|Home | About | Journals | Submit | Contact Us | Français|
Chlamydiae replicate intracellularly within a unique vacuole termed the inclusion. The inclusion circumvents classical endosomal/lysosomal pathways but actively intercepts a subset of Golgi-derived exocytic vesicles containing sphingomyelin (SM) and cholesterol. To further examine this interaction, we developed a polarized epithelial cell model to study vectoral trafficking of lipids and proteins to the inclusion. We examined seven epithelial cell lines for their ability to form single monolayers of polarized cells and support chlamydial development. Of these cell lines, polarized colonic mucosal C2BBe1 cells were readily infected with Chlamydia trachomatis and remained polarized throughout infection. Trafficking of (6-((N-(7-nitrobenz-2-oxa-1, 3-diazol-4-yl) amino)hexanoyl)sphingosine) (NBD-C6-ceramide) and its metabolic derivatives, NBD-glucosylceramide (GlcCer) and NBD-SM, was analyzed. SM was retained within L2-infected cells relative to mock-infected cells, correlating with a disruption of basolateral SM trafficking. There was no net retention of GlcCer within L2-infected cells and purification of C. trachomatis elementary bodies from polarized C2BBe1 cells confirmed that bacteria retained only SM. The chlamydial inclusion thus appears to preferentially intercept basolaterally-directed SM-containing exocytic vesicles, suggesting a divergence in SM and GlcCer trafficking. The observed changes in lipid trafficking were a chlamydia-specific effect because Coxiella burnetii-infected cells revealed no changes in GlcCer or SM polarized trafficking.
Chlamydia trachomatis is the leading bacterial cause of sexually transmitted disease, with over 3 million new cases per year in the USA alone (1) (http://www.cdc.gov/std/stats/tables/table1.htm). Other strains of C. trachomatis are the leading cause of infectious blindness, which remains endemic in parts of Africa and Asia (2). Chlamydia pneumoniae and Chlamydia psittaci are causes of community-acquired pneumonia and specific zoonotic infections, respectively (3).
Chlamydiae are obligate intracellular bacteria that have evolved a unique biphasic developmental cycle. To initiate an infection, the infectious, metabolically dormant form, termed the elementary body (EB), is endocytosed by the host cell and remains within a vesicle termed the inclusion, where it differentiates into a metabolically active, noninfectious reticulate body. Within the inclusion, the organisms acquire essential amino acids, nucleotides and lipids from the host cell (4–6); however, mechanisms for nutrient acquisition are poorly defined.
Early in infection, the inclusion is trafficked to the peri-Golgi microtubule-organizing center, which requires microtubules and dynein (7,8). The chlamydial inclusion does not interact directly with the classical endosomal/lysosomal pathway (9–11) but instead intercepts a subset of Golgi-derived exocytic vesicles containing sphingomyelin (SM) and cholesterol (12–14). Recent studies have identified multivesicular bodies and lipid droplets as additional sources of lipid for chlamydial inclusions (15,16).
In polarized epithelial cells, ceramide is metabolized into glucosylceramide (GlcCer) and SM at the Golgi apparatus and the lipids are exocytosed to the cell surface in a polarized manner. GlcCer is preferentially trafficked apically, while SM is trafficked either equally to both membranes or preferentially basolaterally (17,18). Determining if the chlamydial inclusion preferentially interacts with apical or basolateral directed pathways could help elucidate potential mechanisms of nutrient acquisition and provide important insights into regulation of vesicular trafficking.
Polarized cell models have been used previously to study chlamydial pathogenic processes. These models have utilized polarized endometrial or cervical epithelial cells (immortal and primary) to examine entry and egress (19–22), antibiotic delivery (23–27), effects of hormonal cycle (28,29) and effects of immune cells or cytokines (30–34) on chlamydial development. These previous studies have greatly improved the understanding of chlamydial pathogenesis and biology.
