HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 33, 30325-30336, August 16, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/33/30325    most recent M202626200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wüstner, D. Articles by Maxfield, F. R. Search for Related Content PubMed PubMed Citation Articles by Wüstner, D. Articles by Maxfield, F. R. Social Bookmarking                      What's this?

Rapid Nonvesicular Transport of Sterol between the Plasma Membrane Domains of Polarized Hepatic Cells^*

Daniel Wüstner^§, Andreas Herrmann, Mingming Hao, and Frederick R. Maxfield

From the  Department of Biochemistry, Weill Medical College of Cornell University, New York, New York 10021 and Humboldt-Universität zu Berlin, Mathematisch-Naturwissenschaftliche Fakultät I, Institut für Biologie/Biophysik, Invalidenstrasse 43, D-10115 Berlin, Germany

Received for publication, March 18, 2002, and in revised form, May 10, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We studied the transport of the fluorescent cholesterol analog dehydroergosterol (DHE) in polarized HepG2 human hepatoma cells. DHE delivered via methyl--cyclodextrin was delivered to both the apical and basolateral membranes and became concentrated in the apical membrane within 1 min. Intracellular DHE was targeted mainly to vesicles of the subapical compartment or apical recycling compartment (SAC/ARC), where it colocalized with fluorescent transferrin and fluorescent analogs of phosphatidylcholine and sphingomyelin. In contrast, transport of DHE from the plasma membrane to the trans-Golgi network was found to be very low. Vesicles containing DHE traversed the cells in both directions, but vesicular export of DHE from the SAC/ARC to the plasma membrane domains was low. Disruption of the microtubule cytoskeleton disturbed vesicular transport of DHE but not its enrichment in the apical (canalicular) membrane. Transport of DHE to the canalicular membrane after photobleaching was very rapid (t_1/2 = 1.6 min) and was largely ATP-independent in contrast to enrichment of DHE in the SAC/ARC. Release of DHE from the canalicular membrane was also ATP-independent but slower than the enrichment of sterol in the biliary canaliculus (t_1/2 = 5.4 min). Canalicular DHE could completely redistribute to the basolateral plasma membrane but could not transfer from one cell to the other cell of an HepG2 couplet. We conclude that sterol shuttles rapidly among the plasma membrane domains and other membrane organelles and that this nonvesicular pathway includes fast transbilayer migration.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

One of the hallmarks of polarized epithelial cells is the formation of apical and basolateral plasma membrane domains with distinct lipid and protein compositions. These plasma membrane domains are separated by tight junctions that prevent diffusion of lipids and proteins in the exoplasmic but not in the cytoplasmic leaflet of the membrane bilayer (1, 2). Sorting mechanisms in the biosynthetic, transcytotic, and endocytic recycling pathways ensure the transport of lipids and proteins to the correct plasma membrane domain (3-5). In several types of polarized cells cholesterol has been found to be enriched in the apical membrane along with (glyco)sphingolipids (6). Moreover, establishment of the polarized state in epithelial cells depends on cholesterol (7-9). Cholesterol depletion of polarized cells has been shown to specifically perturb the apical transport of membrane proteins, including various glycosylphosphatidylinositol-anchored proteins (10). This has been attributed to disruption of cholesterol and (glyco)sphingolipid-rich membrane domains in the trans-Golgi network (TGN)¹ (for a review see Refs. 11 and 12).

Despite recognition of the importance of cholesterol in the properties of polarized cells, the transport pathways of cholesterol itself remain obscure. This is largely due to the lack of appropriate cholesterol analogs that would allow one to follow intracellular transport of sterols but which also have physical properties closely resembling those of cholesterol. Dehydroergosterol (DHE) is a natural yeast sterol that does not have a bulky reporter group. Although differences in the desorption rate of DHE compared with cholesterol from lipid monolayers have been reported, DHE shows very similar physicochemical behavior to cholesterol in respect to its lateral and transverse organization in model membranes (13-16). DHE can be used by Caenorhabditis elegans as its only sterol source (17). In pulse-chase studies we demonstrated recently that DHE has an intracellular distribution very similar to cholesterol in CHO cells, with the endocytic recycling compartment as a major sterol pool (18). Thus, DHE can be used to determine the transport pathways of sterol in polarized cells and to study the involvement of cholesterol transport in apically directed membrane trafficking.

Hepatocytes are specifically adapted to secrete cholesterol on their apical (canalicular) membrane into the bile. Biliary secretion of cholesterol and bile salts at the canalicular membrane of hepatocytes along with phosphatidylcholine (PC) is the only pathway of sterol clearance from the body (19). Thus, these cells have to transport cholesterol continuously to their apical membrane, and studying cholesterol dynamics in polarized hepatic cells should set the stage for understanding sterol transport in epithelial cells in general. Several lines of evidence suggest that cholesterol destined for biliary secretion is not synthesized de novo in hepatocytes but is transported to the liver in a process named reverse cholesterol transport (20). During reverse cholesterol transport high density lipoproteins (HDL) carrying cholesterol and cholesteryl esters bind to the basolateral membrane of hepatocytes and deliver their content of free cholesterol to the bile probably independent of uptake and transport of the lipoprotein particle (20, 21). The intracellular trafficking pathways of cholesterol destined for biliary secretion in hepatocytes are unknown. Here we used a well differentiated hepatoma cell line, HepG2, to study the intracellular transport of DHE by fluorescence microscopy. HepG2 cells form an apical vacuole between adjacent cells, which resembles the biliary canaliculus (BC) of hepatocytes in many respects (22). They have been used previously to study the polarized transport of fluorescent sphingolipids as well as the canalicular enrichment of fluorescent PC (23-25). Because the couplets of polarized HepG2 cells are oriented with the apical-basolateral axis lying horizontally, they are well suited for microscopic observation of trafficking between apical and basolateral membranes.

We incorporated DHE initially into the plasma membrane of HepG2 cells and studied its intracellular itineraries and apically directed transport. We found colocalization of DHE with markers of the endocytic pathway and to a small extent with the TGN. Some vesicles containing DHE shuttled bidirectionally between both plasma membrane domains. However, no pronounced vesicle traffic of DHE to either plasma membrane domain could be observed. Enrichment of DHE in intracellular organelles was ATP-dependent, but transport of the sterol between the plasma membrane domains was largely ATP-independent. By using quantitative fluorescence microscopy and image analysis, including fluorescence recovery after photobleaching (FRAP), we show that the transport of DHE to the BC from other membrane components is faster than the observed vesicle traffic and is not slowed by ATP depletion of cells. DHE also redistributed from the canalicular to the basolateral membrane after ATP depletion. DHE could not traverse the BC from one cell to the other in an HepG2 couplet nor shuttle between neighboring cells. These results indicate that vesicular transport of sterols occurs, but nonvesicular pathways appear to predominate in determining sterol distribution in polarized HepG2 cells.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Chemicals

Red fluorescent PC 2-(4,4-difluoro-5-(4-phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s-indacene-3-pentanoyl)-1-hexa-decanoyl-sn-glycero-3-phosphocholine (-BODIPY-PC), N-(4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-pentanoyl)-sphingosine (BODIPY-Cer), N-[6-[(7-nitro-2-1,3-benzooxadiazol-4-yl)amino]hexanoyl]-sphingosine (C₆-NBD-Cer), 2-[6-[(7-nitro-2-1,3-benzooxadiazol-4-y)amino]hexanoyl-1-hexadecanoyl]-sn-glycero-3-phosphatidylcholine, N-[6-[(7-nitro-2-1,3-benzo-oxadiazol-4-y)amino]-hexanoyl]-sphingosyl]-phosphocholine (C₆-NBD-SM), rhodamine-labeled dextran (70 kDa), Alexa-488, Alexa-546, and Alexa-488 phalloidin were purchased from Molecular Probes. Medium 1 contained 150 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 5 mM glucose, and 20 mM HEPES (pH 7.4). Medium 2 was identical to medium 1 except that it contained no glucose but 5 mM sodium azide and 50 mM 2-deoxyglucose for energy depletion of cells (see below). Release medium is Medium 1 supplemented with 25.5 mM citric acid, 24.5 mM sodium citrate, 100 mM deferoxamine mesylate, was adjusted to pH 5.2, and contained 280 mM sucrose instead of glucose (see above). Rhodamine-labeled dextran was dissolved in phosphate-buffered saline and repeatedly dialyzed before use to remove unconjugated dye. Fetal calf serum and Dulbecco's modified Eagle's medium were from Invitrogen. All other chemicals were from Sigma. Transferrin (Tf) was iron-loaded as described previously (26). Succinimidyl ester of Alexa-488 and Alexa-546 were then conjugated to the iron-loaded Tf following the manufacturer's instructions. Texas Red-asialoorosomucoid was a generous gift from Dr. John Murray (Albert Einstein College of Medicine, New York).

Cell Culture

HepG2 cells were grown in Dulbecco's modified Eagle's medium with 4.5 g/liter glucose, supplemented with 10% heat-inactivated fetal calf serum and antibiotics. Cells were routinely passaged in plastic tissue culture dishes. For experiments, cells were plated onto glass coverslips coated with poly-D-lysine and used after reaching the highest degree of polarization as described previously (25).

