The Major Subunit of the Asialoglycoprotein Receptor Is Expressed on the Hepatocellular Surface in Mice Lacking the Minor Receptor Subunit*

The mammalian asialoglycoprotein receptor (ASGPR) is located on the sinusoidal membrane of hepatocytes where it binds and endocytoses galactose-terminated glycoproteins (asialoglycoproteins). ASGPR is composed of two highly homologous subunits, termed he- patic lectin 1 and 2. Despite numerous studies the con-tribution of both subunits to biosynthesis and functional activity of ASGPR in vivo has remained con-troversial. Mice lacking the murine hepatic lectin (MHL)-2 subunit are viable and fertile without obvious phenotypic abnormalities. In the absence of MHL-2, knockout mice express MHL-1 protein at reduced levels. Here, we examine the intracellular fate and function of this remaining subunit. The results show that MHL-1 reaches the hepatocellular surface in knockout mice but is unable to effectively remove any one of three different radiolabeled ligands within 30 min. A small but detect- able residual ligand clearance in knockout mice at 4 h is apparently not mediated by remaining MHL-1. Serum concentrations of galactose-terminating glycoproteins are not elevated in these ASGPR-deficient mice. How- ever, competitive in vitro degradation experiments suggest that other endogenous ASGPR ligands, the nature of which remain to be determined, accumulate in serum of knockout animals. The Thus, our results suggest the presence of as yet unidentified endogenous ASGPR ligand(s), which accumulate(s) in MHL-2 (cid:50) / (cid:50) serum and competitively inhibit(s) 125 I-asialoorosomucoid uptake and degradation by primary cultures of wild type mouse hepatocytes.

The mammalian asialoglycoprotein receptor (ASGPR) is located on the sinusoidal membrane of hepatocytes where it binds and endocytoses galactose-terminated glycoproteins (asialoglycoproteins). ASGPR is composed of two highly homologous subunits, termed hepatic lectin 1 and 2. Despite numerous studies the contribution of both subunits to biosynthesis and functional activity of ASGPR in vivo has remained controversial. Mice lacking the murine hepatic lectin (MHL)-2 subunit are viable and fertile without obvious phenotypic abnormalities. In the absence of MHL-2, knockout mice express MHL-1 protein at reduced levels. Here, we examine the intracellular fate and function of this remaining subunit. The results show that MHL-1 reaches the hepatocellular surface in knockout mice but is unable to effectively remove any one of three different radiolabeled ligands within 30 min. A small but detectable residual ligand clearance in knockout mice at 4 h is apparently not mediated by remaining MHL-1. Serum concentrations of galactose-terminating glycoproteins are not elevated in these ASGPR-deficient mice. However, competitive in vitro degradation experiments suggest that other endogenous ASGPR ligands, the nature of which remain to be determined, accumulate in serum of knockout animals.
The asialoglycoprotein receptor (ASGPR) 1 is a heterooligomeric receptor that is abundantly expressed on the sinusoidal (i.e. basolateral) surface of the hepatic plasma membrane. AS-GPR is an endocytic receptor that rapidly binds and internalizes galactose-terminated glycoproteins (asialoglycoproteins (ASGP)) from the circulation (1,2). The ASGPR in the mouse is composed of two highly homologous subunits, murine hepatic lectin (MHL) 1 and 2, each consisting of a cytosolic NH 2 -terminal domain, a single transmembrane segment (3), a stalk domain, and a Ca 2ϩ -dependent carbohydrate binding domain at the very COOH terminus (4).
Under normal circumstances, the penultimate galactose residues of glycoproteins are masked by terminal sialic acid moieties. Upon enzymatic removal of sialic acid, the now terminal galactose residues constitute the recognition determinants for ASGPR (5,6). Binding of ligands to ASGPR depends on (i) the amount and positioning of terminal galactose residues on the ligand (7-9); (ii) the presence of Ca 2ϩ in an optimal concentration of 0.1-2 mM (10); and (iii) a pH above 6.5 (11).
