Trafficking from CD63-positive late endocytic multivesicular bodies is essential for intracellular development of Chlamydia trachomatis

Chlamydiae are obligate intracellular bacterial pathogens that replicate solely within the confines of a membrane-bound vacuole termed an inclusion. Within this protected organelle, chlamydiae acquire host-cell-derived biosynthetic precursors necessary for intracellular subsistence, yet the mechanisms and pathways responsible for this acquisition remain elusive. The present study identifies an interaction between the chlamydial inclusion and multivesicular bodies, complex organelles pivotal in protein and lipid transport that are positioned along the endosome-lysosome pathway, and intersect the exocytic pathway in various cell types. Resident protein and lipid constituents of multivesicular bodies colocalized with intracellular chlamydiae, with direct delivery of the resident protein CD63 to the chlamydial inclusion. Interruption of trafficking from multivesicular bodies by pharmacological inhibitors and exogenous antibodies subsequently disrupted sphingolipid delivery to the maturing chlamydial inclusion and intracellular bacterial growth. This study identifies a trafficking pathway from CD63-positive multivesicular bodies to the bacterial inclusion, a novel interaction that provides essential lipids necessary for maintenance of a productive intracellular infection.


Introduction
Chlamydiae are etiological agents of several human diseases of medical significance. C. trachomatis is the most commonly encountered sexually transmitted pathogen in the United States and developed countries, and worldwide is the leading cause of infectious blindness (Schachter, 1988). The spectrum of disease manifestations caused by Chlamydiae also encompasses C. pneumoniae, a causative agent of acute respiratory infections and implicated as a potential risk factor of cardiovascular disease (Grayston, 2000).
Chlamydiae are obligate intracellular pathogens that reside exclusively in a membrane-bound organelle called the inclusion. Chlamydiae have evolved a unique biphasic developmental cycle alternating between two distinct morphological forms. Infection of a target host cell is initiated by the metabolically inactive elementary body (EB), which upon internalization differentiates into the noninfectious, metabolically active reticulate body (RB). After multiple divisions within the confines of the inclusion, the RBs redifferentiate back to EBs. Lysis of the inclusion membrane and the host cell results in the release of a multitude of infectious EBs, each capable of eliciting a new round of infection.
The mechanisms that allow chlamydiae to remain sequestered in a protected intracellular niche remain enigmatic. The use of contemporary cell-biological probes for early and late endocytic markers (transferrin and M6PR) Scidmore et al., 2003), as well as lysosomal markers and endocytic tracers (LAMP-1, LAMP-2, Cathepsin D, ATPase, dextran, and lucifer yellow) Scidmore et al., 2003;Taraska et al., 1996) have failed to demonstrate an effective interaction between host pathways and the chlamydial inclusion. However, within the confines of the inclusion, chlamydiae acquire amino acids and nucleotides, as well as eukaryotic lipids that are transported from the host cell and incorporated into the chlamydial inclusion (Hatch, 1975;McClarty, 1994;Wylie et al., 1997). Chlamydiae obtain host cell sphingomyelin and cholesterol from Golgi-derived vesicles destined for the plasma membrane; however, this pathway of acquisition is neither exclusive nor essential for inclusion development (Carabeo et al., 2003;Hackstadt et al., 1996). Acquisition of host cell protein and lipid constituents by other mechanisms or pathways remains elusive.
In the present study, the interaction between the chlamydial inclusion and multivesicular bodies (MVBs) was analyzed. MVBs are complex, heterogeneous late endocytic organelles, seminal in the segregation of proteins and lipids destined for lysosomal degradation, recycling to the Golgi, or plasma membrane exocytosis (Denzer et al., 2000;Piper and Luzio, 2001). Internal vesicles formed by invagination of the limiting membrane of MVBs provide the scaffold for transport and delivery of its contents (Pan et al., 1985;Vidal et al., 1997). These internal membranes are highly enriched in lysobisphosphatidic acid (LBPA) or phosphatidylinositol 3-Chlamydiae are obligate intracellular bacterial pathogens that replicate solely within the confines of a membranebound vacuole termed an inclusion. Within this protected organelle, chlamydiae acquire host-cell-derived biosynthetic precursors necessary for intracellular subsistence, yet the mechanisms and pathways responsible for this acquisition remain elusive. The present study identifies an interaction between the chlamydial inclusion and multivesicular bodies, complex organelles pivotal in protein and lipid transport that are positioned along the endosome-lysosome pathway, and intersect the exocytic pathway in various cell types. Resident protein and lipid constituents of multivesicular bodies colocalized with intracellular chlamydiae, with direct delivery of the resident protein CD63 to the chlamydial inclusion. Interruption of trafficking from multivesicular bodies by pharmacological inhibitors and exogenous antibodies subsequently disrupted sphingolipid delivery to the maturing chlamydial inclusion and intracellular bacterial growth. This study identifies a trafficking pathway from CD63-positive multivesicular bodies to the bacterial inclusion, a novel interaction that provides essential lipids necessary for maintenance of a productive intracellular infection.
phosphate with variable membrane compositions defining a diverse population (Gillooly et al., 2000;Kobayashi et al., 1998). LBPA-rich membranes are thought to play a role in sphingolipid hydrolysis and regulation of intracellular cholesterol transport (Kobayashi et al., 1999;Kolter et al., 1999). The tetraspanin protein CD63 is abundant in MVBs and cycles between these compartments and the exocytic pathway (Kobayashi et al., 2000). Restricted to the limiting membrane of CD63-and LBPA-positive late endosomes is the transmembrane protein metastatic lymph node 64 (MLN64) (Alpy et al., 2001). MLN64 shares homology with the StARrelated lipid transfer (START) domain of the steroidogenic acute regulatory protein (StAR) (Watari et al., 1997). The MLN64 START domain binds cholesterol and stimulates the mobilization of free cholesterol from sterol-rich donor vesicles to acceptor membranes (Kallen et al., 1998).
In this study, candidate MVB-related proteins and lipids were evaluated for their association with intracellular chlamydiae. Demonstration of an intersection between the chlamydial inclusion and MVB-related pathways, identifies these dynamic host cell compartments as a novel source for factors essential for intracellular bacterial growth and inclusion biogenesis.

