ATP depletion causes a reversible redistribution and inactivation of a ...

Communication

THEJOURNALOF BIOLOGICAL CHEMISTRY Vol. 262 No. 5 Issue of February 15 pp. 1942-1945 1987 0 1987 hy ‘fhe American Society of Bioiogical Chemists, Inc. Printed in U.S.A.

ATP Depletion Causes a Reversible Redistribution and Inactivation of a Subpopulation of Galactosyl Receptorsin Isolated Rat Hepatocytes* (Received for publication, August 28, 1986) Douglas D. McAbee and Paul H. WeigelS From the Division of Biochemistry, Department of Human Biological Chemistry and Genetics, The University of Texas Medical Branch, Galueston, Texas77550

Isolated rat hepatocytes, treated with metabolic energy poisons such as NaNs in the absence of exogenous ligand, lose surface galactosyl (Gal) receptor activity (Clarke, B. L., and Weigel, P . H. (1985) J. Biol. Chern. 260, 128-133). We have used ‘2SI-labeledasialo-orosomucoidand affinity-purified anti-receptor IgG to quantitate, respectively, the activity and the amount of Gal receptor protein. Cells were treated with NaN3 at 37 “C and the surface or total (surface and intracellular) binding of these two probes was measured at 4 “C, respectively, in intact cells or in cells permeabilized with digitonin. As a function of NaN, concentration, both surface receptor activity and protein decreased in parallel by 50-80%. Virtually all of the lost surface receptor protein was found inside the cell, but only about 50% of all cellular Gal receptors were active. As determined by equilibrium binding studies, this decreased receptor activity reflected an overall loss of ligand binding sites with little change in binding affinity of the remaining Gal receptors for asialo-orosomucoid. When ATP was restored, normal surface receptor activity and number completely recovered even in the absence of protein synthesis. We conclude that a subpopulation of Gal receptors constitutively recycles and undergoes an inactivation/reactivation cycle. In the absence of ligand, these receptors are normally internalized and then inactivated. Loss of cellular ATP blocks receptor reactivation, prevents the reappearance of receptors at the cell surface and redistributes Gal receptors as inactive receptors accumulate intracellularly.

During its lifetime, a hepatic Gal’ receptor recycles and mediates the uptake of thousands of asialoglycoproteins (14). We have been investigating the energy requirements and mechanism of the receptor recycling pathway. Continuous uptake of ASOR by isolated rat hepatocytes, which is depend* This work was supported by National Institutes of Health Grant GM 30218. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solelyto indicate this fact. $ To whom correspondence should be addressed. The abbreviations used are: Gal, galactosyl; ASOR, asialo-orosomucoid; BSA,bovine serum albumin; IgGR,affinity-purified goat antireceptor antibody; PBS, phosphate-buffered saline; HEPES, 4-(2hydroxyethy1)-1-piperazineethanesulfonicacid.

ent on receptor recycling, is blocked when cellular ATP is reduced by >25%. Internalization of a singlewave of prebound ligand, however, is unaffected by >98% ATP depletion (5). Therefore, receptor recycling, but not internalization per se, requires ATP. Furthermore, in the absence of ligand, metabolically poisoned cells lose about 50% of their surface ASOR bindingactivity (5-8). Similarligand-independent loss of surface Galreceptor activity occurs in hepatocytes treated with colchicine (9), chloroquine (lo), or monensin (11, 12) or equilibrated at temperatures below 37 “C (13). Lysosomotropic amines produce similar effects in human hepatoma cells (14). Based on this indirectevidence, most investigatorshave assumed that these modulated receptors are trapped within the cell and haveconcluded that receptors recycle in the absence of ligand. Alternatively, loss of receptor activitymay occur withoutthenet removal of receptorsfromthe cell surface, as is the case for cultured hepatocytes treated with high concentrations of monensin (12). If Gal receptors constitutively recycle, then energy-dependent steps, such as intracellular receptor processing or translocation back to the cell surface, may be inhibited andbecome rate-limiting when ATP levels are lowered. In this study, we characterized the NaN3-induced changes in thecellular distribution, surface and interior, of Gal receptor activity and protein. lZ5I-IgGRand lZ5I-ASORwere used to quantitate, respectively, Gal receptor proteinandbinding activityinthepresenceor absence of the permeabilizing detergent digitonin. The results provide the first direct evidence for two important conclusions. First, a subpopulation of Gal receptorsrecycles by an ATP-dependentprocess in the absence of ligand. Following ATP depletion these receptors leave the surface andaccumulate inside the cell. Second, constitutive recycling must involve inactivationandthen reactivation of receptors. ATP-depleted cells accumulate inactive receptors intracellularly and do not re-express receptors at thecell surface. EXPERIMENTAL PROCEDURES