Preparation of Lipid Labeling Solutions

Fluorescent Lipid Analogs-- Fluorescent phospholipids and ceramide were stored at 20 °C in chloroform/methanol (1:1, v/v). Analogs were transferred to a glass tube and dried under argon. For labeling of the cells with -BODIPY-PC and C₆-NBD-SM, a previously described protocol was used (25, 27).

DHE-Cyclodextrin Complexes-- A stock solution of DHE (5 mM) was made in ethanol and stored under argon. For labeling cells with DHE, the analog was loaded on methyl--cyclodextrin (MCD) as described previously (18). Briefly, from the stock solution of DHE 750 µl were transferred to a glass tube, and ethanol was evaporated under argon, and a solution of 30 mM MCD dissolved in Medium 1 with 0.1% (w/v) BSA was added to get a MCD/DHE-ratio of 1:8 (mol/mol) (28). The solution was carefully resuspended, vortexed, and bath-sonicated. After centrifugation at 20,000 × g for 20 min, the supernatant DHE/MCD solution was carefully collected and stored at 4 °C under argon. The properties of the labeling solution and reproducibility of the loading procedure was routinely checked by measuring its fluorescence excitation and emission spectra using a Spex Fluorolog spectrofluorometer (Spex Industries, Inc., Edison, NJ).

Intracellular Transport of DHE Derived from DHE/MCD

HepG2 cells were routinely labeled with DHE/MCD for 1 min at 37 °C except for studying time course of labeling (see below). Cells were washed and further incubated at 37 °C for the indicated time points (see Fig. 1).

Colocalization Experiments-- For triple labeling cells with DHE/MCD, -BODIPY-PC and Alexa488-Tf cells were incubated for 10 min at 37 °C with 5 µg/ml Alexa488-Tf, washed with Medium 1, labeled for 1 min with DHE/MCD, washed, and incubated for 2 min at 37 °C with -BODIPY-PC. Cells were washed and subsequently incubated at 37 °C for 30 min. Cells were briefly back-exchanged with BSA (2 times for 5 min, 5% BSA w/v on ice), washed, and fixed with 2% paraformaldehyde (PFA) for 10 min on ice. For triple labeling cells with DHE/MCD, -BODIPY-PC, and C₆-NBD-Cer, the same protocol was used but without preincubation with Alexa488-Tf. After fixation with PFA, cells were washed and incubated with 10 µM C₆-NBD-Cer for 5 min at 37 °C. For colabeling with BODIPY-Cer, cells were labeled first with DHE/MCD as described above, washed, and chased for 30 min at 37 °C with 10 µM BODIPY-Cer present during the last 10 min of incubation. Colocalization studies of DHE with Texas Red-ASOR were performed as described previously for studying transport of fluorescent PC in HepG2 cells (25).

Energy Depletion and Uptake of Transferrin-- To deplete ATP, cells were incubated for 30 min at 37 °C in Medium 2 containing 5 mM sodium azide and 50 mM 2-deoxyglucose. To quantify clathrin-dependent endocytosis, cells were incubated for 10 min with 5 µg/ml Alexa488-Tf, washed, and chased for 30 min at 37 °C. Cells were chilled with ice-cold Medium 1 or 2 and incubated in release medium for 10 min at 2 °C to remove surface-bound Tf by a mild acid wash. Those cells were washed with Medium 1 or 2 to adjust the pH to 7.4, and uptake of Alexa-Tf was quantified on the microscope by measuring fluorescence of Alexa-Tf as intensity per cell (see "Results"). In those experiments designed to determine the energy dependence of transport processes, all incubation steps except labeling with DHE/MCD and rhodamine-dextran were performed in Medium 2.

Incubation with Drugs to Disrupt the Cytoskeleton-- Cells were incubated with 33 µM nocodazole to disrupt microtubules for 30 min at 37 °C, prior to labeling with DHE/MCD as described above. Nocodazole was present during a subsequent chase for 30 min at 37 °C. Visualization of the microtubule cytoskeleton in control and drug-treated cells was performed according to Ref. 25.

Quantification of Cholesterol and Amount of Incorporated DHE-- Cells growing in 6-well cell culture plates were labeled with DHE/MCD as described above or kept non-labeled, washed extensively, and incubated for 5 min at 37 °C in trypsin (0.05% v/v). Cells were washed twice in Medium 1 by centrifugation at 400 × g. Lipids of the cell pellet were extracted in hexane/isopropyl alcohol 3:2 (v/v) as described for cholesterol (29). Solvent was evaporated under argon, and lipids were resolubilized in either reaction buffer (Molecular Probes) for quantification of cellular sterols or in ethanol to measure DHE. DHE was quantified by measuring fluorescence intensity in ethanol at _ex = 326 nm and _em = 370 nm on a Spex Fluorolog spectrofluorometer (Spex Industries, Inc., Edison, NJ). Sterols were quantified using a Quantification Kit from Molecular Probes, which is based on hydrolysis of cholesteryl esters to cholesterol and oxidation of cholesterol by cholesterol oxidase (30).

Time-lapse Fluorescence Microscopy and Fluorescence Recovery after Photobleaching (FRAP)

Time Course of Labeling Cells with DHE and Time-lapse Fluorescence Microscopy of DHE-- HepG2 cells were labeled with 2.5 mg/ml rhodamine-dextran for 1 h or with -BODIPY-PC for 1 min at 37 °C. They were washed and labeled with DHE/MCD for the indicated time points at 37 °C (see Fig. 1 and above). Cells were washed and either immediately placed on a temperature-controlled microscope stage maintained at 35 ± 1 °C of a wide field microscope or incubated for 30 min at 37 °C prior to imaging, respectively. BC were identified by rhodamine-dextran or -BODIPY-PC fluorescence in the red. For studying vesicle-based transport from the SAC/ARC in cells prelabeled with -BODIPY-PC, DHE in the plasma membrane was selectively extracted by incubation for 5 min with Medium 1 containing 5 mM MCD + 10 mM cholesterol-loaded MCD. Cells were washed, and images of DHE fluorescence were acquired every 1 min for a total time of 20 min.

FRAP of DHE-- HepG2 cells were labeled with 2.5 mg/ml rhodamine-dextran for 1 h (canalicular FRAP) or with -BODIPY-PC for 1 min at 37 °C (basolateral FRAP) and subsequently with DHE/MCD as described above. Cells were washed with pre-warmed buffer, and immediately the coverslip dish was placed on a temperature-controlled microscope stage maintained at 35 ± 1 °C of a wide field microscope. BC were identified by rhodamine-dextran fluorescence, and basolateral plasma membrane was indicated by -BODIPY-PC fluorescence in the red channel right after placing the dish on the microscope stage. After acquiring a fluorescence image of rhodamine-dextran or -BODIPY-PC, as well as an image of DHE (see below), the field aperture of the condenser was closed to ~13.3 µm diameter while placing the region of interest (BC or basolateral membrane) in the red channel into the center of the field. Prelabeling of the cells with rhodamine-dextran or -BODIPY-PC, respectively, did not influence the FRAP kinetics measured for DHE. By switching the filter set to DHE excitation (see below) and opening the fluorescence shutter for 1 min, the fluorescence of DHE was selectively bleached in the BC area or in the basolateral membrane. To study the transport of the sterol analogs to the bleached region the shutter was closed, the field aperture opened, and a fluorescence image of DHE was acquired after the indicated time points using a cooled CCD camera. Bleaching of an entire cell was complete after 1 min and did not result in any recovery. For the fluorescence loss in photobleaching (FLIP) experiment, the DHE in the basolateral membrane was bleached repeatedly with 3-min intervals between the bleaches to allow diffusion of DHE. Every bleach in the basolateral region resulted in a fluorescence loss of DHE in the couplet forming a BC of about 28%.

FRAP of -BODIPY-PC-- Cells were labeled for 2 min with -BODIPY-PC at 37 °C, washed, and placed on the microscope stage as described above for DHE. BC were first identified in the red channel and placed in the center of the field using a neutral density filter of 33% in order to prevent bleaching of the fluorophores. For bleaching of the BC and acquisition of images which was performed as described above for DHE, no neutral density filters were used.

Fluorescence Microscopy and Image Analysis

Wide field fluorescence microscopy and digital image acquisition were carried out using a Leica DMIRB microscope with a 63×, 1.4 NA oil immersion objective (Leica Lasertechnik GmbH, Wetzlar, Germany) equipped with a Princeton Instruments cooled CCD camera driven by Image-1/MetaMorph Imaging System software. -BODIPY-PC, Texas Red-ASOR, and rhodamine-dextran were imaged using a standard rhodamine filter set (535-nm (50-nm bandpass) excitation filter, 565-nm long pass dichromatic filter, and 610-nm (75-nm) bandpass) emission filter), whereas Alexa 488-Tf and NBD-lipids were imaged using a standard fluorescein filter set (470-nm, (20-nm bandpass) excitation filter, 510-nm long pass dichromatic filter, and 537-nm (23-nm) bandpass) emission filter). DHE was imaged using a specially designed filter cube obtained from Chroma Technology Corp. (Brattleboro, VT) with 335-nm (20-nm bandpass) excitation filter, 365-nm long pass dichromatic filter, and 405-nm (40-nm bandpass) emission filter as described previously (31). All other components of the microscope were adapted for UV imaging as described previously (18). Image analysis was carried out using the Image-1/MetaMorph Imaging System software (Universal Imaging Inc.). Determination and subtraction of crossover of fluorescence between the channels were performed as described (30, 32).

