Hyperosmolarity Inhibits Galactosyl Receptor-mediated but Not Fluid ...

THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 264,No. 20,Issue of July 15,pp. 1201612024,1989 Printed in U.S.A.

Hyperosmolarity Inhibits Galactosyl Receptor-mediated butNot Fluid Phase Endocytosis in Isolated Rat Hepatocytes* (Received for publication, January 3,1989)

Janet A. Oka, Monique D. Christensen, and PaulH. WeigelS From the Departmentof Human Biological Chemistry and Genetics,The University of Texas Medical Branch, Galveston, Texas 77550

We have investigated the effectsof hyperosmolarity sucrose, sodium chloride, or other osmolites blocks the disinduced by sucrose on the fluid phase endocytosis of sociation of internalized receptor-ligand complexes and comthe fluorescent dye lucifer yellow CH (LY) and the pletely stops receptor recycling (Oka and Weigel, 1988). In endocytosis of '261-asialo-orosomucoid (ASOR) by the the presence of sucrose, the internalization of surface-bound galactosyl receptor systemin isolated rat hepatocytes. ligand was also incomplete. Not all the receptor-ligand comContinuous uptake of LY by cells at 37 "C is biphasic, plexes on the cell surface could be internalized by cells under occurs for 3-4 h, and then plateaus. Permeabilized hyperosmolar conditions. Heuser and Anderson (1987) also cells or crude membranes do not bindLY at 4 or 37 "C. found that exposure of fibroblasts to high concentrations of LY at 4 "C. The rate Intact cells also do not accumulate and extentof LY accumulation are concentration- and sucrose effectively stopped the continuous internalization of energy-dependent,andinternalized LY is released low density lipoprotein. These investigators concluded that from permeabilized cells. Efflux of internalized LY the coated pit pathway involving the recycling of clathrin and from washed cells is also biphasicand occurs with half- the formation of coated pits was inhibited underhyperosmolar times of approximately 38 and 82 min. LY is taken up conditions. In earlier studies, Daukas and Zigmond (1985) into vesicles throughout the cytoplasm and the peri- concluded that hyperosmolarity decreased the endocytosis of nuclear region with a distribution pattern typical of chemotactic peptide in polymorphonuclear leukocytes. Even the endocytic pathway. LY, therefore, behaves as a though peptideuptake was decreased, however, the cells were fluid phasemarker inhepatocytes. LY has no effect on still able to undergo fluid phase endocytosis as judged by their ability to continue to internalize ['4C]sucrose in the presence the uptake of"'I-ASOR at 37 "C. The rate ofLY of high concentrations of sodium chloride. This observation at uptake bycells in suspension is not affected for least 30 min by up to0.2 M sucrose. The rateof endocytosis has not been followed up by other investigators. of '261-ASOR, however, is progressively inhibited by The prevailing understanding in the literatureis that fluid increasing the osmolality of the medium with sucrose phase endocytosis is probably accounted for by the receptor(>98% with 0.2 M sucrose; Oka and Weigel (1988)J. mediated clathrin-coated pit pathway (Marsh and Helenius, Cell. Biochem. 36, 169-183). Hyperosmolarity com- 1980; Steinman et al., 1983). That is, all the fluid uptake in a pletely inhibits endocytosis of 12sI-ASORby the galac- cell could be due to thecoincident volume internalized via the tosyl receptor, whereas fluid phase endocytosisof LY coated pit pathway. According to thisnotion, cells essentially is unaffected. Cultured hepatocytes contained about possess only one mechanism by which vesicles and fluid are 100 coated pits/mm of apical membrane length as as- internalized. This viewof fluid phase endocytosis is based sessed by transmissionelectron microscopy. Inthe presence of 0.4 M sucrose, only 17 coated pitslmm of primarily on studies performed with fibroblasts. Recent results from a number of investigators examining membrane were observed, an 83% decrease. Only a different receptors including those for asialoglycoproteins few percent of the total cellularfluid phase uptake in hepatocytes is due to the coated pit endocytic pathway. (Oka and Weigel, 1983; Weigelet al., 1986), insulin (Smith et al., 1987; McClain and Olefsky, 1988), transferrin (Stein and We conclude that the fluidphase and receptor-mediated endocytic processes mustoperate via twosepa- Sussman, 1986), chemotactic peptide (Zigmond and Tranquillo, 1986), and low density lipoprotein (Edge et al., 1986) rate pathways. indicate that receptor-bound ligands are endocytosed and processed by multiple cellular pathways not just a single pathway (Weigel, 1987). The question of whether these pathMany inhibitors have been used as perturbants to dissect ways all involve coated pits needs to be reexamined. For the multistep pathwaysof receptor-mediated endocytosis and example, insulin uptake in hepatocytes has been shown to receptor recycling (Besterman and Low, 1983;Schwartz, 1984; occur by two separate pathways, one of which involves a Goldstein et al., 1985; Weigel, 1987). Agents such as monensin, coated pit pathway and the other which does not involve lysosomotropic amines, cytoskeletal drugs, and metabolic en- coated pits (McClain and Olefsky, 1988; Smith et al., 1987). ergy poisons have allowed investigators to define and affect Similarly, we have documented over the lastseveral years the discrete processing steps in this pathway. We recently found existence of two separate ligand-processing and receptorthat hyperosmolarity induced by increasing concentrations of recycling pathways in the galactosyl receptor system in isolated rat hepatocytes (Weigel, 1987). In thiscase, it is not yet * This research was supported by National Institutes of Health known whether these are both coated pit pathways. The major Grants GM 30218 and GM 35978. The costs of publication of this galactosyl receptor pathway, which has been studied by other article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accord- investigators as well, is a coated pit pathway (Wall et al., 1980). The purpose of the present study was to determine ance with 18 U.S.C. Section 1734 solely to indicate this fact. whether the receptor-mediated coated pit pathway in hepa$ To whom correspondence should be sent.