MVB constituents localize to the chlamydial inclusion in infected HEp-2 cells.
To evaluate an association between the chlamydial inclusion and MVBs, the distribution of MVB markers was analyzed in C. trachomatis-infected cells by confocal microscopy. CD63 and MLN64, protein constituents of MVBs, and LBPA (a lipid highly enriched in internal vesicles of MVBs) all showed specific localization at the site of the chlamydial inclusion ( Fig. 1). This association was demonstrated in HEp-2 cells infected with C. trachomatis E and labeled by indirect immunofluorescence with antibodies specific to these MVB markers. Confocal analysis of midplane Z-sections through the center of the inclusion confirmed the internal location of all MVB markers. Labeling with antibodies specific to CD63, MLN64 and LBPA all revealed a fine punctate pattern at the site of the chlamydial inclusion ( Fig. 1), which colocalized with TOPRO-3-labeled bacteria within the inclusion. Immunolabeling with anti-LAMP-1 antibody revealed that LAMP-1 was excluded from the inclusion, therefore serving as a negative control. In addition, TOPRO-3-labeled host cell nuclei do not specifically label with antibodies to MVBs and provided in each image an internal control for the level of background label.
Orthogonal analysis confirms the presence of CD63 within the chlamydial inclusion Immunolabeling studies indicated that protein and lipid constituents of MVBs colocalize with chlamydial inclusions. Potential trafficking of MVBs or MVB-related compartments to the chlamydial inclusion was further evaluated using CD63 as a surrogate marker for this interaction. HEp-2 cells were infected with C. trachomatis E for 48 hours and analyzed by indirect immunofluorescence. To precisely identify the boundary of the chlamydial inclusion, cells were labeled with antibody specific to the cytosolic C-terminal domain of the inclusion membrane protein IncA, using conditions that result in labeling of the inclusion membrane exclusively (see Materials and Methods). The inclusion membrane was then detergentpermeabilized and cells were labeled with mouse antibody specific to CD63. Confocal microscopy clearly identified the presence of CD63 within the inclusion (Fig. 2, upper panel) and this colocalized with TOPRO-3-labeled Chlamydia. Orthogonal views unequivocally demonstrated localization of CD63 with TOPRO-3-labeled Chlamydia within the confines of the IncA-labeled inclusion membrane (Fig. 2, lower panel).

Immunoelectron microscopy localizes CD63 to the chlamydial inclusion
The presence of CD63 in the chlamydial inclusion was further investigated at the ultrastructural level using immunoelectron microscopy. Cryo sections of cells that had been infected with Chlamydia for 48 hours and labeled with anti-CD63 antibody revealed the presence of CD63 in multivesicular organelles adjacent to the chlamydial inclusion (Fig. 3). In addition, immunolabeling clearly showed CD63 within the chlamydial inclusion and associated with bacterial membranes. Discrete patches of CD63 were evident in the inclusion membrane itself, suggesting a direct fusion of CD63-positive vesicles with the inclusion (Fig. 3, upper panel). In addition, CD63-positive vesicles were present within the inclusion (Fig. 3, lower panel), further implicating a direct fusion of MVBs with the chlamydial inclusion with release of their internal vesicles into the lumen of the inclusion. Quantitative analysis by immunoelectron microscopy revealed that the cross-sectional labeling-density of anti-CD63 was more than tenfold and eightyfold per area of the inclusion and MVBs, respectively, over background (over labeling of mitochondria and nuclei).

