Hydroxylamine treatment differentially inactivates purified rat hepatic asialoglycoprotein receptors and distinguishes two receptor populations.

We previously showed that two subpopulations of asialoglycoprotein receptors (ASGP-Rs), designated State 1 and State 2 ASGP-Rs, are present in intact cells and that State 2 ASGP-Rs can be inactivated in permeable rat hepatocytes in a temperature- and ATP-dependent manner. These inactivated ASGP-Rs can be quantitatively reactivated by the addition of palmitoyl-CoA (Weigel, P. H., and Oka, J. A.(1993) J. Biol. Chem. 268, 27186-27190). Here we show that 50% of purified rat ASGP-Rs are inactivated by treatment with hydroxylamine under mild conditions. The activity of affinitypurified ASGP-Rs was assessed by measuring the specific binding of I-asialo-orosomucoid (ASOR) in a dot-blot assay after immobilization onto nitrocellulose. Treatment of ASGP-Rs in solution with 0.0125-1.0 M NH2OH, pH 7.4, at 4°C for 4 h resulted in a progressive loss of ASOR binding activity. ASGP-R inactivation with NH2OH occurred more readily at basic pH or at room temperature. Similar treatment with Tris had no effect on ASGP-R activity. The kinetics of ASGP-R activity loss and the dose-response for this inactivation were both biphasic, indicating the presence of two equal populations of ASGP-Rs with different sensitivities to NH2OH. The more sensitive population of ASGP-Rs (50%) was inactivated by treatment with 0.2 M NH2OH (4°C, 4 h) or with 1.0 M NH2OH (4°C, 1 h) without detectable peptide cleavage as assessed by SDS-polyacrylamide gel electrophoresis. State 1 ASGP-Rs, purified from chloroquine- or monensin-treated hepatocytes, showed significantly less sensitivity to NH2OH treatment (both in kinetics and dose dependence). Furthermore, under mild conditions NH2OH caused dissociation and inactivation of 50% of the total ASGP-Rs (State 1 and State 2) that were prebound to ASOR-Sepharose, whereas the same treatment caused dissociation of only <20% of State 1 ASGP-Rs from such preformed complexes. As shown in the accompanying paper (Zeng, F. Y., Kaphalia, B. S., Ansari, G. A. S., and Weigel, P. H.(1995) J. Biol. Chem. 270, 21382-21387) all three RHL subunits of active ASGP-Rs, in fact, contain covalently attached palmitate and stearate. In cultured cells, [3H]palmitic acid is metabolically incorporated into all three subunits. These radiolabeled fatty acids are completely released from purified ASGP-Rs by mild NH2OH treatment. We conclude that the NH2OH-sensitive subpopulation of ASGP-Rs corresponds to the previously described State 2 ASGP-Rs and that these receptors require fatty acylation for their ligand binding activity.

The hepatic ASGP-R 1 mediates the endocytosis of desialylated glycoproteins containing terminal galactose or N-acetylgalactosamine (1,2). In isolated rat hepatocytes, ASGP-Rs endocytose ligand by two functionally different receptor populations; ligand is then intracellularly processed via two distinct pathways (2,3). We have designated these two receptor populations as the State 1 and State 2 ASGP-Rs. Previous studies from our laboratory and others have shown that treatment of hepatocytes with functionally diverse agents such as metabolic energy poisons, microtubule depolymerizing drugs, lysosomatropic amines, vanadate, or proton ionophores causes an inactivation and/or redistribution of cell surface State 2 ASGP-Rs without affecting the State 1 ASGP-Rs (4 -7). Studies with intact cells indicate that the State 2 ASGP-Rs are inactivated and then reactivated during receptor recycling (4). In digitoninpermeabilized cells, the State 2 ASGP-Rs can be inactivated in the absence of cytosol in a temperature-and ATP-dependent manner (8). Recent investigations demonstrated that the ATPinactivated receptor population, which corresponds to the State 2 receptors, was rapidly and quantitatively reactivated in the absence of cytosol when fatty acyl-CoAs were added (9). These studies demonstrated the occurrence of a novel ASGP-R inactivation-reactivation cycle that could regulate receptor activity during endocytosis and receptor recycling (10). Although these above results support the existence of two functionally different populations of ASGP-Rs and the occurrence of an inactivationreactivation cycle for the State 2 ASGP-R population, the structural basis for the two receptor states and the molecular mechanism(s) responsible for this cycle are unknown.
The molecular mass of the rat ASGP-R is approximately 264 kDa in nonionic detergent with a ligand-binding domain of 105-148 kDa (11,12). Based on SDS-PAGE (13), the rat ASGP-R contains three subunits (RHL1, 2, and 3) with molecular masses of ϳ42, 49, and 54 kDa, respectively. These three subunits are the products of two different genes. RHL2 and RHL3 have the same core protein, but differ in the type and extent of post-translational carbohydrate modification (14). The stoichiometry and subunit composition of native rat ASGP-Rs are still unknown, although we have suggested that ASGP-Rs are hetero-hexamers composed of four RHL1 subunits and two subunits of either RHL2 and/or RHL3 (15). That previous studies also demonstrated that the surface and inter-nal ASGP-Rs of the two functionally distinct State 1 and State 2 receptor populations have the same hetero-oliogometric subunit composition (16), indicates the two populations of ASGP-Rs may not be attributed to different subunit compositions.
