INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

SURFACE RECEPTORS TOGETHER with their ligands, as well as other membrane proteins and fluid phase solutes, enter the cells via clathrin-coated or uncoated vesicles that converge in early endosomes. Most of the receptors and a fraction of the solutes are rapidly recycled back to the plasma membrane. In contrast, downregulated receptors, other membrane proteins, and the remaining solutes are delivered from the early to late endosomes and subsequently to lysosomes, where they are degraded. The acidic pH of these compartments, which is maintained by vacuolar-type ATPases (V-ATPases) (12), plays an essential role in the endocytic process. Indeed, dissipation of pH gradients by means of ionophores or weak bases has been shown to prevent fusion between early and late endosomes (6) as well as between late endosomes and lysosomes (35).

Although the recycling and lysosomal delivery of internalized molecules have been extensively investigated, much less is known about the retrograde transport of surface molecules to the secretory pathway. Recent studies have revealed the existence of less prominent routes involved in the delivery of (glyco)proteins and glycolipids from the plasma membrane to the Golgi complex and endoplasmic reticulum (17). The best described of these systems is responsible for the retrieval of proteins such as furin and TGN38 from the plasmalemma to the trans-Golgi network (TGN), where they reside. These proteins are internalized at the plasma membrane via clathrin-coated vesicles and are targeted to the TGN by specific domains of their cytoplasmic tail, which have been defined by mutagenesis studies (13, 14, 36). Like the endocytic pathway, this retrograde route requires luminal acidification for appropriate intracellular traffic. Chapman and Munro (5) showed that delivery of internalized furin and TGN38 to the TGN can be prevented by the weak base chloroquine.

Endogenous retrograde pathways are also utilized by a number of bacterial toxins to reach their target compartment within the cell. Several such toxins share a double-moiety structure: a homopentameric B subunit that binds to cell surface receptors (generally glycolipids) (7) and a toxic A subunit that exerts its activity in the cytosol. After binding to their receptors at the cell surface, these toxins are internalized and retrogradely targeted through the TGN to the Golgi complex and in some cases to the endoplasmic reticulum. The toxic subunit is believed to enter the cytosol by crossing the membrane of these secretory compartments (27). Such is the case for Vibrio cholerae toxin (CT), which specifically binds to the glycolipid monoganglioside 1 (GM₁) at the plasma membrane. After reaching the cytosol, the toxic A₁ fragment of the A subunit of CT catalyzes ADP ribosylation of the -subunit of heterotrimeric G_s proteins, leading to persistent activation of adenylate cyclase (18, 34). Other members of this family are Shiga toxin, the infectious agent in dysentery, which is produced by Shigella dysenteriae (28), and verotoxin (VT). VT, which is synthesized by enterohemorrhagic strains of Escherichia coli, is associated with human vascular diseases such as hemorrhagic colitis and hemolytic uremic syndrome (20). In spite of their different origins, VT and Shiga toxin have a high degree of structural homology and share the same receptor, the glycolipid globotriaosyl ceramide (Gb₃). These two toxins also utilize the same intoxication mechanism: after these toxins reach the endoplasmic reticulum, their A subunits presumably translocate to the cytosol where they inactivate the 60S ribosomal subunits by depurination of a specific adenine residue, causing inhibition of protein synthesis (32).

The mechanism(s) whereby such bacterial toxins gain access to the Golgi complex and endoplasmic reticulum has not been adequately elucidated. CT has been reported to enter the cells through caveolae (22). In contrast, the process underlying VT internalization is not well defined. Initial studies in Daudi cells showed that monodansyl cadaverine prevents endocytosis of the B subunit of verotoxin 1 (VT1B) (15). This alkyl amine competitively inhibits cytosolic transglutaminase activity, an enzyme involved in the cross-linking of ligand-bound cell surface receptors clustered within clathrin-coated pits. The subsequent steps in the retrograde transport of VT, CT, and Shiga toxin to the Golgi complex have not been explored. In particular, it is unclear whether the toxins utilize the same pathway that mediates retrieval of furin and TGN38 to the Golgi complex.