12016

12017

Hyperosmolarity Inhibits Receptor-mediated Not Fluid Phase Endocytosis tocytes solely accounts for the fluid phase uptake capacity of these cells or if there is more than one pathway by which fluid volume can be taken up. The results indicate that hepatocytes take up fluid in a fluid phase pathwayb) ata far greater rate than fluid is taken up by any coated pit pathway. A preliminary report of these results has been presented (Oka and Weigel, 1987). EXPERIMENTALPROCEDURES

Materials-Human orosomucoid from Sigma was desialylated and iodinated as described before (Weigel and Oka, 1982). Na'"1 (10-20 mCi/rg of iodine) was from Amersham Corp. Digitonin from Sigma or Eastman Kodak was dissolved to 25% (w/v) in dimethyl sulfoxide or to 1.4% in 100%ethanol with warming. Lucifer yellow CH (lithium salt) from Sigma was dissolved in medium l/BSA,' usually at a stock concentration of 1 mg/ml, and filtered (0.2-pm pore size). BSA was from Armour (CRG-7) or Sigma (fraction V). Collagenase was from Sigma (type IA), Serva FineBiochemicals Inc. (17449),or Boehringer Mannheim (type D). Medium l/BSA contains a modified Eagle's medium (GIBCO) supplemented with 2.4 g/liter Hepes, pH 7.4,0.22 g/liter NaHC03, and 0.1% BSA. Cell Preparation-Hepatocytes were prepared from male SpragueDawley rats (-250 g) by a modification (Clarke et aL, 1987) of the collagenase liver perfusion procedure of Seglen (1976). Final cell pellets were suspended in medium 1/BSA. The cells were routinely 85-95% single cells and viable as judged by trypan blue exclusion. Viability by trypan blue was not compromised by 0.25 M sucrose for cells in suspension or 0.4 M for cells in culture. Experiments were performed in the absence of serum; 0.1% BSA was present for most experiments. Cells were cultured inWilliam's E medium as described previously (Oka and Weigel, 1987). For experiments with cells in suspension, the cells were first incubated a t 37 "C for 60 min to increase and stabilize the number of surface receptors (Weigel and Oka, 1983). Cell suspensions in different experiments were incubated in 50-250-m1 Erlenmeyer flasks at 2-3 X lo6 cells/ml. The suspensions occupied 10% of the flask volume and were shaken at 100 rpm in a gyratory water bath. Osmolarity-Sucrose was dissolved to a concentration of 1.6 M in double distilled water. Medium 1was concentrated 2-fold in a Buchi rotary evaporator. One volume of the concentrated medium and 1 volume of 1.6 M sucrose and distilled water were combined to give the desired final concentration of sucrose. In this fashion, the osmolarity of the medium could be changed without also altering the concentration of nutrients in the medium. Osmolality was measured using a Wescor Inc. vapor pressure osmometer (model 5100C). The osmolality of medium l/BSA was 270 mmol/kg. 0.2 M sucrose increased the osmolality by 290 mmol/kg. The effects of hyperosmolarity on various steps in the galactosyl receptor system have already been described (Oka and Weigel, 1988). Fluid Phase and Receptor-mediated Endocytosis-The fluorescent dye LY was used as a fluid phase marker. The final concentration of LY in the media was usually 0.2-0.4 mg/ml. Cell samples were diluted 4-8-fold into ice-cold medium l/BSA and centrifuged immediately. The cell pellet was washed again with medium 1/BSA, and the cells were solubilized in 0.5 ml of 0.05% Triton X-100 with 1mg/ml BSA. Simultaneous endocytosis of '261-ASOR was assessed by measuring cell-associated radioactivity in a Packard Multiprias 2-y-spectrometer. The same samples were then transferred to 10 X 75-mm glass tubes for analysis of LY fluorescence. Fluorescence was measured in an Aminco spectrofluorometer (model SPF-125) set at 430 nm excitation and 540 nm emission. Up to 1 mg/ml BSA had no effect on the fluorescence of LY. Electron Microscopy-Cultured hepatocytes were chilled on ice, rinsed once in cold medium 1, twice in cold 0.1 M sodium cacodylate, pH 7.4, and fixed with 3% glutaraldehyde in cacodylate buffer. The cells were incubated for 30 min at room temperature with frequent agitation and then washed three times in cacodylate buffer. They were then incubated in 1%(w/v) Os04 in cacodylate buffer for 1 h at room temperature, washed once, and dehydrated with ethanol. Using the floating sheet method (Arnold and Boor, 1986), the cells were detached from the Falcon dishes by adding 1-2 ml of N-butyl glycidyl ether after the ethanol was removed. The floating sheet was folded,

