Primer pairs for amplification of 16S rRNA and the genes identified in this study (Table 1) were selected using Beacon Designer 2.13 (Premier Biosoft International, Palo Alto, CA). Primers were purchased from Integrated DNA Technology (Coralville, IA). Real time PCR reactions were carried out using iQ Sybergreen supermix (BioRad) in a BioRad iCycler using a two step reaction, 95°C for 10 seconds, 55°C for 1 minute for a total of 40 cycles. Melt curve analysis showed that the accumulation of SYBR green-bound DNA was gene specific and not caused by formation of primer dimers. Transcripts from the 16S rRNA gene of C. caviae served as an endogenous reference and data were analyzed by the 2−ΔΔCT method  using BioRad proprietary software. Each sample was assayed in triplicate and each experiment was performed at least twice. Statistical significance was determined by using Student's t test. A p value of <0.05 was considered significant.
Table 1. PCR primer sets used in this study.doi:10.1371/journal.pone.0030747.t001
In vitro analysis of TLR signaling
The following cell lines were examined: HEK293 cells stably expressing either TLR2 or TLR4/MD2 . X-ray-irradiated preparations of chlamydiae were also assayed using HEK-TLR2 cells transfected with an NF- κB reporter plasmid as previously described . In addition, in vitro infection was performed in murine bone marrow-derived dendritic cells (BMDDCs) cultured following the procedure of Inaba et al. . The TLR2 agonist, Pam3Cys-Ser-(Lys)4 (Axxora, San Diego, CA), the TLR4 agonist LPS (Sigma, St. Louis, MO), and human rTNF-α (R&D Systems, Minneapolis, MN) were used as positive control stimulants. Cells were plated in 24-well tissue culture dishes at a density of ~105 cells/well. Infections were conducted by overlaying cells with a multiplicity of infection of 1 or 3. Cells were incubated for 24 hrs at 37°C, 5% CO2. Supernatants from the HEK cells were harvested and assayed for IL-8 (DuoSet, R&D Systems). The BMDDC supernatants were assayed for IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12p70, IL-13, IL-17, G-CSF, GM-CSF, KC, MIP-1α, TNFα, IFNγ, and MCP-1 via Multiplex bead assay (Millipore Billerica, MA). IL-2, IL-4, IL-5, IFNγ, IL-9, IL-12p70, IL-13, MCP-1 and IL-17 were not detected above media (data not shown). BMDDC data points represent the mean levels of two independent experiments ± SD.
Guinea pigs were inoculated intravaginally with 106 IFU of either GPIC or CC13 suspended in 30 µl SPG intravaginally. Estradiol administration has been demonstrated to potentiate infection and pathology in the guinea pig model  so selected experimental groups were dosed with 1 mg sesame oil-emulsified β-estradiol-3-benzoate (10 mg/ml) (Sigma) daily where indicated, beginning 7 days before infection and continuing until sacrifice. Three separate experiments were conducted with groups of estrogen-treated animals sacrificed on days 9 or 30 and groups of untreated animals sacrificed on day 30. The kinetics of lower genital tract infection was monitored via culture of cervical swabs on L929 cells.
In vivo cytokines
Guinea pig genital tract secretions were collected via vaginal sponges as described previously . Sponges were stored at −70°C until they were eluted individually in 0.5 ml of Eagle minimal essential medium and assayed via ELISA for IL-8 (Human IL-8 DuoSet, R&D Systems).
IgG1, and IgG2a, and IgA antibody to C. caviae in serum was measured by an ELISA as described previously  using gradient purified GPIC elementary bodies as antigen and goat, anti-guinea pig-Ig antibodies obtained from AbdSerotech (Oxford, UK).
Guinea pigs were sacrificed at day 9 or day 30 post infection, and the entire genital tract was removed en bloc, fixed in 10% buffered formalin, and embedded in paraffin. Longitudinal sections (4 µm) were stained with hematoxylin and eosin and evaluated by a pathologist blinded to the experimental design. Each anatomic site (exocervix, endocervix, uterine horn, and oviduct) was independently assessed for the presence of acute inflammation (neutrophils), chronic inflammation (lymphocytes/monocytes), plasma cells, and erosion of the mucosa. Right and left uterine horns and right and left oviducts were evaluated individually. A four-tiered semiquantitative scoring system was used to quantify the inflammation as previously described . Oviduct dilatation scores also include a subjective assessment of flattening of the oviduct epithelial plicae and destruction of the oviduct mucosa.
Statistical comparisons between GPIC and CC13-infected guinea pigs for levels of infection and cytokine production over the course of infection were made using a two-factor (days and strain) RM ANOVA. A post-hoc Tukey test was used as a multiple comparison procedure. Kaplan-Meier survival analysis was used to compare the durations of infection over time. The Fisher-exact test was used for determination of significant differences in frequency of pathological characteristics between groups. One-way ANOVA on ranks was used to determine significant differences in the pathological data between groups. Statistical tests were performed using SigmaStat software with p<0.05 considered significant.
Plasmid-deficient C.caviae CC13 displays normal infectivity
L929 cells infected with C. caviae GPIC were treated with novobiocin (62.5 µg/ml) and plated in a plaque assay as previously described for curing of the plasmid from C. muridarum . Individual plaques were selected at random and each isolate was plaque-purified twice more in order to obtain clones. A total of 28 clones were screened by PCR using primers directed against pgp1, a gene encoded by the resident plasmid. Three plasmid-deficient clones were identified and one was selected for further analysis and designated strain CC13. Amplification and sequencing of a portion of the 16S rRNA gene from CC13 (Fig. 1A, Table 1) confirmed its genetic lineage, and Southern hybridization using a 1.58 Kb probe that spans the open reading frames encoding pGP4-6 of pCpGP1 confirmed the absence of the plasmid from this strain (Fig. 1B). PCR amplicons representing each of the plasmid-encoded open reading frames were detected in C. caviae GPIC only (Fig. S1).