Colocalization and Time-lapse Experiments-- BC were first identified in the red channel by fluorescence of -BODIPY-PC focusing in the equatorial plane of the BC. In the case of studying colocalization with Texas Red-ASOR, a focal position 1 µm above the largest diameter of the cells was chosen as start point for the z-stack. Three serial focal plane images 0.5 µm apart above as well as below this position were acquired subsequently for all channels (i.e. DHE; -BODIPY-PC or Texas Red-ASOR; and Alexa-Tf, C₆-NBD-Cer, or C₆-NBD-SM). The wavelength-dependent variation in the focal position was determined independently by acquiring serial focal plane images of 0.1- and 0.5-µm TetraSpec fluorescent beads (Molecular Probes) mounted in gelvatol on a glass coverslip. No wavelength-dependent variation in the focal position was found between the red and green channels. However the corresponding focal plane for DHE was 1 µm below that of the red and green channel. Thus, a DHE image 1 µm below that of the other fluorescent markers was chosen to compare the distribution of the fluorescent probes. For all acquired stacks a Nearest Neighbor deconvolution algorithm implemented in MetaMorph software (Universal Imaging Inc.) was applied in order to remove out-of-focus fluorescence. This was of particular importance for DHE to improve the spatial resolution. The sum projection of individual stacks from single corresponding planes of each channel was generated using MetaMorph software. In time-lapse experiments vesicle positions were tracked manually by using the method of Barak and Webb (33), and vesicle velocity was calculated from the distance moved between successive images. Release of DHE from the SAC/ARC was quantified by measuring DHE fluorescence in an 8 × 8 pixel box placed in the center of the SAC/ARC and BC, respectively (Fig. 5).

Fluorescence resonance energy transfer (FRET) has been reported to take place between DHE (putative donor) and NBD analogs (putative acceptors) if they reside in the same compartment (34). To test whether FRET between DHE and NBD analogs might interfere with our colocalization studies, fluorescence of NBD analogs in double-labeled cells was bleached as described previously (25). An image of DHE fluorescence was acquired before and after bleaching NBD fluorescence in those cells. Fluorescence of DHE was quantified for both images in regions defined by the fluorescence of the Golgi (for C₆-NBD-Cer) or the SAC/ARC (for C₆-NBD-SM) from the corresponding image of the acceptor (see above) (FRET by sensitized acceptor fluorescence (35)).

FRAP Experiments-- For each image of the time series, a background correction was routinely performed by subtracting the intensity of regions without cells from the whole image. To quantify fluorescence intensity, regions of interest (ROI) were traced manually for the BC according to their geometry in the first image of the time series (i.e. in the prebleach image) and for two cells forming a BC. For basolateral FRAP an image of -BODIPY-PC with closed field aperture was chosen for defining the ROI. The integrated intensity was measured for the defined ROI and for the whole cells. The ratio, R, of fluorescence of ROI and of whole cells was then calculated and normalized as described previously (25). Fluorescence loss of DHE in the canalicular membrane due to basolateral FRAP was calculated by normalizing all fluorescence ratios of a time series to the fluorescence ratio in the first image after bleaching. The data of 9 to 10 or 5 measurements for DHE and fluorescent PC, respectively, were averaged for presentation purposes, whereas all individual recovery kinetics were fitted to a mono-exponential function using Sigma Plot 4.0 (SPSS Inc., Chicago).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Labeling with DHE/MCD Results in Bright and Exclusive Staining of the Plasma Membrane Domains of HepG2 Cells-- To follow the transport of cholesterol analogs from the plasma membrane to intracellular compartments requires (i) sufficient labeling of cells with those analogs and (ii) initial staining exclusively of the plasma membrane. First, the time course of labeling HepG2 cells with DHE was studied. Cells were labeled for 10-60 s at 37 °C, washed, and placed on the microscope stage. DHE rapidly inserts into the plasma membrane of HepG2 cells incubated with the labeling solution containing DHE loaded on MCD (DHE/MCD) (Fig. 1A). After a 10-s pulse, faint staining of the cells was found, and after 20 s DHE was seen in both the basolateral and the canalicular membrane (arrows). After 60 s of labeling, cellular fluorescence intensity of DHE was very bright and could not be increased by longer pulse times (not shown). Thus, in subsequent experiments cells were pulse-labeled for 1 min at 37 °C. Note that images in Fig. 1A have the same intensity scaling to allow visual comparison of the fluorescence intensity of DHE after different times of labeling. The total cellular fluorescence intensity of DHE is more than 5-fold above cellular autofluorescence in the UV region of the spectrum after a 1-min pulse at 37 °C. The amount of incorporated DHE corresponded to 10 mol % of sterols in the cells. Under our labeling conditions about 0.5% of DHE from DHE/MCD became incorporated into the plasma membrane of HepG2 cells. Incubating cells with DHE/MCD did not alter the total sterol content of HepG2 cells which was 20 ± 4 nmol per 10⁶ cells. This indicates that DHE supplied to cells via MCD replaces cholesterol in the plasma membrane without altering cellular sterol homeostasis. We tested uptake of Tf and found no difference between DHE-labeled and non-labeled cells, indicating that clathrin-coated pit internalization, which is altered by cholesterol depletion (36, 37), is not affected by the labeling procedure.

View larger version (73K): [in this window] [in a new window]   Fig. 1.   Transport of DHE after pulse labeling of the plasma membrane with DHE/MCD. A, HepG2 cells were labeled with rhodamine-dextran (70 kDa) for 1 h at 37 °C, washed, and labeled with DHE/MCD for 10 s, 20 s, or 1 min at 37 °C. All images in the upper panel are of the same intensity scaling. The lower panel shows rhodamine-dextran fluorescent BC (arrow) corresponding to the upper panel for DHE. B, cells were labeled with DHE/MCD for 1 min at 37 °C, washed, and incubated for 0, 7, or 30 min at 37 °C. Bright labeling of the basolateral and canalicular membrane (thick arrows) but no intracellular staining was observed after pulse labeling. Already after 7 min of incubation DHE-labeled vesicles (thin arrows) and a perinuclear enrichment of DHE in non-polarized cells (thin arrowheads) were found. Incubation for 30 min at 37 °C resulted in a subapical enrichment of DHE in polarized cells (thick arrowheads). Bar, 20 µm.

In Fig. 1B the time course of uptake of DHE from the plasma membrane is shown. In contrast, to Fig. 1A, the contrast of images was adjusted to distinguish plasma membrane labeling of DHE from its intracellular fluorescence in labeled compartments. Labeling for 1 min at 37 °C resulted in bright, homogeneous staining of the plasma membrane of HepG2 cells with DHE (Fig. 1B, 0 min chase). After a 7-min chase, diffuse cytoplasmic staining as well as vesicles carrying DHE (thin arrows) became apparent (Fig. 1B). After 30 min of chase at 37 °C, additional accumulation of DHE in the perinuclear (thin arrowheads) or subapical region (thick arrowheads) in nonpolarized and polarized cells was found, respectively. At later time points, as the labeling in the perinuclear and subapical regions became brighter, distinct membrane structures (i.e. vesicles and tubules) could be seen. The DHE distribution did not change significantly after 30 min of chase. The DHE-labeled vesicles and the apparently organelle-associated intracellular DHE distribution at steady state indicated that a vesicular pathway might transport DHE between intracellular organelles in HepG2 cells.

DHE Colocalizes with Markers of the Endocytic Recycling System in HepG2 Cells-- To identify the organelles along the intracellular itinerary of DHE, cells were colabeled with DHE/MCD, the fluorescent PC analog -BODIPY-PC (fluorescence emission in red), and Alexa488-Tf (fluorescence in green), chased for 30 min at 37 °C, fixed, and observed by fluorescence microscopy. The trafficking of -BODIPY-PC in HepG2 cells has been described previously (25). In polarized couplets of HepG2 cells, DHE as well as -BODIPY-PC were enriched in the canalicular membrane (thick arrows), and they colocalized in vesicles in the subapical region of the cells (thick arrowheads) and in vesicles scattered throughout the cytoplasm (thin arrowheads) (Fig. 2, A and B). These vesicles were often triple-labeled with both lipid probes and with transferrin (Tf) (Fig. 2C). In addition, vesicles labeled with DHE and -BODIPY-PC but not with Alexa-Tf were found (thin arrows). Colocalization in vesicles in the subapical region is seen more clearly at higher magnification (thick arrowheads) (Fig. 2, D-F). Because the observed vesicles contain markers of the recycling pathway (i.e. fluorescent Tf), they probably belong to a subapical compartment or apical recycling compartment (SAC/ARC) as described in HepG2 cells as well as in the hepatocytic cell line WIF-B (24, 25, 38, 39). Some vesicles and tubules labeled only with DHE were found in the subapical region of the cells (thin arrowheads in Fig. 2D).