The finding that fatty acyl-CoAs such as palmitoyl-CoA can regulate the activity of one receptor population (State 2 ASGP-Rs) led us to suggest that this receptor population may be fatty acylated in vivo (9,10). A palmitoylation/depalmitoylation cycle might regulate the ligand binding activity of State 2 ASGP-Rs. A growing number of membrane proteins have been found to be modified by fatty acids. Fatty acids are often covalently attached to proteins through an amide bond to an N-terminal glycine residue or through an ester or thioester bond to internal serine, threonine, or cysteine (17,18). The thioester linkages of fatty acids such as palmitic acid, are very labile to mild alkaline or NH 2 OH treatment. NH 2 OH, a chemical usually used to release thioester-linked fatty acids from acylated proteins (17,18), has been used to study the role of palmitate in the function of rhodopsin and the influenza virus spike glycoprotein (19,20). Here we report that ϳ50% of the activity of affinity-purified ASGP-Rs is lost by treatment with NH 2 OH under mild conditions and this NH 2 OH-sensitive receptor population corresponds to the State 2 ASGP-Rs. A preliminary report of these results has been presented (21).
Preparation of Hepatocytes-Male Sprague-Dawley rats (200 -300 g) were from Harlan Breeding Laboratories (Houston, TX) and from SasCo (Oklahoma City, OK). Rat hepatocytes, isolated by a modification (23) of the collagenase perfusion procedure of Seglen (24), were suspended in medium 1/BSA (2-3 ϫ 10 6 cells/ml) and incubated at 37°C for 1 h in a gyratory water bath to increase and stabilize the cell surface ASGP-R number (25). To inactivate the State 2 ASGP-Rs, in some experiments following equilibration at 37°C, hepatocytes in medium 1/BSA (2 ϫ 10 6 cells/ml) were incubated at 37°C for 60 min either with 1 mM chloroquine or 50 M monensin (26). Untreated and treated cells were rapidly chilled by transfer to 5 volumes of ice-cold medium 1, and washed twice by centrifugation. 125 I-ASOR binding activity was determined in the presence of 0.055% digitonin as described previously (22).
Purification of Active ASGP-Rs-All steps were carried out at 4°C, unless otherwise stated. ASOR was coupled to CNBr-activated Sepharose 4B to a density of 2-3 mg of ASOR/ml resin according to the manufacturer's guidelines. Active ASGP-Rs were purified from isolated hepatocytes (to prepare total ASGP-Rs) and from chloroquine-or monensin-treated hepatocytes (to prepare State 1 ASGP-Rs) by a modification of a procedure described previously (27). Briefly, rat hepatocytes, prepared as described above, were chilled with 3 volumes of ice-cold buffer 1, centrifuged, and resuspended at 1 ϫ 10 7 cells/ml in permeabilization buffer for 20 min. Permeabilized cells were washed twice with 10 mM EGTA in buffer 1 containing 1 mM PMSF, 1 g/ml leupeptin, and 1 g/ml pepstatin A to remove cytosol and any endogenous ligands. The cells were then resuspended and incubated in extraction buffer (2 ϫ 10 7 cells/ml) for 1-2 h on a rotator. After centrifugation at 10,000 ϫ g for 30 min, the supernatant was filtered sequentially through 1.2-, 0.8-, 0.45-, and 0.2-m filters, and incubated with ASOR-Sepharose for at least 2 h on a rotator. The nonspecifically bound components were removed by extensive sequential washes with BIC10 containing 0.5% Triton X-100, and then 0.05% Triton X-100. Active ASGP-Rs were removed with elution buffer, and immediately neutralized with 1 M Tris-HCl, pH 7.4, and supplemented with CaCl 2 to a final concentration of 10 mM. Individual eluted fractions were subjected to a ligand-binding competition assay to determine the presence of active ASGP-Rs (27). The pooled fractions were concentrated using a Centricon-10 device (Amicon), and stored at 4°C before using.