Materials-Human orosomucoid (ocl-acid glycoprotein), a gift of Dr. M. Wickerhauser of The Plasma Derivatives Laboratory of the American Red Cross, was desialylated and iodinated as described previously (15). Collagenase (type I), BSA (fraction V), CNBr-activated Sepharose 4B, neuraminidase (type X), calf thymus DNA, Percoll, and digitonin were from Sigma. HEPES was from Research Organics, Inc. Bisbenzimide (Hoeschst Dye 33258) wasfrom Behring Diagnostics. Male Sprague-Dawleyrats (150-200 g, bodyweight) were obtained from Timco Breeding Laboratories, Houston, TX. Na’*’I (10-20 mCi/pg iodine) was from Amersham Corp. Allother chemicals were reagent grade. Medium 1 contains modifiedEagle’smedium (Grand Island Biological Co.,catalogue 420-1400) supplemented with 2.4 g/liter of HEPES, pH 7.4, and 0.22 g/liter of NaHC03. Medium 1/BSA is Medium 1 containing 0.1% (w/v) BSA. All binding assays were done in Hanks’ balanced salt solution (16). Hepatocyes-Hepatocytes were prepared by the collagenase perfusion procedure of Seglen (17) as described previously (15, 18).Cells were kept at room temperature during the filtration and differential centrifugation steps and then were suspended in ice-cold Medium 1; they were >90% viable and single cells. Experiments were performed in Medium 1/BSA in the absence of serum. Prior to experiments, cell suspensions (2 X 106/ml in Medium 1/BSA) were incubated at 37 “C to increase and stabilize the number of surface receptors/cell (13). Cell viability was determined by trypan blue exclusion. Antibody Preparation-A male goat was innoculated subcutaneously at multiple sites with 100 pg of affinity-purified Gal receptor