To examine lipid and protein trafficking in polarized cells, several conditions must be satisfied. First, monolayers must form tight junctions, which is indicative of their organization into distinct apical and basolateral domains (35–37). Once a cell is polarized, exocytosis is specifically directed to either the apical or the basolateral membrane, resulting in distinct protein and lipid composition (reviewed in 38). Second, to accurately study the trafficking of a substrate, polarized cells must be able to form a single monolayer of cells. This allows discrete access to a single apical or basolateral surface to follow a substrate in and out of the cell. Finally, it is imperative to develop a model in which normal chlamydial development is supported. In this study, we evaluated seven different cell lines for their suitability as a model to study lipid and protein trafficking and for their ability to support a robust chlamydial infection. Of the cell systems tested, the polarized colonic mucosal epithelial cell line, C2BBe1, best fulfilled the necessary criteria and was utilized to study lipid trafficking in chlamydia-infected cells.
To study polarized trafficking to the chlamydial inclusion, a suitable cell line must exhibit the following characteristics: (i) form tight junctions, indicative of distinct apical and basolateral domains (36,39), (ii) form a single monolayer of cells and (iii) support robust chlamydial growth and development. Seven cell lines were screened for their suitability for these studies. Three endometrial epithelial cell lines, HEC-1-A, and previously described HeLa and HEC-1-B cell lines (20–22,26,34) were included. Human intestinal epithelial cell lines, T84, Caco-2 and C2BBe1 [a clonal derivative of Caco-2 cells (40)], and canine renal epithelial cells, Madin–Darby canine kidney (MDCK1), all of which are commonly used to study lipid and protein trafficking (reviewed in 41,42), were also examined.
Optimal growth conditions on permeable, coated transwell filters for each cell line were determined (Table 1), and transepithelial electrical resistance (TEER) was determined for each cell line for the indicated times (Figure 1). As cells become polarized and tight junctions form, the ability of an electrical charge to pass freely through the monolayers is reduced, which results in an increased TEER (39). Cells were considered weakly polarized at 150 Ωcm2 and fully polarized at 200 Ωcm2(43). Of the seven cell lines examined, five cell lines (HEC-1-A, T84, Caco-2, C2BBe1 and MDCK1) exhibited increased TEER (Figure 1), which is consistent with these cells forming distinct apical and basolateral membrane domains separated by tight junctions. Examination of these cell lines by scanning electron microscopy displayed highly organized microvilli on their apical surfaces (data not shown), consistent with their polarized state. HeLa and HEC-1-B endometrial cells did not exhibit appreciable TEER and, therefore, were eliminated from further study.
Transmission electron microscopy (TEM) was used to evaluate if these cell lines formed a single monolayer of cells (Figure 2). With consideration to the model for intracellular trafficking, a single monolayer of cells permits the ability to track substrates delivered apically or basolaterally. Of the five cell lines examined by TEM, only two cell lines, C2BBe1 and MDCK1, consistently formed single monolayers of cells. These cell lines were then tested for their ability to support C. trachomatis infection.
Polarized or subconfluent C2BBe1 and MDCK1 cells were inoculated with C. trachomatis serovar L2 (L2) for 24 h, and the ability to support chlamydial growth was determined. Subconfluent C2BBe1 and MDCK1 cells supported chlamydial growth similar to HeLa cells grown on coverslips. However, only polarized C2BBe1 cells displayed a similar susceptibility to infection as HeLa cells grown on filters, and polarized MDCK1 cells were resistant to chlamydial infection (Figure 3A). A one-step growth curve analysis revealed that polarized C2BBe1 cells supported normal development of L2 compared with HeLa cells grown on filters (Figure 3B). Infection with L2 does not result in a loss of TEER, indicating that the cells remain polarized during infection (data not shown). Additionally, polarized C2BBe1 cells supported infection with C. trachomatis serovar D, and C. psittaci GPIC, indicating that this cell line is suitable for studying chlamydial pathogenesis of multiple serovars and species (data not shown).
Previous studies in nonpolarized cells have shown that Golgi-derived SM is trafficked to the chlamydial inclusion and retained in the cell walls of the bacteria (13,44,45). At 4°C to 12°C, (6-((N-(7-nitrobenz-2-oxa-1, 3-diazol-4-yl) amino)hexanoyl)sphingosine) (C6-NBD-ceramide) indiscriminately labels biological membranes, but upon temperature shift to 37°C, NBD-ceramide exclusively labels the Golgi apparatus where it is the immediate precursor of GlcCer and SM, which are then transported to the plasma membrane (46). Fluorescent lipid products translocated to the plasma membrane can be quantitatively extracted, or back-exchanged, by lipid acceptors such as defatted BSA in the medium (17). Polarized C2BBe1 cells were infected for 24 h with L2 prior to labeling with NBD-ceramide for 30 min at 12°C. The temperature was then shifted to 37°C, and the C6-NBD-lipid products were back-exchanged into BSA-modified medium for 8 h with one change of medium. Once extracted by BSA, NBD-SM is not recycled to the cell (14). Lipids were then extracted from apical and basolateral media for analysis by thin layer chromatography (TLC). When the cells were examined by fluorescence microscopy after labeling and back-exchange, the majority of the remaining fluorescent lipid was associated with the chlamydial inclusion (Figure 4A), indicating that trafficking to the inclusion is similar between polarized C2BBe1 cells and nonpolarized cells.