' The abbreviations used are: BSA, bovine serum albumin; LY, lucifer yellow;ASOR, asialo-orosomucoid; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonicacid.

transferred to a clean dish, infiltrated and embedded with Spurr resin, and polymerized at 55 "C. Grids were poststained with uranyl acetate and lead citrate. For cells in suspension, glutaraldehyde was added to a final concentration of 1%at 37 "C. After 10 min, samples were further fixed in 2% glutaraldehyde in cacodylate buffer for 1 h at room temperature. The cells were washed by centrifugation, incubated in1%Os04in cacodylate buffer for 30 min, washed, and stained en bloc in 4% uranyl acetate in 70% ethanol for 30 min. The cell pellets were dehydrated with ethanol and infiltrated and embedded in Epon. Samples were sectioned on a PharmaciaLKB Biotechnology Inc. model IV ultramicrotome and examined on a Jeol 100 CX at 60 kV. Membrane perimeter length was determined using a Bio-Quant IV image analysis system. Coated pit profiles were scored independently on coded samples by three individuals. The averaged scores are presented. General-Centrifugations of intact or digitonin-permeabilized cell suspensions were a t 800 or 1000 rpm, respectively, for 2 min in a Sorvall GLC-1 model tabletop centrifuge. Protein was determined by the method of Bradford (1976) using BSA as a standard. RESULTS

Validation of LY as a Fluid Phose Marker inHepatocytesSince LY has not been used previously with hepatocytes, it was necessary to demonstrate that this molecule behaves as a bonu fide fluid phase molecule with these cells. The kinetics of uptake ofLY at 37 "C by cells in suspension is biphasic (Fig. 1). This was more obvious at higher LY concentrations. There is an initial rapid rate of uptake for approximately 20 min and then aslower rate of uptake for 3-4 h until a steadystate intracellular concentration of LY is achieved. The rate of internalization increased linearly with the concentration of LY ( r = 0.985), indicating that theuptake is not saturable at least up to aconcentration of 1.09 mM LY (0.5 mg/ml). Likewise, the extent of LY uptake was linear with increasing concentration ( r = 0.990). LY uptake is, therefore, concentration-independent and time-dependent. LY accumulation only occurred with intact cells at 37 "C (Fig. 2). Cells incubated at 4 "C accumulated virtually no LY, nor did cells that had first been permeabilized with digitonin andthen incubated at either 4 or 37 "C. LY does not bindsignificantly to membranes either in live cells at 4 "C or in permeable cells at 4 or 37 "C.

:I

E 25

15 ",

P

x

0

10

W

5

0

I

10

20

30

40

I

50

I

60

TIME (minutes)

FIG. 1. Continuous uptake of LY. Isolated hepatocytes were incubated at 37 "C in the absence (A)or presence of increasing concentrations (in mg/ml) of LY: 0.025 (A), 0.05 (B),0.1 (0),0.2 (O), 0.4 (0).After washing the cells, cell-associated LY was determined as described under "Experimental Procedures."

HYperosrnolarity Inhibits Receptor-mediated Not Fluid Phase Endocytosis

12018

Other investigators have also concluded that LY is imper- sponded to about 7 x 10"' liters/cell or about 60% of the meant to membranes and does not bind to cells or membranes estimated cellular volume. (Bowman and Tedeschi, 1983; Steinberg et al., 1987; Mir et Reversibility of LY Accumulation-As a further indication al., 1988). A live cell is required in order to obtain the time- that theinternalization of LY represents afluid phase process dependent accumulation at 37 "C. If hepatocytes are incu- and not a partition of LY into membrane or its binding to bated inthe presence of LY at 37 "C, chilled to 4 "C, and then cellular components, the ability of the internalized LY to be washed and permeabilized with the nonionic detergent digi- released from live hepatocytes was examined. Cells were intonin, the cells virtually immediately lose all of the accumu- cubated with LY at 37 "C, washed, and resuspended a t 37 "C lated LY into the medium (Fig. 3). This result indicates that in medium without the dye (Fig. 4A). The rate of release of the internalized LY is not bound to cellular components and LY from the cells was also biphasic and went to completion is free to leave the cell after the vesicular organelles have in approximately 4 h. The efflux kinetics, therefore, correspond closely to the influx kinetics. In a similar series of been made permeable. The initial rate of fluid uptake was calculated from data experiments, cells were allowed to internalize LY for various times at 37 "C, washed, and the rate of efflux at 37 "C was obtained from experiments such as in Fig. 1 to be 5.7 f 2.2 then determined (Fig. 4B).As expected, the biphasic character ( n = 8) pl/cell/h. The estimatedintracellular volume for of the efflux curve was most pronounced when the LY influx isolated rat hepatocytes in suspension (average diameter -28 time had been brief. At long influx times (e.g. 90 min, Fig. pm) is approximately 12 x 10"' liters/cell. This estimated 4B), the efflux curves were less noticeably biphasic. The rate of fluid volume uptake therefore corresponds to approx- rapidly filling/emptying compartment has a relatively small imately 50% of the cell volume per h and is consistent with volume compared with the more slowly exchanging compartvalues obtained in other cell types. As indicated above (Figs. ment. The approximate t H values for release of the inter2 and 3), the rate of internalization decreases with time. The final steady-state extent of internalized fluid volume corre-

0'

I

I

I

I

I

I

I

I

20

40

60

80

100

120

I

I

L

0

I

I

30

60

TIME (minutes)

FIG. 2. LY uptake by intact or permeable cells. Cellswere treated with (m,0) or without (0,0)0.055% digitonin and then incubated at 37 'C (0,m) or 4 "C (0,0 ) with 0.2 mg/ml LY. Cell samples were washed, and cell-associated LY was determined as described under "Experimental Procedures."