Figure 1. C. caviae CC13 is plasmid-deficient but displays no alteration in plaque size or efficiency.
(A) Equal amounts of template were added to PCR reactions with primers directed against the pgp1 coding sequence expressed by pCpGP1 or the C. caviae genome and amplified products were analyzed on 2% agarose gels. (B) Genomic DNA was isolated from GPIC or CC13, digested with SphI and separated on a 1% agarose gel. The DNA was transferred to a nylon membrane and probed for the presence of pCpGP1 DNA using a ~1.6 Kb probe directed against a region spanning pGP4-6. Plasmid pCpGP1 was isolated from GPIC and digested with Sph1 to yield a single, linear fragment of ~7.9 kb which was used as a positive control for hybridization. (C) Plaques formed by CC13 do not differ from those formed by GPIC. C. caviae-infected L929 monolayers were incubated for 5 days at 37 C, 5% CO2, then the media overlay was removed and the plaques visualized using crystal violet.doi:10.1371/journal.pone.0030747.g001
C. caviae GPIC plaques were visible after 4–5 days incubation, and unlike plasmid-cured C. muridarum CM972  or C. trachomatis CTD153 , no reduction in plaque size was observed for CC13 (Fig. 1 C). Furthermore, plaquing efficiency (EOP) in the absence of centrifugation was similar for GPIC (EOP = 2.77x10−2±1.94x10−2) and CC13 (EOP = 2.59x10−2±2.45x10−2), which contrasted with the reduced efficiency we previously observed for plasmid-cured C. muridarum  and C. trachomatis .
Plasmid-responsive transcription by chromosomal loci is partially conserved in C. caviae
Comparison of the transcriptional profiles of plasmid-cured strains of C. muridarum and C. trachomatis revealed a conserved group of plasmid-responsive chromosomal loci (PRCL) whose transcription was altered in the absence of the plasmid . To determine if similar transcriptional changes have occurred in the plasmid cured CC13 we first confirmed that the absence of the plasmid does not impact the growth rate of CC13 when compared to GPIC (Fig. 2A). Furthermore, RT-PCR confirmed that transcripts of all plasmid-encoded ORFs could be detected by 24 hours after infection (Fig. 2B). Interestingly, both strains resembled C. muridarum Nigg  in developmental profile because new GPIC and CC13 EBs were detected in tissue culture by 16 hours post infection. Real time RT-PCR analysis of PRCL homologs expressed by C. caviae revealed significantly decreased transcription of several, but not all, of these genes at 24 hours post infection, including the putative operon encoding CCA00523-525 and CCA00259 (Fig. 2C). Transcription of CCA00416 and CCA00417, that encode orthologs of Pls1 and Pls2  respectively appeared only mildly reduced 24 hours after infection (Fig. 2C). In contrast, transcription of CCA00453, encoding a phospholipase D enzyme  and CCA00924 was not significantly altered in CC13 (Fig. 2C) although all four of these genes were plasmid-responsive in C. muridarum and C. trachomatis . Transcription of glgA, the gene encoding glycogen synthase, is plasmid-dependent in C. trachomatis  but not in C. muridarum . Transcription of glgA by CC13 was mildly increased whether detected by microarray (data not shown) or by quantitative PCR (Fig. 2C). A microarray screen comparing the transcriptional profile of GPIC with CC13, 30 hours after infection confirmed these findings (Fig. S2) and failed to detect additional plasmid-responsive loci with orthologs in C. trachomatis and C. muridarum.
Figure 2. Transcriptional profiling of CC13 via microarray screening and quantitative RT-PCR reveals partial conservation of PRCL but a plasmid-independent glucose response in C. caviae.
(A) CC13 replicates normally during synchronous infection of L929 cells. GPIC and CC13 were inoculated at an MOI ~1 into L929 cells growing in 24 well dishes, then samples were harvested at intervals and titrated to determine IFU/well. (B) pCpGP1 transcripts are detected 24 hours after infection. Total RNA isolated from L929 cells 24 hours after infection with GPIC was treated with reverse transcriptase to generate cDNA and assayed via RT-PCR. Purified genomic GPIC DNA was used as positive control for the reactions while RNA processed for cDNA without reverse transcriptase was used as the negative control. (C) Quantitative RT-PCR confirms partial conservation of candidate PRCL in C. caviae. Total RNA was isolated from infected cells 24 hours after infection and PRCL transcripts were measured by quantitative RT-PCR. CCA00078 (mip) was included as a non-plasmid responsive control because we did not detect a difference in its transcription by microarray screen.doi:10.1371/journal.pone.0030747.g002
Transcription of C. trachomatis PRCL was coordinately reduced in response to glucose limitation . We examined the impact of 2DG treatment on the development of GPIC inclusions. L929 cells infected with GPIC were cultured with medium containing the glucose-6-phosphate inhibitor 2-deoxyglucose (2DG). Inclusions formed by GPIC in the presence of 10mM 2DG appeared smaller, with fewer EBs (Fig. 3B) when compared with those formed in untreated cells (Fig 3A). Occasionally inclusions containing larger aberrant forms were observed (Fig. 3C) that were not noted in the untreated cells. Overall, the number of inclusions was reduced and the infectious yield dropped ~100 fold after treatment (data not shown). Treatment of GPIC-infected cells with 2DG also resulted in significantly reduced transcription of glgA (~6 fold reduction) and other PRCL homologs (4.5–34.4 fold reduction) (Fig. 3D). Transcription of CCA00416, CCA00417, CCA00453 and CCA00924 was also significantly reduced in CC13 in response to treatment with 2DG (Fig. 3E) indicating that glucose-responsiveness by C. caviae was not dependent on the presence of the plasmid. These observations suggested a reduced role for pCpGP1 in regulation of chromosomal loci in C. caviae and indicated that plasmid-deficient C. caviae may not express phenotypes described for plasmid-deficient C. muridarum and C. trachomatis.