When the distribution of lipids between apical, basolateral and cellular fractions was examined in uninfected polarized C2BBe1 cells, the lipids were polarly distributed with GlcCer enriched in the apical membrane (Figure 4B–D). In infected cells, the cells did not retain GlcCer, relative to uninfected cells; however, infected cells displayed subtle, but statistically significant, increased apical trafficking, with a slight decrease in basolateral trafficking of GlcCer (Figure 4B). In contrast, chlamydia-infected cells retained SM, with a coinciding decrease in both apical and basolateral trafficking of SM, with a greater disruption of basolateral trafficking of SM in infected cells (Figure 4C). These results indicate that the chlamydial inclusion intercepts predominately basolaterally trafficked SM-containing vesicles.
The subtle changes observed in GlcCer trafficking in chlamydia-infected cells may be compensatory effects of changes in SM trafficking. To confirm that the inclusions retained only fluorescent SM, polarized C2BBe1 cells were infected with L2 for 24 h, followed by a 14- to 16-h back-exchange to remove unincorporated lipid from the cell. Chlamydial organisms were purified from the polarized cells, and total lipids were extracted for analysis by TLC. Although the cells metabolized NBD-ceramide into both GlcCer and SM, only SM was found in purified L2 EBs (Figure 5). These results confirmed that the SM retained by infected polarized C2BBe1 cells is predominately incorporated into chlamydial organisms, while GlcCer is not, suggesting a divergence of SM and GlcCer trafficking. The ability of purified EBs to incorporate both NBD-GlcCer and NBD-SM, in vitro, implies that it is a lack of GlcCer trafficking to the inclusion that accounts for the absence of GlcCer in EBs rather than an inability to incorporate NBD-GlcCer if it is available (data not shown).
To determine whether inhibition of apically directed NBD-GlcCer trafficking could lead to a redirection of the lipid to the chlamydial inclusion, we examined the effects of 10 µM cyclosporin A (CSA) on NBD-lipid trafficking to the apical and basolateral surfaces as well as to the intracellular chlamydiae. CSA treatment resulted in a 35 ± 5% decrease in apical delivery of NBD-GlcCer but no change in basolateral delivery or alteration of NBD-SM trafficking. Analysis of EBs purified from NBD-ceramide-labeled and CSA-treated C2BBe1 polarized cells indicated no redirection of NBD-GlcCer to the EBs as a result of disruption of apical NBD-GlcCer delivery (data not shown).
As the chlamydial inclusion grows, it physically displaces contents of eukaryotic cells, which could in itself disrupt normal cellular vesicular trafficking. To test the specificity of chlamydial effects on lipid trafficking, Coxiella burnetii Nine Mile phase II was used as a control to study lipid trafficking patterns in cells infected by an obligate intracellular pathogen, which forms a parasitophorous vacuole but interacts with endocytic compartments and does not acquire SM during its life cycle (9). For these studies, polarized C2BBe1 cells were infected with C. burnetii for 72–86 h, labeled with NBD-ceramide and back-exchanged as previously described. The infected cells remained polarized as indicated by TEER measurements >200 Ωcm2 (data not shown). Lipids were then extracted from the apical, basolateral and cellular fractions. In mock-infected cells, GlcCer and SM were again exocytosed in a polarized manner. In comparison of uninfected cells to C. burnetii-infected cells, there were no notable changes in the apical and basolateral trafficking or cellular retention of GlcCer or SM (Figure 6). These studies indicate that the changes seen in lipid trafficking in chlamydia-infected polarized cells are related to organism-specific effects and not simply to the presence of a large parasitophorous vacuole.