1

I

I

120180 150

FIG.3. Digitonin releases internalized LY. Cells were allowed to endocytose 0.1 mg/ml (0, 0 ) or 0.4 mg/ml (0,m) LY at 37 "C. Samples were taken at the indicated times, chilled to 4 "C, washed, and treated with (a, 0 ) or without (0,0)0.055% digitonin for 10 min. Cell-associated LY was then determined after washing the cells.

I " " " " FIG. 4. Efflux of internalized LY. A, cells were allowed to internalize 0.4 mg/ml LY at 37 "C for 2 h. The cells were then chilled, washed twice by centrifugation, resuspended in fresh medium 1/BSA without LY, and put back at 37 "C. Fluorescence in the medium (A),cell-associated fluorescence (m), and their sum (0)were measured at the indicated times. At t = 0, the totalfluorescence measured was equivalent to 1150 ng of LY/106 cells. B , cells were first incubated with LY as in A for 15 (O), 30 (O), 60 (a),or 90 (m) min andthen washed and resuspended without LY at 37 "C. Cell samples were taken, washed, and cell-associated LY determined at the indicated times.

I

90

TIME (minutes)

n

A

51

90 60

30

2

15

0.5-

c30L 30

0

20

40

60

80 100 120 140 160

TIME (minutes)

0.

60

90

120

TIME (minutes)

150

Hyperosmolarity Inhibits Receptor-mediated

Not Fluid Phase Endocytosis

12019

ized LY were determined from semilog plots of the kinetic data (Table I). The average tlr,values for the fast and slow efflux components were, respectively, 37.9 f 2.9 ( n = 3) and 82.3 & 6.6 ( n = 4) min. As the time of internalization increased, the time required for efflux of a constant percentage (e.g. 50%) of the intracellular LY also increased. These characteristics are compatible with the conclusion that LY is taken up by hepatocytes by a fluid phase process. Intracellular Accumulation of LY-Although hepatocytes have a high level of endogenous fluorescence,the LY uptake TABLE I Effect of loading time on the efflux of LY from hepatocytes The data from the early and later portions of the efflux curves shown in Fig. 4B were analyzed by least squares linear regression to estimate the first order rate constants and half-times. Correlation coefficients were all 20.98. Duration of LY uptake min min

89.2

Fast component t,,, min

15 30 60 90

34.5

Mean (S.E.)

37.9 f 2.9

3 82.9

0

I

Slow component tu,

' 0

I

I

1

1

I

I

10

20

30

40

50

60

1

TIME (minutes)

39.8 83.7 82.3 f 6.6

FIG. 5 . Internalization of LY by cultured hepatocytes. Hepatocytes were cultured overnight in William's E medium in 35-mm tissue culture dishes. After two washes, they were put at 37 "C for 30 min in buffer 1/BSA (142 mM NaC1, 6.7 mM KCl, 10 mM HEPES, pH 7.4, and 0.1% BSA) plus ( B ) or minus ( A ) 0.2 mg/ml LY. The cells were then washed five times with buffer 1/BSA and were examined without fixation for fluorescence using a Leitz Orthoplan microscope equipped with a Ploemopak 2 epi-illuminator and an IP fluorescein filter cube. Photographs were taken with Kodak Tri-X Pan black and white 35-mm film. The exposure times and developing and printing conditions were the same. The bur represents 10 pm.

FIG. 6. Effect of sucroseon LY and '261-ASOR endocytosis. Cells were allowed to endocytose continuously 2 pg/ml 'Z51-ASOR( A ) and 0.4 mg/ml LY ( B )simultaneously in the absence (0)or presence of 0.1 (O),0.13 (O), or 0.2 (W) M sucrose in medium 1/BSA.