In the present work, we characterized the internalization and retrograde transport of VT1B and of the B subunit of CT (CTB) in Vero cells. To minimize toxicity, we omitted the A subunits and utilized recombinant B subunits, which dictate the receptor affinity and targeting of the toxins. The results obtained suggest that VT1B can enter the cells by both clathrin-dependent and -independent routes. More importantly, the targeting of VT1B and CTB to the Golgi complex was found to be mediated by a common retrograde pathway, which is distinct from the furin and TGN38 retrieval system.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. Texas red-labeled transferrin and rhodamine-labeled phalloidin were purchased from Molecular Probes (Eugene, OR). Indocarbocyanine (Cy3)-labeled donkey anti-mouse antibodies were from Jackson Immuno Research Laboratories (West Grove, PA). Fluorescein isothiocyanate (FITC)-labeled CTB, filipin, and cytochalasin D were purchased from Sigma Chemical (St. Louis, MO). Concanamycin A was from Kamiya Biochemical (Thousand Oaks, CA), and monodansyl cadaverine was from Fluka Chemie (Buchs, Switzerland). ¹²⁵I-labeled goat anti-human immunoglobulin G (IgG) antibody was from ICN (Costa Mesa, CA). The monoclonal antibody 10E6 was the kind gift of Dr. W. Brown (Cornell University, Ithaca, NY).

Cell culture and incubation with toxins. Vero cells obtained from the American Type Culture Collection (Rockville, MD) were cultured at 37°C in minimal essential medium (MEM; GIBCO, Grand Island, NY) containing 5% heat-inactivated fetal bovine serum (Cansera International, Rexdale, ON, Canada), 0.1% glutamine, vitamins, and 1% penicillin-streptomycin (GIBCO) under 5% CO₂.

Protocols for inhibition of endocytosis. Four different protocols were used to inhibit clathrin-mediated internalization. 1) The first protocol was monodansyl cadaverine treatment: cells were incubated with the appropriate toxin, as described above, in the presence of 500 µM monodansyl cadaverine. 2) The second protocol was K⁺ depletion: cells were subjected to a hypotonic shock by preincubation in 50% diluted culture medium for 5 min at 37°C. The cells were then washed with PBS and incubated in K⁺-free medium [140 mM NaCl, 10 mM N-methyl-D-glucamine chloride, 1 mM MgCl₂, 1 mM CaCl₂, 20 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), 5.5 mM glucose, and 1% bovine serum albumin (BSA; Boehringer Mannheim), pH 7.4] for 10 min at 37°C. Labeling with the appropriate toxin was performed in the K⁺-free medium. 3) The third protocol was hypertonic treatment: cells were preincubated in hypertonic medium (in mM: 350 NaCl, 10 KCl, 1 MgCl₂, 1 CaCl₂, 20 HEPES, and 5.5 glucose and 1% BSA, pH 7.4) for 20 min at 37°C. Labeling with the appropriate toxin was also performed in the hypertonic medium. 4) The fourth protocol was cytoplasmic acidification: cells were preincubated in growth medium supplemented with 5 mM acetic acid (pH 5.0) for 5 min at 37°C. This condition was maintained throughout the labeling of the cells with the appropriate toxin.

Immunocytochemistry and fluorescence microscopy. Cells were fixed in 4% paraformaldehyde in PBS at room temperature for 30 min, washed with 100 mM glycine, and permeabilized with 0.1% Triton X-100 in PBS. To label the cis-Golgi cisternae, fixed and permeabilized cells were then blocked with 5% donkey serum in PBS containing 0.1% BSA and 0.1% Triton X-100 for 20 min at room temperature, washed three times with PBS, and incubated with the monoclonal antibody 10E6 (1:200 dilution in PBS containing 0.1% BSA and 0.1% Triton X-100) for 2 h at room temperature. Samples were subsequently washed three times in PBS containing 0.1% BSA and incubated with Cy3-labeled donkey anti-mouse antibody (1:500 dilution in PBS containing 0.1% BSA and 0.1% Triton X-100) for 1 h at room temperature. After samples were washed three times with PBS-0.1% BSA, samples were treated with Slow Fade (Molecular Probes) and mounted. Control experiments were performed, omitting the primary antibody. F-actin was stained by incubating the fixed and permeabilized cells in the presence of rhodamine-phalloidin (10 U/ml) for 30 min at room temperature.