The chlamydial inclusion is a unique intracellular compartment that, unlike the parasitophorous vacuoles occupied by most intracellular pathogens, is dissociated from the endocytic pathway but intercepts cellular lipids from a Golgi-to-plasma membrane-directed exocytic pathway. SM and cholesterol trafficking to the inclusion is direct from the Golgi, vesicular in nature, temperature and energy dependent and inhibited by brefeldin A or nocodazole (12–14). Chlamydiae acquire both SM and cholesterol from this pathway, yet no cellular proteins are known to be directly delivered in conjunction with these lipids (9–11,47,48). This pathway is thought to represent a subset of exocytic vesicles whose properties are undefined. Although these lipids do not appear to be further metabolized, the incorporation of sphingolipids is important for chlamydial development (49). Because all previous studies were performed in nonpolarized cells, we sought to develop a polarized cell model system to test whether the chlamydial inclusion might preferentially interact with apically or basolaterally directed lipids in an effort to better define the properties of this pathway. Of seven cell systems tested, only one cell line fulfilled all necessary criteria to study lipid sorting and trafficking in chlamydia-infected cells. The polarized colonic mucosal epithelial cell line, C2BBe1, formed a single monolayer with tight junctions and TEER and supported normal chlamydial development. Primary mucosal intestinal epithelial cells, obtained from calf abdominal viscera, had been used previously to examine chlamydial entry and egress (50). The polarized C2BBe1 cell model was utilized in this study to examine lipid trafficking in L2-infected cells. The chlamydial inclusion appears to preferentially intercept and retain SM from basolaterally targeted exocytic vesicles. Consistent with previous observations of C6-NBD-lipid trafficking in infected cells, these changes in lipid trafficking were specific to chlamydiae.
An epithelial cell barrier is often the first line of host defense that potential pathogens must breach. Polarized cell models more closely resemble the in vivo situation encountered by pathogenic microbes thus are highly relevant to studies of pathogenesis (42,51). Infections with most ocular and urogenital strains of C. trachomatis are restricted to the mucosal epithelium. In contrast, after the initial developmental cycle, lymphogranuloma venereum (LGV) strains egress from the epithelial barrier, can survive within macrophages and replicate within subepithelial and lymphatic tissues to cause lymphadenopathy (52,53). The physiological relevance of polarized cells has proven useful in clarifying otherwise confusing aspects of chlamydial biology, such as entry (19–21,26). With respect to chlamydial pathogenesis, a comparison of entry into polarized and nonpolarized epithelia demonstrate that EBs entering polarized cells were more often associated with clathrin-coated pits than in nonpolarized cells (21). Moreover, it has been proposed that the more invasive LGV strains of C. trachomatis preferentially exit through the basolateral domain consistent with their access to the submucosa (20). Access to the basolateral surface has also been proposed to place the LGV EBs at a site enriched in heparan sulfate, which enhances attachment of LGV strains more so than non-LGV biovars (20). The unique properties of the C2BBe1 cells used in this study may prove useful in defining additional characteristics of chlamydia–host cell interactions. Interestingly, polarized MDCK1 cells were found to be resistant to L2 infection either at the stage of entry or development. This cell line also may provide insights into chlamydial biology in a system where access to apical or basolateral membranes can be tightly controlled and manipulated.
Ceramide is the immediate biosynthetic precursor of SM and GlcCer, which are synthesized in the cis Golgi (54,55). These lipids are then believed to be transported to the plasma membrane by a vesicular process in which they are oriented toward the luminal surface of the vesicle such that they are exposed on the outer leaflet of the plasma membrane upon fusion (56–58). In polarized cells, GlcCer and SM are transported asymmetrically such that GlcCer is enriched on the apical surface, whereas SM is equally distributed or slightly enriched on the basolateral surface (17,18). GlcCer and SM are sorted into two distinct vesicle populations (59). Recovery of exocytic vesicles followed by differential centrifugation revealed that the GlcCer was recovered from lower density fractions and SM from higher density fractions (59). In support of these findings, reduced temperature abrogated the transport of GlcCer but not of SM (60). Furthermore, in polarized hepatic cells, trifluoroperazine treatment blocked SM exit from the subapical compartment, while GlcCer transport remained intact (61). Exact mechanisms for post-Golgi sorting of sphingolipids in polarized epithelial cells remain largely undefined. We found that the chlamydial inclusion interrupted preferentially basolaterally trafficked SM. When EBs were purified for lipid analysis after ceramide labeling, however, only NBD-SM but not NBD-GlcCer was associated with the intracellular bacteria. The chlamydial inclusion may thus serve as a unique target organelle to study vectoral lipid trafficking.
