If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Division of Reproductive Biology, Department of Obstetrics and Gynecology, Boston Medical Center, Boston University School of Medicine, Boston, Massachusetts 02118Department of Obstetrics, Gynecology and Reproductive Biology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115
Department of Obstetrics, Gynecology and Reproductive Biology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115Department of Microbiology, Immunology, and Parasitology, Louisiana State University Health Sciences Center, New Orleans, Louisiana 70112
To whom correspondence should be addressed: Division of Reproductive Endocrinology and Fertility, Dept. of Obstetrics, Gynecology and Women's Health, University of Missouri-Columbia School of Medicine, Columbia Regional Hospital, 402 Keene St., Third floor, Columbia, MO 65201. Tel.: 573-499-6044; Fax: 573-499-6063
* This work was supported by National Institutes of Health Grants U19AI061972 and AI046518 (to D. J. S., A. J. Q., and R. B. P.) and KD44319 (to R. S. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Chlamydia trachomatis is an obligate intracellular pathogen that can persist in the urogenital tract. Mechanisms by which C. trachomatis evades clearance by host innate immune responses are poorly described. CD1d is MHC-like, is expressed by epithelial cells, and can signal innate immune responses by NK and NKT cells. Here we demonstrate that C. trachomatis infection down-regulates surface-expressed CD1d in human penile urethral epithelial cells through proteasomal degradation. A chlamydial proteasome-like activity factor (CPAF) interacts with the CD1d heavy chain, and CPAF-associated CD1d heavy chain is then ubiquitinated and directed along two distinct proteolytic pathways. The degradation of immature glycosylated CD1d was blocked by the proteasome inhibitor lactacystin but not by MG132, indicating that degradation was not via the conventional proteasome. In contrast, the degradation of non-glycosylated CD1d was blocked by lactacystin and MG132, consistent with conventional cellular cytosolic degradation of N-linked glycoproteins. Immunofluorescent microscopy confirmed the interruption of CD1d trafficking to the cell surface, and the dislocation of CD1d heavy chains into both the cellular cytosol and the chlamydial inclusion along with cytosolic CPAF. C. trachomatis targeted CD1d toward two distinct proteolytic pathways. Decreased CD1d surface expression may help C. trachomatis evade detection by innate immune cells and may promote C. trachomatis persistence.
Chlamydia trachomatis serovars D–K are common obligate intracellular pathogens that infect the columnar epithelia of the human urogenital mucosa (
). CPAF, composed of N-terminal (29 kDa) and C-terminal (35 kDa) fragments, is secreted from the chlamydial inclusion into the host cytosol and degrades the transcriptional factor RFX5 that would otherwise up-regulate promoters of MHC class I HC and β2m genes (
). Cellular fractionation places CPAF into subfractions that differ from those containing the classic cytosolic proteasome subunits. Further, CPAF-associated proteasomal activity is inhibited by one cytosolic proteasome inhibitor, lactacystin, but not by other proteasome inhibitors (
). These data indicate that CPAF has proteasomal activities distinct from those of the classic cytosolic proteasome. CD1d is an MHC-like glycoprotein that presents lipid antigen to natural killer T (NKT) cells (
). In fact, diverse CD1d HC isoforms, including mature glycosylated (∼48 kDa), immature glycosylated (∼45 kDa), and non-glycosylated (∼37 kDa) forms, are expressed on the cell surface in a cell-type specific fashion (
). Activation of CD1d-restricted invariant NKT cells enhances host resistance to these microbes. CD1d-restricted NKT cells can act directly on infected cells, killing the CD1d-expressing cell. They also promote interferon γ production by conventional NK cells and modulate adaptive immune cells by altering Th1/Th2 polarization. Recognition of CD1d by invariant NKT cells causes rapid release of interleukin-4 and interferon γ from the NKT cell (
Viewing the importance of CD1d in innate immune responses to microbes, we hypothesized that C. trachomatis may alter CD1d-mediated immune pathways and thereby avoid innate immune destruction of the infected cell by the host. Here we demonstrate that surface-expressed CD1d in human urethral epithelial cells is down-regulated by C. trachomatis infection, and this involves both CPAF-mediated and classic cytosolic proteasomal pathways.