The fact that two largely homologous receptor subunits have been conserved in mammalian hepatocytes has led to several investigations into the roles of the respective polypeptides in intracellular transport and functional activity on the cell surface. Using cross-linking experiments on purified rat receptor and hepatocyte membrane, Halberg et al. concluded that the major and minor receptor species form independent homooligomers in the membrane (12). More recently, Bider et al. have shown that H1-overexpressing cells lacking H2 can bind iodinated asialoorosomucoid (13). Substantial binding, however, depended on the expression levels of H1 protein. High levels were required to observe ligand binding, whereas no specific binding was detected at moderate expression levels. Several other groups found that the individual ASGPR subunits have to interact with one another to form a single multicomponent receptor (14 -20).
In the present experiments we used MHL-2-deficient mice generated previously in this laboratory (21). Disruption of the MHL-2 gene results in a complete absence of MHL-2 protein and a substantial reduction in the expression of the major MHL-1 subunit. The knockout mice are viable and fertile, have a normal lifespan under laboratory conditions, and display no obvious phenotypic abnormalities. The current experiments were designed to explore the fate of the remaining major MHL-1 subunit of ASGPR in vivo. In particular, we wanted to determine whether the still detectable fraction of MHL-1 is targeted through the secretory pathway in MHL-2 Ϫ/Ϫ mice. Our results suggest that the major MHL-1 polypeptide is expressed on the surface of hepatocytes in knockout animals but is apparently not involved in mediating a detectable low basal rate of residual ASGP clearance. Although ASGPR deficiency does not result in an increase in the absolute serum concentration of endogenous galactose-terminated glycoproteins, in vitro competition experiments suggest that (an)other ligand(s) accumulate(s) in their circulation. The nature of these putative alternative ASGPR ligands is currently unknown. * This work was supported by Grant HL20948 from the National Institutes of Health and by funds from the Keck Foundation and the Perot Family Foundation. 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 18 U.S.C. Section 1734 solely to indicate this fact.
EXPERIMENTAL PROCEDURES C57BL/6 ϫ 129Sv hybrid mice (age, 6 weeks to 6 months) used in the experiments were bred in house and fed ad libitum during the course of the studies (Teklad 6% Mouse/Rat Diet 7001, Harlan Teklad Premier Laboratories Diets). Animal care and experimental procedures involving animals were conducted in accordance with institutional guidelines. Biochemicals were obtained from Sigma unless indicated otherwise. Human ␣ 2 -macroglobulin activated with methylamine was a gift from Dudley K. Strickland (American Red Cross, Rockville, MD). All 125 Ilabeled ligands were radiolabeled using the IODO-GEN Iodination Reagent (Pierce) according to the manufacturer's recommendations.
Preparation and Immunoblot Analysis of Mouse Liver Membrane Proteins-Membrane proteins were prepared from mouse livers as described previously (22). Protein content of membrane preparations was determined using the Coomassie Plus Protein Assay Reagent (Pierce). 50 g of protein/lane were separated by 9% SDS-PAGE under reducing conditions. Proteins were transferred to nitrocellulose paper at 4°C and incubated with polyclonal rabbit antipeptide antibodies (5 g/ml). Bound IgG was detected using the enhanced chemiluminescence (ECL) system (Amersham Corp.).