While endogenous SM is believed to be transported to the plasma membrane on the luminal surface of exocytic vesicles (62,63), several studies have implicated a role for the multidrug resistance transporter, Mdr1 or p-glycoprotein, in the delivery of short-chain SM derivatives, such as C6-NBD-SM, to the outer leaflet of the plasma membrane (64,65). The widely studied transporter Mdr1 or p-glycoprotein is localized mainly to the apical membranes of polarized cells (66) and functions, in part, to translocate GlcCer across the apical plasma membrane (65,67). Further studies have indicated that Mdr1 functions within the Golgi apparatus, where it is linked to the recycling of SM or GlcCer and exposure of lipid to the luminal or cytosolic face of budding vesicles (68). It is unknown whether a transporter is involved in chlamydial acquisition of SM or other metabolites; however, treatment of polarized cells with CSA or verapamil, which are inhibitors of Mdr1 function, did not inhibit NBD-SM acquisition, whereas inhibitors of vesicle transport, such as 15°C, brefeldin A or monensin, decreased chlamydial acquisition of C6-NBD-SM in either HeLa cells (12,14) or polarized C2BBe1 cells (unpublished data). Additionally, Mdr1 was not detected on the chlamydial inclusion by immunofluorescence analysis (unpublished data). While the Mdr1 transporter appears to be important for C6-NBD-GlcCer transport to the apical surface of intestinal epithelial cells (64,65), inhibitors of Mdr1 had little effect on delivery of C6-NBD-SM to chlamydiae. Furthermore, inhibition of Mdr1 did not lead to a redirection of C6-NBD-GlcCer to the chlamydial inclusion. Perhaps, the dependence of C6-NBD-GlcCer trafficking on multidrug transporters partially explains the distinction in SM and GlcCer trafficking to the chlamydial inclusion. Further study will be required to fully assess any indirect role of Mdr1 in mediating SM acquisition or any role of chlamydiae in modulating Mdr1 activity.
The current model suggests that the chlamydial inclusion intercepts a subset of basolaterally targeted exocytic vesicles. At one time, it was believed that lipids and proteins were trafficked basolaterally by default because they were lacking apical targeting signals (69). However, more recent studies have elaborated on SNARE proteins, which, in part, control apical and basolateral membrane fusion events. Specific SNARE proteins are associated with apical (VAMP7/TI-VAMP, VAMP8 and syntaxin 3) and basolateral (VAMP3 and syntaxin 4) trafficking pathways (70). Additionally, there are cellular proteins expressed exclusively in polarized epithelial cells (AP-1B) that control basolateral trafficking in collaboration with the exocyst complex, which is composed of eight proteins that collectively control aspects of exocytosis (71). In polarized cells, these pathways can be specifically examined in a manner not possible by standard tissue culture models, which may allow unique insights into chlamydial inclusion–host vesicle fusion events. Because SM trafficking to the chlamydial inclusion is chlamydia specific and not because of a generalized dysregulation of trafficking in infected cells, it is likely that chlamydial protein(s) modifies the inclusion membrane to mimic or recruit cellular vesicle fusion machinery to the inclusion. An understanding of the chlamydial and host factors involved may lead to novel means to disrupt chlamydial development.