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Colocalization of DHE with markers of the endocytic system. A-L, cells were labeled with 5 µg/ml Alexa-Tf, DHE/MCD, and with -BODIPY-PC as described under "Experimental Procedures." Cells were washed and subsequently incubated at 37 °C for 30 min, washed, and after a brief back-exchange with ice-cold BSA (2 times for 5 min, 5% BSA w/v on ice), fixed with PFA. Fixed cells were imaged as described under "Experimental Procedures." DHE (A, D, G, and J) as well as -BODIPY-PC (B, E, H, and K) became enriched in the canalicular membrane (thick arrows). All three probes were found in the subapical region (thick arrowheads) (A-C). Additionally, vesicles containing all probes (A-C, thin arrowheads) or only DHE and -BODIPY-PC (A-C, thin arrow) but not Alexa-Tf (C, F, I, and L) were found throughout the cytoplasm. The zoomed regions of A-C show that DHE colocalizes with -BODIPY-PC and Alexa-Tf (thick arrowheads) in vesicles underneath the BC in polarized cells (D-F) and in a perinuclear compartment in nonpolarized cells (J-L), respectively. Vesicles and tubular structures labeled only with DHE were also found (thin arrowheads). M-O, DHE (M) colocalizes with C6-NBD-SM (N) in the canalicular membrane (central ring) and in adjacent vesicles (thick arrowheads); however, also vesicles only labeled with DHE but not with C6-NBD-SM were found (thin arrowhead). O, overlay with DHE in green and C6-NBD-SM in red, colocalizing regions appear orange. Bar, 20 µm.

Colocalization of DHE with -BODIPY-PC and Alexa488-Tf was also found in non-polarized HepG2 cells (Fig. 2, G-L). Those cells are defined by the absence of an apical membrane (i.e. no BC) and the perinuclear localization of recycling endosomes similar to that found in CHO cells (17, 23). In non-polarized HepG2 cells overlap of all markers in the perinuclear region was found (Fig. 2, G-I, arrows). Triple-labeled vesicles were detected in this area (Fig. 2, J-L, thick arrowheads). The degree of colocalization of DHE with the recycling markers seemed to be higher in non-polarized cells than in polarized cells. These results show that DHE partially associates with the recycling pathway in both polarized and non-polarized HepG2 cells.

Previously, it was shown (23) that fluorescent sphingolipids are transcytosed and delivered to the BC via a subapical compartment. As shown in Fig. 2, M-O, DHE colocalizes in the BC as well as in subapically enriched vesicles (thick arrowheads) with C₆-NBD-SM. In addition, vesicles labeled only with DHE (thin arrowheads) were found. These results show that some DHE derived from the plasma membrane is delivered to the SAC/ARC, which contains markers of the recycling as well as the transcytotic pathway.

The TGN Is Not a Major Pool of Intracellular DHE-- One site of intracellular cholesterol accumulation has been proposed to be the TGN (for a review see Refs. 12 and 40). We examined whether DHE derived from the plasma membrane is transported to the TGN. First cells were colabeled with DHE and BODIPY-Cer, which is a vital stain for the TGN (Fig. 3, A-I). DHE colocalized to some extent in tubules with BODIPY-Cer detected in the green channel in polarized cells (Fig. 3, A-C, arrowheads). However, regions with enrichment of BODIPY-Cer, identified by its red emission due to formation of excimers (41), largely lacked DHE (not shown). This suggests that DHE is not highly enriched in the TGN. In nonpolarized HepG2 cells DHE was mostly found in the perinuclear region. Some DHE did overlap in tubules or dots containing BODIPY-Cer in the rim surrounding the perinuclear area (thin arrowheads). A very similar fluorescence pattern was observed in cells that are in the process of forming a BC (semi-polarized cells). Those HepG2 cells are characterized by the diaphragm-like structure on the apical pole of the adjacent cells still segregating the two canalicular membranes (Fig. 3, G-I, arrows). In those semi-polarized cells enrichment of DHE in the two regions resembling the recycling compartment (thick arrowheads) and the TGN (thin arrowheads) was found, but both compartments did not yet have a subapical location like in fully polarized cells (see Fig. 3, A-C). In an independent experiment colocalization of DHE and -BODIPY-PC with C₆-NBD-Cer, another vital stain for the TGN (42), was investigated. The TGN, labeled with C₆-NBD-Cer, was not a major site for accumulation of DHE or -BODIPY-PC (Fig. 3, J-O). The nearest neighbor correction for out of focus fluorescence imaging protocol outlined under "Experimental Procedures" allowed us to clearly distinguish Golgi tubules or stacks from vesicles stained with DHE and -BODIPY-PC. Typically, DHE and -BODIPY-PC were excluded from C₆-NBD-Cer positive Golgi tubules (see outlined tubules in Fig. 3, J-L and M-O for polarized and nonpolarized cells, respectively). However, part of the DHE labeling seemed to overlap with C₆-NBD-Cer in the regions adjacent to the outlined tubules (Fig. 3, J-O). This suggests that some of the sterol is also delivered to the TGN. FRET has been reported to occur between DHE and NBD analogs if they reside in the same compartment in the cell (34), and this could have reduced DHE fluorescence in the TGN. We tested this possibility by bleaching the putative acceptor, C₆-NBD-Cer (35). No fluorescence increase of DHE could be detected after bleaching of C₆-NBD-Cer, so the possibility that part of DHE overlapping with C₆-NBD-Cer is quenched in the TGN can be excluded. For both ceramide analogs we found no increased colocalization in cells labeled with DHE and chased for 1.5 h at 37 °C (not shown). This indicates that DHE residing in the SAC/ARC or the plasma membrane does not redistribute significantly to the TGN after longer incubation. Taken together, the results indicate that DHE is targeted to some extent to the TGN, but they also show that this organelle is not a major pool for DHE derived from the plasma membrane in HepG2 cells.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Transport of DHE to the TGN in polarized and in nonpolarized HepG2 cells. Cells were labeled with DHE/MCD, washed, and incubated for 20 min at 37 °C. Cells were labeled with BODIPY-Cer for 10 min at 37 °C, washed, and imaged. DHE (A, D, and G) labeled the BC (arrow) and colocalized with BODIPY-Cer (B, E, and H) in tubules (thin arrowheads) underneath the BC in polarized cells BC (A-C) and of a perinuclear rim or dot in non-polarized (D-F) and semi-polarized cells (G-I), respectively. DHE additionally stained a clearly segregated region in the perinuclear area of non-polarized (D-F) and semi-polarized cells (G-I) (thick arrowheads). C, F, and I, overlay with DHE in green and BODIPY-Cer in red. Colocalizing regions appear orange. J-O, HepG2 cells were labeled with DHE/MCD and with -BODIPY-PC as described under "Experimental Procedures." Cells were washed and subsequently incubated at 37 °C for 30 min, washed, and after a brief back-exchange (2 × 5 min, 5% BSA w/v on ice), fixed with PFA. Fixed cells were incubated for 5 min at 37 °C with C6-NBD-Cer, washed, and imaged as described under "Experimental Procedures." DHE (J and M) as well as -BODIPY-PC (K and N) became enriched in the canalicular membrane (arrows), in a subapical region in polarized cells (J-L), and a perinuclear region in nonpolarized cells (M-O), respectively. Only partial overlap is found in the central part of those regions whereas most of C6-NBD-Cer (L and O) is in tubules not containing DHE or -BODIPY-PC (see outlined areas). Bar, 20 µm.

Bidirectional Movement of DHE Carrying Vesicles and the Involvement of the Cytoskeleton in Transport of DHE-- DHE derived from the plasma membrane becomes enriched in vesicles of the SAC/ARC as shown by its colocalization with fluorescent Tf, PC, and sphingolipids (Fig. 2). The intracellular route and time course of vesicle-based transport of DHE was next investigated. In time-lapse experiments, we found movement of some vesicles containing DHE (i) from the basolateral region toward the BC, (ii) from the apical pole directed toward the basolateral membrane, and (iii) from the BC toward the SAC/ARC. However, in all acquired video sequences (n = 15) the number of vesicles released from the SAC/ARC pool was found to be very low. A typical time sequence demonstrating all types of movement of DHE-carrying vesicles is shown in Fig. 4A. In the first 4 min of the sequence two DHE labeled vesicles (green arrowheads) moved from the vesicular pool resembling the SAC/ARC (arrow) toward the basolateral membrane. Between 2 and 4 min after starting the sequence, a vesicle moved from the BC toward the SAC/ARC pool (red arrowheads). This suggests apical endocytosis of DHE. Between 9 and 14 min after starting the time sequence another vesicle moved from the basolateral region toward the apical pole of the cell. Longer acquisition was not possible because of artifacts introduced by photobleaching after about 14 min of recording. Thus, we were unable to record a complete vesicle path from one plasma membrane domain to the other. The velocity of the vesicle movements was found to be as fast as 0.18 µm/s as measured by distance moved between two captured images (data from 4 video sequences recorded for 12-14 min). The measured vesicle velocity is similar to that found previously for fluorescent PC and is in line with those observed for motor driven movement along microtubule tracks (25, 43). Further evidence that DHE vesicles move on microtubule tracks was obtained from experiments with drugs disrupting the microtubule cytoskeleton. Incubating cells with nocodazole resulted in scattering of the SAC/ARC (24, 25), so DHE-containing vesicles became dispersed throughout the cytoplasm (Fig. 4C). In contrast, labeling of the basolateral or canalicular membrane with DHE was not affected by nocodazole treatment (Fig. 4, B and C). This indicates that vesicle-based transport of DHE but not its delivery to either plasma membrane domain in polarized HepG2 cells requires an intact microtubule cytoskeleton.