1942

Reuersible InactiuationlRedistribution of Gal Receptors (19). Innoculations, emulsified with either Freund's complete (initial innoculation) orincomplete (subsequent innoculations) adjuvant,and bleedings were alternated every 10 days. Whole IgG fraction was isolated from serum by Na,SO, precipitation. IgGR,characterized in detail elsewhere:was affinity-purified by chromatography on Gal receptor-Sepharose 4B. Purified IgGR was dialyzed against 0.9% NaCl, 50 mM Tris-HC1, pH 7.8, and stored a t -70"C. Iz5I-IgGR, prepared by the method of Fraker and Speck (20), had specific activities of 100-650 cpm/fmol. Western blot analysis (21) of Triton X-100 extracts of hepatocytes (19) using IgGRindicated that most of the antibody was directed against the major 42-kDa subunit; very slight labeling of the two larger minor receptor subunits (22) was observed. As determined by Scatchard analysis of equilibrium binding studies (23), lZ5I-IgGR bound to both intact and digitonin-permeabilized isolated rat hepatocytes with equal, moderately high affinity ( K d = 7.7 1.6 X lo-' and 7.2 f 1.1 X M, respectively), and >90% of the binding was abolished by a 50- to 100-fold excess of unlabeled IgGR, hut not preimmune IgG. Binding of '''l-IgGR to intact and digitonin-permeabilized cells was linear up to, respectively, 2.5 X lo6 and 1.2 X IO6cells/sample using 13 pg of '251-IgGR/ml.Binding assays employed 51.0 X lo6 cells. ATP Depletion-ATP depletion in cells was achieved prior to measuring the binding of lZ5I-ASORand lZ5I-IgGR. Freshly prepared 500 mM stock solutions of NaN, in PBS were used. Cells were treated with NaN3 at 37 "C for 30-45 min (a time at which maximal ATP depletion was obtained), rapidly chilled by addition of 3 volumes of ice-cold Hanks' solution, overlaid on discontinuous Percoll gradients (5), andcentrifuged a t 400 X g for 10 min at 4 "C. The gradients were composed of 0.5 ml of 50%, 10 ml of 40%, and 10 ml of 30% Percoll in PBS in a 50-ml centrifuge tube. For recovery experiments, immediately after NaN3treatment cells were rapidly chilled to 4 "C, gently washed once in Hanks' solution, and incubated at 37 "C in Medium M cycloheximide.Cellular ATP completely l/BSA containing 2 X recovered within 30 s at 37 "C. Samples were rapidly cooled to 4 "C and then centrifuged over discontinuous Percoll gradients as above. Only intact cells (>98% viable), isolated from these discontinuous gradients, were used in binding assays. A modification of the luciferinluciferase procedure of Stanley and Williams (24) was used to quantitate ATP (25). Binding Assays-Viable control or poisoned hepatocytes (-lo6 cells/sample) isolated from Percoll gradients were incubated in Hanks' solution containing either1pg/ml '251-ASOR or13 pg/ml lZ5IIgGRat 4 "C for 60 min, with occasional mixing. Total (surface and intracellular) binding was measured in the presence of 0.055% digitonin (26) added 10 min prior to the addition of lZ5I-ASORor '"1IgGR.Surface binding only was measured in the absence of digitonin. Digitonin at 0.055% permeabilizes cells, releases cytosolic proteins of 2200 kDa without solubilizing Gal receptors, and makes intracellular receptors accessible to added ligand (26). The cells, kept on ice, were then washed twice with Hanks' solution, resuspended in 0.5 ml of 0.1 M NaCl, 50 mM sodium phosphate, pH 7.4, 5 mM EDTA, sonicated for 60 s at 80 watts in a water bath sonicator (Laboratory Supplies Co., Hicksville, NY), and assayed for DNA and cell-associated radioactivity. General-Cellular DNA was determined by the method of Labarca and Paigen (27) using calf thymus DNA as standard. Centrifugation of cell suspensions was at 800 rpm for 2 min at 4 "C in a refrigerated TJ-6 tabletop centrifuge (Beckman Instruments). lZ5I radioactivity was determined using a Packard Multipras 2 gamma spectrometer. Cell suspensions were incubated in Erlenmeyer flasks, in which suspensions occupied 10% of the flask volume, and shaken a t 100 rpm in a gyratory water bath at 37 "C. RESULTS AND DISCUSSION

In thefollowing experiments, only intact cells isolated from discontinuous Percoll gradients after the experiment were used in binding assays. This procedureis absolutely necessary when working with severely ATP-depletedhepatocytes in suspension ( 5 ) . Surface '=I-ASOR binding, assayed at 4 "C, decreases with time when isolated rat hepatocytes are ATPdepleted a t 37 "C, but not a t 15 "C, a temperature which blocks receptor recycling ( 5 ) . If Gal receptors recycle constitutively (5, 6, 9-11, 13, 14), then ATP depletion may trap

* D. D.

McAbee and P. H. Weigel, manuscript in preparation.