Subcellular Fractionation of Mouse Liver Membranes-Mouse liver membranes were fractionated using a modified version of a previously published method (23). Briefly, livers of mice of the indicated genotypes were perfused via the portal vein with 20 ml of ice-cold 0.15 M NaCl, dissected out, and homogenized in 5 volumes of a 0.25 M sucrose buffer solution containing 3 mM imidazole, pH 7.4, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 20 g/ml aprotinin. The homogenate was centrifuged at 12,000 ϫ g for 10 min at 4°C, and the supernatant was collected and subsequently centrifuged at 150,000 ϫ g for 90 min at 4°C. Membranes were resuspended in 1.9 ml of the 0.25 M sucrose buffer/g of wet liver weight. 3-5 ml of the resuspended microsomal membranes were layered on top of 32-ml linear sucrose gradients (density 1.10 -1.25 g/ml) buffered with 3 mM imidazole, pH 7.4, and centrifuged for 18 h at 83,000 ϫ g in a swinging bucket rotor. Thirty fractions (1.2 ml each) were collected from the gradient and subjected to 9% SDS-PAGE under nonreducing conditions. Immunoblot analysis of separated proteins was done as described under "Preparation and Immunoblot Analysis of Mouse Liver Membrane Proteins." Measurement of Circulating Galactose-terminated Glycoproteins-Lyophylized pellets were obtained from ϳ100 l of mouse serum, dissolved in 400 l of distilled water, and stored frozen. 100-l aliquots were diluted to 500 l and centrifuged for 20 min at 5,000 ϫ g on a Centricon filter with a 10,000 M r cut-off (Amicon Co., Beverly, MA). After an additional wash with 500 l of distilled water, the protein retentate was recovered by centrifugation for 2 min at 1,000 ϫ g and diluted to a final volume of 300 l. The protein content of this material was unchanged from that of the original solution as determined by the BCA protein assay (Pierce). Enzymatic hydrolysis of terminal galactose residues was carried out by adding 5 l of 5 M ammonium acetate, pH 5.0, and 5 l of Streptococcus pneumoniae ␤-galactosidase (1 unit/ml, Boehringer Mannheim) to 300 l of the filtered serum. Incubation was continued overnight at 37°C. Released galactose was determined using a Dionex high performance liquid chromatography (Dionex Corp., Sunnyvale, CA) equipped with a pulsed amperometric detector and pellicular anion exchange column (Carbopac PA-1.4 ϫ 250 mm). Elution was carried out isocratically with 20 mM NaOH at a flow rate of 0.8 ml/min, and detector sensitivity was set at 300 nA. Data were analyzed using the Dionex Glycostation software.
Generation of Desialylated Glycoproteins-Asialoorosomucoid was generated by incubating 100 mg of orosomucoid in 10 ml of 0.1 M sodium acetate buffer containing 2 mM CaCl 2 , pH 5, with 1 unit of insoluble neuraminidase type X-A attached to beaded agarose for 4 h at 37°C. After adding another unit of the enzyme, the incubation was continued overnight. The neoglycoprotein (Galacto-BSA) was obtained by incubating ␤-D-galactopyranosylphenylisothiocyanate (Galacto) with bovine serum albumin (BSA) according to a method derived from (24). Briefly, 1 ml of a BSA solution (2 mg/ml in 1 M sodium carbonate, pH 9.0) was incubated with 50 l of Galacto-dimethyl sulfoxide (1 mg/ml in dimethyl sulfoxide) in the dark for 8 h at 4°C. NH 4 Cl was added to 50 mM final, and the solution was incubated for 2 h. Galacto-BSA conjugate was separated from unbound Galacto on a Sephadex G-25 column (PD-10, Pharmacia Biotech Inc.) and stored at 4°C in a lightproof container.
Turnover Experiments-Prior to injection of the radioactive proteins, the animals were anesthetized by intraperitoneal Nembutal injection (ϳ80 g/g body weight). 5 g of the iodinated ligand diluted in 200 l of Tris-buffered saline containing 0.1% BSA were injected within 10 s into the external jugular vein. At the indicated time points blood samples were collected by retro-orbital bleeding, and the amount of trichloroacetic acid-precipitable radioactivity in 20 l of plasma was determined as described (25). Specific activities of the iodinated ligands were: 125 I-asialoorosomucoid, 985-1,325 cpm/ng; 125 I-orosomucoid, 620 -1,288 cpm/ng; 125 I-Galacto-BSA, 1,066 -2,655 cpm/ng; 125 I-BSA, 780 cpm/ng; 125 I-asialofetuin, 455 cpm/ng.
Isolation of Primary Mouse Hepatocytes-Hepatocytes were isolated from adult mice following the two-step portal perfusion method described previously (26) with the following modifications: (i) the enzyme solution contained 100 units/ml of collagenase type I from Clostridium histolyticum (Worthington Biochemical Corp., Freehold, NJ); (ii) following perfusion and transfer of the mouse liver to a sterile Petri dish the cells were separated by combing with a stainless steel flea comb for cats (Lambert Kay, Cranbury, NJ); (iii) passage of the cell suspension through a Nylon mesh after combing was omitted. Cell yield and viability as determined by use of a hematocytometer and trypan blue dye exclusion were consistently 2-4 ϫ 10 7 hepatocytes/ml and greater than 90%, respectively. Freshly isolated mouse hepatocytes were seeded in Primaria surface modified polystyrene plates and dishes (Becton Dickinson Labware, Oxnard, CA). 60 ϫ 15-mm culture dishes (12 ϫ 10 6 cells/dish) and 24-well plates (2 ϫ 10 6 cells/well) were used for degradation experiments and determination of specific degradation, respectively.