Incubation with NH 2 OH-The freshly purified ASGP-Rs were incubated in BIC20 containing 0.025% Triton X-100, and various concentrations of NH 2 OH, pH 7.4, on ice for the indicated times. The reaction was stopped by dilution with BIC20 (to give Ͻ0.005% Triton X-100) and loaded onto a nitrocellulose membrane. Other studies have shown that ASGP-R adsorption to the nitrocellulose is decreased if the Triton X-100 concentration is above 0.005%. 2 Ligand-binding Assay-Untreated and NH 2 OH-treated ASGP-Rs were diluted with BIC20 to give a concentration of Triton X-100 Յ0.005%, then loaded onto a 0.45-m pore size nitrocellulose membrane (0.5 g of ASGP-R/well) using a dot-blot manifold (Schleicher & Schuell). The wells were rinsed twice with 500 l of BIC20, the manifold was disassembled, and the nitrocellulose membrane was blocked with buffer 1 containing 0.1% Tween 20 at 4°C for 90 min. After washing with BIC20 containing 0.02% Tween 20, the sheet was then incubated with 1.0 g/ml 125 I-ASOR (100 -200 cpm/fmol) in BIC20 containing 0.02% Tween at 4°C for 2 h. After extensive washes with BIC20 containing 0.02% Tween 20, individual dots containing immobilized ASGP-R were cut out using a cork borer and transferred into ␥-tubes to measure radioactivity. Nonspecific binding, determined by the bound radioactivity remaining in the presence of a 50-fold excess of nonlabeled ASOR, was less than 5% of the total in all cases. All binding assays were done in duplicate.
Dissociation of Preformed Receptor-Ligand Complexes by NH 2 OH-Active ASGP-Rs, freshly purified as described above, were rebound to ASOR-Sepharose by incubating 10 g of ASGP-Rs with 10 l of ASOR-Sepharose in 50 l of BIC20 containing 0.05% Triton X-100 at 4°C for 2 h on a rotator. After a brief centrifugation, the supernatant fluid was removed using a T-200R flat end tip (RPI, Inc.), and the Sepharose pellet containing ASGP-R⅐ASOR complexes was incubated in BIC20 containing 0.05% Triton X-100, and various concentrations of either NH 2 OH or Tris at 4°C for a time course as indicated in Table I. After centrifugation, the supernatants were transferred into a new tube using a T-200R flat end tip, and the pellet was incubated with 100 l of EGTA-containing elution buffer to remove the bound ASGP-Rs. The protein content in the supernatant and pellet was then determined.
Metabolic Labeling-Hepatocytes were cultured with [ 3 H]palmitic acid and ASGP-Rs were affinity-purified as described in the accompanying paper (28).
General-Protein content was measured by the method of Bradford (29) using BSA as a standard. SDS-PAGE was carried out by the method of Laemmli (30); the samples were boiled for 1 min in the SDS sample buffer with 5% ␤-mercaptoethanol. Protein bands were visualized by silver staining (31). 125 I radioactivity was measured using a Packard multiprias 2␥ spectrometer. Statistical analyses were carried out using the Student's t test.

RESULTS
One Subpopulation of ASGP-Rs Is More Readily Inactivated by NH 2 OH-In order to assess quantitatively the effect of NH 2 OH on the ligand binding activity of affinity-purified rat ASGP-Rs, we used a recently developed dot-blot assay 2 to measure the 125 I-ASOR binding activity of purified ASGP-Rs. This assay permits rapid and reproducible quantitation of changes in the ligand binding activity of both immobilized ASGP-Rs and ASGP-Rs in solution after treatment with NH 2 OH. The use of Tween 20 as a blocking agent and its inclusion in the ligand-binding step improved the 125 I-ASOR binding avidity and reduced nonspecific binding.
To exclude the possible effects of other factors such as pH changes and temperature on ASGP-R activity, the treatment of ASGP-Rs with NH 2 OH or Tris was carried out on ice in BIC20 containing 0.025% Triton X-100 at pH 7.4. In initial studies, we found that purified total ASGP-Rs progressively lost up to ϳ50% activity after storage in BIC10 containing 0.05% Triton X-100 at 4°C for more than 1 week. Furthermore, the rate of ASGP-R activity loss rapidly increased when the temperature was Ն20°C. Freezing and thawing also caused loss of activity. We suspect that the State 2 ASGP-R activity is being lost under these conditions. For this reason, we used only freshly purified ASGP-Rs in all the experiments reported here (Ͻ12 h old) to minimize any initial loss of ASGP-R activity.
Treatment of ASGP-Rs with 1 M NH 2 OH caused ASGP-R inactivation in a time-dependent manner, whereas similar treatment with milder nucleophiles such as 1 M Tris or methylamine (not shown) had no significant effect on ASGP-R activity (Fig. 1A). The kinetics of ASGP-R inactivation in the presence of 1 M NH 2 OH were quite biphasic. The rate of inactivation during the initial 60 min was greater than that during incubation after 60 min. About 50% of the ASGP-Rs were inactivated within 1 h; extended incubation times of up to 18 h resulted in a progressive inactivation of the remaining ϳ50% of ASGP-Rs. A possible reason for ASGP-R inactivation might be that under these conditions NH 2 OH could cause cleavage of ASGP-R subunits that results in the loss of activity. To assess this possibility, the NH 2 OH-treated ASGP-Rs were analyzed by SDS-PAGE under reducing conditions (Fig. 1B). Within the first 2 h (during which time ϳ50% ASGP-R inactivation occurs), NH 2 OH treatment did not degrade any of the three ASGP-R subunits (Fig. 1B). Prolonged incubation times after 2 h indeed caused a progressive increase of ASGP-R degradation: prominent fragments of 40 and 24 kDa were observed. This result indicates that inactivation of the first ϳ50% of ASGP-Rs was not due to peptide cleavage, whereas inactivation of the second ϳ50% ASGP-Rs could be explained by the observed peptide degradation.