Quantitative analysis of toxin binding. To analyze the interaction between toxins, cells were incubated with 10 µg/ml of unlabeled VT1B or 10 µg/ml of unlabeled CTB for 1 h at 4°C, washed, and warmed to 37°C for 15 min to allow internalization. The samples were then incubated with either FITC-CTB or FITC-VT1B, respectively, for 1 h at 4°C and washed. After the unbound toxin was washed, the cells were lysed by incubation in 50 mM tris(hydroxymethyl)aminomethane and 1% Nonidet P-40 (pH 8) for 5 min. After a thorough mixing, debris were sedimented in an Eppendorf 5415 microcentrifuge and the fluorescence intensity of the supernatants was quantified in a Perkin Elmer 650-40 fluorescence spectrophotometer. Background fluorescence was determined using nonlabeled cells and subtracted from all other determinations. The results were expressed as the ratio of the fluorescence intensity of toxin-pretreated and control (nonpretreated) cells.

Fluorescence ratio and Nomarski imaging. Simultaneous imaging of fluorescence and of cell morphology was performed using an inverted microscope (Axiovert 135; Zeiss, Oberkochen, Germany) equipped with epifluorescence optics. Excitation at 440 and 490 nm was provided by a xenon arc lamp via a computer-controlled shutter and filter wheel assembly (Sutter Instruments, Novato, CA), whereas continuous 620-nm illumination was achieved by filtering the transmitted incandescent source. The excitation light was attenuated by a neutral density filter and reflected to the cells by a dichroic mirror (510 nm), while the emitted fluorescence (>510 nm) and the transmitted red light (>620 nm) were separated by an emission dichroic mirror (580 nm). The red light was directed to a video camera, allowing continuous visualization of the cells, while the fluorescent light was directed onto a 542 band-pass 62-nm filter and imaged with a cooled CCD camera (Princeton Instruments). Control of image acquisition was achieved with Metafluor software (Universal Imaging, West Chester, PA), operating on a pentium Dell computer (Dell, Toronto, ON, Canada).

VT1B uptake determinations. Cells were incubated with 10 µg/ml VT1B for 1 h at 4°C, followed by a chase at 37°C for the specified periods of time. Samples were fixed as above and then blocked with 5% donkey serum in PBS containing 0.1% BSA for 20 min at room temperature. After cells were washed three times with PBS, cells were incubated with monoclonal anti-VT1B antibody (3) (1:500 dilution in PBS containing 0.1% BSA) for 1 h at room temperature. Samples were subsequently washed three times in PBS containing 0.1% BSA and incubated with ¹²⁵I-labeled goat anti-mouse antibody (0.1 µCi/ml in PBS with 0.1% BSA) for 1 h at room temperature. After three more washes, the cells were scraped and transferred to counting vials. Radioactivity was quantified using a 1282 Compu-Gamma counter (LKB Wallac, Turku, Finland). Background radioactivity, determined with the omission of the primary antibody, was subtracted from all other determinations.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Time course of VT1B and CTB internalization. The available evidence suggests that VT1B and CTB enter the cells via distinct mechanisms (see introduction). Their rate of internalization and subcellular distribution would therefore be expected to differ. This notion was tested by coincubating Vero cells with the two toxins and monitoring their fate by confocal fluorescence microscopy. The cell surface was labeled with a mixture of RITC-VT1B and FITC-CTB at 4°C (Fig. 1, A-C), the cells were then warmed to 37°C, and samples were taken at increasing times (Fig. 1, D-O). The toxins, which are initially seen lining the surface membrane (Fig. 1, A-C), subsequently assume a punctate distribution and eventually cluster in a tight juxtanuclear complex, likely the Golgi complex. Overlay of the red and green fluorescence (Fig. 1) revealed that the distribution of the two toxins is very similar throughout, suggesting that VT1B and CTB utilize a common pathway.