Cellular Degradation of 125 I-Labeled Ligands-Isolated primary mouse hepatocytes were allowed to attach for 1-3 h at 37°C. The medium was replaced by DMEM (without glutamine) containing 0.2% (w/v) BSA (degradation medium) and the iodinated ligands. Cellular degradation of the 125 I-labeled proteins was determined as described previously (27) and is expressed as nanograms of 125 I-labeled trichloroacetic acid-soluble (noniodide) material released into the culture medium per mg of total cell protein at the indicated time points. Protein content of cell lysates was determined by the Coomassie Plus Protein Assay Reagent (Pierce). Nonspecific degradation of 125 I-asialoorosomucoid was evaluated in the presence of an 100-fold excess of unlabeled asialoorosomucoid after incubation at 37°C for 23 h and was subtracted from total degradation to yield specific degradation. Unlabeled asialoorosomucoid was extensively dialyzed against degradation medium. The specific activities of the ligands were: 125 I-asialoorosomucoid, 863-895 cpm/ng; 125 I-orosomucoid, 1,160 cpm/ng; 125 I-methylamine-activated human ␣ 2 -macroglobulin, 4,433 cpm/ng.
Competitive in Vitro Degradation Assay-Primary MHL-2 ϩ/ϩ hepatocytes were freshly isolated from a single wild type mouse for each of eight separate experiments (n ϭ 8) as described under "Isolation of Primary Mouse Hepatocytes" and used within 1 h after isolation. Immediately prior to incubation of the primary MHL-2 ϩ/ϩ hepatocytes with degradation medium containing 1 g/ml (n ϭ 4), 2 g/ml (n ϭ 3), or 0.5 g/ml (n ϭ 1) 125 I-labeled asialoorosomucoid, respectively, the serum of either MHL-2 ϩ/ϩ or MHL-2 Ϫ/Ϫ was added to degradation medium to yield four different dilutions (1:300, 1:100, 1:30, and 1:10, respectively). Total cellular degradation for each dilution was determined in duplicates (n ϭ 1) or quadruplicates (n ϭ 7) after incubation at 37°C for 8 h as described previously (27). Serum was prepared by obtaining ϳ1,200 l of blood from two MHL-2 ϩ/ϩ and two MHL-2 Ϫ/Ϫ animals, respectively, transferring the blood into 1.5-ml microcentrifuge tubes and letting blood clots form while keeping the samples on ice. The blood samples were then centrifuged at 14,000 rpm for 10 min. Serum was collected, transferred into new 1.5-ml microcentrifuge tubes, pooled according to genotype, and stored on ice until used. Nonspecific degradation of 125 I-labeled asialoorosomucoid was determined in the presence of a 200-fold excess of unlabeled asialoorosomucoid that had been dialyzed overnight against degradation medium. The specific activities of iodinated asialoorosomucoid were 778 -4,410 cpm/ng. RESULTS We have previously reported the generation of mice lacking the minor subunit (MHL-2) of the ASGPR by homologous recombination in embryonic stem cells (21). Knockout animals do not express MHL-2 as judged by Western blot analysis of liver membrane proteins using an MHL-2-specific anti-peptide antibody ( Fig. 1, lanes 2 and 3) and expression of the major MHL-1 subunit is substantially reduced (Fig. 1, lanes 2 and 3). The low density lipoprotein receptor-related protein (LRP), an endocytic receptor that is not functionally or physically related to ASGPR, is not affected by the knockout of the MHL-2 subunit (Fig. 1, lanes 1-4). This result raised the question of whether the apparent lack of phenotypical abnormalities in the MHL-2 Ϫ/Ϫ animals might be due to the fact that a fraction of the residual MHL-1 protein manages to escape intracellular degradation, reach the surface of liver cells, and mediate uptake of ASGP by forming quasi-functional galactose-binding receptors. To explore this issue we first characterized the intracellular processing of the major ASGPR subunit using subcellular fractionation by sucrose density ultracentrifugation to separate intracellular organelles of wild type (MHL-2 ϩ/ϩ ) and MHL-2-deficient (MHL-2 Ϫ/Ϫ ) mouse livers (Fig. 2). The distribution of the indicated proteins was determined by immunoblotting. The processed form of LRP (LRP 85), a protein whose intracellular pathway has been extensively studied and that undergoes proteolytic processing in a post Golgi compartment (28,29), serves as a marker for the endosomal compartment and plasma membrane. Although expression levels are significantly lower no major differences are observed in the distribution profiles of MHL-1 in MHL-2 Ϫ/Ϫ mice compared with control animals (Fig. 2, MHL-2 ϩ/ϩ ). Furthermore, MHL-1 can be localized in the same fractions as LRP 85, strongly suggesting that the major subunit of ASGPR is being delivered to the hepatocyte plasma membrane. In other subcellular fractionation experiments not shown here, MHL-1 subunits also colo-calize with the low density lipoprotein receptor.