CD63-HA localizes to the chlamydial inclusion in transfected cells
Localization of HA-tagged CD63 in transfected cells confirmed the distribution of CD63 and localization to the chlamydial inclusion. HEp-2 cells were transfected with CD63-HA for 24 hours, infected with C. trachomatis E for 48 hours, and subsequently labeled by indirect immunofluorescence using antibodies specific against HA. In transfected cells, anti-CD63 and anti-HA colocalized perfectly, confirming the correct targeting of the HA-fusion protein to CD63-positive compartments (data not shown). Confocal analysis of midplane Z-sections through the center of the Journal of Cell Science 119 (2) inclusion revealed a fine punctate pattern of CD63-HA in transfected cells (Fig. 4,left panel), that colocalized with TOPRO-3-labeled bacteria within the inclusion. No labeling was evident in the inclusion of untransfected cells, which served as a negative control for anti-HA labeling.
Immunoelectron microscopy analysis of CD63-HAtransfected cells further confirmed the presence of CD63 within the chlamydial inclusion. Labeling of cryo sections of Chlamydia-infected cells with anti-HA antibody revealed the presence of CD63-HA in multivesicular organelles adjacent to the chlamydial inclusion (Fig. 4, right panel). In addition, immunolabeling clearly showed CD63-HA within the lumen of the inclusion and often associated with vesicles ( Fig. 4, right lower panel insert), further implicating a direct fusion of MVBs with the chlamydial inclusion.
Internalized exogenous anti-CD63 antibody traffics to the chlamydial inclusion and perturbs intracellular chlamydial development MVBs lie within the endocytic pathway providing a direct route for exogenously added tracers. This route can be exploited by following the fate of endocytosed antibodies that will selectively accumulate after binding to an antigen.  . Immunoelectron microscopy analysis of CD63 association with the chlamydial inclusion. HEp-2 cells were infected with C. trachomatis E for 48 hours and analyzed by cryo immunoelectron microscopy. Infected cells immunolabeled with anti-CD63 antibody (anti-mouse-18 nm colloidal gold) revealed CD63 within the chlamydial inclusion and in compartments adjacent to the inclusion that morphologically resembled MVBs (arrows). CD63-labeling was also evident along the inclusion membrane (upper panel, arrowhead; enlarged in right panel) and in small vesicles within the inclusion lumen (lower panel arrowheads; enlarged in right panel). C, Chlamydia; m; mitochondria. Bar, 0.5 m. For quantification of anti-CD63-labeling, the density of colloidal-gold particles was determined per cross-sectional area of the compartments indicated (particles/m 2 ). The density was also expressed as fold over background, based on the average background-labeling in mitochondria and nuclei.
Previous studies have demonstrated the accumulation of internalized anti-CD63 antibody in late endocytic compartments, with subsequent trafficking to secretory compartments (Kobayashi et al., 2000). Exogenously internalized anti-LBPA antibody traffics to late endocytic compartments, blocking subsequent transport (Kobayashi et al., 1999). In this study, the same approach was used to follow Transfected cells immunolabeled with anti-HA antibody (anti-rabbit-18 nm colloidal gold) revealed CD63-HA within the chlamydial inclusion and in compartments adjacent to the inclusion that morphologically resembled MVBs (arrows). Labeled CD63 was also evident in small vesicles within the inclusion lumen (lower panel, arrowhead; enlarged in insert). C, Chlamydia. The lack of label in mitochondria (m) served as an internal control for background labeling. Bar, 0.5 m. the fate of anti-CD63 antibodies in the Chlamydia-infected cells. Exogenously internalized anti-CD63 antibody accumulated in perinuclear vesicular compartments and in the chlamydial inclusion (Fig. 5). When anti-CD63 was added simultaneously with anti-LAMP-1 antibody or the fluid-phase marker dextran-FITC, anti-CD63 localized to the chlamydial inclusion, while anti-LAMP-1 and dextran were clearly excluded (Fig. 5). The distribution of these markers was shown quantitatively by generating a profile of the intensity distribution of each constituent along a line traversing the chlamydial inclusion (Fig. 5, lower panels). Anti-CD63 antibody clearly colocalized with TOPRO-3-labeled chlamydiae within the profile of the inclusion, while being excluded from the TOPRO-3-labeled host cell nuclei (Fig. 5, lower panels). Anti-LAMP-1 antibody and dextran-FITC displayed high-intensity values peripherally to the inclusion and the host cell nuclei but were clearly excluded from the inclusion itself, as revealed by backgroundintensity levels across this region of the profile (Fig. 5, lower panels). Orthogonal analysis of the entire Z-height of infected cells confirmed the presence of exogenous anti-CD63 within the inclusion (data not shown).
The effect of direct delivery of internalized anti-CD63 antibody on intracellular bacterial development was analyzed. Chlamydia-infected cells, in the absence of exogenous antibody, developed large inclusions containing primarily electron-dense elementary bodies at 48 hours post infection (Fig. 6, upper left panels). In the presence of exogenous anti-CD63 antibody, smaller, less dense inclusions were evident by fluorescence using TOPRO-3 labeling. Electron microscopy revealed that these inclusions contained predominantly RBs (0.6-1 m), indicating an interruption in normal chlamydial maturation (Fig. 6, lower panels). In addition, the treatment of cells with exogenous antibody to CD63 resulted in a marked reduction in infectious progeny when analyzed at 48 hours post infection (Fig. 6, right panel). Equivalent concentrations of anti-LAMP-1 antibodies served as a negative control, with no effect on inclusion development or infectivity (data not shown). The inhibition of chlamydial development by exogenous CD63-specific antibody implies a role for the host protein CD63 or CD63positive compartments in inclusion biogenesis and maturation.