Figure 3. Effect of 2DG treatment on inclusion formation and transcription by C. caviae.
(A) GPIC inclusions in L929 cells 40 hours post infection (A–C). Inclusions formed by GPIC in cells treated with 10 mM 2DG are smaller (B) and may contain aberrant forms (C). Cells were fixed with methanol then stained with an anti-LPS monoclonal antibody and anti-mouse Alexa488 (secondary) to detect chlamydiae and cytoplasm was counterstained with Evans Blue. All inclusions were imaged at 400x magnification. (D) GPIC differentially regulates candidate PRCL transcription in response to 2DG treatment. Total RNA was isolated 24 hours after infection from infected cells treated with 10 mM 2DG and transcripts were quantified as for Fig. 2D. (E) Reduced PRCL transcription is plasmid-independent in C. caviae. L929 cells infected with CC13 were treated with 2DG as for ‘C’ and transcription of plasmid-insensitive loci 24 hours after infection was examined as before. The transcriptional differences are presented as fold change in expression for each gene in a single representative experiment although each experiment was performed independently at least twice and samples were assayed in triplicate. p values for transcripts indicated with * were <0.05.doi:10.1371/journal.pone.0030747.g003
C. caviae CC13 activates TLR2 signaling in vitro
Plasmid-deficient C. muridarum and C. trachomatis do not signal via TLR2, and induce production of significantly lower levels of proinflammatory cytokines than wild-type strains . We infected HEK 293 cells stably expressing TLR2 (HEK-TLR2) or TLR4/MD2 (HEK-TLR4) with C. caviae GPIC or CC13. HEK 293 cells endogenously express TLR1, TLR6 and MyD88, but not TLR2 or TLR4 , thus, specific signaling can be determined by measuring cytokine production in cells stably transfected with TLR2  because TLR-ligand interactions are highly conserved in nature . Analysis of supernatants after 24 hours incubation revealed similar levels of IL-8 production by HEK-TLR2 cells in response to infection with GPIC or CC13 (Fig. 4A). IL-8 production above media was not detected in the supernatant of HEK-TLR4 cells infected with either strain, consistent with the weak immunostimulatory capabilities of chlamydial LPS . These data indicate that despite absence of the plasmid, CC13 was able to induce TLR2-dependent cytokine production at levels equivalent to GPIC. Interestingly, while we observed no difference between TLR2 activation by CC13 or GPIC when we assayed X-ray inactivated preparations of the bacteria using HEK-TLR2 cells transfected with an NF- κB reporter plasmid (Fig. 4B), we noted that the amount of C. caviae required to activate TLR2 in these cells was considerably lower (~100 fold reduced) when compared with C. muridarum Nigg or its plasmid-cured derivative CM3.1, suggesting that the TLR2 ligands of this chlamydial species is extremely potent.
Figure 4. CC13 signals via TLR2 and induces cytokine production at levels similar to wild-type GPIC.
(A) IL-8 was measured in the supernatants of HEK 293 cells transfected with control plasmid, TLR2 or TLR4 and infected with GPIC or CC13 for 24 hrs. At both an MOI of 1 and 3, cytokine production did not differ between HEK-TLR2 cells infected with GPIC or CC13. Neither strain induced IL-8 levels above media for cells transfected with control plasmid or TLR4. The pattern of IL-8 production after stimulation with LPS (TLR4 agonist), TNF (NF-κB agonist), or Pam3Cys (TLR2 agonist) indicated that cytokine production was specific for the transfected TLR. (B) GPIC and CC13 express a potent, plasmid-independent TLR2 stimulating activity. X-ray inactivated chlamydial suspensions at various MOIs were incubated with HEK-TLR2 cells transfected with an NF-κB reporter plasmid. After 24 h of incubation NF-κB-induced secreted alkaline phosphatase activity was assayed using QUANTI-Blue. Bars represent the mean ± SE for three independent experiments. (C) Murine BMDDCs infected with GPIC or CC13 for 24 h did not differentially produce IL-6, (D) IL-1β, or (E) TNF-α. Bars represent the mean ± SD for duplicate wells. The data presented are from a single representative experiment that was performed at least twice.doi:10.1371/journal.pone.0030747.g004
C. caviae is highly pro-inflammatory when compared with C. muridarum in murine BMDDCs
C. muridarum plasmid-deficient strains induce significantly less cytokine production by dendritic cells in vitro than their wild-type parent . We incubated murine bone marrow derived dendritic cells (BMDDCs) with CC13, GPIC, C. muridarum Nigg or CM3.1 to determine if dendritic cell cytokine release was altered by the loss of the plasmid from C. caviae and to compare cytokine release during infection with C. caviae or C. muridarum. After 24 hours, BMDDCs stimulated with GPIC and CC13 secreted IL-1α, IL-1β, IL-6, IL-10, G-CSF, TNF-α, GM-CSF, MIP-1α, and KC. Strikingly, IL-6 (Fig. 4C), IL-1β (Fig. 4D), IL-1α, IL-10, MIP-1α, and KC (data not shown) were secreted at significantly higher levels in response to infection with C. caviae compared to C. muridarum, whereas G-CSF (data not shown), GM-CSF (data not shown), and TNF-α (Fig. 4E) levels were similar between the strains. In all instances, we observed no difference in cytokine response to infection with CC13 compared with GPIC. These data indicate that CC13 stimulated dendritic cell cytokine production to levels equivalent to GPIC and contrasted with C. muridarum where wild-type Nigg induced significantly greater cytokine release than plasmid-deficient CM3.1.