Although it is unclear how the inclusion membrane recruits specific host cell vesicles, chlamydial inclusion membrane proteins (Incs) are good candidates to facilitate such interactions. There are over 40 predicted and known C. trachomatis inclusion membrane proteins (72–75). Of those examined in detail for exposure to the cytoplasm, all were found to be exposed on the cytoplasmic face of the inclusion membrane (76,77). The fact that Inc proteins must be synthesized de novo prior to insertion into the inclusion membrane in conjunction with the requirement for chlamydial protein synthesis to initiate fusion with SM-containing vesicles further implicates this class of proteins (77,78). IncA, which is required for homotypic inclusion fusion (76), has also been shown to form SNARE-like complexes with itself (79). Whether IncA can interact with host cell proteins in this manner is currently unknown. Rab proteins are also involved in regulating trafficking and vesicle fusion (reviewed in 80). Previous studies have established that certain Rab proteins localize to the chlamydial inclusion (81) through interactions with specific Inc proteins (82). Rab11, which localizes to C. trachomatis, Chlamydia muridarum and C. pneumoniae (81), along with Rab25, is involved in apical recycling pathways (83,84). Although a role for Rab8 in chlamydial trafficking has yet to be described, this protein has been linked to modulating basolateral trafficking with the exocyst complex in polarized MDCK II cells (71). The mechanism of chlamydial recruitment of Rab guanosine triphosphatases may be a result of the inclusion targeting specific vesicles or to camouflage the inclusion from other subcellular compartments (85). Although fusion with SM-containing vesicles is common to all strains and species of chlamydiae, there is surprising heterogeneity in the Inc proteins encoded by the different chlamydial species. This follows a distinct pattern of Rab protein recruitment to the different species. One would expect that common themes in host protein recruitment might provide insights into the regulation of vesicle fusion with the chlamydial inclusion.
In this study, we have defined a polarized cell model to study lipid trafficking in chlamydial infected cells. This polarized cell model of infection may also have utility in the study of other processes in chlamydial pathogenesis. Previous studies have shown that the chlamydial inclusion retains Golgi-derived SM (13). In this study, we demonstrate that in polarized C2BBe1 cells, the inclusions retrieved the SM preferentially from a basolateral pathway. The polarized cell model described in this study is a powerful tool to investigate specific apical and basolateral trafficking pathways and their interaction with the chlamydial inclusion to further define elusive mechanisms of chlamydial interactions with host vesicular trafficking pathways.
HeLa 229 cells [American Type Culture Collection (ATCC; CCL-2.1], cultivated in RPMI-1640 (Gibco-BRL) supplemented with 10% FBS (Hyclone) and 10 µg/mL gentamicin (Gibco-BRL), were used to propagate C. trachomatis serovar L2 (LGV 434), serovar D (UW-3/CX) and C. psittaci (C. caviae) GPIC (HC/BW). Infectious EBs were purified from HeLa cells using a Renografin (Braco Diagnostics) gradient as previously described (86). Chlamydial titers were determined as described previously (13,87) by utilizing indirect immunofluorescence with a polyclonal rabbit anti-C. trachomatis L2 EB, followed by an anti-rabbit Alexa fluor 488-conjugated secondary antibody [Invitrogen (Molecular Probes)]. Multiplicities of infection (MOI) for all experiments are based on inclusion forming units determined in HeLa cells. C. burnetii Nine Mile phase II was provided by Bob Heinzen (Rocky Mountain Laboratories, Hamilton, MT, USA) and propagated and purified from Vero cells (ATCC; CCL-81) as previously described (88). MOIs are based on genome equivalents of C. burnetii Nine Mile phase II purified from Vero cells.
All eukaryotic cell lines were cultured at 37°C in 5% CO2. Caco-2 (ATCC HTB-37) and HEC-1-B (ATCC HTB-113) cells were grown in MEM with Earle’s Salts (ATCC) supplemented with 10% FBS, 2 mm L-glutamine (ATCC), nonessential amino acids and 10 µg/mL gentamicin. T84 (ATCC; CCL-248) and MDCK1 cells were grown in DMEM:F12 (1:1) (ATCC) medium supplemented with 5% FBS, 2.5 mm l-glutamine and 10 µg/mL gentamicin. C2BBe1 (ATCC CRL-2102) cells were cultivated in DMEM + 2 mm GlutaMax (Invitrogen) supplemented with 10% FBS, 4 mm l-glutamine, 0.01 mg/mL human transferrin (Invitrogen) and 10 µg/mL gentamicin. HEC-1-A (ATCC HTB-112) cells were cultivated in McCoy’s 5A medium (Gibco) supplemented with 10% FBS, 1.5 mm l-glutamine and 10 µg/mL gentamicin. All eukaryotic cells were passaged based on ATCC-suggested protocols using a 0.25% trypsin and 0.53 mm ethylenediaminetetraacetic acid solution (ATCC).