View larger version (59K): [in this window] [in a new window]   Fig. 1.   Time course of B subunit of verotoxin 1 (VT1B) and B subunit of CT (CTB) internalization. Vero cells were incubated with rhodamine isothiocyanate (RITC) VT1B and fluorescein isothiocyanate (FITC) CTB immediately (10 µg/ml each) for 1 h at 4°C and either fixed immediately (A-C) or incubated at 37°C for the indicated periods of time (D-O). Left column: distribution of VT1B. Middle column: distribution of CTB. Right column: areas of VT1B and CTB colocalization (overlay). Results are representative of 3 similar experiments.

View larger version (54K): [in this window] [in a new window]   Fig. 2.   Internalization of VT1B. A: cells were exposed to VT1B for 1 h at 4°C and immediately fixed and labeled sequentially with anti-VT1B monoclonal (primary) antibody (Ab) followed by indocarbocyanine-labeled goat anti-mouse (secondary) Ab. B: cells were exposed to VT1B at 4°C and then allowed to internalize the toxin for 1 h at 37°C. After fixation, cells were incubated with primary and secondary Abs as in A. C: cells were exposed to VT1B at 4°C and then allowed to internalize the toxin for 1 h at 37°C. Intact cells were next exposed to the primary Ab at 37°C for an additional 1 h. Finally, cells were fixed and labeled with secondary Abs as in A. A-C are representative of 3 similar experiments. D: cells grown on 6-well plates were preincubated with VT1B for 1 h at 4°C, followed by an incubation for the specified periods of time at 37°C (abscissa). Samples were then fixed and incubated sequentially with primary Ab, followed by 125I-labeled goat-anti-mouse Ab, as described in MATERIALS AND METHODS. Radioactivity associated with confluent cells on a well is plotted vs. time. Data are means of 2 determinations. cpm, Counts per minute.

View larger version (88K): [in this window] [in a new window]   Fig. 3.   Endosomal transit of VT1B and CTB. Vero cells were incubated with 5 µg/ml of Texas red-labeled transferrin and 10 µg/ml of either FITC-VT1B or FITC-CTB for 1 h at 4°C. Cells were then washed, and internalization was allowed to occur in MEM for 15 min at 37°C. Cells were fixed, mounted, and visualized by epifluorescence microscopy. A and C: localization of transferrin. B: localization of VT1B in the same cells shown in A. D: localization of CTB in the same cells shown in C. Results are representative of 3 similar experiments.

Clathrin dependence of VT1B internalization. The similarities in the behavior of the two toxins contrast with the notion that they enter via different mechanisms. To better define the pathway utilized by VT, we studied the effects of three protocols routinely used to inhibit clathrin-mediated endocytosis, namely, K⁺ depletion, hypertonic treatment, and cytoplasmic acidification, on the uptake of FITC-VT1B by Vero cells. Whereas K⁺ depletion and hypertonic treatment inhibit internalization by dispersing membrane-associated clathrin lattices, cytoplasmic acidification prevents the budding of clathrin-coated vesicles from the plasma membrane (11). To test the efficiency of these treatments in Vero cells, we assessed the uptake of transferrin, which is typically internalized through clathrin-coated pits and vesicles. Cells were preincubated in the presence of Texas red-labeled transferrin for 1 h at 4°C, washed, and incubated at 37°C for 20 min. The distribution of the fluorophore was then determined by confocal microscopy, and representative x vs. z reconstructions are shown in Fig. 4. Whereas transferrin was normally internalized in untreated cells (Fig. 4A), K⁺-depleted cells displayed only labeling on the surface (Fig. 4B). Virtually identical results were obtained in cells treated hypertonically or with acidic solutions (Fig. 4, C and D).