Inhibitors of MVB biogenesis disrupt chlamydial growth and development
Inhibitors of phosphoinositide 3-kinase (PI 3-kinase) disrupt the biogenesis of MVBs, resulting in swollen organelles that are impaired in membrane trafficking out of this compartment (Fernandez-Borja et al., 1999). The pharmacological agent U18666A inhibits cholesterol transport from late endocytic compartments (Liscum and Faust, 1992) and alters the trafficking of MVB-associated membrane proteins (Higgins et al., 2001;Kobayashi et al., 1999;Kobayashi et al., 2000). In the presence of U18666A and the PI 3-kinase inhibitors 3-Journal of Cell Science 119 (2) methyladenine and LY-294002, maturation of the chlamydial inclusion was disrupted with a marked reduction in inclusion size and yield of infectious progeny (Fig. 7). Electron microscopy of the intracellular bacteria revealed primarily EBs in the untreated control cells but small inclusions that contain typical RBs (0.6 m-1 m) in inhibitor-treated cells, indicating an interruption in RB division and re-differentiation under these conditions. Infected cells treated with MVB inhibitors for 48 hours do not have detectable MVB constituents in their inclusions (data not shown). The concentrations of the inhibitors that disrupted inclusion development also induced swollen CD63-positive MVBs (data not shown), revealing a correlative inhibitory effect on MVB biogenesis and inclusion development. The disruption in inclusion maturation implies an essential role of MVBs or MVB-associated compartments in normal intracellular development of chlamydiae.

Inhibitors of MVB biogenesis disrupt sphingomyelin trafficking to the chlamydial inclusion
The fluorescent sphingolipid-analogue NBD-ceramide is a marker for the Golgi complex (Lipsky and Pagano, 1985b). Within the Golgi, ceramide is processed to sphingomyelin and glucosylceramide before being transported to the plasma membrane (Lipsky and Pagano, 1983). Hackstadt and colleagues used this fluorescent probe to demonstrate endogenous synthesis and transport of sphingomyelin from the Golgi complex to the inclusion and incorporation into the cell wall of chlamydiae (Hackstadt et al., 1995). Because MVBs are pivotal in intracellular sphingolipid transport, the role of these compartments in sphingomyelin delivery to the chlamydial inclusion was analyzed. HEp-2 cells, infected with C. trachomatis for 18 or 48 hours, incorporated the fluorescent probe into both the Golgi complex and the chlamydial inclusion (Fig. 8, left panels). Culturing in the presence of the MVB inhibitors 3-methyladenine and U18666A for 48 hours resulted in smaller inclusions, which failed to incorporate the fluorescent probe (Fig. 8, right panels). However, these cells displayed intense accumulation of the fluorescent probe in swollen perinuclear CD63-positive compartments. The disruption of sphingomyelin transport to the chlamydial inclusion by inhibition of trafficking from MVBs, identifies CD63-positive MVBs within a novel transport pathway of lipid delivery to intracellular chlamydiae.