To screen cell lines for their ability to form polarized monolayers, eukaryotic cells were seeded on 6.5-mm-coated inserts (BD Biosciences) in 24-well plates (Table 1). After 24 h of culture, TEER was measured daily using chopstick electrodes and an EVOM epithelial voltohmmeter (World Precision Instruments). TEER measurements were routinely taken during all experiments to ensure that cells reached and maintained polarity. Optimal insert substrate and seeding density were determined for each cell line; details are provided in Table 1.
Specimens were fixed for 1 h at 4°C with 2.5% glutaraldehyde/4% paraformaldehyde in 0.1 m sodium cacodylate buffer, pH 7.4. Samples were post-fixed for 1 h with 0.5% osmium tetroxide/0.8% potassium ferricyanide and stained overnight with 1% uranyl acetate at 4°C, dehydrated with a graded ethanol series and embedded in Spurr’s resin. Thin sections were cut with an RMC MT-7000 ultramicrotome (Ventana), stained with 1% uranyl acetate and Reynold’s lead citrate prior to viewing at 80 kV on a Philips CM-10 transmission electron microscope (FEI). Digital images were acquired with an AMT digital camera system (AMT).
To visualize chlamydial inclusion retention of fluorescent lipid, C2BBe1 cells were polarized for 4 days on 6.5-mm filters and mock infected or infected with C. trachomatis serovar L2 (MOI 20) for 18–24 h prior to labeling with 5 µm NBD-C6-ceramide (Invitrogen) in Eagle’s Modified Essential Medium (EMEM) plus 0.034% defatted BSA (Sigma) for 40 min at 37°C and 5% CO2. Unincorporated lipids were removed from cells by back-exchange with 5% FBS in EMEM for 7 h. Filters were removed from the insert housing, washed in HEPES-buffered (calcium- and magnesium-free) Puck’s saline and mounted onto a slide. Samples were visualized using a LSM 510 laser module on a Zeiss Axiovert 200M confocal microscope (Carl Zeiss MicroImaging, Inc.).
To examine lipid trafficking, C2BBe1 cells were cultured for 4 days on 24-mm filters until polarized and mock infected or infected with C. trachomatis serovar L2 (MOI 40–50) or C. burnetii Nine Mile phase II (MOI 300–500:1) as described above. After a 24-h infection (L2) or 72- to 88-h infection (C. burnetii), cells were labeled with 5 µm NBD-C6-ceramide as described above. The temperature was shifted to 37°C for 15 min prior to the addition of back-exchange medium, EMEM plus 0.7% defatted BSA. After 4 h of back-exchange at 37°C, apical and basolateral fractions were collected, and fresh back-exchange medium was added. After an additional 4-h back-exchange, additive apical and basolateral fractions were collected. Cellular fractions were obtained by removing whole filters from the insert housing. To ensure equal loading, 0.005 mg/mL NBD-cholesterol (Invitrogen) and/or BODIPY-lactosylceramide (Invitrogen) was added to each sample. Lipids were extracted by Bligh and Dyer chloroform:methanol extraction (89) and dried under a stream of N2. The samples were resuspended in 2:1 chloroform:methanol, spotted onto Silica Gel 60 Å precoated TLC plates (Whatman or Alltech) and resolved using 32.5:12:2 chloroform:methanol:dH2O. The air-dried plates were visualized with ultraviolet light, and densitometry was performed using a Typhoon Phosphorimager and Image Quant (v 5.0) software (Amersham Biosciences). For each sample, the density of SM or GlcCer was determined and normalized to one of the loading controls.
C2BBe1 cells were polarized on 24-mm filters, infected and labeled with NBD-ceramide as described above. Then infected, labeled cells were incubated with EMEM supplemented with 5% FBS for 14–16 h at 37°C. EB purification (86), lipid extraction and sample resolution were performed as described above.
Mean and standard deviation or standard error of the mean were calculated and graphed using GraphPad Prism v. 4.0 (Graph Pad Software). Two-way anova analysis with Bonferroni post tests were used to determine statistical significance (p < 0.001) (GraphPad software).
The authors would like to thank Travis Jewett, Jeff Mital, Olivia Steele-Mortimer and Bob Heinzen for critical review of the manuscript, Janet Sager for technical support and members of the Hackstadt laboratory for helpful comments and suggestions. This study was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases (NIAID) of the National Institutes of Health.