MATERIALS AND METHODS
Epithelial Cell Line—The PURL epithelial cell line was established from penile urethra collected at autopsy under IRB approval. Small pieces of tissue were cultured in supplemented keratinocyte growth media (Invitrogen) and 0.4 mm calcium until epithelial outgrowth occurred. Primary cells were transduced with a retroviral vector (LXSN-16E6E7) (
), selected by resistance to the neomycin analogue G418 and passaged over 20 times prior to experiments. The cytokeratin profile (positive for CK13, -17, and -18), and the expression of the secretory component of the polymeric immunoglobulin receptor (SC) confirmed derivation from penile urethral epithelium.
C. trachomatis Infection—Near confluent PURL cells were overlaid with C. trachomatis serovar F (strain F/IC-Cal-13) elementary bodies suspended in a sucrose-phosphate-glutamate solution at a predetermined dilution that resulted in 80–85% of cells becoming infected. Plates were centrifuged for 1 h, supernatants were aspirated after centrifugation, and cells cultured for up to 45 h at 37 °C in keratinocyte growth media.
Antibodies—For anti-CD1d antibodies, a D5 mAb (mIgG2b (
)) recognizing all isoforms of CD1d HC was used for biochemical experiments. An NOR3.2 mAb (mIgG1, Abcam Inc., Cambridge, MA) recognizing all isoforms was used for flow cytometry and immunofluorescence. A 51.1.3 mAb (mIgG2b (
)) recognizing C terminus fragments of CPAF (kind gifts from Dr. G. Zhong, University of Texas, San Antonio, TX) were used for immunoprecipitation and immunoblotting, respectively. The n54b mAb was also used for immunofluorescence.
Flow Cytometry—PURL cells infected with C. trachomatis were harvested 24 h after infection. Uninfected and infected PURL cells were harvested using accutase (Chemicon, Temecula, CA). Harvested cells were then incubated with an anti-transferrin receptor mAb conjugated to phycoerythrin (Caltag Laboratories) or with an anti-CD1d NOR3.2 mAb for 30 min at 4 °C. CD1d staining was followed by a goat antimouse Ig secondary antibody conjugated to phycoerythrin for 30 min at 4 °C (BD Biosciences, San Jose, CA). Cells were suspended in 1% paraformaldehyde and analyzed using a FACS-Calibur flow cytometry system (BD Biosciences).
Proteasome Inhibitor Treatment—Infected and control PURL cells were cultured for up to 45 h in the presence or absence of two different cytosolic proteasome inhibitors: lactacystin (2 or 10 μm) or MG132 (2 or 10 μm, Sigma-Aldrich) in Me2SO. Control wells included vehicle alone.
Immunoprecipitation and Western Immunoblotting—Harvested PURL cells were lysed in modified radioimmune precipitation assay buffer (1% Nonidet P-40, 1% deoxycholate, 0.1% SDS, 10 mm Tris, 150 mm NaCl, 2 mm EDTA) with protease inhibitors (Amersham Biosciences). Equivalent aliquots of cell lysates were incubated overnight at 4 °C with 5 μg/ml of anti-CD1d D5 mAb or anti-CPAF n54b and 5 μl of Protein-A-Sepharose (Amersham Biosciences). Precipitated proteins were separated by SDS-PAGE using 8 or 10% acrylamide gels and transferred to polyvinylidene difluoride membranes. Purified CPAF c-fragment proteins (a kind gift from Dr. G. Zhong, University of Texas, San Antonio, TX) were included as positive controls. Anti-CD1d D5 mAb, anti-CPAF c100a mAb, or a rabbit anti-β-actin polyclonal antibody (Abcam Inc.) were used as primary reagents for immunoblotting and peroxidase-conjugated goat anti-mouse or anti-rabbit IgG antibodies (Pierce) as secondary reagents. A peroxidase-conjugated mouse anti-ubiquitin antibody (P4D1, Santa Cruz Biotechnology, Santa Cruz, CA) was used to detect ubiquitin. Products in Western immunoblotting experiments were visualized using standard chemiluminescence (Amersham Biosciences). Molecular weights were confirmed by comparison to standard size markers and molecular weight analysis (FluorChem™SP, Alpha Innotech, San Leandro, CA).