View larger version (64K): [in this window] [in a new window]   Fig. 4.   Visualization of vesicle-based transport of DHE in polarized HepG2 cells. A, cells were labeled for 1 min at 37 °C with DHE/MCD, washed, incubated for 30 min at 37 °C, and placed on a temperature-controlled microscope stage of a wide-field microscope maintained at 36 ± 1 °C. Images were acquired every minute using a cooled CCD camera with 2 × 2 binning. Two vesicles (green arrowheads) moved from the vesicle pool resembling the SAC/ARC (arrowhead) toward the basolateral membrane (0, 2, 3, and 4). Those vesicles moved together and are therefore indicated by the same color. Another vesicle moving from the BC (arrow) to the SAC/ARC shown after 2 min (red arrowheads) was found. Moreover, a vesicle moving from the basolateral plasma membrane toward the apical pole of the cell was found between image 9 and 12 min after starting the time lapse (blue arrowheads). The nucleus of the imaged cell with tracked vesicle paths is located below the arrow pointing to the BC. The bright spot in the adjacent nonpolarized cell located left to the observed cell resembles a perinuclear recycling compartment. The observed vesicle pathways are summarized in a schematic in the lower right panel. In this image tracked vesicle paths are shown and colored according to the color of the arrowheads for outlined vesicles (see above). Arrows indicate the direction of observed vesicle movement. Cells were incubated with 33 µM nocodazole (B and C) for 30 min at 37 °C, washed, and labeled with DHE/MCD for 1 min at 37 °C. Cells were washed and either immediately imaged (B) or incubated for 30 min at 37 °C in the presence of the drug (C). In cells treated with nocodazole to disrupt microtubules, DHE labeled the canalicular membrane (arrows) at 0 and 30 min chase as found in control cells (see Fig. 1). However, vesicles were scattered throughout the cytoplasm after 30 min of chase (arrowheads) but not enriched in the subapical or perinuclear region in polarized or nonpolarized cells, respectively. Bar, 20 µm.

Contribution of Vesicular Transport of DHE to Sterol Exchange between the Plasma Membrane Domains Is Low-- To obtain quantitative information about the role of vesicular traffic from the SAC/ARC in transport of DHE between the plasma membrane domains, cells were colabeled with -BODIPY-PC and DHE/MCD and chased for 30 min at 37 °C. Plasma membrane-associated DHE was removed by incubating cells with medium containing 5 mM MCD and 10 mM cholesterol-loaded MCD (Fig. 5A). The intracellular vesicle pool, which is noted as the SAC/ARC adjacent to the BC (outlined by a circle), remained well labeled by DHE (thick arrowheads). Interestingly, the BC identified by the corresponding fluorescence image of -BODIPY-PC (1st panel) does not contain DHE after this extraction procedure (arrows). During the time-lapse sequence some DHE rapidly redistributes to the BC as indicated by the faint staining in this area seen already after 2 min of chase at 37 °C. Vesicles adjacent to the SAC/ARC were occasionally observed (thin arrowheads), but they did not move vectorially toward the BC or basolateral plasma membrane. Little redistribution of DHE to the basolateral or canalicular membrane could be observed. Quantification of these results is shown in Fig. 5B. Only about 10% of initial DHE fluorescence is released during the chase. The measured half-time of this process is 2.9 min, and no further release could be observed during longer incubations. Some DHE was transported to the BC with a half-time of 1.7 min raising its initial fluorescence by about 10%. These results indicate that vesicles of the SAC/ARC containing DHE do not contribute to a large extent to transport of DHE between the basolateral and canalicular membrane.

View larger version (61K): [in this window] [in a new window]   Fig. 5.   Release of DHE from the SAC/ARC and transport to the canalicular membrane studied by time-lapse fluorescence microscopy. Cells were colabeled with -BODIPY-PC and DHE/MCD and chased for 30 min at 37 °C. Plasma membrane-associated DHE was removed by incubating cells with Medium 2 containing 5 mM MCD and 10 mM cholesterol-loaded MCD for 5 min at 37 °C. Cells were washed and placed on a temperature-controlled microscope stage of a wide-field microscope maintained at 35 ± 1 °C. A, the incubation in extraction medium resulted in complete removal of DHE from the basolateral membrane and from the BC. The latter is indicated by a circle between the brightly labeled SAC/ARC (thick arrowheads) as judged from fluorescence of the corresponding image of -BODIPY-PC (not shown) (see corresponding phase contrast image for cell borders and position of BC (arrows)). Some labeled vesicles could be seen after 2 min of incubation after starting the time lapse (thin arrowhead); however, no cross-redistribution of DHE from the SAC/ARC toward the BC or basolateral membrane could be observed. B, quantification of fluorescence of DHE in the SAC/ARC (black circles) and the BC (white circles) normalized to the fluorescence in the first image (i.e. 0 min) after starting the time lapse. The solid lines indicate a monoexponential fit to the data. Data represent mean ± S.E. of 4 measurements. Bar, 20 µm.

Transport of DHE to the Canalicular Membrane in ATP-depleted HepG2 Cells-- Vesicular traffic of proteins and lipids are energy-requiring processes (44, 45). To study the relevance of endocytosis and vesicular traffic in DHE distribution and transport, cells were energy-depleted by incubation in azide and 2-deoxyglucose for 30 min at 37 °C. Under these conditions, uptake of Alexa-Tf was completely blocked, indicating inhibition of clathrin-dependent endocytosis (compare Fig. 6, C and F). Uptake of Alexa-Tf in ATP-depleted cells was 0.7 ± 0.7% that in control cells (mean ± S.E. of three fields with about 6-10 cells per field for each condition). Resupplying glucose to ATP-depleted cells not treated with a mild acid wash to remove surface-bound Tf resulted in endocytosis of Alexa-Tf and its delivery to an endocytic recycling compartment as found in control cells (not shown). Energy depletion blocked the vesicular intracellular enrichment of DHE (Fig. 6D) as well as of -BODIPY-PC (Fig. 6C) as found after 30 min at 37 °C in control cells in the subapical (thick arrowheads) and perinuclear region (thin arrowheads) of the cells (Fig. 6, A and B). In contrast, ATP depletion did not affect the enrichment of both lipids in the BC (arrows) or the diffuse staining of the cytoplasm by both analogs (Fig. 6, D and E). Note that the images shown in Fig. 6 were not deconvolved like those in Fig. 2 which are also sum projections of individual planes (see above and "Experimental Procedures"). This explains the slight difference in intracellular distribution of DHE in Fig. 6A compared with Fig. 2A. In control cells as well as in ATP-depleted cells fluorescence of DHE in the canalicular membrane quantified as fraction of total cellular DHE fluorescence increased by about 15% during the first 1 min after removing the DHE/MCD-labeling solution and remained stable for at least 20 min. This shows that DHE continues to be transported to the BC during the chase in both conditions, and this transport is very rapid.

View larger version (60K): [in this window] [in a new window]   Fig. 6.   ATP depletion blocks vesicular uptake of DHE and -BODIPY-PC but not the canalicular enrichment of both lipid probes. HepG2 cells were washed and incubated for 30 min at 37 °C in Medium 1 (A-C, control cells) or in ATP-depletion medium (D-F, ATP-depleted cells). Cells were colabeled with DHE/MCD (A and D) and -BODIPY-PC at 37 °C (B and E). After incubation for 30 min at 37 °C in control cells, subapical (thick arrowheads) and perinuclear (thin arrowheads) enrichment of DHE and -BODIPY-PC as well as brightly stained BC were found (A and B) (arrows). In ATP-depleted cells only diffuse intracellular fluorescence but no vesicular uptake was found for DHE (D) and -BODIPY-PC (E). However, BC were labeled with both probes (arrows). As a control for ATP depletion, uptake of Alexa-Tf was studied in a separate experiment in which cells were labeled for 10 min at 37 °C with Alexa-Tf (C and F). ATP depletion resulted in complete inhibition of uptake of Alexa-Tf (F), which was in control cells accumulated in a recycling compartment (arrowheads, C). Bar, 20 µm.