Discussion
MVBs are dynamic, heterogeneous, intermediate endocytic organelles important in segregating host-cell protein and lipid constituents (for reviews, see Denzer et al., 2000;Piper and Luzio, 2001). Two important features of MVBs make these compartments compelling candidates for an interaction with the chlamydial inclusion. First, MVBs are central to the intracellular trafficking of sphingolipids and cholesterol, two host-derived constituents acquired by the chlamydial inclusion (Carabeo et al., 2002;Hackstadt et al., 1995;Wylie et al., 1997). Second, MVBs are an active component of the exocytic pathway in various cell types (for reviews, see Denzer et al., 2000;Stoorvogel et al., 2002), the only intracellular trafficking pathway that has been shown to intersect the chlamydial inclusion Hackstadt et al., 1995). The present study demonstrates a dynamic interaction between MVB trafficking pathways and the chlamydial inclusion, essential for intracellular bacterial development. Rerouting of MVB-derived exocytic compartments to the chlamydial inclusion might provide an essential source of biosynthetic macromolecules vital to intracellular chlamydial development.
A hallmark of MVBs is the presence of lipid-enriched internal membranes, which have been proposed to be important in protein sorting and transport of cholesterol and sphingolipids (Denzer et al., 2000;Kobayashi et al., 1999;Piper and Luzio, 2001). Constituents of MVBs include LBPA, a lipid abundant in internal membranes of MVBs (Kobayashi et al., 1998), which is thought to play a role in sphingolipid hydrolysis and regulation of cholesterol transport (Kobayashi et al., 1999). The tetraspanin protein CD63 is present in internal membranes and at the limiting membrane of MVBs (Escola et al., 1998;Kobayashi et al., 2002). MLN64 is a cholesterol-binding protein restricted to the limiting membrane of CD63-and LBPA-positive late endosomes (Alpy et al., 2001).
Using a cell-biological approach, MVB markers, in addition to tracers and inhibitors, were used to analyze the interaction of these dynamic compartments with the chlamydial inclusion. This study demonstrates that CD63, MLN64 and LBPA localize to the chlamydial inclusion, as shown by indirect immunofluorescence (Fig. 1). Because the chlamydial inclusion has been proposed to intersect an exocytic pathway Hackstadt et al., 1995), CD63 -a host protein associated with both MVBs and the exocytic pathway (Kobayashi et al., 2000) -was used as a surrogate marker for additional characterization. Electron microscopy (Fig. 3) and orthogonal analyses obtained from confocal 3D-reconstruction (Fig. 2) substantiated the accumulation of CD63 within the chlamydial inclusion. Localization of HA-tagged CD63 in infected cells confirmed the distribution of CD63 to the chlamydial inclusion (Fig. 4).
The potential importance of MVB-associated CD63 was further explored in experiments based on internalization of exogenous CD63-specific antibodies. Previous studies have shown that endosomal CD63 protein is accessible to antibodies internalized by fluid-phase endocytosis. Exogenously added CD63-specific antibodies partition preferentially to and accumulate in LBPA-positive late endocytic compartments, with subsequent transport to secretory compartments (Kobayashi et al., 2000). Internalized anti-LBPA antibodies traffic to late endocytic compartments, resulting in the impairment of trafficking out of these compartments, as characterized by a massive accumulation of cholesterol (Kobayashi et al., 1999). Here, this approach was used to determine the fate and potential effect of anti-CD63 antibodies on Chlamydia-infected cells. Anti-CD63 antibodies that trafficked to and were incorporated into the chlamydial inclusion demonstrate a direct interaction between a CD63positive compartment and intracellular chlamydiae (Fig. 5). The exclusion of anti-LAMP-1 antibody and dextran from the inclusion indicated that anti-CD63 acquisition was the result of selective vesicular-sorting and -transport from the endocytic pathway or subsequent MVB-related compartments. The disruption in chlamydial development by CD63-specific antibody further implies that this MVB constituent is important for inclusion-biogenesis and -maturation (Fig. 6).
Disruption of MVBs leads to detrimental accumulation of cholesterol and sphingolipids as shown experimentally and in Journal of Cell Science 119 (2) disease states (Kobayashi et al., 1999;Vanier and Millat, 2003). Inhibitors of PI 3-kinases, key regulators of protein sorting, disrupt the biogenesis of MVBs, which results in swollen vesicles that are impaired in membrane trafficking out of these compartments (Fernandez-Borja et al., 1999;Kundra and Kornfeld, 1998). The pharmacological agent U18666A inhibits cholesterol transport from late endocytic compartments (Liscum and Faust, 1992) and alters the trafficking of CD63 (Higgins et al., 2001;Kobayashi et al., 1999;Kobayashi et al., 2000). Because inhibitors can pleiotropically effect cellular function, their potential effect on MVBs was initially confirmed in the HEp-2 cells used in this study. Culturing in the presence of defined concentrations of 3-methyladenine, LY-294002 or U18666A resulted in swollen CD63-positive MVBs. Altered morphology correlated with the disrupted maturation of inclusions when inhibitors where present (Fig. 7), further implicating a role of MVBs in the normal development of chlamydiae.