Interestingly, MHL-1 shows a shift in mobility in SDS-PAGE between fractions 18 and 20 in the knockout animals (Fig. 2, MHL-2 Ϫ/Ϫ ), indicating that the protein migrates at an apparently increased molecular mass in the later fractions enriched in Golgi and endosomal membranes. To investigate whether  2 and 4) or the absence (lanes 1 and 3) of 50 milliunits of neuraminidase type X. All samples were analyzed by 9% SDS-PAGE under reducing conditions. The proteins were transferred to nitrocellulose membranes and incubated with 5 g/ml anti-MHL-1 IgG. Binding of the IgG was detected using the ECL system. A, MHL-2 Ϫ/Ϫ mice. B, MHL-2 ϩ/ϩ mice.   1 and 4) or homozygous for the MHL-2 gene disruption (lanes 2 and 3). Liver membrane proteins (50 g/lane) were separated by 9% SDS-PAGE under reducing conditions and transferred to nitrocellulose filters. Filters were incubated with polyclonal rabbit antibodies (5 g/ml) directed against the carboxyl-terminal sequences of MHL-2, MHL-1, and the 85-kDa subunit of the LRP. Bound IgG was detected using the ECL system (Amersham Corp.). carbohydrate modifications indicative of trans-Golgi processing are responsible for the observed differences in MHL-1 migration, we determined the sensitivity of MHL-1 protein to neuraminidase (Fig. 3). After 1 h of neuraminidase digestion, the protein band in both wild type (Fig. 3B) and MHL-2 knockout mice (Fig. 3A) displays an increased mobility in the gel (Fig. 3,  compare lanes 2 and 4 with lanes 1 and 3). Thus, MHL-1 in the later fractions was carrying N-linked carbohydrates in the neuraminidase-sensitive form, further arguing strongly in favor of the major subunit having traversed the trans-Golgi en route to the cell surface in MHL-2 Ϫ/Ϫ animals.
Upon observation that MHL-1 protein is expressed at the hepatocellular surface of knockout animals, we next tested whether these remaining ASGPR subunits harbor any residual functional activity by exploring the ability of Ϫ/Ϫ animals to clear desialylated ligands from their circulation. As illustrated in Fig. 4A, asialoorosomucoid but not orosomucoid, the corresponding glycoprotein that does not exhibit terminal galactose residues and therefore is not a ligand for ASGPR, is removed from the circulation of MHL-2 ϩ/ϩ mice within 30 min. In contrast, the clearance rate of 125 I-labeled asialoorosomucoid in knockout animals is indistinguishable from that of iodinated orosomucoid (Fig. 4B). We used two other ligands, a synthetically derived neoglycoprotein prepared by chemically coupling BSA to a monosaccharide (galactose) derivative (Fig. 4, C and D) and asialofetuin (Fig. 4, E and F). Comparison of the clearance kinetics of 125 I-Galacto-BSA and unmodified 125 I-BSA in knockout animals reveals a small but reproducible difference. 125 I-Galacto-BSA is consistently removed at a slightly faster rate than 125 I-BSA from plasma of knockout animals (Fig. 4D). To further explore the possibility of a low residual clearance activity of the remaining MHL-1 receptor subunits in MHL-2 Ϫ/Ϫ mice that was not readily detectable within the 30-min interval, we extended the turnover period to 4 h. The results of these experiments show a 75% clearance rate for 125 I-asialoorosomucoid in knockout animals (Fig. 5, right panel, open  triangles) compared with a 54% nonspecific disappearance of 125 I-orosomucoid from the plasma of animals of either genotype after 4 h (Fig. 5, left and right panels, closed triangles).