C. caviae CC13 is not attenuated in the guinea pig genital tract infection model
In order to determine the possible impact of plasmid-deficiency on the virulence of C. caviae in the genital tract infection model, groups of guinea pigs (N = 5) were inoculated intravaginally with either GPIC or CC13 and sacrificed on day 30. The course and intensity of infection was monitored via endocervical swabs. No difference in the magnitude or duration of infection was noted between the strains (Fig. 5A). In addition, IL-8 levels measured in genital secretions obtained during the first 10 days of infection were similar in the two groups (Fig. 5B). Histologic examination of oviduct tissue recovered from these guinea pigs failed to detect any pathology regardless of infecting strain.
Figure 5. Chlamydial burden and levels of IL-8 do not differ between GPIC and CC13 infected guinea pigs during genital tract infection.
Groups of 5 untreated (A, B) or estradiol-treated (C, D) female guinea pigs were intravaginally infected with either GPIC or CC13 and sacrificed on day 30. (A) Monitoring of endocervical swabs during infection of untreated guinea pigs revealed that infectious bacteria were not detectable after day 9 for either strain. (B) Vaginal sponges collected through day 10 revealed similar levels of IL-8 in GPIC and CC13-infected guinea pigs. (C) The course of GPIC infection in estradiol-treated guinea pigs was prolonged compared to CC13. (D) IL-8 levels monitored during the first 10 days of infection were not different in estradiol-treated guinea pigs infected with GPIC and CC13. Bars represent the mean ± SD of five guinea pigs per group and each experiment was performed once.doi:10.1371/journal.pone.0030747.g005
Treatment with estrogen maintains the reproductive tract epithelium and results in dramatically increased bacterial burden , which might facilitate detection of subtle differences in the course of infection and outcome with respect to upper reproductive tract pathology between CC13 and its plasmid-containing parent GPIC. Groups of estradiol-treated female guinea pigs (N = 5) were intravaginally inoculated with GPIC or CC13 and sacrificed on day 9 or day 30 post-infection. Culture of endocervical swabs again revealed no significant difference (two-way RM ANOVA, p = 0.898) in shedding of chlamydiae from the lower genital tract between GPIC and CC13 infected animals (Fig. 5C). However, GPIC infection was prolonged compared to CC13 (Kaplan-Meier, p = 0.02), with infection resolving by day 22 for CC13 and day 30 for GPIC (Fig. 5C). Swabs obtained from the upper uterine horns of guinea pigs sacrificed on day 9 were titrated via plaque assay and the bacterial burden did not differ significantly (p = 0.052, Student's t-test) between the strains (GPIC: 9.8×104±4.1×104; CC13: 4.3×103±2.1×103 PFU/ml). Thus, in estradiol-treated animals, the duration of genital tract infection was shortened and upper genital tract bacterial burden was slightly decreased during infection with CC13.
However, we were unable to detect differences in the cytokine response to infection between estradiol-treated guinea pigs infected with GPIC or CC13. Genital tract secretions collected from CC13 and GPIC infected animals through day 10 contained similar levels of IL-8 (Fig. 5D). Examination of the antibody response to infection in the estradiol-treated animals was performed using serum collected at the time of sacrifice on day 30. Analysis of antibody titers revealed no difference in the mean ± SD log10 titers of IgG1 (GPIC, 3.94±0.38 vs. CC13, 3.51±0.35), IgG2a (GPIC, 3.89±0.34 vs. CC13 3.59±0.46) or IgA (GPIC, 2.81±0.3 vs. CC13, 2.81±0.42) between the strains. These data indicate that despite the mildly prolonged infection observed during infection with GPIC, the humoral immune response did not significantly differ from that of animals infected with CC13.
Estradiol-treated guinea pigs exhibit similar inflammatory scores and pathology during infection with GPIC or CC13
Genital tract tissues harvested from estradiol-treated guinea pigs infected with GPIC or CC13 were analyzed histologically. The groups sacrificed on day 9 post-infection were graded for cellular infiltrates because infection was still present. The groups sacrificed on day 30 were analyzed for cellular infiltrates and for pathology remaining after clearance of infection. On day 9, the levels of inflammation in the exocervix (data not shown), endocervix (data not shown), uterine horns (Fig. 6A), and oviducts (Fig.6B-F) did not differ between the strains. On day 30, histologic examination did not reveal any difference in the degree of inflammation in the exocervix (Fig. 7A), endocervix (Fig. 7B), uterine horns (Fig. 7C) or oviducts (Fig. 7D; Fig. 8 A–F) during infection with GPIC or CC13. In addition, there was no difference in the percentage of animals with oviduct dilatation (Fig. 8A) or degree of dilatation between the groups (Fig. 8B–F).
Figure 6. Cellular infiltrates are similar early during infection in estradiol-treated guinea pigs infected with GPIC or CC13.
Groups of 5 estradiol-treated female guinea pigs infected with GPIC or CC13 were sacrificed on day 9. Pathology scores for inflammatory cells in the (A) uterine horns, and (B) oviducts were similar for GPIC- and CC13-infected animals. Boxes extend from the 25-75 percentiles and whiskers indicate the 5th–95th percentiles. (C, E) Representative oviduct histologic sections at 40X and 200X magnification from a guinea pig infected with GPIC. (D, F) Representative oviduct histologic sections at 40X and 200X magnification from a guinea pig infected with CC13. Data are from one experiment with 5 guinea pigs per group.doi:10.1371/journal.pone.0030747.g006
Figure 7. Cellular infiltrates are similar late during infection in estradiol-treated guinea pigs infected with GPIC or CC13.