In another experiment, purified total ASGP-Rs were incubated with 1 M NH 2 OH on ice for 1 h (this treatment caused an inactivation of ϳ50% ASGP-Rs as shown in Fig. 1A), and NH 2 OH was then removed by repeated dilution of the mixture and subsequent concentration using a Centricon-10 device. Active ASGP-Rs were then separated from inactive ASGP-Rs by purification on ASOR-Sepharose. About 50% of the total ASGP-Rs could still bind to ASOR, the other ϳ50% had lost their ligand binding activity without detectable peptide cleavage (data not shown). This result supports the conclusion that under mild conditions, NH 2 OH inactivates one population of receptors without affecting the other population. Hydroxylamine sensitivity of ASGP-Rs was identical whether a solidphase or solution-based assay was used to assess ligand binding activity.
Another possibility is that the observed ASGP-R inactivation by NH 2 OH is not really due to reactivity of NH 2 OH, but rather to other minor components present in commercial NH 2 OH preparations, such as heavy metal ions that could activate minor contaminating proteases that copurify with ASGP-Rs. To exclude this and several alternative possibilities, other controls were performed. Incubation of purified ASGP-Rs with divalent metal ions including Mg 2ϩ , Zn 2ϩ , Ni 2ϩ , Ca 2ϩ , and Mn 2ϩ did not change their ligand binding activity. Furthermore, NH 2 OH from various sources (Sigma, Aldrich, Fluka, and MC/B) all showed very similar kinetics of ASGP-R inactivation (not shown). The addition of the divalent metal ion chelater EGTA did not affect the ability of NH 2 OH (1 M, 20 h) to inactivate up to ϳ50% of total ASGP-Rs, but did prevent TABLE I Effect of NH 2 OH on dissociation of preformed (ASGP-R)-ASOR complexes The affinity-purified total (State 1 plus State 2) and State 1 ASGP-Rs were rebound to ASOR-Sepharose as described under "Experimental Procedures." The (ASGP-R)-ASOR-Sepharose was subsequently incubated with BIC20 containing 0.05% Triton X-100, and either Tris (pH 7.4, 0.1 or 1 M) or NH 2 OH (pH 7.4, 0.1 or 1 M) at 4°C on a rotator for 1 or 18 h. After centrifugation to separate dissociated ASGP-Rs (supernatant) from those still bound (pellet), the protein contents were determined. Results are averages Ϯ standard error of two (State 1 ASGP-R) or three (total ASGP-R) individual experiments from independent receptor preparations, and are expressed as the percentage of total ASGP-Rs initially bound to ASOR-Sepharose that was dissociated.

FIG. 1. Kinetics of ASGP-R inactivation by NH 2 OH.
Freshly purified, total active ASGP-Rs were incubated in BIC20 containing 0.025% Triton X-100, and either 1 M NH 2 OH (E) or 1 M Tris (q) on ice for 0 -18 h. A, at the indicated times, samples were diluted with 40 volumes of BIC20 and loaded onto nitrocellulose (0.5 g of ASGP-R/ spot) using a dot-blot manifold. 125 I-ASOR binding activity for each sample was determined as described under "Experimental Procedures." B, at the indicated times, samples were immediately mixed with an equal volume of 2 ϫ Laemmli sample buffer (30) with 5% ␤-mecaptoethanol, boiled for 1 min, and then analyzed by SDS-PAGE using a 3.9% stacking gel and a 12.5% resolving gel. The protein bands were visualized by silver staining.
further inactivation and degradation of the remaining 50% of ASGP-R subunits (Fig. 2, A and B). This observation suggests that the activity loss of the less sensitive ASGP-R population is the consequence of subunit cleavage by NH 2 OH. Addition of protease inhibitors such as PMSF, pepstatin, and leupeptin did not reduce the effectiveness of NH 2 OH to inactivate these ASGP-Rs. Furthermore, after incubation of ASGP-Rs with Tris in the presence of Ca 2ϩ for more than 24 h, no detectable protein degradation could be observed (not shown).
The inactivation of ASGP-Rs by NH 2 OH treatment (4°C for 4 h) is also dependent on NH 2 OH concentration and this doseresponse is biphasic (Fig. 3). The extent of ASGP-R inactivation was proportional to NH 2 OH concentration in the range of 0 -0.1 M, corresponding to 0 to ϳ40% inactivation. ASGP-R inactivation was ϳ50% with 0.2 M NH 2 OH. Increasing NH 2 OH concentration in the range of 0.2-1.0 M only slightly increased ASGP-R inactivation. SDS-PAGE analysis showed no degradation products or decrease in size of RHL1, RHL2, and RHL3 after treatment with 0 to 0.2 M NH 2 OH at 4°C for 4 h (not shown).