1943

surfacereceptorswithinthe cell. Accordingly, lZ5I-ASOR binding activity lost from thecell surface would be recovered if intracellular receptor activity was measured. To address this possibility, hepatocytes were treated for 30 min at 37 "C with 15 mM NaN3, which is sufficient to deplete cellular ATP by 60% in 2-3 min ( 5 ) .The cells were then washed and placed back a t 37 "C without NaN3 in the presence of cycloheximide. Cell samples, immediately after NaN3 treatment and during recovery, were isolated on Percoll gradients and assayed at 4 "C for surface and total cell '251-ASOR binding (Table I). As expected, NaN3 treatmentcaused a 63% reduction in "'1ASOR binding at the cell surface. Moreover, surface binding completely recovered within 2 h, independent of protein synthesis, when NaN3 was removed. Surprisingly, NaN, treatment also reversibly reduced the total cellular '251-ASOR binding by 43%. The lostsurface receptor activity, therefore, was not intracellular. Scatchard analysis (23) of equilibrium 1251-ASORbinding to digitoninpermeabilized cells showed that, at saturation, control cells bound 204 fmol of '251-ASOR/~gof DNA with a K d = 1.2 X lo-' M and that NaNs-treated cells bound 123 fmol of Iz5IASOR/pg of DNA with a K d = 2.5 X lo-' M. These data confirm thatreduced '251-ASORbinding toNaN,-treated cells reflected a bona fide loss of ligand binding sites and was not due to changes inGal receptor-ASOR binding affinity. Since NaN3 itself does not inhibit ASOR binding to Gal receptors (5-8), ATP depletion either inactivatedor made inaccessible a subpopulation of receptors. Therefore, affinity-purified antireceptor IgG was used to determine the cellular distribution and content of Gal receptors. Binding of Iz5I-IgGRto 37 "C equilibrated untreatedcells increased 2-4-fold in the presence previously the of 0.055% digitonin (data not shown), similar to reported increase for "'I-ASOR binding (26). This is consistent with the fact that50-90% of all cellular Gal receptors are located intracellularly ( 3 , 4, 26, 28). Hepatocytes were treated a t 37 "C for 45 min with increasing concentrationsof NaN,, isolated on Percoll gradients, and assayed at 4 "C in the presence of digitonin for binding of both 1251-ASOR,to measure receptor activity, and Iz5I-IgGR, to measure receptor protein (Fig. 1).By 30 min, the kinetic change in bindingof these two ligands plateaus and attainsa new steady state.' Theoretically, IZ5I-IgGRbinding topoisoned cells should be constant if Gal receptors remained accessible and if lZ5I-IgGRrecognized both active and inactive Gal receptors equally well (Fig. 1, short-dash line). Alternatively, if receptors eitherwere made inaccessibleor were not recognized TABLEI Depletion and recouery of galactosyl receptor activity after NaN3 treatment Cells were incubated for 30 min at 37 'C in Medium l/BSA with or without 15 mM NaN3, washed, allowed to recover in the presence of 2 X M cycloheximide, which blocked protein synthesis by >98%, and '*'I-ASOR binding to intact and digitonin-permeabilized cells was measured as described under "Experimental Procedures." ATP recovery was complete within 30 s after removing the inhibitor. Treatment

Recovery time

rnin

None

0 60

16 15 mM NaN,

120 0 25

a

60

120 Values are the mean of duplicates.

'*'I-ASOR bound" Surface and Surface intracellular only fmol/& DNA

16 15 6 14 17

28 27 24 16 26

Reversible InactivationlRedistributionof Gal Receptors

1944 1

S

I

I

I

1

1

I

I

'b

'zsI-ASOR BOUND (% of control)

A

6

.,

2 ;$0

*

_""

'%"x"-x"-x l'6

2'4

SODIUM AZIDE (mM)

'