View larger version (42K): [in this window] [in a new window]   Fig. 4.   Effect of K+ depletion, hypertonic treatment, and cytoplasmic acidification on transferrin endocytosis. Vero cells were either untreated (A), depleted of intracellular K+ as described in MATERIALS AND METHODS (B), preincubated in hypertonic medium for 20 min at 37°C (C), or exposed to acidifying medium for 5 min at 37°C (D). Next, cells were incubated with Texas red-labeled transferrin (5 µg/ml) for 1 h at 4°C in the appropriate solution. Cells were then washed, and internalization was allowed to occur in the appropriate media for 15 min at 37°C. Cells were fixed, mounted, and visualized by confocal microscopy. Representative x vs. z (cross-sectional) reconstructions are shown. Results are representative of 3 similar experiments. Bar = 5 µm.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Role of clathrin-mediated endocytosis in VT1B internalization. Vero cells were either untreated (A), depleted of intracellular K+ as described in MATERIALS AND METHODS (B), preincubated in hypertonic medium for 20 min at 37°C (C), or exposed to acidifying medium for 5 min at 37°C (D). Next, cells were incubated with FITC-VT1B (10 µg/ml) for 1 h at 4°C in the appropriate solution. Cells were then washed, and internalization was allowed to occur in the appropriate media for 1 more h at 37°C. Cells were fixed, mounted, and visualized under confocal microscopy. Representative x vs. y scans are shown at top and x vs. z (cross-sectional) reconstructions are shown for each condition at bottom. In this and Figs. 6, 7, and 10, the cross-sectional views were obtained from cells other than those shown in the x vs. y scans. Results are representative of 3 similar experiments.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Effects of monodansyl cadaverine on the internalization of transferrin, VT1B, and CTB. Vero cells were preincubated in medium containing 500 µM monodansyl cadaverine for 30 min at 4°C. Next, cells were incubated with either 5 µg/ml Texas red-labeled transferrin (A), 10 µg/ml FITC-VT1B (B), or 10 µg/ml FITC-CTB (C) for 1 h at 4°C in solution containing monodansyl cadaverine. Cells were then washed and incubated in monodansyl cadaverine-containing medium for 1 more h at 37°C. Cells were finally fixed, mounted, and visualized by confocal microscopy. A: representative x vs. z (cross-sectional) reconstructions. B and C: representative x vs. y scans (B and C) are shown at top and x vs. z are shown at bottom. Results are representative of 3 similar experiments.

View larger version (37K): [in this window] [in a new window]   Fig. 7.   Role of clathrin-mediated endocytosis in CTB internalization. Vero cells were either untreated (A), depleted of intracellular K+ as described in MATERIALS AND METHODS (B), preincubated in hypertonic medium for 20 min at 37°C (C), or exposed to acidifying medium for 5 min at 37°C (D). Next, the cells were incubated with FITC-CTB (10 µg/ml) for 1 h at 4°C in the appropriate solution. Cells were then washed, and internalization was allowed to occur in the appropriate media for 1 more h at 37°C. Cells were fixed, mounted, and visualized by confocal microscopy. Representative x vs. y scans are shown at top and x vs. z (cross-sectional) reconstructions are shown for each condition at bottom. Results are representative of 3 similar experiments.