Endoglycosidase-H Treatment—D5-precipitated CD1d HCs were denatured and incubated overnight at 37 °C with endoglycosidase-H (New England Biolabs, Beverly, MA) in reaction buffer (
). Protein products were analyzed by Western immunoblotting.
Immunoprecipitation by a His-tagged, Synthetic CD1d Cytoplasmic Tail—Three polypeptides with sequences corresponding to the wild-type CD1d cytoplasmic tail (HHHHHH-RFKRQTSYQGVL), to a mutated cytoplasmic tail lacking tyrosine and lysine residues (HHHHHH-RFKRQTSFQGVA), or to a truncated cytoplasmic tail lacking six amino acids (HHHHHH-RFKRQT) (
) were synthesized, His tag conjugated, and purified by high-performance liquid chromatography (New England Peptide, Inc., Gardner, MA). Total cell lysates were incubated with or without the synthesized peptides (0, 20, or 100 μg) in 200 μl of radioimmune precipitation assay buffer for 2 h at 4 °C. Bound proteins were recovered using anti-His tag antibody-conjugated agarose beads (Abcam Inc.), separated by PAGE and immunoblotted with the anti-CPAF c100a mAb.
Fluorescence Deconvolution Microscopy—PURL cells were seeded onto coverslips and infected as above. The ER was visualized using ER tracker Blue-White DPX (Molecular Probes, Eugene, OR) for 30 min at 37 °C. All coverslips were fixed in 4% paraformaldehyde, permeabilized with 0.1% Tween 20, and incubated for 2 h at 37 °C with anti-CD1d 51.1.3 mAb, anti-CD1d NOR3.2 mAb, or anti-CPAF n54b mAb singly, or in combination with anti-chlamydial LPS (clone-3, Accurate, Westbury, NY). Alexa Fluor 568-conjugated anti-mouse IgG1 (51.1.3, NOR3.2, and n54b) or Alexa Fluor 488-conjugated anti-mouse IgG3 (chlamydial LPS) were used as secondary reagents. In some experiments NOR3.2 was directly conjugated with Zenon Alexa Fluor 488 using a mouse IgG1 labeling kit (Molecular Probes) to allow costaining with CPAF. Thereafter, except in ER-Tracker-treated coverslips, cells were counterstained with a 4′,6-diamidino-2-phenylindole (Molecular Probes) nucleic acid stain. Images were obtained with a Leica DMRXA automated upright epifluorescence microscope (Leica Microsystems, Bannockburn, IL); a Sensicam QE charge-coupled device (Cooke Corp., Auburn Hills, MI); and filter sets optimized for Alexa 488 (exciter HQ480/20, dichroic Q495LP, and emitter HQ510/20m), Alexa 568 (exciter 545/30x, dichroic Q570DLP, emitter HQ620/60m), and 4′,6-diamidino-2-phenylindole (exciter 360/40×, dichroic 400DCLP, and emitter GG420LP). Z-axis plane capture, deconvolution, and analysis were performed with Slidebook™ deconvolution software (Intelligent Imaging Innovations, Denver, CO).
To determine how C. trachomatis infection affects CD1d expression, flow cytometry was used to analyze its cell-surface expression on PURL epithelial cells, a cell line we immortalized from penile urethra, the most common site of infection in the male. Cells were infected with C. trachomatis serovar F in this, and all subsequent experiments, because it is one of the most frequent genital isolates. We observed that CD1d was expressed by the majority of uninfected PURL cells; however, by 24 h post infection, CD1d expression was abrogated in >90% of cells in the C. trachomatis-infected cultures (Fig. 1B). The loss of CD1d expression was selective, because PURL cells retained, and in fact showed a slight increase in, their expression of transferrin receptor (Fig. 1C). PURL cells were then infected with C. trachomatis and harvested at various time points (up to 45 h after infection) to biochemically assess the effects of C. trachomatis infection on CD1d HC. All isoforms of CD1d HC were recovered using an anti-CD1d mAb (D5) that recognizes the α-region of the CD1d HC regardless of its association with β2m (
). The anti-CD1d D5 mAb precipitated CD1d HC isoforms of ∼48 kDa (open arrowhead), 45 kDa (closed arrow), and 37 kDa (open arrow) with distinct patterns depending on infection status and time after C. trachomatis infection. (Fig. 2A, panel 1). In non-infected PURL cells, the 48-kDa CD1d HC predominated, whereas the 45- and 37-kDa forms were present in negligible amounts (panel 1, lane 1).