The PI 3-kinase inhibitor 3-methyladenine is also a modulator of autophagy (Seglen and Gordon, 1982); and in the pathogenesis of Chlamydia, a potential role of the host cell's autophagy pathway has been proposed (Al-Younes et al., 2004). Autophagy is a major mechanism for degradation and recycling of intracellular macromolecules (Klionsky and Emr, 2000). Although studies have indicated a lack of autophagosomal fusion with the chlamydial inclusion, the intracellular bacteria were sensitive to 3-methyladenine and other traditional inhibitors of autophagy (Al-Younes et al., 2004). The autophagy pathway is thought to intersect MVBs prior to lysosomal delivery (Lucocq and Walker, 1997). Thus, the effects of inhibitors of autophagy on chlamydial growth are potentially a result of downstream effects, subsequent to autophagosome-MVB interception.
Fundamental questions, regarding intracellular mediators and signals essential for inclusion formation, maturation and sequestration, remain unknown. In contrast to other intracellular pathogens that reside in membrane-bound vacuoles, an association between chlamydiae and endocytic pathways of its host cell has not been identified. However, common among all Chlamydiae species is the fusigenicity of the inclusion with an apparent exocytic pathway Rockey et al., 1996;Wolf and Hackstadt, 2001). Fluorescent analogs of ceramide demonstrate the direct trafficking of an exogenous substrate to the chlamydial inclusion. Ceramide is synthesized in the endoplasmic reticulum and enzymatically converted to sphingomyelin or glucosylceramide at the cis-or medial-Golgi before being transported to the plasma membrane (Futerman et al., 1990;Lipsky and Pagano, 1985a). Labeling infected cells with NBDceramide led to the diversion of fluorescent label to the chlamydial inclusion and the incorporation of NBDsphingomyelin into the chlamydial cell wall (Hackstadt et al., 1995). The acquisition of endogenously synthesized sphingomyelin was disrupted by brefeldin A, an inhibitor of anterograde transport from the Golgi complex. These studies were expanded to analyze cholesterol, a lipid identified in purified EBs and the chlamydial inclusion membrane (Carabeo et al., 2003;Wylie et al., 1997). Acquisition of de novosynthesized and low-density lipoprotein-derived cholesterol occurs through this brefeldin-A-sensitive, Golgi-dependent pathway (Carabeo et al., 2003). However, brefeldin A inhibition of sphingomyelin and cholesterol delivery failed to interrupt bacterial multiplication (Carabeo et al., 2003;Hackstadt et al., 1996), suggesting that, in addition to direct Golgi transport, chlamydiae possess an alternative pathway for lipid acquisition.
The phospholipid composition of purified Chlamydia trachomatis includes both typical prokaryotic lipids (phophatidylserine, phophatidylethanolamine and phophatidylglycerol) and lipids generally associated with eukaryotes (sphinogmyelin, cholesterol, phophatidylcholine, and phophatidylinositol) (Newhall, 1988;Wylie et al., 1997). In addition to de novo phopholipid synthesis, studies utilizing radiolabeled isoleucine demonstrated that host cell phospholipids are trafficked to and modified by chlamydiae, a process that occurs independently of Golgi transport (Wylie et al., 1997). Eukaryotic sphingolipids are required for intracellular bacterial growth and production of infectious progeny (Van Ooij et al., 2000), however, the transfer mechanism of host-derived lipids has not been identified. In the present study, the effect of MVB disruption on the direct delivery of sphingomyelin to the chlamydial inclusion was analyzed using the NBD-ceramide as previously described by Hackstadt (Hackstadt et al., 1995). Loading cells with the fluorescent probe resulted in the labeling of the Golgi complex and also of vesicles that occupied a larger perinuclear area beyond the Golgi that were subsequently identified as CD63positive compartments. Culturing infected cells in the presence of 3-methyladenine and U18666A, the pharmacological inhibitors that prevent trafficking out of MVBs, resulted in loaded probe accumulating in enlarged CD63-positive MVBs. This accumulation was accompanied by the inhibition of delivery of synthesized sphingomyelin to the chlamydial inclusion (Fig. 8). These observations indicate that host lipids are trafficked to the chlamydial inclusion by a pathway that intercepts MVBs. Disruption of this pathway blocks both delivery and bacterial growth, hence, trafficking from MVBs is essential to productive intracellular development of the chlamydial inclusion.
Intersecting the endosome-lysosome- (Van Deurs et al., 1993), autophagy- (Lucocq and Walker, 1997) and exocyticpathways (Denzer et al., 2000;Stoorvogel et al., 2002), MVBs are ideally positioned to interact with the chlamydial inclusion. In addition, the functional attributes of MVBs make these and related organelles an intriguing source of biosynthetic factors that might be essential for inclusion biogenesis and maturation. This study demonstrates a dynamic interaction between MVBderived constituents and the chlamydial inclusion. It is not clear whether a direct interaction exists, or whether MVBrelated or downstream compartments intersect the inclusion. Identification and characterization of chlamydial inclusioninflux pathways and corresponding requisite host factors, will have important implications on the general biology of Chlamydia and the potential development of therapeutic intervention targeting nutrient-acquisition pathways essential for Chlamydia propagation.
Cell culture and propagation of chlamydiae C. trachomatis serovar E (obtained from Harlan Caldwell, Rocky Mountain Laboratories, NIH, NIAID) was propagated in HEp-2 cells (ATCC, Manassas, VA) and elementary bodies (EBs) were purified by Renografin gradient-centrifugation as previously described (Caldwell et al., 1981). HEp-2 cells were maintained at 37°C with 5.5% CO 2 in Iscove's DME medium supplemented with 12.5 mM HEPES, 10% (vol/vol) FBS, and 10 g/ml gentamicin. HEp-2 cells were infected by incubating monolayers with Chlamydia EBs at a multiplicity of infection of 0.5 for 1 hour at 37°C on a platform rocker. Cells were washed twice with PBS and incubated in Iscove's medium for the times indicated.