Effect of cytochalasin B on VT1B and CTB internalization and retrograde transport. Clathrin-independent internalization of plasmalemmal components can occur by macropinocytosis, which is observed at sites of active ruffling. Unlike clathrin-dependent vesicle formation, macropinocytosis requires assembly of filamentous (F) actin and is therefore sensitive to the cytochalasins. To test the possible involvement of macropinosomes in VT1B uptake, Vero cells were pretreated with 5 µM cytochalasin B, which binds to the barbed (plus) end of F-actin, inhibiting its polymerization. The effectiveness of this drug was ascertained by staining F-actin with rhodamine-phalloidin, as illustrated in Figs. 8, A and C, and 9, A and C. Cytochalasin virtually eliminated the stress fibers and decreased the overall content of F-actin, which was largely clustered in irregular aggregates (cf. A and C in Figs. 8 and 9). Figure 8 also shows that VT1B was effectively internalized in cytochalasin-treated cells. It is noteworthy that the morphology of the intracellular compartment where VT1B was accumulated was altered by F-actin disruption. However, this does not reflect a change in the targeting of the toxin but rather an alteration in Golgi morphology. This was determined by comparing the distribution of the toxin with that of a cis-Golgi marker, identified by antibody 10E6 (Fig. 8, E and F). Similar results were obtained using CTB (Fig. 9). These findings indicate that neither macropinocytosis nor other F-actin-dependent processes are important in VT1B internalization. Moreover, they imply that actin polymerization is not required for the retrograde targeting of either VT1B or CTB to the Golgi.

View larger version (41K): [in this window] [in a new window]   Fig. 8.   Effect of cytochalasin B on the internalization and targeting of VT1B. Vero cells were allowed to internalize FITC-VT1B in the absence (A and B) or presence (C-F) of cytochalasin D (Cyto D; 5 µM). A and C: actin filaments were stained with rhodamine-phalloidin. B, D, and F: VT1B fluorescence. E: indirect immunofluorescence using Ab 10E6, directed to an epitope in the cis-Golgi. Results are representative of 3 similar experiments.

View larger version (43K): [in this window] [in a new window]   Fig. 9.   Effect of cytochalasin B on the internalization and targeting of CTB. Vero cells were allowed to internalize FITC-CTB in the absence (A and B) or presence (C-F) of cytochalasin D (5 µM). A and C: actin filaments were stained with rhodamine-phalloidin. B, D, and F: CTB fluorescence. E: indirect immunofluorescence using Ab 10E6, directed to an epitope in the cis-Golgi. Results are representative of 3 similar experiments.

Involvement of caveolae in VT1B internalization. Clathrin-independent internalization can also occur through caveolae, which are 50- to 60-nm plasma membrane invaginations with a characteristic flask shape. These structures are enriched in glycosphingolipids and cholesterol and contain caveolin, a 21-kDa cholesterol-binding protein that cycles between the plasma membrane and the Golgi complex (23). Considering the similar subcellular distribution of VT1B and CTB and their common resistance to antagonists of clathrin-mediated uptake, it appeared likely that VT1B also utilizes caveolae for internalization. This notion was tested with filipin, a cholesterol-binding drug that causes disassembly of caveolae. In cells treated with filipin (5 µg/ml), the internalization of VT1B was partially inhibited (Fig. 10A). The toxin displayed a punctate distribution, with moderate perinuclear accumulation.

View larger version (39K): [in this window] [in a new window]   Fig. 10.   Role of caveolae in VT1B internalization. A: Vero cells were preincubated with FITC-VT1B for 1 h at 4°C in the presence of 5 µg/ml of filipin and then were incubated for another hour at 37°C in filipin-containing solution. B: cells were preincubated with hypertonic medium as in Fig. 1C and then labeled at 4°C with FITC-VT1B in the same medium with added filipin. These conditions were maintained throughout the internalization chase (1 h at 37°C). Representative x vs. y scans and x vs. z reconstructions are shown for each condition in top and bottom, respectively. Results are representative of 3 similar experiments.