Groups of 5 estradiol-treated female guinea pigs were infected with GPIC or CC13 were sacrificed on day 30. Median pathology scores for inflammatory cells in the (A) exocervix, (B) endocervix (B), (C) uterine horns, and (D) oviducts were similar for GPIC- and CC13-infected animals. Boxes extend from the 25–75 percentiles and whiskers indicate the 5th–95th percentiles. Data are from one experiment with 5 guinea pigs per group.doi:10.1371/journal.pone.0030747.g007
Figure 8. Oviduct pathology is similar following infection with GPIC or CC13.
Groups of 5 estradiol-treated female guinea pigs were infected with GPIC or CC13 and sacrificed on day 30. (A) Bars represent the percentage of oviducts with a pathology score of ≥ 1 for degree of infiltration of polymorphonuclear neutrophils (PMNs), lymphocyes/monocytes, plasma cells or oviduct dilatation. (B) Median pathology scores for oviduct dilatation in GPIC- and CC13-infected guinea pigs. Boxes extend from the 25–75 percentiles and whiskers indicate the 5th–95th percentiles. (C) Histologic sections of an oviduct from a representative estradiol-treated guinea pig on day 30 after infection with (C, E) GPIC or (D, F) CC13 shown at 100X and 200X, respectively. The oviduct architecture is intact with minimal inflammation and dilatation. Data are from the same experiment detailed in Fig. 6.doi:10.1371/journal.pone.0030747.g008
The overall degree of oviduct dilatation and oviduct epithelial cell damage resulting from genital tract infection was low. In contrast, we observed severe abdominal pathology in 3 of 5 GPIC-infected and 1 of 5 CC13-infected animals that were treated with estradiol, with fibrous adhesions noted between the genital tract and peritoneum, bowel, and bladder in these animals. These findings had been previously described in estrogen-treated, GPIC-infected guinea pigs by Rank et al  and are reminiscent of Fitz-Hugh-Curtis syndrome observed in a subset of Chlamydia-infected human females . This syndrome results from migration of the bacterium into the abdomen and leads to inflammation and fibrous adhesions in infected women.
Our studies ,  and others  have revealed an important role for the chlamydial plasmid in the expression of key virulence properties by both C. muridarum and C. trachomatis. However the role of the resident plasmid in other Chlamydiaceae has not been investigated. Strains of C. pneumoniae infecting humans generally lack the plasmid, but it is present in strains that infect a diverse range of other mammals including horses  and koalas . Recent studies of C. felis clinical isolates indicate that plasmid carriage is highly conserved  in this species suggesting that the plasmid may be important for virulence, and although the plasmid appears conserved in C. psittaci, plasmid-deficient strains have been described . In this study we investigated the role of the plasmid in C. caviae, a natural pathogen of the guinea pig, by curing GPIC of pCpGP1 to derive strain CC13 and by examining CC13′s ability to cause infection and genital tract disease.
Three plasmid-associated phenotypes have been identified in C. muridarum and are conserved in C. trachomatis: plasmid-deficient strains are unable to accumulate glycogen within the intracellular inclusion during the developmental cycle, display reduced infectivity in vitro , and in vivo  and do not stimulate TLR2 signaling during infection . Whether the effectors of these phenotypes are encoded directly by the plasmid is unknown, but we have identified a conserved group of plasmid-responsive loci encoded on the chromosome that may also contribute to the expression or regulation of these traits . Microarray screening using a custom GPIC array indicated that the transcriptional profile of CC13 very closely resembled that of its parent, but we nevertheless observed that several of the PRCL identified in plasmid-deficient C. muridarum and C. trachomatis were also differentially transcribed in CC13 including CCA00523-525 (orthologous to CT142-44 and TC_419-421), and CCA00259 (orthologous to CT382.1). However, other candidate PRCL such as the CT084 (TC_0357) ortholog CCA00453, and the CT702 (TC_075) ortholog CCA00924 did not differ transcriptionally from GPIC. Mild reduction in transcription of CCA00416-17 (orthologous to CT049-50 and [TC_319-320]) was detected that did not reach significance. Furthermore, transcription of glgA appeared slightly but significantly elevated (~2 fold) in C. caviae CC13. The significance of this observation is unclear because glycogen accumulation within wild-type C. caviae inclusions is not observed and glycogen production by this chlamydial species has not been detected  but indicates that glgA transcription is not plasmid-dependent in C. caviae, more closely resembling what we have previously observed for C. muridarum.
Phenotypic analysis of CC13 in vitro revealed that loss of pCpGP1 did not impact plaque size or plaquing efficiency. GPIC, in common with C. pneumoniae and C. psittaci, does not accumulate glycogen intrainclusionally , so no change in iodine-staining phenotype was anticipated and was not observed (data not shown). Most significantly, CC13 retained the ability to activate TLR2 expressed on stably transfected HEK293 epithelial cells, an observation that contrasted with the plasmid-cured strains CM972 and CTD153 that are unable to stimulate TLR2-dependent signaling in vitro and in vivo , . The overall conservation of chlamydial plasmid organization suggests that the plasmid may not encode a pathogenic TLR2 ligand directly and further, that the conservation of plasmid-responsiveness for CCA00523-25 and CCA00259 indicates that their expression is likely not required for TLR2 activation, reducing the likelihood that their orthologs encode candidate TLR2 ligands in C. trachomatis or C. muridarum. Alternatively, is also possible that GPIC encodes additional TLR2 ligands that are unaffected by the absence of the plasmid, preventing detection of differential TLR2 signaling effects as are observed in plasmid-cured C. trachomatis and C. muridarum. In support of this hypothesis, we observed that GPIC and CC13 activated TLR2 ~100 fold more effectively than C. muridarum Nigg, reflecting the expression of a potent, plasmid-independent TLR2 ligand by C. caviae.