In order to determine whether the low level ligand clearance in MHL-2 Ϫ/Ϫ animals is mediated by residual MHL-1 protein, we performed degradation experiments using isolated primary mouse hepatocytes. As shown in Fig. 6A (open symbols), 125 Ilabeled asialoorosomucoid is efficiently degraded by isolated hepatocytes of MHL-2 ϩ/ϩ mice over 23 h. 125 I-Orosomucoid, the corresponding sialylated ligand, is not degraded by hepatocytes of either genotype (Fig. 6, A and B, closed triangles). The results of 125 I-asialoorosomucoid degradation by isolated MHL-2 Ϫ/Ϫ hepatocytes parallel those obtained by turnover studies in the MHL-2 Ϫ/Ϫ animals in that again a low level rate of ligand uptake can be observed (compare Fig. 6B with Fig. 5, right panel). In control experiments we show that disruption of the MHL-2 gene locus does not result in a general dysfunction of the endocytic apparatus of the mutant cells. As demonstrated in Fig. 6D, methylamine-activated human ␣ 2 -macroglobulin, one of the well-characterized ligands of LRP (30,31), is readily degraded by hepatocytes isolated from MHL-2 Ϫ/Ϫ animals.
To control for nonspecific binding of radiolabel to the membrane surface, the following experiment was performed (Fig. 7). Nonspecific degradation was determined by measuring the degradation of 125 I-asialoorosomucoid over 23 h on addition of an 100-fold excess of unlabeled asialoorosomucoid and subtracting it from the total degradation of iodinated ligand. As illustrated in Fig. 7 (open triangles), no specific degradation by isolated hepatocytes from MHL-2 Ϫ/Ϫ mice is detectable over 23 h. As expected, MHL-2 ϩ/ϩ hepatocytes specifically degrade iodinated asialoorosomucoid (Fig. 7, closed triangles). Taken together, our data clearly demonstrate that the major ASGPR subunit, although expressed on the hepatocellular surface of MHL-2 Ϫ/Ϫ mice, is not able to specifically mediate degradation of ASGP in primary cultures of isolated knockout hepatocytes.
We have previously shown that there was no accumulation of any specific galactose-terminal glycoprotein(s) in plasma of Each symbol represents the mean Ϯ S.E. of two to seven mice. If no range bar can be seen, the S.E. of the individual data points is smaller than or equal to the size of the respective symbol. knockout animals using two-dimensional gel electrophoresis (21). In the current experiments we have further quantitated whether the absence of normal ASGPR function in MHL-2 knockout animals might result in an increase in steady state levels of desialylated plasma glycoproteins terminating in galactose. This was done by measuring the absolute concentration of galactose-terminal proteins in the serum of knockout mice (Table I). Our data do not reveal any significant differences in the absolute levels of circulating glycoproteins terminating in galactose in MHL-2 Ϫ/Ϫ compared with wild type animals. Therefore, we wanted to determine whether ligands other than galactose-terminating glycoproteins accumulate in serum of MHL-2 Ϫ/Ϫ mice. If ligand(s) do accumulate in serum of knockout mice, then it is likely that degradation of a known ligand for ASGPR by wild type hepatocytes will be competitively inhibited in the presence of knockout serum. To test this, we isolated wild type hepatocytes and incubated them with degradation medium containing iodinated asialoorosomucoid (0.5-2.0 g/ml) as well as increasing concentrations of MHL-2 ϩ/ϩ and MHL-2 Ϫ/Ϫ serum, respectively. The presence of MHL-2 Ϫ/Ϫ serum (Fig. 8, open squares) consistently inhibits degradation of 125 I-asialoorosomucoid to a higher extent than does MHL-2 ϩ/ϩ serum (Fig. 8, closed squares) over a wide range of concentrations in eight independent experiments. Statistical analysis of the data shows that this difference reaches significance at 1:10 dilution of degradation medium. Nonspecific degradation of 125 I-asialoorosomucoid has been deter-mined in the presence of a 200-fold excess of unlabeled ligand. Thus, our results suggest the presence of as yet unidentified endogenous ASGPR ligand(s), which accumulate(s) in MHL-2 Ϫ/Ϫ serum and competitively inhibit(s) 125 I-asialoorosomucoid uptake and degradation by primary cultures of wild type mouse hepatocytes.