-

-I

-1 W

FIG. 2. Effect of NaNS on cellular A1P content and lz51FIG. 1. Binding of 'zaI-IgGRto permeabilized cells as a func- ASOR and 'zaI-IgGRbinding to permeabilized cells. Cells were tion of '251-ASOR binding activity. Hepatocyteswere treated treated with various concentrations of NaN3, assayed for lZ5I-ASOR with 0-25 mM NaN3 for45 min at 37 "C, rapidly chilledto 4 'C, and (0, and A) and lZ5I-IgGR (0,0, and A) binding, and processed as centrifuged on discontinuous Percoll gradients, and viable cells were described in Fig.1and under "Experimental Procedures." Each point permeabilized with digitonin and assayed for lZ5I-ASOR and lZ5I-IgGR represents the mean of duplicates. Each symbolrepresents a different binding as described under "Experimental Procedures." Binding of experiment.ATP content (X) isnormalized to cellnumber and lZ5I-ASOR and lZ5I-IgGR is normalized to cellular DNA content and presented as the percent of ATP relative tountreated cells. presented asthe percent bound relative to untreated cells. Each point represents the mean of duplicates. Each symbolrepresents a different experiment.The solid line was calculatedby linear regression analysis ( r = 0.976). Dashed lines are discussed in the text. by the antibody after being inactivated, then loss of lZ5IASOR binding and lZ5I-IgGRbinding would parallel one another (Fig. 1, long-dash line). Of the initial lZ5I-IgGRbinding activity, >92% was retained despite a 50% loss of lZ5I-ASOR binding activity. Scatchard analysis (23) of equilibrium lZ5IIgGRbinding todigitonin-permeabilized cells showed that, at saturation, control cells bound 1127 fmol of lZ5I-IgGR/pgof DNA with a Kd = 6.4 X lo-' M, and NaN3-treatedcells bound 1067 fmol of lZ5I-IgGR/pgof DNA with a Kd = 7.6 x lo-' M. This shows that NaN3-treated hepatocytes exhibited no significant alterationin either binding affinity or the tot'al number of binding sites for lZ5I-IgGR,compared to control cells. Therefore, lZ5I-IgGRbindingisessentiallyindependent of receptor activity and is a valid measure of Gal receptor content. In addition, this resultshows that theloss of lZ5I-ASOR binding in ATP-depleted cells is not due to receptorinaccessibility since binding of 1251-IgGR by these cells in the presence of digitonin was virtually unchanged. Since IgGR competes for lZ5I-ASORbinding to the receptor, it is also unlikely that the loss of receptor activity in NaN3-treated cells is due to the receptor binding endogenous ligands.We conclude, rather, that ATPdepletion inactivates a subpopulation of Gal receptors. The effect of NaN3 concentration on the activity of total cellular Gal receptors clearly showed that only a subpopulation of receptors was affected (Fig. 2). lZ5I-ASORbinding steadily decreased to 50% of control between 3 and 10 mM NaN3, andloss of ATP followed a very similar dose response curve. Severe ATP depletion down t o 4 - 2 % of control with either high concentrations of NaN3 alone or combined with NaF did not further increase receptor inactivation (data not shown). lZ5I-IgGRbinding, however, remained virtually constant across the entire range of NaN3 tested. Therefore,only a portion of the surface and the intracellular Gal receptor activity was modulated by depleting cellular ATP. The above results suggest that the loss of cell surface lZ5IASOR binding after ATP depletion could reflect a redistribution of inactivereceptors from the surface tothe cell interior. Alternatively, surface receptorscould be inactivated, but remain on thecell surface. To test this latterpossibility,

W

1

20

0

-

0

0

0

4

8

12

16

20

SODIUM AZIDE (mM)