Plasmids and transfection
Full length cDNA for CD63 (Genbank accession no. BT007073) was obtained by PCR amplification from human cDNA with KlenTaq LA (Sigma). CD63 was cloned as a C-terminal HA9-epitope fusion in the pcDNA 3.1/V5 His vector (Invitrogen, Carlsbad, CA). HEp-2 cells at subconfluency were transfected with plasmid using FuGENE6 (Roche Molecular Biochemicals, Indianapolis, IN) as described in the manufacturer's protocol. Cells transfected for 24 hours were then infected with Chlamydia as described above.

Immunofluorescence and confocal microscopy
HEp-2 cells were grown on glass coverslips and infected with Chlamydia as described above. For immunofluorescence analysis, cells were fixed for 20 minutes at room temperature in 1% formaldehyde (37% stock contains 10-15% methanol, Fisher Scientific, Hanover Park, IL). Fixation with this concentration of formaldehyde permeabilized the host cell membrane, without permeabilizing intracellular vesicular membranes or the chlamydial inclusion membrane. To subsequently permeabilize these intracellular membranes, cells were treated for 1 minute with 1 mg/ml Zwittergent ® . After blocking in 5% goat serum/PBS, cells were incubated with the indicated primary antibodies for 20 minutes. Cells were washed three times in blocking buffer and subsequently incubated with the indicated fluorophore-conjugated secondary antibody for 20 minutes. Controls omitting the primary antibody were consistently negative at the concentration of fluorophoreconjugated secondary antibodies used in these studies. Incubating with 0.2 M TOPRO-3 (monomeric cyanine nucleic acid stain, Abs 642nm/Em 661nm) for 10 minutes at room temperature labeled both the intracellular bacteria and host cell nuclei. For colocalization studies of host cell markers to the chlamydial inclusion, TOPRO-3-labeling of intracellular bacteria eliminated any potential artifacts that may be encountered with multiple antibody labeling using both chlamydial-and host cell-specific antibodies. Coverslips were mounted on glass slides using ProLong Anti-Fade mounting medium (Molecular Probes). Images were acquired using a Zeiss LSM510 Meta laser scanning confocal microscope (Carl Zeiss Inc., Thornwood, NY) equipped with a 63ϫ, 1.4 numerical aperature Zeiss Plan Apochromat oil objective. Confocal Z-slices of 0.5 m were obtained using the Zeiss LSM510 software. Orthogonal views were generated by acquiring 0.5 m Zslices, with an optimal interval of 0.32 m, from the entire height of the combined Z-sections of infected cells.