was clearly revealed to be intracellular using fluorescence microscopy. Intense punctate staining was observed in vesicles and possibly tubules throughout the cytoplasm after 30 min in the presence ofLY (Fig. 5B). In the absence of LY, the level of fluorescence was much less (Fig.5 A ) . The distribution of labeled vesicles was verysimilar to what has been observed with other endocytic markers such as fluoresceinlabeled dextrans. Prominent staining was often seen close to the nucleus, presumably in the Golgi-lysosomal region. This result confirms the above biochemicalstudies. Effect of Hyperosmolarity on '251-ASORand LY Internalizution-In order to examine the possible differential effect of hyperosmolarity on the uptake of the receptor-mediated ligand uersus the fluid phase ligand LY, hepatocytes in suspension were incubated simultaneously with lZ5I-ASORand LY in the presence of increasing amounts of sucrose (Fig. 6). The continuous uptake of '251-ASOR wasunaffected by the simultaneous presence and uptake of LY (not shown). The continuous uptake of lZ5I-ASORwas progressively inhibited with increasing concentrations of sucrose (Fig. 6 A ) as reported previously (Oka and Weigel, 1988). 0.2 M sucrose gave virtually complete inhibition of endocytosis (99%).At the same time, the internalization of LY was unaffected for at least 20 min by all the concentrations of sucrose tested (Fig. 6B). Only after approximately 30 min was there any change in the rate ofLY accumulation at the highest concentration of sucrose. The effect of increasing osmolarity on the inhibition of '251-ASOR uptake was extremely rapid (Fig. 7A). Hepatocytes continuously internalizing asialoglycoprotein at 37 "C were essentially completely shut down within 1-2 min after the addition of sucrose to a final concentration of 0.2 M. The uptake of LY in these same cells wasunperturbed (Fig. 7B). Effect of Metabolic Energy Poisons on LY Uptake-Hyperosmolarity induced with sucrose at 37 "C did not significantly alter cellular ATP levels. For example, after 30 and 60 min in the presence of 0.2 M sucrose, the ATP content of cells was, respectively, 101 and 89% of the initial control value. We examined previously the sensitivity of 1251-ASORuptake and receptor recycling to ATP depletion in isolated hepatocytes (Clarke and Weigel, 1985) and observed that a single round

Hyperosmolarity Inhibits Receptor-mediated Not Fluid Phase Endocytosis

12020

P

0

10

I

I

20

30

I

40

50

60

I

70

I

SO

TIME(minutes)

I

0

I

I

I

I

8

1

15

30

45

60

75

90

TIME (minutes)

FIG. 8. Effect of sodium azide or nitrogen atmosphere on the endocytosis of lZ6I-ASOR and LY. Cells were incubated at 37'C with 1.5 pg/ml '261-ASOR ( A ) and 0.25 mg/ml LY ( B ) in the absence (0)or presence of 10 mM sodium azide (O),20 mM sodium azide (U), or Nz atmosphere (0).The nitrogen was blown into the

10

20

I

I

I

I

I

I

30

40

50

60

70

SO

TIME (minutes)

flask above the surface of the media. At the indicated times, samples (2.2 x lo6 cells) were layered onto discontinuous Percoll gradients (2 ml of 30%, 2 ml of 40%, and 0.250 ml of 50% Percoll in a 13 X 100mm glass tube) and centrifuged for 8 min at 1500 rpm a t 4 "C (Clarke and Weigel, 1985). Viable (>98%) cells were collected from the bottom of the gradients, lysed in 0.05% Triton X-100,and cell protein, associated LY, and radioactivity were then determined.

FIG. 7. Effect of sucrose on the kinetics of laaI-ASOR and LY uptake. Cells in suspension were allowed to endocytose simulEffect of Intracellular Potassium Ion Depletion on LY Intertaneously 1.5 pg/ml '251-ASOR ( A ) and 0.2 mg/ml LY ( B )at 37 "C. nalization-Larkin et al. (1983, 1985) found that fibroblasts At various times, additions were made of medium 1/BSA alone (0,

depleted of intracellular K+ by hypoosmotic shock in potassium-free buffer were subsequently unable to internalize low density lipoproteins. These investigators concluded that receptor-mediated endocytosis was inhibited under these conditions. The conclusion was that clathrin recycling in the coated pit cycle was interrupted when the cells were depleted of K+. IfLY uptake in hepatocytes is independent of the of uptake of receptor-bound ligand was unaffected even when coated pit pathway as suggested by the above hyperosmolarity ATP poolswere depleted greater than 98%. Recycling of experiments, then LY uptake might continue in cells that galactosyl receptors, however, was inhibited if the cellular have been depleted of potassium. Hepatocytes were first ATP levels were decreased below a threshold of about 60%. washed with a potassium-free medium and then incubated in Recycling was completely inhibited if ATP levels fell to 40% this medium at 37 "C in the presence of increasing concentraor less of controls (Clarke and Weigel, 1985). LY uptake by tions of the K+ ionophore nigericin (Fig. 9). As the nigericin M, the hepatocytes was examined in the presence of a nitrogen concentration was increased from 0.5 to 7.5 X atmosphere or different concentrations of sodium azide (Fig. continuous uptake of lZ6I-ASORwas inhibited by approxi8).As anticipated for a cellular process that involves vesicular mately 80% (Fig. 9A). In the same cells, the simultaneous trafficking and also requires membrane recycling, LY accu- uptake ofLY continued and was relatively insensitive to mulation was inhibited (Fig. 8B).Interestingly, the ability of increasing concentrations of the ionophore. The inhibition of cells to internalize LY was less sensitive to a given ATP LY uptake was about 20% at the highest concentration of depletion thanthe receptor-mediated endocytosis of lZ5I- nigericin (Fig. 9B). Since there may also-be multiple effects receptorASOR (Fig. SA). For example, in the presence of 10 mM of the ionophore on cellular processes other than the sodium azide, which typically depletes hepatocyte ATP levels mediated coated pit cycle, this result is qualitatively identical by approximately 85%, therewas still asignificant rate of LY to thatobserved with hyperosmolarity. We conclude that the accumulation which continued for over 1h. In contrast, during LY uptake in hepatocytes cannot be mediated by a coated pit the same period of time, there was virtually no uptake of lZ5I- pathway. Electron Microscopy of Hepatocytes Exposed to HyperosASOR. Weconclude from this result that LY uptake isenergydependent as expected and that LY and ASOR are taken up molar Sucrose-The effects of sucrose on the morphology of by pathways that have slightly different sensitivities to ATP hepatocytes in suspension or in culture were examined by depletion. Steinman et al. (1974) also observed the relative transmission electron microscopy. Based on the above results, resistance of horseradish peroxidase uptake to ATP depletion one might expect to see fewer or no coated pits in sucrosetreated cells. Although qualitatively this appeared to be the in fibroblasts. A, V, 0)or medium 1/BSA plus sucrose to give a final concentration of 0.2 M sucrose (m, A,V, +). The additions (arrows)were made a t 5 (0,m), 10 (A,A), 15 (V, V), and 20 (0,+) min. One control had no additions (0).Cell samples were washed, and associated LY and radioactivity were determined as described under "Experimental Procedures."