CD1d isoforms were altered in the presence of C. trachomatis infection in a time-dependent, stepwise fashion (Fig. 2A, panel 1). The 48-kDa HC was largely converted to a 45-kDa HC form between 10 to 20 h post infection (p.i.), and this was accompanied by a decrease in β2m protein levels (Fig. 2A, panels 1 and 5). To confirm the glycosylation status of the 45-kDa HC, the isoform was digested by endoglycosidase-H (Fig. 2C). The 45-kDa CD1d HC was sensitive to endoglycosidase-H and represents an immature glycosylated CD1d that may include β2m-unassembled HCs.
The amount of detectable, 45-kDa, glycosylated CD1d HC protein decreased by 20 h p.i. By 30 h p.i., it was no longer detectable. In its place, a 37-kDa non-glycosylated CD1d HC form began to accumulate at 20 h p.i. This isoform predominated at 30 h but was nearly undetectable by the end of the infectious cycle.
We hypothesized that the degradation of CD1d HCs in the presence of C. trachomatis infection may involve the C. trachomatis-specific proteasomal activity, CPAF (
). To address this possibility, we used coimmunoprecipitation (IP) to search for physiologic interactions between CPAF and CD1d HCs (Fig. 2A, panels 2–4). Using combinations of the anti-CD1d D5 mAb and mAbs against the C-terminal (c100a) or N-terminal (n54b) fragments of CPAF we could demonstrate association between CPAF and CD1d and follow these associations through the infectious cycle. CPAF-bound proteins and CPAF itself were immunoprecipitated with the anti-CPAFn n54b mAb, separated by PAGE, and immunoblotted with anti-CD1d D5 (panel 2) or anti-CPAFc c100a (panel 3) mAbs. In turn, CD1d-associated proteins were recovered with the anti-CD1d D5 mAb, separated by PAGE and immunoblotted with anti-CPAFc c100a mAb (panel 4). Experiments using primary IP with anti-CD1d D5 and anti-CPAFn n54b mAbs demonstrated that CD1d HCs and CPAF interacted with each other physiologically (Fig. 2A, panels 2 and 4). Furthermore, CPAF preferentially associated with the 45-kDa immature glycosylated CD1d and 37-kDa non-glycosylated CD1d but not the 48-kDa mature glycosylated CD1d HC (Fig. 2A, panel 2). CPAF binding to CD1d HC was first detected at 20 h p.i. By 30 h p.i., the CPAF-associated immature glycosylated CD1d was undetectable. In parallel, the levels of the CPAF-associated non-glycosylated CD1d increased transiently at 30 h p.i. and decreased to nearly undetectable levels by 45 h p.i. Primary CPAF IPs (panels 3 and 4) confirmed the findings of primary CD1d IPs. CPAF was first detected at 10 h p.i., and the levels of total and CD1d-associated CPAF increased by 20 h p.i. and then decreased by 45 h p.i.
To assess the role of proteasomal activity in C. trachomatis-associated degradation of CD1d HC, C. trachomatis-infected cells were exposed to the cytosolic proteasome inhibitor, MG132 (left panel) and to lactacystin (right panel) (Fig. 2B). Total CD1d HC was detected by IP-IB with the anti-CD1d D5 mAb. In the absence of proteasomal inhibition, CD1d HCs could not be detected at 45 h p.i. In the presence of either MG132 or lactacystin, the 37-kDa non-glycosylated CD1d was detectable, and levels of this HC increased with increasing inhibitor dose. In turn, in the presence of lactacystin, but not MG132, the 45-kDa immature glycosylated CD1d could be detected, and levels of this HC increased with lactacystin dose. The rescued 45-kDa HC form was confirmed to be sensitive to endoglycosidase-H. The immature glycosylated CD1d HC was degraded without deglycosylation (Fig. 2C).