pH dependence of VT1B and CTB retrograde transport. The retrograde transport of furin and TGN38 from the plasma membrane to the Golgi complex requires intraorganellar acidification (see introduction). To determine whether toxins such as VTB utilize a similar pH-dependent pathway, we monitored their distribution under conditions expected to dissipate the electrochemical H⁺ gradient of endomembrane compartments. The luminal pH of endosomes and of components of the Golgi apparatus is maintained as acidic by vacuolar H⁺ pumps, which are selectively inhibited by macrolide antibiotics such as the bafilomycins and concanamycins. Cells were therefore treated with concanamycin A (100 nM) to inhibit active pumping. To ensure elimination of preexisting H⁺ gradients, the cells were additionally exposed to the weak base chloroquine (50 µM). The dissipation of the organellar acidification was initially verified by measuring the pH of endosomes and Golgi complex by imaging the emission of FITC-labeled CTB, a pH-sensitive fluorescent probe. Cells were preincubated with FITC-CTB for 1 h at 4°C, washed, and then incubated for an additional 20 min at 37°C to allow internalization. At this time, the probe is localized largely in endosomes, although some of the juxtanuclear toxin may have entered the TGN (Fig. 11A). Endosomal pH was then monitored by quantifying the fluorescence with excitation at 440 and 490 nm (see MATERIALS AND METHODS). As illustrated in Fig. 11B, the pH in otherwise untreated cells averaged 6.5 ± 0.03 (mean ± SE of 30 determinations from 3 experiments), in the range reported for early and recycling endosomes (8). The combined treatment with concanamycin and chloroquine elevated pH to 7.1 ± 0.09 (mean ± SE of 30 determinations). Alkalinization was similar in the juxtanuclear and submembranous endosomes.

View larger version (35K): [in this window] [in a new window]   Fig. 11.   Effect of concanamycin and chloroquine on the pH of endomembrane compartments. A: Vero cells were preincubated with FITC-CTB for 1 h at 4°C, washed, and then incubated for 20 more min at 37°C to allow internalization of the toxin. A representative fluorescence micrograph, obtained using the imaging setup described in MATERIALS AND METHODS, is illustrated. B: pH of the endomembrane compartments labeled with FITC-CTB as in A was measured by fluorescence ratio imaging in otherwise untreated cells (control) and in cells incubated with 100 nM concanamycin (CCM) and 50 µM chloroquine (CLQ) during the last 30 min of the incubation at 4°C and throughout the incubation at 37°C. Data are means ± SE of 30 determinations.

View larger version (44K): [in this window] [in a new window]   Fig. 12.   pH dependence of VT1B internalization. Vero cells were labeled with FITC-VT1B in the cold and then incubated in the absence (A and B) or presence (C-F) of concanamycin (100 nM) and chloroquine (50 µM) as indicated. At the end of the experiment, cells were fixed, permeabilized, and labeled by indirect immunofluorescence using Ab 10E6. Representative x vs. y confocal scans are illustrated. Results are representative of 3 similar experiments.

View larger version (18K): [in this window] [in a new window]   Fig. 13.   pH dependence of CTB internalization. Vero cells were labeled with FITC-CTB in the cold and then incubated in the absence (A and B) or presence (C-F) of concanamycin (100 nM) and chloroquine (50 µM) as indicated. At the end of the experiment, cells were fixed, permeabilized, and labeled by indirect immunofluorescence using Ab 10E6. Representative x vs. y confocal scans are illustrated. Results are representative of 3 similar experiments.

View larger version (23K): [in this window] [in a new window]   Fig. 14.   pH dependence of furin and CTB internalization. A and B: Chinese hamster ovary (CHO) cells expressing immunoglobulin G (IgG)-furin chimeras were incubated in the absence (A) or presence (B) of concanamycin (100 nM) and chloroquine (50 µM) for 1 h 30 min at 37°C. Fixed and permeabilized samples were then labeled by indirect immunofluorescence using anti-human IgG antibodies. C and D: CHO cells were preincubated with FITC-CTB for 1 h at 4°C in the absence (C) or presence (D) of concanamycin (100 nM) and chloroquine (50 µM), followed by incubation for an additional 1 h at 37°C under the same conditions. Results are representative of 2 similar experiments.