Biphasic kinetic and dose-responses were also observed for ASGP-R inactivation by NH 2 OH using receptor immobilized on nitrocellulose, but inactivation occurred more slowly than in solution (not shown). The inactivation of ASGP-Rs was also pHand temperature-dependent (Fig. 4). At pH 6.0 and 4°C, no significant inactivation was observed with 0.5 M NH 2 OH for 1.5 h. A sharp increase in ASGP-R inactivation was seen between pH 6 and 7 and the extent of inactivation remained constant from pH 7 to 11. In comparison with treatment at 4°C, ASGP-R inactivation occurred more readily at 25°C, especially at basic pH (Fig. 4).
State 1 ASGP-Rs Are Less Sensitive to NH 2 OH Treatment Than State 2 ASGP-Rs-The above results indicate that there are two populations of ASGP-Rs with respect to NH 2 OH sensitivity. One subpopulation of total purified ASGP-Rs is much more sensitive to inactivation by NH 2 OH treatment. Although the percentage of the sensitive population was variable (30 -50%) from preparation to preparation, the biphasic activity loss with NH 2 OH treatment was always observed (n Ͼ 10). Since our previous studies have shown that the State 1 and State 2 ASGP-R populations are roughly equal (50 -50%), we surmised that the NH 2 OH-sensitive subpopulation of ASGP-Rs may be the State 2 ASGP-Rs. State 2 ASGP-R activity has previously been shown to be inactivated by a variety of different treatments in isolated intact rat hepatocytes (26). We, therefore, are able to purify either total active ASGP-Rs (representing State 1 plus State 2) or State 1 ASGP-Rs. Presently, the State 2 ASGP-R population cannot be purified free of State 1 ASGP-Rs.
To confirm that the NH 2 OH-sensitive ASGP-R population corresponds to State 2 ASGP-Rs, we purified State 1 ASGP-Rs, after first inactivating State 2 ASGP-Rs in intact cells with either chloroquine or monensin as described under "Experimental Procedures." Treatment of isolated rat hepatocytes with chloroquine (1 mM, 60 min, at 37°C) or monensin (50 M, 60 min, at 37°C) resulted in 40 -60% inactivation of total ASGP-Rs as assessed by 125 I-ASOR binding with permeable cells. In comparison with untreated cells, about 40 -50% of active ASGP-Rs could be purified from treated cells, verifying that about half of the ASGP-Rs were indeed inactivated. The purified State 1 ASGP-Rs showed very similar subunit patterns to that of total ASGP-Rs by SDS-PAGE under reducing conditions (not shown). We then compared the sensitivity of State 1 ASGP-Rs and total (State 1 plus State 2) ASGP-Rs purified in parallel, to NH 2 OH treatment. Kinetically, State 1 ASGP-Rs showed significantly less sensitivity to NH 2 OH inactivation than the total ASGP-R pool (Fig. 5). During the first 30 min, 30 -35% of the total ASGP-Rs were inactivated, whereas State 1 ASGP-R activity remained essentially unchanged. After 30 min incubation, a slow progressive loss of State 1 ASGP-R activity occurred that reached up to ϳ40% inactivation at 180 min. The difference in the rate of activity loss between State 1 versus State 1 plus State 2 ASGP-Rs was significant within 1 h (p Ͻ 0.05). However, the difference became smaller after 3 h. This latter observation is in agreement with the finding noted above, that extended incubation causes protein degradation.
The dose dependence for ASGP-R inactivation by NH 2 OH also showed a marked difference between the State 1 versus State 1 plus State 2 ASGP-R populations (Fig. 6). At Յ50 mM NH 2 OH, no significant loss of State 1 ASGP-R activity was observed after incubation at 4°C for 4 h, whereas ϳ25% of the activity of the total receptors was lost. The difference in the ability of NH 2 OH to cause inactivation was more marked at a lower concentration range than at higher concentration (Ն400 mM). The purified State 1 ASGP-Rs prepared from either chloroquine-or monensin-treated hepatocytes, showed essentially identical sensitivity to NH 2 OH (both in kinetics and dose dependence; Figs. 5 and 6), indicating that treatment of isolated cells with chloroquine or monensin inactivated the same ASGP-R population (State 2), as reported previously (26). These results strongly support the conclusion that the NH 2 OH-sensitive population is the State 2 receptor population.