Electron microscopy
For analysis of CD63 localization at the ultrastructural level, 48-hour Chlamydiainfected cells were prepared for cryo immunoelectron microscopy. Infected cells were fixed in 4% paraformaldehyde/0.1% glutaraldehyde (Polysciences Inc., Warrington, PA) in 100 mM PIPES pH 7.2/0.5 mM MgCl 2 for 1 hour at 4°C. Samples were then embedded in 10% gelatin and infiltrated overnight with 2.3 M sucrose/20% polyvinyl pyrrolidone in PIPES/MgCl 2 at 4°C. Samples were trimmed, frozen in liquid nitrogen and sectioned with a Leica Ultracut UCT cryoultramicrotome (Leica Microsystems Inc., Bannockburn, IL). 70-nm-sections were blocked with 5% fetal bovine serum/5% goat serum for 30 minutes and subsequently incubated with the indicated primary antibody overnight at 4°C. Sections were then washed in blocking buffer and probed with the appropriate 18 nm colloidal goldconjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 1 hour at room temperature. Sections were washed in PIPES buffer followed by an extensive water rinse and stained with 1% uranyl acetate/1.6% methyl cellulose (Ted Pella Inc., Redding, CA). Samples were viewed with a JEOL 1200EX transmission electron microscope (JEOL USA Inc., Peabody, MA). All labeling experiments were conducted in parallel with controls omitting the primary antibody. These controls were consistently negative at the concentration of colloidal gold-conjugated secondary antibodies used in these studies.
For quantification of anti-CD63 label by electron microscopy, the density of colloidal gold particles per cross-sectional area of the inclusion, MVBs, mitochondria or nucleus was determined. Twenty micrographs of Chlamydiainfected cells were analyzed. The cross-sectional area of the defined structures was determined by using a systematic array of test points as described by Griffiths (Griffiths, 1993).
For analysis of the effect of exogenous antibody and inhibitors on inclusion morphology at the ultrastructural level, infected cells were fixed in 2% paraformaldehyde/2.5% glutaraldehyde (Polysciences Inc.) in 100 mM phosphate buffer pH 7.2 for 1 hour at room temperature. Following three washes in phosphate buffer, cells were postfixed in 1% OsO 4 (Polysciences Inc.) for 1 hour. The samples were then rinsed extensively in dH 2 O before en-bloc staining with 1% aqueous uranyl acetate (Ted Pella Inc.) for 1 hour. Following several rinses in dH 2 O, cells were dehydrated in a graded series of ethanol and embedded in Eponate 12 resin (Ted Pella Inc.). Sections of 70-80 nm were cut, stained with uranyl acetate and lead citrate and viewed by transmission electron microscopy.

Exogenous antibody loading
HEp-2 cells were grown on glass coverslips and infected with Chlamydia as described above. Twenty-four hours post infection, cells were washed with PBS and incubated with Iscove's medium containing 50 g/ml of mouse anti-CD63 antibody .5) and the same concentration of rabbit anti-LAMP-1 antibody (H-228) or 100 g/ml dextran-FITC. At 48 hours post infection, cells were harvested for analysis of antibody uptake by indirect immunofluorescence with anti-mouse Alexa Fluor 594 and anti-rabbit Alexa Fluor 488 antibodies. Effects of exogenous antibody on inclusion development and morphology were determined by infectivity assays and TOPRO-3 labeling, respectively.

Analysis of inhibitors
HEp-2 cells were infected with Chlamydia as described above. One hour post infection, cells were washed with PBS and incubated with Iscove's medium containing 3-methyladenine, LY-294002 hydrochloride or U18666A at the indicated concentrations for the 48-hour time-course of infection. Effects of inhibitors on inclusion development, morphology and sphingolipid trafficking were determined by infectivity assays, immunofluorescence labeling with anti-chlamydial hsp60 antibody and NBD-ceramide labeling, respectively.

Infectivity assays
At 48 hours post infection, infected monolayers cultured in the presence of inhibitors or exogenous antibodies were washed with PBS, scraped from the culture dishes into fresh Iscove's medium, and sonicated to disrupt the HEp-2 cells and release infectious EBs. Dilutions of the disrupted cell suspensions were inoculated onto fresh monolayers of HEp-2 cells as described above. At 48 hours post infection, the monolayers were fixed, immunolabeled with anti-chlamydial hsp60 antibody, and visualized with a Zeiss Axioskop Mot Plus fluorescence microscope to determine the number of infectious forming units.

Ceramide labeling
Chlamydia-infected monolayers were incubated with 5 M fluorescent NBD-C 6ceramide complexed with BSA (1:1 mol:mol) for 30 minutes at 4°C as described previously (Hackstadt et al., 1995;Lipsky and Pagano, 1985a). After washing in HBSS, cells were incubated in Iscove's medium/0.034% dfBSA for 1 hour at 37°C to back-exchange excess label from the plasma membrane and incorporate the probe intracellularly. Cells were fixed, processed for indirect immunolabeling, and analyzed by confocal microscopy as described above.
Sincere thanks to Richard Morrison, Ted Hackstadt, Harlan Caldwell, Toshihide Kobayashi and Jean Gruenberg for generously providing antibodies and reagents, Keliang Tang for technical assistance, L. David Sibley for insightful discussions, and Dan Ory and Thomas Steinberg for comments on this manuscript. This work was supported by an award from the American Heart Association.