FRAP Reveals Rapid Transport of DHE and -BODIPY-PC to the Canalicular Membrane-- To obtain kinetic information about transport of DHE to the BC, a FRAP technique was employed. BC were bleached after pulse labeling at 37 °C, and the fluorescence recovery was measured on a temperature-controlled microscope stage. As shown in Fig. 7A, DHE fluorescence in the BC (arrows) is strongly reduced after the bleach, but the BC fluorescence recovers substantially within 5 min. No vesicular transport of DHE to the canalicular membrane was observed during the measurement, suggesting that most DHE is rapidly transported to the BC by a nonvesicular pathway. The kinetics of recovery of DHE fluorescence in the BC were quantified to provide a rate and a fraction of recovery (Q) as shown in Fig. 7B and Table I. The measured half-time for recovery was 1.6 min for control cells (Fig. 7B, black circles). The fraction of recovery, Q = 0.77, suggests that there is a portion of DHE that cannot be delivered to the BC in the time course of the experiment. Remarkably, in ATP-depleted cells fluorescence recovery was even slightly faster than in control cells (t_1/2 = 1.2 min), whereas the recovery fraction was reduced (Q = 0.55) (Fig. 7B, open circles). Because vesicular traffic of DHE was blocked after ATP depletion (Fig. 6), these results indicate that nonvesicular transport accounts for most of the canalicular transport of DHE.

View larger version (82K): [in this window] [in a new window]   Fig. 7.   Transport of DHE to the canalicular membrane studied by FRAP. Cells were labeled with rhodamine-dextran (70 kDa) for 1 h at 37 °C, washed, and labeled with DHE/MCD for 1 min at 37 °C. After washing cells were placed on a temperature-controlled microscope stage of a wide-field microscope maintained at 35 ± 1 °C. A, after acquiring an image of cells containing a BC (arrows), the BC region was bleached by closing the field aperture to about 13 µm in the object plane as described under "Experimental Procedures." Subsequently the field aperture was opened, and images were acquired at the indicated times. B, normalized fluorescence recovery for DHE in control (black circles) or ATP-depleted cells (white circles) and for -BODIPY-PC (triangles) in control cells. A mono-exponential fit for enrichment of DHE (solid lines) and -BODIPY-PC (dashed line) in the BC are shown. Data represent mean ± S.E. of 9-10 (DHE) or 5 measurements (-BODIPY-PC). Bar, 20 µm.

View this table: [in this window] [in a new window]   Table I Quantification of transport of DHE and -BODIPY-PC as studied by FRAP Cells were labeled, and FRAP of DHE and -BODIPY-PC was performed for the BC or the basolateral plasma membrane as described under "Experimental Procedures." Images of DHE and -BODIPY-PC were acquired prior and after the bleach at the indicated time points (see Figs. 7 and 8). Fluorescence of DHE and -BODIPY-PC in the ROI (BC or basolateral membrane) and in the BC-forming cells was quantified after background subtraction of the images. Recovery curves normalized to the prebleach image were fitted with a mono-exponential function using Sigma Plot 4.0 (SPSS Inc., Chicago).

The kinetics of transport of -BODIPY-PC to the canalicular membrane have been measured previously by quantifying fluorescence of -BODIPY-PC in the BC immediately after pulse labeling of the basolateral plasma membrane with this analog. A half-time for this process at 3.5 min was determined (25). Fig. 7B shows the kinetics of recovery of -BODIPY-PC in the BC using the FRAP method (triangles). A half-time for recovery of 3.0 min was obtained for -BODIPY-PC. The half-time for enrichment of this PC analog in the BC closely resembles that previously measured with an independent method (25). The fraction of fluorescence recovery of -BODIPY-PC (Q = 1.0) points to some differences in canalicular transport of sterols versus a fluorescent PC analog. The result is in line with a nonvesicular transport of fluorescent PC to the canalicular membrane as demonstrated previously (25).

ATP-independent release of DHE from the Canalicular Membrane and Redistribution to the Basolateral Domain-- The surprisingly fast and ATP-independent recovery of DHE fluorescence in the BC raised the question whether this rapid sterol enrichment is a property unique to the canalicular membrane. We next studied the kinetics of transport of DHE from the BC to the basolateral plasma membrane. Cells were labeled with DHE/MCD and -BODIPY-PC. After acquiring an image of DHE, fluorescence of DHE in the basolateral membrane was selectively bleached, and fluorescence recovery in the basolateral as well fluorescence loss in the canalicular membrane were measured. As shown in Fig. 8, A and B, DHE rapidly recovers in the bleached basolateral area. ATP depletion did not affect the extent of fluorescence recovery, which was about 100% (Table I). However, the kinetics of fluorescence recovery in the basolateral region is faster in ATP-depleted cells (Fig. 8B, white circles) (t_1/2 = 1.4 min) compared with control cells (Fig. 8B, black circles) (t_1/2 = 2.8 min). Release of DHE from the BC due to the basolateral bleach was rather low during the time course of the experiment (Fig. 8, A and C). This suggests replenishment of DHE from regions outside the bleach spot in the basolateral membrane. ATP depletion slightly accelerated release of DHE from the BC (t_1/2 = 4.1 min) compared with control cells (t_1/2 = 5.4 min). The results indicate that vesicular transport of DHE is not a major factor in delivery of this sterol to the basolateral membrane.

View larger version (37K): [in this window] [in a new window]   Fig. 8.   Release of DHE from the canalicular membrane and transport to the basolateral domain studied by FRAP. Cells were labeled with -BODIPY-PC for 1 min at 37 °C, washed, and labeled with DHE/MCD for 1 min at 37 °C. After washing cells were placed on a temperature-controlled microscope stage of a wide-field microscope maintained at 35 ± 1 °C. A, after acquiring an image of cells containing a BC (arrows), the basolateral region (outlined circle) was bleached by closing the field aperture as described under "Experimental Procedures." Subsequently the field aperture was opened, and images were acquired at the indicated times. B, normalized fluorescence recovery for DHE in control (black circles) or ATP-depleted cells (white circles) in the basolateral plasma membrane. C, normalized fluorescence of DHE in the BC during the basolateral recovery in control (black circles) or ATP-depleted cells (white circles). A mono-exponential fit for DHE fluorescence (solid lines) in control or ATP-depleted cells is shown. Data represent mean ± S.E. of 8-13 measurements. Bar, 10 µm.

DHE in the Canalicular Membrane Can Completely Exchange with DHE in the Basolateral Plasma Membrane of the Same Cell-- The slower release of DHE from the BC compared with its recovery in the basolateral membrane might have been explained by replenishment of DHE from the second unbleached cell of a couplet, while being continuously released from the canalicular membrane. Alternatively, DHE could be released slowly from the BC and then transported to the basolateral membrane. To distinguish between these possibilities, we performed a FLIP experiment. As shown in Fig. 9, DHE fluorescence throughout one cell was destroyed due to repeated bleaching, whereas it appeared stable in the second cell forming the BC (thin arrows). Interestingly, DHE fluorescence could be destroyed in the canalicular membrane of the bleached, but not the unbleached, cell as indicated by the brightly labeled half-BC in images after bleaching (Fig. 9, thick arrows). The nonpolarized cell adjacent to that forming the couplet was also completely bleached because it was also illuminated by the bleach beam (see outlined circle). However, no fluorescence loss of DHE in neighboring unbleached cells was observed, indicating that DHE does not diffuse rapidly from one cell to the next. The experiments demonstrate that DHE does not cross the BC from one cell to the other of an HepG2 couplet, nor is it released into the BC lumen. However, DHE associated with the canalicular membrane can completely redistribute to the basolateral membrane. The slow loss of DHE from the BC observed following a single bleach of the basolateral membrane (see above and Fig. 8) is therefore not due to continuous sterol transport from one cell of a couplet to the other. We conclude that DHE associates preferentially with the canalicular membrane but that it can be transported from this membrane to the basolateral membrane by a nonvesicular mechanism.

View larger version (82K): [in this window] [in a new window]   Fig. 9.   Dynamic exchange of DHE between the canalicular and basolateral membrane studied by FLIP. Cells were labeled with -BODIPY-PC for 1 min at 37 °C, washed, and labeled with DHE/MCD for 1 min at 37 °C. After washing cells were placed on a temperature-controlled microscope stage of a wide-field microscope maintained at 35 ± 1 °C. After acquiring an image of cells containing a BC (arrows) the basolateral region was bleached as described under "Experimental Procedures." The field aperture was opened, an image of DHE fluorescence was acquired, and DHE fluorescence was allowed to recover in the bleached region for 3 min. The field aperture was closed and cells were bleached again. This cycle was repeated until no further fluorescence loss in the bleached cell could be observed. Fluorescence loss including the canalicular membrane was observed for the bleached cell but not for the neighboring cell of the couplet (thin arrow). Fluorescence of DHE in the BC (thick arrow) decreased only half, i.e. in the bleached but not the non-bleached cell. Neighboring cells also forming a BC (arrowhead) appeared bright during the whole experiment (see text for further explanations). Bar, 20 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cholesterol is crucial for establishment and maintenance of the polarized state in epithelial cells. However, the intracellular transport pathways of cholesterol in polarized cells are not well understood. We have studied the intracellular itineraries and dynamics of a cholesterol analog, DHE, in polarized HepG2 cells. We find that DHE transport occurs by a combination of several pathways. Some DHE is found associated with organelles and is seen to move through the cytosol in ways that are consistent with vesicular traffic. However, the majority of sterol transport into and out of the apical and basolateral plasma membranes occurs by an ATP-independent, nonvesicular process. The results of this study are summarized in a proposed model of sterol transport in polarized HepG2 cells as shown in Fig. 10.