NH 2 OH Treatment of ASGP-Rs Bound to ASOR-Sepharose Readily and Selectively Dissociates the State 2 Population of ASGP-Rs-
As an alternative way to assess the effect of NH 2 OH treatment on ASGP-Rs, we determined the ability of NH 2 OH at pH 7.4, 4°C, to cause dissociation and inactivation of preformed ASGP-R⅐ASOR-Sepharose complexes in the presence of CaCl 2 . Treatment with Tris (0.1 M, 18 h; or 1 M, 1 h) did not significantly dissociate purified ASGP-Rs from ASOR-Sepharose using either total (State 1 plus State 2) or State 1 ASGP-Rs (Table I), whereas NH 2 OH treatment (0.1 M, 18 h; or 1 M, 1 h) dissociated ϳ40% of State 1 plus State 2 ASGP-Rs but Ͻ20% of State 1 ASGP-Rs from receptor-ASOR complexes. The dissociated ASGP-R subunits showed no degradation or decreased size as determined by SDS-PAGE (Fig. 7). To verify that the NH 2 OH treatment reduced the affinity of ASGP-R for ASOR, which then caused the dissociation of the complexes, the ligand binding activity of the dissociated ASGP-Rs was assessed by using a dot-blot assay and by measuring their ability to rebind to ASOR-Sepharose. Both assays showed that ASGP-Rs, dissociated in the presence of NH 2 OH, were subsequently inactive. Thus, receptor-ligand dissociation was due to inactivation of ASGP-Rs, rather than to an effect on the ASOR-Sepharose or to a transient reduction of the receptor affinity for ASOR.
In contrast to total purified ASGP-Rs, only 14 -19% of the State 1 ASGP-Rs were dissociated and inactivated by treatment with NH 2 OH under the same conditions (Table I). The difference in inactivation between total versus State 1 ASGP-Rs was about 25% (p Ͻ 0.005) for both treatment conditions tested (0.1 M NH 2 OH, 18 h; 1 M NH 2 OH, 1 h). That only 14 -19% of receptors were still released from State 1 ASGP-R⅐ASOR complexes by NH 2 OH under conditions that removed almost 50% of total ASGP-Rs, is in agreement with the results shown in Figs. 5 and 6, and could be due to the presence of a small amount of State 2 ASGP-Rs in our State 1 ASGP-R preparations. We find that the extent of ASGP-R inactivation in isolated rat hepatocytes treated with chloroquine or monensin varies (40 -60%) from cell preparation to preparation, even using the same conditions (26). At present, we cannot determine the percentage of State 2 ASGP-Rs that copurifies with State 1 ASGP-Rs by this method. Nonetheless, our results FIG. 4. Effect of pH and temperature on ASGP-R inactivation by NH 2 OH. Freshly purified, total active ASGP-Rs were incubated on ice or at 25°C for 90 min with 0.5 M NH 2 OH (f, ) or 0.5 M Tris (q) in 10 mM Hepes at the indicated pH containing 142 mM NaCl, 6.7 mM KCl, 20 mM CaCl 2 , and 0.025% Triton X-100. The treated samples were diluted with 100 volumes of BIC20, loaded onto nitrocellulose, and 125 I-ASOR binding activity was assessed as described in the legend to Fig. 1.

FIG. 5. Kinetic difference between State 1 and total (State 1 plus State 2) ASGP-R inactivation by NH 2 OH treatment.
Active State 1 ASGP-Rs ( , Ⅺ), freshly purified either from chloroquinetreated ( ), or monensin-treated hepatocytes (Ⅺ), and active total ASGP-Rs (q), freshly purified from nontreated cells, were incubated in BIC20 containing 0.025% Triton X-100 and 0.5 M NH 2 OH on ice for 0 -3 h. At the indicated times, the 125 I-ASOR binding activity of samples was assessed. Each point is the mean Ϯ S.E of four to six independent experiments. Values significantly different from the control (total ASGP-Rs) are indicated: *, p Ͻ 0.001; ϩ, p Ͻ 0.005; #, p Ͻ 0.05. demonstrate the existence of two receptor populations in purified ASGP-Rs.
One potential disadvantage of using the dot-blot assay to quantitate changes in ASGP-R activity after treatment with NH 2 OH is that only about 10% of immobilized ASGP-Rs are capable of binding ASOR (assuming that one ASGP-R binds one ASOR). 2 The majority (ϳ90%) of immobilized receptors are unable to bind ASOR, perhaps due to conformational inflexibility of immobilized molecules or inaccessibility to ligand. Differential adsorption or preferential conformational changes of active or inactivated ASGP-Rs during immobilization onto nitrocellulose could complicate interpretation of the results. Several controls were performed to address this possibility. Using specific antireceptor antibody to quantitate nontreated or NH 2 OH-treated ASGP-Rs after immobilization, we found that NH 2 OH treatment did not change the adsorption of receptor. Most importantly, the finding that ϳ50% of ligand-bound ASGP-Rs are more readily dissociated and inactivated by NH 2 OH (Table I) corroborates the results obtained from the dot-blot assay. These results also argue against a difference in adsorption ability or in conformational changes of immobilized active versus inactivated ASGP-Rs.