CC13 and GPIC both displayed a strongly pro-inflammatory profile in dendritic cells. Consistent with our observation that C. caviae expresses a potent, plasmid-independent TLR2 ligand we detected strong induction of both TNF-α and IL-6 by BMDDCs in response to incubation with GPIC and CC13. C. caviae strongly induced IL-1β production by BMDDCs while C. muridarum did not. Prestimulation with TLR ligands is required for IL-1β production during infection of macrophages with C. muridarum . The high levels produced by BMDDCs infected with GPIC or CC13 suggests C. caviae is able to independently prime and induce release of IL-1β in these cells, unlike C. muridarum. Thus, it appears that in addition to a lack of plasmid-control for TLR2 activation, stimulatory pathways are activated by C. caviae differentially when compared to C. muridarum.
In light of our discovery that TLR2 signaling and infectivity were unimpaired in CC13 despite the absence of pCpGP1, it was not surprising that we were unable to demonstrate any significant attenuation in the guinea pig model of genital tract infection. Only with the potentiating effects of estradiol treatment were we able to detect a minor shortening in the course of infection and a slight reduction in upper genital tract bacterial load. Ultimately, these differences were not sufficient to drive differences in the development of oviduct immunopathology. Intra-abdominal adhesions were noted in animals in both groups and were likely the result of prolonged inflammation caused by the enhanced bacterial burden associated with administration of estradiol.
Comparisons of the effects of plasmid-curing on C. trachomatis, C. muridarum and C. caviae gene expression, regulation and virulence indicate significant differences that may be important for understanding the outcome of genital tract infection with these pathogens (Fig. 9). The overall similarity of both plasmid organization and sequence homology is high. Recent phylogenetic analysis of the plasmids expressed by 6 of the 9 chlamydial species indicates that they group distinctly, in a manner greatly resembling their genomes, with the plasmids obtained from C. pneumoniae strains most closely related to each other and to a lesser extent to those carried by C. psittaci, C. felis and C. caviae, and finally C. muridarum and C. trachomatis isolates . With such genetic similarity, how can we account for the differences in phenotype that we have observed in plasmid-cured strains? If the plasmid encodes effectors of chlamydial virulence directly, why aren't these phenotypes conserved in all species that carry the plasmid? Carlson et al.  proposed that the plasmid encodes a transcriptional regulator, and we have identified a conserved sub-population of chromosomal loci that are under its control . However, we observed only partial conservation of plasmid-dependence for these genes in C. caviae and if these are effectors of these phenotypes in C. muridarum and C. trachomatis, then this is a likely explanation of the failure of plasmid-curing to alter C. caviae virulence.
Figure 9. The impact of plasmid-curing on Chlamydia sp. gene expression, regulation and virulence.
This diagram outlines the gene expression and regulation and virulence phenotypes of plasmid-cured Chlamydia sp., and indicates the phylogenetic relationships of the plasmid carried by these species or strains and other genera. Data regarding phenotypic and transcriptional responses were obtained in the course of this or previous studies , , . NT indicates where phenotype was not examined. The dendrogram was constructed via CLUSTAL  and Neighbor-joining analysis was performed using plasmid sequences from C. trachomatis D/UW-3/Cx (pCTDLC1) , L2/434/Bu (pL2) , C. muridarum (pMoPn) , C. caviae (pCpGP1) , C. psittaci pCpA1 (Lusher,M.E., Gregory,J., Storey,C.C. and Richmond,S.J., unpublished; Genebank: NC_002117, C. felis (pCfe1)  and the non-human pathogens C. pneumoniae LPCoLN  and N16 (pCpnE1)  to determine branching order. The outcome of this analysis did not differ from the more detailed analysis published by Mitchell et al. .doi:10.1371/journal.pone.0030747.g009
Interestingly, we observed significant reduction of all candidate PRCL transcription in GPIC in response to 2DG treatment, indicating that C. caviae, like C. trachomatis, alters gene expression in response to an environment in which glucose is limiting. Furthermore, this response persisted in the cured CC13 strain indicating that this process is plasmid independent. Consequently, glgA expression was plasmid-insensitive but glucose-limited by C. caviae, a novel transcriptional profile that contrasts with our observations for both C. trachomatis where glgA transcription is both plasmid and glucose sensitive and C. muridarum where glgA transcription is unaltered in response to plasmid loss or glucose limitation . This may reflect a glucose-responsive regulatory pathway evolving within Chlamydiaceae to facilitate modulation of non-essential, plasmid-associated, virulence gene expression. This is a common theme in bacterial pathogens where expression of virulence loci may be tightly controlled in response to environmental signals such as temperature , nutrient limitation , carbon availability , or phosphate homeostasis . In the context of such a model, it appears that C. muridarum has not co-evolved or has dispensed with the transcriptional controls exerted by this pathway with the result that transcription of PRCL remains constitutively active in glucose-limiting conditions . If true, this may explain the high incidence of upper reproductive tract pathology observed in the mouse model  while human infection is predominantly sub-clinical and reproductive tract sequelae relatively uncommon . Indeed, it may be worth noting that neither C. caviae nor C. muridarum are natural pathogens of the genital tract, infecting the eye and respiratory tract of their respective hosts, so coordination of virulence-associated gene expression in response to glucose availability might not be relevant for these sites. Our preliminary observations indicate that TLR2 signaling by C. trachomatis but not by C. muridarum is impaired when the chlamydiae have been cultured under glucose-restricted conditions , but until the pathogenic TLR2 ligand(s) have been identified, limited or altered expression of the TLR2 ligand(s) cannot be confirmed. If correct, then identification and characterization of the regulatory factor(s) encoded by the plasmid and those involved in the chlamydial carbon response will greatly advance our understanding of this important virulence process.
Primers directed against the predicted open reading frames encoded on pCpGP amplify predicted fragments from C. caviae GPIC but not from CC13. Primers pairs directed against each ORF are detailed in Table 1 and amplification conditions are described in Methods.