Hyperosmolarity Inhibits Receptor-mediated

0

Not Fluid Phase Endocytosis

12021

e 10

20

30

40

50

60

TIME (minutes)

FIG.9. Effect of nigericin and K+ depletion on the endocytosis of "'I-ASOR or LY. Cells were incubated a t 37 "C with 1.5 pg/ml lZ5I-ASOR( A ) and 0.2 mg/ml LY ( B ) in the absence (0)or presence of 0.5 (O),1.0 (O), 2.0 (W), 4.0 (A), or 7.5 (A)pM nigericin. Medium 1 for this experiment was formulated as in the GIBCO catalog but without K'.

case, it could not be reliably quantitated for cells in suspension. In the presence of 0.2 M sucrose, the surface of these cells showedextensive microvilli and membrane convolutions (Fig. 10B)compared with the untreated cells (Fig. 1OA). The pericellular region in the sucrose-treated cells also contained numerous vesicles and larger vacuoles. Coated pits were evident on the control cells but were not apparent on the treated cells, whichoften contained broad diffusely coated regions. In fact, the cell surface morphology made quantitation essentially impossible. The morphology of the cultured hepatocytes exposed to sucrose was verydifferent. Receptor-mediated endocytosis in cultured cells was more resistant to increasing sucrose concentration, requiring about 0.4 M sucrose to achieve maximal inhibition. The cells attached to the substratum were, therefore, exposed to a 50% greater osmolality compared with hepatocytes in suspension in the presence of 0.2 M sucrose (-850 versus -560 mmol/kg). Despite this osmotic stress, the cell surface morphology of the cultured hepatocytes (Fig. 1OD) was verysimilar to theuntreated cells (Fig.1OC).The sucrosetreated cultured cells didnot have extensive microvilli, surface membrane convolutions, or dramatically increased intracellular vesicles and vacuoles. With cells in suspension or in culture, an obvious difference wasseen in the appearance and organization of mitochondria in the presence of sucrose. It was possible to quantitate the frequency of normal coated pits in cultured hepatocytes even in the presence of sucrose (Table 11). Control cells after culture overnight had 99.6 f 18.8 coated pits/mm of apical membrane length. Coated pits were found over the whole apical cell surface (Fig. ll), often clustered near regions of cell-cell contact (Fig. lla). Coated pits at the base of a microvillus were very common (Fig. 11,c and cl). The basolateral surface was essentially devoid of coated pits. In the presence of 0.4 M sucrose, the number of coated pits was 16.9 f 6.2, an 83% decrease (Fig. 11, e-i). Hyperosmolarity,therefore, interferes with the normal coated pit cycle in hepatocytes as it does in fibroblasts (Heuser and Anderson, 1987).

FIG. 10. Effect of hyperosmolarity on hepatocytes in culture and in suspension.Hepatocytes in suspension (2 X lo6 cells/ml) or after overnight in culture (1 X 106/35-mm dish) were incubated a t 37 "C for 25 min with medium l/BSA alone or containing, respectively, 0.2 or 0.4 M sucrose. The cells were then fixed, washed, and processed for electron microscopy as described under "Experimental Procedures." A , suspension cells minus sucrose; B , suspension cells plus sucrose; C, cultured cells minus sucrose; D, cultured cells plus sucrose. The bar is 1 pm. TABLE I1 Effect of hyperosmolarity on coated pit frequency in cultured rat hepatocytes Hepatocytes were cultured overnight, washed, and incubated inthe absence of ASOR or LY in medium 1/BSA with or without 0.4 M sucrose a t 37 "C for 25 min. The cells were then fixed, prepared for electron microscopy, and sections were scored for the frequency of coated pits as described under "Experimental Procedures." Treatment

None 0.4 M Sucrose

Length of apical Coated surface analvzed

pits

pm

no./mm

1001

99.6 f 18.8 (100%)

809

16.9 2 6.2 (17%)

DISCUSSION

LY has been used as a small fluorescent molecule to demonstrate electrical coupling between cells (Stewart, 1978), to stain neurons (Zimmerman, 1986), and to labelselectively certain cell types in developing embryos (Sarthy andHilbush, 1983). More recently, LY has beenused to measure fluid phase endocytosis in macrophages (Swanson et al., 1985, 1987),yeast (Riezman, 1985),proximal tubular cells (Goligorsky and Hruska, 1986), and CV-1 kidney cells (Doxseyet aL, 1987). LY seems well suited for this use since it is highly fluorescent, is impermeable to membranes (Bowman and Te-