DISCUSSION
The current experiments were designed to examine the fate of the MHL-1 subunit of the heterooligomeric ASGPR in MHL-2 knockout mice and further characterize their phenotype. Our results indicate that MHL-1 receptor subunits can be transported to the hepatocellular surface in the absence of MHL-2. This conclusion is supported by two lines of evidence: (i) When hepatocytes from knockout animals were fractionated into early and late secretory pathway (Fig. 2), MHL-1 protein was detectable in the same light-density fractions as LRP 85, a proteolytic maturation product of the 600-kDa precursor LRP known to be produced in the trans-Golgi or trans-Golgi network (28). (ii) The MHL-1 polypeptide exhibited maturation of carbohydrate side chains characteristic of Golgi-processing, i.e. it acquired sialic acid residues and thus became sensitive to treatment with neuraminidase (Fig. 3). In the past, several investigators have addressed the role of the individual receptor subunits in intracellular transport of ASGPR in vitro using different transfection assays (for review, see Ref. 2). Our results are in agreement with the findings of Shia and Lodish After incubation at 37°C for the indicated periods of time, the amount of radiolabeled degradation products secreted into the medium was determined. Each value represents the mean of duplicate incubations from one animal. (17), who demonstrated that in stably transfected NIH 3T3 fibroblasts the major subunit of ASGPR is able to traverse through the Golgi complex to the cell surface in the absence of minor receptor subunit. Because proper folding and oligomerization of various membrane proteins are major prerequisites for their transport from the endoplasmic reticulum to the cis-Golgi complex (32), our data further suggest that a fraction of MHL-1 in the knockout animals might have escaped potential intracellular degradation, possibly by forming homooligomers. It has been shown previously that oligomerization of the major ASGPR subunit can occur as judged by chemical cross-linking experiments (12,17). However, the finding that a small amount of MHL-1 reaches the hepatocellular surface in MHL-2-deficient mice does not rule out the possibility that a subfraction of the major ASGPR subunit is degraded in an early compartment of the secretory pathway. This consideration is supported by the fact that MHL-1 expression is profoundly reduced in the knockout animals, although MHL-1 mRNA levels are essentially unaffected by the gene disruption (21). Thus, MHL-2 seems to be required for post-translational stability of MHL-1.
The observation that MHL-1 protein is expressed at the surface of liver cells in MHL-2 Ϫ/Ϫ animals raised the possibility that residual functional activity of the remaining receptor subunits might account for the lack of obvious phenotypic ab-normalities in the knockout mice. Earlier work has shown that individual ASGPR subunits are, in fact, expressed alone (33,34) and can function as independent galactose binding proteins (12,13,35). Our evaluation of the functional status of the MHL-1 subunits in MHL-2 Ϫ/Ϫ animals showed a low level clearance of iodinated ASGP (Fig. 4D and Fig. 5, right panel), which could be explained by either one or a combination of the following mechanisms: (i) residual functional activity of MHL-1 subunits in knockout mice, albeit at a markedly reduced level; (ii) existence of other mammalian galactose-specific recognition systems such as the lectin-like receptor on Kupffer cells (36) and the homologous galactose/GalNAc-specific lectin expressed on macrophages (37), which are not affected by the disruption of the MHL-2 gene locus; and (iii) nonspecific binding phenomena. In order to distinguish between these possibilities, we performed degradation experiments using isolated mouse hepatocytes from animals of either genotype. Our results clearly demonstrate that MHL-2 Ϫ/Ϫ hepatocytes do not specifically degrade iodinated asialoorosomucoid (Figs. 6 and 7). Thus, normal ASGPR function is totally ablated in homozygous MHL-2 deficient mice and the observed residual clearance most likely due to nonspecific ligand binding and/or uptake by non-ASGPR members of the C-type lectin superfamily. Our data do not exclude the possibility that MHL-1 subunits per se might be able to bind ASGP in vivo. Bider et al. have shown that H1 subunits bind 125 I-labeled asialoorosomucoid in vitro with high affinity in the absence of H2 (13). However, the H1-overexpressing cells were only able to bind iodinated asialoorosomucoid at very high levels of H1 expression. Thus, it is likely that the density of residual MHL-1 receptor subunits on the hepatocellular surface of MHL-2 knockout mice might be too low to allow detectable ligand binding.