View larger version (38K): [in this window] [in a new window]   Fig. 10.   Model of the transport pathways of sterol between the plasma membrane domains of polarized HepG2 cells. DHE incorporated into the exoplasmic leaflet of the plasma membrane of HepG2 cells can be transported in vesicles, summarized as early endosomes (EE), to the SAC/ARC and from there to the opposite plasma membrane domain (1). Additionally, DHE traverses the plasma membrane by rapid transbilayer migration, indicated by the zoomed region of both plasma membrane domains, and traffics to the opposite membrane domain by lateral diffusion along the inner plasma membrane leaflet (2) or through the cytosol bound to a cytosolic protein carrier (3). All transport pathways (arrows) are bi-directional. Tight junctions (TJ) might prevent lateral diffusion between the membrane domains in the exoplasmic plasma membrane leaflet. Sterol-enriched compartments are black (i.e. the basolateral and canalicular membrane and the SAC/ARC), and those with low or intermediate sterol content judged by poor or absent labeling with DHE are shown in gray (i.e. the Golgi or TGN and other endocytic compartments named here as EE and the nucleus/endoplasmic reticulum membrane system (Nu)). Sterol transport is not possible from one cell of a couplet to the other by crossing either the inter-membrane space along the basolateral membrane or by crossing the biliary canaliculus by release from the canalicular membrane into the BC lumen (crossed arrows). See text for further explanation.

We demonstrate that DHE is transported to the SAC/ARC as shown by its colocalization with -BODIPY-PC, with the recycling marker Alexa-Tf and especially with C₆-NBD-SM. Interestingly, vesicular structures were found in the region of the SAC/ARC in HepG2 cells that were labeled only with DHE but not with -BODIPY-PC, Alexa-Tf, or C₆-NBD-SM (Fig. 2). This points to the heterogeneous nature of the organelles in the SAC/ARC region as suggested earlier (46-48). The cholesterol content of late endosomes and lysosomes was reported to be 6-10% of total cellular cholesterol (49). In order to test whether DHE is transported also to these compartments, HepG2 cells were colabeled with DHE and Texas Red-ASOR, a marker of late endosomes and lysosomes as described under "Experimental Procedures." Notably, no colocalization of DHE with Texas Red-ASOR was found, suggesting that the amount of DHE in late endosomes/lysosomes is too low to be detected (not shown).

The fluorescence of DHE in the SAC/ARC reached a steady state after about 30 min of chase from the plasma membrane at 37 °C (Fig. 1). Note that cholesterol derived from the plasma membrane gets esterified at a rate of 0.25% per min in hepatoma cells (50). Moreover, a lag time due to sterol transport from the plasma membrane to intracellular compartments prior to the onset of sterol esterification has been demonstrated (51). If DHE is esterified with the same kinetics as cholesterol, maximally 7.5% of cholesterol or DHE derived from the plasma membrane should become esterified during the time course of our experiments.

It has been proposed that the Golgi complex is a major cholesterol-containing organelle, playing an important role in polarized sorting of membrane components in the biosynthetic pathway (40, 52). Cholesterol and (glyco)sphingolipids have been proposed to form ordered microdomains or rafts in the TGN (for a review see Ref. 12). By using DHE as a marker of cholesterol in the plasma membrane, we found that most vesicular DHE resides in vesicles belonging to the SAC/ARC and that only a small fraction is found in the TGN at steady state. This suggests that the sterol content of the TGN is lower than that of the plasma membrane or of recycling endosomes, as we found previously in CHO cells (18, 31). We cannot rule out the presence of high levels of sterol in small regions of the TGN nor did we investigate the fate of newly synthesized sterol. It might be that de novo synthesized cholesterol accumulates in the TGN creating the lipid environment in this compartment that is required for polarized sorting. However, although several studies (48, 53, 54) have shown that the recycling compartment is enriched in cholesterol, the published data on cholesterol content of the Golgi apparatus is contradictory and depends on the method used for purification of Golgi membranes (18, 55-57). It has been reported that de novo synthesized cholesterol reaches the plasma membrane with a half-time of about 10 min and that this transport only partially involves the Golgi complex (58, 59). According to our results, synthesized cholesterol exported to the plasma membrane would be reinternalized with a similar half-time but delivered to the SAC/ARC keeping the sterol concentration in the TGN low. Strikingly, we did not find targeting of DHE to the TGN even after incubation at 37 °C for 1.5 h (not shown).

We found that ATP depletion inhibited delivery of DHE to the SAC/ARC and other intracellular membrane organelles but did not interfere with the rapid transport of DHE to the BC. In addition, DHE associated with the canalicular membrane could redistribute to the basolateral plasma membrane in an ATP-independent manner (Figs. 7 and 8). Although some directed vesicle movement was observed (Fig. 4), DHE associated with membranes of the SAC/ARC did not redistribute to a large extent to either plasma membrane domain (Fig. 5). These results indicate that sterol can move between the plasma membrane domains of polarized cells independent of vesicular transport but that vesicular transport largely contributes to the delivery of sterols to intracellular compartments. There is a preferential incorporation of DHE into the apical membrane that is reflected in the relative rates of exchange out of versus into the apical membrane (compare Figs. 7 and 8 and Table I). This preferential incorporation of DHE into the apical membrane is apparently a consequence of the composition of the membrane. Our kinetic measurements for movement of DHE between the plasma membrane domains (Figs. 7 and 8 and Table I) reflect the steady state distribution of cholesterol that was found to be enriched 1.5-2-fold in the canalicular membrane compared with the basolateral membrane in hepatocytes (60, 61). Strikingly, we did not find release of DHE into the lumen of the BC (Fig. 9) suggesting that additional factors like continuous secretion of bile salts and transport of the bile fluid are required for release of cholesterol from the canalicular membrane into the bile in hepatocytes.

What might be the mechanism of nonvesicular sterol transport in polarized cells? We propose that sterols incorporated into the outer leaflet have to flip first to the cytoplasmic leaflet of the plasma membrane in order to circumvent the tight junction diffusion barrier that has been shown to block diffusion of phospholipid analogs (1, 2). DHE would have to traverse the plasma membrane very rapidly to explain the measured rapid transport kinetics. However, measured flip-flop rates and the steady state transversal distribution of sterols in biological membranes are contradictory (62-64). Leventis and Silvius (65) found an upper limit for the half-time of cholesterol flip-flop in model membranes of 1 min. The half-time of transbilayer migration of DHE in vesicle membranes is on the order of 1 min but depends on the fluidity of the bulk lipid phase (66). Thus, it is reasonable to assume that a low energy barrier exists for flipping sterols to the inner leaflet of the plasma membrane as suggested in our ATP depletion experiments. DHE could then diffuse laterally to the BC in the cytoplasmic leaflet of the plasma membrane. Measured diffusion coefficients for sterol are on the order of 10⁸ cm²/s which would be sufficient to explain the observed rapid movement of DHE between the plasma membrane domains (67). Alternatively, sterol could be transported through the cytoplasm via sterol-carrier proteins such as SCP2 or bound to many cytosolic proteins with a hydrophobic cavity (18, 68). Sterol transport through the cytosol would be similar to transport of fatty acids, which have been shown to be transported through the cytoplasm on various carrier proteins (69).

We found a small but consistent effect of ATP depletion on the kinetics and extent of recovery of DHE fluorescence in the canalicular and the basolateral plasma membrane. It is possible that an ATP-dependent step is involved in moving DHE from the cytoplasmic to the exoplasmic leaflet, especially in the canalicular membrane. Partition of sterol into the (glyco)sphingolipid-rich exoplasmic leaflet could be an ATP-driven process, and sterols once associated with this membrane could form hydrogen bonds and hydrophobic contacts preferentially with sphingolipids in the lumenal leaflet of the canalicular membrane (for a review see Refs. 12 and 70). Stable association of sterol with lipids in the canalicular membrane is supported by our observation that release of DHE from the BC was 3-4-fold slower than apical enrichment of this sterol and slightly accelerated by ATP depletion (Fig. 8). An ATP-dependent transporter such as the ATP-binding cassette (ABC) transporters which are highly expressed in the apical membrane of hepatic cells, including HepG2 cells, might affect the transbilayer distribution of sterol, perhaps by transporting another substrate from the cytoplasmic to the lumenal leaflet of the canalicular membrane (for a review see Ref. 71). This has been suggested for the multidrug resistance (MDR) protein 3 (MDR3), which translocates PC to the luminal leaflet of the BC (72), and for MDR protein 1 (MDR1), which translocates also analogs of glycosphingolipids (73). However, the high rate of spontaneous sterol flip-flop in model membranes supports that sterol rapidly equilibrates in response to phospholipid translocation in biological membranes (74).

It has been shown in Madin-Darby canine kidney cells that endosomes of the ARC containing Tf are enriched in "raft markers" like cholesterol and sphingolipids but also in phosphatidylserine (48). We hypothesize that raft components like (glyco)sphingolipids play a major role in the apical membrane and in recycling endosomes and that extensive vesicle traffic of other lipids as well as of certain proteins to and from the SAC/ARC might be important for creation of the sterol-accommodating properties of this compartment. Recent work by Edidin and colleagues (75, 76) suggests that vesicle traffic in concert with lateral diffusion barriers are required for maintenance of dynamic membrane domains or patches in the plasma membrane. We suggest that creation of a sterol-accepting environment in the apical plasma membrane of polarized cells is a dynamic and ATP-dependent process requiring vesicle traffic (e.g. transcytosis) as well as diffusion barriers (tight junctions).