12022

Hyperosmolarity Inhibits Receptor-mediated Not Fluid

Phase Endocytosis

FIG. 11. Effect of hyperosmolarity on coated pits in cultured hepatocytes. Hepatocytes cultured overnight were incubated at 37 "C in medium 1/ BSA in the absence (a-d) or presence (e-i) of 0.4 M sucrose for 25 min. The cells were fixed and processed for electron microscopy as described under "Experimental Procedures."The bar is 1 pm. Examples of coated pits are indicated by arrowheads.

deschi, 1983), and does not bind to membranes (Mir et al., 1988; Steinberg et al., 1987). Other fluid phase markers including sucrose, fluorescein-labeled dextran, horseradish peroxidase, and polyvinylpyrrolidone have been used in different cell systems. Different cell types may bind a possible fluid phase marker, which then precludes its use. For example, although horseradish peroxidase has been widely used and in many cases is a suitable fluid phase marker, Esfahani et al. (1986) showedthat it behaved differently in its uptake characteristics compared with LY in human monocytes or the macrophage-like cell line U937. Presumably, problems arise if cells have a mannose or N-acetylglucosamine receptor that can interactwith the horseradish peroxidase. Klein and Satre (1986) determined that amebae were able to accumulate fluorescein-labeled dextran to a concentration greater than that in the medium. Therefore, a given fluid phase probe must first be shown to be suitable in a given celltype. In thestudies presented here, the evidence that LY is taken up in a fluid phase manner by hepatocytes was the following. 1)Crude membranes or permeabilized cells did not bind LY at 4 or 37 "C. 2) Intact cells did not accumulate LY at 4 "C, only at 37 "C. 3) The rate and extent ofLY accumulation were not saturable and were energy-dependent. 4) Internalized LY was accumulated ina typical endocytic pattern, particularly localized in the perinuclear region and could be rapidly released from permeable cells. Using LY, the calculated initial rateof fluid uptake by hepatocytes was 5.7 & 2.2 pl/cell/h ( n= 8). The reported values using different markers for fluid phase uptake in isolated rat hepatocytes range from 0.04 to 30.0 pl/cell/h (Munniksma et al., 1980; Oseet al., 1980; Gordon et al., 1987; Scharschmidt et aL, 1986; Sasaki et al., 1987). The reason for this wide range in rate of fluid uptake is unclear, although it is known that thegrowth state of cells can dramatically influence their fluid phase uptake capability

(Wiley and Cunningham, 1982). It has also been reported that fluid phase endocytosis capability changes in the ratwith age (Horbach et al., 1986). It is possible that the wide variation from different laboratories also reflects the different methodologies used to isolate or culture hepatocytes. The complex process of fluid phase endocytosis involving membrane vesicle formation, fusion, and recycling continued in cells under a severe osmotic stress. Despite a 100-200% increase in the osmolarity of the medium, the cells remained viable for hours with normal ATP levels, and the rateof fluid phase uptake was virtually unchanged. This is thefirst report of a treatment that completely blocks a receptor-mediated pathway (>98%) without affecting (52% inhibition) a fluid phase pathway. Hepatocytes may bea unique cell type in this regard. Thirion and Wattiaux (1988)observed that 10 p~ monensin had no effect on the fluid phase uptake of sucrose in rathepatocytes but thatsucrose-labeled asialofetuin uptake was inhibited about 70%. The kinetics of coated pit recycling cessation in the presence of hyperosmolar sucrose was extremely rapid. ASOR uptake was stopped within 1-2 min after making the medium hyperosmolar. Under thesesame conditions, the uptake of LY was virtually unaffected, suggesting that fluid phase endocytosis occurs by a different pathway than the receptor-mediated uptakeof the ligand. Hyperosmolarity inhibits receptor-mediated endocytosis of transferrin (Bowen and Morgan, 1988), asialoglycoprotein (Oka and Weigel, 1988), low density lipoprotein (Heuser and Anderson, 1987), and chemotactic peptide (Daukas and Zigmond, 1985). 0.2 M Sucrose added to complete medium increases the osmolality approximately 2-fold (from 264 to 555 mmol/kg) and completely inhibits continuous endocytosis of lZ5I-ASORby hepatocytes in suspension (Oka and Weigel, 1988). All of the effects on cells in suspension could also be