FIG. 3. Effect of NaNs on '261-ASORand lzaI-IgGRbinding to the surface of intact cells. Cells were treated with 0-20 mM NaN3 for45 min at 37 "C and assayed at 4 "C for the binding of lZ5IASOR (0)and lZ5I-IgGR (0)in the absence of digitonin as described in Fig.1and under "Experimental Procedures." Each point represents the mean of duplicates. cells were treated with increasing concentrations of NaN3 and assayed for '251-ASOR and lZ5I-IgGRbinding at 4 "C in the absence of digitonin (Fig. 3). Between 2 and 12 mM NaN3, cellsurfacelZ5I-ASOR and lZ5I-IgGRbindingdramatically decreased, respectively,80 and 70%. The close correspondence of these dose responsecurves indicates that the loss of surface receptor activityis accompanied by the loss of surface receptor protein. Similar experiments routinely produced losses of cell surface Gal receptor activity and protein between 50 and 70%. Since lZ5I-IgGRbinding to NaN3-treatedcells in thepresence of digitonin remained constant (Fig. 2), ATP depletionredistributed surface Gal receptors to an intracellular compartment(s) at 37 "C. The mechanism of receptor inactivation and itsrole in the function(s) of the Gal receptor remain to be explained. Gal receptor inactivation is notdue to inaccessibility of receptors to added ligand, nor is it likely that receptors are occupied with endogenousligand.Amore likely possibility isthat