NH 2 OH Treatment Releases [ 3 H]Palmitate from
Metabolically Labeled ASGP-Rs-Consistent with the possibility that State 2 ASGP-Rs are sensitive to NH 2 OH treatment because they are fatty acylated, we determined whether receptors could be radiolabeled when hepatocytes are cultured in the presence of [ 3 H]palmitic acid. As shown in the accompanying paper (28), all three RHL subunits are metabolically labeled. Virtually all of the [ 3 H]palmitate was released from RHL1, RHL2, and RHL3 by the same NH 2 OH treatment conditions that inactivates State 2 ASGP-Rs (Fig. 8). These data indicate that ASGP-Rs are fatty acylated, predominantly at Cys residues, and that the loss of State 2 ASGP-R ligand binding activity occurs when the receptor is deacylated with hydroxylamine.

DISCUSSION
Our previous studies have demonstrated that rat hepatocytes contain two functionally different receptor populations: State 1 ASGP-Rs and State 2 ASGP-Rs (2). Although present in approximately equal numbers, the State 2 ASGP-Rs mediate the large majority (ϳ80%) of ligand processing (i.e. endocytosis, segregation, and degradation) in hepatocytes. Functionally therefore, State 2 receptors are roughly about 4 times more active than State 1 receptors. In contrast to State 1 ASGP-Rs, the State 2 ASGP-Rs can be modulated in their cellular distri-bution and/or ligand binding activity by treating cells with low temperature or with metabolic energy poisons, microtubule drugs, monensin, chloroquine, or vanadate (4,5,25,32,33). Many of these agents cause selective inactivation of the State 2 ASGP-Rs, which led us to propose that this receptor population undergoes an intracellular inactivation-reactivation cycle during receptor-mediated endocytosis and receptor recycling (34).
Since the inactivation of State 2 ASGP-Rs is transient and reversible, we suggested that receptor inactivation could be the mechanism by which cells achieve efficient ligand dissociation and subsequent segregation of ligand from receptor (34). During receptor-mediated endocytosis, the concentration of ligand in endosomes can be increased up to 10 4 -fold over the extracellular concentration (2). Cell surface ASGP-Rs are also concentrated about 50-fold in endosomal membranes. Under these conditions of such high concentrations, dissociated ligand molecules would likely rebind to active receptors even at the lower pH of early endosomes. Any ligand rebound to receptor would not be shuttled to lysosomes for degradation, but rather nonproductively recycled back to the cell surface. Such "futile" receptor cylces would, of course, be wasteful. Most significantly, futile ligand recycling would increase as the extracellular ligand concentration increased, and the endocytic machinery would function less efficiently when the need for ligand clearance was greatest. This situation, which could be physiologically deleterious, is avoided by inactivating ASGP receptors so that ligand rebinding does not occur.
The cumulative evidence indicates that the ASGP-R is a hetero-oligomeric complex, composed of RHL1 and RHL2/3 subunits. (i) Both RHL1 and RHL2/3 gene products are required for expression of a functional ASGP-R capable of binding ASOR in transfected cells (35). (ii) Subunit-specific antibodies coimmunoprecipitate all three RHL subunits of the ASGP-R (16,36). (iii) Studies with a chemical affinity derivative of 125 I-ASOR showed identical cross-linking patterns for State 1 or State 2 ASGP-Rs (16). (iv) Affinity-purified ASGP-Rs from cells with active State 1 or State 1 plus State 2 ASGP-Rs show identical subunit patterns in SDS-PAGE. 3 These above results indicate that functional differences between State 1 and State 2 ASGP-R are not due to different subunit compositions.
This and the accompanying paper (28) establish for the first time a structural difference between the two populations of receptor and the possible role of this structural difference in the functional differences between State 1 and State 2 ASGP-Rs. The molecular basis for the inactivation-reactivation cycle of State 2 ASGP-Rs is related to fatty acylation. We previously 3  showed that palmitoyl-CoA rapidly and quantitatively reactivates inactivated State 2 ASGP-Rs, suggesting that either ASGP-R subunits or other unknown regulatory proteins are palmitoylated and that a palmitoylation-depalmitoylation cycle may regulate ASGP-R activity (9). These results prompted us to examine more directly if ASGP-Rs are modified by fatty acylation. As detailed in the accompanying paper (28), all three RHL subunits contain covalently attached palmitate and stearate. These fatty acids were found using gas chromatographymass spectroscopic analysis of purified RHL subunits after SDS-PAGE of active ASGP-Rs. Our results in the present study provide further evidence to support this conclusion; metabolic labeling with [ 3 H]palmitate also confirms that RHL1, RHL2, and RHL3 are fatty acylated in intact cells.
Surprisingly, only the State 2 ASGP-Rs, not State 1 ASGP-Rs, are palmitoylated in a metabolic labeling experiment. Essentially all of the palmitic acid metabolically incorporated into RHL1, RHL2, and RHL3 is released by mild hydroxylamine treatment that inactivates only the State 2 ASGP-Rs. We conclude that State 2 ASGP-Rs are clearly fatty acylated. However, further studies are needed to determine whether State 1 ASGP-Rs are fatty acylated; it is quite possible that they are not.