As illustrated in panels 3 and 4 of Fig. 2A, CD1d-associated CPAF was no longer detectable by 45 h p.i., despite the continued presence of some CPAF protein. Indeed, CD1d-associated CPAF was rescued by both MG132 and lactacystin, indicating that at least the classic cytosolic proteasome is involved in the degradation of CPAF together with the CD1d HC (Fig. 3A).
Fractionation experiments have demonstrated that CPAF is localized to the cytosol and is active in this location (
). Still, CD1d and CPAF interact physiologically. We hypothesized that CPAF associates with CD1d HCs via a site on the CD1d cytoplasmic tail. To address this hypothesis, we synthesized three peptides with amino acid sequences corresponding to the 12 amino acids that comprise the entire wild-type CD1d cytoplasmic tail (RFKRQTSYQGVL), to a mutated cytoplasmic tail lacking tyrosine and lysine residues (RFKRQTSFQGVA), and to a truncated cytoplasmic tail lacking six amino acids (RFKRQT) (
). All peptides were conjugated to a His tag. Total cell lysates from infected and control PURL cells were incubated with the synthetic His-tagged CD1d cytoplasmic tail, and peptide-associated proteins were recovered with an anti-His tag antibody. Western immunoblotting of the precipitated proteins demonstrated that CPAF interacted physically with the CD1d cytoplasmic tail, and these interactions occurred in a dose-dependent manner (Fig. 3B). In comparison to the CPAF band precipitated by the wild-type peptide (wt), CPAF was barely detectable after precipitation by the point-mutated peptide (mut1) and undetectable after precipitation with the truncated peptide (mut2) (Fig. 3C). These data suggest that CPAF interacts directly or indirectly to CD1d HC via sites on the CD1d cytoplasmic tail.
The classic cellular proteasome requires ubiquitination of host N-linked glycoproteins prior to their degradation (
). PURL cells were infected with C. trachomatis and exposed to proteasome inhibitors. Total cell lysates harvested at 30 h pi were subjected to IP with either anti-CD1d D5 (Fig. 4A) or anti-CPAFn n54b (Fig. 4B) mAbs. Immunoprecipitants were separated by PAGE and immunoblotted with a peroxidase-conjugated anti-ubiquitin antibody. Ubiquitinated CD1d HCs were not detected in non-infected cells (Fig. 4A, lane 1) and were barely detectable in C. trachomatis-infected cells that were not exposed to proteasome inhibitors (Fig. 4A, lane 2). In contrast, in the presence of proteasome inhibitors, ubiquitinated proteins accumulated in C. trachomatis-infected cells. Use of protein size markers and size analysis confirmed that the molecular weights of the ubiquitinated CD1d and CPAF bands were as expected. Ubiquitinated CD1d HCs were observed as a ladder of signals with molecular masses greater than ∼45 kDa (Fig. 4A, lanes 3 and 4). The 37- and 43-kDa bands may represent ubiquitinated proteins coprecipitated with CD1d, possibly including ubiquitinated CPAF N and C terminus fragments (Fig. 4A). Ubiquitinated CPAF-associated proteins, including CD1d HCs, were observed as a ladder of signals with molecular masses greater than ∼45 kDa (Fig. 4B, lanes 3 and 4). The amounts of ubiquitinated CPAF or CPAF-associated proteins, including CD1d HCs, were greater after exposure to lactacystin than to MG132 (Fig. 4, A and B).