Scatter plot illustration of microarray comparison of the transcriptional profile of C. caviae GPIC and its plasmid-cured derivative CC13 30 hours after infection.
Conceived and designed the experiments: CO TD RB LF. Performed the experiments: YA KCK MM LF CO CA MZ. Analyzed the data: CO LF TD RB RI MZ YA. Wrote the paper: CO LF TD.
- 1. Stephens RS, Myers G, Eppinger M, Bavoil PM (2009) Divergence without difference: phylogenetics and taxonomy of Chlamydia resolved. FEMS Immunol Med Microbiol 55: 115–119.
- 2. Comanducci M, Ricci S, Ratti G (1988) The structure of a plasmid of Chlamydia trachomatis believed to be required for growth within mammalian cells. Mol Microbiol 2: 531–538.
- 3. O'Connell CM, Nicks KM (2006) A plasmid-cured Chlamydia muridarum strain displays altered plaque morphology and reduced infectivity in cell culture. Microbiology 152: 1601–1607.
- 4. O'Connell CM, Ingalls RR, Andrews CW Jr, Skurlock AM, Darville T (2007) Plasmid-deficient Chlamydia muridarum fail to induce immune pathology and protect against oviduct disease. J Immunol 179: 4027–4034.
- 5. Russell M, Darville T, Chandra-Kuntal K, Smith B, Andrews CW Jr, et al. (2011) Infectivity acts as in vivo selection for maintenance of the chlamydial “cryptic” plasmid. Infect Immun: 79: 98–107.
- 6. Matsumoto A, Izutsu H, Miyashita N, Ohuchi M (1998) Plaque formation by and plaque cloning of Chlamydia trachomatis biovar trachoma. J Clin Microbiol 36: 3013–3019.
- 7. Darville T, O'Neill JM, Andrews CW Jr, Nagarajan UM, Stahl L, et al. (2003) Toll-like receptor-2, but not toll-like receptor-4, is essential for development of oviduct pathology in chlamydial genital tract infection. J Immunol 171: 6187–6197.
- 8. O'Connell CM, Abdelrahman YM, Green E, Darville HK, Saira K, et al. (2011) TLR2 activation by Chlamydia trachomatis is plasmid-dependent and plasmid-responsive chromosomal loci are coordinately regulated in response to glucose limitation by C. trachomatis but not by C. muridarum. Infect Immun 79: 1044–1056.
- 9. Rank RG, Bowlin AK, Reed RL, Darville T (2003) Characterization of chlamydial genital infection resulting from sexual transmission from male to female guinea pigs and determination of infectious dose. Infect Immun 71: 6148–6154.
- 10. Rank RG, Sanders MM, Bowie WR, Caldwell HD, Jones RB, et al. (1990) Ascending genital tract infection as a common consequence of vaginal inoculation with the guinea pig inclusion conjunctivitis agent in normal guinea pigs. Chlamydial infections. New York: Cambridge University Press. pp. 249–252.
- 11. Batteiger BE, Rank RG, Soderberg LSF, Bowie WR, Caldwell HD, et al. (1990) Immunization of guinea pigs with isolated chlamydial outer membrane proteins. Chlamydial infections. New York: Cambridge University Press. pp. 2650–2268.
- 12. Rank RG, Sanders MM (1992) Pathogenesis of endometritis and salpingitis in a guinea pig model of chlamydial genital infection. Am J Pathol 140: 927–936.
- 13. Gordon FB, Quan AL (1972) Susceptibility of Chlamydia to antibacterial drugs: test in cell cultures. Antimicrob Agents Chemother 2: 242–244.
- 14. Rockey DD, Fischer ER, Hackstadt T (1996) Temporal analysis of the developing Chlamydia psittaci inclusion by use of fluorescence and electron microscopy. Infect Immun 64: 4269–4278.
- 15. Fan VS, Jenkin HM (1970) Glycogen metabolism in Chlamydia-infected HeLa-cells. J Bacteriol 104: 608–609.
- 16. Kelly KA, Robinson EA, Rank RG (1996) Initial route of antigen administration alters the T-cell cytokine profile produced in response to the mouse pneumonitis biovar of Chlamydia trachomatis following genital infection. Infect Immun 64: 4976–4983.
- 17. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: 402–408.
- 18. Latz E, Visintin A, Lien E, Fitzgerald KA, Monks BG, et al. (2002) Lipopolysaccharide rapidly traffics to and from the Golgi apparatus with the toll-like receptor 4-MD-2-CD14 complex in a process that is distinct from the initiation of signal transduction. J Biol Chem 277: 47834–47843.
- 19. Inaba K, Inaba M, Romani N, Aya H, Deguchi M, et al. (1992) Generation of large numbers of dendritic cells from mouse bone marrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J Exp Med 176: 1693–1702.
- 20. Pasley JN, Rank RG, Hough AJ Jr, Cohen C, Barron AL (1985) Effects of various doses of estradiol on chlamydial genital infection in ovariectomized guinea pigs. Sex Transm Dis 12: 8–13.
- 21. Darville T, Andrews CW Jr, Laffoon KK, Shymasani W, Kishen LR, et al. (1997) Mouse strain-dependent variation in the course and outcome of chlamydial genital tract infection is associated with differences in host response. Infect Immun 65: 3065–3073.
- 22. Patterson TL, Rank RG (1996) Immunity to reinfection and immunization of male guinea pigs against urethral infection with the agent of guinea pig inclusion conjunctivitis. Sex Transm Dis 23: 145–150.
- 23. O'Connell CM, Ionova IA, Quayle AJ, Visintin A, Ingalls RR (2006) Localization of TLR2 and MyD88 to Chlamydia trachomatis inclusions. Evidence for signaling by intracellular TLR2 during infection with an obligate intracellular pathogen. J Biol Chem 281: 1652–1659.