NH 2 OH treatment decreases ASGP-R activity in a time-and dose-dependent manner ( Figs. 1 and 2). Inactivation of ASGP-Rs is kinetically biphasic; ϳ50% of ASGP-Rs are very sensitive to NH 2 OH and the other 50% are much less sensitive. During the loss of NH 2 OH-sensitive ASGP-R activity, no protein degradation occurs, indicating that loss of activity is not due to peptide cleavage. Although the percentage of NH 2 OHsensitive ASGP-Rs ranged from 30 to 60% with different ASGP-R preparations, the phenomenon of biphasic ASGP-R inactivation by NH 2 OH was always observed. This observation is in agreement with our previous results in intact cells, that the percentage of State 2 ASGP-Rs that are sensitive to treatment with diverse agents varies from 40 to 60% with different hepatocyte preparations (32,33). The studies using purified State 1 ASGP-Rs confirm that the NH 2 OH-sensitive population corresponds to the State 2 ASGP-Rs. Furthermore, the purification process may also cause inactivation of some State 2 ASGP-Rs, as we find that a small percentage (5-10%) of ASGP-R activity is lost after elution from ASOR-Sepharose. At present, we cannot quantitate the amount of receptor that may be inactivated during purification, due to the lack of specific antibodies that can differentiate between active and inactivated ASGP-Rs.
Our results in this study suggest that ASGP-Rs are acylated in vivo. Many membrane proteins and receptors are palmitoylated, often at Cys residues near transmembrane domains (37)(38)(39)(40)(41)(42). The subunits of the rat and human ASGP-Rs have a Cys-Ser sequence close to the cytoplasm membrane junction (14,43). This same sequence in the transferrin receptor (44) and the HLA-D-associated invariant chain (45) is palmitoylated at Cys. It is possible, therefore, that the same position in one or more of the ASGP-R subunits is also palmitoylated. Inactivation of ASGP-R by NH 2 OH is likely due to the removal of covalently bound fatty acids from one or more of the ASGP-R subunits. Since all three RHL subunits contain covalently bound fatty acids (palmitic acid and stearic acid) and ASGP-Rs are hetero-oligomeric, there are many possible partially deacylated receptor species whose ligand binding activity could be affected.
We also find that NH 2 OH treatment, but not Tris treatment, results in the formation of dimeric RHL subunits (based on SDS-PAGE analysis under nonreducing conditions) indicating that deacylation with NH 2 OH generates free thiol groups that then form disulfide bonds in a time-dependent way. 4 This formation of new disulfide bonds upon NH 2 OH treatment has been demonstrated in other palmitoylated proteins such as vesicular stomatitis virus G glycoprotein (38) and human tissue factor (46). NH 2 OH has been widely used to remove thioester bond-linked palmitate from many palmitoylated proteins, such as transferrin receptor (37), rhodopsin (19), and virus glycoproteins (38). Under mild conditions, depalmitoylation by NH 2 OH did not greatly affect the conformational structure of rhodopsin as determined by circular dichroism (39). Although NH 2 OH has usually been used as a specific chemical to release ester and thioester-linked fatty acids from fatty acylated proteins (40,41), it also has a number of other chemical effects on proteins. NH 2 OH (usually Ն1 M, pH Ն 9.0; Ն37°C) cleaves susceptible Asn-Gly bonds in many proteins (47). 2 M NH 2 OH at pH 9 and at 45°C has been used specifically to cleave at the C-terminal side of succinimides in proteins (48). NH 2 OH is also known to remove O-acetyl groups from tyrosine (49). Our results also show that 1 M NH 2 OH cleaves all three RHL subunits after extended incubation (Ͼ4 h). Loss of the ligand binding activity of the less NH 2 OH-sensitive State 1 ASGP-R population is most probably the consequence of this subunit cleavage.
The functional significance of palmitoylation is not known for the majority of proteins containing this modification (41,42). Since the turnover of fatty acyl groups is generally much faster than that for the protein, the potential exists to have a cyclic, or reversible, alteration of some function of the protein tied to its fatty acylation. Acylation/deacylation cycles may serve as a general mechanism for controlling the subcellular distribution and/or function of many proteins. For example, the activation-induced depalmitoylation results in activationinduced translocation of the G-protein ␣ subunit into cytoplasm (50). Regulation of the palmitoylation state of p21 N-ras , a key component of growth factor receptor signaling pathways, affects its avidity for binding to membranes and therefore regulates its ability to interact with other proteins (51). Palmitoylation of the neuronal growth cone protein, GAP43, alters its ability to stimulate G o protein (52). Palmitoylation of the transferrin receptor may regulate the receptor-mediated endocytosis of diferric transferrin (53).
Based on the results here, in the accompanying paper (28), and our previous results that palmitoyl-CoA reactivates inactive State 2 ASGP-Rs (9, 10), we propose that a dynamic deacylation/acylation cycle in vivo is the molecular basis for the inactivation/reactivation cycle that State 2 ASGP-Rs undergo during receptor mediated endocytosis.