To visually document the effect of C. trachomatis infection on CD1d intracellular trafficking, immunofluorescence microscopy was first performed with an anti-CD1d mAb (NOR3.2) that reacts with total CD1d HCs and either an ER-specific marker (ER tracker) or an anti-chlamydial LPS mAb (Fig. 5). In non-infected PURL cells, NOR3.2-reactive CD1d was detected throughout the intracellular space, with increased accumulation near the cell surface (Fig. 5A, upper images). In contrast, the majority of CD1d molecules in C. trachomatis-infected PURL cells localized to the perinuclear area near the ER (Fig. 5A, lower images). In addition, CD1d was present in infected cells in a distinct ring-shaped intracellular distribution that correlates morphologically with the chlamydial inclusion. CD1d and ER signals partially colocalized in the perinuclear area, suggesting that some forms of CD1d are present within the ER, but the majority of CD1d localized to the cytosol surrounding the ER (Fig. 5A, right images). Dual labeling for CD1d and chlamydial LPS verified the colocalization of CD1d and chlamydial elements within the chlamydial inclusion (Fig. 5B, green to red pseudocolor). These immunofluorescence microscopy studies thus support our flow cytometry and biochemical data that CD1d HC in C. trachomatis-infected cells fails to traffic efficiently to the cell surface. Rather, by 24 h p.i., the majority of CD1d HCs can be found in the cytosol near the ER, although some HCs localize within the ER and around chlamydial inclusion. CD1d HC appears to be targeted toward two degradation pathways: one ER-associated and one Chlamydia-mediated.
The anti-CD1d 51.1.3 mAb preferentially recognizes a conformational epitope associated with CD1d maturity (
). This allowed us to discriminate the effects of C. trachomatis infection on properly folded mature glycosylated CD1d (51.1.3-reactive) from the effects on total CD1d HCs (NOR3.2-reactive) (Fig. 6). A pseudocolor rendering demonstrated moderate colocalization of CPAF and NOR3.2-reactive CD1d HC (Fig. 6B, bottom left panel) and confirmed our biochemical data showing the physiologic interaction of cytosolic CPAF with the CD1d HC. The distribution of the CPAF-CD1d complex was similar to that of total CD1d HC shown in Fig. 5. In the right panel of Fig. 6A, 51.1.3-reactive mature glycosylated CD1d is again noted throughout the intracellular space and on the cell surface in uninfected cells. By 24 h p.i., mature glycosylated CD1d localized almost exclusively to the perinuclear area (Fig. 6B, right panels). Signals for CPAF also localize to the perinuclear area and in the area of the chlamydial inclusion. Pseudocolor rendering (bottom right panel) indicated that there is no colocalization of CPAF with 51.1.3-reactive mature glycosylated CD1d. These patterns differed significantly from those for NOR3.2-reactive CD1d HC (left panels) and demonstrate that mature glycosylated CD1d is neither bound to CPAF nor associated with the chlamydial inclusion.
This study demonstrates that CD1d molecules are decreased on the surface of C. trachomatis-infected cells, although CD1d mRNA levels were not altered when compared with those in non-infected cells (data not shown). The three described isoforms of CD1d HC protein were here observed in distinctive patterns that depended upon the infection status and the time after C. trachomatis infection. The 45-kDa glycosylated CD1d HC, rather than the 48-kDa mature glycosylated CD1d HC, is the predominant isoform present between 10 and 20 h p.i. This was accompanied by a decrease in β2m protein levels. The reduction in recoverable β2m in C. trachomatis-infected cells has been reported to result from degradation of the transcription factor, RFX5, by CPAF (
). The exact timing of the secretion of CPAF in our experiments may be characteristic of our infected cell type and the infecting C. trachomatis serovar and may therefore differ from the timing seen in similar experiments performed by others. Here, CPAF secretion started at 10 h p.i. and accumulated in the cytosol by 24 h p.i. as demonstrated visually in Fig. 6. Because the effect of CPAF on β2m through RFX5 is very rapid (within 30 min) (
), the appearance of CPAF at 10 h p.i. could result in our observed decrease in β2m between 10 to 20 h p.i. In β2m-deficient cells, surface-expressed CD1d HCs are glycosylated, but their carbohydrate side chains are incompletely modified. These CD1d HCs migrate at ∼45 kDa (
). This effect is consistent with our immunohistochemical data. The delayed exit of immature glycosylated CD1d from the ER should facilitate CPAF-associated direction of CD1d to degradation. Finally cell-surface CD1d was clearly down-regulated in C. trachomatis-infected cells analyzed by flow cytometry.