The total absence of normal ASGPR function does not entail a measurable increase in the steady state concentrations of galactose-terminating glycoproteins in plasma of knockout FIG. 7. Specific degradation of 125 I-labeled asialoorosomucoid by primary cultures of isolated mouse hepatocytes. Wells with isolated hepatocytes from mice either wild type (closed symbols) or homozygous for the MHL-2 gene disruption (open symbols) received 500 l of DMEM (without glutamine) containing 0.2% (w/v) BSA and the indicated concentrations of 125 I-labeled asialoorosomucoid. After incubation for 23 h at 37°C, the total amount of radiolabeled degradation products secreted into the medium during steady state conditions was measured. Nonspecific degradation was determined in the presence of a 100-fold excess of unlabeled asialoorosomucoid and subtracted from total degradation to yield the specific degradation. Each symbol represents the mean of duplicate determinations of specific degradation from one animal.  8. Competitive inhibition of specific degradation of 125 Ilabeled asialoorosomucoid by primary MHL-2 ؉/؉ hepatocytes. MHL-2 ϩ/ϩ hepatocytes (2 ϫ 10 6 /well) received 500 l of DMEM (without glutamine) containing 0.2% (w/v) BSA, 125 I-asialoorosomucoid (concentrations used in eight independent experiments: 0.5 g/ml (n ϭ 1); 1 g/ml (n ϭ 4); 2 g/ml (n ϭ 3), and MHL-2 ϩ/ϩ (closed squares) or MHL-2 Ϫ/Ϫ (open squares) serum in the indicated dilutions. After incubation for 8 h at 37°C, the total amount of iodinated asialoorosomucoid secreted into the medium was measured and expressed relative to the degradation of 125 I-asialoorosomucoid in the absence of any serum (100%). Nonspecific degradation was determined in the presence of a 200-fold excess of unlabeled asialoorosomucoid. Each symbol represents the mean Ϯ S.E. of duplicate (n ϭ 1) or quadruplicate (n ϭ 7) determinations of specific degradation from eight independent experiments. The asterisk denotes statistical significance (p ϭ 0.0023) as determined by Mann-Whitney U test within each dilution and Bonferroni adjustment of p values for multiple testing (StatView 4.51 software; Abacus Concepts, Berkeley, CA). mice (Table I). This finding further supports the previously stated hypothesis that ASGPR is not involved in normal turnover of serum glycoproteins (Refs. 38 -40; for review, see Ref. 41). During acute surges in the plasma concentration of desialylated glycoproteins, however, the alternative galactose-specific recognition systems are not able to compensate for the loss of ASGPR function in knockout mice. Such excess levels of ASGP might occur during certain infectious diseases known to be associated with increased sialidase activity (42) or experimentally (Figs. 4 and 5). Taken together, our data are in agreement with a model in which ASGPR might function to prevent acute increases in the concentration of desialylated soluble or particulate galactose-terminating glycoproteins that might be harmful in mammals (43).
The results of our competitive degradation experiments (Fig.  8) are of particular interest considering that no endogenous ligands for ASGPR have been identified despite its discovery almost 30 years ago (44). The generation of MHL-2 knockout mice allowed us to test for the presence of an accumulation of potential ASGPR ligand(s) in their serum regardless of knowledge about ligand structure. When primary cultures of MHL-2 ϩ/ϩ hepatocytes were incubated with medium containing different concentrations of knockout serum, we consistently observed a greater inhibition of specific 125 I-asialoorosomucoid degradation compared with wild type serum (Fig. 8). These findings are consistent with a model in which cross-competing endogenous ASGPR ligands accumulate in the serum of MHL-2 Ϫ/Ϫ mice. Thus, ASGPR-deficient animals will provide an important tool to further analyze the nature of the endogenous ligand(s) of ASGPR, thereby advancing our understanding of the physiological function of this major endocytic receptor system.