Are VT1B and CTB receptors associated? Association among lipids with similar functional and/or structural characteristics in microdomains often called "rafts" has been shown or postulated to exist in different systems. Such lipid-rich microdomains are seemingly involved in sorting and targeting processes (4). Because the receptors for VT1B and CTB (namely, Gb₃ and GM₁) are both glycolipids, we considered the possibility that they are physically associated in macromolecular complexes, possibly rafts. We reasoned that, if stable association exists, internalization of one of the glycolipids would drive the uptake of the other and vice versa. This was tested experimentally by inducing the uptake of one of the lipid receptors by addition of its unlabeled cognate toxin and measuring the degree of surface exposure of the other with the respective tagged toxin. To confirm that a sizable fraction of the first lipid was internalized, we initially performed experiments measuring the effect of each toxin on the ability of a second pulse of the same toxin to bind to the surface. Pretreatment with unlabeled CTB was followed by incubation for 15 min at 37°C, to allow internalization, and then by exposure to FITC-CTB, which was finally quantified fluorometrically (see MATERIALS AND METHODS). As shown in Table 1, upwards of 60% of the GM₁ became inaccessible after incubation with 10 µg/ml CTB. Similarly, nearly 55% of the Gb₃ was inaccessible after treatment with 10 µg/ml of VT1B. In contrast, pretreatment with VT1B had no detectable effect on the amount of GM₁ available at the plasma membrane and CTB only modestly decreased the surface-exposed Gb₃, as measured by VT1B binding. These observations indicate that GM₁ and Gb₃ are internalized independently and not as part of tightly coupled macromolecular complexes.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Endogenous retrograde pathways directing molecules from the plasma membrane to compartments of the secretory pathway can be subverted by bacterial toxins, which travel through endosomes to the Golgi complex and sometimes as far back as the endoplasmic reticulum and nuclear envelope. Several lines of evidence indicate that such retrograde translocation is essential for the toxins to exert their biological effects. In particular, it has been shown that cells are protected against the toxic effects of these bacterial polypeptides by brefeldin A (30) and by overexpression of inactive mutants of Rab1, Sar1, and Arf1 (33), which preclude retrograde transport between the Golgi and the endoplasmic reticulum.

Although different aspects of the uptake of toxins have been investigated (29, 31), the nature of the retrograde route mediating their transport from the plasma membrane has not been elucidated. In this report, we compared the internalization and targeting of VT1B and CTB and showed that they share steps of a common retrograde pathway that differs, at least in part, from the one employed by furin and TGN38.

VT1B internalization. VT1B appears to be internalized in Vero cells by two distinct pathways: one that is clathrin dependent and one (or more) clathrin-independent route(s). Endocytosis was partially inhibited by filipin, suggesting the involvement of caveolae, whereas the combination of this drug with hypertonicity resulted in complete inhibition, implying a contribution from clathrin-coated vesicles. We were unable to determine whether both routes are constitutively active or whether the inhibition of one of them caused upregulation of the other. In addition, the specificity of the pharmacological agents used may not be absolute, so that alternative pathways cannot be ruled out. It is clear, however, that macropinocytosis and other F-actin-dependent processes do not contribute importantly to toxin uptake in Vero cells.

Characterization of VT1B and CTB targeting to the Golgi complex. Glycosphingolipids have been shown to form microdomains at the plasma membrane and TGN. In polarized cells, TGN microdomains are involved in sorting and specific targeting of secreted proteins through the formation of rafts (4). However, the involvement of such rafts in retrograde transport from the plasma membrane has not been determined. Indeed, comparatively little is known about the lipid retrograde transport pathways. Martin and Pagano (21) showed that, in fibroblasts, glucosylceramide fluorescent analogs are internalized from the plasma membrane by an ATP-independent and temperature-insensitive saturable mechanism to the Golgi complex, where they can be further processed. In contrast, internalization of fluorescent analogs of sphingomyelin is temperature and energy dependent. These results suggest that retrograde transport of glycolipids from the plasma membrane can be mediated by both endocytic and nonendocytic pathways and show that these molecules can be targeted to the Golgi complex. In the case of Gb₃ and GM₁, uptake is clearly temperature sensitive and mediated by endosomes.