The classic cellular proteasome requires removal of N-linked glycans from aberrant cytosolic glycoprotein targets prior to their degradation (
). However, in our experiments, lactacystin rescued a 45-kDa CD1d HC that remained sensitive to endoglycosidase-H. This immature glycosylated CD1d HC appears to be degraded without deglycosylation. Because MG132 was not able to rescue this immature glycosylated CD1d isoform, we propose that it is degraded by CPAF rather than by the classic cellular proteasome.
In contrast, the 37-kDa non-glycosylated isoform of CD1d was rescued by both proteasome inhibitors. Because the CPAF-mediated proteolytic pathway can't be inhibited by MG132 (
), the non-glycosylated 37-kDa CD1d HC must be degraded at least partially by the classic cytosolic proteasome. The degradation process for this isoform, including ubiquitination and deglycosylation, suggest it represents an intermediate in the classic proteolytic pathway for degrading aberrant N-linked glycoproteins (
). The C. trachomatis-infected epithelial cells shown here represent the first model system to allow detection of all three isoforms of the CD1d HC. The experimental system suggests that all isoforms may be pathophysiologically relevant.
Immunostaining data support a model in which CPAF binds to immature glycosylated CD1d and non-glycosylated CD1d and dislocates them into the cytosol surrounding the ER and the chlamydial inclusion. Here, it is degraded in a CPAF-dependent manner that can be distinguished from pathways for classic cytosolic proteasomal degradation.
Interestingly, our ubiquitination experiments demonstrated that the amounts of ubiquitinated CPAF-associated proteins were greater after exposure to lactacystin than after exposure to MG132. However, there was no difference in the amount of rescued CPAF between exposure to MG132 and lactacystin (Fig. 3A). These data suggest that CPAF target proteins could be ubiquitinated prior to degradation by a CPAF-associated proteolytic pathway. Ubiquitinated CD1d HC could be degraded by both classic and CPAF-associated pathways, because those were rescued equally by MG132 and lactacystin. In initial descriptions of chlamydial proteasome-like activity, the authors noted the possibility that C. trachomatis could also co-opt cellular cytosolic proteolytic pathways to degrade host transcription factors (
). Our data clearly demonstrate that C. trachomatis not only provides its own mechanism for CD1d degradation but also uses the classic proteolytic pathway for degrading aberrant N-linked glycoproteins to inhibit CD1d trafficking to the cell surface. CPAF is involved in both pathways.
We propose the following model for CD1d proteolysis in C. trachomatis-infected cells (Fig. 7). Cytosolic CPAF interacts with CD1d via cytoplasmic tail of the CD1d HC. This triggers dislocation of the CD1d HC into the cytosol where it is further processed along two distinct pathways. In one, glycosylated HC in the cytosol is ubiquitinated and deglycosylated. Ubiquitinated, deglycosylated CD1d and CPAF are directed toward degradation by the classic cytosolic proteasome. Immature glycosylated CD1d HC is degraded by CPAF-associated mechanisms that are distinct from those of the cytosolic proteasome.
Does C. trachomatis target CD1d HC for degradation as a means to evade immune recognition? In responses to some microbes, the rapid effects of CD1d-restricted NKT cells do not require recognition of microbial specific antigens (
). Certainly, a reduced expression of CD1d at the cell surface could prevent C. trachomatis-infected cells from such an attack. It was recently demonstrated that Kaposi sarcoma-associated herpesvirus reduces cell-surface expression of CD1d via ubiquitination of the CD1d HC on its cytoplasmic tail (
K. Kawana and D. J. Schust, unpublished observation.
Regardless of whether chlamydial antigens are presented by CD1d, the disruption of CD1d expression in C. trachomatis-infected cells may interfere with rapid and essential innate immune responses.
We are grateful to Dr. G. Zhong for his kind gifts of anti-CPAF antibodies and purified CPAF c-fragment proteins, the Morphology and Imaging Core of the Louisiana State University Health Sciences Center Gene Therapy Program for facilitation of the fluorescent deconvolution microscopy, Dr. J. Nichols and J. Niles at University of Texas Medical Branch for facilitation of the flow cytometry studies, to C. D. McGahan and Dr. L. S. Graziadei for editorial assistance, and to Dr. Priscilla Wyrick for critical reading of the manuscript.