Our results set the stage for understanding the mechanism of biliary secretion of cholesterol in hepatocytes. Rapid nonvesicular transport of HDL-derived cholesterol delivered to the basolateral membrane as here reported for DHE might be an efficient mechanism to deliver sterol to the canalicular membrane for its subsequent release into the bile. In an independent set of experiments, we have incubated HepG2 cells with HDL₃ labeled with Alexa488 on apolipoprotein A-I (apoA-I) and additionally with DHE. We found colocalization of Alexa488-apoAI with Tf but no transport to the BC as described previously (21). In contrast, delivery of DHE to the BC occurred already after 1 min of pulse labeling with the double-labeled HDL.² Further studies are warranted to clarify the mechanisms of biliary sterol transport in hepatocytes.

	ACKNOWLEDGEMENTS

We thank Timothy E. McGraw for critical reading of the manuscript and John Murray for kindly providing Texas Red-ASOR.

	FOOTNOTES

^* This work was supported in part by Deutsche Forschungsgemeinschaft Grant GK 268 (to A. H.), National Institutes of Health Grant DK27083, and the Ara Parseghian Medical Research Foundation grant (to F. R. M.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Supported by a postdoctoral fellowship for Biomedical Research from the Charles Revson Foundation.

To whom correspondence should be addressed: Dept. of Biochemistry, Rm. E-215, Weill Medical College of Cornell University, 1300 York Ave., New York, NY 10021. Fax: 212-746-8875; E-mail:frmaxfie@med.cornell.edu.

Published, JBC Papers in Press, June 5, 2002, DOI 10.1074.jbc.M202626200

² D. Wüstner and F. R. Maxfield, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: TGN, trans-Golgi network; Alexa- Tf, Alexa 488-labeled transferring; ARC, apical recycling compartment; BC, biliary canaliculus, biliary canaliculi; C₆-NBD-Cer N-[6-[(7-nitro-2-1, 3-benzooxadiazol-4-y)amino]caproyl]-sphingosine; C₆- NBD-SM, N-[6-[(7-nitro-2-1,3-benzooxadiazol-4-y)amino]caproyl]-sphingosyl]-phosphocholine; BODIPY-Cer, N-(4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-pentanoyl)-sphingosine; -BODIPY-PC, 2-(4,4-difluoro-5-(4-phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s-indacene-3-pentanoyl)-1-hexadecanoyl-sn-glycero-3-phosphocholine; BSA, bovine serum albumin; CCD, charge-coupled device; CHO cells, Chinese hamster ovary cells; DHE, dehydroergosterol; DHE/MCD, DHE loaded on methyl--cyclodextrin; FLIP, fluorescence loss in photobleaching; FRET, fluorescence resonance energy transfer; FRAP, fluorescence recovery after photobleaching; MCD, methyl--cyclodextrin; MDR, multidrug resistance; PC, phosphatidylcholine; PFA, paraformaldehyde; ROI, region of interest; SAC, subapical compartment; Tf, transferrin.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. Mondal, B. Mesmin, S. Mukherjee, and F. R. Maxfield Sterols Are Mainly in the Cytoplasmic Leaflet of the Plasma Membrane and the Endocytic Recycling Compartment in CHO Cells Mol. Biol. Cell, January 1, 2009; 20(2): 581 - 588. [Abstract] [Full Text] [PDF]

				N. H. Petersen, N. J. Faergeman, L. Yu, and D. Wustner Kinetic imaging of NPC1L1 and sterol trafficking between plasma membrane and recycling endosomes in hepatoma cells J. Lipid Res., September 1, 2008; 49(9): 2023 - 2037. [Abstract] [Full Text] [PDF]

				B. P. Atshaves, A. L. McIntosh, H. R. Payne, A. M. Gallegos, K. Landrock, N. Maeda, A. B. Kier, and F. Schroeder SCP-2/SCP-x gene ablation alters lipid raft domains in primary cultured mouse hepatocytes J. Lipid Res., October 1, 2007; 48(10): 2193 - 2211. [Abstract] [Full Text] [PDF]

				J. Leyt, N. Melamed-Book, J.-P. Vaerman, S. Cohen, A. M. Weiss, and B. Aroeti Cholesterol-sensitive Modulation of Transcytosis Mol. Biol. Cell, June 1, 2007; 18(6): 2057 - 2071. [Abstract] [Full Text] [PDF]

				D. Wustner Plasma Membrane Sterol Distribution Resembles the Surface Topography of Living Cells Mol. Biol. Cell, January 1, 2007; 18(1): 211 - 228. [Abstract] [Full Text] [PDF]

				K. Aravindhan, C. L. Webb, M. Jaye, A. Ghosh, R. N. Willette, N. J. DiNardo, and B. M. Jucker Assessing the effects of LXR agonists on cellular cholesterol handling: a stable isotope tracer study J. Lipid Res., June 1, 2006; 47(6): 1250 - 1260. [Abstract] [Full Text] [PDF]

				L. Yu, S. Bharadwaj, J. M. Brown, Y. Ma, W. Du, M. A. Davis, P. Michaely, P. Liu, M. C. Willingham, and L. L. Rudel Cholesterol-regulated Translocation of NPC1L1 to the Cell Surface Facilitates Free Cholesterol Uptake J. Biol. Chem., March 10, 2006; 281(10): 6616 - 6624. [Abstract] [Full Text] [PDF]

				M. K. Wu, H. Hyogo, S. K. Yadav, P. M. Novikoff, and D. E. Cohen Impaired response of biliary lipid secretion to a lithogenic diet in phosphatidylcholine transfer protein-deficient mice J. Lipid Res., March 1, 2005; 46(3): 422 - 431. [Abstract] [Full Text] [PDF]

				D. Wustner Mathematical Analysis of Hepatic High Density Lipoprotein Transport Based on Quantitative Imaging Data J. Biol. Chem., February 25, 2005; 280(8): 6766 - 6779. [Abstract] [Full Text] [PDF]

				I. Navarro-Lerida, M. M. Corvi, A. A. Barrientos, F. Gavilanes, L. G. Berthiaume, and I. Rodriguez-Crespo Palmitoylation of Inducible Nitric-oxide Synthase at Cys-3 Is Required for Proper Intracellular Traffic and Nitric Oxide Synthesis J. Biol. Chem., December 31, 2004; 279(53): 55682 - 55689. [Abstract] [Full Text] [PDF]

				S. C.D. van IJzendoorn, J. M. van der Wouden, G. Liebisch, G. Schmitz, and D. Hoekstra Polarized Membrane Traffic and Cell Polarity Development Is Dependent on Dihydroceramide Synthase-Regulated Sphinganine Turnover Mol. Biol. Cell, September 1, 2004; 15(9): 4115 - 4124. [Abstract] [Full Text] [PDF]

				R. E. Soccio and J. L. Breslow Intracellular Cholesterol Transport Arterioscler. Thromb. Vasc. Biol., July 1, 2004; 24(7): 1150 - 1160. [Abstract] [Full Text] [PDF]

				D. Hoekstra, D. Tyteca, and S. C. D. van IJzendoorn The subapical compartment: a traffic center in membrane polarity development J. Cell Sci., May 1, 2004; 117(11): 2183 - 2192. [Abstract] [Full Text] [PDF]

				M. Hao, S. Mukherjee, Y. Sun, and F. R. Maxfield Effects of Cholesterol Depletion and Increased Lipid Unsaturation on the Properties of Endocytic Membranes J. Biol. Chem., April 2, 2004; 279(14): 14171 - 14178. [Abstract] [Full Text] [PDF]

				D. Wustner, M. Mondal, A. Huang, and F. R. Maxfield Different transport routes for high density lipoprotein and its associated free sterol in polarized hepatic cells J. Lipid Res., March 1, 2004; 45(3): 427 - 437. [Abstract] [Full Text] [PDF]

				H. A. Scheidt, P. Muller, A. Herrmann, and D. Huster The Potential of Fluorescent and Spin-labeled Steroid Analogs to Mimic Natural Cholesterol J. Biol. Chem., November 14, 2003; 278(46): 45563 - 45569. [Abstract] [Full Text] [PDF]

				P. L. TUMA and A. L. HUBBARD Transcytosis: Crossing Cellular Barriers Physiol Rev, July 1, 2003; 83(3): 871 - 932. [Abstract] [Full Text] [PDF]

				D. Hoekstra, O. Maier, J. M. van der Wouden, T. A. Slimane, and S. C. D. van IJzendoorn Membrane dynamics and cell polarity: the role of sphingolipids J. Lipid Res., May 1, 2003; 44(5): 869 - 877. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/33/30325    most recent M202626200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wüstner, D. Articles by Maxfield, F. R. Search for Related Content PubMed PubMed Citation Articles by Wüstner, D. Articles by Maxfield, F. R. Social Bookmarking                      What's this?