Hyperosmolarity Inhibits Receptor-mediated Not Fluid Phase Endocytosis observed with cells in culture. In both cases, the same qualitative results were obtained in terms of the sensitivity of ligand uptake and processing steps to increasing osmolarity. However, cells in suspension were moresusceptible to increasing osmolarity than were cells in culture. For cells in suspension, the optimal sucrose concentration necessary to get a maximum inhibition was 0.2 M, whereas 0.4 M sucrose was required for cells in culture. The same results were observed for the uptake of low density lipoprotein in fibroblasts by Heuser and Anderson (1987), who concluded that hyperosmolarity interfered with receptor-mediated endocytosis by disrupting the coated pit cycle and causing the generation of abnormal nonfunctional coated pits. Our finding that coated pit frequency decreased by >80% in cultured hepatocytes treated with sucrose agrees with this conclusion. Since hyperosmolarity decreased the number of coated pits on the cell surface, it is also likely that the coated pit cycle is constitutive. The relative resistance of cultured cells to the effect of hyperosmolarity may be due to their attachment to the substratum or the higher surface:volume ratio. The anchored cells tolerated the physical stress with less deformation of the cell surface. At the electron microscopy level, the plasma membrane of the suspension cells had numerous microvilli and pericellular vacuoles induced by the hyperosmolar sucrose. This was not the case with the cultured cells. Depletion of intracellular K' has also been used to arrest coated pit formation and to stop receptor-mediated endocytosis in fibroblasts (Larkin et al., 1983, 1985). In the latter studies, potassium-depleted fibroblasts or hepatocytes had 70% fewer coated pits than the control cells. Fibroblasts had 30 coated pits/mm of membrane, whereas hepatocytes in perfused livers contained -150 coated pits/mm of sinusoidal membrane. We found -100 coated pits/mm of apical membrane in cultured hepatocytes. The difference in coated pit frequency between hepatocytes and fibroblasts may berelated to a greater amount of endo/exocytic activity in hepatocytes or to a different surface:volume ratio. The mechanism by which hyperosmolarity induced by sucrose blocks the coated pit recycling pathway is not known. Hyperosmolarity decreases the transmembrane potential in hepatocytes (Howard and Wondergem, 1987). Hepatocyte volume was essentially constant at 37 "C but not at4 "Cwith up to a 50% increase in osmolality. Henderson et al. (1986) linked a decreased membrane potential in hepatocytes with intracellular acidification secondarily caused by inhibition of a membrane Na+/H' exchange activity. Sandvig et al. (1987) showed that cytoplasmic acidification induced by several techniques blocks receptor-mediated uptake via coated pits, although the number of coated pits was not reduced. Heuser et al. (1987) reported that nigericin-induced K+ depletion, hyperosmolarity, and ATP depletion all cause cytoplasmic acidification to pH6.2-6.5. They proposed that thelow pH induces aberrant clathrin assembly to give microcages on the plasma membrane and in the cytoplasm, thus stopping the coated pit cycle. The reason that the coated pit number decreases with hyperosmolarity but not necessarily with cytoplasmic acidification remains to be reconciled. Both the internalization and the externalization, or efflux, of LY were biphasic, suggesting the involvement of more than one kinetic step in these processes. Multiple kinetic steps in fluid phase endocytosis have also been described in macrophages and fibroblasts (Bestermanet al., 1981; Murphy, 1985; Swanson et al., 1985). Our results are consistentwith the idea that theuptake kinetics represent a gradual sequential filling of intracellular compartments further intothe cell with time.

12023

The ability of the internalized LY to be released indicates that the transfer of material among these various compartments and their communication with the extracellular medium are reversible. The results indicate that a large compartment that requires a long time to fill may equilibrate slowly with a smaller compartment that fills more quickly. The larger compartment in turn takes alonger time to empty. Alternatively, the results are explained by two parallel compartments: a small one that fills rapidly and a large one that fills slowly. Cells can have separate pathways for the uptake of fluid components and membrane-adsorbed components (Gonnella and Neutra, 1984; Gonatas et al., 1984). Brown et al. (1987) reported that endocytosis of horseradish peroxidase by kidney collecting duct intercalated cells occurred by a nonclathrincoated vesicle pathway. The alternative pathway in this cell type also involves a cytoplasmic coat on the inner side of the plasma membrane, but the protein involved is not clathrin. Other studies suggest that the inhibition of the coated pit recycling pathway by depletion of intracellular K+ does not necessarily block the uptakeof other molecules. For example, ricin but not diphtheria toxin was transported into HepG2 cells after hypotonic shock and K+ depletion (Moya et al., 1985). Under the same conditions, these cells were unable to internalize transferrin orlow density lipoprotein. In thiscase, ricin is not a fluid phase marker but a membrane marker, presumably an adsorbed membrane component following bulk membrane recycling. Many of the above studies suggest the operation of alternate pathways for both membrane-bound and fluid phase components otherthan a receptor-mediated coated pit pathway. The relative contribution of different pathways within agiven cell type has been investigated in only a few cases. The coated pit pathway appears to account for approximately 50% of the fluid phase uptake capacity in CV-1 African green monkey kidney cells (Doxsey et al. 1987) and about 16% in polymorphonuclear leukocytes (Daukas and Zigmond, 1985). Intracellular K+ depletion markedly decreased the uptake of horseradish peroxidase in human fibroblasts (Larkin et al., 1983), indicating that virtually all of the fluid uptake may be accounted for by the coated pit pathway. Marsh and Helenius (1980) calculated that thefluid uptake into coated vesicles in baby hamster kidney-21 cells is similar to the fluid phase uptake rates in macrophages and fibroblasts. Although this conclusion was supported by indirect evidence, it may nonetheless apply to thesituation infibroblasts. The major finding in the present study was that conditions that almost completely inhibited galactosyl receptor-mediated endocytosis via the coated pit pathway did not decrease total fluid phase uptake in hepatocytes (