Reversible InactivationlRedistribution of Gal Receptors

1945

2. Tanabe, T., Pricer, W. E., and Ashwell, G. (1979) J. Biol. Chem. inactive and then active Gal receptors are generated during 254,1038-1043 receptor recycling by reversible covalent modification. For 3. Steer, C. J., and Ashwell, G. (1980) J. Biol. Chem. 2 5 5 , 3008instance,phosphorylation of theepidermal growth factor 3013 receptor by protein kinaseC inhibits epidermalgrowth factor 4. Warren, R., and Doyle, D. (1981) J. Biol. Chem. 2 5 6 , 1346-1355 receptor binding activity (29,30). Drickamer and Mamon (31) 5. Clarke, B. L., and Weigel, P. H. (1985) J. Biol. Chem. 2 6 0 , 128133 have shown that the chicken hepatic receptor is phosphoryl6. Weigel, P. H.,Clarke, B. L., and Oka, J. A. (1984) J. Cell Biol. ated on the cytoplasmic domain. Likewise, the mammalian 99,372a Gal receptoris a phosphoprotein (32,33)with a serine residue 7. Scarmato, P., Durand, G., Agneray, J., and Feger, J. (1986) Biol. suitable for phosphorylation located on a cytoplasmic domain Cell 56,255-258 near the membrane junction(22). Therefore, reversible phos8. Tolleshaug, H., Kolset, S. O., and Berg, T. (1985) Biochem. phorylation of the Gal receptor could modulate receptor acP h a r m o l . 34,1639-1645 9. Kolset, S. O., Tolleshaug, H., and Berg, T. (1979) Exp. Cell Res. tivity. The involvement of ATP at one or more steps in the 122,159-167 recycling pathway is a thermodynamic requirement and coupling this to the regulationof receptor activity may provide 10. Tolleshaug, H., and Berg, T. (1979) Biochem. Pharmacal. 28, 2919-2922 additional controlof the system(5). For example, inactivation 11. Berg, T., Blomhoff, R., Naess, L., Tolleshaug, H., and Drevon, C. of receptor could facilitate receptor-ligand dissociation and A. (1983) Exp. Cell Res., 148, 319-330 effect efficient segregation of receptor and ligand into their 12. Fiete, D., Brownell, M. D., and Baenziger, J. U. (1983) J. Biol. Chem. 258,817-823 respective intracellular routes. In summary, we conclude that ATP depletion by NaN3 in 13. Weigel, P. H., and Oka, J. A. (1983) J. Biol. Chem. 2 5 8 , 50895094 .". the absence of added ligand redistributes Gal receptors from 14. Schwartz, A. L., Bolognesi, A., and Fridovich, S. E. (1984) J. Cell the cell surface to thecell interior. Furthermore,loss of ATP Biol. 98, 732-738 inactivates about 50% of the total Gal receptor population. 15. Weigel, P. H., and Oka, J. A. (1982) J. Biol. Chem. 257, 1201Kineticanalyses suggest thatreceptorinactivation occurs 1207 intracellularly after active receptors areinternalized.* Recep- 16. Hanks, J. H.,and Wallace, R. E. (1949) Proc. SOC.Exp. Biol. Med. 7 1 , 196-200 tor reactivation follows restoration of cellular ATP and does 17. Seglen, P. 0. (1973) Exp. Cell Res. 82,391-398 not require de nouo protein synthesis. Our interpretation is Weigel, P. H. (1980) J. Biol. Chem. 2 5 5 , 6111-6120 that a subpopulation of receptors, which we have designated 18. 19. Ray, D. A., and Weigel, P. H. (1985) Anal. Biochem. 1 4 5 , 37-46 State 2 Gal receptors (34,35), constitutively recycles and that 20. Fraker, P. J., and Speck, J. C. (1978) Biochem. Biophys. Res. inactivation/reactivation normallyoccurs duringreceptor Commun. 80,849-857 processing in this receptor pathway. We have proposed (35) 21. Burnette, W. N. (1981) Anal. Biochem. 112, 195-203 that State1and State2 Gal receptors function to internalize, 22. Drickamer, K., Mamon, J . F., Binns, G., and Leung, J. 0. (1984) J. Biol. Chem. 2 5 9 , 770-778 dissociate, and degrade asialoglycoproteins in hepatocytes by 23. Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 5 1 , 660-672 two separate parallel pathways. State 2 Gal receptor reacti- 24. Stanley, P. E., and Williams, S. G. (1969) Anal.Biochem. 2 9 , vation requires ATP. Hence,ATP depletion by NaN3 inhibits 381 "_ -392 _" receptor reactivation and inactive receptors accumulate intra25. Weigel, P. H., and Englund, P. T. (1975) J. Biol. Chem. 250, 8536-8542 cellularly. This latter result contrasts with the Gal receptor modulation induced by 100 ~ L Mmonensin, in which inactive 26. Weigel, P. H., Ray, D.A., and Oka, J. A. (1983) Anal. Biochem. receptors remained on the cell surface (12). There is no reason, 133,437-449 27. Labarca, C., and Paigen, K. (1980) Anal. Biochem. 102,344-352 however, to expect thesetwo agents to have identical effects. 28. Pricer, W. E., and Ashwell, G. (1971) J. Biol. Chem. 246, 4825Indeed, these two inhibitors should proveuseful in future 4833 efforts todefine the multiple steps in the constitutive receptor 29. Fearn, J. C., and King, A. C. (1985) Cell 40,991-1000 30. Davis, R. J., and Czech, M. P. (1986) Biochem. J. 233. 435-441 recycling pathway.

Acknowledgments-We thank Deborah Baudy for technical assistance, Betty Jackson for help preparing the manuscript, Janet Oka for doing the artwork, Dr. Darryl Ray for help preparing the antibody, and Dr. Benjamin Clarke for stimulating discussions. REFERENCES 1. Regoeczi, E., Debanne, M. T., Hatton, M.W. C., and Koj, A, (1978) Biochim. Biophys. Acta 541,372-384

31. Drickamer, K., and Mamon, J. F. (1982) J . Biol. Chem. 257, 15156-15161 32. Schwartz, A. L. (1984) Biochem. J. 223, 481-486 33. Takahashi, T., Nakada, H., Okumura, T., Sawamura, T., and 126, Tashiro, Y. (1985) Biochem. Biophys. Res.Commun. 1054-1060 34. Weigel, P. H.,Clarke, B. L., and Oka, J. A. (1986) Biochem. Biophys. Res. Commun. 140,43-50 35. Weigel, P. H. (1987) in Vertebrate Lectins (Olden, K., and Parent, B., eds) pp. 65-91, Van Nostrand/Reinhold, New York