- 24. Jorgensen I, Valdivia RH (2008) Pmp-like proteins Pls1 and Pls2 are secreted into the lumen of the Chlamydia trachomatis inclusion. Infect Immun 76: 3940–3950.
- 25. Koo IC, Walthers D, Hefty PS, Kenney LJ, Stephens RS (2006) ChxR is a transcriptional activator in Chlamydia. Proc Natl Acad Sci USA 103: 750–755.
- 26. Carlson JH, Whitmire WM, Crane DD, Wicke L, Virtaneva K, et al. (2008) The Chlamydia trachomatis plasmid Is a transcriptional regulator of chromosomal genes and a virulence factor. Infect Immun 76: 2273–2283.
- 27. Ingalls RR, Rice PA, Qureshi N, Takayama K, Lin JS, et al. (1995) The inflammatory cytokine response to Chlamydia trachomatis infection is endotoxin mediated. Infect Immun 63: 3125–3130.
- 28. Rank RG, Barron AL, Mardh PA, Holmes KK, Oriel JD, et al. (1982) Prolonged genital infection by GPIC agent associated with immunosuppression following treatment with estradiol. Chlamydial infections: Elsevier Biomedical Press, New York. pp. 391–394.
- 29. Kobayashi Y, Takeuchi H, Kitade M, Kikuchi I, Sato Y, et al. (2006) Pathological study of Fitz-Hugh-Curtis syndrome evaluated from fallopian tube damage. J Obstet Gynaecol Res 32: 280–285.
- 30. Thomas NS, Lusher M, Storey CC, Clarke IN (1997) Plasmid diversity in Chlamydia. Microbiology 143(Pt 6): 1847–1854.
- 31. Mitchell C, Hovis K, Bavoil P, Myers G, Carrasco J, et al. (2010) Comparison of koala LPCoLN and human strains of Chlamydia pneumoniae highlights extended genetic diversity in the species. BMC Genomics 11: 442.
- 32. Harley R, Day S, Di Rocco C, Helps C (2010) The Chlamydophila felis plasmid is highly conserved. Vet Microbiol 146: 172–174.
- 33. McClenaghan M, Honeycombe JR, Bevan BJ, Herring AJ (1988) Distribution of plasmid sequences in avian and mammalian strains of Chlamydia psittaci. J Gen Microbiol 134 (Pt 3): 559–565.
- 34. Ojcius DM, Degani H, Mispelter J, Dautry-Varsat A (1998) Enhancement of ATP levels and glucose metabolism during an infection by Chlamydia. NMR studies of living cells. J Biol Chem 273: 7052–7058.
- 35. Prantner D, Darville T, Sikes JD, Andrews CW Jr, Brade H, et al. (2009) Critical role for interleukin-1beta (IL-1beta) during Chlamydia muridarum genital infection and bacterial replication-independent secretion of IL-1beta in mouse macrophages. Infect Immun 77: 5334–5346.
- 36. Klinkert B, Narberhaus F (2009) Microbial thermosensors. Cell Mol Life Sci 66: 2661–2676.
- 37. Somerville GA, Proctor RA (2009) At the crossroads of bacterial metabolism and virulence factor synthesis in staphylococci. Microbiol Mol Biol Rev 73: 233–248.
- 38. Lucchetti-Miganeh C, Burrowes E, Baysse C, Ermel G (2008) The post-transcriptional regulator CsrA plays a central role in the adaptation of bacterial pathogens to different stages of infection in animal hosts. Microbiology 154: 16–29.
- 39. Lamarche MG, Wanner BL, Crépin S, Harel J (2008) The phosphate regulon and bacterial virulence: a regulatory network connecting phosphate homeostasis and pathogenesis. FEMS Microbiol Rev 32: 461–473.
- 40. Haggerty CL, Gottlieb SL, Taylor BD, Low N, Xu F, et al. (2010) Risk of sequelae after Chlamydia trachomatis genital infection in women. J Infect Dis 201: Suppl 2S134–155.
- 41. Larkin M, Blackshields G, Brown N, Chenna R, McGettigan P, et al. (2007) Clustal W and Clustal X version 2.0. Bioinformatics 23: 2947–2948.
- 42. Sturdevant GL, Kari L, Gardner DJ, Olivares-Zavaleta N, Randall LB, et al. (2010) Frameshift mutations in a single novel virulence factor alter the in vivo pathogenicity of Chlamydia trachomatis for the female murine genital tract. Infect Immun 78: 3660–3668.
- 43. Thomson NR, Holden MT, Carder C, Lennard N, Lockey SJ, et al. (2007) Chlamydia trachomatis: Genome sequence analysis of lymphogranuloma venereum isolates. Genome Res 18: 161–71.
- 44. Read T, Brunham R, Shen C, Gill S, Heidelberg J, et al. (2000) Genome sequences of Chlamydia trachomatis MoPn and Chlamydia pneumoniae AR39. Nucleic Acids Res 28: 1397–1406.
- 45. Read TD, Myers GS, Brunham RC, Nelson WC, Paulsen IT, et al. (2003) Genome sequence of Chlamydophila caviae (Chlamydia psittaci GPIC): examining the role of niche-specific genes in the evolution of the Chlamydiaceae. Nucleic Acids Res 31: 2134–2147.
- 46. Azuma Y, Hirakawa H, Yamashita A, Cai Y, Rahman M, et al. (2006) Genome sequence of the cat pathogen, Chlamydophila felis. DNA Res 13: 15–23.
- 47. Thomas N, Lusher M, Storey C, Clarke I (1997) Plasmid diversity in Chlamydia. Microbiology 143: 1847–1854.